RELIGION and

BR AZILIAN
DEMOCR ACY
Mobilizing the People of God
Amy Erica Smith
Religion and Brazilian Democracy

As Brazilian democracy faces a crisis of legitimacy, political divisions

grow among Catholic, evangelical, and nonreligious citizens. What has
caused religious polarization in Brazilian politics? Does religious poli-
tics shore up or undermine democracy? Religion and Brazilian
Democracy: Mobilizing the People of God uses engaging anecdotes
and draws on a wealth of data from surveys and survey experiments
with clergy, citizens, and legislators to explain the causes and conse-
quences of Brazil’s “culture wars.” Though political parties create
culture-war conflict in established democracies, in Brazil’s weak party
system religious leaders instead drive divisions. Clergy leverage legisla-
tive and electoral politics strategically to promote their own theological
goals and to help their religious groups compete. In the process, they
often lead politicians and congregants. Ultimately, religious politics
pushes Brazilian politics rightward and further fragments parties. Yet
Religion and Brazilian Democracy also demonstrates that clergy-led
politics stabilizes Brazilian democracy and enhances representation.

Amy Erica Smith is Associate Professor of Political Science at Iowa State

University. Smith’s research has attracted funding from the National
Science Foundation, Fulbright, Mellon, and Templeton, and the Award
for Early Achievement in Research at Iowa State University. Her work
on democracy in developing countries has appeared in top political-
science journals and in the Portuguese-language book Legitimidade
e qualidade da democracia no Brasil: Uma visão da cidadania (2011,
with Lucio Rennó, Matthew Layton, and Frederico Batista Pereira).
Cambridge Studies in Social Theory, Religion, and Politics

Editors
David E. Campbell, University of Notre Dame
Anna M. Grzymala-Busse, Stanford University
Kenneth D. Wald, University of Florida, Gainesville
Richard L. Wood, University of New Mexico
Founding Editor
David C. Leege, University of Notre Dame

In societies around the world, dynamic changes are occurring at the intersection of religion
and politics. In some settings, these changes are driven by internal shifts within religions; in
others, by shifting political structures, institutional contexts, or by war or other upheavals.
Cambridge Studies in Social Theory, Religion, and Politics publishes books that seek to
understand and explain these changes to a wide audience, drawing on insight from social
theory and original empirical analysis. We welcome work built on strong theoretical framing,
careful research design, and rigorous methods using any social scientific method(s)
appropriate to the study. The series examines the relationship of religion and politics
broadly understood, including directly political behavior, action in civil society and in the
mediating institutions that undergird politics, and the ways religion shapes the cultural
dynamics underlying political and civil society.

Mikhail A. Alexseev and Sufian N. Zhemukhov, Mass Religious Ritual and Intergroup
Tolerance: The Muslim Pilgrims’ Paradox
Luke Bretherton, Resurrecting Democracy: Faith, Citizenship, and the Politics of a Common
Life
David E. Campbell, John C. Green, and J. Quin Monson, Seeking the Promised Land:
Mormons and American Politics
Ryan L. Claassen, Godless Democrats and Pious Republicans? Party Activists, Party Capture,
and the “God Gap”
Darren W. Davis and Donald Pope-Davis, Perseverance in the Parish? Religious Attitudes
from a Black Catholic Perspective
Paul A. Djupe and Christopher P. Gilbert, The Political Influence of Churches
Joel S. Fetzer and J. Christopher Soper, Muslims and the State in Britain, France, and
Germany
François Foret, Religion and Politics in the European Union: The Secular Canopy
Jonathan Fox, A World Survey of Religion and the State
Jonathan Fox, Political Secularism, Religion, and the State: A Time Series Analysis of
Worldwide Data
Anthony Gill, The Political Origins of Religious Liberty
Brian J. Grim and Roger Finke, The Price of Freedom Denied: Religious Persecution and
Conflict in the Twenty-First Century
Kees van Kersbergen and Philip Manow, editors, Religion, Class Coalitions, and Welfare States
Mirjam Kunkler, John Madeley, and Shylashri Shankar, A Secular Age beyond the West:
Religion, Law and the State in Asia, the Middle East and North Africa

(continued after index)

Religion and Brazilian Democracy
Mobilizing the People of God

AMY ERICA SMITH

Iowa State University
University Printing House, Cambridge cb2 8bs, United Kingdom
One Liberty Plaza, 20th Floor, New York, ny 10006, USA
477 Williamstown Road, Port Melbourne, vic 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
79 Anson Road, #06–04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781108482110
doi: 10.1017/9781108699655
© Amy Erica Smith 2019
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2019
Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A.
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
names: Smith, Amy Erica, 1976– author.
title: Religion and Brazilian democracy : mobilizing the people of God / Amy Erica Smith.
description: Cambridge, UK ; New York, NY : Cambridge University Press, 2019. | Series:
Cambridge studies in social theory, religion and politics | Includes bibliographical references
and index.
identifiers: lccn 2018045104 | isbn 9781108482110 (hardback)
subjects: lcsh: Christianity and politics – Brazil. | Evangelicalism – Political aspects – Brazil.
| Catholic Church – Political activity – Brazil. | Democracy – Brazil. | Social change – Political
aspects – Brazil. | Brazil – Politics and government – 1985– | BISAC: POLITICAL SCIENCE /
Government / General.
classification: lcc br115.p7 s56 2019 | ddc 261.70981–dc23
LC record available at https://lccn.loc.gov/2018045104
isbn 978-1-108-48211-0 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
For Tibi, Oscar, and Adam
Com muito amor
Ş i cu drag
Contents

List of Figures page viii

List of Tables x
Acknowledgments xi

part i introduction 1
1 Introduction 3
2 Clergy, Congregants, and Religious Politicians 26
3 Methods and Case Studies 46
part ii what clergy think, say, and do 59
4 What Clergy Think and Say: Religious Teachings and Political Views 61
5 What Clergy Do: Encouraging Partisan and Electoral Politics 80
part iii how congregants respond 97
6 Church Influence on Citizens’ Policy Views and Partisanship 99
7 Church Influence on Voting Behavior 113
8 Church Influence on Citizen Support for Democracy 130
part iv representation 147
9 The Representational Triangle 149
10 Conclusion: Mobilizing the People of God 162

Afterword 176
Appendix A. List of Focus Groups and Church Observations 178
Appendix B. Focus-Group Protocol 181
Appendix C. Variable Coding and Information on Studies 183
References 184
Index 203

vii
Figures

1.1 Religious affiliation in Brazil page 11

1.2 Characteristics of Brazilians choosing non-Catholic
religious affiliations 15
2.1 Relationships of influence and representation within
religious groups 44
3.1 Membership changes, outreach, and competition (clergy
reports) 52
3.2 Core religious beliefs at the sites of the quantitative
congregational study 56
4.1 Priorities in policy-related teachings, by religious tradition 65
4.2 Competitive threat affects core and policy-related
teachings – but only among Catholics 69
4.3 Perceived state neutrality or bias 75
4.4 Determinants of perceived neutrality or bias of the political system 76
5.1 Church-leader discussion of election campaigns 83
5.2 Competitive threat affects legislative advocacy and candidate
endorsements 90
5.3 The quasi-experimental impact of the grievance treatment
on clergy support for political activism 95
6.1 Religious affiliation and policy preferences 103
6.2 Religious affiliation and support for the Workers’ Party 105
6.3 Clustering in policy views at eight congregations and
community sites 108
6.4 Determinants of variance in policy views 111
7.1 Campaigning and electoral discussion in eight
congregations, Juiz de Fora, 2014 118
7.2 Congregational messages and turnout in Juiz de Fora,
2008, and Brazil, 2014 121

viii
List of Figures ix

7.3 Experimental impact of clergy campaigning and candidate

issue stances on citizens 125
7.4 The impact of candidate religious characteristics, by
respondent secular norms 127
8.1 Attitudes toward the political system within eight
congregations 134
8.2 Determinants of attitudes toward the state and democratic regime 135
8.3 Religious affiliation and intergroup attitudes 138
8.4 Religion, church attendance, and political tolerance 140
8.5 Protest participation by religious affiliation, over time 144
9.1 Policy views of federal legislators, by religion 154
9.2 Differences between Catholics and evangelicals in policy
attitudes, for clergy, citizens, and legislators 156
9.3 The perceived electoral bases of federal legislators, by religion 159
Tables

3.1 Sample statistics, clergy study page 51

5.1 Characteristics associated with clergy support for
political activity (all religious traditions combined) 93

x
Acknowledgments

As always in a project of this scope, I have acquired legion debts to people and
institutions over the course of the research and writing of this book. First,
I gratefully acknowledge a number of sources of external funding. A Fulbright
Postdoctoral Fellowship funded my research stay in Brazil in July–November,
2014, and a Small Research Grant from the American Political Science
Association helped to fund the quantitative and qualitative studies I executed
in that period. A Regional Faculty Research-Travel Award from the Center for
Latin American and Caribbean Studies at the University of Wisconsin-
Milwaukee funded a planning visit in 2014, prior to my extended research
trip. A National Science Foundation Doctoral Dissertation Improvement
Grant (Award Number 0921716) helped to support both the 2008 Local
Elections Study and the 2010 Brazil Electoral Panel Study.
Straddling the line between internal and external funding is the support
I have received from the Kellogg Institute for International Studies at the
University of Notre Dame, where I was a Visiting Fellow in the 2016–2017
academic year. The Kellogg Institute provided material support for writing and
research over the course of that year, as well as opportunities to present and
workshop my work. The Institute also funded a book workshop at Notre Dame
in April, 2017. I am tremendously grateful to Director Paolo Carozza and
Associate Director Sharon Schierling, and to the Institute’s truly exceptional
and much-loved professional staff, including Denise Wright, Judy Bartlett,
Karen Clay, Therese Hanlon, and Elizabeth Rankin.
Finally, internally to Iowa State University (ISU), this research has been
supported by three Faculty Small Grant Awards from the College of Liberal
Arts and Sciences (2012–2013, 2014–2015, and 2016–2017), and by two
Foreign Travel Grants from the Provost’s Office. For help weaving all of these
resources together, I am very grateful to my chair, Mack Shelley, who has been
committed to . . . well, as one would say in Portuguese, “dar um jeito” – Mack
has always been sure we can find a way to make all the funding sources and

