Jean-Paul Carvalho, Sriya Iyer, Jared Rubin - Advances in The Economics of Religion (2019, Springer International Publishing - Palgrave Macmillan)

ADVANCES IN
THE ECONOMICS
OF RELIGION
Edited by
Jean-Paul Carvalho,
Sriya Iyer,
Jared Rubin
IEA Conference Volume No. 158

International Economic Association Series
More information about this series at
http://www.palgrave.com/gp/series/13991
Other titles from IEA:
INSTITUTIONS, GOVERNANCE AND THE CONTROL OF CORRUPTION

Edited by Kaushik Basu and Tito Cordella
INEQUALITY AND GROWTH: PATTERNS AND POLICY

Volume II: Regions and Regularities
Edited by Kaushik Basu and Joseph E. Stiglitz
INEQUALITY AND GROWTH: PATTERNS AND POLICY

Volume I: Concepts and Analysis
Edited by Kaushik Basu and Joseph E. Stiglitz
THE MIDDLE EAST ECONOMIES IN TIMES OF TRANSITION

Edited by Galal Ahmed and Ishac Diwan
TOWARDS A GENERAL THEORY OF DEEP DOWNTURNS

Presidential Address from the 17th World Congress of the International Economic Association in 2014
Joseph E. Stiglitz
CONTEMPORARY ISSUES IN MACROECONOMICS

Lessons from the Crisis and Beyond
Edited by Joseph E. Stiglitz
CONTEMPORARY ISSUES IN MICROECONOMICS

Edited by Joseph E. Stiglitz
CONTEMPORARY ISSUES IN DEVELOPMENT ECONOMICS

Edited by Timothy Besley
TAMING CAPITAL FLOWS

Capital Account Management in an Era of Globalization
Edited by Joseph E. Stiglitz and Refet Gurkaynak
LIFE AFTER DEBT

The Origins and Resolutions of Debt Crisis
Edited by Joseph E. Stiglitz and Daniel Heymann
INCOME CONTINGENT LOANS

Theory, Practice and Prospects
Edited by Joseph E. Stiglitz, Bruce Chapman and Timothy Higgins
THE INDUSTRIAL POLICY REVOLUTION I

The Role of Government Beyond Ideology
Edited by Joseph E. Stiglitz and Justin Lin Yifu
THE INDUSTRIAL POLICY REVOLUTION II

Africa in the 21st Century
Edited by Joseph E. Stiglitz, Justin Lin Yifu and Ebrahim Patel
THE CHINESE ECONOMY

A New Transition
Edited by Masahiko Aoki and Jinglian Wu
INSTITUTIONS AND COMPARATIVE ECONOMIC DEVELOPMENT

Edited by Franklin Allen, Masahiko Aoki, Nobuhiro Kiyotaki, Roger Gordon, Joseph E. Stiglitz and Jean-Paul Fitoussi
COMPLEXITY AND INSTITUTIONS: MARKETS, NORMS AND CORPORATIONS
Edited by Masahiko Aoki, Kenneth Binmore, Simon Deakin and Herbert Gintis
CORPORATE SOCIAL RESPONSIBILITY AND CORPORATE GOVERNANCE

The Contribution of Economic Theory and Related Disciplines
Edited by Lorenzo Sacconi, Margaret Blair, R. Edward Freeman and Alessandro Vercelli
IS ECONOMIC GROWTH SUSTAINABLE?

Edited by Geoffrey Heal
KEYNE’S GENERAL THEORY AFTER SEVENTY YEARS

Edited by Robert Diman, Robert Mundell and Alessandro Vercelli
CORRUPTION, DEVELOPMENT AND INSTITUTIONAL DESIGN

Edited by János Kornai, László Mátyás and Gérard Roland
MARKET AND SOCIALISM

In the Light of the Experience of China and Vietnam
Edited by János Kornai and Yingyi Quian
INSTITUTIONAL CHANGE AND ECONOMIC BEHAVIOUR

Edited by János Kornai, László Mátyás and Gérard Roland
INTERGENERATIONAL EQUITY AND SUSTAINABILITY

Edited by John E. Roemer and Kotaro Suzumura
PSYCHOLOCY, RATIONALITY AND ECONOMIC BEHAVIOUR

Challenging Standard Assumptions
Edited by Bina Agarwal and Alessandro Vercelli
MULTINATIONALS AND FOREIGN INVESTMENT IN ECONOMIC DEVELOPMENT

Edited by Edward M. Graham
POST-CONFLICT ECONOMIES IN AFRICA

Edited by Paul Collier and Augustin Kwasi Fosu
STRUCTURAL REFORM AND MACROECONOMIC POLICY

Edited by Robert M. Solow
THE PAST, PRESENT AND FUTURE OF THE EUROPEAN UNION

Edited by Alan V. Deardorff
LATIN AMERICAN ECONOMIC CRISES

Trade and Labour
Edited by Enrique Bour, Daniel Heymann and Fernando Navajas
ADVANCES IN MACROECONOMIC THEORY

Edited by Jacques H, Drèze
EXPLAINING GROWTH
A Global Research Project
Edited by Gary McMahon and Lyn Squire
TRADE, INVESTMENT, MIGRATION AND LABOUR MARKET ADJUSTMENT

Edited by David Greenaway, Richard Upward and Katherine Wakelin
INEQUALITY AROUND THE WORLD

Edited by Richard B. Freeman
MONETARY THEORY AND POLICY EXPERIENCE

Edited by Axel Leijonhufvud
MONETARY THEORY AS A BASIS FOR MONETARY POLICY
Edited by Axel Leijonhufvud
ECONOMIC DEVELOPMENT IN SUBSAHARAN AFRICA

Proceedings of the Eleventh World Congress of the International Economic Association, Tunis
Edited by Ibrahim Elbadawi and Beno Ndula
International Economics Association

Series Standing Order ISBN 978-0-3337-1242-9 (Hardback)
978-0-3338-0330-1 (Paperback)
(outside North America only)
You can receive future titles in this series as they are published by placing a standing order. Please contact your bookseller or,
in case of difficulty, write to us at the address below with your name and address, the title of the series and one of the ISBNs
quoted above.
Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke, Hampshire RG21 6XS, England
Jean-Paul Carvalho • Sriya Iyer • Jared Rubin
Editors
Advances in the
Economics of Religion
Editors
Jean-Paul Carvalho Sriya Iyer
University of California, Irvine Faculty of Economics
Irvine University of Cambridge
CA, USA Cambridge, UK
Jared Rubin
Chapman University
Orange
CA, USA
International Economic Association Series

ISBN 978-3-319-99336-2 ISBN 978-3-319-98848-1 (eBook)
https://doi.org/10.1007/978-3-319-98848-1
Library of Congress Control Number: 2019930427
© The Editor(s) (if applicable) and The Author(s) 2019

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does
not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions
that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.
This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Foreword
This book comes out of a roundtable on the Economics of Religion organized

with the International Economic Association (IEA), of which I was President
from 2014 to 2017. The book is a veritable “who’s who” of contributors to
the field. The IEA is immensely grateful to the organizers. For this, we were
able to harness the enthusiasm of Jean-Paul Carvalho, Sriya Iyer, and Jared
Rubin who conceived the project and brought it to fruition including this
excellent volume. The volume is destined to become a locus classicus for those
who are interested in this burgeoning field. I hope that people are inspired by
the contributions in the volume and that the field grows even more strongly.
Economic methods and ways of thinking are able to provide insights into
religion. That is not to say that economics has a monopoly of insight, but
it is now firmly established that having economists working in this area adds
value. Another aspect of the volume and the field which ties to the IEA’s
mandate is that it is truly global. Different societies have distinctive patterns
of religiosity with particular historical and cultural influences. However, there
are trends and common patterns worth understanding. Moreover, the search
for common understanding among religions and the social science behind it
brings people together. Like the study of religion in economics, the landscape
is always changing as new phenomena and challenges emerge. Religions are
dynamic like the societies and economies that spawn them. The volume
also highlights the potential for the field to blend insights from theory and
empirics. Finding new ways of conceptualizing how religion matters is just as
important as seeking new data. I was personally privileged to have participated
in the meeting that led to this volume including the panel on growth and
religion. The sense of excitement among those present was palpable. I hope
vii
viii Foreword
that this comes over in this volume so that those who thumb its pages
appreciate what an exciting field of economics this has become.
LSE Tim Besley

President of the International Economic
Association, 2014–2017
Contents
1 Introduction 1
Jean-Paul Carvalho, Sriya Iyer, and Jared Rubin
Part I Theoretical Advances in the Economics of Religion 23
2 Religious Clubs: The Strategic Role of Religious Identity 25

Jean-Paul Carvalho
3 Spatial Models of Religious Market Competition:

A Critical Assessment 43
Michael McBride
4 When Average Is Irrelevant: Computational Modeling

of Religious Groups 61
Michael D. Makowsky
5 The Intelligent Design of Religious Beliefs 73

Gilat Levy
6 Religion and Segregation 89

Ronny Razin
7 Religious and Cultural Leaders 103

Anja Prummer
ix
x Contents
8 Intermediated Social Preferences: Altruism in an

Algorithmic Era 119
Daniel L. Chen
Part II Empirical Advances in the Economics of Religion 139
9 Religion and Demography 141

Sriya Iyer
10 Economics and Church State Research: Past, Present, Future 155

Daniel Hungerman and Timothy Weninger
11 Protestants and Catholics and Educational Investment

in Guatemala 169
Rachel M. McCleary and Robert J. Barro
12 The Religious Factor in Private Education in the United States 197

Danny Cohen-Zada and Moshe Justman
13 How Luther’s Quest for Education Changed German

Economic History: 9+5 Theses on the Effects
of the Protestant Reformation 215
Sascha O. Becker and Ludger Woessmann
14 Hindu-Muslim Violence in India: A Postscript from the

Twenty-First Century 229
Anirban Mitra and Debraj Ray
15 Religion and Conflict: A Quantitative Approach 249

José G. Montalvo and Marta Reynal-Querol
16 Why Are Some Societies More Religious Than Others? 265

Jeanet Sinding Bentzen
Contents xi
17 Socioeconomic Inequality Across Religious Groups:

Self-Selection or Religion-Induced Human Capital
Accumulation? The Case of Egypt 283
Mohamed Saleh
18 Religion and the European Union 295

Benito Arruñada and Matthias Krapf
Part III Advances in Religion and Political Economy 309
19 The Political and Economic Consequences of Religious

Legitimacy 311
Jared Rubin
20 Religious Legitimacy and the Joint Evolution of Culture

and Institutions 321
Alberto Bisin, Avner Seror, and Thierry Verdier
21 Strategic Interactions Between Religion and Politics: The

Case of Islam 333
Jean-Philippe Platteau
22 State and Religion: An Economic Approach 347

Metin M. Coşgel and Thomas J. Miceli
23 A Great Academic Re-awakening: The Return

to a Political Economy of Religion 361
Anthony Gill
24 The State, Toleration, and Religious Freedom 377

Noel Johnson and Mark Koyama
25 Religiosity and Economic Performance: The Role of

Personal Liberties 405
Laura Mayoral and Joan Esteban
xii Contents
26 Causes and Consequences of Monotheism in the Supply

of Religion 423
Murat Iyigun
27 Religion, Political Power and Human Capital Formation:

Evidence from Islamic History 437
Eric Chaney
28 Islam, Trade, and Innovation 449

Alireza Naghavi
Index 461
Notes on Contributors
Benito Arruñada is Professor of Business Organization at Universitat Pompeu Fabra,

Barcelona, Spain. Former President of the Society for Institutional and Organizational
Economics, his research lies in the conjunction of law, economics, and organization
and focuses on the organizational conditions that facilitate impersonal exchange,
covering from moral systems to property titling and business formalization. He is
the author of Institutional Foundations of Impersonal Exchange (2012).
Robert J. Barro is the Paul M. Warburg Professor of Economics at Harvard Uni-
versity, a visiting scholar at American Enterprise Institute, and a research associate
of the National Bureau of Economic Research. He has a PhD in Economics from
Harvard University and a BS in Physics from Caltech. Barro is the co-editor of
Harvard’s Quarterly Journal of Economics and was previously President of the Western
Economic Association, Vice President of the American Economic Association, a
viewpoint columnist for Business Week, and a contributing editor of The Wall
Street Journal. His noteworthy research includes empirical determinants of economic
growth, economic effects of public debt and budget deficits, and the economics
of religion. His research focuses on the impact of rare disasters on asset markets
and macroeconomic activity, with recent applications to environmental protection,
quantities of safe assets, and pricing of stock options. His books include The Wealth of
Religions (with Rachel McCleary, forthcoming), Economic Growth (2nd edition, with
Xavier Sala-i-Martin), Macroeconomics, Nothing Is Sacred: Economic Ideas for the New
Millennium, Determinants of Economic Growth, Getting It Right: Markets and Choices
in a Free Society, and Education and Modernization Worldwide, from the 19th to the 21st
Century (with Jong-Wha Lee).
Sascha O. Becker is Research Director of the Centre for Competitive Advantage
in the Global Economy (CAGE) and Professor of Economics at the University of
Warwick. He obtained his PhD from the European University Institute, Florence,
Italy, in 2001. He was an assistant professor in Munich (2002–2008) before moving to
xiii
xiv Notes on Contributors
Scotland (2008–2010). His research has appeared in international journals, including

The American Economic Review and the Quarterly Journal of Economics.
Jeanet Sinding Bentzen is Associate Professor of Economics at the Department of
Economics, University of Copenhagen. She is a research affiliate at CEPR (EH), an
external associate at CAGE (Warwick), and a board member of the Association for the
Study of Religion, Economics, and Culture. Her main research interests are religion,
cultural values, and economic growth.
Alberto Bisin is Professor of Economics at New York University. He is an elected
fellow of the Econometric Society. He is also a fellow of the NBER, CESS at NYU,
and the CEPR. He is Associate Editor of the Journal of Comparative Economics,
Economic Theory, and Research in Economics. He is the co-organizer of the annual
NBER Meeting on Culture and Institutions. He holds a PhD from the University of
Chicago, obtained in 1994. His main academic contributions are in the fields of social
economics, financial economics, and behavioral economics. He has published widely
in economics journals. He co-edited the Handbook of Social Economics and is in the
process of co-editing the Handbook of Historical Economics. Finally, he is the founding
editor of noiseFromAmerika.org and contributes op-eds for the Italian newspaper la
Repubblica.
Jean-Paul Carvalho is Associate Professor of Economics at the University of Cal-
ifornia, Irvine, and Interim Director of the Institute for Mathematical Behavioral
Sciences. His work is on the economics of culture, identity, and religion. He is an
associate of the Network for Economic Research on Identity, Norms and Narratives
(ERINN), a faculty fellow of the Association for Analytic Learning about Islam and
Muslim Societies (AALIMS), and a faculty affiliate of the Center for Global Peace and
Conflict Studies (CGPACS). He was educated at the University of Oxford (DPhil,
MPhil), as a Monash scholar, and at the University of Western Australia.
Eric Chaney is Associate Professor of Economic History at the University of Oxford.
His research focuses on the economic history and long-run development of the Islamic
world and Western Europe and has been published in leading journals in economics
and political science. He has held fellowships at the Institute for Advanced Study
(Princeton) and CASBS (Stanford).
Daniel L. Chen is Directeur de Recherche at the Centre national de la recherche
scientifique (CNRS) and Professor at the Toulouse School of Economics and Institute
for Advanced Study in Toulouse. He has a PhD in Economics from MIT, a JD
from Harvard Law School, and a BA and an MS in Applied Mathematics/Economics
from Harvard College (summa cum laude). Chen was previously Chair of Law and
Economics, a co-founder of the Center of Law and Economics at ETH Zurich, and
Assistant Professor of Law, Economics, and Public Policy at Duke University.
Danny Cohen-Zada is Associate Professor of Economics at Ben-Gurion University,
Israel, and a research fellow in IZA. He completed his PhD in Economics at Ben-
Notes on Contributors xv
Gurion University and did his postdoc in the Graduate School of Education at the
University of Pennsylvania. His research focuses mostly on the economics of religion,
the economics of education, and sports economics. Among other issues, it analyzes
patterns of school choice and explores, both theoretically and empirically, the role
of education in preserving religious identity. In addition, it deals with the effects of
religion and religiosity on economic and political outcomes. He has published his
work in international journals such as the International Economic Review, The Journal
of Law and Economics, Journal of Urban Economics, Journal of Economic Behavior and
Organization, Regional Science and Urban Economics, and the Economics of Education
Review.
Metin M. Coşgel is Professor of Economics at the University of Connecticut. He
is the author of The Economics of Ottoman Justice: Settlement and Trial in the Sharia
Courts (with Boğaç Ergene), Cambridge University Press (2016). He has published
widely on the economic history of the Ottoman Empire, political economy of religion,
and economics of social institutions. His research interests include the origins and
economic effects of religious diversity and the historical roots of comparative devel-
opment in the Middle East and Eastern Europe. Website: www.cosgel.uconn.edu.
Joan Esteban is a research professor at the Institut d’Anàlisi Económica (CSIC)
and at the Barcelona GSE. He received his PhD from the Universities of Oxford
and Autònoma de Barcelona. He has been Director of the IAE (CSIC) (1989–1991,
2001–2006), Secretary General (2011–2013) of the International Economic Association,
and President of the Society for the Study of Economic Inequality (ECINEQ) (2007–
2009). He is working on social polarization and conflict, and in public economics. His
research has been published in journals such as The American Economic Review, Amer-
ican Political Science Review, Annual Review of Economics, Econometrica, Economics
Letters, Economics of Governance, European Economic Review, International Economic
Review, Journal of Economic Behavior and Organization, Journal of Economic Inequality,
Journal of Economic Theory, Journal of the European Economic Association, Journal of
Income Distribution, Journal of Peace Research, Journal of Political Economy, Journal of
Public Economics, Regional Science and Urban Economics, Science, Social Choice and
Welfare, and Theory and Decision.
Anthony Gill is Professor of Political Science at the University of Washington and
Distinguished Senior Fellow at Baylor University’s Institute for Studies of Religion.
He obtained his PhD from the University of California, Los Angeles. He authored
Rendering unto Caesar: The Catholic Church and the State in Latin America (Chicago)
and The Political Origins of Religious Liberty (Cambridge), as well as numerous articles
on religious economies and liberty.
Daniel Hungerman is Professor of Economics at the University of Notre Dame; he is
also a research associate at the National Bureau of Economic Research. Hungerman’s
work has been published in various journals in the economics profession and has
been funded by various organizations such as the NIH and the John Templeton
xvi Notes on Contributors
Foundation. He uses both economic theory and applied microeconomic methods

to study topics related to the social effects of religiosity and the impact of various
public policies on religious practice. He received his PhD in Economics from Duke
University and holds an undergraduate degree in Economics from Miami University.
Sriya Iyer is Janeway Fellow in Economics and Affiliated Lecturer in the Faculty
of Economics, University of Cambridge, and Bibby Fellow and College Lecturer
at St Catharine’s College, Cambridge. She researches in the economics of religion,
demography, education, and development economics. She serves on the Board of
Directors of the Association for the Study of Religion, Economics, and Culture
and on the Editorial Board of the Journal of Religion and Demography, is a research
fellow at the Institute for Labor Economics (IZA), and was awarded a University of
Cambridge Pilkington Prize in 2014. She has published two books: Demography and
Religion in India (2002) and The Economics of Religion in India (2018). She has also
published articles in economics journals including the Journal of Political Economy,
Journal of Economic Literature, The Review of Economics and Statistics, and the Journal
of Development Economics.
Murat Iyigun is Professor of Economics at the University of Colorado at Boulder
specializing in the economics of religion, economics of the family, economic devel-
opment, and cliometrics. He received his PhD from Brown University in 1995. He is
a research fellow at the Institute for the Study of Labor (IZA), a founding fellow of
the Association for Analytical Learning on Islam and Muslim Societies (AALIMS),
and the co-editor of the Journal of Demographic Economics (JODE) published by
Cambridge University Press. Prior to joining the CU-Boulder faculty in 2000, he
served as a staff economist at the Federal Reserve Board in Washington, DC. Some
of Iyigun’s research has been published in The American Economic Review, Quarterly
Journal of Economics, The Review of Economic Studies, The Economic Journal, the
International Economic Review, and other leading journals. His general-interest book,
War Peace and Prosperity in the Name of God, was published by the University of
Chicago Press in March 2015.
Noel Johnson is an associate professor in the Economics Department at George
Mason University. He is also a member of the Center for the Study of Public Choice
and a research fellow at the Mercatus Center. He earned his PhD at Washington
University in St Louis. His interests lie at the intersection of economic history,
development, and the new institutional economics. His recent research has focused
on questions such as “how do states build administrative and fiscal capacity?”, “what
is the relationship between state capacity and growth enhancing economic and
social outcomes such as religious tolerance or free trade?”, and “how has the disease
environment and climate affected economic and social outcomes under different
institutional settings?” This research has been published in outlets including The
Journal of Law and Economics, Explorations in Economic History, The Economic Journal,
and the Journal of Development Economics.
Notes on Contributors xvii
Moshe Justman holds a PhD in Business Economics from Harvard University

and is Professor and Dean of Economics and Business Administration at Ruppin
Academic Center in Israel. Previously, he was Professor of Economics at Ben-Gurion
University and held fractional and visiting positions at the University of Melbourne,
Carnegie Mellon University, and Renmin University. He is a past president of the
Israel Economics Association. His research, centered on the economics of education,
combines theoretical, empirical, and policy perspectives on the balance between
private and public education, the role of education in shaping religious and cultural
identities, macroeconomic perspectives on education, intergenerational mobility,
gender and STEM, and measurement in education. He has published extensively
in international journals, including The American Economic Review, The Review of
Economic Studies, The Review of Economics and Statistics, and Journal of the European
Economic Association, and co-authored a book, The Political Economy of Education
(MIT Press).
Mark Koyama is Associate Professor of Economics at George Mason University at
the Center for the Study of Public Choice and a senior scholar at the Mercatus Center
where he is part of the F. A. Hayek Program for Advanced Study in Philosophy,
Politics, and Economics. He obtained his PhD from the University of Oxford.
His work has appeared in a variety of journals including The Economic Journal,
International Economic Review, the Journal of Development Economics, The Journal of
Law and Economics, and Explorations in Economic History. His book with Noel D.
Johnson Persecution and Toleration: The Long Road to Religious Freedom is published
by Cambridge University Press.
Matthias Krapf in his recent research has been studying the effect of wealth taxation
on reported wealth, as well as how wealth accumulation and income risk vary along
the wealth and income distributions. He is affiliated with the University of Basel
in a position funded by the Swiss National Science Foundation and was previously
affiliated with the Universities of Vienna, Zurich, and Lausanne. He holds a doctoral
degree from the University of Konstanz. His previous research has been published
in journals such as the Journal of International Economics, Economic Inquiry, and the
Journal of Economic Behavior and Organization.
Gilat Levy from the Economics Department has a track record of publishing in top
journals in economics and in political science. She has worked before on a host of
questions relating to religious organizations, the effect of religiosity on redistribution,
the sustainability of religious beliefs, barriers to social mobility resulting from segrega-
tion and discrimination, and information aggregation in democracies, among others.
She was awarded two ERC research grants, one for the study of religious organizations
and one for a project on cognitive biases in information processing. Levy served in the
Council of the Royal Economic Society; is a member of the Council of the European
Economic Association, the regional standing committee of the Econometric Society;
and was a board member in many journals such as The American Economic Review,
xviii Notes on Contributors
The Review of Economic Studies, and Theoretical Economics. She is the managing editor
of The Economic Journal.
Michael D. Makowsky is Associate Professor in the John E. Walker Department of
Economics at Clemson University. He obtained his PhD from George Mason Uni-
versity in 2008. He specializes in the economics of religion and club goods, crime and
law enforcement, and the application of agent-based computational modeling. Prior
to joining Clemson, he served as a faculty in the Johns Hopkins University Center for
Advanced Modeling. Makowsky’s work has been published in The American Economic
Review, The Review of Economic Studies, The Journal of Law and Economics, Journal of
Economic Behavior and Organization, and other leading journals.
Laura Mayoral is a research associate professor at the Institut d’Anàlisi Económica
(IAE) in Barcelona and an affiliated professor at the Barcelona Graduate School of
Economics and the Department of Economics of the University of Gothenburg. She
has also been a visiting professor at INSEAD, Paris School of Economics, and NYU.
Her fields of interest are the study of social conflict, long-term development, and the
economics of culture and religion. Her research has been published in journals such as
Science, The American Economic Review, Econometrica, Journal of Monetary Economics,
International Economic Review, and Journal of Public Economics, among others.
Michael McBride is Professor of Economics, Logic and Philosophy of Science, and
Religious Studies at the University of California, Irvine, and Founding Director of the
Experimental Social Science Laboratory. He received his PhD from Yale University.
He uses game theory and experimental methods to study collective action, conflict,
and religion. His research has appeared in the Journal of Economic Theory, the Journal
of Public Economics, the American Journal of Sociology, the Journal of Economic Behavior
and Organization, and Rationality and Society.
Rachel M. McCleary is a lecturer at Harvard University in the Faculty of Arts
and Sciences. Her research is interdisciplinary focusing on how religion interacts
with economic performance and the political and social behavior of individuals and
institutions across societies. She has written four books: Seeking Justice: Ethics and
International Affairs (1992), Dictating Democracy: Guatemala and the End of Violent
Revolution (University Press of Florida, 1999—English; Artemis-Edinter 1999—
Spanish), Global Compassion: Private Voluntary Organizations and U.S. Foreign Policy
Since 1939 (2009, and winner of the 2010 AFP Skystone Ryan Research Prize), and
The Oxford Handbook of the Economics of Religion (2011). McCleary and Robert Barro
are co-authoring a book, The Wealth of Religions: Economic Implications of Believing
and Belonging (Princeton University Press, Spring 2019).
Thomas J. Miceli is Professor of Economics at the University of Connecticut, where
he has taught since 1987. He received his PhD in Economics from Brown University,
and his research is primarily in the area of applied microeconomics, with an emphasis
on law and economics, the economics of religion, and sports economics. He has
Notes on Contributors xix
published extensively in these areas and serves as an associate editor at the International
Review of Law and Economics.
Anirban Mitra is Lecturer (Assistant Professor) in Economics at the University
of Kent, Canterbury, UK. His research interests lie in the areas of development
economics and political economy—often in their overlap. He has been interested
in the role of economic factors behind ethnic and civil conflicts particularly in the
context of Hindu-Muslim violence in India and the Maoist conflict in Nepal. His
other works relate to institutional design and its ramifications on elements of public
expenditure and income distribution. His research articles have been published in the
Journal of Political Economy, The Economic Journal, and Economica. He is a research
affiliate of the CESifo Research Network (Munich). Prior to joining Kent, he was
employed at the University of Oslo. He received his PhD from New York University
(USA) in 2012 and MA in Economics from JNU (New Delhi, India) in 2005.
José G. Montalvo is Professor of Economics at Universitat Pompeu Fabra and
Research Fellow at the Institució Catalana de Recerca i Estudis Avançats (ICREA-
Academia). He is also a research professor at the Barcelona Graduate School of
Economics and the IVIE. He holds a PhD in Economics from Harvard University and
a bachelor’s degree from the University of Valencia. He has worked as a consultant
for the OECD, the European Union, the Inter-American Development Bank, and
the World Bank. He has served as the Chairman of the Economics and Business
Department and Vice Chancellor of Science Policy at the UPF. His areas of research
include economic development, the effects of ethnic diversity, social conflicts, and
terrorism. He is the author of 15 books and over 100 articles in scholarly journals,
including The American Economic Review, The Economic Journal, and The Review of
Economics and Statistics, among many others.
Alireza Naghavi is Full Professor of Economics at the University of Bologna and
an adjunct professor at Johns Hopkins University School of Advanced International
Studies Bologna Center and Dickinson College Bologna Center for European Studies.
He holds a PhD from University College Dublin. He has published in international
journals such as The Economic Journal, Journal of Economic Growth, Journal of
International Economics, and Journal of Development Economics. He was the global
scientific coordinator of the European Commission’s FP7 project INGINEUS on
global innovation networks and three Italian projects at the national level since. His
research interests focus on international trade and institutions and include topics
such as intellectual property rights, firm organization, innovation, migration, and the
economics of religion.
Jean-Philippe Platteau is Professor Emeritus at the University of Namur and an
active member of the Centre for Research in Economic Development (CRED), which
he founded at the same university. He is the author of numerous journal articles as
well as several books, including Culture and Development: New Insights into an Old
Debate (2010) and Islam Instrumentalized: Religion and Politics in Historical Perspective
xx Notes on Contributors
(2017). Most of his works have been concerned with the understanding of the role
of institutions in economic development and the processes of institutional change.
The influences of non-economic factors and other frontier issues at the interface
between economics and sociology are a central focus of his work. He co-founded
the European Development Network (EUDN) and is presently the co-director, with
François Bourguignon, of an international research program devoted to Economic
Development and Institutions (EDI), funded by the Department of International
Development (DFID) of the British government.
Anja Prummer is a lecturer at Queen Mary University of London. She specializes in
the areas of social networks and political economy, with an emphasis on culture and
gender. She focuses on how social interactions can improve or hinder an individual’s
economic success. Through the combination of theoretical and empirical work, she
has been able to identify suboptimal network patterns and highlight why these
patterns persist. Relatedly, she is interested in how political, cultural, and religious
leaders can influence individuals, taking into account their social network. Before
joining Queen Mary University of London, she was a postdoctoral fellow at the
Cambridge-INET Institute. She obtained her PhD at the European University
Institute, Florence, and visited the University of Wisconsin–Madison, University of
California, Berkeley, as well as New York University for extended periods.
Debraj Ray is the Julius Silver Professor in the Faculty of Arts and Science and
Professor of Economics at New York University. He received his BA from the
University of Calcutta and his PhD in 1983 from Cornell University. He is a fellow of
the American Academy of Arts and Sciences, a fellow of the Econometric Society,
a fellow of the Society for Advancement of Economic Theory, and Guggenheim
Fellow. He holds an honorary degree from the University of Oslo. He is the co-
editor of The American Economic Review. Among Ray’s teaching awards are the Dean’s
Award for Distinguished Teaching at Stanford and the Golden Dozen Teaching
Award from New York University. Ray’s (1998) textbook, Development Economics
(Princeton University Press), was described by The Chronicle of Higher Education as
“a revolutionary textbook that takes the field by storm.”
Ronny Razin from the Department of Economics, London School of Economics,
has been doing research on a range of topics that sit in between economics and political
science. Razin has top publications both in Economics and Political Science journals on
topics such as information aggregation in elections, theories of religious organizations,
and group and individual decision making. In recent years Razin’s research focus
centers on developing new behavioral models of how individuals process information
and form beliefs and studying the implications of this to both economic and political
outcomes. Razin has been awarded and has been a team member on numerous
research grants from the ESRC, from the ERC, and from the National Science
Foundation in the US. Razin is an associate editor in The American Economic Review
and has previously served on the board of Games and Economic Behavior.
Notes on Contributors xxi
Marta Reynal-Querol is ICREA Research Professor at the Department of Eco-

nomics and Business at Universitat Pompeu Fabra (UPF), Research Professor and
Affiliated Professor of the Barcelona GSE, and Director of IPEG. She is also Director
of the Master in Economics at UPF. She is a research fellow at the CEPR and at the
CESifo and a full member at the EUDN. She is a fellow of the EEA. She was a member
of the Council of the European Economic Association (EEA) between 2011 and 2015.
She is a member of the Editorial Board of the JCR and in the past of the EJPE. She
won an ERC-Consolidator grant in 2014 and an ERC-Starting grant obtained in the
first call of the European Research Council. She won the Banco Herrero prize 2011
awarded annually to a Spanish social scientist under 40 years old. She worked at the
World Bank between 2001 and 2005.
Jared Rubin is Associate Professor of Economics at Chapman University. He is the
author of Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not (2017), which explores the role that politics and the religious institutions of Islam
and Christianity played in the long-run “reversal of fortunes” between the economies
of the Middle East and Western Europe. He graduated with a PhD in Economics
from Stanford University in 2007 and BA from the University of Virginia in 2002.
Rubin is a co-director of Chapman University’s Institute for the Study of Religion,
Economics, and Society and serves in various roles for the Association for the Study
of Religion, Economics, and Culture.
Mohamed Saleh is an assistant professor at Toulouse School of Economics (TSE)
and Institute for Advanced Study in Toulouse (IAST), France. His research interests
are in economic history, political economy, and development economics, with a focus
on the economic history of the Middle East and North Africa. His research agenda has
thus far focused on understanding the historical origins of the socioeconomic differ-
ences between religious groups in the Middle East, the effects of state industrialization
and public mass education on these differences, the historical role of the Islamic tax
system in the formation of religious groups, and the institutions of labor coercion
and land distribution in the pre-colonial period. He approaches these questions using
novel micro data constructed from both archival and secondary sources, including
population censuses, agricultural censuses, and papyrological records.
Avner Seror is a postdoctoral fellow at the Smith Institute for Political Economy
and Philosophy at Chapman University, California. He completed his PhD at the
Paris School of Economics in 2018.
Thierry Verdier is Ingenieur Général des Ponts et Chaussées (Ecole des Ponts
ParisTech) and Professor of Economics at the Paris School of Economics (France) and
at PUC-Rio (Brazil). He is also a research fellow and a former program director at the
Centre for Economic Policy Research (CEPR; London), research fellow at CESifo,
and a fellow of the European Economic Association. His research interests are in the
fields of international trade and globalization, the political economy of institutions,
and the analysis of cultural change at the interface between economics, evolutionary
xxii Notes on Contributors
anthropology, and economic sociology. His work has been published in international
academic journals, such as The American Economic Review, the Journal of Political
Economy, The Review of Economic Studies, the Quarterly Journal of Economics, the
Journal of Economic Development, and the Journal of International Economics.
Timothy Weninger is an assistant professor at the University of Notre Dame where
he directs the Data Science Group and is a member of the Interdisciplinary Center
for Networks Science and Applications (ICENSA). He has authored over 60 research
publications in the areas of social media, machine learning, and network science. The
key application of his research is to identify how humans generate, curate, and search
for information in the pursuit of knowledge. He uses properties of these emergent
networks to reason about the nature of relatedness, membership, and other abstract
and physical phenomena. He is a recipient of the NSF CAREER Award and the Army
Research Office Young Faculty Award and has received research grants from the Air
Force Office of Scientific Research, DARPA, and the John Templeton Foundation.
He is an inaugural member of the ACM’s Future of Computing Academy and serves
on numerous scientific program committees and editorial boards.
Ludger Woessmann is Professor of Economics at the University of Munich and
Director of the ifo Center for the Economics of Education. He was National Fellow
at the Hoover Institution at Stanford University and spent extended research visits at
Harvard University and the National Bureau of Economic Research. He is a fellow
of the International Academy of Education and a member of the German National
Academy of Sciences Leopoldina. He is the co-editor of the Handbook of the Economics
of Education. His main research interests are the determinants of long-run prosperity
and of student achievement, focusing on the economics of education, economic
growth, economic history, and the economics of religion. His work was rewarded with
the Gossen Award and the Stolper Award of the German Economic Association, the
Young Economist Award of the European Economic Association, and the Choppin
Memorial Award of the International Association for the Evaluation of Educational
Achievement.
List of Figures
Fig. 2.1 (a) Multiple steady-state levels of religious capacity. (b) A

temporary shock to social mobility induces a large, permanent
increase in religious participation/capacity. Source: Binzel and
Carvalho (2017) 32
Fig. 3.1 Affiliations by religious consumers. (a) Uniform religious
demand. (b) Symmetric, single-peaked religious demand.
(c) Asymmetric, single-peaked religious demand 45
Fig. 3.2 Barros and Garoupa analysis. (a) Monopolist, a = b.
(b) Monopolist, a = 2 > b = 1. (c) Monopolist without
adjustment to non-affiliation, a = 2 > b = 1. (d) Monopolist
with adjustment to non-affiliation, a = 2 > b = 1 47
Fig. 3.3 McBride analysis. (a) Low entry cost. (b) High entry cost.
(c) High entry cost, population increase 49
Fig. 8.1 Experimental design 123
Fig. 8.2 Proposed split by groups 127
Fig. 8.3 Distribution of bonus splits 127
Fig. 8.4 Distribution of bonus splits 128
Fig. 9.1 Size of major religious groups in 2010—Percentage of the
global population. *Includes followers of African traditional
religions, Chinese folk religions, Native American religions
and Australian aboriginal religions. **Includes Bahai’s, Jains,
Sikhs, Shintoists, Taoists, followers of Tenrikyo, Wiccans,
Zoroastrians and many other faiths. Source: Pew Research
Center’s Forum on Religion & Public Life, Global Religious
Landscape, December 2012. Also reproduced in Iyer (2016),
p. 402 142
xxiii
xxiv List of figures
Fig. 9.2 Median age of religious group in 2010. Notes: (1) Folk
Religionists includes followers of African traditional religions,
Chinese folk religions, Native American religions and
Australian aboriginal religions. (2) Other Religions includes
Bahai’s, Jains, Sikhs, Shintoists, Taoists, followers of Tenrikyo,
Wiccans, Zoroastrians and many other faiths. (3) Percentage
may not add to 100 due to rounding. Source: Adapted from
the Pew Research Center’s Forum on Religion and Public
Life—Global Religious Landscape, December 2012 144
Fig. 10.1 Fraction of Americans who never pray. Source: General Social
Survey 159
Fig. 10.2 An example of a church bulletin. The above are two pages
from a sample bulletin 163
Fig. 10.3 Occurrence of democratic/republican words in church
bulletins, by week 165
Fig. 11.1 Literacy rate and indigenous share across departments of
Guatemala in 1950. See Table 11.4 for acronyms of departments 175
Fig. 11.2 Literacy rate and indigenous share across departments of
Guatemala in 2011. See Table 11.4 for acronyms of departments 175
Fig. 12.1 Private share in elementary and secondary school enrollment,
1955–2016. Source: US Census Bureau, Current Population
Survey, 1955 to 2016 198
Fig. 12.2 Distribution of private school enrollment by school type.
Source: Digest of Education Statistics 1980, 1990, 2015 198
Fig. 12.3 The distribution of households among school types. P denotes
public school enrollment; R denotes religious private school
enrollment; and N denotes non-sectarian private school
enrollment 201
Fig. 14.1 Conflict and per-capita expenditure. Panel (a) plots the
residual of casualties against (log) Muslim expenditures after
region, time, and (log) Hindu expenditure effects have been
removed, in the five-year period following expenditures. Each
line segment connects five data points for a region. Panel (b)
plots the analogous graph for (log) Hindu expenditures 237
Fig. 14.2 Conflict and per-capita expenditure: without Ahmedabad.
Panel (a) plots the residual of casualties against (log) Muslim
expenditures after region, time, and (log) Hindu expenditure
effects have been removed, in the five-year period following
expenditures. Each line segment connects five data points for
a region. Panel (b) plots the analogous graph for (log) Hindu
expenditures 242
List of figures xxv
Fig. 16.1 Average religiosity across the globe. Notes: Country averages
across all waves 1981–2014 of the pooled WVS-EVS.
Religiosity is measured using the Strength of Religiosity Scale
in the left maps and answers to the question “How important
is God in your life?” in the two maps to the right. The upper
panels show the simple country averages without control
variables. The lower panels show the within-denomination
differences, that is, the residuals of regressions where the
religiosity measures are regressed on the five major religious
groups: Christianity, Islam, Buddhism, Hinduism, and Other 269
Fig. 16.2 Impact of disaster risk on religiosity across religious
denominations. Notes: The figure shows the parameter
estimate on earthquake risk in a regression on religiosity,
accounting for country-by-year fixed effects and a dummy for
actual earthquakes during the past year 273
Fig. 18.1 Religious denomination versus December 2011 bond yields 302
Fig. 18.2 Religious denomination versus control over corruption in 2016 303
Fig. 20.1 Societal policy game 323
Fig. 20.2 Joint dynamics of culture and institutions 324
Fig. 22.1 The decline of state religion in different religious traditions 353
Fig. 25.1 Legal evolution of the personal liberties index, 1960–2013 410
Fig. 25.2 Cross-country std. deviation of the personal liberties index,
1960–2013 411
List of Tables
Table 8.1 Summary statistics by treatment group 126

Table 8.2 Altruism by treatment group 128
Table 11.1A Panel A: Regressions for literacy rates 180
Table 11.1B Panel B: Estimated department fixed effects in Panel A,
column 2 181
Table 11.2 Regressions for school-enrollment ratios 182
Table 11.3 Means and standard deviations of variables 183
Table 11.4 Departments of Guatemala 184
Table 14.1 The effect of Hindu and Muslim expenditures on regional
conflict: Poisson FE regressions 239
conflict: Poisson FE regressions (urban households) 240
conflict: Poisson FE regressions (excluding the region
containing Ahmedabad) 243
conflict: Poisson FE regressions (urban households,
excluding the region containing Ahmedabad) 245
Table 18.1 Religious denomination and macro indicators 301
Table 18.2 Religious denomination and institutional quality 303
Table 18.3 Religion and AfD vote share in the German federal election
2017 306
Table 25.1 Baseline: effort, religious intensity, and liberties, OLS and
2SLS 414
xxvii
1
Introduction
Jean-Paul Carvalho, Sriya Iyer, and Jared Rubin
Over the past 20 years, economists have come to realize that religion is linked
inextricably to their subject matter, from fertility choices in the household,
to risk-sharing schemes in a village, to large-scale political movements, and to
economic growth. A deeper understanding of religion is perhaps now more
important than ever before.
The International Economic Association (IEA) Roundtable on the Eco-
nomics of Religion was held at St Catharine’s College, Cambridge, on 10–11
July 2017 to take stock of developments to date in the economics of religion
and to chart new directions for the field. Organized by Sriya Iyer (Cambridge),
Jared Rubin (Chapman), and Jean-Paul Carvalho (UC Irvine), the Roundtable
brought together some of the world’s most distinguished economists and other
scholars to debate and discuss the role of religion in society today.
The IEA Roundtable began with an opening address by Bishop Marcelo
Sánchez Sorondo, Chancellor of the Pontifical Academies of Sciences and
Social Sciences from the Vatican, who spoke about key points from the
J.-P. Carvalho
Department of Economics and Institute for Mathematical Behavioral Sciences,
University of California, Irvine, Irvine, CA, USA
S. Iyer
Faculty of Economics, University of Cambridge, Cambridge, UK
J. Rubin ()
Chapman University, Orange, CA, USA
e-mail: jrubin@chapman.edu
© The Author(s) 2019 1

J.-P. Carvalho et al. (eds.), Advances in the Economics of Religion, International Economic
Association Series, https://doi.org/10.1007/978-3-319-98848-1_1
2 J.-P. Carvalho et al.
Laudato si, the Papal Encyclical document on environmental protection and

climate change, highlighting the need for effective action in these areas. The
highlight was a panel discussion on the role of state religion, religious freedoms,
institutions, and growth featuring Robert Barro, Timur Kuran, Tim Besley,
and Sascha Becker and a second panel on how religion had shaped the world
with Partha Dasgupta, Rachel McCleary, David Maxwell, Larry Iannaccone,
and Michael McBride. The first panel explored questions such as religion
and innovation, religious education, the role of women, findings on religious
participation and economic growth, the role of state religion, the debates about
the clash of civilizations, and the role of Islamic and Christian fundamentalism
in the world today. The second panel keenly discussed Pentecostalism in Africa
and Latin America, how to think about definitions of religion, the relationship
between religion and the natural world, and how this relates to economic
development. The individual sessions ranged from exploring economic theory
and religion, history and religion, religious giving, and reformations and
religious freedoms. To that end, the dialogue about religion at the Roundtable
was truly wide-ranging and inter-disciplinary.
In order to bring together and to consolidate the many interesting debates
and lively discussions at the Roundtable, and more widely, this volume
documents the tremendous growth in research on religion in economics. The
aim is to provide an introduction to the economics of religion for academics,
professionals, and students and to chart and to possibly shape the rapidly
expanding frontiers of the field.
Although Adam Smith devoted an entire chapter to religious organizations
in his foundational work, The Wealth of Nations, the economic analysis of
religion was largely neglected for nearly two centuries. This began to change
in the late 1980s. As with many movements, the growth in the economics of
religion is due to both personal and impersonal forces. In personal terms, the
field owes an immense debt to Larry Iannaccone for his seminal contributions
(e.g. Iannaccone 1988, 1990, 1992), most importantly his canonical model
of religious clubs, as well as his organizational work and tireless advocacy for
the field. Larry was brave enough to venture into this territory when it was
considered unfashionable and unprofitable to do so. He can rightly be called
the father of the economics of religion.
By 1998, there were enough papers in religion written by social scientists
to warrant a review piece, Iannaccone’s “Introduction to the Economics of
Religion” in the Journal of Economic Literature. Most work in the field up
to that point—inspired by Iannaccone as well as sociologists Rodney Stark
1 Introduction 3
and Roger Finke—attempted to explain religious phenomena via the rational

choice framework. This framework proved extremely flexible, permitting
social scientists to explore a variety of religious phenomena such as church
attendance (Azzi and Ehrenberg 1975), medieval Church doctrine (Ekelund
et al. 1989, 1996), religious markets (Finke and Stark 1992; Montgomery
1996; Iannaccone et al. 1997), religious organizations (Iannaccone 1992,
1994), religious practice (Finke and Stark 1988; Iannaccone 1990), and the
underpinnings of church and sect (Iannaccone 1988).
The two decades following Iannaccone’s 1998 survey have seen a vast
acceleration in the quantity and quality of articles in the field. The current
state of research in the field was brought up-to-date by Sriya Iyer’s Journal
of Economic Literature survey on the new economics of religion in 2016
(Iyer 2016). There have been edited volumes by leading scholars in the field
(McCleary 2011). There are also major research monographs that focus on
the Middle East and on South Asia (Berman 2009; Kuran 2011; Iyigun 2015;
Rubin 2017; Platteau 2017; Iyer 2018). Other works that have attempted to
review the literature are a survey on Islam and contemporary performance
(Kuran forthcoming) and recent social scientific work on the Reformation
(Becker et al. 2016). Articles aside, one important pathway for disseminating
research in this field has been the growth of the ASREC (Association for the
Study of Religion, Economics, and Culture) Annual Conference, which is now
an important event in the field.
Another development of note is the significant increase in the number of
articles in the field being published in top economics and political science
journals. Publications in top journals encourage graduate students and young
scholars to undertake research in the field, seeing it as a fruitful career path.
And indeed, within economics and political science alone, the top journals
have become much more receptive to research on religion. For instance,
between 2000 and 2007 there were a total of 20 articles on religion in the
top 8 economics and top 3 political science journals. However, between 2016
and 2018 alone, 31 articles were either published or accepted by these journals
(calculations by Rubin). This research covers a wide range of topics related
to religion. Fifty-one of the 115 articles published between 2000 and 2018
in top journals take religion as the dependent variable and 64 take religion
as the independent variable. Of those that take religion as the dependent
variable, important recent findings (since 2012) include the economic and
political causes of increased Islamization (Carvalho 2013; Hegghammer 2013;
Gould and Klor 2016; Valdez 2016; Binzel and Carvalho 2017; Nellis and
Siddiqui 2018; Brooke and Ketchley 2018), economic causes of inter-religion

violence (Jha 2013; Mitra and Ray 2014), and the causes of religious change
(with much attention focused on the Reformation; Cantoni 2012; Rubin
2014). Of those that take religion as the independent variable, important
recent findings include the (sometimes negative, sometimes positive) effect of
religion on economic growth (Almond et al. 2015; Campante and Yanagizawa-
Drott 2015), the effects of religion on political participation (McKenzie and
Rouse 2013; Gaskins et al. 2013a, b; Stegmueller 2013; McClendon and
Riedl 2015), the long-run economic and political effects of pre-modern Jewish
economic activity and persecutions (Voigtlaender and Voth 2012; Pascali 2016;
Anderson et al. 2017; Spenkuch and Tillmann 2018), the long-run effect of
Christian missionary activity on economic and political outcomes (Woodberry
2012; Castelló-Climent et al. forthcoming), the economic and social effects
of the Reformation (Cantoni 2015; Becker and Woessmann 2018; Cantoni
et al. 2018), and the effects of religion on the mental health of adolescents
(Fruehwirth et al. forthcoming).
There are numerous reasons for these developments. First, social scientific
studies in religion have taken full advantage of the “empirical revolution” that
has pervaded economics and political science since the early 2000s. Tech-
niques such as instrumental variables, regression discontinuity, and natural
experiments are perfectly suited for identifying aspects of religion that were
considered to be in a “black box” as recently as the 1990s. These techniques
take advantage of new sources of data made available through digitization
to answer questions previously thought unanswerable (see, e.g. Gruber and
Hungerman 2008; Iyigun 2008; Becker and Woessmann 2009; Rubin 2014;
Anderson et al. 2017; Cantoni et al. 2018). Second, in large part due to
the empirical techniques described above, studies in religion have seen an
acute resurgence in two fields: development and economic history. It has
become increasingly clear that religion (and culture) plays a critical role in eco-
nomic development—or lack thereof—and understanding the mechanisms
through which religion affects underdeveloped economies, and economic
growth more generally, has gained first-order importance in the literature
(Barro and McCleary 2003; Guiso et al. 2003; Noland 2005; Durlauf et al.
2011; Woodberry 2012; Meyersson 2014; Campante and Yanagizawa-Drott
2015). Meanwhile, economic historians have “rediscovered” religion, in the
process shedding light on all sorts of issues such as long-run economic
divergence (Kuran 2011; Rubin 2017), human capital accumulation (Becker
and Woessmann 2009; Botticini and Eckstein 2012; Chaney 2016; Cantoni et
al. 2018), and religious/political movements (Iyigun 2008, 2015; Chaney 2013;
Rubin 2014; Johnson and Koyama 2019; Michalopoulos et al. forthcoming;
1 Introduction 5
Iyer and Shirvastava 2018). Third, developments in game theory, in particular

the application and absorption of game theory into microeconomics and
new evolutionary models of social norms (e.g. Young 1998), have provided
sophisticated tools for modeling religious behavior (Bisin and Verdier 2000;
Carvalho 2013; Bénabou et al. 2016; Seror 2018).
The goal then of this volume is to take stock of recent advancements in the
economics of religion. This is no small task; as we have already noted, the field
is developing rapidly, with advances being made on numerous fronts. This
volume focuses on three sets of contributions which we believe covers most of
the cutting-edge research in the field. Part I of our volume, titled “Theoretical
Advances in the Economics of Religion,” reviews the theoretical foundations
of the economics of religion, which was built in the 1990s on models of
religious clubs and religious markets. These “new” foundations include models
of religious belief formation, cultural transmission, and religious extremism.
Most of the recent growth in the economics of religion has come from
empirical research by labor, development, and public economists. Part II,
titled “Empirical Advances in the Economics of Religion,” provides the reader
with a view of this new empirical research in areas as diverse as demography,
education, philanthropy, conflict, and international integration. Of the many
promising applications of the economics of religion being explored today,
political economy stands out. We devote Part III, titled “Advances in Religion
and Political Economy,” to the role of religion in political economy, from a
contemporary and historical perspective.
Theoretical Advances in the Economics of Religion

Part I of this volume is devoted to recent theoretical contributions to the
economics of religion. The theoretical foundations of the economics of religion
lie in religious club models which explain the internal structure of reli-
gious organizations, and religious market models which examine competition
among religious organizations for members. Both approaches are formulated
in response to an important puzzle. Religious club models, beginning with
Iannaccone (1992), explain why strict religious groups impose seemingly
bizarre and unproductive rules on their members. Religious sacrifice and
stigma limit free riding and promote the efficient production of club goods in
an organization. Religious beliefs and preferences for various forms of religious
experience and sacrificial rituals are not required. Strict entry requirements and
participation rules are rationalized on purely “economic grounds.” Religious
market models (Barros and Garoupa 2002; Montgomery 2003; McBride
2008, 2010) explain why the United States has higher rates of religious belief
and participation than Western Europe. Following Adam Smith’s original
analysis in The Wealth of Nations, the differences are attributed to religious
market structure. Countries with low religious participation, such as the
United Kingdom and Sweden, have established state-funded churches which
provide weak incentives for effort and innovation on the part of clergy. In
contrast, the United States with its separation of church and state has a
proliferation of religious organizations and denominations which compete
vigorously for members and produce religious innovations that are often
exported to other countries (e.g. the Pentecostal movement). Recent work
builds on and expands the frontiers of the economics of religion beyond the
original religious club and religious market models.
Chapter 2 by Jean-Paul Carvalho is titled “Religious Clubs: The Strategic
Role of Religious Identity.” It examines the internal organization of religious
clubs and explores the connections between the economics of religion and
recent work on the economics of identity (Akerlof and Kranton 2000, 2010).
Stigmatizing requirements in terms of dress, diet and behavior not only
demarcate community boundaries but also provide strategic benefits to group
members. In the original club model by Iannaccone (1992), these requirements
mitigate free-rider problems through (1) screening out free riders and (2)
inducing substitution from outside activity to group participation. Carvalho
reviews four new strategic roles for religious identity, which add a behavioral
element to the economic structure of religious club models. First, religious
costs promote social sorting, that is, matching of agents with similar social and
personality traits (Carvalho forthcoming). One such sorting mechanism may
be “exotic” or “weird” religious beliefs. Members may be attracted to a group
not out of deep adherence to its belief system, but because they can entertain
its beliefs at low cost and know only those with similar characteristics can
do so. Second, religion may play a psychological role in esteem maintenance.
Building on advances in behavioral economics, religious participation can be
modeled as a means of coping with unfulfilled aspirations and other forms
of relative deprivation (Binzel and Carvalho 2017). Third, costly forms of
religious identity, such as veiling among Muslim women, can serve as a religious
commitment to observing conservative norms of behavior (Carvalho 2013).
Fourth, religious participation rules and restrictions on outside interactions are
critical to the cultural transmission of religious beliefs and preferences (Carvalho
2016). All of the strategic functions of religious club membership reviewed in
this chapter are examples of the theory of the second best (Lipsey and Lancaster
1956).
1 Introduction 7
Chapter 3 by Michael McBride, titled “Spatial Models of Religious Market

Competition: A Critical Assessment,” reviews applications of spatial com-
petition models (Hotelling 1929) to religious markets. In the basic model
introduced by Barros and Garoupa (2002), religious organizations compete for
adherents by setting a level of religious strictness, that is, choosing a location
in strictness space, usually assumed to be linear. Observing the strictness
of each group, religious consumers decide which group to join. Religious
consumers receive a benefit from affiliating with a religious group and pay
a cost which is increasing in the distance between their ideal strictness and
the strictness of the group. The equilibrium structure has the feature that
groups want to locate away from their competitors to maximize market share,
but not too far away as to invite entry by a new religious entrepreneur. The
key properties of the equilibrium structure are religious pluralism, defined
as the number of active groups, and religious participation, defined as the
proportion of consumers who are affiliated. Comparative statics focus on how
religious pluralism and participation depend on the distribution of consumer
preferences over strictness, market size, entry costs, and attractiveness of secular
alternatives. Various models also assume different objective functions on the
part of religious organizations, most commonly maximizing members’ welfare
and maximizing group size. McBride points to an important missing feature
of existing work: dynamics. In these static models, the distribution of religious
preferences is exogenous, as is the size of the religious market. McBride
encourages scholars studying religious competition to incorporate religious
capital formation (Iannaccone 1990), cultural transmission (Bisin and Verdier
2000), and endogenous fertility (McBride 2015) into models. Works by Chen
et al. (2018) and Carvalho and Sacks (2018) have already made progress in this
direction.
The main obstacle to fully incorporating endogenous religious preferences,
fertility, and other factors into models of religion is analytical tractability.
Chapter 4 by Michael Makowsky, titled “When Average Is Irrelevant: Com-
putational Modeling of Religious Groups,” reviews computational approaches
to modeling religion which overcome some of the limitations of analytic
models. Computational modeling has been largely neglected in microeco-
nomics, despite its usefulness in discovering complex social behavior and
serving as a form of data collection for analytical work. In the economics
of religion, computational models have been used to study complex patterns
of religious club formation and competition. Makowsky’s chapter shines a
spotlight on deep agent heterogeneity (in preferences, spatial location, etc.)
and especially the role of extremists in driving religious participation and
market structure. Heterogeneous populations exhibit qualitatively different
religious behavior to standard analytic models with homogenous agents. For

example, moderate groups can survive in heterogeneous populations using
a weak form of screening that permits some free riding. One measure of
interest is the distribution of zealots and free riders within and between groups.
Makowsky discusses how existing models of church-sect cycles and religious
capital formation in clubs could be enriched by computational approaches,
including the explicit incorporation of social network structure. He too calls
for development of a research program on endogenous religious preferences
and describes ways in which computational models can contribute to this
program.
Chapter 5 chapter by Gilat Levy, titled “The Intelligent Design of Reli-
gious Beliefs,” systematically explores the role of religious beliefs in religious
organizations, reviewing work by Levy and Razin (2012, 2014a, b) and others.
Religious beliefs are often modeled as either a preference or a good that
is consumed. Levy treats religious beliefs in a manner more familiar to
economists as a probability distribution over states of the world which is
updated based on empirical evidence. The content of religious beliefs varies
depending on the context. In the rewards-punishment model, individuals have
different beliefs concerning the connection between their actions and negative
payoff shocks they experience. Religious believers believe that cooperative
behavior in social interactions leads to positive payoff shocks (i.e. a reward
from a supernatural entity) and cheating leads to negative payoff shocks (i.e.
supernatural punishment). Non-believers believe that payoff shocks are inde-
pendent of their actions. Hence religious beliefs are instrumental in promoting
cooperative behavior. In the predestination model, individuals update their
beliefs about whether they are part of a predestined elect based on experience.
Costly religious rituals have an important social function, signaling religious
belief to partners in social interactions and self-signaling membership in an
elect. Common negative payoff shocks (e.g. natural disasters) can undermine
religious belief and lead to a collapse in cooperation. Levy proposes several
ways in which religious organizations “intelligently design” institutions which
make religious beliefs robust to such shocks. These include forgiveness and
atonement (which blur the link between actions and rewards/punishment),
personal responsibility, miracles, doctrine adaptation, and motivated beliefs.
Chapter 6 by Ronny Razin, titled “Religion and Segregation,” examines
the causes and consequences of geographical segregation of religious com-
munities. It has long been known that dietary restrictions serve to demarcate
community boundaries and prevent interactions across groups. Razin focuses
on geographical segregation between religious communities. Segregation has
several functions. First, geographic proximity enables group members to attend
1 Introduction 9
communal rituals which signal the religiosity of those participating. Second,

insofar as this signal promotes cooperation among group members, living close
to co-religionists enables members to enjoy the benefits of greater in-group
cooperation. Third, segregation permits intensive monitoring which further
stabilizes in-group cooperation. Such monitoring institutions are in place in
many strict groups including Calvinists and Mormons. Fourth, locating in
a remote place can itself deter entry by non-believers and non-cooperators.
Fifth, segregation mitigates religious conflict. Finally, segregation reinforces
religious beliefs through concentrated interactions among fellow believers and
preventing updating of beliefs through observation of the outcomes of others.
Chapter 7 by Anja Prummer, titled “Religious and Cultural Leaders,”
reviews theoretical work on the role of leaders in cultural transmission.
Bisin and Verdier (2000) contribute to the cultural transmission literature
by modeling socialization choice by parents. This changes the qualitative
properties of cultural dynamics, stabilizing cultural diversity, among other
things. Prummer surveys recent work on the role of cultural leaders in shaping
this process (e.g. Hauk and Mueller 2015; Verdier and Zenou 2015, 2018;
Prummer and Siedlarek 2017). This is highly relevant to religious beliefs and
preferences which are formed not only through socialization in the family and
through social interactions but also by exposure to leaders who have their
own cultural objectives. Prummer outlines a model in which homogeneity
prevails in the absence of cultural leaders. It is leaders that generate persistent
cultural diversity. Prummer shows how the intensity of religious preferences
is decreasing in labor market opportunities and the cost of providing club
goods to members and increasing in government transfers and community
size. Similar results emerge under competition among religious leaders. Finally,
Prummer discusses conditions under which we would expect religious and
cultural leaders to emerge.
Chapter 8 by Daniel Chen, titled “Intermediated Social Preferences: Altru-
ism in an Algorithmic Era,” examines the effect of institutions on social
preferences, an issue closely related to the economics of religion. Through-
out history, religious organizations have played a dominant role in shaping
preferences and intermediating social interactions. Today, new sources of
intermediation are arising, and it is an open question how these technologies
will shape human psychology. Chen presents original experimental evidence
that intermediation by algorithms reduces altruism. Workers are recruited
through Amazon Mechanical Turk to complete a translation task. After
workers complete a task, they grade the work of another worker and decide
how to split a bonus between that worker and themselves (i.e. a dictator game).
When the implemented split is partially controlled by a computer, the dictat-

ing worker is significantly more selfish than when the implemented split is
partially controlled by other workers. Hence computer intermediation affects
social preferences. This has important implications for models of religion,
suggesting among other things that altruism toward in-group members may
be contingent on religious institutions involving personal interaction.
Empirical Advances in the Economics of Religion

Part II of the book presents the key developments in empirical approaches to
the economics of religion. If there is one key development in the economics
of religion over the past ten years, it is the growth of high-quality empirical
research that uses a range of data sources, both historical and contemporary,
as well as a range of methodological approaches to investigate the influence of
religion on society (Iyer 2016). The fields covered by this research include edu-
cation, demography, public finance, conflict, inequality, and macroeconomic
policy. This part of the book demonstrates both the diversity and the high
quality of this research.
Chapter 9 by Sriya Iyer explores the interactions between religion and
demography. Iyer discusses why we should be interested in the question of
religion and demography and recent trends in the world’s religious demogra-
phy such as the growth of Christianity in Africa, Hinduism in South Asia, and
Islam worldwide. She also locates this within the wider context of what major
world religions think theologically about demographic decisions, highlighting
the various mechanisms through which religion might affect fertility behavior,
as well as the channels through which religion affects mortality, especially
physical and mental health behaviors. These channels include particularized
theology, characteristics, or minority group status.
Chapter 10 by Dan Hungerman and Timothy Weninger explores the
relationship between church and state, an important theme in the economics
and sociology of religion. The chapter explores their recent work on policies
such as school vouchers in the United States and how this affects the market
for religion. They discuss new data based on weekly church bulletins, showing
that the study of church-state relations continues to be an exciting area of
research. They suggest that policies which concern secular forces can often
motivate methodologically rigorous studies of religious practice in situations
where no better methodology is available. They argue that as religion continues
to decline in the United States, the importance of secular forces may only grow.
1 Introduction 11
Chapter 11 by Robert Barro and Rachel McCleary examines investment

in education across various types of Protestantism in Guatemala and also
compares it to Catholicism. The work explores various theological distinctions
between Mainline Protestant denominations and premillennialist movements
such as Evangelical and Pentecostal movements. The authors argue that these
denominations placed less emphasis on investment in education, so that
literacy is helped more by Mainline Protestant schools than by Catholic
schools. The chapter suggests that research on Protestantism also needs to take
into account different types of Protestantism in Guatemala and how they differ
from Mainline Protestants.
Chapter 12 by Danny Cohen-Zada and Moshe Justman discusses the
religious factor in private education in the United States. Private education
accounts for 8% of enrolment in primary and secondary schools in the United
States, but 80% of these private school students attend religious schools. The
chapter combines theoretical modeling with structural empirical estimation,
and some calibrated policy analyses. The authors show that as a result of the
highly decentralized structure of public education in the United States, the
share of public, private-secular, and private-religious schools varies by school
district as well as spending per student in public schools. Secular private
schools charge more in fees than public schools spend on their students, and
religious private schools charge less, on average. Students in religious private
schools come from all socioeconomic strata.
Chapter 13 by Sascha Becker and Ludger Woessmann discusses the con-
sequences of the Protestant Reformation for education of the population
and historical development in Germany. The authors show the importance
of analyzing historical data with modern methods. They summarize recent
research on the effects of the Protestant Reformation on education and
various other outcomes, including economic development, fertility, suicide,
and secularization. Based on these insights, they also challenge several classic
hypotheses of social science including Marx’s argument about religion being
the opium of the people and Weber’s thesis about how the Protestant ethic
affected economic development in Europe.
In Chap. 14, Anirban Mitra and Debraj Ray illustrate an economic approach
to Hindu-Muslim conflict in India. Using extended data they show that
economic factors matter for religious conflict in India. Hence, increases in
Muslim expenditures are followed by greater levels of religious casualties,
while increases in Hindu expenditures dampen such violence. They highlight
the growth of strong political motives which manifest themselves in BJP
legislators. They argue that there is some evidence of a positive association
between the BJP’s share of Lok Sabha (Lower House of Parliament) seats in a
region and the extent of Hindu-Muslim conflict. They also conduct a careful
study of the city of Ahmedabad in Gujarat in these contexts.
Chapter 15 by José Montalvo and Marta Reynal-Querol examines religion
and conflict, with a background, as the authors write, that civilization fault
lines are a source of conflict. They argue that while religion should not be
considered the main cause of violence and conflict, it cannot be excluded
from accounts of international relations, having an impact on both inter-state
relations and domestic politics. The chapter assesses the Modifiable Areal Unit
Problem (MAUP), that the relationship between variables at one scale may
be distorted when analyzed using another scale. They suggest that the lack
of explanatory power of religion on economic development may be because at
high levels of aggregation, the positive effects of religion are less important than
its negative consequences. They suggest that further research of the impact
of MAUP on the relationship between religious diversity, polarization, and
conflict is needed.
In Chap. 16, Jeanet Bentzen discusses why some societies are more religious
than others, including a measure of religiosity that is comparable across
countries. The measure is constructed to compare religiosity within countries,
especially across sub-national districts. It also describes how to construct
variables from spatial data. The chapter shows that one reason for the large
differences in religiosity across the globe today is differences in unpredictabil-
ity, caused by differences in the risk and actual occurrence of earthquakes,
tsunamis, and volcanic eruptions. The chapter suggests that it is important to
link data on religiosity at a global level with other data available at the sub-
national district level.
Chapter 17 by Mohamed Saleh examines socioeconomic inequality across
religious groups, suggesting a deeper examination of the historical formation
process of groups that may be caused by self-selection on socioeconomic char-
acteristics. The chapter examines socioeconomic inequality between Copts
(Egyptian Christians) and the Muslim majority in Egypt also using two
nationally representative individual-level samples from the 1848 and 1868
population censuses. The chapter suggests that self-selected conversions that
were induced by the tax system are arguably a plausible explanation of the
Coptic-Muslim socioeconomic inequality that emerged in early medieval
Egypt. State policies in industrialization and the growth of modern education
in 1800–1950 altered the Coptic-Muslim socioeconomic gap in favor of Copts.
It was not until the arrival of public mass education in the 1950s that the gap
declined. Since we still know very little about the formation processes of other
non-Muslim minorities in the region, such as Greeks, Armenians, Karaite Jews,
1 Introduction 13
Rabbinic Jews, and Levantine Christians, this chapter allows us to consolidate

our understanding of non-Muslim minorities.
In Chap. 18, Benito Arruñada and Matthias Krapf adopt a more macro
perspective that deals with religion and the European Union. The chapter
first reviews the literature on cultural differences across Euro member states,
highlighting that cultural differences between Protestants and Catholics may
explain cross-country differences. The authors argue that a confessional culture
explains why Catholic countries have weaker institutions but are more open
to economic and political integration. They show that economic outcomes
and measures of institutional quality systematically differ across Catholic and
Protestant countries in the Euro zone and that this might have implications
for European integration. They also suggest that religious denomination may
be a major determinant of the Euro crisis because Catholic societies may be
more willing to give up national sovereignty in favor of European integration,
whereas Protestant countries have national institutions that are of higher
quality. They suggest an institutions-integration trade-off that is resolved
differently by Catholics and by Protestants.
In summary, research in the economics of religion has made very strong
strides in empirical economics. By combining both contemporary and his-
torical data with cutting-edge econometric and other techniques, to address
important public policy questions, this sub-field is truly vibrant and thriving.
Advances in Religion and Political Economy

Our final part overviews recent theoretical and empirical advances in the
role that religion plays in political economy. We devote an entire part to
this line of research because it is one of the fastest growing in the field.
Indeed, the last five years have seen a flowering of theoretical (Bénabou et al.
2016; Michalopoulos et al. forthcoming), empirical (Meyersson 2014; Nellis
and Siddiqui 2018; Brooke and Ketchley 2018), and economic history work
(Chaney 2013, 2016; Iyigun 2015; Cantoni 2015; Rubin 2017; Platteau 2017;
Spenkuch and Tillmann 2018; Johnson and Koyama 2019; Cantoni et al. 2018)
at the cross section of religion and political economy. This part contains ten
chapters from some of the leading figures conducting research in this area.
The first four chapters in this part focus on one common theme: political
legitimacy. In Chap. 19, entitled “The Political and Economic Consequences
of Religious Legitimacy,” Jared Rubin overviews ideas laid out in his recent
book (2017) Rulers, Religion and Riches: Why the West Got Rich and the Middle
East Did Not. Rubin provides a framework for understanding why the use of
religious legitimacy may differ across societies and what this means for political
and economic outcomes. As in his book, he focuses on the implications
of these political economy outcomes for the long-run economic divergence
between Western Europe and the Middle East. This chapter concludes with
implications for contemporary Middle Eastern political economy, with a
special focus on Iran, Saudi Arabia, and Turkey.
In Chap. 20, entitled “Religious Legitimacy and the Joint Evolution of
Culture and Institutions,” Alberto Bisin, Avner Seror, and Thierry Verdier
provide the theoretical underpinnings for the role religion plays in political
legitimacy. This chapter embeds a model of religious legitimacy into one
studying the joint evolution of culture and institutions (as in Bisin and Verdier
2017). Bisin, Seror, and Verdier show how legitimacy affects the evolution and
distribution of political power between secular and religious elites and thus the
implementation of policies related to religious beliefs and values. They apply
this model to understand political economy differences between the Middle
East and Western Europe, shedding light on why the former has more strongly
religious states than the latter.
In Chap. 21, Jean-Philippe Platteau focuses specifically on the political
economy of Islam. This chapter, entitled “Strategic Interactions Between
Religion and Politics: The Case of Islam,” lays out insights made in Platteau’s
(2017) recent book Islam Instrumentalized: Religion and Politics in Historical
Perspective. Platteau highlights the historical role of Islam as a “handmaiden to
politics” and analyzes what this means for the type of policies enacted in the
“politico-religious equilibrium” as well as the (in)stability of autocratic, Islamic
rule. This chapter also points out that political instability is exacerbated due
to (Sunni) Islam’s decentralized institutional structure, since it permits rival
clerics from expressing their dissatisfaction with the regime.
Chapter 22 is the final chapter focusing on religion and political legitimacy.
Metin M. Coşgel and Thomas J. Miceli’s “State and Religion: An Economic
Approach” presents a theory of the relationship between religion and the
state, providing intuition for the conditions under which the state suppresses
religion, endorses an official religion, or permits religious competition. This
chapter builds on the theory laid out in Coşgel and Miceli (2009), which
studies the trade-offs rulers face between religious legitimation, tax collection,
and “religious good provision.” This theory permits intuition for a wide range
of phenomena, including the long-run decline of state religions, the appeal of
theocracy, and suppression of religious minorities.
The next set of chapters in this part discuss some of the broader implications
of religion and political economy. In Chap. 23, “A Great Academic Re-
awakening: The Return to a Political Economy of Religion,” Anthony Gill
1 Introduction 15
provides an overview of the intellectual history of this field. As is suggested

by the chapter’s title, Gill is (rightfully) optimistic about recent work in the
political economy of religion as well as its future. Gill notes that this had
long been a relatively neglected field within political science, and certainly
within economics, but the recently revitalized literature has analyzed the
political economy of religion from two different perspectives. The “ideational”
perspective considers the importance of theological ideas on political economy
outcomes, while the “economic” perspective employs rational choice analyses
to understand the actions of religious players in political economy settings.
In Chap. 24, Noel Johnson and Mark Koyama offer an account of religious
freedom that is tied to the rise of modern states. This chapter, entitled “The
State, Toleration, and Religious Freedom,” builds on arguments laid out
in their (2019) book Persecution and Toleration: The Long Road to Religious
Freedom. Johnson and Koyama suggest that it was only with the rise of states
capable of enforcing the rule of law—applicable to both minority groups and
those in power—that religious freedom and peace were allowed to spread.
In weak states—states that were unable to always uphold the rule of law—
religious minorities were subject to occasional persecution because the state
could treat minorities differently from others when it felt necessary. It was only
with the rise of the rule of law, which is applied “impersonally” to all subjects,
that the state’s hands were tied with respect to persecuting one minority
religious group in favor of another.
The theme of religion and freedom is also the subject of Chap. 25, Laura
Mayoral and Joan Esteban’s “Religiosity and Economic Performance: The Role
of Personal Liberties.” This chapter notes that most religions historically placed
some types of constraints on personal liberties, often sanctioned by the state.
These sanctions have been loosening over the last half century in Europe,
especially with respect to laws pertaining to abortion, women’s rights, LGBT
rights, and divorce. Mayoral and Esteban show how this loosening affects the
labor supply, with more secular types providing more labor as they are less
restricted while more religious types provide less labor.
The final three chapters consider the role of religion in large-scale political
economy outcomes. Chapter 26, Murat Iyigun’s “Causes and Consequences
of Monotheism in the Supply of Religion,” builds off his 2015 book War,
Peace, and Prosperity in the Name of God. Iyigun shows how monotheistic
religious leaders were in a better position to legitimate rule than polytheistic
leaders. He shows how monopoly religion/gods carried all sorts of social and
political advantages, permitting somewhat uniform social order, increased
political stability, economies of scale, longer time horizons (due to belief in the
afterlife), greater unity in the face of external threats, and limits on free riding.
Iyigun elaborates how all of these “evolutionary” advantages enabled the long-
run success of monotheistic religions, even if it also meant they frequently
confronted each other.
In Chap. 27, Eric Chaney considers how the interaction between religion
and political economy affected human capital formation in the Islamic world.
This chapter, entitled “Religion, Political Power and Human Capital Forma-
tion: Evidence from Islamic History,” appeals to his own voluminous empirical
work on the rise and fall of Muslim science (Chaney 2016). In this chapter,
Chaney shows that the rise of secular bureaucratic institutions in early Islam
led to a flourishing of scientific output (in the period known as the “Golden
Age of Islam”), as the emergence of rationalized forms of administration was
associated with an increased demand for secular knowledge. However, as the
political power of religious leaders increased—beginning especially in the late
eleventh and twelfth centuries, a period known as the “Sunni Revival”—
these bureaucratic structures collapsed and as a result so did Muslim scientific
output.
The book’s final chapter is Alireza Naghavi’s “Islam, Trade, and Innova-
tion.” Naghavi seeks answers to a similar question posed in Eric Chaney’s
chapter, namely, “why did innovation slow down in the Islamic world despite
such an early fluorescence?” Naghavi notes that this question is especially
puzzling since early Islamic innovation was linked closely to trade (which
played a key role in the development of early Islamic principles, as noted in
Michalopoulos et al. forthcoming), but trade continued to flow while innova-
tion did not. Naghavi considers numerous explanations which may account for
both phenomena, including changes in political economy, tolerance, diversity,
and institutions.
In short, the “political economy of religion” is a vibrant field, with scholars
tackling questions on all aspects of political economy. While the field currently
has a slight overemphasis on the Muslim world, this is not completely surpris-
ing given the important role that Muslim religious authorities play in politics.
Indeed, there is still much low-hanging fruit for scholars to pick. There has
been little social scientific work on the role of religion in societies in which
non-Abrahamic faiths dominate. This includes Eastern religions (Buddhism,
Hinduism, Shintoism, etc.) as well as indigenous religions. The latter may
be especially important when studying places like sub-Saharan Africa and
Latin America, where indigenous religions have been increasingly confronted
by Christianity and Islam in the twentieth and twenty-first centuries. The
consequence for political economy in these states is not obvious, and it is
certainly worthy of investigation.
1 Introduction 17
References
Akerlof, George A., and Rachel E. Kranton. 2000. Economics and Identity. Quarterly
Journal of Economics 415 (3): 715–753.
———. 2010. Identity Economics: How Our Identities Shape Our Work, Wages and
Wellbeing. Princeton: Princeton University Press.
Almond, Douglas, Bhashkar Mazumder, and Reyn van Ewijk. 2015. In Utero
Ramadan Exposure and Children’s Academic Performance. Economic Journal 125
(589): 1501–1533.
Anderson, Robert Warren, Noel D. Johnson, and Mark Koyama. 2017. Jewish
Persecutions and Weather Shocks: 1100–1800. Economic Journal 127 (602): 924–
958.
Azzi, Corry, and Ronald G. Ehrenberg. 1975. Household Allocation of Time and
Church Attendance. Journal of Political Economy 83 (1): 27–56.
Barro, Robert J., and Rachel M. McCleary. 2003. Religion and Economic Growth
Across Countries. American Sociological Review 68 (5): 760–781.
Barros, Pedro Pita, and Nuno Garoupa. 2002. An Economic Theory of Church
Strictness. Economic Journal 112 (481): 559–576.
Becker, Sascha O., and Ludger Woessmann. 2009. Was Weber Wrong? A Human
Capital Theory of Protestant Economic History. Quarterly Journal of Economics
124 (2): 531–596.
———. 2018. Social Cohesion, Religious Beliefs, and the Effect of Protestantism on
Suicide. Review of Economics and Statistics 100 (3): 377–391.
Becker, Sascha O., Steven Pfaff, and Jared Rubin. 2016. Causes and Consequences of
the Protestant Reformation. Explorations in Economic History 62: 1–25.
Bénabou, Roland, Davide Ticchi, and Andrea Vindigni. 2016. Forbidden Fruits: The
Political Economy of Science, Religion, and Growth. NBER Working Paper 21105.
Berman, Eli. 2009. Radical, Religious and Violent: The New Economics of Terrorism.
MIT Press.
Binzel, Christine, and Jean-Paul Carvalho. 2017. Education, Social Mobility and
Religious Movements: The Islamic Revival in Egypt. Economic Journal 127 (607):
2553–2580.
Bisin, Alberto, and Thierry Verdier. 2000. Beyond the Melting Pot: Cultural Trans-
mission, Marriage, and the Evolution of Ethnic and Religious Traits. Quarterly
———. 2017. On the Joint Evolution of Culture and Institutions. National Bureau of
Economic Research Working Paper 23375.
Botticini, Maristella, and Zvi Eckstein. 2012. The Chosen Few: How Education Shaped
Jewish History, 70–1492. Princeton: Princeton University Press.
Brooke, Steven, and Neil Ketchley. 2018. Social and Institutional Origins of Political
Islam. American Political Science Review 112 (2): 376–394.
Campante, Filipe, and David Yanagizawa-Drott. 2015. Does Religion Affect Eco-
nomic Growth and Happiness? Evidence from Ramadan. Quarterly Journal of
Economics 130 (2): 615–658.
Cantoni, Davide. 2012. Adopting a New Religion: The Case of Protestantism in 16th
Century Germany. Economic Journal 122 (560): 502–531.
———. 2015. The Economic Effects of the Protestant Reformation: Testing the
Weber Hypothesis in the German Lands. Journal of the European Economic Associ-
ation 13 (4): 561–598.
Cantoni, Davide, Jeremiah Dittmar, and Noam Yuchtman. 2018. Religious Compe-
tition and Reallocation: The Political Economy of Secularization in the Protestant
Reformation. Quarterly Journal of Economics 133 (4): 2037–2096.
Carvalho, Jean-Paul. 2013. Veiling. Quarterly Journal of Economics 128 (1): 337–370.
———. 2016. Identity-Based Organizations. American Economic Review Papers and
Proceedings 106 (5): 410–414.
———. Forthcoming. Sacrifice and Sorting in Clubs. Forum for Social Economics.
Carvalho, Jean-Paul, and Michael Sacks. 2018. The Economics of Religious Commu-
nities: Social Integration, Discrimination and Radicalization. Working Paper, UC
Irvine.
Castelló-Climent, Amparo, Latika Chaudhary, and Abhiroop Mukhopadhyay. Forth-
coming. Higher Education and Prosperity: From Catholic Missionaries to Lumi-
nosity in India. Economic Journal.
Chaney, Eric. 2013. Revolt on the Nile: Economic Shocks, Religion and Political
Power. Econometrica 81 (5): 2033–2053.
———. 2016. Religion and the Rise and Fall of Islamic Science. Working Paper.
Chen, Tongzhou, Michael McBride, and Martin B. Short. 2018. Dynamics of Religious
Group Growth and Survival. Manuscript.
Coşgel, Metin M., and Thomas J. Miceli. 2009. State and Religion. Journal of
Comparative Economics 37: 402–416.
Durlauf, Steven, Andros Kourtellos, and Chih Ming Tan. 2011. Is God in the Details?
A Reexamination of the Role of Religion in Economic Growth. Journal of Applied
Econometrics 27 (7): 1059–1075.
Ekelund, Robert B., Robert F. Hébert, and Robert D. Tollison. 1989. An Economic
Model of the Medieval Church: Usury as a Form of Rent Seeking. Journal of Law,
Economics, and Organization 5 (2): 307–331.
Ekelund, Robert B., Robert F. Hébert, Robert D. Tollison, Gary M. Anderson, and
Audrey B. Davidson. 1996. Sacred Trust: The Medieval Church as an Economic Firm.
New York: Oxford University Press.
Finke, Roger, and Rodney Stark. 1988. Religious Economies and Sacred Canopies:
Religious Mobilization in American Cities, 1906. American Sociology Review 53
(1): 41–49.
———. 1992. The Churching of America, 1776–1990: Winners and Losers in Our
Religious Economy. New Brunswick, NJ: Rutgers University Press.
1 Introduction 19
Fruehwirth, Jane, Sriya Iyer, and Anwen Zhang. Forthcoming. Religion and Depres-
sion in Adolescence. Journal of Political Economy.
Gaskins, Ben, Matt Golder, and David A. Siegel. 2013a. Religious Participation and
Economic Conservatism. American Journal of Political Science 57 (4): 823–840.
———. 2013b. Religious Participation, Social Conservatism, and Human Develop-
ment. Journal of Politics 75 (4): 1125–1141.
Gould, Eric D., and Esteban F. Klor. 2016. The Long-Run Effect of 9/11: Terrorism,
Backlash, and the Assimilation of Muslim Immigrants in the West. Economic
Journal 126 (597): 2064–2114.
Gruber, Jonathan, and Daniel M. Hungerman. 2008. The Church Versus the Mall:
What Happens When Religion Faces Increased Secular Competition? Quarterly
Guiso, Luigi, Paola Sapienza, and Luigi Zingales. 2003. People’s Opium? Religion
and Economic Attitudes. Journal of Monetary Economics 50 (1): 225–282.
Hauk, E., and H. Mueller. 2015. Cultural Leaders and the Clash of Civilizations.
Journal of Conflict Resolution 59: 367–400.
Hegghammer, Thomas. 2013. Should I Stay or Should I Go? Explaining Variation
in Western Jihadists’ Choice Between Domestic and Foreign Fighting. American
Political Science Review 107 (1): 1–15.
Hotelling, Harold. 1929. Stability in Competition. Economic Journal 39: 41–57.
Iannaccone, Laurence R. 1988. A Formal Model of Church and Sect. American Journal
of Sociology S94: 241–268.
———. 1990. Religious Practice: A Human Capital Approach. Journal for the
Scientific Study of Religion 29 (3): 297–314.
———. 1992. Sacrifice and Stigma: Reducing Free-Riding in Cults, Communes, and
Other Collectives. Journal of Political Economy 100 (2): 271–297.
———. 1994. Why Strict Churches Are Strong. American Journal of Sociology 99 (5):
1180–1211.
———. 1998. Introduction to the Economics of Religion. Journal of Economic
Literature 36: 1465–1496.
Iannaccone, Laurence R., Roger Finke, and Rodney Stark. 1997. Deregulating Reli-
gion: The Economics of Church and State. Economic Inquiry 35 (2): 350–364.
Iyer, Sriya. 2016. The New Economics of Religion. Journal of Economic Literature 54
(2): 395–441.
———. 2018. The Economics of Religion in India. Cambridge, MA: Harvard Univer-
sity Press.
Iyer, Sriya, and Anand Shrivastava. 2018. Religious Riots and Electoral Politics in
India. Journal of Development Economics 131: 104–122.
Iyigun, Murat. 2008. Luther and Suleyman. Quarterly Journal of Economics 123 (4):
1465–1494.
———. 2015. War, Peace, and Prosperity in the Name of God. Chicago: University of
Chicago Press.
Jha, Saumitra. 2013. Trade, Institutions, and Ethnic Tolerance: Evidence from South
Asia. American Political Science Review 107 (4): 806–832.
Johnson, Noel D., and Mark Koyama. 2019. Persecution & Toleration: The Long Road
to Religious Freedom. New York: Cambridge University Press.
Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle
East. Princeton: Princeton University Press.
———. Forthcoming. Islam and Economic Performance: Historical and Contempo-
rary Links. Journal of Economic Literature 56.
Levy, Gilat, and Ronny Razin. 2012. Religious Beliefs, Religious Participation and
Cooperation. American Economic Journals: Microeconomics 4 (3): 121–151.
———. 2014a. Rituals or Good Works: Social Signalling in Religious Organizations.
Journal of European Economic Association 12 (5): 1317–1360.
———. 2014b. Calvin’s Reformation in Geneva: Self and Social Signalling. Journal
of Public Economic Theory 16 (5): 730–742.
Lipsey, R.G., and K. Lancaster. 1956. The General Theory of Second Best. Review of
Economic Studies 24 (1): 11–32.
McBride, Michael. 2008. Religious Pluralism and Religious Participation: A Game
Theoretic Analysis. American Journal of Sociology 114: 77–108.
———. 2010. Religious Market Competition in a Richer World. Economica 77: 148–
171.
———. 2015. Why Churches Need Free-Riders: Religious Capital Formation and
Religious Group Survival. Journal of Behavioral and Experimental Economics 58:
77–87.
McCleary, Rachel, ed. 2011. Oxford Handbook of the Economics of Religion. Oxford:
Oxford University Press.
McClendon, Gwyneth, and Rachel Beatty Riedl. 2015. Religion as a Stimulant of
Political Participation: Experimental Evidence from Nairobi, Kenya. Journal of
Politics 77 (4): 1045–1057.
McKenzie, Brian D., and Stella M. Rouse. 2013. Shades of Faith: Religious Foun-
dations of Political Attitudes Among African Americans, Latinos, and Whites.
American Journal of Political Science 57 (1): 218–235.
Meyersson, Erik. 2014. Islamic Rule and the Empowerment of the Poor and Pious.
Econometrica 82 (1): 229–269.
Michalopoulos, Stelios, Alireza Naghavi, and Giovanni Prarolo. Forthcoming. Trade
and Geography in the Spread of Islam. Economic Journal.
Mitra, Anirban, and Debraj Ray. 2014. Implications of an Economic Theory of
Conflict: Hindu-Muslim Violence in India. Journal of Political Economy 122 (4):
719–765.
Montgomery, James D. 1996. Dynamics of the Religious Economy: Exit, Voice, and
Denominational Secularization. Rationality and Society 8 (1): 81–110.
———. 2003. A Formalization and Test of the Religious Economies Model. American
Sociological Review 68: 782–809.
1 Introduction 21
Nellis, Gareth, and Niloufer Siddiqui. 2018. Secular Party Rule and Religious Violence
in Pakistan. American Political Science Review 112 (1): 49–67.
Noland, Marcus. 2005. Religion and Economic Performance. World Development 33
(8): 1215–1232.
Pascali, Luigi. 2016. Banks and Development: Jewish Communities in the Italian
Renaissance and Current Economic Performance. Review of Economics and Statistics
98 (1): 140–158.
Platteau, Jean-Philippe. 2017. Islam Instrumentalized: Religion and Politics in Historical
Perspective. Cambridge: Cambridge University Press.
Prummer, A., and J.-P. Siedlarek. 2017. Community Leaders and the Preservation of
Cultural Traits. Journal of Economic Theory 168: 143–176.
Rubin, Jared. 2014. Printing and Protestants: An Empirical Test of the Role of Printing
in the Reformation. Review of Economics and Statistics 96 (2): 270–286.
———. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East
Did Not. New York: Cambridge University Press.
Seror, Avner. 2018. A Theory on the Evolution of Religious Norms and Economic
Prohibition. Journal of Development Economics 134: 416–427
Spenkuch, Jörg L., and Philipp Tillmann. 2018. Elite Influence? Religion and the
Electoral Success of the Nazis. American Journal of Political Science 62 (1): 19–36.
Stegmueller, Daniel. 2013. Religion and Redistributive Voting in Western Europe.
Journal of Politics 75 (4): 1064–1076.
Valdez, Inés. 2016. Nondomination or Practices of Freedom? French Muslim Women,
Foucault, and The Full Veil Ban. American Political Science Review 110 (1): 18–30.
Verdier, Thierry, and Yves Zenou. 2015. The Role of Cultural Leaders in the
Transmission of Preferences. Economics Letters 136: 158–161.
———. 2018. Cultural Leaders and the Dynamics of Assimilation. Journal of Eco-
nomic Theory 175: 374–414.
Voigtlaender, Nico, and Hans-Joachim Voth. 2012. Persecution Perpetuated: The
Medieval Origins of Anti-Semitic Violence in Nazi Germany. Quarterly Journal
of Economics 127 (3): 1339–1392.
Woodberry, Robert D. 2012. The Missionary Roots of Liberal Democracy. American
Young, H. Peyton. 1998. Individual Strategy and Social Structure: An Evolutionary
Theory of Institutions. Princeton: Princeton University Press.
Part I
Theoretical Advances in the Economics
of Religion
2
Religious Clubs: The Strategic Role
of Religious Identity
Jean-Paul Carvalho
Introduction
In the summer of 1967, college students from around the United States
descended on the Haight-Ashbury district of San Francisco, squatting, playing
music in the park, and providing free food, clothing, and medical care.
This became known as the Summer of Love. The aim of this quasi-religious
utopian movement was to create a community that was at once inclusive and
cooperative. By all accounts participants had a wonderful time. Why then was
it the Summer of Love, and not the Summers of Love?
By the end of the summer, the movement had made national news. The next
year, the college students arrived once more. But this time they were joined by
people with no spiritual or ideological connection to the movement—the free
riders there for the free food, free rent, and free love of the flower children.
The movement collapsed under the weight of its own success.
I am grateful for comments by Larry Iannaccone, Mark Koyama, Mike McBride, Pete Richerson, Stergios
Skaperdas, and participants at the IMBS conference on Identity, Cooperation and Conflict (2018) and
the UCI Program in Religious Studies lecture series.
J.-P. Carvalho ()
Department of Economics and Institute for Mathematical Behavioral Sciences,
University of California, Irvine, Irvine, CA, USA
e-mail: jpcarv@uci.edu

26 J.-P. Carvalho
A successful social movement requires the production of both spiri-

tual/ideological and material goods. Free riders undermine both forms of
production. First, they consume the material goods collectively produced by
the group, without contributing, depleting the group’s existing resources and
undermining reciprocal altruism. Second, they erode the spiritual/ideological
life of the community with an infectious lack of faith in the movement’s
mission. Without a way to screen out these free riders, such communities are
self-undermining. Their success begets failure.
A great deal can be learned from religious communities in this regard.
Religious organizations are among the most stable and long-lived institutions
in human history. They have a remarkable ability to survive even while at odds
with their host society. For example, Sosis and Bressler (2003) find that over
the nineteenth century religious communes were stricter and far longer-lived
than secular communes. The economics of religion provides insights into why
religious groups succeed where other social movements fail (see Iannaccone
1998; Iyer 2016). This chapter reviews and extends the theory of religious clubs
developed by Iannaccone (1992, 1994).
We depart from the original literature by making extensive reference to
religious identity. The economics of identity (Akerlof and Kranton 2000,
2010) is connected to the economics of religion in important and largely
unrecognized ways. In models of identity, individuals are assigned (and often
assign themselves) to one or more social categories, based on observable
characteristics. These characteristics can be ascriptive: sex, skin color, physical
build. Or they could be chosen: speech (accents, dialects), styles of dress,
religious symbols. Social identity includes gender, race, nationality, and many
other things. Religious identity, more specifically, comes from religious group
membership and markers thereof. Like other forms of identity, religious group
membership can cut across existing social ties (e.g. families, neighborhoods)
and reshape an individual’s pattern of social interactions. This turns out to be
important.
Central to the ubiquity and longevity of religious organizations is the
development of various means of screening out free riders and stabilizing
in-group cooperation. The club structure of religious groups plays a critical
role in this process. Clubs are groups that impose costly entry requirements
and participation rules. Religious clubs produce religious identity through
observance of these rules which distinguish members from nonmembers and
define group boundaries. Conspicuous markers of religious identity include
unconventional forms of dress, speech, dietary, and sexual practices, all of
which stigmatize members in the broader society. This stigma is obviously
costly, but it plays a strategic role. In this chapter, we will examine six strategic
functions of religious identity, the most basic of which mitigate the free-rider
problem inherent in collective production.
The strategic role of religious identity rationalizes seemingly bizarre, ineffi-
cient, and irrational religious beliefs and practices. All six functions examined
in this chapter are examples of the theory of the second best (Lipsey and Lancaster
1956). The theory of the second best receives little attention but is at least as
important as the two fundamental theorems of welfare economics. It warns
against (improper) reductionism in economics. Rather than the stylized world
of the welfare theorems, most economic systems are characterized by deep
inefficiency with multiple causes. Acting locally to remove one inefficient
behavior can increase the inefficiency of a system, as this behavior may have
mitigated other sources of inefficiency. In the religious context, inefficiencies
due to imperfect monitoring, observability of types, commitment, and so on
produce (compensating) institutions which may seem unproductive without
the right theoretical lens and proper ecological knowledge.
The Canonical Model

At the core of the economics of religion lies the model of religious clubs
developed by Iannaccone (1992, 1994), with notable contributions by Berman
(2000), McBride (2008, 2010, 2015), and Aimone et al. (2013). Communities
face severe free-rider problems (Olson 1965; Cornes and Sandler 1986).
Members would like to join without contributing to the various club goods
produced by the community. But if all members did so, production would
collapse, inducing exit en masse from the community. In the club model of
religion, costly entry requirements and participation rules play two important
strategic roles in mitigating these free-rider problems:
1. Screening: by raising the cost of membership, stigmatizing forms of religious

identity ensure that only cooperators and true believers join the group.
2. Substitution: by lowering the payoff from outside activity, they induce
members to divert time and money from non-group activities to group
activities.
The strategic costs of religious membership are sometimes referred to as

“sacrifice” to distance them from more standard solutions to the free-rider
problem such as punishment, exclusion, fees, and so on, which are often
infeasible in religious groups due to monitoring constraints and taboos on the
use of financial incentives. Aimone et al. (2013) test the religious clubs model
28 J.-P. Carvalho
experimentally, finding that individuals playing a public goods (VCM) game

voluntarily joined groups that taxed private investment vis-à-vis contribution
to the group. This endogenous group formation produced substitution and
especially screening effects, which increased members’ welfare.
Hence the costs of religious identity and club membership are designed
(or emerge) to solve incentive problems associated with collective production
by stigmatizing members in mainstream society and segregating them from
nonmembers. We see this with many strict religious sects, including ultra-
Orthodox Jews (Berman 2000; Carvalho and Koyama 2016), Mormons
(McBride 2007), and strict Muslims (Carvalho 2013). The beauty of this
theory is that it provides a simple and powerful explanation for many types
of religious behavior that may be considered “irrational”.
The basic club model has been extended in various ways. Carvalho and
Koyama (2016) allow members to choose between time and money contri-
butions to the group. In low-wage environments, the standard substitution
effect applies. Religious groups impose stigmatizing membership rules that
“tax” outside activity in order to increase time contributions to the religious
club good. In high-wage environments, however, religious groups will lib-
eralize and lower stigmatizing requirements to encourage social integration
of members and benefit from their financial contributions. Carvalho and
Koyama also characterize conditions under which an expansion in outside
economic opportunities leads to stricter forms of religiosity and religious
polarization. Based on this analysis, they present an historical account of
Jewish emancipation in nineteenth-century Europe. Emancipation suddenly
exposed isolated and relatively homogenous communities to different eco-
nomic environments. Community responses differed not just quantitatively,
but qualitatively, and in a manner predicted by the theory. In Germany, a
liberal reform movement developed in response to emancipation, while ultra-
Orthodox Judaism emerged in Eastern Europe which was even stricter and
more isolationist than traditional Judaism.
McBride (2015) introduces to religious clubs models the notion of religious
capital (see Iannaccone 1990). Religious preferences, particularly willingness
to contribute to the group, are cultivated through religious participation. As
such, religious groups must grant new members a period of participation
without contribution, so that new members can develop religious preferences
and contribute more intensively at a later stage. This explains why many
religious groups tolerate some free riding by new members and emphasizes
the importance of this strategy to the success of the group.
The most recent work on religious clubs by Chen et al. (2018) and
Carvalho and Sacks (2018) adds competition and dynamics to the standard
religious clubs model. Chen et al. examine competition among groups with
heterogeneous objectives and endogenous fertility, whereas Carvalho and
Sacks introduce forward-looking religious leaders and cultural transmission
of religious preferences.
Religious Identity: Further Strategic Functions

The religious clubs model is a simple, economic rationalization for seemingly
inefficient and bizarre religious practices. However, one can go further than
the screening and substitution effects studied in the first generation of club
models by exploring a richer set of social and psychological motivations for
costly forms of religious identity. Recent research identifies four additional
strategic functions of religious costs:
3. Social sorting
4. Esteem maintenance
5. Religious commitment
6. Cultural transmission
Social Sorting
In the standard club model of religion, religious groups require various

sacrifices in order to screen out free riders and promote efficient production
of club goods. This screening function can be described as the “economic
role” of religious sacrifice and stigma. Carvalho (forthcoming) proposes a
complementary “social role” in religious clubs based on social sorting. Sorting
differs from the usual screening function of religious clubs in that it operates
on traits that do not directly affect club goods production.
Like many other types of groups, one of the primary functions of religious
clubs is to facilitate social interactions among members. Suppose individuals
vary in their ethnolinguistic background, taste in art and music, hobbies, and
so on. These can be termed “economically neutral” traits, because unlike coop-
erativeness or commitment to the group, these traits are unlikely to directly
affect an individual’s productivity in the group. Instead these social traits play
an important role in social interactions. In Carvalho’s model, individuals prefer
to interact with people possessing similar traits. This preference is referred to
as own-type bias.
30 J.-P. Carvalho
Consider a population in which individuals have one of two (economically

neutral) social types τ ∈ {A, B}. Suppose type A is the minority. Each player
joins one of two groups. The payoff to a type-τ agent who joins group k is
a function of (1) private consumption, (2) club good consumption in group
k, (3) the proportion of group k members of type τ (own-type bias). The
strength of own-type bias is parameterized by α. One may think that because
individuals prefer to interact with their own type, they will be perfectly sorted
into groups, with one group composed exclusively of A types and the other
of B types. But this is not necessarily the case. This is where the social sorting
theory of clubs intersects with the standard club goods model and its focus on
screening out free riders. It turns out that the usual free-rider problem in club
goods production can inhibit sorting.
In the absence of membership costs, a perfect sorting equilibrium exists if
and only if own-type bias is large: α ≥ α for some threshold α > 0. For
α ∈ (0, α), sorting breaks down. Suppose all A types join group 1 and all
B types group 2. Because group 1 is smaller (as A is the minority type), it
will face a less severe free-rider problem in collective production and better
produce club goods including mutual insurance (Olson 1965). To gain access
to these club goods, B types would deviate and join group 1, breaking sorting.
Suppose now that group 1 imposes a membership cost c > 0 on all group
members (e.g. stigmatizing form of identity). Because the group is initially
composed exclusively of A types and due to own-group bias, A types value
membership in group 1 more highly than deviating B types. Hence there
exists a positive (uniform) membership cost that deters B types from joining
group 1 while retaining A types, stabilizing a perfect sorting equilibrium. The
minimum membership cost required for perfect sorting is strictly decreasing
in the share of the minority population. Costly sacrifices are demanded not by
groups catering to the majority but by those attracting rare/exotic types. The
rarer the type, the more costly the sacrifice required to achieve sorting.
One costly sacrifice that is neglected in the literature is the cost of enter-
taining “weird” beliefs. Many religious groups have elaborate and exotic belief
systems. Carvalho’s theory of religious clubs suggests that these belief systems
may play a strategic role in social sorting. The more exotic the belief system,
the more costly it may be to entertain for some types relative to others.
Hence individuals may join a group not because they adhere deeply to its
belief system, but because they can hold its beliefs at low cost and know
only those with similar characteristics can do so. Thus religious beliefs come
to be associated with certain traits and become a basis for sorting in social
interactions.
Esteem Maintenance
Club models of religion rationalize strict and unusual religious rules on

economic grounds. Binzel and Carvalho (2017) add a psychological role for
these rules to religious club models, proposing that religion is part of the
“psychological immune system” (Bénabou and Tirole 2002). This conception
of religion has a long history, with Marx’s claim that religion is “the opium
of the people” being only its most famous statement. In particular, suppose
people care about relative performance, including comparisons with others
(envy) and with their own expectations (unfulfilled aspirations). Building on
recent advances in behavioral economics, Binzel and Carvalho develop the
first model connecting religion to unfulfilled aspirations and other forms of
relative deprivation. The principal idea is that religion boosts self-esteem by
minimizing the psychological loss that occurs when consumption falls below
an endogenous expectations-based reference point. For example, religion can
adjust downward an individual’s reference point or shift attention to other
(non-material) dimensions of comparison (see also Akerlof (2017), on the
production of esteem). This approach is consistent with empirical evidence
that religious people have higher self-esteem, life satisfaction, and ability to
withstand traumatic experiences (Smith et al. 2003).
In contrast to work in other fields which assumes that relative deprivation
automatically increases religiosity, Binzel and Carvalho show that individuals
face a tradeoff. They can work harder and participate less in religion to live up to
expectations, or they can immerse themselves in religious activity to better cope
with unfulfilled aspirations. It turns out there is a simple condition governing
which way individuals go: the coping effect dominates when expected social
mobility, inequality, and poverty are high.
They proceed to show how these psychological motivations for religiosity
interact with organizational dynamics. Suppose religious organizations have
capacity for providing standard club goods, such as leisure activities, social
insurance, and labor market contacts. Organizational capacity evolves with
religious participation. Denote religious organizational capacity by Kt in
period t. The evolution of religious capacity is depicted in Fig. 2.1. A stable
steady state occurs where the function cuts the 45◦ line from above. As
illustrated in panel (a), there could be multiple steady states: K ∗ and K ∗∗∗
are asymptotically stable, while K ∗∗ is not.
In periods of high expected social mobility and inequality, a negative eco-
nomic shock to individuals with high aspirations raises religious participation
through the coping effect. This is represented by the upward shift of the
32 J.-P. Carvalho
+1 +1
45° 45°
+1 +1
∗ ∗∗ ∗∗∗ ∗ ∗∗ ∗∗∗
(a) Equilibrium structure (b) Rise in religious participation
Fig. 2.1 (a) Multiple steady-state levels of religious capacity. (b) A temporary shock to
social mobility induces a large, permanent increase in religious participation/capacity.
Source: Binzel and Carvalho (2017)
function in panel (b). This builds up the organizational capacity of religious

groups, eliminating the low-religiosity steady states and tipping society into
the basin of attraction of the high-religiosity steady state K ∗∗∗ . Thus a negative
shock to social mobility can produce a large, rapid rise in religious participation
and organizational capacity, which persists even after the shock has dissipated
and expectations have adjusted. In this way, the psychological function of
religion can attenuate the free-rider problem faced by religious clubs.
Binzel and Carvalho apply their theory to the Islamic revival in Egypt—the
rise in religious piety and participation in the last quarter of the twentieth cen-
tury. They present evidence of a sudden and economically significant decline
in social mobility for educated Egyptians in the 1980s, due to rapid expansion
of education combined with a contraction in public sector employment. On
this basis, they argue that coping with unfulfilled aspirations was one of the
factors behind Egypt’s Islamic revival, a movement led by the educated middle
class.
Religious Commitment
In the canonical model, religious identity acts as a tax on outside activity. But
not all outside activity is created equal from a religious perspective. Except in
the strictest religious groups (e.g. the Amish, ultra-Orthodox Jews), economic
activity (e.g. education, labor force participation) is not as threatening as
forms of social interaction that violate religious norms (e.g. dating, eating with
nonmembers). Moderately strict religious groups would like members to enjoy
the economic benefits of integration while avoiding religiously prohibited

forms of consumption. Carvalho (2013) proposes that conspicuous religious
markers can facilitate such a partial integration strategy by committing group
members to religious norms of behavior even while they are outside the
monitoring range of their community.
A leading example is veiling. By veiling, we mean the various concealing
forms of dress, especially headcovering, that are worn today by Muslim
women. Amid current concerns over cultural assimilation and social integra-
tion of immigrants, veiling by Muslim women has become a symbol of cultural
rejection and separation, prompting various bans in France, Belgium, the
Netherlands, and parts of Italy, Spain, Switzerland, and Russia. The cultural
rejection/separation interpretation of veiling is in line with the basic religious
clubs model, but not with the version discussed here.
Let us place veiling in its proper context. Earlier forms of veiling were
aristocratic markers, signaling that a man could provide for his wife (or
wives) without them having recourse to work (Amer 2014). The new veiling
movement is as the name suggests a late twentieth-century innovation rather
than a preservation of ancient tradition (El Guindi 1981). In 1923, there was
a deveiling movement in Cairo initiated by an Egyptian woman named Huda
Sharawi after she attended a meeting of feminists in Rome. After that middle-
and upper-class women stopped veiling and indeed anthropologists report
virtually no veiling in public in 1970 (Abu-Lughod 1971). By 2000, however,
around 80% of women in Cairo wore some form of headcovering (Bayat
2007), in styles distinct from those of the past.
Carvalho (2013) addresses three unanswered questions: (1) Why do women
veil? (2) Why has there been a rise in veiling among Muslim women? (3) What
are the consequences of banning veiling?
The basic idea comes from the economics of identity. Being assigned to
a social category, either through self or social categorization, imposes cer-
tain behavioral prescriptions, that is, identity-specific ideals and expectations
(Akerlof and Kranton 2000, 2010). If a woman adopts a conspicuous religious
symbol by veiling, she is categorized as a strict Muslim and the ideals and
expectations of a strict Muslim are applied. Thus veiled women face different
behavioral prescriptions: they are expected to more strictly adhere to religious
rules and face greater stigma for violating religious proscriptions than unveiled
women. In this way, veiling acts as a commitment to observing religious norms.
One cannot walk into a bar wearing a headscarf; it is simply incongruous. One
attracts different kinds of friends and social encounters. Moreover, veiling is a
signal of this commitment to one’s community (see Patel 2012).
34 J.-P. Carvalho
This links the new veiling movement to rising economic opportunities for
Muslim women. When one takes a close look at the movement, one sees
educated, urban, working, middle-class women in its vanguard (e.g. El Guindi
1981). Whether it be among migrants to Cairo from rural Egypt or migrants
from Muslim societies to Europe, Muslim women face an expansion in eco-
nomic opportunities, especially educational and labor market opportunities.
In religiously conservative communities, women’s behavior is subject to intense
social scrutiny and judgment. Exploiting these new economic opportunities,
and in doing so venturing outside the monitoring range of one’s community,
can mean a loss of community esteem, a deterioration of marriage market
options, and access to community resources. By committing an individual to
religious norms of behavior outside of her community, veiling can be seen
as a kind of partial integration strategy, a way of participating in economic
and social life while minimizing temptation to violate religious norms and
preserving esteem in one’s community. In this sense, some forms of veiling
may not be regressive.
This changes our view of the consequences of banning veiling. Banning
veiling in public spaces may actually inhibit social integration, as women
segregate in the home or local community as a costly substitute. Even more sur-
prisingly, when embedding this in a model of intergenerational transmission
of preferences, Carvalho (2013) finds that bans on veiling can lead to higher
levels of religiosity in these communities. The mechanism is not blowback,
but is more subtle. Suppose a parent wants to control the behavior of her
child. She can use both internal and external control instruments to do so.
Banning veiling (an external instrument) can lead to substitution toward
internal control methods such as religious education. A parent will invest
more in religious education when the behavior of a religious child is very
different to the behavior of a non-religious child. Suppose, in response to a
ban on veiling, non-religious women integrate while not veiling and religious
women segregate in the home/community. This drives a wedge between the
behavior of religious and non-children women and increases the willingness
of parents to invest in religious education and other means of transmitting
religious preferences to their children. This can be viewed as a form of cultural
resistance. Thus bans on veiling aimed at integrating and secularizing Muslim
communities could in fact have the opposite outcome.
Veiling is a complex phenomenon and there are many potential reasons
for adopting the practice. Recent empirical work by sociologists, however,
supports Carvalho’s (2013) commitment-signaling theory of veiling (Aksoy
and Gambetta 2016). In particular, religious women tend to exhibit higher
rates of veiling when they are more exposed to modern influences including
education, urbanization, and contact with non-Muslims.
Let us return to the standard club model, in which religious identity
stigmatizes members of a religious group and segregates them from mainstream
society. We have distinguished here between two domains in which stigma
can operate, the economic and the social. If in the extreme a religious leader
wishes to set up a fully segregated community, she needs to impose rules such
as avoiding money and modern technology that lead to economic as well as
social stigma and separation. Suppose instead a religious leader wants to build
a partially integrated community, with economically productive members
that adhere to non-mainstream behavioral norms. Then she might impose
requirements such as veiling that do not induce discrimination in education
and the labor market, but do induce social stigma. Veiling functions as a partial
integration strategy due to the combination of these two factors: (1) tolerance
of veiling in education and the labor market, combined with (2) stigma in
social interactions.
If either discrimination appears in the economic domain or veiling is
normalized (i.e. not stigmatized) in the social domain, then veiling loses its
power as a partial integration strategy. The response is likely to be extreme:
either community members fully integrate if economic returns are sufficiently
high or fully segregate otherwise.
Cultural Transmission
Economic models tend to treat tastes/preferences as primitive and fixed.

However, there are important cases in which preferences are endogenous to
the system being analyzed (Stigler and Becker 1977; Bowles 1998). Religion
in particular provides fertile ground for analyzing preference/belief formation.
Iannaccone (1990) models changes in religious preferences over an individual’s
lifetime as a product of conscious accumulation of religious capital. This model
can be extended to analyze the intergenerational transmission of religious
preferences, which would explain why the strongest determinant of individ-
ual religious affiliation and participation is parental religious affiliation and
participation (Iannaccone 1998). To model cultural transmission of religious
preferences more broadly, including oblique transmission (non-parent adult to
child) and horizontal transmission (child to child), however, one has to engage
with the literature on cultural transmission in evolutionary biology (Cavalli-
Sforza and Feldman 1981), evolutionary anthropology (Boyd and Richerson
1985, 2005), and economics (Bisin and Verdier 2000, 2001). The major break-
through of Bisin and Verdier (2000, 2001) is to make cultural transmission
36 J.-P. Carvalho
probabilities endogenous by allowing parents to invest in socialization of their

children. Among other things, they find that cultural diversity persists under
a larger set of conditions than in previous analyses with fixed transmission
probabilities.
One neglected aspect of the cultural transmission process is the fact that
it entails its own free-rider and externality problems. Suppose one wishes to
acquire a particular system of metaphysical and moral beliefs and makes costly
investments to do so (e.g. in education). These beliefs, if successfully acquired,
can be rapidly undermined by contact with nonbelievers. This is the externality
problem. Likewise, beliefs can be acquired through contact with believers,
without much personal investment. This is the free-rider problem. It is natural
that organizations emerge to solve and possibly exploit these problems.
Carvalho (2016) examines identity formation in groups, where (personal)
identity is modeled as a socially transmitted trait. We will interpret this
model in terms of the formation of religious beliefs. Unlike work on cultural
leaders (Hauk and Mueller 2015; Verdier and Zenou 2015; Carvalho et al.
2017; Prummer and Siedlarek 2017), group membership is endogenous in
Carvalho (2016). Religious leaders do not directly transmit cultural traits in
society, but rather indirectly shape cultural transmission in their own groups
through club membership, participation, and club goods production. In this
model, religious organizations cultivate religious belief by (1) imposing rules
of participation in belief-producing activities (e.g. communal prayer, scriptural
study, rituals) and (2) excluding nonmembers from social interactions. Because
belief formation is based on the exclusion of nonmembers, religious beliefs can
be viewed as a club good. Religious membership costs play a strategic role in
production of this psychic club good.
Consider a population composed of A and B types. A (B) types want
to acquire belief system A (B). Let B be the mainstream or default belief
system, which can be acquired outside of a group (e.g. through the education
system and media). Individuals can join one of two groups, 1 and 2, or can
be unaffiliated. Participation in group 1 (2) cultivates belief system A (B).
There are two stages of belief formation. Suppose an A type joins group 1
and chooses to spend proportion x of her time in the group on costly (belief-
forming) activities. With probability x, she becomes a carrier of trait A. With
probability, 1 − x direct transmission of beliefs fails and she acquires the
mainstream trait B. In the second stage, this individual is randomly matched
with another member of her group (nonmembers are excluded to avoid the
externality problem) and acquires the trait of her partner. Thus, the likelihood
that she ends up with trait A depends on the average participation level
in the group. Hence there is a severe free-rider problem. In the absence of
participation requirements, production of belief system A would collapse as

no group 1 member would choose a positive participation level. To mitigate
this free-rider problem in belief formation, groups choose a level of strictness,
which is a minimum participation requirement imposed on members.
In equilibrium, no type B individual is affiliated because they can acquire
their preferred belief system at no cost by interacting in mainstream society.
As such, group 2 attracts no members. In contrast, group 1, which cultivates
the oppositional belief A, can attract members as long as its strictness is
not too low as to only weakly produce beliefs and not too high as to make
membership prohibitively costly. How far can group 1 push strictness? There is
a natural concept of “religious tension” in this model. Tension is an important
concept in the sociology of religion. Stark and Finke (2000) propose that
“[a]ll religious groups can be located along an axis of tension between the
group and its sociocultural environment”, where tension is defined in terms of
“distinctiveness, separation, and antagonism” [p. 143]. Tension in Carvalho’s
model can be defined as the distaste for the mainstream belief system B by A
type individuals. When tension is high, it is more costly for A types to remain
unaffiliated and acquire a mainstream identity. Thus, group 1 can push its
members further in terms of participation and strengthen belief formation
under conditions of high tension. Of course, the results switch if A becomes
the mainstream belief system.
This theory has two main implications. First, it provides a link between
religious doctrine and religious participation that is missing from the eco-
nomics of religion, which has focused largely on religious practice. Specifically,
religious organizations that cultivate belief systems that are further from the
mainstream can be more extreme in terms of their practical demands of
members. Episcopalian congregations, for example, may be unable to make
strict demands of members because beliefs prevailing in mainstream society
are a close substitute for their belief system. Second, the theory provides a link
between religious doctrine and collective action. Religious organizations are
exceptionally effective at various forms of collective action, including public
good provision, political opposition, and violent rebellion (e.g. Gruber and
Hungerman 2007; Berman 2009). Berman (2009) identifies the strict club
structure of religious groups as the reason why they are more successful than
secular groups in collective action and in particular terrorism. Carvalho’s
(2016) theory attributes this strict club structure to belief systems that are at
odds with mainstream thought (e.g. apocalyptic beliefs). In light of this theory,
it is unsurprising that religious organizations which expend an extraordinary
38 J.-P. Carvalho
amount of resources on cultivating metaphysical and moral beliefs are so

effective at collective action.
Further analysis of the role of costly religious sacrifices in cultural transmis-
sion is conducted by Carvalho and Koyama (2016), Carvalho et al. (2017),
and Carvalho and Sacks (2018). Among other things, Carvalho and Sacks
(2018) study various dynamic radicalization strategies by religious groups.
One such strategy is to begin by forming a small group of highly committed
individuals and use the intensive participation of these members to cultivate
religious beliefs in subsequent generations. In this way, the group becomes
both stricter and larger over time. This radicalization strategy works only in
the absence of religious competition and when extremists in the community
have disproportionate influence over cultural transmission.
Conclusion
This chapter examined the strategic role of sacrifice and stigma in religious
clubs. Religious clubs impose costly entry requirements and participation rules
on members, including stigmatizing forms of dress, speech, and diet. These
entry requirements produce religious identity by defining group boundaries.
In the canonical model (Iannaccone 1992), religious prohibitions and pro-
scriptions are designed (or emerge) to solve incentive problems associated with
collective production by (1) screening out non-cooperators and (2) inducing
substitution from outside activity to group activity. Recent research explores
new social and psychological motivations for religion and identifies four
additional strategic functions of religious identity: (3) social sorting, (4) esteem
maintenance, (5) religious commitment, and (6) cultural transmission. This
produces new explanations for exotic religious beliefs, oppositional identity,
political mobilization by religious groups, and religious radicalization. All
strategic functions of religious identity are examples of the theory of the second
best.
References
Abu-Lughod, J. 1971. Cairo: 1001 Years of the City Victorious. Princeton, NJ: Princeton
University Press.
Aimone, J.A., L.R. Iannaccone, M.D. Makowsky, and J. Rubin. 2013. Endogenous
Group Formation Via Unproductive Costs. Review of Economic Studies 80: 1215–
1236.
Akerlof, G.A., and R.E. Kranton. 2000. Economics and Identity. Quarterly Journal of
Economics 415 (3): 715–753.
———. 2010. Identity Economics: How Our Identities Shape Our Work, Wages and
Wellbeing. Princeton, NJ: Princeton University Press.
Akerlof, R. 2017. Value Formation: The Role of Esteem. Games and Economic Behavior
102: 1–19.
Aksoy, O., and D. Gambetta. 2016. Behind the Veil: The Strategic Use of Religious
Garb. European Sociological Review 32 (6): 792–806.
Amer, S. 2014. What is Veiling?. Chapel Hill, NC: UNC Press Books.
Bayat, A. 2007. Making Islam Democratic. Stanford, CA: Stanford University Press.
Bénabou, R.J., and J. Tirole. 2002. Self-confidence and Personal Motivation. Quar-
terly Journal of Economics 117 (3): 871–915.
Berman, E. 2000. Sect, Subsidy, and Sacrifice: An Economist’s View of Ultra-
Orthodox Jews. Quarterly Journal of Economics 115 (3): 905–953.
———. 2009. Radical, Religious, and Violent: The New Economics of Terrorism.
Cambridge, MA: MIT Press.
Binzel, C., and J.-P. Carvalho. 2017. Education, Social Mobility and Religious Move-
ments: The Islamic Revival in Egypt. The Economic Journal 127 (607): 2553–2580.
Bisin, A., and T. Verdier. 2000. Beyond the Melting Pot: Cultural Transmission,
Marriage, and the Evolution of Ethnic and Religious Traits. Quarterly Journal of
Economics 115 (3): 955–988.
———. 2001. The Economics of Cultural Transmission and the Dynamics of
Preferences. Journal of Economic Theory 97: 298–319.
Bowles, S. 1998. Endogenous Preferences: The Cultural Consequences of Markets
and Other Economic Institutions. Journal of Economic Literature 36 (1): 75–111.
Boyd, R., and P.J. Richerson. 1985. Culture and the Evolutionary Process. Chicago, IL:
University of Chicago Press.
———. 2005. The Origin and Evolution of Cultures. New York, NY: Oxford Univer-
sity Press.
Carvalho, J.-P. 2013. Veiling. Quarterly Journal of Economics 128 (1): 337–370.
———. 2016. Identity-Based Organizations. American Economic Review 106 (5):
410–414.
———. Forthcoming. Sacrifice and Sorting in Clubs. Forum for Social Economics.
Carvalho, J.-P., and M. Koyama. 2016. Jewish Emancipation and Schism: Eco-
nomic Development and Religious Change. Journal of Comparative Economics 44
(3): 562–584.
Carvalho, J.-P., M. Koyama, and M. Sacks. 2017. Education, Identity, and Commu-
nity: Lessons from Jewish Emancipation. Public Choice 171 (1–2): 119–143.
Carvalho, J.-P., and M. Sacks. 2018. The Economics of Religious Communities: Social
Integration, Discrimination and Radicalization. Working Paper UC Irvine.
Cavalli-Sforza, L.L., and M.W. Feldman. 1981. Cultural Transmission and Evolution:
A Quantitative Approach. Princeton, NJ: Princeton University Press.
40 J.-P. Carvalho
Chen, T., M. McBride, and M.B. Short. 2018. Dynamics of Religious Group Growth
and Survival. Working Paper UC Irvine.
Cornes, R., and T. Sandler. 1986. The Theory of Externalities, Public Goods and Club
Goods. Cambridge: Cambridge University Press.
El Guindi, F. 1981. Veiling Infitah with Muslim Ethic: Egypt’s Contemporary Islamic
Movement. Social Problems 28 (4): 465–485.
Gruber, J., and D.M. Hungerman. 2007. Faith-Based Charity and Crowd-Out
During the Great Depression. Journal of Public Economics 91 (5–6): 1043–1069.
Journal of Conflict Resolution 59 (3): 367–400.
Iannaccone, L.R. 1990. Religious Practice: A Human Capital Approach. Journal for
the Scientific Study of Religion 29: 297–314.
Other Collectives. Journal of Political Economy 100(2): 271–291.
———. 1994. Why Strict Churches are Strong. American Journal of Sociology 99
(5): 1180–1211.
Literature 36 (3): 1465–1495.
Iyer, S. 2016. The New Economics of Religion. Journal of Economic Literature 54
(2): 395–441.
Lipsey, R.G., and K. Lancaster. 1956. The General Theory of Second Best. The Review
of Economic Studies 24 (1): 11–32.
McBride, M. 2007. Why Churches Need Free-Riders: Religious Capital Formation
and Religious Group Survival. Working Paper, University of California-Irvine.
———. 2008. Religious Pluralism and Religious Participation: A Game Theoretic
Analysis. American Journal of Sociology 114 (1): 77–108.
———. 2010. Religious Market Competition in a Richer World. Economica 77: 148–
171.
Religious Group Survival. Journal of Behavioral and Experimental Economics
58: 77–87.
Olson, M. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups.
Cambridge, MA: Harvard University Press.
Patel, D.S. 2012. Concealing to Reveal: The Informational Role of Islamic Dress in
Muslim Societies. Rationality and Society 24 (3): 295–323.
Smith, T.B., M.E. McCullough, and J. Poll. 2003. Religiousness and Depression:
Evidence for a Main Effect and the Moderating Influence of Stressful Life Events.
Psychological Bulletin 129 (4): 614–636.
Sosis, R., and E.R. Bressler. 2003. Cooperation and Commune Longevity: A Test of
the Costly Signaling Theory of Religion. Cross-Cultural Research 37 (2): 211–239.
Stark, R., and R. Finke. 2000. Acts of Faith: Explaining the Human Side of Religion.
Berkeley, CA: University of California Press.
Stigler, G.J., and G.S. Becker. 1977. De gustibus non est disputandum. American
Economic Review 67 (2): 76–90.
Verdier, T., and Y. Zenou. 2015. The Role of Cultural Leaders in the Transmission of
Preferences. Economics Letters 136: 158–161.
3
Spatial Models of Religious Market
Competition: A Critical Assessment
Michael McBride
Introduction
Understanding the nature of religious competition has been a primary goal of
the emerging field of the economics of religion (Iannaccone 1998; Iyer 2016).
Though discussion of religious competition dates back centuries (Anderson
1988), the first conceptualizations of religious competition in this new field
came not from economists but rather from sociologists inspired by economic
theories (Stark and Bainbridge 1985, 1987; Finke and Stark 1988, 1992;
Iannaccone and Stark 1994). Economists eventually rejoined the fray and over
the last 15 years have used spatial-location models (Hotelling 1929) to examine
the competition for adherents among religious groups.
The first task in applying a spatial model is defining the dimension in
which competition occurs. The most common assumption is to assume
that product differentiation and competition occur in a single dimension of
religious strictness; see Barros and Garoupa (2002), Barro and McCleary
(2005), McBride (2008), Ferrero (2008), McBride (2010), Gaskins et al.
(2013), and North and Gwin (2014). Religious strictness is the degree to which
a religious group requires absolutism and conformity to exclusive theologies
and behavioral codes. The higher the strictness, the more the outside world is
repudiated and the more distinctive the group. The dimension of competition
M. McBride ()
Department of Economics, University of California, Irvine, Irvine, CA, USA
e-mail: mcbride@uci.edu

44 M. McBride
is conceived in other ways by Poutvaara and Wagener (2010), Eswaran (2011),

Raynold (2013), Reda (2012), Iyer et al. (2014), and Berg et al. (2016). It can
also be intentionally described with some amount of flexibility even though
it retains concrete mathematical structure, which is the approach used by
Montgomery (2003).
This chapter critically assesses these spatial-location models of religious
competition using three standards.1 First, a good model concisely captures
essential features of the phenomenon studied. Second, a good model is flexible
enough to be applied to a variety of research questions. Third, a good model
generates new insight into religious competition.
When held to these standards, the spatial models do well but are not without
shortcomings. A spatial model of religious competition incorporates hetero-
geneity in religious demand while also accounting for product differentiation
by suppliers. Moreover, product differentiation can also be formally grounded
in the influential club theory of religious group production. The basic model is
also sufficiently flexible to study different religious regulatory structures, thus
enabling a theorist to use the model to make testable predictions. However,
the spatial models have not been used to their full potential. I identify how
future research should incorporate dynamic elements including the cultural
transmission of religious demand, variation in birth rates, and other adaptation
by religious suppliers.2
Differentiation in Religious Strictness

A Basic Model
Religious strictness is the degree of absolutism in theology and conformity to

behavioral codes for a religious group. Because spatial models with differen-
tiation in strictness are most common, I here provide some details about their
mathematical structure. Consider the following two-stage dynamic game. In
stage 1, each religious group g from a set of groups G chooses where to locate
in strictness space. In stage 2, each individual religious consumer i from a
set of individuals I decides with which religious group to affiliate or to not
affiliate at all.
1 Attention is restricted to research using explicit mathematical spatial models instead of verbal descriptions
of the spatial models (e.g., Blake 2014).
2 Hungerman (2010) describes other shortcomings of the literature on religious competition without
specific focus on spatial models.
3 Spatial Models of Religious Market Competition: A Critical Assessment 45
Let each religious consumer have utility function

⎧
⎨ y − a(sg − si ), if i affiliates with g, sg > si ,
ui = y − b(si − sg ), if i affiliates with g, si > sg , (3.1)
⎩
−si , if i does not affiliate.
Parameter y ≥ 0 is a fixed value of affiliation, and additional benefits of

affiliation depend on the distance between the individual’s ideal strictness
si ∈ [0, s] and the strictness of the group sg ∈ [0, s] with s > 0. Parameters
a > 0 and b > 0 capture asymmetry in the disutility when affiliating with
a group above or below the ideal. If a = b, then i’s optimal affiliation
is the group whose strictness is closest to her ideal strictness, but a > b
(a < b) implies the disutility is steeper from joining a group with higher
(lower) strictness.
A distribution of ideal strictnesses represents the religious demand in the
market and, when combined with the groups that constitute the religious
supply, provides a complete depiction of a religious market. Figure 3.1a
depicts one such market with y = 0, a = b, a uniform distribution of ideal
NA m1 m2
0 s1 s2 1
(a)
NA
m1 m2
0 s1 s2 1
(b)
NA m1 m2
0 s1 s2 1
(c)
Fig. 3.1 Affiliations by religious consumers. (a) Uniform religious demand. (b) Symmet-
ric, single-peaked religious demand. (c) Asymmetric, single-peaked religious demand
46 M. McBride
strictnesses over [0, 1] (i.e., so that s = 1) with mass n individuals at each ideal
strictness, and two groups located at s1 and s2 , respectively. In the subgame
equilibrium for stage 2, each individual affiliates with the closest option so
that all individuals in the area marked m1 have affiliated with group 1, all
individuals in area m2 affiliated with group 2, and all others close to 0 chose
to not affiliate (NA). Of course, we might expect the distribution of ideal
strictnesses to take a different shape. Stark and Finke (2000) suppose that it
might be symmetric and single-peaked as in Fig. 3.1b, but other distributions,
such as the one in Fig. 3.1c, could arise.
Now consider stage 1. Religious groups locate in strictness space in
anticipation of the affiliation decisions just described. Where the groups
locate—and, hence, the exact properties of the equilibrium—will depend on
several factors, such as how many groups are in G, the sequencing of those
groups’ entry in the market, whether consumers can choose to not affiliate with
a group, the preferences of the groups, the suppliers’ cost of locating (or entry),
and more. Indeed, there is so much variation in assumptions and equilibrium
possibilities among the prior studies that a full summary of all possibilities is
impractical here. However, a closer look at two applications of this model
demonstrates their value.
Two Applications
Barros and Garoupa (2002) provide the first formal spatial model of religious
competition. They use the Euclidean utilities defined above, assume that ideal
strictnesses are uniformly distributed with s = 1, and assume that each group
maximizes the sum of utilities of its members. They then examine settings
with one and two churches (i.e., G = {1} or G = {1, 2}) and with and
without non-affiliation options for the consumers.
If there is a monopoly group and consumers cannot choose to not affiliate,
then the monopolist’s utility is
s 1
V = (y − a (s − si )) dsi + (y − b (si − s)) dsi
0 s
1 1 1
= y − b − as 2 − bs 2 + bs.
2 2 2
Solving the first-order condition yields s1 = a+b

b
as the monopolist’s optimal
location. If a = b, the monopolist locates at the median s1 = 12 , as in
1
m1
0 s 1 = 1/2 1
(a)
1
m1
0 s 1 = 1/3 1
(b)
NA m1
0 (as -y )/(a +1) 1

(c)
NA m1
0 s' 1 1
(d)
Fig. 3.2 Barros and Garoupa analysis. (a) Monopolist, a = b. (b) Monopolist, a =
2 > b = 1. (c) Monopolist without adjustment to non-affiliation, a = 2 > b = 1. (d)
Monopolist with adjustment to non-affiliation, a = 2 > b = 1
Fig. 3.2a, giving equal weight to the disutilities of those with ideal strictnesses
above and below its location. With a = b, the group will shift its strictness
toward those who have higher disutility. Figure 3.2b illustrates this equilibrium
with a = 2 and b = 1 so that it is more costly for a consumer to join a
group above one’s ideal strictness than below it. This asymmetry induces the
monopolist to lower its strictness to s1 = 13 .
If the consumers are given the option to not affiliate but the group remains
fixed at 13 , then we have the scenario depicted in Fig. 3.2c where those with
low ideal strictness choose to not affiliate (NA). The non-affiliation option
now provides a good alternative for those with low ideal strictnesses. It can be
directly found that an individual with ideal strictness si = as−y
a+1
is indifferent
as−y
between not affiliating and affiliating, all those with si < a+1 do not affiliate,
48 M. McBride
as−y
and all with si > a+1
affiliate. The monopolist’s utility function is now
s 1
V = (y − a (s − si )) dsi + (y − b (si − s)) dsi
as−y
a+1 s
= −(b − 2y + a 2 b + as 2 + bs 2 − ay 2 − 2a 2 y + 2ab − 2bs − 4ay

−2y 2 + 2abs 2 − 2a 2 bs + 2a 2 sy + a 2 bs 2 − 4abs + 4asy)
1
× .
2 (a + 1)2
Solving the first-order condition reveals the monopolist’s optimal location

2
b−a 2 y+2ab−2ay
to be s1 = b+aa+b+a 2 b+2ab . Some additional derivation reveals s1 < s1
when y > a+b ab
and that s1 > s1 when y < a+b ab
so that the monopolist’s
optimal response to the competition from the non-affiliation option depends
on the fixed value of its good. If the fixed value is high, then the group
aggressively competes with the non-affiliation option by lowering its strictness,
as seen in Fig. 3.2d. If the fixed value is low, then the group raises its strictness
and specializes in serving those with high ideal strictness (not shown). Barros
and Garoupa consider other alternatives as well, for example, a setting where
the first group is a Stackelberg leader and the second group a Stackelberg
follower.
Barros and Garoupa’s focus on monopoly and duopoly settings fits histori-
cally regulated religious markets in Europe. They show the conditions under
which a monopolist lowers its strictness (i.e., secularizes) under pressure from
secular alternatives, and they demonstrate why a monopolist may increase its
religious strictness to compete with a new religious group entrant. They also
illuminate our understanding of other cases. They demonstrate, for example,
why a change from compulsory to voluntary membership results in a liberaliz-
ing trend in the monopoly church. Moreover, it provides a competition-based
logic to differences between the western Catholic church and its eastern
counterpart. Because the western church was a strong monopolist while
the eastern church faced competition from Islam and local Greek religious
movements, the western church could maintain its strict religious character,
while the eastern church could not.
McBride (2008) changes the basic model in two important ways. While
fixing y = 0 and a = b, he keeps a large set of groups and assumes that
each group maximizes its membership size (rather than the sum of members’
utilities) and then considers two-period endogenous entry. In the first period
of entry, each religious group decides whether to locate and at what locations to
locate. Entry entails a fixed cost c for the group, and the group has an outside
option that yields payoff 0. A group will thus only want to enter the market
and produce religious goods and services if doing so produces a large enough
membership to make entry worthwhile. Any group that does not locate in the
first period observes the groups that located and, then, in the second period,
decides whether to locate and at what location.
McBride examines an equilibrium in which the firms that locate in the first
period do so optimally in a way to prevent additional entry in the second
period. The number of entrants is endogenous and reflects how religious
markets with low barriers to entry can function like contestable markets where
both current entrants and the threat of future entry constrain the behavior
of incumbent suppliers. There are typically multiple equilibria, yet every
equilibrium shares some important properties: the groups must be sufficiently
close together to prevent an entrant from entering between a group and its
nearest competitor, but the groups must also be sufficiently far away from
their nearest competitors to make it worthwhile to remain in the market.
Figure 3.3a depicts a market equilibrium with four groups. Notice that the
groups are not necessarily equally sized, and they are close enough together
1
NA m1 m2 m3 m4
0 s1 s2 s3 s4 1
(a)
1
NA m1 m2 m3
0 s1 s2 s3 1
(b)
1.5
1
NA m1 m2 m3 m4
0 s1 s2 s3 s4 1
(c)
Fig. 3.3 McBride analysis. (a) Low entry cost. (b) High entry cost. (c) High entry cost,
population increase
50 M. McBride
to prevent new entry in between any two groups. Figure 3.3b depicts the
market equilibrium with a higher cost. Fewer groups find it worthwhile to
remain in the market, and the groups that did enter can now spread out
farther from each other. The higher entry cost thus implies a lower level of
religious pluralism. Yet, even with the higher cost, if there is growth in the
population, then pluralism can increase as groups have more consumers to
attract. Figure 3.3c depicts an increase in the population and an equilibrium
with more religious groups than in Fig. 3.3b, holding fixed the high entry cost.
This endogenous-entry setting is ideal for examining the richness of highly
competitive and pluralistic religious markets like those in the United States.
The model demonstrates how religious markets with low entry costs and
heterogeneous religious consumer preferences will manifest high levels of
religious diversity with churches spread across the strictness spectrum. We
see such diversity in the open and competitive American religious market.
However, the model can account for a wide variety of market structures by
merely changing model parameters. Religious markets with large populations
and low barriers to entry with have high pluralism and participation (large
cities in the United States), those with high entry costs (cities in Saudi Arabia)
or low populations (small towns) will have lower pluralism, and those with
inhibited secular alternatives (cities in Saudi Arabia) will have low pluralism.
Importantly, this model reveals how the interplay of both supply-side (e.g., the
entry cost) and demand-side (e.g., population size) factors determines both
the level of pluralism and the rate of participation. The model thus provides
a concise framework for explaining and interpreting cross-country patterns of
religiosity.
Assessment
The strictness spatial models have several merits. First, strictness spatial
models capture several key features of actual religious markets. They allow
for variation in religious demand and supply, two key features of actual
religious markets. They also identify the dimension of competition with
the conceptual notion of strictness, a concept understood as key to under-
standing differentiation in religious services as well as the nature of collective
production within religious groups. The club theory of religious production
(see Iannaccone 1992, 1994) has established that strictness provides a means
to confront free-riding, thus enabling stricter churches to better succeed in
providing collectively produced goods. Assuming strictness as the dimension
of competition is thus not merely for convenience but actually includes

substantive interpretational value.
As shown by McBride (2008, 2010), this link can be derived from primi-
tives. Let sg ≥ 0 denote the strictness of group g ∈ G. Individual i ∈ I has
type wi with utility function

R(sg ) + Z(wi , sg ), if affiliate with group g,
ui (wi , sg ) =
Z(wi , 0), if not affiliate.
The R(sg ) term captures the religious benefits from affiliating and participat-
ing with group g. According to the club theory, religious groups are collective
organizations that must solve free-rider problems, and they can use religious
strictness as a screening and sorting mechanism. Stricter groups impose more
extensive behavioral requirements on their members, thereby raising the cost
of entry into the group and reducing the influx of free-riders into the group.
Assuming religious benefits increase in strictness, Rs > 0, concisely captures
this religious production technology because stricter groups produce more
benefits for their members. The Z(wi , sg ) term captures the simple trade-
off between religious and non-religious pursuits. The parameter wi , with
Zw > 0, represents the individual’s secular productivity, so that a higher wi
corresponds to higher secular benefits. Higher conformity as measured by
higher sg implies less success in secular pursuits, Zs < 0. Under standard
assumptions about second-order and cross-partial derivatives, each i, given her
type wi , will have a uniquely defined ideal strictness si , which is the strictness
of a group that, should she join, gives i her highest utility over joining any other
group with any other strictness. Because each i has a unique ideal strictness,
a distribution of individuals with different types will map one-to-one into a
distribution of ideal strictnesses, thus allowing for graphical analysis as with
the simple Euclidean preferences.
This derivation from primitives also highlights an important trade-off
between secular and religious that is captured by the spatial model. A crude
interpretation is that wi is i’s market wage, sg is the time spent by i in group
g, and 1 − sg is time spent by i pursuing secular goods. Several studies have
found a negative correlation between income and time spent toward religious
pursuits, and time is naturally bounded between 0 and an upper limit. A
broader interpretation is that sg represents a more general notion of effort and
resources that an individual contributes to the group. The more resources
contributed to the group, the less in secular benefits obtained by the individual
but the larger the benefits produced by and received from the group.
52 M. McBride
A second commendable feature of the strictness spatial models is that

they are flexible enough to be applied to a variety of research questions. In
addition to two papers previously examined, strictness spatial models have
been used to study secularization (McBride 2010), the transformation of
Christianity from sect to church (Ferrero 2008), the likelihood that a country
will have a state church (Barro and McCleary 2005), and more. That such
a wide range of topics has been studied using the strictness spatial model is
a testament to the flexible nature of spatial models more generally but also
to the fundamental role of strictness in religion more specifically. Because
strictness is a fundamental way in which actual religious groups distinguish
themselves from one another, a model that captures strictness succinctly has
huge potential to study several aspects of religious competition.
A third commendable feature is that these models generate new insight into
our understanding of religious markets. In some cases, the model provides
new insight into previously known empirical patterns or historical cases.
Such is the case with Barros and Garoupa (2002) showing how response
of a monopolist to rising competition from secular competition depends
on how valuable are the religious goods, McBride (2008) demonstrating
how conflicting findings in the empirical literature on religious competition
have been misinterpreted because they did not properly account for the
level of analysis, and Ferrero (2008) interpreting theological choices made by
Roman Christianity in response to competition from paganism and Judaism.
In other cases, the model generates new predictions that are then tested
econometrically. This combination of theory and data analysis can be found
in Barro and McCleary’s (2005) linkage between the population share of a
religious group and likelihood a country has a state church, North and Gwin’s
(2014) examination of population homogeneity and the presence of a state
church, and Gaskins et al.’s (2013) investigation of how religious participation
changes with the level of economic development.
Of course, the models are not without their shortcomings. One criticism is
that the simplifying assumptions limit the realism and, hence, applicability
of the models. Yet, many simplifications provide analytical convenience
without losing substantive content. Assuming a uniform distribution of ideal
strictnesses, for example, will typically enhance clarity without reducing the
substantive merits of the analytical results. The same is likely true when
assuming that all entrants use an identical religious technology. Given the
high rate at which new religious groups fail, it is unrealistic to assume that
all religious suppliers are equally effective at overcoming the coordination,
collective action, and moral hazard problems inherent in collective production,
and differences in effectiveness will lead to entry, churn, and exit in the
religious market. Nonetheless, the model can be understood as capturing

competition between the most effective religious suppliers. That is, the model
can be understood as capturing the long-run market trajectory rather than
the short-run dynamics. Criticisms about modeling simplifications are less
compelling even if literally accurate.
Other criticisms are more compelling and demonstrate limits of the strict-
ness spatial models currently found in the literature. One criticism is that
the focus on only a single dimension misses key aspects of competition when
multiple dimensions may be equally relevant. For example, the desire to
worship with people of similar traits may be just as important for some
individuals as worshipping in a group that is closest to one’s ideal strictness.
“The most segregated hour of Christian America,” as the Reverend Martin
Luther King, Jr. once called the time when Americans are at church on Sunday
mornings, may be so segregated not just because of social and governmental
efforts to separate people by race but also by people’s desire to worship with
similar others. This possibility especially applies in cities and towns where one
religious denomination has a large presence because different congregations
of the same denomination often specialize by adopting a particular worship
style or appealing to a particular demographic. This additional product
differentiation creates inter-congregation competition even within the same
denomination. It is unclear how robust will be the predictions of a single-
dimension spatial model once other dimensions of product differentiation are
added.
Another glaring omission from the current strictness spatial models is the
disregard of various dynamic aspects of religious markets. Religious groups
can and do change their strictness levels in response to both internal and
external pressures but do so in a variety of ways (e.g., Mauss 1994; Stark
and Finke 2000). They can also merge and split apart. The strictness spatial
models account for external competition in the form of secular and religious
substitutes, but they do not account for the internal dynamics that drive
many changes in religious groups’ behavior. Neither do they account for the
dynamic formation of preferences in religious populations. Neither religious
capital formation (Iannaccone 1990) nor cultural transmission (Bisin and
Verdier 2001) nor differential fertility across groups (McBride 2015) has been
incorporated into spatial models that can be analytically solved. Agent-based
modeling provides one way to directly examine the dynamics of religious
competition when formal analysis is too complicated (Montgomery 1996;
Makowsky 2011; Chen et al. Forthcoming).
54 M. McBride
Differentiation Similar to Strictness

In three papers, the religious groups are competing for adherents as in the
strictness spatial models, and the dimension is not labeled strictness even
though the dimension is either directly or indirectly related to strictness.
Montgomery (2003) is conspicuously agnostic about labeling the dimen-
sion, though he explains that it could refer to a close relative of religious
strictness called religious tension (p. 784). His model, which includes clergy
effort as an additional factor affecting the quality of religious services, is
actually inspired by the same sociological work that inspired the first spatial
model in Barros and Garoupa’s (2002). However, Montgomery’s research
question is quite different from the other papers: he constructs and applies
a new measure of competition that avoids some of the problems of the oft-
used Herfindahl index. This derivation of a new measure from a spatial model
makes Montgomery’s paper unique in the literature on religious competition.
When he applies his measure, he finds that cities with more groups have lower
participation. This findings runs counter to economic intuition but can still be
explained by variation in religious demand across those cities (McBride 2008).
Reda (2012) refers to the single dimension in his spatial model as “religious
investment,” that is, the amount of money and time the individual is willing to
contribute to religious goods. He then examines how government subsidies
to religious groups affect the location of religious groups in this dimension;
attention is restricted to a setting with two groups. In particular, the funds
allow the group to dedicate more of its resources to proselytizing efforts,
thereby shifting the ideal points of the consumers. By deriving the ideal
points from more primitive assumptions instead of directly asserting Euclidean
preferences, Reda is able to identify how religious preferences may shift as
the result of religious regulations, a clear contribution to the literature. The
relabeling of the dimension from strictness to religious investment is less
significant. The club theory of religious production already teaches us that
high strictness generates higher contributions, so Reda’s dimension could easily
be labeled strictness without any substantive change in interpretation.
In Iyer et al. (2014), individuals differ in “religiousness,” and each of the
two groups chooses a “focal point” in the dimension of religiousness as its
location. The groups choose their focal points simultaneously to maximize
their adherents, yet unlike other models the groups also provide non-religious
services. The analysis reveals that the groups differentiate from each other
when there is competition in the provision of non-religious goods, and
that provision of non-religious services by the groups increases as economic
inequality increases. Econometric analysis using data from India supports both
predictions. The notion of religiousness is sufficiently flexible to account for
various aspects of religiosity, including strictness. In some countries like the
United States, the correlation between certain forms of religious belief and
strictness is positive and strong. As such, Iyer et al.’s model can been applied
beyond India.
Because the models just discussed place competition in dimensions similar
to strictness, the strengths and weaknesses of the strictness spatial models will
apply to these models as well. These models capture key aspects of religious
markets, are flexible enough to examine a range of questions and issues related
to religious competition, and provide new insight into our understanding of
religious competition. The models also ignore many of the dynamic aspects
of both religious supply and religious demand in real-world religious markets.
Differentiation Dissimilar to Strictness

Unlike the above three papers, Eswaran (2011) conceptualizes competition in
a novel manner that does not directly relate to strictness, even though his
motivating research question (i.e., how does religiosity depend on market
structure?) is similar to some of the previously discussed papers. Eswaran
assumes that each religious group seeks to increase the piety of its members
and does this by encouraging members to donate both time and money to
the group. Interestingly, the group can distort its message by adjusting how
members mix these two resources. Consumer preferences are identified by a
point on a circle that represents “theology or its practical implementation”
rather than strictness, with utility decreasing in the distance between the
individual’s ideal point and the point on the circle where the group located.
Eswaran identifies conditions under which a monopoly would dilute its
message to increase its profits, but he also shows that state subsidization of
religion can increase the quality of the spiritual product, while competition
can lead groups to focus on monetary donations at the expense of piety. By
separating time and money in the analysis, he takes the spatial-competition
analysis in a new direction and offers different insight to questions previously
asked.
Two other papers use the structure of a spatial model but ask quite
different questions, thus necessitating a very different conceptualization of
the dimension of competition. Raynold (2013) envisions competition in
religious risk mitigation rather than in strictness, and although the two may
be related, the distinction allows Raynold to derive new predictions. Drawing
56 M. McBride
from Iannaccone’s (1995) insight that congregational structures are designed

in part to reduce the perceived risk of specializing in one religious group,
Raynold’s dimension of product differentiation is a “fellowship ratio.” This
ratio reflects how much total time spent in religious activity is devoted to
group activities rather than private, independent activities. Raynold shows
how an individual’s optimal fellowship ratio increases in her degree of aversion
to religious risk and that belief in “one true path” and monotheism are both
positively related to fellowship ratios. By reconceptualizing the dimension of
religious competition, Raynold’s spatial model connects aspects of religious
doctrine more directly to market structure.
Berg et al. (2016) examine segmentation in the market for Islamic financial
institutions (IFIs). In all of the above papers, religious competition occurs
between religious groups typically best understood as congregations or denom-
inations, yet Berg et al. demonstrate that religious competition occurs between
other kinds of religious organizations. In their model, the groups are not
religious congregations or denominations but are instead IFIs competing for
customers who vary in how much they value an IFI’s Shariah compliance. This
variation in preference for Shariah compliance thus serves as the dimension of
product differentiation. The authors demonstrate how market segmentation
occurs with one IFI catering to high-piety types and another catering to low-
piety types.
A key novelty of Poutvaara and Wagener (2010) is that both consumers and
suppliers are ex ante differentiated. Consumers are distributed along a single
dimension that reflects their willingness to pay for services, while suppliers
are ex ante distinguished by their degrees of persuasiveness. Having both
dimensions of differentiation creates a rich setting for religious competition
and a rich set of results. Poutvaara and Wagener show that multiple equilibria
abound, some with more sects, higher welfare, and lower membership costs,
and secular societies with sects with higher membership costs. An implication
of their findings is that cross-country differences in patterns of religiosity
need not be directly mapped into fundamental differences in demand and
supply or informational asymmetries, but may instead be merely different
realized equilibria in a world where very different equilibria are possible.
This insight challenges the more widely held view in the literature that
cross-country differences in religious regulation drive national differences in
religiosity (Iannaccone and Stark 1994; McBride 2008).
By reconceptualizing the dimension of product differentiation, these papers
further reveal the general flexibility of spatial models to study religious
competition. New questions can be asked and new dimensions of religious
competition can be explored. Again, these papers confine their attention to
one-shot or static analysis, thus limiting their applicability to studying many

of the dynamics inherent in religious markets.
Conclusion
The last 15 years has witnessed a blossoming literature that uses spatial-
location models to study religious competition. My review of this literature
has identified several strengths and limitations of the literature. The models
capture several important features of actual religious markets, are sufficiently
flexible to study a variety or religious phenomena, and provide new insight
into those phenomena. A model that considers differentiation in strictness
as the dimension of analysis has the best claim to being the canonical model
of religious competition. Half of the papers reviewed (7 of 14) specifically
mentioned that the dimension could or should be understood as strictness,
and three of the remaining seven described the dimension of differentiation as
something similar to strictness. I also note that I have used a strictness spatial
model when teaching undergraduate students about religious competition for
more than a decade and have found it to be a very effective teaching tool.
Spatial models have long had a secure home in economic analysis, so it
should not be surprising that they have found a place in the economics of
religion literature. However, the value of these models going forward will
depend on researchers’ continued ability to incorporate new and novel features.
Perhaps a particularly fruitful area of future work would be to more directly
combine the spatial models with models of cultural transmission (e.g., Chen
et al. Forthcoming). The spatial models have as their particular strength
the ability to identify competitive responses by religious suppliers, but they
have largely ignored the dynamics of the demand side. Conversely, models
of cultural transmission, which are blossoming in the economics of religion
literature (e.g., Carvalho 2013; Carvalho et al. 2017), have as their strength
the characterization of how cultural values are distributed throughout society
and change as societal conditions change, although they do not provide much
insight into the supply side. A fully fledged combination of the cultural
transmission in a spatial framework with endogenous entry would draw from
the strengths of each.
58 M. McBride
References
Anderson, Gary. 1988. Mr. Smith and the Preachers: The Economics of Religion in
the Wealth of Nations. Journal of Political Economy 96: 1066–1088.
Barro, Robert, and Rachel McCleary. 2005. Which Countries Have State Religions?
Quarterly Journal of Economics 120: 1331–1370.
Barros, Pedro, and Nuno Garoupa. 2002. An Economic Theory of Church Strictness.
Economic Journal 112: 559–576.
Berg, Nathan, et al. 2016. Market Segmentation and Non-uniform Shariah Standards
in Islamic Financing. Journal of Economic Behavior and Organization 132: 39–49.
Bisin, Alberto, and Thierry Verdier. 2001. The Economics of Cultural Transmission
and the Dynamics of Preferences. Journal of Economic Theory 97: 298–319.
Blake, Jonathan. 2014. Stability in Sacred Competition? Pentecostals in the Global
Religious Marketplace. Working Paper.
Carvalho, Jean-Paul. 2013. Veiling. Quarterly Journal of Economics 128: 337–370.
Carvalho, Jean-Paul, et al. 2017. Education, Identity, and Community: Lessons from
Jewish Emancipation. Public Choice 171: 119–143.
Chen, Tongzhou, et al. Forthcoming. Dynamics of religious group growth and
survival. Journal for the Scientific Study of Religion.
Eswaran, Mukesh. 2011. Competition and Performance in the Marketplace for
Religion: A Theoretical Perspective. B.E. Journal of Economic Analysis & Policy
(Contributions) 11: 1–36. Article 14.
Ferrero, Mario. 2008. The Triumph of Christianity in the Roman Empire: An
Economic Interpretation. European Journal of Political Economy 24: 73–87.
Finke, Roger, and Rodney Stark. 1988. Religious Economies and Sacred Canopies:
Religious Mobilization in American Cities, 1906. American Journal of Sociology 53:
41–49.
———. 1992. The Churching of America, 1776–1990: Winners and Losers in Our
Religious Economy. New Brunswick: Rutgers University Press.
Gaskins, Ben, Matt Golder, and David Siegel. 2013. Religious Participation and
Economic Conservatism. American Journal of Political Science 57: 823–840.
Hotelling, Harold. 1929. Stability in Competition. Economic Journal 39: 41–57.
Hungerman, Daniel. 2010. Rethinking the Study of Religious Markets. In The Oxford
Handbook of the Economics of Religion, ed. Rachel McCleary, 257–275. New York:
Iannaccone, Laurence. 1990. Religious Practice: A Human Capital Approach.
Journal for the Scientific Study of Religion 29: 297–314.
———. 1992. Sacrifice and Stigma: Reducing Free-Riding in Cults, Communes,
and Other Collectives. Journal of Political Economy 100: 271–291.
———. 1994. Why Strict Churches Are Strong. American Journal of Sociology 99:
1180–1211.
———. 1995. Risk, Rationality, and Religious Portfolios. Economic Inquiry 33:
285–295.
———. 1998. An Introduction to the Economics of Religion. Journal of Economic
Iannaccone, Laurence, and Rodney Stark. 1994. A Supply-Side Reinterpretation
of the “Secularization” of Europe. Journal for the Scientific Study of Religion 33:
230–252.
Iyer, Sriya. 2016. The New Economics of Religion. Journal of Economic Literature
54: 395–441.
Iyer, Sriya, Chander Velu, and Melvyn Weeks. 2014. Divine Competition: Religious
Organizations and Service Provision in India. Cambridge Working Papers in
Economics 1409.
Makowsky, Michael. 2011. A Theory of Liberal Churches. Mathematical Social
Sciences 61: 41–51.
Mauss, Armand. 1994. The Angel and the Beehive: The Mormon Struggle with
Assimilation. Urbana: University of Illinois Press.
McBride, Michael. 2008. Religious Pluralism and Religious Participation: A Game
Theoretic Analysis. American Journal of Sociology 114: 77–108.
———. 2010. Religious Market Competition in a Richer World. Economica 77:
148–171.
77–87.
Montgomery, James. 1996. The Dynamics of the Religious Economy: Exit, Voice,
and Denominational Secularization. Rationality and Society 8: 81–110.
———. 2003. A Formalization and Test of the Religious Economies Model.
American Sociological Review 68: 782–809.
North, Charles, and Carl Gwin. 2014. The Political Economy of Church and State.
Working Paper.
Poutvaara, Panu, and Andreas Wagener. 2010. The Invisible Hand Plays Dice:
Multiple Equilibria in Sects Markets. Public Choice 145: 483–502.
Raynold, Prosper. 2013. Fellowship, Social Network Externalities, and Management
of Religious Risk. Rationality and Society 25: 229–260.
Reda, Ayman. 2012. Religious Charities and Government Funding. International
Advances in Economic Research 18: 331–342.
Stark, Rodney, and Roger Finke. 2000. Acts of Faith: Explaining the Human Side of
Religion. Berkeley: University of California Press.
Stark, Rodney, and William Bainbridge. 1985. The Future of Religion. Berkeley:
University of California Press.
———. 1987. A Theory of Religion. Bern: Peter Lang.
4
When Average Is Irrelevant: Computational
Modeling of Religious Groups
Michael D. Makowsky
Consider a thought experiment. Two groups are competing for members in

a small town religious market. Each of the groups, the Alphas and Omegas,
count 100 members on its respective registry, but while the Alphas can boast 90
members in its pews every Sunday, the Omegas must demure to only having
60. Which group is more successful? Which receives more in yearly tithing?
Which is more likely to still exist in 20 years?
While the Alphas would seem the immediately more likely pick, let me now
present a slightly richer depiction of their registries:
Alphas Omegas
10 (daily) zealots 30 zealots
80 (weekly) regular attendees 30 regular attendees
10 (yearly) free riders 40 free riders
As is often the case when the dimensionality of the problem is increased, the
solution is far less clear. What is the value of zealotry? What costs (or benefits)
do free riders bring to the group? Depending on the type of club good being
I thank Charles North for useful comments and suggestions.

M. D. Makowsky ()
John E. Walker Department of Economics, Clemson University, Clemson, SC, USA
e-mail: mdmakow@clemson.edu

62 M. D. Makowsky
produced, it may be only the output of zealots that matter (Iannaccone and
Berman 2006). Free riders, ostensibly the bane of all club and public goods,
may in fact be a net positive in the ledger (McBride 2015). Perhaps regular
tithing is only realized from the stable, job holding regular attendees, 20% of
whom account for 80% of the groups operating budget (Iannaccone 1997).
The heterogeneity introduced by this relatively coarse typology has radically
increased the predictive power of the model, but is also already beginning to
threaten the tractability of the problem. Just as importantly, we haven’t actually
modeled the emergence of this typology, we’ve simply assumed it—we’ve
defined agents as regular members and zealots, without assigning a meaningful,
quantifiable, metric. Modeling a population of deeply heterogeneous agents
offers the possibility of building a theory of behavior beyond the norm, theory
predicting both the magnitude of deviation and its representation in broader
society. It is the promise of accommodating some additional degree of agent
and population richness, while retaining sufficient theoretic parsimony, which
motivates the introduction of computational modeling techniques to the study
of religious groups.
In this chapter, I discuss prior use of computational modeling in the
economic study of religion, as well as opportunities to extend and elaborate
some of the core theoretical models underlying current research. Along the
way, we will revisit influential models and discuss the theoretical assumptions,
necessary for analytic tractability, that computational methods allow to be
relaxed and, in some cases, specifically targeted as explanatory mechanisms
for observable phenomenon.
Agent-Based Computational Models

At risk of disagreement over methodological taxonomy, the majority of the
contributions discussed here can fall under the rubric of agent-based com-
putational models. As in virtually any microeconomic model, a set of agents
is individually constructed with attributes and objective functions. These
agents will typically exist within a defined setting, from as simple as a linear
projection or circle to as complex as a fully realized replication of a geographic
or historical location. The set of agents and their environment will, with
rare exception, exist within a dynamic simulation defined by a progression
of time, either continuous or discretized into steps. The empirically testable
predictions of the model are typically generated via Monte Carlo simulation,
exogenously varying parameters while simulating stochastic “runs” of the
model, and tracking the individual agent and population outcomes of interest.
The nomenclature of “simulation” is appropriate here. The model remains a
4 When Average Is Irrelevant: Computational Modeling of Religious Groups 63
relatively sparse abstraction, but a run of the model has a distinct beginning
and end—the arrival of an “equilibrium” is a possible, but not necessary,
outcome. Rather, steady-state conditions, broken ergodicity, random walks,
or cyclical dynamics all exist as potential model outcomes.
My colleagues in sociology, anthropology, or evolutionary biology would
likely take a different view, but within the context of the economics of religion,
it is worth emphasizing that these computational models are not especially
different from standard microeconomic models.1 What will distinguish the
computational models discussed (and suggested) here is their abandonment
of the representative agent. The unifying, theoretically salient character of
these models is their deep agent heterogeneity, manifested through fully realized
distributions of agent attributes, endowments, strategies, social networks, and
geographic location. The ability to not just cope with deep agent heterogeneity,
but leverage it as a socially generative mechanism, has offered, and will
continue to offer, useful and exciting insights into the scientific study of
religious groups. Modeling n agents, of m ≤ n types, brings with it a new set of
issues regarding analytic tractability, particularly when those richly distributed
attributes cannot be collapsed to their statistical moments without loss of con-
tent. It is within these contexts in which agent-based computational simulation
offers a compelling alternative. Given the character of religious groups—
their association with charismatic leaders, dedicated zealots, entrepreneurial
heretics, and strictures of often incredible nuance—there can be scarcely few
popular phenomena where the average agent is less sufficient to generate
testable predictions of social outcomes.
As point of emphasis, the ambition here is not to just “weaken” the
assumptions of prior theories for the sake of elegance or generalizability.
Rather, it is to identify questions that can be newly asked or better answered
when a dimension previously assumed away (or collapsed to homogeneity)
becomes the subject of interest or explanation. In much of the discussion that
follows, we will find that a variable previously homogenous across agents—
adherence to parental beliefs, geographic location, network connectivity—
becomes the fulcrum by which the model generates new and interesting
testable predictions.
1 While computational models in the economic study of religion has been relatively sparse, there have been a
handful of models constructed within other social scientific disciplines, including sociology, anthropology,
and evolutionary biology (Bainbridge 1995, 2006; Doran 1998; Upal 2005; Chattoe 2006; Dow 2008).
In fact, cognitive models of religious groups and theology, well still early stage theoretical contributions,
are perhaps the first to computationally model beliefs, symbolism, and the structure of ritual (Whitehouse
2002; Whitehouse et al. 2012). In general, the consideration of beliefs in economic models of religion has
been notably sparse (Montgomery 1996a, b).
64 M. D. Makowsky
The Club Theory of Religion

Iannaccone’s (1992) theory of sacrifice and stigma was a watershed moment for
the economic study of religion, serving as the seminal research introducing the
“club theory” of religious groups (Iyer 2016). In classic Beckerian fashion the
club theory of religion applies not just broad economic intuition to the study of
religious groups but also the specific mechanical tools of the economic theorist,
such as utility maximization, the representative agent, and Nash equilibria.
The Iannaccone (1992) analytic equilibria hinges on a club of homogenous
agents making simultaneous decisions based on an assumption of identical
Nash equilibrium behavior. The core insight serves as an invaluable existence
proof of unproductive costs as a utility enhancing mechanism for club mem-
bers, but it comes with a nontrivial blind spot: the corner solutions of the
strictness space—requirements of zero or absolute sacrifice of private goods—
are the only viable clubs. Any club of intermediate strictness is inferior to those
permitting either all secular production or those permitting none.
This “corner solution only” outcome is a product of the (analytically
necessary) assumption of homogenous agents.2 The means by which moderate
strictness clubs might offer a utility maximizing club good for a dominant
subset of the population is demonstrated in Makowsky’s (2011) computational
cellular automata model. The viability of moderate groups is a product of
extending the club theory to accommodate agent heterogeneity, in terms of
both income and social network, in both the broader population and within
groups. Moderate sacrifice rates sufficiently mitigate member free riding by
acting as a weak screening device that filters in a range of agent income types
into the group. The resulting within-group heterogeneity in the opportunity
costs of tithing and activity participation (i.e. time) allows agents to benefit
from the differing comparative advantages of their fellow members. This
results in both an interior optimum sacrifice requirement for any one group
and range of simultaneously viable group types recruiting different member
types from the broader population.
2 The “corner solution” issue is accommodated in different manners in the club theory of religion literature.
In his formal model of church and sect, Iannaccone (1988) assumes that the returns to personal religious
conduct always have an interior maximum. Berman (2000), in point of contrast, simplifies the sacrifice
to purely screening mechanism (omitting secular good-club, good substitution effects). Subsequent
laboratory research found the sacrifice mechanism effective in a purely non-religious context, but also
found the screening mechanism dominated (nontrivial) price effects (Aimone et al. 2013).
Fundamentally Extreme
The dominance of corner solutions in the original club-theoretic model is a
more subtle concern, but no less costly, for insight into extremist groups. If,
within the assumptions of the model, all successful groups impose extreme
high sacrifice requirements, then there is no room within the model to gain
insight under which groups become relatively stricter, or in which the relatively
strictest gain market share in the religious marketplace. Computational models
of religious groups, by simulating the behavior of large-scale populations of
heterogeneous agents, can provide insight into the mechanisms generating
rich and varied religious markets. New steady-state outcomes and insights
are not limited to macroscopic phenomena—new behavioral patterns can be
observed within the groups themselves. Within each club there is an ecology of
members: a full distribution of zealots, free riders, and everyone in between.
We can observe the market share for extremism within the population, but
also within different types of groups—from which types of groups extremists
are likely to emerge and the conditions under which groups, subgroups, and
individuals become more or less committed.
Identifying religious extremism demands context. Absent reference points,
identifying someone or something as extreme is an exercise in normative,
frequently pejorative, analysis. A useful model of extremists requires modeling
everyone else who is not extreme. Makowsky (2012) and Friedman et al. (2018)
both build club-theoretic agent-based computational models of religious
groups, within which populations of richly attributed religious agents are
realized and deeply heterogeneous distributions of religiosity are generated.
While their respective models leverage different model attributes to generate
their predictions, both models are only able to gain tractability via Monte
Carlo simulation of populations of heterogeneous agents.
Both models are illustrative of the importance of deep agent heterogene-
ity. A standard dynamic stochastic general equilibrium (DSGE) model of
heterogeneous agents would collapse the dimensions of agent heterogeneity
to some finite number of statistical moments (Krusell and Smith 1998),
typically just the mean. This “aggregate approximation” works well if the mean
sufficiently captures the impact of the broader distribution on population
dynamics (Krusell and Smith 1998; Winberry 2018). The population means
(and standard deviations) of income, social networks, and familial religious
66 M. D. Makowsky
history are highly unlikely to be sufficient to capture the population dynamics

that underlie behavior and prevalence of religious extremists.3
Collapsing the Makowsky (2012) model to the moments of the income
distribution and local religiosity would preclude the conditions under which
the most extreme religious groups can demand the highest levels of member
sacrifice. Such simplifying assumptions would obscure the prediction gener-
ated by the model that extremist groups can be undermined through targeted
public good provision. Friedman et al. (2018) simulate the emergence of
fundamentalist groups in tension with the broader population. They find that
outcomes are sensitive to the influence of peer groups relative to individual
interpersonal interactions. Approximately aggregated social network distribu-
tions would likely preclude the delineation of the impact of group interaction
from individual interactions in a model of dynamically evolving peer groups.
The Church-Sect Cycle

In James Montgomery’s influential (1996) paper on the dynamics of religious
markets, he extends the club theory to incorporate both dynamic entry and
exit of groups within a religious market, concurrent with dynamic movement
of members across different types of groups. Returning to the thought exper-
iment, his model introduced agents choosing among heterogeneous groups
whose own institutions of strictness were dynamically serving the median
member agent within the group.
The model is notable for several features. First, it employs three income
classes of agents (low, medium, and high) to produce simultaneous demand
for low, medium, and high strictness religious groups. Religious capital and
denominational affiliation is passed on from parent to child. Combined
with random assignment of income class, agent denominational strictness
may be suboptimal relative to their income. This tension leads to outcomes
where populations can secularize, de-secularize, or end up in a cycle of group
formation and dissolution analogous to the “Church-Sect Cycle” theory of the
American religious market (Finke and Stark 1992).
To see how the Montgomery model and its insights might be extended fur-
ther in a computational context, we can turn to a different model. Iannaccone
and Makowsky (2007) present an agent-based model of religious geographic
3 One of the themes in Krueger’s (2007) empirical investigation of terrorism is the insufficiency of average
rates of education and income in trying to empirically predict terrorism rates and other national level
outcomes.
regionalism. Their model is not club-theoretic, in the sense of Friedman

et al. (2018) or Makowsky (2012). Rather, it instantiates m groups to which
n agents are assigned initial membership and parameterizes a simple utility
maximizing exercise by agents, each endowed with heterogeneous preferences
for (1) the size of their religious community, (2) continuing with their current
religious identity, and (3) continuing in the religious identity of their parents.
Their model demonstrates the possibility for steady-state religious regionalism,
without tipping toward a uniform population, uniform regions, or pure
mixing. It also demonstrates the counterintuitive population stabilizing effects
of deep agent heterogeneity.
Integrated within the Iannaccone-Makowsky framework, the Montgomery
model of the Church-Sect Cycle takes on entirely new context. The addition
of a spatial setting, in this case a two-dimensional lattice, would allow for
the possibility for groups to “move” across the terrain, following their target
populations, or for the types of groups to take root (e.g. the evangelical south
and the irreligious west), and wait for their would-be converts to come to
them. Perhaps both. A computational setting would also allow for the three
income class agent typology to be expanded to a full distribution of income
more representative of the United States or another population of interest. The
shape parameters, particularly those related to a growing divide between the
highest and lowest earners, could generate predictions about future sectarian
movements and their evolving bundle of proscriptions. Keeping in mind that
tithing tends to follow a classic 80-20 Pareto Rule (Iannaccone 1997), the
incentives for religious groups to retain their highest earners are likely to
only grow stronger with income inequality, with important implications for
churches trying to optimally balance their attractiveness to those most willing
to dedicate their time and energy to the group and those able to donate
financial resources.
The Value of Free Riders

Since economists first began thinking about public goods, it has begged the
question of how to solve, or at least cope with, the problem of free riders.
While the club theory of religious groups posits unproductive costs as a way
of mitigating free riding, McBride’s (2007, 2015) model extends the theory to
contexts in which free riders, within bounds, can prove a net lifetime gain for
the club good when group-specific capital is acquired through exposure over
time. Free riders are the manner in which a group continues to grow its flock
68 M. D. Makowsky
and expand its influence—rather than just a cost to be minimized, free riders
instead become an opportunity to be optimized.
McBride’s model, for the sake of analytic tractability, simplifies religious
capital to two states: low and high. This capital is also entirely group specific—
there is no prospect for moving to a lower or higher strictness group and
bringing some non-zero religious capital. Ported to a computational setting,
the model could be enriched in two interesting dimensions. First, individual
religious capital could be made a continuous variable that is dependent on
both the group(s) and individuals exposed to and the length of their exposure.
This would allow the agents to steadily grow from low capital free riders
to valued contributors and community leaders. Second, groups could adjust
their practices to increase or decrease the specificity of the capital formed to
maximize the balance of retention (increasing the cost of leaving for capital-
committed members) and recruitment (reducing the costs of switching for
prospective recruits from rival groups). The optimal specificity of religious
capital and tolerance of free riders is likely to vary greatly for new versus estab-
lished religious groups and could speak to the theories of optimal “tension”
with established churches for cults and other new religious movements (Stark
et al. 1985).
Deep agent heterogeneity, in the context of geography and social networks,
would also allow for new dimensions in which the admittance of free riders can
be club-enhancing. McBride (2007) examines the role of welcoming free riders
with low religious capital in Mormon churches affiliated with the Church of
Latter-Day Saints. Mormons, along with Jehovah’s Witnesses and many evan-
gelical churches, are known for the missionary and door-to-door recruitment.
Within most simple cost-benefit models, the success rates of these operations
would seem insufficient to justify them in any terms other than additional
unproductive costs to reduce member free riding. In a model of heavily
clustered social networks, however, with cross-cluster connections limited by
geography, cost, and technology, these largely futile missionary endeavors offer
the potential for beachheads in previously untapped religious markets. In a
population of agents heterogeneous in income, group-specific religious capital,
rates of geographic displacement, and social connectivity, optimal patterns of
missionary activity are likely to diverge greatly from simpler models. What
might otherwise look like the radically counterproductive efforts of religious
zealots might prove instead to be the uncannily sophisticated efforts of an
organization whose growth seems unconstrained by political, cultural, or
economic boundaries.
Religion as a Network Good

While the club theory of religion is a rich source of useful insight and testable
predictions, there can be little doubt the content of the “club” good has room
for useful enrichment as well. Retaining our focus on purely social dimensions,
religion is very clearly a network good. Membership provides access to a social
network, but any one agent’s place in a network will vary greatly, and will be
heterogeneous with regard to the density and importance of their connections.
Founders, leaders, and charismatic figures will hold vastly different positions
of influence and connectivity within any religious network. Further, they will
vary not only in their endowment of connections but also their ability to form
new connections and retain existing ones. The value of these attributes in
building and maintaining religious movements will depend on the attributes
of the networks they are embedded in, the distribution of members within
those networks, and the network-relevant technologies at their disposal.
Extending Rubin (2014), Kuran (1987), and Makowsky and Rubin (2013)
embed agents within a social network and spatial landscape and assess the
impact of technologically extended and enhanced social networks on the
influence of centralized (i.e. government) and decentralized (i.e. religious)
authorities on the public expression of private preferences. The television,
radio, and internet are all technologies which are likely to have had important
influences on the nature and reach of religious group identity as a network
good. As increasingly innovative religious movements seek to expand their
membership and their broader social influence via social media, new insights
from models of religious markets are likely to be had from redesigning
the religious club good as an explicit network good, defined by nodes of
heterogeneous influence, connectivity, and endowment.
The Cults Are Alright

So much of the magic of religion is lost when the possibility of outsiders,
heretics, and outsized characters exists beyond the scope of the theoretic
construct. A new religion is, by definition, a cult, a term that seems to have
carried a pejorative context since its inception. Cults are made feasible by the
unmet social and spiritual needs of individuals on the periphery of existing
social networks. When cults are successful, it often requires credit to found-
ing individuals with extraordinary personal attributes finding themselves in
extraordinary social settings. Regardless of what constitutes the representative
70 M. D. Makowsky
agent in a standard economic model, it seems unlikely to represent the

prospective members of a newly constituted religious group. Recruitment to
groups varies with background religious affiliation, education, and specific
social network conditions. Low-density networks, broader social isolation,
and weak ties to the group in question all correlate to receptiveness to cult
recruitment (Stark et al. 1985). Research on cults, as much as anything, is an
opportunity to extend the rigor and testable predictions of economic models of
religion to a subject where the barriers to formal modeling are the necessity of
incorporating network properties and deep agent heterogeneity. Agent-based
computational modeling methods offer the theorist a construct in which the
weird, the ostracized, and the maybe-a-little-bit-scary outliers are no longer
relegated to the “apologetically assumed away” footnote. Sometimes they’re
even the star of the show.
Final Thoughts
Economic studies of non-market behavior often assume that behavior is
uniformly rational and self-interested, while differences in behavior are a
product of variations in exogenously determined endowments and preferences.
Attention is seldom paid to how differences in preferences came to be.
Computational models can attempt to grow, or evolutionarily select for,
those preferences. Models of endogenous religious preferences, as products
of historic (even pre-historic) agent networks, geographic limitations, and
resource distributions, are a natural next step for the social scientific study
of religion. Religious preferences constitute too much of the story to remain
relegated to the realm of exogenously assigned parameters whose only job is
to facilitate mathematical elegance and expository simplicity. Given standard
disciplinary modeling norms, it is not surprising that some of the earliest
computational models of emergent belief and ritual structures have come
from outside economics and mathematical sociology (Whitehouse 2002;
Whitehouse et al. 2012).
As with any technique, the time and need for methodological advocacy has
quickly passed for agent-based computational modeling. The economic study
of religion and religious groups is a perfect example of a field of study where
computational models can explicitly build from the existing body of theory,
opening up models along dimensions previously intractable, and contribute
new avenues of theoretical insight and testable predictions.
References
Aimone, J.A., L.R. Iannaccone, M.D. Makowsky, and J. Rubin. 2013. Endogenous
Group Formation and Unproductive Costs. Review of Economic Studies 80 (4):
1215–1236.
Bainbridge, W.S. 1995. Neural Network Models of Religious Belief. Sociological
Perspectives 38 (4): 483–495.
———. 2006. God from the Machine: Artificial Intelligence Models of Religious
Cognition. Rowman Altamira.
Berman, E. 2000. Sect, Subsidy, and Sacrifice: An Economist’s View of Ultra-
Orthodox Jews. Quarterly Journal of Economics 65 (3): 905–953.
Chattoe, E. 2006. Using Simulation to Develop Testable Functionalist Explanations:
A Case Study of Church Survival. The British Journal of Sociology 57 (3): 379–397.
Doran, J. 1998. Simulating Collective Misbelief. Journal of Artificial Societies and
Social Simulation 1 (1): 1–3.
Dow, J. 2008. Is Religion an Evolutionary Adaptation? Journal of Artificial Societies
and Social Simulation 11 (2): 2.
Finke, R., and R. Stark. 1992. The Churching of America, 1776–1990: Winners and
Losers in Our Religious Economy. New Brunswick, NJ: Rutgers University Press.
Friedman, D., J. Fan, J. Gair, S. Iyer, B. Redlicki, and C. Velu. 2018. How Funda-
mentalism Takes Root: A Simulation Study. Working Paper No. 1681, Faculty of
Economics, University of Cambridge.
Iannaccone, L.R. 1988. A Formal Model of Church and Sect. American Journal
of Sociology 94 (Supplement: Organizations and Institutions: Sociological and
Economic Approaches to the Analysis of Social Structure): S241–S268.
Other Collectives. Journal of Political Economy 100 (2): 271–291.
———. 1997. Skewness Explained: A Rational Choice Model of Religious Giving.
Journal for the Scientific Study of Religion 36 (2): 141–157.
Iannaccone, L., and E. Berman. 2006. Religious Extremism: The Good, the Bad, and
the Deadly. Public Choice 128 (1–2): 109–129.
Iannaccone, L., and M. Makowsky. 2007. Accidental Atheists? Agent-Based Explana-
tions for the Persistence of Religious Regionalism. Journal for the Scientific Study of
Religion 46 (1): 1–19.
Iyer, S. 2016. The New Economics of Religion. Journal of Economic Literature 54 (2):
395–441.
Krueger, A.B. 2007. What Makes a Terrorist: Economics and the Roots of Terrorism.
Princeton University Press.
Krusell, P., and Anthony A. Smith Jr. 1998. Income and Wealth Heterogeneity in the
Macroeconomy. Journal of Political Economy 106 (5): 867–896.
Kuran, T. 1987. Preference Falsification, Policy Continuity and Collective Conser-
vatism. Economic Journal 97: 642–665.
72 M. D. Makowsky
Makowsky, M.D. 2011. A Theory of Liberal Churches. Mathematical Social Science 61

(1): 74–87.
———. 2012. Emergent Extremism in a Multi-Agent Model of Religious Clubs.
Economic Inquiry 50 (2): 327–347.
Makowsky, M.D., and J. Rubin. 2013. An Agent-Based Model of Centralized Insti-
tutions, Social Network Technology, and Revolution. PLoS One 8 (11): e80380.
McBride, M. 2007. Club Mormon: Free-Riders, Monitoring, and Exclusion in the
LDS Church. Rationality and Society 19 (4): 395–424.
77–87.
Montgomery, J.D. 1996a. Contemplations on the Economic Approach to Religious
Behavior. The American Economic Review 86 (2): 443–447.
———. 1996b. The Dynamics of the Religious Economy: Exit, Voice, and Denom-
inational Secularization. Rationality and Society 8 (1): 81–110.
Rubin, J. 2014. Centralized Institutions and Cascades. Journal of Comparative Eco-
nomics 42 (2): 340–357.
Stark, R., W.S. Bainbridge, and The Seymour Martin Lipset Collection. 1985. The
Future of Religion: Secularization, Revival, and Cult Formation. Berkeley: University
of California Press.
Upal, M.A. 2005. Simulating the Emergence of New Religious Movements. Journal
of Artificial Societies and Social Simulation 8 (1).
Whitehouse, H. 2002. Modes of Religiosity: Towards a Cognitive Explanation of the
Sociopolitical Dynamics of Religion. Method & Theory in the Study of Religion 14
(3): 293–315.
Whitehouse, H., K. Kahn, M.E. Hochberg, and J.J. Bryson. 2012. The Role for Sim-
ulations in Theory Construction for the Social Sciences: Case Studies Concerning
Divergent Modes of Religiosity. Religion, Brain & Behavior 2 (3): 182–201.
Winberry, T. 2018. A Toolbox for Solving and Estimating Heterogeneous Agent
Macro Models. Quantitative Economics. Forthcoming.
5
The Intelligent Design of Religious Beliefs
Gilat Levy
Introduction
There is no doubt that one of the first words that comes to mind when
one mentions religion is “belief ”. Religion is viewed as a system of beliefs
and religious individuals are often called “believers”. Moreover, religious
organizations heavily invest in creating and maintaining beliefs. The Catholic
Church’s Index Librorum Prohibitorum shows that the church would like its
believers to avoid getting information that may change their beliefs. Church-
approved miracles is another way in which beliefs in God may increase.
Religious rituals, sermons, and mass are obvious ways to maintain believers’
devotion and strong beliefs.
The empirical literature in economics has also shown the importance of
religious beliefs, as, for example, in Barro and McCleary (2003), Huber
(2005), and Guiso et al. (2003, 2006). Guiso et al. (2003, 2006) show that
The research presented in this chapter is based mainly on Levy and Razin (2012, 2014a,b). This research
was supported by the ESRC (grant number RES-000-22-1856) and the ERC (grant number 210385).
G. Levy ()
London School of Economics (LSE), London, UK
e-mail: g.levy1@lse.ac.uk

74 G. Levy
religious beliefs are associated with more trust and better economic attitudes
and that these effects are larger for Protestants than for Catholics.1 Barro and
McCleary (2003) show that beliefs in heaven and hell are positively associated
with higher growth, while participation in rituals is associated with lower
growth.
There are several explanations or justifications of religious beliefs in the
anthropological literature, as well as in psychology or evolutionary biology.
For example, some claim that religious beliefs help to understand the world
and thus reduce anxiety or allow to cope with difficult situations as God will
help or that it is all God’s will.2 In this short survey, I will further the view that
beliefs are instrumental in the sense that they affect the behavior of individuals
in the social sphere and their daily life.
Specifically, one important feature of religious beliefs is that they induce
good behavior in social interactions. Here I will mainly follow Levy and
Razin (2012) in considering a system of beliefs that is based on reward and
punishment for good and bad behavior, respectively.3 I will argue that such a
system is flexible enough to also relate to theologies of predestination. I will
show how this system of beliefs relates to rituals, trust, and the behavior of
religious individuals toward one another and toward non-affiliated or non-
religious. Finally, I will illustrate how an intelligent design of beliefs can make
them immune to changes and facilitate the survival of religion.
By enabling good behavior, religious beliefs and religious organizations may
also induce individuals to signal their ethics and religiosity to others. Adam
Smith observed that religions tend to produce and distribute moral informa-
tion about their members which allows traders to assess the risk involved in
conducting business with them.4 Weber (1906) writes of the social pressure
in American Protestant communities, “Unqualified integrity, evidenced by,
for example, a system of fixed prices in retail trade…appears as the specific,
indeed, really the only, form by which one can demonstrate his qualification
as a Christian and therewith his moral legitimation for membership in the
1 La Porta et al. (1997) show that countries with hierarchical religions perform comparatively worse on
a wide range of outcomes, which accords with Putnam (1993) who suggests that such religions deter
formation of trust.
2 See the survey in Boyer (2001).
3 There are other ways to model religious beliefs. Benabou and Tirole (2006) assume that agents differ
in their beliefs with respect to how much hard work is rewarded, in this or in the afterlife, and actively
choose to maintain such beliefs. In Scheve and Stasavage (2006), on the other hand, religious beliefs allow
for a psychic benefit in bad times, and hence such beliefs negatively correlate with preferences for social
insurance.
4 See Anderson (1988).
sect…admittance into the Baptist congregation was primarily of decisive

importance…because of the on-going inquiries about moral and business
conduct.”
The above implies a relation between three important aspects of religion:
beliefs, behavior in the social sphere, and religious rituals or sacrifices. Reli-
gious rituals and sacrifices are typically public and, as in Iannaccone (1992),
allow individuals to signal how devout they are to each other.5 Rituals, as well
as previous behavior in the social sphere, can then signal to others how strong
are one’s beliefs in reward and punishment. Strong beliefs can indicate to others
that an individual will be cooperative in the social sphere and thus to facilitate
cooperation, enabling societies to achieve better economic outcomes.
Indeed, there is by now a large literature focusing on the link between
religious beliefs and religious practices (Huber 2005) and how such beliefs and
practices affect various parameters ranging from micro data such as individuals’
well-being and individual behavior (Gruber 2003; Sosis and Ruffle 2003) to
macro parameters such as growth (Barro and McCleary 2003; Guiso et al.
2003) and the public provision of social insurance (Scheve and Stasavage
2006). Thus, the starting point of this chapter is that such features of religious
organizations—beliefs as well as public rituals—enable individuals to increase
cooperation in society.
Systems of Religious Beliefs

Many daily interactions demand trust; we imagine here such a situation
commonly known as a prisoner’s dilemma (PD). In such environments it
would be best for all to cooperate; this would yield the highest social surplus.
However, it is also the case that when others cooperate, players are tempted to
defect or “free ride” and enjoy the utility provided by others. For example, if
individuals need to provide a public good by contributing time and money,
they would rather others do so while they enjoy the good provided. This
typically yields situations in which all individuals shy away from investing and
providing the good. In other words, all “defect”, and society ends up with a
lower surplus overall.
Religious beliefs however typically place high value on pro-social behavior
which is what we consider here as “cooperate”. One way to think about this
is to assume that an individual believes that if she cooperates she is rewarded
5 See also Chwe (2003).

76 G. Levy
with a benefit from a supernatural entity such as God. This can be perceived
as a spiritual benefit of better luck in the afterlife or an actual benefit that may
accrue in this life. Individuals who think that this benefit is high will always
cooperate even if they think the other players defect. Individuals who think
that this benefit is moderate will cooperate if the others cooperate too, while
individuals who think that the benefit is low (weak believers) will then always
defect as in the basic case without any added “spiritual” benefit.
This implies that the religion will potentially enable cooperation in society
as we will see below. Before we do so, we discuss two theologies that can induce
such incentives to cooperate in social interactions, namely, belief in reward
and punishment, and Calvin’s predestination. These two systems of beliefs are
explored, respectively, in Levy and Razin (2012, 2014b).
Beliefs in Reward and Punishment
Examples of systems of religious beliefs that connect rewards and punishments

by a supernatural entity to behavior in the social sphere are abundant in the
anthropological literature. Beliefs in rewards and punishments, whether in this
life or the afterlife, are rife in many ancient and modern religions and create
an incentive to properly behave in a social context. One famous example is
the Nuer religion in Sudan, about which Evans-Pritchard (1956) writes: “
…and in any argument about conduct the issue is always whether a person
has conformed to the accepted norms of social life…the Nuer are of one voice
in saying that sooner or later good will follow right conduct and ill will follow
wrong conduct.” Reward and punishment are prevalent in Abrahamic religions
as well as in some Eastern religions. According to the Wycliffe Dictionary of
Theology, “it is plain from the Bible that sin will be punished.” In Hinduism,
a basic concept is that of Karma—the accumulated sum of one’s good or bad
deeds. More generally, in Buddhism and the Hindu traditions, the state in
which one is reincarnated may depend on one’s ethical behavior in the present
life.
To incorporate this within an economic model, as in Levy and Razin (2012),
we can assume that individuals are exposed to shocks throughout their lives,
positive and negative. Non-believers think that the probability of a shock
does not depend on their own actions. Thus, when they observe a natural
disaster or an economic boom, they just think that these arise disregarding
their actions. However, believers think that good social behavior (cooperation
in social interactions, contributions of time and effort in the public sphere) is
more likely to induce a positive shock, in this or in the afterlife. Thus, when
they observe an earthquake, they think this was a punishment from God due
to deviant behavior, while success in life will be attributed to God as a reward
for good behavior. I specify this formally in the technical appendix.
A related system is explored in Benabou and Tirole (2006). In their work
individuals have an anticipatory utility for an afterlife shock which, according
to their belief, is likely to be higher when they work harder.6 They focus on how
such beliefs interact with the political economy of taxation and redistribution.
Predestination
Even Calvinistic beliefs which emphasize predestination, as Weber (1904)

first recognized, may constitute an incentive for good works in the social
sphere; this arises because an individual wishes to glean information about
whether she will be saved and doing good deeds provides a positive signal
about the individual’s future. On this Weber (1904) writes: “The question:
Am I one of the elect? must sooner or later have arisen for every believer.”
Those belonging to the elect embody Jesus within themselves and hence are
the epitome of good behavior. Calvin’s followers may wish to know whether
they have been identified by God to be part of the elect. Individuals will
be encouraged then to “test” themselves and examine their behavior. A self-
signaling mechanism can induce individuals to cooperate (motivated by the
behavioral prescriptions put forward through the stories on Christ in the
scriptures) in order to assure themselves that they belong to the elect. In that
sense, the reward and punishment system can be perceived as positive and
negative anticipatory utility from belonging to the elect. Calvin has therefore
made good works and moral behavior the center of religious life by creating a
sense of anxiety over individuals’ salvation.
Again, this can be incorporated in an economic model (which I specify in
the technical appendix), as shown in Levy and Razin (2014b). Individual’s
anticipation of belonging to the elect provides some spiritual utility. We
assume that each individual believes that he belongs to the elect with some
probability, where individuals can differ in this belief or their “conviction.”
The heterogeneity of beliefs in the population represents different reactions to
exposure to theological material, such as reading the Bible or the publications
of Calvin.
6 Such a system of beliefs can be interpreted both as the simple reward and punishments and as the
“Protestant work ethic” belief, tying hard work for benefits in the afterlife.
78 G. Levy
In Calvin’s theology, sanctification implies that individuals who are in the

elect behave as Jesus Christ would. Accordingly, in our model, to hold and
maintain their beliefs individuals look back to their previous social behavior
to infer the likelihood of belonging to the elect. We assume that individuals
believe that those who belong to the elect always cooperate, whereas all other
individuals cooperate only with some probability. Self-signaling then amounts
to an individual revisiting her previous actions and noting whether she had
cooperated or not. If an individual cooperated in the first period, then her
updated beliefs that she belongs to the elect increase. On the other hand, if she
defects, she will believe that she does not belong to the elect as these individuals
never defect. It is then easy to show that if individuals have an anticipatory
utility from believing that they belong to the elect, then they are induced to
cooperate in the PD game.
Note that while Calvin advocates justification by the grace of God, Luther
focuses on justification by faith. Luther offered individuals personal certitude
of salvation already in this life, provided only that they have faith. These
beliefs reduce the anxiety about salvation but also imply that Luther’s theology
reduces the motivation for good works (McGrath 1990).
How Rituals and Beliefs Combine to Enhance

Cooperation
Above we had considered systems of beliefs that provide incentives for indi-
viduals to cooperate in social interactions. It is easy to show that in societies
in which individuals are simply endowed with these beliefs, then in such
interactions the more devout will cooperate and the less devout will defect.
Thus, the beliefs themselves can increase the level of trust in society.
However, religion is not only about beliefs and typically such beliefs
arise through organized teaching and preaching. It is therefore important to
consider the joint role of beliefs and institutions. We can show that this can lead
to better outcomes in society.
Specifically, consider the institution of public rituals. Rituals such as praying
together, or wearing specific clothes, are observable. Moreover, such rituals are
costly. Praying takes time from other activities, while wearing special clothes
can be a burden vis-à-vis the outside society. However, as in Iannaccone (1992),
partaking in such costly activities can imply to others that the individual is
sufficiently religious or, in our model, that she strongly believes that she is
rewarded for cooperative behavior.
The argument for this is as follows. An individual who believes she will
be strongly rewarded if cooperative is more likely to cooperate. But also
this implies that it is more important for her that her partner in the social
interaction cooperates as well. Thus an individual would like to signal to others
that she has these strong beliefs; this shows others that she is more likely to
cooperate and induces them to cooperate too. Public and observable costly
rituals do not only instill beliefs, but can then serve as ways to signal to others
strong beliefs.
We can then show how religions can arise as a stable social organization,
when individuals choose optimally whether to participate in rituals, and how
to behave in the social interaction given the affiliation decisions of their
partners. While rituals are costly, individuals may be rewarded with more
cooperative behavior from others, and this trade-off allows to shape the
composition and size of religions.
Specifically, if we call the set of individuals who participate in rituals as
“religious”, while those that choose not to participate as “non-religious”, we
can provide a taxonomy of the type of religions that arise in society for
different costs of rituals. Specifically, religions with very costly rituals attract
only the very devout. These will be small and composed of individuals that
will cooperate vis-à-vis all—the religious as well as the non-religious. Religions
with less costly rituals will also attract the less devout. As a result, they would
be composed of some individuals that will cooperate vis-à-vis members of
the same religion but may defect against those outside their religion. We
would therefore observe more in-group cooperation. Moreover, when more
than one sect exists, members of the stricter sect may be less cooperative
toward others from a less strict sect. Rituals can then signal the level of the
expected cooperation of its members among themselves and toward others.
The taxonomy of religious organizations is such that religious membership
decreases in the cost of rituals, while internal and external cooperation both
increase when sacrifice is higher.
The benefit from such a combined system of rituals and beliefs is then
twofold. It allows individuals to increase their cooperation level, which
increases their anticipatory or spiritual utility, as well as their material welfare,
as more individuals cooperate with them. Still, costly rituals will reduce this
material utility. It is therefore aligned with the empirical findings of Barro and
McCleary (2003) who show that economic growth responds positively to the
extent of religious beliefs, notably beliefs in hell and in heaven, but negatively
to church attendance. The model shows that beliefs are indeed conducive for
good economic outcomes and that costly and wasteful rituals are the main
determinant behind the lower growth of religious communities.
80 G. Levy
We have discussed above how public rituals and a system of beliefs can
induce higher levels of cooperation in society. We have also explained how
cooperation is more likely to be in-group cooperation for moderate level of
rituals. We now proceed to discuss how religious beliefs are designed so that
such systems are more likely to survive in the long term.
Intelligent Design of Religious Beliefs

While many religions stress reward and punishment in the afterlife, many
stress reward and punishment in this life. Moreover, there is plenty of evidence
that following natural and other disasters, individuals and religious leaders
often re-evaluate, or feel a need to justify, their religious stances (e.g., Jewish
theology after the holocaust). In turn, this re-evaluation affects individuals’
decisions of whether to stay/become religious. For example, Chen (2010)
suggests that following the financial crisis in Indonesia, religious participation
had increased.
We can incorporate in our model the possibility that individuals update
their beliefs following the shocks they experience. Individuals may not be
sure about the exact probability of receiving a reward when they behave
cooperatively; this means that when they do so and experience a positive shock
to their utility, they may think that this probability is higher than initially
perceived (again, exact technical details are in the appendix). Similarly, if a
cooperative individual experiences a negative shock such as a natural disaster,
she may think that the probability of a reward for cooperation is lower.
Thus, in response to a shock, a short-term effect will be that some indi-
viduals will switch their affiliation; some will leave their faith while some will
“find God”. Individuals that leave the faith are more likely to be those with
relatively weak beliefs, which might have defected. Thus leaving the faith is
more likely to arise when there is a positive shock such as an economic boom,
while a negative shock such as a natural disaster or a financial crisis can also
induce individuals to “find God”, in line with Chen (2010) findings.
If individuals keep on updating and learning about the relation between
their action and the system of reward and punishment, they may find out
over time that their beliefs are wrong. How can religions survive in the long
term if they are based on wrong beliefs? The title of this chapter refers to how
religions manage to maintain and sustain their system of beliefs even in such
environments.
One obvious way of creating an irrefutable system of beliefs is to base them
only on afterlife events. Anthropologists such as Rappaport (1999) and Sosis
and Alcorta (2003) have already discussed this unverifiability of beliefs. Still,
even religions that stress the afterlife have believers questioning their faith after
major events such as wars or natural disasters. Moreover, there are other ways
to maintain beliefs even when those are based on this-life experiences. We now
consider features of religions that allow them to maintain strong beliefs in the
face of events which seem to contradict them.
Atonement, Repentance, and Forgiveness
The concept of forgiveness, with its twin concepts of atonement and repen-
tance, is an important concept in many religions. According to the model
above, forgiveness blurs the relation between one’s action and utility shocks.
Intuitively, if one sinned and received a positive utility shock, she can believe
she was forgiven rather than update her beliefs.
Assume that individuals know that if they are forgiven, then the probability
of receiving a positive or a negative shock is the same, disregarding their action.
In addition, assume that a person believes she is potentially forgiven (although
she is not sure about that). If not forgiven, as above, one believes that the
probability of getting a positive shock conditional upon cooperating is higher
than conditional on defecting.
There are two implications to the above. The first one, as desired, is a
(de)learning effect: Learning becomes slower and beliefs are not updated as
quickly as in the case in which there is no forgiveness. This is because a person
cannot attribute the fact that she was not punished to her actions as maybe she
was forgiven. This implies that individuals do not learn that potentially sins
are not costly and so may maintain their religious beliefs.
However, there is a second important effect. The concept of forgiveness (as
well as indulgences sold by the Catholic Church in medieval times) will reduce
the incentives of individuals to cooperate. Intuitively, knowing that you will
be forgiven implies a lower incentive to cooperate. This implies that for any
cost of rituals, the number of religious adherents will be smaller. Thus, while
forgiveness induces a long-term gain in terms of slower learning, it induces a
short-term cost as it leads to smaller religions.
There are other ways to blur the lines between actions and rewards. One
notable example is the biblical story of Job: Even a righteous person can be
punished, so one should not try and understand God’s ways. Another way to
blur the lines is to have a complicated system in which reward and punishment
also depend on others’ behavior, the effects of which we discuss below.
82 G. Levy
Religious Intolerance due to Pressures to Survive
In the model described above, the relation of one to God was individualistic.
However, individuals may believe that they have no personal relation with
God and that the relation between actions and rewards is the same or at least
similar for all individuals, or that punishments are collective to some degree.
This implies that learning about this relation will be even faster as they can
look at others’ actions and rewards.
This leads to two conclusions. First, in general, it would be better for
religions to foster the idea of personal relation to God, so learning will be
slower.
Second, it provides a rationalization why religious individuals would like
to impose religious codes of behavior on the rest of society. This is indeed
what we observe: In many countries, religious individuals and organizations
advocate for religious prohibitions (marriage laws, divorce, sexual behavior,
restrictions on alcohol, dressing, and food consumption) to be maintained
by law. There are abundant examples of massive demonstrations against the
broadening of some of these personal liberties. In some cases violence is being
used toward those that exercise such personal liberties, as, for example, in
what is now defined as “anti-abortion terrorism”.7 Esteban et al. (2016a,b)
analyze individual labor choices as well as political choices over legal levels of
personal liberties when religious individuals suffer negative externalities when
such liberties are practiced in society.
The model above rationalizes such behavior of religious organizations. First,
religious individuals may fear collective punishment. Second, if non-religious
“sin”, and are not necessarily be punished for their sins, religious individuals
may learn from this that behaving in this way carries no punishment. Thus
the basis for the theological system of rewards and punishments will weaken.
Others “sinning” creates negative information externalities for the religious. If
no one practices such sinful behavior, learning will be curtailed.
7 According to statistics gathered by the National Abortion Federation (NAF), an organization of abortion
providers since 1977 in the United States and Canada, there have been 17 attempted murders, 383 death
threats, 153 incidents of assault or battery, 100 butyric acid stink bomb attacks, 373 physical invasions, 41
bombings, 655 anthrax threats, and 3 kidnappings committed against abortion providers.
Miracles
Each year more than 6 million pilgrims visit the Marian shrine at the town of
Lourdes, renowned for its perceived miracle cures. However, only 69 miracles
were approved by the Catholic Church to have occurred in the place since
1858. A miracle needs to be approved by a committee which comprises 30
medical specialists. Clearly, the church goes through a lot in order to gain
credibility. Such credibility will allow believers to increase their beliefs that
God exists.
In our model above, we think of the incentive to cooperate as induced by
the belief that some supernatural entity rewards and punishes individuals on
the basis of their actions. But of course individuals will be more devout if they
think that there is a higher likelihood that such entity exists. Miracles allow
individuals to increase such a belief therefore if perceived as credible. This
provides a rationale for the elaborate medical committee approving miracles
at Lourdes.
Doctrine Adaptation
Benabou et al. (2016) assume that religious beliefs imply a taste for religious
goods. They analyze the relation between religious beliefs, religious organi-
zations, and scientific innovations. Their key assumption is that scientific
innovation might potentially contribute to the decline of beliefs, and to avert
that, religious organizations can attempt a shift of doctrine. A new religious
doctrine will attempt to explain why new scientific innovation is compatible
with the old religious beliefs, as theories of intelligent design attempt to
make evolution and creationism compatible. This is another way for religious
organizations to ensure survival of religious beliefs.
Motivated Beliefs
In Benabou and Tirole (2006) mentioned above, individuals also learn about
the nature of rewards and benefits, as in Levy and Razin (2012), but can choose
to suppress information. In some environments individuals would prefer to
suppress information so as to keep themselves motivated to work and believe
that they will get future rewards for their hard work. Thus, such beliefs can
sustain adverse information if individuals choose to ignore them.
84 G. Levy
Conclusion: Joint Design of Beliefs

and Institutions
The survey above shows how religious beliefs induce cooperation, together
with other features of religion such as costly observable rituals. To focus
on this, I have simplified the belief systems to focus on common feature,
the motivation to behave well in social interactions. But of course there
are many differences in theologies; they might affect individuals’ beliefs
through different psychological channels and might not be easily compared.
For example, the self-signaling interpretation by Weber (1904) of Calvinistic
beliefs might induce incentives for good works, but such a mechanism is rather
indirect. Also, Luther and Calvin encouraged their supporters to go back to
the scriptures and to read the Bible by themselves (enabled by the advent of
printing and higher levels of literacy). Such a practice may lead to weaker or
stronger beliefs than when one participates in rituals conducted by priests,
depending both on the individuals and the priests in question. On the other
hand, mechanisms such as forgiveness and indulgences (the system of exchange
between money and redemption) that have evolved in the Catholic Church
might erode the connection between beliefs and good works to some degree as
I have shown. It would be interesting to analyze these more nuanced systems
of beliefs.
Let me note though that the differences between these two theologies might
be consistent with the differences in the institutional structure (although the
causality between institutions and theology is not obvious). Specifically, in
the Catholic Church, good works alone do not suffice; according to Thomas
Aquinas, three are required for salvation: direct reliance on the church and
its sacraments, the free turning of the will to God and away from sin, and
the remission of the guilt incurred by sin by priestly absolution.8 In medieval
times, this had evolved into a heavy load of public rituals and an impressive
system of rent extraction. More generally, this theology easily lands itself to
a hierarchical structure in which priests have to certify which actions provide
rewards and can possibly deliver forgiveness. In the absence of free will, as in
Calvin’s theology, such a role of the Church’s hierarchy, which is reinforced by
rituals, is reduced.
Given the above, future research should focus on the relation between
systems of beliefs and the religious organizations and how these maximize the
utility and benefits of both religious leaders and religious affiliates.
8 Thomas Aquinas, summa theologica. p. 39.

Technical Appendix
Beliefs: Suppose that an individual i believes that when she cooperates she
receives a benefit γ i . This implies that the PD game can be written as follows,
for a general γ :
C D
C d + γ,d + γ c + γ,b
D b, c + γ a, a
whereas assume as usual that b > d, a > c, and moreover that the game
exhibits strategic complementarities so that d − b > c − a.
In the reward and punishment system, we can further refine γ i to be as
follows. Let then pix be the probability of receiving a positive utility shock
when taking action x ∈ {c, d}. Then we can think of the reward for
cooperation defined above, γ i , as:
γ i = pic − pid .
In the predestination system explained in the text, we can think of the

incentive to cooperate as follows. We assume that each individual i believes
that he belongs to the elect with some probability, initially α i ∈ [0, 1].
Belonging to the elect provides some spiritual utility ε. We assume that the
individual’s conviction is private information. Let the initial types in the
population be distributed uniformly on [0, 1]. We assume that individuals
believe that those who belong to the elect always cooperate, whereas all other
individuals cooperate only with some probability, κ ∈ [0, 1].9 If an individual
with beliefs α cooperated in the first period, her posterior beliefs are that she
αi
belongs to the elect with probability αi +(1−α i )κ
. On the other hand, if she
defects she will believe that she does not belong to the elect with certainty.
It is then easy to show that if individuals have an anticipatory utility from
believing that they belong to the elect, then they are induced to cooperate in
the PD game.
9 The value of κ is determined in equilibrium.

86 G. Levy
Cooperation: Suppose that individuals are matched randomly in pairs to

play the PD game. We then have:
Lemma 1: There exists a value γ ∗ such that all above some γ ∗ cooperate while
all below γ ∗ defect.
Assume now that individuals can participate in rituals which entail a cost r
and that these rituals are observable. We then have10 :
Proposition 1: Let r ≤ r̄. (i) If r > r , then only those with the highest values
of γ join the religion and they cooperate against all. (ii) If r ∈ {r , r }, then also
those with moderate values of γ join the religion and these moderate individuals
cooperate with fellow religious individuals and defect against non-religious. The
non-religious defect against all.
Learning: Let pix be the probability of receiving a positive utility shock

when taking action x ∈ {c, d}. Then we can think of the reward for
cooperation defined above, γ i , as follows:
γ i = E(pic ) − E(pid ),
where pic taken from fic (.), pic taken from fid (.). This allows us to model
learning: If an individual cooperated (defected) and received a positive utility
shock (+), we have:

pf x (p)
E (p|+) =
x
p x dp > pf x (p)dp = E x (p)
p f (p )dp
Similarly, if an individual cooperated (defected) and received a negative utility

shock (−), she will weaken (strengthen) her religious beliefs, so that E x (p) >
E x (p|−).
We can now model forgiveness. Assume that individuals know that if they
are forgiven, then the probability of receiving a positive or a negative shock is
the same, at p̄, disregarding their action. In addition, assume that a person
believes she is forgiven with probability φ. If not forgiven, as above, one
believes that the probability of getting a positive shock upon cooperating
(defecting) is pic (pid ) and this probability is taken from fic (fid ).
10 We assume in Levy and Razin (2012) that participation in rituals also instills beliefs. See the proof there.
To see that learning is slower, note now that when a person defects, then
following a shock, updating is:

(1 − φ)pf d (p)
Eφd (p|+) = p dp < E d (p|+)
(1 − φ) p f d (p ) + φ p̄

(1 − φ)(1 − p)f d (p)
Eφd (p|−) = p dp > E d (p|−)
(1 − φ) (1 − p )f d (p ) + φ(1 − p̄)
On the other hand, individuals will now have a lower γ i as:
γ i = (1 − φ i )(E(pic ) − E(pid ))
This implies that for any cost of rituals, religions will be smaller.
Finally to see how to model miracles, simply let
γ i = ζ i (pic − pid )
Miracles allow individuals to increase ζ i if perceived credible.
References
Anderson, G. 1988. Mr. Smith and the Preachers: The Economics of Religion in the
Wealth of Nations. The Journal of Political Economy 96(5): 1066–1088.
Barro, R., and R. McCleary. 2003. Religion and Economic Growth across Countries.
American Sociological Review 68(5): 760–781.
Benabou, R., D. Ticchi, and A. Vindigni. 2016. Forbidden Fruits: The Political
Economy of Science, Religion, and Growth, mimeo.
Benabou, R., and J. Tirole. 2006. Belief in a Just World and Redistributive Politics.
Quarterly Journal of Economics 121(2): 699–746.
Boyer, P. 2001. Religion Explained. New York, NY: Basic Books.
Chen, D. 2010. Club Goods and Group Identity: Evidence from Islamic Resurgence
During the Indonesian Financial Crisis. Journal of Political Economy 118(2): 300–
354.
Chwe, M.S. 2003. Rational Ritual: Culture, Coordination, and Common Knowledge.
Princeton: Princeton University Press.
Esteban, J., G. Levy, and L. Mayoral. 2016a. Personal Liberties, Religiosity and Effort,
mimeo.
———. 2016b. The Political Economy of Religion, Individual Liberties, and Redistribu-
tion, mimeo.
88 G. Levy
Gruber, J. 2003. Religious Market Structure, Religious Participation, and Outcomes: Is

Religion Good for You?, mimeo.
Guiso, L., P. Sapienza, and L. Zingales. 2003. People’s Opium? Religion and Eco-
nomic Attitudes. Journal of Monetary Economics 50: 225–282.
———. 2006. Does Culture Affect Economic Outcomes? Journal of Economic
Perspectives 20: 23–48.
Huber, J.D. 2005. Religious Belief, Religious Participation, and Social Policy Attitudes
Across Countries, mimeo.
Iannaccone, L. 1992. Sacrifice and Stigma: Reducing Free-Riding in Cults, Com-
munes and Other Collectives. Journal of Political Economy 100(2): 271–291.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R.W. 1997. Trust in Large
Organizations. American Economic Review 87(2): 333–338.
Levy, G., and R. Razin. 2012. Religious Beliefs, Religious Participation and Cooper-
ation. American Economic Journal: Microeconomics 4(3): 121–151.
———. 2014a. Rituals or Good Works: Social Signalling in Religious Organizations.
Journal of European Economic Association 12(5): 1317–1360.
———. 2014b. Calvin’s Reformation in Geneva: Self and Social Signalling. Journal
of Public Economic Theory 16(5): 730–742.
McGrath, A. 1990. A life of John Calvin. Oxford: Oxford University Press.
Putnam, R. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Prince-
ton, NJ: Princeton University Press.
Rappaport, R. 1999. Ritual and Religion in the Making of Humanity. Cambridge:
Cambridge University Press.
Scheve, K., and D. Stasavage. 2006. Religion and Preferences for Social Insurance.
Quarterly Journal of Political Science 1: 255–286.
Sosis, R., and B.J. Ruffle. 2003. Religious Ritual and Cooperation: Testing for a
Relationship on Israeli Religious and Secular Kibbutzim. Current Anthropology
44(5): 713–722.
Sosis, R., and C. Alcorta. 2003. Signaling, Solidarity, and the Sacred: The Evolution
of Religious Behavior. Evolutionary Anthropology 12(6): 264–274.
Weber, M. 1904. The Protestant Ethic and the Spirit of Capitalism. New York: Scribners
(translated by T. Parsons, 1930).
———. 1906. “Churches” and “Sects” in Northern America: An Ecclesiastical
Sociopolitical Sketch”, Trans. Colin Loader. Sociological Theory 3: 7–13 (1986
translation).
6
Religion and Segregation
Ronny Razin
In recent years economists have started taking an interest in the increasing

levels of segregation of Western societies. For example, in the US, since 1970
residential segregation has been on the rise.1 Moreover, this trend in residential
choices seems to be correlated with important economic and political vari-
ables. For example, Reardon and Bischoff (2011) study the relation between
income inequality and segregation according to income in the US. They
find that residential segregation and income inequality have been following
a remarkably similar trend from 1970 to 2000. Chetty et al. (2014) look
at the relation between segregation in the US and social mobility. They
show large gaps between different localities, so that the more segregated areas
have much lower social mobility. Alesina and Ferrera (2005) survey the large
literature that studies the relation between segregation and economic outcomes
such as growth. A large theoretical and empirical literature has analysed the
effects of ethnic diversity on different measures of social cohesion, conflict,
and social attitudes (see Field et al. 2008 and Uslaner 2012). The effects of
segregation on political outcomes are also explored, ever since Bishop and
Cushing (2008) coined the term “The Big Sort”, describing the patterns of
1A similar trend is found in the UK. For example, see https://www.opendemocracy.net/wfd/ted-cantle-

and-eric-kaufmann/is-segregation-on-increase-in-uk.
R. Razin ()
London School of Economics, London, UK
e-mail: r.razin@lse.ac.uk

90 R. Razin
residential segregation in the US and its effects on polarisation of political

beliefs.
Throughout history there seems to be a tight connection between religion
and residential segregation. In ancient times, cities were divided into quarters,
often based on religious affiliation.2 Smith (2010) surveys the archaeological
literature on neighbourhoods and districts in ancient cities. He shows that
cities with a plurality of religions typically exhibit spatial segregation by
religion such as Jewish, Muslim, and Christian groups in Near Eastern cities
or Jewish communities within European cities. Stark (1996) follows the spread
of Christianity by showing how first adapters of the new religion came from
communities that were physically connected to the Jewish diaspora at the time.
Closer to our times, Berman (2000) discusses the emergence of the ultra-
Orthodox sect of Judaism in Eastern Europe at the time of the emancipation
of the Jews in Europe. He shows that the ultra-Orthodox, in contrast to other
Jews, intentionally eschewed opportunities available at the time to integrate
with the rest of society.
But even today, religion plays an important role in the patterns of residential
location we observe. Berman (2000) and Razin (2017) document how the
ultra-Orthodox Jews in Israel segregate away from the secular Jewish pop-
ulation both physically and also through their lack of participation in the
labour market and the compulsive military service. In current-day London,
Brimicombe (2007) finds that “The landscape of religion is found to be more
highly segregated in contrast to the landscape of ethnicity”.
The nature of religious segregation might be different to other types of
segregation, such as segregation according to income. Some religions have
specific rules or prescriptions that are directly related to the ability or to the
induced preferences of the religious to live side by side with non-believers.
One example is dietary restrictions (such as the rules of keeping Kosher or
Halal) that limit the ability of members of the religion to socially interact with
individuals who do not share these restrictions. Other restrictions might be
imposed on other social activities that involve non-believers, such as restric-
tions on participation in different cultural events, restrictions on drinking
alcohol, or restrictions on driving a car on certain days. These features that
are special to religious organisations might imply more than just segregation
in the broad sense, but living in neighbourhoods that exclude others. This
tendency for exclusive religious neighbourhoods shows up when estimating
2 See Arnon (1992) for the history of residential segregation in Jerusalem during the Ottoman Empire.
religion-based segregation. For example, Field et al. (2008) find that more
than 70 % of the population in Ahmedabad in 2002 lived in completely
homogeneous neighbourhoods.
There is also evidence of persistence in religious segregation over time.3
In Belfast, where there has been a bitter conflict between Protestants and
Catholics for many years, neighbourhoods are still divided, years after peace
has been restored (see Lloyd and Shuttleworth 2012). In the US, the “religious
landscape” has been remarkably stable over the years. Given the high mobility
rates of the US labour and residential markets, this provides a challenge for
researchers who have predicted the decline of religious convergence in the past
(see Iannaccone and Makowsky 2007).
In this chapter we survey economists’ current understanding of religious
organisations with an eye to understand the relation between religion and
segregation. We start by pointing out that the economic literature that
explicitly looks at the connection between religion and segregation is still in its
infancy.4 Nevertheless, we show that many of the existing economics models
of religious organisation implicitly contain insights about this relation. We
highlight these insights one by one to suggest new avenues for future research
on this important topic.
The chapter is organised as follows. We go over different models of religious
organisation, teasing out the implications for religious segregation. We start
with models of religious organisations as pure club goods, models in which
religious beliefs are not explicitly modelled. We then discuss different models
of religious beliefs and discuss how these models lead to new insights about
religious segregation. Finally, we discuss how an important aspect of religious
organisations is to create incentives and to use different means to help members
signal their conviction to one another. In particular, we distinguish between
religions that emphasise ritualistic intensity as a signal of religiosity to ones in
which good social behaviour is the means by which conviction is signalled to
others. As we show, each of these mechanisms becomes more effective with
segregation.
3 An insightful recent New Yorker article, “Where the Small-Town American Dream Lives On”, tells the
story of a small town, Orange City, Iowa, that has a strong sense of community which it owes to its heritage
of a group of Dutch protestant immigrants who settled there in the end of the nineteenth century. See
https://www.newyorker.com/magazine/2017/11/13/where-the-small-town-american-dream-lives-on.
4 This is true for both theoretical and empirical work.
92 R. Razin
Religion as a Club Good

The canonical model of religion in economics, presented in Iannaccone
(1992), is the club good model of religion. This theory of religion takes the
enjoyment of religious goods as a given and focuses on the organisation of
a group of individuals, who enjoy taking religious activities together, into a
club. Iannaccone (1992) gives an example of the type of activities that the
religious members enjoy and the complimentarities in doing these together:
“…the pleasure and edification I derive from Sunday services depends not
just on my own inputs but also on the inputs of others: how many others
attend, how warmly they greet me, how well they sing, how enthusiastically
they read and pray, how deep their commitment, and so forth”. The club
nature of a religion allows the group members to enjoy these public goods.
In order to increase the quality of their religious activities, the religious might
want to impose restrictions on other, non-religious, activities and perhaps
exclude believers with weak conviction that may not contribute as others to
the religious activities.
Segregation is not modelled explicitly in the club good theory of religion,
but the theory provides some implications for the members’ residential choice.
The club good theory builds on the notion that individuals enjoy meeting
together physically and practising religious actions. The emphasis of the theory
is on religious activities that take place on some regular basis, “…the collective
side of religion is most evident at congregation activities: sermons, scripture
studies, testimony meetings, liturgies, worship services, hymn singing, com-
munion, and the like”. Given that every now and then the members meet
and practise religion together, the club good model of religion implies some
constraints on how far they live from the place they meet.
Berman (2000) builds on Iannaccone (1992) to explain the emergence of
the ultra-Orthodox sect of East European Jews in the late nineteenth century.
To do this, Berman (2000) adds to the club good framework additional social
benefits such as mutual insurance and charity activities. Berman documents
that
The degree of mutual insurance practiced within these communities …is

believed to be unprecedented in Jewish history. No sick member is without
visitors and no single member is without an arranged match. For example, …tens
of in-kind free loans advertised in a flyer written by neighbourhood rabbis in the
Bayit Vegan neighbourhood of Jerusalem, ranging from Torah scrolls to wedding
gowns to playpens. The same flyers list free services available, including visits to
the sick, logistic support and advice for mourners, and frozen meals for the sick,
the elderly and for mothers after childbirth.
As in Iannaccone (1992) while not explicitly modelled, the theory implies

that to benefit fully from the club good, individuals must live closely together.
Note that these two theories of religion imply that to take advantage of
the benefits of being a religious member, the religious should live close
together. These theories, however, do not necessarily imply that neighbour-
hoods in which the religious choose to live should be exclusive to their
group.
In the club good model of religion, religious beliefs are assumed to be behind
the desire of members to partake in religious activities. So in this model, the
religious beliefs determine what the activities are that members enjoy together.
Once the set of activities is determined, the economic model of club goods
provides the theory of organisation of the religion. In the next section we
discuss models that explicitly model religious beliefs and how they shape the
ways in which religions organise themselves.
The Nature and Role of Religious Beliefs

It is hard to think of religion without thinking of religious beliefs and their
role. A recent line of research attempts to explicitly model how different
theologies have implications to the way religious groups are organised. In
particular here we discuss Levy and Razin (2012, 2014a,b) who introduce
religious beliefs into the model of religion. The religious beliefs under this
methodology make connections between social behaviour and supernatural
outcomes. Levy and Razin (2012) model a theology of reward and punishment
in relation to behaviour in the social sphere. In particular, the theology of the
religion makes a connection between the actions of a deity and the behaviour
that individuals take in their day-to-day social interactions. Holding these
beliefs allows the religious to sustain cooperative outcomes that may not be
available otherwise. One example is to view social interactions as a “prisoner’s
dilemma” game. Without religious beliefs, individuals might converge to play
the non-cooperative Nash equilibrium. But holding religious beliefs “changes
the game” to a coordination game in which they can achieve the cooperative
outcome.
Levy and Razin (2014a,b) model religious beliefs according to the protestant
theology. Following Luther’s theology, under this belief system there is no
direct relation between social behaviour and reward and punishment. The
belief system is instead centred around the existence of a subgroup of society,
94 R. Razin
the elect, who will be saved.5 Luther offered individuals personal certitude
of salvation already in this life, provided only that they have faith. These
beliefs reduce the anxiety about salvation but also imply that Luther’s theology
reduces the motivation for good works (McGrath 1990).
Calvin, in contrast, increases the anxiety of believers by advocating justifi-
cation by the grace of God. This implies that individuals do not know if they
are saved or not. As Weber (1904) describes, “The question: Am I one of the
elect? must sooner or later have arisen for every believer and have forced all
other interests into the background. And how can I be sure of this state of
grace?”.
The answer to the question of the individual’s salvation lies in another
innovation in Calvin’s theology: “It is faith in Christ which makes him live
in me and move in me and act in me…faith receives Christ’s good works; love
performs good works for the neighbours” (cited in Green 1964). Calvin has
therefore made good works and moral behaviour the centre of religious life by
creating a sense of anxiety over individuals’ salvation. The only thing known
about the elect is that those in the elect “have Jesus within them” and so should
expect to see the types of behaviours that Jesus has been known to have. Levy
and Razin (2014a,b) show that this theology motivates people to behave in
pro-social ways in order to relieve their anxieties about whether they belong
to the elect or not. By observing their own behaviour and seeing good works,
they are able to sustain their belief that they might be chosen.
These theologies of religion will have several implications to the choices of
where the religious chose to live. Some of these implications, as above, come
from the particular activities that religious beliefs are related to. For example,
to enjoy the increased trust and cooperation between fellow members, religious
individuals have an incentive to live close to each other and interact socially.
The activities that religious beliefs usually relate to are everyday social inter-
actions. This implies that to recoup the benefits of being religious one would
want to live close to and to interact daily with fellow religious members. In
addition, if there are non-believers in their social circle, the religious might
worry about being taken advantage of. This gives religious individuals an extra
incentive to actively shy away from out-of-group social interactions.
The explicit modelling of the religious beliefs has further implications for a
more exclusive segregation. As Levy and Razin (2012, 2014b) show, the beliefs
of the religious are not static, they evolve given the personal experiences of
5 “For all man are not created in equal condition rather eternal life is foreordained for some, eternal
damnation for others” (Calvin, Institutes, III. xxi 5).
the believer. Religious beliefs must be maintained and protected if they are
to sustain in the long run. To sustain religious beliefs individuals should be
guarded from observing behaviours and outcomes that do not agree with their
belief system. The trust that exists between the religious implies that they
behave cooperatively. On the other hand, when interacting with out-of-group
members, mistrust can imply that they themselves partake in asocial behaviour.
This makes sustaining their beliefs harder as they might not experience any of
the expected punishments they believe might follow after asocial actions.
Moreover, in Levy and Razin (2012), by observing the fortunes of others, the
religious can learn about the general connection between social behaviour and
reward and punishment. By segregating in closed communities the religious
can sustain their beliefs avoiding the learning involved from their own bad
actions (that they reduce substantially) and by not observing the fortunes of
the asocial. Some religious beliefs also involve collective punishments which
provide extra incentives to make sure there are no non-believers in your
community.
An additional aspect of religious beliefs is that they are often affected by the
beliefs of those that surround us. This aspect of beliefs has several implications
about the reciprocal relation between segregation and religion. Below we
devote a separate section to discuss this socialisation effect and its relation to
segregation.
The above discussion shows how the nature of religious segregation could
yield different outcomes to those of other types of segregation. These insights
can inform the large literature that focuses on the relation between diversity
and conflict. This literature has often looked at this relation without distin-
guishing the source of diversity or segregation. If we are blind to the exact
nature of segregation, the literature has provided conflicting evidence, some
showing that diversity relaxes tensions and improves trust while others showing
the opposite. Above we have illustrated a dual effect of religion; it creates
segregated communities for the goal of increasing trust and cooperation.
Field et al. (2008) investigate religious segregation and violence in Ahmed-
abad, where “residential arrangements in this city are remarkably segregated:
By 2002, 71 percent of the population in our sample lived in completely
homogeneous neighbourhoods, a fact that presumably reflects increasing
intolerance for living with members of another religion”. Indeed, Field et al.
(2008) show that in the 2002 riots in the city, incidents of violence were
more likely to occur in integrated neighbourhoods rather than homogenous
ones.
96 R. Razin
The Importance of Signalling Religiosity in How

Religions Are Organised
The above discussion highlighted the importance that members of the religion
put on distinguishing between true believers and non-believers. If members of
a religious organisation engage in some form of signalling, this will enable
them to both tailor their behaviour and location decisions towards other
fellow religious members. The above models of religious organisation give
rise to two different forms of signalling that we now discuss. The first
model highlights the importance of rituals as signals or religiosity. The
second model focuses on social behaviour itself as the way to signal your
conviction.
Rituals as Signals of Religiosity Iannaccone (1992) and later Berman (2000)

first suggested that religious rituals and the intensity with which they are
practised can be used as signals of religious conviction. The intuition is that
of “costly signalling”, that is, those who believe are those who have the highest
incentive to engage in the costly ritual to gain trust from others. Many religious
rituals are indeed costly and so might serve as a costly and useful way to convey
your religious conviction.
Given the need to recognise who is a believer and who is not, segregation
can be viewed as a substitute to the number and intensity of rituals. If
religious members are fully segregated, they are then able to maintain their
social interaction just amongst themselves. The need to further signal their
religiosity is therefore diminished. For this logic to work, non-believers should
be somehow dissuaded from moving in as well. This can be achieved in
different ways such as by locating in remote and somewhat undesirable
locations (this is the case for many monasteries such as the Monastery of
Ostrog in Montenegro and Metéora Monastery in Greece). Another way
is to use legal or non-legal means to make the non-believers unwelcome
by imposing local restrictions. One example of that is how many ultra-
Orthodox neighbourhoods in Israel and in Western countries close their roads
on Saturdays.
Signalling Your Religiosity by Behaviour Weber (1906) who studied the

protestant communities in the US writes of the social pressures members
experience, “Unqualified integrity, evidenced by, for example, a system of fixed
prices in retail trade…appears as the specific, indeed, really the only, form by
which one can demonstrate his qualification as a Christian and therewith his
moral legitimation for membership in the sect…admittance into the Baptist
congregation was primarily of decisive importance…because of the on-going

inquiries about moral and business conduct”.
In the context of Calvin’s version of the protestant theology, Levy and
Razin (2014a,b) show how a new form of signalling arises. In particular, the
protestant religion has drastically reduced the number of nature of rituals. But
as we have seen above, predestination and Calvin’s emphasis on the grace of
God imply that individuals use good behaviour as a way to ease their anxiety
about belonging to the elect.
Calvin’s innovation was not just suggesting a different theology, he also insti-
tuted innovative institutional features for the organisation of the church. In his
Reformation in Geneva, Calvin shifted the church’s emphasis to discipline. In
his second spell in the city, Calvin initiated the institution of the consistory to
monitor, discipline, and publicise individuals’ behaviour. In the Ecclesiastical
Ordinances (1541),6 Calvin drew up the structure of his well-ordered church
in Geneva. The most distinctive and controversial aspect of this organisation
was the consistory. It was formed in 1542, “ their office is to have oversight
of the life of everyone…There were to be twelve of them, chosen from the
members of the three councils, to keep an eye on everybody”.
To be able to do so, the consistory members met once a week and visited
the homes of each Genevan family twice a year, which allowed them to be
well informed. Deviant behaviour—religious and civil alike—was punished
by public scolding, sometimes by Calvin himself. When other communities
in Switzerland and France decided to adopt Calvin’s religion, he insisted
on the formation of local consistories, which may be better at collecting
local information. A great deal of its function was devoted to resolving civil
disputes within families, neighbours, and business associates. In fact, Arrunada
(2010) shows that to this day, Protestants better monitor each other’s conduct
compared with Catholics.7
Such Protestant communities use good behaviour, not rituals, as a way to
judge the conviction of fellow members. This particular trait of such commu-
nities has important implications to segregation. Together with incentives to
congregate to take advantage of good behaviour, to better monitor others, one
has to have a close relationship with fellow worshippers. This implies that such
communities will have stronger incentives to segregate. Blanchard (2007) uses
county-level data from the 2000 Census of Population and Housing and the
6 SeeGilbert (1998).
7 McCleary (2007) also shows that Protestants tend to trust or place obligations on others as they do with
family members.
98 R. Razin
2000 Churches and Church Congregations data in the US to analyse religious

segregation. He finds that counties with a large number of conservative
Protestant congregations exhibit higher levels of residential segregation along
multiple dimensions.
The Role of Socialisation

So far we have explored how different aspects of religious organisation can
lead to the geographical concentration of religious groups. But there is also
the opposite effect to consider; how does segregation impact religious practice
in general and religious beliefs in particular? Recent research in economics
has shown that the patterns of our social interactions are not just affected
by our beliefs and attitudes but are also impacting the way we see the
world. For example, Dustmann and Preston (2001) show how segregation
in neighbourhoods affects attitudes towards minorities. Algan et al. (2015)
find that students’ political opinions converge with those of their friends, and
Boisjoly et al. (2006) and La Ferrara et al. (2014) find that living in mixed-race
housing lowers the number of prejudiced beliefs among students.
What are the implications of this socialisation channel on religious beliefs?
Iannaccone and Makowsky (2007) introduce socialisation to their empirical
approach to study the “paradox of persistent mobility and persistent regional-
ism” we discussed above. The challenge is to explain the stability of the regional
religiosity in the US together with high levels of social mobility. Their “Multi-
Agent Religion Simulation (MARS) traces the magic of persistent regionalism
back to the combined impact of personal experience and social conformity.
If newcomers routinely adapt to their new social environment, then a region
can remain ‘ blue’ despite a large influx of ‘ reds.’ As newcomers change their
colours, the West acquires ‘accidental’ atheists and the South gains ‘accidental’
enthusiasts”.
New methodologies of information diffusion in social networks and new
models of persuasion allow us to do tap in to the reciprocal relation between
beliefs and segregation. Two recent examples are Levy and Razin (2017a,b),
who introduce and use a new heuristic of socialisation to explicitly model the
co-evolution of segregation and beliefs. Levy and Razin (2017a) show this in
a model that studies the relation between segregation in schooling and labour
market discrimination. They show how segregation and polarisation can be
sustained in the long run, due to the process of socialisation. Levy and Razin
(2017b) study the reaction of a prejudiced society to immigration waves. The
model highlights the importance of socialisation that together with segregation

implies long-run polarisation of prejudice in societies.
The heuristic of socialisation is called Bayesian Peer Influence (BPI). In
this model of socialisation, beliefs are modelled as probability distributions
over some unknown parameters that describe the world. When individuals
socialise, they exchange their beliefs with each other. They then update beliefs
in a semi-rational way. That is: (1) they ignore selection biases in their informa-
tion and only update on the basis of what they hear, as in the network literature;
(2) they believe that the information sources that led to this information were
conditionally independent. With these two assumptions, when an individual
with beliefs q meets N others with beliefs q = (q1 , . . . , qn ), the BPI heuristic
implies that his beliefs change to:

q qi
i∈N
q BP I
(q, q) =
q qi + (1 − q) (1 − qi )
i∈N i∈N
The BPI has some simple useful characteristics. First, individuals with
more conviction (stronger beliefs) are more persuasive. Second, the change of
belief is monotone; individuals’ beliefs increase in their peers’ beliefs. Finally,
socialisation exhibits polarisation; a group of like-minded individuals will have
more conviction after socialising.
Further research is needed to explore the above mechanisms in relation to
religion. As we tried to illustrate in this chapter, segregation is an important
aspect of the study of religion. There is an intimate and reciprocal relationship
between the evolution of religions and the patterns of residential and other
types of segregation in society. New methodologies, both theoretical and
empirical, imply that we can now have a new generation of studies to push
the frontier and shed more light on the dynamic evolution of religious beliefs
and organisations.
References
Alesina, A., and E. La Ferrara. 2005. Ethnic Diversity and Economic Performance.
Journal of Economic Literature 43(3): 762–800.
Algan, Yann, Quoc-Anh Do, Nicolò Dalvit, Alexis Le Chapelain, and Yves Zenou.
2015. How Social Networks Shape Our Beliefs: A Natural Experiment Among
Future French Politicians. Unpublished Manuscript.
100 R. Razin
Arnon, Adar. 1992. The Quarters of Jerusalem in the Ottoman Period. Middle Eastern
Studies 28(1): 1–65.
Arrunada, B. 2010. Protestants and Catholics: Similar Work Ethic, Different Social
Ethic. Economic Journal 120(547): 890–918.
Berman, E. 2000. Sect, Subsidy and Sacrifice: An Economist’s View of Ultra-
Orthodox Jews. Quarterly Journal of Economics 115(3): 905–953.
Bishop, B., and R.G. Cushing. 2008. The Big Sort: Why the Clustering of Like-minded
America is Tearing Us Apart. Boston, MA: Houghton Mifflin Harcourt.
Blanchard, T.C. 2007. Conservative Protestant Congregations and Racial Residential
Segregation: Evaluating the Closed Community Thesis in Metropolitan and
Nonmetropolitan Counties. American Sociological Review 72(3): 416–433.
Boisjoly, Johanne, Greg Duncan, Michael Kremer, Dan Levy, and Jaque Eccles. 2006.
Empathy or Antipathy? The Impact of Diversity. American Economic Review 96:
1890–1905.
Brimicombe, A. J. 2007. Ethnicity, Religion, and Residential Segregation in London:
Evidence from a Computational Typology of Minority Communities. Environment
and Planning B: Planning and Design 34(5): 884–904.
Chetty, R., N. Hendren, P. Kline, and E. Saez. 2014. Where is the Land of Opportunity?
The Geography of Intergenerational Mobility in the United States (No. w19843).
National Bureau of Economic Research.
Dustmann, C., and I. Preston. 2001. Attitudes to Ethnic Minorities, Ethnic Context
and Location Decisions. Economic Journal 111: 353–373.
Field, E., M. Levinson, R. Pande, and S. Visaria. 2008. Segregation, Rent Control, and
Riots: The Economics of Religious Conflict in an Indian City. American Economic
Review: Papers & Proceedings 98(2): 505–510.
Gilbert, W. 1998. Renaissance and Reformation, Electronic edition. Lawrence, KS:
Kansas University.
Green, V.H.H. 1964. Luther and the Reformation. New York: Putnam.
Iannaccone, Laurence R. 1992. Sacrifice and Stigma: Reducing Free-Riding in Cults,
Communes, and Other Collectives. Journal of Political Economy 100(2): 271–291.
Iannaccone, L.R., and M.D. Makowsky. 2007. Accidental Atheists? Agent-Based
Explanations for the Persistence of Religious Regionalism. Journal for the Scientific
Study of Religion 46(1): 1–16.
La Ferrara, E., J. Burns, and L. Corno. 2014. Interaction, Prejudice and Performance:
Evidence from Randomly Assigned Peers in South Africa. Presentation in the CEPR
Public Economics Conference, Warwick University, December 2014.
Levy, G., and R. Razin. 2012. Religious Beliefs, Religious Participation and Cooper-
ation. American Economic Journal: Microeconomics 4(3): 121–151.
———. 2014a. Calvin’s Reformation in Geneva: Self and Social Signalling. Journal of
Public Economic Theory 16(5): 730–742.
———. 2014b. Rituals or Good Works: Social Signalling in Religious Organizations
(with Gilat Levy). Journal of European Economic Association 12(5): 1317–1360.
———. (2017a). The Coevolution of Segregation, Polarised Beliefs and Discrim-

ination: The Case of Private vs. State Education. American Economic Journal:
Microeconomics 9(4): 141–170.
———. 2017b. Immigration into Prejudiced Societies: Segregation and Echo Chambers
Effects, working paper.
Lloyd, C.D., I. Shuttleworth. 2012. Residential Segregation in Northern Ireland in
2001: Assessing the Value of Exploring Spatial Variations. SAGE Journals 44(1):
52–67.
McCleary, R.M. 2007. Salvation, Damnation and Economic Incentives. Journal of
Contemporary Religion 22(1): 49–74.
McGrath, A. 1990. A life of John Calvin. Oxford: Oxford University Press.
Razin, A. 2017. Israel and the World Economy: The Power of Globalization. Cambridge,
MA: MIT Press.
Reardon, S.F., and Bischoff, K. 2011. More Unequal and More Separate: Growth in
the Residential Segregation of Families by Income, 1970–2009. Report prepared
for the Russell Sage US2010 Project. www.s4.brown.edu/us2010/Data/Report/
report111111.pdf .
Smith, M.E. 2010. The Archaeological Study of Neighborhoods and Districts in
Ancient Cities. Journal of Anthropological Archaeology 29(2): 137–154.
Stark, R. 1996. The Rise of Christianity How the Obscure Marginal Jesus Movement
Became the Dominant Religious Force in the Western World in a Few Centuries.
Princeton, NJ: Princeton University Press.
Uslaner, E. 2012. Segregation and Mistrust: Diversity, Isolation, and Social Cohesion.
Cambridge: Cambridge University Press.
Weber, M. 1904. The Protestant Ethic and the Spirit of Capitalism. New York: Scribners
(Translated by T. Parsons, 1930).
———. 1906. “Churches” and “ Sects” in Northern America: An Ecclesiastical
Sociopolitical Sketch”, trans. Colin Loader. Sociological Theory 3: 7–13 (1986
translation).
7
Religious and Cultural Leaders
Anja Prummer
JEL Classification J15, Z10, D02
Introduction
It is widely assumed that religious and cultural leaders have an impact on
the attitudes and behaviors of their followers and on their community as
a whole. An instance where religious leaders were systematically employed
to adjust people’s attitude is Turkey under Mustafa Kemal Atatürk. First,
Atatürk used Imams to mobilize the Turkish people in the Turkish War of
Independence. Later, he created an institution that employed Imams, which
ensured that a moderate Islam was taught, legitimizing the objectives of the
state, often against the wishes of the Imams themselves. In this tradition,
the Turkish government today employs Imams in other countries, such as
Germany, securing its influence in Turkish enclaves. In an excellent case study,
Ceylan (2010) goes as far as to argue that the convictions of the Turkish
Imams determine the attitude of Turkish migrants to Germany, as well as their
integration into German society.
A. Prummer ()
Queen Mary University of London, London, UK
e-mail: a.prummer@qmul.ac.uk

104 A. Prummer
The impact of cultural leaders on attitudes and values has also been
documented through several case studies by Portes (1987) and Portes and
Sensenbrenner (1993). They highlight the role of community leaders in ethnic
enclaves. Community leaders can be religious leaders, such as priests, rabbis,
and imams, but also foreign language media, which are described as cultural
leaders.1
In this chapter, we survey the growing theoretical literature on religious and
cultural leaders that is motivated by these case studies. We first discuss the
possible motives of the leaders that have been brought forward, before turning
to the impact these leaders have. Existing work focuses on the effect of leaders
on cultural diversity, with an emphasis on whether cultural assimilation occurs
to either a mainstream society or another religious or culturally distinct group.
Remarkably, leaders ensure distinct cultural traits and prevent assimilation,
irrespective of their exact motives. Next, we focus on the interplay between
two leaders, where the leaders can either belong to the same group or different
ones, before turning to an analysis of the environment under which religious
and cultural leaders emerge. The emergence of leaders has only received
limited attention so far, and therefore, we conclude with a discussion of open
questions.
Leader Objectives
As the notion of a leader, who has an impact on the values and norms,
the religious beliefs, cultural traits, and, more generally, the identities of his
followers, has not yet become a common concept in Economics, it seems
natural to start with a discussion of a leader’s objective.
First, leaders may be interested in increasing the number of their followers.
A payoff function that captures this motive has been suggested by Hauk and
Mueller (2015) and Verdier and Zenou (2015, 2018). In their setting, leaders
aim for increasing the number of religious followers, to maximize the spread of
their cultural trait in society. Each agent in the population is characterized by
a trait, implying that each agent is either religious or not. In order to convince
agents to become religious, leaders exert a socialization effort. This effort can
1 For
more details on how foreign language media influence norms and value, see Subervi-Velez (1986),
Zhou and Cai (2002).
take the form of providing a public good. In this case, the leader’s payoff is
specified as
π L = R(q) − C(G), (7.1)
where q denotes the share of individuals of the religious trait in the popu-
lation.2 R(q) is an increasing, positive function of the share of agents with
a religious trait. The leader provides a public good G, which comes at cost
C(G), but induces individuals to join the community. Put differently, a higher
provision of the public good leads to more followers. The goal is to then find
the optimal level of the public good, depending on the costs and benefits
derived from a larger community. This formulation therefore emphasizes the
missionary zeal religious leaders may display.
An alternative specification is motivated by early literature which thought
of leaders as managers of a church, which was equated with a firm (Ian-
naccone 1998). While religious leaders certainly provide their congregation
with religious goods and services, seeing a leader as purely business oriented
seems to fall short of a realistic description of the workings of a religious
community. Prummer and Siedlarek (2017) incorporate both features of
religious leaders, namely, their zeal in spreading a mission, in creating a belief,
while simultaneously having entrepreneurial traits, in the leader’s objective
function.
They argue that leaders influence the beliefs and attitudes of all the members
i in the community, denoted by pi ∈ [0, p̄], with i ∈ {1, 2, . . . , n},
by establishing a set of rules and norms. Religious leaders can propagate
these rules and their beliefs through sermons and in interactions with their
community. Note that the cultural trait here is not binary, but a continuous
value, which implies that religious leaders care about the extent to which
the religious trait is displayed, or rather the leader defines the religious trait.
Moreover, they also care to some extent about the material well-being, wi ∈
[0, w̄], of each community member. While religious leaders can influence the
attitudes within their community, they cannot affect the economic well-being
directly. However, an individual’s economic situation depends also on which
norms he follows, as these can affect economic outcomes both positively and
2 The payoff provided here departs slightly from all the payoffs specified in Hauk and Mueller (2015) and
Verdier and Zenou (2015, 2018). Hauk and Mueller (2015) suggest a payoff function that maximizes the
share of individuals with the same trait and an alternative which aims for a maximization of conversions the
leader is part of. Verdier and Zenou (2015) consider a time-dependent version, whereas Verdier and Zenou
(2018) develop a continuous-time model. For simplicity, we consider here the simplest, most general
specification.
106 A. Prummer
negatively. To investigate this in more detail, we state the religious leader’s

payoff:

n
πL = π (pi , wi ) (7.2)
i=1
The payoff of a religious leader is increasing in both religious beliefs and

economic well-being, implying that a religious leader wants his community to
identify with the values of norms they propagate. Simultaneously, the leader
would like his community to be wealthy.
This payoff function is motivated by ample evidence that a leader has
an interest in followers obeying the rules and teachings of the religion and
doing so visibly. This can be easily observed in daily life when considering the
norms that different religious groups display. Payot and headscarfs identify
the religion of the wearer and thus shape their religious identity. Moreover,
the influence of religious leaders on the display of religious symbols among
their followers and thus on creating specific identities is widely documented,
see Breton (1964) and Carvalho and Koyama (2016).
Furthermore, leaders care about the financial well-being of their commu-
nity. This might be due to a need to maximize donations (Ceylan 2010;
Bekkers and Wiepking 2010) or from a sense of prestige or altruism. Every
religious organization we are aware of asks for donations. A religious leader
benefits from these donations as this allows, for example, to build a better
church, synagogue, mosque, or temple or, more generally, allows them to
provide a public good (Verdier and Zenou 2018). Further, religious leaders
care both about the spiritual well-being of their community members and their
material well-being: most religions do want their followers to be able to at least
afford food, clothing and housing.
Naturally, if leaders can encourage their community to identify with the
prescribed norms and values while simultaneously generating high incomes,
then leaders have an incentive to do so. A prominent example of a community,
which is both extremely religious and very wealthy, are the Jewish Diamond
Traders in New York described by Coleman (1990). Here, membership in
a religious community leads to better business opportunities. However, it
is more common in modern society that economic activities are conducted
outside their own community. In these instances, displaying strong religious
norms may be a hindrance to success.3
3 To be more precise, consider the impact of being part of a religious community on the attitude toward as
well as actual female labor force participation. Generally, views that see men as the main breadwinners and
Therefore, in many instances, leaders face a trade-off: they can either

emphasize the importance of adhering to religious norms and values, resulting
in higher identification with the community and its beliefs. Alternatively, they
foster moderate religious values, which allows for a successful labor market
integration of their followers, with higher income that may result in higher
donations.4
A similar trade-off has also been suggested by Carvalho and Koyama (2016),
which is formally a special case of the more general payoff analyzed in Prummer
and Siedlarek (2017). However, their interpretation differs: Carvalho and
Koyama (2016) argue that religious leaders face a trade-off between time spent
to produce a religious public good and time spent on market activities, which
generate income.
Last, it is worth connecting the two specifications outlined given in (7.1)
and (7.2): by allowing for the religious trait to be continuous, rather than
present or not, it is straightforward to introduce a cutoff point such that for
individuals with a religious identity or belief above the cutoff, they can be
considered as religious, whereas those below the cutoff can be seen as non-
religious. Then, it is also for the leader with payoff (7.2) optimal to maximize
the spread of religiosity. However, how religiosity is spread differs in these
two complementary setups. In one setting, the leader sets an identity, which
impacts earnings, in the other, the leader chooses a public good. Despite these
differences, the leader affects in both cases how attractive the membership in
the community is due to its economic consequences, which arise either due to
the public good provided or the earnings realized.
Leaders and Cultural Diversity