xi
xii Acknowledgments

travel plans and research leaves work out. I am also indebted to two highly
capable administrative support professionals in the Department of Political
Science at ISU, Shirley Barnes and Donna Burkhart, who have worked with
me – sometimes in spite of me – to try to make sure people get paid.
I also owe debts of gratitude to several people and institutions for access to
data. Thanks to the Latin American Public Opinion Project (LAPOP) and its
major supporters (the United States Agency for International Development, the
Inter-American Development Bank, and Vanderbilt University) for making the
AmericasBarometer data available. I am especially grateful for advanced access
to the 2017 Brazilian survey. Thanks also to Tim Power and César Zucco for
early access to the 2013 Brazilian Legislative Studies (BLS), and for introducing
me and Taylor Boas to those data by way of an invitation to a 2014 workshop at
Oxford. Last, Lucas Mingardi, Rafael Mucinhato, and Sergio Simoni, doctoral
candidates at the University of São Paulo, kindly shared with me data they had
painstakingly compiled on the religious affiliations of all Brazilian deputies
from the return to democracy through the 2007–2010 legislative period.
In Brazil, I have been very fortunate to have the help and friendship of many
people. My longest-running debt is to Ana Paula Evangelista Almeida, with
whom I have been working since 2008. Ana Paula and I first crossed paths when
she and Rafaela Reis – at the time both highly competent undergraduates in the
Department of Social Sciences at the Federal University of Juiz de Fora (UFJF) –
worked as my research assistants supervising a team of other undergraduate
interviewers in a study of the 2008 local election in Juiz de Fora. When
I returned to the same city in 2014 and 2017, Ana Paula worked as my
research coordinator and fieldwork supervisor. Much of the data presented in
this book bears Ana Paula’s thumbprints. Ana Paula is supersmart,
intellectually curious, hardworking, creative, funny, generous, and frugal with
other people’s money. Our work has been a true partnership, and I have learned
much from her. I am delighted that she will soon be finishing her own PhD in
rural sociology. She is already an excellent professor.
I have benefited from many other excellent research assistants in Brazil. Two
who stand out are Rafaela Reis (the abovementioned supervisor in 2008, and
herself now a professor at the UFJF) and Mariana Gonzales, who worked on the
2014 study. I am particularly grateful to Mariana for her extraordinarily good
summaries/transcripts of the focus groups we conducted. The 2014 research
also benefited from the research assistance of Pedro Makla, Suelen Masson,
Linecker Mauler, Marlon Moreira, and Aylla Prata in Juiz de Fora; Paulo Vítor
Del Rey da Silva and Júlia Vieira dos Santos in Rio de Janeiro; and Suiany Silva
de Moraes, Tuany Sousa, and Erivaldo Teixeira in Fortaleza. Thanks to Jakson
Alves de Aquino, professor at the Federal University of Ceará, and to Cesar
Zucco, professor at the Instituto Getúlio Vargas (IGV) in Rio, for their help
finding student research assistants in those two cities. Thanks also to Cesar and
Dani for their hospitality at the IGV.
Acknowledgments xiii

In the preliminary stages of this research, several students helped with

literature reviews; thanks to Luíza Helena Almeida at the University of
Brasília and Joe Gettemy and Marisa Wilson at Iowa State University. Joe’s
work was distinguished by intellectual curiosity and excellent language ability.
At the UFJF, I am grateful for intellectual community and practical help from
Magali Cunha (a professor at the Methodist University of São Paulo who visited
UFJF), as well as UFJF professors Emerson Sena Silveira and Marta Mendes.
I have three collaborators with whom I have been writing and thinking about
Brazilian politics for much of my academic career. Barry Ames (my dissertation
advisor), Taylor Boas, and Matt Layton have each shaped the way I think about
political behavior in Brazil. My debts to these three are practical as well as
intellectual. Matt and Barry both read and workshopped the manuscript.
Taylor was my collaborator in collecting the data for the two online studies
conducted via Facebook recruitment in 2012 and 2014, and he and I have
coauthored a paper using the BLS data, exploring opinion congruence
between evangelical voters and politicians. Taylor and I, in particular, have
been walking parallel and sometimes overlapping paths as we work to
understand religion and politics in Brazil.
Beyond Barry and Matt, four other people participated in my book
workshop at Kellogg. I am very grateful to Mark Brockway, Fr. Bob Dowd,
Emma Rosenberg, and Guillermo Trejo for reading the manuscript, and for
their insightful advice.
I am also deeply indebted to the ever kind and smart Dave Campbell, as well
as his coeditors within this Cambridge series; and to Sara Doskow at
Cambridge, who has been a model of professionalism. Two anonymous
reviewers helped me improve this manuscript tremendously – thank you for
your patience and insight.
Parts of what would eventually become this book have been presented at
a number of institutions, including talks at Oxford University, Purdue
University, the University of Notre Dame, the University of Western Ontario,
and the Inter-American Development Bank. I have also presented this work in
a conference format at the 2012 meeting of the Latin American Studies
Association, the 2013 Annual Meetings of the American and Southern
Political Science Associations, and the 2015 Meeting of the Midwest Political
Science Association. I workshopped this work in reading groups at ISU and
Kellogg, and in the online Gender and Political Psychology group.
Beyond those I’ve already mentioned, at Kellogg I suffered (yes, suffered!)
from feedback from many smart people, including Juan Albarracín, Márcio
Bahia, Rodrigo Castro Cornejo, Hernán Flom, Lauren Honig, Stuart Kirsch,
Victor Maqque, Ann Mische, Maggie Triyana, George Tsebelis, and Samuel
Valenzuela. Outside Kellogg, I am grateful for comments at various points in
this long project from Paul Djupe, Cornelia Flora, Ken Greene, Fran Hagopian,
Jonathan Hassid, Erin Heidt-Forsythe, Greg Love, Dave Peterson, Tim Power,
Rachel Riedl, Heather Rice, Robert Urbatsch, and L. J. Zigerell. Aaron Javsicas
xiv Acknowledgments

and Robin Globus Veldman gave extensive and very useful feedback on an early
version of the prospectus, and Jay McCann and Liz Zechmeister provided
insightful comments on a later version of the same. More generally, Jay, Liz,
and Paul have been exemplars of friendship and mentorship over these many
years. Though David Samuels has not commented on this project, discussions
with him about partisanship and voting behavior in Brazil shaped my thinking
as I was writing the book.
Earlier versions of pieces of this book were previously published in two
articles: “When Clergy Are Threatened: Catholic and Protestant Leaders and
Political Activism in Brazil,” Politics and Religion 9 (3): 431–55, in 2016; and
“Democratic Talk in Church: Religion and Political Socialization in the Context
of Urban Inequality,” World Development 99 (November 2017): 441–51.
Thanks to Cambridge University Press and to Elsevier.
Many people opened their homes and hearts to me and my family in Brazil.
José Luiz Britto Bastos and Dalila Freitas were extraordinarily helpful and
hospitable with my 2014 research visit, and became my children’s “Brazilian
grandparents.” Lilián Costa Magalhães, whom I met in 2014, became a good
friend, and she and seven-year-old Luiza were excellent hosts in 2017. Finally,
I am grateful for my time with Elisângela Andrade, a dear friend whom I met in
2008 and who became my children’s nanny when I returned in 2014. Elisângela
passed away in July 2015.
And last, my “real” family. Tibi Chelcea has been with me all the way from
my first visit to a Catholic parish in the city of Juiz de Fora in 2008. He spent
nine months with me in Brazil in 2008–2009, and four months in 2014. He has
managed with remarkable good humor my research trips away, my
conferencing, and my general distractedness and workaholic tendencies, and
he has never failed in his confidence in me and my work. I am so lucky to have
suckered this kind, cheerful feminist into putting up with me and taking care of
our two boys all these years. My mom, Esther Smith, and mother-in-law, Adina
Chelcea, have also been pillars of our little village, helping with childcare, cat-
sitting, plant-sitting, etc. Oscar and Adam joined our family on this journey,
and came along for the ride to Juiz de Fora in 2014 and South Bend in
2016–2017; they are the most wonderful distractions from work. This book is
dedicated to Tibi, Oscar, and Adam.
part i

INTRODUCTION
1

Introduction

On November 7, 2017, conservative opponents of “gender ideology” burned

the American feminist theorist Judith Butler in effigy on the street in front of the
art institute Sesc Pompeia in the city of São Paulo, while Butler herself was inside
giving a lecture. As protesters hoisted a life-sized doll in a pink brassiere and
witch’s hat over their heads and set it on fire, they chanted, “Burn the witch!”
The protest apparently included both conservative Catholics and evangelicals.
Though Catholic crucifixes were on prominent display during the
demonstration, evangelical groups built much of the momentum behind the
protests. In the days leading up to the talk, a Facebook group and website led by
Assembly of God clergy from the city of Ilha Solteira (São Paulo state) drove
traffic to an online petition that gathered 366,000 signatures, opposing Butler’s
visit (J. Gonçalves 2017). While the protest was cast in the media as an attempt
to shut down the conference, a survey conducted with protesters at the event
itself found that most did not aim to stop Butler’s talk (Calegari 2017). Rather,
they hoped to stimulate a debate over gender, sexuality, and the role of public
schools in sexual education.
Gender and sexuality have become perhaps the most important issues driving
a recent period of religiously motivated democratic conflict in Brazil – what I
term Brazil’s “culture wars.” Protestant clergy, congregants, and
representatives are far to the right of Catholics and the nonreligious on
matters such as transgender rights or public-school sex education.
Meanwhile, religious conservatives and secular voters battle over whether to
entirely outlaw or fully legalize abortion, which is presently legal only under
conditions of rape or danger to the mother’s life. On this issue, Catholics are
sometimes to the right of evangelicals. And Catholics, evangelicals, and the
nonreligious each take opposing stances on a third issue: the rights of religious
communities in the context of a formally secular state. Conflicts spill into
elections, as a growing evangelical voting bloc favors religious conservatives,
and especially coreligionist candidates.