Having specified the payoff functions of leaders, we turn to the impact they
have on the religiosity of their community members. A key result common to
all work on leaders is that they increase religiosity, compared to the setting
women as homemakers are strongly influenced by religious ideology (Algan and Cahuc 2006; Guiso et al.
2003). This is also evidenced by a case study on Imams in Oslo (Predelli 2004), which documents that
these Imams do not forbid women to work in the public sphere, but that if women were to take outside
jobs they are encouraged to work in education or medical care. This seems to indicate that religious leaders
have indeed an impact on attitudes that are relevant for labor market outcomes. These attitudes also have
an impact on actual labor market outcomes, as perceptions of women as homemakers are closely associated
with women’s labor market outcomes (Fortin 2005).
4 For an explicit discussion of the role of religious organizations in discouraging education, see Carvalho
et al. (2017).
108 A. Prummer
without leaders. However, the leader’s efforts in increasing religiosity are

constrained by the economic situation of their community members or by
their ability to provide public goods. Prummer and Siedlarek (2017) document
that there can be three possible outcomes, namely, the leader chooses the
most extreme level of religiosity and intermediate levels or he refrains from
encouraging his community to display any particular religious behaviors
whatsoever. Which outcome emerges depends on the economic environment
and the interplay between earnings and religious beliefs in the leader’s payoff
function. We therefore discuss how productivity and discrimination, govern-
ment transfers, as well as community structure affect cultural assimilation.
Last, we highlight how results are affected if community size varies and describe
the cultural adjustment process.
Cultural Extremism, Distinction, and Integration
Consider the simplest case, where the leader aims to maximize the weighted
average of beliefs and earnings, as specified in Carvalho and Koyama (2016).
Then, if an increase in beliefs and religious identity is met by a smaller
decrease in terms of earnings and wages, the leader will always pursue a policy
of extremism. If, on the other hand, such an increase in religious identity
affected earnings negatively and significantly so, the leader will choose to
pursue a policy of assimilation which leads to the abandonment of religious
rituals. Carvalho and Koyama (2016) connect their theory to the rise of Ultra-
Orthodox Judaism. After the Emancipation of Jews, that is after they were
granted full citizenship, they were no longer restricted in their economic
activities. As a response to the Emancipation, some Jewish communities
shed many of their traditions, while other leaders created stricter rules the
communities had to abide by. Carvalho and Koyama (2016) document that in
areas in which economic opportunities were poor, Rabbis implemented stricter
religious rules, leading to Ultra-Orthodox communities, while in prosperous
areas, religious traits and behaviors were no longer openly displayed. In poorer
regions, it was not worthwhile to shed visible religious characteristics as this
would not have increased income sufficiently, whereas in the wealthier areas,
incomes were high enough to compensate for the now less distinct religious
teachings.
The picture becomes somewhat more nuanced if leaders do not only
maximize a weighted sum of religious beliefs and earnings, but if earnings and
religiosity enter multiplicatively in the leader’s payoff function. In this case,
there will never be a full suspension of religious values and norms, but rather an
intermediate level of religiosity. Religious leaders will choose to abolish some

restrictive religious practices, if doing so results in a large enough economic
gain for their parishioners.
The result that leaders prevent homogeneity within a society also emerges in
Hauk and Mueller (2015) and Verdier and Zenou (2015). Hauk and Mueller
(2015) show that the leader of a religious community, who faces a mainstream
society, will strengthen the cultural, non-pecuniary benefits a community
member derives from the group and create a perception that mainstream
society cannot provide these cultural benefits. In this sense, the leaders select
distinct cultural values that continue to differ from those of the wider society.
Similarly, Verdier and Zenou (2015) show that cultural leaders ensure cultural
diversity in a society. All of these papers, which employ distinct models, but
reach very similar conclusions, highlight the importance of cultural leaders in
obtaining culturally distinct groups.
Comparative Statics
We now turn to a discussion of different economic and social variables that can
affect the leader’s choice: (1) the economic gain from working, (2) transfers to
community members and (3) the extent of heterogeneity within the group. All
three of them affect the extent to which religious beliefs are propagated; for
more details, see Prummer and Siedlarek (2017).
Economic Gains If members of a religious community decide to enter the

main stream labor market, they will be able to weakly increase their earnings.
The argument is a simple one: if agents are not restricted in their job search
to their own community anymore, they may match with better paid jobs,
increasing their income. However, their income also depends on the economic
environment they face. Both the general economic situation and the payoffs
that the members of a specific community can generate, that is, the extent
of discrimination they face, matter for their decision to enter the general
labor market. If the economic situation is good and discrimination is of
little concern, then agents have a greater incentive to look for work outside
their community. However, a leader who observes group members taking up
employment outside their own community may fear for his influence and for
the cohesion of the group. Group members who work outside their parish are
more likely to be exposed to worldly influences, which can shape their attitudes
toward their own religion and which may lead to a decreased identification
with their original norms and values. Therefore, religious leaders may have an
110 A. Prummer
incentive to impose stricter rules, to propagate more distinct cultural values,

in order to counterbalance the influences of the outside world. Therefore, if
agents face a better economic situation, they will have higher earnings, but
at the same time, they display more religious values. This finding highlights
how a policy analysis that does not take the presence of a religious leader into
account can yield misleading results: without any religious leader, religious
community members eventually integrate if it is worthwhile for them to join
the mainstream labor market. In the presence of a leader, they will never fully
shed their religious traits and attitudes. One controversial implication of this
is that if communities face a labor market which allows for high earnings,
but only to those who do not display any religious traits or symbols, then
religious leaders will allow for a greater assimilation toward the mainstream
society. Clearly, this result is an extreme one, and it does not seem plausible
that we as a society should strive for complete homogeneity. But in some
Western societies that aim for a reduction of extremism, a restriction of obvious
religious symbols may be a path toward more homogeneity as it forces more
assimilation, although the cost is entirely born by the members of the religious
community. Remarkably this result does not only emerge in the framework
suggested by Prummer and Siedlarek (2017) but also in the distinct model of
Hauk and Mueller (2015). There, leaders choose more distinct cultural values,
that is, they pursue a strategy of alienation if discrimination of the mainstream
is low, the logic being exactly the same as described above.
Government Transfers The role of government transfers in fostering cultural

integration of different religious communities has also received significant
attention (Koopmans 2010). Assuming that these transfers can be modeled
as lump sum transfers, Prummer and Siedlarek (2017) show that this leads
to a reduction in incentives to join the mainstream labor market. Instead,
leaders have an incentive to increase the extent of religiosity, having now the
certainty that even if their community members do not join the mainstream
labor market, they have sufficient funds. In this respect, the transfers reduce the
incentives to weaken the differences in norms and values between the religious
groups and the outsiders. However, the religious community is better off: they
are able to keep their religious values and are simultaneously wealthier.
Community Heterogeneity Last, the extent to which a religious commu-
nity displays different values and norms depends on how tightly knit the
community is, where the community members differ in their productivity.
In each community, there are members who are more productive and who
have greater incentives to integrate in the mainstream labor market and thus in
turn to shed some of their religious characteristics. Less productive community
members stand little to gain from seeking employment outside the group and
are therefore more identified with the norms and values of a given group. If
community members are more connected, that is, they have a greater influence
on each other, then their religious beliefs are more aligned. This implies that
the differences in productivity are moderated. In this case the religious leader
can set a higher level of religious norms and values, as he caters to a more
homogeneous group. This result hinges on the fact that more religious group
members are influenced to a great extent by the preaching and input of the
leader, relative to more moderate members in the community. Therefore, a
more extreme norm by the leader induces a greater loss from the religious
members due to their significantly reduced income, compared to the moderate
ideological gain from the moderate members. This effect is ameliorated in
a tight-knit community, as the more religious members are relatively less
religious, but the moderate members are relatively more religious compared
to the case in which their mutual influence is limited. Thus, if community
members influence each other more, then the most religious members become
less identified with the leader, which is countered by a greater level of cultural
distinction.
These comparative statics results shed light on what measures societies may
have to undertake in order to foster cultural integration, if this is a goal they
strive for. Additionally, it highlights why some groups are more likely to display
more extreme and persistent differences in cultural traits—the tightness of
their community.
Changes in Community Size and Transition Process
Changes in Community Size So far we have focused on a leader that takes

the community size as given. However, as outlined previously, it might also be
the case that a leader has a missionary zeal and aims to increase the community.
If this is the case, then he will aim to be as appealing to the mainstream
society as possible, by either providing cultural, non-pecuniary payoffs to them
(Hauk and Mueller 2015) or by providing a larger public good. This makes
the community more attractive to them, and thus they may be inclined to
become community members themselves. Such a strategy is made possible as
the leader can appeal to those outside a community, potentially with a public
good. Therefore, the leader aims to make their own community as appealing as
possible and the outside community as off-putting as possible, considerations
absent in Prummer and Siedlarek (2017).
112 A. Prummer
Transition Process All findings here are based on a steady-state analysis,

where the leader sets one level of religiosity. However, transition phases are
also of importance to politicians who aim to implement policies. Verdier and
Zenou (2018) analyze this feature in depths, allowing for leaders to adjust
the benefits of participating in the community. Recall that in their setting,
which assumes binary religious traits, leaders reap benefits from the share of
individuals who display the religious trait they propagate. They induce agents
to join their community by providing a public good that comes at a cost
to them.5 If it becomes marginally more costly for the leader to provide the
public good, then the leader reacts with an under-provision of the public good
until a new steady state is reached. Therefore, a policy aimed at increasing
cultural integration may initially seem more promising than it actually is. A
full evaluation of a measure will only be possible once the new steady state has
been reached, after a certain amount of time has passed.
Leader Competition
So far, we have restricted attention to the case with one religious leader who
sets values and norms or decides on a public good. In many contexts, this
seems an adequate description. When discussing the cultural norms of the
minority of Muslims in Western countries, it is sensible to keep the culture
in the mainstream society fixed. However, if there are two religious groups of
somewhat equal size, an explicit characterization of how the leaders interact,
or rather compete, is in order. This analysis has been conducted by Verdier and
Zenou (2015, 2018) as well as Hauk and Mueller (2015). Verdier and Zenou
(2015) consider two leaders who myopically invest in their own cultural trait
and then the cultural traits of the agents in the population are determined
by a contest success function, which depends on the effort of the leaders.
The key result is that the presence of the leaders ensures cultural distinction
and heterogeneity in the population, which would not be the case without
the leaders, a result mirroring the findings with one leader and thus ensuring
their robustness. However, their finding crucially depends on the assumption
that leaders are myopic. This is relaxed in Verdier and Zenou (2018), who
consider competition between perfectly forward-looking religious leaders.
These leaders can differ in terms of how patient they are. If both evaluate
the future identically, then both select the same levels of public goods. This
5 The cost can be either paid by an external power or by the community itself through donations.
implies that the long-run outcomes for the communities are the same as if no
leaders were present—the effort of the leaders simply cancels each other out.
If, however, one leader is more patient than the other, then the less patient
leader does not provide any public good, essentially becomes inactive. The
notion of leaders displaying different levels of patience can be interpreted
as differences in institutional stability. If a leader is part of an organization
that has a strong base, then such a leader can naturally be more forward-
looking. Verdier and Zenou (2018) provide two distinct examples: (1) the more
patient leader represents the mainstream trait of a society, whereas the less
patient leader heads a minority group; and (2) the more patient leader has
followers who display a strong religious commitment, without any political
motivations, whereas the less patient leader is a secular leader, who faces a lot
of political uncertainty. In the latter case, the secular leader may then refrain
from opposing the more patient, religious leader, and thus the model provides
an explanation for why extremist leaders may face too little opposition.
Up until now, leader competition assumed that leaders head different
religious communities. In addition to this external competition between two
leaders, Hauk and Mueller (2015) consider internal competition. Internal
competition would arise if there are two or more leaders who compete to
propagate the same religious trait. Internal competition will lead to moder-
ation as leaders now have an incentive to increase the utility of their followers.
One way to do so is to invoke less fear within their community regarding the
mainstream society.
Emergence of Leaders
Up to this point, we have taken the existence of one or two leaders as
given. However, it is natural to ask under what circumstances religious leaders
emerge. This seminal question has been addressed by Verdier and Zenou
(2018). In this setting, leaders have an incentive to increase the size of their
community, by spreading a religious trait. In absence of a leader, traits are
transmitted by parents in the standard socialization framework developed in
Bisin and Verdier (2000). The leader can improve on what the parents do,
by providing additional means of transmitting a cultural trait. His decision
to become active depends crucially on the benefits he obtains from leading a
group. If these benefits are too low, then naturally a leader will never emerge.
If the benefits are high however, the decision to become active depends on the
share of agents that are already characterized by the religious trait. If this share
is too low or too high, then the leader chooses to remain inactive, and only
114 A. Prummer
for an intermediate group size will he provide a public good. This surprising
non-monotonicity emerges for the following reason: if the community size is
small, then even if the leader becomes active, the leader’s effort cannot ensure a
lasting increase in the share of agents with the trait he propagates. Similarly, if
the community size is large, then the leaders effort are unnecessary, compared
to the parents efforts. In this case, parents’ and leaders’ efforts display the
usual cultural substitutability effect (Bisin and Verdier 2000). Only for an
intermediate group size the leader becomes active and his efforts complement
those of the parents.
This non-monotonicity in the leader’s decision to become active or not also
has interesting implications for policy measures. The share of agents with the
leader’s religious trait depends on the utility they derive from such a religious
trait. If a government that aims to foster a greater degree of cultural integration
impacts the utility derived from the religious trait negatively, then the response
of the community can be fundamentally different if a leader is present (whether
he is active or passive) compared to a setting where a leader cannot emerge.
In particular, such an assimilationist policy may induce a leader to become
active, which can be interpreted as a form of cultural resistance. Additionally,
a migratory shock can have lasting implications, as it may cause the emergence
of a religious leader and thus to a significantly larger, religious community than
previously anticipated.
Verdier and Zenou (2018)’s analysis is based on the natural assumption
that there always exists a leader who may become active or not. This is a
natural assumption, for example, in the context of Turkish Imams, deployed
to foreign countries by the Turkish government. In other instances however,
communities themselves choose their religious leaders. This is, for example,
common among British Muslims (Geaves 2008). In such circumstances, a
religious leader may have a different function. In his seminal paper, Iannaccone
(1992) has argued that churches are like clubs everyone contributes to. In any
club there are externalities and thus every congregation suffers from free-riding
problems. While churches have the means to address free-riding, for example,
through requiring sacrifices (Iannaccone 1992), it may be beneficial, especially
as a community grows, to have a leader, a priest, who is paid to ensure that
everyone makes a contribution, someone who ensures that the community
sticks together. This argument has been formalized in a different context
by Acemoglu and Wolitzky (2015), who show under what circumstances a
system with a specialized enforcer outperforms a system with peer punishment.
Additional evidence that central institutions can indeed outperform peer
enforcement has also been documented in a different context by Greif (1994).
Future Research
This brings us to future research on the topic of religious and cultural leaders.
First, a formal analysis of the impact of a leader, who has been called in
by his own community, on this community seems in order. Do the leaders
merely serve to anchor already present religious convictions or do they indeed
increase the religiosity of their followers? Under what circumstances is this
leader called in? In order to answer these questions, a model that endogenizes
both the leaders’ choice of the extent of religiosity and the decision of the
community members and their adjusted behavior due to their participation
in the community is required.6 Moreover, an analysis of the different tools
a religious leader has available, both to keep his community together and to
increase its appeal to members on the outside (see also Hauk and Mueller
(2015)), seems to be in order.
While the role of cultural and religious leaders has received some attention
from a theoretical perspective, there is next to no empirical work on the impact
of a religious leader. A notable exception to this is Nteta and Wallsten (2012),
who show that religious leaders have an impact on their parishioners regarding
their attitude toward immigration laws. Empirical research on how influential
leaders are or how important they can be in shaping attitudes, values and
norms seems fundamentally important in order to understand more generally
the impact an individual can make, which can matter for the implementation
of policies.
References
Acemoglu, D., and A. Wolitzky. 2015. Sustaining Cooperation: Community Enforce-
ment versus Specialized Enforcement. National Bureau of Economic Research.
Algan, Y., and P. Cahuc. 2006. Job Protection: The Macho Hypothesis. Oxford Review
of Economic Policy 22: 390–410.
Bekkers, R., and P. Wiepking. 2010. A Literature Review of Empirical Studies of
Philanthropy: Eight Mechanisms that Drive Charitable Giving. Nonprofit and
Voluntary Sector Quarterly 40: 924–973.
Bisin, A., and T. Verdier. 2000. “Beyond the Melting Pot”: Cultural Transmission,
Marriage, and the Evolution of Ethnic and Religious Traits. The Quarterly Journal
of Economics 115: 955–988.
6 A model that incorporates many of these features, without addressing the role of the leader specifically,
has been made by Carvalho (2016).
116 A. Prummer
Breton, R. 1964. Institutional Completeness of Ethnic Communities and the Personal

Relations of Immigrants. American Journal of Sociology 70: 193–205.
Carvalho, J.-P. 2016. Identity-Based Organizations. American Economic Review 106:
410–414.
Carvalho, J.-P., and M. Koyama. 2016. Jewish Emancipation and Schism: Economic
Development and Religious Change. Journal of Comparative Economics 44: 562–
584.
Carvalho, J.-P., M. Koyama, and M. Sacks. 2017. Education, Identity, and Commu-
nity: Lessons from Jewish Emancipation. Public Choice 171: 119–143.
Ceylan, R. 2010. Die Prediger des Islam Imame–Wer sie sind und was sie wirklich wollen.
Freiburg: Herder.
Coleman, J. 1990. Foundations of Social Theory. Cambridge, MA: Belknap Press of
Harvard University Press.
Fortin, N.M. 2005. Gender Role Attitudes and the Labor-Market Outcomes of
Women across OECD Countries. Oxford Review of Economic Policy 21: 416–438.
Geaves, R. 2008. Drawing on the Past to Transform the Present: Contemporary
Challenges for Training and Preparing British Imams. Journal of Muslim Minority
Affairs 28: 99–112.
Greif, A. 1994. Cultural Beliefs and the Organization of Society: A Historical
and Theoretical Reflection on Collectivist and Individualist Societies. Journal of
Political Economy 102: 912–950.
Guiso, L., P. Sapienza, and L. Zingales. 2003. People’s Opium? Religion and Eco-
nomic Attitudes. Journal of Monetary Economics 50: 225–282.
Journal of Conflict Resolution 59: 367–400.
Iannaccone, L.R. 1992. Sacrifice and Stigma: Reducing Free-Riding in Cults, Com-
munes, and Other Collectives. Journal of Political Economy 100: 271–291.
Koopmans, R. 2010. Trade-offs Between Equality and Difference: Immigrant Integra-
tion, Multiculturalism and the Welfare State in Cross-National Perspective. Journal
of Ethnic and Migration Studies 36: 1–26.
Nteta, T.M., and K.J. Wallsten. 2012. Preaching to the Choir? Religious Leaders and
American Opinion on Immigration Reform. Social Science Quarterly 93: 891–910.
Portes, A. 1987. The Social Origins of the Cuban Enclave Economy of Miami.
Sociological Perspectives 30: 340–372.
Portes, A., and J. Sensenbrenner. 1993. Embeddedness and Immigration: Notes on
the Social Determinants of Economic Action. American Journal of Sociology 98:
1320–1350.
Predelli, L.N. 2004. Interpreting Gender in Islam: A Case Study of Immigrant
Muslim Women in Oslo, Norway. Gender & Society 18: 473–493.
Subervi-Velez, F.A. 1986. The Mass Media and Ethnic Assimilation and Pluralism: A
Review and Research Proposal with Special Focus on Hispanics. Communication
Research 13: 71–96.
Verdier, T., and Y. Zenou. 2015. The Role of Cultural Leaders in the Transmission of
Preferences. Economics Letters 136: 158–161.
———. 2018. Cultural Leaders and the Dynamics of Assimilation. Journal of Eco-
nomic Theory 175: 374–414.
Zhou, M., and G. Cai. 2002. Chinese Language Media in the United States:
Immigration and Assimilation in American Life. Qualitative Sociology 25: 419–441.
8
Intermediated Social Preferences: Altruism
in an Algorithmic Era
Daniel L. Chen
JEL Classification B51, C93, D63, D64, J15, K00
Introduction
The role of markets in moral behavior is poorly understood. Economists
and philosophers usually study the boundary between markets and gov-
ernment, rather than the issue of what should and should not be on the
market. It has been hypothesized that market interactions corrode moral
values (Shleifer 2004; Radin 1987). In societal transition to market economies,
human economic mentalities were changed, and people became more eco-
nomically rational, behaving as neoclassical economic theory would pre-
dict (Polanyi 1944). Prior to the transition, people based their economies
Most recent version at: http://users.nber.org/∼dlchen/papers/Intermediated_Social_Preferences.pdf. I

thank Aroha Bahuguna and Veronika Bordas for research assistance. Work on this project was conducted
while I received financial support from the European Research Council (Grant No. 614708), Swiss
National Science Foundation (Grant Nos. 100018-152678 and 106014-150820), Agence Nationale de la
Recherche, John M. Olin Foundation, Ewing Marion Kauffman Foundation, and Templeton Foundation
(Grant No. 22420).
D. L. Chen ()
Toulouse School of Economics, Institute for Advanced Study in Toulouse,
University of Toulouse Capitole, Toulouse, France
e-mail: daniel.chen@iast.fr

120 D. L. Chen
on reciprocity and redistribution and were not rational utility maximiz-

ers. This chapter investigates one mechanism of the great transformation—
intermediation (Judge 2012)—by which market forces can shape morality.
Other aspects of market interactions—commodification (market alienabil-
ity) and competition—have been explored elsewhere (Chen 2016), so this
chapter explores the role of intermediation and does so with an experi-
ment.
Broadly speaking, the proponents of the doux commerce thesis (a theory
popularized by eighteenth-century political philosophers) have proposed that
markets, with its disruptive effect on geographical and tribal isolation, will
actually have morally improving effects, increasing our care for and under-
standing of others. On the negative side, intermediation gives individuals the
possibility to justify taking egoistic actions while maintaining a feeling of
morality. People want to believe that they are moral—and may be motivated
less about outcomes or actions and more about revealing themselves as moral
beings. In experiments where individuals could choose to allocate either a
pleasant or an unpleasant task to a partner and keep the other task, but had
the opportunity to flip a coin to do the allocation, 90% of individuals who
chose to flip the coin assigned the unpleasant task to the partner and felt more
moral (Batson et al. 1997). In other experiments, individuals want to avoid
revealing to themselves what are the payoffs to the other so they could avoid
feeling moral disutility when choosing a selfish outcome (Dana et al. 2007). In
vignette studies, moral judgment of a litigation scenario changes if the subject
knew which side they would be assigned to (Babcock et al. 1995). These and
other studies of self-serving interpretations on what is right or wrong lend
possibility to intermediated moral decisions becoming less pro-social.
In general, measuring how market experiences affect moral values is dif-
ficult. Inferring causality from correlation is difficult since the causality can
go in both directions. Organizations may foster dramatic changes in value
orientations (Kohn 1986), but value orientations can also foster economic
change (Katz and Goldin 2000; Eriksson and Villeval 2008). Pre-existing
traits may drive selection into competitive environments. Such self-selection
makes it difficult to ascertain causal link between market conditions and
moral behavior (Fletcher and Nusbaum 2008; Ford and Richardson 1994;
Detert et al. 2008; Dubinsky and Ingram 1984). Cross-sectionally, market
integration was found to be positively associated with moral behavior in
the analysis of economic games measuring social preferences (Henrich et al.
2001). In laboratory experiments with strategic settings, however, someone
who chooses to intermediate moral responsibility is less likely to be punished
by norm-enforcing third parties (Coffman 2011). This study investigates the
8 Intermediated Social Preferences: Altruism in an Algorithmic Era 121
behavioral question: does intermediating moral responsibility change moral

behavior even in the absence of third party punishment? If so, what aspect of
intermediation matters? In particular, does algorithmic intermediation differ
from social intermediation?
To study this question, I use a labor market intermediary (LMI), using
a methodology similar to what the author has employed in other studies
(Chen 2016; Chen and Yeh 2014; Chen and Horton 2016). An advantage of
contextualizing is that natural field experiments mitigate potential Hawthorne
effects relative to lab experiments (Orne 1962; Titchener 1967). Workers
are recruited through Amazon Mechanical Turk. The LMI is designed to
recruit a large number of workers in a short amount of time. Through an
interface provided by the LMI, registered users perform tasks posted by buyers
for money. The tasks are generally simple for humans to do yet difficult
for computers. Common tasks include captioning photographs, extracting
data from scanned documents, and transcribing audio clips. The LMI also
allows a researcher to implement randomization although randomization is
not inherent to the LMI. Although most buyers post tasks directly on the LMI
website, they are also able to host tasks on an external site. I use this external
hosting method; I post a single placeholder task containing a description of the
work at the LMI and a link for workers to follow if they want to participate.
The subjects are then randomized, via stratification in the order in which they
arrived at the job, to one of several treatment conditions. Treatment is not
revealed at this early stage. All workers see identical instructions.
The experimental approach eliminates omitted variable biases where certain
types of individuals select into different market experiences. After workers
complete data entry, I ask them to grade the work of another. I observe the
accuracy of the workers’ own data entry as well as their evaluations of the
accuracy of their co-workers’ data entry. Workers are asked to propose a split
of a 50-cent bonus with the other worker. The split of a 50-cent bonus is a
contextualized dictator game. I use this split to measure a workers’ altruism
(or moral behavior, conditional on the number of perceived errors of the co-
worker). The hypothesis is that increasingly intermediating responsibility for
moral decisions can lead to more selfish behavior, even if workers rate the data
entry task equally well. This is likely to be true when intermediation is more
asocial—maintaining or increasing isolation.
The results also relate a literature on whether altruism is deontological
(Chen and Schonger 2016, 2017) or egoist (Becker 1976). Under a deontologi-
cal view, altruistic behavior should not change with the circumstances. We can
think of the thought experiment from Kant. In a classic vignette, a murderer
asks you whether your friend is hiding in your house (Kant 1797). In the
122 D. L. Chen
categorical imperative, Kant would say, “You must not lie.” No matter what the
consequences are, you must tell the truth. Under the egoist view, individuals
are altruistic only because they get some benefit from being perceived as being
altruistic. This would suggest that, as the circumstances of how an individual’s
actions are perceived change, the agent’s altruism will change as well. Another
related literature is the impact of algorithms and machine learning on judicial
decision-making. A large collection of findings on the malleability of moral
reasoning or decision-making by judges can be modeled as shifts in reference
points about what is the just and fair decision.1 In an era when algorithms may
be used in lieu of legal actors (Amaranto et al. 2018), an open question is how
this historical shift may impact judicial decisions.2
Methodology
The LMI can be used to implement anything from a natural field experiment
to a laboratory experiment (Harrison and List 2004). Workers come to the
marketplace naturally and are unaware they are in an experiment at the time
of arrival, and this lack of awareness alleviates the Hawthorne effects. Even if
people become aware of an experiment, they are unaware that other subjects
receive different treatment conditions. The behavior of subjects in this labor
market intermediary is comparable to the behavior of subjects in a laboratory
and may be comparable to subjects in a real labor market (Barankay 2010).
The experimental design is shown in Fig. 8.1.
I ask workers to transcribe paragraphs from a Dutch translation of Adam
Smith’s The Wealth of Nations. This task is sufficiently tedious that no one is
likely to do it “for fun,” and it is sufficiently simple that all market participants
can do the task. The source text was machine-translated to prevent subjects
from finding the text elsewhere on the Internet. Time and money are the
1 Malleabilityof moral reasoning by judges has been documented in US federal circuit judges (Ash et al.
2016; Chen 2017b; Chen et al. 2016d), federal district judges (Chen 2017a; Barry et al. 2016), immigration
judges (Chen et al. 2016c), sentencing judges (Chen and Prescott 2016; Chen and Philippe 2017), military
judges (Chen 2017c), and juvenile judges (Eren and Mocan 2016). Some of these findings can be attributed
to snap judgments whether from analysis of the first three seconds of oral arguments (Chen et al. 2016a,
2017a) or from early predictability of judicial decisions based on race or nationality (Chen et al. 2017;
Chen and Eagel 2017).
2 Outside the lab, the malleability of injunctive norms to formal institutions such as the law (Chen and Yeh
2016b, 2014; Chen et al. 2017b) or markets (Chen 2015b; Chen and Lind 2016; Chen 2016) is suggestive
of the impact of broader historical shifts in human rights (Chen 2005), sexual harassment (Chen and
Sethi 2016), and free speech (Chen 2015a). This chapter also shares the experimental approach to measure
normative commitments (Chen et al. 2016b; Shaw et al. 2011).
Pool of
subjects
Worker agrees to perform 3 paragraph

transcriptions at a 10 cent piece rate.
Sample paragraph is shown.
S SA SR CS CM
Worker Worker Worker Worker Worker
completes 3 completes 3 completes 3 completes 3 completes 3
transcriptions transcriptions transcriptions transcriptions transcriptions
Worker Worker Worker Worker Worker

grades other grades other grades other grades other grades other
worker’s worker’s worker’s worker’s worker’s
transcription. transcription. transcription. transcription. transcription.
Grader Two other Two other With 2/3 Computer

determines graders. graders. One probability, modifies
split. Average of of three splits computer grader’s split.
splits implemented split is
Short survey implemented. at random. implemented. Short survey
Short survey Short survey Short survey
End
Fig. 8.1 Experimental design
most cited reasons for participation in Mechanical Turk.3 Because subjects are
unaware of an on-going experiment, differential attrition may arise at the time
3 http://behind-the-enemy-lines.blogspot.com/2008/03/mechanical-turk-demographics.html. Some
workers do it out of need. A disabled former United States Army linguist became a Turk Worker for
various reasons, and in nine months he made 4000 dollars (New York Times, March 25, 2007). Some
drop out of college to pursue a full-time career with these disaggregated labor markets (Web Worker
Daily, October 16, 2008, Interview with oDesk CEO). For more information about the motivation and
demographics of Mechanical Turk workers, see, for example, Paolacci et al. (2010).
124 D. L. Chen
treatment is revealed (Reips 2001). I minimize attrition through a commitment

mechanism. In all treatment conditions, workers face an identical “lock-in”
task in order to minimize differential attrition before the treatment is revealed.
This lock-in successfully reduces attrition (Chen 2016).
The payment for each paragraph is 10 cents. They are told there are three
paragraphs, all of which must be completed to be paid. A paragraph takes
about 100 seconds to enter so the offered payment of 10 cents per paragraph is
equivalent to $86.40 per day. The current US federal minimum base wage for
tipped waiters is $17.20 per day, and the federal minimum wage in the United
States is $58/day. In India, it depends on the type of work done, although the
“floor” for data entry positions appears to be about $6.38/day.4 An example
paragraph is displayed on the first page of the external hosting site so workers
are aware of the high payment before beginning work.
Paragraph Sample: De jaarlijkse arbeid van elk volk is het fonds die oorspronke-
lijk levert hij met alle benodigdheden en conveniencies van het leven die het
jaarlijks verbruikt, en die altijd bestaan, hetzij in de onmiddellijke produceren
van die arbeid, of in wat wordt gekocht met die van andere landen. Volgens
dus, als deze producten, of wat is gekocht met het, draagt een grotere of kleinere
verhouding tot het aantal van degenen die zijn om te consumeren, het volk zal
beter of slechter geleverd met alle de benodigdheden en conveniencies waarvoor
zij gelegenheid. Maar dit deel moet in elk volk worden geregeld door twee
verschillende.
Each treatment group had different specifications about how the split would
be implemented, increasingly distancing the subject from his action by making
the proposed split less clear to an outside party.
After a lock-in task of three paragraphs, subjects are then asked to grade the
task of another worker. They compare a scanned text with another worker’s
answer and enter the number of errors found and the general assessment
of worker quality (very low/low/fair/high/very high). The task comes from
data entry experiments involving summaries of court decisions (Chen and Yeh
2016a,b). Treatment is revealed at the next stage. They are told they and the
4 Payscale, Salary Snapshot for Data Entry Operator Jobs, http://www.payscale.com/research/IN/Job=

Data_Entry_Operator/Salary?, accessed June 17, 2011.
worker whose work they just graded will receive a bonus and that they need
to share a ω cent bonus with the worker. The treatment groups are:
G1 Split (Control): Workers are told that their split of the bonus will be
implemented.
G2 Peer Average Split: Workers are told that two others are also grading
the same work. The average of the three workers’ proposed splits will be
implemented.
G3 Peer Random Split: Workers are told that two others are also grading
the same work. One of the three workers’ proposed splits will be chosen at
random to be implemented.
G4 Computer Random Split: Workers are told that, with 2/3 probability,
a computer algorithm designed to optimize among data entries will implement
its proposed split instead of the worker’s split.
G5 Computer Modification: Workers are told that a computer algorithm
designed to optimize among data entries will modify their proposed splits of
the bonus.
G2 and G3 provide a human component to intermediation, while G4 and
G5 provide non-human intermediation. When workers prospectively think
about peers, they become more pro-social (Shaw et al. 2011). Thinking about
one’s peers may be a mediating factor for the effects of intermediation on moral
behaviors. G3 is closest to G4 except the randomly chosen split is among peers
or with a computer. Note that this treatment differs from standard diffusion
of responsibility settings (volunteer dilemma) since it is not the case that any
of the graders can ensure a fair outcome (Dana et al. 2007; Darley and Latané
1968). G5 differs from G4 in that a computer will definitely as opposed to
possibly deviate from the proposed split. It serves to some extent as a control for
G4 in the event that simply reminding individuals about computer algorithms
may have a priming effect separate from intermediation. The computer’s split
in both G4 and G5 is a uniform random number. The exact wordings of the
instructions are provided in Appendix.
Experimental Results
Demographic characteristics are balanced across treatment groups, consistent
with the randomization of workers across treatment. Table 8.1 displays sum-
mary statistics by treatment interaction. Males comprise 36% of the sample.
A total of 36% and 35% are from the United States and India, respectively. A
total of 25% are Christian, 28% are Hindu, and 16% are atheist. The average
age is 30. The average religious attendance is between once a year and once a
126 D. L. Chen
Table 8.1 Summary statistics by treatment group

Peer Peer Computer
average random random Computer
Control split split split modification Overall
Bonus split 20.9 21.6 22.4 17.8 22.2 20.9
(8.8) (10.3) (7.9) (9.7) (10.8) (9.6)
Male 0.5 0.5 0.4 0.4 0.4 0.4
(0.5) (0.5) (0.5) (0.5) (0.5) (0.5)
Age 29.0 28.9 31.1 30.6 30.4 30.0
(9.2) (9.3) (11.8) (9.9) (9.7) (10.0)
American 0.4 0.4 0.5 0.4 0.4 0.4
(0.5) (0.5) (0.5) (0.5) (0.5) (0.5)
Indian 0.4 0.4 0.4 0.5 0.4 0.4
(0.5) (0.5) (0.5) (0.5) (0.5) (0.5)
Christian 0.3 0.3 0.4 0.3 0.3 0.3
(0.4) (0.5) (0.5) (0.5) (0.4) (0.5)
Hindu 0.3 0.3 0.4 0.4 0.3 0.3
(0.5) (0.4) (0.5) (0.5) (0.5) (0.5)
Atheist 0.3 0.3 0.2 0.1 0.2 0.2
(0.4) (0.4) (0.4) (0.3) (0.4) (0.4)
Religiousness 2.4 2.6 2.5 2.8 2.6 2.6
(1.4) (1.5) (1.3) (1.4) (1.4) (1.4)
Respect for 4.5 4.7 4.8 4.9 4.8 4.8
Parents (1.0) (0.8) (0.6) (0.3) (0.8) (0.7)
Respect for 4.2 4.3 4.4 4.3 4.1 4.3
Supervisor (1.1) (0.9) (0.8) (1.0) (1.0) (1.0)
Respect for 4.0 4.2 4.3 4.3 4.2 4.2
Police (1.1) (1.1) (0.9) (0.9) (1.1) (1.0)
Grades 4.2 4.2 4.1 4.0 4.0 4.1
(0.8) (0.7) (0.8) (0.8) (0.7) (0.8)
Errors found 2.7 1.7 1.5 1.5 1.9 1.8
(12.2) (2.9) (1.0) (1.0) (1.5) (5.5)
Own errors 55.9 54.8 42.3 60.9 38.1 50.5
(114.8) (145.4) (78.9) (128.8) (107.0) (116.8)
Observations 67 79 76 88 76 386
month. The average respect for parents and authority is between “a little” and
“a lot.” After work has been completed, according to the original expiration
date listed on the LMI, bonuses are calculated and workers are notified of their
earnings.
The empirical specification examines the effect of treatment on donation:
Donationit = β 0 + β t1 T reatment it + β 2 Xit + εit (8.1)

Fig. 8.2 Proposed split by groups

50
All Others
Computer Random Split
4030
Percent
20
10
0
0 10 20 30 40 50
Split
Fig. 8.3 Distribution of bonus splits
T reatment it represents the treatment group for individual i in treatment t

and Xit represents individual demographic characteristics. Figure 8.2 displays
the raw data means of each treatment condition. Individuals assigned to
Computer Random Split condition donated the least. Figure 8.3 plots the
distribution of the raw data, comparing the distribution of donations in the
128 D. L. Chen
60
Control
Computer Random Split
40
Percent
20
0
0 10 20 30 40 50
Split
Fig. 8.4 Distribution of bonus splits
Table 8.2 Altruism by treatment group

Bonus split
Full sample Male Female
(1) (2) (3) (4)
Peer average split 0.34 0.89 −0.58 0.067
(0.21) (0.55) (−0.21) (0.33)
Peer random split 1.42 1.35 −1.56 3.02
(0.87) (0.84) (−0.55) (0.15)
Computer random split −2.81* −2.65* −5.94** −0.87
(−1.77) (−1.64) (−2.10) (−0.45)
Computer modification 0.53 1.91 −0.15 1.74
(0.33) (1.17) (−0.05) (0.88)
Controls N Y Y Y
Observations 386 356 154 202
Notes: t statistics in parentheses. Control is the omitted category
*p < 0.10, **p < 0.05
Computer Random Split condition with the distribution of donations in all

other treatment conditions. Figure 8.4 plots the distributions only comparing
the Control (G1) and Computer Random Split (G4) conditions.
Table 8.2 reports ordinary least squares regressions. I also conduct a
Wilcoxon-Mann-Whitney test, a non-parametric analog to the independent
samples t-test, which can be used when the dependent variable is not normally
distributed, but at least ordinal.
The Wilcoxon-Mann-Whitney test reveals that the Computer Random

Split condition is statistically significantly different at the 1% level. Similar and
statistically significant results obtain when I compare only Control with the
Computer Random Split condition in the regression and in the Wilcoxon-
Mann-Whitney test. In a linear regression, the Computer Random Split
condition is also a statistically significant determinant of the proposed split, at
the 10% level. Workers in this group contribute about 2.8 cents less than the
control group. Column 1 displays the results without demographic controls.
Column 2 adds controls, which results in the loss of some observations due
to non-response on some demographic questions, but the effect remains
quantitatively similar. Column 2 also controls for the perceived error of the
recipient.
In results available on request, I find a number of patterns in the data that
are consistent with previous findings: workers deemed as providing higher
quality work are given higher bonuses (Eckel and Grossman 1996; Ruffle 1998)
and men are less generous than women (Andreoni and Vesterlund 2001; Eckel
and Grossman 2008). These partial correlations, however, disappear when the
complete set of demographic controls is included.
Notably, among male subjects, the treatment effect is accentuated. Men
donate 5.9 cents less than the control group when assigned the Computer
Random Split condition. The effect is statistically significant at the 5% level.
When examining only Americans or only Indians, subjects assigned to the
Computer Random Split condition donated the least.
Workers randomly assigned to peer treatment conditions (G2 and G3)
display no effect of intermediation on donations, which is consistent with
online workers becoming more pro-social when prospectively thinking about
peers (Shaw et al. 2011). No effect is found for direct computer modification
(G5).
The great transformation of human mentalities into an economically ratio-
nal cost-benefit perspective has been attributed to the societal transition to
market economies (Polanyi 1944). This experiment has found that when
asocial decision-intermediation eroded indirect reciprocity norms and gen-
erosity toward others. In an algorithmic era, as interactions become more
asocially intermediated, decisions may become less social and other-regarding.
This might suggest that market interactions corrode moral values only the
interactions are asocial.
130 D. L. Chen
Conclusion
While economists primarily focus on efficiency, philosophers worry about how
markets leave their mark on social norms. Two oft-stated concerns are whether
people become corrupted or degraded when an aspect of the human experience
is commodified and whether market competition makes people immoral.
Economists and psychologists have begun to approach the issue: documenting
repugnance of certain market transactions, negotiating around taboo tradeoffs,
and questioning why some normative arrangements are repugnant. However,
little empirical and no experimental research has been conducted on the issue
of market inalienability nor on the issue of market intermediation despite
the potential role for intermediation in explaining the financial crisis (Judge
2012). This chapter takes a first step in investigating the causal effect of
intermediation on moral behavior, which may have implications for behaviors
in other domains, be they religion, law, or ethics.
Appendix Section 1: Placeholder Task at Amazon

Mechanical Turk
Transcribe Text
Instructions:
• After you have read the instructions, go to this site to begin work:
Please Right Click Here (to open job in a new window).
• Copy text exactly as it appears in the scanned image.
Payment:
• You will receive 10-cent reward for completing the first paragraph. You can
earn much more in bonus.
• When you complete the survey at the end, you will receive a completion
code in order to receive payment.
You MUST keep this window open in order to enter the completion code.
Bonuses will be paid after the HIT expires or after the work has been
completed.
Enter completion code here:
Appendix Section 2: First Page at External Host

Introduction
Task:
You will be presented with three (3) text paragraphs. Please enter the para-
graphs word for word in the text box below each paragraph, ignoring hyphen-
ation. For example, if a word is split over two lines, that is, “cup-cake,” type
“cupcake.” Once you have transcribed as many paragraphs as you would like,
hit “next,” leaving the text boxes blank—you will eventually get to the last
questions.
Payment:
You must complete at least three paragraphs to have your work accepted. A
sample paragraph is shown below. Note: Once you click “Next” you will not
be able to navigate to previous pages.
De jaarlijkse arbeid van elk volk is het fonds die oorspronkelijk levert hij met
alle benodigdheden en conveniencies van het leven die het jaarlijks verbruikt,
en die altijd bestaan, hetzij in de onmiddellijke produceren van die arbeid,
of in wat wordt gekocht met die van andere landen. Volgens dus, als deze
producten, of wat is gekocht met het, draagt een grotere of kleinere verhouding
tot het aantal van degenen die zijn om te consumeren, het volk zal beter
of slechter geleverd met alle de benodigdheden en conveniencies waarvoor
zij gelegenheid. Maar dit deel moet in elk volk worden geregeld door twee
verschillende
132 D. L. Chen
Appendix Section 3: Sixth Page at External

Host—Task to Check
Below is the transcription work of another Mechanical Turk Worker.
Please compare the scanned text and their answer and assess their work.
Consistently since Kihlberg the Supreme Court has upheld as conclusive on
both parties decisions made under finality clauses of government contracts
by applying to such contracts the same principles as are applied to contracts
between private persons. Decisions under finality clauses of government
contracts have been upheld, regardless of whether the decision was on a simple
question of fact, such as the mileage between two points, or a decision calling
for the application of expert knowledge and experience, such as an appraisal or
accounting determination, or a complicated mixed question of law and facts,
such as the interpretations of contract specifications. Before any appeal boards
were ever established, the Supreme Court upheld decisions under finality
clauses regardless of whether made on-the-spot by the government officer
directly involved in the dispute without any right of appeal or made by a higher
level review or appellate authority. The conclusive effect of such decisions
did not turn on the independence, impartiality or disinterestedness of the
government officer who made the decision, the opportunity of the contractor
to present evidence in support of his position, or what evidence the officer
making the decision had to support his decision.
Consistently since Kihlberg the Supreme Court has upheld as conclusive on
both parties decisions made under finality clauses of government contracts
by applying to such contracts the same principles as are applied to contracts
between private persons. Decisions under finality clauses of government
contracts have been upheld, regardless of whether the decision was on a simple
question of fact, such as the mileage between two points, or a decision calling
for the application of expert knowledge and experience, such as an appraisal or
accounting determination, or a complicated mixed question of law and facts,
such as the interpretation of contract specifications. Before any appeal boards
were ever established, the Supreme Court upheld decisions under finality
clauses regardless of whether made on-the-spot by the government officer
directly involved in the dispute without any right of appeal or made by a higher
level review or appellate authority. The conclusive effect of such decisions
did not turn on the independence, impartiality or disinterestedness of the
government officer who made the decision, the opportunity of the contractor
to present evidence in support of his position, or what evidence the officer
making the decision had to support his decision.
How many errors did you find?

How would you rate the quality of the other worker’s work?
Very low
Low
Fair
High
Very high
Appendix Section 4a: Seventh Page at External

Host—Sharing Bonus (Split—Treatment Group 1
(Control))
You and the worker whose work you have just graded will receive a bonus. You
will determine how the bonus is divided.
How much of a 50-cent bonus would you like to share with the other worker?
Appendix Section 4b: Seventh Page at External

Host—Sharing Bonus (Peer Average
Split—Treatment Group 2)
You and the worker whose work you have just graded will receive a bonus. For
accuracy, two other people have also graded the worker’s assignment. Each
grader will propose how to share a 50-cent bonus with the worker. We will
average your evaluations and distribute the bonus accordingly.
Appendix Section 4c: Seventh Page at External

Host—Sharing Bonus (Peer Random
You and the worker whose work you have just graded will receive a bonus. For
accuracy, two other people have also graded the worker’s assignment. Each
grader will propose how to share a 50-cent bonus with the worker. One of
your proposed splits will be chosen at random to be implemented.
134 D. L. Chen
Appendix Section 4d: Seventh Page at External

Host—Sharing Bonus (Computer Random
will propose a split of the 50-cent bonus. Your work helps us validate a new
computer grading program that is still rough. There is a 2/3 probability that
the program’s proposed split will be implemented instead of your proposal.
Appendix Section 4e: Seventh Page at External

Host—Sharing Bonus (Computer
Modification—Treatment Group 5)
will propose a split of the 50-cent bonus. Your work helps us validate a new
computer grading program that is still rough. The program will modify your
proposed split, which will then be implemented.
Appendix Section 5: Seventeenth Page at

External Host—Demographic Survey
What is your gender?
What is your age?
What country do you live in?
How much do you respect:
Your parents, the police, your boss/supervisor at work?
Not at all
Not much
Some
A little
A lot
What is your religion?
How often do you attend religious services? (answers may be approximate)

Never
Once a year
Once a month
Once a week
Multiple times a week
Please click on this link to get your completion code (it will open as a new
window):
Enter the code below AND on the Mechanical Turk website.
References
Amaranto, Daniel, Elliott Ash, Daniel L. Chen, Lisa Ren, and Caroline Roper. 2018.
Algorithms as Prosecutors: Lowering Rearrest Rates Without Disparate Impacts and
Identifying Defendant Characteristics ‘Noisy’ to Human Decision-makers, Technical
report.
Andreoni, James, and Lise Vesterlund. 2001. Which is the Fair Sex? Gender Differ-
ences in Altruism. The Quarterly Journal of Economics 116: 293–312.
Ash, Elliott, Daniel L. Chen, and Suresh Naidu. 2016. Ideas Have Consequences: The
Impact of Law and Economics on American Justice, Technical report.
Babcock, Linda, George Loewenstein, Samuel Issacharoff, and Colin Camerer. 1995.
Biased Judgments of Fairness in Bargaining. The American Economic Review 85:
1337–1343.
Barankay, Iwan. 2010. Rankings and Social Tournaments: Evidence from a Field
Experiment, Working Paper, University of Pennsylvania, Mimeo.
Barry, Nora, Laura Buchanan, Evelina Bakhturina, and Daniel L. Chen. 2016. Events
Unrelated to Crime Predict Criminal Sentence Length, Technical Report.
Batson, C. Daniel, Diane Kobrynowicz, Jessica L. Dinnerstein, Hannah C. Kampf,
and Angela D. Wilson. 1997. In a Very Different Voice: Unmasking Moral
Hypocrisy. Journal of Personality and Social Psychology 72: 1335.
Becker, Gary S. 1976. Altruism, Egoism, and Genetic Fitness: Economics and Socio-
biology. Journal of Economic Literature 14: 817–826.
Chen, Daniel, and Arnaud Philippe. 2017. Clash of Norms: Judicial Leniency on
Defendant Birthdays, Technical Report, Mimeo.
Chen, Daniel, Yosh Halberstam, and Alan Yu. 2017a. Covering: Mutable Character-
istics and Perceptions of Voice in the U.S. Supreme Court. Review of Economic
Studies invited to resubmit, TSE Working Paper No. 16–680.
Chen, Daniel, Yosh Halberstam, and Alan C.L. Yu. 2016a. Perceived Masculinity
Predicts U.S. Supreme Court Outcomes. PLOS One 11: 1–20, e0164324.
136 D. L. Chen
Chen, Daniel L. 2005. Gender Violence and the Price of Virginity: Theory and Evidence
of Incomplete Marriage Contracts, Working Paper, University of Chicago, Mimeo.
———. 2015a. Can Markets Overcome Repugnance? Muslim Trade Response to Anti-
Muhammad Cartoons, Working paper, ETH Zurich, Mimeo.
———. 2015b. Can Markets Stimulate Rights? On the Alienability of Legal Claims.
RAND Journal of Economics 46: 23–65.
———. 2016. Markets, Morality, and Economic Growth: Competition Affects Moral
Judgment, TSE Working Paper No. 16–692.
———. 2017a. Mood and the Malleability of Moral Reasoning, TSE Working Paper
No. 16–707.
———. 2017b. Priming Ideology: Why Presidential Elections Affect U.S. Judges.
Journal of Law and Economics 60(3): 479–496.
———. 2017c. The Deterrent Effect of the Death Penalty? Evidence from British
Commutations During World War I. American Economic Review. Resubmitted,
TSE Working Paper No. 16–706.
Chen, Daniel L. 2016. Implicit Egoism in Sentencing Decisions: First Letter Name Effects
with Randomly Assigned Defendants, Technical report.
Chen, Daniel L., and Jasmin K. Sethi. 2016. Insiders, Outsiders, and Involuntary
Unemployment: Sexual Harassment Exacerbates Gender Inequality. Invited to
resubmit, TSE Working Paper No. 16–687.
Chen, Daniel L., and Jess Eagel. 2017. Can Machine Learning Help Predict the
Outcome of Asylum Adjudications? In Proceedings of the ACM Conference on AI
and the Law.
Chen, Daniel L., and Jo Thori Lind. 2016. The Political Economy of Beliefs: Why Fiscal
and Social Conservatives/Liberals (Sometimes) Come Hand-in-Hand. Under review,
TSE Working Paper No. 16–722.
Chen, Daniel L., and John J. Horton. 2016. Are Online Labor Markets Spot
Markets for Tasks? A Field Experiment on the Behavioral Response to Wage Cuts.
Information Systems Research 27: 403–423, TSE Working Paper No. 16–675.
Chen, Daniel L., and Martin Schonger. 2016. Social Preferences or Sacred Values?
Theory and Evidence of Deontological Motivations. American Economic Journal:
Microeconomics. Invited to resubmit, TSE Working Paper No. 16–714.
———. 2017. A Theory of Experiments: Invariance of Equilibrium to the Strategy
Method of Elicitation, TSE Working Paper No. 16–724.
Chen, Daniel L., Martin Schonger, and Chris Wickens. 2016b. oTree—An Open-
Source Platform for Laboratory, Online, and Field Experiments. Journal of Behav-
ioral and Experimental Finance 9: 88–97.
Chen, Daniel L., Matt Dunn, H. Sirin and Levent Sagun. 2017. Early Predictability
of Asylum Court Decisions. In Proceedings of the ACM Conference on AI and the
Law.
Chen, Daniel L., and Susan Yeh. 2014. The Construction of Morals. Journal of
Economic Behavior and Organization 104: 84–105.
———. 2016a. Government Expropriation Increases Economic Growth and Racial

Inequality: Evidence from Eminent Domain. Economic Journal. Invited to resub-
mit, TSE Working Paper No. 16–693.
———. 2016b. How Do Rights Revolutions Occur? Free Speech and the First Amend-
ment, TSE Working Paper No. 16–705.
Chen, Daniel L., Tobias J. Moskowitz, and Kelly Shue. 2016c. Decision Making
Under the Gambler’s Fallacy: Evidence from Asylum Judges, Loan Officers, and
Baseball Umpires. The Quarterly Journal of Economics 131: 1181–1242.
Chen, Daniel L., Vardges Levonyan, and Susan Yeh. 2017b. Do Policies Affect
Preferences? Evidence from Random Variation in Abortion Jurisprudence. Journal
of Political Economy, TSE Working Paper No. 16–723. Under review.
Chen, Daniel L., Xing Cui, Lanyu Shang, and Junchao Zheng. 2016d, Forthcoming.
What Matters: Agreement Among U.S. Courts of Appeals Judges. Journal of
Machine Learning Research. TSE Working Paper No. 16–747.
Coffman, Lucas C. 2011. Intermediation Reduces Punishment (and Reward). Ameri-
can Economic Journal: Microeconomics 3: 77–106.
Dana, Jason, Roberto A. Weber, and Jason Xi Kuang. 2007. Exploiting Moral Wiggle
Room: Experiments Demonstrating an Illusory Preference for Fairness. Economic
Theory 33: 67–80. Symposium on Behavioral Game Theory.
Darley, John M., and Bibb Latané. 1968. Bystander Intervention in Emergencies:
Diffusion of Responsibility. Journal of Personality and Social Psychology 8: 377–383.
Detert, James R., Linda Klebe Trevino, and L. Vicki Sweitze. 2008. Moral Disengage-
ment in Ethical Decision Making: A Study of Antecedents and Outcomes. Journal
of Applied Psychology 93: 374–391.
Dubinsky, Alan J., and Thomas N. Ingram. 1984. Correlates of Salespeople’s Ethical
Conflict: An Exploratory Investigation. Journal of Business Ethics 3: 343–353.
Eckel, Catherine C., and Philip J. Grossman. 1996. Altruism in Anonymous Dictator
Games. Games and Economic Behavior 16: 181–191.
———. 2008. Differences in the Economic Decisions of Men and Women: Experi-
mental Evidence. In Handbook of Experimental Economics Results, vol. 1, Chap. 57,
ed. C. Plott and V. Smith. Amsterdam: Elsevier, 509–519.
Eren, Ozkan, and Naci Mocan. 2016. Emotional Judges and Unlucky Juveniles, Working
Paper.
Eriksson, Tor, and Marie-Claire Villeval. 2008. Other-Regarding Preferences and
Performance Pay–An Experiment on Incentives and Sorting. Journal of Economic
Behavior and Organization 68: 412–421.
Fletcher, Thomas D., and David N. Nusbaum. 2008. Trait Competitiveness as a
Composite Variable: Linkages with Facets of the Big-Five. Personality and Indi-
vidual Differences 45: 312–317.
Ford, Robert C., and Woodrow D. Richardson. 1994. Ethical Decision Making: A
Review of the Empirical Literature. Journal of Business Ethics 13: 205–221.
Harrison, Glenn W., and John A. List. 2004. Field Experiments. Journal of Economic
138 D. L. Chen
Henrich, Joseph, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr, Herbert
Gintis, and Richard McElreath. 2001. In Search of Homo Economicus: Behavioral
Experiments in 15 Small-Scale Societies. The American Economic Review 91: 73–78.
Judge, Kathryn. 2012. Fragmentation Nodes: A Study in Financial Innovation,
Complexity, and Systemic Risk. Stanford Law Review 64: 657–726.
Kant, Immanuel. 1797. Über ein vermeintes recht aus menschenliebe zu lügen.
Berlinische Blätter 1: 301–314.
Katz, Lawrence F., and Claudia Goldin. 2000. The Power of the Pill: Oral Contracep-
tives and Women’s Career and Marriage Decisions. The Journal of Political Economy
110: 730–770.
Kohn, Alfie. 1986. No Contest: The Case Against Competition. Boston: Houghton
Mifflin.
Orne, Martin T. 1962. On the Social Psychology of the Psychological Experiment:
With Particular Reference to Demand Characteristics and Their Implications.
American Psychologist 17: 776–783.
Paolacci, Gabriele, Jesse Chandler, and Panagiotis G. Ipeirotis. 2010. Running Exper-
iments on Amazon Mechanical Turk. Judgment and Decision Making 5: 411–419.
Polanyi, Karl. 1944. The Great Transformation: The Political and Economic Origins of
Our Time. Beacon Paperbacks. Boston: Beacon Press.
Radin, Margaret Jane. 1987. Market-Inalienability. Harvard Law Review 100: 1849–
1937.
Reips, Ulf-Dietrich. 2001. The Web Experimental Psychology Lab: Five Years of Data
Collection on the Internet. Behavior Research Methods, Instruments, and Computers
33: 201–211.
Ruffle, Bradley J. 1998. More Is Better, But Fair Is Fair: Tipping in Dictator and
Ultimatum Games. Games and Economic Behavior 23: 247–265.
Shaw, Aaron D., John J. Horton, and Daniel L. Chen. 2011. Designing Incentives for
Inexpert Human Raters. In Proceedings of the ACM 2011 Conference on Computer
Supported Cooperative Work, CSCW ’11. New York: ACM, 275–284.
Shleifer, Andrei. 2004. Does Competition Destroy Ethical Behavior? The American
Economic Review 94: 414–418.
Titchener, James L. 1967. Experimenter Effects in Behavioral Research. Archives of
Internal Medicine 120: 753–755.
Part II
Empirical Advances in the Economics
of Religion
9
Religion and Demography
Sriya Iyer
Religion and demography is a contested space. Every country in the world

is interested in how its demographic numbers are changing over time, and
yet they are acutely conscious of the rise of religion (Iyer 2016; Lehrer 2009).
Whether there were early concerns about a ‘population bomb’ or worries about
‘minority fertility rising’, the subject of religion and fertility has not gone away
and indeed seems to reappear in different periods of world history repeatedly.
In this chapter, I discuss why we should be interested in the question of
religion and demography. I provide an overview of recent trends in the world’s
religious demography and then locate this within the historical context of
what major world religions continue to think theologically about factors that
affect demographic decisions. The focus is on the various mechanisms through
which religion might affect fertility behaviour. I then distinguish between
the various channels through which religion also affects mortality, especially
through its effect on physical and mental health behaviours. I put forward the
view that religion continues to have important consequences for fertility and
mortality decisions which collectively affect a country’s religious demography
overall.
S. Iyer ()
Faculty of Economics, University of Cambridge, Cambridge, UK
e-mail: sriya.iyer@econ.cam.ac.uk

142 S. Iyer
Historical and Contemporary Trends in Religious

Demography
The original interest in religion and demography stemmed from Catholic-
Protestant fertility differentials in the USA in the 1960s and 1970s (Westoff
and Jones 1979; Sander 1995). This was sparked by early studies such as the
Princeton Fertility Project, which first suggested the influence of diffusion
effects and highlighted some cultural aspects of fertility decline (McQuillan
2004). However, as the Christian demographic differences between groups in
the USA became less pronounced over time, scholars too were less interested
in the subject of religion and demography in Western societies. In recent
decades though, we have seen a shift back to interest in this topic, with many
more studies of religion and demography and more interest now than ever
before especially in fertility patterns in the Muslim world both in North Africa
and the Middle East generally, and for Hindu populations in South Asia
(Goldscheider 1971; Obermeyer 1992; Iyer 2002).
Trends in religious demography based on census and survey data on religious
affiliation and religious beliefs are being conducted across the world. A 2012
Pew Research Center study of over 230 countries and 2500 censuses and
surveys showed that 84% of the world’s population, approximately 5.8 billion
people, currently report a religious affiliation. This is shown more clearly in
Fig. 9.1.
Other
Religions** Folk
Jews 0.8% Religionists*
0.2% 5.9%
Buddhists
7.1%
Christians
31.5%
Hindus
15.0%
Unaffiliated
Muslims 16.3%
23.2%
Fig. 9.1 Size of major religious groups in 2010—Percentage of the global population.
*Includes followers of African traditional religions, Chinese folk religions, Native
American religions and Australian aboriginal religions. **Includes Bahai’s, Jains, Sikhs,
Shintoists, Taoists, followers of Tenrikyo, Wiccans, Zoroastrians and many other faiths.
Source: Pew Research Center’s Forum on Religion & Public Life, Global Religious
Landscape, December 2012. Also reproduced in Iyer (2016), p. 402
The world’s major religious groups are Christians (32%), Muslims (23%)
and Hindus (15%). What is most striking though is that there are 16% of
respondents who describe themselves as ‘religiously unaffiliated’. This is not
to suggest that they do not believe in a higher power or are atheists, simply
that they do not affiliate themselves to any particular religious group. Lutz
and Skirbekk (2012) carried out a study of the world’s global population
in which they report that about a hundred years ago, the distribution of
the world’s population in 1910 was 34.8% Christian, 12.6% Muslim, 7.8%
Buddhist and 12.7% Hindu. If one were to compare these two studies, then
the headline message is that the biggest change from 1910 to the present day is
the rapid growth of Islam worldwide, the rise of the religiously unaffiliated
especially in the Western world and the growth of Christianity in Africa.
Geographically, the current distribution of the world’s religious population
is varied. The region that has experienced the most significant change over
the last century is Sub-Saharan Africa, which is now 60% Christian, from a
level of 9% Christians in 1900 compared to African religions which were 76%
in 1900 and 13% in Sub-Saharan Africa in 2012 (Lutz and Skirbekk 2012).
These changes to the world’s religious demography over the last century are
unprecedented historically and very striking indeed.
There is also some evidence to suggest that the world’s religious population
is increasingly youthful, which implies that religion will continue to be
important for many decades to come. Figure 9.2 depicts this as it illustrates
the median age of the world’s religious population, broken down by religious
group. Figure 9.2 shows that the median age for the world’s population overall
is 28 years. The median age for Muslims is 23 years, it is 26 years for Hindus,
30 years for Christians, and for Buddhists and Jews it is much higher at 34 years
and 36 years, respectively. This again seems to suggest that the Muslim and
Hindu populations, on average, are much younger than other religious groups,
which of course may also have implications for the future evolution of these
religions, demographically speaking, and their role in societies worldwide.
Explaining the Relationship Between Religion

and Demography
Given the distribution of the world’s religious population, why should we
be interested in the question of religion and demography? The interest in
religion and demography is not just because of this relationship alone but also
because of how these factors interact with other factors thought important for
144 S. Iyer
Global Median 28
Jews 36
Buddhists 34
Unaﬃliated 34
Folk Religionists 33
Other Religions 32
Chrisan 30
Hindu 26
Muslim 23
0 5 10 15 20 25 30 35 40
Fig. 9.2 Median age of religious group in 2010. Notes: (1) Folk Religionists includes
followers of African traditional religions, Chinese folk religions, Native American
religions and Australian aboriginal religions. (2) Other Religions includes Bahai’s, Jains,
Sikhs, Shintoists, Taoists, followers of Tenrikyo, Wiccans, Zoroastrians and many other
faiths. (3) Percentage may not add to 100 due to rounding. Source: Adapted from the
Pew Research Center’s Forum on Religion and Public Life—Global Religious Landscape,
December 2012
economic development. For instance, another key relationship of interest is

how religion and demography interact with gender and education. In some
cases, it is not that religion has a direct theological effect on fertility, but that
it does reinforce low status for women in some religions, and this then feeds
through to higher fertility rates because of factors such as lower education
levels among women compared to men (Obermeyer 1992; Iyer 2002). This
is a channel through which religion might affect demographic outcomes.
This is more difficult to address as economists, as we may not be able to
separate out easily the effect of gender from the effect of education and/or
religious theology on demographic behaviour. One way to address this is to
come up with more methodologically innovative ways of examining religion
and demography for example through new instrumental variable strategies,
difference-in-difference approaches, or natural experiments, which would be
important methodological areas for future research in this field. In addition
to education, the relationship between religion and demography is affected by
other characteristics such as women’s labour force participation. But it does
suggest that the relationship between these factors is not simple and that more
research is sorely needed on their interactions.
In terms of mechanisms, there are many hypotheses that explain the

relationship between religion and fertility. Chamie (1977: 365–382) initially
articulated these—first, the ‘particularised theology’ hypothesis, that religion
affects fertility irrespective of socio-economic and demographic contexts.
The second is the ‘characteristics’ hypothesis which suggests that the socio-
economic characteristics of individual members of religious groups might be
salient. The third is the ‘minority group status’ hypothesis, which suggests
that the insecurity faced by minority religious groups increases their fertility
compared to the majority group. Of course, minority fertility can also go either
way—it can be higher to increase numerical preponderance or it can be lower
if socio-economic assimilation is easier to manage with lower fertility.
The particularised theology argument stems from the values that religions
transmit especially with respect to the acceptability of birth control and norms
about family size. Religious rituals often provide distinct roles for children (as
is the case with religious roles for sons in South Asia). Religions can impose
religious vows and practices of celibacy, either lifelong or outside marriage.
Religions may take positions on the acceptability of contraception. And some
religions may encourage literacy in order to read the scriptures, which in turn
may lead to indirect effects on fertility.
Empirical studies show that Catholics show different fertility than Protes-
tants and that this may be due to differences in the content of their religious
beliefs, although this too is debated (Westoff and Jones 1979). Catholic norms
on contraception and family size are believed to influence fertility. Similarly, it
is sometimes argued that the particular philosophical content of Islam affects
demographic behaviour (Obermeyer 1992).
Islam and Hinduism

Since the rise of Islam is one of the main trends in religious demography
worldwide, the main issue of interest today concerns Muslim fertility, so it
is worth discussing the relationship between Islam and fertility in some depth.
Empirically, most studies of Islam and fertility concern North Africa and
West Asia, with a focus on countries such as Egypt, Sudan, Morocco, Algeria,
Iraq, Saudi Arabia, Syria and Yemen. The connection between Islam and
fertility is a research area that is fraught with difficulty mainly because there
are many differing perspectives on this matter, based on different schools of
Islamic jurisprudence. Within Islam, the values associated with the religion
are important, along with Koranic prescriptions (Gellner 1981). The Sharia
or Islamic law, derived from the Koran, and the Sunna (the interpretations
146 S. Iyer
of the sayings of the Prophet Muhammad) govern family relations (Ragab

1980). The Sharia consists of interpretations of Muslim scholarship by the
ulema (Muslim clerics) that both encompass the Koran, the Sunna and their
formal codification which took place until the eleventh century.
From these sources, it is clear that the content of religion in Islam may affect
Muslim women’s fertility, despite general injunctions in the religion which do
emphasise that wives should be treated fairly and equitably and that divorce
should be a considered decision (Coulson and Hinchcliffe 1978). There are
three channels though which Islam is hypothesised to affect fertility directly—
first, through its stance on the acceptability of birth control; second, by its
views on marriage and polygyny; and third, by its views on divorce (Qureshi
1980). One commonly held argument is that Islam permits polygyny, does
not discourage high fertility and permits unilateral divorce. Collectively, this
would encourage high fertility in Muslim societies compared to non-Muslim
ones.
From a scholarly perspective, Koranic prescriptions on marriage do encour-
age marriage, as well as for the widowed and the divorced (Youssef 1978). The
legal age at marriage, as with other religions, varies in most Muslim countries,
although marriages are still contracted between those below the minimum age,
and these are also considered legal. As with Hinduism, a feature of the Islamic
marriage contract is the mahr, sadaq or dower, which is paid to the bride, and
is a woman’s formal right to property. Although this also is changing over
time, parents and guardians may exercise control over the choice of marriage
partners and there may be seclusion restrictions both before and after marriage
(Youssef 1978).
Both polygyny and divorce are a contentious and much-debated area
of Islamic religious discourse. Traditionally, the Muslim husband does not
need to obtain permission to contract another marriage, although in practice
permission from first wives may be sought and may be written into the
marriage contract. While taking more than one wife is not actively encouraged
by the Islamic religion, this is subject to interpretation by the man’s own
conscience. In Pakistan, for example, men need to at least obtain permission
from a special arbitration council, so this protects the first wife’s interests. In
practical terms, despite the theology, there is not much empirical evidence
that polygyny is widespread in Muslim societies. But there is evidence that
the provision for unilateral divorce does exist in Muslim societies and that
this provision resides with men. There have also been issues in some countries
about the provisions for alimony that a woman might have rights to. A further
issue is that a divorced or widowed woman retains custody of her children
only for a time-limited period after the divorce and then legally custody passes
over to a male relative. This again makes the Muslim woman’s position under
Islamic law extremely vulnerable, compared with women of other religions.
Muslim law also lists prescriptions about birth control, as in the case
of Catholicism. While the popular perception is that birth control is not
permitted in any situation, which is true of the Maliki school of thought,
the reality again is a mixed picture. Obermeyer has suggested that how we
think about Islam and birth control does depend to a large extent on the
interpretation of the different schools of Islamic jurisprudence (Iyer 2002). As
also discussed there, Obermeyer argues that Sunni and Shia positions on birth
control originate from Al-Ghazali, who argues that there are many contexts in
which birth control is permitted, including “if one of the partners is afflicted
with a disease which may be passed on to children; or if there is concern for
the effect of too many pregnancies on the wife’s health” (Obermeyer 1992:
43). Abortion is a more contested issue, with most schools allowing abortion
until a foetus is considered ‘ensouled’—this may be 40th, 80th or 120th
day of pregnancy after which abortion is prohibited in all cases and schools
(Obermeyer 1992). This position is reinforced by a Muslim woman’s position
in Islamic law, which also includes rules about gender segregation. Coulson
and Hinchcliffe (1978) pointed out first that “Islamic law has continued to
reflect the patriarchal and patrilineal nature of a society based on the male
agnatic tie. Within the scheme of family law which developed in this way,
woman, whether as daughter, wife, or mother, occupied an inferior position”
(Coulson and Hinchcliffe 1978: 38). High fertility is also encouraged by
direct pronatalist tendencies in Islam because women achieve status with
motherhood, even if they are divorced or overlooked with the arrival of a
second wife. Collectively, many of these laws, prescriptions and proscriptions
may encourage traditional values and pronatalism in Islam because these
provisions encourage lower autonomy for women. However, there also appears
to be a great deal of variation with respect to Islamic jurisprudence. And
whether this effect is distinct to other religions such as Christianity and
Hinduism is highly debatable.
As discussed earlier, among the main trends in contemporary religious
demography is the growth of the Hindu population in South Asia. This too
stems from Hindu religious norms of various kinds. In the case of Hinduism,
a variety of sources put forward Hindu beliefs and the associated code of
conduct. One aspect of this is the ordering of society and caste endogamy
which influences Hindu demographic behaviour. Marriage is endorsed widely
in Hinduism. Polygyny is sanctioned theologically in that most Hindu gods are
depicted with multiple consorts, but for mere mortals monogamous marriages
are repeatedly idealised in the scriptures. Although divorce is permitted in
148 S. Iyer
limited circumstances, it is discouraged by the Hindu religion overall, reflected

in low overall rates of divorce in South Asia in general, compared to other
parts of the world. There is little in Hinduism about birth limitation or
abortion, which has been attributed to notions of purity and pollution in
the religion (see Iyer 2002 for a more detailed discussion of this matter). The
control of births is addressed only indirectly in the context of norms about
abstinence. Women in the family are considered very important, establishing
a household is considered a fundamental duty, and the wife is central to it.
Iyer (2002) argues that compared with Islam, where the role for women is
mainly to ‘uphold family honour’, the role for women in Hinduism is much
more ‘equal in every respect’, although this is really only in the domestic
sphere. Consequently, this lack of autonomy affects women’s fertility. This
is reinforced by a greater role for sons compared to daughters in the Hindu
home, and especially as a source of old-age support, which leads to skewed
birth outcomes and which provides a religious motive for high fertility (and
high female mortality at birth) in South Asian and other societies which have
Hindu populations.
For the question of religion and fertility, neither Hinduism nor Islam
theologically assigns an independent role for women outside the family. They
both encourage the importance of marriage, children and especially sons. The
religions differ most on birth control and abortion. As Iyer (2002) also argues,
Hinduism does not explicitly express any scriptural opinion for or against
birth control or abortion. In contrast, schools of Islamic thought suggest
birth control is permissible but only in restricted situations, leaving room for
many alternative interpretations by individuals in real-life situations. This is
also influenced by the institutions through which these religions operate: for
example, in matters relating to birth control or abortion, the Muslim clerics are
in a position to interpret Islam for followers, whereas in the case of Hinduism,
the lack of explicit scriptural injunctions may mean that Hindu priests are not
consulted on demographic matters as much.
This part of the paper reviewed evidence for the particularised theology
hypothesis in the context of a couple of major world religions, given the
increased size of Muslim and Hindu populations worldwide, compared to
a century ago. A second hypothesis that links religion and fertility is the
characteristics hypothesis, which I turn to now.
The Characteristics Hypothesis

Religious theology is one channel through which religion has an impact on
demography, but many economists argue that the socio-economic character-
istics of individual members of a religious group may be as important, or
even more important, in explaining religious differences in fertility behaviour.
So religion may also affect fertility through the ‘characteristics hypothesis’. If
fertility differentials between populations reflect socio-economic differences
between members of different religious groups, then the fertility of one
religious community may differ significantly from that of another because
of land ownership, wealth status or educational characteristics. Both Iyer
(2002) and Lutz and Skirbekk (2012) show that there are important fertility
differences within religions, for those who have high education compared to
a low education, and that this is especially true among Hindus, Muslims and
Christians in India, so that highly educated women, regardless of religion, have
significantly lower fertility than their less educated contemporaries. This is also
true of many Muslim-majority countries. The characteristics effect could also
affect other factors such as religious discrimination in access to health services.
In addition, there could be an interaction between the theology and
characteristics channels if say low autonomy for women translates into lower
educational opportunities for them relative to men. Another hypothesis that
links religion and demography is minority group status, which can either
increase or reduce fertility depending on how it interacts with other factors.
For example, if a minority community feels threatened by the majority
community, then this can induce pronatalist tendencies (Van Heek 1966). This
could also work through differential access to education for those belonging to
different religions. Studies on the USA, Philippines and Sri Lanka support the
minority group hypothesis. If we examine the balance of the evidence on the
theology hypothesis and the characteristics hypothesis, then there was a large
demographic literature which documented the divergence between 1945 and
the 1960s and then the convergence in the 1970s in Catholic and Protestant
fertility in the USA (Sander 1995), which attributed that convergence to char-
acteristics that also converged over time. Education, income and urban/rural
residence have also been suggested as leading causes for Catholic-Protestant
differentials in Europe (Galloway et al. 1994).
The ‘particularised theology’ hypothesis, the ‘characteristics’ hypothesis and
the ‘minority group status’ hypothesis are three different mechanisms for
thinking about how religion might influence demography. In many societies,
these three mechanisms are not mutually exclusive; indeed at different times
150 S. Iyer
in the history of religious communities, some channels may work better, or

be more salient, than others. The theology effect influences norm and values.
The characteristics effect may be influenced by economic policy more widely.
Eventually, there could well be convergence in fertility rates, at least among
those with similar socio-economic status, and we have seen this in many
Muslim-majority countries around the world. So it is worthwhile bearing
in mind that how religion affects demography is also influenced by the
institutions in a country, whether it is the state, the power of the church or
other religious institutions. The social organisations of religious groups and
the social status of religious communities then become very important here
and mediate the relationship between religion and demography.
Religion, Mortality and Health

Demographic trends are influenced both by fertility and by mortality pat-
terns. Compared to the research which has been devoted to religion and
fertility, there is comparatively less scholarship on the links between religion
and mortality. Again, socio-economic characteristics may explain religious
differences in mortality, but in addition, religion can affect mortality rates due
to attendance at religious services and how religion impacts on individuals’
health behaviours (Sullivan 2010). This suggests that the links between religion
and mortality may be quite complex. In a study of the USA, for example,
Sullivan (2010) argues that whites may constitute a higher proportion of
Mainline Protestant and Jewish faiths in the USA while Catholics may make
up a greater proportion of Latin American immigrants than other religions.
Immigrants tend to have lower mortality than the local population of the
USA (Sullivan 2010). He also shows that the religiously unaffiliated and Black
Protestants have higher proportions of those who are never-married than other
religious groups. This study also shows that Jews and Mainline Protestants
have better income and education compared to Evangelical Protestants. The
case of the USA here is illustrative of a single country only, but such variation
is commonplace in many religiously diverse societies.
A second channel through which religion affects mortality is through its
effect on health behaviours. This is a channel which needs further research,
and this is true for both research on physical health and mental health.
For example, recent work has shown that religion affects mental health
especially when dealing with depression in adolescents in the USA (Fruehwirth
et al. 2019), and religion may also affect physical health behaviours such
as alcohol consumption and smoking patterns (Fletcher and Kumar 2014).
Those who attend religious services, for example, may demonstrate better
health behaviours than those who do not, although the evidence on this
is mixed (Sullivan 2010). Then there is the question of why religion might
matter here—either because of theological prescriptions and proscriptions or
because of the effect of peer behaviours of those who belong to the same
religious denomination. This effect may also interact with socio-economic
characteristics, which suggests that much more research does need to be
done on how religion affects mortality either due to its effect on individuals
directly through religion’s value-enforcing channels or through peers through
its norm-enforcing channels. The limited evidence that does exist suggests
that religiously driven peer effects here are likely to be strong and may
be even stronger than the influence of socio-economic characteristics alone
(Fruehwirth et al. 2019).
Conclusion
In summary, religion affects demography through its effects on fertility and
mortality. In terms of fertility, this can happen through particularised theology,
characteristics or minority group status. For mortality, religion can have an
impact through socio-economic characteristics or through its effects on health
behaviours. This may affect religious denominations differently as the effect
may work both directly on individuals psychologically and/or through social
context with the influence of peer groups. For economists, identifying these
effects statistically will always be a challenge, but probably very important
to do, in order to demonstrate a first-order effect of religion on fertility and
mortality. Presently, the world is getting more religious, but as rich countries
are getting more secular, with large proportions of the religiously unaffiliated
rising day by day, poorer countries are becoming increasingly more religious.
This will have long-term effects on the religious demography of these societies.
If the trends of the last century are anything to go by, Christianity has increased
rapidly in Africa and Hinduism in South Asia, and Islam has increased its
influence worldwide. All of this has future implications for the evolution of
the world’s demography, both fertility and mortality trends. Going forward,
religion and demography will continue to be a much-contested space.
152 S. Iyer
References
Chamie, J. 1977. Religious Differentials in Fertility: Lebanon, 1971. Population Studies
31 (2): 365–382.
Coulson, N., and D. Hinchcliffe. 1978. Women and Law Reform in Contemporary
Islam. In Women in the Muslim World, ed. L. Beck and N. Keddie, 37–49.
Cambridge: Harvard University Press.
Fletcher, J., and S. Kumar. 2014. Religion and Risky Health Behaviors among U.S.
Adolescents and Adults. Journal of Economic Behavior & Organization 104 (C):
123–140.
Fruehwirth, J., S. Iyer, and A. Zhang. 2019. Religion and Depression in Adolescence.
Journal of Political Economy 127 (3).
Galloway, P.R., E.A. Hammel, and R.D. Lee. 1994. Fertility Decline in Prussia, 1875–
1910: A Pooled Cross-Section Time Series Analysis. Population Studies 48 (1): 135–
158.
Gellner, E. 1981. Muslim Society. Cambridge: Cambridge University Press.
Goldscheider, C. 1971. Population, Modernization, and Social Structure. Boston: Little,
Brown and Company.
Iyer, S. 2002. Demography and Religion in India. Delhi: Oxford University Press.
———. 2016. The New Economics of Religion. Journal of Economic Literature 54
(2): 395–441.
Lehrer, E.L. 2009. Religion, Economics and Demography: The Effects of Religion on
Education, Work, and the Family. Routledge Press.
Lutz, W., and V. Skirbekk. 2012. Universal Rights in a World of Diversity: The Case
of Religious Freedom. Pontifical Academy of Social Sciences, Acta 17, 2012.
McQuillan, K. 2004. When Does Religion Influence Fertility? Population and Devel-
opment Review 30 (1): 25–56.
Obermeyer, C.M. 1992. Islam, Women and Politics: The Demography of Arab
Countries. Population and Development Review 18 (1): 33–60.
Pew Research Center. 2012. Forum on Religion and Public Life. Report on the Global
Religious Landscape, December 2012.
Qureshi, S. 1980. Islam and Development: The Zia Regime in Pakistan. World
Development 8: 563–575.
Ragab, I.A. 1980. Islam and Development. World Development 8: 513–521.
Sander, W. 1995. The Catholic Family. Boulder: Westview Press.
Sullivan, A.R. 2010, December. Mortality Differentials and Religion in the U.S.:
Religious Affiliation and Attendance. Journal for the Scientific Study of Religion 49
(4): 740–753.
Van Heek, F. 1966. Roman Catholicism and Fertility in the Netherlands: Demo-
graphic Aspects of Minority Status. Population Studies 20: 125–138.
Westoff, C.F., and E.F. Jones. 1979. The End of “Catholic Fertility”. Demography 16
(2): 209–217.
Youssef, N.H. 1978. The Status and Fertility Patterns of Muslim Women. In Women
in the Muslim World, ed. L. Beck and N. Keddie, 69–99. Cambridge: Harvard
University Press.
10
Economics and Church State Research: Past,
Present, Future
Daniel Hungerman and Timothy Weninger
Introduction
This chapter concerns scholarship on church/state relations, an area of long-
standing interdisciplinary interest. We will discuss past work, some especially
interesting recent work, and some recent work looking at especially timely
policies, such as school vouchers, and especially noteworthy new data, such as
new data based on weekly church bulletins. We will focus on research in the
discipline of economics, where work on this topic goes back several centuries.
In fact, in the seminal The Wealth of Nations, Adam Smith himself considers
the appropriate relation between the church and the state at some length.
Writing about the difference between a society with a state-supported church
and one without, Smith argues:
In a country where the law favored the teachers of no religion more than those of
another, it would not be necessary that any of them should have any particular
This work was supported by a FRSPI grant from the University of Notre Dame. We thank Chungeun Yoon
for excellent research assistance, along with Kathleen Ryan and Eric Fein for terrific help in assembling
our bulletins data. Mackenzie Jones and Mitchell Murphy also provided helpful assistance. The project
was developed while Professor Hungerman was working on a project funded by the John Templeton
Foundation, but this work was not funded by Templeton.
D. Hungerman () · T. Weninger
University of Notre Dame, Notre Dame, IN, USA
e-mail: dhungerm@nd.edu

156 D. Hungerman and T. Weninger
or immediate dependency upon the sovereign or executive power, or that he

should have anything to do, either in appointing, or in dismissing them from
their offices. In such a situation he would have no occasion to give himself any
concern about them. Smith (1994, Ch 1, PT 3, p. 856)
For the present purposes, a key observation on the above passage is that it
focuses on a setting where a state-supported church is possible and something
practical to consider. But there are many places today where there is no state
church and there is not going to be a state church anytime soon. In some
countries, such as the United States, the government is in fact prohibited
from making laws regarding the establishment of religion altogether. But this
observation marks the beginning of the story, rather than the end.
Beyond State-Supported Religion

Certainly, work on the direct state sponsorship (or, alternately, persecution)
of religion remains quite worthwhile (e.g., Barro and McCleary 2005; or
Anderson et al. 2017). But we argue that church/state work is also relevant
in places without any possibility of a state-sponsored church, and indeed
that in recent years this type of church/state work has produced a number
of particularly relevant and exciting insights.
How can church/state issues arise when the government cannot make
laws respecting religion? One straightforward answer is that religious groups
interact with secular forces, and these secular forces are themselves often the
subject of regulation or other policy intervention. Government intervention
into even purely secular topics can thus, through this circuitous channel, affect
religious practice. Scholarship looking at this circuitous channel is important
for several reasons; here we focus on two.
First, work of this nature often highlights important and sometimes quite
unexpected policy implications. Most economists are (or at least claim to be)
interested in policy, so that this channel can lead to work interesting to scholars
and to policy makers. Second, work of this nature can also enable us to learn
things about religion that would be very hard to learn otherwise. We discuss
each of these two issues in the next section.
Unexpected Policy Implications: The Case

of Vouchers
As an example of how work in this area can generate unexpected policy
effects, consider school vouchers in the United States. Like many countries,
the United States distinguishes between public and private education, and
like many countries, religious groups play an important and well-defined role
in the education market. Finally, like many countries, the United States has
undergone a school-choice-based revolution in how education is financed.1
Vouchers and other voucher-style programs are a prominent example of
these choice-style programs. In the United States, the public education system
is entirely secular, and the private education system is dominated by religious
schools. (According to the Private School Survey, more than four out of
five private school students in the United States in the 2011/2012 school
year attended a religiously affiliated school.) Private schools have traditionally
received no funding from the government. Vouchers have changed this,
by allowing government funds to subsidize attendance at private schools,
including religious ones. Here, we will use the term vouchers to refer to private
school subsidies directly paid by a voucher along with similar programs that
provide such a subsidy through the tax code (e.g., through an income-tax
credit). Vouchers are currently state- or city-run programs, so that the details,
restrictions, and size of vouchers can vary. Many voucher programs provide
reasonably large subsidies (such as several thousand dollars per student), and
subsidies are often (but not always) restricted to a certain type of student (such
as a student from a low-income family, or a student initially in public school).
A large literature has developed to study the impact of vouchers, largely
focused on how vouchers affect educational outcomes.2 This is certainly an
important and worthwhile topic. But very little attention has been given to
the fact that vouchers subsidize schools that are often run by churches, and so
vouchers could transform the financial landscape for churches in the United
States.3
1 For an international discussion of school-choice policies, see OECD (2012).

2 See Figlio (2009) and Hoxby (2003) for discussion of work in this area.
3 Like any subsidy, it is possible that in fact the voucher here does not mainly benefit churches but instead
would benefit the families using the voucher—this is the standard issue of incidence one discusses with
subsidies and taxes in principles of microeconomics. Hungerman and Rinz (2016) explore this issue and
show that vouchers generate large revenue gains for schools (and thus potentially the churches that run
then), but interestingly, the incidence of subsidies can vary greatly depending on whether subsidies are
restricted to certain students or available for all students.
In a recent working paper (Hungerman et al. 2017a), some of the authors

here partnered with the Catholic Church in the city of Milwaukee, Wisconsin,
to study this issue. We gained access to the financial data not just of Catholic
schools in the Milwaukee area but also the data of the churches running those
schools. Milwaukee is home to one of the most famous voucher programs
in the United States and a program that has received significant academic
attention as well (e.g., Rouse 1998). We then explored the role of vouchers in
church finances and what happens to church finances when vouchers expand.
There were two major findings of our study. The first did not even require
running a single regression. Looking at basic means, the average church in
Milwaukee running a voucher-accepting school now gets more money from
vouchers—that is, the government—than from any other source, including
its own worshippers. Our data is for Milwaukee but the implications of this
result go far beyond that city. Voucher programs have proliferated across the
country, with billions of dollars subsidizing millions of students in recent years.
If someone 15 years ago had told us that in the near future there would soon be
hundreds, perhaps thousands, of churches getting most of their revenue from
the government, we would have found that ridiculous. But that appears to be
reality today.
The implications of this for the future of American religion should not
be underestimated. The United States is currently experiencing a decline in
organized—and, also, not-organized—religious practice. It might have been
possible a few years ago to dismiss this as specific to mainline protestants,
or to denominational churches more generally, or to dismiss the well-known
rise of the “nones” (Hout and Fisher 2002) as pertaining to outward practice
rather than a change in inner spirituality. We do not think these explanations
bear up against modern data, however. As just one example, consider Fig.
10.1, the fraction of respondents in the General Social Survey (a nationally
representative US survey) who say they never pray. In just the last 10 years, the
fraction has more than doubled, and in 20 years it has increased (from a low
base) by an order of magnitude. Perhaps one could squint and see a “leveling
out” of the trend in recent years, or (to drag out the excuse of the nones) argue
that lots of people never prayed, they only just became comfortable admitting
it recently. The issue is tangential here and we will not take it up in depth,
but suffice it to say we do not think such stories can explain the widespread
declines observed in modern American society or even in one’s casual day-
to-day experience in the country. In America, most arrows point down for
religion.
Vouchers, however, are an arrow that points up. As churches struggle to
stay open and attract new families, vouchers may not only support churches
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
2000 2002 2004 2006 2008 2010 2012 2014 2016
Fig. 10.1 Fraction of Americans who never pray. Source: General Social Survey
that currently run voucher schools but may encourage churches to open new
schools. Vouchers could significantly affect the organization of congregations
across the United States. Given current market forces, it is hard to imagine
otherwise, in fact. Of course, we cannot say what the coming decades hold,
and, for example, the constitutionality of vouchers may be called into question
in the United States. To use a phrase sometimes attributed to Niels Bohr,
prediction is very difficult, especially when it concerns the future. That said,
the future of American religion in the coming decades may be vouchers, and
the future of vouchers in America may have nothing to do with education but
everything to do with religion.
However, a second finding of the Hungerman, Rinz and Frymak study is
that an increase in financial revenue from the government may not lead to
greater religious activity in churches (as measured by, e.g., offertory donations).
The study does not find strong evidence that more revenue from vouchers
leads to more religious activity and may lead to less. Vouchers may thus affect
what churches close, which ones stay open, and which ones run schools. But
further, they may affect what happens inside of churches, and here the evidence
to date does not support a simple vouchers-are-good-for religion conclusion.
Clearly, however, more research is needed on this topic, in the United States
and elsewhere.
Using Secular Phenomena to Learn About

Religious Practice
Religious work can also gain critical new insights by focusing on the interaction
between religious groups and secular phenomena. There are many papers that
undertake this general strategy in recent years, some of them ranking among
the best economics-of-religion works done during this time period.
As an extremely incomplete list, Meyersson (2014) looks at electoral out-
comes in Turkey to study the role of women in Islam, and Carvalho (2013)
looks at wardrobe restrictions to study the role of veiling as a commitment
mechanism among Muslim women. In a historical study, Rubin (2014) uses
the invention of the printing press to study the role of inexpensive access to
books in promoting the reformation. In the modern US context, Gruber and
Hungerman (2008) and Cohen-Zada and Sander (2011) each use blue laws
that prohibit commerce on Sunday to study how religion affects drinking and
happiness, respectively. Finally, Frank and Iannaccone (2014) and Chen and
Lind (2016) each consider a variety of policy interventions in Europe and the
United States in the past century to further our understanding of religious
practice.
At first, the above list may simply seem like a grab-bag of (worthwhile)
studies, but the fact that they all look at policies or “secular” variation to
improve our understanding of religion may seem more coincidence than
anything else. Is there any underlying, unifying theory that suggests why the
study of the secular may be crucial to the study of religion?
Fortunately, the answer is yes, as is discussed in a recent study (Hungerman
2014). That study begins with a big question for economists interested in
religion: why do religious adherents seem different (e.g., less likely to commit
crimes) than others? On the one hand, many religious groups have rules that
proscribe certain behavior. So, perhaps adherents respond to these rules by
changing what they do. On the other hand, perhaps it is the case the people
who “like” religion coincidentally tend to not like the prohibited things, so
that religious proscriptions don’t matter all that much.
Often economists would attempt to solve a problem like this using an
instrumental variable: something that affects religion, but would not otherwise
matter for a person’s decision to commit a crime (or drink, or vote, or be happy,
etc.). It has proven very hard to find instruments for religion, however. A few
clever studies (such as Benjamin et al. 2016) try to experimentally affect one’s
religiosity and in so doing learn about religion. This is a very good idea and
these studies are valuable. However, there may be important aspects of faith
that are not easily subjected to experimental variation.
How can causation be established, if one cannot necessarily appeal to
instruments or experiments? By appealing to theory. Hungerman’s work
revisits the canonical model of religion, developed famously by Iannaccone
(1992). It turns out that, in some settings, this well-studied model makes
distinct empirical predictions for what should happen if proscriptions matter
compared to what should happen when they do not. The model provides
a test for whether religious groups causally affect behavior. Importantly, the
key exogenous variation needed for this test does not involve incentives to be
religious, but rather involves exogenous variation in secular phenomena. This
latter sort of variation is much easier to find in the real world—especially when
it is driven by policy. The model not only provides a method for testing the
causal effect of religion but provides a test that is easy to use. Hungerman then
lays out several tests and concludes that religious participation does indeed
matter for secular behavior.
Scholars interested in studying the effects of religion but lacking good
instruments for religion (which is to say, most scholars studying religion)
would do well to consider whether secular phenomena could be used. Often,
by understanding how policies and other secular forces impinge on religiosity,
we can learn things about religion that would otherwise prove quite difficult.
Case Study: Church Bulletins

The above examples lay out a rationale for studying church-state issues well
beyond the traditional state-sponsored-church discussion. In this section, we
discuss a final example of recent work that highlights some of these ideas, based
on data from church bulletins.
Our use of bulletins happened accidentally. When working with the Arch-
diocese of Milwaukee (as discussed above), we asked church officials if any
of the information we sought was publicly available. To our surprise, they
said it was. Specifically, parishes (i.e., Catholic churches) in Milwaukee were
required to report financial information in their worship bulletins, and these
bulletins, beyond being disseminated during worship, were almost always put
on a parish’s website online. After some reflection, it occurred to us that we
could visit parish websites—for all parishes, not just those in Milwaukee—to
collect this information, giving us a vibrant and high-frequency (weekly) view
to church life for the largest denomination in the United States.
We began our collection efforts in 2014 and they continued until 2015, at
which point we ran out of server space. The collection effort and other details
about the bulletins are discussed in Hungerman et al. (2018).
Figure 10.2 gives an example of a bulletin. The figure shows two pages
(pages 1 and 5 in this case). The name of the church, and the pastors, has
been redacted, as has the address. But, of course, knowing the exact location
of a parish can be quite valuable, as can the date (in this case, March 29). On
the right, the image of page 5 of this bulletin shows the collection amount
taken from the week before; this is something reported in almost all bulletins.
This provides a novel way to study weekly church activity—measuring weekly
donations collected. One drawback of this method is that, since bulletins
rarely report attendance, one cannot distinguish between changes in donations
driven by attendance (more people at worship) and a change in the generosity
of donors (the same number of people, but they give more). But the ability
to explore weekly donations, for a parish whose exact location is known,
represents an unpredicted opportunity to study the day-to-day (or, more
accurately, week to week) life of hundreds or even thousands of congregations.
Beyond collection amounts, we can also study the content of the bulletins.4
Given our interest in church/state relations, we thought it interesting to
look for politically loaded words or phrases. The recent, well-known paper
by Gentzkow and Shapiro (2010) provided such a list (in fact the list is in
the appendix to their paper). They examine the text of speeches given by
Congressional Democrats and Republicans and identify words and phrases
(or “words,” for short) overwhelmingly used by one party versus the other.
Admittedly, their list may be somewhat dated by today’s standards, so that
some words that have political meaning today (such as “Sad!”, or “death
panels”) would not have had the same political meaning then. However, clear
benefits of using Gentzkow and Shapiro’s list are that it was independently
developed, well-known, and readily available.
Figure 10.3 shows an average number of republican leaning words, and
democratic leaning words, that appear in bulletins from week to week during
the year 2015.5 There are a great number of things one could take away from
4 The words in the bulletin are not easily searched (we started with pdf bulletins, and OCR technology
did not always find text in them successfully), but we converted the text in the bulletins to ASCII files
(using a program in Unix) and then searched the many thousands of files for particular words (using a
program in Matlab).
5 Some examples of the most common republican leaning words/phrases in bulletins are “boy scouts,”
“human life,” and “ten commandments.” Some examples of the most common democratic leaning
words/phrases include “senior citizens,” “African American,” and “low income.”
Fig. 10.2 An example of a church bulletin. The above are two pages from a sample
bulletin
this picture. First, note that the number of political words used on average
is somewhat modest, with the average bulletin having less than one of these
words appearing in a given week. Second, there are more republican words
used than democratically leaning words in nearly every week of the year.
Fig. 10.2 (Continued)

Pope Visits
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
Thanksgiving
0
Republican
Fig. 10.3 Occurrence of democratic/republican words in church bulletins, by week
Third, the lowest week of the year in terms of political language used is the
US holiday of thanksgiving. On the one hand, the idea the bulletins may use
less political language on this holiday might make sense. But we do not think it
is ex-ante obvious that this holiday would generate less political language than,
say, Valentine’s day, or even a religiously focused holiday such as Christmas.
Fourth, there is a large amount of republican language used at the start of
the year and then a massive spike in the fall. The language used at the start
of the year appears to often be related to Boy Scout troop activity (e.g., “this
year our boy scout troop will meet in a new room”). We initially took the
incidence of this language at the start of the year as non-political in nature and
indeed the equivalent of statistical noise. However, several colleagues pointed
out to us that our interpretation of this boy-scout-based discourse as politically
neutral was perhaps naïve. We agree that such language could be interpreted
in different ways, but we have little else to say about it thus far.
The large spike in the fall occurs when Pope Francis visited the United
States, and our reading of the bulletins confirms that his visit prompted this
increase and that much of the language generated in response to his visit
concerns abortion. We find this to be remarkable—and, to our knowledge,
somewhat unprecedented—evidence of the immediate and dramatic impact
the visit by a leading figure in the church can have on the language used within
the church. But, next, notice that this usage on language is remarkably brief
in duration—immediately upon the Pope’s departure the usage of republican

leaning words returns to normal.
Finally, there are several other spikes in the summer in republican word use,
one in late June and two larger ones in August. The smaller spike in June
occurs the week gay marriage was legalized in the United States. The larger
spikes happen during the start of the republican-presidential-primary debates,
but we cannot say with certainty whether or not this is simply coincidence.
Can these data be brought to bear on current church state work, along
the lines of the research described earlier? Hungerman et al. (2017b) take this
issue up by matching weekly bulletin data to data on political campaigns. (As
discussed in that paper, no central dataset on political campaign stops existed
for the 2017 election, and so we constructed one ourselves). Fortunately, many
people ran for president in 2015, so that we have many campaign stops we can
include in our data. Further, our bulletin data allows us to pursue a number
of strategies, such as creating 52-week differences within parishes over time, to
address the fact that political campaign stops are clearly non-random in nature
and that communities receiving stops may have different levels of church
collections for any number of reasons.
After addressing this concern, we find that stops lead to a short-term increase
in church collections, but this effect becomes indistinguishable from zero
after only a week or two. The results do not support a story where political
gatherings encourage church activity by, for example, providing people with
information about the locations, or political inclinations, of neighborhood
churches (because in that case the effects would likely persist). Moreover, we
find that the effects are largest for Catholic candidates (but do not appear to
vary based on the political language found in the church bulletins) suggesting
that allowing churches to campaign for candidates could affect the fortunes of
different churches (and different candidates) in different ways. This touches
on the theme of unexpected policy implications discussed earlier, as currently
the federal government in the United States is considering altering the Johnson
Amendment, which traditionally has discouraged churches from campaigning
for candidates (Goodstein and Shear 2017).
Overall then, this discussion of church bulletins highlights the value of
studying church state relations in a context such as the uses for the reasons
mentioned above—it uncovers unexpected policy implications, provides new
insights into religious practice, and utilizes a novel source of data that could
prove useful to future scholarship.
Conclusions
The study of church state issues remains a vibrant and exciting area of work in
economics as well as other fields. The discussion here has focused on church-
state research in contexts where policy impinges mainly on secular forces. Two
benefits of this approach were mentioned: it can highlight important policy
implications and can motivate methodologically rigorous study of religious
practice in situations where no other attractive methodology is available. While
some of the work described considers the United States in the present day, we
have attempted to repeatedly point out that the benefits of this work, and many
recent good examples of work in this vein, concern the practice of religion in
a wide variety of places and times.
Looking forward, what can one say about work of this nature? As religion
continues to decline in the United States and many other (not all) countries,
the relevant importance of secular forces may only grow. This is not to say
that religion will disappear altogether, of course, but that today if one wishes
to understand the religious, one must increasingly look to the secular for
answers.
References
Anderson, Robert Warren, Noel Johnson, and Mark Koyama. 2017. Jewish Persecu-
tions and Weather Shocks, 1100–1800. Economic Journal 127 (602): 924–958.
Barro, Robert, and Rachel McCleary. 2005. Which Countries have State Religions?
Quarterly Journal of Economics 120 (4): 1331–1370.
Benjamin, Daniel J., James J. Choi, and Geoffrey Fisher. 2016. Religious Identity and
Economic Behavior. The Review of Economics and Statistics 98 (4): 617–637.
Chen, Daniel, and J. Lind. 2016. The Political Economy of Beliefs: Why Fiscal and Social
Conservatives/Liberals (Sometimes) Come Hand-in-Hand. TSE Working Paper No.
16-722.
Cohen-Zada, Danny, and William Sander. 2011. Religious Participation Versus Shop-
ping: What Makes People Happier? Journal of Law and Economics 54 (4): 889–906.
Figlio, David. 2009. Voucher Outcomes. In Handbook of Research on School Choice,
ed. M. Brends, M.G. Springer, D. Ballou, and H.J. Walberg. New York: Routledge.
Franck, Rapheal, and Laurence Iannaccone. 2014. Religious Decline in the 20th
Century West: Testing Alternative Explanations. Public Choice 159 (3–4): 385–
414.
Gentzkow, Matthew, and Jesse M. Shapiro. 2010. What Drives Media Slant? Evidence
from U.S. Daily Newspapers. Econometrica 78 (1): 35–71.
Goodstein, Laurie, and Michael Shear. 2017. Trump’s Order on Religious Liberty
Pleases a Few, but Lets Down Many Conservatives. The New York Times, May 4.
Gruber, Jonathan, and Daniel Hungerman. 2008. The Church Versus the Mall: What
Happens When Religion Faces Increased Secular Competition? Quarterly Journal
of Economics 123 (2): 831–862.
Hout, Michael, and Claude Fischer. 2002. Why More Americans Have No Religious
Preference: Politics and Generations. American Sociological Review 67 (2): 165–190.
Hoxby, Caroline. 2003. School Choice and School Productivity: Could School
Choice Be a Tide that Lifts All Boats? In The Economics of School Choice, ed. C.M.
Hoxby. University of Chicago Press.
Hungerman, Daniel. 2014. Do Religious Proscriptions Matter? Evidence from a
Theory-Based Test. Journal of Human Resources 49 (4): 1053–1093.
Hungerman, Daniel, and Kevin Rinz. 2016. Where Does Voucher Funding Go? How
Large-Scale Subsidy Programs Affect Private School Revenue, Enrollment, and
Prices. Journal of Public Economics 136: 62–85.
Hungerman, Daniel, Kevin Rinz, and Jay Frymark. 2017a. Beyond the Classroom: The
Implications of School Vouchers for Church Finances. NBER Working Paper 23159.
———. 2017b. Political Campaigns and Church Contributions. Journal of Economic
Behavior and Organization 155: 403–426.
Hungerman, Daniel, Kevin Rinz, Tim Weninger, and Chungeun Yoon. 2018. Reli-
gious Leadership, Religious Language, and Religious Giving: A Study of Church
Bulletins. NBER Working Paper 24374.
Iannaccone, Laurence. 1992. Sacrifice and Stigma: Reducing Free-Riding in Cults,
Communes, and Other Collectives. Journal of Political Economy 100 (2): 271–291.
Meyersson, Erik. 2014. Islamic Rule and the Empowerment of the Poor and Pious.
OECD. 2012. School Choice and Equity: Current Policies in OECD Countries
and a Literature Review. Directorate for Education Working Paper 66. http:/
/www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=EDU/
WKP(2012)3&docLanguage=En
Rouse, Cecelia. 1998. Private School Vouchers and Student Achievement: An Evalu-
ation of the Milwaukee Parental Choice Program. Quarterly Journal of Economics
113 (2): 553–602.
in the Reformation. The Review of Economics and Statistics 96 (2): 270–286.
Smith, Adam. 1994. The Wealth of Nations. New York: The Modern Library.
11
Protestants and Catholics and Educational
Investment in Guatemala
Rachel M. McCleary and Robert J. Barro
Introduction
Beginning in the early nineteenth century, British and US Protestant missions
invested worldwide in traditional forms of human capital, namely, education
and healthcare.1 Recent empirical studies comparing Protestant with Catholic
missions in Asia, the Indian subcontinent, Oceania, and Africa support this
assessment (Grier 1997; Woodberry and Shah 2004; Woodberry 2009, 2012;
Nunn 2014; Bai and Kung 2012).
These studies are part of a broader discussion of the Protestant Reformation
(1517–1555) and its long-term effects on literacy and mass education in Europe
(Becker and Wöessmann 2009; Schaltegger and Torgler 2009; Becker et al.
2011; Boppart et al. 2013, 2014; Cantoni 2015).2 Within the context of the
Reformation, mass literacy and formal education became the hallmarks of
Protestantism. Protestant sects stressed individual reading of the Bible as the
basis of one’s salvation (sola scriptura-sola fide). The leaders of Protestantism—
1 We are grateful to Sriya Iyer, Jared Rubin, and Jean-Paul Carvalho for their comments on the draft of
our chapter.
2 For a review of the literature on the relationship between Protestantism and human capital, see Becker
et al. (2016).
R. M. McCleary () · R. J. Barro
Harvard University, Cambridge, MA, USA
e-mail: rachel_mccleary@harvard.edu

170 R. M. McCleary and R. J. Barro
Martin Luther, John Calvin, and Huldrych Zwingli—emphasized reading

the Bible as well as broader educational goals (Boppart et al. 2014). The
Bible became the source of personal spiritual life, communal worship, and
evangelization. Through vernacular translations, printing, and international
distribution networks, Protestants created a public demand for reading (Becker
and Wöessmann 2009; Becker et al. 2011; Boppart et al. 2014; Cantoni
2015). Protestantism is often linked with economic growth because education
is a major factor in the development of countries, particularly with regard
to productivity, technological innovation, and participation in commercial
enterprises.
Research on Protestant foreign missions during the nineteenth century
shows that missionaries established schools and hospitals in countries such as
China and Korea, and in British colonies such as India.3 Therefore, it is not
surprising that a strong positive correlation exists between the historical pres-
ence of British and US Protestant missionaries and educational attainment.
These empirical studies have limitations. First, they emphasize comparisons
of Protestant with Catholic investment in human capital. In this view, Protes-
tant missions in geographic regions where Catholicism had a monopoly, for
example, in Latin America, succeeded due to an indolent Catholic hierarchy
and a paucity of national priests. Official prohibitions on vernacular Bibles
coupled with the reading of the Latin Vulgate restricted to ordained clergy or,
in their absence, a native maestro cantor, created a demand for access to reading
materials including the Bible.
The Catholic reaction to the Reformation was to extend its control over the
publication and translation of the Bible so as to limit heretical interpretations.4
The Council of Trent (1545–1563) emphasized equally apostolic traditions
and the Bible. For Catholics, the Bible was a communal, liturgical resource
whose message was mediated by clerical authority. The goal was to convert
people into the Catholic Church where they would receive instruction through
baptismal catechism and the lectionary. A culture of discouraging public access
to reading the Bible was actively enforced by the Catholic Church until Vatican
II (1962–1965).
A second, important limitation of recent studies is that Protestantism is
treated as synonymous with Mainline Protestant denominations and their
3 For China, see Cohen (1963), Bai and Kung (2012), and Woodberry (2012). For Korea, see Shearer
(1965), Lee (1989), Woodberry (2007), and McCleary (2013). For Africa, see Nunn (2009, 2010, 2014).
4 There were many exceptions prior to the Council of Trent; see Crehan (1963, 199–237).
11 Protestants and Catholics and Educational Investment in Guatemala 171
missions.5 Mainline denominations share a cluster of beliefs, such as viewing

the Bible not only as the inspired word of God but also as an historical
document subject to textual analysis and criticism. Henry Sloane Coffin, one
of the more famous US Mainline ministers in the early twentieth century,
succinctly expressed this secular view. The Bible should be subject to “the
methods of ascertaining and imparting religious truth which are employed
in all departments of human knowledge” (Coffin 1915).
Woodberry (2012, 244) referred to Mainline or liberal Protestant mission-
aries as “conversionary Protestant” and defined them as those who “(1) actively
attempt to persuade others of their beliefs, (2) emphasize lay vernacular Bible
reading, and (3) believe that grace/faith/choice saves people, not group mem-
bership or sacraments.” Notably, this definition does not include Evangelical,
Pentecostal, neo-Pentecostal, and charismatic groups of believers.6 That is, the
definition does not take into account the supernatural manifestations of the
holiness and Pentecostal movements beginning in the late nineteenth century,
as well as the development of neo-Pentecostal and independent charismatic
churches in the second half of the twentieth century.
In the early twentieth century, Mainline Protestant theology was in crisis
and under sharp attack. New groups emerging—holiness (to become known
as Evangelical) and Pentecostal—were theologically and organizationally dis-
tinct from Mainline and preached premillennial theology. Their biblically
based missiology required evangelizing to the world before Christ’s imminent
return. Premillennialist missions steadily increased as Mainline Protestant
(post-millennial) missions declin. The focus on evangelizing coupled with
premillennialist urgency meant that mission work could not wait for mission-
aries to receive a college and graduate seminary education. Biblical prophetic
passages supported the view that the gospel had to be preached before Christ’s
coming. “So near in fact that there was scarcely time to do more than preach,
throw up cheap ‘sheds’ in which to preach, and then to preach. . . . The
5 Hutchison (1989) referred to seven Mainline or liberal Protestant groups as the Seven Sisters of
American Protestantism. The seven are United Methodist Church, Evangelical Lutheran Church in
America, Presbyterian Church, Episcopal Church USA, United Church of Christ, American Baptist
Churches USA, and Disciples of Christ. In his earlier work, Hutchison (1976) identified as Mainline
Methodist, Presbyterian, Congregational, Episcopal/Anglican, Baptist, and Disciples of Christ. The Pew
Research Center for Religion and Public Life identifies as Mainline Methodist, Lutheran, Presbyterian,
Baptist, Episcopalian/Anglican, Congregational, Disciples of Christ, Restorationist, Anabaptist, Friends
(Quakers), Reformed, and other small churches. For a discussion of the history of Mainline denominations
in the United States, see Thuesen (2002, 27–53).
6 The term “charismatic” refers to those who share with Pentecostals and neo-Pentecostals a belief in
Holy Spirit manifestations. Some charismatics formed independent churches, while others remained as
members of Mainline Protestant denominations (Synan 1975, 1–4; Johnson 2014, 274–276).
urgency of the case demands shortcut methods in preparing for the work”
(Wynkoop 1976, 33–34). Further, many premillennialist preachers viewed
colleges and seminars as too secularizing, dulling one’s faith and spiritual zeal.
The famous holiness preacher, Edward Franklin Walker, evangelized one night
at the annual convention of the Pentecostal Mission: “ . . . the tendency of the
times is to trust too much in learning. . . . We have seen uneducated, frail
women, filled with the Spirit, more powerful in winning souls than half a
dozen cultured doctors of divinity” (Wynkoop 1976, 34). Faith was all one
needed.
By the early 1950s, the premillennial Evangelical and Pentecostal missions
were in the majority among Protestants in many countries. Therefore, a
satisfactory approach to Protestant missions and their impact on human
capital must include an analysis of this growing segment of Protestantism—
Evangelicals and Pentecostals.
This study compares investment in education across types of Protestantism
(Mainline, Evangelical, Pentecostal, and neo-Pentecostal) as well as between
Protestant and Catholic. We hypothesize that educational investment in
Mainline Protestant churches and schools exceeds that in Evangelical and
Pentecostal churches and schools. In Guatemala and some other parts of Latin
America, the connection with education is likely to be weakest for Catholic
because the ouster of Catholic orders and schools in the liberal reforms of
the 1870s had a persisting influence. Guatemala is a particularly good place
to test these hypotheses because of the historical dominance of Catholicism,
followed by sharp growth of Protestantism, notably among Evangelicals and
Pentecostals.
All but one of the original Protestant missions in Guatemala were Evangel-
ical and Pentecostal. In 1964, the last year the Guatemalan national census
reported religious affiliation, 8.2 percent of Guatemalans were Protestant,
with the majority of those adhering to Evangelical and Pentecostal faiths. In
2001, an estimated 30 percent of Guatemalans were Protestant; in 2007, it
was 35.5 percent.7 In 2010, an estimated 40 percent of Guatemalans were
Protestant, with 56 percent of Protestants being Pentecostal and a strong
majority falling into the broad Evangelical category. The World Religion
Database estimates that in 2010, 50.3 percent of Guatemalans professed
belief in the manifestations of the Holy Spirit (Pentecostal, charismatic, or
independent charismatic).
7 SeeGallup Survey, gallup.com/poll/101905/gallup-poll.aspx, 2007 and Latino Barometer, latino-

barometro.org, 2007.
Forms of Protestantism in Guatemala

Starting at the end of the nineteenth century, US Protestant missions—
Mainline, holiness (which came to be called Evangelical), and later
Pentecostal—entered Guatemala to evangelize for the purpose of conversion.8
Holiness and Pentecostal missionaries had minimal education, some attended
a Bible or missionary training institute. Holiness missionaries practiced John
Wesley’s creed to be “ . . . at full liberty, simply to follow the Scriptures and
the Primitive Church” (Telford 1931, 237–238). The holiness movement, out
of which Evangelicals came, believed the church was formed by the calling
together of the people (ecclesia). This approach contrasts with the Mainline
Protestant emphasis on the authority of ordained ministers. For Evangelicals,
the spiritual calling—el llamado—rather than an academic degree became the
predominant requirement for the ministry.
An important trend in the early twentieth century was the rise of Pente-
costalism.9 The key distinguishing feature of Pentecostalism is the belief in
speaking in tongues (glossolalia), which accompanies a second act of grace,
Holy Ghost baptism. Pentecostals emphasize Holy Spirit manifestations,
such as prophesying, miracle healing, and exorcism. Similar to Evangelicals,
Pentecostals emphasize the imminent return of Christ on earth, an idea that
focuses human efforts and resources on evangelization. For our purposes, the
critical point is that Pentecostals share with Evangelicals a stress on rapid
evangelizing with a de-emphasis on institution-building or investment in
education and health (Svelmoe 2008, 67).
In contrast to Evangelicals and Pentecostals, Mainline Protestant denomina-
tions invested heavily in civilizing aspects of religion: establishing schools, liter-
acy campaigns, and vernacular translations of the Bible and religious materials
(McCleary 2013).10 This approach did not accord well with the prominence
8 According to Marsden (1987, 1991), an Evangelical is defined as believing in (1) the supreme authority of
inspired Scripture for faith and practice, (2) the divinity of Jesus Christ as incarnate God, (3) Jesus Christ
as savior and the only means of saving sinful humanity, (4) the importance of personal conversion as the
central criterion for salvation, and (5) a commitment to sharing the transforming “good news” of new life
in Jesus Christ, which comes by God’s Grace alone through faith in the crucified and risen savior. In this
chapter, the term “Evangelical” encompasses holiness churches and denominations that formed during the
first decades of the twentieth century, including Keswickian faith missions such as the Central American
Mission (CAM), known today as Camino Global, Wycliffe Bible Translators, and its sister organization
the Summer Institute of Linguistics, and Christian and Missionary Alliance.
9 The accepted history of Pentecostalism is that it became a recognized movement after the 1906 Azusa
Street revival in California (Wacker 1984, 354).
10 For a discussion of this emphasis on literacy, see McLuhan (1964), Ong (1982), Schmidt (2000), and
Eisenstein (1997).
of oral communication among indigenous communities in Guatemala. Only

two indigenous languages, K’iche and Kaqchikel, had a Latin alphabet with
colonial Spanish orthography. Thus, the services conducted by Evangelicals
and Pentecostals typically appealed more to indigenous people because of
the use of oral communication methods with charismatic manifestations.
Evangelicals and Pentecostals particularly valued shared communal and spon-
taneous worship involving singing, music, clapping, praying out loud, spiritual
exhortations, shouting, crying, testimonials, and glossolalia (Spittler 1988,
411–416).
Reading a Bible or hymnal—a typical feature of Mainline Protestantism—
has minimal utility in a church service of this type. Although the Bible affects
the religious service even for Evangelicals and Pentecostals, it is central not
because it is written and read, but because it is the inerrant word of God.
Biblical authority, not biblical interpretation, guides and justifies the believer’s
actions (Spittler 1988, 420). How believers receive God’s word is usually oral
(McCleary 2017).
The traditional Mainline Protestant emphasis on individual reading of the
Bible was shared to some extent by Evangelicals and Pentecostals. However,
Mainline Protestantism’s promotion of the broader goal of formal education
and mass literacy was much weaker for Evangelicals and even less so for
Pentecostals. In particular, the large numbers of conversions to Protestantism
over the decades in Guatemala cannot primarily be attributed to the reading of
the Bible. Illiteracy rates among indigenous peoples (measured in Spanish) in
the late nineteenth century and still today are particularly high in departments
with a majority indigenous population and where several indigenous languages
are spoken. In fact, a strong negative relationship exists between literacy rates
and the indigenous share of the population. Figures 11.1 and 11.2 show this
relationship across 22 departments of Guatemala in 1950 and 2011.
The dominant argument from scholars of Protestantism in Guatemala has
been that Evangelical missions positively influenced human capital, especially
through education. US missions purportedly championed the cause of uni-
versal literacy in Guatemala (Garrard 1986, 74–75, 77, 81), with investment
in education perceived as a means of social mobility (Sherman 1997, 43;
Martin 1990, 89; Rose and Shultze 1993, 415–451; esp. 428) and material
prosperity (O’Neill 2010, 10, 15; Althoff 2014, 369–370). Support for these
claims of causation from evangelicalism to education and prosperity are in need
of empirical investigation, particularly in the form of long-term quantitative
analysis. As we demonstrate with long-term data for Guatemala, Evangelicals,
.7
.6 GUA - 50
.5 PET - 50
Literac y Rate
SAC - 50
IZA - 50
.4
ESC - 50
QUE - 50
.3 ELP - 50
RET - 50
ZAC - 50
STR - 50 SNM
SUC - 50- 50
JUT - 50 JAL - 50
.2 CHM - 50
CHQ - 50
BAJ - 50 HUE - 50 TOT - 50
SOL - 50
.1
QUI - ALT
50 - 50
.0
0.0 0.2 0.4 0.6 0.8 1.0
IndigenousShare of Population
Fig. 11.1 Literacy rate and indigenous share across departments of Guatemala in 1950.
See Table 11.4 for acronyms of departments
.9
SAC - 11
.8
ESC - 11
ELP - 11GUA - 11 CHM - 11
STR - 11
RET - 11 QUE - 11
JUT - 11
Literacy Rate
IZA - 11
.7 ZAC - 11 SNM - 11
PET - 11
SUC - 11
JAL - 11 TOT - 11
.6 HUE - 11
SOL - 11
CHQ - 11 BAJ - 11
.5 ALT - 11
QUI - 11
.4
0.0 0.2 0.4 0.6 0.8 1.0
IndigenousShare of Population
Fig. 11.2 Literacy rate and indigenous share across departments of Guatemala in 2011.
See Table 11.4 for acronyms of departments
Pentecostals, and neo-Pentecostals were different from Mainline Protestants in

terms of investing in human capital. In particular, schools run by Evangelicals,
Pentecostals, and neo-Pentecostals contributed less than Mainline Protestant
schools to literacy.
Data for Departments of Guatemala

To test our hypotheses on the connection between religion and schooling, we
constructed two data sets. The first data set enumerates Protestant churches by
type in Guatemala from 1882, the year that Protestant churches first entered
Guatemala, until 2010. This data set contains 12,861 churches, geocoded to the
village and street level using the ArcGIS World Geocode service. The churches
are classified by type (Mainline, Evangelical, Pentecostal, neo-Pentecostal)
across the 22 current departments of Guatemala. Second, we constructed a
data set on the numbers of Protestant schools by type from 1882 to 2010,
applying to the 22 current departments of Guatemala. Prior to 1882, the
numbers of Protestant churches and schools were zero.
We have data on numbers of Catholic schools and parishes by department
and over time. A parish is a geographically designated portion of a diocese
to which at least one priest is assigned. A diocese, currently numbering
29 in Guatemala, is a geographic unit that falls under the authority of
a bishop. The sources of our information are annual publications by the
Vatican, Annuario Pontificio (2017/2018), and the Archbishop of Guatemala,
Annuario Eclesiástico de la Iglesia Católica en Guatemala (1924/2008). We have
incomplete information by department on numbers of Catholic priests and
were unable to gather reliable data on numbers of Catholic churches.
The Liberal reforms instituted by the governments of Miguel García
Granados (1871–1873) and Justo Rufino Barrios (1873–1885), including the
constitution of 1879, altered the religion market in Guatemala by introducing
religious freedoms and breaking the monopoly of the Catholic Church.
Catholic religious orders that operated schools, including the Jesuits, Domini-
cans, and Franciscans, were expelled from the country.11 The government
also expropriated Catholic properties (monasteries, convents, seminaries) and
converted them to public use, especially for public schools. The Liberal govern-
ment abolished the Ley Pavón, which emphasized religion as the foundation of
a good education. Schools under that law were administratively linked to the
local parish. In place of the Ley Pavón, the Liberal government established a
national public educational system providing free, universal, and obligatory
primary schooling without religious instruction. Between 1872 and 1879,
11 PaulBurgess, a Presbyterian missionary, in his biography of Barrios, provides a detailed description of

how Barrios used his position as General of the Guatemalan Army under President Granados to round up
the Jesuits and force them onto a boat to Panama (Burgess 1946, 89–90). For a more detailed discussion of
these events, see Miller (1969). Priests who wished to remain in the country were required to give up their
vows and become secular priests, thereby submitting to the authority of the Archbishop of Guatemala.
legislation was enacted to secularize and centralize education in the newly

created Ministry of Education (Miller 1966, 251–263).
At the same time that Barrios was expelling Catholic religious orders
and expropriating Church property, he actively encouraged the entry of US
Protestant missions into the country. Presbyterian, Friend’s (Quaker), and
Nazarene missions established schools under the direction of the government.
Barrios’s children attended the fledgling Presbyterian school in Guatemala
City. Paul Burgess, in his biography of Barrios, states that Liberals throughout
Latin America favored Protestantism over Catholicism as a means of creating
political space for the secular state (Burgess 1946, 196–197, n. 3). Barrios was
also an admirer of the United States and sought to introduce US progressive
ideas and modern technology into Guatemala (Rippy 1942, 595–605).
A 1925 report that we obtained from the Vatican Secret Archives, signed
by Guatemalan Archbishop Muñoz y Capurón, indicates that 108 Catholic
parishes existed in the country in 1925. Half of these lacked a parish priest.
This pattern is corroborated by the data in the Annuario Eclesiastico Catolico
de Centro America (1925). The report from the Vatican Secret Archives also
mentions that there were 70 secular priests in Guatemala in 1925, concentrated
in Guatemala City and mainly serving the Catholic Church’s administration.
Two years later, the Annuario reported only 58 parish priests, with most
concentrated in the capital.
Barrios’s successors maintained the Liberal reforms until December 1933,
when the Pope appointed the first Papal Nuncio for Guatemala. With the
agreement of the Guatemalan government, Catholic religious order reentered
Guatemala in 1937 (Jesuits), 1943 (American Maryknolls), 1947 (Franciscans),
1955 (Spanish Franciscans of the Order of the Sacred Heart), 1959 (Belgian
Catholics), and 1962 (American Benedictines). The constitution of 1955
legally recognized the juridical personality of religious institutions, including
the Catholic Church. Correspondingly, the official presence of the Catholic
Church expanded greatly in Guatemala. The number of dioceses went from
1 in 1921 to 3 in 1937, 7 in 1955, and 29 in 2010. The number of parishes
expanded from 87 in 1955 to 494 in 2010, and the number of Catholic schools
went from 36 in 1955 to 340 in 2010. Guatemala had 4484 Catholic priests
in 1995 and 26,696 in 2010. In our data set, the number of Catholic schools
in each department is entered as zero from 1880 to 1939. That is, only as of
1940 was there a start of the revival of Catholic education.
With respect to education, we use two indicators that are available over
the long term and across departments—literacy rates and primary-school
enrollment ratios. Literacy is an outcome measure (based on answers to
questions on censuses or surveys), whereas enrollment is an input measure,
based on whether children in the appropriate age range are reported to be

attending primary school.12 One problem with the enrollment numbers is that
children may be listed as attending primary school without actually attending
or learning much. A limitation of the literacy numbers is that they capture only
an early stage of acquisition of human capital—without considering numeracy
and more advanced forms of education.
We gathered data on literacy and population by municipality and depart-
ment from 1880 to 2011 from government censuses and surveys. For these
data, we used the national censuses as well as information from the Ministry
of Education, the bilingual literacy program (DIGEBI), and the adult literacy
program (CONALFA). The census and survey data on literacy rates go back to
1880 and are available by department for 11 dates: 1880, 1893, 1921, 1940, 1950,
1964, 1973, 1981, 1994, 2002, and 2011.13 The corresponding data on primary-
school enrollment ratios are available by department starting in 1973, covering
the five years: 1973, 1981, 1994, 2002, and 2011. Therefore, the literacy data
provide information over a much longer period.
National census data on ethnic composition by municipality and depart-
ment were collected from 1880 to 2011. We focus on the fraction of the
population identified as indigenous. As already mentioned, the relationship
between literacy and the indigenous share is strongly negative. Particularly
since the Peace Accords of 1996, the government of Guatemala has expressed
the objective of bringing literacy among indigenous peoples up to the levels
attained by Ladinos. The implementation of bilingual programs in designated
public schools has been one method employed by the government. However,
we do not find reliable statistical evidence that the negative relationship
between literacy and the indigenous share has been weakening over time.
12 The government of Guatemala measures literacy in two ways. First, the national censuses ask for children
aged 7 to 15 years, “Can you read?” or “Do you know how to read?” The same question is sometimes
targeted toward Mayan languages. Second, literacy information is gathered through survey data using
national sampling techniques (Fernando Rubio, Director and Chief of Party, United States Agency for
International Development, Guatemala, e-mail correspondence with Rachel McCleary, August 5, 2013).
In recent years, primary and secondary education is mandatory and offered free as public schooling for
children between the ages of 7 and 14.
13 These are the years of national population censuses, except for 2011. The last national census was in
2002.
Long-Term Evidence on Religion and Human

Capital in Guatemala
Our central empirical objective is to isolate the effects of Protestantism and
Catholicism on the accumulation of human capital in Guatemala. We focus
on the roles of religious schools and churches.
Tables 11.1A, 11.1B, and 11.2 contain regressions to assess determinants
of human capital across 22 departments of Guatemala for years in which
the data are available. The sample for literacy rates has 239 observations
(22 departments, 11 dates14 ) and that for school-enrollment ratios has 110
observations (22 departments, 5 dates). Table 11.3 shows the sample means
and standard deviations of the variables used in Tables 11.1A, 11.1B, and 11.2.
Table 11.4 lists the departments.15
An important advantage of the regression framework is that it allows an
examination of effects of religious schools and churches on human capi-
tal while holding constant important background variables. These variables
include the indigenous share of the population and population density (urban-
ization). These background variables are correlated with the human-capital
measures and also with the numbers of religious schools and churches.
Therefore, if these variables were not held constant, effects on literacy from
religious schools and churches might reflect these correlation patterns. For
example, Mainline Protestant schools per capita may seem to have a positive
linkage with literacy because these schools are higher when the indigenous
share of the population is lower, and this share has a negative relation with
literacy.
More generally, we would like to isolate the exogenous parts of the variations
in the numbers of religious schools and churches—of which we consider Main-
line Protestant, Other Protestant (Evangelical, Pentecostal, neo-Pentecostal),
and Catholic. Then we could clearly isolate causation from religious schools
and churches to human capital, rather than the reverse. In the regression
analysis, we take several approaches to assessing potential endogeneity with
regard to numbers of religious schools.
Key elements in the cross-department locations of religious missionaries,
churches, and schools are the comity agreements reached in the early twentieth
14 Data prior to 1940 for El Progreso are unavailable because this area was included with other departments
until 1934.
15 All departments are included, except for Amatitlán, which was a separate department until 1935, after
which it became part of the Department of Guatemala. Amatitlán is excluded throughout because of
missing data. El Progreso is included as a distinct department since 1934.
Table 11.1A Panel A: Regressions for literacy rates
180
(1) (2)
No dept. fixed effects Dept. fixed effects
Constant 0.575 (0.024)*** 0.580 (0.028)***
Mainline Protestant schools per 1000 persons 6.86 (1.96)*** 2.91 (2.50)
Other Protestant schools per 1000 persons 0.93 (0.48)* 0.56 (0.42)
Catholic schools per 1000 persons 0.28 (0.25) 0.16 (0.22)
Mainline Protestant churches per 1000 persons 0.014 (0.012) −0.006 (0.013)
Other Protestant churches per 1000 persons 0.012 (0.012) 0.008 (0.011)
Catholic parishes per 1000 persons 0.65 (0.17)*** 0.20 (0.16)
Indigenous share of population −0.254 (0.017)*** −0.044 (0.039)
Population density (1000s per square mile) 0.290 (0.038)*** 0.216 (0.043)***
Population density squared −0.072 (0.012)*** −0.064 (0.011)***
1880 dummy −0.342 (0.029)*** −0.423 (0.028)***
R. M. McCleary and R. J. Barro
1893 dummy −0.380 (0.029)*** −0.476 (0.028)***