3
4 Introduction

However, one typical source of religious and political division is

conspicuously absent: partisanship. In Brazil’s famously weak and fragmented
party system – with twenty-eight parties elected to the lower house of Congress,
the Chamber of Deputies, in 2014 – neither Catholics nor evangelicals have
much of a partisan home. The correlation between religion and partisanship is
near zero. This contrasts markedly with the United States. Though Hunter
(1992) first popularized the term “culture wars” to describe conflict between
religious and secular citizens in the United States, later scholarship
demonstrates that cultural polarization there is tightly bound to parties.
This book is about the causes and consequences of Brazil’s culture wars.
How has conflict developed, absent partisan leadership? And how have the
culture wars affected Brazil’s post-1985 democracy? Most urgently, does
religious politics either threaten or help to shore up a democracy now facing
grave challenges to its legitimacy? I will argue that the answers lie not in parties,
but in clergy, interacting with and sometimes leading congregants and
politicians. Religious leaders have a complex mix of motivations for getting
involved in politics, including religious and political ideals, but also anxiety
over religious competition. Ultimately, religious politics polarizes Brazilian
politics and pushes it to the right, while contributing to partisan
fragmentation; yet it also enhances democratic representation and stabilizes
democracy by giving religious leaders a stake in the system.
Before we proceed, some definitions are in order. What is a “culture war”?
How do we know whether Brazil (or any other country) is in the midst of one?
I define “culture wars” as pervasive and prolonged democratic conflicts within
polities, between social groups who perceive their worldviews as fundamentally
mutually incompatible.1 By “democratic” conflict, I mean that culture warriors
primarily use democratic arenas and weapons – elections, policy debates, and
persuasion – to influence public opinion. This does not preclude physical
violence, but outbreaks of violence are usually peripheral and nonstrategic. By
“pervasive and prolonged,” I mean that an isolated skirmish does not constitute
a culture war. Rather, many groups in society choose sides, levels of hostility are
elevated, and the conflict extends across various battles. Some groups may aim
for ultimate social or political dominance; others may want discrete policy
changes.
Who are culture warriors? The competing worldviews driving culture wars
are typically delineated by religions – meaning sets of ideas and practices that
communities develop to describe transcendent forces, and to derive
prescriptions for human behavior. Culture wars often involve conflict

1
The second half of this definition is similar to that of Hunter (1992), emphasizing competing
groups defined by fundamental worldviews. However, thinking about the culture wars in a
comparative context reveals assumptions that are likely implicit in Hunter’s definition. To wit,
I differ from Hunter in emphasizing the methods – democratic politics – and the extension of the
conflict.
Introduction 5

between the two poles on a continuum of religious devotion: “seculars” vs.

“religious conservatives.” However, culture wars can also occur between
members of different religious communities – for instance, evangelicals and
Catholics, or Muslims and Christians. In Brazil, the culture wars take place
on two fronts simultaneously: between religious and secular citizens, and
between evangelicals and Catholics. Sometimes evangelicals and Catholics are
allies, and at other times they are in conflict.
In Brazil, culture-war opponents agree about many issues beyond sexuality,
the family, and church–state relations. Yet though disagreement may be fairly
narrow in scope, it is deep. Views on contested issues are deeply held and
expressed in sacred, stark, black-and-white terms. Culture warriors perceive
issues such as gay rights or sex education as existential threats to their own
group, perhaps involving a struggle against supernatural evil forces.
Disagreement intensifies into conflict as it is reinforced by the fault lines of
religious identity.
To return to our first puzzle: what drives Brazil’s culture wars, if not parties?
This book takes a clergy-centered approach. Two shocks have triggered clergy
activism. First was a leftward shift in Brazilian society and public policy on
issues related to sexuality, gender, and family roles. Conservative religious
leaders perceive policies such as the high court’s legalization of same-sex
marriage, in a pair of decisions in 2011 and 2013, as deeply threatening to
social order. The second shock entails an increasing fragmentation of the
religious landscape – a shift away from monolithic Catholicism, toward both
religious nonadherence and evangelicalism – that has intensified interreligious
competition for what I will term “souls and money.”
Both ideas and group interests motivate Catholic and evangelical clergy. On the
one hand, the experimental evidence presented in Part II shows that Catholics and
evangelicals both hold inflexible stances on one all-important issue, homosexuality,
anchoring Catholics to the center and evangelicals to the right on this issue.
Abstract support for the democratic regime is also high and unbudging. On the
other hand, competition to attract and keep souls also affects clergy behavior and
attitudes. In the two-front culture wars, Catholic clergy contend with both
secularism and evangelicalism; strategic calculations in response to the threat of
membership loss sometimes push Catholics to de-emphasize certain “culture-war”
issues. At other times, membership pressures draw both evangelicals and Catholics
into activism, or repress both evangelical and Catholic speech when clergy fear
controversy. Meanwhile, clergy who perceive their group to be unfairly treated lose
faith in the legitimacy of the political system.
Do clergy influence congregants’ behaviors and attitudes? If so, how? Part III
of this book shows that clergy can shape citizens’ issue attitudes, turnout, voting
behavior, and democratic dispositions. Nonetheless, influence is partial –
affecting some attitudes and behaviors more than others – and asymmetric –
affecting some citizens more than others. The great majority of citizens hold
6 Introduction

secular democratic norms that lead them to resist some types of clergy influence.
Clergy are more influential on issues seen as core religious concerns, such as
those related to sexuality and the family, and less effective in guiding other
attitudes as well as vote choice. Doctrinally conservative citizens and
congregations are more readily influenced than others, yet even in the most
politically effective Pentecostal denomination, the Universal Church of the
Kingdom of God (UCKG), influence is far from automatic. In the early 2000s,
UCKG leaders projected that just 20 percent of “their” voters supported in-
group candidates (Conrado 2001). The partial and asymmetric influence of
clergy pushes Brazilian politics to the right, as religious conservatives are most
likely to be influenced.
Clergy also affect citizens’ democratic attitudes. On the one hand, clergy
convey their own robust support for democracy to citizens. On the other, clergy
who perceive the political system as biased against their religious in-group
undermine congregants’ confidence in that system. In addition, clergy who
promote dualistic, good-versus-evil visions of social conflict can contribute to
intolerance of out-groups such as atheists and gays.
Part IV argues that religious leaders also have substantial leverage over
politicians whom they choose to support. In Brazil’s highly permissive party
and electoral systems, hundreds of candidates run in most legislative districts,
and religious leaders have great latitude to get their chosen candidates onto
ballots. Furthermore, religious candidates attribute their electoral support more
to their grassroots religious base than to mass partisanship, elite party
organization, or ties to wealthy social groups. Thus, when religious
candidates get elected, they are strongly tied to religious patrons. Religious
institutions’ influence is intensified when “their” elected representatives are
themselves religious professionals.
This discussion provides many of the tools needed to address our second
puzzle: how are the culture wars shaping Brazilian democracy? Clergy-driven
politics both enhances and dulls representation, I argue. In classic theories in
political science, party leaders are supposed to help citizens understand how
issues fit together: “what goes with what” (Converse 1964). Parties may be
largely incapable of this kind of opinion leadership in Brazil, outside narrow
wedges of strong party identifiers (but see Samuels and Zucco 2014, 2018). By
acting as opinion leaders, however, clergy can help to align the views of religious
conservatives – both voters and legislators (Boas and Smith in press). At the
same time, they strengthen Brazil’s right more generally (Power and Rodrigues-
Silveira 2018). And in the context of the massive “Operation Car Wash”
corruption scandals (named after one money-laundering site) that have
unfolded across Brazil since 2014, religious middlemen have another positive
externality. Reliance on clergy as electoral intermediaries reduces candidates’
need for large campaign donations from wealthy individuals – the kinds of
transfers that feature prominently in corruption scandals, and that lead to
overrepresentation of business interests.
Introduction 7

Yet clergy-driven politics also has troublesome implications for

representation. Recall that clergy have partial influence, shaping some
attitudes but not others. By contrast, clergy are more closely aligned with the
range of views of religious politicians, as we will see in Part IV. Thus, on many
issues, religious politicians are arguably better representatives of the clergy who
constitute middlemen than of the religious citizens who voted for them.
The clergy-driven nature of Brazil’s culture wars also likely exacerbates
partisan fragmentation. When each religious leader has his (or occasionally
her) own built-in base of support, there are few incentives for coordination.
Commentators often note that evangelical elites’ alliances are “pulverized”
across a very large number of candidates and parties – ones chosen based
more on personalistic ties, than on clear ideological criteria (Dantas 2011;
Freston 1993; Lisboa 2010). As Power and Rodrigues-Silveira put it, “in
partisan terms, Pentecostals are highly diasporic” (2018). Evangelical
organizers recognize the benefits they could achieve through collective action,
especially the ability to win elections to executive office, where one needs to
assemble majority coalitions of voters. Every election cycle features many calls
for evangelical unity, and even for the creation of unified evangelical parties.
However, the “pulverization” of evangelical candidate support reflects the
“pulverization” of evangelical religious institutions. The problem is not just
that there are no incentives for resolving the evangelical collective action
problem. Rather, the nature of evangelical institutions actually creates
disincentives for coordination, since religious groups that subordinate their
own identity or “brand” to a broader evangelical collective may hurt their
long-term prospects for competitive church growth.
The clergy-driven culture wars have yet broader democratic implications.
When clergy feel that their religious group is unfairly treated, their congregants
come to perceive the political system as less legitimate. Church politics can also
erode tolerance of the political rights of atheists and gays. Yet clergy-driven
politics also helps steady Brazil’s stressed and fractured democratic regime.
Democratic competition provides religious leaders a stake in the system, and
clergy convey their high levels of support for democracy to their congregants.
They also encourage many forms of electoral, non-electoral, and civil-society
participation. As citizen confidence in democracy, the political system, and
elections has plummeted in recent years, trust in religious authorities who are
invested in the rules of the democratic game is helping to maintain the stability
of the democratic system. Over time, the declining credibility of politicians
could lead citizens to give greater credence to the political views of clergy.
This book contributes to scholarship on representation, partisanship, and
religion and politics. First, it elucidates the causes and consequences of the
culture wars by examining how such conflicts developed in an institutional
and religious context that has not been explored before, and which is very
different from those examined in previous studies. The great majority of
academic work on the culture wars has focused on the United States. Some
8 Introduction