1921 dummy −0.357 (0.027)*** −0.452 (0.026)***
1940 dummy −0.219 (0.025)*** −0.299 (0.024)***
1950 dummy −0.271 (0.026)*** −0.340 (0.024)***
1964 dummy −0.301 (0.027)*** −0.342 (0.023)***
1973 dummy −0.201 (0.027)*** −0.243 (0.023)***
1981 dummy −0.087 (0.024)*** −0.125 (0.019)***
1994 dummy −0.029 (0.022) −0.061 (0.018)***
2002 dummy 0.038 (0.021)* 0.016 (0.017)
R-squared 0.922 0.955
Standard error of regression 0.069 0.055
Number of observations 239 239
***Statistically significant at 1 percent level; **Statistically significant at 5 percent level; *Statistically significant at 10 percent level
Standard errors of estimated coefficients are in parentheses. Estimation is by panel least squares, using 22 departments of Guatemala
in years with available data. The 11 dates are 1880, 1893, 1921, 1940, 1950, 1964, 1973, 1981, 1994, 2002, and 2011. Standard errors are
clustered by department. The left-out dummy variable is for 2011
Statistics on the dependent and independent variables are in Table 11.3
Table 11.1B Panel B: Estimated department fixed effects in Panel A, column 2

Department Estimated coefficient (s.e.)
Alta Verapaz −0.117 (0.031)***
Baja Verapaz −0.064 (0.014)***
Chimaltenango −0.014 (0.023)
Chiquimula −0.043 (0.017)**
El Progreso 0.069 (0.026)***
Escuintla 0.073 (0.018)***
Guatemala (capital) 0.193 (0.029)***
Huehuetenango −0.086 (0.016)***
Izabal 0.090 (0.019)***
Jalapa −0.020 (0.015)
Jutiapa 0.028 (0.021)
Petén 0.113 (0.024)***
Quetzaltenango 0.007 (0.026)
Quiche −0.111 (0.027)***
Retalhuleu 0.043 (0.019)**
Sacatepéquez 0.070 (0.019)***
San Marcos 0.009 (0.016)
Sololá −0.149 (0.020)***
Santa Rosa 0.021 (0.024)
Suchitepéquez −0.022 (0.020)
Totonicapán −0.109 (0.024)***
Zacapa 0.021 (0.022)
***Statistically significant at 1 percent level; **Statistically significant at 5 percent
level; *Statistically significant at 10 percent level
Note: The sum of the estimated fixed-effects coefficients is constrained to be zero. The
p-value for the hypothesis that these coefficients are all equal (so that each department
has the same intercept, corresponding to the constant term) is 0.000
century. The first two Protestant missions to Guatemala were the Mainline
Presbyterians and the Evangelical Central American Mission (CAM). These
two groups differed greatly in evangelizing styles and theology. In order to
moderate competition between the groups, they came to an agreement in
1903 to divide Guatemala City into respective mission zones. In addition,
they divided up various departments outside of the capital into distinct, non-
competing evangelizing territories. This agreement was extended in 1907 to
include another Evangelical enterprise, the California Yearly Friends’ Society
mission. In 1916, a new comity agreement was reached to incorporate addi-
tional Evangelical groups—the independent Brethren medical mission and
the Church of the Nazarene. This agreement included reaffirmation of the
territory reserved for CAM and the Friends’ Society. A final comity agreement,
in 1935, included the Presbyterians along with the various Evangelical groups,
with the addition of the Evangelical Primitive Methodists. The main point
is that the locations of these Protestant religious groups reflected political
Table 11.2 Regressions for school-enrollment ratios

(1) (2)
No dept. fixed effects Dept. fixed effects
Constant 0.876 (0.031)*** 0.849 (0.085)***
Mainline Protestant schools 1.85 (2.21) 2.43 (4.44)
per 1000 persons
Other Protestant schools per 0.32 (0.60) −0.81 (1.05)
1000 persons
Catholic schools per 1000 −0.04 (0.34) 0.46 (0.45)
persons
Mainline Protestant churches 0.142 (0.050)*** 0.190 (0.082)**
per 1000 persons
Other Protestant churches per 0.022 (0.017) 0.075 (0.036)**
1000 persons
Catholic parishes per 1000 −0.03 (0.41) −0.81 (0.80)
persons
Indigenous share of −0.193 (0.028)*** 0.007 (0.111)
population
Population density (1000s per 0.148 (0.043)*** 0.148 (0.123)
square mile)
Population density squared −0.045 (0.012)*** −0.047 (0.023)**
1973 dummy −0.391 (0.030)*** −0.447 (0.056)***
1981 dummy −0.346 (0.025)*** −0.395 (0.046)***
1994 dummy −0.106 (0.022)*** −0.143 (0.031)***
2002 dummy −0.022 (0.021) −0.048 (0.023)**
R-squared 0.893 0.922
Standard error of regression 0.069 0.067
Number of observations 110 110
***Statistically significant at 1 percent level; **Statistically significant at 5 percent
level; *Statistically significant at 10 percent level
Standard errors of estimated coefficients are in parentheses. Estimation is by panel
least squares, using 22 departments of Guatemala in years with available data. The
five dates are 1973, 1981, 1994, 2002, and 2011. Standard errors are clustered by
department. The left-out dummy variable is for 2011
Statistics on the dependent and independent variables are in Table 11.3
decision-making and were unlikely to reflect substantial reverse causation from

human capital to religious missionaries, churches, and schools.
The comity agreements of the early twentieth century omitted Pentecostal
(and, later, neo-Pentecostal) groups. Pentecostal missionaries focused their
initial entry in 1901 on a few departments (notably Totonicapán, El Quiché,
Alta Verapaz, and Zacapa) but subsequently spread throughout Guatemala.
Pentecostal groups competed aggressively with other forms of Protestants and
other groups of Pentecostals, as well as with Catholics. The locations and
numbers of Pentecostal missionaries were driven by the desire to proselytize
and convert people as rapidly as possible. The evangelizing campaigns by
Table 11.3 Means and standard deviations of variables

Variable Mean Std. dev.
Literacy-rate sample: 22 departments, 11 years, 239 observations
Literacy rate 0.378 0.236
Indigenous share of population 0.484 0.302
Protestant schools per 1000 persons 0.0124 0.0178
Evangelical schools per 1000 persons 0.0088 0.0133
Mainline schools per 1000 persons 0.00063 0.0022
Neo-Pentecostal schools per 1000 persons 0.00061 0.0022
Pentecostal schools per 1000 persons 0.0024 0.0080
Non-Mainline Protestant schools per 1000 persons 0.0118 0.0176
Protestant churches per 1000 persons 0.788 0.635
Evangelical churches per 1000 persons 0.292 0.313
Mainline churches per 1000 persons 0.156 0.406
Neo-Pentecostal churches per 1000 persons 0.021 0.036
Pentecostal churches per 1000 persons 0.319 0.373
Non-Mainline Protestant churches per 1000 persons 0.632 0.556
Catholic schools per 1000 persons 0.0185 0.0271
Catholic parishes per 1000 persons 0.0441 0.0324
Population density (1000s per square mile) 0.265 0.418
Population density squared 0.245 1.207
Population (1000s) 252.5 341.5
Area (square miles) 1926 2795
Enrollment-ratio sample: 22 departments, 5 years, 110 observations
Primary-school enrollment ratios 0.699 0.198
Literacy rate 0.574 0.163
Indigenous share of population 0.418 0.320
Protestant schools per 1000 persons 0.0221 0.0195
Mainline schools per 1000 persons 0.00085 0.0025
Non-Mainline Protestant schools per 1000 persons 0.0212 0.0193
Protestant churches per 1000 persons 0.885 0.456
Mainline churches per 1000 persons 0.086 0.131
Non-Mainline Protestant churches per 1000 persons 0.799 0.425
Catholic schools per 1000 persons 0.0318 0.0302
Catholic parishes per 1000 persons 0.0405 0.0208
Population density (1000s per square mile) 0.429 0.559
Population density squared 0.494 1.748
Population (1000s) 413.6 441.2
Area (square miles) 1911 2788
Pentecostals were particularly aggressive in the 1940s and 1950s, resulting in

thousands of conversions. For our purposes, the central point is that we do
not expect significant causation from human capital (gauged, for example,
by literacy and school enrollment) to numbers of Pentecostal churches and
schools. Pentecostal leaders were basically uninterested in promoting human
capital.
Table 11.4 Departments of Guatemalaa

Department Acronym
Alta Verapaz ALT
Baja Verapaz BAJ
Chimaltenango CHM
Chiquimula CHQ
El Progreso ELP
Escuintla ESC
Guatemala GUA
Huehuetenango HUE
Izabal IZA
Jalapa JAL
Jutiapa JUT
Petén PET
Quetzaltenango QUE
Quiche QUI
Retalhuleu RET
Sacatepéquez SAC
San Marcos SNM
Sololá SOL
Santa Rosa STR
Suchitepéquez SUC
Totonicapán TOT
Zacapa ZAC
a Amatitlánwas a separate department until 1935 (after which it became part of
Department of Guatemala), but Amatitlán is excluded throughout as a separate
department because of missing data. El Progreso is included as a distinct department
since 1934
Regression Results for Literacy Rates
The regression results for literacy rates are in Table 11.1A, Panel A. These results
apply to the 22 departments of Guatemala for 11 dates from 1880 to 2011.16 The
error terms are clustered by department. The setting includes time effects (year
dummies), which capture the sharp long-term upward trend in literacy. The
basic setting in column 1 excludes fixed effects for departments. Therefore,
these regression coefficients reflect information from the 11 cross sections—
with levels of literacy related to levels of religious schools and churches and the
other variables. The key assumption is that the error terms are independent of
these level variables.
16 Because ofmissing data, the numbers of churches and religious schools per capita in 2011 were assumed
to be the same as those observed in 2010.
Results on Background Variables
We first assess the results in Table 11.1A, Panel A, column 1, for the background
variables, starting with the indigenous share of the population. The overall
relationship between literacy and the indigenous share is strongly negative,
with an estimated coefficient of −0.254 (s.e. = 0.017). Thus, a one-standard-
deviation rise in the indigenous share (by 0.30 in Table 11.3) associates with
a decline in the literacy rate by 0.08, compared with the mean of the literacy
rate of 0.38 (Table 11.3).
An important question is whether the relation of literacy rates to the
indigenous share has weakened over time. We first broke up the sample into the
six time observations from 1880 to 1964 versus the five from 1973 to 2011. With
this division, there is no evidence of a weakening relationship.17 The estimated
coefficient on the indigenous share for the 1880–1964 period is −0.230
(s.e. = 0.024), whereas that for the 1973–2011 period is −0.276 (0.023).
This pattern suggests, if anything, a strengthening in the inverse relationship
between literacy and the indigenous share. However, the difference between
the estimated coefficients is not statistically significantly different from zero at
usual significance levels (p-value = 0.16). Therefore, the results accord with a
stable relationship over the full sample from 1880 to 2011.
As already noted, the government of Guatemala has tried particularly since
the signing of the Peace Accords in December 1996 to raise literacy among
the indigenous population by providing extra bilingual educational resources
to areas with high indigenous shares of the population.18 The 2003 Law
of National Languages, while affirming Spanish as the official language of
the country, states that the government “recognizes, respects, and promotes
the development and usage of indigenous languages including non-Maya
Garifuna and Xinka.” The Directorate for Bilingual Education (Dirección
General de la Educación Bilingüe or DIGEBI), operating out of the Ministry
of Education, overseas school bilingual programs as a means of transitioning
to the official language, Spanish. DIGEBI operates in departments with
large numbers of indigenous-speaking children. There are also a number of
17 In this system, the coefficients of all variables other than the indigenous share were constrained to be
the same across the two sub-periods.
18 The term “bilingual education” as used in Guatemala refers to learning in the first language to achieve
competency in a second language (transitional), or developing communication skills in the first language
while achieving competency in a second (parallelism). See Consejo Nacional de la Educación Maya (2003)
and Richards and Richards (1996, 208–221).
other bilingual education programs with international funds operating at the

department level.
If government promotion of bilingual education were successful, we would
expect that the negative relation between literacy rates and the indigenous
share would be weaker in the last two periods compared with that in the earlier
periods. We find that the estimated coefficient on the indigenous share for
the nine dates for 1880–1994 is −0.263 (s.e. = 0.019), whereas that for the
two dates for 2002–2011 is −0.220 (0.035). This pattern is consistent with a
weakening relationship between ethnicity and the indigenous share, but the
difference in estimated coefficients is not statistically significantly different
from zero (p-value = 0.27). Hence, there is no reliable statistical evidence even
since 2002 that the government has diminished the negative relation between
literacy and the indigenous share.
The regression for literacy in Table 11.1A, Panel A, column 1, includes
population density and its square as explanatory variables. The results show
a statistically significant, non-linear relationship. The estimated coefficients
imply that the literacy rate rises with population density until this density
becomes very high—2014 persons per square mile, compared with the mean
of 265 (Table 11.3). Beyond that—that is, in the most urban areas—the
estimated relationship becomes negative. The rising segment likely reflects
weak educational systems in departments that are extremely rural. However,
the positive linkage between population density and literacy tends to weaken
as density rises, likely because of diminishing effects of urbanization on
the availability and quality of schooling. When density becomes extremely
high, the estimated relationship becomes negative, possibly because of adverse
effects from crowding and crime. We also found that the estimated non-linear
relationship between literacy rates and population density is stable over time.
The hypothesis that the relationship in the six periods from 1880 to 1964 is
the same as that for the five from 1973 to 2011 is accepted with a p-value of
0.20.19
The regression in Table 11.1A, Panel A, column 1, also includes a dummy
variable for each time period, with the 2011 period being the left-out variable.
Thus, each coefficient shows estimated literacy (for 1880, 1893, . . . ), com-
pared to that in 2011, for given values of the other explanatory variables. The
negative coefficient estimates (except for 2002) indicate the overall long-term
trend toward higher literacy rates, given the values of the explanatory variables.
19 In this case, the coefficients of all variables other than population density and its square were constrained
to be the same for the two sub-periods.
Results on Religious Schools
The results of most interest for our study concern the connections between
literacy rates and numbers of religious schools per capita. Our main thesis is
that Mainline Protestant schools will be more favorable than Other Protestant
schools toward literacy. Given the government’s historical antagonism toward
Catholic orders and schools, we also anticipate that Catholic schools will have
the weakest effect. For the Protestant data, we separate Mainline from Other,
which comprises Evangelical, Pentecostal, and neo-Pentecostal (see Table 11.3).
For religious schools, one finding in Table 11.1A, Panel A, column 1, is
the large and significantly positive coefficient, 6.9 (s.e. = 2.0), on Mainline
Protestant schools per 1000 persons. This result compares with the signifi-
cantly positive (at the 10 percent level) but smaller estimated coefficient, 0.93
(0.48), for Other Protestant schools. These two estimated coefficients differ
with a p-value of 0.003.
If we separate Evangelical schools from Pentecostal/neo-Pentecostal schools,
we get respective coefficients of 0.91 (s.e. = 0.65) and 0.95 (0.74). These two
estimated coefficients do not differ significantly (p-value = 0.97). That is, we
lack evidence that Evangelical schools differ from Pentecostal/neo-Pentecostal
in terms of impacts on literacy.
The estimated coefficient on Catholic schools per 1000 persons, 0.28 (0.25),
is positive but insignificantly different from zero.20 This coefficient differs
from that of Mainline Protestant schools with a p-value of 0.001 but does
not differ significantly from that of Other Protestant schools (p-value = 0.28).
Consistent with our expectations, the point estimates reveal a ranking whereby
literacy is enhanced most by Mainline Protestant schools, then by Other
Protestant schools, and then by Catholic schools.
We now consider whether the estimated effects of religious schools on
literacy rates are stable over the long term. We estimated one set of the three
coefficients for the first six sub-periods, from 1880 to 1964, and another over
the last five periods, from 1973 to 2011.21 For Mainline Protestant schools, the
estimated coefficient is 7.23 (s.e. = 3.28) for the first interval and 5.88 (2.32)
for the second. For Other Protestant, the respective estimated coefficients are
1.24 (0.78) and 0.79 (0.60). For Catholic, they are 0.86 (0.53) and 0.15 (0.28).
20 Recall that the numbers of Catholic schools are entered as zero in each department prior to 1940.
However, the inclusion of these years—for 1880, 1893, and 1921—does not have much influence on the
results. If the sample begins only in 1940, the estimated coefficient of Catholic schools per capita is 0.28,
s.e. = 0.26.
21 The coefficients of other variables were constrained to be the same over the two sub-periods.
A joint test that these three pairs of coefficients are the same across the two
intervals is accepted with a p-value of 0.44. Therefore, the data are consistent
with a stable connection over the long history between religious schools and
literacy rates.
Results on Churches
The results for churches in Table 11.1A, Panel A, column 1, are weaker than
those for religious schools. The estimated coefficients on Mainline Protestant
and Other Protestant churches per 1000 persons are each positive but statisti-
cally insignificantly different from zero.22
For Catholic, we lack data on numbers of churches and, therefore, use
instead the number of Catholic parishes per capita. In Table 11.1A, Panel
A, column 1, “Catholic parishes per 1000 persons” has a positive estimated
coefficient, 0.65 (s.e. = 0.17), that is statistically significantly different from
zero. It is not meaningful to compare the size of this coefficient with those for
Protestant churches because a parish often has more than one church.
Department Fixed Effects for Literacy Rates
A common practice in regressions with panel data, particularly to try to deal

with endogeneity issues,23 is to add cross-sectional (in our case, department)
fixed effects. These effects capture long-term differences in levels of the depen-
dent variable (literacy rates) across departments. Therefore, the remaining
basis for identification of effects of independent variables on the dependent
variable comes from differences across departments in the trends (changes)
in the independent variables. The idea—not necessarily correct—is that these
differential trends in the independent variables may be more exogenous than
the differences in levels.
The panel regression estimates for literacy rates with the inclusion of
department fixed effects is in Table 11.1A, Panel A, column 2. In this case, the
estimated coefficients associated with religious schools and churches become
22 If we separate Evangelical churches from Pentecostal/neo-Pentecostal churches, we get respective

coefficients of −0.001 (s.e. = 0.019) and 0.022 (0.014). These two estimated coefficients do not differ
significantly (p-value = 0.34).
23 Hickman (2016) attempts to deal with endogeneity by using as instruments for Protestant presence in
the 1990s the historical locations of Protestant missionaries and of Catholic parishes. However, the former
variable turns out to lack explanatory power for Protestant presence in the 1990s. More Catholic parishes
do predict lower Protestant presence.
statistically insignificantly different from zero. That is, the identification based
on differential trends in these schools and churches across departments is too
weak to obtain statistically significant results. Importantly, however, the broad
pattern of point estimates for religious schools is the same as that in column
1. That is, the estimated coefficient is largest for Mainline Protestant schools,
next largest for Other Protestant schools, and smallest for Catholic schools.
Table 11.1B, Panel B, shows the estimated fixed-effects coefficients for
the 22 departments.24 These estimated coefficients are jointly statistically
significantly different from zero (p-value = 0.000). That is, the departments
exhibit clear differences in average levels of literacy, given the values of the
independent variables. Notably, the department of Guatemala (containing
the national capital) has a fixed-effects coefficient of 0.193 (s.e. = 0.029) so
that, other things equal (including population density), the literacy rate in the
capital is higher than that in the typical department by 19 percentage points.
We further attempted to deal with endogeneity concerns by using the
method developed by Sims (1972). Consistent with his approach, we included
as explanatory variables a ten-year lag and lead of the numbers of each type of
religious school, along with the contemporaneous values already included.25
Sims’s idea is that statistical significance for the lag suggests causation from
school numbers to literacy, statistical significance for the lead suggests the
opposite direction of causation, and statistical significance contemporaneously
could reflect a mixture of effects. We carried out this estimation in the setting
with department fixed effects.
For Mainline Protestant schools, the estimated coefficients are 6.53
(s.e. = 3.96) for the lag, −8.08 (5.14) contemporaneously, and 6.17 (3.43) for
the lead. The corresponding estimates for Other Protestant schools are 1.27
(0.52) for the lag, −0.77 (0.78) contemporaneously, and 0.50 (0.82) for the
lead. For Catholic schools, the results are 0.08 (0.37) for the lag, 0.26 (0.34)
contemporaneously, and 0.25 (0.26) for the lead. The p-values for the joint
statistical significance of each set of three estimated coefficients are 0.15 for
Mainline Protestant schools, 0.11 for Other Protestant Schools, and 0.34 for
Catholic schools.
For Other Protestant schools, the significantly positive estimated coefficient
on the ten-year lag, along with the statistically insignificant estimates for the
contemporaneous and lead values, suggests causation from school numbers
24 The regression includes a constant term, and the sum of the estimated department fixed effects is
constrained to be zero.
25 The sample now uses nine dates for the literacy rate, from 1893 to 2002. Results are similar with five-year
lags and leads.
to literacy. This result accords with our informal discussion about the loca-
tion of non-Mainline Protestant schools and churches—dictated by political
agreements and the desire to evangelize rather than to promote human capital.
Despite the lack of major concern by these religious authorities with human
capital, these schools do appear to promote literacy.
For Mainline Protestant schools, the estimated ten-year lag coefficient is
significantly positive with a p-value of 0.10, and the estimated ten-year lead
coefficient is significantly positive with a p-value of 0.07. Therefore, the results
suggest a mixture of effects whereby Mainline Protestant schools promote
human capital and are also more likely to locate in areas with high human
capital. Although the evidence on causation is not clear cut, the results do
indicate a much stronger overall interaction between religious schools and
literacy for Mainline Protestant than for Other Protestant.
For Catholic schools, the results suggest little interaction in either direction
between schools and literacy. Again, this finding likely reflects the elimination
of Catholic schools and orders in the 1870s and the delay in reintroduction of
these schools and orders until around 1940.
Regression Results for School-Enrollment Ratios
The regression results for primary-school enrollment ratios are in Table 11.2.
These results apply for the 22 departments of Guatemala for 5 dates from
1973 to 2011. As before, the error terms are clustered by department, and
the regression includes time effects (year dummies). The basic setting in
column 1 excludes fixed effects for departments. The overall influences of
the background variables—indigenous share of the population, population
density, and the dummies for time periods—accord with those in the literacy-
rate regression (Table 11.1A, Panel A, column 1).
One difference from the results for literacy rates arises in the relation
between school-enrollment ratios and the indigenous share. The overall rela-
tion, shown in Table 11.2, column 1, is significantly negative, with an estimated
coefficient of −0.193 (s.e. = 0.028). However, this relationship has weakened
in recent periods. The coefficient estimated for the three periods from 1973
to 1994 is −0.289 (0.028), whereas that for the two periods since 2002 is
−0.058 (0.033). These two estimated coefficients differ significantly from
each other (p-value = 0.000), and the one since 2002 differs significantly from
zero only at the 10 percent level. That is, the government of Guatemala seems
to have succeeded in eliminating most of the inverse relationship between
primary-school enrollment and the indigenous share, likely because of legal
requirements for attendance of children at primary school. It is, therefore,

surprising that the success in getting indigenous children to attend school
(as reported) has been unsuccessful in promoting literacy among indigenous
people.
For religious schools, we find no statistically significant effects on atten-
dance at primary schools (Table 11.2, column 1). These results may arise
because attendance at religious primary schools is mostly a substitute for
attendance at non-religious schools. For churches, the one statistically signif-
icant effect is a positive impact on primary-school attendance from Main-
line Protestant churches per 1000 persons. The estimated coefficient, 0.142
(s.e. = 0.050), differs significantly from zero. This estimated coefficient differs
significantly from that on Other Protestant churches (0.022 [0.017]) with a p-
value of 0.017. The estimated coefficient on Catholic parishes per capita is
essentially zero. Thus, the main inference is that Mainline Protestant churches
have promoted overall attendance at primary schools.
Table 11.2, column 2, shows the regression results for school-enrollment
ratios with the inclusion of department fixed effects. In this case, the estimated
fixed-effects coefficients are jointly insignificantly different from zero (p-
value = 0.26). As a consequence, the overall pattern of results for school
enrollment is not greatly affected by the addition of department fixed effects.
One new finding, however, is that Mainline Protestant and Other Protestant
churches per 1000 persons both have statistically significant positive effects on
school-enrollment ratios. The estimated coefficient for Mainline Protestant is
larger than that for Other Protestant, and the p-value for equal coefficients is
0.17.
Conclusion
A good deal of recent empirical research on the relation of religion to
human capital has focused on Mainline Protestantism versus Catholicism.
Our research emphasizes differential investment in education across types
of Protestantism. In addition, we compare Protestantism with Catholicism.
Our research was motivated by theological differences between Mainline
Protestant denominations and later premillennialist movements (Evangelical,
Pentecostal) that arose in Guatemala at the end of the nineteenth century.
Premillennialist denominations came to dominate missions in the twentieth
century. These religious groups placed less emphasis than Mainline Protestants
on investment in education. Our findings bear out this idea. Specifically,
literacy is enhanced more by Mainline Protestant schools than by Other
Protestant schools. Catholic schools have the weakest effect, likely because
of the ouster of Catholic orders and schools from Guatemala in the Liberal
reforms of the 1870s.
Our findings suggest that future research on Protestantism and human-
capital investment should take into account types of Protestantism. Evangel-
icals, Pentecostals, and neo-Pentecostals—which now dominate the religious
landscape in Guatemala and many other countries—differ greatly from Main-
line Protestants.
References
Althoff, Andrea. 2014. Divided by Faith and Ethnicity: Religious Pluralism and the
Problem of Race in Guatemala. Berlin and Boston: Walter de Gruyter GmbH.
Annuario Eclesiástico Católica de Centro America. 1925. Arquidiócesis de Santiago de
Santa Ana, El Salvador.
Annuario Eclesiástico de la Iglesia Católica en Guatemala. 1924/2008. Arquidiócesis de
Santiago de Guatemala, Guatemala City.
Annuario Pontificio. 2017/2018. Libreria Editrice Vaticana, Vatican City.
Bai, Ying, and James Kai-sing Kung. 2012. Diffusing Knowledge While Spreading God’s
Message: Protestantism and Economic Prosperity in China, 1840–1920. Unpublished,
April.
Becker, Sascha O., and Ludger Wöessmann. 2009. Was Weber Wrong? A Human
124 (2): 531–596.
Becker, Sascha O., Erik Hornung, and Ludger Wöessmann. 2011. Education and
Catch-Up in the Industrial Revolution. American Economic Journal: Macroeco-
nomics 3 (3): 92–126.
the Protestant Reformation. Warwick Economics Research Paper Series No. 1105.
Coventry, UK: University of Warwick, January.
Boppart, Timo, Josef Falkinger, Volker Grossmann, Ulrich Woitek, and Gabriela
Wüthrich. 2013. Under Which Conditions Does Religion Affect Educational
Outcomes? Explorations in Economic History 50 (2): 242–266.
Boppart, Timo, Josef Falkinger, and Volker Grossmann. 2014. Protestantism and
Education: Reading (the Bible) and Other Skills. Economic Inquiry 52 (2): 874–
895.
Burgess, Paul. 1946. Justo Rufino Barrios: A Biography. Guatemala: Quetzaltenango.
Cantoni, Davide. 2015, August. The Economic Effects of the Protestant Reformation:
Testing the Weber Hypothesis in the German Lands. Journal of the European
Economic Association 13 (4): 561–598.
Coffin, Henry Sloan. 1915, May 18. The Practical Aims of a Liberal Evangelicalism;
Closing Address. New York: Union Theological Seminary.
Cohen, Paul A. 1963. China and Christianity: The Missionary Movement and the
Growth of Chinese Antiforeignism, 1860–1870. Cambridge: Harvard University
Press.
Consejo Nacional de la Educación Maya. 2003. Las Escuelas Mayas: Una Experiencia
Educativa en Guatemala. www.edbiguatemala.org/articleview/8511/121.
Crehan, F.J. 1963. The Bible in the Roman Catholic Church from Trent to the Present
Day. In The Cambridge History of the Bible v.3: The West from the Reformation to
the Present, ed. S.L. Greendale, 199–237. Cambridge: Cambridge University Press.
Eisenstein, Elizabeth L. 1997. The Printing Press as an Agent of Change: Communication
and Cultural Transformations in Early-Modern Europe. Cambridge: Cambridge
University Press.
Garrard, Virginia. 1986. A History of Protestantism in Guatemala. Doctoral thesis,
Tulane University.
Grier, Robin. 1997. The Effect of Religion on Economic Development: A Cross
National Study of 63 Former Colonies. KYKLOS 50 (1): 47–62.
Hickman, Peter. 2016. Is the Protestant Ethic Alive in Latin America? An Empirical
Assessment of the Economic Impacts of Religion and Missionaries in Guatemala. Senior
thesis, Harvard University.
Hutchison, William R. 1976. The Modernist Impulse in American Protestantism.
———. 1989. Between the Times: The Travail of the Protestant Establishment in
America, 1900–1960. Cambridge: Cambridge University Press.
Johnson, Todd M. 2014. Counting Pentecostals Worldwide. Pneuma 36: 265–288.
Latino Barometer. 2007. Santiago de Chile: Latinobarómetro Corporation. http://
www.latinobarometro.org.
Lee, Sungho. 1989. The Emergence of the Modern University in [the Republic of ]
Korea. Higher Education 18 (1): 87–116.
Marsden, George M. 1987. Evangelical and Fundamental Christianity. In The Encyclo-
pedia of Religion, ed. Mircea Iliade, vol. 15, 190–197. New York: Collier Macmillan.
———. 1991. Understanding Fundamentalism and Evangelicalism. Grand Rapids, MI:
William B. Eerdmans.
Martin, David. 1990. Tongues of Fire: The Explosion of Protestantism in Latin America.
Oxford and New York: Blackwell Press.
McCleary, Rachel M. 2013. Protestantism and Human Capital in Guatemala and the
Republic of Korea. Asian Development Bank Economics Research Paper, Series 332.
Manila: Asian Development Bank.
———. 2017. Protestant Innovative Evangelizing to Oral Cultures in Guatemala.
In Oxford Handbook of Latin American Christianity, ed. O.P. Edit David Orique,
Susan Fitzpatrick Behrens, and Virginia Garrard-Burnett. Oxford University Press.
McLuhan, Marshall. 1964. Understanding Media: The Extensions of Man. New York:
McGraw-Hill.
Miller, Hubert J. 1966. Positivism and Educational Reforms in Guatemala, 1871–

1885. Journal of Church and State 8 (2): 251–263.
———. 1969, January. The Expulsion of the Jesuits from Guatemala in 1871. The
Catholic Historical Review 54 (4): 636–654.
Nunn, Nathan. 2009. Christians in Colonial Africa. Unpublished manuscript, March.
———. 2010, May. Religious Conversion in Colonial Africa. American Economic
Review 100 (2): 147–152.
———. 2014. Gender and Missionary Influence in Colonial Africa. In Africa’s
Development in Historical Perspective, ed. Emmanuel Akyeampong, Robert H.
Bates, Nathan Nunn, and James A. Robinson, 489–512. Cambridge: Cambridge
University Press.
O’Neill, Kevin Lewis. 2010. City of God: Christian Citizenship in Postwar Guatemala.
Berkeley, CA: University of California Press.
Ong, Walter. 1982. Orality and Literacy: Technologizing of the Word. London: Rout-
ledge.
Richards, Julia Becker, and Michael Richards. 1996. Maya Education: A Historical and
Contemporary Analysis of Mayan Language Education Policy. In Maya Cultural
Activism in Guatemala, ed. Edward F. Fischer and R. McKenna Brown, 208–221.
Austin, TX: University of Texas Press.
Rippy, J. Fred. 1942, November. Relations of the United States and Guatemala During
the Epoch of Justo Rufino Barrios. The Hispanic American Historical Review 22 (4):
595–605.
Rose, Susan D., and Quentin Shultze. 1993. The Evangelical Awakening in
Guatemala: Fundamentalist Impact on Education and Media. In Fundamentalisms
and Society: Reclaiming the Sciences, the Family, and Education, ed. Martin E. Marty
and R. Scott Appleby, vol. 2, 415–451. Chicago and London: The University of
Chicago Press.
Schaltegger, Christopher A., and Benno Torgler. 2009. Was Weber Wrong? A Human
Capital Theory of Protestant Economic History: A Comment on Becker and Woess-
mann. Center for Research in Economics, Management and the Arts. Working
Paper No. 2009-06. Basel: CREMA.
Schmidt, Leight Eric. 2000. Hearing Things, Religion, Illusion, and the American
Enlightenment. Cambridge, MA: Harvard University Press.
Shearer, Roy E. 1965. Wildfire: Church Growth in Korea. Grand Rapids, MI: Eerd-
mans.
Sherman, Amy. 1997. The Soul of Development: Biblical Christianity and Economic
Transformation in Guatemala. New York: Oxford University Press.
Sims, Christopher A. 1972. Money, Income, and Causality. American Economic Review
62 (September): 540–552.
Spittler, Russell P. 1988, October. Implicit Values in Pentecostal Missions. Missiology:
An International Review 16 (4): 411–416.
Svelmoe, William Lawrence. 2008. A New Vision for Missions: William Cameron
Townsend, the Wycliffe Bible Translators, and the Culture of Early Evangelical Faith
Missions, 1896–1945. Tuscaloosa, Alabama: The University of Alabama Press.
Synan, Vinson, ed. 1975. Aspects of Pentecostal-Charismatic Origins. Plainfield, NJ:
Logos International.
Telford, John. ed. 1931. The Letters of the Rev. John Wesley (vol. 9, pp. 237–238).
London: Epworth Press. Volume 7.
Thuesen, Peter J. 2002. The Logic of Mainline Churchliness: Historical Background
Since the Reformation. In The Quiet Hand of God: Faith-Based Activism and the
Public Role of Mainline Protestantism, ed. Robert Wuthnow and John H. Evans,
27–53. Berkeley: University of California Press.
Wacker, Grant. 1984. The Functions of Faith in Primitive Pentecostalism. The
Harvard Theological Review 77 (3/4): 353–375.
Woodberry, Robert D. 2007. The Social Impact of Missionary Higher Education. In
Christian Responses to Asian Challenges, ed. Philip Yuen Leung and Peter Tze Ming,
99–120. Hong Kong: Center for the Study of Religion and Chinese Society.
———. 2009. Dividing Elites: Religious Liberty, Protestant Competition, and Democ-
racy in the Global South. Working Paper 002, Project on Religion and Economic
Change.
———. 2012, May. The Missionary Roots of Liberal Democracy. American Political
Science Review 106 (2): 244–274.
Woodberry, Robert D., and Timothy S. Shah. April 2004. Christianity and Democ-
racy: The Pioneering Protestants. Journal of Democracy 15 (2): 47–61.
World Religion Database. World Religion Database. www.worldreligiondatabase.org.
Wynkoop, Mildred Bangs. 1976. The Trevecca Story: 75 Years of Christian Service.
Nashville, TN: Trevecca Press.
12
The Religious Factor in Private Education
in the United States
Danny Cohen-Zada and Moshe Justman
Introduction
This chapter provides a brief overview of our theoretical and empirical
contributions to understanding the dominant role of the religious factor in
private education in the United States. Private, fee-paying education today
accounts for 8% of enrollment in primary and secondary schools in the United
States, down from a high of 14%, 50 years ago (Fig. 12.1); and about 80% of
these private school students attend religious schools, down from almost 90%,
30 years ago (Broughman and Swaim 2013; Fig. 12.2). The low overall rate of
private education is largely a consequence of the historically dominant role of
local school districts in funding public education in the United States, coupled
with socio-economic geographic segregation, which allows for substantial
variation in the quality of public schools, and the general absence of tax credits
D. Cohen-Zada ()
Department of Economics, Ben-Gurion University, Beer Sheva, Israel
e-mail: danoran@bgu.ac.il
M. Justman
Department of Economics, Ben-Gurion University, Beer Sheva, Israel
School of Economics and Business Administration, Ruppin Academic Center,
Emek Hefer, Israel
e-mail: justman@bgu.ac.il

198 D. Cohen-Zada and M. Justman
16.0%
14.0%
12.0%
10.0%
8.0%
6.0%
1950 1960 1970 1980 1990 2000 2010 2020
Fig. 12.1 Private share in elementary and secondary school enrollment, 1955–2016.
Source: US Census Bureau, Current Population Survey, 1955 to 2016
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
1975 1980 1985 1990 1995 2000 2005 2010 2015
Catholic Other religious Nonsectarian
Fig. 12.2 Distribution of private school enrollment by school type. Source: Digest of
Education Statistics 1980, 1990, 2015
for private school tuition.1 The further recent decline in private enrollment
likely reflects the growth of publicly funded charter schools,2 as well as court-
mandated funding reforms that increased state support for poorer school
districts, and a decline in demand for Catholic education.
1 The use of school vouchers to fund private schooling is generally limited to special circumstances: low-
income inner-city families with inadequate local public education, families in sparsely populated rural
areas, and so on.
2 Enrollment in charter schools in 2011/2012 exceeded two million students, a sixfold increase from
1999/2000.
The dominant religious factor in private education is a direct result of the

strict constitutional separation of Church and State, which bars the use of
public funds to directly support schools operated by organized religion, and
of the access of religious schools, especially Catholic schools, to subsidized
inputs.3 The recent decline in the share of religious education in private
schools in the last three decades reflects the decline in Catholic school
enrollment; schools affiliated with other Christian denominations slightly
increased their enrollment in this period.
We open our survey with a brief exposition of the theoretical models
that motivate our research. They describe a two-stage political-economic
equilibrium. In the first stage, democratic voting determines spending on
public education, with voters anticipating their subsequent individual deci-
sions on private education. Aggregate outcomes are determined by the joint
distribution of household income and religious motivation and by the religious
composition of the community. We then survey our empirical findings on
these issues and consider their methodological and policy implications. An
appendix surveys the data sources we used to study these issues.
Theoretical Perspectives
Our approach to characterizing the religious factor in private education is
structural, grounded in formal, context-specific models of the demand for
and supply of public, private religious, and private secular education in
the United States, which combine communal, political decisions on public
education with individual school choice. The parameters of the model are then
either estimated or calibrated. A full model of these individual and political
interactions would be intractable, and so each analysis emphasizes some aspects
of the problem and glosses over others. The exposition here is necessarily brief,
with further details available in the papers we cite.
Individual Choice
We begin with a simple model of education choice, after an education tax rate
has been set, which draws on Cohen-Zada and Justman (2005), and forms
3 Hoxby (1998) estimated that 50% of costs in Catholic elementary schools were subsidized by donations
from household and dioceses, and by teachers working for less than the going wage. See also Zelman v.
Simmons-Harris (2002, footnote 15).
the basis for more complex models. It takes the tax rate that finances public
education as given, assumes that private education is provided competitively
(at any desired level of spending per student), and abstracts from the variety of
religious denominations. Parents who send their children to private schools do
not receive tax credits for their school fees and therefore “pay twice” for their
children’s education, paying the education tax that funds public education in
their school district as well as their private school fees. Demand for private
school stems either from a desire to send one’s child to a school with much
higher spending per student than public schools offer, presumably offering a
higher quality of education in some dimension, or from a desire to provide
one’s child with a religious education. As average school fees in religious
schools are substantially lower than spending per student in public schools,
demand for religious education must be driven by religious sentiment.4
In a formal setting, assume for simplicity a continuous community of
households of measure 1, each comprising one parent and one child. Parents
derive utility from the quality of education they provide their children and
from consuming their remaining income. Each household i is characterized
by its income yi and its degree of religious attachment ki ≥ 0, a higher value
indicating stronger attachment. It may include the benefit to religious schools
from lower costs, for example, because they have access to highly motivated
teachers in religious orders willing to work for less pay. Denote by t the
proportional income tax rate that funds public education, and denote the share
of children attending public education by q ≤ 1.
If household i opts for public education, it consumes (1 − t) yi and enjoys
school quality t yi /q. If it opts for private secular education, let si denote the
school fees it pays. It then consumes (1 − t) yi − si and enjoys school quality
si . If it opts for private religious education and pays si in school fees, it again
consumes (1 − t) yi − si , but enjoys school quality ki si . Households maximize
utility by choosing a school type, determining if and how much to spend
on private schooling, conditioned on an assumed level of public enrollment.
Under reasonable conditions there is an equilibrium level of public enrollment,
on which we focus, that equates anticipated and actual enrollment rates.
Figure 12.3 describes the distribution of households among the three school
types—public, religious, and private non-sectarian—as a function of income
y and the religious factor k. Households that attribute greater advantage
to religious schooling have a lower-income threshold for switching from
4 In 2011/2012, average spending per student in public schools was almost $13,000, while annual tuition
averaged under $7000 in Catholic schools, and over $21,000 in non-sectarian schools (NCES 2013, Tables
201.10, 205.50).
k(y)
Religious senƟment
R
1
N
0 yn y
Income
Fig. 12.3 The distribution of households among school types. P denotes public school
enrollment; R denotes religious private school enrollment; and N denotes non-sectarian
private school enrollment
public schooling to private religious schooling. The threshold level of religious

preference k between public and religious education, which separates the areas
marked P and R in Fig. 12.3, is a decreasing function of yi . Private religious
enrollment includes lower-income households with strong religious sentiment
as well as higher-income households, but lower-income households require
stronger religious motivation to choose religious education. The threshold
income yn between public and non-sectarian private education corresponds to
the vertical line separating the areas marked P and N in Fig. 12.1. Religious
schools are more heterogeneous, in terms of parental income, than non-
sectarian private schools.
Communal Decisions and Influences
The model presented above describes the second stage of a broader model
in which the education tax rate is set in the first stage, by a majority vote
of these same households, anticipating the different second stage equilibria
that will result from different first-stage outcomes. We focus on political
equilibria defined as a tax rate such that no other tax rate is preferred to it
by a majority of households. Voters with above-average income pay more for
public schooling—as taxes are proportional while the quality of education is
uniform—and this substitution effect leads them to prefer a lower tax rate,
while the income effect implies that households with higher incomes prefer
a higher tax rate. Which of these effects is stronger determines whether an
equilibrium exists—existence of a political equilibria in models with public

and private education is not assured in general—and the form it takes.
The literature highlights two possibilities: one, in which low-income voters
prefer higher tax rates, and the median-income voter is decisive, a median-
income equilibrium; and the other one, in which middle-income voters
prefer the highest tax rate, an “ends against the middle” equilibrium. Which,
if either, equilibrium holds depends on the parameters of the household
utility function: a median-income equilibrium holds if the income elasticity
of demand for education is less than the elasticity of substitution between
education and other expenditures; an ends-against-the-middle equilibrium
holds if the income elasticity is greater. Calibrating the model to observed tax
rates and school-type enrollment shares allows us to estimate the size of the
perceived religious advantage and determine the type of equilibrium that holds
(Cohen-Zada and Justman 2003, 2005). Cohen-Zada and Justman (2012)
and Cohen-Zada and Elder (2017) relax the assumption of a single religious
denomination, allowing the data to indicate the degree of tension or affinity
between denominations. We review these findings in the next section.
Finally, we note that the religious factor in demand for private education
may also be influenced by the religious composition of the community.
The direction of influence depends on whether the efforts of family and
community to preserve religious identity are complements, in which case
parents have more incentive to socialize their children as their religion’s
market share in the local population grows, or substitutes, in which case the
opposite holds (Cohen-Zada 2006; Cohen-Zada and Sander 2008; Cohen-
Zada and Elder 2017). This sheds light on the debate between the traditional
interpretation of religious dynamics, which argues that hegemony promotes
religiosity, and the revisionist view, grounded in the comparison between the
United States and European countries with state religions, that competition
among religious denominations promotes religiosity. We expand on this, too,
below.
Empirical Findings
Our initial contribution to the empirical literature on the religious factor
in private education is Cohen-Zada and Justman (2003), which specifies a
political economy model of education choice that, unlike previous studies,
allows private tuition dollars (mostly religious) to have a greater (or lesser)
perceived purchasing power than tax dollars spent on public education, which
we calibrate. We specify a Stone-Geary utility function, which allows both
the income elasticity and the elasticity of substitution to vary.5 Calibration

reveals that the income elasticity is slightly less than one and the elasticity
of substitution slightly greater than one, so that the median-income voter
is decisive; and values are near enough to one that using the more tractable
logarithmic specification seems reasonable. It also reveals that private tuition
dollars have a perceived relative advantage of just over 30% at the margin, over
tax dollars spent on public education.
Cohen-Zada and Justman (2005) further refine this analysis by distinguish-
ing between non-sectarian private schools and subsidized religious private
schools. While the motive for sending a child to a non-sectarian private school
is a higher level of spending per student, the motive for sending a child to
religious schooling may be mainly religious, recognizing also that religious
schools often subsidize their tuition from external sources. In addition, the
model allows household to differ not only in their incomes but also in
the advantage they attribute to religious schooling (over secular schooling).
Households with a stronger commitment to religious values are more willing
to incur the added expense of sending their children to private schools
while continuing to pay the taxes that support public education. Thus, very
religious households may send their children to religious schools though their
income is lower than that of some households that attend public schools. We
calibrate the model to state-level data on the mean and variance of household
income, public spending per pupil on education, and the enrollment shares of
public, private non-sectarian, and religious schools to obtain estimates of the
subjective advantage of religious schools in each state, and find considerable
variation.
The calibrated model was then applied to analyze the effect on enrollment of
means-tested vouchers that can be applied toward tuition in either religious or
non-sectarian schools. The conclusions from this analysis accorded with early
experience with such voucher programs in Cleveland and Milwaukee.6 When
means-tested vouchers are offered in sums that fall substantially below public
school spending levels—as they were in Cleveland—many eligible parents
5 Logarithmic utility implies that both the income elasticity and the elasticity of substitution equal one;
CES utility implies that the income elasticity is one while the elasticity of substitution can vary.
6 The Cleveland School District Scholarship Program and the Milwaukee Parental Choice Program were
established to provide disadvantaged families poorly served by public schools with vouchers that could be
applied to other educational choices. They were the first such programs to include religious schools, from
1995, challenging earlier interpretations of the First Amendment, as barring the use of publicly funded
vouchers to pay for tuition at religious schools. Both programs were challenged in State and Federal courts
and allowed to stand, notably in the landmark Supreme Court decision on the Cleveland Program, in the
case of Zelman v. Simmons-Harris (2002).
remain within the public system, and those who decide to opt out of the public
system choose religious schools. When larger vouchers are offered—as they
were in Milwaukee—a sizable minority chooses non-sectarian private schools,
but most parents still prefer religious schools. These results highlight the
significance of including subsidized religious schools in means-tested voucher
programs, especially when the amount of the voucher falls substantially below
spending per pupil in public schools.
Cohen-Zada and Sander (2008) use individual data from the General Social
Survey (GSS) on frequency of church attendance as a proxy for religiosity,
which they show to be a significant source of variation in the magnitude of
the religious factor in demand for private education. Specifically, they build
a Random Utility Model of school choice in which households can choose
between public, Catholic, Protestant, and non-sectarian private schooling.
In their model, households differ not only in their income levels but also
in their religion and religiosity levels. The model is used to derive the
probabilities that a household attends various types of schools. They then
use data from the General Social Survey to estimate these probabilities using
multinomial logit regressions. Both household and community-level effects of
religion and religiosity on the demand for private schooling are considered in
their estimation. They find that both religion and religiosity have important
effects on the demand for private, Catholic, Protestant, and non-sectarian
schools. Specifically, they show that Catholic religiosity increases the demand
for Catholic schools and has no effect on the demand for other types of
private schooling; fundamentalist Protestant religiosity increases the demand
for Protestant schools and has no effect on the demand for other types of
private schooling; and other Protestant religiosity increases the demand for
non-sectarian private schools and has no effect on the demand for other
types of private schooling. These results suggest that religiosity is a key factor
that affects who attends private schools and who might respond to voucher
initiatives. They imply that previous research on the impact of Catholic
schooling on test scores that did not control for religiosity were likely to over-
estimate the treatment effect of Catholic schools.
A further extension of this approach, elaborated in Cohen-Zada and
Justman (2012) and Cohen-Zada and Elder (2017), recognizes that the many
Christian denominations in the United States exhibit different degrees of affin-
ity or tension toward each other, differences that are reflected in their choice
of religious education. Cohen-Zada and Justman (2012) develop a simplified
model of intergenerational cultural transmission through education with dif-
ferentiated religious groups, which indicates that the light in which adherents
of religious denomination j regard another denomination, k, should have an
effect on how the share of denomination j in the local population affects local
enrollment in private religious schools affiliated with denomination k. They
then regress private school enrollment shares, classified by religious affiliation,
on the local distribution of adherents among denominations, across counties
in the United States to identify patterns of tension or affinity.
The patterns they observe among the different denominations are generally
consistent with the earlier findings of expert panels reported in Hoge and
Roozen (1979) and Iannaccone (1994). They find a bilateral tension between
Catholics and evangelical Protestants, evident both in the multiplicity of neg-
ative effects associated with evangelical Protestant groups regarding Catholic
education and in the negative effect associated with Catholics regarding enroll-
ment in evangelical Protestant schools. This stands in contrast to the strong
affinity they find among Episcopalians regarding Catholic education and the
affinity of Catholics to mainline Protestant education. The generally low
tension associated with mainline Protestant groups is evident in the absence
of significant negative effects on the part of any denomination regarding
mainline Protestant schooling. This variation in affinity or tension among
denominations suggests that recognizing these distinctions can contribute to
a more nuanced understanding of demand for religious education and of the
channels through which religious identity is transmitted from one generation
to the next.
Another important line of research, pursued in Cohen-Zada (2006),
Cohen-Zada and Sander (2008), and Cohen-Zada and Elder (2017), considers
the effect of the religious composition of the community on demand for
religious schooling. The direction of influence depends on whether the efforts
of family and community to preserve religious identity are complements, in
which case parents have more incentive to socialize their children as their
religion’s market share in the local population grows, or substitutes, in which
case the opposite holds. This sheds light on the debate between the traditional
interpretation of religious dynamics, which argues that hegemony promotes
religiosity (Berger 1967), and the new paradigm, grounded in the comparison
between the United States and European countries with state religions, that
competition among religious denominations promotes religiosity (Iannaccone
1991; Finke and Stark 2005; Stark and McCann 1993).
Applying this to demand for private education, the argument is that
parents choose religious education for their children to protect them from
the external influences of other religions, to preserve their religious identity.
Thus, they attribute a higher perceived advantage to religious schooling of their
denomination if outside influences of other religions are more threatening.
This implies that when the share of a religious minority in the local population
grows—and outside influences become less threatening—the demand for sep-

arate religious schooling among the members of the religious group decreases.
This pattern implies concavity in the relationship between enrollment in
private and religious schooling and the share of the religious group in the
population.
Cohen-Zada (2006) provides evidence from US county data on Catholic
and private school enrollment that strongly supports this theory. Specifically,
he regresses the ratio of Catholic enrollment to Catholic adherents, in each
county, on the share of Catholics in the local population and estimates the
relationship between local private and Catholic school enrollment rates and
the share of Catholics in the local population. The results reveal that the
ratio of Catholic enrollment to Catholic adherents is a decreasing function
of the share of Catholics in the local population. They also show that the
relationship between both the Catholic and private enrollment rates and the
share of Catholics in the population is strongly concave, with a threshold level
of the Catholic share, beyond which private and Catholic enrollment rates no
longer increase with the Catholic share and may fall slightly. These findings
are robust to several econometric specifications. Taken together, they strongly
support the theoretical model and shed new light on the demand for private
religious education in the United States. These results imply that in localities
where registration in public schools is dominated by coreligionists, parents
are less inclined to opt for private education, viewing a public school with
many coreligionists as a reasonable, free substitute for costly private religious
education.
Cohen-Zada and Elder (2017) extend the analysis and findings of Cohen-
Zada (2006). First, it characterizes the conditions (on the primitives) under
which the share of students that attend schools of a given denomination (its
“enrollment share”) is a concave function of the share of its adherents in the
local population (its “market share”). Second, it shows that if the motivation to
preserve religious identity is sufficiently strong, a denomination’s enrollment
share is an inverse u-shaped function of its market share. Third, it tests
the predictions of the model not only for Catholics but also for mainline
Protestants and evangelical Protestants. Fourth, it provides evidence not only
from aggregate county-level data but also from individual survey data from the
National Educational Longitudinal Study of 1988 (NELS:88) and the Early
Childhood Longitudinal Study, Kindergarten Class of 1998–1999 (ECLS-K).
Fifth, it extends Cohen-Zada’s (2006) two-group model—“religious” and
“non-religious”—to multiple denominations, thus allowing an analysis of how
different denominations respond to changes in their market share. Cohen-
Zada and Elder (2017) show, through a combination of calibration and
estimation, that, on average, preserving religious identity is much less impor-

tant for mainline Protestants than for Catholics and evangelical Protestants,
a finding consistent with Hoge and Roozen’s (1979) and Iannaccone’s (1994)
systematic polls of expert panels. Sixth, Cohen-Zada and Elder (2017) shed
light on the relationship between religious pluralism and overall religious
activity, predicting that under weak regularity conditions, the fraction of
children who attend the schools of a given denomination is a positive function
of its share in the population and of religious pluralism. Their empirical
analysis supports this prediction and demonstrates that failing to control for
the local level of religiosity may lead to the erroneous conclusion that religious
pluralism has a negative effect on participation.
Cohen-Zada and Elder (2017) also analyze the dynamic implications of
their school choice model for the evolution of denominational market shares
across generations and the distribution of religions in steady state. While Bisin
and Verdier (2000) show that if vertical and horizontal socialization efforts
are substitutes then religious pluralism exists indefinitely, with minorities
never completely assimilating, Cohen-Zada and Elder (2017) show that this
result may not hold for schooling. They show that even if horizontal (school,
neighborhood) socialization and vertical (parental) socialization efforts are
substitutes, if attending a secular school preserves a secular family’s non-
religious identity as effectively as attending a religious school preserves a
religious family’s identity, then secularism ultimately dominates. This is
because public schools are secular in nature, which implies that, in each
generation, all secular families send their children to secular schools while
only a portion of religious parents send their children to religious schools.
Thus, the probability of children of secular parents preserving their non-
religious identity is higher than the probability of children of religious parents
preserving their religious identity. This dynamic property is unique to the
context of schooling.
Conclusion
Private education accounts for 8% of enrollment in primary and secondary
schools in the United States today, and about 80% of these private school
students attend religious schools. Our contributions to understanding the
dominant role religious schools play in private education in the United
States, reviewed in this chapter, combine theoretical modeling with structural
empirical estimation, and some calibrated policy analyses.
As a result of the highly decentralized structure of public education in

the United States, the share of public, private secular, and private religious
schools (affiliated with various denominations) varies widely from one school
district to another, as does spending per student in public schools. Yet certain
regularities are apparent. Secular private schools generally charge much more
in fees than public schools spend on their students, and religious private
schools charge less, on average; and where the large majority of students in
secular private schools come from affluent homes, students in religious private
schools come from all socio-economic strata.
The models we use to represent and analyze these phenomena posit a
demand structure with household-specific preferences for religious education
that are independent of their demand for secular dimensions of school quality
determined by spending per pupil. Households first vote on the public school
budget, then make their individual education choices, where their voting
anticipates these choices. These models readily accommodate the regularities
that characterize public and private school finance and enrollment in the
United States, described above.
We use these models to identify the parameters of education demand
through structural empirical estimations and calibration exercises, which lead
us to conclude that the income elasticity of demand for education spending
is slightly less than one while the elasticity of substitution between education
spending and other spending is slightly greater than one. This implies that
demand for public spending on education—the preferred education tax—
declines with income and hence that a political equilibrium exists in which
the tax rate desired by the median-income household is decisive. Moreover,
as both elasticities are close to unity, this suggests that the highly tractable
logarithmic utility function, where both elasticities are fixed and equal to one,
is a reasonable approximation. Applying this to simulate educational voucher
schemes shows clearly that typical school vouchers—with stringent means tests
and in values that are considerably less than public spending per pupil—are not
effective unless they can be redeemed in private religious, especially Catholic,
schools.
Our findings further show that a family’s religious affiliation, its religiosity,
and the religious composition of a given area are all important factors that
shape demand for private education in the United States. Households with
a strong religious identity generally prefer religious schools of their own
denomination, but will send their children to a convenient religious school
affiliated with a different denomination if the tension between the two
denominations is low. This allows us to elicit the level of tension between
denominations from patterns of school choice. At the same time, the degree
of household religiosity primarily increases demand for religious schooling

of its own denomination, with little effect on the demand for other types
of religious schooling. Finally, parents choose religious education for their
children to protect them from the external influences of other religions, to pre-
serve their religious identity. Consequently, they attribute a higher perceived
advantage to religious schooling of their denomination if outside influences
of other religions are more threatening. This implies that when the share of a
religious minority in the local population grows—and outside influences are
less threatening—individual demand for separate religious schooling among
members of the religious group decreases.
Understanding the dominant role of the religious factor in private education
in the United States sheds light on demand for education, in general, as well
as on the broader role of religion in the United States, today. It has far-
reaching implications for the social fabric of society in the United States as
well as for achieving broad-base economic growth through school-based public
policy. As our theoretical understanding of these phenomena deepens and
as the relevant data sources improve—especially individual data on religious
affiliation, religiosity, and schooling—further progress becomes possible.
Appendix: Data Sources

Some aspects of demand for religious schooling in the United States can
be estimated from individual data, while for others only aggregate data are
available, at the school district, county, or state level. Each has its advantages
and disadvantages. This section describes these sources of data, their uses, and
limitations.
Individual Level Data
Estimating the probability that a student will choose to attend each of the
various types of schools requires data on the type of school students attend,
their parents’ religious affiliation and religiosity, and the local distribution of
religious affiliation.
The largest database that includes data on enrollment in private schooling in
the United States draws on census data. It provides information on enrollment
in public, private, and home schooling, but does not distinguish between
religious and non-religious schooling. Another disadvantage of census data
is that it does not provide information on the religion and religiosity of the
student. However, Hofrenning and Chiswick (1999) developed an algorithm

for estimating the probability distribution of a respondent’s religion from
such microdata that includes data on ancestry but not religion, applying the
frequency distribution of religion by ancestry from the General Social Survey.
These limitations highlight the advantages of surveys that include informa-
tion on both school enrollment and the religion and religiosity of the student’s
family. These are much smaller datasets, but they distinguish between different
types of private schooling and provide explicit information on religion and
religiosity. The first of these surveys is the General Social Survey (GSS), a
cross-sectional national survey conducted since 1972. It includes information
on the type of school to which respondents send their children (public,
Catholic, Protestant, non-Christian religious, private non-sectarian, home
schooling); on their religious affiliation (Catholic, fundamentalist Protestant,
non-fundamentalist Protestant); and on their religiosity, measured as fre-
quency of attendance at religious services (nine categories from “never” to
“more than once a week”). Importantly, the GSS also reports the primary
sampling unit (PSU) in which respondents reside, allowing individual data
from the GSS database to be merged with local demographic variables.
A second survey of this type is the National Education Longitudinal Study
of 1988 (NELS:88), conducted by the US National Center for Education
Statistics (NCES). It comprises a nationally representative sample of 24,599
students from 1032 schools who attended eighth grade in 1988, with sub-
samples of these respondents resurveyed in 1990, 1992, 1994, and 2000
follow-ups. The survey provides information on household and individual
backgrounds, including parental religious affiliation, and on the type of school
attended (public, private non-sectarian, Catholic, non-Catholic religious).
It also includes frequency of church attendance and other measures of the
respondent’s religiosity and allows identification of the zip code in which the
respondent attended eighth grade.
A third is the Early Childhood Longitudinal Study, Kindergarten Class of
1998–1999 (ECLS-K), a longitudinal study of 18,644 children in over 1000
schools. This survey includes the same information as NELS:88 regarding type
of schooling as well as information on the school’s zip code, which permits
merging of the survey data with information on the within-county religious
distribution of the population and other county-level variables described.
However, it does not include information on the religion or religiosity of the
child. Follow-up surveys were administered in spring 1999, fall 1999, spring
2000, spring 2002, spring 2004, and spring 2007 (when the survey cohort
attended eighth grade).
Aggregate Data
School Enrollment. County-level data on private school enrollment by school

type can be obtained from the Private School Survey, a cross-sectional survey of
private elementary and secondary schools that reports enrollment by grade and
identifies the school’s religious affiliation, using a finely defined classification
(United States Department of Education, 2001a).7 This data can be combined
with enrollment data in public schools from the Public Elementary and
Secondary School Universe Survey to obtain a full characterization of county-
level enrollment shares (United States Department of Education, 2001b).
Local Distribution of Religions. The most complete data on religious affili-
ation by county in the United States is available from the 2010 US Religion
Census: Religious Congregations and Membership Study (Grammich et al.
2012), designed and conducted by the Association of Statisticians of American
Religious Bodies (ASARB). It compiles data on the number of congregations
and adherents of 236 religious groups and the share of each religious group
in each county in the United States, totaling 344,894 congregations and
150,686,156 adherents. This publication is the latest in a series of four
decennial religious censuses conducted since 1980, with two earlier censuses in
1952 and 1971. Historical data on membership by denomination is available
for the years 1890–1936 in the American Religious Data Archive (United States
Bureau of the Census 1890, 1906, 1916, 1926, 1936).
References
Berger, Peter. 1967. The Sacred Canopy: Elements of a Sociological Theory of Religion.
New York: Anchor.
Bisin, Alberto, and Thierry Verdier. 2000. Beyond the Melting Pot: Cultural Trans-
mission, Marriage, and the Evolution of Ethnic and Religious Traits. Quarterly
Journal of Economics 115: 955–988.
Broughman, Stephen, and Nancy Swaim. 2013. Characteristics of Private Schools in
the United States: Results from the 2011–2012 Private School Universe Survey. NCES
2013-316, U.S. Department of Education.
7 The types of religiousschools are Catholic, Calvinist, Disciples of Christ, Episcopal, Friends, Evangelical
Lutheran Church in America, Methodist, Presbyterian, African Methodist Episcopal, Amish, Assembly of
God, Baptist, Brethren, Christian (no specific denomination), Church of Christ, Church of God, Church
of God in Christ, Lutheran Church—Missouri Synod, Wisconsin Evangelical Lutheran Synod, Other
Lutheran, Mennonite, Pentecostal, Seventh-Day Adventists, Greek Orthodox, Islamic, Jewish, Latter-Day
Saints, and “all others not listed above.”
Cohen-Zada, Danny. 2006. Preserving Religious Values Through Education: Eco-

nomic Analysis and Evidence from the US. Journal of Urban Economics 60 (3):
372–398.
Cohen-Zada, Danny, and Todd Elder. 2017. Religious Pluralism, Religious Market
Shares and the Demand for Religious Schooling. Journal of Economic Behavior and
Organization. https://doi.org/10.1016/j.jebo.2018.01.015.
Cohen-Zada, Danny, and Moshe Justman. 2003. The Political Economy of School
Choice: Linking Theory and Evidence. Journal of Urban Economics 54: 277–308.
———. 2005. The Religious Factor in Private Education. Journal of Urban Economics
57: 391–418.
———. 2012. Affinity and Tension Between Religious Denominations: Evidence
from Private School Enrollment. Regional Science and Urban Economics 42 (6):
950–960.
Cohen-Zada, Danny, and William Sander. 2008. Religion, Religiosity and Private
School-Choice: Implications for Estimating the Effectiveness of Private Schools.
Journal of Urban Economics 64 (1): 85–100.
Finke, Roger, and Rodney Stark. 2005. The Churching of America, 1776–2005: Winner
and Losers in Our Religious Economy. New Brunswick, NJ: Rutgers University Press.
Grammich, Clifford, Kirk Hadaway, Richard Houseal, Dale Jones, Alexei Krindatch,
Richie Stanley, and Richard Taylor. 2012. 2010 U.S. Religion Census: Religious
Congregations and Membership Study. Lenexa, KS: Association of Statisticians of
American Religious Bodies.
Hofrenning, Stella, and Barry Chiswick. 1999. A Method for Proxying a Respondent’s
Religious Background: An Application to School Choice Decisions. Journal of
Human Resources 34: 193–207.
Hoge, Dean, and David Roozen. 1979. Technical Appendix to Understanding Church
Growth and Decline, 1950–1978. Hartford, CN: Hartford Seminary Foundation.
Hoxby, Caroline. 1998. What Do America’s ‘Traditional’ Forms of School Choice
Teach Us About School Choice Reforms? Federal Reserve Bank of New York
Economic Policy Review 4: 47–59.
Iannaccone, Laurence. 1991. The Consequences of Religious Market Structure: Adam
Smith and the Economics of Religion. Rationality and Society 3: 156–177.
———. 1994. Why Strict Churches Are Strong. The American Journal of Sociology 99
(5): 1180–1211.
National Center for Educational Statistics (NCES). 2013. Digest of Education Statis-
tics. https://nces.ed.gov/programs/digest/d13/tables/dt13_201.10.asp and https://
nces.ed.gov/programs/digest/d13/tables/dt13_205.50.asp.
Snyder, Thomas, and Charlene Hoffman. 1991. Digest of Education Statistics 1990
(NCES 91-660). Washington, DC: National Center for Education Statistics, U.S.
Department of Education.
Snyder, Thomas, Cristobal de Brey, and Sally Dillow. 2016. Digest of Education
Statistics 2015 (NCES 2016-014). Washington, DC: National Center for Education
Statistics, Institute of Education Sciences, U.S. Department of Education.
Stark, Rodney, and James McCann. 1993. Market Forces, Catholic Commitment:
Exploring the New Paradigm. Journal for the Sociology of Religion 32: 111–124.
United States Census Bureau and Bureau of Labor Statistics. 1955–2016. Current
Population Survey. https://www.census.gov/programs-surveys/cps.html.
United States Department of Education. 2001a. Common Core of Data. Public
Elementary/Secondary School Universe Survey, National Center for Education
Statistics, 1999–2000.
———. 2001b. Common Core of Data. Private School Survey, National Center for
Education Statistics, 1999–2000.
Vance, Grant, and Leo Eiden. 1980. Digest of Education Statistics 1980. Washington,
DC: National Center for Education Statistics, U.S. Department of Education.
Zelman et al. v. Simmons-Harris et al. United States Supreme Court No. 00-1751,
Decided June 27, 2002. http://laws.findlaw.com/us/000/00-1751.html.
13
How Luther’s Quest for Education Changed
German Economic History: 9+5 Theses
on the Effects of the Protestant Reformation
Sascha O. Becker and Ludger Woessmann
Five hundred years ago, according to legend, Martin Luther nailed his 95
theses on the door of the castle church of Wittenberg—and changed the
course of history. Cliometric research over the past years has generated several
new insights about the consequences of the Protestant Reformation. One
can observe a veritable digitization boom which changed the way researchers
approached the analysis of economic history. This boost in historical research
with econometric methods has also contributed to the recent growth in
research on the economics of religion (see Iyer 2016). Research into the long-
run effects of the Reformation benefited particularly from the fact that—in
the heartland of the Reformation—the Prussian Statistical Office, and later
the Statistical Office of the German Empire, collected vast amounts of census
data, ever since the first population census in 1816 (see Becker et al. 2014). Most
of this is at the level of counties, some at the more disaggregated city level
and some at the more aggregated province level. Using this newly digitized
data, researchers have uncovered new insights or given statistical grounding
S. O. Becker ()
CAGE, University of Warwick, Coventry, UK
Department of Economics, University of Warwick, Coventry, UK
e-mail: S.O.Becker@warwick.ac.uk
L. Woessmann
Department of Economics, University of Munich, Munich, Germany
ifo Center for the Economics of Education, Munich, Germany

216 S. O. Becker and L. Woessmann
to proposed relationships about the influence Luther had on the course of

German economic history.
Significantly short of 95 theses, in this chapter we propose 9+5 theses
derived from the new insights: nine theses on the importance of the Protestant
Reformation for the education of the population and historical development
in Germany plus five additional theses on what the insights tell us about
the empirical relevance of five classic hypotheses of social science. This is
possible because the analysis of historic data with modern methods opens up
the fascinating possibility to re-assess the hypotheses of the great intellectuals.
They were constrained in their limited access to detailed data from across large
geographic areas and were thus forced to form hypotheses based on proposed
relationships without the possibility to test them in the same rigorous empirical
way that is possible today. It turns out that some of the well-known hypotheses
of famous classics have to be at least re-written or re-interpreted.
Thesis 1: The Reformation Triggered

an Unprecedented Push for Education
in the Population
At the time of Luther’s Reformation, at best 1% of the German population
was able to read and write (Engelsing 1973). Against this background, it
is highly noteworthy that, in 1524, Luther published his pamphlet “To the
Councilmen of All Cities in Germany That They Establish and Maintain
Christian Schools” (Luther 1524), in which he pressured the Protestant rulers
to build and maintain schools. He thus tried to foster the supply side of
education. But he also addressed the parents, in a sermon published in 1530
under the title “A Sermon on Keeping Children in School” (Luther 1530),
trying to encourage the demand side of education.1
For Luther, the Bible was the Word of God and directly related humankind
to God. For this reason, he wanted every good Christian to be able to read
God’s word by themselves. But since being able to read the Word of God
required being able to read at all, the young Protestant Church had to make
sure that the population became more literate. And thus, in regular church
visitations, the Protestant reformers, led by the “Educator of Germany” (Prae-
ceptor Germaniae) Philip Melanchthon, regularly inspected the Protestant
1 Note that the spread of these sermons, as well as of the Reformation in general, was helped by the use of
the recently emerging technology of the printing press (Rubin 2014).
13 9+5 Theses on the Effects of the Protestant Reformation 217
cities and congregations to make sure that they had introduced a proper
education system. Dittmar and Meisenzahl (2018) show that many Protestant
cities introduced city-level ordinances that regulated many aspects of life,
including public education, and that these “constitutions” mattered a lot.
As a consequence, there was a push for education. Using Prussian data,
Becker and Woessmann (2010) show that in 1816, the year of the first Prussian
population census, the school enrollment rate in Protestant-majority counties
was at about two thirds. In Catholic-majority counties, the school enrollment
rate was at less than 50%. Protestantism led to more schooling already before
the Industrial Revolution, typically dated to have started in Prussia around
1830.
In 1871, when the Prussian Statistical Office first collected data about the
literacy rates of the population, they found literacy rates to be at 90% in
Protestant-majority counties, 8 percentage points higher than in Catholic-
majority counties. Since this finding could be the result of reverse causality—
the more educated or education-prone counties were the ones adopting
Protestantism—Becker and Woessmann (2009) use an instrumental-variable
strategy to disentangle cause and effect. As instrument, they use distance
to Wittenberg, the birthplace of the Reformation, to predict the share of
Protestants in a county in 1871. Then, using only the part of the variation
in the share of Protestants predicted by the instrument, they find that indeed
higher shares of Protestantism cause higher literacy rates.2
Thesis 2: Girls Benefited from the Protestant

Education Push Even More Than Boys
As early as 1520, three years after the Protestant Reformation started, Luther
published one of his main pamphlets: “To the Christian Nobility of the
German Nation Concerning the Improvement of the Christian Estate”
(Luther 1520). He writes: “And would to God that every town also had a
girls’ school, in which the girls were taught the Gospel for an hour each day.”
This call was followed up with local church ordinances. In particular, Johannes
Bugenhagen, one of the leading Protestant reformers, in his Church Ordinance
for the city of Braunschweig in 1528, which provided for girls to learn reading,
was to set the standard for subsequent systems. In this ordinance, Bugenhagen
2 Using Swiss data, Boppart et al. (2013) confirm a persistent positive effect of Protestantism on several
education indicators, in particular in areas with conservative milieus, and Boppart et al. (2014) show that
the effect of Protestantism was particularly large for reading skills but also existed in other subjects.
requested that the city should have both four boys’ schools and four girls’
schools. In his Church Ordinance for Wittenberg, he extended the request for
girls’ schooling to writing and calculating. The church ordinances in turn seem
to have led to effective changes in the provision of schooling, as documented by
Green (1979) for the example of the visitations by church officials to the local
parishes in the Electorate (Kurfürstentum) of Brandenburg. In this core state
of what later became Prussia, the number of schools for girls increased more
than tenfold between 1539 (when the Reformation was introduced) and 1600.
The request to also foster education for girls is quite in contrast to Catholic
thinking even a century later: for instance, in Bavaria, the biggest Catholic
state in Germany at the time, there were still strong objections against schools
in the countryside in general as late as 1614 (see Gawthrop and Strauss 1984).
As a consequence of the Protestant Reformers’ attempt to educate also
girls that was unmatched by Catholics, it is not surprising that, in the
nineteenth century, Prussian data show that Protestantism led to a decrease
in the gender gap in basic education in 1816 as well as in the gender gap
in adult literacy in 1871 (Becker and Woessmann 2008). Interestingly, while
compulsory schooling laws closed the gender gap in primary education for
both denominations over the course of the nineteenth century, the pattern
of effects of religious denomination on the gender gap then continues to
show up in secondary and tertiary education in the twentieth and twenty-
first centuries. For instance, in the first year when women were admitted to
university in Prussia in 1908, there were more than eight times as many female
students of Protestant denomination than of Catholic denomination. Even in
contemporary Germany, according to data from the German Socioeconomic
Panel, Protestant women continue to stay ahead of Catholic women in
educational attainment.
Thesis 3: The Education Advantage of Protestants

Accounts for Most of Their Economic Lead
The Protestant quest for education also had economic consequences. A more
highly educated workforce made Protestants more productive. Becker and
Woessmann (2009) find that Protestant counties had higher salaries, higher
income tax receipts, and a higher share of the workforce in manufacturing and
services. They assess how much of these differences in economic outcomes
can be accounted for by the education differences that can be traced back
to the exogenous part of the spread of the Protestant Reformation. Their
point estimates suggest that once income differences are adjusted for literacy
differences, the remaining difference is no longer systematically related to

Protestantism. If education had the same effect on economic outcomes here as
it has been shown to have in other settings, their results suggest that the higher
literacy of Protestant regions can account for at least some of their economic
advantage over Catholic regions, and they are consistent with the hypothesis
that literacy can account for most, or even all, of the advantage.
These consequences of the Protestant Reformation were probably not fully
intended, but rather a side effect of the Reformation. But they show long-
lasting consequences which can even be documented in modern-day data. In
individual-level German data in the year 1997, Protestants have higher levels
of education and, as a consequence, higher incomes.
+1 Human Capital May Be an Alternative to Max Weber’s Specific

Protestant Work Ethic in Explaining Protestant Economic Success
In his celebrated work, The Protestant Ethic and the Spirit of Capitalism, Max
Weber (1904/1905) proposed the thesis that a specific Protestant work ethic
promoted economic success. The results in Becker and Woessmann (2009) suggest
that the alternative hypothesis of a Protestant emphasis on education receives
more empirical support in the data for nineteenth-century Prussia. Once differ-
ences in literacy rates between Protestant and Catholic counties are controlled
for, little difference (if any) remains in economic outcomes. This implies that
there is little room for alternative explanations that are more in line with the
Weber thesis, such as differences in work effort and in the propensity to save.
In the Prussian heartland of the Protestant Reformation, human capital seems to
explain most of the economic success of Protestants compared to Catholics.
Using data from different countries and time periods, others have still found
some support for differences in work ethic between Protestants and Catholics,
though. For instance, Basten and Betz (2013) address the Weber thesis using data
from modern-day Switzerland. Using votes in federal referenda during the late
twentieth century, they show that Protestantism reduces referenda voting for
more leisure, which is consistent with Weber’s theory regarding the effect of
Protestant doctrine on labor-leisure trade-off preferences. Using micro-data from
the German Socioeconomic Panel, Spenkuch (2017) shows that in contemporary
Germany Protestants work longer hours, although he does not find an effect on
wages.
Also in the historical German context, the effect of Protestantism on economic
outcomes appears to have been restricted to areas outside the major cities.
Cantoni (2015) does not find an effect of the Reformation on the growth of city
sizes in Germany, consistent with an interpretation that literacy was relatively
high in cities independent of religious denomination. One thing is clear: the
debate about Weber’s Protestant ethic is going to fascinate scholars for years
to come, as it has done over the last century.
Thesis 4: Education of the Population Played

an Important Role in the Industrial Revolution
Did the Protestant quest for education have consequences beyond the question
of whether Protestant areas were more developed than Catholic areas? The
Industrial Revolution was arguably the biggest transformation of the econ-
omy and has received a lot of attention among economic historians. While
education is considered a key driver of economic development in modern
days, scholars studying the Industrial Revolution, which first started in Britain,
have generally found formal education to be of no importance. Does this
finding equally apply in a follower nation like Prussia which had the luxury of
being able to adopt the discoveries first made in Britain? Becker et al. (2011)
employ a unique school enrollment and factory employment database linking
334 counties from pre-industrial 1816 to two industrial phases in 1849 and
1882. Using pre-industrial education as an instrument for later education and
controlling extensively for pre-industrial development, they find that basic
education is significantly associated with non-textile industrialization in both
phases of the Industrial Revolution. Specifically, they find that the degree of
industrialization in nineteenth-century Prussia would have been lower by one
third had all counties in Prussia only achieved the education level of the least
educated county.
So, it seems that different from the technological leader Britain, the educa-
tion level of the population did in fact play an important role during the first
and second phases of the Industrial Revolution in Prussia.
+2 Joel Mokyr’s Proposition That Education Did Not Play a Role During
the Industrial Revolution in England Does Not Seem to Extend to Fol-
lower Countries
The finding that education played a role in the adoption of the Industrial
Revolution in Prussia might require some qualification of the view quite com-
monly held among economic historians that education was not important during
the Industrial Revolution. The famous quote by Joel Mokyr (1990, p. 240)—“If
England led the rest of the world in the Industrial Revolution, it was despite,
not because of, her formal education system”—has often been taken to apply to
industrialization more broadly. However, in line with Nelson and Phelps (1966, p.
69), the distinction between leader and follower nations seems to be important:
“But probably education is especially important to those functions requiring
adaptation to change. Here it is necessary to learn to follow and to understand
new technological developments.”
Thesis 5: The Immigration of Well-Educated

Huguenots Contributed to Technological
Development Over the Long Run
Economic development not only benefits from a better-educated native
population. Immigration is another contributing factor. In 1685, religiously
persecuted French Huguenots—well-educated Protestants—settled in
Brandenburg-Prussia and compensated for population losses due to plagues
during the Thirty Years’ War. Many of them worked there in textile
production.
Hornung (2014) shows that textile manufacturers in the Prussian cities
where Huguenots had settled were more productive a hundred years later.
The historical incident documents that the immigration of highly qualified
people can sustainably improve technological development and economic
productivity and in fact did so in the Prussian heartland, well before the
Industrial Revolution.
Thesis 6: The Increase in Education Contributed

to Decreasing Fertility and Thereby
to the Demographic Transition
The historical evolution from stagnation to modern economic growth has also
been closely linked to the demographic transition in which the number of
children per family declined significantly (Galor 2011). In the pre-industrial
world described by Thomas Malthus in the late eighteenth century, economic
progress merely increased fertility, so that per capita income stagnated at the
subsistence level.
The demographic transition which led the economy from stagnation to
growth was in turn promoted by the rising education of the population. For
one, a family can “afford” fewer children for a given budget if they invest more
in each child’s education. This inverse relationship between the number of
children and education per child (the so-called child quantity-quality trade-
off ) can be found already before the demographic transition. Becker et al.
(2010, 2012) show that causation between fertility and education runs both
ways, based on separate instrumental-variable models that instrument fertility
by sex ratios and education by landownership inequality and distance to
Wittenberg. Furthermore, education in 1849 predicts the fertility transition
in 1880–1905.
Education matters not only in the context of the child quantity-quality

trade-off, but for an additional reason: better-educated mothers chose to have
fewer children (Becker et al. 2013).
Thesis 7: Protestantism Was Associated

with a Significantly Larger Suicide Rate
While we have so far highlighted positive effects of the Protestant Reformation
and the ensuing increase in education, there are also negative aspects. One
might call this the dark side of the Reformation.3 Becker and Woessmann
(2018) show that suicide rates are significantly larger in Protestant compared to
Catholic counties in Prussia, in 1816–1821 when the Prussian Statistical Office
first collected suicide statistics, and in 1869–1871 when they expanded those
statistics. Using distance to Wittenberg as an instrumental variable, the results
suggest that these effects are causal.4
Taken together with the findings of various positive long-run effects of
the Reformation, the effect of Protestantism on well-being seems to be
neither uniformly positive nor uniformly negative and may affect the average
population differently than the very select subgroup of people who are in a
suicidal state of mind.
+3 Émile Durkheim’s Thesis That Protestant Individualism Increased

Suicide Propensity via Reduced Social Integration Is Corroborated
Émile Durkheim’s book Le Suicide, published in 1897, is regarded by many as a
classic of empirical social science research. Durkheim (1897) not only documented
the higher suicide rates in Protestant countries and regions but also put forward
the thesis that the finding was explained by the religious individualism of
Protestantism. While their unified community embedded Catholics socially, the
Protestant doctrine promoted independent thinking and meant that Protestants
relied less on their religious community but instead sought a more direct connec-
tion with God.
Becker and Woessmann (2018) proposed and analyzed an alternative thesis:
Catholic doctrine emphasized more than the Protestant one that the deadly sin
of suicide prevented access to paradise. Ultimately, however, the results confirm
(continued)
3 Becker et al. (2016), in a survey of causes and effects of the Protestant Reformation, document further
dark sides of the Reformation.
4 Similar results (without instrumentation) are found for Switzerland, in a study by Torgler and Schaltegger
(2014).
Durkheim’s sociological rather than the theological thesis. Among other things,
the suicidal tendency of Protestants is more pronounced in areas with low church
attendance. The strongest effect is thus more likely to be found in areas with
little social integration rather than in areas with high devotion to the Protestant
doctrine. In addition, modern data show that while Protestants still have a
higher suicide rate than Catholics, it is highest among people without a religious
affiliation who are not subject to theological doctrine.
Thesis 8: As a Result of the Reformation, Many

University Graduates Have Chosen Secular Rather
Than Religious Subjects and Occupations
The Reformation not only broke the monopoly of the Catholic Church
(Ekelund Jr. et al. 2002). In a recent study, Cantoni et al. (2018) show that
the Reformation led to a massive reallocation of resources from religious to
secular purposes. Secular authorities acquired enormous amounts of wealth
from monasteries closed during the Reformation. This is particularly true in
Protestant regions and had important consequences.
Graduates of Protestant universities increasingly took secular, especially
administrative, occupations. Protestant university students increasingly stud-
ied secular subjects, especially degrees that prepared students for public sector
jobs, rather than church sector-specific theology.
This effect on the allocation of upper-tail human capital is complemented
by a shift in the sectoral composition of fixed investment. New construction
shifted from religious toward secular purposes, especially the building of
palaces and administrative buildings, which reflected the increased wealth and
power of secular lords, particularly in Protestant regions. Reallocation was
not driven by pre-existing economic or cultural differences. The findings by
Cantoni et al. (2018) indicate that the Reformation played an important causal
role in the secularization of the West.
Thesis 9: The Educational Boost Contributed

to the Decline of Church Attendance and Thus
to Secularization
While Cantoni et al. (2018) give unique insights into secularization in the
immediate aftermath of the Reformation, it is insightful to look at the long-
run consequences of higher education for church attendance as a way to
understand the role the church still has in the nineteenth and early twentieth
centuries. Unique data about participations in Holy Communion collected as
part of the statistics of the Protestant regional churches of Germany show a
significant decline in the Protestant population in church attendance at the
turn of the twentieth century.
Becker et al. (2017) show that an important cause of this decline in the
importance of the church in everyday social life was the increasing education
of the population in secondary schools. It can be shown that the educational
boost preceded the decline in church attendance, and not vice versa. In
this respect, the expansion of education of Protestantism may have laid the
foundation for the decline of churchliness in its own ranks.
+4 The Empirical Findings Are Consistent with the Views of Sigmund

Freud on the Relationship Between Education and Religiosity
In his book The Future of an Illusion, Sigmund Freud (1927)—like David Hume
and others before him—took the view that education and science would lead
to a decline in religion (see McCleary and Barro 2006). While previous analyses
questioned this connection between education and secularization, the new
findings by Becker et al. (2017) support this thesis.
The relationship between educational expansion and decline in church atten-
dance is as strong for classical humanist grammar schools as it is for the newly
emerging upper secondary schools whose curriculum focuses strongly on natural
sciences. In this sense, general critical thinking (which may undermine belief in
the institutionalized church) seems to have been a more important mechanism
for the context than specific scientific knowledge of facts of natural sciences.
+5 Religion Did Not Function as the “Opium of the People” Propagated

by Karl Marx in the Sense That Rising Income Would Have Contributed
to Declining Church Attendance
The effect of education on secularization contrasts with the fact that Becker and
Woessmann (2013) do not find a causal effect of rising income on the decline of
church attendance. The missing effect of improved material conditions suggests
that church services were not only in demand to provide religious consolation. In
this sense, religion was probably not, as propagated by Karl Marx (1844, p. 72),
the “opium of the people” which was needed only to alleviate the misery of poor
economic circumstances.
The Protestant Reformation continues to throw its shadows. The past

decade has seen a huge surge in research studying the short- and long-run con-
sequences of the Reformation on socioeconomic outcomes. We complement
Luther’s 95 theses with a set of 9+5 new “theses.” Nine “theses” summarize
recent research on the effects of the Protestant Reformation on education and
various other outcomes, including economic development, fertility, suicide,
and secularization. Based on these insights, five “theses” additionally challenge
several classic hypotheses of social science, ranging from Marx’s “Opium of
the people” to Durkheim’s “Suicide” and Weber’s “Protestant ethic.” There
is no doubt that the future will bring further insights in one of the most
transformative episodes of European history.
References
Basten, Christoph, and Frank Betz. 2013. Beyond Work Ethic: Religion, Individual,
and Political Preferences. American Economic Journal: Economic Policy 5 (3): 67–91.
Becker, Sascha O., and Ludger Woessmann. 2008. Luther and the Girls: Religious
Denomination and the Female Education Gap in Nineteenth-Century Prussia.
Scandinavian Journal of Economics 110 (4): 777–805.
———. 2009. Was Weber Wrong? A Human Capital Theory of Protestant Economic
History. Quarterly Journal of Economics 124 (2): 531–596.
———. 2010. The Effect of Protestantism on Education Before the Industrialization:
Evidence from 1816 Prussia. Economics Letters 107 (2): 224–228.
———. 2013. Not the Opium of the People: Income and Secularization in a Panel
of Prussian Counties. American Economic Review 103 (3): 539–544.
———. 2018. Social Cohesion, Religious Beliefs, and the Effect of Protestantism on
Suicide. Review of Economics and Statistics 100 (3): 377–391.
Becker, Sascha O., Francesco Cinnirella, and Ludger Woessmann. 2010. The Trade-
Off Between Fertility and Education: Evidence from Before the Demographic
Transition. Journal of Economic Growth 15 (3): 177–204.
Becker, Sascha O., Erik Hornung, and Ludger Woessmann. 2011. Education and
Catch-Up in the Industrial Revolution. American Economic Journal: Macroeco-
nomics 3 (3): 92–126.
Becker, Sascha O., Francesco Cinnirella, and Ludger Woessmann. 2012. The Effect
of Investment in Children’s Education on Fertility in 1816 Prussia. Cliometrica 6
(1): 29–44.
———. 2013. Does Women’s Education Affect Fertility? Evidence from Pre-
Demographic Transition Prussia. European Review of Economic History 17 (1):
24–44.
Becker, Sascha O., Francesco Cinnirella, Erik Hornung, and Ludger Woessmann.
2014. iPEHD – The ifo Prussian Economic History Database. Historical Methods
47 (2): 57–66.
Becker, Sascha O., Markus Nagler, and Ludger Woessmann. 2017. Education and
Religious Participation: City-Level Evidence from Germany’s Secularization Period
1890–1930. Journal of Economic Growth 22 (3): 273–311.
Boppart, Timo, Josef Falkinger, Volker Grossmann, Ulrich Woitek, and Gabriela
Wüthrich. 2013. Under Which Conditions Does Religion Affect Educational
Outcomes? Explorations in Economic History 50 (2): 242–266.
Boppart, Timo, Josef Falkinger, and Volker Grossmann. 2014. Protestantism and
Education: Reading (the Bible) and Other Skills. Economic Inquiry 52 (2): 874–
895.
Cantoni, Davide. 2015. The Economic Effects of the Protestant Reformation: Testing
the Weber Hypothesis in the German Lands. Journal of the European Economic
Association 13 (4): 561–598.
Cantoni, Davide, Jeremiah Dittmar, and Noam Yuchtman. 2018. Reformation and
Reallocation: Religious and Secular Economic Activity in Early Modern Germany.
Dittmar, Jeremiah, and Ralf Meisenzahl. 2018. Public Goods Institutions, Human
Capital, and Growth: Evidence from German History. Review of Economic Studies.
Forthcoming.
Durkheim, Émile. 1897. Le suicide: étude de sociologie. Paris: Félix Alcan. (Suicide: A
Study in Sociology. Trans. John A. Spaulding and George Simpson. Glencoe, IL:
The Free Press, 1951).
Ekelund, Robert B., Jr., Robert F. Hébert, and Robert D. Tollison. 2002. An
Economic Analysis of the Protestant Reformation. Journal of Political Economy 110
(3): 646–671.
Engelsing, Rolf. 1973. Analphabetentum und Lektüre: Zur Sozialgeschichte des Lesens in
Deutschland zwischen feudaler und industrieller Gesellschaft. Stuttgart: Metzler.
Freud, Sigmund. 1927 [1961]. The Future of an Illusion. Ed. J. Strachey. New York:
W. W. Norton. [Original Version (in German) Published in: Psychoanalytischer
Verlag, 1927].
Galor, Oded. 2011. Unified Growth Theory. Princeton, NJ: Princeton University Press.
Gawthrop, Richard, and Gerald Strauss. 1984. Protestantism and Literacy in Early
Modern Germany. Past and Present 104: 31–55.
Green, Lowell. 1979. The Education of Women in the Reformation. History of
Education Quarterly 19 (1): 93–116.
Hornung, Erik. 2014. Immigration and the Diffusion of Technology: The Huguenot
Diaspora in Prussia. American Economic Review 104 (1): 84–122.
(2): 395–441.
Luther, Martin. 1520. An den christlichen Adel deutscher Nation von des christlichen
Standes Besserung (To the Christian Nobility of the German Nation Concerning
the Reform of the Christian Estate). In Dr. Martin Luthers Werke: Kritische
Gesamtausgabe, vol. 6. Weimar: Verlag Hermann Böhlhaus Nachfolger, 1888.
———. 1524. An die Ratsherren aller Städte deutschen Landes, dass sie christliche
Schulen aufrichten und halten sollen (To the Councilmen of All Cities in Germany
That They Establish and Maintain Christian Schools). In Dr. Martin Luthers
Werke: Kritische Gesamtausgabe, vol. 15. Weimar: Verlag Hermann Böhlhaus Nach-
folger, 1899.
———. 1530. Eine Predigt, daß man Kinder zur Schule halten solle (A Sermon on
Keeping Children in School). In Dr. Martin Luthers Werke: Kritische Gesamtaus-
gabe, vol. 30, Part 2. Weimar: Verlag Hermann Böhlhaus Nachfolger, 1909.
Marx, Karl. 1844. Zur Kritik der Hegel’schen Rechtsphilosophie: Einleitung. In
Deutsch-Französische Jahrbücher, ed. Arnold Ruge and Karl Marx, 71–85. Paris:
Bureau der Jahrbücher.
McCleary, Rachel M., and Robert J. Barro. 2006. Religion and Economy. Journal of
Economic Perspectives 20 (2): 49–72.
Mokyr, Joel. 1990. The Lever of Riches: Technological Creativity and Economic Progress.
Oxford: Oxford University Press.
Nelson, Richard R., and Edmund S. Phelps. 1966. Investment in Humans, Tech-
nological Diffusion, and Economic Growth. American Economic Review 56 (2):
69–75.
in the Reformation. Review of Economics and Statistics 96 (2): 270–286.
Spenkuch, Jörg L. 2017. Religion and Work: Micro Evidence from Contemporary
Germany. Journal of Economic Behavior & Organization 135: 193–214.
Torgler, Benno, and Christoph Schaltegger. 2014. Suicide and Religion: New Evi-
dence on the Differences Between Protestantism and Catholicism. Journal for the
Scientific Study of Religion 53: 316–340.
Weber, Max. (1904/1905). Die protestantische Ethik und der “Geist” des Kapi-
talismus. Archiv für Sozialwissenschaft und Sozialpolitik 20: 1–54 and 21: 1–110.
Reprinted in: Gesammelte Aufsätze zur Religionssoziologie, 1920: 17–206 [English
Translation: The Protestant Ethic and the Spirit of Capitalism. Trans. Talcott Parsons,
1930/2001, London: Routledge Classics].
14
Hindu-Muslim Violence in India: A Postscript
from the Twenty-First Century
Anirban Mitra and Debraj Ray
Introduction
The importance of religious conflict today can hardly be overstated. The
appalling situation in Syria—stemming from the activism of the Islamic State
(ISIS)—is just one cruel reminder of the utter devastation that religious
extremism can wreak. The tragic humanitarian crisis in Myanmar involving
the displacement of the Rohingya Muslims has strong religious overtones, as
do the attacks on Muslims and Christians by Buddhist nationalist groups in
Sri Lanka.
India, of course, is no exception. Following the election of the National
Democratic Alliance in 2014, headed by the Bharatiya Janata Party (or BJP),
violence—or the threat of violence—fomented by a sense of Hindu religious
nationalism has been in the news on a daily basis; for an example, look no
further than the spate of lynchings that have followed on the suspicion of
Ray acknowledges funding from the National Science Foundation under grant number SES-1629370.
A. Mitra ()
University of Kent, Canterbury, UK
e-mail: A.Mitra@kent.ac.uk
D. Ray
New York University, New York, NY, USA
University of Warwick, Coventry, UK

230 A. Mitra and D. Ray
beef consumption or trade. Consider the lynching of Mohammad Akhlaq,

a Muslim man, in Dadri (a place barely 50 km from the Indian capital) on
September 28, 2015. A panel from the National Commission for Minorities
in India investigated this event and concluded that this lynching—over rumors
that he kept and consumed beef—was premeditated. Additionally, it said a
Hindu temple had been used to plan the attack. Mr Akhlaq’s killing sparked
furious debate about religious tolerance, with some criticizing Prime Minister
Narendra Modi (from the BJP) for not condemning the attack sooner.
Two weeks later he called it “sad and undesirable”.1 Certainly, the country
isn’t a newcomer to religious violence by any means. Recurrent episodes of
Hindu-Muslim conflict in India (going back to the Partition of the Indian
subcontinent in 1947 and earlier) have continued through the second half of
the twentieth century, accounting for over 7000 deaths over 1950–2000, and
many more at the time of Partition. So—and beginning with the violence
unleashed in the state of Gujarat in 1992—twenty-first-century India is just
more of the same, and in its pervading sense of menace and repression, possibly
much more.
Scholars from various disciplines have approached the issue of Hindu-
Muslim violence from different perspectives. There is the view espoused by
Huntington and others that it is all about “a clash of civilizations”. Hopeless
and dispiriting as it may be, this primordialist view of violence is hard to fully
escape in the wake of the sheer irrational brutality that has been on display in
some of the violent encounters. But a bewildered focus on irrationality alone
runs the risk of obscuring more calculated realities. For instance, Wilkinson
(2004) has explored how the degree of political competition in state-level
elections can affect the extent to which (religious, ethnic) minorities become
important electorally that, in turn, can and does influence the attitude of state
governments toward riot containment. Varshney (2003) has highlighted the
role of social capital in a multi-cultural society in influencing the pattern of
such sectarian violence. In a related vein, Jha (2013) provides evidence that the
degree to which medieval Hindus and Muslims could provide complementary,
non-replicable services and a mechanism to share the gains from exchange has
resulted in a sustained legacy of ethnic tolerance in some South Asian towns.
Specifically, he finds that medieval ports, despite being more ethnically mixed,
were five times less prone to Hindu-Muslim riots between 1850 and 1950, two
centuries after Europeans disrupted Muslim overseas trade dominance, and
remained half as prone between 1950 and 1995.
1 See the following link for more details: http://www.bbc.com/news/world-asia-india-34599849.

14 Hindu-Muslim Violence in India: A Postscript from the Twenty-First Century 231
In our earlier work (Mitra and Ray 2014), we have followed the line
espoused in Engineer (1984, 1987, 1994) and others (see, e.g., Upadhyaya
1992; Rajgopal 1987; Khan 1992; Bagchi 1990, and Das 2000) by highlighting
an economic component to Hindu-Muslim conflict. We model inter-group
conflict driven by economic changes within groups and make the following
prediction: if group incomes are low, increasing group incomes raises violence
against that group and lowers violence generated by it. Using data collected by
Varshney and Wilkinson from 1979 to 2000, we show that regional Hindu-
Muslim violence rises in response to an increase in the regional average of
Muslim incomes. The opposite is true when regional Hindu incomes rise.2
These empirical findings are of interest in their own right, but we can go
further by using the theory as a device for the interpretation of the empirical
patterns. In our work, we suggest that Hindu groups have largely been the
aggressors in Hindu-Muslim violence in India, or at least in Hindu-Muslim
violence driven by instrumental, specifically economic, considerations.
In this chapter, we revisit and extend the core issues studied in Mitra and
Ray (2014). The main reason behind this retrospection is to check if the robust
empirical patterns recorded in Mitra and Ray (2014) persist once we consider
a longer time frame extending into the twenty-first century. This is not simply
driven by a concern for greater external validity—there is a more subtle issue
at play. In the Indian context, the new millennium witnessed a marked change
in the political arena at both the national and sub-national levels. Gone were
the days of hegemony of the Indian National Congress (INC). The challenge
to INC’s dominance chiefly came from the Bharatiya Janata Party (henceforth,
BJP) and its allies. BJP’s persistent appeal to “rejuvenate” Hindu religious
pride most certainly put a strain on the civic engagement between Hindus
and Muslims in several parts of the country. Now this could potentially affect
the dynamics of Hindu-Muslim violence; in particular, this could alter—or,
potentially, nullify—the economic effects uncovered in Mitra and Ray (2014).
Therefore, in this steadily changing political climate, it becomes imperative to
re-assess the potency of the previously identified economic effects.
There is a related secondary goal, which is to explore the role of political
factors in shaping the patterns of Hindu-Muslim conflict. To be sure, we did
account—to an extent—for the influence politics has on religious riots. But
that was in the spirit of a robustness exercise—only to investigate if the links
2 The effect holds even after controlling for literacy, inequality within the groups, and urbanization levels
for each region.
between Hindu-Muslim incomes and inter-religious violence were salient,

controlling for any potential links between politics and such violence.
Specifically, in this exercise, we investigate if there is a correlation between
the rise of the BJP and its allies on the national political scene and religious
violence. It is worth reiterating that over the period 1979–2000, the presence of
“Hindu” political activism was highly localized and relatively minor in scale. It
is only in the last 15 years or so that the BJP and its alliance partners have been
prominent on the national political scene. We are in a position to make some
preliminary observations on these newer developments, thanks to an extended
dataset which now goes all the way up to 2010.3
We explore the role of the political factors in the following two ways:
(a) We create a measure of BJP presence in each region and directly include it
in our regression analysis, and
(b) we perform the analysis with and without the district of Ahmedabad
(situated in the state of Gujarat) and contrast the results.
The idea of measuring BJP presence (details below) is to create an indicator

for the pro-Hindutva political forces operating in the region. It should be
noted that the BJP is not the sole political player in terms of the ideology and
rhetoric relating to the establishment of a “Hindu-rashtra” (roughly translates
as “nation of the Hindus”). But as a first pass, focusing on the BJP does seem
reasonable. We briefly discuss two recent studies whose findings further justify
our approach.
Nellis et al. (2016) analyze data between 1962 and 2000 to show that the
election of Congress legislators in close elections contained riot incidents by
about 10% over this period. They focus on assembly constituencies that faced
close contests and then compare differences in rioting in constituencies where
the Congress won to those constituencies where the Congress lost elections to
arrive at their findings. They also argue that rioting polarizes voters and harms
the electoral prospects of the Congress party in ensuing elections, while it has
the opposite effect on the fortunes of its principal rival, the BJP, which gains
vote share following such polarization. Iyer and Shrivastava (2018) analyze the
effect of Hindu-Muslim riots on state election results. They combine data on
geo-coded riots with data on state elections and on demographics and public
goods provision for the period 1981 through 2001. To safeguard against the
3 We are grateful to Lakshmi Iyer for making the data on Hindu-Muslim violence for the period 2001–2010
available to us.
possible endogeneity of riot occurrence to various ambient conditions, they

introduce a new instrument which draws upon the random variation in the
day of the week that Hindu festivals fall on, as set by a lunar calendar. The
probability of a riot increases if a Hindu festival falls on a Friday, the holy day
for Muslims. In this manner, they are able to isolate the causal effect of riots
on electoral results. They find that riots occurring in the year preceding an
election increase the vote share of the BJP by a minimum of five percentage
points.
The reason behind isolating the district of Ahmedabad stems from the
conjunction of two factors—the place has been a stronghold of the BJP since
the party assumed prominence on the national political scene, and it has been
the hotbed of Hindu-Muslim rioting for several decades with an intensification
in the early twenty-first century. Of course, one could argue that this is true
for all of the state of Gujarat and not necessarily Ahmedabad. Our counter-
argument is this: first, not all of Gujarat is equally politically inclined toward
the BJP as Ahmedabad is, and second, the extent of religious violence in some
places of Gujarat (like Surat, see Jha 2013) is hardly above the national average
and hence much lower than that in Ahmedabad.
In the next section, we discuss the data and the empirical strategy and then
present our findings.
Empirical Analysis
We begin by describing our various datasets.
Data
The baseline Varshney-Wilkinson data (Varshney and Wilkinson 2004) on

Hindu-Muslim violence runs from 1950 to 1995. We extended this dataset
by a period of five years, that is, from 1996 to 2000. Subsequently, the data
have been extended by another ten years and hence bring us to the very end of
2010.4 All in all, we have conflict data from 1950 to 2010 at our disposal, but
we limit ourselves to the sub-period 1979–2010. The main reason for doing
4 The data on Hindu-Muslim violence for the period 2001–2010 has been obtained from Lakshmi Iyer. In
conducting these extension exercises, the same data collection protocol has been adhered to as followed in
the construction of the original dataset, and the source has remained limited to newspaper reports from
The Times of India.
so is the non-availability of reliable data on economic conditions (by religious

group) for earlier years.
We use three different count measures from the dataset: the number of
people killed or injured (“casualties”), the number of people killed, or the
number of riot outbreaks over the period. In all cases, we take aggregates over
a five-year period in each location.
As for economic data, large-scale household surveys are conducted quin-
quennially in India as part of the National Sample Surveys (NSSs). The survey
rounds cover the entire nation and capture monthly expenditure incurred by
sample households for domestic consumption.5 The earliest “thick” round
that provides spatially disaggregated economic information by religion is
the 38th round (1983).6 We use five such “thick” rounds: the 38th (1983),
the 43rd (1987–1988), the 50th (1993–1994), the 55th (1999–2000), and
the 61st (2004–2005). For all of these rounds, there is information on the
religious affiliation of the household, or more precisely, that of the head of the
household. We can therefore compute the per-capita monthly expenditures for
Hindu and Muslim households.
We are further restricted by the relative lack of spatial disaggregation in the
38th and the 50th rounds, which do not permit identification of the surveyed
households all the way down to the district level. To use all five rounds (and
thereby exploit the panel structure), we aggregate the Varshney-Wilkinson
dataset up to the regional level in India. “Regions” are areas that lie midway
between the state and the district. We do so for 55 such regions, which together
span 14 major Indian states and account for more than 90% of the Indian
population.7
Apart from the economic variable, we seek some measure of BJP strength in
each region. To do so, we use as a measure the share of regional Lok Sabha seats
won by the BJP. This is the same variable as in Mitra and Ray (2014). There
5 Unfortunately, a well-known problem in the case of the NSS is that we do not have income data on a
nationwide scale, and expenditure is the closest we can get.
6 NSSs which occur annually utilize smaller samples and hence are referred to as “thin” rounds. However,
the rounds performed quinquennially draw upon larger samples (about 120,000 households per survey),
hence the term “thick”.
7 We leave out border states with their own specific sets of problems: Jammu and Kashmir and Himachal
Pradesh in the north, and the north-eastern states of Assam, Arunachal Pradesh, Manipur, Meghalaya,
Nagaland, Sikkim, and Tripura. There are two specific issues with these areas: (a) NSS does not survey
all regions within these states (owing to hilly terrain, safety issues, national security reasons due to border
skirmishes, etc.), and (b) for the border states, it is sometimes difficult to tell whether a reported riot
is indeed civilian in nature or due to the Army clashing with extremist groups. In addition, the north-
eastern states (which happen to be sparsely populated) have an insignificant Muslim population: they are
primarily Hindus, Christians, Buddhists, and Scheduled Tribes. So even in the violence dataset there are
almost no reports of riots there.
the variable had no significant effect on conflict. To ensure consistency, we

have used the same variable here. The idea is to check if this variable behaves
any differently now owing to the addition of the later rounds. Additionally,
given that the Lok Sabha is a national body and given that the BJP is a national
party with a distinct ideology, it makes sense to see how the BJP MPs affect the
climate for riots. It is of course possible to use other measures—for example,
the state assembly shares—but then there are two distinct issues which one
needs to account for:
(a) the party’s ideology/stance/policy could vary from state to state especially
given the caste and religious composition of the electorate. In other words,
BJP at the center and BJP at the state could mean different things for riots,
and
(b) state-level competition, by itself, can have an independent effect on riots (a
la Wilkinson (2004)). So when using BJP’s assembly seatshare, one could
be picking up some state-level competition effects. It might be difficult to
isolate a pure “BJP effect” from the state-level competition effect.
Therefore, keeping these considerations in mind, we proceed with this rather

conservative measure of BJP presence, namely, the share of Lok Sabha seats
won by the party in the region.
We also employ a number of controls: population by region; religious
polarization across Hindus and Muslims; the literacy rate; the completion rate
for primary education; urbanization, calculated as the percentage of urban
households in the region; and Gini coefficients as controls for expenditure
inequality among Hindu and Muslims.
Empirical Specification
We are interested in the effect of changes in Hindu and Muslim expenditures

within an area on different measures of Hindu-Muslim conflict within that
same area. This motivates the baseline Poisson specification that we use:
E(Counti,t |Xit , γ i ) = γ i exp(Xit β + τ t )
where i indexes region while t indexes time, X is the vector of explanatory

variables and controls, and γ i and τ t stand for region and time fixed effects in
the panel regressions.
We also use the Negative Binomial and OLS specifications as robustness

checks. In the case of OLS, to avoid losing observations in cases where reported
conflict is zero, we add a very small number (0.01) to the total count variable,
so that the dependent variable for the OLS regressions is actually ln(count +
0.01). That is, our OLS specification is:
ln(countit + 0.01) = γ i + τ t + Xit β + errorit .
Among the important variables in X are, of course, Muslim and Hindu per-
capita expenditures (our proxies for per-capita income), and in some variants
their ratio. We will also pay special attention to the regional share of the BJP in
the Lok Sabha. Apart from these core variables, population and some measure
of Muslim presence are always included as controls in every specification
(despite the region fixed effects, these are important variables that could
potentially vary with time). Muslim “presence” is measured in two ways: we use
either the share of Muslim households in the region, or a measure of Hindu-
Muslim polarization along the lines proposed by Esteban and Ray (1994) and
Montalvo and Reynal-Querol (2005).8 To be sure, in all the regressions we
either control for Muslim percentage or religious polarization but never both
simultaneously. The correlation between these two variables is very high (about
0.97).9
We also control for inequality (in terms of consumption expenditure)
among Hindu and Muslims, as our predictions pertain to balanced increases
in income for either group.
The basic controls are constructed using the data from the NSS rounds. In
some specifications, we also use an expanded set of controls, to be described
below. In all the specifications, expenditures and population are entered
logarithmically, and all other controls are brought in linearly.
We look at the effect of group expenditure variables on Hindu-Muslim
conflict starting the year right after the corresponding expenditure round.
Specifically, expenditures from the 38th round (1983) are matched with con-
flict during 1984–1988, the 43rd round (1987–1988) expenditures are matched
with conflict during 1989–1993, the 50th round (1993–1994) expenditures are
matched with conflict during 1994–1998, the 55th round (1999–2000) are
8 The degree of religious polarization for a region is defined by 4 sj2 (1 − sj ) for j = H, M where H
denote Hindus and M Muslims and sj denotes the population share of j in the region.
9 In some areas, there are other dominant religious groups (like Sikhs in Punjab), so that Muslim percentage
and Hindu-Muslim polarization measure different things. But these cases are exceptions rather than the
rule.
052
● 222 ●
● 052
121
●
−17.5 224
053 ●
●
053 ●
091 053 121

093 123 124 022 222 ●●
●
222 123114 ●
● ●
●
●
043113 ●
055 ●
055 052
● ● 121
052 121
●
●
●
022 ●
●
● 223 125122 ●● ●
●
●
125 232 042 051 ●
121 223 ●
●
● 042 ● 042 093 ●
● ●
28 124094222224 101
123 182
055 ●
● ●
●
●
●
094 022 123
●
116233121 055 093

● ●
094 ●
094 ●
● ●
052
log Casualties (residual)

125 182 ● ●
223 114
●
●
022 184 203 ● ● ●
●
113 055114 233 ●
224 182093
● ● ● ●
●
123222
182 093 163 114
022
122
021 ● ● ●
● ●
●
−20.0 116 ● 053
●
●
222 022
●
● ●
●
●
052 121
052 094 042 051
● 101 ●
● ●
●
●
●
092 223
● ●
042 051 201 091

● ● ●
051 123
054 224 ● ●
161 201 041 124 182

●
222 181
●
● ● ●
●
181 051
●
223 ● ● ● ● ●
093 124
●
124224
●
●
●
041 041 ● 123
●
●
●
054 ● ●
●
023 201 184 222 ● ● ●
222 ● 162 ●
●
233124102 203
●
●
201 094 ●
●
●
●
113 041 204 223 181 125 224 ●

● ●
●
●
043 053 113

● ●
● ●
042 163 123
116 053
● ●
053 022 182093

233
093022 114 051
092 ●
●
●
116 091 114 ● ● ● ●
122 102
●
●
●
●
● ● ● ●
● ●● ● 222 091 053 ●
● ●
●
094
223 021 ●
125 022 ●
●
●
122 041 117 ●●

181 111 ●
●
051 121 223

●
24 224 124 042 ●

201184
●
043
041 ●
● ●
234 162
● ● ●
123 ●
125
● ●
●
●
−22.5 092204 101 ●
093 ● ●
●
●
224 ● 102 224 054 054
125 171 182
233091 182
● ● ●
121
223 055 181 091 ● ● ●
●
055 183 183 114
022
● ●
● ●
● ●
● ●
●
094 123 092161

●
● ●
●
●
●
024
●
184 055 055 121 ●
●
●
●
041 203122
● ● ●
232 114 ●
●
●
● ● 101 051 052 ●
●
●
113 ●
052
●
203
● ●
116 171 043 093

124102
● ●
042
● 054
116233 111 ● ●
● ● ● ● 061 223054181 122 ● ●
●
201 ●
●
203 ●
●
● 021 ●
●
●
●
041
061 183
043 ● 092 021 043 181 051 ●
●
101 113 ● ●
125234055
●
051 ●
●
●
043 ●
●
●
181
054
●
●
203 041 ●
●
023
●
● 041 ●
●
●
021183 ● −25.0 052 224 101 ● ●
●
092
●
117 021
20 024 ●
●
●
●
043
● ● 054
9 10 11 12 9 10 11 12
log Muslim expenditure log Hindu expenditure
(a) Conflict and Muslim Expenditure (b) Conflict and Hindu Expenditure
Fig. 14.1 Conflict and per-capita expenditure. Panel (a) plots the residual of casualties
against (log) Muslim expenditures after region, time, and (log) Hindu expenditure
effects have been removed, in the five-year period following expenditures. Each line
segment connects five data points for a region. Panel (b) plots the analogous graph for
(log) Hindu expenditures
matched with conflict during 1999–2003, while the 61st round (2004–2005)
expenditures are matched with conflict during 2004–2008.
All variations utilize both region and time fixed effects, just as the baseline
Poisson model does.
Before turning to the results from the regression analysis, we would like to
draw the reader’s attention to Fig. 14.1. These graphs essentially represent the
key correlations we seek to explore in our regression framework. In panel (a)
of the figure, we look at the association between “Casualties” (which is the
number of people either killed or injured) and average Muslim expenditure at
the regional level. To arrive at that regional plot, we first estimate a Poisson
regression where Casualties are regressed on Muslim and Hindu expenditures
(log) and region and time dummies. Using the estimates from this regression,
we are able to remove region, time, and (log) Hindu expenditure effects from
the actual Casualties variable so as to arrive at the residual of Casualties. As
shown in panel (a), this residual is positively correlated with (log) Muslim
expenditures. Panel (b) is created by a similar process where we now remove
(log) Muslim expenditure effects rather than (log) Hindu expenditure effects
to arrive at the residual of Casualties. As the figure depicts, the correlation
between this residual of Casualties and average Hindu expenditure is negative.
It is precisely these asymmetric relationships (between conflict and expendi-
tures) that we seek to explore more carefully in our empirical analysis.
Results
First we present some results using all 55 regions (as in Mitra and Ray 2014)
and using all 5 NSS rounds.
Table 14.1 contains some Poisson fixed effects regressions where the depen-
dent variable is “Casualties” (= killed+injured). In columns 1 and 2, the
coefficients on Muslim expenditure are positive and significant at the 5%
and 10% levels, respectively. However in the specification with the full set
of controls (column 3), the expenditure terms are no longer significant. The
next three columns have the ratio of Muslim-to-Hindu expenditures in place
of Muslim and Hindu expenditures. In columns 4 and 5, the coefficients
on the ratio of Muslim-to-Hindu expenditures are positive and statistically
significant—much like their counterparts in columns 1 and 2. However, with
the full set of controls (column 6), the statistical significance on the ratio of
Muslim-to-Hindu expenditures is well below 10%. For brevity, we do not
report the results from the Negative Binomial and OLS models. They are
qualitatively similar.10
Hence, there seems to be some flavor of the results from Mitra and Ray
(2014) but it is somewhat diluted.
Urban Households
Next we conduct the analysis by restricting attention to only urban house-

holds. Here we eliminate instances of conflict which are explicitly stated to
have occurred in rural areas. However, it should be noted that such cases
are relatively few in the violence dataset. Some of the results are presented
in Table 14.2.
In Table 14.2, the dependent variable in all the regressions is “Casualties”
which is the number of people killed or injured. In the first three columns,
the coefficient on Hindu expenditures is negative—in addition, it is also
statistically significant in columns 2 and 3. In contrast, the coefficient on
Muslim expenditure is positive in all three columns; additionally, it is statisti-
cally significant in columns 1 and 3. Therefore, in the specification with the
full set of controls (column 3), an increase in Muslim expenditures exerts a
positive and statistically significant effect on Hindu-Muslim conflict, while
rising Hindu expenditures tend to drive it down. In columns 4–6, the main
variable of interest is the ratio of Muslim-to-Hindu expenditures. In each of
10 These results are available from the authors upon request.

Table 14.1 The effect of Hindu and Muslim expenditures on regional conflict: Poisson FE regressions
[1] [2] [3] [4] [5] [6]
H pce −1.742 −1.243 −2.298
(0.290) (0.450) (0.115)
M pce 2.259** 1.683* 1.425
(0.028) (0.083) (0.282)
M/H 2.210** 1.644* 1.483
(0.027) (0.084) (0.231)
Average per-capita exp. 0.613 0.444 −0.886
(0.688) (0.771) (0.628)
Pop −0.552 −0.218 −0.202 −0.599 −0.231 −0.186
(0.765) (0.904) (0.908) (0.750) (0.899) (0.915)
RelPol 0.363 0.355 0.746 0.399 0.379 0.699
(0.918) (0.916) (0.781) (0.909) (0.910) (0.798)
Literacy 0.053** 0.051** 0.053** 0.051**
(0.021) (0.014) (0.021) (0.014)
Urban −0.006 −0.012 −0.006 −0.012
(0.569) (0.445) (0.578) (0.447)
Gini H 5.063 5.042
(0.342) (0.343)
Gini M 0.864 0.952
(0.786) (0.768)
BJP LS seatshare 0.310 0.387 0.356 0.319 0.386 0.358
(0.616) (0.468) (0.491) (0.602) (0.470) (0.490)
Log-likelihood −6934.71 −6576.82 −6475.35 −6931.57 −6576.64 −6475.09

Number of observations 230 230 230 230 230 230
Sources and Notes. Varshney-Wilkinson dataset on religious riots, National Sample Survey 38th, 43rd, 50th, 55th, and 61st rounds. All
counts over a five-year period starting immediately after the expenditure data. Dependent variable is regional casualties (killed+injured).
14 Hindu-Muslim Violence in India: A Postscript from the Twenty-First Century
Robust standard errors clustered by region, corresponding p-values in parentheses. Time dummies included in all regressions. *Significant
at 10%, **significant at 5%, ∗∗∗ significant at 1%
239
240
Table 14.2 The effect of Hindu and Muslim expenditures on regional conflict: Poisson FE regressions (urban households)
[1] [2] [3] [4] [5] [6]
H pce −1.876 −2.691* −3.162***
(0.171) (0.060) (0.008)
M pce 1.402* 1.493 1.844*
(0.077) (0.109) (0.085)
M/H 1.446* 1.662* 1.952**
(0.062) (0.061) (0.046)
Average per-capita exp. −0.704 −1.395 −1.631
(0.547) (0.264) (0.183)
A. Mitra and D. Ray
Pop −0.774 0.158 0.172 −0.763 0.194 0.207

(0.517) (0.899) (0.890) (0.529) (0.878) (0.868)
RelPol 0.056 1.246 1.442 −0.048 1.089 1.250
(0.974) (0.520) (0.386) (0.978) (0.568) (0.458)
Primary edu. 0.079*** 0.080*** 0.080*** 0.080***
(0.003) (0.003) (0.003) (0.002)
Gini H 2.991 3.007
(0.533) (0.513)
Gini M −1.092 −0.538
(0.586) (0.797)
BJP LS seatshare 0.523 0.977 0.939 0.536 0.987 0.946
(0.463) (0.115) (0.146) (0.458) (0.115) (0.155)
Log-likelihood −6922.78 −6318.98 −6273.87 −6908.62 −6289.13 −6245.32

Observations 229 229 229 229 229 229
at 10%, **significant at 5%, ***significant at 1%
the three columns, we find that this coefficient is positive and statistically
significant. These results resonate with the main findings of Mitra and Ray
(2014).
The Case of Ahmedabad
The state of Gujarat has been witness to many incidents—both small and large
in scale—of Hindu-Muslim animosity over several decades. The district of
Ahmedabad, in particular, has achieved considerable notoriety in this regard.
Field et al. (2008) write:
Recurring communal violence between Hindus and Muslims has become

increasingly common in Indian cities over the last two decades of the twentieth
century (Varshney 2002), and nowhere is it more acute than in Ahmedabad.
Correspondingly, residential arrangements in this city are remarkably segregated:
by 2002, 71% of the population in our sample lived in completely homogeneous
neighborhoods, a fact that presumably reflects increasing intolerance for living
with members of another religion.
As observed (briefly) earlier, Ahmedabad has also been a political stronghold

of the BJP and pro-Hindutva allies starting from the early 1990s—but this
has intensified since the late 1990s.11 Given this state of affairs, we isolate
Ahmedabad from our sample and perform our analysis afresh while excluding
the region containing the Ahmedabad district for all the five time periods.
Like before, prior to examining the results from the regression analysis, we
turn to a graphical depiction of the key relationships. This is presented in
Fig. 14.2. These graphs essentially are the counterparts to the ones presented
in Fig. 14.1. The only difference is that the sample of regions here does not
include the region containing Ahmedabad.
So in panel (a) of the figure, we look at the association between “Casualties”
(which is the number of people either killed or injured) and average Muslim
expenditure at the regional level. Once again, we first estimate a Poisson
regression where Casualties are regressed on Muslim and Hindu expenditures
(log) and region and time dummies. Using the estimates from this regression,
we are able to remove region, time, and (log) Hindu expenditure effects from
11 One illustration of this is the fact that the state government of Gujarat has always been headed by a
BJP Chief Minister since 1998; prior to that periods of BJP dominance were interspersed with either
Congress or some other party’s rule. Additionally, the Member of Parliament from any constituency in
the Ahmedabad district has never been from outside of the BJP since the 1990s.
121
●
−17.5 222 ●
121
●
053 224 ●
053 ● 053
●
123 124 022 ●
●
091
●
●
●
093
●
223 125122
232 222 ●● 121
123114
● ● ●
222101 ● 042 ● 043113

093
●
●
222 022
093 055
●
121 ●
●
●
●
● ●
−20.0 094
● ●
123
124 125233 055 ●
●
●
121
121 223
●
125 182 055 ● ● ●
113042
055114 051
●
182122
● ● ● ●
● ●
●
●
116 ●
022 ●
●
055
● ●

224 094 ● ●
45 094 114
022 051 223
092
●
184 114042
● ●
● ● ●
121 022 116 181051
163 021
● 054 224 182093
233
● ●
●
222
182 203 101 ●
●
● ●
●
●● 041 ● 053
●
●
●
● ●
●
093 ●
● ●
●
023 201 ●
094 123
●
123 091 ●
●
233123 ●
● ●
094 201 ●
102 161 ● ●
222 051 042 181 051 ●

●
●
●
−22.5 041 201 ●
●
● ●
223
●
●
●
203 184 222

042 124223 124 022 ●
● ● ●
222 224 204 054 021 ●

●
●
181 181 041 124 223

224 ●
● ●
●
●
053 ● ● ●
182 ● ●
●
102091
● ●
093 233
182
●
116
122
●
093 124
● ●
022 093
●
053 022 113 114 041 092 ● ●
●
111
●
●
●
●
●
051 043 162
● 113 054
● ●
●
● ●
● ●
234 183 171 201 ●
122043 223 ●
041 ●
●
●
● ●
125
054
●
●
●
094 184
102
●
224 117
041 ●
●
043 055 163
123 ● ●
124 ● 201 ●
●
●
● ● ●
024
●
183 ●
222 ●
● ●
125 222 123

●
123
224 125 042
● 055 181 ● 101 051 094 053 ● ●
●
●
223
●
224
●
091 ● ●
−25.0 204042 125 224 ● ●
●
●
092 122 ●
092 121101
094 123 171
● ● ●
161 ●
● ●
● 061 116233
182 053
●
●
● ● ●
40 232
●
113 114
102
041 203 ●
● ●
●
● 091 114
114
022
●
●
●
●
● ●
● ●
092 021 055091
022 ●
●
● ●
● ●
233
054
124 ● ● 121 ● ●
●
201 116061 203 ● 162 051

223 ● ●
●
042 116
101 111
● ● ● ●
●
● ● ●
223 093 093 ●
●
182122 ● ●
051 ●
041 181125234055 ●
043 ●
●
●
● 043 043 184
181 055 ● ●
●●
● 121
183 ●
●
203 041
● ● ●
023 054 054 203 ● ●

●
● 021 ●
● −27.5 181
●
092 224 021 ●

●
183 ●
●
113021051
●
024 ●
●
●
●
041 ●
117 ● ● 101
043
● ● 054
35 −30.0
9 10 11 12 9 10 11 12
log Muslim expenditure log Hindu expenditure
(a) Conflict and Muslim Expenditure (b) Conflict and Hindu Expenditure
Fig. 14.2 Conflict and per-capita expenditure: without Ahmedabad. Panel (a) plots
the residual of casualties against (log) Muslim expenditures after region, time, and
(log) Hindu expenditure effects have been removed, in the five-year period following
expenditures. Each line segment connects five data points for a region. Panel (b) plots
the analogous graph for (log) Hindu expenditures
the actual Casualties variable so as to arrive at the residual of Casualties. As

shown in panel (a), this residual is positively correlated with (log) Muslim
expenditures. Panel (b) is created by a similar process where we now remove
(log) Muslim expenditure effects rather than (log) Hindu expenditure effects
to arrive at the residual of Casualties. As the figure depicts, the correlation
between this residual of Casualties and average Hindu expenditure is negative.
Therefore, these asymmetric relationships (between conflict and expenditures)
uncovered in our earlier work persist in the sample without Ahmedabad.
Next, we present our results using all households (urban and rural); sub-
sequently, we will restrict our attention to just urban households and urban
conflict.
Table 14.3 contains some results where the dependent variable is “Casu-
alties”. This table is the exact counterpart of Table 14.1. All the difference
stems from the fact that the sample here excludes the region containing
Ahmedabad.12
A column-by-column comparison of Tables 14.1 and 14.3 reveals that the
basic patterns are similar. However, the significance on the expenditure terms
is much enhanced here. Specifically, in the first three regressions reported
(columns 1–3), the coefficients on both Hindu and Muslim expenditures are
highly significant. As before, an increase in Muslim expenditures is associated
with rising conflict, while an increase in Hindu expenditures has a mitigating
effect on conflict.
12 That the sample size is reduced by 5 (one for each period) is easily checked regression by regression.
Table 14.3 The effect of Hindu and Muslim expenditures on regional conflict: Poisson FE regressions (excluding the region containing
Ahmedabad)
[1] [2] [3] [4] [5] [6]
H pce −3.659* −3.357* −3.500**
(0.053) (0.052) (0.013)
M pce 2.747*** 2.464*** 3.267***
(0.008) (0.006) (0.004)
M/H 2.835*** 2.546*** 3.267***
(0.003) (0.002) (0.002)
Average per-capita exp. −1.108 −1.130 −0.483
(0.595) (0.566) (0.778)
Pop −0.245 0.403 0.597 −0.117 0.567 0.719
(0.919) (0.859) (0.782) (0.962) (0.807) (0.747)
RelPol 5.808*** 5.478*** 5.901*** 5.800*** 5.460*** 5.892***
(0.001) (0.003) (0.003) (0.001) (0.003) (0.003)
Literacy 0.035** 0.033* 0.036** 0.033*
(0.046) (0.056) (0.043) (0.056)
Urban −0.003 −0.003 −0.004 −0.003
(0.856) (0.856) (0.838) (0.846)
Gini H −0.838 −0.682
(0.815) (0.853)
Gini M −3.440* −3.329
(0.092) (0.123)
BJP LS seatshare 0.700 0.713 0.738 0.694 0.708 0.730
(0.204) (0.146) (0.131) (0.216) (0.154) (0.139)
Log-likelihood −4850.79 −4726.48 −4662.41 −4843.12 −4717.37 −4659.73

243

Next, we investigate the effects of Muslim-to-Hindu expenditure ratios on

Hindu-Muslim violence. Some of these results are collected in columns 4–6
of Table 14.3. Note, this is the exact counterpart of columns 4–6 in Table 14.1.
A column-by-column comparison of Tables 14.1 and 14.3 depicts the analogy
in the core patterns. Additionally, the significance on the expenditure ratio
terms is much enhanced here. In all the three regressions (columns 4–6), the
coefficient on the Muslim-to-Hindu expenditure ratios are highly significant.
Simple eye-balling also suggests that the magnitude of the effects are also
larger—judging by the size of the coefficients in all the six regressions—but
we refrain from making a definitive claim without further empirical testing.
Next, we report results for urban households (and conflict) alone while
excluding all observations from the region containing Ahmedabad.
Table 14.4 is the exact counterpart of Table 14.2. A column-by-column
comparison between the two tables tells us that the core patterns are nearly
identical. We use the term “nearly” since there is one notable exception: the
coefficient on BJP’s share of Lok Sabha seats in Table 14.4 is positive and
significant in all specifications. We shall return to this issue shortly.
In sum, the economic effects seem to be more robust—both in terms of
statistical significance and in magnitude—in the sample of regions excluding
Ahmedabad. It is not hard to understand why. After all, Ahmedabad is one
of the hotbeds of Hindu-Muslim violence, and if anything, this feature has
intensified in the twenty-first century. If primordialism is at play anywhere,
Ahmedabad would be a contender. Therefore, removing Ahmedabad from the
sample facilitates in shifting the focus from innate hatreds to the systematic—
but no less pernicious—role of economic forces.
The Role of Politics
Here we briefly discuss the role the presence of the BJP has had on communal
violence. The careful reader would have noticed that in each of the regressions
in each of our tables (Tables 14.1, 14.2, 14.3, and 14.4), we have used BJP’s
share of Lok Sabha seats as a regressor. The coefficient on this variable is
positive throughout. However, it is statistically significant only in some of
the regressions, particularly, in the sample without Ahmedabad and when
restricted to urban households (and conflict). This might suggest that the role
of BJP vis-a-vis religious conflict is limited only to urban areas. While this may
well be true, it is worthwhile to point out that in most of the other empirical
specifications (NB and OLS), we have high statistical significance for the BJP
variable.
Table 14.4 The effect of Hindu and Muslim expenditures on regional conflict: Poisson FE regressions (urban households, excluding the
region containing Ahmedabad)
[1] [2] [3] [4] [5] [6]
H pce −3.420*** −4.076*** −3.460**
(0.007) (0.003) (0.015)
M pce 1.662** 1.793** 2.010*
(0.027) (0.025) (0.053)
M/H 1.874*** 2.097*** 2.051**
(0.008) (0.003) (0.019)
Average per-capita exp. −2.266** −2.772** −2.419
(0.027) (0.023) (0.139)
Pop 0.240 1.141 1.156 0.333 1.246 1.251
(0.831) (0.294) (0.281) (0.768) (0.249) (0.241)
RelPol 2.306** 3.745*** 3.732*** 2.122* 3.551*** 3.574***
(0.038) (0.000) (0.000) (0.070) (0.000) (0.001)
Primary edu. 0.087*** 0.087*** 0.088*** 0.089***
(0.006) (0.007) (0.005) (0.005)
Gini H −2.213 −1.699
(0.520) (0.593)
Gini M −1.406 −0.317
(0.551) (0.896)
BJP LS seatshare 1.260** 1.637*** 1.621*** 1.319** 1.705*** 1.710***
(0.037) (0.003) (0.003) (0.032) (0.002) (0.002)
Log-likelihood −4875.09 −4361.15 −4325.55 −4784.98 −4259.42 −4247.07

245
So why might the “BJP effect” be more subdued when we include Ahmed-
abad? After all this is the place where primordialism should be more at play,
obfuscating if not obliterating the economic channels. There are at least two
explanations for this. First, it is clear that in the rest of India (i.e., without
Ahmedabad), the BJP is not as dominant. There may be a threshold effect in
operation—only in places where the BJP has less than a certain level of political
presence does additional BJP control wield any impact on religious violence.
Second, it may be possible that the dynamics of rioting in Ahmedabad may
rely more on the presence of the BJP in more localized government bodies;
hence Lok Sabha presence may not be the appropriate measure to study. We
leave such questions open to further probing.
Concluding Remarks
From our vantage point in the early twenty-first century, and equipped with
extended data on religious conflict, is there good reason to alter our views on
the dynamics of Hindu-Muslim violence in India? Based on our findings, we
offer the following observations:
(a) There is a clear economic component to violence, especially if one excludes

Ahmedabad, and it is roughly along the lines of Mitra and Ray (2014). So,
increases in Muslim expenditures are followed by greater levels of religious
casualties, while increases in Hindu expenditures dampen such violence.
(b) There is a new aspect which is assuming salience now—namely, a strong
political component which is manifesting itself through the presence of
BJP legislators. There is some evidence of a positive association between
BJP’s share of Lok Sabha seats in a region and the extent of Hindu-Muslim
conflict therein.
(c) Ahmedabad exemplifies the ascendancy of this political component. As
discussed earlier, Ahmedabad has been a focal point for Hindu-Muslim
violence and even more so in the last couple of decades. Additionally, there
has been a concomitant rise in the dominance of the BJP in that region;
hence the need to isolate Ahmedabad for a careful study of the economic
and political forces shaping religious antagonisms.
References
Bagchi, A. 1990. Predatory Commercialization and Communalism in India. In
Anatomy of a Confrontation, ed. S. Gopal. New Delhi: Penguin.
Das, S. 2000. The 1992 Calcutta Riots in Historical Continuum: A Relapse into
‘Communal Fury’. Modern Asian Studies 34: 301.
Engineer, A. 1984. The Causes of Communal Riots in the Post-Partition Period
in India. In Communal Riots in Post-Independence India, ed. A. Engineer, 33–41.
Hyderabad: Sangam Books.
———. 1987. Meerut: the Nation’s Shame. Economic and Political Weekly 22:
969–973.
———. 1994. Communal Violence in Kanpur. Economic and Political Weekly 29:
473–474.
Esteban, Joan-Maria, and Debraj Ray. 1994. On the Measurement of Polarization.
Field, E., M. Levinson, R. Pande, and S. Visaria. 2008. Segregation, Rent Control,
and Riots: The Economics of Religious Conflict in an Indian City. The American
Economic Review, Papers and Proceedings 98:505–510.
Iyer, S., and A. Shrivastava. 2018. Religious Riots and Electoral Politics in India.
Journal of Development Economics, 131: 104–122.
Jha, S. 2013. Trade, Institutions and Ethnic Tolerance: Evidence from India. American
Khan, D. 1992. Meerut Riots: An Analysis. In Towards Understanding Communalism,
ed. P. Kumar, 465. Chandigarh: Centre for Research in Rural and Industrial
Development.
Mitra, A., and D. Ray. 2014. Implications of an Economic Theory of Conflict: Hindu-
Muslim Violence in India. Journal of Political Economy 122 (4): 719–765.
Montalvo, J., and M. Reynal-Querol. 2005. Ethnic Polarization, Potential Conflict,
and Civil Wars. The American Economic Review 95:796–813.
Nellis, Gareth, Michael Weaver, and Steven C. Rosenzweig. 2016. Do Parties Matter
for Ethnic Violence? Evidence from India. Quarterly Journal of Political Science 11
(3): 249–277.
Rajgopal, P. 1987. Communal Riots in India. New Delhi: Uppal Publishing
House/Centre for Policy Research.
Upadhyaya, A. 1992. Recent Trends in Communal Violence: A Case Study of
Varanasi. In Proceedings of the International Conference on Conflict and Change.
Portrush/Tokyo: Ulster/United Nations University.
Varshney, Ashutosh. 2002. Ethnic Conflict and Civic Life: Hindus and Muslims in
India. Yale University Press.
Varshney, A. 2003. Ethnic Conflict and Civil Life: Hindus and Muslims in India. New
Haven, CT: Yale University Press.
Varshney, A., and Steven Wilkinson. 2004. Dataset on Hindu–Muslim Violence in
India, Version 2. October 8, 2004.
Wilkinson, S. 2004. Votes and Violence: Electoral Competition and Ethnic Riots in India.
15
Religion and Conflict: A Quantitative
Approach
José G. Montalvo and Marta Reynal-Querol
This chapter surveys the literature on the economics of religion, with a

particular emphasis on its association with conflict. There is a long tradition
of work on the relationship between conflict and ethnic diversity culminating
with Horowitz’s seminal Ethnic Groups in Conflict. The effect of religion on
conflict has generated less attention. Recently there have been several reviews of
the topic. See, for instance, Iyer (2016), Finke (2013), or Silvestri and Mayall
J. G. Montalvo
Department of Economics, Universitat Pompeu Fabra, Barcelona, Spain
ICREA-Academia, Barcelona, Spain
Barcelona Graduate School of Economics, Barcelona, Spain
IPEG, Barcelona, Spain
IVIE, Valencia, Spain
e-mail: jose.garcia-montalvo@upf.edu
M. Reynal-Querol ()
Department of Economics and Business at Universitat Pompeu Fabra, Barcelona,
Spain
ICREA, Barcelona, Spain
IPEG, Barcelona, Spain
Barcelona Graduate School of Economics, Barcelona, Spain
e-mail: marta.reynal@upf.edu

250 J. G. Montalvo and M. Reynal-Querol
(2015). However, prior to 2000, few attempts tried to include religion and
culture into the larger body of research and theory on social conflict. Samuel
P. Huntington’s “Clash of Civilizations” (1996) thesis is the basic reference
of this literature, the same way as Weber’s classic The Protestant Ethics and
the Spirit of Capitalism (1905) is the focal point of the empirical study of the
relationship between religion and economic development. Borrowing partly
from an idea put forward by British-American historian, Bernard Lewis (1990),
Huntington became the most prominent voice claiming that religious and
cultural identities would be the main driver of international conflict in the new
world order following the end of the Cold War. At the core of Huntington’s
clashing civilizations lay religion. He argued that the civilization of Western
Christianity is different from that of Eastern Orthodox Christianity; Eastern
Christianity is distinct from Islam; Islam represents a fundamentally distinct
civilization from Hindu; and so forth. The “clash of civilizations” occurs
at two levels. One level points to the civilization divides across countries
and regions, the other refers to the “fault lines between civilizations” within
countries or territories. Thus, the civilizational fault line(s) within countries
leads to conflicts just as they do across countries. Huntington recognizes that
the argument is over-simplified, yet he concludes that “countries with similar
cultures are coming together” while “countries with different cultures are
coming apart.” He argues that civilizations compete on the international scene
and that this competition can turn into violent conflict, most importantly
because of the different religions that have formed these civilizations. In other
words, civilization fault lines are a source of conflict; civilization homogeneity
is a source of unity and peace (Huntington 1996).
Traditionally, the literature on the relationship between religion and conflict
was generated by fields like politics and international relations, peace and
conflict studies, theology, sociology, history, and security or terrorism studies.
In many of these disciplines, the phenomenon is analyzed using a qualitative
approach, usually based on case studies. More recently, there has been a flurry
of data collection on religious diversity around the world, and a new body
of more quantitative research has emerged. With the availability of new data,
many economists examine the effect of religion in varied historical time periods
and countries, drawing lessons from this experience for religious diversity in
contemporary developing societies. This chapter develops a quantitative view
of the relationship between religion and conflict.
The Measurement of Religious Diversity: Indices

of Fractionalization and Polarization
In order to evaluate the relationship between religiosity and potential conflict
in a particular geographic area, there are two basic measures of religious
diversity. The index of religious fractionalization (FRAC) can be interpreted as
the probability that two randomly selected individuals in a country will belong
to different religious groups. The form of this indicator is the following:
J

nij 2
F RAC i = 1 −
j =1
Ni
where nij /Ni is the proportion of people affiliated to religion j in country i.1
By construction FRAC increases when the number of groups increases.
An alternative indicator of religious diversity is the index of religious
polarization of Montalvo and Reynal-Querol (2000a, b):
J

0.5 − π ij 2
P OLi = 1 − π ij
j =1
0.5
where π ij is equal to nij /Ni . The index POL ranges from 0 to 1. Contrary to
what happens with the fragmentation index, polarization reaches its maximum
when there are two religious groups of equal size. In this type of index, what
matters is not only how many groups there are but also if they view other
groups as a potential threat for their interests. For a given number of groups,
the higher the threat is, the larger is the size of another group relative to
the size of the reference group. Therefore, the polarization index can reflect
potential religious conflict in a society better than the fragmentation index.
There are several theoretical justifications for the discrete polarization index.
Rent-seeking models point out that social costs are higher, and social tensions
emerge more easily, when the population is distributed in two groups of
equal size. In fact, Montalvo and Reynal-Querol (2005b) have shown that
their polarization index can be derived from a simple rent-seeking model.
Therefore, the index of polarization captures basically how far the distribution
1 The ethnolinguistic fragmentation index used in many empirical growth studies belongs to this class of
indices.
of the groups is from a bimodal distribution, while the fragmentation index

increases monotonically with diversity. It is also the case that potential conflict,
measured by the polarization index, can reduce social capital and affects, also
through this channel, economic growth.
The Measurement of Religious Diversity: Data

There are many different sources of data on religious diversity, including ad hoc
surveys conducted for specific research purposes2 and standard surveys that
include some questions on religious beliefs and formal religious services like the
World Value Survey (WVS), the International Social Survey Program (ISSP),
or the Gallup Millennium Survey. These surveys are limited by the number
of countries that are included and their temporal extension. They are mostly
used for microeconomic analysis of the determinants of religiosity or the
relationship between religious beliefs and other socioeconomic characteristics
(trust, etc.).
However, the main source of data for religious diversity at the country level
comes from Barrett’s (1982, 2000) World Christian Encyclopedia (WCE). The
proportion of each religion is constructed from the answer to the question
“What is your religion?” in public polls. As this question does not have
multiple responses, every individual is supposed to have only one possible
religious affiliation.
This Encyclopedia is descended from a long series of some 40 major surveys and
atlases of Christianity and missions. The genesis of this encyclopedia goes back to
the World Christian Handbook, and Anglican and Protestant publication which
appeared on average every five years from 1949 to 1968, and to the similar Roman
Catholic publication3 which had editions in 1958 and 1964. The volume of 1982,
embodies the traditions of both these former publications. This study shares with
its more recent predecessors a critical, scholarly, and scientific approach to data
describing the Christian world. In the main, it uses existing data collected by the
churches for their own purpose.
The WCE has the advantage of being a cross section of time series,
providing information for 1970, 1975, and 1980. However, this source has
several shortcomings. First, and probably the most important, the data do not
2 Dowd (2016).
3 Bilan du monde: encyclopedie catholique du monde chretien.
consider the possibility of double practice, very common in Latin American

countries. Moreover, comparing these data to other sources of information,
Montalvo and Reynal-Querol (2002) realize the data are biased toward the
Christian religion. A clear example is the case of Zaire in which the distribution
of religions is considered to be similar to Spain or Italy. This basically means
that this source does not compute the followers of syncretic cults in Latin
American countries and that the information on the followers of animist
cults or traditional religion in sub-Saharan Africa is always lower than the
percentages provided by national sources or not computed. The distribution
of religious groups between 1970 and 1980 does not change in many countries.
If there are changes, they usually occur in countries where there is double
practice, and they usually imply an increase in the percentage of Christians
and a reduction in the size of animist followers in many African countries or
the syncretic cults in Latin America. Because of these reasons, the data coming
from the WCE should be handled carefully.
Montalvo and Reynal-Querol (2002) notice the limitations of the WCE
as a reliable source of data on religion, and they construct our dataset using
essentially two sources of information. The first source is “L’Etat des Religions
dans le monde” (ET), which takes the data from the “World Christian
Encyclopedia” (WCE) and corrects them using national sources, providing
information on the proportions of followers of animist and syncretic cults.
There are two other religious sources that provide information on animist and
syncretic cult followers based on national sources: the Statesman’s Yearbook
(ST) and the World Factbook (WF). The proportions on animist and syncretic
cults reported by these two other sources are very similar to the proportions
reported by ET. ST is not as complete as the ET or WCE but is completely
based on national sources. For this reason, it gives very detailed information
on animist followers in African countries even though sometimes it does
not provide enough information on the proportion of followers of other
religions. The WF is more comprehensive than ST, but less than the ET or
the WCE. It gives information on the proportions of animist and traditional
religions mainly in African countries. However, it has the disadvantage, as with
the WCE, that it does not consider the double practice in Latin American
countries and therefore does not report the proportion of syncretic cult
followers. In many cases the two basic sources coincide and they take that value
to be the correct one. In some other cases, the ST does not provide enough
disaggregation and it uses the information of ET. The great advantage of the
ST is the extreme detail on animist religions.4 The use of these sources allows
Montalvo and Reynal-Querol (2002) to consider the percentages of followers
of animist and traditional religions, the followers of syncretic cults, and the
percentage of the different Muslim subgroups.
Using the previous methodology, Montalvo and Reynal-Querol (2002)
consider the following religious groups: Jews, Christians, Muslims, Buddhism,
Hinduism, Taoism, Confucianism, Chinese Religion, Bahaism, syncretic cults,
animist religions, other religions, and no religion. They also have information
on the different Christian subgroups (Catholics, Protestants, Orthodox, and
other Christians) and also on the different Muslim subgroups (Sunnites and
Shiite). The animists are followers of traditional religions which practice magic
and the veneration of a large number of gods and spirits. Finally, other religions
include small collectives as “black church” or “spiritual groups.” The following
table summarizes the basic differences among these sources of religious data:
Source Traditional religions Syncretic cults

World Christian Encyclopedia (WCE) Some countries No
Etat des Religions dans le Monde (ET) Yes Yes
Statesman (ST) Yes Sometimes
World Factbook (WF) Yes No
National sources Yes Yes
Some examples illustrate the main differences among these data sources:
Following the WCE, in Angola 80.5% of the population are Christians and
only 19.4% animist. However, following ET and ST Angola has 64% of
Christian followers and 34% of animists. The WCE reports that in Bolivia
95.3% of the population are Christians, while ET and ST report that only
43% of the population are Christians and 44% are followers of syncretic
traditional religions. In Burundi, the WCE reports that 74% are Christians
and 25% animist, and ET and ST report that only 60% are Christians and
39% are animist. In Central African Republic, the WCE reports that 76.5%
are Christians and only 20.3% are animist, while ET and ST report that 35%
are Christians and 57% animist. In Congo the WCE reports that 92% are
Christians, and ET and ST report that only 78% are Christians and 19% are
animist. In Zaire, the WCE reports that 90.3% are Christians, while the ET
and ST report that only 69% are Christians and that 30% of the population
follow animist cults. In the Dominican Republic, the WCE reports that 98.9%
4 Montalvo and Reynal-Querol (2002) confronted the data with national sources in order to improve the
reliability of this information.
of the population are Christians, while ET and ST report that only 48.9%
are Christians and the 51% follow syncretic cults. In Guatemala the WCE
reports that 99.3% are Christians, while the ET and ST report that 73.9% are
Christian, while 25.2% follow syncretic cults.5
Economics, Religion, and Conflict

Economists have considered religion as a determining factor of democracy
and economic development (Barro 1997; Sala-i-Martin 1997; Tavares and
Wacziarg 2001). In these early attempts to evaluate the impact of religion on
development, the authors use simple indicators like a dichotomous variable
for the majority religion in a country or the proportion of each religious group
among the population of a country.
Initially academic economists used religion as a control variable in the anal-
ysis of the determinants of economic growth. Barro (1997) uses the WCE and
distinguishes nine religious groups: Catholics, Protestants, Muslims, Hindus
(include Jains and Sikhs), Buddhists, miscellaneous eastern religions (Chinese
folk religions, Confucianism, and new religionists), no professed religion, and
other religious groups. Barro (1997) shows, with some cross tabulations, that
religions are correlated with the democratic stance of the countries, although
the regression analysis shows less support for this association. Sala-i-Martin
(1997) studies the robustness of the determinants of economic growth using
the religious variables of Barro (1997). Sala-i-Martin (1997) concludes that
the religious indicators belong to the group of variables that are strongly
correlated with growth: the fraction of Confucius, Buddhist, and Muslim
positively and the fraction of Protestant and Catholic negatively. It is unclear
if the interpretation of these variables has to do with the implication of
religion in those countries or if they represent specific geographical areas. Sala-
i-Martin et al. (2004), using Bayesian averaging of models, conclude that
the only religious variable that is strongly and robustly related (positively)
with growth is the fraction of Confucians although its high correlation with
Hong Kong and the Asian Tigers may explain its high explanatory power
for growth. Sala-i-Martin et al. (2004) find that the fraction of protestants
is robustly related with growth, showing a negative effect. Montalvo and
Reynal-Querol (2003) use the growth equation specification of Mankiw et
5 More recently, Grim and Finke (2007) have used data collected by the Association of Religion Data
Archives (ARDA, www.theARDA.com).
al. (1992), and they find that religious polarization has a negative effect on
economic development. Barro and McCleary (2003, 2006), using three cross
sections of countries, find that monthly attendance at religious services and
belief in hell are statistically significant (negatively and positively, respectively)
in the explanation of economic growth. Other religious variables such as
belief in God or whether people self-identify as religious do not have any
statistical power. They also find that the religious population shares are jointly
statistically significant.
Religious Diversity and Polarization as the Source

of Conflict
Montalvo and Reynal-Querol (2002) argue that one of the most impor-
tant causes of conflict are characteristics and differences that are relatively
immutable. Therefore while it may be possible to reach an agreement in polit-
ical or economic issues, it is much more difficult, for instance, to compromise
in religious matters. More than other dimensions, religion discriminates and
differentiates humans in a sharp and exclusive way. A person can be half French
and half Saudi Arab and, at the same time, be a citizen of both countries.
However, it is not possible to be half Catholic and half Muslim. In the old
Soviet Union, communists could become democrats or poor people could
become rich. However, it is unlikely that Muslim will become Christian or
vice versa. In the class and ideological conflicts, the key question is, “which
side are you fighting with?” and people can decide and can change sides. In
the conflict between religions, the question is “who are you?” and you cannot
change sides easily.
Horowitz (1985) points out that in plural societies in Asia, Africa, and
the Caribbean, parties tend to be organized along ethnic lines. In Western
Europe and North America, religion, social class, and language are the basic
dimensions of the situation of a political party. Lijphart (1984) found, in a
sample of 22 democratic regimes, that the two dimensions that most frequently
differentiate systems are the socioeconomic and the religious.
These authors, among others, claim that religious differences are more
important than language differences as a social cleavage that can develop into
a conflict. There are two basic reasons why religious differences can generate
more violence than other social cleavages. First, there is no doubt about the
exclusivity of religion. One can speak two or more languages, but you usually
have one religion. Religion can be used as a sign of identity, stronger than
language in the sense that you exclude absolutely the ones from other religion,
while speaking two languages diffuses the division line among groups. Second,
religious differences, which are the base of the differences among civilizations,
imply different ways of understanding the world, social relationships, and so
on. Even if different groups speak different languages, they could share the
same way of understanding the world and the relationships if they belong to
the same civilization. This is more difficult for people of different religions.
However, from a theoretical perspective, using the proportion of people
professing a particular religion as the literature on religion and economic
development has done implies that what matters for economic development is
the identity and size of each religious group and not the potential conflictual
relationship among them. If we want to analyze the effect of religion on
conflict, it is unclear why any of these two quantitative indices (majority
religion and proportion of believers) can be a good indicator of conflict.
The empirical studies of the relationship between conflict and religion
started almost simultaneously to the analysis of the effect of religion on
development. However, many authors have found that even though religious
fractionalization seems to have some explanatory power, although not very
robust, for economic growth, it is not significant in the explanation of civil
wars and other kinds of conflicts. These results led many authors to disregard
ethnicity as a source of conflict and civil wars. Fearon and Laitin (2003)
and Collier and Hoeffler (2004) find that neither ethnic fractionalization nor
religious fractionalization6 had any statistically significant effect on the onset
of a civil war. Collier and Hoeffler (2004) conclude that ethnic dominance
has a significant impact on the probability of the onset of a civil war, but they
did not find any significant effect of religious fractionalization.7 Collier and
Hoeffler (2004) conclude also that religious fractionalization does not have a
significant effect on the duration of civil wars.
Reynal-Querol (2002) uses the same dataset to show that religious polar-
ization is the most important ethnic dimension in explaining ethnic civil
wars. Fox (2001) specifically examines the role of religion in conflicts in the
Middle East and their resulting characteristics, based on the Minorities at Risk
dataset (which contains information on 267 politically active ethnic minorities
throughout the world) and religious factors. This study uses an empirical
method to provide a perspective on the issue different from the comparative
approach commonly used in the literature. The author finds that religion plays
a disproportionately important role in ethno-religious conflicts in the region,
6 They use the same classification of Barro (1997).

7 Collierand Hoeffler (1998) include an index of ethnolinguistic fractionalization in their study of the
causes of civil wars but do not consider any religious indicator.
more so than in the non-Middle Eastern states with Muslim majorities. States
in the Middle East are also disproportionately more autocratic than in other
regions. However, despite the unique importance of religion, Fox argues that
the prevalence of religious conflict is not explained by either the Islamic or
autocratic character of the states, and in reality the ethno-religious conflicts
in the Middle East are not significantly different from similar ethnic struggles
around the world.
In a further study also based on the Minorities at Risk dataset, Fox (2004)
analyzes the role of religious ties in the spread of ethnic conflict across borders.
The findings show that religious conflict is more contagious than nonreligious
conflict; however, only violent conflicts cross borders, while non-violent ones
do not. The author argues that one possible explanation for this is the
argument that violence is an intrinsic element of religion. This can explain why
religious contagion is stronger than nonreligious contagion and why religious
conflicts cross borders only when they are violent ones.
Fox (2003) focuses on the nature of grievances and demands in a conflict.
He argues that “when religious issues are important, they will change the
dynamics of the conflict.” This can be attributed both to the role of religious
institutions within the state and to the way in which religion influences inter-
national intervention in ethnic conflict. Internally, religious institutions tend
to facilitate a reaction if the grievances have religious importance; however,
if they have no religious importance, the religious institutions often inhibit
protest. Svensson (2007) explores the conditions for negotiated settlements.
The study develops Toft’s (2007) distinction between whether religion plays a
peripheral or central role in armed conflicts. However, whereas Toft measures
religious dimensions of civil wars on the level of analysis of conflict, this
study disaggregates the analysis to the dyadic level. Svensson argues that across
religions, where the grievances or demands are based on explicit religious
claims, the negotiated settlement of conflict is less likely to succeed than if
there are no religious claims. He demonstrates that the chances for negotiated
settlement are not affected if the conflicting parties are from different religious
traditions.
Montalvo and Reynal-Querol (2005a) find that when ethnolinguistic polar-
ization is included as an explanatory variable, neither religious polarization nor
religious fractionalization has a significant effect on the incidence of civil wars.
However, Montalvo and Reynal-Querol (2005b) use a system of equations to
analyze the direct and indirect effect of religion in the economy. The first
equation is a traditional growth regression “a la Barro.” The systems add
three additional equations for investment, civil war incidence, and government
expenditure. Montalvo and Reynal-Querol (2005b) find that neither religious
fractionalization nor religious polarization has a direct impact on growth.

However, they find that religious polarization has an indirect negative impact
on growth due to its positive effect on public expenditure and civil wars and
it negative impact on investment.
Finally, there appears to be a strong correlation between the emergence of
religious conflict and situations of state failure or collapse. Fox (2007), for
instance, tracks state failures between 1960 and 2004, identifying the shifts in
the role of religion and state failure. Using data from the state failure dataset,
he identifies an increase in state failure related to religion as a proportion of
all state failures during this period and finds that it became the most common
kind of state failure in 2002, after which he identifies religion as an element
in the majority of all conflicts that relate directly to state failure.
Religious Restrictions as the Source of Conflict
Finke (2013) challenges Huntington’s theory. Finke claims that restrictions

placed on religions, and not religious or cultural pluralism, are far more
important in understanding violence closely tied to religion. The author
focuses on the causes and consequences of religious restrictions and on their
non-trivial relationship with social conflict. Finke (2013) argues that the
motives for states restricting or supporting religious freedoms are many and
varied. States can restrict freedoms in an effort to support the religion(s)
that hold an alliance with the state.8 Or the state might impose restrictions
because religions are perceived challenging the state’s ideology, threatening
government stability, and being a potential harm to public order and the
welfare of local citizens. Finally, states can have economic and political motives
for supporting such freedoms. Also, even when the state has few motives for
denying freedom and holds no strategy for restricting religious freedoms, the
inaction of the state can lead to high levels of restrictions. When the state
is weak and the judiciary is embedded within other government or religious
institutions, the tyranny of the majority and the actions of religious, political,
and social movements can quickly deny the religious freedoms of others. At
the individual level, these restrictions change the incentives and opportunities
both for religious producers and consumers. At the organizational and market
level, the restrictions change how organizations operate and alter the structure
of the market.
8 Barro and McCleary (2005) provide a discussion of the importance of state religions.
Fox and Tabory (2008) study the impact of religious competition on

religiosity by looking at state support for religion as a structural factor affecting
religious pluralism. Their dependent variables include measures of attendance
at religious services, religious beliefs, and self-categorization as a religious
person (World Values Survey and the International Social Survey Program).
The independent variable consists of a series of six measures that deal with
state support for religion from the Religion and State (RAS) database for
the 1990–2002 period. The results indicate that state regulation of religion is
significantly and negatively correlated with religiosity. This is consistent with
predictions that religious monopolies will reduce participation but not belief.
As Finke (2013) notes, religious restrictions, as well as subsidies, have
consequences that go far beyond the religious institutions and activities
being targeted. Although restrictions are most frequently targeted at religious
minorities, they alter the entire market structure and revise both how religious
institutions operate and whom they serve. More specifically, Grim and Finke
(2007) bring new evidence on the effects of civilizational homogeneity within
countries when applied to religious persecution. Building on a theory of
religious economies and using a new source of data collected by the Association
of Religion Data Archives, the paper finds that the cultural and religious
pluralism of a country is far less important than the state’s response to this het-
erogeneity. Theoretically, they argue that not only does less regulation reduce
the grievances of religions, it also decreases the ability of any single religion
to wield undue political power. When a religious group achieves a monopoly
and holds access to the temporal power and privileges of the state, the ever-
present temptation is to openly persecute religious competitors. Empirically,
the authors find that civilizational divides have only an indirect relationship
with religious persecution, whereas the coefficient from government regulation
was direct, powerful, and highly significant. They conclude that to the extent
that governments ensure religious freedoms for all, religious persecution is
reduced.
Conclusions
In the light of the research presented in this chapter, the consensus seems
to be that while religion should not be taken for granted as the main
driving force of violence and conflict, it cannot be excluded from accounts
of international relations, impacting both interstate relations and domestic
politics. In addition, Dowd (2016) has argued, using the case of Nigeria,
that religious diversity, tolerance, and conflict have important subnational
variation. Montalvo and Reynal-Querol (2017) have shown that the lack of
explanatory power of ethnic diversity on economic development is due to the
fact that at high levels of aggregation (country level) the positive effects of
diversity are less important than its negative consequences. This is an example
of the well-known modifiable areal unit problem (MAUP). Essentially, MAUP
implies that the relationship between variables at one scale may be distorted
when analyzed using another scale. This could be also the case in the analysis of
the effect of religion on conflict. Further research should evaluate the potential
impact of MAUP in the relationship between religious diversity, polarization,
and conflict.
References
Barrett, D. 1982, 2000. World Christian Encyclopedia. Oxford: Oxford University
Press.
Barro, R. 1997. The Determinants of Economic Growth. Cambridge, MA: MIT Press.
Barro, R., and R. McCleary. 2003. Religion and Economic Growth. American
Sociological Review 68 (5): 760–781.
Barro, R.J., and R.M. McCleary. 2005. Which Countries Have State Religions?
Barro, R., and R. McCleary. 2006. Religion and Economy. Journal of Economic
Perspectives 20 (2): 49–72.
Collier, P., and A. Hoeffler. 1998. On Economic Causes of Civil War. Oxford Economic
Papers 50: 563–573.
———. 2004. Greed and Grievance in Civil Wars. Oxford Economic Papers 56: 563–
595.
Dowd, R. 2016. Religious Diversity and Religious Tolerance: Lessons from Nigeria.
Journal of Conflict Resolution 60 (4): 617–644.
Fearon, J., and D. Laitin. 2003. Ethnicity, Insurgency, and Civil War. American
Political Science Review 97: 75–90.
Finke, R. 2013. Presidential Address Origins and Consequences of Religious Free-
doms: A Global Overview. Sociology of Religion 74 (3): 297–313.
Fox, J. 2001. Are Middle East Conflicts More Religious? Middle East Quarterly 8 (4):
31–40.
———. 2003. Counting the Causes and Dynamics of Ethnoreligious Violence.
Totalitarian Movements and Political Religions 4 (3): 199–144.
———. 2004. Is Ethnoreligious Conflict a Contagious Disease? Studies in Conflict
and Terrorism 27 (2): 89–106.
———. 2007. The Increasing Role of Religion in State Failure 1960–2004. Terrorism
and Political Violence 19: 395–414.
Fox, J., and E. Tabory. 2008. Contemporary Evidence Regarding the Impact of State
Regulation of Religion on Religious Participation and Belief. Sociology of Religion
69: 245–271.
Grim, B.J., and R. Finke. 2007. Religious Persecution in Cross-National Context:
Clashing Civilizations or Regulated Religious Economies? American Sociological
Review 72 (4): 633–658.
Horowitz, D. 1985. Ethnic Groups in Conflict. Berkeley, CA: University of California
Press.
Huntington, S.P. 1996. The Clash of Civilizations and the Remaking of the Modern
World. New York: Simon and Schuster.
Iyer, S. 2016. The New Economics of Religion. Journal of Economic Literature 54 (2):
395–441.
Lewis, Bernard. 1990. The Roots of Muslim Rage. The Atlantic, September.
www.theatlantic.com/magazine/archive/1990/09/the-roots-of-muslim-rage/
304643.
Lijphart, A. 1984. Democracies, Patterns of Majoritarian and Consensus Government in
Twenty-One Countries. New Haven, CT: Yale University Press.
Mankiw, G., P. Romer, and D. Weil. 1992. A Contribution to the Empirics of
Economic Growth. Quarterly Journal of Economics 107 (2): 407–437.
Montalvo, J.G., and M. Reynal-Querol. 2000a. The Effect of Ethnic and Religious
Conflict on Growth. IVIE WP-EC 2000-04.
———. 2000b. The Effect of Ethnic and Religious Conflict on Growth. IVIE WP-EC
2000-04. An Updated Version Can Be Found in http://www.wcfio.harvard.edu/
programs/prpes.
———. 2002. Why Ethnic Fractionalization? Polarization, Ethnic Conflict and Growth.
Mimeo.
———. 2003. Religious Polarization and Economic Development. Economic Letters
80: 201–210.
———. 2005a. Ethnic Polarization, Potential Conflict and Civil Wars. American
Economic Review 95 (3): 796–816.
———. 2005b. Ethnic Diversity and Economic Development. Journal of Develop-
ment Economics 76: 293–323.
———. 2017. Religion and Conflict: A Review from the Post 9–11 Era. Mimeo.
Reynal-Querol, M. 2002. Ethnicity, Political Systems and Civil Wars. Journal of
Conflict Resolution 46 (1): 29–54.
Sala-i-Martin, X. 1997. I Just Ran Two Million Regressions. American Economic
Review, American Economic Association 87 (2): 178–183.
Sala-i-Martin, X., G. Doppelhofer, and R.I. Miller. 2004. Determinants of Long-
Term Growth: A Bayesian Averaging of Classical Estimates (BACE) Approach.
American Economic Review 94 (4): 813–835.
Silvestri, S., and J. Mayall. 2015. The Role of Religion in Conflict and Peacebuilding.
London: British Academy.
Svensson, I. 2007. Fighting with Faith: Religion and Conflict Resolution in Civil
Wars. The Journal of Conflict Resolution 51 (6): 930–949.
Tavares, J., and R. Wacziarg. 2001. How Democracy Fosters Growth. European
Economic Review 45: 1341–1378.
Toft, M. 2007. Getting Religion? The Puzzling Case of Islam and Civil War.
International Security 31 (4): 97–131.
Weber, Max. 1905 [1930]. The Protestant Ethic and the Spirit of Capitalism. London:
Allen & Unwin.
16
Why Are Some Societies More Religious
Than Others?
Jeanet Sinding Bentzen
Introduction
The world of today sees vast differences in religiosity. The most religious
countries are Algeria and Pakistan, where 100% of the population believe
in God.1 At the opposite end of the spectrum lies China with 20% of the
population believing in God. These differences appear to matter for important
socioeconomic outcomes such as health, wealth, labor force participation,
and education choices.2 It therefore seems relevant to ask what explains these
differences in religiosity across the globe. Providing answers may even help us
understand why religion has not declined in many places of the world today
as the secularization hypothesis otherwise suggests.3
1 According to the most recent waves of the pooled World Values Survey and European Values Study.
2 See Guiso et al. (2003), Scheve and Stasavage (2006), McCleary and Barro (2006), Gruber and Hunger-
man (2008), and Campante and Yanagizawa-Drott (2015) for empirical investigations or Iannaccone
(1998), Lehrer (2004), and Kimball et al. (2009) for reviews.
3 The secularization hypothesis predicts that religiosity falls as societies modernize. It has received mixed
support, though. Norris and Inglehart (2011) show that while religion has become less important in many
Western countries, it has increased in importance in other parts of the world, leading to a net increase in
the number of people with traditional religious views during the past 50 years. See also Stark and Finke
(2000) and Iannaccone (1998) for discussions and Becker et al. (2017) for an empirical investigation of
the influence of education on the secularization process.
J. S. Bentzen ()
Department of Economics, University of Copenhagen, København, Denmark
e-mail: Jeanet.Bentzen@econ.ku.dk

266 J. S. Bentzen
While scholars have attempted to answer the question for centuries and
numerous theories have been put forward, only recently has data availability
and advances in programming technology made it possible to empirically test
some of the theories.4 One of the key empirical challenges has been issues
regarding external validity. For instance, conclusions from a study of Catholics
in one country cannot necessarily be extended to the world at large. Since all
societies were most likely religious at some point in their past (e.g., Murdock
1965; Brown 1991; Peoples et al. 2016), a theory explaining differences in
religiosity in general is most useful if it holds across all religious denominations
and countries.
This chapter introduces a measure of religiosity that is globally comparable,
which enables validity checks of the existing theories within all major religious
denominations and countries, thereby solving the challenge of external valid-
ity. The religiosity measure is constructed in a way so that one does not have
to compare religiosity across countries, but can stick with comparison within
countries (across subnational districts). The chapter also shows a method to
solve the second key empirical challenge identification via ArcGIS program-
ming. This program makes it possible to construct variables from all thinkable
spatial data. In particular, natural experiments can be exploited, providing
exogenous variation in key variables of interest. The empirical insights will
be based primarily on the data and analysis by Bentzen (forthcoming). The
next section first provides a brief overview of the main theories for differences
in religiosity.
Theories
Theories abound for why religion emerged in the first place. These theories,
however interesting and important they are, may not necessarily explain cur-
rent differences in religiosity. Take, for instance, the theory that religion arose
as a solution to cooperation problems (e.g., Norenzayan 2013).5 According to
this theory, beliefs in an almighty punishing god solved the problem of free-
riding in pre-modern societies; God was believed to punish deterrents, thus
inducing cooperation. Inhabitants in societies that developed punishing gods
4 The particular programming technology referred to here is ArcGIS programming, which makes it feasible
to exploit the spatial dimension of the data better than ever.
5 For other evolutionary theories of the origins of religion, see Boyer (2008). Another theory is that major
religions arose as a tool for power legitimization (e.g., Bentzen and Gokmen (2017) for an empirical
investigation).
were better able to cooperate and thus more likely to survive and multiply.
Eventually, evolution selected societies that held beliefs in punishing gods. In
keeping with this theory, the invention of formal policing institutions reduced
the need for God as a policing institution, thus reducing the importance of
religion (also emphasized by Norenzayan (2013)). This theory, therefore, has
no clear-cut prediction as to whether societies that developed religion earlier
than others are more or less religious today.
Social scientists have applied microeconomic theory to explain patterns
of current religious behavior among individuals, groups, and cultures.6 This
work began with the model by Azzi and Ehrenberg (1975), where individuals
allocate their time and goods among religious and secular commodities to
maximize their lifetime and afterlife utility. Within this framework, the reasons
for differences in religiosity can be grouped into demand- and supply-side
factors (e.g., Finke and Stark 2005).
One supply-side theory is that religious congregations compete for fol-
lowers, thus increasing the quality of the religious services provided, which
in turn increases religious participation (e.g., Finke and Stark 2005; Olson
2011). However, empirical investigation of the supply-side drivers of religious
intensity provides mixed results (see reviews by Chaves and Gorski (2001)
and Hungerman (2010)).7 Furthermore, instead of competing with other
congregations, a more important competitor from the viewpoint of the partic-
ular congregation could be secular organizations (e.g., Hungerman 2010 and
Hungerman 2005). In particular, Gruber and Hungerman (2008) show that
the legalization of retail activity on Sundays led to lower church attendance
and church donations across US states.
Demand-side theories point to factors that elevate the demand for religion,
increasing the extent of religious engagement (e.g., Norris and Inglehart
2011). The main proposed demand-side factors are the extent of stress and
uncertainty in society, attempts to understand the world by referring to
religion, or material aid obtained through the church. One demand-side
theory that has received support in the data is the idea that individuals use
their religion to cope with stress and uncertainty. I will return to this theory
and the empirical investigations in section “Testing One Theory: Religious
Coping”. Among the demand-side theories is the secularization hypothesis,
6 See reviews by Iannaccone (1998) and Iyer (2016).

7 Another supply-side-based story is that religious organizations create incentives to encourage stigmatizing
behaviors in order to screen out potential free-riders. Iannaccone (1992) considers this.
268 J. S. Bentzen
where the idea is that religion will die out as countries develop. This, however,
has received mixed support in the data.8
Another set of theories regard differences in the type of religious affiliation.
For instance, scholars have documented various socioeconomic differences
between Protestants and Catholics or between Christians and Muslims (e.g.,
Becker and Woessmann 2009; Weber 1905; Rubin 2017). One could imagine
that differences in religiosity could be explained by differences in people’s
religious denomination. While we shall see below that indeed Muslims are
on average more religious than others, it turns out that differences in terms of
religious denominations explain a miniscule part of differences in religiosity.
The data used to investigate these tendencies and other theories is presented
below.
Data on Global Religiosity

A first-order issue that applies to most of the empirical studies reviewed so far
is the problem of external validity: Most existing studies investigate a sample
of people from one or a few religious denominations in one country. This
issue was also emphasized in a recent review of the economics of religion
literature by Iyer (2016). A necessary (but not sufficient) fix is a globally
comparable measure of religiosity. The World Values Survey and European
Values Study provide just this for a total of 500,000 individuals from 109
countries interviewed over the period 1981–2014. These two surveys can be
appended to one another, which has been done in the following.
One major concern with these data is that self-reported religiosity is
affected by many things—for instance, individuals’ religious denomination
and national institutions—which complicates comparison across religious
denominations and countries. To make the religiosity measures comparable
across religious denominations, one of the fathers behind the values surveys,
Ronald Inglehart, identified, together with political scientist Pippa Norris,
six measures of religiosity that span global religiosity (Inglehart and Norris
2003). These measures include answers to the questions “How important
is God in your life?”, “Are you a religious person?”, “How often do you
8 Rather, religion seems to be on the rise in many societies, which some see as a rejection of the
secularization hypothesis (e.g., Iannaccone 1998; Finke and Stark 2005; Norris and Inglehart 2011). Some
scholars have viewed rising religiosity in the US as a counterexample of the secularization hypothesis.
However, Voas and Chaves (2016) document that religiosity in the US has declined over the past decades
when cohort effects are accounted for.
Fig. 16.1 Average religiosity across the globe. Notes: Country averages across all waves
1981–2014 of the pooled WVS-EVS. Religiosity is measured using the Strength of
Religiosity Scale in the left maps and answers to the question “How important is God
in your life?” in the two maps to the right. The upper panels show the simple country
averages without control variables. The lower panels show the within-denomination
differences, that is, the residuals of regressions where the religiosity measures are
regressed on the five major religious groups: Christianity, Islam, Buddhism, Hinduism,
and Other
attend religious services?”, “Do you get comfort and strength from religion?”,
“Do you believe in God?”, and “Do you believe in a life after death?”
These questions are answered by 268,859–477,843 individuals from 82–105
countries, where the first three questions are answered by the largest amount of
respondents. Inglehart and Norris suggest a composite measure constructed by
factor component analysis, which they term the Strength of Religiosity Scale.
This composite measure is available for 221,249 individuals interviewed in 80
different countries.
The country averages of the Strength of Religiosity Scale are shown in
the upper left panel of Fig. 16.1. The upper right panel shows one of the
religiosity measures available for the full sample: Answers to the question “How
important is God in your life?”. Both measures are scaled to lie between 0 and 1.
The two measures show a very similar pattern of the spread of religiosity across
the globe.
270 J. S. Bentzen
These patterns in religiosity cannot be explained by differences in individ-

uals’ religious denominations. The lower panels show again country averages,
but now after differences across the major religious denominations have
been removed.9 That is, these are the within-denominations differences. The
picture is nearly unchanged. The largest change occurs in the Middle East,
Indonesia, and the Americas, where relative religiosity falls in the former two
and increases in the Americas after holding religious denominations constant.
This is particularly due to high religiosity among Muslims.10 One can investi-
gate the importance of religious denominations more formally using analysis
of variances, which shows that only 3.4% of the total variation in religiosity
across the globe can be attributed to individuals’ religious denomination.
Equivalently, 96.6% of the variation in global religiosity comes from variation
within religious denominations. This means that when we are searching for
the reasons for differences in religiosity across the globe, we can disregard
explanations concerning differences in the type of religion. The pattern is
also unchanged if we take basic individual characteristics into account, such
as gender, age, and marital status.
Another concern when constructing a global measure of religiosity is that
individuals’ understanding of the particular survey questions is potentially
influenced by the general culture or national institutions in the country in
which they live. This makes comparison across countries difficult. Bentzen
(forthcoming) addresses this concern by exploiting information on the sub-
national district in which the individuals were interviewed. This information
is available for 90 of the countries in the pooled World Values Survey and
European Values Study dataset shown above. There are ten districts on average
per country. Exploiting this information enables comparison of religiosity
within countries, instead of having to compare across countries.11 Thus,
differences in individuals’ understanding of the questions across countries are
not a problem in such an analysis. Further, unobserved country-level factors,
such as national institutions and culture, can be removed from the analysis.
9 The religious denominations accounted for are Christianity, Islam, Hinduism, Buddhism, and others.
Before aggregating the data, the two religiosity measures are regressed on religious denominations fixed
effects. The residuals are saved, scaled between 0 and 1, and aggregated to the country level.
10 The global average of the Strength of Religiosity Scale across all waves 1981–2014 is 0.78, which covers
0.59 for Buddhists, 0.76 for Christians, 0.78 for Hindus, 0.87 for Muslims, and 0.80 for others. These
differences are statistically different from one another.
11 Note that within-country analysis can be done without exploiting the subnational districts, if the
analysis is restricted to information available in the pooled WVS-EVS. The subnational districts become
particularly useful when the analysis involves linking the WVS-EVS data to data from other sources. See
section “Testing One Theory: Religious Coping” for an application.
One could be concerned that we end up with too little variation in religiosity
when removing the between-country variation. Lack of variation is an issue if
we want to do econometric analysis, investigating, for instance, the reasons
for varying religiosity levels as is the focus here. This does not seem to be
a large problem, though. In fact, the within-country variation in religiosity
amounts to 71% of the total variation in religiosity.12 Thus, less than a third
of the variation in religiosity is lost when throwing away the variation across
countries.
Testing One Theory: Religious Coping

Equipped with these globally comparable measures of religiosity, we can return
to the question of why religiosity differs across societies. The focus here is on a
major theory within the demand-side models.13 Examples of so-called religious
coping are seeking a closer relationship with God, praying, or finding a reason
for the event by attributing it to an act of God. The religious coping hypothesis
states that individuals draw on religious beliefs and practices to understand and
deal with unbearable and unpredictable situations.14
Numerous empirical studies show that individuals hit by various adverse
life events, such as cancer, heart problems, death in close family, alcoholism,
divorce, or injury, are more religious than others.15 In addition, prayer is often
chosen by various hospitalized patients as a coping strategy above seeking
information, going to the doctor, or taking prescription drugs (Conway 1985).
This literature faces the major challenge that being hit by adverse life events
is most likely correlated with unobserved individual characteristics (such as
lifestyle), which in turn may matter for the individual’s inclination to be
religious.
12 Calculated using analysis of variance, where the unit of analysis is individuals and the groups are
countries.
13 The reasons for focusing on the demand-side are the following. First, the supply-side theories have not
received strong support in the data. Second, the supply-side theories that have received support in the
data (such as the theory based on secular competition) are most likely mainly suitable for development of
religiosity in the US. This chapter explores global differences in religiosity. Third, the surveyed data allows
for a test of the demand-side, not the supply-side. Last, when asked, the religious state that one of the
main purposes of religion is to provide buffering against life stressors (see, e.g., Clark 1958 and Pargament
2001).
14 For example, Pargament (2001), Cohen and Wills (1985), Park et al. (1990), Williams et al. (1991).
The terminology “religious coping” stems from psychology, but other labels have been used. For instance,
religious buffering, the religious comfort hypothesis, and psychological social insurance.
15 See, for example, Ano and Vasconcelles (2005) and Pargament (2001) for reviews.
272 J. S. Bentzen
Norenzayan and Hansen (2006) addressed the endogeneity concern in a

controlled experiment of 28 undergraduate students from the University of
Michigan. They obtained exogenous variation in thoughts of death by asking
half of the students questions such as “What will happen to you when you
die?” After the experiment, the students primed with thoughts of death were
more likely to reveal beliefs in God and to rank themselves as more religious.
While solving the endogeneity issue, the conclusions based on 28 students
in Michigan cannot necessarily be extended to the world at large. The study
cannot tell us whether elderly from California or students from Pakistan
would respond in the same way. Yet, the theory is that religious coping is not
something peculiar to Christianity. For instance, Pargament (2001) notes that
(p. 3) “While different religions envision different solutions to problems, every
religion offers a way to come to terms with tragedy, suffering, and the most
significant issues in life.”16 Performing lab experiments for a representative
global population is rather tedious and costly. Instead one can exploit natural
experiments to obtain exogenous variation in the extent to which individuals
experienced an unpredictable adverse event.
Endogeneity
One adverse and unpredictable event is natural disasters, especially earth-

quakes. Indeed, the belief that natural disasters carried a deeper message
from God was the rule rather than the exception before the Enlightenment
(e.g., Hall 1990; Van De Wetering 1982). Later, the famous 1755 Lisbon
earthquake has been compared to the Holocaust as a catastrophe that trans-
formed European culture and philosophy. Penick (1981) documents more
systematically that US states hit by massive earthquakes in 1811 and 1812 saw
church membership increase by 50% in the following year, compared to an
increase of only 1% in remaining states. More recently, Sibley and Bulbulia
(2012) found that conversion rates increased more in the Christchurch region
after the large earthquake in 2011, compared to the remaining four regions
of New Zealand. Other disasters may have left an imprint on religiosity. For
instance, Ager et al. (2016) find that church membership increased in counties
affected by the Mississippi River flood of 1927.
While the mentioned studies certainly solve the endogeneity issue relating
to adverse life events, their conclusions are potentially not externally valid.
Bentzen (forthcoming) solves this issue by combining the mentioned globally
16 See also Feuerbach (1957), Freud (1927), and Marx (1867) for similar generalizations across all religions.
0.25
0.2
0.15
0.1
0.05
0
Average Christians Catholics Protestants Muslims Hindus Buddhists Others
-0.05
-0.1
-0.15
-0.2
Fig. 16.2 Impact of disaster risk on religiosity across religious denominations. Notes:
The figure shows the parameter estimate on earthquake risk in a regression on religios-
ity, accounting for country-by-year fixed effects and a dummy for actual earthquakes
during the past year
comparable data on religiosity at the subnational level with data on earthquake

risk and actual earthquakes.17,18
The analysis by Bentzen (forthcoming) first documents that individuals
living in districts more frequently hit by earthquakes are indeed more religious
than those living in areas with fewer earthquakes. Figure 16.2 shows the
impact of earthquake risk on religiosity for the world on average and for
each major religious denomination. Religiosity increases for Christians (both
Protestants and Catholics), Muslims, Hindus, and others, but is statistically
indistinguishable from zero for Buddhists. Note the very large standard errors
for Hindus and Buddhists, reflecting few Hindus and Buddhists in the sample.
In addition to being statistically large, the impact of earthquake risk on
religiosity is also economically large: The size of the effect amounts to 70%
17 The data on earthquake risk measures the risk of getting hit by an earthquake of a certain size within the
next 50 years. The data on earthquake events measures the exact location of actual earthquakes of various
strengths. Larger earthquakes increase religiosity more. See more details in Bentzen (forthcoming).
18 Note, though, that increased religiosity after an earthquake could be due to other things unrelated to
religious coping, which I will return to in section “The Mechanism”.
274 J. S. Bentzen
of the well-established gender difference in religiosity.19 The result is robust to

various changes (see Bentzen (forthcoming) for details). Similar results obtain
for other unpredictable major disasters such as volcanic eruptions and tsunamis
and for different measures of earthquake risk.
A central concern is that important district-level factors are left out of the
analysis, biasing the results. For instance, earthquakes are more likely along
the coast where tectonic plates often meet. At the same time, being close
to the coast may influence religiosity independently. In that case, omitting
distance to the coast in the regression would create a spurious relation between
earthquakes and religiosity. This is the reason why Bentzen (forthcoming)
includes a control for distance to the coast throughout. Broader geographic
confounders are controlled for by including country-fixed effects and absolute
latitude throughout. Additional district-level controls include recent actual
earthquakes, population density, light intensity, the share of arable land,
average temperature, average and variance of precipitation, district area, and a
dummy equal to one if the district is often hit by earthquakes. Since the data
is available at the individual level, various individual-level confounders can
also be accounted for, such as age, gender, marital status, income, education,
employment status, and various measures of other cultural values.
Unobservable time-varying factors are not accounted for in the cross-section
analysis. To account for these factors, the time dimension of the data can
be exploited. The same individuals are not followed over time, but instead
Bentzen (forthcoming) exploits that a third of the subnational districts are
followed over time. This enables constructing a so-called synthetic panel,
where the districts are the panel dimension. It turns out that district-level
religiosity increases when an earthquake hit in between the years of interview in
keeping with the religious coping hypothesis. Investigating the effect of disaster
on the individual measures of religiosity reveals that the religious become more
religious, while evidence for conversion into religion is somewhat weaker.20
Consistent with a literature on dynamic effects of various shocks on cultural
values, the short-term spike in religiosity after an earthquake abates with time.
Regarding the surprise element, an earthquake in a district that is otherwise
rarely hit increases religiosity more than an earthquake in a district that is often
hit. The phenomenon that earthquakes can still, in the modern world, affect
19 This means that the standardized parameter estimate on earthquake risk amounts to 70% of the
standardized parameter estimate on a gender dummy. It is well known in the literature that women are
more religious than men, for example, Miller and Hoffmann (1995).
20 The result is robust to adding country-by-year fixed effects, individual- and district-level controls, and
rather comforting, future earthquakes have no impact on current levels of religiosity.
belief is further illustrated by a Gallup survey conducted in the aftermath of

the great 1993 Mississippi River floods, which asked Americans whether the
recent floods were an indication of God’s judgment upon the sinful ways of
the Americans. A total of 18 % answered in the affirmative (Steinberg 2006).
Persistence in Religiosity
The event analysis eliminates unobservable district-level factors, which is a

great improvement of macroeconomic research in general, where eliminating
country-level unobservables is even rarely achieved. However, one caveat of
the event analysis is that it can say nothing about the long-term effects on
religiosity. Indeed, the fact that the effect abates with time speaks for an
investigation of whether elevated uncertainty (earthquake risk) has a lasting
effect on religiosity. If a lasting effect exists, parents must be transmitting
religiosity to their children. In a model of cultural transmission, parents will
choose to transmit a particular cultural trait to their children if this grants
utility to either parents or children (e.g., Bisin and Verdier 2001). Empirical
evidence suggests that religiosity may be such a trait: Religion is likely to
improve mental health, life satisfaction, abilities to cope with adverse life
events, and deter deviant behavior.21 Thus, it seems theoretically likely that
parents might choose to transmit their religion to their children.
Bentzen (forthcoming) investigates whether the impact of earthquake risk
transmits across generations by combining data on earthquake risk with a
dataset with information on children of migrants currently living in Europe,
but whose parents came from various countries across the globe.22 It turns
out that children of migrants whose parents came from countries with high
earthquake risk are more religious than those from low earthquake risk areas,
independent of actual earthquake risk and level of religiosity in their current
country of residence. It seems that living in high earthquake risk areas instigates
a culture of religiosity that is passed on to future generations like many other
cultural values.
21 For instance, Miller et al. (2014), Campante and Yanagizawa-Drott (2015), Clark and Lelkes (2005),
and Lehrer (2004). See also reviews by Smith et al. (2000) and Pargament (2001).
22 This analysis is based on the European Social Survey. The methodology used was coined the epidemi-
ological approach by Fernandez (2011).
276 J. S. Bentzen
The Mechanism
An effect of earthquakes on religiosity does not necessarily confirm the reli-

gious coping hypothesis. Instead of religion serving as psychological insurance,
it could also act as physical insurance. Likewise, the effect of earthquakes on
religiosity could cover that atheists move out in response to an earthquake or
that earthquakes reduce development levels, increasing religiosity, and so on.
Additional characteristics of the data can be exploited to disentangle different
explanations. It turns out that while other factors can explain some results, the
psychological story is the only mechanism that explains all results.
For instance, if elevated religiosity levels after an earthquake is simply
driven by people going to church for material needs, church going should
increase after an earthquake. On the other hand, the literature on religious
coping finds that people mainly use intrinsic religiosity (ones’ personal relation
to God) to cope with adversity and to a lesser extent extrinsic religiosity
(going to church).23 Bentzen (forthcoming) shows that only intrinsic reli-
giosity increases in response to a recent earthquake, while church going is
not affected in the short term. Church going is, however, over the long
term. Therefore, physical insurance might explain some of the long-term
results, but not the short-term results. Further, if the mechanism is physical
insurance, other disasters that pose the same material losses should also increase
religiosity. The degree of predictability should not matter much. On the
other hand, the religious coping hypothesis states that individuals use religion
more when faced with adverse unpredictable events and less with predictable
ones.24 Major geophysical and meteorological disasters can be grouped in
terms of predictability. For instance, meteorologists have a much easier time
predicting storms than seismologists have predicting earthquakes.25 Also,
23 For example, Johnson and Spilka (1991) or review by Pargament (2001). Koenig et al. (1988) found
that the most frequently mentioned coping strategies among 100 older adults dealing with 3 stressful
events were faith in God, prayer, and gaining strength from God. Social church-related activities were less
commonly noted. Similarly, a medical study by Miller et al. (2014) found that individuals for whom
religion is more important experienced reduced depression risk (measured by cortical thickness), while
frequency of church attendance was not associated with thickness of the cortices.
24 For example, Norris and Inglehart (2011), Sosis (2008), Park et al. (1990). See also Mattlin et al.
(1990) on how practical everyday problems are less likely to trigger religious coping compared to large
bad events. Skinner (1948) found that this reaction to unpredictability extends into the animal world.
Pigeons subjected to an unpredictable feeding schedule were more likely to develop inexplicable behavior,
compared to the birds not subject to unpredictability. Since Skinner’s pioneering work, various studies
have documented how children and adults in analogous unpredictable experimental conditions quickly
generate novel superstitious practices (e.g., Ono 1987).
25 The US Geological Survey (USGS) notes that earthquakes cannot be predicted (https://www2.usgs.
gov/faq/categories/9830/3278). See also this post about our ability to forecast storms and their paths,
earthquakes can be grouped into more or less surprising ones. Consistent with
the religious coping literature, surprising disasters increase religiosity more
than less surprising ones for equal amount of damage (Bentzen forthcoming).
For instance, elevated risk of earthquakes, tsunamis, and volcanic eruptions
increase religiosity, while storm risk has no effect on religiosity. Storms result
in comparable material and personal losses and thus should instigate the same
effect on religiosity if the explanation was physical insurance. Also, earthquakes
in areas frequently hit by earthquakes affect religiosity less than earthquakes in
areas otherwise rarely hit.
The effect of earthquakes on religiosity could also be due to atheists
moving out in the face of disaster. This theory can potentially explain the
short-term effect per se, but not the tendency for the effect to abate with
time. Explaining this tendency with population movements would mean that
atheists move out in the immediate aftermath of the earthquake, but then
choose to move into the district again after 6–12 years, only to move out again
when the next earthquake hits. Perhaps a more plausible alternative story is
that reconstruction workers move into the district after an earthquake. This
tendency can explain the uncovered results if the reconstruction workers are
more religious than the average person and if they stay for 6–12 years before
moving out again. While this is plausible, it probably does not drive the entire
effect. On the other hand, the fall in religiosity after a while is reconcilable
with the idea that religion provides stress relief, reducing the need for religion
after a while.
If the effect is due to a direct impact of earthquakes on income, the effect
should fall when accounting for personal or regional development levels.
This is not the case. Last, if religiosity is just part of the characteristics of a
different type of people developing in earthquake areas, the effect should fall
when controlling for other cultural characteristics stressed as important in the
literature (e.g., trust, independence, thriftiness, preference for hard work, etc.).
This is also not the case.
To sum up, some of the alternative explanations involving physical insur-
ance, direct economic loss, migration/selection, or a special culture evolving
in high-risk areas can explain some of the results. Thus, each individual set
of results is probably due to a combination. But the only explanation that
can explain all uncovered results across all three analyses (cross-section, event
study, and cross-generational) is religious coping.
as opposed to our inability to forecast earthquakes: https://www.tripwire.com/state-of-security/risk-

based-security-for-executives/risk-management/hurricanes-earthquakes-prediction-vs-forecasting-in-
information-security/.
278 J. S. Bentzen
Conclusion
Religion may matter for important socioeconomic outcomes. Thus, identi-
fying its causes seems relevant. Economics of religion has taken us far in the
understanding of the theoretical foundations. The next step is to disentangle
and test the theories empirically. Supply-side theories have obtained only
mixed empirical support. Likewise for the secularization hypothesis. However,
these empirical investigations have been compromised by lack of global data
and lack of proper identification.
Equipped with globally comparable data on religiosity and a novel iden-
tification strategy, Bentzen (forthcoming) found evidence consistent with
one of the major demand-side theories for differences in religiosity today.
Religiosity increases in response to earthquakes, and the impact of living in
high earthquake risk areas transmits across generations in the form of elevated
religiosity. The reason is most likely that religion is used for psychological
comfort, meaning that individuals hit by adverse and unpredictable life events
can use their religion to gain comfort and understanding. In conclusion, one
reason for the large differences in religiosity across the globe today is differences
in unpredictability, caused in particular by differences in the risk and actual
occurrence of earthquakes, tsunamis, and volcanic eruptions.
A path for future research is to link the presented globally comparable
data on religiosity to other data available at the subnational district level in
order to test remaining theories. Furthermore, once the causes of differences
in religiosity have been identified more fully, a next step will be to properly
investigate its socioeconomic consequences.
References
Ager, P., C.W. Hansen, and L. Loenstrup. 2016. Church Membership and Social
Insurance: Evidence from the American South. University of Southern Denmark
Discussion Papers on Business and Economics No. 7/2016.
Ano, Gene G., and Erin. B. Vasconcelles. 2005. Religious Coping and Psychological
Adjustment to Stress: A Meta-Analysis. Journal of Clinical Psychology 61 (4): 461–
480.
Azzi, Corry, and Ronald Ehrenberg. 1975. Household Allocation of Time and Church
Attendance. Journal of Political Economy 83 (1): 27–56.
Becker, S.O., and L. Woessmann. 2009. Was Weber Wrong? A Human Capital
Theory of Protestant Economic History. The Quarterly Journal of Economics 124
(2): 531–596.
Becker, Sascha O., Markus Nagler, and Ludger Woessmann. 2017. Education and
Religious Participation: City-Level Evidence from Germany’s Secularization Period
1890–1930. Journal of Economic Growth 22 (3): 273–311.
Bentzen, Jeanet Sinding. Forthcoming. Acts of God? Religiosity and Natural Disasters
Across Subnational World Districts. The Economic Journal.
Bentzen, Jeanet Sinding, and Gunes Gokmen. 2017. The Power of Religion: Political
Origins of Religion Across the Globe. Unpublished manuscript.
and the Dynamics of Preferences. Journal of Economic Theory 97 (2): 298–319.
Boyer, Pascal. 2008. Religion Explained. London: Random House.
Brown, D.E. 1991. Human Universals. Philadelphia: Philadelphia Temple University
Press.
nomic Growth and Happiness? Evidence from Ramadan. The Quarterly Journal of
Economics 130 (2): 615–658.
Chaves, Mark, and Philip S. Gorski. 2001. Religious Pluralism and Religious
Participation. Annual Review of Sociology 27 (1): 261–281.
Clark, Andrew, and Orsolya Lelkes. 2005. Deliver Us from Evil: Religion as Insurance.
Papers on Economics of Religion 603: 1–36.
Clark, Walter Houston. 1958. How Do Social Scientists Define Religion? The Journal
of Social Psychology 47 (1): 143–147.
Cohen, S., and T. A. Wills. 1985. Stress, Social Support, and the Buffering Hypothesis.
Psychological Bulletin 98 (2): 310–357.
Conway, K. 1985. Coping with the Stress of Medical Problems Among Black and
White Elderly. The International Journal of Aging and Human Development 21 (1):
39–48.
Fernandez, R. 2011. Does Culture Matter? The Netherlands: North-Holland.
Feuerbach, Ludwig. 1957. The Essence of Christianity. New York: Barnes & Noble
Publishing.
Finke, Roger, and Rodney Stark. 2005. The Churching of America, 1776–2005:
Winners and Losers in Our Religious Economy. Piscataway: Rutgers University Press.
Freud, Sigmund. 1927. The Future of an Illusion. Peterborough: Broadview Press.
Gruber, Jonathan, and Daniel M. Hungerman. 2008. The Church Versus the
Mall: What Happens When Religion Faces Increased Secular Competition? The
Guiso, L., P. Sapienza, and L. Zingales. 2003. People’s Opium? Religion and
Economic Attitudes. Journal of Monetary Economics 50 (1): 225–282.
Hall, D.D. 1990. Worlds of Wonder, Days of Judgment: Popular Religious Belief in Early
New England. Harvard: Harvard University Press.
Hungerman, Daniel M. 2005. Are Church and State Substitutes? Evidence from the
1996 Welfare Reform. Journal of Public Economics 89 (11): 2245–2267.
———. 2010. Rethinking the Study of Religious Markets. In Handbook of the
Economics of Religion, 257–275. Oxford: Oxford University Press
280 J. S. Bentzen

Literature 36 (3): 1465–1495.
Inglehart, R., and P. Norris. 2003. Rising Tide: Gender Equality and Cultural Change
Around the World. Cambridge: Cambridge University Press.
(2): 395–441.
Johnson, Sarah C., and Bernard Spilka. 1991. Coping with Breast Cancer: The Roles
of Clergy and Faith. Journal of Religion and Health 30 (1): 21–33.
Kimball, Miles S., Colter M. Mitchell, Arland D. Thornton, and Linda C. Young-
Demarco. 2009. Empirics on the Origins of Preferences: The Case of College Major
and Religiosity. Technical Report. National Bureau of Economic Research.
Koenig, H.G., L.K. George, and I.C. Siegler. 1988. The Use of Religion and Other
Emotion-Regulating Coping Strategies Among Older Adults. The Gerontologist 28
(3): 303–310.
Lehrer, Evelyn L. 2004. Religion as a Determinant of Economic and Demographic
Behavior in the United States. Population and Development Review 30 (4):707–726.
Marx, Karl. 1867. Capital: A Critique of Political Economy. London: The Penguin
Group.
Mattlin, Jay A., Elaine Wethington, and Ronald C. Kessler. 1990. Situational
Determinants of Coping and Coping Effectiveness. Journal of Health and Social
Behavior 31 (1): 103–122.
McCleary, R.M., and R.J. Barro. 2006. Religion and Economy. The Journal of
Economic Perspectives 20 (2): 49–72.
Miller, Alan S., and John P. Hoffmann. 1995. Risk and Religion: An Explanation of
Gender Differences in Religiosity. Journal for the Scientific Study of Religion 31 (4):
63–75.
Miller, Lisa, Ravi Bansal, Priya Wickramaratne, Xuejun Hao, Craig E. Tenke,
Myrna M. Weissman, and Bradley S. Peterson. 2014. Neuroanatomical Correlates
of Religiosity and Spirituality: A Study in Adults at High and Low Familial Risk
for Depression. JAMA Psychiatry 71 (2): 128–135.
Murdock, G.P. 1965. Culture and Society: Twenty-Four Essays. Pittsburgh: University
of Pittsburgh Press.
Norenzayan, A., and I.G. Hansen. 2006. Belief in Supernatural Agents in the Face of
Death. Personality and Social Psychology Bulletin 32 (2): 174–187.
Norenzayan, Ara. 2013. Big Gods: How Religion Transformed Cooperation and Conflict.
Norris, Pippa, and Ronald Inglehart. 2011. Sacred and Secular: Religion and Politics
Worldwide. Cambridge: Cambridge University Press.
Olson, Dan. 2011. Toward Better Measures of Supply and Demand for Testing
Theories of Religious Participation. In The Oxford Handbook of the Economics
of Religion. Oxford: Oxford University Press.
Ono, K. 1987. Superstitious Behavior in Humans. Journal of the Experimental Analysis

of Behavior 47 (3): 261.
Pargament, K.I. 2001. The Psychology of Religion and Coping: Theory, Research, Practice.
New York: Guilford Press.
Park, C., L.H. Cohen, and L. Herb. 1990. Intrinsic Religiousness and Religious
Coping as Life Stress Moderators for Catholics Versus Protestants. Journal of
Personality and Social Psychology 59 (3): 562.
Penick, J.L. 1981. The New Madrid Earthquakes. Columbia: University of Missouri.
Peoples, Hervey C, Pavel Duda, and Frank W. Marlowe. 2016. Hunter-Gatherers and
the Origins of Religion. Human Nature 27 (3): 261–282.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle
East Did Not. Cambridge: Cambridge University Press.
Scheve, K., and D. Stasavage. 2006. Religion and Preferences for Social Insurance.
Quarterly Journal of Political Science 1 (3): 255–286.
Sibley, Chris G., and Joseph Bulbulia. 2012. Faith After an Earthquake: A Longitudi-
nal Study of Religion and Perceived Health Before and After the 2011 Christchurch
New Zealand Earthquake. PloS One 7 (12): e49648.
Skinner, B.F. 1948. ’Superstition’ in the Pigeon. Journal of Experimental Psychology 38
(2): 168.
Smith, Bruce W., Kenneth I. Pargament, Curtis Brant, and Joan M. Oliver. 2000.
Noah Revisited: Religious Coping by Church Members and the Impact of the
1993 Midwest Flood. Journal of Community Psychology 28 (2): 169–186.
Sosis, R. 2008. Pigeons, Foxholes, and the Book of Psalms. In Evolution of Religion,
ed. Bulbulia, Joseph, et al., 103–109. Santa Margarita: Collins Foundation.
Steinberg, T. 2006. Acts of God: The Unnatural History of Natural Disaster in America.
Van De Wetering, M. 1982. Moralizing in Puritan Natural Science: Mysteriousness
in Earthquake Sermons. Journal of the History of Ideas 43 (3): 417–438.
Voas, David, and Mark Chaves. 2016. Is the United States a Counterexample to the
Secularization Thesis? American Journal of Sociology 121 (5): 1517–1556.
Weber, M. 1905. The Protestant Ethic and the Spirit of Capitalism. London: Penguin
Books.
Williams, D.R., D.B. Larson, R.E. Buckler, R.C. Heckmann, and C.M. Pyle. 1991.
Religion and Psychological Distress in a Community Sample. Social Science &
Medicine 32 (11):1257–1262.
17
Socioeconomic Inequality Across Religious
Groups: Self-Selection or Religion-Induced
Human Capital Accumulation? The Case
of Egypt
Mohamed Saleh
Socioeconomic inequality across religious groups, such as between Protestants

and Catholics in Western Europe, Hindus and Muslims in India, Jews and
non-Jews in the US and Europe, has been the subject of a voluminous literature
in social sciences and, more recently, economics. Perhaps the most well-known
explanation of the phenomenon dates back to Max Weber (1930 [1905]),
who traced the Protestant-Catholic socioeconomic gap to Protestantism’s
culture of work ethic and individualism. Extending his thesis to Asia, Weber
hypothesized in a similar vein that Asiatic religions such as Hinduism and
Buddhism were less conducive to capitalism. The more recent economics of
religion literature, while acknowledging the potential endogeneity of religion,
attempted to disentangle the causal impact of religious beliefs on socioe-
conomic outcomes, first in cross-country regressions (Barro and McCleary
2003) and then in single-country studies (Borooah and Iyer 2005; Becker and
Woessmann 2009; Chaudhary and Rubin 2011). A common narrative in the
This chapter draws extensively on Saleh, Mohamed (2018), “Taxation, Conversions, and the Coptic-
Muslim Socioeconomic Gap in Medieval Egypt,” Journal of Economic History. I refer the reader to this
article for a more complete analysis of the topic. Copyright © 2018 The Economic History Association.
Used with permission.
M. Saleh ()
Toulouse School of Economics and Institute for Advanced Study in Toulouse,
Toulouse, France
e-mail: mohamed.saleh@tse-fr.eu

284 M. Saleh
latter line of literature is that some religions put more emphasis than others
on the accumulation of human capital.
While the recent economics of religion literature has put socioeconomic
inequality across religious groups back to the center of debate in economics,
and social sciences more generally, the literature in its attempt to estimate
the causal effect of religious beliefs on socioeconomic outcomes has largely
overlooked another (equally) plausible explanation of the phenomenon: that
of sorting or self-selection on socioeconomic status (henceforth, SES) into reli-
gions. This hypothesis was first put forward by Weber himself when he noted
that conversions to Christianity and Islam in India were more concentrated
among the lower Hindu castes (Weber 1996 [1958], p. 6). Another example
of the self-selection hypothesis is the study by Maristella Botticini and Zvi
Eckstein (2005), who argued that Rabbinic Judaism with its emphasis on
literacy to be able to read the Torah and the Talmud drove Rabbinic Jews
with a lower taste for education out of Judaism, leading the remaining Jews
to shrink into a better-off minority. Their thesis seems to combine the two
hypotheses. On the one hand, there is self-selection on socioeconomic status
into religion, since less educated (likely poorer) Jews systematically converted
out of Judaism. On the other hand, there is a causal impact of religious beliefs
on the accumulation of human capital, since investment in a child’s human
capital is driven here by belief in Judaism, and this higher preference for human
capital is in turn the primary cause of the positive selection on education of
the remaining (non-convert) Jews.
Inspired by the selection conjecture, Saleh (2018) argues that an
examination of socioeconomic inequality across religious groups requires
a deeper examination of the historical formation process of each group,
which may (or may not) have been characterized by self-selection on SES.
Indeed, religious groups are not fixed ahistorical categories, but are necessarily
formed via a process of conversion of an initial population. The study
of socioeconomic inequality between Copts (Egyptian Christians), who
constitute around 6 percent of Egypt’s population, and the Muslim majority
is illuminating in this regard. Using a novel data source, two nationally
representative individual-level samples from the 1848 and 1868 population
censuses that I digitized from the original manuscripts at the National Archives
of Egypt (NAE) (Saleh 2013) and that are among the earliest pre-colonial
censuses from any non-Western country, I documented that Copts were over-
represented among white-collar workers (mostly, mid-low bureaucrats) and
artisans in 1848 and 1868. Among adult employed men, 50 percent of Copts
were white-collar workers and artisans, compared to only 20 percent among
17 Socioeconomic Inequality Across Religious Groups: Self-Selection… 285
Muslims.1 This phenomenon is striking if we take into account that Egypt

was Coptic Christian before the Arab conquest of Egypt in 641, and since
in- and out-migration were limited, Egypt’s “Copts” and “Muslims” are both
descendants of the pre-641 “Coptic” population that either chose to remain
Coptic or to convert to Islam.2 Bearing this fact in mind, Saleh (2018) argues
that Copts’ conversion to Islam was characterized by selection on SES due
to the tax system. Upon the conquest, Arabs imposed an annual poll tax on
every adult free Coptic male, which was enforced until 1856. As conversion
freed Copts from the poll tax liability, and since the conversion incentive
was decreasing in income owing to the (quasi) lump-sum feature of the tax,
I hypothesize that the tax caused the conversion of poorer Copts, leading
Copts to shrink into a better-off minority. Restrictions on apprenticeships
and schooling imposed by each religious group led the initial selection to
perpetuate thereafter.
There are a few distinguishing features of this hypothesis in comparison
to the existing studies that emphasize self-selection in religious conversions.
Whereas Weber (1996 [1958]) did not specify a selection mechanism of
converts, selection is caused here by an economic incentive, the tax exemption.
This incentive is distinct from the religious incentive to read the Bible in
Botticini and Eckstein (2005). And unlike the latter article, the persistence
of the Coptic-Muslim SES gap is explained by group restrictions on acquiring
skills, in essence a supply-side explanation, and not by Copts’ higher preference
for human capital (i.e. the demand side), a point that I will come back to later.
The study of taxation, conversions, and the Coptic-Muslim SES inequal-
ity contributes to a century-long debate on the Middle East. Inspired by
major papyri discoveries from early Islamic Egypt, pioneering early-twentieth-
century historians and papyrologists such as Julius Wellhausen (1927 [1902]),
Carl Becker (1902), Harold Bell (1910), and Adolphus Grohmann (1932)
emphasized the tax incentive of conversion to Islam under the early Arab
caliphate. The hypothesis of tax-induced conversions triggered fierce debates
among historians, however. While scholars such as Gladys Frantz-Murphy
(2004) and Yossef Rapoport (2004) endorsed the hypothesis, others such
as Daniel Dennett (1950), Kosei Morimoto (1981), and Petra Sijpesteijn
1 Copts were not a political elite minority though. In 1848–1868 half of Copts were farmers and unskilled
workers, and Muslims (mostly Turks) monopolized top political elite positions. The highest white-collar
positions that Copts reached were limited to mid-low bureaucracy such as scribes, accountants, and land
tax collectors. However, the hypothesis of self-selected conversions seeks to understand why Copts were
richer, on average, than Muslims.
2 Converts could not switch back to Christianity due to the death penalty of apostates in Islam.
286 M. Saleh
(2013) took a cautious stance in light of the growing papyri discoveries that
suggested that conversions in Egypt may not have started until the mid-
eighth century. A third group argued that conversions occurred even later
for other causes including the suppression of Coptic tax revolts in the ninth
century (Al-Maqrizi 2002 [1500]; Mikhail 2004) and state persecution in
1250–1517 (El-Leithy 2005), while a fourth group contended that it was
Islam’s appeal that attracted converts (El-Shayyal 1966). Despite this large
body of scholarship, and partially due to data limitations, the “conversion”
literature did not address the SES advantage of the surviving non-Muslim
minorities, a task that was left to a separate body of (qualitative) literature
(Tagher 1998 [1951]; Issawi 1981; Courbage and Fargues 1997), and so the
impact of taxation on the inter-religious SES differences in the region has
remained a black box. As a result of this omission, neither the regressivity
of the poll tax nor the possibility of selection on SES of converts, a logical
consequence of tax regressivity, was examined by this literature, with the
exception of a conjecture by Courbage and Fargues (1997, pp. 22–23). Notice
that the argument here is that taxation led to self-selected conversions, and
not simply, conversions, as suggested by the early-twentieth-century scholars.
This distinguishes taxation from the other causes of conversion that did not
necessarily trigger self-selection of converts. In particular, I do not claim that
taxation was the sole cause of conversions, but that, compared to the other
causes, taxation offers a consistent answer to both conversions and the Coptic-
Muslim SES gap.
The long-term trends of the poll tax, Copts’ population share, and the
Coptic-Muslim SES gap are broadly consistent with the selection hypothesis.
To construct these trends, I draw on novel data sources including a village-
level dataset on Christian churches and monasteries in 1200 and 1500 and an
individual-level dataset on occupations and religion in 641–969 (N = 402)
from Egyptian papyri, in addition to the population census samples from 1848
and 1868. The trends suggest that the higher poll tax rate before 1250 was
correlated with a decline in Copts’ population share and the emergence of
a Coptic-Muslim SES gap as farmers and unskilled Copts were more likely
to convert, but that conversions subsided afterward as the tax rate declined.
Since taxes were administered at the local level, the econometric evidence
on the hypothesis is based on exploiting the cross-district variation in the
average poll tax that is observed in the extant papyrological individual-level
poll tax payment records in 641–1100. Tax papyri are subject to certain caveats,
however. They survived for only 4 out of 42 kuras (Egypt’s administrative units
in 641–1036) that map into 11 out of 76 districts in 1848–1868, all located in
the Nile Valley, and most papyri are dated within a range, such as a century,
rather than a specific date. There are two outcomes of interest. The first is
Copts’ population share, which I measure in 1200 and 1500 by the village-
level presence of at least one Coptic church or monastery and in 1848–1868 by
the individual-level religious affiliation in the population census samples. The
second outcome is the Coptic-Muslim SES (occupational) gap, which I am
able to observe at the district level in only 1848–1868, but not before. The
findings lend support to the selection hypothesis. I document that compared
to Copts in “low-tax” districts, Copts whose origin is in a “high-tax” district
are relatively fewer in 1200, 1500, and 1848–1868, but differentially more likely
to be artisans and white-collar workers in 1848–1868. Since all districts were
(almost) 100 percent Coptic in 641, the findings suggest that high-tax districts
witnessed relatively more conversions and a more extensive selection on SES
that resulted in a greater Coptic-Muslim SES gap.
The empirical evidence indicates that the initial positive selection of
non-convert Copts between 641 and 1200 persisted for over a millennium.
I argue that this is likely due to group restrictions on apprenticeships and
schooling. As conversions sorted Copts and Muslims on occupations, each
group then attempted to exclude the other from the artisanal and white-collar
occupations in which it was over-represented, via limiting apprenticeships
and schooling within group members. Copts restricted access to skills that
were required for jobs in mid-low bureaucracy. While Coptic elementary
schools taught arithmetic and geometry in order to train Coptic children
for jobs in mid-low bureaucracy, Muslim schools did not provide this
training (Heyworth-Dunne 1938, pp. 2–7, 84–92). However, it was primarily
apprenticeships, not schools, that trained Coptic children for bureaucratic
jobs. In Fatimid Egypt (969–1171), “the persistence of Coptic administrative
personnel [was because] the agrarian administration was very complex and
not easily mastered. In it the Copts played an important role at the local level
as well as at the central offices in the capital . . . The administrative knowledge
was passed on by the officials in their families when fathers employed their
sons, thus maintaining the hold of the family over posts” (Samir 1996, p. 190).
In the words of Lord Cromer, the British consul of Egypt in 1883–1908, the
Coptic accounting system was “archaic” and “incomprehensible to anyone but
themselves” (Tagher 1998 [1951], p. 213). Copts used fractions and “ambiguous
abbreviations” in accounting based on units of measurement in use in rural
Egypt. Group effects on acquiring human capital were not limited to Copts in
mid-low bureaucracy though. Copts were legally banned from the judiciary,
military, police, and clergy, and these jobs were thus monopolized by Muslims.
Muslims were banned from brewing that became a Coptic specialization.
The 1848 and 1868 census samples reveal that Copts were over-represented
288 M. Saleh
among jewelers, dyers, carpenters, weavers, and tailors, whereas Muslims

were over-represented among blacksmiths, sawyers, bakers, and butchers.
Raymond (1973, pp. 544–551) suggests that the reason for the persistence of
this occupational specialization was restricting apprenticeships.
There are other theories that can explain both the emergence and the
persistence of Copts’ economic advantage over Muslims. I refer the reader to
Saleh (2018) for a discussion of these theories, but I highlight a few theories
here. A well-known explanation of the socioeconomic advantage of local
non-Muslim minorities in the Middle East attributes the phenomenon to
the rising European influence in the Ottoman Empire after 1800. Charles
Issawi (1981) and Timur Kuran (2004) postulated that the privilege of non-
Muslim minorities emerged in the nineteenth century (and not before) because
Europeans favored non-Muslims or because non-Muslims adopted European
legal structures through the capitulations system. Under this system, non-
Muslims in the Ottoman Empire were permitted to use European laws in
conflict resolution, whereas Muslims were obliged to resort to Islamic law. The
role of the expansion of European schools in the rise of non-Muslim minorities
has also been explored (Salama 1963; Abecassis 2000). But although these
theories may indeed account for part of the SES advantage of non-Coptic
Christians (such as Levantines, Armenians, Greeks) and Jews and for the
(possibly) widening SES gap between Copts and Muslims in Egypt after 1850,
it does not fully account for the findings. It does not explain why the Coptic-
Muslim SES gap emerged by 969 (as indicated by the papyrological evidence),
long before the rise of Europe. Furthermore, the 1848–1868 censuses record
the “protégé” status that individuals purchased from European consulates in
return for having access to European legal structures. The share of protégés was
negligible among Copts in both 1848 and 1868 and was only sizable among
non-Coptic Christians and Jews; the two groups constituted only 6 percent of
non-Muslims in Egypt and were mostly urban. These differences are likely due
to occupational specialization. While non-Coptic Christians and Jews excelled
in commerce, a profession that involved transactions with Europeans where
access to European laws mattered, Copts’ advantage stemmed from artisanship
and bureaucracy where European laws had less of an influence.
Nevertheless, even if the emergence of the Coptic-Muslim SES gap can be,
at least partially, traced to the initial selection on SES due to, say, the tax
system, the persistence of inequality can still be attributed to other factors.
These include, most importantly, the intergenerational transmission of inter-
group cultural differences that are caused by differences in religious beliefs.
For example, one can hypothesize, following Botticini and Eckstein (2005),
that Coptic Christianity, like Rabbinic Judaism, encouraged the accumulation
of human capital which pushed Copts with a lower preference for education
to convert to Islam. And even if self-selected conversions were triggered
by the tax system, one can still argue that Coptic Christianity encouraged
human capital accumulation once non-convert Copts shrank into a minority.
This theory is unlikely to hold though as there is no literacy requirement
under Coptic Christianity and illiteracy among adult male Copts in 1986
was 34 percent. Furthermore, Coptic schools were purely religious in 641
when conversions started, and their shift toward teaching secular subjects by
1700 was possibly a result of selected conversions on SES. Another cultural
explanation is hypothesized, inspired by Weber (1930 [1905]), that Copts
had a stronger work ethic. Yet, Coptic Christianity shared with the Egyptian
Muslim Sufi culture a mystical outlook on life that attributed materialistic
success to metaphysical factors rather than to hard work. Moreover, the fact
that Copts’ advantage stemmed from bureaucracy and artisanship, and not
from commerce, indicates that Coptic Christianity was not more conducive
to capitalism than Islam.
This is not to say that there were no other historical processes, besides group
restrictions on skills, that affected the Coptic-Muslim SES gap, however. To be
sure, state policies throughout Egypt’s history and European influence starting
from 1800 had their effects. For example, the Coptic-Muslim differences in
human capital were altered in favor of Copts with state industrialization in the
nineteenth century and with the expansion of European schooling after 1850
and then in favor of Muslims, with the introduction of public mass education
a century later in 1951–1953.
Saleh (2015) examines the impact of nineteenth-century Egypt’s state
industrialization on the Coptic-Muslim occupational differences. Muhammad
Ali Pasha, the autonomous Ottoman viceroy of Egypt, and his successors led an
ambitious state industrialization program that occurred in two waves in 1816–
1848 (mostly, textiles) and 1848–1868 (mostly, transportation). Although
the program did not produce economic development in the sense of rapid
sustainable growth in GDP per capita, it triggered a social transformation. In
1848, state firms employed 8 percent of the adult active male urban population
and in 1868 3 percent, forming the nucleus of the Egyptian working class.
But did state industrialization act as a “melting pot,” reducing occupational
inequality between Copts and Muslims? Using the 1848 and 1868 population
census samples, I argue that it did not. In 1848 state firms had relatively
fewer medium-skilled Muslim and Christian workers than the traditional
sector, with a stronger effect on Christians in urban Egypt. For Muslims the
effect was compounded by a drop in the likelihood of being a high-skilled
worker, resulting in a “de-skilling” effect on Muslims. For Christians, the
290 M. Saleh
dearth of medium skill jobs in state firms was offset by a rise in the likelihood
of being a high-skilled worker, perhaps to fill in the administrative jobs in
state firms, resulting in an “up-skilling” effect on Christians. In contrast,
state industrialization in 1868 (equally) decreased the likelihood of being a
low-skilled worker among both Muslims and Christians, resulting in an “up-
skilling” effect on both groups. But while the “up-skilling” of Muslims came
from an increase in the likelihood of being a medium-skilled worker and the
likelihood of being a high-skilled worker, the “up-skilling” of Christians was
solely due to an increase in the likelihood of being a high-skilled worker, which
was greater than the corresponding increase among Muslims.
The introduction of modern education, whether public or foreign, after
1800, did not reduce inequality either. Up to 1848, the vast majority of
students, whether Muslims or Copts, were enrolled in elementary religious
schools (kuttabs) or higher religious institutions such as al-Azhar. A tiny
percentage of Muslim students were enrolled in modern public schools that
were introduced by Ali in 1818 following the European school model. But
although Coptic students were not allowed to enter public schools until 1873,
these schools did not push Muslim student enrollment rate up to converge
with that of Copts, because it targeted a tiny Muslim population. More
importantly, Copts were much faster than Muslims in switching from religious
to modern schools. In 1868, 40 percent of Coptic students were already
enrolled in modern schools, compared to only 4 percent among Muslim
students. But unlike Muslim students who enrolled in public schools, Coptic
students in modern schools were all enrolled in private schools, both Coptic
and foreign (French, American, and English). And by 1907/1908, almost all
Coptic students were enrolled in modern schools. By contrast, as late as in
1948/1949, 70 percent of Muslim students were still enrolled in “elementary”
(awwaliya) schools (a modernized version of the kuttabs that was created in
1916).
Saleh (2016) examines the effect on the Coptic-Muslim educational and
occupational inequality of the expansion of public mass modern education
in the 1950s. After a military coup that overthrew the rule of Muhammad
Ali’s dynasty in 1952 and ended the de facto British colonization in 1956,
Egypt embarked on a pioneering state-led development program in the region.
In the educational arena, the post-1952 regime expanded on the public mass
education policies that were first introduced by Taha Hussein, a prominent
liberal intellectual and Egypt’s minister of education from 1950 to 1952.
In 1951–1953, as the final stroke in a century-long process of unifying the
dichotomous traditional and modern educational systems, the government
transformed traditional “elementary” (awwaliya) schools into public modern
“primary” (ibtida’iya) schools (Boktor 1963, pp. 27–28). Prior to the reform
modern primary schools were de facto the sole route to secondary schools,
university education, and white-collar jobs, whereas awwaliya schools qualified
their graduates to only religious (Muslim or Christian) higher education
institutes (Boktor 1936, p. 123; Harby and El-Azzawi 1960). Although the
reform entailed partial improvement of equipment and facilities in awwaliya
schools to match those in primary schools and the construction of new
primary, preparatory, and secondary schools, these measures were limited in
scope. Similarly, the reform barely altered the curriculum in awwaliya schools,
which almost converged to that of primary schools by 1949. Instead, the reform
essentially meant the “re-labeling” of awwaliya schools as public modern
primary schools, loosening the entry requirements to post-primary education,
and increasing class size in primary, preparatory, and secondary schools in order
to absorb the surge in student enrollment. In addition to these policies, as
documented by Ragui Assaad (1997), the government abolished tuition fees
in public universities in 1961 and introduced in 1961–1964 an employment
guarantee in the government and public sectors for graduates of secondary
schools and universities, a policy that lasted until the 1983 graduates. As a result
of these policies, the supply of modern schools (per 1000 population of age 5
to 19) more than doubled from 0.44 in 1951/1952 to 0.96 in 1959/1960, and
primary enrollment rose from 1 million students in 1952 to nearly 3.5 million
in 1965/1966, and preparatory and secondary enrollment expanded even faster,
multiplying sixfold and threefold, respectively, from 1956 to 1961.
Using the 1986, 1996, and 2006 individual-level population census samples
that are available on IPUMS-International and the decennial village-/quarter-
level 1897–1986 population censuses that are available on CEDEJ (2003),
I first documented that the Christian-Muslim differences in educational and
occupational attainment fell during the second half of the twentieth century. I
then exploited the variation across cohorts and districts of birth in the intensity
of the reform, by matching two novel data sources: the individual-level 10-
percent 1986 population census sample and the district-level 1951/1952 school
census. I found that the 1951–1953 reform had a positive impact on Muslims’
educational and occupational attainment but had no statistically significant
impact on Christians. However, the estimated impacts of the reform on the
Christian-Muslim differences with respect to these outcomes are imprecisely
estimated and statistically insignificant.
To conclude, self-selected conversions that were induced by the tax system
are arguably a plausible explanation of the Coptic-Muslim socioeconomic
inequality that emerged in early medieval Egypt. And while group restric-
tions on white-collar and artisanal skills contributed to the perpetuation of
292 M. Saleh
inequality for over a millennium, state policies in industrialization and the

growth of modern education in 1800–1950 altered the Coptic-Muslim SES
gap in favor of Copts. It was not until the arrival of public mass education in
the 1950s that the gap declined. It may be tempting to generalize this argument
to other parts of the whole Arab world with sizable non-Muslim minorities
(Iraq and the Levant in particular), since the poll tax on non-Muslims was
imposed throughout the entire Arab caliphate and, in fact, in all Muslim-
ruled territories. However, I refrain from generalizing the hypothesis beyond
Egypt at this stage, since we still know very little about the formation processes
of other non-Muslim minorities in the region, such as Greeks, Armenians,
Karaite Jews, Rabbinic Jews, and Levantine Christians. There are unexplored
data sources that can help to explore this research question including papyri
(98 percent of which are still unpublished), the sixteenth-century Ottoman
tax registers, and the Ottoman population censuses in 1891–1914. These
sources may reshape our understanding of non-Muslim minorities in the
region. Understanding the origins of Copts’ privilege over Muslims in Egypt
is valuable though because Copts are the largest non-Muslim minority (in
absolute number) in the region, because unlike other non-Muslims who were
mostly urban traders, Copts’ spatial and occupational distributions exhibited
enough variation to test the selection hypothesis, and because using the
medieval papyri advances our knowledge of this phenomenon instead of
relying on often subjective narratives.
References
Abecassis, Frédéric. 2000. L’enseignement étranger en Egypte et les élites locales 1920–
1960: Francophonie et identités nationales. PhD diss., Université d’Aix-Marseille I,
Marseille.
Al-Maqrizi. 2002 [1500]. Al-mawa’iz wal i’tibar fi zhikr al-khitat wal athar (Sermons
and Considerations in Examining Plans and Monuments). Ed. A.F. Sayyid, Vol. 4.1.
London: al-Furqan Islamic Heritage Foundation.
Assaad, Ragui. 1997. The Effects of Public Sector Hiring and Compensation Policies
on the Egyptian Labor Market. World Bank Economic Review 11 (1): 85–118.
Barro, Robert J., and Rachel M. McCleary. 2003. Religion and Economic Growth
Across Countries. American Sociological Review 68 (5): 760–781.
Becker, Carl H. 1902. Beiträge zur Geschichte Ägyptens unter dem Islam. Strasbourg:
Verlag Von Karl J. Trübner.
Becker, Sasha O., and Ludger Woessmann. 2009. Was Weber Wrong? A Human
124 (2): 531–596.
Bell, Harold I. 1910. Greek Papyri in the British Museum: Catalogue with Texts. Vol. IV.
London: The Aphrodito Papyri.
Boktor, Amir. 1936. School and Society in the Valley of the Nile. Cairo: Elias’ Modern
Press.
———. 1963. The Development and Expansion of Education in the United Arab
Republic. Cairo: The American University in Cairo Press.
Borooah, Vani, and Sriya Iyer. 2005. Vidya, Veda, and Varna: The Influence of
Religion and Caste on Education in Rural India. Journal of Development Studies
41: 1369–1404.
Botticini, Maristella, and Zvi Eckstein. 2005. Jewish Occupational Selection: Educa-
tion, Restrictions, or Minorities? Journal of Economic History 65 (4): 922–948.
Centre d’Etudes et de Documentation Economiques, Juridiques, et Sociales (CEDEJ).
2003. Century Census CD-ROM: Egypt 1882–1996. Cairo: CEDEJ.
Chaudhary, Latika, and Jared Rubin. 2011. Reading, Writing, and Religion: Institu-
tions and Human Capital Formation. Journal of Comparative Economics 39 (1):
17–33.
Courbage, Youssef, and Philippe Fargues. 1997. Christians and Jews Under Islam.
Trans. J. Mabro. London-New York: I. B. Tauris Publishers.
Dennett, Daniel C. 1950. Conversion and the Poll Tax in Early Islam. Cambridge, MA:
Harvard University Press.
El-Leithy, Tamer. 2005. Coptic Culture and Conversion in Medieval Cairo 1293–
1524 AD. PhD diss., Princeton University, Princeton.
El-Shayyal, Gamal. 1966. Tarikh misr al-islamiya (History of Islamic Egypt). Cairo: Dar
Al-Ma’arif.
Frantz-Murphy, Gladys. 2004. Conversion in Early Islamic Egypt: The Economic
Factor. In Muslims and Others in Early Islamic Society, ed. Lawrence Conrad, 323–
330. Burlington, VA: Ashgate.
Grohmann, Adolphus. 1932. Aperçu de Papyrologie Arabe. Etudes de papyrologie 1:
23–95.
Harby, Mohammed Khayri, and El-Sayyed Mohammed El-Azzawi. 1960. Education
in Egypt (U.A.R) in the 20th Century. Cairo: Ministry of Education: Education
Documentation Centre of U.A.R.
Heyworth-Dunne, James. 1938. An Introduction to the History of Education in Modern
Egypt. London: Luzac and Co.
Issawi, Charles. 1981. The Arab World’s Legacy. Princeton: The Darwin Press.
Kuran, Timur. 2004. The Economic Ascent of the Middle East’s Religious Minorities:
The Role of Islamic Legal Pluralism. Journal of Legal Studies 33: 475–515.
Mikhail, Maged S.A. 2004. Egypt from Late Antiquity to Early Islam: Copts, Melkites,
and Muslims Shaping a New Society. PhD diss., University of California Los
Angeles, Los Angeles.
Morimoto, Kosei. 1981. The Fiscal Administration of Egypt in the Early Islamic Period.
Kyoto: Dohosha.
294 M. Saleh
Rapoport, Yossef. 2004. Invisible Peasants, Marauding Nomads: Taxation, Tribalism

and Revolt in Mamluk Egypt. Mamluk Studies Review 8 (2): 1–22.
Raymond, André. 1973. Artisans et Commerçants au Caire au XVIIIe Siècle. Damascus:
Institut Français de Damas.
Salama, Girgis. 1963. Tarikh al-Ta’lim al-ajnabi fi misr fil qarnayn al-tasi’ ‘ashar wal
‘ishreen (History of Foreign Education in Egypt in the Nineteenth and Twentieth
Centuries). Cairo: Supreme Council for Art, Literature, and Social Sciences.
Saleh, Mohamed. 2013. A Pre-Colonial Population Brought to Light: Digitization
of the Nineteenth Century Egyptian Censuses. Historical Methods: A Journal of
Quantitative and Interdisciplinary History 46 (1): 5–18.
———. 2015. The Reluctant Transformation: State Industrialization, Religion, and
Human Capital in Nineteenth-Century Egypt. Journal of Economic History 75 (1):
65–94.
———. 2016. Public Mass Modern Education, Religion, and Human Capital in
Twentieth-Century Egypt. Journal of Economic History 76 (3): 697–735.
———. 2018. On the Road to Heaven: Taxation, Conversions, and the Coptic-
Muslim Socioeconomic Gap in Medieval Egypt. Journal of Economic History 78
(2): 394–434.
Samir, Samir Khalil. 1996. The Role of Christians in the Fatimid Government Services
of Egypt to the Reign of Al-Hafiz. Medieval Encounters 2 (3): 177–199.
Sijpesteijn, Petra M. 2013. Shaping a Muslim State: The World of a Mid-Eighth-Century
Egyptian Official. Oxford: Oxford University Press.
Tagher, Jacques. 1998 [1951]. Christians in Muslim Egypt: A Historical Study of the
Relations Between Copts and Muslims from 640 to 1922. Altenberge: Oros Verlage.
Weber, Max. 1930 [1905]. The Protestant Ethic and the Spirit of Capitalism. Trans.
T. Parsons. New York-London: Charles Scribner’s Sons, George Allen and Unwin
Ltd.
———. 1996 [1958]. The Religion of India: The Sociology of Hinduism and Buddhism.
Trans. H. Gerth and D. Martindale. New Delhi: Munshiram Manoharlal Publish-
ers.
Wellhausen, Julius. 1927 [1902]. The Arab Kingdom and Its Fall. Trans. M.G. Weir.
Calcutta: University of Calcutta.
18
Religion and the European Union
Benito Arruñada and Matthias Krapf
[Helmut] Kohl was a devout catholic; and I have heard him say that the Europe
he dreamt of was one as united as the catholic Europe of pre-reformation times; a
space where students, cathedral-builders, abbots, minstrels and troubadours could
roam the continent looking for work, for wisdom, or for adventure. Klau (2018)
JEL Classification Z12, F15
We are grateful to Adrian Chadi, for helping putting together data on the German federal election 2017,
and to Johannes Buggle, Mircea Epure, as well as the volume editors for comments. Arruñada wishes to
acknowledge support from the Spanish Ministry of the Economy, Industry and Competitiveness through
grant ECO2017-85763-R.
B. Arruñada ()
Universitat Pompeu Fabra and Barcelona GSE, Barcelona, Spain
e-mail: benito.arrunada@upf.edu
M. Krapf
University of Basel, Basel, Switzerland
e-mail: matthias.krapf@unibas.ch

296 B. Arruñada and M. Krapf
Introduction
The European Union (EU) seems to be preferred by Catholics. Among the six
original members of the EU’s predecessor “European Economic Community,”
formed in 1957, four were predominantly Catholic and the remaining two had
substantial Catholic minorities. European countries with greater Protestant
influence seem more reluctant to pass on responsibilities and jurisdiction to
European institutions. Protestant Scandinavian EU member countries Sweden
and Denmark decided not to initially join the Euro. Norwegians voted not to
join the EU in a referendum in 1994, and the United Kingdom is likely to
leave the EU after a referendum in 2016.1
We start by reviewing the recent literature on cultural differences across
Euro member states, in particular Guiso et al. (2016a), Alesina et al. (2017), and
Brunnermeier et al. (2016), and then suggest a complementary interpretation
of their findings. While all of these are excellent pieces of research, they
largely ignore religious denomination and, thus, fail to explain the origins of
the explored cultural differences. Our argument builds on previous work by
Arruñada (2009, 2010), analyzing the different social ethics of Catholic and
Protestant values, and Chadi and Krapf (2017), showing how Protestant values
have differentially affected political attitudes of German citizens with respect
to the Euro crisis.
We show that economic outcomes and measures of institutional quality
systematically differ across Catholic and Protestant countries in the Euro
zone and explore how greater consideration of these differences should help
us understand and organize European integration. We argue that religious
denomination is a major determinant of the Euro crisis and that a deeper
understanding of this relationship is crucial for its resolution. In the long run,
and even at current levels of intercountry diversity, the Euro zone can probably
become as successful economically as national states that have both Catholic
and Protestant populations (e.g., Germany, the Netherlands, Switzerland).
Currently, with little coordination of fiscal policies, it appears that Catholic
countries have a harder time adjusting to the rules required for joint monetary
policy. Europe does not have the luxury of exploring different options to
address this problem. Decisions must be based on a limited set of information,
and future historical judgments will be based on the single decision path
finally chosen. However, the hybrid policies adopted in response to the
economic crisis may be well-adjusted in two dimensions. First, they are close
1 See also Nelsen and Guth (2015).

to the political equilibrium and may therefore be more sustainable than more
ambitious technocratic integration plans, which may easily trigger nationalistic
backlash. Second, they are also close to median preferences at the citizens’ level,
especially when considering the mixed cultural composition of some countries
such as Germany.
Our interpretation of cultural differences between Catholics and Protestants
also differs from the literature that emphasizes elite influence. Rubin (2017),
for example, argues that Catholic and Protestant countries developed differ-
ently because of self-serving actions of religious leaders. Nowadays, however,
religious leaders are not as influential as they used to be. Yet, Catholics and
Protestants are still different and in some dimensions keep diverging. This
is, however, not because of the actions of contemporary religious leaders but
because of cultural differences that were shaped by the actions of religious
leaders centuries ago.
This chapter is organized as follows. Section “Literature Review: European
Integration” critically assesses the recent literature on European integration.
Section “Christian Denominations and the Economy of the Euro Zone”
presents and interprets evidence that the European sovereign debt crisis can be
seen as a cultural conflict between Protestants and Catholics. Section “Religion
and National Identity” argues that a clearer understanding of the link between
religious denomination and national identities is key for the road ahead for
the EU. Lastly, section “Discussion and Outlook” discusses how Europe can
benefit from its cultural diversity.
Literature Review: European Integration

A stream of recent works discusses the role of cultural values in European
integration, pointing out tensions and ambiguities, such as lack of convergence
in values and the strength of national identities. These analyses would benefit
from greater consideration of religious factors that are likely a major root
of differences in values but also a mitigating force of centrifugal tendencies,
including that of nationalism.
Guiso et al. (2016a) study different cultural norms in Euro member coun-
tries and develop model in which the desirability of centralization increases in
cultural diversity. They present several descriptive statistics that are very similar
to statistics in Arruñada (2010). The difference is that where, similarly to Guiso
et al. (2003), Arruñada compares Catholics and Protestants within countries,
Guiso et al. (2016a) compare countries. For example, whereas Arruñada reports
that Catholics are more likely to cover up for friends, Guiso et al. report that,
for instance, Greeks are less likely than Germans to call the police, identify
criminals, or testify against them. If these really are distinctive differences
between Northern and Southern Europe, they may, again, be entirely due to
differences between Catholics and Orthodox on the one hand, and Protestants
on the other hand. Indeed, in a footnote, Guiso et al. (2016a) allude to the
possibility that religious denomination and confessional cultures may play a
role, but largely ignore the implications.
Asserting that in many areas there has been divergence rather than conver-
gence, Guiso et al. (2016b) argue that Europe is trapped; it has gone so far that
it cannot go back; going further is difficult because of national interests; and
the current setting with common monetary policy and national fiscal policies
is unsustainable. So, they claim that, inevitably, there will be a catastrophe.
It is true that Europe put itself in a hard position. But, like for a swimmer
in the middle of a strong current, it could be suicidal to swim head on,
accelerating political or economic union. Better to “navigate” the current,
moving along it, saving energy, and focusing all efforts into deviating bit
by bit toward the chosen direction. Building effectively on Jean Monnet’s
idea of relying on centralization in certain areas, hoping that there would be
convergence in other areas, requires choosing carefully what to centralize and
when.
This choice must consider institutional and cultural constraints. In our
view, Guiso et al. (2016a) underestimate the role of moral hazard that takes
place at the level of governments. In Table 18.2 of their paper, they report
survey evidence showing much higher support among Germans for Spain to
remain part of the Euro zone in July 2011 than for Greece, where moral hazard
was a more important factor. Arguably, such moral hazard at the governmental
level is caused by poor resolution of the public good problem between citizens
and governments.
This interaction between governments and citizens is heavily affected by
citizens’ values. Based on arguments about social ethics and compliance with
uniform rules laid out in Arruñada (2010) and Chadi and Krapf (2017), we
consider culture, or more precisely religious denomination, a main driver
of moral hazard. For example, both Arruñada and Chadi and Krapf show
that German Catholics are more likely than German Protestants to consider
tax fraud as morally justifiable. In the same direction, they observe greater
willingness of Catholics to cover up rule-cheating friends. Overall, Catholics
are more protective of their personal relations, including their families, but
less willing to contribute to solving public good problems at a community or
institutional level. Focusing on country differences hides these differences and
precludes attempts to develop institutionally well-suited solutions.
Alesina et al. (2017) argue that cultural divergence within the EU between
1980 and 2009 endangers its future. However, their measure of cultural
diversity suffers potential weaknesses. First, divergence disappears when a
different metric (cosine distance) is considered (Alesina et al. 2017, footnote
22). Second, some of the variables that they consider for identifying cultural
traits (e.g., opinions on abortion) seem relatively unimportant with respect to
decisions on European integration, whereas they omit others arguably more
relevant (e.g., views on immigration and EU-wide distribution).
In any case, more deeply, cultural divergence between countries in (for
cultural matters) such short horizons might well be a minor factor compared
with deeper and more permanent cultural differences. Alesina et al.’s finding
that within-country differences are larger than cross-country differences points
in this direction, too. They also sideline the origins of such deeper cultural
differences—including religion, which holds potential to explain cultural
diversity.
Alesina et al. (2017) do find that between 1980 and 2009 (different waves
of EVS), Europeans have become more different in their attitudes toward
religiosity. Both, Northern Europeans and Southern Europeans, have become
more secular. But in the South this trend was much slower and the gap
increased. In principle, this divergence may be explained by different trends
among Catholics and Protestants. It is well documented that Protestants have
been more likely to quit the church than Catholics. It is possible that the gap
could have remained constant or even decreased if Alesina et al. had controlled
for religious denomination. They show that there was cultural divergence
within countries, too. For example, in Germany and the Netherlands, which
have both had Protestant and Catholic populations for centuries, this within-
country divergence may also have been driven by different trends among
different religious denominations.
Other aspects of Alesina et al. are in line with a denomination-based
interpretation, too. For example, their analysis of cultural clusters suggests that
France is an outlier in Europe and most different from the other countries.
This may be related to France’s distinctive religious history. France is still
on paper predominantly Catholic but was secularized early and profoundly.
On the other hand, Germany, which may be a good mirror for Europe as a
whole because it has both Protestant and Catholic populations, is closest to
the centroid. Differences between France and Germany are also emphasized
in Brunnermeier et al. (2016), a recent book that investigates persistent cultural
traits behind the Euro crisis.
Brunnermeier et al. (2016) divide the EU into “French” and “German”
countries, and emphasize how the predominance of different economic
ideas in French- and German-type countries may hinder further integration.