scholars have also examined the international activism of US-based religious

conservatives, particularly in Africa (e.g., Bob 2012; Kaoma 2014). More
germane to the present study, a rich but relatively small literature traces how
parties, religious activists, and political elites shape policy debates on issues such
as abortion and homosexuality in a wide range of wealthy, highly
institutionalized democracies (Ang and Petrocik 2012; Bean 2014b; Engeli,
Green-Pedersen, and Larsen 2013; Grzymala-Busse 2015; T. A. Smith and
Tatalovich 2003; Studlar and Burns 2015). A key conclusion emerges from
this latter body of work: party institutions and political elites strongly affect the
outcomes of potential religious and cultural conflicts. When political parties,
elected officials, or high-level bureaucrats largely ignore orthodox-progressive
cleavages in the electorate, those cleavages are less likely to shape policy. By
contrast, when one or more groups of elites ally themselves with orthodox or
progressive forces, latent issue cleavages are more likely to manifest in electoral
divides or policy changes.
However, this insight provides a poor explanation of the Brazilian case, in
which political parties have largely failed to build strong linkages to religious
groups, with the important exception of the linkages between the leftist
Workers’ Party (PT, or Partido dos Trabalhadores) and the Catholic Church
in the 1980s. Furthermore, in Brazil’s secular, pluralistic policymaking context,
no faction of bureaucrats has captured the policy process to benefit a single
religious group. The clergy-driven approach I develop in this book better
explains Brazil’s recent period of political and social conflict. At the same
time, it suggests broader lessons about the mutual influence of religious and
electoral conflict. Just as parties can capture and exacerbate latent social
cleavages for electoral gain, in countries with permissive party systems such as
that of Brazil, religious groups use the tools of democratic politics to aid in
interreligious competition.
The book also has implications for the long-running debate among scholars
of comparative politics over the causes of multipartisanship. In broad strokes,
the debate has revolved around two potential explanations: one focused on the
nature and number of fundamental social cleavages (e.g., Sartori 1976), and the
other on the mechanical functioning of electoral institutions, as well as
the incentives they create for strategic behavior (e.g., Duverger 1972). I do not
assume that social cleavages automatically create parties. Nonetheless, I suggest
that when competing civil-society organizations are not simply allies to
preexisting parties, but actually coordinate candidacies, the organizations’
incentives for disunity at the level of civil society can undermine incentives to
electoral collective action. This argument thus brings together elements of both
cleavage-based and competitive-incentive-based approaches to understanding
party systems.
Finally, the book contributes to scholarship on comparative religion and
politics, synthesizing approaches in several domains. First, prior scholarship
distinguishes between “demand-side” and “supply-side” explanations of clergy
Introduction 9

behavior – that is, between explanations focusing on the social and political
circumstances stimulating doctrinal changes, and those focusing on the strategic
calculations of clergy. By contrast, I argue that explaining clergy political
activity requires us to consider the interaction between the religious supply
and demand sides; the strategic calculations of clergy respond to changes in
social and political conditions. Second, previous scholars have debated the
relative explanatory power of theologically based policy ideas and
institutional interests as incentives for clergy behavior. I argue, however, that
both ideas and institutional interests matter to clergy. Moreover, ideas shape
calculations of group interests by constraining the range of alternatives that can
be considered. Third, the richest studies of religion and politics in Latin America
have generally developed micro-level explanations of the political behavior
either of Catholic or of Protestant clergy; rarely have scholars incorporated
these two groups’ motivations and behaviors within a single study. Fully
understanding the ideological and institutional incentives clergy face,
however, requires incorporating the two groups within a single theoretical
framework.
But before we go further, let us introduce the actors who are the protagonists
of this story. What are Brazil’s major religious groups? Which citizens join
which groups? How have they taken part in Brazilian politics? Most of the
remainder of this chapter takes up these questions.

the protagonists: evangelical and catholic individuals

and groups
Winds of change. At the large, middle-class Vila Bela Methodist Church, a
visiting African preacher was giving a sermon on a Wednesday night. The
sanctuary was full. Doors were open to the street. Electric fans located high
along the walls near the ceiling kept a cool breeze circulating through the room,
and breathed some life into the colorful pendants decorating the walls of the
sanctuary in honor of the guest. At the end of the preacher’s hour-long sermon
on fighting the devil, he called all the congregants up to the front of the room for
individual blessings. Long, single-file lines snaked through the sanctuary as
ecstatic music played, pendants waved, and the visiting preacher blessed each
person. About ten people fell to the floor in shaking trances when they were
blessed. Attendants, obviously at the alert, jumped up each time a person fell to
make sure he or she was arranged comfortably and was securely out of the way
of foot traffic [CO2].2
A few weeks earlier, about thirty people had met for a prayer group in the
sanctuary of the São Ignácio Catholic Parish in a working-class neighborhood in
Juiz de Fora for two-plus hours on a Thursday night. There was no formal

2
Throughout the text, the numbered codes beginning “CO” and “FG” denote specific visits to
congregations and other field sites, as listed in Appendix A.
10 Introduction

service or preacher, though one of the participants stood up to give a long

reflection on how the Holy Spirit had changed his life. Mostly, though, the
group just sang and prayed, hands stretched upward. About an hour in, a breeze
picked up through the open doors and windows, a relief on a hot night in the
middle of a drought. And then there was a crack of thunder, and the sudden
onslaught of rain drumming on the roof, and the dusty smell of ozone refreshed
the air. As the prayer group ended, the rain let up a bit, but by the time I climbed
down the hill to my bus stop, I was thoroughly wet.
This is not your grandmother’s Methodism, and it is not your grandfather’s
Catholicism. Methodism is typically classified as a “mainline Protestant”
denomination in the United States, and the common image of both Catholic
and Methodist worship services is fairly staid. A month before the visiting
preacher’s sermon at the Vila Bela Methodist Church, though, I had asked an
affiliated Methodist pastor how he classified the congregation, whether as
“traditional Protestant” or “evangelical” or “Pentecostal.” He responded,
“most people see our church as a traditional Protestant church, but today it’s
very Pentecostal.” The pastor had come to believe that Pentecostalism was more
“biblically justified.” He pointed out that Pentecostalism had changed even
Catholicism [CO23].
So what are these groups? Throughout this book, I follow Brazilian usage in
applying the term “evangelical” (evangélico) to Brazil’s largest and most
politically important religious minority. This highly diverse set of religious
communities includes those termed historical Protestant, evangelical, and
Pentecostal. Historical Protestant denominations – often called “mainline
Protestant” in the United States – arise from the Protestant Reformation and
the subsequent fragmentation of denominations over the course of several
centuries. Examples include Lutherans, Methodists, Presbyterians, Anglicans,
Congregationalists, and Baptists.3 More recently, some Protestant
congregations have chosen to be nondenominational.
Academics also use the term “evangelical” in a narrower sense, to refer to a
subset of Protestants identified by their beliefs and behavior, rather than their
denomination. Bebbington (1989) influentially defined evangelicalism based on

3
Within historical/mainline Protestantism, large religious traditions such as Presbyterianism or
Methodism have tended to be fragmented into many denominations, each with their own
organizational identity and hierarchical decision-making structures. For instance, three of the
many Presbyterian denominations in North and South America include the Presbyterian Church
of the USA (PCUSA), the Covenant Order of Evangelical Presbyterians (which largely operates in
the United States, in competition with the PCUSA), and the Presbyterian Church of Brazil.
Constituent denominations within one tradition often vary greatly in their theological and
political stances. In some traditions, constituent denominations continue largely to maintain a
shared identity and to work together within a larger federation. In the Church of England, for
example, denominations are for the most part organized territorially (by country or group of
countries); national Anglican denominations take part in a global body known as the Anglican
Communion that has no legal existence, but maintains a unified identity and shares much of its
doctrine.
Introduction 11

Religious Affiliation in the Brazilian Census

100
99.7 98.9
95.0 93.5 92.0 90.4
87.4
80

83.3
Percentage of Population

73.7
65
60
40

22.2
20

15.4
9.0
5.1 6.5
0 2.6 3.4 4.0
1.0
0

1872 1890 1940 1950 1960 1970 1980 1991 2000 2010

Catholic Evangelical/ No Religion/ Other

Protestant Unidentified Religion

figure 1.1 Religious affiliation in Brazil

four beliefs: “conversionism” (belief in the need to have an adult affirmation of

religious commitment); “crucicentrism” (emphasis on Jesus’ death and
resurrection to save humanity); “Biblicism” (special regard for the Bible); and
“activism” (belief in the need to actively express one’s faith and convert others).
Robbins (2004) further discusses two distinctive behavioral commitments:
moral asceticism, or emphasis on conservative codes of behavior; and
continuous evangelism (“activism” in Bebbington’s scheme).4 Protestant
individuals, congregations, and denominations may or may not be
evangelical. In Brazil, the overwhelming majority of Protestant clergy, both
ones in nondenominational congregations and those affiliated with historical
Protestant traditions, adhere to these traits. They also emphasize
supernaturalist aspects of Christian faith, and envision the divine as an
agentic, interventionist presence in society (Bohn 2004; Mariz and Machado
1997; Pew Research Center 2006).
Figure 1.1 shows that the sporadic censuses of the nineteenth century
registered extremely few evangelicals. At the time Brazil gained its
independence from Portugal, Roman Catholicism was officially a religious
monopolist, in that it was the religion of state. This is not to say everyone was

4
Robbins (2004) discusses these as defining traits of Pentecostalism and charismatic Protestantism.
However, they apply to evangelicalism more broadly.
12 Introduction