Whereas the German approach to the Euro crisis is characterized by rules,
accountability, and structural reforms, the French approach favored discretion,
solidarity, and Keynesian stimulus. Again, the dichotomy between Germany
and France applies only superficially. Chadi and Krapf (2017) show that there
is substantial heterogeneity within Germany, which is linked to religious
denomination. Alesina et al. (2017) find that within-country heterogeneity
exceeds cross-country heterogeneity and suggest that France itself may be an
outlier among the Southern European countries that Brunnermeier et al. label
as “French.”
Brunnermeier et al. note that “nothing is written in stone” meaning that
cultural attitudes in a country can also change. For example, they argue that
Germany was Keynesian and interventionist until WWII and that German
policy making became more rule-based afterward in order to prevent a reemer-
gence of populism and fascism. While this reasoning is certainly intriguing, it
is based on anecdotes and leaves many open questions. Austerity was already
popular in Weimar Germany (Galofré-Vilà et al. 2017). And others, such
as Fratzscher (2014) and Krugman (1999), trace German’s affection for rules
back to Kantian philosophy, which stresses legal institutions and, thus, follows
Protestant traditions.
Christian Denominations and the Economy

of the Euro Zone
Macroeconomic performance and measures of institutional quality across
EU countries are significantly correlated with shares of different religious
denominations, which also correlate with different attitudes toward the Euro
crisis.2 Table 18.1 and Fig. 18.1 display correlations between denominational
shares and macroeconomic performance.
In the vein of Arruñada (2010), we characterize Catholicism as intermediate
between Orthodox Christianity and Protestantism with respect to their views
on penance and confession, which, we claim, are fundamental for institutional
2 We include all countries that were members of the Euro zone and EU member states in 2010 if the relevant
data are available. We retrieved country-specific population shares of religious denominations from the
World Religion Database. These shares are defined such that they assume values between 0 and 1. The
16 countries included in our analysis are Austria, Belgium, Cyprus, Finland, France, Germany, Greece,
Ireland, Italy, Luxembourg, Malta, the Netherlands, Portugal, Slovakia, Slovenia, and Spain. Monaco, San
Marino, and the Vatican also used the Euro in 2010, but were not EU member states. Where available,
we used shares based on wave 4 of the European Values Survey, which took place in 2008/2009. The
Table 18.1 Religious denomination and macro indicators

(1) (2) (3) (4) (5) (6)
Bond yields 12/2011 Public debt 2015
Share Catholic −1.649 8.072** −18.638 55.007*
(8.389) (2.669) (39.451) (26.427)
Share Catholic and 12.751** 85.226**
Orthodox
(5.231) (31.464)
Population and Yes Yes Yes Yes Yes Yes
constant
Greece and Cyprus Yes No Yes Yes No Yes
included
Observations 14 13 14 16 14 16
R2 0.0606 0.4707 0.4619 0.0478 0.3048 0.3491
Notes. Data were retrieved from the World Religion Database, the OECD, and the
World Bank. Long-term bond yields are not available for Cyprus and Malta. Robust
standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1
and economic performance.3 We, therefore, provide results for three different
regression specifications. First, we include all Euro countries and look at how
share of Catholics affects the outcome of interest. Then, we exclude Greece and
Cyprus (if data for Cyprus are available), for which the distinction between
Protestant and Catholic does not apply, from our analysis. Finally, we display
results using the combined shares of Catholics and Orthodox Christians, again
for the whole sample. We do not show results for share of Protestants, which
are never statistically significant.
As expected, we find positive correlations between share of Catholics among
the populations of Euro zone member states and bond yields as well as public
debt if we only consider countries that have traditionally been either Protestant
or Catholic (see regression specifications (2) and (5) in Table 18.1). According
only exception is Italy, for which we used shares based on the Pew Global Attitudes Project 2007. Long-
term bond yields in percent were retrieved from the OECD. We use data for December 2011, the last
month of the year examined in Chadi and Krapf (2017), but the results are robust to choice of the month
for which we use data. Data on public debt were also retrieved from the OECD, but are missing for
Cyprus, Estonia, Lithuania, and Malta. For these four countries, we filled in numbers published by the
World Bank. Our measures of institutional quality, “rule of law” and “control over corruption,” were
retrieved from the World Bank Governance Indicators. We use debt and institutional quality data for
2015, the latest year for which they were available. Debt is measured relative to gross domestic product
(GDP) in percent. All indicators of macroeconomic performance and institutional quality are available
online from https://data.oecd.org/interest/long-term-interest-rates.htm, http://data.worldbank.org/data-
catalog/world-development-indicators, and http://info.worldbank.org/governance/wgi. Bond yields were
not available for Cyprus and Malta.
3 For details on Christian Orthodox confessional culture, see Ware (1993), pp. 288–90.
Greece
20
Long−term interest rates Dec 2011
15
Portugal
10
Ireland
Slovenia Italy
Spain Slovakia
5
Belgium
France Austria
Finland Netherlands Luxembourg
Germany
0
0 .2 .4 .6 .8 1
Combined share Roman Catholic and Orthodox
Fig. 18.1 Religious denomination versus December 2011 bond yields
to our results, an increase in the share of Catholics from 0 to 1 is associated with

an increase in long-term interest rates by 8 percentage points and an increase
in total public debt by 55 percent of GDP.
The link between religious denomination and the two outcomes under
consideration becomes even stronger if we combine the shares of Catholics
and Orthodox Christians. Corresponding regression results are displayed in
specifications (3) and (6) of Table 18.1. We also visualize the relationship with
bond yields in Fig. 18.1. An increase in the combined share of Catholics and
Orthodox Christians from 0 to 1 is then associated with an increase in long-
term interest rates by 12.75 percentage points and an increase in total public
debt by 85 percent of GDP.4
That Catholics view corruption as less of a bad thing has been shown before
using micro data, for example, in Arruñada (2010). Colvin and McCracken
(2017) further corroborate Arruñada’s findings and provide additional evi-
dence that Protestants have higher confidence in the rule of law. Table 18.2 and
Figure 18.2 show that these differential attitudes translate into cross-country
differences. According to the World Bank’s World Governance Indicators,
4 The observed link between religion and economic outcomes may be related to Bordo and Istrefi’s (2018)
finding that, controlling for other characteristics, Protestant members of the Federal Reserve’s Open
Market Committee tend to be more hawkish, Jews more dovish, and Catholics centrist.
Table 18.2 Religious denomination and institutional quality

(1) (2) (3) (4) (5) (6)
Rule of law 2015 Control over corruption 2015
Share Catholic −4.804 −25.202*** −6.985 −31.897***
(11.947) (6.445) (14.895) (7.918)
Share Catholic −28.891*** −35.919***
and Orthodox
(6.613) (8.708)
Population and Yes Yes Yes Yes Yes Yes
constant
Greece and Yes No Yes Yes No Yes
Cyprus
included
Observations 16 14 16 16 14 16
R2 0.0317 0.4464 0.4805 0.0322 0.4395 0.4611
Notes. Data were retrieved from the World Religion Database, the OECD, and the World
Bank. Robust standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1
100
Finland
Luxembourg
Netherlands
Germany
Ireland
WGI Control of Corruption 2015
Belgium
90
Austria
France
80
Portugal Cyprus
Malta
Slovenia
70
Spain
Slovakia
60
Italy
Greece
50
0 .2 .4 .6 .8 1
Combined share Roman Catholic and Orthodox
Fig. 18.2 Religious denomination versus control over corruption in 2016
Protestant European countries score better than Catholic countries according

to their rule of law measures.
Chadi and Krapf (2017) show that cultural differences between Protestants
and Catholics also translate into different attitudes toward the Euro crisis.
Members of the German parliament, in particular the ones affiliated with the
ruling Christian-conservative party, were more likely to vote against a Greek
bailout if they came from more Protestant constituencies. Moreover, concerns

about the Euro crisis adversely affected the personal wellbeing of Protestants,
while the wellbeing of Catholics and others remained virtually unaffected.
All these outcomes are consistent with arguments based on the different
theologies of penance and confession of sins which were the basis for the major
split in Christianity. For centuries, auricular confession of sins, discontinued in
Protestant areas soon after the Reformation, was maintained in Catholic and
Orthodox communities. Thus, Chadi and Krapf (2017) argue that Catholics
are more likely to forgive sins (and debt) because of the lenience associated with
the ages-old practice of confession of sins, a possibility that is also mentioned
in Guiso et al. (2016a). Arruñada (2009, 2010) makes a more subtle argument:
Whereas Protestantism is based on uniform moral standards that apply to all
individuals and at all times, confession predisposed Catholic (and Orthodox)
believers to adapt moral standards to different circumstances and personalities.
Different confessional cultures have thus emerged, each having different
advantages and disadvantages. Crucially, different cultures can only survive
if they each offer economic benefits. Catholics tend to accept heterogeneous
moral standards that differ across individuals assuming that individuals have
different moral capacities. Protestants, on the other hand, apply uniform
moral standards that can be easy to follow for some individuals but hard for
others. From an organizational point of view, none of the two cultures strictly
dominates the other.
But the two cultures may lead to very different institutions shaped by
the societies’ different moral codes. While Protestant society is built around
uniform rules, Catholic society is more socially cohesive but also more
conducive to moral hazard. Given that, in Catholicism, standards can be
adapted to circumstances, it may not come as a surprise that Catholic countries
tend to be more likely to abandon the Maastricht criteria when they seem
inconvenient even though they agree that they may be good in principle.
Protestant countries, however, based on their more uniform moral standards,
are less likely to make exceptions.
Our theory of confessional cultures explains why a more integrated Europe
finds more support among Catholics, whereas Protestants tend to prefer a
multi-speed union. Just as Protestantism’s uniform moral standards imply that
an individual can only be part of a social community if they are able to comply
with its moral standards, a country can, following this logic, only be part of
a political community if it is able to comply with its economic and political

rules.5
Religion and National Identity

Confessional culture may help explain not only the recent past but also shed
some light on the EU’s future, in particular about the emerging threat of
different forms of nationalism. A unifying theme of populist movements in
Europe is their anti-EU stance. Their recent rise is linked to a combination of
concerns about the economy and about national sovereignty. It may, however,
be no coincidence that economic issues are relatively more serious in the
Catholic south, whereas nationalism and sovereignty concerns matter more in
the Protestant north. As national identity remains the main obstacle to further
European integration (Alesina et al. 2017), weaker Catholic preferences for
national identity and sovereignty may play a viable countervailing role.
The Catholic Church’s long history of fighting national political powers
supports the view that Catholicism is less conducive to the consolidation of
both national identities and institutions. Moreover, the church has tradition-
ally performed many functions as a supra-national entity that could serve
as substitutes for national institutions and were in some way similar to the
ones performed nowadays by the EU. For example, medieval universities,
mostly linked to the church, with Latin as a common language enhanced
active movement of human capital. The church also served as a court of
last resort for many litigations and coordinated ‘common-defense’ crusades.
5 The issue of integrated versus multi-speed Europe goes back at least to 1994, when Protestant Wolfgang
Schäuble, at the time leader of the CDU/CSU parliamentary group in the German Bundestag, proposed
that there should be a “Two-Speed Europe.” This concept would have divided the EU into a more
advanced and homogeneous “Core Europe” that progresses more quickly and into periphery countries
that might join the core later based on certain criteria. The idea was instantaneously rejected by Catholic
German Chancellor and CDU party leader Helmut Kohl. But the creation of the Euro was a first step
toward such a multi-speed Europe with Protestant countries such as the United Kingdom, Sweden, and
Denmark not immediately joining the common currency. A concept for EU reform outlined by the
President of the European Commission Jean-Claude Juncker, a Catholic, in a speech to the EU parliament
on September 13, 2017, reignited this discussion. Among Juncker’s proposals was the extension of the
Euro area to all members of the EU. The opt-outs of the past may indeed have been a brake on European
integration, and bringing all member states closer together may help implement the reforms that are
necessary to solve the Euro crisis. Juncker’s speech received much applause but also immediately faced
opposition, notably from Protestant Dutch and Danish Prime Ministers Mark Rutte and Lars Løkke
Rasmussen.
Table 18.3 Religion and AfD vote share in the German federal election 2017
AfD vote share in percent
Germany Except Bavaria Bavaria
Share Protestant 2.346** 5.091*** −5.283***
(1.110) (1.097) (1.766)
Other or no religion −0.527 2.362 −13.110***
(2.011) (1.883) (3.758)
Federal state dummies Yes Yes No
GDP pc and unemp and const Yes Yes Yes
Observations 299 253 46

R2 0.8155 0.8459 0.4040
Notes. Dependent variable is AfD vote share in percent. Shares of denominations
between 0 and 1 are from the census 2011; baseline is Catholics. Unit of observation is
Germany’s 299 electoral districts. ***p < 0.01, **p < 0.05, *p < 0.1
The Reformation may, to some extent, have been a rebellion against Rome’s
influence.6
Recent research by Cantoni et al. (2017) showed that Catholics were less
likely than Protestants to vote for Germany’s contemporary right-wing party
AfD in state elections after it started espousing more nationalist and anti-
immigrant policies in 2015. In Table 18.3, we consider the AfD’s results in
the federal election held in September 2017 using the shares of Catholics and
Protestants in Germany’s constituencies used in Chadi and Krapf (2017). We
find that, overall, Protestants are still more likely to vote for the AfD than
Catholics.7
6 Referring to a vast literature on the causes and consequences of the Reformation (Becker et al. 2016),
Stark (2014), for example, argues that, nearly without exception, autocrats opted in favor of Lutheranism
in places in which the church had the greatest local power. Local princes such as Frederick the Wise of
Saxony may have been inclined to support Martin Luther because, according to his teachings, the church
should not wield any secular power. Similarly, the Reformation may have failed in countries like Spain
because the Spanish kings had already won over Rome and reformed their churches. In the fifteenth
century, the Spanish grandparents of Emperor Charles V (Charles I in Spain), King Ferdinand II and
Queen Isabella, had been able to deprive Rome of its secular powers in Spain. But he was not able to
achieve the same in the Holy Roman Empire.
7 A possible explanation for the Bavarian paradox is that the influx of asylum seekers via the Balkan route
in Lower Bavaria, one of Germany’s most Catholic regions, caused a disproportionate surge in AfD votes
there.
Discussion and Outlook

Catholic societies appear more willing to give up national sovereignty in favor
of European integration, whereas Protestant countries tend to have national
institutions that have higher quality. This suggests an institutions-integration
trade-off that is resolved differently by Catholics and by Protestants. We have
developed an argument, according to which these different approaches are in
line with different confessional cultures.
But Europe can transform its cultural heterogeneity from a weakness into
a strength. Protestant values reduce moral hazard and improvidence and
improve public goods provision. Catholic values, on the other hand, balance
against exceeding austerity and, possibly, against nationalism.
Whether fiscal centralization is economically beneficial for all involved par-
ties is not immediately obvious. People in Protestant countries fear that fiscal
transfers would only flow from the Protestant north to the Catholic south. But
the currency union is not prepared to deal with the challenges of the future
and may not survive without joint unemployment insurance, a banking union,
and commonly issued bonds that serve as a cushion against idiosyncratic
regional shocks. Without further centralization, Protestant countries would
risk forgoing the benefits of integration.
It may, however, be a mistake to rush such reforms too quickly. Many
authors, like Guiso et al. (2016a), suggest that centralization may not only
be optimal from a purely economic but also from a cultural point of view. At
least in the medium term, however, ignoring national identities may trigger
serious conflicts and a risk of disintegration. Recent policies adopted by the
EU may, therefore, have been more appropriate than drastic centralization.
These policies also accounted for moral hazard by punishing Greece more
severely than Spain. They may not have been fully satisfactory to anyone,
but provided a compromise between tough Protestant and lenient Catholic
positions. In fact, this compromise may have been representative of median
values and therefore most widely accepted in countries like Germany and
the Netherlands, which traditionally have a mix of Protestant and Catholic
populations.
References
Alesina, Alberto, Guido Tabellini, and Francesco Trebbi. 2017. Is Europe an Optimal
Political Area? Brookings Papers on Economic Activity (Spring): 169–213.
Arruñada, Benito. 2009. Specialization and Rent-Seeking in Moral Enforcement: The
Case of Confession. Journal for the Scientific Study of Religion 48 (3): 443–461.
———. 2010. Protestants and Catholics: Similar Work Ethic, Different Social Ethic.
Economic Journal 120 (547): 890–918.
the Protestant Reformation. Explorations in Economic History 62 (C): 1–25.
Bordo, Michael D., and Klodiana Istrefi. 2018. Perceived FOMC: The Making of
Hawks, Doves and Swingers. Working Paper 24650, NBER.
Brunnermeier, Markus K., Harold James, and Jean-Pierre Landau. 2016. The Euro
and the Battle of Ideas. Princeton: Princeton University Press.
Cantoni, Davide, Felix Hagemeister, and Mark Westcott. 2017. Persistence and Acti-
vation of Right-Wing Political Ideology. Munich Discussion Paper 2017-14, LMU
Department of Economics.
Chadi, Adrian, and Matthias Krapf. 2017. The Protestant Fiscal Ethic: Religious
Confession and Euro Skepticism in Germany. Economic Inquiry 55 (4): 1813–1832.
Colvin, Christopher L., and Matthew McCracken. 2017. Work Ethic, Social Ethic, No
Ethic: Measuring the Economic Values of Modern Christians. Journal of Applied
Econometrics 32 (5): 1043–1053.
Fratzscher, Marcel. 2014. Die Deutschland-Illusion. Munich: Hanser Verlag.
Galofré-Vilà, Gregori, Christopher M. Meissner, Martin McKee, and David Stuckler.
2017. Austerity and the Rise of the Nazi Party. Working Paper 24106, NBER.
Guiso, Luigi, Helios Herrera, and Massimo Morelli. 2016a. Cultural Differences and
Institutional Integration. Journal of International Economics 99 (S1): 97–113.
Guiso, Luigi, Helios Herrera, Massimo Morelli, Paola Sapienza, and Luigi Zingales.
2003. People’s Opium? Religion and Economic Attitudes. Journal of Monetary
Economics 50 (1): 225–282.
Guiso, Luigi, Paola Sapienza, and Luigi Zingales. 2016b. Monnet’s Error?. Economic
Policy 31 (86): 247–297.
Klau, Thomas. 2018. Helmut Kohl’s Theory of Europe. https://www.linkedin.com/
pulse/helmut-kohls-theory-europe-thomas-klau. Accessed: 2018-06-21.
Krugman, Paul. 1999. Why Germany Kant Kompete. Fortune Magazine, July 1999.
Nelsen, Brent F., and James L. Guth. 2015. Religion and the Struggle for European
Union: Confessional Culture and the Limits of Integration. Washington, DC: George-
town University Press.
Stark, Rodney. 2014. How the West Won: The Neglected Story of the Triumph of
Modernity. New York: Open Road Media.
Ware, Timothy. 1993. The Orthodox Church: New Edition. New York: Penguin.
Part III
Advances in Religion and Political Economy
19
The Political and Economic Consequences
of Religious Legitimacy
Jared Rubin
This essay assesses the political and economic consequences of religious

legitimacy. I employ the framework proposed in Rubin (2017) to shed light on
why religious legitimacy might matter for political and economic outcomes
and how it affects economies differently from other mechanisms that rulers
use to stay in power. I proceed to provide some historical support for the
framework before concluding with some speculative thoughts on the role of
religious legitimacy on economic outcomes in the twenty-first century.
The framework employed in Rubin (2017) is straightforward, and it yields
an intuitive explanation linking religious legitimacy to economic outcomes.
It begins with a consideration of the desires of a political authority and the
available means of satisfying those desires. These will obviously differ by the
context; presidents, sultans, kings, premiers, tribal leaders, and so on often
have very different desires and face different types of constraints. Yet, it is
possible to boil this problem down to one general idea: rulers want to stay in
power. In general, they have two means of doing so: legitimacy and coercion.
These are not mutually exclusive, of course. Coercion is the degree to which
rulers can employ violence (or the threat of violence) to achieve their desired
ends. A ruler does not actually employ most of the violence himself—although
there are plenty of warlords in the historical record who were part of a coercive
regime—but must rely on others in society with some edge in producing
J. Rubin ()
e-mail: jrubin@chapman.edu

312 J. Rubin
violence. Rubin (2017) calls such individuals coercive agents, because they
provide coercion for the ruler. These may include police, warlords, military,
armed gangs, or militias. On the other hand, a political ruler has legitimacy
when his subjects believe he or she has the right to rule.1 They may believe this for
many reasons. One exogenous reason relevant in some societies is the ruler’s
birth status; for instance, the eldest male child of the previous ruler often
has the “right” to succeed his father in many societies, including monarchies.
Of course, the requirements for legitimacy differ by society; being the eldest
son of the previous US president or UK prime minister hardly qualifies one
for office. More importantly, another mechanism exists through which rulers
derive legitimacy: the blessing of those in the society who people follow for one
reason or another. Rubin (2017) denotes such individuals legitimating agents.2
They might be religious authorities, nobility, tribal leaders, or celebrities.
Thinking about rulers’ desires and possible action sets in this manner allows
us to analyze political economy outcomes in the terms of a bargaining game.
Rulers want to stay in power, and there are agents in society that can help them
do so. Meanwhile, legitimating and coercive agents do not support the ruler for
free—they have some set of policy goals which they would like implemented
in return for their support. The outcome of the bargain—the society’s laws
and policies—are therefore determined by (1) the bargaining power of rulers
and their agents and (2) the players’ preferences.
This general framework yields two insights that can be applied to specific
situations. First, the identity of the legitimating and coercive agents matters,
since their identity is indicative of their preferences. Religious authorities,
military elite, local magnates, tribal chiefs, and democratically elected leaders
all have very different preferences. To the extent that they have bargaining
power vis-à-vis their rulers, these preferences will be reflected in the society’s
laws and policies. Second, the choice of legitimating agents among many
possible ones reflects the costs (i.e., how much the ruler gives up in the bargain)
and the benefits (i.e., the legitimacy or coercion the ruler attains in the bargain)
associated with each agent.
Where do these costs and benefits come from? This is where history matters.
In some societies, religious authorities are better at legitimating rule than
others. The degree to which this is the case is a result of historical processes
that have been institutionalized over time. Likewise, military authorities are
1 For related, but alternative, views of legitimacy, see Lipset (1959) and Greif and Tadelis (2010).
2 Legitimating agents have been shown to be important in determining all types of political and economic
outcomes. For some examples, see Gill (1998), Platteau (2011), Coşgel et al. (2012), Chaney (2013), and
Greif and Rubin (2018).
19 The Political and Economic Consequences of Religious Legitimacy 313
“cheaper” to placate in some societies, while other societies have powerful tribal
leaders that are particularly effective at legitimating rule. Hence, in order to
understand where laws and policies come from and why they differ across
societies, one must know the historical processes that make different forms of
coercion and legitimacy more or less expensive for a ruler, and one must also
consider the preferences of the agents that the ruler ends up choosing.
We can therefore use this framework to understand why religious legitimacy
may have an impact on economic outcomes. First, consider the preferences of
religious authorities. They often desire such things as exemption from taxation
(for themselves), upholding of religious laws, and suppression of rival religions
(Gill 1998). Importantly, these desires are largely orthogonal to the type of
policies that benefit economic development, such as protection of property
rights, investment in public goods, moderate taxation, impartial rule of law,
and the like. Some of the policy outcomes religious authorities desire benefit
economic growth, such as poor relief, while other desires are detrimental, such
as restrictions on usury (interest). But most of their preferences simply have
little effect on economic growth in either direction. This suggests that when
religious authorities have an important seat at the political bargaining table,
it is not necessarily the case that economically harmful laws and policies will
be agreed upon, but it is also unlikely that policies promoting growth will be
near the top of the agenda either.
The degree to which religious legitimacy was effective in keeping rulers
in power in different societies at different times provides insight into the
economic constraints facing the various key players in those societies. Given
the above logic, societies in which religious authorities had significant capacity
to legitimate rule (for some historical reason or another) should have laws and
policies that are somewhat neutral with respect to economic development.
This insight raises two questions: what type of legitimation does entail eco-
nomic growth? Why might the manner in which rulers stay in power change
over time?
Rubin (2017) tackles these questions, using European and Middle Eastern
history to shed light on why legitimating arrangements vary over time and
place and what this means for long-run economic growth. In the Middle East,
Islam was particularly useful at legitimating rule due to the circumstances
surrounding its origins in the Arabian Peninsula (and beyond). In short,
Islam formed conterminously with a rapidly expanding empire—the largest
empire the world had ever seen. Islamic doctrine therefore formed in a manner
conducive to legitimating political rule, which was useful for Muhammad
and his followers as they expanded their rule over an ever-increasing territory
(Crone and Hinds 1986, ch. 1, 4, 5; Hallaq 2005). On the other hand,
314 J. Rubin
Christianity was born in the Roman Empire, with its well-functioning legal
and political institutions. The Romans did not need nor desire legitimation
from the Church. Indeed, Christians were mostly persecuted during the first
three Christian centuries. Therefore, early Christian doctrine is largely silent
on the concept of religious legitimation of rule. Where it does have something
to say, it is often in favor of a separation of church and state (e.g., Jesus’s
famous saying in Matthew 22:21 “Render unto Caesar the things which are
Caesar’s . . . ”). In short, from their inceptions Islam was more conducive than
Christianity to legitimating rule.
Yet, Christian rulers still employed religious legitimacy, if for no other reason
than it was extremely cost effective (Gill 1998). For instance, Charlemagne was
famously crowned “Emperor of the Romans” by Pope Leo III on Christmas
Day, 800 CE, in one of the most famous public acts of religious legitimation
in Christian history. Likewise, Holy Roman Emperor Henry IV was willing
to stand in deep snow for three days outside the castle of Canossa in a plea
to Pope Gregory VII to rescind his excommunication. Over time, however,
the use of religion in legitimating rule diverged between Western Europe and
the Middle East. There were two phenomena responsible for this. The first
was the growth of Western European commerce during what is known as the
“Commercial Revolution” from the tenth through thirteenth centuries. With
the growth of commerce and a new bourgeoisie class, there was incentive for
European rulers to bring this class to the political bargaining table (if, for no
other reason, as a promising source of tax revenue). The bourgeoisie’s rise to
prominence, largely in the burgeoning parliaments, came at the expense of
the Church. Such a process never really happened in the Middle East, despite
the fact that it was economically far ahead of Western Europe for centuries
following the spread of Islam. The reason that widespread Middle Eastern
commerce failed to undermine the role of religious authorities in legitimating
rule follows directly from the above framework: if Muslim religious authorities
were particularly effective at legitimating rule, any actions which might
undermine them—including bringing alternative sources of legitimacy to the
bargaining table—could threaten their capacity to legitimate rule. In other
words, the costs of losing religious legitimacy were greater for Middle Eastern
rulers than they were for Western European ones. Hence, on the margin,
Western European rulers benefitted more from bringing commercial elites to
the political bargaining table.3
3 A related argument is put forth in a series of articles by Auriol and Platteau (2017a, b), who consider
the role that centralization of religious authority plays in the relationship between religious and political
The second event that Rubin (2017) argues affected European legitimating
arrangements is the Reformation. Where the Reformation spread (England,
Scotland, Netherlands, Scandinavia, and parts of the Holy Roman Empire
and Switzerland), religious legitimacy was no longer an option for rulers.
This is largely due to the fact that reformers dismantled the existing Catholic
institutions wherever they were successful, replacing them with Reformed
churches, many of which were under the thumb of the state (for instance, state
churches in England and Sweden formed conterminously with the spread of
the Reformation). Such churches were in no position to provide legitimacy, as
their word was not independent of the state and thus added no information
regarding one’s “right to rule” (Greif and Rubin 2018). In place of the Church,
Protestant leaders tended to turn to parliaments to legitimate their rule (and
provide revenue).
The implications of the transition from religious legitimation to parlia-
mentary legitimation were enormous and warrant greater discussion. Most
importantly, parliaments were disproportionately comprised of individuals
who benefitted from commercial expansion: not only merchants, but those
invested in overseas commercial enterprises (Jha 2015). Much like religious
authorities (or any other type of legitimating or coercive agents), these
individuals pursued their own self-interest in the bargain over laws and
policies. The key difference between the economic-parliamentary elite and
other types of agents is that what was in the self-interest of the former was often
(although certainly not always) consistent with broader economic success.
For instance, these individuals generally desired policies protecting property
rights and providing public goods (such as transportation and communication
infrastructure). To be clear, they desire these policies out of their own self-
interest, not necessarily because they view them as beneficial for the public
good.
The historical record bears out this assertion. Soon after the Reforma-
tion took hold in the Dutch Republic, in conjunction with the Dutch
Revolt, numerous policies that benefitted commercial activity were enacted:
for instance, the establishment of the United East India Company and
the West India Company, investment in canals and other inland transport,
and increased poor relief. A similar chain of events transpired in England,
where soon after the Reformation took hold Parliament re-organized English
property rights under the Statute of Uses and the Statute of Wills, the latter
authorities. Their argument suggests that more decentralized religions, like Sunni Islam, will be less willing
to implement reforms and will be associated with less stable regimes.
316 J. Rubin
of which provided unprecedented clarity for the rights of (wealthy) property

holders and upended key parts of the feudal property rights regime.
In the wake of the Reformation, the Protestant regions of Europe took a
decisively different economic path not only from the Middle East but also from
those parts of Europe that remained Catholic. Rubin (2017) argues that this
was in part, though not solely, due to the different manner in which Protestant
rulers were legitimated. They relied much more heavily on parliaments, whose
self-interest was more consistent with the types of policies that portend
economic growth, than Catholic or Muslim rulers, such as those in Spain or the
Ottoman Empire, who continued to rely on religious legitimacy. This insight
therefore provides a twist on Weber’s (1905 [2002]) famous “Protestant Work
Ethic” hypothesis. Yes, Protestant areas did end up doing better economically
than their Catholic counterparts, and significantly better than the rest of
the world. Indeed, it is no coincidence that soon after the Reformation, the
world’s leading economy has been predominantly Protestant—first the Dutch
Republic, then England, then the United States. But this was not due to some
unique Protestant work ethic or, for that matter, anything about Protestant
doctrine. Instead, it was a result of Protestant political economy, whereby
religious authorities at best had a minimal seat at the political bargaining table.
Therefore, if Rubin (2017) is correct, religious legitimacy (or lack thereof ) had
a huge impact on the “great divergence” between Western Europe and the rest
of the world and the “little divergence” between Northwestern Europe and the
rest of Europe.
While Rubin’s (2017) hypothesis has clear implications for the “great
divergence” literature, I would like to focus the rest of this essay on some
of the implications of Rubin’s framework for the twenty-first century Middle
East. This is inherently an exercise in speculation, but the framework proposed
by Rubin provides a mental model that may be useful for such an endeavor.
In short, I believe that the framework portends good news and bad news
for modern Middle Eastern economies. The good news is the economic
divergence was not a result of Islam itself (except for the role it plays in
legitimating political rule). Laws and policies can and do change in all parts
of the world when it is in the incentive of powerful players. This is as true of
Middle Eastern history as it is of Western European history.
The trick is getting those incentives in place. And that’s where the bad
news, at least for the moment, lies. It is unclear how Middle Eastern societies
will transition to a setting where the economic elite have a sufficiently strong
seat at the political bargaining table. Even Turkey, a nation which for nearly
a century flirted with such a system, is regressing rapidly in the face of an
increasingly autocratic Erdoğan regime. Meanwhile, in the oil-rich Middle
Eastern countries, twentieth-century rulers did not have to bargain with the
economic elite for revenue. They therefore stayed in power so long as they had
a monopoly on coercion (e.g., Gaddafi’s Libya or Saddam’s Iraq) or religious
legitimacy (e.g., Saudi Arabia or post-Revolution Iran). These rulers had little
incentive to provide basic economic protections which encourage investment
and entrepreneurship that are so vital to long-run economic prosperity. As
a result, many of them are blowing the one-time chance provided by oil, and
they are likely candidates for political and economic chaos when the oil money
dries out.
The overarching problem is that the appeal of Islamic religious legitimacy
is strong, even to democratically elected rulers. Religious legitimacy permits
rulers to take otherwise unpopular stances or undermine democratic norms,
and it is therefore appealing to an erstwhile autocrat. This legacy of Islamic
religious legitimation has therefore stifled democracy in the Middle East while
serving as a useful crutch for rulers hoping to impose otherwise unpopular
policies (also see Platteau 2011; Auriol and Platteau 2017a).
Turkey is an especially instructive case. Once the great Middle Eastern
democratic hope, it rapidly slid into autocracy under the Erdoğan regime in the
2010s. A formerly staunchly secular republic, religious legitimacy has recently
become an important part of President Erdoğan’s menu of options for securing
his rule. This “Turkish blueprint” is steeped in Middle Eastern (and certainly
Turkish) history; the Ottomans frequently resorted to the religious elite to
burnish their pious credentials when attempting some controversial action. I
call this the Turkish blueprint because Erdoğan’s actions have revealed that
even when some semblance of democracy is established, there is always the
lure of religious legitimacy for a Muslim (and especially Turkish) ruler who
wants to pursue unpopular policies or a rival who wants to unseat the regime
in power.
This is not just an issue in Turkey, however. Across the Middle East,
democracy has long been a pipe dream. In the post-colonial period, democracy
failed because the power vacuum left by the West was largely filled with
military-backed autocrats. Yet, even in the wake of the Arab Spring, democracy
still has not flourished. Why has democracy consistently failed in the Middle
East? Can the framework laid out above lend some insight into this important
question?
When religion is a particularly powerful source of legitimation, as it is
in Islam, it can insulate a ruler who wants to pursue otherwise unpopular
policies. The quid pro quo between Muslim rulers and religious authorities
has long been a powerful one; religious authorities get protection from rivals,
purview over wide swaths of the law (including many important aspects of
318 J. Rubin
commercial law), and social and economic prestige. In return, they permit
rulers wide discretion over laws and policies. One upshot is that appeals to
religious authority permit rulers to undermine democratic norms.
Even more damningly, such appeals can stop a transition to democracy
in its tracks. What prevents a rival to a democratically elected leader from
claiming religious legitimacy in order to dethrone his rivals? What prevents
a democratically elected leader hungry for more power from appealing to
religion to justify a power grab? The recent history of the three largest Middle
Eastern economies—Saudi Arabia, Iran, and especially Turkey—make this
point crystal clear.
The House of Saud famously ascribes to the ultraconservative Wahhabi
brand of Islam. Combined with its vast oil revenues, this gives the Saudi
crown immense leverage to take any action to its liking, so long as such
actions can find justification from leading clerics. The key point here is not
whether Wahhabism itself is good or bad for the Saudi people (although it is
obviously not so great for women), but that it insulates the crown from popular
discontent. Democracy is unlikely to take hold in such a setting. Even if the
seeds of democracy begin to sprout, the lure of religious legitimacy would be
strong at the first sign of instability. Like Turkey, it is difficult to imagine a
successful and lasting transition to democracy occurring any time in the near
future of Saudi Arabia.
The Iranian situation is more vexing. On the one hand, the Iranian populace
is among the most cosmopolitan and Western-looking in the Middle East.
However, since the 1979 revolution, Iran has been a theocracy ostensibly ruled
by a Supreme Leader. The clerical class has given the Iranian leadership cover
to pursue policies that harmed the vast majority of its citizens—the nuclear
buildup of the early 2010s and ensuing international sanctions chief among
them. Iranian clerics clearly have the power and the will to thwart any potential
transition to democracy. It is thus unlikely, absent another revolution, that
Iran will make the democratic transition. There are simply too many powerful
players within the legitimating regime that stand to lose from such a transition.
Turkey, once a secular democracy on the verge of EU membership, is
the most obvious case. In the second decade of the twenty-first century,
President Erdoğan has increased religious rhetoric as his rule has become
more autocratic. Thousands of dissidents have been imprisoned or worse while
Erdoğan tightens his grip on power. He accomplished this with help from the
religious class, who insulate him from what would otherwise be unpopular
policies. This is the Turkish blueprint. Even when democracy has ostensibly
succeeded for a given time, the lure of religious legitimacy will always be there
for a ruler who is losing grip on power or who wants to pursue controversial
policies.
Is there any hope for democracy in the Middle Eastern future? Or is
Islam simply incompatible with democracy? The Saudi, Iranian, and especially
Turkish experiences highlight two conflicting issues. On the one hand, a
completely secular government is susceptible to threats from rivals claiming
religious legitimacy, as happened during the Iranian revolution. On the other
hand, a ruler who is strongly backed by the religious establishment may be so
insulated from popular opinion that he can enact unpopular policies without
fear of discontent. Neither of these bode well for the future of Middle Eastern
democracy.
What does this mean for the future of Middle Eastern economies? In the
short term, many of them will be insulated from turmoil due to their vast
petroleum reserves. But the medium- to long-term forecasts look grimmer,
as the rise of alternative fuels drives down demand for oil and technological
progress increases supply coming from outside the region—both of which
will, of course, drive down oil prices. It is almost certain that oil revenues
will decrease, probably dramatically, for Middle Eastern states in the third and
fourth decades of the twenty-first century. How will political leaders react to
these changes, and what will their reactions mean for their economies?
There are two reasons to be more pessimistic than optimistic about the
long-run economic and political future of the Middle East. First, much of
the economic opportunity offered by the one-time resource boom has been
squandered. Although the UAE recently diversified its economic portfolio,
most of the oil-rich Middle Eastern states are still, as of 2018, too reliant
on oil revenues to survive in a low-oil revenue world. Efforts to diversify
their economies will be key in this process. Second, as oil revenues dry up
and rulers have less capacity to buy support via subsidies and graft, the
odds of them leaning even more heavily on religious legitimacy—using the
Turkish blueprint—are high. This means the exclusion from the political
bargaining table of precisely the type of people who would push for policies
that portend long-run growth: leaders in manufacturing, tourism, finance,
higher education, and so forth. Whether the Middle Eastern democratic
transition occurs or not is one of the keys to the twenty-first-century prosperity
of the region, especially once oil revenues run dry. Yet, it is unclear how
a sustainable political and economic situation—one that does not have the
specter of religious legitimacy hanging over it—could work.
Unfortunately, these insights are just a diagnosis of the problem, not a cure.
But they do shed light on the constraints faced along the long path to more
functional Middle Eastern political and economic systems. It is important to
320 J. Rubin
understand these constraints. If we simply blame “Islam” or any other cause

that is not at the root of economic and political stagnation, any “cure” will be
misinformed.
References
Auriol, Emmanuelle, and Jean-Philippe Platteau. 2017a. Religious Seduction Under
Autocracy: A Theory Inspired by Theory. Journal of Development Economics 127:
395–412.
———. 2017b. The Explosive Combination of Religious Decentralisation and
Autocracy: The Case of Islam. Economics of Transition 25 (2): 313–350.
Chaney, Eric. 2013. Revolt on the Nile: Economic Shocks, Religion, and Political
Power. Econometrica 81 (5): 2033–2053.
Coşgel, Metin M., Thomas J. Miceli, and Jared Rubin. 2012. The Political Economy
of Mass Printing: Legitimacy, Revolt, and Technology Change in the Ottoman
Empire. Journal of Comparative Economics 40 (3): 357–371.
Crone, Patricia, and Martin Hinds. 1986. God’s Caliph: Religious Authority in the First
Centuries of Islam. Cambridge: Cambridge University Press.
Gill, Anthony. 1998. Rendering Unto Caesar: The Catholic Church and the State in
Latin America. Chicago: University of Chicago Press.
Greif, Avner, and Jared Rubin. 2018. Political Legitimacy and the Institutional Founda-
tions of Constitutional Government: The Case of England. Mimeo.
Greif, Avner, and Steven Tadelis. 2010. A Theory of Moral Persistence: Crypto-
Morality and Political Legitimacy. Journal of Comparative Economics 38 (3): 229–
244.
Hallaq, Wael B. 2005. The Origins and Evolution of Islamic Law. Cambridge:
Jha, Saumitra. 2015. Financial Asset Holdings and Political Attitudes: Evidence from
Revolutionary England. Quarterly Journal of Economics 103 (3): 1485–1545.
Lipset, Seymour Martin. 1959. Some Social Requisites of Democracy: Economic
Development and Political Legitimacy. American Political Science Review 53 (1):
69–105.
Platteau, Jean-Philippe. 2011. Political Instrumentalization of Islam and the Risk of
Obscurantist Deadlock. World Development 39 (2): 243–260.
East Did Not. New York: Cambridge University Press.
Weber, Max. 1905 [2002]. The Protestant Ethic and the ‘Spirit’ of Capitalism. New
York: Penguin.
20
Religious Legitimacy and the Joint Evolution
of Culture and Institutions
Alberto Bisin, Avner Seror, and Thierry Verdier
JEL Classification D02, Z10, P16, P48
Avner Seror and Thierry Verdier acknowledge the financial support from the ERC Grant TECTACOM
324004.
A. Bisin ()
NYU and NBER, New York, NY, USA
e-mail: alberto.bisin@nyu.edu
A. Seror
e-mail: seror@chapman.edu
T. Verdier
PSE, Paris, France
ENPC-ParisTech, Champs-sur-Marne, France
PUC-Rio, Rio de Janeiro, RJ, Brazil
CEPR, London, UK
e-mail: thierry.verdier@ens.fr

322 A. Bisin et al.
Introduction
If Phillip of Valois is - as he affirms - the true king of France, let him prove the
fact by exposing himself to hungry lions; for lions never attack a true king; or let
him perform the miraculous healing of the sick, as all other true kings are wont
to do. Edward III; in Marc Bloch (1924, p. 16)
The importance of religious legitimacy in supporting the power of French

and English kings in the Middle Ages is emphasized by Bloch (1924). But the
history of the Muslim world can also be seen as driven crucially by the role
of religious clerics in providing legitimacy to autocratic rulers; see Platteau
(2017) and Rubin (2017). Indeed, the political power that the clerics gained
in the preindustrial Middle East may contribute to explain to a large extent
the comparatively limited socio-economic growth of the Middle East during
the second millennium of the current era; see Rubin (2017) and Kuran (2012)
who refer to this phenomenon as to the long divergence.
This short chapter provides some preliminary insights on the emergence of
religious legitimacy as a general phenomenon in the history of political econ-
omy, relying on the study of the dynamic interaction of political institutions
and religious culture, as in Bisin and Verdier (2017).
The next section introduces the general framework we adopt to study the
joint evolution of culture and institutions. Section “Religious Legitimacy”
illustrates how this framework can be fruitfully applied to the issue of religious
legitimacy. Section “Some Historical Evidence on the Long Divergence”
discusses some historical evidence on the long divergence and section “Con-
clusions” concludes.
The Joint Evolution of Culture and Institutions

Bisin and Verdier (2017) formalize the evolution of institutions and culture
and study their joint dynamics. In line with Acemoglu and Robinson (2000),
Acemoglu (2003), and Acemoglu and Robinson (2006), institutions are
conceptualized as mechanisms through which social choices are delineated and
implemented. More specifically, institutional change represents an effective
commitment mechanism on the part of the political elites to imperfectly
and indirectly internalize the lack of commitment and the externalities which
plague social choice problems. In line with Bisin and Verdier (1998, 2000b,a,
2001), culture is instead conceptualized as preference traits, norms, and
20 Religious Legitimacy and the Joint Evolution of Culture and Institutions 323
attitudes which can be transmitted across generations by means of various

socialization practices or can be acquired through socio-economic interactions
between peers.
To illustrate, consider a society populated by two groups i ∈ {E, S} (E for
elite and S for the rest of society at large), with distinct cultural traits, objectives,
and technologies. At each period t, a societal policy game is played between
private individuals and some hierarchical public authority (the state) which
controls socio-economic policies. Individuals in each group i ∈ {E, S} are
characterized by an objective function V i = U i (a i , p, A) that depends on
the “cultural” type i of the individual, private actions a i by that individual,
a policy vector p implemented by the state during the period, and some
aggregator measure of socio-economic outcomes A = A a, p, q that captures
the interactions between the private agents and the public authorities. A
naturally depends on the aggregate vector a of actions by individuals of the
two groups, the public policy vector p, and the distribution of cultural types
in the population (captured by the frequency distribution q of types in society).
A benevolent state chooses socio-economic policies in accordance with
the distribution of political power between the two groups encoded and
represented by institutions. Specifically, the institutional system may be char-
acterized by weights β E and β S = 1 − β E associated to the two groups E and
S in the decision-making problem of the state relative to the policy vector p. As
illustrated in Fig. 20.1, a specific institutional setup β = β E , β S , therefore,
State Any individual

with weight U i (a i, p, A )
βi > 0 chooses
chooses action/
policy behavior
p ai
Equilibrium
outcomes:
p*, a*
Lack of internalization
of economic and
political externalities
A=A(a,p,q)
Fig. 20.1 Societal policy game

324 A. Bisin et al.
induces a set of policies p∗ and actions a∗ , as the equilibrium of the societal

policy game between individuals and the public authorities. It is important to
note that this societal policy game and its equilibrium outcomes p∗ and a∗ are
clearly affected by the distribution of cultural traits q in the population. In such
a setting, even when a state is benevolent in its policy choice problem, societies
are fundamentally characterized by economic and political externalities that are
not fully accounted for by private and public decisions. Externalities typically
arise because of socio-economic or political imperfections associated with the
existence of asymmetric/incomplete information, matching frictions, limited
rationality and cognitive biases, strategic behaviors associated with market
power, private opportunism, and lack of political commitment. In any of
these situations, the equilibrium outcomes a∗ and p∗ of the societal policy
game do not fully
internalize their implications on aggregate social outcomes
A = A a, p, q , and as a result inefficient policies and social allocations are
implemented.
The prevailing institutions at any point in time might then have an incentive
to change the distribution of political power in the future, to internalize
the externalities responsible for the inefficiencies at equilibrium. This is the
fundamental driver of institutional change in society. It induces as a general
principle that the political group most likely to internalize the externality is the
group receiving more residual decision rights along the institutional dynamics,
that is, the group having a higher political weight in the state policy choice
problem.
As schematically illustrated in Fig. 20.2, for any cultural population profile
at a given time t, qt , this mechanism provides a mapping from the institutional
Institutional Institutional
design design
Current institutions Future institutions
with weights t with weights t+1
t+1 t+2
pt , at pt+1 , at+1
Cultural Cultural
Society Evolution Society Evolution
with cultural with cultural
distribution (qt ) distribution (qt+1)
(qt+1) (qt+2 )
Fig. 20.2 Joint dynamics of culture and institutions

system at t, β t , into the one at t + 1, β t+1 . Similarly, the cultural profile of

society qt evolves over time, driven by cultural diffusion and social selection
processes within and across generations and influenced by the status of the
institutional system, β t . As shown in Bisin and Verdier (2001, 2017), in the
presence of some degree of paternalistic motivations for cultural transmission,
parental socialization is generally stronger for cultural minorities and, ceteris
paribus, for the group which is relatively favored at the equilibrium outcome
of the policy game. This mechanism induces a mapping from qt into qt+1 ,
which represents the cultural dynamics.
Cultural and institutional dynamics can then reinforce or hinder a specific
socio-economic equilibrium pattern, for example, one which is induced by
an exogenous institutional shock to the dynamical system. In this context,
the concept of cultural multiplier is identified in Bisin and Verdier (2017) as
central to the analysis. It is defined as the ratio of the total effect of institutional
change and its direct effect, that is, the counterfactual effect which would have
occurred had the distribution of cultural traits in the population remained
constant after the institutional change. Whether the cultural multiplier is
positive or negative depends, it is shown, on whether culture and institutions
are dynamic complements or substitutes. To see this, consider, for instance,
the case of complementarity, and take an exogenous shock to the system that
makes more salient the existence of an externality or a political commitment
issue to be internalized. Such a shock triggers then an institutional response
and aimed at internalizing the externality and/or committing policy choices.
The institutional response implies a larger weight to the political group
gaining relatively more from policy changes which lead to an (at least partial)
internalization of the externality and/or to a greater commitment. Suppose the
strength of the institutional response is positively related to the frequency of
the cultural traits carried by that group. Complementarity then obtains when
the policy change also affects positively the relative socialization incentives
of the aforementioned group. Indeed in such a case, we should expect a
larger diffusion of the cultural traits specific to the group that effectively
gets more power through the institutional response. Conversely, a larger
diffusion of those traits makes it more likely to have some further institutional
change giving even more political power to the group, as a way to solve the
internalization policy problem. Over time, institutional and cultural dynamics
reinforce each other and therefore act as dynamic complements.
326 A. Bisin et al.
Religious Legitimacy
The logic of the framework we have illustrated in the previous section can
be fruitfully applied to the study of religious legitimacy, to better understand
its implications in terms of institutional and cultural change in a society. We
say an elite is legitimate when the people believe in its right to rule. Such a
belief is ingrained into a set of values and normative statements describing
how society should be organized.While some legitimacy principles can be
derived from rational premises, legitimacy often takes its roots in the existence
of internalized values and worldviews provided by specific organizations or
individuals. Given the nature of the beliefs and values that they promote,
religious institutions and their members (priests, clerics, etc.) are therefore
important agents contributing to the construction of legitimate orders which
elites can leverage for their ruling and policy making.
Relating to the conceptual framework of the previous section, the basic
principles driving the joint evolution between institutions and culture in the
context of religious legitimacy can be delineated as follows. First, legitimacy
helps (secular) elites address the political economy problems associated with
the implementation of their policy choices. This in turn leads to specific
institutional changes and an evolution of the distribution of political power
between the elite and the religious clerics.
Second, the capacity of the religious clerics to contribute to the solution
of the elite’s policy implementation problems relies fundamentally on how
the religious values promoted by the clerics are disseminated in society. Insti-
tutional changes associated with legitimacy therefore depend on the cultural
profile of society in terms of religious beliefs and values.
Third, the diffusion of the religious values is in turn facilitated by institu-
tions that entrust more political power to the clerics. The institutional system
reflecting the structure of power between the elite and the religious clerics
impacts strongly on the dynamics of cultural diffusion of religious values in
the population. Bisin et al. (2018) provide a formal model of these different
elements and study the implications of religious legitimacy for the joint
evolution of culture and institutions.1 Here we present an informal outline
of the main arguments.
Consider a society composed of political elite, clerics, and civil society
(merchants, workers, popular masses, etc.). The political power of the religious
clerics reflects their relative control of policies and reforms. It relates, for
1 See also Coşgel and Miceli (2009) for a related model.

instance, to their control of the judicial administration, the local police, and
taxation. It is also linked to their presence in key institutions providing social
services to the population. Clerics care about the provision of a religious good,
for example, places of worship or of religious study, which they control and
they extract (material or immaterial) rents from.
The interaction between rulers, clerics, and civil society involves a number
of externalities and commitment issues not internalized by individual deci-
sions. For instance, the religious good constitutes a public good for individuals
in society, in that it facilitates individual participation to religious activities. In
turn, participating to religious activities results in psychological and emotional
benefits, as a consequence of a (perceived) closer relationship with the divine.
Favoring a more intense participation in religious practices, the religious good
provided for by the clerics legitimizes the political control of the elites, for
example, reducing at the margin the psychological cost associated with taxation
and other forms of extraction. Less directly, participating to religious activities
increases the scope of social interactions between religious individuals. This
can have positive effects on productivity, since it potentially improves pro-
social attitudes, for example, coordination and cooperation; also, it promotes
informal information networks in principle very useful in trading and other
economic activities. At the same time, religious beliefs and restrictions may
also prevent or increase the costs of economic activities and transactions.
Religious regulations on occupational choices in labor markets, restrictions on
credit markets at a positive interest rate, or prohibitions of adoption of new
technologies are typical examples distorting the allocation of resources away
from efficiency.
In this environment, the public provision of the religious good controlled by
the clerics will generally not internalize these social externalities and, depend-
ing on the state of society and its organizational features, be inefficiently low or
high. The elites controlling political power cannot internalize the public good
aspect of the religious good; they cannot commit ex ante to its provision. Non-
clerical elites, in particular, would gain from committing to higher religious
good provision inasmuch as the legitimacy it provides allows them to extract
more (and more efficiently) resources from society. Consequently, when the
legitimacy effect of the religious clerics is sufficiently strong, institutional
change would favor an increase in the relative institutional weight of the clerics
in society. Religious legitimacy induces therefore a shift in the structure of
power towards religious clerics. In turn, clerics exercise this power by providing
the religious good in larger quantities, which in turn favors religious practices
and activities, propagating beliefs within the population that justify the ruling
and extractive capacity of the political elite.
328 A. Bisin et al.
The political empowerment of religious clerics has important consequences

in terms of institutional trajectory and cultural change. In particular, since
the clerics care about the provision of religious services, their prevalence in
society should increase religious parents’ direct inculturation efforts, that is,
vertical socialization. Religious legitimacy then makes culture and institutions
complement: institutional change devolving political power to the clerics rein-
forces the incentives of the religious member of society to transmit their values,
while in turn a higher fraction of religious individuals augments the political
incentive to credibly change the institutional structure so as to empower the
clerics. When the legitimacy effect of the religious clerics is sufficiently strong,
the complementarity between the empowerment of the clerics and the cultural
transmission of religious values creates dynamic feedback effects between the
institutional and cultural dynamics, leading plausibly to multiple stationary
states in the joint dynamics.
Specifically, the framework can explain the emergence of two distinctive
types of society: strong religious states characterized by the wide diffusion
of religious norms and influential clerics imposing religious restrictions that
facilitate the extractive power of political rulers (eventually at some cost in
terms of economic efficiency), and alternatively secular states where religious
norms do not diffuse, clerics become steadily less influential on economic and
political aspects of social life, and civil society (merchants, workers, or popular
masses) eventually gains control over production and redistribution.
Interestingly, our analysis predicts that the joint evolution of religious values
and institutions will crucially depend on the initial conditions. When religious
values are initially sufficiently diffused in society, the institutional dynamics
will tend to steadily increase the political power of the clerics. Indeed, when
religious values are largely diffused, empowering the clerics is the most efficient
mechanism for reducing the externalities in policy choice problems. In turn,
institutional changes devolving more power to the cleric reinforce both the
religious values and eventually the extractive capacity of the political elite.
Alternatively, when religious values are initially not largely diffused in society,
institutions steadily decrease the power of the clerics and the political elites,
while society becomes less religious.
Furthermore, observe that the impact of the initial political strength of
the rulers on the institutional dynamics closely depends on the nature of
the externalities associated with the social policy game played between the
different parts of society. A strong political elite imposing important ineffi-
ciencies related to the lack of commitment may find particularly beneficial to
relinquish a fraction of its power to the religious clerics in order to pre-commit
to policies which reduce such inefficiencies. At the same time, a strong political
elite may also have access to other means than those provided by religious
clerics to resolve its governance problems and therefore would not need to
relinquish power to the religious clerics.
Some Historical Evidence on the Long Divergence

Our general framework provides an interpretation for two major historical
phenomena related to the growth of the Islamic world. First, it explains
the militarization of government structures in the Islamic world and the
concomitant increase in the political power of religious clerics. Second, our
framework relates the emergence of strong religious states in the Middle East
to the lower constraints that initially weighted on the power of the sovereign
relative to Western Europe.
In the first centuries after the death of the Prophet Muhammad, the
emergence of Islamic commercial institutions complemented the diffusion of
religious values through conversions in conquered lands; see Bulliet (1979).
Indeed, a significant share of the Islamic early doctrine is characterized by the
pursuit of commercial efficiency, and Islamic partnership law carried com-
mercial regulations into places previously without written laws. Furthermore,
the spread of Arabic facilitated communication, exchange, and cooperation,
and the commercial expansion of the Middle East triggered widespread
conversions; see Kuran (2012, p. 70) and Michalopoulos et al. (forthcoming),
which show that proximity to the pre-600 CE trade network is a robust
predictor of today’s Muslim adherence across countries and ethnic groups in
the Old World.
The introduction of slave armies in the Islamic world began approximately
in the ninth century. In this period, (1) secular bureaucracies contracted,
suggesting lower constraints on the power of the sovereign; and (2) Madrasa
spread and religious leaders emerged as social and political elites across the
Islamic world; see Chaney (forthcoming). This process of institutional change
culminated the Sunni Revival in the second half of the twelfth century, which
saw a surge in the endowments of waqfs, or pious organizations (Hodgson
1974, p. 51), as well as an increase in the political power of religious leaders;
see Chaney (2016).
The evolution of early Islamic societies is then characterized by a form
of complementary dynamics between religious values and Islamic commer-
cial institutions. Around the ninth century, religious values may have been
sufficiently diffused that the sovereign could increase its power by seeking
legitimacy from religious authorities. The introduction of slave armies, the
330 A. Bisin et al.
contraction of state bureaucracies, and the rise to political power of religious

clerics all occurred starting from the ninth century, in accordance with this
narrative.
Why did strong religious states form in the Middle East and not in Western
Europe? A narrative which contributes to an explanation of the long divergence
can be constructed along similar lines. The collapse of the Western Roman
Empire led to the emergence of a landed aristocracy in Europe. Indeed, in
Western Europe, military shocks—in particular the Germanic invasions of the
fifth century—may have contributed to the emerging autonomy of military
leaders at the expense of the sovereign; see Chaney (2011). In contrast, no
separation of powers occurred in the Middle East. Blaydes and Chaney (2013)
provide evidence that while constraints on the sovereign gradually increased
across Europe from the ninth century through the fifteenth century, this has
not been the case in this same period in the Islamic world. This may provide
a rationale, along the lines of the framework illustrated in the last section, as
to why rulers consolidated their power by relying on slave armies and religious
legitimacy in the Middle East and not in Western Europe.
Conclusions
In this short chapter, we have provided some preliminary insights on the
emergence of religious legitimacy in the context of the general theory of the
evolution of institutions and culture.
We modeled religious legitimacy as a phenomenon that induces a change
in the structure of institutional power between political elite, religious clerics,
and civil society. As religious legitimacy helps the political elite to internalize
some basic governance problems in terms of extractive capacity, religious clerics
may get more influence in the institutional structure. In turn, clerics exercise
this power by providing religious goods and services in larger quantities,
which then favors religious practices and activities, propagating beliefs within
the population that in turn justify the ruling of the political elite. Religious
legitimacy therefore helps maintaining if not increasing the power of the elite.
We finally discussed how this narrative can help explaining the long divergence
between the Middle East and Western Europe.
Of course much more needs to be done to better understand the key
interactions between the evolution of cultural norms that legitimize autocratic
rulers and the dynamics of institutions. Along with Greif and Rubin (2015),
however, this chapter could represent a starting point for the analysis of the
political economy role of religious legitimacy in this context.
References
Acemoglu, Daron. 2003. Why Not a Political Coase Theorem? Social Conflict,
Commitment, and Politics. Journal of Comparative Economics, 31: 620–652.
Acemoglu, Daron, and James A. Robinson. 2000. Why Did the West Extend the
Franchise? Democracy, Inequality, and Growth in Historical Perspective. The
———. 2006. Economic Origins of Dictatorship and Democracy. Cambridge: Cam-
bridge University Press.
Bisin, Alberto, Avner Seror, and Thierry Verdier. 2018. Religious Legitimacy and the
Long Divergence. Culture and Institutions in the West and the Muslim World. Working
paper.
Bisin, Alberto, and Thierry Verdier. 1998. On the Cultural Transmission of Prefer-
ences for Social Status. Journal of Public Economics 70: 75–97.
Bisin, Alberto, and Thierry Verdier. 2000a. A Model of Cultural Transmission, Voting
and Political Ideology. European Journal of Political Economy 16: 5–29.
Bisin, Alberto, and Thierry Verdier. 2000b. Beyond the Melting Pot: Cultural
Transmission, Marriage, and the Evolution of Ethnic and Religious Traits. The
and the Dynamics of Preferences. Journal of Economic Theory 97: 298–319.
Bisin, Alberto, and Thierry Verdier. 2017. On the Joint Evolution of Culture and
Institutions. Working Paper 23375, National Bureau of Economic Research.
Blaydes, Lisa, and Eric Chaney. 2013. The Feudal Revolution and Europe’s Rise:
Political Divergence of the Christian West and the Muslim World Before 1500
CE. American Political Science Review 107: 16–34.
Bloch, Marc. 1924. Les rois thaumaturges. Paris: Editions Gallimard.
Bulliet, Richard W. 1979. Conversion to Islam in the Medieval Period: An Essay in
Quantitative History. Cambridge: ACLS Humanities E-Book, Harvard University
Press.
Chaney, Eric. 2011. Separation of Powers and the Medieval Roots of Institutional
Divergence Between Europe and the Islamic Middle East. In IEA Papers and
Proceedings.
Chaney, Eric. 2016. Religion and the Rise and Fall of Islamic Science. Working paper,
Harvard University.
Chaney, Eric. Forthcoming. Religion, Political Power and Human Capital Formation:
Evidence from Islamic History. In Advances in the Economics of Religion, ed. Jean-
Paul Carvalho, Jared Rubin, and Sriya Iyer. Basingstoke: Palgrave Macmillan.
Coşgel, Metin, and Thomas Miceli. 2009. State and Religion. Journal of Comparative
Economics 37: 402–416.
332 A. Bisin et al.
Greif, Avner, and Jared Rubin. 2015. Endogenous Political Legitimacy: The English
Reformation and the Institutional Foundations of Limited Government. Working
Paper.
Hodgson, Marshall G.S. 1974. The Venture of Islam, Volume 2: Conscience and History
in a World Civilization. NONE Series. Chicago: University of Chicago Press.
Kuran, Timur. 2012. The Long Divergence: How Islamic Law Held Back the Middle
East. Princeton: Princeton University Press.
Michalopoulos, Stelios, Alireza Naghavi, and Giovanni Prarolo. Forthcoming. Trade
and Geography in the Spread of Islam. The Economic Journal. https://doi.org/10.
1111/ecoj.12557
Platteau, Jean-Philippe. 2017. Islam Instrumentalized. Religion and Politics in Historical
Perspective. Cambridge: Cambridge University Press.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the
Middle East Did Not. Cambridge Studies in Economics. Cambridge: Cambridge
University Press.
21
Strategic Interactions Between Religion
and Politics: The Case of Islam
Jean-Philippe Platteau
Introduction
To what extent and in what sense is Islam responsible for the problems
encountered by the countries in which it dominates? Foremost among such
problems are high political instability and the postponement or reversal of
social reforms conducive to long-term development: reform of the family
code and measures to improve women’s status, or modernization of school
curricula and measures to minimize rote learning of religious and other texts,
for example, clearly involve high costs in terms of growth opportunities
foregone. How to explain the simultaneous presence of these two problems
is the question addressed in this chapter. Our central argument rests on two
propositions. First, we disagree with the essentialist view according to which
Islam is a major obstacle to modern development because it has always been
associated with a merging of religion and the state, or a fusion between the
spiritual and political spheres of life. Second, we reckon that Islam possesses a
special feature in the form of a highly decentralized structure. It makes politics
comparatively unstable yet, by buying off religious clerics, autocratic rulers can
mitigate instability at the cost of fewer institutional reforms. Radicalization of
the clerics nevertheless makes this co-option strategy more costly.
J.-P. Platteau ()

Université de Namur, Namur, Belgium
e-mail: jean-philippe.platteau@unamur.be

334 J.-P. Platteau
Our focus on the relationship between autocratic politics and religion is

reasonable because modern states tend to be non-democratic during the crucial
phase of their formation and consolidation, as attested by the importance
of absolute monarchies in Western Europe before the Industrial Revolution.
The legitimacy of autocratic regimes cannot rest on the principles of democ-
racy, and they therefore need to rely on other sources. In the context of
traditional societies in which literacy levels are low and religious authorities
have a monopoly or quasi-monopoly in the transmission of knowledge, these
authorities exert an important influence by conferring legitimacy through
loyalty upon the ruling autocracy. The subjects are encouraged to believe that
the ruler has the right to rule and the power to provide protection and other
public goods and thus the right to collect taxes. In sum, given the power of
religious belief, the word of the religious authorities—through sermons and
speeches delivered as part of their official function—could provide “a single,
coherent, and effective source of legitimacy” (Coşgel et al. 2012, p. 362).
Our argument on the strategic relationship between Islam and politics
follows a political economy approach that section “A Political Economy
Approach to Islam and the Predicament of Muslim Countries” sketches.
Within that framework, section “Decentralized Versus Centralized Religions”
contrasts the effects on political stability of decentralized and centralized
religions. Section “Conclusion” concludes. The reader interested in obtaining
more details about the various steps involved as well as the empirical basis
behind them can refer to Platteau (2017). The underlying formal model is
expounded in Auriol and Platteau (2017a, b).
A Political Economy Approach to Islam

and the Predicament of Muslim Countries
The Analytical Framework
The starting idea is that only in the times of the Prophet have religion
and politics been truly merged in the history of Islam. After the death of
Muhammad, violent confrontations between different power-seeking factions
became the order of the day, and each faction claimed legitimacy for its own
version of inheritance from him. Politics thus took precedence over religion,
and military men often occupied the commanding positions, whether at the
center or behind the stage. The implication is that Islam is separable from
politics. Religious clerics must therefore be conceptualized as actors separate
from the state, and they must decide how they want to relate to it.
21 Strategic Interactions Between Religion and Politics: The Case of Islam 335
Not only has autocracy persisted as the dominant political system in

Muslim countries since the times of the Umayyads and the Abbasids, but
submission of the clerics to the autocrat quickly became a general rule of
conduct. Not infrequently, it amounted to slavish obedience. The idea that
religion is the handmaiden of politics and that religious clerics are expected to
cooperate with absolute monarchs in a subordinate position has been justified
in principle despite the professed aim of Islam to establish a righteous world
order and provide guarantees against despotic rule. In this canonical situation,
epitomized by Safavid Iran, pre-colonial Morocco, the Ottoman and most of
the Mughal empires, the autocrat succeeds in wielding complete control over
the religious clerics, and the political regime is stable. The strategy chosen
by the autocratic ruler is called an Opposition Suppression Strategy. The
alternative equilibrium is obtained when the ruler chooses the Opposition
Confrontation Strategy in which only a fraction of the religious clerics is
brought into submission, carrying the risk of a popular rebellion led by them.
A political crisis may then arise as the endogenous outcome of the autocrat’s
choice. A clue to a better understanding of how the autocratic ruler may choose
to have partial rather than complete (or near complete) control over the clerical
body lies in the existence of a trade-off between political stability and the
autocrat’s realization of his preferred policies and selfish interests. This trade-
off is critically influenced by the behavior of the clerics.
Religious clerics have two special features that distinguish them from other
elite members: (1) they hold values regarding social justice and human rights,
or regarding proper behavior, that they derive from their religion; (2) as
representatives of the supernatural world and as wise men possessing deep
knowledge (theological and philosophical, in particular), they have a natural
prestige and influence on the population. Because of these two traits, the clerics
are susceptible of playing a role as political actors or social leaders, especially
in societies where most people are uneducated and strongly believe in the role
of supernatural forces. At the same time, however, the clerics are vulnerable to
corruption, meaning that they can be “bought off”, seduced or corrupted, by
the autocrat. The price of their submission increases with the distance between
their values and the policies or practices followed by the autocrat.
Because the clerics attach different weights to the values inspired by their
religion as compared to the weights attached to material achievements (their
preferences are heterogeneous), the autocrat chooses the proportion of them
whom he wants to co-opt. This he does with the knowledge that those left out
may become opposition leaders, thereby representing a threat to the stability
of his autocratic rule. How many clerics to bribe with material privileges is not
the only instrument in the hands of the autocrat. He may also choose policies
336 J.-P. Platteau
that are more or less acceptable to them: for given levels of perquisites received
from the autocrat, policies that have strong disequalizing effects and involve a
great measure of elite corruption, or those that hurt religious values or interests,
tend to arouse more opposition from the clerical body. When determining his
policies and the extent of co-option of religious clerics, that is, when selecting
his “policy mix”, the autocrat pursues his own interests conceptualized as his
expected income. This implies that he pays attention to both his income and
his probability of political survival, which are influenced by the extent of
religious co-option: indeed, not only does co-option involve costs that must
be subtracted from the gross income of the autocrat (and his clique), but more
extensive co-option also reduces the risk of popular rebellion.
Inside the religious body, the co-option strategy may create a divide between
the official clerics, co-opted by the autocrat, and those who stand outside the
ambit of the state. They either belong to rather independent institutions run by
the ulama themselves, or they are self-appointed clerics and firebrands who act
outside any kind of organization. Clerics of the latter type tend to be radicals
susceptible of organizing popular rebellions. The division between official and
radical clerics is possible in the world of Islam because no hierarchy exerts
authority over the whole clerical profession. No church establishment exists,
and, as a result, the clerics operate in a rather decentralized way, pronouncing
their own fatwas as they deem fit.
Characterizing Policy Choices and Their Effects on Political

Stability
The canonical politico-religious equilibrium is brought about when the auto-

crat’s “policy mix” consists of wide co-option of religious clerics combined with
moderately popular policies. An unstable autocracy prevails when the opposite
mix is chosen: the autocrat follows policies that blatantly favor his inner
circle and the surrounding elite, surrender national sovereignty to external
powers, and/or antagonize the sort of (traditional) values cherished by religious
representatives. If an open rebellion occurs that succeeds in overthrowing the
autocrat or in severely limiting his ruling capacity, a crisis situation arises.
Rebellious clerics then successfully enter the political stage in order to protect
ordinary people or rescue the nation from colonial domination. Under these
circumstances, socio-economic and cultural grievances tend to be expressed
in the language familiar to most people, a religious idiom depicting a fateful
struggle between the forces of good and the forces of evil, and promising to
bring justice to the oppressed. Led by Khomeini, the 1978–1979 ayatollah
revolution in Iran provides a remarkable illustration of this scenario: a rather

massive uprising of clerics, rooted in continuous struggles that stretched back
to the early nineteenth century and were directed against corruption and the
surrendering of the shah to foreign imperialist forces, occurred in spite of
the tradition of quietism characteristic of Shi’ism (Keddie 2003, chap. 1–4;
Platteau 2017, pp. 165–175; Amanat 2017).
Antagonizing traditional values upheld by the religious elite and reforming
Islamic institutions may generate political instability in the same manner as
inequitable policies. Nonetheless, pervasive corruption, cynicism, aloofness,
and callous indifference of the elite surrounding the autocrat appear far more
damaging for political stability than progressive institutional changes ushered
in by an honest, equitable, and dedicated ruler. It is when the former rather
than the latter situation is observed that religious clerics tend to be more prone
to rebellion and that their opposition move is more likely to resonate among
the masses. Positions of religious dignitaries in traditional religious institutions
can indeed be exchanged against positions in new state structures, whereas self-
appointed clerics do not care much about the way the ruler treats institutions
of the “high Islam”, insofar as they are excluded or have excluded themselves
from these. Moreover, the most unstable political situation is created when the
autocrat simultaneously pursues progressive institutional reforms and socially
unjust policies. By contrast, an enlightened autocrat is wise enough to adopt
inclusive growth policies to accompany institutional change, and to introduce
such change in a manner that does not openly confront the “low Islam” of the
uneducated masses. He thus achieves the best compromise between political
stability and long-term development of the country.
The situation most often observed in Muslim countries since World War II
is best depicted as unstable autocracy. It is characterized by the combination
of a partial co-option of the religious elite, which results in a division between
official and non-official clerics, and pervasive elite corruption, somehow
mitigated by social subsidies and moderate institutional reforms. Eager to
preserve their privileges and to forestall anti-corruption measures and political
reforms, Muslim autocrats have mobilized Islam in order to buttress their
legitimacy and condone their unjust policies. Because of this strategic choice,
most public debates and controversies are framed in religious terms. On the
one hand, by presenting progressive and secular opposition forces as apostates
and enemies of Islam, the regime not only prevents any serious discussion of its
policies, but it also justifies its harsh crackdown on these forces. On the other
hand, the opposition, gradually deprived of its secular and leftist components,
becomes dominated by self-appointed religious leaders who blame the autocrat
and his clique for their corruption, cynical opportunism, and hypocritical
338 J.-P. Platteau
behavior. Such domination is easily established because in traditional societies

leftist ideas do not have a large appeal: ordinary people are therefore not ready
to come to the rescue of leftist militants when they are the victims of brutal
repression.
The Risk of Religious Outbidding
In many countries, the political stage has thus been largely dominated by,
on one side, official clerics who pronounce fatwas vindicating the regime’s
religious legitimacy and, on the other side, rebellious clerics from the low Islam
who pronounce counter-fatwas accusing the ruling clique of being miscreants
who transgress Islamic values and pervert the original message of pure Islam.
The former type of clerics are faithful to a deep-rooted Islamic tradition
prescribing that, to avoid chaos and disorder, Muslims should obey their
sovereign regardless of the despotic character of his rule. The only condition
is that he acts as a pious Muslim on the superficial basis of his official gestures
and postures. As for the second type, they are deviant clerics who have entered
into open rebellion against the official religious establishment.
What the autocratic authority is thus sparking is a dangerous religious war
in which both the regime and the opposition try to outbid each other in
their claim to be the most legitimate bearer of Islamic values and principles.
Intransigent discourses and a winner-take-all attitude come to invade the
political space where arguments are replaced by anathema and confrontation
takes on the form of a Manichean struggle between the forces of good and
evil. Some strand of religious opposition, typically of urban origin, may take
on the shape of Puritan movements preaching return to the pristine form of
Islam. More moderate groups clamor for the replacement of state laws by the
sharia deemed to be the only way to coax the despotic sovereign to end blatant
corruption and oppression (originally, sharia means a way of promoting the
well-being of the individual and the community). An “obscurantist deadlock”
is thereby created in unstable autocracies, and how it ends up is an open
question. A possible outcome is chaos caused by the elimination, physical
or political, of the autocrat. In Egypt, for example, Anwar al-Sadar was
assassinated by a member of the violent Islamist al-Jihad group (in 1981), and
Mubarak was overthrown (in 2011) by a wide coalition of opposition forces
(Cook 2012). This may be followed by a takeover of political power by the
army acting in support of autocracy or by the coming to the frontline of
politics of religious leaders determined to restore social order in the name of
Islam. The final outcome is often the emergence of a secular regime resting
on the use of coercion and repression (as illustrated by al-Sisi rise to power
in 2013). The corruption and cynicism of often secular despotic rulers are
largely to blame for this sobering association between secularism and force. It is
utterly disappointing for all those who believe that secularism should promote
a democratic order and an inclusive society based on tolerance, fair access to
economic opportunities, and peaceful cooperation among people.
When autocrats opt for a wide religious support in order to stabilize their
regime or make up for their lack of legitimacy, they may have to rely on the
allegiance of religious family dynasties which lead big Sufi brotherhoods and
wield considerable local political influence due to their moral authority and
their patronage power. In these cases, the co-option of clerics goes beyond
the world of high Islam to reach out to lower rungs. The rulers are then
automatically tempted to enact laws or adopt measures that reflect erstwhile
tribal customs and not only the preferences and values of the high-level urban
ulama. The consequence is that tribalism and clannism are consolidated, as
attested by the late rule of Saddam Husayn in Iraq and by Zia ul-Haq and
his successors’ regimes in Pakistan (see Baram 2014 for the former and Gazdar
2000 for the latter country).
The Role of the International Factors
To understand why many Muslim autocracies are politically unstable since

the end of World War II, we need to bring the role of the international
context into the picture. One important channel of influence goes through
the supply of Islamist ideologies, the propagation of which is facilitated by
the abundant oil wealth of Saudi Arabia, the Iranian Islamist Revolution,
unresolved problems of statehood in Pakistan and Afghanistan, and the ready
availability of effective mass communication technologies. On the other side,
the demand for ideologies stressing the victimhood of Muslim people and
demonizing the Western civilization has been stimulated by the one-sided
meddling of advanced Western countries in the regional conflicts of the
Middle East. Such meddling has been associated with colonial interventions,
the priority given by Western countries to their own geopolitical interests
in the context of the Cold War, the struggle against emerging left-wing and
nationalist movements in developing countries, and the unflinching support
to Israel by the US in particular.
Factors pertaining to the supply of/demand for Islamist ideologies, plus
the threats and challenges arising from the pressure to catch up with the
rapidly developing economies of the West, have modified the trade-off faced
340 J.-P. Platteau
by Muslim autocrats. Religious clerics, at least those relatively sensitive to

social injustice, become harder to buy off (they are more “expensive” than
before), making the autocratic regime more unstable. Confronted with a
growing threat, autocrats are not expected to remain passive, however. Their
predicted response consists of moderating their controversial policies and ill-
famed practices or of backtracking on those institutional reforms that reduced
the influence of Islam in public life. The policy shift ought to be large enough
to ensure that they eventually regain the support of the clerics even at the cost
of creating new barriers against economic growth and development.
Islamist movements, which tend to appeal to educated or semi-educated
people with dislocated life experiences, are born of deep-seated frustrations
caused by the behavior of both political and religious elites. Their struggle
tends to be especially fierce and determined when, as a result of the cor-
ruption and/or incompetence of the political autocracy, national interests are
surrendered to foreign powers. The proclivity of these movements to adopt
Puritan scripturalist interpretations of the Prophet’s message is the outcome
of two circumstances: the association of corruption with the values of material
individualism and atheism and the obsequious attitude of religious dignitaries
accused of being “lackeys of the prince” hurting the “dignity of Islam” (Kepel
2005, p. 101). This second feature is especially important because it reinforces
the idea that what matters are deeds and not talks: speaking in the name of
Islam even with the apparently highest credentials may just be a trick to conceal
a devious cooperation with those who treat the state as their personal fiefdom
and as a set of arrogated privileges that can be sold to foreign foes. Official
Islam is thus seen as a debased version of the primeval faith, justifying the
need for the latter’s restoration. In that sense, Islamist radicalism resembles
Christian radical movements, millenarist movements in particular, such as
Savonarole’s short-lived Christian republic in Florence (ended in 1498), the
Protestant Anabaptists in Münster, or the totalitarian rule of Thomas Müntzer
in Mühlhausen (see Ory 2017, chap. 5).
In the postwar pattern of autocratic polity, the ruler instrumentalizes a
portion of the clerical body constituting the set of so-called official clerics.
Since those left out stand beyond his control, they are a potential threat to the
regime. As revealed by the recent experiences of several Arab countries (e.g.,
Syria, Algeria, Yemen, and Egypt), things may be more complicated: indeed,
the “deep state” consisting of various forces obeying the Ministry of Interior,
the intelligence services, and top brass in the army may act behind the public
stage of autocratic regimes to defeat left-wing secular movements, through
thuggish groups if needed. Both the state and radical religious groups then play
the same vicious and fateful game of extremism, justifying means by the ends.
Intelligence and police forces may nurture and encourage radicalized religious
movements on the condition that they directly attack leftist or other secular
opposition movements operating on university campuses, inside trade unions
and professional associations. In this case, the autocrat relies on two kinds
of religious forces to buttress his regime: official clerics serving as its “clean”
partners and violent Islamist organizations which are his covert and shameful
ally in the struggle against secular opposition. Political cynicism and perversion
are at their highest when the autocratic regime simultaneously supports violent
outfits and misrepresents moderate religious opposition forces. This is done
by demonizing the latter, conflating them with the former so as to justify
the harshest repression against moderate opposition in the name of anti-terror
struggle.
It is a distressing fact that Western powers themselves did not hesitate
to cynically support radical Islamist movements when it suited their own
geopolitical and economic interests. This was especially manifest in their
protracted struggle against Arab nationalism and in their tactic of using radical
Islamists as proxy allies against communism and the Soviet Union. They thus
played a “devils’ game” that later proved to be disastrous when Islamism
became a force that turned against its sponsors (see Curtis 2010 for the UK
and Dreyfuss 2005 for the US).
Decentralized Versus Centralized Religions

Under the decentralized organization of Islam, clerics unhappy about the
coziness of the arrangement of their official co-religionists with political
power tend to opt for independence. This leaves them the possibility of
leading popular rebellion against the despised autocratic regime (and religious
officialdom). When the religion is centralized, by contrast, clerics dissatisfied
with the compromising attitude of their church typically try to reform it from
within, as was evident even during the Protestant Reformation. It is only in
extreme circumstances that they choose to leave it and even join revolutionary
movements, such as happened in Russia during the last two Romanov reigns
(Obolonsky 2003, pp. 110, 136–137) and in Latin America during the 1960s
and 1970s. Unlike dissident Muslim clerics, however, they then lose their
religious credentials and their ability to claim supernatural legitimacy (they
are excommunicated).1
1 In the case of Russia, the main revolutionary movement had characteristics strongly evoking a centralized
religion with its counter-church. Marxist or communist ideology actually supplanted the Orthodox faith
342 J.-P. Platteau
A major advantage of our analytical approach is that it does not treat

religions in an undifferentiated manner. It actually allows us to bring out both
similarities and differences between them, in particular between decentralized
and centralized religions, such as Islam for the former and Christianity (with
the exception of American Protestantism) for the latter. Similarities originate
from the fact that, like Islam, Christianity cannot be properly analyzed outside
a setup featuring the state, and this applies to the periods preceding and
succeeding the rise of national churches. The central difference between the
two faiths lies in their internal organization: unlike Islam, Catholic and Eastern
Orthodox Christianity are endowed with a hierarchical structure under the
form of a centralized church. Something akin to a war of fatwas and the
resulting disorder cannot therefore occur: either the church cooperates or it
is in conflict with political power.
A key prediction of the theory is that political instability is stronger with a
decentralized than with a centralized religion, thus illustrating Ory’s statement
that “the strength of established churches lies in their ability to contain radical
energy” (2017, p. 121). Supporting empirical evidence is found in Blaydes and
Chaney (2013) who compared durations of autocratic rule for the Christian
West and the Muslim world. They found that, from the tenth century onward,
Christian kings were increasingly long-lived compared to Muslim sultans.
Divergence in ruler duration reflects a change in political stability, as indicated
by the inverse relationship between ruler duration and the probability of being
overthrown. Ruler duration therefore appears as a reasonable proxy for political
stability, suggesting that over time, rulers in Western Europe were significantly
less likely to be deposed than their Muslim counterparts.2 An immediate
implication of the theory is that, whenever this proves possible, such as when a
new state is formed, autocratic rulers should try hard to establish a centralized
national church. This is precisely what Safavid rulers strove to do in early
seventeenth-century Iran, thus influencing the modern evolution of the Shi’a
brand of Islam.
“which had discredited itself by means of its complete submission to crown” (Obolonsky 2003, p. 166). It
operated as a new faith antithetical to Christianity, with its temptation to turn stones into bread, to make
social miracles, and to build an eternal kingdom on earth. It had its own cult of saints and its own holy
legends and own dogmas. Any doubt, criticism or disrespect regarding these symbols of the new faith was
sanctioned by excommunication, and even the horrors of inquisition were imposed on the people living
under the new faith (pp. 166–167).
2 Blaydes and Chaney put forward a different explanation than ours in order to account for their result.
Their explanation rests on the contrast between the use of mercenary (slave) armies in the lands of Islam
and the use of loyal armies at the service of the autocrat in the lands of Christianity.
What bears emphasis, however, is that in our framework something close to

a consensus among clerics may arise as the endogenous outcome of particular
circumstances, yet cannot be taken for granted.3 That centralization is not an
organic feature of Shi’ism is evident from the fact that the Shia approach to
Islamic rulings does not necessarily lead to a consensus at any given point
of time because a plurality of interpretations is considered to be normal
(Berkey 2003, p. 136). This explains why charismatic leadership came to play a
significant role in Shi’ism, with the prestige of the religious leader being more
important than the relationship of his rulings to the Quran or the sunna.
As historical evidence attests, once the leader passes away, and perhaps even
before, volatile and plural opinions are quick to return (Lee 2014, p. 186).
Conclusion
During the critical period corresponding to the formation of their central-
ized modern states, European countries have been particularly successful in
building a cooperative relationship with national church establishments. By
contrast, the decentralized character of Islam makes the same task much more
arduous in countries with large Muslim populations: in these countries the
state tends to follow the path of increasingly unstable autocracies. A sudden
shift toward democracy appears quite difficult in such conditions. Yet, even a
less radical and more realistic transition involving a shift from kleptocratic or
predatory autocracy to liberal autocracy has proven to be problematic in most
Muslim countries. In a kleptocratic autocracy, the despot not only ensures
law and order but also provides exclusive and unjustified economic privileges
to his clique. Law and order therefore appear to be aimed at defending
these privileges rather than at protecting ordinary people faced with threats
against their physical security and their day-to-day livelihood. By contrast, a
liberal autocracy is a regime in which the economic space is open to genuine
competition instead of being controlled by the autocrat’s circle for its own
benefit. Because it provides fair access to economic opportunities for all the
population groups, the authoritarian methods used to establish law and order
are then considered acceptable by a large majority of people.
By treating the left as the most dangerous threat to their political survival,
predatory autocrats have ended up suppressing the progressive forces that
3 If we go back to the time of its origin, we find that Shi’ism actually began “as a rather vague and undefined
movement of support for the leadership of someone from the family of the Prophet” (Berkey 2003, p.
136).
344 J.-P. Platteau
could have compelled them to become gradually more accountable to their

people. The pressure exerted by these forces would have hopefully involved the
discontinuation of the most blatant privileges granted to an elite exclusively
concerned with its own well-being. Not only would the scope of corruption
have been reduced but also, and more importantly, its form would have
hopefully evolved from the noxious type of prebendary taxation, expropriation
and racketeering for purely patronage purposes to a system of privileges
conditioned upon socially useful achievements, such as has been observed in
East Asia (Amsden 1989; Khan and Jomo 2000). The prospect of a transition
from patrimonial to liberal autocracy is certainly more conducive to economic
and social development than running the risk of a religious takeover in order
to leave unchanged a system of rent capture that exclusively benefits a narrow
ruling clique. Since religions stress moral absolutes and thereby create aversion
to political compromises, a cleric-led opposition is quite unlikely to stop
short of a revolution which, if successful, will reproduce autocracy rather than
establish an accountable political regime. The crushing of progressive, left-
leaning forces therefore appears as a real tragedy for a large part of the Muslim
world.
References
Amanat, A. 2017. Iran: A Modern History. New Haven & London: Yale University
Press.
Amsden, A. 1989. Asia’s Next Giant: South Korea and Late Industrialization. Oxford:
Auriol, E., and J.P. Platteau. 2017a. Religious Seduction Under Autocracy: A Theory
Inspired by Theory. Journal of Development Economics 127: 395–412.
———. 2017b. The Explosive Combination of Religious Decentralisation and
Autocracy: The Case of Islam. Transition Economics 25 (2): 313–350.
Baram, A. 2014. Saddam Hussein and Islam, 1968–2003: Ba’thi Iraq from Secularism
to Faith. Baltimore: Johns Hopkins University Press.
Berkey, J.P. 2003. The Formation of Islam: Religion and Society in the Near East, 600–
1800. Cambridge: Cambridge University Press.
Blaydes, L., and E. Chaney. 2013. The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian and Muslim Worlds Before 1500 CE. American
Cook, S.A. 2012. The Struggle for Egypt: From Nasser to Tahrir Square. Oxford: Oxford
University Press.
Coşgel, M.M., T.J. Miceli, and J. Rubin. 2012. The Political Economy of Mass
Printing: Legitimacy and Technological Change in the Ottoman Empire. Journal
of Comparative Economics 40: 357–371.
Curtis, M. 2010. Secret Affairs: Britain’s Collusion with Radical Islam. London:
Serpent’s Tail.
Dreyfuss, R. 2005. Devil’s Game: How the United States Helped Unleash Fundamentalist
Islam. New York: Holt Paperbacks.
Gazdar, H. 2000. State, Community, and Universal Education: A Political Economy
of Public Schooling in Rural Pakistan. Asia Research Centre, London School of
Economics.
Keddie, N. 2003. Modern Iran: Roots and Results of Revolution. New Haven & London:
Yale University Press.
Kepel, G. 2005. The Roots of Radical Islam. London: Saqi Books.
Khan, M.H., and K.S. Jomo, eds. 2000. Rents, Rent-Seeking and Economic Develop-
ment. Cambridge: Cambridge University Press.
Lee, R.D. 2014. Religion and Politics in the Middle East: Identity, Ideology, Institutions,
and Attitudes. 2nd ed. Boulder, CO: Westview Press.
Obolonsky, A.V. 2003. The Drama of Russian Political History: System Against Individ-
uality. Austin: Texas A & M University Press.
Ory, P. 2017. Peuple souverain: De la révolution populaire à la radicalité populiste. Paris:
Gallimard.
Platteau, J.P. 2017. Islam Instrumentalized: Religion and Politics in Historical Perspective.
22
State and Religion: An Economic Approach
Metin M. Coşgel and Thomas J. Miceli
State and religion, two of the oldest institutions known to mankind, have
historically had a complex relationship with each other. At times rulers
have suppressed religion altogether, at others they have treated religion as
independent of the state, and at still others they have preferred one religion
over others or even endorsed one as the official religion. A survey of 177
countries in the year 2008 reveals a similar diversity of attitudes. There were 16
countries in the survey (9%) which exhibited a hostile attitude toward religion;
43 countries (24%) had a neutral attitude; 77 countries (44%) clearly favored
certain religions; and 41 countries (23%) endorsed one or more religions as
the official state religion.1
The current fraction of states with an official religion is actually low by
historical standards. Throughout history, states typically had an established
religion, an arrangement that often went back hundreds or even thousands
of years. At the beginning of the seventeenth century, for example, almost
1 The data come from Round 2 of the Religion and State dataset (http://www.religionandstate.org),
collected under the supervision of Jonathan Fox. The original data uses 15 categories to measure the formal
relationship between religion and the state. To simplify, we have aggregated these categories as hostility
(coded as categories 0–3 in the original data), neutrality (4–5), favoritism (6–9), and state religion (10–14).
See Fox (2008) for a comprehensive analysis of the relationship between state and religion based on the
Round 1 of the dataset.
M. M. Coşgel () · T. J. Miceli
Department of Economics, University of Connecticut, Storrs, CT, USA
e-mail: Metin.Cosgel@uconn.edu

348 M. M. Coşgel and T. J. Miceli
90% of the territories corresponding to today’s nation-states were under a

state religion (Coşgel et al. 2018). The fraction fell sharply thereafter, as states
increasingly severed official relationships with religion, moving instead toward
varying types and degrees of separation or even hostility.
In order to make sense of this complex set of institutional structures, the
next section develops a simple analytical framework for thinking about the
relationship between state and religion from an economic perspective. This
framework then serves as the point of departure for surveying the literature in
this area of research.
Analytical Framework
This section briefly describes an analytical framework first developed in Coşgel
and Miceli (2009) to describe the economic relationship between state and
religion. The model consists of three agents or sectors: the political authority or
ruler, the religious sector (the church or religion market), and a representative
citizen. The citizen has a utility function that depends on a religious good,
q,2 and a composite consumption good, x (the numeraire), according to the
simple, quasi-linear form:
U = x + v(q), (22.1)
where v > 0 and v < 0. The citizen is endowed with wealth of W,3 which he
or she spends on x; taxes, which we assume are assessed by the ruler as a lump
sum T ; and the religious good. The budget constraint is therefore
W = x + pq + T , (22.2)
where p is the price of the religious good, assuming for now that it is
purchased in an independent “religion market.” The demand for q is found
2 We do not specify the exact nature of this good. Generally, it is meant to reflect those goods and services
that a church provides to its followers, whether material (worship, charity) or non-material (forgiveness
of sins, promise of salvation).
3 In a more general setting, we would allow the citizen to engage in production.
by maximizing (22.1) subject to (22.2), which yields the first-order condition
v (q) = p. (22.3)
This defines the demand function under an independent religion. (Note that
the function is downward sloping given diminishing marginal utility.)
The state/ruler sets the amount of taxes, T, to extract the maximum wealth,
subject to the citizen’s reservation utility, U , which represents the level of
utility associated with subsistence, or that just avoids a popular revolt.4 Setting
U = U in (22.1) and using (22.2) to substitute for x, we obtain the “tax
capacity” function:
T (q) = W + v(q) − pq − U, (22.4)
which represents the maximum ability of citizens to pay.

While the expression in (22.4) is the amount of taxes the citizen is able to
pay, it does not necessarily represent the amount that the ruler actually collects
because of tax collection costs, which reflect citizens’ resistance to taxation. We
capture this by assuming that for each dollar of taxes assessed, a fraction δ is
lost. Thus, the revenue collected is

T (q) (1 − δ) = W + v(q) − pq − U (1 − δ) . (22.5)
Note that the consumer surplus from the religious good, v(q) − pq, contributes
positively to this amount. This reflects the pacifying function of religion
as recognized by Marx. Thus, even an independent religion that is not
sympathetic to the state can be useful to the latter by raising the tax capacity
of the citizenry.
The expression in (22.5) implies that the organization of the religion market
will also affect the ruler’s tax collection through its effect on the consumer
surplus from the religious good. Suppose first there is a single, dominant
religion, which, as Stark (2007, 120–122) notes, has been the usual state of
religion markets throughout history. In this case, the monopolistic church
will choose q to maximize its profit and will therefore be able to divert some
potential consumer surplus to profit, thereby reducing the amount that the
state can extract in taxes. In contrast, if the religion market is organized
competitively in the sense that there is free entry of religions, competition will
4 We assume that W > U , for otherwise, the citizen’s endowment would not be sufficient to cover
subsistence.
cause the output of the religious good to increase to the level that maximizes
consumer surplus. In this case, the ruler will be able to extract a larger amount
of taxes from citizens as compared to the monopolistic market structure. It
follows that when religion is independent of the state, rulers would prefer a
competitive religion market because it maximizes the tax capacity of citizens.
(This assumes, however, that religious teachings are either neutral toward the
state or not so antithetical that the latter would find it desirable to suppress
religion altogether, a point to which we return below.)
Economists have long observed that the organization of religion market
would have direct consequences for the incentives of clergy and the behavior of
individuals. In The Wealth of Nations, for example, Adam Smith (1776) laid the
foundation for an economic analysis of religion by using an analogy between
the remuneration systems of clergy and public-school teachers, arguing that
a competitive market would enhance the performance of clergy and public
welfare. It remains true that one of the most controversial questions in the
social-scientific study of religion has been the effect of competition and
pluralism on religiosity. Sociologists of religion, for example, have traditionally
argued that diversity in the religion market would undermine religiosity
because a multiplicity of religions would reduce the plausibility of each belief
(Berger 1967). The advocates of an economic approach have challenged this
view, however, arguing that diversity would increase religiosity by enhancing
competition among groups, thereby offering believers a greater variety and
quality of choices.
Building on Smith’s insights, the seminal study by Iannaccone (1991)
explored implications for the effect of market structure on religious beliefs and
participation. Based on data from developed Western nations, he argued that
the rates of church attendance and religious belief were significantly higher in
competitive markets than in those monopolized by established churches. (Also
see Iannaccone et al. 1997.)
Despite the large number of subsequent studies devoted to this question,
the results have been mixed and inconclusive because of omitted variables
affecting both diversity and religiosity. To mitigate the endogeneity problem,
Coşgel et al. (forthcoming) have proposed a novel identification strategy that
exploits the variation among nations in their geographic proximity (cost of
travel) to the “capitals” of the universal religions of the world (Buddhism,
Christianity, Islam).5 Their results show that direct effect of diversity, after
5 The specific centers are Lumbini, Nepal (Buddhism); Wittenberg, Germany (Protestantism); Istanbul,
Turkey (Orthodox Christianity); Karbala, Iraq (Shia Islam); Mecca, Saudi Arabia (Sunni Islam); and
Vatican City (Roman Catholicism). These are the centers of universal religions, or their sub-branches,
controlling for endogeneity, is positive, a finding that is consistent with the

economic prediction that enhanced competition in the religion market would
increase religiosity.
Alliance of Religion and State: State Religion

The above model predicted that, because consumer surplus is higher under
competition than under monopoly, the state would be able to extract more
tax revenue when the religion market is organized competitively, all else equal.
This conclusion provides one motivation for the state to seek an alliance with a
monopolistic religious organization—namely, to capture some of the revenues
that the religion is able to divert to its own use. For such an alliance to be
consensual, however, the religious organization must receive some benefit in
return. One such benefit is for the state to use its monopoly on force to prevent
entry of competing religions. The joint returns to the two parties under such
an alliance would consist of the sum of the tax revenues in (22.5) and the
profit of the monopolistic religion, subject to the constraint that the religious
organization chooses q, the level of the religious good, to maximize profit.
We can think of this as a modified principal-agent problem in which the two
parties divide the available surplus according to some contractual arrangement
reflecting their relative bargaining strengths, subject to the church’s incentive
compatibility constraint.
There is also a second benefit that religion can provide to the state under
such an alliance. That is to legitimize the ruler, which has the effect of lowering
the tax collection cost, δ (Coşgel and Miceli 2009). For example, citizens will
be less resistant to paying taxes if the religious authorities declare the ruler to be
divine, or divinely inspired.6 Legitimacy increases the joint returns by bringing
taxes collected (the expression in (22.5)) closer to the tax capacity of citizens
(the expression in (22.4)). The more effective is this legitimizing function, the
more favorable is an alliance between state and religion.
Historically, the typical alliance between state and religion has been in the
form of an official state religion.7 Barro and McCleary (2005, 1332) note, for
that have historically expanded out from their birthplaces, eventually becoming main religions in other
territories.
6 The same proposition, stated conversely by Adam Smith, is that “When the authorized teachers of
religion propagate . . . doctrines subversive of the authority of the sovereign, it is by violence only, or
by the force of a standing army, that he can maintain authority” (Smith [1776] 1965, 749).
7 For recent historical analyses of the relationship between state and religion, see Mueller (2013) and Vaubel
(2017). See also Coşgel et al. (2009) for the evolution of state-religion relationship in Islamic history.
example, that state religion has been “probably . . . the single most important
form of state monopoly in existence.” Using a Hotelling-style model of spatial
competition and cross-national data, they identify factors contributing to the
likelihood of state religion in the years 1970 and 2000. Their analysis shows
that the likelihood rose with the adherence rate to the main religion, fell under
communist regimes, had a nonlinear relation with population, and was not
significantly related to per capita GDP.
Coşgel and Miceli (2009) develop a broader model of the state-religion
relationship which centers on the ability of religion to confer legitimacy on
political rulers. This approach allows different configurations of this relation-
ship, including suppression and independence, as special cases emerging from
a single framework that is driven by the interaction between a self-interested
ruler and citizens, as detailed above. Their empirical analysis, which is based
on cross-national data, shows that the likelihood of a state-religion alliance at
the beginning of the twenty-first century was negatively related to the levels
of democracy and was positively related to the rates of religious loyalty and
concentration in the religion market, as predicted by the model.
Since the presence of an official state religion would eliminate or reduce
competition in the religion market, it can influence religiosity in society
through the interactions of mechanisms discussed above. The empirical results
on this question, however, are mixed. Based on cross-national data on religios-
ity, North and Gwin (2004) have found that the presence of state religion
reduces religiosity, consistent with Smith’s arguments regarding the positive
incentives of clergy under greater competition, and the findings of modern
economists who have observed a positive influence of competition on reli-
giosity. Using the method of instrumental variables to remedy the potential
endogeneity between religiosity and per capita GDP, McCleary and Barro
(2006) have reached the opposite conclusion regarding the effect of state
religion, namely, that state religion actually increased religious beliefs and
participation. Their interpretation is that this may be due to the subsidies that
the state provides to official religion.
Decline of State Religions

Although the alliance between state and religion has been dominant in history,
the disestablishment of state religions has been one of the most dramatic
institutional transformations of the modern era. Coşgel et al. (2018) offer a
systematic analysis of the decline of state religions based on a political economy
approach that again centers on the notion of legitimacy. Adopting the view of
religion as a legitimizing force for political leaders, they consider the factors
affecting the cost and benefits of alternative sources of legitimacy.
For an empirical analysis of forces influencing the secularization of states,
they develop a new cross-national time-series dataset for the relationship
between state and religion since the year 1000. Using this dataset, they
constructed Fig. 22.1, which shows the evolution of state religion in different
religious traditions during this period. Note that all of the major religions
experience a long period of constantly high rates of state religion until about
seventeenth century, followed by a steady decline thereafter. (The rise in the
fraction of state religion among Buddhist and Muslim nations between 1950
and 2000 corresponds to choices made by many new states, previously under
colonial or communist rule in 1950, to adopt an official religion after gaining
independence.)
To address endogeneity concerns regarding the causal relationship between
religious concentration and state secularism, the authors again exploit the
variation among countries in their geographic distance to religious capitals
of the world as an instrument. Their results show that concentration in
the religion market influenced the secularism of states negatively, again as
predicted by the economic model. However, the effect was reversed and rose
in magnitude if the ruler had a different religion than the general population,
Fig. 22.1 The decline of state religion in different religious traditions

indicating that the presence of such a difference in a monopolized market

likely had a de-legitimizing effect by facilitating resistance against the ruler.
Consequently, the net effect of religious concentration on secularism was
positive in such circumstances. The analysis shows that the historical inertia
of a state and the prevailing political regime also influenced the adoption of a
secular regime.
Merger of Religion and State: Theocracy

We next turn to the question of when religion and state will fully merge to
form a theocracy. This is partly a question of semantics—that is, when is a
religion merely a state religion, and when are religion and state sufficiently
coterminous that we refer to the merged entity as a theocracy? We will treat
this as an industrial organization question along the lines of that studied by
Coase (1937) and Grossman and Hart (1986) regarding the boundary of the
firm. In this light, we will define theocracy here to mean a situation where
the state and religion are fully merged, or integrated, as opposed to the above
discussion of a mere “alliance,” or contractual relationship, between the state
and an independent religious organization, a situation that we equate instead
to existence of a “state religion.”
In terms of the theory, the difference between the two organizational
structures is that the theocracy (the integrated state and religion) is able to
choose the quantity of the religious good, q, to maximize its overall return,
consisting of the combined revenue from taxes and the finances of the religious
organization; whereas in the case of a state religion, we assumed that the
independent religious organization chose q to maximize its own profit. This
“agency cost” represents an inherent advantage to theocracy as compared
to mere alliance. The offsetting disadvantage of integration is that religious
financing now comes under the umbrella of overall state financing, so the
religious organization’s revenue is subject to tax collection costs, which is
not true of the independent religious organization.8 The choice between
the two organizational forms depends on the effectiveness of religion as
a legitimizing force, with a theocracy being preferred as that effectiveness
becomes sufficiently large.
Surprisingly, despite the large and growing literature on the economics
of religion in general, economists have paid little attention to theocracy,
8 Specifically, we assumed that payment to the church in that case was voluntary, as in a market exchange.
with a couple of notable exceptions. In a collection of papers edited by

Ferrero and Wintrobe (2009), the contributing authors discuss various aspects
of theocracy, including the separation of powers, dictatorship, evolution of
morals, and economic development. Allen (2009), for example, interprets
theocracy as an institutional screening mechanism that facilitates cooperation
when people’s willingness to cheat on one another is unobservable. Ferrero
(2013) takes a different approach by developing a principal-agent model in
which the church (or religious authority) is the principal and a secular ruler
is the agent (thus reversing our structure). The church chooses between a
theocratic regime wherein it provides government services itself and a secular
regime wherein it contracts out government services to a ruler. The trade-off
is between corruption/incompetence under a theocracy (because the theocrat
does not have specialized knowledge required for governing) and agency costs
under secular rule. Building on the model’s implications, the author offers
evidence to explain the rise of theocracies in Ancient Israel and among Muslim
countries in the twentieth century.
Coşgel and Miceli (2013) propose a political economy approach to theoc-
racy along the lines of the framework described here and then test the
predictions of the model. Using data from historical polities and today’s
nation-states regarding the relationship between these authorities, they exam-
ine empirically why in some states religious and political authorities have
maintained independence, while in others they have chosen to integrate. In an
interesting extension of the emphasis on the legitimizing relationship between
religious and political authorities, Naghavi and Pignataro (2015) explain how
the potential consolidation of state and religion in a theocracy impacts the
outcome of economic sanctions on a regime’s policies and behavior.
State Suppression of Religion

We have assumed so far that the religious organization legitimizes the ruler,
but in some cases religious leaders may oppose the state or engage in activities
that may result in negative consequences for the ruler. In this case, religious
teachings will actually increase δ or “de-legitimize” the state. As noted above,
however, it is not necessarily the case that the state will act to suppress a
de-legitimizing religious organization. This is true, first, because suppression
would be costly, but, second, because the pacifying function of religion
increases tax capacity even when religious doctrine is in opposition to the
state. This trade-off is seen in (22.5), where religious goods increase net taxes
by raising consumer surplus, v(q) − pq, but negative teachings associated
with those goods simultaneously raise δ. The state will therefore only find
it desirable to suppress the religious organization when the net de-legitimizing
effect outweighs the cost of suppression. This shows why some moderate
religious opposition to the state is tolerated.
In this same vein, Johnson and Koyama (2013) have studied the relationship
between the centralization of state institutions and increased religious tolera-
tion in medieval and early modern Europe. Developing a model to identify
the conditions under which legal centralization raises the cost of religious
persecutions, they have offered an argument for the birth of the secular state in
Europe and supported it by historical evidence consisting of two case studies
drawn from French history. Their analysis shows that the process of legal
standardization gave rise to greater de jure religious toleration when religious
beliefs were sufficiently salient and diverse.
Religious suppression is often directed against minorities. With a historical
focus on the persecution of minorities, Anderson et al. (2017) have stud-
ied forces causing some states to persecute minorities more than others in
pre-industrial Europe and why this persecution declined during the period
between 1500 and 1800. Using panel data regarding the persecution of the
Jews in medieval and early modern Europe, they examine the way persecutions
were associated with weather shocks. They find that the probability that a
community would be persecuted increased by 1 to 1.5 percentage points (rel-
ative to a baseline of 2%) corresponding to a one standard deviation decrease
in average growing season temperature (about one-third of a degree Celsius)
in the previous five-year period. The empirical relationship was strongest in
regions controlled by weak states and with poor soil quality. The rise of more
centralized states and better integrated markets played an important role in
reducing the vulnerability of minority groups to economic shocks in the long
run. Religious persecution remains a major problem, however, even today.
Using cross-national data from the 2003 “International Religious Freedom
Reports,” Grim and Finke (2007) have found that a state’s regulation of
religion is a strong predictor of the incidence of religious persecution.
State Regulation of Technology and Economic

Decisions
The state’s calculation regarding possible suppression of religious groups can
extend to other economic choices, such as its decision of whether or not to
adopt new technologies. Indeed, the same sort of trade-off described above
potentially emerges concerning how a new technology would affect the ruler’s
net revenue, taking account of its effect on the legitimizing function of
religious leaders and other agents. Once again consider expression (22.5).
Generally, a new technology would raise the tax capacity of citizens by allowing
more efficient production (i.e., it would increase W ), but an overall assessment
of its impact would also depend on how the technology would affect the ruler’s
legitimacy. If, like adverse religious teachings, it would raise tax collection costs
(i.e., increase δ), the ruler would have to weigh that detrimental effect against
the gain in the productivity of citizens to determine the net effect on revenues.
Coşgel et al. (2012a, b) contend that this logic explains the puzzling delay of
nearly three centuries before the Ottoman Empire fully adopted moveable type
printing. They specifically argue that, despite the new technology’s obvious
advantages for commerce, the delay was based on the fact that the new
technology could have been used to undermine the legitimacy being provided
by religious leaders. In contrast, the Ottoman’s quick embrace of new military
technologies like firearms, which enhanced the government’s ability to collect
taxes, shows their general receptivity to favorable technologies.
A similar calculation confronts the state in implementing laws that regulate
other economic choices and institutions that have implications for political
legitimacy. Consider, for example, laws that restrict the charging of interest
(usury). Such restrictions prevailed in Christianity and Islam in the medieval
period and continue to form the basis for the modern financial institution
called “Islamic Banking.” To explain the origins and differential persistence
of such economic restrictions, Rubin (2011) investigates the way Christian
and Islamic states differed in the degree to which political authorities derived
legitimacy from religious authorities. He argues that interest restrictions lasted
longer in the Islamic world because of the greater degree to which the
legitimacy of political authorities depended on conforming to the dictates of
religious authorities.
State restrictions on economic choices that involve its relationship with
religion extend to laws regulating consumption choices, such as women’s
clothing and school textbooks. For example, Carvalho (2013) has studied Mus-
lim women’s choices of dress and headcovering, known as “veiling.” Obviously
a matter of private choice, veiling has nevertheless been a controversial topic
for public debate and policy in countries such as France and Turkey because of
its implications for the state’s relationship with religion. Modeling veiling as
a commitment mechanism that restricts temptations to deviate from religious
norms of behavior, Carvalho (2013) explores the way it can actually enable
women’s integration and allow them to take advantage of outside economic
opportunities. His analysis accounts for the puzzling aspects of the dramatic
and widespread rise of veiling among Muslim women since the 1970s and
the way the laws regulating veiling have affected the salience of Islam and
religiosity around the world in unexpected ways.
Similar controversies concern choices regarding schooling, such as the
effects of the rise of public schools and mass education after the eighteenth
century, and also restrictions on the choice of textbooks in public schools.
West and Woessmann (2010) study the historical resistance to public schooling
based on religious preferences and estimate the long-term effects of this resis-
tance on student achievement that emanates from rising private competition.
The textbook controversy in the United States follows from the different
standards and procedures that states use in choosing textbooks for their public
schools. Investigating these differences, Phillips (2014) uses a model to explain
the degree to which states impose limits on the selection of textbooks by local
school districts. Based on empirical analyses of textbook selection policies,
she finds a strong link between religious fundamentalism and policies at the
district and state level.
Conclusion
This chapter has offered an economic perspective on the relationship between
two of humanity’s oldest institutions, religion and state. We have specifically
sought to develop a unifying framework that allows us to organize the various
strands of research on this complex topic based on a few key principles. These
are, first, the ability of religion to enhance the ability of citizens to pay taxes
according to the so-called Marx effect, which provides an important benefit
to rulers whose primary objective is to extract resources from the population,
and second, the ability of religion to legitimize (or possibly de-legitimize) the
state as a ruling entity. We feel that these factors, when incorporated into a
rational model of organizational choice, go a long way toward explaining the
nature and variety of institutional forms that have characterized the church-
state relationship throughout history, extending up to modern times. Far from
fully resolving questions in this area, however, this approach serves instead as
a starting point for continued research into this fascinating and timely issue.
References
Allen, Douglas W. 2009. Theocracy as a Screening Device. In The Political Economy
of Theocracy, ed. Mario Ferrero and Ronald Wintrobe, 181–201. Houndmills and
New York: Palgrave Macmillan.
Anderson, Robert W., Noel D. Johnson, and Mark Koyama. 2017. Jewish Persecutions
and Weather Shocks: 1100–1800. Economic Journal 127 (602): 924–958.
Barro, Robert J., and Rachel M. McCleary. 2005. Which Countries Have State
Religions? Quarterly Journal of Economics 120 (4): 1331–1370.
Berger, Peter L. 1967. The Sacred Canopy; Elements of a Sociological Theory of Religion.
Garden City, NY: Doubleday.
Coase, Ronald. 1937. The Nature of the Firm. Economica 4 (16): 386–405.
Comparative Economics 37 (3): 402–416.
———. 2013. An Economic Theory of Theocracy. Working Paper, The University of
Connecticut.
Coşgel, Metin M., Thomas J. Miceli, and Rasha Ahmed. 2009. Law, State Power, and
Taxation in Islamic History. Journal of Economic Behavior and Organization 71 (3):
704–717.
Coşgel, Metin M., Thomas J. Miceli, and Jared Rubin. 2012a. The Political Economy
of Mass Printing: Legitimacy and Technological Change in the Ottoman Empire.
Journal of Comparative Economics 40 (3): 357–371.
———. 2012b. Political Legitimacy and Technology Adoption. Journal of Institutional
and Theoretical Economics 168 (3): 339–361.
Coşgel, Metin M., Jungbin Hwang, Thomas J. Miceli, and Sadullah Yıldırım.
forthcoming. Religiosity: Identifying the Effect of Pluralism. Journal of Economic
Behavior and Organization.
Coşgel, Metin M., Matthew Histen, Thomas J. Miceli, and Sadullah Yildirim. 2018.
State and Religion Over Time. Journal of Comparative Economics 46 (1): 20–34.
Ferrero, Mario. 2013. The Rise and Demise of Theocracy: Theory and Some Evidence.
Public Choice 156 (3–4): 723–750.
Ferrero, Mario, and Ronald Wintrobe, eds. 2009. The Political Economy of Theocracy.
Houndmills and New York: Palgrave Macmillan.
Fox, Jonathan. 2008. A World Survey of Religion and the State. Cambridge and New
York: Cambridge University Press.
Grim, B.J., and R. Finke. 2007. Religious Persecution in Cross-National Context:
Clashing Civilizations or Regulated Religious Economies? American Sociological
Review 72 (4): 633–658.
Grossman, Sanford J., and Oliver D. Hart. 1986. The Costs and Benefits of Owner-
ship: A Theory of Vertical and Lateral Integration. Journal of Political Economy 94
(4): 691–719.
Iannaccone, Laurence R. 1991. The Consequences of Religious Market Structure:

Adam Smith and the Economics of Religion. Rationality and Society 3 (2): 156–
177.
Iannaccone, Laurence R., Roger Finke, and Rodney Stark. 1997. Deregulating Reli-
gion: The Economics of Church and State. Economic Inquiry 35 (2): 350–364.
Johnson, Noel D., and Mark Koyama. 2013. Legal Centralization and the Birth of the
Secular State. Journal of Comparative Economics 41 (4): 959–978.
McCleary, R.M., and R.J. Barro. 2006. Religion and Political Economy in an
International Panel. Journal for the Scientific Study of Religion 45 (2): 149–175.
Mueller, Dennis C. 2013. The State and Religion. Review of Social Economy 71 (1):
1–19.
Naghavi, Alireza, and Giuseppe Pignataro. 2015. Theocracy and Resilience Against
Economic Sanctions. Journal of Economic Behavior and Organization 111: 1–12.
North, Charles M., and Carl R. Gwin. 2004. Religious Freedom and the Unintended
Consequences of State Religion. Southern Economic Journal 71 (1): 103–117.
Phillips, Michelle A. 2014. State Involvement in Limiting Textbook Choice by School
Districts. Public Choice 160 (1–2): 181–203.
Rubin, Jared. 2011. Institutions, the Rise of Commerce and the Persistence of Laws:
Interest Restrictions in Islam and Christianity. Economic Journal 121 (557): 1310–
1339.
Smith, Adam. 1776 [1976]. An Inquiry into the Nature and Causes of the Wealth of
Nations. Oxford: Clarendon Press.
Stark, Rodney. 2007. Discovering God: The Origins of the Great Religions and the
Evolution of Belief. New York: Harper One.
Vaubel, Roland. 2017. The Making of State Religion: Political Economy and Histor-
ical Evidence. Critical Research on Religion 5 (1): 9–33.
West, Martin R., and Ludger Woessmann. 2010. ‘Every Catholic Child in a Catholic
School’: Historical Resistance to State Schooling, Contemporary Private Compe-
tition and Student Achievement Across Countries. Economic Journal 120 (546):
F229–F255.
23
A Great Academic Re-awakening: The Return
to a Political Economy of Religion
Anthony Gill
Religion and Politics: Scholarly Blind Spots

and a Great Awakening
Ever since the dawn of civilization, human beings have been mixing religion
and politics. And irrespective of various attempts to “separate church and state”
or “secularize” society by confining matters of spiritual belief to the private
sphere, the realms of religion and politics have remained intertwined and will,
in all likelihood, continue to be. Indeed, speaking of “separate realms” may
be misleading given that people with religious convictions take their beliefs
and values into the political arena as a way of informing their choices and
behavior. Political authorities, irrespective of their own faith, must factor the
existence of people with religious beliefs. Even self-proclaimed atheistic rulers
such as Vladimir Lenin and Josef Stalin had to devise policies of how to deal
with the remnants of the Orthodox Church that they could not extinguish.
Even in North Korea, history’s most-recent paragon of secular totalitarianism,
President Kim Jong-Un must repeatedly enforce restrictions on Christians who
refuse to vanish. Religion and politics will continue to mix and it behooves
social scientists to understand why and how.
Academically speaking, the study of religion and politics has been, at best,
a minor sub-field in the discipline of political science. At worst, it has been
A. Gill ()
University of Washington, Seattle, WA, USA
e-mail: tgill@u.washington.edu

362 A. Gill
utterly ignored. While most political science departments at top research

universities have multiple professors studying Congress, ethnic conflict, or
social revolutions, finding anyone devoted to understanding the influence of
religion on politics would be difficult. And among the faculty at secular schools
who did note the presence of religion in politics, the reference often would
be to note it as some outlying curiosity that probably did not have much
importance.
More to the neglect of the study of religion and politics has been a reluctance
to use economic (or “rational choice”) concepts and theory to understand the
political economy of religion. The bulk of research on religion and politics
emanates from an ideational perspective (cf. Philpott 2001), emphasizing the
role of beliefs as a primary motivating factor for political behavior. Opinion
surveys of what religious adherents believe are de rigueur in the field (cf. Green
2007). But advancements in the field of the economics of religion have spurred
greater interest in studying the political economy of religion with a focus on
how governments regulate the spiritual marketplace and what consequences
that has for a greater range of socio-political behavior (cf. Finke 1990; Kalyvas
1996; Gill 1998; Warner 2000). While a seemingly new phenomenon, this
perspective harkens back to Adam Smith, who economists of religion hail as
their forefather, but who in reality had a comprehensive political economy
view of the spiritual marketplace. The following chapter lays out where the
study of religion and politics has been and why a political economy of religion
offers a fruitful path of future research. We will begin with a brief conceptual
framework defining the key elements within the study of religion and politics
that will help provide a road map for where we have been and where we have
yet to go.
An Evolving Understanding of Religion

and Politics, Church and State
While the study of politics dates back to the time of Aristotle, the modern
academic discipline of political science only emerged with the increasing
specialization of the social sciences during the late 1800s and early 1900s.
What follows is a general survey of how the topic of religion has figured
into our understanding of politics and government since that time with the
realization that secularization theory tended to permeate the field in general.
Most of the attention will be placed on scholars within the United States
who tended to be more interested in religion during this period than their
23 A Great Academic Re-awakening: The Return to a Political Economy… 363
European counterparts. At the time when political science departments were

being created in modern universities, Europe was seeing a stagnation of
religious practice (not to mention two major wars that disrupted everything).
As one might expect, European scholars have devoted even less attention to
religion and politics than their American counterparts. With the renewed
study of religion and politics, there have emerged two general camps into
which analyses have fallen into: an ideational perspective emphasizing the
importance of theological ideas and values and an economic school of thought
that examines the interests motivating both clergy and laity. While these
two schools of thought are not necessarily mutually exclusive—both ideas
and interests can motivate humans—they have nonetheless remained fairly
separate. We will examine each general paradigm separately with an eye toward
possible points of connection in the future.
Religion and Politics from an Ideational Perspective
Given that religions and religious organizations trade in the currency of ideas
about God and other philosophical questions, it is not surprising that those
within the political science profession would assume the most interesting
aspect of religion with respect to politics would be in the area of ideas and
values. Theologies shape our understanding of the world, including our core
values, and thus humans will take these values to the ballot box or barricades in
an effort to shape their political surroundings. Ideational perspectives, putting
beliefs at the center of human action, seem custom-made for those studying
religion and politics. The nature of their knowledge, and how it was imparted
to them via cultural institutions such as religious organizations, will determine
what choices are made and the political results thereof. It would be expected
that individuals raised in a Catholic culture will hold different views of the
world than Jews or Muslims. Those differences will be translated into the types
of political institutions and policies that they prefer, and how they go about
advocating for those institutions and policies. Changing political outcomes,
therefore, will be a function of changes in worldviews and theologies. Thus,
from this perspective a great deal of emphasis is placed upon intellectual history
as well as the stated religious preferences of individuals via survey research.
Of particular interest early on was the role that religious thought played
in the founding of the new American republic. As the first constitutional
democracy to appear in modern time, and with a great deal of interest being
paid to the rise of democracy (or its demise), a number of political scientists
and historians emphasized the role that Puritanism played in shaping the
364 A. Gill
US Constitution and America’s subsequent political culture. Earlier scholars

such as Perry Miller (1956) emphasized how Calvinism shaped the emphasis
on individual rights that permeated revolutionary rhetoric in the late 1700s,
eventually becoming codified in the US Constitution. All of this was in the
growing tension with Enlightenment philosophy and deism, and these scholars
emphasized how intellectual debates between these two worldviews shaped the
structure of government. These arguments have been furthered recently by a
number of other scholars who have argued that the institutional framework
governing the United States emanated from the religious ideas of a broad set of
constitutional architects who have gone unrecognized, such as Roger Sherman
or John Jay (Dreisbach et al. 2009; den Hartog 2015). These authors write in
contradistinction to others who claimed that the rise of secular Enlightenment
ideas combined with a denominationally neutral deism was more foundational
in determining the governmental structure of the new nation (cf. Kramnick
and Moore 1997). For all of these authors, the critical role that religion plays
comes through the ideological milieu of the time; religion influences politics
via the avenue of intellectual ideas and debate.
Another branch of the ideational perspective on religion and politics arose
with the “behavioral revolution” in political science during the mid-twentieth
century, dominated by the increasing cost-effectiveness and use of public
opinion polling. The logic here is that if we ask individuals what their political
preferences are and match them with a set of characteristics, we would be
able to find patterns in the data suggesting explanations. Taking a more
quantitative approach to research, these studies used correlation as a first step
in determining causation. Researchers would ask a battery of questions about
political issues followed by a set of “usual suspect” demographics including
age, gender, income, education, and religion. The general logic behind these
studies was that if a substantial portion of Catholics, for example, tend to
hold similar preferences for candidates or policies, and other factors are held
constant (via regression analysis), then there must be something innate to
Catholicism as a theology that prompts such views. Indeed, early survey
research dating back to the 1950s, combined with studies linking voting pat-
terns to ethno-religious precincts, did show a fairly strong connection between
denominational affiliation and political preferences, with Protestants leaning
toward the Republican Party and Catholics and Jews favoring Democrats
(Fowler 1985). These partisan affiliations have not been static, though, and
one of the more interesting findings within this literature has been a significant
political realignment among religious adherents that began manifesting itself
in the 1970s. Whereas voting and policy preferences were drawn largely along
denominational lines, religious attendance and intensity of belief are now more
closely aligned with political divisions creating a new secular-religious divide.

Atheists, agnostics, and individuals who do not attend church very often
tend to lean Democrat, whereas regular churchgoers and those that consider
religion to be an important aspect of their life favor Republicans (cf. Wald
and Calhoun-Brown 2010). In other words, Catholics who attend services
weekly are as likely to vote Republican as are regular-attending Pentecostals.
Explanations for this phenomenon usually date the divide to the controversial
Roe v. Wade Supreme Court case over abortion rights and the Democratic
Party’s embrace of a more secular platform (Green 2007). The continued
relevance of religiosity has spawned a number of other scholars to investigate
the ideological influence wielded by clergy members and religious networks
(e.g., Campbell et al. 2014; Djupe and Gilbert 2009) expanding the realm
of research in religious and political attitudes to examine how religious and
political ideas are actually formed within churches. Their research has taken
the study of religious ideas beyond mere correlation and inductive guessing and
has started to examine the underlying theoretical linkages as to how spiritual
attitudes change and can have an influence on political behavior.
While most of the early work on the connection between religious ideas and
political outcomes focused on America, a new wave of scholars have extended
this perspective into global affairs, explaining both the politics within and
between other countries. The surprise success of the 1979 Islamic Revolution
in Iran prompted political scientists to think more carefully about religion,
including some individuals who had previously looked at the world exclusively
through a secular lens. In writing about that event, Theda Skocpol, a noted
sociologist/political scientist of social revolutions from a secular-materialist
standpoint, admitted that “the networks, the social forms, and the central
myths of Shi’a Islam helped to coordinate urban mass resistance and to give
it the moral will to persist in the face of attempts at armed repression” (1982,
275). The Islamic Revolution in Iran remained an odd curiosity, though, and
the influence of religion on world affairs was not consider much beyond that
case until another famous scholar who had previously ignored religion fired
the next salvo. Samuel Huntington’s 1993 article in Foreign Affairs, entitled
“A Clash of Civilizations?,” set off a firestorm of controversy in claiming
that political conflict would be defined by relatively incompatible cultural
worldviews (or civilizations), most notably between Islam and Christianity
(1993, 31–32). While coming under intense criticism for his assertion, the
attacks of September 11, 2001, and subsequent “war on terror” (including
the current rise of the Islamic State in Syria and Iraq) may have proven
Huntington to be prophetic. Prompted further by a major research project
on the causes and consequences of the global resurgence in fundamentalist
366 A. Gill
religion spearheaded by Martin Marty and R. Scott Appleby (1996), more

scholars began examining the role that theology played in promoting social
tension.
During the mid-1990s, a group of sociologists, political scientists, and
religious studies scholars began to examine the connection between religious
fundamentalism and political violence. In examining the events of the Middle
East, Mark Juergensmeyer (1995) and James Piscatori (1994) advanced the
argument that while secular nationalism had been used to create nation-
states in Western Europe, similar attempts to do so during the latter half
of the twentieth century in Europe failed dramatically. Regimes such as
Abdel Nasser’s Egypt and the Baathist systems in Syria and Iraq were unable
to deliver on the promises of improved socio-economic conditions among
populations that were intensely religious. This gave rise to a “confrontation of
two ideologies of order,” wherein secular nationalists battled with Islamists to
define the ideological framework of governance (Juergensmeyer 1995, 379).
The Pan-Arabism of the 1960s and 1970s eventually gave way to the Pan-
Islamism of the 1980s and beyond. Juergensmeyer extended his analysis in
both his seminal article (1995) and his later book (2000) to include Hindu
and Christian fundamentalism, with the general thesis being that secular
modernization will prompt intense backlash from religious citizens who felt
left behind in the march of economic progress nationally or worldwide.
With entrenched autocrats still wielding power in many of these areas, and
globalization reminding people of their economic backwardness relative to
the West, it wasn’t surprising that the fundamentalist response to secular rule
manifested itself in political violence (cf. Barber 1996).
The observations made by Huntington, Juergensmeyer, Barber, and others
necessarily prompted more scholarly soul searching as to whether or not certain
theologies were compatible with democracy and a peaceful civil society. Not
surprisingly given where most of the violent action was, the attention was
turned toward Islam. Bernard Lewis, a prominent Princeton professor, is the
most well-known author to argue that Arab Islamic culture is difficult to square
with modern notions of liberal democracy and capitalism (Lewis 2003). His
analysis largely echoes the themes of Huntington, as is evidenced by the title of
his 2003 book What Went Wrong?, but the exact causal mechanisms between
a general religious worldview and the implementation of actual political and
economic institutions remain relatively vague.1 John Esposito and John Voll
1 Aswill be noted below, Timur Kuran (2004, 2011) and Jared Rubin (2017), who fit more squarely in
the “economic perspective” on religion, also acknowledge the roles that ideas matter and examine how
historical and contemporary Islamic ideas of economics and law can affect governing outcomes. However,
(1996), who actually picked this question up earlier, argue that there is nothing
inherent to Islam that would make it inimical to democratic governance. It
is not the theology, per se, that has led primarily to autocratic outcomes in
post-colonial Muslim states, but rather the historical legacies of imperialism
that have been hard to overcome, particularly in the Arab world (cf. Stepan
2000) or by actions of rulers who have through repressive policies politicized
Islam in a way that gives it the appearance of autocracy (Feldman 2008; Cesari
2014). Many of these scholars have noted that democracy has been relatively
successful in some Muslim nations such as Turkey, Malaysia, and Indonesia.
Recent work by González (2013) goes to a more grassroots level and notes
that Islam is not incompatible with democracy, per se, pointing out that there
are rumblings from society advocating for greater social freedoms in various
demographic pockets, including the promotion of women’s rights, though not
necessarily in the manner in which Western feminists may expect.
The study of Islam has not been the only area where scholars have been using
ideational models to explore the connection between religion and politics
around the world. Ideological battles that have been fought in the past have had
a significant impact in structuring church-state relations throughout Europe
demonstrating that religious ideas are fairly malleable and can accommodate
different forms of government. How religious and political leaders are willing
to understand the relationship between church and state in Europe has also
affected the ability to accommodate new faiths. Fetzer and Soper (2005)
argue that Britain’s ideological preference for handling religious affairs at the
local level has allowed for greater accommodation of Muslims than has the
French system which seeks to impose a national ideology of laicism uniformly
across the entire country. Kuru (2009) echoes this claim with a sophisticated
historical analysis showing how different conceptualizations of “secularism”
have led to policies that are more accommodating to religious minorities
in the United States, but less accommodating in France and, ironically, to
the majority Muslim population in Turkey. And Daniel Philpott (2001) has
extended the examination of religious ideas beyond the realm of domestic
politics to show how the Protestant Reformation intellectually shaped the
system of international relations coming out of the Peace of Westphalia. His
work has led a number of scholars to look critically at how religion and
both scholars situate the role of theologies into a more institutional context wherein different individuals
have various incentives to employ some ideas over another. Their work is probably the best example we
have of the blending of ideational and economic perspectives of religion, and they are among the most
exciting scholars to study the topic of religion.
368 A. Gill
theology matter in the realm of foreign affairs (Owen 2015) and civil and
international conflict (Toft et al. 2011).
The ideational perspective on religion and politics has blossomed in the
past decade, moving beyond simple correlational analysis of voting behavior
to look seriously how religious ideas have affected everything from political
protest to domestic governmental institutions to the structure of international
relations. The work of recent scholars such as Philpott, Owen, Hall, and Kuru
has rooted many of their analysis in a deep historical context showing how
ideas do shape the political options available to various actors. The one crucial
weakness that many of these studies offer, particularly those that work on
a very broad level (e.g., Huntington 1993; Lewis 2003), is that they never
explain clearly why certain ideas win out over others. While it might seem
logical that the best idea always wins out in a free and fair intellectual debate,
rarely is the playing field level. Individuals who have a self-interested stake in
certain ideas and have the power to coerce others to their position may often
win the debate irrespective of any intellectual merits of the arguments being
proffered. For many centuries, the Catholic Church was able to tamp down
theological dissenters either by buying them off with their own monasteries or
burning them at the stake (cf. Stark 2003). Likewise, the imposition of certain
ideas about Islamic theology in different countries, such as Saudi Arabia and
its preference for Wahhabism, may also revolve around the exercise of self-
interested power by well-situated individuals. As such, an understanding of
the progress of ideas and how they shape the world of religion and politics
is not complete without an examination of the incentives faced by different
religious actors. Enter we now into the world of political economy.
Religion and Politics from an Economic Perspective
While ideational perspectives on religion and politics have tended to reign

supreme in the political science literature, a new source of theoretical inspira-
tion began emerging in the late 1980s from an unlikely source—economics.
During the 1950s and 1960s, the economic profession itself witnessed a
broadening of its scope, away from narrow topics such as banking and
employment statistics and toward using economic theories to examine all sorts
of human behavior including criminal behavior, marriage, and religion. While
many individuals were involved in this endeavor, Nobel Laureate Gary Becker,
who pioneered the use of economic reasoning in “non-traditional” economic
topics, deserves the most credit. Becker’s theoretical approach inspired one
of his students Larry Iannaccone to make the explicit link between religion
and rational choice theory, a perspective that was also being developed at
the time by Rodney Stark, who popularized the term “religious economies”
(Stark 1983) and set in motion a stream of research that would reshape the
way scholars understood religion and, eventually, religion and politics.2 Along
with his former student Roger Finke, another early and critical player in
this new paradigm, they wrote what could be considered the seminal work
summarizing these new ideas in their book Acts of Faith: Explaining the Human
Side of Religion (2000). Stark, Finke, and Iannaccone wrote on a wide array
of topics related to religion including the decline of Catholic vocations, the
success of strict sects, the attraction of certain cults to secularists, and religious
intermarriage to name just a few of the topics they put under the microscope.3
The temporal concurrence of Stark, Finke, and Iannaccone writing and
collaborating on similar topics helped to shape a new direction in the study of
religion and politics. One of their most important contributions to this effort
came with the observation that religious pluralism and competition coincided
with greater religious practice (Stark et al. 1995; Stark 1992; Stark and
Iannaccone 1994). While new to the world of twentieth-century scholarship,
this finding actually reinforced an assertion made by none other than the
eighteenth-century economist Adam Smith, undoubtedly the first economist
of religion, who noted that when granted a religious monopoly over a citizenry,
church leaders could be found “reposing themselves on their benefices” and
became negligent in cultivating the faith of the populace (1976 [1776], 789;
see also Iannaccone 1991). It was Adam Smith, a political economist at heart,
who clearly understood that government policy conditioned the competitive
environment in which the clergy of various faiths practiced their trade, just as
butchers, bakers, and brewers did. Smith established the basis for a political
economy of religion when he wrote quite simply and clearly:
The laws concerning corn may every where be compared to the laws concerning
religion. The people feel themselves so much interested in what relates either
to their subsistence in this life, or to their happiness in a life to come, that
government must yield to their prejudices, and, in order to preserve the publick
2 See Witham (2010), particularly Chapter 6, for a colorful history of how the economics of religion became
a major paradigm in the study of religion.
3 The ideas advanced by these scholars and a handful of others in late 1980s and early 1990s represented a
Cambrian explosion of new thinking on religion, leading some scholars to call it the emergence of a new
paradigm. The amazing scope of the topics studied by these researchers is too broad to cover here. Indeed,
it would take a book itself just to document the contributions each of these individuals made. Our focus
here is instead on how advanced the football in political science by inspiring a new generation of scholars
who were willing to take the risk of deviating from the dominant stream of thinking in the profession.
370 A. Gill
[sic] tranquility, establish that system which they approve of. It is upon this
account, perhaps, that we so seldom find a reasonable system established with
regard to either of those two capital objects. (1976 [1776], 539)
Enter politics into the world of the economics of religion, wherein church and
state found it useful to use one another for their own purposes.
At the time when this small group of sociologists (Stark and Finke) and
economists (Iannaccone) were shifting the paradigm on our understanding
of religious behavior, a number of political science graduate students were
picking up the scent of these economic insights and bringing them over
to their discipline. Stathis Kalyvas (1996) integrated rational choice theory,
which underpinned the work of Iannaccone, Stark, and Finke, to explain the
rise of Christian Democratic parties in a number of countries throughout
Europe. He argued that the interests of Catholic leaders in protecting their
institutions amid growing calls for democratic governance prompted them to
become engaged in the political process by founding and supporting auxiliary
organizations that funneled lay leaders into forming political parties. While
Christian Democratic parties have distinct ideologies, it was the reactive
interests of the Catholic hierarchy that shaped the political landscape. Interests
held sway over theology.
Carolyn Warner (2000) expanded such an interest-based analysis in her
examination of the fates of the Catholic Church’s efforts to engage in partisan
politics in France, Germany, and Italy. Using a rational choice approach,
Warner explored how the Church operated as an interest group seeking to
bolster its institutional power and influence over society, but strategic choices
made during and after World War II acted as significant constraints upon the
ability of hierarchs to build political alliances. The interesting thing to note
about her innovative use of historical institutional analysis was that although
the theology of the Church was essentially constant across all three cases,
the differing outcomes in the three nations were determined much more by
cost-benefit constraints. The importance of this work stands to remind us
that the personal and institutional incentives of clergy and their followers
are theoretically similar to individuals acting in another context; the tools of
interest-based analyses can be extended to the realm of religion and politics to
great effect.
A third scholar and present author of this chapter, Anthony Gill, also
directly applied the insights of Stark, Finke, and Iannaccone to understand the
varying Catholic responses to dictatorship in Latin America during the 1970s
and 1980s. Whereas the dominant theory was that the theological changes
introduced during Vatican II and the corresponding rise of a Marxist-based
liberation theology changed the mindset of Church leaders and set them in
opposition to right-wing military dictatorships, Gill (1998) noticed that such
opposition was not uniform across the region as the ideational perspective
would predict. Some bishops adopted the rhetoric of Vatican II, while others
continued with their decades-long support of the ruling regimes. He adopted
the religious competition model of Stark and Iannaccone to argue that where
religious competition between Protestants and Catholics was most intense
(viz., Brazil and Chile), bishops needed to signal their concern for the weakest
in society by opposing the excesses of dictators and undertake a preferential
option for the poor. Indeed, the tactics of the Catholic Church at the pastoral
level were similar to the efforts of Protestant missionaries, thereby providing
additional evidence to the religious pluralism equals religious vitality thesis,
but the institutional history of the Catholic Church required an additional
political response to break with oppressive governments.
Gill further explored the micro-economic rationale behind religious com-
petition in Latin America by noting, in accordance with the predictions of
Iannaccone (1991), that Protestant competition to Catholic hegemony in
Latin America was most likely to arise in countries that had deregulated
religious markets. In other words, religious liberty gave rise to religious plural-
ism/competition, which in turn prompted greater amounts of religious activity
in society. All these works presented a significant challenge to the age-old
secularization theory that claimed the decline of religion was a demand-side
phenomenon (i.e., parishioners stopped believing). Instead, it was the ability of
organizations to supply religious services to the population that mattered more
in creating a more religiously dynamic environment, and government policy
played an important role in shaping the nature of the supply, be it monopolistic
(bad for religious faith) or pluralistic (good for it). Inspired by a seminal article
by Roger Finke (1990), Gill (2008) introduced a theory of why governments
would ever deregulate religious markets when it would be seemingly in their
interests not to. As compared to earlier theories of religious liberty that linked
greater freedoms to intellectual innovations during the Enlightenment, Gill
argued that politicians are constantly tinkering with the rules and regulations
that affect churches and that this tinkering is motivated by political ambition
and economic concerns. Whereas previous studies of religion and politics
tended to ignore the influential role that public policy had on the overall
level of religious practice and belief, Gill put this connection at center stage.4
4 While this assertion may seem a bit pompous on the part of the current author, he is extremely proud of
having made this case for two full decades amid great criticism that this type of economic analysis should
372 A. Gill
And this research has now expanded into a more vigorous examination of the
effects religious liberty—the deregulation of the religious market—has had
a number of socio-economic variables such as economic growth and overall
human flourishing (Fox 2008; Grim and Finke 2010; Gill 2013).
The political economy approach to religion has not caught on as rapidly
as rational choice approaches to religion have in sociology and economics.
However, a few younger scholars are beginning to take this perspective in new
directions by examining how Russia and China have regulated their religious
minorities (Koesel 2014), how dictatorships in general have restricted the rights
of religious minorities (Sarkissian 2015), and examining the role of religion in
ethnic conflict (Isaacs 2016). Ironically, it has been economists and sociologists
who have more readily picked up the torch of studying political influences on
religion. Fenggang Yang (2012) examines how the revival of religion in com-
munist China, a major challenge to ideational perspectives and secularization
theory in-and-of-itself, is largely a matter of how the government regulates
the religious marketplace, allowing religious groups that don’t challenge the
CCP’s hegemony to operate in a “gray market” while crushing those that
appear as a mobilizing threat against the regime. And Kollman (2013) also
examines the internal organizational dynamics of the Vatican and how it
may affect its overall political placement in the world. Finally, the innovative
work of economic historians Timur Kuran (2011) and Jared Rubin (2017)
has demonstrated how legal restrictions placed on entrepreneurs by Islamic
religious leaders have shaped the differential economic outcomes observed in
the Middle East as compared to Western Europe. Their scholarship is unique
in that both scholars show that ideas do matter, but in the context that they
can affect the economic and political incentives of various actors who have an
influence in enacting those ideas in policy.
In contrast to ideational perspectives, a more economic (interest-based)
approach has expanded our knowledge of religion and politics by putting
forth the observation that people acting in a religious context are not uniquely
different from any other situation. Clergy and laity exist within a world of
scarcity and must make difficult trade-offs in how to use resources to best
achieve their goals. Such goals might include such lofty objectives as seeking
spiritual enlightenment, but they also often involve accumulating wealth and
power, just as any organization must do. Priests, rabbis, and imams must all
contend with resource scarcity and power relations within their own religious
not be done. As such, I feel a little touchdown celebration in warranted. Consider this my spiking of the
theoretical football.
organizations and must also manage relations with political actors that seek
to manipulate these religious organizations for their own purposes. And, of
course, spiritual leaders have been known to appeal to the political powers that
are for their own objectives, often restricting religious competitors or seeking
subsidies to help them carry out their evangelical mission. This is not to say
that ideas do not matter in the grand scheme of things. Sociologists who have
made great use of an economic approach to religion, such as Rodney Stark,
have repeatedly noted that ideas do matter and have incorporated them into
their studies (cf. Stark 2003, 2007). However, political scientists have yet to
plow this field, which could yield a great harvest for any budding graduate
student so inclined to take on that challenge.
Conclusion: A Great Awakening

Political science has undergone a “great awakening” when it comes to the topic
of religion. Since 2001, not only has there been growing interest in studying the
relation between religious extremism and terrorism, but many scholars are now
aware that the influence of religious beliefs and institutions extends to all facets
of social and political life. Not only has the empirical scope of our attention
increased tenfold, but the theoretical tools used to understand the religious-
political nexus have expanded as well, encompassing the insights of economics
and taking us full circle to the prescient thinking of Adam Smith. No longer is
religion just understood as a mindset among increasingly irrelevant believers
who may occasionally bring their values to bear on political action, but we
are open to approaches that see religious actors and organizations influenced
by a whole host of incentive structures, many of which have commonalities
with other political phenomena (e.g., political parties, protest movements),
but which also may have a certain uniqueness to them. As for the latter, we
now recognize sincerely that individuals and groups may be motivated by more
than concerns over wealth and power and that the pull of worshiping one’s
God may be of equal or greater importance in many situations.
References
Barber, Benjamin. 1996. Jihad vs. McWorld: Terrorism’s Challenge to Democracy. New
York: Ballantine Books.
Campbell, David E., John C. Green, and J. Quin Monson. 2014. Seeking the Promised
Land: Mormons and American Politics. Cambridge: Cambridge University Press.
374 A. Gill
Cesari, Jocelyne. 2014. The Awakening of Muslim Democracy: Religion, Modernity, and
the State. Cambridge: Cambridge University Press.
den Hartog, Jonathan J. 2015. Patriotism and Piety: Federalist Politics and Religious
Struggle in the New American Nation. Charlottesville: University of Virginia Press.
Djupe, Paul A., and Christopher P. Gilbert. 2009. The Political Influence of Churches.
New York: Cambridge University Press.
Dreisbach, Daniel, Mark David Hall, and Jeffry H. Morrison, eds. 2009. The
Forgotten Founders on Religion and Public Life. South Bend: University of Notre
Dame Press.
Esposito, John L., and John O. Voll. 1996. Islam and Democracy. Oxford: Oxford
University Press.
Feldman, Noah. 2008. The Fall and Rise of the Islamic State. Princeton: Princeton
University Press.
Fetzer, Joel S., and J. Christopher Soper. 2005. Muslims and the State in Britain, France,
and Germany. Cambridge: Cambridge University Press.
Finke, Roger. 1990. Religious Deregulation: Origins and Consequences. Journal of
Church and State 32: 609–626.
Fowler, Robert Booth. 1985. Religion and Politics in America. Metuchen: Scarecrow
Press.
Fox, Jonathan. 2008. A World Survey of Religion and the State. Cambridge: Cambridge
University Press.
Gill, Anthony. 1998. Rendering Unto Caesar: The Catholic Church and the State in
Latin America. Chicago: University of Chicago Press.
———. 2008. The Political Origins of Religious Liberty. Cambridge: Cambridge
University Press.
———. 2013. Religious Liberty and Economic Development: Exploring the Causal
Connections. The Review of Faith and International Affairs 11: 5–23.
González, Alessandra L. 2013. Islamic Feminism in Kuwait: The Politics and Paradoxes.
New York: Palgrave Macmillan.
Green, John C. 2007. The Faith Factor: How Religion Influences the Vote. Westport,
CT: Praeger Press.
Grim, Brian, and Roger Finke. 2010. The Price of Freedom Denied: Religious Persecution
and Conflict in the Twenty-First Century. Cambridge: Cambridge University Press.
Huntington, Samuel P. 1993. The Clash of Civilizations? Foreign Affairs 72: 22–49.
Iannaccone, Laurence R. 1991. The Consequences of Religious Market Structure:
Adam Smith and the Economics of Religion. Rationality and Society 3: 156–177.
Isaacs, Matthew . 2016. Faith in Contention: Explaining the Salience of Religion in
Ethnic Conflict. Comparative Political Studies 50 (2): 200–231.
Juergensmeyer, Mark. 1995. The New Religious State. Comparative Politics 27: 379–
391.
———. 2000. Terror in the Mind of God: The Global Rise of Religious Violence.
Berkeley: University of California Press.
Kalyvas, Stathis. 1996. The Rise of Christian Democracy in Europe. Ithaca: Cornell
University Press.
Koesel, Karrie J. 2014. Religion and Authoritarianism: Cooperation, Conflict, and the
Consequences. Cambridge: Cambridge University Press.
Kollman, Ken. 2013. Perils of Centralization: Lessons from Church, State, and Corpora-
tion. Cambridge: Cambridge University Press.
Kramnick, Isaac, and Laurence Moore. 1997. The Godless Constitution: The Case
Against Religious Correctness. New York: W.W. Norton & Company.
Kuran, Timur. 2004. Islam and Mammon: The Economic Predicaments of Islamism.
———. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Kuru, Ahmet. 2009. Secularism and State Policies Towards Religion: The United States,
France, and Turkey. Cambridge: Cambridge University Press.
Lewis, Bernard. 2003. What Went Wrong? The Clash Between Islam and Modernity in
the Middle East. New York: Harper Perennial.
Marty, Martin, and R. Scott Appleby, eds. 1996. Fundamentalisms and the State:
Remaking Polities, Economies, and Militance. Chicago: University of Chicago Press.
Miller, Perry. 1956. Errand into the Wilderness. Cambridge: Belknap Press.
Owen, John M. 2015. Confronting Political Islam: Six Lessons from the West’s Past.
Philpott, Daniel. 2001. Revolutions in Sovereignty: How Ideas Shaped Modern Interna-
tional Relations. Princeton: Princeton University Press.
Piscatori, James. 1994. Accounting for Islamic Fundamentalisms. In Fundamentalisms
and the State: Remaking Polities, Economies, and Militance, ed. Martin Marty and
R. Scott Appleby. Chicago: University of Chicago Press.
Rubin, Jared. 2017. Rulers, Religion, & Riches: Why the West Got Rich and the Middle
Sarkissian, Ani. 2015. The Varieties of Religious Repression: Why Governments Restrict
Religion. Oxford: Oxford University Press.
Skocpol, Theda. 1982. Rentier State and Shi’a Islam in the Iranian Revolution. Theory
and Society 11: 265–283.
Smith, Adam. 1976 [1776]. An Inquiry into the Nature and Causes of the Wealth of
Nation. Indianapolis: Liberty Fund.
———. 1983. Religious Economies: A New Perspective. Paper Delivered at the Confer-
ence on New Directions in Religious Research, University of Lethbridge, Stepan.
———. 1992. Do Catholic Societies Really Exist? Rationality and Society 4: 261–271.
———. 2003. For the Glory of God: How Monotheism Led to Reformations, Science,
Witch-Hunts, and the End of Slavery. Princeton: Princeton University Press.
———. 2007. Discovering God: The Origins of the Great Religions and the Evolution
of Belief. New York: HarperOne.
376 A. Gill
Stark, Rodney, and Laurence R. Iannaccone. 1994. A Supply-Side Reinterpretation of

the ‘Secularization’ of Europe. Journal for the Scientific Study of Religion 33: 230–
252.
Stark, Rodney, Roger Finke, and Laurence R. Iannaccone. 1995. Pluralism and Piety:
England and Wales, 1851. Journal for the Scientific Study of Religion 34: 431–444.
Stepan, Alfred. 2000. Religion, Democracy, and the ‘Twin Tolerations’. Journal of
Democracy 11: 37–56.
Toft, Monica Duffy, Daniel Philpott, and Timothy Samuel Shah. 2011. God’s Century:
Resurgent Religion and Global Politics. New York: W.W. Norton & Company.
Wald, Kenneth D., and Allison Calhoun-Brown. 2010. Religion and Politics in the
United States. 6th ed. Lanham, MD: Rowman & Littlefield.
Warner, Carolyn M. 2000. Confessions of an Interest Group: The Catholic Church and
Political Parties in Europe. Princeton: Princeton University Press.
Witham, Larry. 2010. Marketplace of the Gods: How Economics Explains Religion.
Yang, Fenggang. 2012. Religion in China: Survival and Revival Under Communist Rule.
24
The State, Toleration, and Religious Freedom
Noel Johnson and Mark Koyama
Introduction
The relationship between religion and the state is among the most important
and contentious themes in world history. It remains important today, whether
we are considering political turmoil in the Middle East or issues such as same-
sex marriage and the veil in Western democracies. This chapter provides an
overview of the relationship between religion and the state in Europe from
antiquity to the Industrial Revolution. In so doing, we provide a novel account
of how the West got religious freedom, one that will help us better understand
the present.
Philosophers and political scientists have long recognized that the rise
of religious freedom played a crucial role in the rise of liberalism more
generally (e.g. Rawls 1993).1 Recognition of the right to worship as one pleased
was foundational for the appreciation of other important personal liberties.
Economic and political freedoms rest on freedom of conscience and freedom
of religion (see Schmidtz and Brennan 2010). However, accounts of the rise of
1 The political philosopher Chandran Kukathas provides a statement that can stand in lieu of a definition:
‘Liberalism does not care who has power; nor does it care how it is acquired. All that matters is that the
members of society are free to pursue their various ends, and that the polity is able to accommodate all
peacefully’ (Kukathas 2003, 253). See Tomasi (2012) for a similar view.
N. Johnson · M. Koyama ()
George Mason University, Fairfax, VA, USA

378 N. Johnson and M. Koyama
religious freedom have often centered on the intellectual history of religious

toleration, focusing on the writings of thinkers such as Pierre Bayle, Baruch
Spinoza, and John Locke (e.g. Jordan 1932, 1936; Lecler 1960; Zagorin 2003).
Important as these works are, such an approach does not directly address the
question of most interest to social scientists which is: what factors ensured
that it was in the interest of policy makers to move toward greater religious
freedom? In recent years historians have developed nuanced and sophisticated
approaches to this question. But they focus either on a single period like
early modern Europe (Kaplan 2007) or on a specific country such as England
(Sowerby 2013) or France (Wilson 2011). Here we provide a political economy
approach to the question of how political and economic institutions changed
in such a way so as to give rise to religious freedom. As social scientists, we
develop a conceptual framework and then take this framework to historical
evidence to see how well it can explain the outcomes that we observe.
A growing literature in the economics of religion studies the relationship
between church and state today (Iannaccone 1998; Barro and McCleary 2005;
Hungerman 2005). This literature focuses on the effects of a state religion
on religiosity. Monopoly state religion has been found to reduce religiosity
because monopoly religious providers offer an inferior product in comparison
with the competitive religious economies that emerge in the absence of a state
religion. Our concern in this chapter is with the evolution of the state and
church relationship over the long course of history from the emergence of
early states to around 1800.
In our book, Persecution and Toleration: The Long Road to Religious Freedom
(2019), we seek to build on and go beyond existing accounts of the origins
of religious freedom. This chapter summarizes and expands upon the more
detailed narrative we provide there. We are not the first to consider these
questions. Gill (2008) makes the case for the political origins of religious
liberty. Rubin (2017) and Platteau (2017) develop highly complimentary
accounts of the historical role religion played legitimating authority in the
Islamic world. Studying Colonial America, Latin America, and the Former
Soviet Union, Gill (2008) proposes a rational choice account for why some
societies have adopted religious freedom while others have not. We share
with Gill (2008) an emphasis on the incentives facing political actors and
a skepticism toward ideational accounts of the rise of religious liberty. We
go back to antiquity to provide an explanation for why religious and secular
authorities were so often such close partners and how this partnership shaped
the character of religious toleration in the premodern period.
First we outline the relationship between religion and the state in the
premodern era. We describe it as a conditional toleration equilibrium. This
conditional toleration equilibrium was characterized by an absence of religious

freedom and the occasional persecution of religious heterodoxy. We explain
how this conditional toleration equilibrium was necessitated by a reliance
on identity rules in a world of weak states. We illustrate the nature of this
equilibrium by documenting the treatment of Jews in medieval Europe and the
persecution of heretics. Next, we examine how this relationship slowly changed
in the early modern period as a result of the Reformation and technological
and military developments which led to the rise of stronger states. We
outline how movements toward religious liberty arose in the Dutch Republic,
England, and France. These developments caused the conditional toleration
equilibrium to unravel—the conflicts usually known as the Wars of Religion
were symptomatic of this unraveling. Subsequent attempts to establish states
that governed via general rules and relied on non-religious sources of legitimacy
were what enabled the emergence of religious freedom. We also consider what
this legacy meant for the establishment of religious freedom in the United
States. Finally, we conclude by considering the implications of our argument
for religious liberty and the rule of law today.
Religious Legitimation and the State

Why is religion a natural source of political legitimacy? Since the earliest
recorded history, religion has been the predominant source of legitimation.
Religion is universal in human society. It provides narratives that satisfy the
human desire for transcendence. Religious narratives can also make sense of,
and hence justify, the way the world works. Religion can support or undermine
political authority. In early history, however, the tendency was for religion
to validate existing political authority. Early religious deities were powerful
beings so that it was natural to associate worldly power as wielded by kings
and chiefs, with them (Bellah 2011, 231–232). Religion became increasingly
important in validating political authority as political structures became more
complex and stratified. Following the rise of agriculture and the invention of
writing, religion came to play a crucial role in supporting political authority
in Pharaonic Egypt, Babylon, and China.
Anthropologists and psychologists such as Joseph Henrich and Ara Noren-
zayan speculate that the rise of religions with moralizing “high gods” helped to
expand the scope of human cooperation.2 High gods can punish bad behavior
2 Shariffand Norenzayan (2007); Shariff et al. (2009); Atran and Henrich (2010); Laurin et al. (2012);
Slingerland et al. (2013); Norenzayan (2013); Purzycki et al. (2016); Norenzayan et al. (2016).
and thereby prevent individuals from deviating from cooperative ventures.

Such high gods can also validate political authority.
An important reason for the ubiquitous role of religion in validating
political authority is outlined by Coşgel and Miceli (2009); Coşgel et al.
(2012); Greif and Rubin (2015) and Rubin (2017), and this is that religious
legitimation is cheap for rulers. Governments face high costs in governing.
It is expensive to pay officials, bureaucrats, soldiers, and tax collectors. It is
still more costly to monitor them and to ensure that they are not behaving
opportunistically. Providing public goods such as education, roads, or welfare
is similarly extremely expensive, and in fact most premodern states were unable
to provide many public goods beyond defense and basic law and order (see
Johnson and Koyama 2017). In this context, reliance on religion provides rulers
with an easily available and low cost source of political legitimacy.
Religious authorities also played an important role in all premodern societies
in providing many of the public goods that states were unable to provide. The
Catholic Church in medieval Europe, for instance, provided welfare, health-
care, and education.3 As a consequence, the Church was able to influence their
congregations and encourage obedience to the political authorities. In a world
where religion is the primary source of legitimacy, the ruling coalition in any
society will be comprised of secular and religious authorities. Their position
within this coalition will depend on their relative strengths.
The need for religious legitimation had important consequences for pre-
modern polities. In return for providing legitimacy, religious authorities could
demand services in return from secular rulers. Monotheistic proselytizing
religions such as Christianity and Islam have traditionally sought converts
and limited apostasy. One common bargain, therefore, was for a sufficiently
strong secular ruler to promise to enforce religious conformity in return for
legitimacy. This bargain was usually in the interests of the secular powers
too as they did not wish competing religions to generate political instability.
This was not, however, the only bargain that could be struck. Weaker rulers
might have to obtain legitimacy from the religious authority by granting the
religious authority land or resources as occurred in early medieval Europe. As
we explain below, however, even when secular authorities did not enforce a
religious monopoly, there was no true religious freedom.
3 Religiousorganizations excel at the provision of many public goods because they have evolved institu-
tional practices such as strict rituals and rules that enable them weed out free-riders and resolve moral
hazard and adverse selection problems. See for a discussion Berman (2009). The canonical model is, of
course, Iannaccone (1992).
Naturally, the ability of a given religion to provide legitimacy to the state

varied. Rubin (2017), for instance, argues that Christianity was a less powerful
vehicle for religious legitimation than Islam because Christianity emerged as
a fringe sect and only became the religion of a powerful empire 300 years
after its founding, whereas Islam was formed as a religion that would bind
together the Arabian peninsula under a new, powerful, and expansionist state.
Platteau (2017) further emphasizes that Islam was more easily instrumentalized
by political authorities because of its decentralized character.
The ability of a state to enforce religious conformity will also depend on
the level of religious heterogeneity. A bargain in which the secular authority
persecutes religious dissent in return for legitimation from the religious
authority may be enforceable in a world where religious beliefs are fairly
homogenous. But it may become unenforceable if religious beliefs become
more heterogeneous.
Conditional Toleration and Identity Rules

The absence of religious freedom does not mean that individuals were rou-
tinely persecuted for their religious views. Rather, the situation that char-
acterized the vast majority of premodern societies was one of what we call
conditional toleration. Under conditional toleration, there are a range of
permitted and prohibited religious views. Individuals generally do not face
persecution if they do not stray beyond the bounds or conditions of toleration.
These bounds varied by society and over time. Thus the ancient Romans were
in general more tolerant than were later Christian Europeans, though they did
frequently expel “foreigner” cults, and notoriously persecuted Christians.
Our framework helps elucidate the relationship between the absence of
religious freedom and reliance on identity rules in polities reliant on religious
legitimacy. Identity rules are rules where either the form of the rule or
its enforcement depends on the social identity of the parties involved (e.g.
religion, race, or language). In contrast, impersonal rules are rules where both
the form of the rule and its enforcement are independent of the identity or
status of individuals. Identity rules are inconsistent with the enforcement of
the liberal rule of law.
Identity rules differentiate between individuals on the basis of their religion,
ethnicity, or class. Reliance on identity rules was ubiquitous throughout the
premodern world. Across Europe, for example, the nobility were often exempt
from taxation and from the most heinous forms of capital punishment. City
dwellers typically had different rights than did peasants. Jews were often
prohibited from hiring Christian servants and forced to wear badges or hats
to distinguish themselves from the Christian population. In the Middle East,
both Christians and Jews paid special taxes and proselytizing to Muslims was
punishable by death. In Japan, commoners could be executed by samurai for
the smallest perceivable slight. In India each caste had its prescribed role and
status in society.
Weak secular authorities both rely on religion as a source of political
legitimacy and on identity rules to govern. In both cases this is a reflection of
their weakness. They depend on religious sources of legitimacy because they
lack the means to provide the public goods that would provide an alternative
source of legitimation. In particular, they lack the administrative and legal
capacity to enforce general rules and to ensure legal equality so they rely instead
on the least costly form of governance—rules that leverage preexisting religious
or ethnic identities.
Identity rules were ubiquitous for a reason—they are a low cost way of
providing governance. In medieval Europe Jewish communities were often
given the authority to police their own members. Only for capital cases was
the punishment carried out by the Christian authorities. Another instance
of the role of identity rules in medieval Europe were the guilds. Forming a
layer of self-governance between the individual and the state, guilds provided
crucial functions for their members.4 Merchant guilds played an important
role in enforcing contracts and regulating long distance and overseas trade.
Craft guilds controlled who could be employed as a skilled worker in any of the
major industries including brewing, baking, tanning, smithing, construction,
glasswork, and many other more specialized occupations.5 Guilds exemplified
the importance of identity rules in medieval Europe as the benefits they
conferred were for members—outsiders, foreigners, Jews, and women were
excluded (Ogilvie 2011, 2014).
Another example of conditional toleration was the dhimmi system which
characterized premodern Islamic empires and became fully developed under
the Ottoman empire. Non-Muslims were typically not persecuted or oppressed
so long as they paid a special tax—the jizya. This tax was often collectively
imposed on Christian or Jewish communities within the empire. On the one
4 SeeHanawalt (1984); Richardson (2001, 2005); Stabel (2004).

5 Craftguilds were often based on the worship of a particular patron saint. As Richardson and McBride
(2009) document, after the Black Death, guilds came to play an important religious role in establishing
chantries that would pray for the soul of a member to speed their soul through purgatory. An important
example of an religiously enforced identity rule was the prohibition on usury as discussed in Koyama
(2010) and Johnson and Koyama (2019).
hand, Christian and Jews were legal and social inferiors to Muslims. Attempts
to proselytize to Muslims were punishable by death. On the other hand, these
religious minorities were granted considerable autonomy and self-governance,
including the right to maintain their own legal systems (Braude and Lewis
1982).6
In a world of conditional toleration, religious peace was the usual state
of affairs but religious peace was always fragile and conditional. Persecution
could and did occur, and the possibility of religiously inspired violence shaped
individuals’ lives and actions even if it was infrequent. For example, historians
often describe Islamic Spain as tolerant. It is certainly true that religious
diversity was greater in Islamic Spain than it was in Christian Europe at the
time and that religious violence was less frequent or intense. Nevertheless,
Islamic Spain was characterized by conditional toleration and persecutions
did occur. From the twelfth century onwards, Jews and Christians faced not
only discrimination and high taxes but also the periodic threat of violence and
expulsion.
We call it the conditional toleration equilibrium because it describes a
situation in which none of the constituent players had an incentive to deviate
from their actions. As such, it was self-reinforcing: as states become more
reliant on identity rules to collect taxes and administer justice, they also faced
lower incentives to invest in the fiscal and legal institutions that would lead to
higher state capacity. This, in turn, makes them more likely to rely on identity
rules and less able to enforce general rules of behavior. Low state capacity and
identity rules reinforced one another.
Identity rules are a low cost way of providing order. But reliance on identity
rules has pernicious economic consequences. Identity rules functioned because
they generated economic rents for insiders. Their cost was that this excluded
outsiders and resulted in static inefficiency as measured by deadweight loss
triangles. Identity rules also limited the scope of trade and markets. As the
scope of the market determines the extent of the division of labor, a reliance
on identity rules impeded specialization and deterred innovation, thereby
acting as a drag on the ability of an economy to generate economic growth
over time. Innovation was the key to the onset of modern economic growth.
Innovation did occur under the equilibrium of identity rules and conditional
toleration, but the incentives to innovate were subdued, and the ability for
new information to disseminate among groups of individuals was limited.
6 AsKuran (2004, 2010) has argued, this choice of legal system gave Christian and Jewish minorities
and economic advantage once trade with western Europe opened up in the eighteenth and nineteenth
centuries.
In contrast, the Industrial Revolution was driven by an upsurge in innovative

activity which occurred in eighteenth-century England, after the dismantling
of the identity rules equilibrium (Mokyr 2009).
The Treatment of Jews Under the Conditional

Toleration Equilibrium
We have argued that a reliance on religious legitimation and identity rules
formed a self-reinforcing equilibrium. The situation of the Jews in western
Europe provides a perfect example of this. Jewish communities have existed
in Europe since Roman times, and their history is extremely well documented
in comparison to any other minority groups. The treatment of the Jews, thus,
reveals the conditional toleration equilibrium in all its details.
Following St. Augustine, the Church taught that the Jews were not to be
subject to forced conversions or violence but that they were to be maintained
in an inferior state so that they could serve as “witnesses” to the errors of
their ancestors who had turned away Christ and so that their conversion could
herald his Second Coming. In this respect their situation was different from
Christian heretical groups.
Jews were often welcomed to settle in towns and cities. They obtained letters
of protection from secular rulers and in some cases the terms of settlement were
fairly good. But Jews also faced discriminatory laws which prohibited them
from carrying weapons and sometimes required that they wear distinguishing
clothes or badges. The situation of Jews in medieval Europe thus illustrates
the nature of conditional toleration. Jews were tolerated but this toleration was
conditional.
Throughout the Middle Ages, Jews had higher levels of human capital
than their Christian neighbors and excelled as merchants, doctors, and mon-
eylenders (Botticini and Eckstein 2012). This determined the nature of the
toleration that they obtained. Rulers offered Jews protection from antisemitic
violence. In return they exploited the comparative advantage of the Jews in
finance and moneylending in order to use Jewish lending to extract rents from
the blossoming commercial economy—by taxing its profits (Koyama 2010).
This comparative advantage thus became as much of a curse as a blessing for
Europe’s Jews.
In part because of the role they played in this system of fiscal extraction,
the positions of Jews became increasingly fragile over time. Pogroms and
expulsions became more common, often fueled by tropes such as accusations
of blood libels or of host desecration (Stacey 1998).
Despite promises of protection from secular rulers, persecutions and expul-

sions became increasingly common after 1300. These expulsions were some-
times carried out by medieval rulers who wanted to demonstrate their commit-
ment to Christianity or wished to expropriate the Jewish community in one
go. But violence against the Jews was also often local and driven by economic
shocks.
Anderson et al. (2017) provide evidence that the conditional toleration
equilibrium broke down under economic stress. During lean times, it became
harder for rulers to credibly uphold their promise to protect the Jews. A one
standard deviation decrease in temperature during the growing season was
associated with a 50% increase in the baseline probability of persecution.
This medieval equilibrium we have described also faced serious challenges
at the end of the Middle Ages. The Black Death represented a massive
demographic and economic shock to medieval society—about 40% of the
population died. But it was also a shock to the religious economy of medieval
Europe.7 Finley and Koyama (2018) focus on the Holy Roman empire during
the period of Black Death, which saw the worst pogroms in premodern history
(see Breuer 1988; Cohn 2007). They show that the persecution of Jews was
more violent in communities governed by bishoprics, archbishoprics, and
imperial free cities while Jews were less vulnerable in territories ruled either by
the emperor or by one of the major secular electors. They argue this was the
case because in areas where the authority of the emperor was contested by local
political authorities, rents from Jewish moneylending were more likely to be
dissipated. Rulers faced a problem of “overfishing” from the fiscal commons.
Local rulers had less of an incentive to protect Jews where these rents were
exhausted in the event of a crisis like the Black Death. Looking across all of
Europe, Jedwab et al. (2017) find that Black Death pogroms were more likely
to take place in areas with preexisting antisemitism and when the plague hit
a city during religiously sensitive months of the year such as around Easter or
after Christmas.
In summary, conditional toleration meant that Jewish communities could
survive throughout the Middle Ages and in some cases grow and flourish. But
it also left them vulnerable to growing antisemitism and to fiscal exploitation
from secular rulers (Baron 1967). Moreover, though Jewish religious culture
flourished at times, conditional toleration impeded the ability of Jewish
communities to fully contribute to medieval European culture of civilization.
7 See Richardson and McBride (2009).

The economic contribution that Jews could have made to the European
economy was also stunted by the existence of restrictions and regulations that
confined them to a small number of occupations.
The worsening position of Jews in western Europe led to a migration
to Poland-Lithuania where, initially at least, they faced less discrimination
and persecution (Weinryb 1972, 51). Nevertheless, the conditions of Jewish
settlement in eastern Europe were also characterized by conditional toleration,
and pogroms and persecutions eventually became a reoccurring feature of
Jewish life in eastern Europe as it had been in western Europe.
The Suppression of Heretics in Medieval Europe

We now use our framework to explain the persecution of religious dissenters
in medieval and premodern Europe.
When medieval states were fragmented and weak in the early Middle Ages
(c. 400 AD to 1000 AD), religious dissent was not investigated or punished
by secular authorities. During this time the institutional power of the Church
was still dispersed and political authority was weak and fragmentary. This
fragmentation in political authority was accompanied by a fragmentation in
local patterns of religiosity as cults of local saints took central stage in the
imaginations of believers. Far from being religiously uniform, early medieval
Europe was characterized by a variety of what Peter Brown terms “micro-
Christendoms” (Brown 2013, 13).
Given popular views of a benighted “dark ages”, it is perhaps surprising
that there were no heresy trials in western Europe between the end of the
Roman empire and 1022 and, even after this, no heresy trials resulted in
executions between 1022 and 1143 (Moore 1987, 13–23). Heresy was an
observer-dependent phenomenon. It required a state and Church strong
enough to encounter and investigate heterodoxy.
After the fall of the Roman empire, political authority devolved and became
local. As a consequence, the legal systems of early medieval Europe were
fragmentary and overlapping. Marc Bloch observed that “Each human group,
great or small, whether or not it occupied a clearly defined area, tended to
develop its own legal tradition”. Family law he noted might be common to an
entire region, whereas “agrarian law, on the other hand, conformed to usages,
peculiar to their community”. Moreover, “[a]mong the obligations with which
they were burdened, some, which they incurred as tenants, were fixed by the
custom of the manor whose limits did not always coincide with those of the
village’s agricultural lands” (Bloch 1961, 112). Identity rules became the norm.
And reliance on these rules, in turn, impeded the rise of stronger states.
In the centuries after the fall of the Roman empire, both secular authorities
and the Church lacked the capacity to impose religious uniformity. While
it remained ideologically wedded to religious persecution and the physical
punishment of heretics, the Church understood that secular rulers lacked
the ability to enforce religious conformity. Instead, the Church relied on
persuasion. This attitude not only applied to heretics; it also shaped the
Church’s attitudes to those who believed in paganism and in witchcraft.
The period after 1200, however, saw the rise of more powerful states.
The rediscovery of Roman Law led to the emergence of a systematic legal
system that was capable of handling contract disputes in a way that preexisting
Germanic legal system was incapable of doing (Berman 1983; Bellomo 1995).
Markets expanded and there was a tremendous expansion of cities and urban
life.
The new monarchies mobilized religious legitimacy as they sought to
build state power. In France, Philip Augustus (1179–1223) developed a “royal
religion” in order to forge a more powerful and centralized French monarchy.
He advised his successors to “honor God and the Holy Church, as I have
done. I have drawn great usefulness from his, and you will obtain just as
much” (quoted in Goff 2009, 552). The combination of reliance on religious
legitimacy with an expansion in the scale of the state meant that medieval rulers
increasingly were incentivized to defend religious orthodoxy by campaigning
against newly discovered heresies and religious dissenters—giving rise to what
Moore (1987) termed the birth of the persecuting society.
As we document in more detail in Johnson and Koyama (2013, 2019), the
French monarchy first burnished its Catholic credentials by supporting the war
declared by the Papacy against heretics in Languedoc known as the Cathars.
Traditional historical accounts view the Cathar heresy as a distinctive dualist
religion influenced by eastern heresies and by older Gnostic and Manichean
ideas.8
8 Traditional historical accounts emphasize that the Cathars believed that the devil had created the material
world and that Christ could not therefore have been born in human flesh and have suffered on the cross.
Following the writings of contemporary Cistercians chroniclers, historians saw the Cathars as developing
a church-like organization in opposition to the Catholic Church. See, for example, Oldenbourg (1961,
28–81), Lambert (1998), Hamilton (1999), and Barber (2000). This view sees the Cathars as decisively
influenced by the Bogomil heresy which arose in Macedonia in the tenth century. More recent historical
scholarship casts doubt on the existence of Catharism as a distinctive set of beliefs prior to the crusade
and the introduction of the Inquisitorial method (e.g. Pegg 2001).
Johnson and Koyama (2013) develop a formal model in which religious

legitimacy increases the ability of the state to collect taxes. The ability of
the state to obtain religious legitimation, however, depends on the state
enforcing religious homogeneity. As rulers come to govern more disparate
populations with different religious beliefs, it becomes increasingly costly to
enforce religious conformity. This model predicts that attempts by a ruler to
build state capacity and particularly to implement legal centralization can give
rise to spikes in religious persecution in the short run but can give rise to greater
religious toleration in the long run.
Johnson and Koyama (2013) use this model to explain why the expansion
of the French state “revealed” the existence of long-entrenched heretics. The
administrators and churchmen of northern France encountered a different
kind of Christianity in southern France: “Local saints and their festivals,
commemorations, and customs were cherished with corresponding fervor.
Doctrine, on the other hand, cannot have been at all clearly or precisely
disseminated or understood among lay people, and was doubtless subject to a
good deal of local variation in its expression” (Moore 2012, 120). Emmanuel Le
Roy Ladurie noted that the peasants of the region were ignorant concerning
many details of Christianity. Guillaume Baille excused his failure to fast by
noting that “I do not know what the fast days of the Church are, apart from
Lent and Friday” (Ladurie 1978, 314). Ladurie observes that “[w]hile first
communion and marriage acted as rites of passage, other sacraments seem
to have been almost unknown in the upland villages. For example, there are
no instances of confirmation. There is good reason for this: the Bishop, who
would have performed the ceremony, rarely left Pamiers and his inquisitorial
tasks, and in any case was not eager to travel among the mountainous areas of
his diocese” (Ladurie 1978, 313). Given this widespread ignorance of doctrine,
beliefs and practices naturally deviated from Catholic orthodox. Hence as
inquisitors were sent out to investigate “heresy”, they naturally “uncovered”
it among the peasantry.
The French state used the suppression of the Cathar heresy to incorporate
Languedoc into France and to strengthen monarchical power. The Church
benefited from the suppression of religious dissent and the standardization of
religious practices. The losers where those individuals in southern France who
followed the Cathar preachers—many of those were killed and many more
investigated by the newly established Roman Inquisition.9 The beliefs of the
9 The medieval inquisition was a personal office and not a permanent institution, unlike the later Spanish
and Roman Inquisitions (see Kelly 1989; Kieckhefer 1995). The inquisitors did not employ torture until
1256.
heretics found in Languedoc were not cohesive enough, and individuals could
be assimilated into the Catholic Church. Medieval heretics, moreover, lacked
the technology to widely disseminate their critique of the existing religious
order. Church and state were together able to successfully suppress religious
dissent. Heretical groups including the Waldensians, the Spiritual Franciscans,
and the Lollards were extirpated or driven underground. The bargain between
Church and state remained robust.
From Crisis to Religious Freedom: Europe After

the Reformation
The Papacy’s moral leadership and reputation was damaged by the relocation
to Avignon for much of the fourteenth century where they were seen as pawns
of the French crown and as venally corrupt. In the wake of the Black Death,
Papal authority took a further blow with the subsequent Great Schism (1378–
1417), the defeat of the last major Crusade at Nicopolis in 1396, and the
failure of the Conciliar movement to achieve Papal reforms. As a result, there
emerged heretical groups like the Lollards and Hussites. They prefigured the
Protestant Reformation in challenging the authority of the Catholic Church.
The religious equilibrium we described in section “Conditional Toleration and
Identity Rules” was thus already increasingly fragile before 1500. However,
what came after 1517 proved to be a shock on an unprecedented scale.
The Reformation was inaugurated by Martin Luther’s attack on the sale of
Papal Indulgences contained in his Ninety-Five Theses published in 1517. But
the scope of the attack on the Catholic Church and the existing order soon
went far beyond this. Violence erupted across many parts of Europe giving
rise to a series of conflicts in the sixteenth and seventeenth centuries in which
religion played an important role, including civil wars in France, the Thirty
Years’ War, the Wars of the Three Kingdoms,10 These conflicts were not solely
religious in origin, but they stemmed from the shock that the Reformation
had delivered to the medieval religious equilibrium. But what emerged from
these crises were states that were on their way to being recognizably modern.
The Reformation was a critical juncture in European history. It shattered
the religious and political equilibrium of medieval Europe. Once broken
this equilibrium could not be restored. Attempts to recreate the medieval
10 See Nexon (2009). It is largely a misnomer to call these wars “wars of religion” as their cause was typically
not solely or predominantly religious (see Cavanaugh 2009).
equilibrium, that is, to restore religious unity—such as Louis XIV’s expulsion

of the Protestants from France—were doomed to failure (Wilson 2011). Span-
ish policies to Jewish and Muslim converts were similarly self-destructive.11
There is no need to explore all of the causes and consequences of the
Reformation here.12 What matters for our account is how the Reformation
fused with other economic and political developments in Europe. Together
these developments shattered the conditional toleration equilibrium.
Several important developments accompanied the Reformation. First, the
scope and scale of warfare increased. This necessitated the rise of larger and
more powerful states with greater fiscal capacity (Parker 1976, 1988; Gennaioli
and Voth 2015; Johnson and Koyama 2017). Second, state consolidation
increased perceived, or actual, religious heterogeneity. The rise of stronger and
larger territorial states after 1500 meant that rulers governed populations with
more heterogeneous religious beliefs. But the rise of the state also led to a
greater perception, or realization, of preexisting religious heterodoxy.
Historians have documented the ignorance of most peasants when it came
to the basic fundaments of Christian belief in the Middle Ages. Absent
the legal and administrative capacity to investigate heresy, the potpourri of
half-remembered heterodox and orthodox beliefs held by most European
peasants went unnoticed by rulers and the Church. This changed after 1500.
Groups like the Waldensians who had escaped persecution by fleeing to the
mountainous Pays de Vaud region of southwestern France faced renewed and
much more effective persecution after 1500. Four thousand of them are said
to have been killed in the 1540s (Roelker 1996, 212).
Lastly, the Reformation coincided with the discovery of the Americas and
the rise of Atlantic trading networks. The opening up of a “new world”
unmentioned by either the classics or the bible set in motion intellectual
challenges to the status quo (Wooton 2015). Meanwhile the riches available
due to Atlantic trade would eventually strengthen commercial and mercantile
elites, particularly in northwest Europe (Acemoglu et al. 2005).
The Reformation reinforced and interacted with these developments: it
weakened the legitimizing powers of the Catholic Church; it increased diver-
sity of religious belief; and it strengthened secular states.
Through the medium of the printing press, reformers like Luther were able
to disseminate their critique of the Catholic Church, and this shock to the
11 See Monter (1994); Roth (1995); Ruiz (2007). For a study of the economic impact of the expulsion of
the Moriscos, see Chaney and Hornbeck (2016).
12 For a survey, see Becker et al. (2016). Recent synoptic treatments include MacCulloch (2003) and Eire
(2016).
religious beliefs of many Catholics, particularly in the German-speaking lands,

gave rise to an increase in religious diversity. The Reformation also offered an
opportunity to secular rulers who used it to seize Church lands and strengthen
secular states. The financial demands placed on states after the military
revolution obliged them to invest in greater fiscal capacity. As states gained
fiscal and administrative capacity, they became less reliant on identity rules and
increasingly capable of enforcing more general rules. Permanent bureaucracies
were established. Systems of taxation and law became more legible to use James
Scott’s phrase (1999). Finally, rulers turned to economic elites as represented
in parliaments as an alternative source of political legitimacy (Rubin 2017).
The net result was an abandonment of the identity rules which had governed
the conditional toleration equilibrium.
The Dutch Republic
Traditional historical accounts all note that it was in the Dutch Republic
that moves toward both the establishment of modern states and religious
freedom began. The Dutch Republic was the most religiously tolerant state
of the seventeenth century. But as we have noted, toleration is not the same
as religious freedom. Recent historical accounts note that “as far as religion is
concerned, even in the United Provinces state policies were not guided by the
notion that diversity was unavoidable or even desirable, but by the conviction
that concord was necessary. The existence of various religious sub-cultures was
regarded at best as an unforeseen and unfortunate result of the Reformation
and the Dutch Revolt” (van Eijinatten 2003, 3). Religious toleration was
a regrettable policy that had to be pursued because one of the results of
the Dutch Revolt was that it left the newly formed Protestant Republic in
possession of lands where Catholics remained either as a substantial minority
or in places as a local majority.13 Religious diversity increased when the
Dutch Republic offered permanent protection to Jews, with large numbers
of so-called crypto-Jews arriving in 1593. The rights of Jews to practice their
religion was codified in 1619. A split within Dutch Protestantism between
the Reformed Church and the Arminians further increased the importance
of religious toleration as means of ensuring religious peace.
But religious restrictions remained in place in the Netherlands. Pierre Bayle,
living in the Dutch Republic having had to flee France, was censored for
his Dictionnaire Historique and Critique. The Dutch equilibrium came to be
13 During the seventeenth century, Catholics comprised 35% of the population.

known as Pillarisation. Calvinists, Catholics, Lutherans, and Jews could coexist

with one another in the Dutch Republic, and the example this set for the rest
of the world did help lay the foundations for religious freedom. But the Dutch
Republic was not itself a fully religiously free society. It maintained many
aspects of the conditional toleration equilibrium. Furthermore, Dutch society
ossified and the economy stagnated, particularly after 1750 as the Dutch ceded
commercial and economic leadership to England.
England
England also only gradually progressed toward an appreciation of, first,

religious toleration and then, much later, full religious freedom. Elizabeth I (r.
1558–1603) did not favor “making windows into men’s souls”; but she believed
in making a single monopoly church compulsory for the entire population.
Thus her government founded an Anglican Church that was intended to be
broad enough to encompass both Calvinists and those who favored Catholic
practices and traditions. Recusants and dissenters were forced to conform or
pay hefty fines or face imprisonment. While this appeared to work during
her reign, it left a troubled legacy that would erupt during the reigns of her
successors.
Conflicts between the Anglican Church, crypto-Catholics, Presbyterians,
various Puritan denominations, and groups like the Quakers destabilized
English politics throughout the seventeenth century. Religious conflicts within
England were greatly exacerbated by the fact that the monarch of Anglican
England also ruled Catholic Ireland and after 1603 were also kings of Presby-
terian Scotland. From the Bishops War of 1639 which began with Charles I’s
attempt to impose the Anglian communion on Scotland to Ireland, the civil
wars that afflicted the British Isles were, in many respects, conflicts sparked and
exacerbated by religion. The Stuart kings sought to use the Church of England
to legitimate their rule, but they governed a set of kingdoms that were much
more religious diverse than the one that their medieval ancestors had ruled.
The restoration of Charles II in 1660 did not solve these problems. Though
he was personally tolerant and skeptical (to the extent that he was religious,
he leaned toward Catholicism), his Cavalier Parliament ruthlessly suppressed
Protestant dissenters. The Quakers, in particular, experienced harsh persecu-
tion: as many as 11,000 were imprisoned between 1660 and 1685 and many
fled to America (Marshall 2006). The laws against Catholics were not lessened,
and the Popish Plot of 1678 saw mass executions of Catholics for an alleged
conspiracy against the crown.
Resolution only came after 1689 with the Glorious Revolution and the Edict
of Toleration. Though this brought only partial toleration—and was in fact
less tolerant than the Toleration Act passed by James II in 1687 (see Sowerby
2013)—it produced a religious peace in England. This peace was crucial as it
allowed space for the state to develop alternative source of legitimation. Over
time, the treatment of Catholics and dissenting Protestants improved. Religion
gradually receded in prominence from political life.
Intellectual arguments for religious freedom played an important role here.
Bayle and Locke wrote influential treatises calling for greater religious tolera-
tion. But intellectual arguments do not exist in a vacuum. Bayle and Locke’s
arguments took on greater relevance in the wake of the backlash against Louis
XIV’s (1685) Revocation of the Edict of Nantes and the dispute over religious
toleration in James II’s England. Powerful as the arguments made by Bayle
and Locke were, their political influence was cemented by the failure of an
intolerant Louis XIV to dominate Europe and by the successes of states like the
Dutch Republic that were perceived to be religiously tolerant. Had Louis XIV
triumphed, then it is possible that Bishop Bossuet’s critique of Protestantism
might have been as influential as Locke’s Second Letter Concerning Toleration.14
This was not the case in part because after 1713 the French model of Catholic
absolutism was seen as decisively defeated.
Accidents of history thus played a role together with longer term economic
and political developments that meant that by the eighteenth century, the
leading European powers were on the way toward establishing modern states,
rule of law, and religious freedom. This, necessarily, entailed the abandonment
of religious legitimation and identity rules.
Even after 1700, however, progress was gradual and there were countless
setbacks. Popular opinion often favored religious repression as evidenced by
the Gordon Riots that took place in England in 1780 and saw Protestant mobs
devastate predominantly Irish and Catholic slums in London and the repeal
of the Jewish Naturalization Act of 1753 (Katz 1994). Political elites remained
more inclined to religious toleration than the population at large. Catholic
emancipation only came in 1828. Jews obtained the right to sit in Parliament
in 1858. These reforms came as part of a package which gradually dismantled
the Hanoverian warfare state and greatly liberalized the British political and
legal system.
14 Jacques-Bénigne Bossuet wrote The History of the Variations of the Protestant Churches in 1688. It
contained an influential critique of Protestantism and the tendency for Protestant movements to fragment
and a defense of an organized, monopoly church.
France
France was also racked by a civil war in the late sixteenth century in which
religious differences between Catholics and Protestants became intermingled
in a conflict between noble families over the French crown. The Protestant
Henri IV was able to bring about peace by converting to Catholicism and
signing the Edict of Nantes, which guaranteed the rights of Protestants. As in
the Dutch Republic, this measure of toleration was successful but it was seen
as a regrettable concession, agreed to only because the costs of religious conflict
were too great. Louis XIV therefore sought to re-Catholicize his country. The
condition of Protestants was made worse, and Louis and his government hoped
that they could ensure the conversion of the remaining Protestants. Finally, in
1685 he revoked the Edict of Nantes.
The Revocation was a failure. A far larger numbers of Protestant fled abroad
than had been anticipated, while the remaining Protestants laid low, many only
passing as Catholics. Converts had to pay special taxes and their behavior was
monitored. It was the responsibility of the Intendants to investigate these new
converts and to stamp out secret Protestant worship. In theory at least the
punishment for obdurate Protestants were harsh: “ranging from death or a life
term rowing the King’s galleys to confiscation of property or loss of inheritance
rights” (Adams 1991, 35). But once the attention of the French state was
distracted by the War of the Spanish Succession (1701–1714), many Protestants
returned to their old religion. Despite the Revocation, a Protestant minority
continued to exist. By the 1760s they numbered around 700,000. Although
their religion remained illegal until 1787, “administrative laxity mitigated the
severity of the laws” (Merrick 1990, 139). All in all, Louis XIV’s attempt to
turn back the clock to the Middle Ages was a dramatic failure.
The failure of all attempts to restore religious unity paved the way for
the development of states which no longer relied on religion for legitimacy
and which sought to govern by general rules rather than identity rules. In
England these developments took place gradually as we have seen. In France
they occurred in a rush at the end of the eighteenth century. In 1787 Protestants
received religious toleration and in 1791 full legal recognition.
Jews were also emancipated by the French Assembly in that year. Emanci-
pation spread across Europe with French armies during the Revolutionary and
Napoleonic wars (Berkovitz 1989, 111–114). The defeat of Napoleon in 1815,
however, saw a reversal in much of Germany (Jersch-Wenzel 1997). In cities
like Frankfurt, Jews were emancipated and then sent back into the Ghetto.
Full Jewish emancipation across all of Germany was only completed in 1871.15
In the Russian empire, it only took place in 1917.
It was therefore only gradually over the course of the nineteenth century
that the liberal ideal of a rule of law state came close to being established in
Britain, France, and Germany. Religious freedom became a key principle for
liberal states alongside the right to property and to free expression. However, as
history makes clear, as late as 1900 these freedoms were fairly recent and fragile.
The experience of Europe in the mid-twentieth century demonstrates how easy
it was to reverse this process. Jews in some parts of Germany only possessed
full rights for approximately 60 years before the National Socialists returned
Germany to a world of conditional toleration and then outright persecution
and genocide.
It is possible for liberal societies to revert to a reliance on identity rules rather
than general rules. When they do so, liberties such as religious freedom become
endangered and the risk of returning to a world of conditional toleration
increases.
North America
We now turn to developments on the other side of the Atlantic. Numerous

accounts of the rise of religious freedom give pride of place to the United States.
It is true that the American colonies saw important experiments in religious
freedom in Rhode Island and Pennsylvania, but from the long historical
perspective taken in this chapter, we view the American contribution to the
rise of religious freedom as an important but late addition to the major
institutional developments which took place in Europe.
The American colonies were founded by Puritan settlers seeking to escape
from the imposition of Anglican conformity in the Stuart England. The
Puritans themselves did not favor freedom for non-Puritans. Nevertheless,
as colonial America came to be peopled by a range of different religious
traditions (Anglicans in Virginia, Puritans in Massachusetts, Quakers in
Pennsylvania), the value of religious liberty became increasingly apparent.
Traditional accounts emphasize the influence of the writings of Locke on
America. No doubt Lockean discourse was influential, particularly in elite and
enlightened circles. But Locke himself did not argue for the disestablishment of
the Anglican Church and other facts on the ground made Americans receptive
15 Jewswere granted partial but not full civic rights in Frankfurt in 1824 (Adler 1960). The majority of
German Jews were effectively emancipated by the 1850s, but in Bavaria emancipation was only granted
in 1871.
to Lockean themes. Rather, as Gill argues, it seems that “the primary factors
driving the movement to deregulate religion were immigration, trade, internal
migration, and the continued growth of pluralism (due to the difficulties in
enforcing conformity), which meant a rise in new constituencies demanding
tax relief from general religious assessment” (Gill 2008, 91).
The American experience is consistent with our argument for the rise
of religious freedom in Europe and with the model developed in Johnson
and Koyama (2013). The main justification for religious freedom in the
American context was the desire to protect the church from the state rather
than to protect the state from religious influence. Fear of state intervention
in religion made religious freedom increasingly attractive in late eighteenth-
century America. In this the American settlers were the inheritors of a long
European evolution of economic and political institutions. The authors of
the Constitution wanted an institutional framework that would preserve both
existing liberties and at the same time ensure that the state was strong enough
to ward off foreign powers (Hamilton et al. 1788/2004).
Placed in its appropriate historical context, what led to the First Amend-
ment was, less the secularism of Thomas Jefferson, but rather the realization
that the creation of a powerful federal government posed renewed challenges
to local religious establishments across America. For Feldman the “motivating
political reality that pushed liberty of conscience onto state and then federal
agendas: the sudden increase in religious diversity that resulted from bring-
ing the states together into federal union” (Feldman 2005, 26). “Religious
diversity”, he notes, “drove this push for a constitutional amendment on
religious liberty. The new form of government under consideration was
intended to bind together the states into a union that was more complete—
‘more perfect’—than under the Articles of Confederation. The resulting
bound-together union would contain a degree of religious diversity much
greater than existed in any of the several states. Under these conditions,
various religious groups worried about the possibility—unlikely, to be sure—
of the federal government coming under the control of some other particular
denomination” (Feldman 2005, 43–44). Religious diversity between states
both made a national established church impossible and made an explicit
statement of religious separation from the state desirable. The guarantee
of religious freedom therefore accompanied the birth of a modern state in
North America. The establishment of a modern state which was increasingly
committed to the enforcement of general rules meant that established churches
would wither away in the first half of the nineteenth century, and by the end
of the nineteenth century, America would come to embrace its commitment
to religious freedom as a core part of its political system.
Concluding Comments
As the rise of religious freedom was a crucial stepping stone to the rise of
liberalism in general, our argument has implications for our understanding
of how states which respect the rule of law emerged in Europe during the early
modern period.
North et al. (2009, 2) describe the rise of modernity as the transition
from natural states to open access orders. Natural states limit access in order
to generate rents which in turn are used to create political order. Limited
access orders, as ongoing work by John Wallis makes clear, rely on personal
enforcement and on identity rules—rules where either the form of the rule
or its enforcement depends on the social identity of the parties involved (e.g.
religion, race, or language). In contrast, impersonal rules are rules where both
the form of the rule and its enforcement are independent of the identity or
status of individuals. Identity rules are incompatible with the generality norms
associated with the rule of law and political liberalism. Open access orders, in
contrast, do not depend on personal enforcement or on identity rules.
According to this thesis: the development of the modern state rested on
this transition to a political order that was no longer reliant on personal
enforcement and on identity rules. Missing from North et al. (2009), however,
is a recognition of the central role played by religion in premodern society. Reli-
gion, as we have argued, was crucial to political legitimacy in the premodern
world and to the maintenance of political order.
Rules based on religious identity played a vital role in maintaining order
in Europe for many centuries. Separating individuals into different legal
categories was a means of ensuring peace, but it prevented individuals from
reaping the benefits that come from trading and sharing ideas across religious
boundaries.
Once we recognize the ubiquity of identity rules, we can understand why
religious freedom is such a recent historical development. It is the product
of modernity. Modernity means the possibility of sustained economic growth
and the existence of technologies capable of disseminating information rapidly.
But modernity also means modern states—states capable of enforcing general
rules.16 Modern states offer a threat to liberalism because they have the
power to both crush individual liberties and to suppress intermediate groups
and organizations that are independent of the state. But modern states also
16 Of course other aspects of modern societal and economic organization such as the rapid dissemination
of information, greater social fluidity, and sophisticated division of labor also played a crucial role.
offer perhaps the only way to ensure that individual rights and freedoms are
enforced. Liberal states have to strike a balance between recognizing private
individuals and organizations that are separate from the state and ensuring
that individuals are treated equally (see Levy 2015).
We have offered a bird’s-eye view of the history of religion and the state
in western Europe. We have documented the age-old reliance of political
authorities on religion as a source of political legitimacy. In western Europe
after the Fall of the Roman empire, this relationship gave rise to a particular
medieval equilibrium, according to which the Church gave the state religious
legitimacy and administrative services in return for land, power, and the
enforcement of religious homogeneity.
This conditional toleration equilibrium governed the religious affairs of
Europe for more than a millennia. It is still in place in many parts of the
world including the Middle East. Rulers ruled on the basis of identity rules:
different rules applied to different religious and ethnic groups. These identity
rules generated economic rents that helped rulers maintain political power. On
occasion reliance on identity rules resulted in large-scale religious violence, but
in normal times it helped bring about peace and a form of de facto religious
toleration. Nevertheless, it was incompatible with religious freedom and the
liberal rule of law and retarded economic development.
After 1500 successive shocks disrupted the conditional toleration equilib-
rium. The religious changes brought about by the Reformation interacted
with changes in military technology to bring about the rise of larger and more
powerful states that were less reliant on religion for political legitimation. As
states built their own apparatus for the collection of taxes and the enforcement
of laws, they were forced to abandon identity rules and became increasing
reliant on more general rules of behavior. This process was a gradual one. It
was only the failure of attempts to restore the medieval equilibrium that gave
rise to experiments in religious toleration. These were envisioned as temporary
and as regrettable acknowledgements of political reality, but they opened the
eyes of policy makers to the possibility of a different form of government—one
dependent upon secular rather than religious justifications for rule, capable of
governing via general rules and subject to the rule of law.
References
Acemoglu, Daron, Simon Johnson, and James Robinson. 2005. The Rise of Europe:
Atlantic Trade, Institutional Change, and Economic Growth. American Economic
Review 95 (3): 546–579.
Adams, Geoffrey. 1991. The Huguenots and French Opinion, 1685–1787: The Enlighen-
ment Debate on Toleration. Toronto: Wilfrid Laurier University Press.
Adler, H.G. 1960. The Jews in Germany. Notre Dame: University of Notre Dame
Press.
Anderson, R. Warren, Noel D. Johnson, and Mark Koyama, Jewish Persecutions and
Weather Shocks 1100–1800. Economic Journal 127 (602): 924–958.
Atran, Scott, and Joseph Henrich. 2010. The Evolution of Religion: How Cognitive
By-Products, Adaptive Learning Heuristics, Ritual Displays, and Group Compe-
tition Generate Deep Commitments to Prosocial Religions. Biological Theory 5:
1–13.
Barber, Malcome. 2000. The Cathars. Harlow: Longman.
Baron, Salo W. 1967. A Social and Religious History of the Jews, Vol. XI: Citizen or
Alien Conjurer. New York: Columbia University Press.
Barro, Robert J., and Rachel M. McCleary. 2005. Which Countries Have State
Religions? The Quarterly Journal of Economics 120 (4): 1331–1370.
Becker, Sascha O., Steven Pfaff, and Jared Rubin. 2016. Causes and consequences of
Bellah, Robert N. 2011. Religion in Human Evolution. Cambridge: Harvard University
Press.
Bellomo, Manlio. 1995. The Common Legal Past of Europe. Washington: The Catholic
University of America Press.
Berkovitz, Jay A. 1989. The Shaping of Jewish Identity in Nineteenth-Century France.
Detroit: Wayne State University Press.
Berman, Eli. 2009. Radical, Religious, and Violent. Cambridge: MIT Press.
Berman, Harold J. 1983. Law and Revolution, the Formation of the Western Legal
Tradition. Cambridge: Harvard University Press.
Bloch, Marc. 1961. Feudal Society. Vol. 1. Chicago: University of Chicago Press.
Translated by L.A. Mayon.
Botticini, Maristella, and Zvi Eckstein. 2012. The Chosen Few. Princeton: Princeton
University Press.
Braude, Benjamin, and Bernard Lewis. 1982. Introduction. In Christians and Jews in
the Ottoman Empire, ed. Benjamin Braude and Bernard Lewis, vol. I, 1–37. New
York: Holmes & Meier.
Breuer, Mordechai. 1988. The “Black Death” and Antisemitism. In Antisemitism
Through the Ages, ed. Shmuel Almog, Vidal, 139–151. Oxford: Sassoon Interna-
tional Center for the Study of Antisemitism, Pergamon Press.
Brown, Peter. 2013. The Rise of Western Christendom, the tenth anniversary revised ed.
Malden: John Wiley & Sons.
Cavanaugh, William T. 2009. The Myth of Religious Violence. Oxford: Oxford
University Press.
Chaney, Eric, and Richard Hornbeck. 2016. Economic Growth in the Malthusian
Era: Evidence from the 1609 Spanish Expulsion of the Moriscos. Economic Journal
126 (594): 1404–1440.
Cohn, Samuel K. 2007. The Black Death and the Burning of Jews. Past and Present
196 (1): 3–36.
Comparative Economics 37 (3): 402–416.
Coşgel, Metin M., Thomas J. Miceli, and Jared Rubin. 2012. The Political Economy
of Mass Printing: Legitimacy and Technological Change in the Ottoman Empire.
Eire, Carlos. 2016. Reformations. New Haven: Yale University Press.
Feldman, Noah. 2005. Divided by God. New York: Farrar, Straus and Giroux.
Finley, Theresa, and Mark Koyama. 2018. Plague, Politics, and Pogroms: The Black
Death, Rule of Law, and the Persecution of Jews in the Holy Roman Empire.
Journal of Law and Economics 61 (2): 254–277.
Gennaioli, Nicola, and Hans-Joachim Voth. 2015. State Capacity and Military
Conflict. Review of Economic Studies 82 (4): 1409–1448.
Gill, Anthony. 2008. The Political Origins of Religious Liberty. Cambridge: Cambridge
University Press.
Goff, Jacques Le. 2009. Saint Louis. Notre Dame: University of Notre Dame Press.
Translated by Gareth Evan Gollrad.
Greif, Avner, and Jared Rubin. 2015. Endogenous Political Legitimacy: The English
Reformation and the Institutional Foundations of Limited Government. Memo.
Hamilton, Alexander, James Madison, and John Jay. 1788/2004. The Federalist. New
York: Barnes & Noble Books.
Hamilton, Bernard. 1999. The Albigensian Crusade and Heresy. In New Cambridge
Medieval History, ed. David Abulafia, vol. 5, 164–182. Cambridge: Cambridge
University Press.
Hanawalt, Barbara A. 1984. Keepers of the Light: Late Medieval English Parish
Guilds. Journal of Medieval and Renaissance Studies 14 (1): 21–37.
Hungerman, Daniel M. 2005. Are Church and State Substitutes? Evidence from the
1996 Welfare Reform. Journal of Public Economics 89 (11–12): 2245–2267.
Literature 36 (3): 1465–1495.
Jedwab, Remi, Noel D. Johnson, and Mark Koyama. 2017. Negative Shocks and Mass
Persecutions: Evidence from the Black Death. Mimeo.
Jersch-Wenzel, Stefi. 1997. Legal Status and Emancipation. In German-Jewish History
in Modern Times, ed. Michael A. Meyer, vol. 2, 5–49. New York: Columbia
University Press.
Johnson, Noel D., and Mark Koyama. 2013. Legal Centralization and the Birth of the
Secular State. Journal of Comparative Economics 41 (4): 959–978.
———. 2017. States and economic growth: Capacity and constraints. Explorations in
Economic History 64: 1–20.
———. 2019. Persecution and Toleration: The Long Road to Religious Freedom.
Jordan, W.K. 1932. The Development of Religious Toleration in England. Vol. I.
———. 1936. The Development of Religious Toleration in England. Vol. II. Cambridge:
Kaplan, Benjamin. 2007. Divided by Faith, Cambridge: Harvard University Press.
Katz, David S. 1994. The Jews in the History of England, 1485–1850. Oxford: Claren-
don Press.
Kelly, Henry Ansgar. 1994. Inquisition and the Persecution of Heresy: Misconceptions
and Abuses. Church History 58 (4): 439–451.
Kieckhefer, Richard. 1995. The Office of Inquisition and Medieval Heresy: The
Transition from Personal to Institutional Jurisdiction. The Journal of Ecclesiastical
History 46 (1): 36–61.
Koyama, Mark. 2010. The Political Economy of Expulsion: The Regulation of
Jewish Moneylending in Medieval England. Constitutional Political Economy 21 (4):
374–406.
Kukathas, Chandran. 2003. The Liberal Archipelago: A Theory of Diversity and
Freedom. Oxford: Oxford University Press.
Kuran, Timur. 2004. The Economic Ascent of the Middle East’s Religious Minorities:
The Role of Islamic Legal Pluralism. Journal of Legal Studies 33: 475–515.
———. 2010. The Long Divergence. Princeton: Princeton University Press.
Ladurie, Emmanuel Le Roy. 1978. Montaillou, London: Scolar Press. Translated by
Barbara Bray.
Lambert, Malcolm. 1998. The Cathars. Oxford: Basil Blackwell.
Laurin, K., A. F. Shariff, J. Henrich, and A. C. Kay. 2012. Outsourcing Punishment
to God: Beliefs in Divine Control Reduce Earthly Punishment. Proceedings of the
Royal Society B-Biological Science 79: 3272–3281.
Lecler, Joseph. 1960. Toleration and The Reformation. Vol. II. New York: Association
Press.
Levy, Jacob. 2015. Rationalism, Pluralism, and Freedom. Oxford: Oxford University
Press.
MacCulloch, Diarmaid. 2003. Reformation, Europe’s House Divided 1490–1700. Lon-
don: Allen Lane.
Marshall, John. 2006. John Locke, Toleration and Early Enlightenment Culture. Cam-
bridge: Cambridge University Press.
Merrick, Jeffrey W. 1990. The Desacralization of the French Monarchy in the Eighteenth
Century. Baton Rouge: Louisiana State University Press.
Mokyr, Joel. 2009. The Enlightened Economy. New Haven: Yale University Press.
Monter, E. William. 1994. The Death of Coexistence: Jews and Moslems in Christian
Spain, 1480–1502. In The Expulsion of the Jews: 1492 and After, ed. Raymond B.
Waddington and Arthur H. Williamson, 1–5. New York: Garland Publishing.
Moore, Robert I. 1987. The Formation of a Persecuting Society: Power and Deviance in
Western Europe, 950–1250. Oxford: Basil Blackwell.
———. 2012. The War on Heresy. London: Profile Books Ltd.
Nexon, Daniel H. 2009. The Struggle for Power in Early Modern Europe. Princeton:
Norenzayan, Ara. 2013. Big Gods: How Religion Transformed Cooperation and Conflict.
Norenzayan, Ara, Azim F. Shariff, Will M. Gervais, Aiyana K. Willard, Rita A. McNa-
mara, Edward Slingerland, and Joseph Henrich. 2016. The Cultural Evolution of
Prosocial Religions. Behavioral and Brain Sciences 39: 001.
North, Douglass C., John Joseph Wallis, and Barry R. Weingast. 2009. Violence and
Social Orders: A Conceptual Framework for Interpreting Recorded Human History.
Ogilvie, Sheilagh. 2011. Institutions and European Trade: Merchant Guilds, 1000–1800.
———. 2014. The Economics of Guilds. Journal of Economic Perspectives Fall 28 (4):
169–92.
Oldenbourg, Zoe. 1961. Massacre at Montségur. London: Phoenix Press.
Parker, Geoffrey. 1976. The “Military Revolution”, 1560-1660–a Myth? The Journal
of Modern History 48 (2): 195–214.
———. 1988. The Military Revolution: Military Innovation and the Rise of the West,
1500–1800. Cambridge: Cambridge University Press.
Pegg, Mark Gregory. 2001. The Corruption of Angels, Princeton: Princeton University
Press.
Platteau, Jean-Philippe. 2017. Islam Instrumentalized. Cambridge: Cambridge Uni-
versity Press.
Purzycki, Benjamin Grant, Coren Apicella, Quentin D. Atkinson, Emma Cohen,
Rita Anne McNamara, Aiyana K. Willard, Dimitris Xygalatas, Ara Norenzayan,
and Joseph Henrich. 2016. Moralistic Gods, Supernatural Punishment and the
Expansion of Human Sociality. Nature 530 (7590): 327–330.
Rawls, John. 1993. Political Liberalism. New York: Columbia University Press.
Richardson, Gary. 2001. A Tale of Two Theories: Monopolies and Craft Guilds in
Medieval England and Modern Imagination. Journal of the History of Economic
Thought 23 (2): 217–242.
———. 2005. Craft Guilds and Christianity in Late-Medieval England: A Rational
Choice Analysis. Rationality and Society 17: 139–189.
Richardson, Gary, and Michael McBride. 2009. Religion, Longevity, and Coopera-
tion: The Case of the Craft Guild. Journal of Economic Behavior and Organization
71 (2): 172–186.
Roelker, Nancy Lyman. 1996. One King, One Faith. Berkeley: University of California
Press.
Roth, Norman. 1995. Conversos, Inquisition, and the Expulsion of the Jews from Spain.
Madison: The University of Wisconsin Press.
Rubin, Jared. 2017. Rulers, Religion, and Riches Why the West Got Rich and the Middle
Ruiz, Teofilo F. 2007. Spain’s Centuries of Crisis: 1300 - 1474. Oxford: Basil Blackwell.
Schmidtz, David and Jason Brennan. 2010. A Brief History of Liberty. Oxford: Wiley-
Blackwell.
Scott, James C. 1999. Seeing like a state: How Certain Schemes to Improve the Human
Condition Have Failed. The Institution for Social and Policy St Series. New Haven:
Yale University Press.
Shariff, A., A. Norenzayan, and J. Henrich. 2009. The Birth of High Gods: How
the Cultural Evolution of Supernatural Policing Agents Influenced the Emergence of
Complex, Cooperative Human Societies, Paving the Way for Civilization. Mahwah:
Lawrence Erlbaum Associates.
Shariff, Azim F., and Ara Norenzayan. 2007. God Is Watching You. Psychological
Science 18 (9): 803–809.
Slingerland, Edward, Joseph Henrich, and Ara Norenzayan. 2013. The Evolution
of Prosocial Religions. In Cultural Evolution: Society, Technology, Language and
Religion, ed. Peter J Richerson and M. H. Christiansen. Cambridge: MIT Press.
Sowerby, Scott. 2013. Making Toleration: The Repealers and the Glorious Revolution.
Stabel, Peter. 2004. Guilds in Late Medieval Landers: Myths and Realities of Guild
Life in an Export–Oriented Environment. Journal of Medieval History 30: 187–212.
Stacey, Robert C. 1988. From Ritual Crucifixion to Host Desecration: Jews and the
Body of Christ. Jewish History 12 (1): 11–28.
Tomasi, John. 2012. Free Market Fairness. Princeton: Princeton University Press.
van Eijinatten, Joris. 2003. Liberty and Concord in the United Provinces. Leiden: Brill.
Weinryb, Bernard D. 1972. The Jews of Poland. Philadelphia: Jewish Publication
Society of America.
Wilson, Christie Sample. 2011. Beyond Belief: Surviving the Revocation of the Edict of
Nantes in France. Bethlehem: Lehigh University Press.
Wooton, David. 2015. The Invention of Science. London: Allan Lane.
Zagorin, Perez. 2003. How the Idea of Religious Toleration Came to the West. Princeton:
25
Religiosity and Economic Performance:
The Role of Personal Liberties
Laura Mayoral and Joan Esteban
JEL Classification Z12, J22
Introduction
What is the effect of religion on economic outcomes? Religion is a complex
social phenomenon affecting individual choice and behavior through numer-
ous channels. This contribution summarizes some of these effects. We focus
on the causality direction from religiosity to economic performance, therefore
Laura Mayoral and Joan Esteban gratefully acknowledge the financial support from the Spanish Ministry
of Economy and Competitiveness Grant, through the Severo Ochoa Programme for Centres of Excellence
in R&D (SEV-2015-0563) and grant number ECO2015-66883-P, Generalitat de Catalunya project
number 2017SGR1359, and the National Science Foundation Grant SES-1629370.
L. Mayoral
Institut d’Anàlisi Económica, CSIC and Barcelona GSE, Barcelona, Spain
University of Gothenburg, Gothenburg, Sweden
e-mail: laura.mayoral@iae.csic.es
J. Esteban ()
Institut d’Anàlisi Económica, CSIC and Barcelona GSE, Barcelona, Spain
e-mail: joan.esteban@iae.csic.es

406 L. Mayoral and J. Esteban
skipping the potential influence that the economic environment and personal
realizations can have on religious beliefs.1
The economics literature has emphasized several mechanisms that relate
religion and economic outcomes in different directions. We shall focus on
three dimensions: (1) religion has an [ambivalent] effect on human capital
formation; (2) religion fosters a set of attitudes that are beneficial for growth;
and (3) religious values may be detrimental for economic development as
they might hinder innovation as well as reduce incentives for individual
effort. We shall discuss these three issues in section “Religiosity and Economic
Outcomes”.
Sections “Constraints on Personal Liberties” and “Religiosity, Personal
Liberties, and Labor Supply” focus on an alternative mechanism, recently
introduced by Esteban et al. (2018a, ELMa hereinafter), through which
religion can also be detrimental for economic outcomes, notably for individual
labor supply. Religions typically impose on their affiliates a code of conduct
that constrains their behavior in a number of areas: this includes issues
such as divorce, women’s rights, abortion, gay sexuality, marriage, euthanasia,
and so on. Religious institutions have successfully managed to influence the
state so that legal caps on those issues were established, implying that both
religious and secular individuals have been effectively constrained by those
laws. However, over the past 50 years many countries have significantly
changed such regulations, giving rise to a rights revolution (Hitchcock et al.
2012). Section “Constraints on Personal Liberties” quantifies the extent of
the changes in regulation of personal liberties in Europe from 1960 to the
present. These changes are likely to have affected in a different way religious
and secular individuals: while the latter now face a more lax environment, the
religious are still constrained by their own religion’s code of conduct. Section
“Religiosity, Personal Liberties, and Labor Supply” analyzes the impact of the
expansion of the legal cap on liberties on individual effort and argues that the
former increase (contracts) labor supply among secular (religious) individuals.
Section “Religiosity, Liberties, and Voting” examines the implications of
these results on voting over redistribution and over the expansion of the
cap on liberties (Esteban et al. 2018b, ELMb henceforth). Finally, section
1 This direction of causality has been the focus of numerous theories. Of particular relevance is the
secularization hypothesis that states that economic growth decreases religious beliefs and participation
in religious services and reduces the influence of religious institutions on politics and governance. Marx
(1843, 1976) viewed religion as resulting from poverty, the opiate of the have-nots. Solt et al. (2011) have
obtained empirical evidence supporting Marx’s view: inequality breeds religiosity.
25 Religiosity and Economic Performance: The Role of Personal Liberties 407
“Concluding Remarks” concludes by pointing at potentially fruitful new lines

of research.
Religiosity and Economic Outcomes

Several strands of research have stressed the positive effects of religion on
economic development. A first channel highlights the contribution of religion
to human capital formation. Access to schooling depends on the availability of
educational infrastructure and, in many countries, that infrastructure is based
on facilities originally developed by religious organizations to promote educa-
tion and spread the faith. Becker and Woessmann (2009) argue that Protestant
Reformation promoted literacy because of its contention that everyone needed
to read the Bible and show that Protestants’ higher literacy accounts for the
whole gap in economic prosperity between Protestant and Catholic regions.
A similar argument has been made with respect to Judaism (Botticini and
Eckstein 2012). Other authors have highlighted the role of missionaries in
spreading mass education in Africa or America (Nunn 2010, 2014 and Valencia
Caicedo 2018). The universal validity of this channel, however, has been
contested. More recent evidence points toward the opposite direction: higher
family religiosity is linked to poorer education performance and lower human
capital. For instance, Darnell and Sherkat (1997) and Sherkat and Darnell
(1999) document a strong negative effect of fundamentalism on educational
achievement, net of other social background variables, and Berman (2000)
uses a “club good” model to explain the very low labor supply by ultra-
Orthodox Jews.
A second mechanism linking religion and economic outcomes argues that
religion fosters a set of attitudes that are beneficial for growth. This argument
was pioneered by Weber (1904), whose central thesis is that Protestantism
promotes attitudes such as work ethic and honesty that contributed to the
success of the capitalism.2 Guiso et al. (2003) have shown that more religious
people are more trusting (of other people and of public institutions and
market outcomes), as well as more trustworthy (less willing to break the law).
Using cross-country regressions, Barro and McCleary (2003) and McCleary
2 This hypothesis has been examined in numerous papers (see Grier 1997, Ekelund et al. 2006, and
Delacroix and Nielsen 2001) obtaining mixed results. More recently, Cantoni (2015) has used a “natural
experiment”—the forced imposition of religious denominations as a consequence of the Peace of Augsburg
(1555)—to investigate the differences in growth patterns across the Protestant and Catholic parts of the
Holy Roman Empire, finding no positive effect of Protestantism on economic growth over the very long
run (1300–1900).
and Barro (2006) find that for a given level of church attendance, increases
in religious beliefs (particularly belief in heaven, hell, and an afterlife) have
a positive impact on economic growth.3 From a theoretical perspective,
Bénabou and Tirole (2006, 2011) and Levy and Razin (2012) have argued that
beliefs in divine rewards and punishments can induce individuals to behave
less opportunistically and more cooperatively.
A third mechanism points toward a set of values and beliefs associated to
religion that, contrary to the previous channel, can be detrimental to economic
progress. Christianity and science often have come into conflict with each
other. Bénabou et al. (2015a) argue that a key pillar of long-run economic
growth is innovation and have shown that the more religious people are, in
general, the less inclined to it. Bénabou et al. (2015b) also show that there is
a negative correlation between religiosity and patents per capita using both
international and US data. Barro and McCleary (2003) and McCleary and
Barro (2006) provide evidence showing a negative correlation between church
attendance and economic growth and income. Campante and Yanagizawa-
Drott (2015) exploit exogenous variation in religious practices—differences in
the length of the Ramadan fasting period in the north and south hemisphere
due to the rotating Islamic calendar—and find that Ramadan fasting has a
robust negative effect on output growth in Muslim countries. They also report
evidence showing that a higher exposure to religious practices has an impact
on individual preferences that leads to a decrease in labor supply beyond the
month of Ramadan itself, which in turn leads to a decrease in economic
performance. The result that religiosity is associated with lower effort or labor
supply has also been attested by abundant theoretical and empirical literature.4
ELMa have proposed an alternative channel through which religion can be
detrimental for economic outcomes, notably for individual labor supply. As
3 The findings by Barro and McCleary are based on cross-country regressions and they are subject to
numerous criticisms. See Durlauf et al. (2012) for a critique.
4 The mechanism explored in the literature is typically based on the assumption that religious individuals
value relatively more religious than economic activities. This takes different specifications. Iannaccone
(1992) initiated a stream of papers in which this lower relative valuation of material benefits is introduced
by assuming that the utility from consumption of material goods is subject to some sort of “psychological
tax”. This makes individuals to devote more time and effort to religious activities. A complementary
explanation, also due to Iannaccone (1992, 1998), follows from the view of religious practice as a problem
of demand for a “club good”. In this case, members of a religious club experience an externality because
their individual enjoyment depends on the intensity of commitment of the other members. Hence, it
may be in the interest of the club members to impose highly restrictive rules of behavior in order to leave
out the individuals with lower levels of religiosity. These restrictions would be geared toward making
more attractive the allocation of time to religious activities, at the expense of productive activities. Besides
Iannaccone, Berman (2000) and Carvalho and Koyama (2016), among others, also argue that religions
strategically induce individuals to participate in costly rituals and hence reduce their material productivity.
mentioned in the “Introduction”, religions impose on their members a number

of constraints on their behavior: this includes what to eat, what to drink, how
to dress, the use of free time, gender roles, the possibility of divorce, abortion,
gay sexuality, marriage, and so on. Many of these aspects have typically
been regulated by laws that restrict their use for both religious and non-
religious individuals. Over the past 50 years, however, many governments have
significantly changed such laws: women’s rights to education or employment
and the right over their bodies had expanded, along with gay rights and
individuals’ rights over ending their life. These changes in the regulations are
likely to have affected in a different way the restrictions faced by religious
and secular individuals: while the latter now face a more lax environment,
the religious are still constrained by their own religion’s code of conduct. The
following section examines the evolution of the regulation of personal liberties
in Europe, while section “Religiosity, Personal Liberties, and Labor Supply”
analyzes its impact on individual labor supply among secular and religious
individuals.
Constraints on Personal Liberties

Most religions instill to their affiliates a number of constraints on the use of
the above-mentioned personal liberties. Although religious individuals are free
not to use those liberties (regardless whether the law forbids them or not),
most if not all religions have successfully obtained the legal sanctioning of the
religious norms by the state. Such legal caps on personal liberties have the effect
of helping to implement the religious norms within the religious community.
But they also have the effect of constraining the decisions of non-religious
individuals. Thus, prohibiting by law the use of such liberties suggests that
individuals have some utility externality by which religious individuals feel
offended (and seculars satisfied) from living in a world in which such liberties
are permitted.5
In order to quantify the change in the legal caps on personal liberties, ELMa
construct an index of personal liberties based on the legal evolution of certain
personal liberties from 1960 to 2013 that are or have been controversial in the
5 The partial lifting of the legal caps often finds an active opposition of religious and conservative groups. In
many countries religious organizations advocate the tightening of laws regulating such personal liberties.
In some cases violence is being used toward those that exercise such liberties, as, for example, in what is
now defined as “anti-abortion terrorism”. See the statistics for the US gathered by the National Abortion
Federation (NAF).
recent past. The data reflect legislation on abortion, divorce, women’s rights,
LGBT rights, and euthanasia for 34 European countries and are assembled
from various sources such as the UN, the EU Parliament, World Bank, the
Human Rights Project, Pew Research Center, Freedom to Marry, and so on.
See ELMa for details on its construction. Figure 25.1 displays the evolution of
the index of liberties for European countries from 1960 to the present, while
Fig. 25.2 depicts the cross-country standard deviation of the index for each year
in the period 1960–2013. A higher value of the index indicates more liberal
laws regulating personal liberties.
The main message from Figs. 25.1 and 25.2 is that legal caps on personal
liberties have greatly expanded in most European countries. The value of the
index was relatively low and homogeneous across European states during the
1960s. As shown in Fig. 25.2, the 1970s witnessed an important increase in the
variability of the index, reflecting the legal changes occurring in some, but not
all, of the countries in the dataset. During the 1980s and 1990s, most of the
countries in the sample kept introducing legal reforms, so the variability of the
index decreased as countries started to catch up. A new liberalization wave in
the 2000s has led the variability of the index to a new maximum. Differences
across countries are substantial, as Fig. 25.2 shows. The most conservative
Fig. 25.1 Legal evolution of the personal liberties index, 1960–2013

.18
.16
Std. Dev. Liberties
.14
.12
.1
1960 1970 1980 1990 2000 2010

year
Fig. 25.2 Cross-country std. deviation of the personal liberties index, 1960–2013
countries according to their liberties index are Ireland (with an average value
of the index of 0.15 across the period), Israel (0.20), and Poland (0.21). On the
other side of the spectrum, the most liberal countries are Sweden at the top
(0.70 on average), followed by Norway and Denmark (0.56). The following
section analyzes how this unprecedented change in regulation has affected both
religious and secular individuals.
Religiosity, Personal Liberties, and Labor Supply

While the legal caps on personal liberties have considerably loosened over
the last decades, religious norms have remained fairly stable over the same
period. The amplitude and the strictness of these norms vary through faiths
and also across religious communities. And of course the extent to which
such norms effectively condition individual behavior also depends on the
personal degree of religiosity. The literature has modelled the incidence of
religion’s rules following two main approaches. One is that norms directly
shape individual preferences. The second approach is to model religious norms
as a pure constraint on choices. For instance, religious individuals may enjoy
some types of meat but accept the religious prohibition on its consumption.
ELMa follow the first approach: religion instills negative utility from violat-
ing the norms on these personal liberties. As in Becker (1996) they assume that
religious norms are internalized as preferences. Individual preferences differ
only in the valuation of the use of private liberties. As liberties broaden, secular
individuals have a higher utility [all equal] and religious individuals a lower
one. Then, under standard assumptions in consumer theory, the marginal
utility of material consumption experienced by the secular should be higher
than that by the religious [the rest all equal]. From this it follows that a lifting
of the cap on liberties will be a positive (negative) incentive to work for the
secular (religious) individuals. Hence, the wider the legal cap on liberties, the
more divergent will be the choices made by religious and secular.
The core of ELMa’s model is as follows. Assume there are three goods which
individuals can potentially enjoy. Two of them are private goods, consumption
c and leisure l ∈ [0, 1]. Individuals need to exert effort to obtain consumption,
which is defined as e = 1−l. There is also a public good, personal liberties ∈
[0, M ]. The maximum liberties accessible, M , is determined by law. The cap
on liberties M has two effects. First, it establishes the limit to what is accessible
to individuals. Second, it may produce an externality because individuals may
dislike to being in a society where some liberties are permitted, independently
of whether or not they will personally use them.
ELMa represent the effect of liberties on an individual as the combination
of the personal use of them, , and the maximum legally permitted, M , this
being multiplied by the parameter α ∈ [0, 1]. This parameter indicates the
weight assigned to the externality effect.6
Individuals are endowed with a “religiosity” index x ∈ [0, 1]. ELMa
parameterize the difference in the individual preferences over liberties by
assuming that the utility function over consumption, liberties, and leisure can
be written as

u c, (x − x)[ + α M ], 1 − e , (25.1)
where u(·, ·, ·) is common to all individuals and x is the threshold level of

religiosity separating those that value liberties positively with x ≤ x (we call
them secular) from those that value them negatively, with x > x (the religious).
Notice that the higher the degree of religiosity, the lower the utility [all equal].
Under standard assumptions in consumer theory, higher religiosity produces
6 The analysis is robust to the externality arising instead from the actual practice of wider liberties by some
individuals in society. It is also robust to considering different values for religious and secular individuals.
lower marginal utility of consumption and of leisure.7 If the effect on the

marginal utility of consumption is larger [in absolute terms] than on leisure,
the supply of effort is decreasing as religiosity increases. In view of the assumed
utility function, it is immediate that the difference in labor supply induced by
religiosity is amplified by the width of personal liberties M .
ELMa show that if the effect on the marginal utility of consumption is larger
than on leisure8, then
[1] religiosity has a negative effect on effort and the effect is larger the higher
the cap of liberties afforded, that is, the interaction of religiosity and
liberties has a negative effect on labor supply; and
[2] the effect of liberties on labor supply is positive for secular and negative
for religious.
Empirical Evidence
Using data for European countries, ELMa obtain that [1] and [2] hold
empirically. The analysis employs the above-described liberties index as well
as data from the European Social Surveys on religiosity, labor supply, and
other controls.9 Based on the personal liberties index mentioned above, ELMa
construct an individual-specific index that reflects the level of liberties at
different stages of the individual’s life cycle, as some decisions are specific to
some particular stage, like decisions on education or birth control. To that
effect, ELMa average the values of the liberties index corresponding to the
years when that individual was between 18 and 40 years of age.
ELMa estimate the following equation:

ei,j,s = β 0 + β 1 Reli,j,s + β 2 Reli,j,s × i,j,s + β 3 i,j,s + Xi,j,s β + γ Zj
+ δYs + μZj × Ys + i,j,s , (25.2)
where i, j , and s denote individual, country, and survey year, respectively, e

denotes individual labor supply, Rel is a proxy of individual religiosity (x),
is the individual-specific index of liberties (lib40 ), X contains individual
controls, and Z and Y are country and survey dummies, respectively.
7 ELMa assume the standard properties on u(·, ·, ·): the utility increases in all arguments, satisfies concavity
with respect to each argument, and has non-negative cross derivatives.
8 If the opposite holds, then the implications are reversed.
9 Data from rounds 2002, 2004, 2006, 2008, 2010, and 2012 are employed for all countries available (at
most 34, including Turkey and Israel).
ELMa results imply that β 2 < 0 and β 3 > 0, that is, the marginal impact
of an increase in the cap on liberties on effort is positive for low levels of
religiosity but it becomes negative when religiosity is high. The turning point
of this equation allows us to identify x, the threshold separating religious and
seculars, which is given by x = −β 3 /β 2 . They use the estimated coefficients
to identify the value of x. A final implication of this theory is that the marginal
effect of religiosity on effort is overall negative. ELMa estimate this equation
by OLS (ordinary least squares) and by 2SLS (two stage least squares) using
an instrument for individual religiosity described in detail there.
Table 25.1 contains the baseline results. The dependent variable is the total
amount of hours worked per week (in main job), including any paid or
Table 25.1 Baseline: effort, religious intensity, and liberties, OLS and 2SLS
[1] [2] [3] [4] [5] [6]
RELINT −0.551 3.355*** 3.794*** −0.904 6.820** 7.089**
(0.185) (0.007) (0.004) (0.398) (0.013) (0.014)
LIB40 0.220 3.189** 3.414** 0.450 5.752*** 5.788***
(0.887) (0.036) (0.030) (0.767) (0.004) (0.005)
RELINT ×LIB40 −8.669*** −9.226*** −16.451** −16.929**
(0.002) (0.002) (0.011) (0.013)
AGE 0.063 0.082 0.131* 0.051 0.089 0.131**
(0.377) (0.230) (0.050) (0.487) (0.196) (0.046)
AGE2 −0.001 −0.001 −0.002** −0.001 −0.001 −0.002**
(0.299) (0.165) (0.027) (0.402) (0.129) (0.021)
GENDER −7.666*** −7.676*** −7.677*** −7.586*** −7.633*** −7.620***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
CHILDREN −0.555** −0.720***
(0.030) (0.002)
EDUYRS 0.067 0.102*
(0.230) (0.060)
HHSIZE −0.063 0.033
(0.568) (0.716)
HEALTH −0.019 −0.010
(0.852) (0.926)
EDU-PTNR −0.060 −0.035
(0.258) (0.470)
c 43.009*** 41.219*** 43.688*** 52.155*** 48.080*** 46.827***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
R2 0.148 0.149 0.150 0.149 0.150 0.151
Obs 98,200 98,200 96,448 91,055 91,055 89,418
K-P test 499 272 270
Cragg-Donald 22,815 15,942 10,687
Anderson-Rubin test 0.4052 0.028 0.031
Notes. Dependent variable is hoursworked. All models contain country, survey, and
country-survey dummies. Columns 1–3 have been estimated by OLS, while columns 4–6
by 2SLS. There are 34 countries. Robust standard errors clustered at the country level
have been computed. p-values are in parentheses
*p-value < 0.10; **p-value < 0.05; ***p-value < 0.01
unpaid overtime (hoursworked).10 Column 2 regresses hoursworked on

religious intensity (relINT ), our individual-specific index of personal liberties
(lib40 ), the interaction of these two variables, as well as some exogenous
individual characteristics (gender and age) and country, survey, and country-
survey dummies.11 The coefficient of the interacted term is highly significant,
as our theory predicts, and is negative, suggesting that the negative effect of
religiosity on effort is amplified by the availability of liberties. The coefficient
associated with lib40 , that captures the effect of liberties on effort when relINT
is close to zero, is positive and significant at the 5% level which means that
liberties are an incentive for the secular. On the other hand, the effect of relINT
is positive and significant suggesting that for moderate values of lib40 the
overall effect of relINT is ambiguous. Column 3 shows that the results continue
to hold when other individual controls are introduced in the regression.12
As there are good reasons to believe that relINT can be endogenous, we
have reestimated columns 1–3 by 2SLS using an instrument for religiosity. To
construct an instrument for relINT , ELMa exploit the fact that religiosity is a
cultural trait that transcends national borders in the European environment.
That is, people belonging to a particular religious denomination are likely to
have similar beliefs and attitudes toward topics regulated by their religions.
They construct an instrument for relINT , relIV INT , as follows: for an individual
i from country j with (current or past) religious denomination r, we consider
the people of the same religious denomination, gender, and age bracket living
in the countries that share a border with country j .13 We consider neighboring
countries so that similarities between national customs are more pronounced.
Next, we compute the average value of relINT for those individuals. The
resulting quantity is the value of relIV INT for individual i in country j . As
10 Notice, however, that hoursworked is an imperfect measure of the willingness to work, as it reflects
attitudes toward effort as well as the characteristics of the environment (i.e., rigidity of the labor market).
To address this issue, ELMa also consider a variable that reports the number of hours that the respondent
would like to work (desiredhours).
11 Religious intensity is computed as the average of three variables: (1) the respondent’s monthly frequency
of praying, (2) the self-reported degree of religiosity, and (3) the respondent’s monthly frequency of
attendance of religious services.
12 These additional controls are: whether the respondent lives with a partner (cohab), years of completed
education (eduyrs), a subjective measure of own’s health (health), whether there are children in the
household (children), the size of the household (hhsize), and a measure of partner’s education (edu-
ptnr).
13 We define three age brackets: from 18 to 35, from 36 to 60, and from 60 onwards. The results are robust
to changes in the definition of the brackets.
the instrument is the average of other people’s religiosity, by construction it

is uncorrelated with individual i’s “innate” characteristics, such as ability.14
Columns 4–6 in Table 25.1 reestimate columns 1–3 by 2SLS, instrumenting
religious intensity with the above-described instrument. The results are very
similar, and for the sake of brevity, we focus on column 6, our baseline
specification henceforth, which reestimates column 3 in the same table. This
column shows that the interaction of religiosity and liberties has a negative and
significant effect on labor supply, while the coefficient on religiosity is positive
and significant. As mentioned before, the overall effect of religiosity is then,
ambiguous, especially for moderate values of lib40 .
The magnitude of the effect of religiosity on effort is quite sizeable among
individuals that have access to liberties. Focusing again on column 6 in
Table 25.1, an increase in one standard deviation in the intensity of religious
beliefs is associated with a decrease in the number of hours worked per week
of 0.75, 1.39, or 1.8 hours for individuals with a value of lib40 in the 75th,
95th, and 99th percentile, respectively.
The second prediction of the model states that liberties provide differential
incentives to religious and seculars. As noted above, the value of x, the
threshold separating religious from seculars, is x = −β 3 /β 2 , where β 2 and
β 3 are the coefficients associated to relINT ×lib40 and lib40 , respectively.
Using the values of these coefficients from column 6 of Table 25.1, a value
of x = 0.34 is obtained. This value corresponds to the percentile 57 of the
distribution of relINT , which means that an increase in lib40 translates into
an increase in the amount of hours worked for almost 60% of the population
in Europe (the secular). For the remaining 40% (the religious), an expansion
of the legal cap on liberties has a negative effect on their labor supply.
Finally, the magnitude of the effect of liberties on effort is large for both
seculars and religious. We provide two examples, for individuals in the 20th
and in the 80th percentile of the distribution of relINT . In the former case, an
increase in one standard deviation in lib40 increases by 0.72 hours per week
the number of hours worked. In the latter, a similar increase reduces by 0.92
hours the weekly amount of hours worked.
14 We construct the instrument for the interaction term of religious intensity and liberties simply by
replacing relint by reliv
int in the product.
Religiosity, Liberties, and Voting

From the assumption that religiosity shapes the valuation of individual liber-
ties, we have derived that the interaction between religiosity and liberties is
a key variable in determining labor supply. Wider liberties enlarge the gap in
labor supply—hence income—between religious and secular. Consequently,
the distribution of religiosity over the population and the legal cap on
liberties influences the extent of income inequality and hence the demand for
redistribution. But at the same time, the outcome of the vote over liberties will
change the incentives of the individuals and hence have an effect on who the
median voter is. This is the problem studied by ELMb.
The link between religiosity and the preferences of poor individuals for
low taxation is documented in, among others, Guiso et al. (2006). Huber
and Stanig (2009) argue that the right-wing positioning of poor religious
individuals is driven by a “forced choice” provoked by the specific policy mix
furnished by conservative parties. Chen and Lind (2016), Ceyhun et al. (2013),
and Huber and Stanig (2011) argue that the lower support for redistribution by
the religious results from the preference for social assistance and redistribution
within their own religious community. In Scheve and Stasavage (2006)—
see also Gill and Lundsgaarde (2004) and Clark and Lelkes (2005)—the
psychic benefit from religion allows individuals to cope with bad states which
reduces the need for social insurance, and hence religious individuals prefer less
redistribution and social assistance. In Bénabou and Tirole (2006) religion is
a way of manipulating one’s beliefs in order to motivate continued effort, and
therefore religious agents will work harder and demand less taxes. A similar
argument is made by Huber and Stanig (2011) and Bénabou et al. (2015a,b).
Roemer (1998) formalizes the voting over redistribution and religious provi-
sions and shows that all parties may propose relatively moderate taxes as a result
of the two-dimensional competition.
ELMb contribute to this literature by studying the vote on redistribution
and personal liberties by individuals that differ in their productivity and in
their esteem for liberties—their religiosity. Besides, the problem also has the
novelty that the vote for the cap on personal liberties has a side incentive effect
on labor supply and hence on inequality and the demand for redistribution.
How does this effect of religiosity and liberties on income inequality
translate into political voting? Using a similar model as in the previous section,
ELMb examine the outcome of sequential voting, first on the cap on liberties
and next on the marginal tax rate—for the family of linear tax functions. While
changes in the tax schedule can take place yearly together with the voting of
the government’s budget, changes in the legislation on liberties are far less
frequent.
Voters are assumed to be aware that the outcome of the vote on liberties
defines a set of incentives for work that will determine each individual’s
preference for redistribution, as well as who is the median voter in the second
stage. In order to make the problem tractable, ELMb simplify the environment
to two productivity types, high and low, and two religiosity levels, high
(religious) and low (secular). It is assumed that low productivity individuals
are in a majority.
ELMb’s argument can be summarized as follows. Suppose that the secular
are in a majority. Then they will impose full liberties in the first stage of the
voting because it is their best choice independently of the continuation. In the
second stage, the tax chosen is the one preferred by the median voter, given
that there are full liberties. Nothing unexpected arises in the case of a secular
majority.
When the religious are in a majority, there is more room for strategic
behavior in voting. Observe that the religious with low productivity will always
earn the lowest income and that the secular with high productivity will earn
the highest. Let us now consider the religious poor. They face a dilemma with
respect to their potential support for liberties, when taking into account that
they will not be pivotal in the second stage. Because of their religiosity, they
will experience a utility loss if wider liberties are chosen. However, under full
liberties work incentives for the secular will be the largest, and this results in a
larger tax base and larger net transfers to the lowest incomes.
Which one dominates depends on the gap in religiosity between secular
and religious relative to the gap in productivities between low and high
productivity. The incentive effect of liberties on earned income by the secular
depends on how wide the religious gap is. Therefore, under a religious majority,
if differences in productivity are large relative to differences in religiosity, the
low productivity secular will be pivotal in the second stage and will demand
substantial redistribution. In this case, the religious of low productivity might
vote for liberties in order to have redistribution in the second stage. However,
if the gap in religiosity is large enough, the religious majority will vote
for repression in the first stage, knowing that under repression all the low
productivity individuals will vote for the same level of redistribution.
Summing up, ELMb result is that the outcome of the voting process
critically depends on whether the majority is made of secular or of religious
individuals and also depends on the relative size of the religious and pro-
ductivity gaps. When the secular are in a majority, the political outcome is
full liberties and the tax preferred by the low productivity secular. When
the religious are in a majority, we have that: (1) when the religiosity gap is
sufficiently wide, the religious majority represses liberties and also votes for
low redistribution; (2) when the productivity gap is dominant but small, the
religious vote for liberties and high taxation; and (3) when the productivity
gap is large enough, the religious majority chooses full liberties and moderate
redistribution.
Concluding Remarks
In the last 50 years, Western countries have witnessed an unprecedented
change in the legal regulations pertaining to individual liberties, such as
women’s rights for education or employment, abortion, divorce, LGBT rights,
euthanasia, and so on. Those regulations have evolved from almost mimicking
the religion codes of conduct to becoming far more tolerant. To the extent
that such legal limits were an effective constraint for the most secular part
of the population, the sequential lifting of these constraints must have had
important effects on individual choices. These effects have been both direct—
since the lifting of the legal constrains permits actions that were forbidden
before—and indirect, because a more tolerant legislation can be perceived as a
positive or a negative externality, depending on individuals’ preferences. That
is, irrespective of the individual use of all or some of those liberties, individuals
may like or dislike to live in a more permissive society.
There are many interesting open questions that deserve attention. Why this
change in legislation has taken place? Is it that religiosity among the population
has changed and this has mapped into a more liberal legislation? If so, which
factors have driven such a change in religiosity? Is it the case that the change
in the legal caps—and the induced change in behavior of some—has had an
effect on the social interpretation of the norms and codes of conduct? Have
the changes in the legislation further induced secularization of society? There
is little doubt that this is a phenomenon in which causality is likely to operate
both ways.
The work in ELMa,b is but a first step in the study of the causes and
consequences of the “human rights revolution”. Their results can be inter-
preted as providing a possible explanation for our first question: why has
legislation changed so much? Their point is that in order to relax the regulation
on liberties, we don’t need to appeal to the vote of a potential majority of
the secular, a change that in most cases—notably in the US—apparently has
not taken place. ELMb show that an economic shock that widens the gap
in productivities—thus, increasing income inequality—is enough to induce
the low income religious individuals to support a more relaxed legislation on

personal liberties, even when the religious are in a majority. This of course
should be supplemented with the study of the subsequent effect of the change
in behavior of part of society on personal levels of religiosity.
ELMb also start exploring in the direction proposed by Becker (1996),
suggesting that religiosity might be manipulated by the elite. They find that
the elite prefers that the low productivity individuals are split into secular
and religious, rather than all being religious (or secular). This is so because
either type of homogeneity, by making all the poor alike, would lead to higher
demands for redistribution. But if the low productivity are split, the secular
individuals will impose lower demands for redistribution because their income
is higher than the religious poor.
Our last remark concerns the significant role of externalities in shaping
individual preferences and behavior, according to our empirical results. The
externality effect of liberties can be viewed as a cultural factor specific to
each religion. In our model, religions that are more individualistic (such as
Protestantism that focuses on the personal relationship with god, independent
of what the others might be doing) could induce a less negative effect of
wider liberties on labor effort. Hence, differences across religions in the weight
assigned to the externality of liberties could explain differences in individual
and aggregate labor effort across countries. This would provide an explanation
complementary to Weber’s (1904) as to why Protestantism became more
conducive to economic growth than Catholicism. We leave for future research
the precise identification and measurement of this externality factor across
different cultures and religious affiliations.
References
Barro, Robert, and Rachel McCleary. 2003. Religion and Economic Growth across
Countries. American Sociological Review 68: 760–81.
Becker, Gary. 1996. Norms and the Formation of Preferences. In Accounting for Tastes,
ed. G. Becker. Chicago: University of Chicago Press.
Becker, Sascha O., and Ludger Woessmann. 2009. Was Weber Wrong? A Human
124: 531–596.
Bénabou, Roland, Davide Ticchi, and Andrea Vindigni. 2015a. Forbidden Fruits: The
Political Economy of Science, Religion, and Growth. NBER Working Paper 21105.
———. 2015b. Religion and Innovation. American Economic Review, P&P 105 (5):
346–351.
Bénabou, Roland, and Jean Tirole. 2006. Belief in a Just World and Redistributive
Politics. Quarterly Journal of Economics 12: 699–746.
———. 2011. Identity, Morals and Taboos: Beliefs as Assets. Quarterly Journal of
Economics 126: 805–855.
Berman, Eli. 2000. Sect, Subsidy and Sacrifice: An Economist’s View of Ultra-
Orthodox Jews. Quarterly Journal of Economics 115: 905–953.
Botticini, Maristella, and Zvi Eckstein. 2012. The Chosen Few: How Education Shaped
Jewish History, 70–1492. Princeton: Princeton University Press.
nomic Growth and Happiness? Evidence from Ramadan. Quarterly Journal of
Economics 130: 615–658.
Cantoni, Davide. 2015. The Economic Effects of Protestant Reformation: Testing
the Weber Hypothesis in the German Lands. Journal of the European Economic
Association 13: 561–598.
Carvalho, Jean-Paul, and Mark Koyama. 2016. Development and Religious Polar-
isation: The Emergence of Reform and Ultra-Orthodox Judaism. Journal of
Comparative Economics 44: 562–584.
Ceyhun, Elgin, Turkmen Goksel, Mehmet Gurdal, and Cuneyt Orman. 2013. Reli-
gion, Income Inequality, and the Size of the Government. Economic Modelling 30:
225–234.
Chen, Daniel L., and Jo Thori Lind. 2016. The Political Economy of Beliefs: Why
Fiscal and Social Conservatives/Liberals (Sometimes) Come Hand-in-Hand? TSE
Working Paper, n. 16–722.
Clark, A., and O. Lelkes. 2005. Deliver Us From Evil: Religion as Insurance. Paris:
Mimeo, DELTA.
Darnell, Alfred, and Darren E. Sherkat. 1997. The Impact of Protestant Fundamen-
talism on Educational Attainment. American Sociological Review 62: 306–316.
Delacroix, J., and F. Nielsen. 2001. The Beloved Myth: Protestantism and the Rise of
Industrial Capitalism in Nineteenth-Century Europe. Social Forces 80: 509–553.
Durlauf, S., A. Kourtellos, and C.M. Tan. 2012. Is God in the Details? A Reexamina-
tion of the Role of Religion in Economic Growth. Journal of Applied Econometrics
27: 1059–1075.
Ekelund, R.B.Jr., R.F. Hébert, and R.D. Tollison. 2006. The Market for Christianity.
Cambridge: MIT Press.
Esteban, Joan, Gilat Levy, and Laura Mayoral. 2018a. Personal Liberties, Religiosity,
and Effort. Unpublished Working Paper.
———. 2018b. Liberté, Égalité, et … Religiosité. Journal of Public Economics 164:
241–253.
Gill, Anthony, and Erik Lundsgaarde. 2004. State Welfare Spending and Religiosity.
A Cross-National Analysis. Rationality and Society 16: 399–436.
Grier, Robin. 1997. The Effect of Religion on Economic Development: A Cross
National Study of 63 Former Colonies. Kyklos 50: 47–62.
Guiso, Luigi, Paola Sapienza, and Luigi Zingales. 2003. People’s Opium? Religion
and Economic Attitudes. Journal of Monetary Economics 50: 225–282.
———. 2006. Does Culture Affect Economic Outcomes? Journal of Economic

Perspectives 20: 23–48.
Hitchcock, William, Akira Iriye, and Petra Goedde. 2012. The Human Rights Revolu-
tion: An International History. New York: Oxford University Press.
Huber, John, and Piero Stanig. 2009. Individual Income and Voting for Redistribu-
tion Across Democracies. Unpublished Working Paper.
———. 2011. Church-State Separation and Redistribution. Journal of Public Eco-
nomics 95: 828–836.
Communes and Other Collectives. Journal of Political Economy 100: 271–291.
Levy, Gilat, and Ronny Razin. 2012. Religious Beliefs, Religious Participation and
Cooperation. American Economic Journals: Micoreconomics 4: 121–151.
Marx, Karl. 1843, 1976. Introduction to A Contribution to the Critique of Hegel’s
Philosophy of Right, vol. 3. Collected Works. New York.
McCleary, Rachel M., and Robert J. Barro. 2006. Religion and Economy. Journal of
Economic Perspectives 20: 49–72.
Nunn, Nathan. 2010. Religious Conversion in Colonial Africa. American Economic
Review 100: 147–152.
———. 2014. Gender and Missionary Influence in Colonial Africa. In Africa’s
Development in Historical Perspective, ed. E. Akyeampong , R. Bates , N. Nunn,
J.A. Robinson, 489–512. New York: Cambridge University Press.
Roemer, John E. 1998. Why the Poor Do Not Expropriate the Rich: An Old Argument
in New Garb. Journal of Public Economics 70: 399–424.
Scheve, Keneth, and David Stasavage. 2006. The Political Economy of Religion and
Social Insurance in the United States, 1910–1939. Studies in American Political
Development 20: 132–159.
Sherkat, D.E., and A. Darnell. 1999. The Effects of Parents’ Fundamentalism on
Children’s Educational Attainment. Journal for the Scientific Study of Religion 38:
23–35.
Solt, Frederick, Philip Habel, and J. Tobin Grant. 2011. Economic Inequality, Relative
Power, and Religiosity. Social Science Quarterly 92: 447–465.
Valencia Caicedo, Felipe. 2018. The Mission: Human Capital Transmission, Eco-
nomic Persistence, and Culture in South America. The Quarterly Journal of
Economics, qjy024. https://doi.org/10.1093/qje/qjy024.
Weber, Max. 1904. The Protestant Ethic and the Spirit of Capitalism. New York:
Scribners.
26
Causes and Consequences of Monotheism
in the Supply of Religion
Murat Iyigun
Introduction
Human history is a testament to the strong complementarities between
political and religious authority in conferring upon sovereignties the legitimacy
to sustain, expand, and prolong their political rule.1 From Charles V, who
was crowned Holy Roman Emperor by Pope Clement VII in Bologna as
the last emperor to receive a papal coronation, to Yavuz Sultan Selim, who
conquered Mecca in 1517 enshrine the Ottoman Sultanate with the title of
Sunni Caliph, all the way to contemporary nation-states such as Iran and
Saudi Arabia that are ruled by sharia law, we have countless historical examples
of ecclesiastical and secular political authority being combined to bolster the
sovereign legitimacy necessary for social and political order.
In this handbook chapter, I first revisit and elaborate on my hypothesis
that ecclesiastical market power was immensely aided by the birth, spread, and
success of monotheisms. On that basis, I argue that religious authorities were
able to confer political legitimacy more credibly and efficiently when they were
associated with a monotheism. Based on some of my ongoing research with
Miles Kimball, I then put forward the claim that there were even more elemen-
1 Iyigun (2015) and Rubin (2017).

M. Iyigun ()
University of Colorado Boulder, Boulder, CO, USA
e-mail: iyigun@colorado.edu

424 M. Iyigun
tary fundamentals that account for the spread and sustenance of monotheisms
in human history. In particular, I illustrate that monotheistic ecclesiasticisms
had a demonstrable evolutionary advantage that helped with their favorable
selection over time. I then conclude this chapter with suggestions for further
analytical and empirical research in the economics of religion. Specifically, I
make a case for exploring in some detail whether or not monotheisms that
evolved to become more congregational or clerical had a further leg up in their
capacity to confer political legitimacy.
Religion, Social Order, and Political Stability

The idea that religion and faith, in general, help to promote social order
and stability is more than a century old. Progressive thinkers of the post-
Enlightenment era, such as Émile Durkheim, Auguste Comte, and David
Hume, began to emphasize the positive social functions of faith and religion.
According to Hume (1911), for example, benevolence and moral considerations
associated with religion are the pillars of social harmony and stability. And
Durkheim (1915) saw the manifestations of religious practices, norms, and
rituals in social cohesion.
In the 1930s, the structural-functionalist school, led by Talcott Parsons,
began to assert that the cohesion of societies depended on their members
sharing a common purpose, conceptions of morality and an identity. In this,
they were adhering to Durkheim, who saw these social necessities in religion.
Other twentieth-century thinkers such as Reinhold Niebuhr (1932, p. 20)
honed in on less abstract functions of religious beliefs and adherence, such as
the sustenance of social order via individual moral self-restraint as the channels
through which religion aided social order and stability. He articulated, for
instance, that societies’ survival depended on their efficacy in eliminating—
or, at the very least, holding in check—all forms of self-serving violence.
Human beings are genetically encoded to survive, which is manifested in
their will to live. But according to Niebuhr, that same survival instinct easily
morphs into “will to power” because individuals recognize that social, political,
and economic power can enhance survival. Since social stability requires the
elimination of violence and anarchy and the abuse of various types of power,
one way to attain that stability and equilibrium is through the moral self-
restraint that religions offer:
Essentially religion is a sense of the absolute. When, as is usually the case,

the absolute is imagined in terms of man’s own highest ethical aspirations,
26 Causes and Consequences of Monotheism in the Supply of Religion 425
a perspective is created from which all moral achievements are judged to be

inadequate. Viewed from the relative perspectives of the historic scene, there is
no human action which cannot be justified in terms of some historic purpose or
approved in comparison with some less virtuous action. The absolute reference
of religion eliminates these partial perspectives and premature justifications . . . .
The history of religion is proof of the efficacy of religious insights in making
men conscious of the sinfulness of their preoccupation with self. (Niebuhr 1932,
pp. 52, 55)
Wright (2009, pp. 55–57) attributes the rise of religion to its ability to
enhance social stability and argues that religions it began to evolve in the
transition from hunter-gatherer societies to settled agriculture:
[Laissez-faire] law enforcement is a shakier source of social order in chiefdoms

than in hunter-gatherer societies. In a hunter-gatherer village, you know every-
one and see them often and may someday need their help. So the costs of
getting on someone’s bad side are high and the temptation to offend them
commensurately low. In a chiefdom, containing thousands or even tens of thou-
sands of people, some of your neighbors are more remote, hence more inviting
targets for exploitation . . . . In this phase of cultural evolution, a supplementary
force of social control was called for. Religion seems to have responded to the
call . . . . Whereas religion in hunter-gatherer societies didn’t have much of a
moral dimension, religion in the Polynesian chiefdoms did; it systematically
discouraged anti-social behavior . . . . Believing that anyone you mistreat might
haunt you from the grave could turn you into a pretty nice person.
The Greatest Revelation: Monotheism

Monotheisms, Social Order, and Political Stability
The upshot here is that the conventional literature on the sociopolitical and
economic functions of religion is quite extensive and established. By contrast,
there is a much thinner relevant literature and next no explicit acknowl-
edgment of whether monotheistic religions differed from other ecclesiastical
creeds in the efficacy of their roles in sociopolitical and economic life.
The first two chapters of Iyigun (2015) are intended to bridge this gap
in some systematic and explicit fashion. As I have elaborated in some detail
there, a review of the related literature on the topic indicates that there isn’t
much consensus on whether monotheistic faiths structurally differ in their
functional impact on their adherent societies. Some scholars, however, did see
426 M. Iyigun
that monotheistic creeds imbued societies with a host of additional social and
political advantages.
The three Abrahamic monotheisms of Judaism, Christianity, and Islam
were all born between 606 BCE and 622 CE in the Middle East. And the
subsequent diffusion of the One God faiths in North Africa, Asia, and Europe
was not only rapid and remarkable but also accompanied by the rise of
centralized government.
In the words of Diamond (1997, pp. 266–267), “ At the end of the last
Ice Age, much of the world’s population lived in [hunter-gatherer societies]
and no people then lived in a much more complex society. As recently as
1500 CE, less than 20 percent of the world’s land area was marked off by
boundaries into states run by bureaucrats and governed by laws. Today, all land
except Antarctica’s is so divided. Descendants of those societies that achieved
centralized government and organized religion earliest ended up dominating
the modern world. The combination of government and religion functioned,
together with germs, writing, and technology, as one of the four main sets of
proximate agents leading to history’s broadest pattern.”
In fact, as impressive as the spread and sustenance of organized religion was
over time, it paled in comparison with that of Abrahamic monotheisms: As I
show in Iyigun (2010, 2015, pp. 33–35), societies that predominantly adhered
to one of the three Abrahamic monotheisms were roughly 15 percent of the
total in the Old World as late as the eighth-century CE. But, by the year 2000,
161 countries around the globe subscribed predominantly to one or more of
the three monotheistic faiths, representing 86 percent of the 188 countries and
close to 3.3 billion people—roughly 55 percent of the world population.
While Judaism, Christianity, and Islam are not the only three monotheistic
faiths that have emerged in history, they have had and continue to have the
most adherents by far. These monotheisms have also been geographically
dominant in the Middle East, Europe, and the Near East ever since their
advent, and the spread of Christianity and Islam in particular in those regions
has been quite extraordinary by historical standards.
As monotheisms spread and sustained, especially the Abrahamic versions
came to be closely intertwined with governments and stable societies. Not only
did many historical civilizations subscribe to and promote Judaism, Christian-
ity, and Islam but also the political elite in these societies often derived their
temporal authority from their respective monotheist ecclesiastical institutions.
In turn, the latter derived substantial financial and political benefits from being
associated with One God. Thus the stability of preindustrial societies came
to be linked with their respective ecclesiastical institutions. Stable polities,
societies, and economies might not have been sufficient for ushering in human
enlightenment, industrialization, and sustainable economic progress but seem

to have been the precursors of all those (e.g., Diamond 1997; North et al. 2008;
Mokyr 2010).
So if monotheistic faiths were especially adept at sustaining sociopolitical
stability, what were their common features that mattered in this regard?
Although these are not necessarily exclusive to the three main monotheisms,
there are at least five salient traits of monotheistic faiths regarding their impact
on the economic and sociopolitical realms.2
Social Advantages of Economies of Scale
Judaism, Christianity, and Islam all acknowledge and promote the “oneness”
of God, and the barriers to entry in the religion market are substantially lower
when there are many gods. By nature, this introduces monopoly power and
elements of increasing returns to scale in the provision of religious services
when entry barriers are high (as they are under One God).
Ekelund et al. (1996, 2002) and Barro and McCleary (2003, 2005) claim
that the fixed setup costs of religion influence the equilibrium number of faiths
that can be sustained by a society or state. Most germane to what I advocate
here is the idea that a state religion is more likely to emerge when the cost
of establishing a religion is high. Monotheisms entail relatively high costs of
entry into the religion market.
In fact, Armstrong (1993, p. 49) identifies that monotheistic faiths were
unique in their mutual exclusivity, especially with respect to the belief in
One God. On the one hand, it is true that the effective monotheism of
any particular One God faith is a matter of degree, as a single deity shares
ecclesiastical power with angels, saints, prophets, and other holy people. On
the other hand, however, the hierarchy between the one deity and other
ecclesiastical actors is very clear. Along these lines, Stark (2001, pp. 19, 34)
draws a critical distinction between individuals’ relationship with One God
under monotheism and that with multiple deities in polytheism, according to
which competition between various divine beings shortened and limited the
interactions between the adherents and their gods.
Taken together, the One God/One Faith duality inherent in Abrahamic
monotheisms made them mutually exclusive vis-à-vis other faiths. That
mutual exclusivity, in turn, enabled monotheistic faiths to more than likely
emerge as the socially dominant or even the state religion once they gained
2 For further details, see Iyigun (2015, pp. 6–12).