Catholic in practice. Many people were nonobservant, of course. More

significantly, many practitioners of African-influenced, traditional, and
syncretic slave and former-slave religions such as Umbanda and Candomblé
told census-takers they were Catholic. Still, the census numbers tell the story of
Roman Catholicism’s political and social power.
Through most of the twentieth century, the proportion of the population
identifying as Roman Catholic slowly fell, while waves of Protestant,
evangelical, and Pentecostal missionaries gradually established footholds in
various parts of the country (Mendonça 2006; Oro 2006). In the last three
decades of the twentieth century, though, the pace of change quickened. As
global Christianity’s “center of gravity” has shifted southward in recent decades
(Jenkins 2002, 2), Pentecostalism and evangelicalism have swept across Brazil
as well as the developing world (Freston 2004; A. B. Fonseca 2008; Garrard-
Burnett 2009; Pew Research Center 2006).
Figure 1.1 depicts a striking demographic change whose beginnings
observers generally date from about 1970 – a sustained shift away from
Roman Catholicism and toward evangelical, Pentecostal, and Protestant
denominations (Levine 2009, 2012). An observer in the 1990s claimed that
10,000 Brazilians were abandoning Catholicism each day (Stockwell 1995). In
1970, ninety out of every hundred Brazilians told the census they were Roman
Catholics; forty years later, only sixty-five out of every hundred did so. In the
same time period, the number of evangelicals, Protestants, and Pentecostals
quadrupled. Thus, there were eighteen Catholics for every Protestant,
evangelical, or Pentecostal in Brazil in 1970; in 2010, there were three. While
Catholics will continue to outnumber Protestants in Brazil for many years,
simply extrapolating the trend lines suggests that the 2020 census may register
about two Catholics for every Protestant in the country.
The earliest Protestant missionaries to Brazil in the nineteenth century were
non-Pentecostal revivalists and evangelicals. Brazilian Protestantism continued to
grow slowly during two early waves of Pentecostal conversion in the early and
mid-twentieth century (Anderson 2004). Pentecostalism arose out of revivalist
movements in the United States in the early part of that century (Robbins 2004).
The term “Pentecostal” derives from a miracle described in the Book of Acts in
the Bible, which occurred shortly after Jesus’ death, on the Jewish feast day of the
Pentecost; the Holy Spirit descended on an assembly of disciples, who suddenly
began to speak “in tongues,” in a multitude of foreign languages simultaneously.
Theologically, Pentecostals emphasize the immediate presence of the Holy Spirit,
and promote spiritual “gifts” or miracles such as healing, casting out the devil,
and speaking in tongues (Chesnut 2003b; Gaskill 2002; Pew Research Center
2006; Steigenga and Cleary 2007). In practical terms, Pentecostal worship
services are very lively affairs. Beginning in the 1970s, neo-Pentecostal
denominations – a further offshoot of Pentecostalism – such as the UCKG
began to spread rapidly. In the 1970s and 1980s, neo-Pentecostalism made
theological contributions to global Christianity, two of which are particularly
Introduction 13

significant. First, “prosperity theology” holds that believers will be divinely

rewarded with material blessings, including physical and mental health as well
as financial prosperity; in poorer countries, the doctrine generally has a greater
emphasis on the supernatural (Autero 2015; B. Martin 2006; Mora 2008; Offutt
2015). This doctrine is highly controversial even within Pentecostalism, and
many older Pentecostal denominations reject it forcefully. Second, the notion of
“spiritual warfare” posits that daily struggles result from encounters with locally
contained evil spirits or demons, who can be cast out (Robbins 2004). The deities
of non-Christian religions – including, in the Brazilian case, those of Umbanda
and Candomblé – are often reframed as evil spirits.
Technically speaking, some Pentecostal individuals and congregations are
not evangelical, in the second, narrower sense just defined (Løland 2015; Pew
Research Center 2011). For practical purposes, though, we will see that
Pentecostal and evangelical/mainline Protestant clergy tend to think and act in
very similar ways. In surveys of citizens it is hard reliably to distinguish
historical Protestants/evangelicals from Pentecostals because many adherents
use these terms inconsistently or consider themselves to be a part of both
groups.
Such has been the success and cultural importance of Pentecostalism that
both non-Pentecostal Protestant groups and parts of the Catholic Church have
become “Pentecostalized” (Bom 2015; Chesnut 2003a, 2003b). The Catholic
Charismatic Renewal (CCR) constituted the Church’s most successful new
product in response to Pentecostalism (Chesnut 2003b; A. R. de Souza 2007).
Chesnut estimated that by 2000, more than half of practicing Catholics were
Charismatics (2009). The movement was responsible for almost all of the
Church’s rapidly expanded media presence, which was oriented around
popular musician priests such as the handsome Padre Marcelo Rossi
(Carranza 2006; Mariz 2006). This is the transformation to which the pastor
cited in the introduction to this section was referring. Worship services in
Pentecostalized religious traditions have become livelier, and there is a greater
emphasis on miracles and the immediacy of ecstatic, spiritual experiences.
Meanwhile, some historical Protestant groups have also adopted prosperity
theology.
Yet another reason why evangelicalism and Pentecostalism have impacts
beyond what their proportion of adherents in the population might suggest is
the relative devotion of their followers. In the 2014 AmericasBarometer survey,
56 percent of evangelicals/Pentecostals said they attended worship service more
than once a week, while just 14 percent of Catholics did so.5 Fully 80 percent of
evangelicals/Pentecostals were in church at least weekly, compared to 37

5
The AmericasBarometer is a periodic (typically biannual) series of surveys of public opinion,
democratic attitudes, and voting behavior, conducted in approximately two dozen countries of
North and South America by the Latin American Public Opinion Project (LAPOP). More
information as well as data downloads are available at www.americasbarometer.org.
14 Introduction

percent of Catholics. Based on these numbers, A. E. Smith (2018) estimates that

the average evangelical spends well over twice the number of hours a year in
church as the average Catholic.
Beyond the two major groups of Christians considered in this study, the analysis
includes an extremely diverse category of “other religions,” incorporating just
under 5 percent of the Brazilian population in 2010. This category includes two
very small non-Protestant branches of Christianity, Jehovah’s Witnesses and the
Latter-Day Saints, plus other major world religions such as Judaism, Islam, and
Buddhism. The largest proportion of “others,” though, belongs to syncretic
religions that are largely local creations, including Spiritism, Candomblé, and
Umbanda. These last two African-influenced religions developed in slave
communities and are culturally important in Brazil, yet their representation is
low, in part because many of their practitioners also claim adherence to
Christianity. With the rising political power of evangelicals, these religions have
increasingly been the target of discrimination by evangelicals (Phillips 2015).
Finally, the rate of religious nonadherence has been slowly rising in Brazil; 8
percent of respondents in the 2010 census reported having no religion. Still, the
overwhelming majority of people in this group say they believe in God, and some
even attend church regularly.
Demographers conclude that Brazilians of all ages and generations are
becoming more evangelical and less Catholic (Coutinho and Golgher 2014;
Jacob, Hees, and Waniez 2013). Still, change has been particularly accelerated
among young people – a leading indicator. In the 2007 AmericasBarometer, 65
percent of 16–25-year-olds identified as Catholic. Just seven years later, the
2014 AmericasBarometer recorded that only 51 percent of them did so. In the
year the majority of the data presented in this book were collected, Catholicism
was on the cusp of becoming a minority religion among young adults. By 2017,
45 percent of that age group identified as Catholic, as did 44 percent of 26–35-
year-olds.
Who chose an affiliation other than Catholicism? The answer to this question
will be key in the following chapters, as we seek to understand how the need to
attract and keep members affects religious leaders’ behavior. Figure 1.2 analyzes
data from the 2008, 2010, 2012, 2014, and 2017 rounds of the
AmericasBarometer. Along the side of the figure are the independent variables –
the characteristics of individuals, their places of residence, and the survey year – that
may influence which religious affiliation a person reported. All independent
variables here (and throughout the book) are coded on a scale from 0 to 1 to
make it easier to compare the relative sizes of the effects of each variable. When
the dot corresponding to each independent variable is to the right of the 0 line, that
independent variable makes it more likely that someone will report the given
religious affiliation, rather than Catholic; when the dot is to the left of the 0 line,
it makes it less likely that they will choose that affiliation over being Catholic. The
lines surrounding the dot (or whiskers) represent 95 percent confidence intervals.
Introduction 15

Protestant/Pentecostal No Religion Other Religion

Year (2008–2017)
Age Group
North Region
Northeast Region
Southeast Region
South Region
Size of Locality
Female
Afro-Descendant
Household Wealth
Primary Education
Secondary Education
Higher Education
Leftist
Rightist
Interpersonal Trust
Political Interest

–2 –1 0 1 2 –2 –1 0 1 2 –2 –1 0 1 2
Coefficients estimate each variable’s impact on the likelihood of reporting a religious affiliation, versus Catholic.

figure 1.2 Characteristics of Brazilians choosing non-Catholic religious affiliations

When the 95 percent confidence interval does not overlap the 0 line, we say that the
variable is statistically significant.
Context – geography, time, and age group – is the most important factor
shaping who is likely to choose a religious affiliation other than Catholicism:
whether evangelical, nonreligious, or “other” (a heterogeneous category largely
comprised of small, non-Christian groups). The top row of coefficients in the
figure shows that over these nine years, respondents became much more likely
to report that they were all three non-Catholic religious affiliations. Age group
is coded in categories from “16–25” to “over 65,” and then recoded to run from
0 to 1. Older people were less likely to say they were Protestant, Pentecostal,
evangelical, or nonreligious rather than Catholic; they were more likely to be
members of other religions, such as Candomblé or Spiritist. The region of the
country also strongly influences religious affiliation. And switching away from
Catholicism is a decidedly urban phenomenon – people living in larger localities
were much more likely to choose all three non-Catholic religious affiliations
than people living in small and rural areas.
The bottom variables in the figure include a series of demographic
characteristics and personal attitudes. These variables are relatively
unimportant in determining who switched away from Catholicism. Women
were a little more likely than men to choose Protestantism, Pentecostalism, and
other religions over Catholicism, while they were less likely than men to say they
did not have a religion. People who identified as Afro-descendants were slightly
more likely to identify as Protestant/Pentecostal, over Catholic. Household wealth
did not matter, but those with university educations were slightly more likely to
say they were nonreligious or were a member of some other religion. Finally, those
who switched away from Catholicism held similar general political and social
16 Introduction

attitudes as those who remained; ideological identification, trust in other people,

and political interest did not matter much for one’s choice of religion.

the drama

Political Activism among Evangelicals and Catholics under Brazil’s

Democratic Regime
Though evangelicals and Catholics have been present in Brazil for centuries, the
curtain lifts on the present story in 1986. Two events, or shocks, in the 1970s
and 1980s set the stage. First, in the late 1970s a burst of religious creativity
produced neo-Pentecostalism, and led to the establishment of denominations
such as the UCKG. This innovation forced older religious groups to adapt in
order to compete. Second, Brazil returned to democracy in 1985, following a
twenty-one-year period of rule by the military regime. While the Brazilian state
has been constitutionally secular in the sense of lacking an official religion since
the late nineteenth century, in the recent democratic period religious groups
have faced an increasingly level playing field. In Brazil’s ever freer religious
marketplace, the political positions clergy adopt and promote have been driven
by strategic competition among Catholic and evangelical congregations for
adherents and financial resources – what I term “souls and money” in the
next chapter.
Until about the time the curtain lifts on our period of study, Brazil’s
evangelicals had historically been viewed as clientelistic and apolitical
(Burdick 1993b; Chesnut 1999; Corten 1999; A. D. Fonseca 2014; Garrard-
Burnett 2009; Pierucci and Prandi 1995; Robbins 2004; Santos 2009). In the
1980s and 1990s, however, the classic evangelical saying that “Believers
[evangelicals] don’t mess with politics” gave way to a new political slogan:
“Brother votes for brother” (Freston 1993).6 Evangelicals emerged as a political
force in the November 1986 elections to the National Constituent Assembly
(held 1987–88), which wrote Brazil’s current democratic constitution (Bohn
2007; Freston 1993; Mariano and Pierucci 1992; Oro 2006). The Constituent
Assembly included eighteen Pentecostal and fifteen non-Pentecostal evangelical
representatives (out of 559 total), prominently featuring members of the oldest
and largest Pentecostal denomination, the Assembly of God. This first cohort
began for the first time to function as a “bancada evangélica” (evangelical
caucus) promoting evangelical positions (Dantas 2011). Subsequently, in the
1990s the UCKG began to develop a highly disciplined method of campaigning
within congregations, utilizing religious symbolism as well as keen electoral
strategizing to maximize electoral impact (Conrado 2001; Oro 2003a, 2003b).
The trajectory of political activism of the Roman Catholic Church in Brazil
differs markedly from that of evangelical groups. At Brazil’s transition to

6
In Portuguese, “Crente não mexe em política” and “Irmão vota em irmão.”
Introduction 17

democracy, Catholicism was known for leftist political activism. Though the
National Conference of Bishops of Brazil had initially supported the military
regime, by the mid-1970s it turned against it and served as the most important
opposition force in civil society (Serbin 2000). Not only did Catholic leaders
campaign actively for democracy, but priests inspired by liberation theology
organized “ecclesiastical base communities” (comunidades eclesiais de base, or
CEBs) to promote political consciousness and organizing among the poor
(Bruneau 1980, 1982; Gill 1994, 1998; Mainwaring 1986). Some scholars
suggest that this activism was a response to incipient competition from
Protestants (Gill 1994, 1995; Hagopian 2008). Gill argues, “To prevent
nominal Catholics from choosing competitors, the episcopacy has advocated
(or at least tolerated) innovative reforms that better serve these individuals”
(1995, 405). During this period, the Church developed ties to the PT (Keck
1992; Mainwaring 1986; Mir 2007).7
Following the transition to democracy, however, the National Conference of
Bishops of Brazil became more quiescent. Gill (1995) argues that this transition,
too, was stimulated by institutional pressure to shore up state support for the
Church’s institutional overhead expenses. Catholic leaders today avoid overtly
partisan stances. Well-enforced norms prohibit clergy from running for office
and discourage politicking during campaigns, though pastoral letters
disseminated in parishes commonly promote nonpartisan civic norms such as
turnout and informed voting. The Church has also quietly maintained positions
of political power. For instance, it has been the guiding force behind the
interfaith National Forum for Religious Education, which advocates policy
related to public religious education under the 1988 Constitution (L. A.
Cunha 2009).

Battle Lines: Issues, But Not Parties

Just a handful of issues dominate Brazil’s culture wars. Gender and sexuality are
perhaps most prominent, as an anecdote of local evangelical activism illustrates.
Participants in two different focus groups described evangelical congregations’
recent lobbying effort against the city of Juiz de Fora’s long-standing annual “Miss
Gay” pageant. In 2014, local papers announced that the event had been canceled
due to a “lack of public funds.” Yet behind the scenes, congregations and clergy
played a key role. As one participant said, “It was the union of evangelical
churches that prevented Juiz de Fora from becoming a gay city.” Another
remarked, “the whole evangelical community prayed a lot for it” [FG3, FG4].8

7
See also the unpublished 2018 working paper by Guadalupe Tuñón, “When the Church Votes
Left: How Progressive Religion Hurts Gender Equality.”
8
All English translations of Portuguese-language sources, interviews, surveys, and other commu-
nications in this book are my own.
18 Introduction

On the national stage, public-school curricula related to gender and sexuality

have been at the top of agendas. In 2011, center-leftist president Dilma Rousseff
was forced to rescind what opponents derided as a “gay kit”: a high-school
curriculum promoting tolerance of homosexuality. Six years later, the Ministry
of Education adopted a new national core curriculum, which – after much
lobbying from all sides – ultimately eliminated all references to gender and
sexuality in early drafts: both language intended to counter homophobia and
provisions for religious education on gender and sexuality (Ferreira and Mariz
2017). Clergy, legislators, and citizens also continue to debate many other issues
related to gender and sexuality: from anti-hate-speech legislation to same-sex
marriage and adoption (Dantas 2011; Vianna 2015). As one indicator of
growing intolerance, murders of LGBT individuals spiked 30 percent to an
all-time high of 445 in 2017, among the highest rates in the world (Cowie
2018; Jacobs 2016).
Abortion also motivates religious groups’ political participation. Here,
though, evangelical views have historically been more diverse than those of
Catholics. For instance, Bishop Edir Macedo, founder and head of the UCKG,
supports abortion rights. As he told his biographers, “Yes, I’m in favor of
abortion. The Bible is too” (Lemos and Tavolaro 2007). This stance may
partially explain Macedo’s support for the PT between 2002 and 2014. Yet
Macedo is an outlier. Even most clergy within the UCKG disagree with him on
this topic. Macedo’s nephew, UCKG Bishop Marcelo Crivella, affirmed his
opposition to abortion in successful campaigns for senator from the state of
Rio de Janeiro in 2002 and 2008, and then for mayor of the city of Rio in 2016.
Crivella has also asserted that botched abortions produce homosexual children
(Polêmica Paraíba 2016).9 Meanwhile, opposition to abortion is solid in the
Roman Catholic National Conference of Bishops of Brazil. At the level of
public opinion, Macedo’s stance is similarly exceptional. Fewer than one in
ten citizens – under one in twenty evangelicals and Pentecostals – think the law
should be liberalized. Moreover, though there has historically been little
difference between evangelicals and Catholics on this issue, evangelicals are
rapidly becoming more conservative.10
A third issue regularly polarizing religious groups’ discussions of politics
involves the public roles and rights of religious groups (Mariano 2011;
Ranquetat Júnior 2016). Restrictions on evangelization are a frequent irritant.
As just one example, in 2014, Federal Senator Magno Malta, a Baptist pastor,
led a successful charge against a proposed regulatory change that would have
prohibited churches from proselytizing to participants in publicly funded drug-
treatment programs (Malta 2014). Religious groups also fight over tax policy. For
instance, in 2015, evangelical legislators successfully negotiated churches’
exemption from a controversial financial tax known as the CPMF (Contribuição

9
A further example of UCKG clergy disagreement with Macedo’s stance is found in [CO42].
10
Results from the 2014 Brazilian Electoral Panel Study. See Chapter 6 for further discussion.
Introduction 19

Provisória sobre a Movimentação Financeira, or Provisional Contribution on

Financial Movements) (Neto 2015); meanwhile, secular groups question
churches’ exemption from income, social security, and sales taxes.
At the same time, non-evangelicals increasingly fear attacks by evangelicals.
For instance, the aforementioned mayor of Rio, Bishop Marcelo Crivella, has
implemented new event-scheduling rules that, critics charge, enable him to
discriminate against Afro-Brazilian religious events. An anecdote of local
evangelical activism illustrates intergroup tension. An evangelical focus group
participant recounted a story about her friend’s Methodist congregation in a
different city. A neighboring Catholic congregation began constructing a
moderately large crucifix outside their property. The image of Jesus on the
cross violated the Methodists’ religious sensibilities; the congregation
organized to get the city council to prevent the construction of the crucifix, on
the grounds that it would be a traffic hazard [FG3].
Beyond the triumvirate of abortion, homosexuality, and church–state
relations, other issues may become salient. Evolution education may be poised
to become another fault line of religious-political conflict, as creationist ideas
are increasingly making their way into public-school curricula (Oliveira and
Cook 2018). In May 2017 the Mackenzie Presbyterian University announced
that it would open a research center promoting the theory of intelligent design;
support came from the Discovery Institute, a think tank known for its anti-
evolution activism in the United States (Demartini 2017). This institute’s
opening could be the harbinger of future religious conflicts over education
policy.
Yet parties have been at best minor players in these dramas. At the mass level,
religious and party affiliation are largely uncorrelated, despite the PT’s ties to
the Catholic left in its early years. Since 2015, the older linkage between
partisanship and religious affiliation has resurfaced to a limited extent. In the
midst of the Operation Car Wash corruption scandal, the PT lost about half its
mass-level support. With the impeachment of President Dilma Rousseff of the
PT, evangelicals were somewhat quicker to jump off the PT bandwagon than
were Catholics and adherents of other religions.11 Still, the extreme weakness of
the renewed Catholic–PT association – in 2017 11 percent of Catholics but 7
percent of Protestants/evangelicals identified with the party – only underscores
the general absence of religious-partisanship linkages.