428 M. Iyigun
traction within a society. A monotheistic state religion, then, involved fairly

difficult to surmount entry barriers for other faiths, precisely due to the mutual
exclusivity inherent in its creed and its association with and endorsement of
state political authorities.
Accountability Cum Personalized Spiritual Exchange
In polytheistic faiths, there are multiple deities who rule various aspects of
temporal life, but there exists none with the omnipotence to control all aspects
of temporal and spiritual existence. In contrast, monotheistic faiths involve one
omnipotent divine being who has not only control over the whole universe but
also desires he wishes humans to fulfill, which he can communicate to them.
Stark (2001, pp. 15–19) provides a utilitarian analysis of the psychological
and sociological effects of monotheism. He draws attention to the fact that
individual accountability toward One God is a unique feature of monotheistic
faiths and that, since the relationship between the deity and individual is per-
sonalized and extended into the afterlife, there is a strong element of beneficial
exchange on the basis of personal commitment. A relevant assertion he makes
is that, by personalizing spiritual exchange and reinforcing accountability,
theology and faith provide a very effective means to deal with human wants
and desires, such as survival, health, financial security, which are often fleeting
and inherently in short supply.
By nature, otherworldly rewards are compensation for individuals’ temporal
deeds, and to an extent, they substitute for material necessities that are in short
supply or entirely unavailable. As a consequence, they introduce an element
of continuity and commitment to the exchange between the deity and the
individual—two features that are entirely lacking in atheism and polytheism
(Stark 2001, p. 19).
Longer Time Horizons Due to Belief in Afterlife
The belief in afterlife is not unique to monotheisms, but the Judgment Day,
when individuals are held accountable for their deeds and are judged by
God accordingly, is. By contrast, religions that include reincarnation, such as
Hinduism, lack a Judgment Day—the determination of how an individual is
to be reborn being a particular judgment on the merit of the life just lived.
As Stark (2001, pp. 15–19) explains, the fact that individuals are held
accountable by one god for their temporal deeds and that his rewards are
often delayed until after death “is a major factor allowing Godly religions
to generate the long-term levels of commitment necessary to sustain strong
religious organizations.”
External Threats Beget Social Unity
Political scientists have long been interested in determinants of conflict and,

in particular, how religious identity and affiliation influences belligerence. For
instance, based on data on more than 300 violent conflicts around the world
between 1820 and 1949, Richardson (1960, p. 239) was the first to reveal that
differences of religion, especially those of Christianity and Islam, have been the
causes of wars and that, to a weaker extent, “Christians fought Christians more
than would be expected from their population.” He found that none of the
observed conflicts in his data arose because the two sides identified with the
same religion; he also did not find any conflicts that were restrained because
the involved parties adhered to different religions. Instead, conflicts arose
and persisted due to religious differences, or they were subdued or eventually
contained primarily because the actors involved coreligionists. Using much
more extensive data on conflicts in Europe, the Middle East, and North Africa
over the period between 1400 and 1900 CE, I verified these empirical patterns
as well (Iyigun 2008, 2013, 2015, pp. 15–22).
The One True Faith provided the galvanizing elements for confronting
external foes when they did not subscribe to it. Judaism, Christianity, and
Islam are all built on fundamental beliefs of One God and One Religion, a
concept of duality labeled particularism. Various theologians, such as Rodney
Stark (2001) and Karen Armstrong (1993, 2006), have suggested that this is
an important reason why, as monotheisms spread and dislocated other faiths
and belief systems in the Old World, faith-based conflicts and confrontations
began to take hold in Europe, North Africa, Asia, and the Middle East.
The upshot of all this is that real or perceived external foes aided and abetted
social cohesion, compromise, and unity internally. In Iyigun (2008, 2015, pp.
89–107), I discuss and empirically document how the Ottomans’ European
military motives and conquests led, at least in part, to the survival and
eventual recognition of Lutheran Protestants in Europe. And, in Iyigun (2015,
pp. 153–163), I review and discuss the myriad ways in which the European
rivalries with the Ottomans helped to shape the continents’ sociopolitical and
institutional history. Finally, wars and violent conflict were the two main if
not the sole drivers of lower inequality historically, precisely because they
430 M. Iyigun
set in motion economic, social, and political concessions and compromise in

societies (Scheidel 2017).3
Effective Limits on Free Ridership
A final important advantage inherent to monotheisms is that they possessed

the credential basis for minimizing the free-ride problems associated with
religion. As Ekelund et al. (1996) and Ekelund and Tollison (2011) have argued,
religion is a metacredence good. These are goods whose quality cannot be
ascertained objectively and certainly even after purchase. And the long-term
survival and success of belief systems depend on how easily they can establish
credence.
Religion, in fact, is a collective consumption good with large network
externalities—that is, individuals’ satisfaction from its use depends on others’
participation and adherence (Janeba 2007). It is generally difficult to subsidize
activities that generate these kinds of network externalities. In the language
of economics, religion is a cultural good for which free riding could be an
impediment for its growth and sustainability. Religions have typically over-
come such problems by establishing norms and rituals that penalize or inhibit
alternative uses of time and resources (Iannaccone 1992, 1994; Berman 2000;
Berman and Laitin 2008). The sustenance of such norms and rituals not only
helps screen out the less committed free riders but also forms the cornerstone
of a common social “culture.” Such norms and rituals appear to be crucial
for the Durkheimian argument. In this regard, all sociopolitical innovations
that raise accountability and loyalty among the adherents help limit free-rider
problems and bolster credence. And those belief systems whose credence is
more effectively established among the populace (i.e., monotheisms) would
also be the ones with a comparative advantage in serving the sociopolitical
stability function.
The Role of Middlemen
Monotheisms differ from one another in the extent to which they are “cler-
ical” or “congregational,” although, in this regard, the heterogeneity within
Christianity is unique. In Islam, which by construction is congregationalist,
3 This is not to dismiss the fact that wars and violent conflicts also raised real wages and income from labor
while lowering real interest rates and income from capital.
the caliphate was heavily used to legitimize political authority.4 Naturally, the
clerical system enabled more of a fusion between ecclesiastical authority and
temporal political power. The extent to which the clergy had ecclesiastical
authority often influenced the political sphere because the clergy could use
their powers to bolster or undermine the legitimacy of secular authorities. As
a particular example of such interactions between secular and ecclesiastical
political power, some scholars have documented interdependence between
political and religious authorities on Islamic sociopolitical and economic
development (Rubin 2008, 2010, 2017).
Nonetheless, if the distinction I have begun to make immediately above
holds any water, it is plausible that religious authorities were less effective in
conferring legitimacy upon political rulers in relatively more congregationalist
sects or religions where the clergy had less influence over their followers. In
other words, my main—although fairly speculative—conjecture here is that
political authorities derived presumably more “ecclesiastical legitimacy” when
monotheist clerics were directly involved in bestowing secular rulers with
their stamp of approval and recognition. And, by extension, they derived
less legitimacy when the prevalent monotheism was more congregational,
involving fewer, if any, temporal ecclesiastical intermediaries.
On the one hand, one could argue that the prolific variety of denomi-
nations within Christianity and, to a much lesser extent, Islam contradicts
the argument that monotheisms are easier sustained as monopoly religions.
However, as Ekelund et al. (1996) have shown, most of the denominational
pluralism within Christianity occurred because the Roman Catholic Church
began to operate as a heavily price-discriminating monopolist. Presumably this
left room for Lutheranism and its various offshoots to eventually enter the
European religion market. In contrast, the Shi’a-Sunni split within Islam is
an artifact of the disagreement over who was the legitimate successor of the
prophet Mohammed.
On the other hand, monotheisms with less intermediation by the clergy
ought to have been more efficient, ceteris paribus. From an economics stand-
point and in light of the main theme of this chapter, this should have meant
an additional advantage for less clerical monotheist sects that involved less
intermediation. As such, the various sociopolitical and institutional advantages
we reviewed above might have been even more pronounced for less clerical
monotheisms.
4 Of course, this is true in a relativist sense as there are variations within Islam too in this respect, with
Shi’a Islam being less congregationalist than Sunni Islam.
432 M. Iyigun
Taken together, the two particular potential effects of monotheisms that are
congregationalist versus clerical along the lines we have just reviewed could
have well come to bear on how monotheisms impinged upon a society’s
politics, economics, and institutions. Specifically, if on net the legitimacy-
conferring effect of clerical sects dominated the efficiency gains involved with
more congregationalist sects, then the benefits of adhering to a relatively
more clerical monotheism ought to have been amplified (and vice versa).
Consequently, one would expect to find some systematic evidence for quan-
titatively and statistically meaningful differences in the speed and magnitude
with which relatively more congregationalist sects spread and endured vis-à-vis
more clerical monotheistic sects.
The Evolutionary Roots of Monotheism
What I’ve advocated thus far is that while religion did serve useful sociopo-
litical, institutional, and individual purposes from a purely functionalist
perspective, monotheistic creeds supplied unique advantages that dominated
polytheistic faiths historically. Furthermore, I’ve discussed a related literature
as well as some of my earlier empirical work which suggest such advantages
might well account for the very rapid spread, resilience, and sustenance of
monotheistic societies in the long run.
Nevertheless, in more recent research with Miles Kimball, we argue that
there are even more rudimentary reasons for the historical triumph of
monotheisms over other religions, especially polytheisms (Iyigun and Kimball
2018).
The essence of our argument is very simple: Consider a fixed number of
total agents at any given point in time t, t ≥ 0 whose mass is normalized
to one. Among these agents, Mt are monotheist and the rest, 1 − Mt = Pt ,
are polytheist. Here, agents could be individuals or societies. In every period
t, each agent engages in a one-to-one interaction with another agent where
the interaction assignments are all random. We are deliberately abstract what
these interactions entail, although one can envision them to be religious
proselytization and dialogue, commercial trade, or even violent conflicts.
In any given period t, an agent of type M would match with another
agent of the same type with probability of Mt and with an opposite-type
agent with the probability of Pt = 1 − Mt . And conditional on match
probabilities, she’d have a p probability of conversion to polytheism. By
contrast, an agent of type P would match with another agent of the same type
with probability of Pt and with an opposite-type agent with the probability
of Mt = 1 − Pt . And, conditional on match probabilities, she’d have an m

probability of conversion to monotheism. Hence, aggregating over the mass
of polytheists and monotheists, one would then derive the net conversion
flows into monotheism and polytheism, respectively, as a = 2(m − p)Mt Pt
and −a = −2(m − p)Mt Pt .
Were there fundamental differences in how monotheistic and polytheistic
creeds interacted with others of a different faith? Yes. As Karen Armstrong
(1993, p. 49) has clearly articulated, polytheistic religions were historically
more tolerant than monotheistic creeds: “Paganism was essentially a tolerant
faith: provided that old cults were not threatened by the arrival of a new deity,
there was always room for another god alongside the traditional pantheon.”
By contrast, all monotheistic traditions agree on the oneness and uniqueness
of God, but they are in inherent disagreement over who the messenger of
God is and, accordingly, which ecclesiastical clerical institutions ought to
have domain over religious matters. Sociologists brand this aspect of religion
particularism—the idea that Judaism, Christianity, and Islam are all built on
fundamental beliefs in One God and One Religion. Moreover, scholars such
as Rodney Stark (2001) and Karen Armstrong (1993, 2006) have suggested
particularism as an important cause of faith-based conflicts and confrontations
historically.
When the interactions we contemplate involve violent conflicts, it is more
readily apparent that the net conversion flows favored monotheisms over
polytheisms. After all, history is replete with cases in which monotheistic
societies forcefully or through economic, political, or social incentives and
disincentives converted the polytheistic societies or those that subscribed to
other monotheisms (Iyigun 2015, Chaps. 3, 8, 9).5 Nonetheless, even when
these interactions were of a peaceful nature, such as trade or proselytizing, one
can find in particularism we reviewed immediately above the basis for a greater
probability of conversion from P to M than from M to P.
On this basis, consider now that whenever a monotheist (M ) agent ran-
domly interacts with a polytheist (P) agent, the probability of agent P convert-
ing to M is greater than agent M converting to P. That is, let 1 ≥ m > p ≥ 0.
Consequently, the net conversion probability into monotheism, which we
defined above as a, would be strictly positive, leading to a strictly positive net
flow into monotheism of a = 2(m − p)Mt Pt = 2(m − p)Pt (1 − Pt ) > 0.6
5 By contrast, recall that polytheist societies were more tolerant of other gods and religions (Armstrong
1993, p. 49).
6 By default, if an agent of either type matches with someone of the same type, then there is no conversion.
Of course, it is conceivable and historically much relevant that if two type M agents interact and
434 M. Iyigun
Given such interactions, the equations of motion for the evolutions of P

and M over time are given by the following, respectively:
Pt +1 = −aP t (1 − Pt ) < 0 and Mt +1 = −Pt +1 = aM t (1 − Mt ) > 0
(26.1)
Based on (26.1), it is easy to verify that M and P would be constant over
time if m = p = 0. Most fundamentally, however, one can easily verify that all
agents would become type M asymptotically if a = m − p > 0 and provided
that a monotheism emerges at some point in time, that is, M0 > 0.
Conclusion
In this handbook chapter, I lay out the economic argument and historical
support for the case that monotheistic religions had very fundamental and
basic evolutionary advantages over other religions. Furthermore, I revisit some
of the earlier literature on the topic to assert that their unique doctrinal creeds
and monopolistic construction provide further efficacy in their sociopolitical,
economic, and institutional functions in society.
As potentially fruitful future research in this area, I make the case for
systematic empirical inquiry into whether or not there is any evidence for
quantitatively and statistically meaningful differences in the speed and mag-
nitude with which relatively more congregationalist versus more clerical
monotheistic sects spread and endure.
References
Armstrong, K. 1993. A History of God: The 4,000-Year Quest of Judaism, Christianity
and Islam. New York, NY: Ballantine Books.
———. 2006. The Great Transformation: The Beginning of Our Religious Traditions.
New York: Anchor Books.
they subscribe to different monotheisms, one of them would convert to the monotheism of the other.
Nonetheless, they’d both remain type M.
Also, one could enrich the dynamics of the model by allowing for spontaneous conversions (those
that occur without any interaction) as well as deliberate and strategic shunning. The former would lead to
steady states in which, even when the net conversion probabilities aren’t virtually zero, polytheism won’t
become fully extinct, and the latter would bolster this. See Iyigun and Kimball (2017) for the complete
details.
Barro, R., and R. McCleary. 2003. Religion and Economic Growth. American
Sociological Review 68 (5): 760–781.
———. 2005. Which Countries Have State Religions? Quarterly Journal of Economics
120 (4): 1331–1370.
Berman, E. 2000. Sect, Subsidy and Sacrifice: An Economist’s View of Orthodox
Jews. Quarterly Journal of Economics 115 (3): 905–953.
Berman, E., and D.D. Laitin. 2008. Religion, Terrorism and Public Goods. Journal
of Public Economics 92 (10–11): 1942–1967.
Diamond, J. 1997. Guns, Germs and Steel: The Fate of Human Societies. New York: W.
W. Norton.
Durkheim, E. 1915. The Elementary Forms of the Religious Life. Trans. Joseph Swain.
London: George Allen and Unwin.
Ekelund, R., and R.D. Tollison. 2011. Economic Origins and Monopolization of
Christianity. Chicago: University of Chicago Press.
Ekelund, R., R.D. Tollison, G.M. Anderson, R.F. Hebert, and A.B. Davidson. 1996.
Sacred Trust: The Medieval Church as an Economic Firm. New York: Oxford
University Press.
Ekelund, R., R. Hebert, and R. Tollison. 2002. An Economic Analysis of the
Protestant Reformation. Journal of Political Economy 110 (3): 646–671.
Hume, D. 1911. A Treatise of Human Nature. London: J. M. Dent and Sons.
Iannaccone, L.R. 1992. Sacrifices and Stigma: Reducing the Free-Riding in Cults,
Communes and Other Collectives. Journal of Political Economy 100 (2): 271–291.
———. 1994. Why Strict Churches Are Strong. American Journal of Sociology 99 (5):
1180–1211.
Iyigun, M. 2008. Luther and Suleyman. Quarterly Journal of Economics 123 (4): 1465–
1494.
———. 2010. Monotheism (From a Sociopolitical and Economic Perspective). IZA
Working Paper No. 3116, November.
———. 2013. Lessons from the Ottoman Harem on Ethnicity, Religion and War.
Economic Development and Cultural Change 61 (4): 693–730.
———. 2015. War, Peace & Prosperity in the Name of God. Chicago, IL: The University
of Chicago Press.
Iyigun, M., and M. Kimball. 2018. Markov Modeling of Cultural Evolution: An
Application to Theistic Dynamics. University of Colorado at Boulder, Unpublished
Mimeo, In Progress.
Janeba, E. 2007. International Trade and Consumption Network Externalities. Euro-
pean Economic Review 51 (4): 781–803.
Mokyr, J. 2010. The Enlightened Economy: An Economic History of Britain 1700–1850.
New Haven, CT: Yale University Press.
Niebuhr, R. 1932. Moral Man and Immoral Society: A Study in Ethics and Politics.
Kentucky: Westminster John Knox Press.
436 M. Iyigun
North, D., J.J. Wallis, and B.R. Weingast. 2008. Violence and Social Orders: A
Conceptual Framework for Interpreting Recorded History. Cambridge: Cambridge
University Press.
Richardson, L.F. 1960. Statistics of Deadly Quarrels. Pittsburgh, PA: Boxwood Press.
Rubin, J. 2008. The Lender’s Curse: A New Look at the Origin and Persistence of
Interest Bans Throughout History. Journal of Economic History 68 (2): 575–579.
———. 2010. Bills of Exchange, Interest Bans, and Impersonal Exchange in Islam
and Christianity. Explorations in Economic History 47 (2): 213–227.
———. 2017. Rulers, Religion and Riches. Cambridge: Cambridge University Press.
Scheidel, W. 2017. The Great Leveler: Violence and the History of Inequality from the
Stone Age to the Twenty-First Century (The Princeton Economic History of the Western
World). Princeton, NJ: Princeton University Press.
Stark, R. 2001. One True God: Historical Consequences of Monotheism. Princeton, NJ:
Wright, R. 2009. The Evolution of God. New York: Little, Brown and Company.
27
Religion, Political Power and Human Capital
Formation: Evidence from Islamic History
Eric Chaney
The flowering of scientific production in the medieval Islamic world has

received increased scholarly attention in recent years (e.g. Saliba 2007; Dallal
2010). For some, this historical episode demonstrates the compatibility of
Islam and science. For others, it represents a temporary deviation from Islam’s
obscurantist pull. These conflicting views stem partly from the fact that the
reasons behind the rise and subsequent decline in scientific production remain
poorly understood.
This chapter draws on recent and ongoing research to locate the “Golden
Age” of scientific advances in the Islamic world’s institutional history. In
particular, the rise of scientific production is linked to the waxing of secular
bureaucratic structures in the late eighth-century CE, whereas the decline
is attributed to the waning of these structures in the eleventh and twelfth
centuries.
How did the rise of a secular bureaucracy encourage scientific development?
The emergence of rationalized forms of administration increased the demand
for secular knowledge and politically empowered secular elites such as admin-
istrators, astrologers, doctors and secretaries (Crone 2006, pp. 23–24). These
I thank Jared Rubin for helpful comments. Any remaining errors are mine.
E. Chaney ()
University of Oxford, Oxford, UK
e-mail: eric.chaney@economics.ox.ac.uk

438 E. Chaney
individuals joined rulers in their patronage of scientific works both for practical
(Sabra 1996, p. 662) and for ideological reasons (Gutas 1998; David 2008).
A period of institutional change in the late eleventh and twelfth centuries,
traditionally known as the Sunni Revival, resulted in the decline of the secular
bureaucracy. This empowered religious leaders (e.g. Gibb 1982a, p. 24) who
molded the institutional environment to discourage scientific production that
undermined their societal influence.
I hypothesize that an increasingly militarized government undermined the
secular bureaucracy and allowed religious leaders to take control of civil society.
This in turn contributed to the emergence of educational institutions such
as madrasas that decreased the relative payoff to producing science. As the
payoff structure shifted in favor of the production of religious knowledge,
talent increasingly flowed away from the study of scientific topics, leading to a
decline in both the quality and quantity of scientific works produced (Baumol
1990; Murphy et al. 1991).
Conceptually, this historical account is related to the growing literature
highlighting the relationship between state capacity and economic outcomes
(e.g. Gennaioli and Rainer 2007; Michalopoulos and Papaioannou 2013;
Acemoglu et al. 2015; Dell et al. 2017). In particular, it suggests that secular
state structures can help constrain rent-seeking religious elites within civil
society. In this sense, it is consistent with Acemoglu and Robinson (2017)
who argue that growth-promoting state structures constrain (and in turn are
constrained by) civil society.
The chapter also complements the growing literature arguing that “religion
matters” in understanding differences in human capital formation rates across
societies (e.g. Mokyr 2002; Becker and Woessmann 2009; Botticini and
Eckstein 2012; Benabou et al. 2013). Consistent with Cantoni and Yuchtman
(2013) and Cantoni et al. (2018), I argue that where religious elites are more
powerful, they will favor an institutional and educational framework that
discourages human capital accumulation that could detract from their control
over the population (see Acemoglu and Autor 2012, for a related discussion).
This view suggests that better understanding these political constraints may
help clarify why some religions historically encouraged human capital forma-
tion to a greater extent than others.
Finally, this chapter is related to studies investigating the economic rise
of the Western world. The literature increasingly highlights the role of the
technological and scientific dynamism of the West as a key driver of the
divergence both within the West and between the West and the rest of the
world (e.g. Mokyr 2005; A’Hearn et al. 2009; Buringh and van Zanden
2009; Squicciarini and Voigtlander 2015). It is often forgotten, however,
27 Religion, Political Power and Human Capital Formation: Evidence… 439
that the Islamic world outstripped the West technologically and scientifically
for much of the Middle Ages. The analysis in this chapter suggests that
future research investigating how the political equilibrium in the West placed
constraints on religious leaders may provide insights into the greater scientific
and technological development of the West in the run-up to the Industrial
Revolution.
The remainder of the chapter proceeds as follows: section “Science in
the Medieval Islamic World: A Brief Background” provides a brief overview
of scientific development in the medieval Islamic world. Section “Scientific
Development in Institutional Perspective” locates these advances within the
Islamic world’s institutional history, and section “Conclusion” concludes.
Science in the Medieval Islamic World: A Brief

Background
Scholars broadly agree that for much of the medieval period Islamic societies
led the world in both technology and science (e.g. Huff 2003, p. 48). While the
factors that led to this surge in intellectual output remain a topic of debate,
its timing is reasonably well known. From the rise of Islam in the seventh
century until the start of the Abbasid Caliphate in 750 CE, the nascent Islamic
world produced relatively little scientific output.1 Abbasid Caliphs over the
following century sponsored a translation movement aimed at rendering every
available scientific text into Arabic (Gutas 1998). This translation movement
coincided with and served as a catalyst for the explosion of scientific output
that occurred in the Islamic world over the following centuries. Scientists
during this period made important advances in fields as varied as astronomy,
mathematics, medicine and optics (Kennedy 1970, p. 337).2 Indeed, many
scientific works from the medieval Islamic world were eventually translated
into Latin and are believed to have played a central role in the scientific
1 In the century following the death of the Prophet Muhammad in 632 CE, Arab-Islamic armies conquered
a vast territory reaching from modern-day France to Pakistan. The immediate successors to Muhammad
(known as the Rashidun Caliphs) were followed by the Umayyad dynasty in 661 which was replaced,
in turn, by the Abbasid dynasty in 750. Contrary to what is implied in many studies, this dynasty was
relatively short-lived, at least in its ability to directly control territory. Generally speaking, after 945 the
Abbasid Caliphs no longer controlled territory and primarily provided their blessing upon the true holders
of power across the Islamic world. In 1258, this state of affairs came to an end when the last Abbasid Caliph
was killed by the Mongols.
2 The term scientist is, admittedly, anachronistic. Throughout, I use this term in place of others such as
natural philosopher for expositional ease.
440 E. Chaney
development of Western Europe during the late medieval and early modern
periods (e.g. Lindberg 1978).
According to Brentjes (2009, p. 305), there were “two major periods for
the patronage of scientific knowledge,” the first spanning roughly the eighth
to the twelfth centuries and the second running from the twelfth to the
nineteenth centuries. During the first period, rulers, bureaucrats and wealthy
urban groups funded scientific output and also established institutions such
as libraries where scientific topics were studied (e.g. Brentjes 2009, p. 305).
Rulers and other wealthy individuals patronized scientists for both prestige
(David 2008) and for the “practical benefits promised by the practitioners of
medicine and astronomy and astrology and applied mathematics” (Sabra 1996,
p. 662).
In the second period, funding for intellectual pursuits shifted to religious
institutions such as madrasas. Unlike the direct patronage system of the first
period, in the second period rulers and wealthy individuals endowed religious
institutions dedicated to knowledge production. Prospective scholars who
wished to be appointed to these posts generally had to specialize in the
production of religious knowledge (Makdisi 1981, p. 285).
One line of scholarship implies that this shift in the patronage structures
was the product of the Sunni Revival which began sometime in the eleventh
century. Traditionally, scholars have stressed the ideological aspects of this shift
arguing that the Revival marked the final triumph of traditionalist religious
leaders in their battle against rationalism (e.g. Makdisi 1973, p. 168). As
traditionalists gained societal influence, madrasas replaced institutions such
as libraries focused on scientific research (Makdisi 1981, p. 10).3
Historical evidence suggests that traditionalist religious leaders worked
to orchestrate these institutional changes. These elites had long criticized
rationalist interpretations of Islam, claiming that they led to a loss of belief
(e.g. Kraemer 1986, p. 72). Madrasas are generally believed to have served the
traditionalist agenda by discouraging rationalist strains of Islam, unrestricted
study of scientific topics and innovation more broadly (e.g. Talbani 1996).4
Other explanations for the medieval decline of scientific output compete
with the Revival hypothesis in the literature. For centuries, scholars have
3 While there are isolated examples of madrasas solely dedicated to funding scientific endeavors, most
available posts were dedicated to the production of religious knowledge (e.g. Brentjes 2009, p. 313).
4 Rationalism, innovation and the unrestricted study of science were believed to lead individuals to a
loss of belief. Consistent with this view, Stroumsa (1999) provides examples of the deist/atheist figures
that emerged during the Golden Age. Also, see Glaeser and Sacerdote (2008) for evidence of a negative
relationship between religious beliefs and human capital formation.
stressed the role that the Mongol invasions of the thirteenth century played
in bringing about the medieval decline. Although the mechanisms through
which these shocks affected scientific output remain a topic of debate, the
existing scholarship often stresses the importance of the destruction of physical
and human capital as well as institutional and cultural changes (for one critical
review of this hypothesis, see Saliba 2007, p. 235).
In recent years, a new line of scholarship has emerged challenging the entire
medieval decline narrative. For example, Saliba (2007) provides overwhelming
evidence that important advances continued to be produced in the Islamic
world long after the supposed decline of scientific production. He suggests
that the decline did not begin until the sixteenth century.
Chaney (2016) provides new empirical evidence suggesting that scientific
output declined in the late eleventh century. Such evidence casts doubt
on both Mongol and revisionist narratives, suggesting the importance of
better understanding the mechanisms through which the Revival dampened
scientific production.
Scientific Development in Institutional

Perspective
Existing studies of scientific development in the Islamic world tend to abstract
from the region’s institutional and political history. In this section, I argue
that this political history can explain both the rise and subsequent decline of
scientific activity during the medieval period.
The basic premise underlying this hypothesis is consistent with recent
evidence linking the rise of scientific activity in the medieval Islamic world to
institutional changes. For example, Gutas (1998) recounts how the emergence
of the Abbasid Caliphate led to a resurgence of pre-Islamic institutional forms
which led to the translation movement and subsequent scientific activity. In
particular, his analysis suggests that the rise of the secular bureaucracy during
this period was “instrumental” in stimulating scientific development and that
state functionaries constituted “one of the most important social groups”
patronizing scientific activities (Gutas 1998, p. 128).5
Scholars have long noted that the rise of the Abbasid Caliphate represented a
break with previous institutional arrangements. Importantly, the Abbasid Rev-
olution politically empowered non-Arab (initially largely Persian) bureaucratic
5 The term secular bureaucracy has been used by Goitein (1957, p. 597), among others.
442 E. Chaney
elites who aimed to remold institutions on pre-Islamic Iranian patterns (Gibb

1982b, p. 66). The increased political power of this group, in turn, stemmed
from the fact that both Persianized Arabs and Persians played a central role
in bringing the Abbasids to power (Gutas 1998, p. 29). Thus, the Abbasid
Revolution led to a marked change in both the size and political influence of
the secular bureaucracy (e.g. Biddle 1972, p. v), and by the start of the ninth
century, this bureaucracy had become “a major political and cultural force”
(Kennedy 2004, p. 146).
How did the rise of a secular bureaucracy encourage scientific develop-
ment? The emergence of rationalized forms of administration increased the
demand for secular knowledge and politically empowered secular elites such
as administrators, astrologers, doctors and secretaries (Crone 2006, pp. 23–
24). These individuals joined rulers in their patronage of scientific works
both for practical (Sabra 1996, p. 662) and for ideological reasons (Gutas
1998; David 2008). Saliba (2007, pp. 58–64) argues that competition between
bureaucrats also worked to stimulate scientific output.6 As the bureaucracy
expanded and competition increased, incumbent bureaucrats “resorted to
higher specialization through the translation of the more advanced sciences” to
secure their positions (Saliba 2007, p. 65). Regardless of the exact mechanisms
at work, the educational needs of these bureaucrats are generally viewed as
a central impetus for scientific development (e.g. Gutas 1998, p. 115). For
example, algebra is believed to have been developed, at least in part, to address
engineering and irrigation problems bureaucrats faced at the time (Gutas 1998,
p. 113).
Although the rise of the secular bureaucracy had benign consequences for
scientific development, it also reflected an attempt to build autocratic state
structures. Opposition to this project began early on as disaffected groups
began to organize opposition to government policy around religious ideals.
These groups are generally viewed as proto-ulama or the first of the religious
elites that were to play a central role throughout much of Islamic history
(Hodgson 1974a, p. 250).
As Arab elites in the Umayyad dynasty became increasingly autocratic, the
distance between these religious elites and the ruling class grew. This process
continued under the Abbasids and played a central role in the development
of Islamic law (Hallaq 2009, pp. 128–129). Some scholars have viewed the
development of this law—which largely developed outside state structures—as
6 It is worth noting that Saliba (2007) sees the rise of the translation movement in this competition, which
in his view is rooted in the bureaucratic reforms of the late Umayyad period.
an attempt to constrain increasingly autocratic caliphs (e.g. Crone and Hinds

2003, pp. 108–109).
The process is believed to have come to a head during the Abbasid Caliphate
during what is known as the Mihna (833–848). During this period, the
Abbasid Caliph attempted to undermine this growing opposition by forcing
religious leaders to adhere to rationalist interpretations of Islam. After this
attempt failed, Abbasid Caliphs introduced the slave soldier system that
characterized much of the Islamic world over the following centuries (Crone
1980).
The timing of the failure of the Mihna and the introduction of slave soldiers
suggests that these two events may have been related. In particular, it seems
plausible that the introduction of slave soldiers marked an attempt by caliphs
to circumvent the growing constraints that the autonomous development of
Islamic law threatened to place on their rule (Crone and Hinds 2003, pp. 108–
109). Recent analyses support this claim (Blaydes and Chaney 2013; Chaney
2013; Blaydes and Chaney 2016).7
Slave Soldiers, Militarization of Government Structures

and Scientific Decline
The introduction of armies staffed by slaves marked a “major innovation in

Middle Eastern history” (Lapidus 2014, p. 86). Although the use of military
slaves began in the ninth century, this institution did not fully develop until the
tenth century (Kennedy 2004, p. 204).8 Although the introduction of slave
soldiers is often viewed as an attempt to create a loyal military, these slaves “had
a disconcerting tendency to take over the state and tear it apart from within”
(Tor 2008, p. 223).
The historian Miskawayh (who died in 1030 CE) records how the introduc-
tion of slave soldiers undermined bureaucratic structures. After the introduc-
tion of slave soldiers, he notes how “the bureaucracy was shut down and the
bureaucratic secretaries and tax collectors disappeared [… instead the land was
managed by] slave soldiers and agents who did not take care of what was in
their power or manage the land in a fruitful manner” (Miskawayh 1915 [1030],
pp. 97–98). Lapidus (2014, p. 179) sums up such evidence, noting how these
soldiers “destroyed established elites.”
7 More broadly, the idea that slave soldiers were introduced to reduce constraints on the caliphs is consistent
with both Fukuyama (2011, p. 192) and Lapidus (1982, p. 717).
8 See Blaydes and Chaney (2013) for an in-depth overview of this institution.
444 E. Chaney
This process culminated in the institutional changes known as the Sunni

Revival in the second half of the eleventh century. During this period, the
secular bureaucracy contracted in what has been described as the militariza-
tion of final governmental power (Hodgson 1974b, p. 64). As the secular
bureaucracy contracted, religious leaders filled the void, emerging as social and
political elites in addition to religious ones (Gilliot et al. 2012). Thus, during
the Revival, civil society “came under the purview of Islam” as religious scholars
“took charge of judicial administration, local police, irrigation, public works
and taxation” (Lapidus 1996, pp. 13–14).
During this period the madrasa spread across much of the Islamic world, in
what is widely viewed as a sign of the increased political influence of religious
elites. As religious leaders gained political power, they could have continued
to support scientific research in madrasas or other waqf institutions had they
desired. The available evidence, however, suggests that while some scholars
employed in madrasas continued to work on scientific topics, the sciences were
“banished from the regular course of institutionalized education” under the law
of waqf (Makdisi 1981, p. 283).
The evidence suggests that religious elites used their increased influence to
limit the study of science because they believed that the unrestricted study of
science led to religious skepticism which undermined, in turn, their control
over the population. Such skepticism had been widespread among the secular
bureaucrats who were “often disinclined to defer to religious scholars” (Crone
2006, pp. 23–24). To prevent such skepticism, religious elites advocated
restricting the study of science to individuals who could be trusted to not
go astray (in practice, primarily religious leaders themselves).9 This desire to
eradicate skepticism appears to have guided the crafting of the post-Revival
education system. Thus, in addition to limiting the study of science, this
system is believed to have discouraged “questioning, verifying, criticizing, eval-
uating and making judgements” while encouraging “the uncritical acceptance
of authority” (Halstead 2004, p. 526).
As madrasa-employed religious leaders worked with military elites to pre-
vent “unfettered private inquiry by persons not trained to come to the right
answers,” the “liberty of opinion” that had characterized the previous centuries
came to a close (Hodgson 1974b, p. 192). Chaney (2016) provides empirical
9 For example, al-Ghazālı̄ notes that those who study sciences (in this case geometry and arithmetic often)
“have gone beyond their study into innovations (i.e. non-orthodox beliefs) and thus the weak (in faith)
should be barred from their study like the boy is barred from the bank of the river so that he does not
fall in” (al-Ghazālı̄ 2005, p. 31). Sabra (1987, p. 237) notes that after the Revival, scientists “were imbued
with Muslim learning and tradition.”
evidence consistent with this historical narrative, showing that scientific

production experienced a significant drop during the Revival.
Conclusion
This chapter has argued that the Islamic world’s Golden Age of scientific pro-
duction was closely linked to the region’s secular bureaucratic structures. I have
provided evidence that these institutions encouraged scientific production by
both increasing demand for scientific output and by constraining religious
leaders opposed to the study of scientific topics. The introduction of slave
soldiers and the eventual militarization of government led to the collapse of
this bureaucracy and the empowerment of religious leaders. These religious
elites worked to restrict scientific production that undermined their societal
position.
The evidence presented in this chapter weighs against claims that Islam
is uniquely inimical to scientific development. Instead, and consistent with
Rubin (2017), it argues that secular state structures can play a fundamental
role in constraining rent-seeking religious elites aiming to restrict scientific
research. In this sense, it suggests strengthening secular institutions across the
Islamic world may encourage scientific production today.
References
Acemoglu, D., C. Garcia-Jimeno, and J. Robinson. 2015. State Capacity and Eco-
nomic Development: A Network Approach. American Economic Review 105 (8):
2364–2409.
Acemoglu, D., and D. Autor. 2012. What Does Human Capital Do? A Review
of Goldin and Katz’s The Race between Education and Technology. Journal of
Economic Literature 50 (2): 426–463.
Acemoglu, D., and J. Robinson. 2017. The Emergence of Weak, Despotic and
Inclusive States. NBER Working Paper 23657.
A’Hearn, B., J. Baten, and D. Crayen. 2009. Quantifying Quantitative Literacy: Age
Heaping and the History of Human Capital. Journal of Economic History 69 (3):
783–808.
al-Ghazālı̄, A. (2005). Ih.yā’ ’ulūm ad-dı̄n, 1111. Beirut: Dār Ibn H
. azm.
Baumol, W. 1990. Entrepreneurship: Productive, Unproductive, and Destructive. The
Journal of Political Economy 98 (5): 893–921.
Becker, S., and L. Woessmann. 2009. Was Weber Wrong? A Human Capital Theory
of Protestant Economic History. Quarterly Journal of Economics 124 (2): 531–596.
446 E. Chaney
Benabou, R., D. Ticchi, and A. Vindigni. 2013. Forbidden Fruits: The Political
Economy of Science, Religion, and Growth. New York: Mimeo.
Biddle, D. 1972. The Development of the Bureaucracy of the Islamic Empire During
the Late Umayyad and Early Abbasid Period. PhD dissertation, The University of
Texas at Austin.
Blaydes, L., and E. Chaney. 2013. The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian West and the Muslim World before 1500 CE. American
Blaydes, L., and E. Chaney. 2016. The Political Economy Legacy of Institutions from
the Classical Period of Islam. In The New Palgrave Dictionary of Economics, ed.
S. Durlauf and L. Blume. London: Palgrave Macmillan.
Botticini, M., and Z. Eckstein. 2012. The Chosen Few: How Education Shaped Jewish
History. Princeton: Princeton University Press.
Brentjes, S. 2009. Patronage of the Mathematical Sciences in Islamic Societies. In
The Oxford Handbook of the History of Mathematics, ed. E. Robson and J. Stedall.
Buringh, E., and J.L. van Zanden. 2009. Charting the “Rise of the West”: Manuscripts
and Printed Books in Europe, A Long-Term Perspective from the Sixth through
Eighteenth Centuries. Journal of Economic History 69 (2): 409–445.
Cantoni, D., J. Dittmar, and N. Yuchtman. 2018. Religious Competition and Real-
location: The Political Economy of Secularization in the Protestant Reformation.
The Quarterly Journal of Economics 133 (4): 2037–2096.
Cantoni, D., and N. Yuchtman. 2013. The Political Economy of Educational Content
and Development: Lessons from History. Journal of Development Economics 104:
233–244.
Chaney, E. 2013. Revolt on the Nile: Economic Shocks, Religion and Political Power.
Econometrica 81 (5): 2033–2053.
Chaney, E. 2016. Religion and the Rise and Fall of Islamic Science. Working Paper.
Crone, P. 1980. Slaves on Horses. Cambridge: Cambridge University Press.
Crone, P. 2006. Post-Colonialism in Tenth-Century Islam. Der Islam 83 (1): 2–38.
Crone, P., and M. Hinds. 2003. God’s Caliph: Religious Authority in the First Centuries
of Islam. Cambridge: Cambridge University Press.
Dallal, A. 2010. Islam, Science, and the Challenge of History. New Haven: Yale
University Press.
David, P. 2008. The Historical Origins of ‘Open Science’: An Essay on Patronage,
Reputation and Common Agency Contracting in the Scientific Revolution. Cap-
italism and Society 3 (2): 1–103.
Dell, M., N. Lane, and P. Querubin. 2017. The Historical State, Local Collective
Action, and Economic Development in Vietnam. Working Paper.
Fukuyama, F. 2011. The Origins of Political Order. New York: Farrar, Straus and
Giroux.
Gennaioli, N., and I. Rainer. 2007. The Modern Impact of Precolonial Centralization
in Africa. Journal of Economic Growth 12 (3): 185–234.
Gibb, H.A.R. 1982a. An Interpretation of Islamic History. In Studies of the Civilization

of Islam, ed. S. Shaw and W. Polk, 3–33. Princeton: Princeton University Press.
Gibb, H.A.R. 1982b. The Social Significance of the Shuubiya. In Studies of the
Civilization of Islam, ed. S. Shaw and W. Polk, 62–73. Princeton: Princeton
University Press.
Gilliot, C., R. Repp, K. Nizami, M. Hooker, C. Lin, and J. Hunwick. 2012. Ulama.
In Encyclopedia of Islam, ed. P. Bearman, T. Bianquis, C. Bosworth, E. van Donzel,
and W. Heinrichs, 2nd ed. Leiden: Brill.
Glaeser, E., and B. Sacerdote. 2008. Education and Religion. Journal of Human
Capital 2 (2): 188–215.
Goitein, S.D. 1957. The Rise of the Near-Eastern Bourgeoisie in Early Islamic Times.
Journal of World History 3 (3): 583–604.
Gutas, D. 1998. Greek Thought, Arabic Culture: The Graeco-Arabic Translation Move-
ment in Baghdad and Early ‘Abbāsid Society. New York: Routledge.
Hallaq, W. 2009. Shari’a. Cambridge: Cambridge University Press.
Halstead, J.M. 2004. An Islamic Concept of Education. Comparative Education 40
(4): 517–529.
Hodgson, M. 1974a. The Venture of Islam, vol. 1. Chicago: University of Chicago Press.
Hodgson, M. 1974b. The Venture of Islam, vol. 2. Chicago: University of Chicago
Press.
Huff, T. 2003. The Rise of Early Modern Science: Islam, China and the West. Cambridge:
Kennedy, E.S. 1970. The Arabic Heritage in the Exact Sciences. Al-Abhath 23:
327–344.
Kennedy, H. 2004. The Prophet and the Age of the Caliphates. Harlow: Pearson
Education.
Kraemer, J. 1986. Humanism in the Renaissance of Islam. Leiden: E.J. Brill.
Lapidus, I. 1982. Review of Slave Soldiers and Islam: Genesis of the Military System
by Daniel Pipes. Journal of Interdisciplinary History 12 (4): 716–718.
Lapidus, I. 1996. State and Religion in Islamic Societies. Past and Present 151: 3–27.
Lapidus, I. 2014. A History of Islamic Societies, New York: Cambridge University Press.
Lindberg, D. 1978. The Transmission of Greek and Arabic Learning to the West. In
Science in the Middle Ages, ed. D. Lindberg. Chicago: The University of Chicago
Press.
Makdisi, G. 1973. The Sunni Revival. In Islamic Civilization, 950-1150, ed. D.H.
Richards, 155–168. Oxford: Bruno Cassirer.
Makdisi, G. 1981. The Rise of Colleges. Edinburgh: Edinburgh University Press.
Michalopoulos, S., and E. Papaioannou. 2013. Pre-Colonial Ethnic Institutions and
Contemporary African Development. Econometrica 81 (1): 113–152.
Miskawayh, A. 1915 [1030]. Tajārib al-Umam. Cairo: al-Tamadun al-Sina’iya.
Mokyr, J. 2002. The Gifts of Athena. Princeton: Princeton University Press.
Mokyr, J. 2005. The Intellectual Origins of Modern Economic Growth. Journal of
Economic History 65 (2): 285–351.
448 E. Chaney
Murphy, K., A. Shleifer, and R. Vishny. 1991. The Allocation of Talent: Implications
for Growth. The Quarterly Journal of Economics 106 (2): 503–530.
Rubin, J. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East
Did Not. New York: Cambridge University Press.
Sabra, A.I. 1987. The Appropriation and Subsequent Naturalization of Greek Science
in Medieval Islam: A Preliminary Statement. History of Science 25: 223–243.
Sabra, A.I. 1996. Situating Arabic Science: Locality versus Essence. Isis 87 (4):
654–670.
Saliba, G. 2007. Islamic Science and the Making of the European Renaissance. Cam-
bridge: The MIT Press.
Squicciarini, M., and N. Voigtlander. 2015. Human Capital and Industrialization:
Evidence from the Age of Enlightenment. Quarterly Journal of Economics 130 (4):
1825–1883.
Stroumsa, S. 1999. Freethinkers of Medieval Islam. Leiden: Brill.
Talbani, A. 1996. Pedagogy, Power, and Discourse: Transformation of Islamic Educa-
tion. Comparative Education Review 40 (1): 66–82.
Tor, D.G. 2008. The Mamluks in the Military of the Pre-Seljuq Persianate Dynasties.
British Institute of Persian Studies 46: 213–225.
28
Islam, Trade, and Innovation
Alireza Naghavi
Introduction
Trade fosters innovation. This is by now a well-established fact in the eco-
nomics literature. The seminal theoretical work on trade, innovation, and
growth goes back to Grossman and Helpman (1989, 1990), among others.
Theories that follow suggest that trade liberalization essentially affects inno-
vation through two channels. The first channel is improved access to foreign
markets. Higher profits brought about by a larger market create incentives
to innovate (Acemoglu and Linn 2004). Trade also opens the gate to a more
competitive marketplace, which in turn creates a threat to profits and induces
firms to innovate new varieties to continue to benefit from monopoly rents
(Aghion et al. 2005). Grossman and Helpman (2018) also show an adverse
effect of competition as successful innovators must share the market with
foreign competitors. They, however, show that the market size effect offsets
this negative effect and that open economies tend to innovate and grow more
than closed ones. Furthermore, Coelli et al. (2016) show striking evidence that
indeed around seven percent of the increase in global knowledge over the 1990s
can be attributed to freer trade.
The trade and innovation nexus tends to be absent in the Muslim world
today. Naghavi and Pignataro (2018) show that countries with the highest
A. Naghavi ()
University of Bologna, Bologna, Italy
e-mail: alireza.naghavi@unibo.it

450 A. Naghavi
percentage of Muslim population tend to produce the lowest level of patents

worldwide. These countries practically include the core of the Islam, or
historically the greater Muslim empire, and consist of Iran, Saudi Arabia,
Syria, Tunisia, Egypt, Yemen, Algeria, Jordan, Pakistan, Morocco, Bangladesh,
Tajikistan, Uzbekistan, Turkey, and Sudan. At the same time, we know that
Islam is a religion of trade. Prophet Muhammad was himself a Meccan
merchant. A large part of those involved in the gradual manifestation of Islamic
law were also craftsmen or merchants or at least came from such backgrounds
(Cohen 1970). Also, trade as a profession held a more favored position in the
Islamic world than in Europe (Skeen 2008). Michalopoulos et al. (2016) argue
that trade played a critical role in the formation of Islamic principles. A main
argument they put forth is the role of Islam as an agreement to secure safe
passage across the desert against raids and ensure redistribution required a
commitment device to make cooperation between tribes viable.
Trade and Islamic Principles

Islam is particularly known to have conferred economic benefits to indi-
viduals through the institutional advantages it created by facilitating trade.
Michalopoulos et al. (2016) describe the role of the waqf as a dynamic
redistribution system that spread wealth inter-generationally and stalled its
concentration, making Islam work as a self-enforcing cooperation treaty that
unified tribes of different natures into an Islamic community. Ensminger
(1997) offers case studies comparing trade within Muslim and non-Muslim
indigenous African societies, stressing how the institutional guarantee of Islam
provided an additional impetus for trade by facilitating the flow of credit. An
important advantage of Islam with respect to previous arrangements was the
fact that it offered a powerful ideology with built-in sanctions that created
non-material interest in holding to the terms of contracts. The existence of
informal and formal punishments in Islam, such as those related to ridda
(apostasy), gave Islam an edge over similar pre-existing, short-lived attempts.
Although the focus of Michalopoulos et al. (2016) was the mitigation of
conflict between trading merchants and Bedouin tribes in arid lands though
the Islamic institutional complex, a reduced form variant of the theory can be
used to illustrate the role of Islam as a commitment device in the following
game.
Think of merchants traveling on trade routes who must cross the desert
and the possibility of their caravans being raided by Bedouins, who lack
merchandise to engage in long-distance trade themselves. A confrontation
between the trading caravan and the Bedouins may take place in the desert and
result of a loss of their cargo. There is, however, the possibility of redistributing
a part of the potential profits from trade to the Bedouins. An agreement could
be made between the two sides to furnish a redistribution of rents from those
who can benefit from trade to the less-advantaged portion of the society and
secure safe passage across the route. It is easy to show a game, in which the
unique subgame perfect Nash equilibrium would be to not redistribute by
the merchants and raid by the Bedouins, resembling a prisoner’s dilemma
interaction where both sides end up in a socially inefficient equilibrium. This
situation illustrates why the pre-Islamic redistribution attempts between tribes
in the Arabian Peninsula, such as the case of ilaf, were temporary, highlighting
the need for a credible commitment device to achieve the socially efficient
outcome. This was the role played by Islam as a religion. Indeed, the terms
of Muhammad’s agreements with the tribes were always the same: the tribe
agreed to pay the zakat and to refrain from attacking other Muslims and their
allies (Lewis 1993). Adding an apostacy cost or, in general, a cost of deviating
from Muslim law that proposes a (Redistribute, Not Raid) equilibrium serves
as a credible punishment device to make the cooperative solution the unique
equilibrium outcome which could also be Pareto optimal.
Intellectual Property and Islamic Law

The mechanism is not the same, if not the opposite, with innovation in Islam.
Naghavi and Pignataro (2018) highlight how it is exactly the fundamentals
in Islamic principles that could lead to the demise of innovation in Muslim
countries. This occurs due to the existence of a fallacy with respect to the
protection of intellectual property in Islamic law. Although the concept of
protection is more emphasized in Islam than other religions, the enforcement
is left to the morality of the agents in the society creating a situation of adverse
selection. The intellectual property environment is different in other religions,
where legal institutions rather than religious scripts take charge in enforcing
intellectual property rights.
The Sharia law includes clauses that resemble modern intellectual property
considerations. The Caliphs in early Islam, for example, bought books and
only made copies after paying an adequate compensation to the author
(Malkawi 2013). Indeed, Sharia draws close connections between intellectual
property and Islamic law, albeit not explicitly regulating intellectual property
rights with precise rules as it does for tangible property rights. In short,
ownership of property can be acquired by appropriation, under which one
452 A. Naghavi
may acquire ownership by developing it and making it productive to reward

individuals for the fruits of their effort. The Sharia text does not limit such
ownership to tangible assets and is hence compatible with offering economic
incentive to intellectual property owners by entitling them to exclusive right
to enjoy the benefits of their creativity.
Nevertheless, the prohibition of intellectual property rights piracy in Islam
has been limited to the Muslim judges’ use of qiyas, though without a clear
punishment such as that for material theft. An interesting analogy regarding
the separation of tangible and intangible assets in Islam lies in the Hedaya,
which suggests that one does not amputate the hand of a thief for stealing
a book because the thief’s intention is not to steal the book as paper, but
the ideas in the book, which is not tangible property (Al-Marghinani 1791).
The existence of this gray area when partial enforcement is left to Sharia has
impeding consequences for business relationships and investment incentives
that are key for innovation and economic development. Naghavi and Pignataro
(2018) show that formal enforcement of intellectual property rights through
legal institutions that are not linked to religious guidance could break the
fallacy that emerges in a world of asymmetric information by encouraging
moral agents to engage in technological effort in sectors less sensitive to
intellectual property rights while preventing expropriation by the amoral
agents in sectors with more valuable technologies.
Trade as the Powerful Gravity of Islam

Michalopoulos et al. (2018) provide evidence that a main factor in the fast
and expansive spread of Islam was the interests of merchants who wanted
to gain access to the widest possible network of trade routes. Indeed, they
show Islam predominantly spread across trade routes in the Old World. They
argue that after the initial phase of the expansion of Islam throughout the
first century after its birth within a certain radius around Mecca through
military conquests, the religion gained control over a vast strategic location
from North Africa passing through the Middle East all the way to the Indian
sub-continent. This put the Islamic region in the center of a network of the
most important traditional trade routes that connected the Old World and
included the whole Western half of the Silk Road. It also linked maritime
trade of the Mediterranean and the Indian Ocean. Merchants in those regions
therefore had extra incentives to adopt the religion to gain advantages via
cooperation and favorable laws reserved only within Muslim trading networks
and access to valuable contacts to expand their trade. Islam at the same time
spread along trade routes through propagation as Muslims themselves traveled

long distance along the routes when they engaged in trade.
It is likely that trade and innovation went together and reinforced each other
in the advent of Islam, during the flourishing years of the Muslim territories.
The Islamic empire stretched from Spain to the borders of China in inner Asia,
creating a unique unified economic region, facilitating movement of not only
merchants and goods but also craftsmen and scholars. The ease of trade in
the Islamic also allowed new ideas, techniques, and inventions to travel freely.
Skeen (2008) provides a simple example of a rug made in China that could
be traded as far as Algeria within a year, and the technique of its manufacture
could be copied by local craftsmen. Last, but not least, trade also nurtured a
swift movement of books across the Islamic world.
The Big Bang and the Big Crunch of Innovation

in the Islamic World
Science and technology in the medieval Muslim world surpassed that in the
West and China by far (Huff 2003). Several very practical and important
innovations took place during the golden age of Islamic civilization under
the Abbasid dynasty, especially in agriculture. Improved methods of irrigation
allowed more land to be cultivated, and new types of mills and turbines were
used to reduce the need for labor. The cultivation of rice and the production of
textile were among major innovations of the era. Crops and farming techniques
were adopted from remote neighboring cultures such as India, China, and
Africa. Islamic farmers contributed to eventually bringing crops such as sugar
and citrus fruits to the West and distributing the cultivation of cotton to Iraq
and Egypt (Watson 1983). Other major scientific inventions associated with
the Abbasids were in the fields of mathematics (algebra and geometry), science,
astronomy, medicine, and anatomy. In medicine, the ninth century saw over
800 doctors in Baghdad, and many discoveries with regard to anatomy and
diseases were achieved. Also, the invention of techniques to make thicker,
more useful paper was an achievement of Baghdad (The Saylor Foundation
n.d.; Syed et al. 2011).
The reborn Umayyad caliphate in Córdoba, after being overthrown by
the Abbasid revolution and their relegation to al-Andalus of Spain, also
represented a peak of scientific and artistic achievements of the Muslim
world. The Umayyads wanted to be seen as intellectual competitors of the
Abbasids who led to their collapse, and pushed Córdoba to have libraries and
454 A. Naghavi
educational institutions to rival those of Baghdad. Despite the rivalry between

the two powers, which created competition, freedom to travel between the
two Caliphates was allowed and helped spread fresh ideas and innovations
over time. There was, however, a radical switch in the link between Islam and
innovation in the aftermath of the Umayyads. This occurred when smaller
unstable independent Taifas first took power for about 50 years, after which
they were forced to call for the help of North African (Berber) caliphates.
The Almoravids came to their rescue but soon took control themselves and
were succeeded by the Almohads. These dynasties were more aggressive, and
at the same time more unstable (and hence under threat), lasting 70 years
each and losing more and more territory over time during the Reconquista
period. In principle, they were more of a ruling military elite in al-Andalus
than intellectuals. Under the Almoravids and Almohads, there was a decline in
cultural and social exchange and increased persecution of religious minorities,
with a switch to more fundamentalist forms of Islam (Jayyusi 1994).
Innovation as a Sacrifice for Power

The great economic divergence described in depth in Kuran (2010) has been
related to the nature of Islam discouraging curiosity and preventing risk-taking
activities by the Muslim society. In line with the narratives on religion and
science in medieval Spain, Cinnirella et al. (2018) look directly at intellectual
achievements in Spanish municipalities based on the number of years they
have been under Muslim rule. Looking at literacy rate in 1900 and the
number of historical patents achieved during the period of 1878–1939 in each
municipality, they find that both measures of human capital or innovation are
lower for municipalities that have been under Muslim rule for a longer period.
There are several factors and historical narratives that can explain the negative
correlation between religion, or particularly Islam, and innovation. I pick three
recent interrelated analyses that are relevant for the argument in this chapter
and provide a brief description of each below.
A recent analysis on the potential role of Islam in impeding innovation
comes in Chaney (2016). Collecting data on intellectual production (from
Harvard’s library collection and a catalog of books from seventeenth-century
Istanbul), he establishes that Islamic societies flourished with scientific and
technological production in the medieval period, but produce a dispropor-
tionately small share of world scientific output today. He focuses on the Sunni
Revival, namely, the eleventh-century surge in political power of religious lead-
ers, as the decisive historical landmark. The hypothesis argues that increased
political power came together with a replacement of scientific institutions

that contributed to human capital development with religious educational
centers such as madrasas. The elite promoted religious institutions as centers
of knowledge production by requiring scholars to specialize in religious
knowledge as a prerequisite to being appointed to any posts. This reduced
the payoff to the production of scientific knowledge and shifted preferences
in favor of production of religious knowledge. The evidence revolves around
the newly empowered religious leaders and how they intentionally limited the
production of scientific knowledge to avoid tampering with religious beliefs in
society. The hypothesis is in line with previous argument in Mokyr (2002) that
human capital formation and religion can be interrelated. In this context, we
are talking about a political economy setting where religious elites believe that
human capital accumulation could weaken their grip on power and control
over the population.
Bènabou et al. (2015) is another contribution in literature that refers to the
clash between innovation and religious doctrines in general. They show that
the number of patents filed is decreasing in both religiosity and belief in god
both across countries and across US states. Here the blame goes to a state
that incorporates religious rule, which may repress new discoveries to protect
religious beliefs. The model provided views that the diffusion of scientific
discoveries brings productivity gains, but also neglects religious beliefs by
contradicting the doctrine. The key mechanism suggested in this study is
the potential belief-eroding characteristic of innovation. The government can
impede innovation from spreading by blocking the spread of belief-eroding
science. Muslim authorities started perceiving innovations as a potential threat
in the beginning of the twelfth century; the anti-science attitudes are still
observable in the Muslim world today. The most striking case of knowledge
blocking in the case of Islam is the decree in the Ottoman Empire that ruled the
practice of printing punishable by death until the beginning of the nineteenth
century.
Muslim religious authorities used their political power to advocate laws that
inhibited economic development, such as regulations that inhibited the spread
of knowledge (against the printing press) and human capital development
(against mass education). Conservative Islamic clerics in the background who
enjoyed a strong political seat saw such important advancements as a threat to
their power, and the state which was strongly tied to religion saw it in their
own interest to rule in their favor (Rubin 2017). The crux of the argument
behind the Muslim anti-printing attitude has been described to be the reliance
of political authorities under Islam on legitimacy by religious authorities,
the attainment of which depended on their conforming to the correspond-
456 A. Naghavi
ing religious rules. This environment reduced enthusiasm for introducing

novel technologies, including the printing press (Coşgel et al. 2012). More
specifically, the power of religious authorities depended on their role in the
transmission of knowledge, an essentially oral process in early Ottoman society
prior to the introduction of the printing press. Skepticism in Islam with regard
to unsupervised writings as reliable means of communication granted religious
authorities a monopoly position in the transmission of knowledge. They used
their position to provide Ottoman rulers with legitimacy by producing loyalty,
in return for compensations such as the incorporation of religious institutions
into the state. Ottoman rulers banned printing because they believed it
may undermine the comparative advantage of religious authorities to confer
legitimacy, thereby increasing the costs of tax collection. The Ottomans
eventually granted permission to printing in Arabic script in the eighteenth
century only upon the emergence of new sources of legitimacy (Rubin 2017).
The Missing Role of Tolerance

Tolerance is a concept often associated with the impact of religion on inno-
vation. The argument on tolerance is not confined to Islam, but works for
any belief by creating diversity and synergy between diverse people (Cinnirella
and Streb 2017). The Muslim empire is known to have been tolerant toward
followers of other religions (Lapidus 2002). Also, the output of Arab science
during the Islamic Golden Age shows that there is no fundamental barrier
to tolerance and progress in the Islamic world (Ofek 2011). Iyer et al. (2017)
show that even when Muslims were engaged in destroying holy sites of other
religions, they did not oppose their presence. The unique evidence found in
their research shows that Muslims desecrated temples to diminish the authority
of the enemy, the Hindu state, during war.
Interestingly, Jha (2013) shows that a factor that has led to religious tolerance
in Islam is trade. Due to advantages enjoyed by Muslims brought about by
full control over Indian Ocean shipping, ethnic complementarities between
them and Hindus for exchange and provision of services were highest in
medieval trading ports. This led to development of institutional mechanisms
that supported interethnic cooperation, and reduced probability of riots and
violence between different ethnic groups of the two religions. The role of
trade on religious tolerance has also been given several alternative compatible
explanations. Clingingsmith et al. (2009) for instance connects tolerance to
Islam through the role of the Hajj pilgrimage, a pillar of Islam. Their study
argues that a global gathering as such creates unity within an Islamic world
through exposure to global Islamic practices, with a spillover of increased

belief in peace, harmony, and equality among adherents of different religions.
This concept indirectly links with trade and innovation as Hajj gatherings
also facilitated economic trade and the diffusion of economic, cultural, and
scientific ideas (Bose 2006).
Then why has a tolerant Muslim religion led to the great economic
divergence with respect to adherents of other religions? As mentioned above,
a switch in the level of religiosity and its view of new inventions could have
taken place after the start of the Crusades. Once again, an environment of war
may have changed the perspective of Muslims and led to a less tolerant version
of Islam (Gutas 1998). This may have also led to intolerance in other religions
toward Islam and hence reduced diversity and inter-religion interaction. An
example of this is the reduced human capital and innovation in areas of Spain
longer exposed to Islam after the periods of Reconquista that was accompanied
by forced conversions and expulsion of Muslims (Cinnirella et al. 2018). As
for countries with a Muslim majority today, Bosker et al. (2013) relate the
economic and intellectual stagnation of the Muslim world to reasons not
related to tolerance, namely, a different choice of main transport mode (camel
versus ship) and failure to develop local participative governments as in Europe
that made cities less dependent on centralized decision-making states. They
show that Islamic lands failed to keep up with the rest of the world as the
methods of trade and urbanization eventually limited the potential for growth.
This is in line with Michalopoulos et al. (2016), who argued that Islamic
principles designed to limit wealth inequality and escape a state of constant
feuding blocked the accumulation of capital indispensable for growth in the
era of large-scale shipping trade and industrialization.
Conclusion
Theory and empirical evidence in the economics literature has reached a
consensus that trade and innovation reinforce each other in nurturing devel-
opment. Although this was also observed throughout the Muslim world in
its flourishing years of the preindustrial era, innovation in Islamic countries
whose economies continue to be based on commerce is lagging behind with
respect to the world technology frontier. Historical narratives that have also
been recently put to data show that trade has been a catalyst for the spread of
Islam. The Islamic Golden Age saw some of the most important inventions
in history. The progress in the Muslim world however came to a halt, even if
trade continued to be the core instrument in the further spread to Sub-Saharan
458 A. Naghavi
Africa, Inner Asia, and East and Southeast Asia. A number of studies relate the
interruption in innovation and growth in Islam to the behavior of the ruling
elite to maintain power and rent, and the influence of religion in the state. Lack
of religious tolerance and diversity of ethnic groups and ideas are often blamed
for the lack of enthusiasm in the society for innovation, though evidence has
shown Islam to be a tolerant religion toward other beliefs. Some recent works
have also linked the lack of technological progress to the gaps and fallacies
with regard to intellectual property in Islamic law, and even their choice of
transport, or the centralization of the state as opposed to powerful smaller
local governments that, for example, emerged in Europe. Despite several recent
empirical analyses that support alternative explanations of the relationship
between Islam, innovation, and trade, economic research on this front is in
its infancy and we call for further in-depth study of the issue.
References
Acemoglu, D., and J. Linn. 2004. Market Size in Innovation: Theory and Evidence
from the Pharmaceutical Industry. Quarterly Journal of Economics 119 (3): 1049–
1090.
Aghion, P., N. Bloom, R. Blundell, R. Griffith, and P. Howitt. 2005. Competition
and Innovation: An Inverted U Relationship. Quarterly Journal of Economics 102
(2): 701–728.
Al-Marghinani, B. 1791. The Hedaya: C. Hamilton Translation. London: T. Bensley.
Bènabou, R., D. Ticchi, and A. Vindigni. 2015. Forbidden Fruits: The Political
Economy of Science, Religion, and Growth. NBER Working Paper No. 21105.
Bose, S. 2006. A Hundred Horizons: The Indian Ocean in the Age of Global Empire.
Bosker, M., E. Buringh, and J.L. van Zanden. 2013. From Baghdad to London:
Unraveling Urban Development in Europe, the Middle East, and North Africa,
800–1800. Review of Economics and Statistics 95 (4): 1418–1437.
Chaney, E. 2016. Religion and the Rise and Fall of Islamic Science. Mimeo, Harvard
University.
Cinnirella, F., and J. Streb. 2017. Religious Tolerance as Engine of Innovation. CEPR
Discussion Paper 12466.
Cinnirella, F., A. Naghavi, and G. Prarolo. 2018. Islam, Human Capital, and Innova-
tion in Historical Spain. University of Bologna, Mimeo.
Clingingsmith, D., A.I. Khwaja, and M. Kremer. 2009. Estimating the Impact of
the Hajj: Religion and Tolerance in Islam’s Global Gathering. Quarterly Journal of
Economics 124 (3): 1133–1170.
Coelli, F., A. Moxnes, and K.H. Ulltveit-Moe. 2016. Better, Faster, Stronger: Global
Innovation and Trade Liberalization. CEPR Discussion Paper 11506.
Cohen, Hayyim J. 1970. The Economic Background and the Secular Occupations of
Muslim Jurisprudents and Traditionists in the Classical Period of Islam (Until the
Middle of the Eleventh Century). Journal of the Economic and Social History of the
Orient 13: 16–61.
Coşgel, M.M., T.J. Miceli, and J. Rubin. 2012. The Political Economy of Mass
Printing: Legitimacy, Revolt, and Technology Change in the Ottoman Empire.
Ensminger, J. 1997. Transaction Costs and Islam: Explaining Conversion in Africa.
Journal of Institutional and Theoretical Economics 153 (1): 3–29.
Grossman, G.M., and E. Helpman. 1989. Product Development and International
Trade. Journal of Political Economy 97 (6): 1261–1283.
———. 1990. Trade, Innovation, and Growth. American Economic Review 80 (2):
86–91.
———. 2018. Growth, Trade, and Inequality. Econometrica 86 (1): 37–83.
Gutas, D. 1998. Greek Thought, Arabic Culture: The Graeco-Arabic Translation Move-
ment in Baghdad and Early Abbasid Society. New York: Routledge.
Huff, T. 2003. The Rise of Early Modern Science: Islam, China and the West. Cambridge:
Iyer, S., R. Tucku, and A. Shrivastava. 2017. Holy Wars? Temple Desecrations in
Medieval India. Cambridge Working Papers in Economics 1705.
Jayyusi, S.K. 1994. The Legacy of Muslim Spain. 2 Vols. Leiden: Brill.
Jha, S. 2013. Trade, Institutions, and Ethnic Tolerance: Evidence from South Asia.
American Political Science Review 107 (4): 806–832.
Kuran, T. 2010. The Long Divergence: How Islamic Law Held Back the Middle East.
Lapidus, I.M. 2002. A History of Islamic Societies. Cambridge: Cambridge University
Press.
Lewis, B. 1993. The Arabs in History. 6th ed. Oxford: Oxford University Press.
Malkawi, B.H. 2013. Intellectual Property Protection from a Sharia Perspective.
University of St. Thomas Law Journal 16: 87–122.
Michalopoulos, S., A. Naghavi, and G. Prarolo. 2016. Islam, Inequality and Pre-
Industrial Comparative Development. Journal of Development Economics 120:
86–98.
———. 2018. Trade and Geography in the Spread of Islam. Economic Journal.
Forthcoming.
Mokyr, J. 2002. The Gifts of Athena. Princeton: Princeton University Press.
Naghavi, A., and G. Pignataro. 2018. The Fallacy of Intellectual Property Rights in
Islamic Law. University of Bologna, Mimeo.
Ofek, H. 2011. Why the Arabic World Turned Away from Science. The New Atlantis
30: 3–23.
460 A. Naghavi
Rubin, J. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East
Did Not. Cambridge: Cambridge University Press.
Skeen, B.A. 2008. Trade and Exchange in the Medieval Islamic World. In Encyclopedia
of Society and Culture in the Medieval World, ed. Pam J. Crabtree. New York: Facts
on File, Inc.
Syed, M.H., S.S. Akhtar, and B.D. Usmani. 2011. A Concise History of Islam. New
Delhi: Vij Books India Private Limited.
The Saylor Foundation. n.d. The Abbasid Dynasty: The Golden Age of Islamic Civiliza-
tion. http://www.saylor.org/hist101.
Watson, A.M. 1983. Agricultural Innovation in the Early Islamic World: The Diffusion
of Crops and Farming Techniques, 700–1100. Cambridge: Cambridge University
Press.
Index
A Ager, P., 272

Abbasid Caliphate, 439, 439n1, 441, Ahmedabad, Hindu–Muslim violence
443 in, 241–244
Abbasid dynasty, 453 Aimone, J.A., 27
Abbasid Revolution, 442 Alcorta, C., 81
Abortion, see Birth control Alesina, A., 296, 299, 300
Abrahamic monotheisms, 426 Algan, Y., 98
Abrahamic religions, reward and Allen, D.W., 355
punishment in, 76 Almoravids, 454
Abrahamic versions, 426 Alta Verapaz, 182
Accountability cum personalized Altruism, 106
spiritual exchange, 428 in algorithmic era, 9–10, 119–135
Acemoglu, D., 114, 322, 438 Ámatitlan, 179n15
AfD, 306 Amazon Mechanical Turk, 121
Africa placeholder task at, 130–131
growth of Christianity in, 10 American colonies, 395
Pentecostalism in, 2 Amish, 32
Agent-based computational models, Anderson, G., 74n4
62–63 Anderson, R.W., 356, 385
Agent-based model, 53 Anglican Church, 392
of religious geographic regionalism, Annuario Eclesiastico Catolico de Centro
66–67 America, 177
Note: Page numbers followed by “n” refers to notes.

Association Series, https://doi.org/10.1007/978-3-319-98848-1
462 Index
Annuario Eclesiástico de la Iglesia Católica Awwaliya schools, 290–291

en Guatemala, 176 Al-Azhar, 290
Annuario Pontificio, 176 Azzi, C., 267
Anti-abortion terrorism, 82, 82n7,
409n5
Apocalyptic beliefs, 37 B
Appleby, R.S., 366 Baathist systems, 366
Aquinas, T., 84 Bai, Y., 170n3
Arabian Peninsula, 451 Baille, G., 388
Arab nationalism, 341 Bargaining, 312
ArcGIS programming, 266, 266n4 Barrett, D., 252
ArcGIS World Geocode service, 176 Barrios, J.R., 176
ARDA, see Association of Religion Data Barro, R.J., 2, 11, 43, 73, 74, 79, 169,
Archives 255, 256, 257n6, 259n8, 351,
Aristotle, 362 352, 407–408, 427
Armenians, 12 Barros, P.P., 7, 43, 46–48, 52, 54
Arminians, 391 Basten, C., 219
Armstrong, K., 427, 429, 433 Bavarian paradox, 306n7
Arnon, A., 90n2 Bayesian Peer Influence (BPI), 99
Arrunada, B., 13, 97, 296–298, 300, Bayle, P., 391, 393
302, 304 Becker, G., 368, 412, 420
Articles of Confederation, 396 Becker, S.O., 2, 11, 169n2, 215, 217–221,
ASREC, see Association for the Study 222n3, 224, 407
of Religion, Economics, and Bedouin tribes in arid lands, 450–451
Culture Behavioral economics, 6, 31
Assaad, R., 291 Behavioral revolution in political
Association for the Study of Religion, science, 364
Economics, and Culture Behaviour, signalling religiosity by,
(ASREC) Annual Conference, 96–98
3 Belgium, religious commitment, 33
Association of Religion Data Archives Belief systems, 430
(ARDA), 255n5 Bènabou, R., 74n3, 77, 83, 408, 417,
Atatürk, M.K., 103 455
Atonement, 8, 81 Bentzen, J.S., 12, 266, 270, 272–276,
Augustus, P., 387 278
Auriol, E., 314n3, 334 Berg, N., 44, 56
Autocratic authority, 338 Berman, E., 27, 37, 64n2, 90, 92, 96,
Autocratic politics and religion 408n4
relationship, 334 Besley, T., 2
Autocratic polity, postwar pattern of, Betz, F., 219
340 Bharatiya Janata Party (BJP), 11–12,
Autocratic regimes, legitimacy of, 334 229–235, 241n11, 246
Autocrats, 338, 339 The Bible, 169–171, 174, 216
Index 463
The Big Sort, 89–90 C

Bilingual education, 178, 185n18 Cai, G., 104n1
Binzel, C., 31, 32 California Yearly Friends’ Society, 181
Birth control Caliphs, 451
in Hinduism, 148 Calvin, J., 94, 170
in Islamic religious discourse, 147 Calvinism, 9, 364
Bischoff, K., 89 Calvinistic beliefs, 84
Bishop, B., 89 predestination and, 77
Bishops War of 1639, 392 CAM, see Central American Mission
Bisin, A., 9, 14, 35, 113, 207, 322, 325, Camino Global, 173n8
326 Campante, F., 408
BJP, see Bharatiya Janata Party Canada, anti-abortion terrorism, 82n7
Black Death, 382n5, 385, 389 Canonical model, 27–29, 32
Blanchard, T.C., 97 of religion, 92
Blaydes, L., 330, 342, 342n2 of religious clubs, 2
Bloch, M., 322, 386 Canonical politico-religious
Bohr, N., 159 equilibrium, 336
Boisjoly, J., 98 Cantoni, D., 219, 223, 306, 438
Boppart, T, 217n2 Capitalism, liberal democracy and, 366
Bordo, M.D., 302n4 Carvalho, J.-P., 1, 6, 7, 25, 31, 32,
Bosker, M., 457 106–108, 115n6, 160, 357,
Bossuet, J.-B., 393n14 408n4
Botticini, M., 284, 285, 288 Cathar heresy, suppression of, 388
Boyer, P., 74n2 Cathars, 387, 387n8
BPI, see Bayesian Peer Influence Catholic Church, 176, 177, 223, 305,
Brentjes, S., 440 368, 370, 371, 387n8
Bressler, E.R., 26 authority of, 389
Brethren medical mission, 181 critique of, 390
Breton, R., 106 forgiveness, 81, 84
Brimicombe, A.J., 90 Index Librorum Prohibitorum, 73
Britain, see United Kingdom indulgences, 84
British Muslims, 114 in medieval Europe, 380
Brown, P., 386 miracles and, 83
Brunnermeier, M.K., 296, 299, voucher programs, 158
300 Catholics, 97, 266, 296, 300, 301, 304,
Buddhism, 16, 283 364, 365, 394
reward and punishment in, 76 birth control, 147
Buddhist nation, 353 cultural differences between
Buddhists, 273 Protestants and, 297, 303–304
Bulbulia, J., 272 culture, 363
Bulliet, R.W., 329 divergence in, 299
Burgess, P., 176n11, 177 educational investment, in
Gautemala, 169–192
464 Index
enrollment to Catholic adherents, Christian Europeans, 381

ratio of, 206 Christianity, 16, 272, 304, 313–314,
and evangelical Protestants, bilateral 342, 380, 381, 426, 427, 429
tension between, 205 in Africa, growth of, 10
forgive sins, 304 birth control, 147
investment in education and, 11 commitment to, 385
and Orthodox differences, 298 in India, 284
protective of personal relations, 298 Islam and, 365
and Protestants, 299, 394 and science, 408
conflict between, 91 in Southern France, 388
cultural differences between, 13 Christian missionary activity on
values, 296 economic and political
religious beliefs, 74 outcomes, 4
religious denomination, 302 Christian-Muslim differences
values, 307 in educational and occupational, 291
Central American Mission (CAM), impacts of reform on, 291
173n8, 181 Christians, 273
Centralization, 307 denominations and economy, Euro
of state institutions, 356 zone, 300–305
Centralized religion, decentralized doctrine, 314
religion vs., 341–343 fundamentalism, 2, 366
Ceyhun, E., 417 and Muslims, socioeconomic
Ceylan, R., 103 differences between, 268
Chadi, A., 296, 298, 300, 301n2, 303, “up-skilling” effect on, 290
304, 306 in urban Egypt, 289
Chamie, J., 145 Church attendance, 3
Chaney, E., 16, 330, 342, 342n2, 441, impact of education on, 223–225
444, 454 Church bulletins, 161–166
Characteristics hypothesis of fertility, Church of England, 392
145, 149–150 Church of Latter-Day Saints, 68
Charismatics, 171n6 Church of the Nazarene, 181
Charlemagne, 314 Church Ordinance for Wittenberg, 218
Charles V, 423 Church-Sect Cycle, 8, 66–67
Chaves, M., 268n8 Church–state relationship, 10, 155–167
Chen, D.L., 9, 80, 119, 160, 417 Chwe, M.S., 75n5
Chen, T., 7, 28, 29 Cinnirella, F., 454
Chetty, R., 89 Clannism, 339
China, Protestant foreign missions in, Clash of civilizations, 2, 230, 250
170 “Clash of Civilizations” (Huntington),
Chiswick, B., 210 250, 365
Christian and Missionary Alliance, Cleveland School District Scholarship
173n8 Program, 203n6
Christian Democratic parties, 370 Clingingsmith, D., 456
Index 465
Club good model of religion, 92–93 Conditional toleration equilibrium,

Club theory of religion, 64 378–379, 383, 398
“corner solution” issue, 64, 64n2 Jews treatment under, 384–386
Club theory of religious production, 50 Confessional cultures, 304, 305
Coase, R., 354 Constitutional democracy, 363
Coşgel, M.M., 14, 348, 350, 352, 355, Consumer preferences, 55
357, 380 Cooperative behavior, rituals/beliefs
Coelli, F., 449 and, 78–80
Coercion, 311, 312 Coptic Christian, 285, 288–289
Coercive agents, 312 Coptic-Muslim occupational
identity of, 312 differences, 289
legitimating and, 312 Coptic-Muslim SES gap, 285–289,
Coffin, H.S., 171 292
Cohen, P., 170n3 Coptic-Muslim SES inequality, 285
Cohen-Zada, D., 11, 160, 197, 199, Coptic–Muslim socioeconomic
202–206 inequality, 12, 291–292
Coleman, J., 106 Coptic schools, 289
Collier, P., 257, 257n7 Coptic tax revolts, suppression of, 286
Colvin, C.L., 302 Copts, 284–285, 285n1
“Commercial Revolution,” 314 children for bureaucratic jobs, 287
Commitment-signaling theory of and Muslims, 287–288
veiling, 34 “Core Europe,” 305n5
Commodification, 120 Corruption, 339
Community heterogeneity, 110–111 Coulson, N., 147
Community leaders, 104 Council of Trent, 170, 170n4
Community size, changes in, 111 Courbage, Y., 286
Compulsory schooling laws, 218 Craft guilds, 382, 382n5
Computational cellular automata Crehan, F.J., 170n4
model, 64 Cromer, L., 287
Computational modeling of religious Cross-national time-series dataset, 353
groups, 7–8, 61–70 Cults, 69–70
agent-based computational models, Cultural assimilation, 33
62–63 Cultural diffusion, 325
Church-Sect Cycle, 66–67 dynamics of, 326
club theory of religion, 64 Cultural distinction, 108–109
cults, 69–70 Cultural divergence, 299
free riders, value of, 67–68 Cultural diversity
religion, as network good, 69 leaders and, 107–112
religious extremism, 65–66 community heterogeneity,
CONALFA (adult literacy program), 110–111
178 cultural extremism, diversity, and
Conciliar movement, 389 integration, 108–109
Conditional toleration, 381–384 economic gains, 109–110
466 Index
government transfers, 110 Directorate for Bilingual Education

measure of, 299 (Dirección General de la
Cultural dynamics, 325 Educación Bilingüe or
Cultural extremism, 108–109 DIGEBI), 178, 185
Cultural integration, 108–109 District-level factors, 274
Cultural leaders, 103–115 Dittmar, J., 216
Cultural multiplier, 325 Divorce
Cultural resistance, 34 in Hinduism, 147–148
Cultural transmission model, 7, 35–38, in Islamic religious discourse,
275 146–147
of religious beliefs and preferences, 6 Doctrine adaptation, 83
role of leaders in, 9 Dominant religion, 349
Cushing, R.G., 89 Dominicans, 176
Cynicism, 339 Dowd, R., 252n2, 260
DSGE, see Dynamic stochastic
general equilibrium model of
D heterogeneous agents
Darnell, A., 407 Durkheim, É., 222
Dasgupta, P., 2 Durkheimian argument, 430
Decentralized religion vs. centralized Dustmann, C., 98
religion, 341–343 Dutch Protestantism, 391
Deep agent heterogeneity, 7, 63, 67, 68, Dutch Republic, 391–392, 394
70 Dutch Revolt, 315, 391
Deep-rooted Islamic tradition, 338 Dynamic stochastic general
Degree of predictability, 276 equilibrium (DSGE) model of
(De)learning effect, 81 heterogeneous agents, 65
Demand-side factors, 267
Demand-side theories, 267–268,
278 E
Democracy, 367 Early Childhood Longitudinal Study,
Middle East, 317–319 Kindergarten Class of
Democrats, 364, 365 1998–1999 (ECLS-K), 206
Demography–religion relationship, 10, Earthquakes, 272–273
143–145 direct impact of, 277
See also Religious demography elevated risk of, 277
Diamond, J., 426 literature on dynamic effects,
Dietary restrictions, 90 274–275
DIGEBI, see Directorate for Bilingual and religiosity, 274, 276, 277
Education (Dirección General risk of, 273–274, 273n17, 275
de la Educación Bilingüe or standardized parameter estimate on,
DIGEBI) 274n19
Digitization, 215 Eastern Europe, ultra-Orthodox
“Dignity of Islam,” 340 Judaism, 28
Index 467
Eastern Orthodox Christianity, 250 “Emperor of the Romans,” 314

Eastern religions, reward and Empirical advances, in economics of
punishment in, 76 religion, 10–13
East European Jews, ultra-Orthodox Empirical analysis
sect of, 92 cross-national data, 352
Ecclesiastical authority, 431 secularization of states, 353
Ecclesiastical legitimacy, 431 Empirical evidence, 275
Ecclesiastical Ordinances (1541), 97 Endogeneity, 272–275
Eckstein, Z., 284, 285, 288 Endogenous fertility, 7, 29
ECLS-K, Early Childhood Longitudinal Engineer, A., 231
Study, Kindergarten Class of England, 392–393
1998–1999 Ensminger, J., 450
Economic gains, 109–110 Erdoğan, R.T., 316–318
Economic growth, effects of religion on, Esposito, J., 366–377
4 Esteban, J.-M., 15, 82, 236, 406
Economic-parliamentary elite, 315 Esteem maintenance, 6, 31–32
Economics, and religious conflict, Eswaran, M., 44, 55
255–256 ET, see L’Etat des Religions dans le
Economics literature, 406 monde
Economies of scale, social advantages of, Ethnic Groups in Conflict (Horowitz),
427–428 249
Edict of Nantes, 394 EU reform, 305n5
Edict of Toleration, 393 Euro crisis, 296, 300, 303, 304
Education, fundamentalism on, 407 Euro member countries, cultural norms
Educational inequality, Coptic-Muslim, in, 297
290 Europe
Educational investment, in Guatemala, cultural heterogeneity
169–192 transformation, 307
Egypt issue of integrated vs. multi-speed,
Coptic Christian, 285 305n5
state industrialization, 289 personal liberties in, 406
Ehrenberg, R., 267 populist movements in, 305
80-20 Pareto Rule, 67 secular phenomena, for learning
Eisenstein, E.L., 170n10 religious practice, 160
Ekelund, R., 427, 430, 431 European Economic Community,
Elder, T., 202, 204–207 296
Elementary school enrollment, private European integration, 296–300
share in, 197, 198 European legitimating, 315
Elizabeth I, 392 European religion market, 431
ELMa, 408–409, 412–414 European Union (EU), 296–297, 306
ELMb, 417–420 European integration, 297–300
El Quiché, 182 Euro zone, Christian denominations
Emancipation of Jews, 108 and economy, 300–305
468 Index
and national identity, 305–306 Formal policing institutions, invention

religion and, 13, 305–306 of, 267
Euro zone, 296, 298, 300n2, 301 Fox, J., 257–260
Christian denominations and France, 394–395
economy, 300–305 Germany vs., 299–300
Evangelical movements, 11 religious commitment, 33
Evangelical Primitive Methodists, 181 signalling religiosity by behaviour, 97
Evangelical Protestants and Catholics, Franciscans, 176
bilateral tension between, 205 Franck, R., 160
Evangelicals, 171–174 Fratzscher, M., 300
Externality problem, 36 Free-rider problems, 6, 25–27, 36, 37,
External threats beget social unity, 51
429–430 value of, 67–68
Extrinsic religiosity, 276 Free ridership, effective limits on, 430
Freud, S., 224
Friedman, D., 65–67
F Frymark, J., 159
Fargues, P., 286 The Future of an Illusion (Sigmund
Fearon, J., 257 Freud), 224
Fellowship ratio, 56
Ferrero, M., 43, 52, 355
Fertility G
characteristics hypothesis, 145, Gallup Millennium Survey, 252
149–150 Garoupa, N., 7, 43, 46–48, 52, 54
and education, relation between, Gaskins, B., 43, 52
221–222 General Social Survey (GSS), 158
in Islamic religious discourse, on frequency of church attendance,
145–146 204
‘minority group status’ hypothesis, Gentzkow, M., 162
145, 149 Germanic legal system, 387
‘particularised theology’ hypothesis, German Protestants, 298
145, 149 Germany
and religion, relationship between, AfD, contemporary right-wing party,
142, 145 306
Fetzer, J.S., 367 France vs., 299–300
Field, E., 91, 95, 241 Imams in, 103
Figlio, D., 157n2 liberal reform movement, 28
Finke, R., 3, 37, 46, 249, 255n5, 259, Protestant Reformation on
260, 356, 371 education, 11
Finley, T., 385 Gilbert, W., 97n6
First-order issue, 268 Gill, A., 14–15, 370, 371, 378, 396
Floods, 275 Glaeser, E., 440n4
Forgiveness, 81, 84 Global religiosity, data on, 268–271
Index 469
Glorious Revolution, 393 Health behaviours, 150–151

Golden Age of Islam, 16 Hedaya, 452
Gonzalez, A.L., 367 Helpman, E., 449
Gordon Riots, 393 Henrich, J., 379
Government transfers, 110 Henry IV (Holy Roman Emperor), 314,
Granados, M.G., 176 394
Greece, Meteora Monastery in, 96 Heretics suppression in medieval
Greeks, 12 Europe, 386–389
Green, L, 218 Herfindahl index, 54
Greif, A., 114, 330, 380 Hickman, P., 188n23
Grim, B.J., 255n5, 260, 356 “High Islam,” 337
Grossman, G.M., 449 Hinchcliffe, D., 147
Grossman, S.J., 354 Hinduism, 16, 147–148, 273, 283, 428
Group identity formation, 36 birth control, 148
Gruber, J., 160, 267 divorce, 147–148
GSS, see General Social Survey fertility, 148
Guatemala fundamentalism, 366
churches in, 188 marriage, 147
data for departments of, 176–178 polygyny, 147
educational investment in, 169–192 reward and punishment in, 76
investment in education and in South Asia, growth of, 10
Protestantism, 11 Hindu–Muslim conflict in India,
literacy rates economic approach to, 11–12
department fixed effects for, Hindu–Muslim violence, in India,
188–190 229–246
regression results for, 184 empirical analysis
long-term evidence on religion and data, 233–235
human capital in, 179–191 empirical specification, 235–237
Ministry of Education, 177, 178, 185 results
religious schools in, 187–188 Ahmedabad case, 241–244
school-enrollment ratios, 182, politics, role of, 244–246
190–191 urban households, 238, 240–241
Guiso, L., 73, 296–298, 304, 307, 407, Hoeffler, A., 257, 257n7
417 Hofrenning, S., 210
Gutas, D., 441 Hoge, D., 205–207
Gwin, C.R., 43, 52, 352 Holocaust, 272
Hornung, E, 221
Horowitz, D., 249, 256
H Hotelling-style model, 352
Hansen, I.G., 272 House of Saud, 318
Hart, O.D., 354 Hoxby, C., 157n2, 199n3
Hauk, E., 104, 105n2, 109, 110, 112, 113 Huber, J.D., 73, 417
Hawthorne effects, 121, 122 Huguenots, immigration of, 221
470 Index
Human capital formation, 16 Indulgences, 84

Human rights revolution, 419 Industrial Revolution, 334, 384
Hungerman, D.M., 10, 44n2, 155, role of education in, 220
157n3, 159–162, 166, 267 Inglehart, R., 265n3, 268, 276n24
Hunter-gatherer societies, 425 Innovation
Huntington, S.P., 250, 259, 365 in Islamic world, big bang and big
Hussein, T., 290 crunch of, 453–454
Hutchison, W.R., 171n5 religion and, 2, 6, 16
as sacrifice for power, 454–456
Institutional dynamics, 325, 328
I Institutional perspective, scientific
Iannaccone, L.R., 2, 3, 5, 6, 26, 27, 35, development in, 441–443
56, 64, 64n2, 66–67, 75, 78, Institutional quality
92, 93, 96, 98, 114, 160, 205, macroeconomic performance and
207, 350, 371, 408n4 measures of, 300
Ideational perspective, politics, outcomes and measures of, 296
363–368, 372 religious denomination and, 303
Identity rules, 381–384, 398 Institutional reforms, 333
religious legitimation and, 384 Institutions, effect on social preferences,
ubiquity of, 397 9–10
IEA, see International Economic Integrated rational choice theory, 370
Association Intellectual property, 451–452
IFIs, see Islamic financial institutions Intellectual pursuits, funding for, 440
Illiteracy, Coptic Christianity and, 289 Intelligent design of religious beliefs, 8,
Imams, 103, 107n3 73–87
INC, see Indian National Congress atonement, 81
Income inequality, segregation and, 89 doctrine adaptation, 83
Index Librorum Prohibitorum (Catholic forgiveness, 81
Church), 73 miracles, 83
India motivated beliefs, 83
British colonies in, 170 religious intolerance due to pressures
Christianity and Islam in, 284 to survive, 82
economic approach to Hindu– repentance, 81
Muslim conflict, 11–12 rituals, 78–80
Hindu–Muslim violence in, 229–246 systems, 75–78
National Commission for Minorities, predestination, 77–78
230 reward and punishment,
National Sample Surveys, 234, 76–77
234n7, 236, 238 Intermediated social preferences,
Indian National Congress (INC), 231, 119–135
232 demographic survey, 134–135
Individual liberties, 419 experimental results, 125–129
Indonesia, religious participation in, 80 methodology, 122–125
Index 471
placeholder task at Amazon Islamic law, 442, 451–452

Mechanical Turk, 130–131 Islamic principles, trade and, 450–451
sharing bonus, 133–134 Islamic religious legitimacy, 317
task, checking, 132–133 Islamic Revolution in Iran, 365
Intermediation, 120 Islamic societies, evolution of, 329–330
International Economic Association Islamic Spain, 383
(IEA) Islamic State (ISIS), 229
Roundtable on the Economics of Islamic theology, 368
Religion, 1–2 Islamic values, 338
International factors, 339–341 Islamic world, 329, 439
International Social Survey Program Golden Age of scientific production,
(ISSP), 252 445
Inter-religion violence, 4 scientific development in, 441
Intrinsic religiosity, 276 Islam Instrumentalized: Religion and
Investment in education, 11 Politics in Historical Perspective
Iran, S., 335 (Jean-Philippe Platteau), 14
Iranian Islamist Revolution, 339 Islamism, 341
Islam, 3, 10, 16, 145–147, 320, 333, 366, Islamist ideologies, supply of/demand
380, 381, 426, 427, 429 for, 339
birth control, 147 Islamist movements, 340
and Christianity, 365 Islamist radicalism, 340
decentralized character of, 343 Islamization, 3
decentralized organization of, 341 Israel, ultra-Orthodox neighbourhoods
divorce, 146–147 in, 96
fertility, 145–146, 148 Issawi, C., 288
Golden Age of, 16 ISSP, see International Social Survey
in India, 284 Program
marriage, 146 Istrefi, K., 302n4
in Middle East, 313 Italy, religious commitment, 33
political economy approach to, Iyer, S., 1, 3, 10, 44, 54, 55, 141, 142,
334–341 148, 149, 232, 249, 268
and politics relationship, 334 Iyigun, M., 15–16, 425, 426, 429
polygyny, 146
in public life, influence of, 340
rationalist interpretations of, 440 J
study of, 367 James II, 393
Islamic Banking, 357 Jay, J., 364
Islamic commercial institutions, 329 Jedwab, R., 385
Islamic Egypt, 285 Jefferson, T., 396
Islamic farmers, 453 Jehovah’s Witnesses, 68
Islamic financial institutions (IFIs), 56 Jesuits, 176
Islamic fundamentalism, 2 Jewish Diamond Traders, New York,
Islamic Golden Age, 456, 457 106
472 Index
Jewish Naturalization Act of 1753, Koyama, M., 15, 106–108, 356, 382n5,
393 385, 387, 388, 396, 408n4
Jews, 381–383 Krapf, M., 13, 296, 298, 300, 301n2,
civic rights in Frankfurt, 395n15 303, 304, 306
settlement in eastern Europe, 386 Krueger, A.B., 66n3
treatment under conditional Krugman, P., 300
toleration equilibrium, Kung, J.K., 170n3
384–386 Kuran, T., 2, 69, 288, 366n1, 372,
Jha, S., 230, 456 383n6, 454
Jizya, 382 Kuru, A., 367
Johnson, N.D., 15, 356, 382n5, 387,
388, 396
Johnson, S.C., 276n23 L
Joint evolution La Ferrara, E., 98
of culture and institutions, 322–325 La Porta, R., 74n1
of religious values, 328 Labor market intermediary (LMI), 121,
Journal of Economic Literature, 2, 3 122, 126
Judaism, 284, 407, 426, 427, 429 Labor supply, 411–413
Judgment Day, 428 empirical evidence, 413–416
Juergensmeyer, M., 366 Ladinos, 178
Juncker, J.-C., 305n5 Ladurie, E.L.R., 388
Justification Laitin, D., 257
by faith, 78 Latin America, 16
by the grace of God, 78 Pentecostalism in, 2
Justman, M., 11, 197, 199, 202–204 Protestant with Catholic investment
in human capital, 170
Latino Barometer, 172n7
K Laudato si, 2
Kalyvas, S., 370 Law enforcement, 425
Kant, I., 121–122 Law of National Languages (2003),
Kaqchikel, 174 185
Karaite Jews, 12 Leader(s)
Karma, 76 community, 104
K’iche, 174 competition, 112–113
King, M.L., Jr., 53 cultural, 103–115
Kleptocratic autocracy, 343 and cultural diversity, 107–112
Koenig, H.G., 276n23 emergence of, 113–114
Kohl, H., 305n5 objectives of, 104–107
Kollman, K., 372 religious, 103–115
Koran, 145 Lee, S., 170n3
Korea, Protestant foreign missions in, Legitimacy principles, 326
170 Legitimating
agents, 312, 312n2
Index 473
and coercive agents, 312 Marx, K., 11, 406n1

identity of, 312 Marx effect, 358
Le Suicide (Émile Durkheim), 222 Marxist-based liberation theology,
L’Etat des Religions dans le monde 370–371
(ET), 253–255 Mass education, rise of, 358
Levantine Christians, 13 Mattlin, J.A., 276n24
Levy, G., 8, 73, 74, 76, 77, 83, 86n10, MAUP, see Modifiable areal unit
93–95, 97, 98, 408 problem
Lewis, B., 250, 366 Maxwell, D., 2
Ley Pavón, 176 Mayall, J., 249
Liberal autocracy, 343, 344 Mayoral, L., 15
Liberal democracy and capitalism, 366 McBride, M., 2, 7, 27, 28, 43, 48, 49,
Liberalism, 397 51, 52, 67, 68, 382n5
Liberties, 417–419 McCleary, R.M., 2, 11, 43, 73, 74, 79,
Lijphart, A., 256 97n7, 169, 170n3, 256, 259n8,
Lind, J.T., 160, 417 351, 352, 407–408, 427
Lisbon earthquake, 272 McCracken, M., 302
Literature, in religion economics, 378 McLuhan, M, 173n10
LMI, see Labor market intermediary Means-tested vouchers, enrollment of,
Locke, J., 393 203–204
Long divergence, 322, 329–330 Medieval Church doctrine, 3
Louis XIV, 390, 393, 394 Medieval Europe
“Low Islam,” 337 Catholic Church in, 380
Luther, M., 170, 215–225 heretics suppression in, 386–389
Lutheranism, 306n6 identity rules in, 382
Lutz, W., 143, 149 Jewish communities, 382
Jews in, 379, 384
religious and political equilibrium of,
M 389
Madrasas, 438, 440, 440n3, 444 Medieval heretics, 389
Mainline Protestantism, 174, 175 Medieval Islamic world
denominations, 11, 170–171, 173 science in, 439–441
missionaries, 171 scientific production in, 437
Makowsky, M.D., 7, 61, 64–67, 69, 98 Meisenzahl, R., 216
Malthus, T., 221 Melanchthon, P., 216
Mankiw, G., 255 “Melting pot,” 289
Marriage Mental health of adolescents, effects of
in Hinduism, 147 religion on, 4
in Islamic religious discourse, 146 Meteora Monastery, Greece, 96
MARS, see Multi-Agent Religion Meyersson, E., 160
Simulation Miceli, T.J., 14, 348, 352, 355, 380
Marsden, G.M., 173n8 Michalopoulos, S., 450, 457
Marty, M., 366 Microeconomic theory, 267
474 Index
Middle East, 3 Mokyr, J., 220, 455

democracy, 317–319 Monastery of Ostrog, Montenegro, 96
economic and political future of Monnet, J., 298
economies, 319 Monopolistic religious organization, 351
history of, 313 Monopoly state religion, 378
Islam in, 313 Monotheism, in supply of religion,
non-Muslim minorities, 15–16, 423–424
socioeconomic advantage accountability cum personalized
of, 288 spiritual exchange, 428
political turmoil in, 377 effective limits on free ridership, 430
religious demography in, 142 evolutionary roots of, 432–434
socio-economic growth of, 322 external threats beget social unity,
and Western Europe 429–430
long-run economic divergence longer time horizons due to belief in
between, 14 afterlife, 428–429
political economy differences middlemen, role of, 430–432
between, 14 social advantages of economies of
Middlemen, role of, 430–432 scale, 427–428
Mihna, 443 social order, and political stability,
Miles Kimball, 423 424–427
Militarization, of government structures Monotheist ecclesiastical institutions,
and scientific decline, 443–445 426
Miller, H.J., 176n11 Monotheistic and polytheistic,
Miller, L., 276n23 fundamental differences in,
Miller, P., 364 433
Milwaukee Parental Choice Program, Monotheistic ecclesiasticisms, 424
203n6 Monotheistic faiths, 427
‘Minority group status’ hypothesis of Monotheistic proselytizing religions,
fertility, 145, 149 380
Miracles, 83 Monotheistic state religion, 428
Miskawayh, A., 443 Montalvo, J.G., 12, 236, 249, 251,
Mississippi River floods, 275 253–256, 258, 261
Mitra, A., 11, 229, 231, 234, 238, 241, Monte Carlo simulation, 62, 65
246 Montenegro, Monastery of Ostrog in,
Modern education, 290 96
Modern state Montgomery, J.D., 44, 54, 66, 67
birth of, 396 Moore, R.I., 387
development of, 397 Morality, 120
establishment of, 391, 393, 396 Mormons, 9, 28, 68
threat to liberalism, 397 Mortality, 150–151
Modi, N., 230 Motivated beliefs, 83
Modifiable areal unit problem (MAUP), Mueller, H., 104, 105n2, 109, 110, 112,
12, 261 113
Index 475
Muhammad Ali Pasha, 289, 290 National identity, religion and, 305–306
Multi-Agent Religion Simulation Nationalism, 305
(MARS), 98 National Sample Surveys (NSSs), 234,
Muñoz y Capurón (Guatemalan 234n7, 236, 238
Archbishop), 177 Natural disasters, 272
Muslim, 28, 270, 273, 285, 287–289 Nellis, G., 232
accommodation of, 367 NELS:88, see National Educational
anti-printing attitude, 455 Longitudinal Study of 1988
authorities, 455 Nelson, R., 220
autocracies, 339 Neo-Pentecostals, 175
and Christians, socioeconomic The Netherlands, 307
differences between, 268 religious commitment, 33
Copts and, 287–288 Network good, religion as, 69
empire, 456 Niebuhr, R., 424
medium-skilled, 289 9+5 theses, effects on Protestant
nations, 353, 367 Reformation, 215–225
population, 450 education advantage of protestants,
schools, 287 218–219
society, 454 fertility–education relationship,
“up-skilling” of, 290 221–222
women, 357, 358 girls vs. boys, 217–218
world, 322, 342, 344 graduates of Protestant universities,
Muslim autocrats, 337 223
trade-off, 339–340 Huguenots, immigration of, 221
Muslim countries, 337 Industrial Revolution, role of
dominant political system in, 335 education in, 220
output growth in, 408 reading and writing, 216–217
predicament of, 334–341 secularization, 223–225
suicide rate and Protestantism,
222–223
N Non-Coptic Christians, 288
NAF, see National Abortion Federation Non-Coptic Jews, 288
Naghavi, A., 16, 355, 449, 451, 452 Non-market behavior, economic studies
Nash equilibrium, 64, 93, 451 of, 70
Nasser, A., 366 Non-Muslim minorities, 292
National Abortion Federation (NAF), socioeconomic advantage of, 288
82n7 Non-Muslims, 382
National Archives of Egypt (NAE), 284 Non-sectarian private schools vs.
National Commission for Minorities subsidized religious private
(India), 230 schools, 203
National Democratic Alliance, 229 means-tested vouchers, enrollment
National Educational Longitudinal of, 203–204
Study of 1988 (NELS:88), 206 Norenzayan, A., 272, 379
476 Index
Norris, P., 265n3, 276n24 Pargament, K.I., 272, 276n23

North, C.M., 43, 52, 352 Park, C., 276n24
North, D.C., 397 Parsons, T., 424
North Africa, religious demography in, ‘Particularised theology’ hypothesis of
142 fertility, 145, 149
North America, 395–396 PD, see Prisoner’s dilemma
Northern Europeans, 299 Peace, 15
NSSs, see National Sample Surveys Peace Accords of 1996, 178, 185
Nteta, T.M., 115 Peace of Augsburg, 407n2
Nunn, N., 170n3 Penick, J.L., 272
Pentecostalism, 6, 11, 173–175, 182, 183
in Africa, 2
O in Latin America, 2
Obermeyer, C.M., 147 Persecution and Toleration: The Long
Occupational inequality, 289 Road to Religious Freedom (Noel
Coptic-Muslim, 290 Johnson and Mark Koyama),
OECD, 157n1 15
Official clerics, 340 Personal liberties, 15, 406, 411–413
One God/One Faith duality inherent in constraints on, 409–411
Abrahamic monotheisms, 427 empirical evidence, 413–416
“Oneness” of God, 427 regulation of, 409
One True Faith, 429 Personal liberties index
Ong, W., 173n10 cross-country std. deviation of, 411
Opposition Confrontation Strategy, 335 legal evolution of, 410
Opposition Suppression Strategy, 335 Pew Research Center, 142
Organizational capacity of religious Phelps, E.S., 220
groups, 31–32 Philippines, minority group hypothesis
Organized religion, spread and of fertility, 149
sustenance of, 426 Phillips, M.A., 358
Orthodox Christianity, 300 Philpott, D., 367
Orthodox Christians, 301 Physical insurance, 276
religious denomination, 302 Pignataro, G., 355, 449
Ottoman Empire, 288, 357, 382, 455 Pillarisation, 391–392
Ottoman rulers, 456 Piscatori, J., 366
Ottoman Sultanate, 423 Platteau, J.-P., 14, 314n3, 322, 334, 378,
Overarching problem, 317 381
Own-type bias, 29 Polarization, as source of conflict,
256–259
Political community, advances in
P religion and, 13–16
Paganism, 433 Political cynicism, 341
Pan-Arabism, 366 Political economy approach, 355, 372,
Paolacci, G., 123n3 378
Index 477
analytical framework, 334–336 Prestige, 106

characterizing policy choices and Preston, I., 98
political stability effects, Princeton Fertility Project, 142
336–338 Principal-agent problem, 351
international factors role, 339–341 Prisoner’s dilemma (PD), 75, 78, 85, 86
religious outbidding risk, Private education, religious factor in, 11,
338–339 197–211
Political economy model of education communal decisions and influences,
choice, 202 201–202
Political economy of religion, 14–16 data sources
Political instability, 333 aggregated data, 211
Political legitimacy, 379, 382 individual level data, 209–210
Political participation, effects of religion empirical findings, 202–207
on, 4 individual choice, 199–201
Political power, distribution of, 324 Private school enrollment by school
Politics type, distribution of, 198
from economic perspective, 368–373 Private School Survey, 157, 211
from ideational perspective, 363–368 Prophet Muhammad, 329, 334,
role in Hindu–Muslim violence, 439n1
244–246 Protestant-Catholic socioeconomic gap,
scholarly blind spots and great 283
awakening, 361–362 The Protestant Ethic and the Spirit of
Poll tax, 286 Capitalism (Max Weber), 219,
on non-Muslims, 292 250
Polygyny Protestantism, 217, 217n2, 300, 304,
in Hinduism, 147 407n2
in Islamic religious discourse, 146 effects on economic outcomes, 219
Polytheistic faiths, 428 in Gautemala, 169–170, 173–175
Pope Francis, 165 investment in education and, 11
Pope Gregory VII, 314 suicide rate and, 222–223
Pope Leo III, 314 Protestant missionaries, 371
Populist movements in Europe, 305 Protestant Reformation, 169, 341, 367,
Portes, A., 104 389
Post-colonial Muslim states, 367 on education, 11
Post-enlightenment era, progressive Protestants, 97
thinkers of, 424 and Catholics, 394
Poutvaara, P., 44, 56 conflict between, 91
Powerful gravity of Islam, trade as, cultural differences between, 13,
452–453 297, 303–304
Prayer, 271 countries in Euro zone, 296
Predestination, 77–78 divergence in, 299
Preindustrial Middle East, 322 educational investment, in
Premillennialist movements, 11 Guatemala, 169–192
478 Index
European countries with, 296 Ray, D., 11, 229, 231, 234, 236, 238,
German Protestants, 298 241, 246
multi-speed union, 304 Raynold, P., 44, 55, 56
nationalism and sovereignty Razin, R., 8–9, 74, 76, 77, 83, 86n10,
concerns, 305 89, 90, 93–95, 97, 98, 408
religious beliefs, 74 Reardon, S.F., 89
society, 304 Rebellious clerics, 338
values, 296, 307 Reciprocity, 120
work ethic belief, 77n6 Reconquista period, 454, 457
Protestant Scandinavian EU member Reda, A., 44, 54
countries, 296 Redistribution, 120
Prummer, A., 9, 103, 105, 107–111 Reformation, 304, 306, 315, 316,
Prussia, role of education in Industrial 389–391, 398, 407
Revolution, 220 causes and consequences of,
Prussian Statistical Office, 215, 217, 306n6
222 economic and social effects of, 4
Psychological immune system, 31 Reformed Church, 391
Psychological tax, 408n4 Relative religiosity, 270
Public debt, 301 Religion, 54–55
Public schools, rise of, 358 affiliation, 142
Punishment authorities, 313, 315, 334, 380
peer, 114 behavior, 370
religious beliefs in, 74–77, 80–82, change, 4
85, 93, 95 commitment, 6, 32–35
Putnam, R., 74n1 communities, 26
dissent, 386
economies, 369
Q extremism, 65–66
Quasi-religious utopian movement, 25 fractionalization index, 251–252
good, 327, 330
institutions, 406
R intensity, 415n11
Rabbinic Jews, 13 intolerance, due to pressures to
Rabbinic Judaism, 284, 288–289 survive, 82
Radical Islamist movements, 341 investment, 54
Radicalized religious movements, 341 leaders, 103–115
Ramadan, 408 literature, economics of, 268,
Random Utility Model of school choice, 284
204 market, 348–352
Rappaport, R., 80 market models, 5–6
Rasmussen, L.L., 305n5 narratives, 379
Rationalized forms of administration, outbidding, risk of, 338–339
emergence of, 442 participation, 6, 7, 28, 80, 161
Index 479
peace, 383 political power of, 326–327

pluralism, 7, 207 Religious club models, 5, 27–29, 31, 35
prohibitions, 82 Religious clubs
radicalization, 38 role of, 6
restrictions, as source of conflict, strategic role of religious identity.,
259–260 25–38
sacrifice, 5 Religious competition model, 371
sociologists of, 350 Religious concentration
stigma, 5 on secularism, 354
suppression, 355–356 and state secularism, 353
tension, 37, 54 Religious conflict, 392
theology, 144, 149 in India, 11
toleration, 391 quantitative approach to, 249–261
Religiosity, 12, 28, 31, 32, 34, 50, 55, economics, 255–256
56, 65, 66, 107–110, 112, 115, polarization, as source of conflict,
204, 273, 411–413, 417–419 256–259
across globe, 269 restrictions, as source of conflict,
distribution of, 417 259–260
and economic outcomes, 407–409 segregation and, 9
empirical evidence, 413–416 Religious coping, 271–272, 276
global measure of, 270 literature on, 276
impact of disaster risk on, 273 Religious countries, 265
patterns in, 270 Religious demography, 141–151
persistence in, 275 characteristics hypothesis, 149–150
signalling, 96–98 Hinduism, 147–148
within-country variation in, 271 historical and contemporary trends
Religious beliefs, 5, 381 in, 142–143
atonement, 81 Islam, 145–147
cultural transmission of, 6 mortality and health, 150–151
forgiveness, 81 Religious denomination, 270, 270n9,
intelligent design of, 8, 73–87 302
nature and role of, 93–95 vs. control over corruption, 303
repentance, 81 and institutional quality, 303
rituals and, 78–80 Religious diversity, 391, 396
segregation and, 9 measurement of
systems of, 75–78 data, 252–255
predestination, 77–78 fractionalization and polarization
reward and punishment, 76–77 indices, 251–252
Religious capital, 28, 35, 66, 68 Religious education, 2, 34
formation, 7, 8 in private schools, 199
Religious clerics, 334–336, 340 Religious freedom, 2, 15, 381, 395
legitimacy effect of, 328 guarantee of, 396
political empowerment of, 328 intellectual arguments for, 393
480 Index
rise of, 377–378, 397 and belief system, relation between,

Religious groups 84
computational modeling of, cultural transmission by, 36
61–70 Religious polarization, 28
median age of, 143, 144 index, 251–252
membership, 26–28, 30 Religious practice, 3, 369, 371
size of, 142 stagnation of, 363
socioeconomic inequality across, Religious preferences, 5, 7, 28
12–13, 283 cultural transmission of, 6, 35
Religious identity, 25–38 intensity of, 9
cultural transmission, 35–38 Religious rituals, 73–75, 78–81, 84, 86,
esteem maintenance, 31–32 86n10, 87, 145
religious commitment, 32–35 as signals of religiosity, 96
rules based on, 397 Religious schools, 211n7
social sorting, 29–30 in Guatemala, 187–188
Religious legitimacy, 313, 315, 326–330 Religious segregation, 8–9, 89–99
Christian rulers, 314 club good model of religion, 92–93
importance of, 322 religious beliefs, nature and role of,
and joint evolution of culture and 93–95
institutions, 14 signalling religiosity, 96–98
political and economic consequences socialisation, role of, 98–99
of, 13–14, 311–320 See also Segregation
transition implications to Religious strictness, 7, 43
parliamentary legitimation, differentiation dissimilar to, 55–57
315 differentiation in, 44–53
Religious legitimation, 380, 388 applications, 46–50
and identity rules, 384 assessment, 50–53
and state, 379–381 basic model, 44–46
Religious market, 3, 371 differentiation similar to, 54–55
deregulation of, 372 Religious values
Religious market competition, spatial diffusion of, 326
models of, 7, 43–57 joint evolution of, 328
differentiation dissimilar to religious Rent extraction, 84
strictness, 55–57 Rent-seeking model, 251
differentiation in religious strictness Repentance, 81
applications, 46–50 Representative agent, 64
assessment, 50–53 Residential segregation, 89, 90, 90n2
basic model, 44–46 Revival, 440, 441, 444
differentiation similar to religious Revocation, 394
strictness, 54–55 Reward, religious beliefs in, 74–76, 93,
Religious organizations, 2, 3, 7, 26, 31, 95
82, 90, 91, 380n3 Rewards-punishment model, 8
Index 481
Reynal-Querol, M., 12, 236, 249, 251, School-enrollment ratios, 182, 190–191
253–258, 261 School vouchers, 198n1, 208
Richards, J.B., 185n18 in United States, 10, 157–159
Richards, M., 185n18 Scott, J., 391
Richardson, G., 382n5 Secondary school enrollment, private
Richardson, L.F., 429 share in, 197, 198
Rinz, K., 157n3, 159 Secular bureaucracy, 437, 438, 441n5,
Robinson, J.A., 322, 438 442, 444
Roemer, J.E., 417 Secularism, 367
Roe v. Wade, 365 and force, 339
Rohingya Muslims, 229 Secularization, 223–225
Roman Empire, 313–314 of states, 353
fall of, 386, 387 Secularization hypothesis, 265n3,
Roman Law, rediscovery of, 387 268n8, 278, 406n1
Roozen, D., 205–207 Secularization theory, 362
Rubin, J., 1, 3, 13–14, 69, 160, 297, Secular phenomena, for learning
311–313, 315–317, 322, 330, religious practice, 160–161
357, 366n1, 372, 378, 380, Secular regime, emergence of, 338–339
381, 445 Secular states, 328
Rule of law, 15 Segmentation, religion and, 56
Rulers, Religion and Riches: Why the West Segregation
Got Rich and the Middle East and income inequality, 89
Did Not (Jared Rubin), 13–14 religion and (see Religious
Russia, religious commitment, 33 segregation)
Rutte, M., 305n5 residential, 89, 90, 90n2
social mobility and, 89
Self-reported religiosity, 268
S Semantics, 354
Sacerdote, B., 440n4 Sensenbrenner, J., 104
Sacks, M., 7 “A Sermon on Keeping Children in
al-Sadar, A., 338 School” (Martin Luther), 216
St. Augustine, 384 Seror, A., 14
Sala-i-Martin, X., 255 Seven Sisters of American Protestantism,
Saleh, M., 12, 284, 285, 288–290 171n5
Saliba, G., 441, 442, 442n6 Shapiro, J.M., 162
Sanctification, 78 Sharawi, H., 33
Sander, W., 160, 204, 205 Sharia, 145, 146
Saudi Arabia Shariah compliance, 56
oil wealth of, 339 Sharia law, 451
religious strictness, 50 Shearer, R.E., 170n3
Schaltegger, C., 222n4 Sherkat, D.E., 407
Scheve, K., 74n3, 417 Sherman, R., 364
Schmidt, L.E., 173n10 Shi’ism, 343
482 Index
Shintoism, 16 Socioeconomic status (SES)

Shrivastava, A., 232 initial selection on, 288
Sibley, C.G., 272 self-selection on, 284
Siedlarek, J.-P., 105, 107–111 of surviving non-Muslim minorities,
Signalling religiosity, 96–98 286
by behaviour, 96–98 Sociopolitical innovations, 430
rituals, as signals of religiosity, 96 Solt, F., 406n1
Silvestri, S., 249 Soper, J.C., 367
Sims, C.A., 189 Sorondo, M.S., 1–2
Sin, 82 Sosis, R., 26, 80, 276n24
Skeen, B.A., 453 South Asia, 3
Skepticism, 444 divorce in, 148
in Islam, 456 growth of Hinduism in, 10
Skinner, B.F., 276n24 religious demography in, 142
Skirbekk, V., 143, 149 Southern Europeans, 299
Skocpol, T., 365 Sovereignty, 305
Slave armies, 329–330 Spain, religious commitment, 33
Slave soldiers, 443–445 Spanish municipalities, 454
Smith, A., 2, 6, 74, 122, 155–156, 350, Spatial-competition analysis, 55
351n6, 352, 369, 373 Spatial models, of religious market
Smith, M.E., 90 competition, 43–57
Social advantages of economies of scale, differentiation dissimilar to religious
427–428 strictness, 55–57
Social “culture,” 430 differentiation in religious strictness
Social interactions, 93 applications, 46–50
Socialisation, role in religious assessment, 50–53
segregation, 98–99 basic model, 44–46
Social mobility, segregation and, 89 differentiation similar to religious
Social preferences, effect of institutions strictness, 54–55
on, 9–10 Spatial models of religious market
Social sorting, 6, 29–30 competition, 7
Social stability, 424 Spenkuch, J., 219
Societal policy game, 323–324 Spilka, B., 276n23
Society, types of, 328 Sri Lanka, minority group hypothesis of
Socio-economic equilibrium pattern, fertility, 149
325 ST, see Statesman’s Yearbook
Socioeconomic inequality Stanig, P., 417
across religious groups, 12–13, 283 Stark, R., 2, 37, 46, 90, 306n6, 349,
between Copts, study of, 284 369, 371, 373, 427–429, 433
religion literature economics, 284 Stasavage, D., 74n3, 417
Socioeconomic outcomes, 265 State
religious beliefs effect on, 283, 284 and religion, 347–348, 351–352
Socio-economic policies, 323 alliance of, 351–352
Index 483
analytical framework, 348–351 Sweden, religious participation in, 6

decline of, 352–354 Switzerland
economic approach to, 14 religious commitment, 33
merger of, 352–355 signalling religiosity by behaviour, 97
suppression of religion, 355–356 Synthetic panel, 274
technology and economic
decisions regulation, 356–358
religious legitimation and, 379–381 T
State–church relationship, 10, 155–167 Tabory, E., 260
State-funded churches, 6 Tax collection, 349, 351, 357
State industrialization program, 289 Tax-induced conversions hypothesis,
State institutions, centralization of, 356 285
State secularism, religious concentration Tax papyri, 286–287
and, 353 Tax revenues, 351
Statesman’s Yearbook (ST), 252–255 Tax system, 291–292
State-supported church, 155–156 Theocracy, 352–355
State-supported religion, 156 Theologians, 429
Statistical Office of the German Empire, Theoretical advances, in economics of
215 religion, 5–10
Stigma, 26, 38 Theories, 266–268
Stone-Geary utility function, 202–203 Theory of sacrifice and stigma, 64
Storms, 277 Theory of the second best, 27
Strength of Religiosity Scale, global Thirty Years’ War, 221
average of, 269, 270n10 Thuesen, P.J., 171n5
Strong religious states, 328–330 Time-varying factors, 274
Stroumsa, S., 440n4 Tirole, J., 74n3, 77, 83, 408, 417
Structural-functionalist school, 424 Tolerance, missing role of, 456–457
Subervi-Velez, F.A., 104n1 Toleration Act, 393
Sub-Saharan Africa, 16 Tollison, R.D., 430
religious demography in, 143 Torgler, B., 222n4
Subsidized religious private schools vs. “To the Christian Nobility of the
non-sectarian private schools, German Nation Concerning
203 the Improvement of the
means-tested vouchers, enrollment Christian Estate” (Martin
of, 203–204 Luther), 217
Sudan, Nuer religion in, 76 “To the Councilmen of All Cities in
Sullivan, A.R., 150 Germany That They Establish
Summer Institute of Linguistics, 173n8 and Maintain Christian
Summer of Love, 25 Schools” (Martin Luther), 216
Sunna, 145–146 Totonicapán, 182
Sunni Revival, 16, 438, 440, 444, 454 Trade
Supply-side theories, 267, 271n13 and Islamic principles, 450–451
Svensson, I., 258 as powerful gravity of Islam, 452–453
484 Index
Trade fosters innovation, 449 school vouchers in, 10, 157–159,

Trade-off, 339–340, 355–357 198n1, 208
Traditional societies, 334 secular phenomena, for learning
Transition process, 112 religious practice, 160
Translation movement, 439 signalling religiosity by behaviour,
Tribalism, 339 96–98
Turkish War of Independence, 103 Summer of Love, 25
“Two-Speed Europe,” 305n5 Unobservable district-level factors,
275
Unobservable time-varying factors,
U 274
Ulema (Muslim clerics), 146 Urban households, 238, 240–241
Ultra-Orthodox Judaism, 28, 32, 108 US, see United States
Umayyad dynasty, 439n1, 442, 453 US Constitution, 364
United Kingdom Utility maximization, 64
Industrial Revolution, role of
education in, 220
religious participation in, 6 V
United States (US, USA) Varshney, A, 230, 231
anti-abortion terrorism, 82n7 Vatican II, 170
characteristics hypothesis of fertility, Veiling, 33–35
149 Verdier, T., 9, 14, 35, 104, 105n2, 109,
Christian denominations in, 112–114, 207, 322, 325
204–205 Violence, 389
church bulletins, 161–166 Voas, D., 268n8
Church-Sect Cycle, 67 Voll, J., 366–377
General Social Survey, 158 Voting, 417–419
minority group hypothesis of fertility,
149
private education, religious factor in, W
197–211 Wagener, A., 44, 56
Private School Survey, 157, 211 Wahhabism, 318
public education vs. private Wallis, J., 397
education, 157 Wallsten, K.J., 115
regional religiosity, stability of, 98 Waqf, 444, 450
relative religiosity in, 270 Warner, C., 370
religious demography in, 142 “War on terror,” 365
religious factor in private education, War, Peace, and Prosperity in the Name of
11 God (Murat Iyigun), 15
religious participation in, 6 Wars of Religion, 379
religious segregation in, 91 WCE, see World Christian Encyclopedia
religious strictness, 50 The Wealth of Nations (Smith), 2, 6,
residential segregation in, 89, 90 122–123, 155, 350
Index 485
Weber, M., 11, 74, 77, 84, 94, 96, 219, Woodberry, R.D., 170n3, 171
250, 283–285, 289, 316, 407, World Christian Encyclopedia (WCE),
420 252–255
Weninger, T., 10, 155 World Factbook (WF), 253
West, M.R., 358 World Religion Database, 172
Western Christianity, 250 World Value Survey (WVS), 252
Western Europe Wycliffe Bible Translators, 173n8
absolute monarchies in, 334 Wycliffe Dictionary of Theology, 76
and Middle East
long-run economic divergence
between, 14 Y
political economy differences Yanagizawa-Drott, D., 408
between, 14 Yang, F., 372
nation-states in, 366 Yavuz Sultan Selim, 423
rulers in, 342 Yuchtman, N., 438
scientific development of, 439–440
Western world, economic rise of, 438
WF, see World Fact book Z
Wilcoxon-Mann-Whitney test, 128, 129 Zacapa, 182
Wilkinson, S., 230, 231, 235 Zelman et al. v. Simmons-Harris et al,
Wintrobe, R., 355 199n3, 203n5
Within-country analysis, 270n11, 271 Zenou, Y., 104, 105n2, 109, 112–114
Woessmann, L., 11, 215, 217–219, 222, Zhou, M., 104n1
224, 358, 407 Zwingli, H., 170
Wolitzky, A., 114

Jean-Paul Carvalho, Sriya Iyer, Jared Rubin - Advances in The Economics of Religion (2019, Springer International Publishing - Palgrave Macmillan)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jean-Paul Carvalho, Sriya Iyer, Jared Rubin - Advances in The Economics of Religion (2019, Springer International Publishing - Palgrave Macmillan)

Uploaded by

Copyright:

Available Formats

ADVANCES IN

IEA Conference Volume No. 158

Other titles from IEA:

INSTITUTIONS, GOVERNANCE AND THE CONTROL OF CORRUPTION

INEQUALITY AND GROWTH: PATTERNS AND POLICY

INEQUALITY AND GROWTH: PATTERNS AND POLICY

THE MIDDLE EAST ECONOMIES IN TIMES OF TRANSITION

TOWARDS A GENERAL THEORY OF DEEP DOWNTURNS

CONTEMPORARY ISSUES IN MACROECONOMICS

CONTEMPORARY ISSUES IN MICROECONOMICS

CONTEMPORARY ISSUES IN DEVELOPMENT ECONOMICS

TAMING CAPITAL FLOWS

LIFE AFTER DEBT

INCOME CONTINGENT LOANS

THE INDUSTRIAL POLICY REVOLUTION I

THE INDUSTRIAL POLICY REVOLUTION II

THE CHINESE ECONOMY

INSTITUTIONS AND COMPARATIVE ECONOMIC DEVELOPMENT

CORPORATE SOCIAL RESPONSIBILITY AND CORPORATE GOVERNANCE

IS ECONOMIC GROWTH SUSTAINABLE?

KEYNE’S GENERAL THEORY AFTER SEVENTY YEARS

CORRUPTION, DEVELOPMENT AND INSTITUTIONAL DESIGN

MARKET AND SOCIALISM

INSTITUTIONAL CHANGE AND ECONOMIC BEHAVIOUR

INTERGENERATIONAL EQUITY AND SUSTAINABILITY

PSYCHOLOCY, RATIONALITY AND ECONOMIC BEHAVIOUR

MULTINATIONALS AND FOREIGN INVESTMENT IN ECONOMIC DEVELOPMENT

POST-CONFLICT ECONOMIES IN AFRICA

STRUCTURAL REFORM AND MACROECONOMIC POLICY

THE PAST, PRESENT AND FUTURE OF THE EUROPEAN UNION

LATIN AMERICAN ECONOMIC CRISES

ADVANCES IN MACROECONOMIC THEORY

TRADE, INVESTMENT, MIGRATION AND LABOUR MARKET ADJUSTMENT

INEQUALITY AROUND THE WORLD

MONETARY THEORY AND POLICY EXPERIENCE

ECONOMIC DEVELOPMENT IN SUBSAHARAN AFRICA

International Economics Association

International Economic Association Series

Library of Congress Control Number: 2019930427

© The Editor(s) (if applicable) and The Author(s) 2019

This book comes out of a roundtable on the Economics of Religion organized

LSE Tim Besley

Part I Theoretical Advances in the Economics of Religion 23

2 Religious Clubs: The Strategic Role of Religious Identity 25

3 Spatial Models of Religious Market Competition:

4 When Average Is Irrelevant: Computational Modeling

5 The Intelligent Design of Religious Beliefs 73

6 Religion and Segregation 89

7 Religious and Cultural Leaders 103

8 Intermediated Social Preferences: Altruism in an

Part II Empirical Advances in the Economics of Religion 139

9 Religion and Demography 141

10 Economics and Church State Research: Past, Present, Future 155

11 Protestants and Catholics and Educational Investment

12 The Religious Factor in Private Education in the United States 197

13 How Luther’s Quest for Education Changed German

14 Hindu-Muslim Violence in India: A Postscript from the

15 Religion and Conflict: A Quantitative Approach 249

16 Why Are Some Societies More Religious Than Others? 265

17 Socioeconomic Inequality Across Religious Groups:

18 Religion and the European Union 295

Part III Advances in Religion and Political Economy 309

19 The Political and Economic Consequences of Religious