Battle Lines: The 2010 and 2014 Presidential Elections

Despite the absence of partisanship, issue positions shape electoral behavior.
The 2010 national presidential campaign mobilized both evangelical and
Catholic communities. In the final month of the first-round campaign, videos

11
Between the 2012 and 2017 AmericasBarometer waves, support for the PT dropped from 17
percent to 9 percent.
20 Introduction

surfaced of an interview in which Dilma Rousseff, then the front-runner from

the PT, appeared to support decriminalizing abortion. The information quickly
spread through evangelical and Catholic media, sermons, DVDs distributed in
houses of worship, chain emails, and YouTube videos (Lisboa 2010; Moraes
2010). Though Dilma had looked to be a shoo-in to win the first round outright,
she unexpectedly came in with less than a majority of the popular vote, while
Marina Silva, a born-again member of the Assemblies of God, had an
unexpectedly strong showing at 19 percent.12 The campaign went into a
second-round runoff against José Serra from the Party of Brazilian Social
Democracy (PSDB).
In the weeks leading up to the runoff, the Pope weighed in, directing bishops
to instruct Catholics to consider carefully candidates’ stances on abortion
(Lisboa 2010). Violating electoral law, a popular Catholic television station
played a sermon preaching against voting for Dilma (Abril.com 2010; Borges
2010). Yet other religious groups – prominent among them, the UCKG –
endorsed Dilma. In the Folha Universal, the church’s weekly newspaper,
distributed in congregations nationwide, the UCKG accused the Catholic
Church of “trying to interfere” in the elections, and of “taking part in an
aggressive and defamatory campaign against Dilma” (Folha Universal 2010,
13). The same newspaper featured a prominent pull quote from former
President Fernando Henrique Cardoso of the PSDB, implying that his co-
partisan Serra might be subject to supernatural evil forces: “Serra has some
demons in him that sometimes even he can’t control” (16).13 As clergy
endorsements poured in on both sides, Dilma met with top religious leaders,
promising not to legalize abortion (Sant’Anna 2010). She easily won the second
round.
Much of the analysis in this book is based on a case study of the October
2014 general election. Though the campaign was officially launched in July,
both politicians and civil-society groups had kicked into gear by June. Media
outlets widely described the elections as a watershed for evangelicals. At the
legislative level, 2014 saw a 47 percent rise over 2010 in the number of
congressional candidates running using titles indicative of evangelical
religious leaders, such as “pastor” (Tavares 2014). Evangelicals’ legislative
success was exemplified by the election of Eduardo Cunha, a member of the
Sara Nossa Terra denomination, as president of the Chamber of Deputies in
January 2015. At the presidential level, two evangelical members of the
Assembly of God were candidates: Marina Silva, a born-again leftist and

12
I adopt the Brazilian custom of referring to most politicians by their first names. Polling results
from “Tracking vox populi/Band/iG: Dilma tem 53% dos votos válidos,” Último Segundo,
October 2, 2010.
13
F. H. Cardoso almost certainly intended the word “demon” metaphorically, but the UCKG
authors of this newspaper article very likely intended for the word to be taken literally, given
UCKG cosmology. Thanks to Taylor Boas for sharing this newspaper, which he found in the
Columbia University archives.
Introduction 21

environmentalist, and the conservative Pastor Everaldo Dias Pereira.

Incumbent president Dilma Rousseff remained comfortably in the lead
throughout the campaign. Outside of first place, though, there was a great
deal of volatility, stimulated in part by the tragic death of Socialist candidate
Eduardo Campos in a plane crash a month and a half before the first-round
election. The evangelical Marina Silva, his running mate, quickly rocketed into
second place when the Socialist Party confirmed her as his replacement. The
candidate Aécio Neves from the center-right PSDB overtook her in the last few
days before the election, however, and Marina failed to go on to the second
round. Dilma won the second round three weeks later.
Evangelicals’ growing political presence was also felt among campaign
activists and voters. Clergy and citizens skirted the 2014 electoral rules,
which prohibited some forms of campaigning on church property (Tribunal
Superior Eleitoral 2014). The Superior Electoral Tribunal forbade campaigns’
use of loudspeakers or amplifiers within 200 meters of church property “when
[the churches are] in operation” (205). In addition, churches were covered by
rules prohibiting the “transmission of political advertising” such as signs,
flags, or banners in “goods of common use” such as churches, parks, and
public gymnasiums (206).14 Nonetheless, penalties for violation were minor
or nonexistent. Electoral law stipulated that churches caught using
loudspeakers or amplifiers would simply be made to turn them off (Tribunal
Superior Eleitoral 2012). Those caught posting illegal advertising would have
forty-eight hours to remove the advertising before facing a fine ranging
between R$2,000 and R$8,000 (about US$1,000 to US$4,000 at the then-
current exchange rate).
Nonetheless, as in other recent elections, many high-profile pastors made
endorsements in the media. The media also widely reported on campaigns
within evangelical congregations (Mali 2014). Fines were minor. Clergy
communicated their preferences in both subtle and overt ways, and
coreligionist candidates often attended services. Evangelicals highlighted the
need to elect coreligionists and to combat legislative initiatives involving gay
and transgender rights, which they saw as threatening the traditional family
(S. D. de Souza 2013, 2014). Still, evangelical influence was hampered by the
fact that evangelicals were far from united. In addition to many endorsements
of the two evangelical candidates, pastors also endorsed both Dilma Rousseff
and Aécio Neves.

Democratic Troubles: 2013 to 2018

By the 2014 election, the fabric of the Brazilian polity was showing fraying
edges. A long period of sustained economic growth had ended in the early

14
“Nos bens . . . de uso comum . . . é vedada a veiculação de propaganda de qualquer natureza,
inclusive pichação, inscrição a tinta, fixação de placas, estandartes, faixas e assemelhados.”
22 Introduction

2010s, leading to growing pessimism about the PT and dissatisfaction with the
incumbent president Dilma Rousseff. In June 2013, what Brazilians call the
“June Protests” began, as citizens nationwide took to the streets – many for
the first time in their lives – to take a stand in favor of public funding for social
services and basic infrastructure, and against large public-works projects in
preparation for the World Cup (Alonso and Mische 2016; Moseley and
Layton 2013). These protests were marked by rejection of parties and
partisanship (Alonso and Mische 2016). Then in late 2014, hints of the
Operation Car Wash corruption scandal began to appear. In this context, the
incumbent president squeaked over the finish line in first place in October 2014.
Within the first year of Dilma Rousseff’s second term, a congressional
movement for her impeachment arose. The long, political partisan process
culminated in seven-month proceedings that polarized civil society. On the
streets of most big cities, red-clad petistas (PT supporters) protesting what
they called a rightist “coup” clashed with pro-impeachment protesters decked
out in the colors of the Brazilian flag. People who made the mistake of wearing
the color red without political intent sometimes found themselves the target of
public ire. In April 2017, nine months after the completion of the impeachment,
Brazilians remained highly polarized over the legitimacy of the proceedings. In
that month, the AmericasBarometer asked Brazilians whether they thought it
had been fair. Responses were bimodal. More than half of respondents gave the
impeachment the very highest or very lowest rating: 21 percent reported a “1,”
indicating they strongly disagreed, and 32 percent reported a “7,” indicating
that they strongly agreed that the impeachment was fair. While there were small
differences between religious groups – with evangelicals on average slightly
more supportive of impeachment than Catholics, and the nonreligious slightly
less supportive – responses were bimodal within every religious group.
When President Michel Temer, Dilma’s former vice president and one of the
masterminds of the impeachment proceedings, took office in September 2016,
he was already highly unpopular. In 2017, his fiscal austerity policies and labor
reforms became the impetus for renewed protest. By June 2017, when the
AmericasBarometer went back into the field and I revisited many
congregations for follow-up interviews, Brazilians had spent two years in a
state of constant elite-level political crisis. The Operation Car Wash corruption
scandal had engulfed a large percentage of elected politicians of all parties.
Brazil was in many ways a changed country.
Brazil’s democratic troubles registered in many quantitative indicators. For
instance, the Varieties of Democracy project noted a drop in Brazil’s “Liberal
Democracy Index” from 0.78 on a 0–1 scale in 2014, to 0.57 in 2017.15 Indices
from Freedom House and the Economist have likewise registered democratic
declines since 2015. Meanwhile, public support for democracy among Brazilian
citizens has eroded even more rapidly. In 2012, 69 percent of Brazilians agreed

15
See https://www.v-dem.net/en/analysis/.
Introduction 23

with the statement that “democracy may have problems, but it is better than the
alternatives” – a level of support that had been essentially constant in 2007,
2008, and 2010. By 2014, support for democracy had dropped to 63 percent,
and by 2017 to 52 percent.16 Meanwhile, an index of citizens’ perceptions of the
legitimacy of the political system dropped from 0.50 on a 0–1 scale in 2010, to
0.34 by 2017. Levels of both partisanship overall and petismo (sympathy for the
PT) have declined in tandem with the legitimacy of democracy and the political
system.
What role have clergy, congregations, and other religious actors played?
Have they bolstered the political system? Or instead contributed to the slow,
partial erosion of Brazilian democracy? Throughout the book, we will find that
clergy have played mixed roles in this drama.

plan of the book

What has produced Brazil’s culture wars? And what are the consequences for
democracy? The next chapter develops an explanation of the culture wars’
causes and consequences that focuses on the behavior of clergy, conceived as
rational actors who maximize a combination of theological and material
institutional objectives related to financial resources and membership bases.
Chapter 3 then describes the book’s empirical strategy. Much of the evidence
comes from quantitative and qualitative studies of citizens and clergy in the city
of Juiz de Fora (state of Minas Gerais).17 In 2008/2009, I conducted five months
of dissertation fieldwork in the city during the mayoral election campaign,
which featured a front-runner lesbian university professor who was opposed
by evangelical groups. I returned to the city in July–November 2014 and June
2017 for an in-depth study of mobilization within congregations. This
fieldwork was augmented with contemporaneous surveys of clergy in Rio de
Janeiro and Fortaleza.
The book also includes evidence from nationally representative studies of
Brazilian adults by the AmericasBarometer, the Pew Forum, and the 2010 and
2014 Brazilian Electoral Panel Studies (Ames et al. 2013, 2016); a panel study of
political socialization in two cities between 2002 and 2006 (Baker, Ames, and
Renno 2006); two online survey-experimental studies that recruited samples
through Facebook (Boas and Smith 2015); and the Brazilian Legislative Surveys
of legislators’ attitudes (Power and Zucco 2012). These other data sets are

16
Data on support for democracy and the legitimacy of the political system are from the
AmericasBarometer. In the AmericasBarometer, the legitimacy of the political system is based
on responses to five questions asking about citizens’ perceptions of whether the courts, institu-
tions, and political system generally protect basic rights (see discussion in Online Appendix C).
17
This city has become a focus for a number of academic studies of social influence, analogous to
the US political-socialization studies focused on South Bend (Baker, Ames, and Renno 2006;
A. E. Smith 2016a).
24 Introduction

briefly introduced in the chapters in which they are used, and further details
presented in Appendix C.
Throughout the book, survey experiments help us get a better handle on how
congregational politicking works. A rapidly growing body of work applies
experimental methods to core questions in religion and politics.18
Experiments improve on individuals’ self-reports of what their clergy say;
they enable researchers to tease apart potential causal mechanisms for
observed correlations; and they can improve causal inference. The present
study is among the first to apply survey-experimental methods to study
religious elites (see also Calfano and Oldmixon 2016; Calfano, Michelson,
and Oldmixon 2017; Calfano, Oldmixon, and Suiter 2014). Experimental
methods are especially revealing in studying religious elites because clergy are
particularly likely to self-censor, given their leadership roles. For instance,
though very few clergy openly admitted to declining membership, priming
clergy to think about the threat of competition nonetheless affected the way
they responded to subsequent questions.
The “meat” of the book centers around six chapters describing how religious
elites, congregants, and politicians construct the culture wars and shape
Brazilian democracy. Part II focuses on clergy ideas and behavior. Chapter 4
examines what clergy think and say about policy issues, finding large gaps
between evangelical and Catholic clergy on one issue, sexuality, and smaller
gaps on a wide variety of issues. Chapter 5, then, shows that clergy widely
discuss neutral democratic norms related to political engagement. However,
they discuss candidates more sparingly.
Part III asks, how does congregational politics shape citizens’ attitudes and
behaviors? Chapter 6 finds that at the citizen level, religious gaps in policy views
are limited to gender, sexuality, and, to a growing extent, abortion. Chapter 7
then shows that churchgoers exposed to civic messages in church are more likely
to go to the polls, and that evangelical clergy’s campaigning can sway
evangelical voters. However, clergy influence is far from automatic: influence
is stronger among individuals and congregations with higher levels of doctrinal
conservatism. Chapter 8 demonstrates that religious groups have multivalent,
ambivalent impacts on congregants’ attachment to democracy.
Finally, Part IV assesses the upshot for Brazil’s representative democracy. As
shown in Chapter 9, evangelical legislators improve representation in certain
policy areas, since elected legislators as a whole have generally been
unrepresentative of Brazilian citizens on issues such as homosexuality and
abortion. However, evangelical legislators deviate from the interests of
evangelical citizens in some issue areas, promoting conservative positions on
economic policy and race that are not aligned with the evangelical base. In these

18
For a sampling of this new field, see Albertson (2011); Ben-Nun Bloom, Arikan, and
Courtemanche (2015); Boas (2014); Djupe and Calfano (2014); Glazier (2013); McClendon
and Riedl (2015); Weber and Thornton (2012).
Introduction 25

areas, the views of evangelical legislators more closely resemble those of clergy.
These policy deviations may be facilitated by personalistic ties forged between
politicians and citizens forged in church, and perhaps by clientelism.
Nonetheless, the concluding chapter argues that while Brazil’s culture wars
create polarization and push politics to the right, they may ultimately also
help to stabilize democracy by giving important civil-society groups a stake in
the Brazilian electoral game.
2

Clergy, Congregants, and Religious Politicians

In 1831, the twenty-six-year-old French nobleman Alexis de Tocqueville

obtained permission from the French crown to travel to the United States for
nine months, ostensibly to write a report on penitentiaries. Returning to France,
Tocqueville published the first major comparative case study of republican
government in the modern era. He argued that religion helped to make
democracy work. “When I arrived in the United States,” he said, “it was the
religious aspect of the country that first struck my eyes. As I prolonged my
journey, I noticed the great political consequences that flowed from [it]”
(Tocqueville 2010, 479). “Religion,” he said, “must be considered as the first
of their political institutions; for . . . it singularly facilitates their use of
[liberty]” (475).
Yet Tocqueville thought churches supported democracy best when kept out
of the electoral fray. He warned that politicized religion “must adopt maxims
that are applicable only to certain peoples . . . So, therefore, in allying itself with
a political power, religion increases its power over some and loses the hope of
reigning over all” (483–484). He contrasted the United States with the ongoing
culture war in postrevolutionary France: “Unbelievers in Europe pursue
Christians as political enemies, rather than as religious adversaries; they hate
faith as the opinion of a party much more than as a mistaken belief; and in the
priest they reject the representative of God less than the friend of power.”
Centuries later, scholars argue that the alignment between evangelicalism and
Republicanism in the United States has fostered rejection of Christianity (Djupe,
Neiheisel, and Conger 2018; Margolis 2016, 2018; Putnam and Campbell
2012), just as young people may be driven away from versions of Islam
endorsed by the state (Wainscott 2017).
In the last three decades, religious congregations in Brazil have played
a growing role in democratic politics. On the one hand, they have contributed
to what Tocqueville would call morés, or “habits of the heart,” that foster civic
participation and support for the democratic regime. On the other hand, they
have also come to create an ideological Brazilian right that did not previously

26
Clergy, Congregants, and Religious Politicians 27

exist within the party system. What leads the culture wars to develop in certain
times and places? And what are the democratic consequences when religious
groups get involved in politics? This chapter takes up these questions.

explanations of the culture wars in a cross-national

context
Brazil’s recent period of religious and political polarization partakes in a cross-
national wave. Hunter (1992) argued that the “culture wars” in the United
States were initially triggered by a backlash against the political and cultural
changes of the 1960s. In the past fifty years, the winds of globalization have
carried the seeds of culture wars across most of the globe. In the 1960s, 1970s,
and 1980s, scientific and technological advances made abortion and birth
control more accessible and effective throughout the world, while public
health gains produced drops in mortality. These shifts, combined with
changes in the structure of work, led to fertility declines, demographic
transitions, and increases in women’s labor-force participation. Scholars
argue that this social restructuring triggered changes in fundamental values,
especially among young people (Flanagan and Lee 2003; Inglehart 1990, 1997;
Inglehart and Norris 2003; Norris and Inglehart 2004). In part as a result of
these value changes, global networks of gay-rights activists and feminists arose,
spreading frames, globalizing new norms, and disseminating repertoires of
contention. In Brazil in particular, pro-gay social movements began to form in
the late 1970s and 1980s, with substantial cross-pollination from international
networks (Facchini 2010; Green 2015; Green et al. 2010). The Brazilian
movement in turn contributed to the global push for gay rights, when Brazil
wrote and championed an ultimately unsuccessful “Resolution on Human
Rights and Sexual Orientation” at the United Nations Commission on
Human Rights in 2003 (Sampaio 2016). By the 2000s, one key policy target
of gay-rights activists in many Western democracies had become the
legalization of same-sex marriage; Brazil’s Supreme Federal Tribunal legalized
same-sex marriage in two decisions in 2011 and 2013. Across contexts, what
Hunter (1992) would call “orthodox” or conservative groups have reacted
predictably against changing social structures and norms. The culture wars
have arisen – sometimes – when the seeds of the culture wars have fallen on
certain kinds of soil, where culturally important pockets of religious and social
conservatism persisted.
Still, conservative resistance to change is not sufficient to explain the
culture wars. Economic, demographic, and cultural changes have taken
place across most of the globe; pockets of conservatism are ubiquitous.
This tension has only sporadically produced extensive alliances of political
and religious actors polarized by their responses to those changes. One
illustrative comparison is between evangelicals in the United States and
28 Introduction

Canada. While evangelicals in the two countries share similar attitudes on

issues such as abortion and homosexuality, only in the United States have
those issue positions become the basis for a political movement. The forces
scholars identify as producing the culture wars in the United States but not
in Canada range from the party system and agenda-setting by political
leaders (Ang and Petrocik 2012; Schwartz and Tatalovich 2009), to interest-
group mobilization (Rayside and Wilcox 2011), to national and in-group
identity construction among congregants and clergy (Bean 2014a, 2014b;
Bean, Gonzalez, and Kaufman 2008).
What turns latent or potential cultural conflicts into culture wars? One
answer that has featured prominently in studies of developing countries has
focused on the international activism of conservatives from wealthier
countries (J. Anderson 2011; Bob 2012; Kaoma 2009, 2014; McCrudden
2015; Oliver 2012; Stroop 2016). However, an overemphasis on this
“global right” can under-appreciate the agency of domestic religious and
political actors (J. Anderson 2011; Offutt 2015). As Robbins argues,
What Stoll (1990, 327) labels the conspiracy theory view – that [Pentecostal/
Charismatic] churches are largely funded and ideologically shaped by the North
American new right – has met with little scholarly support and critics stress that
[Pentecostal/Charismatic] churches are usually run by local leaders whose own agendas
resist cooptation. (2004, 135)
The notion that Brazil’s culture wars might have been driven by international
actors is simply not plausible. Both progressive and religious groups in Brazil
are embedded in international networks, and for that matter the Catholic
Church is the quintessential transnational institution. Nonetheless, the major
actors in Brazil’s culture wars are Brazilian.
A very different explanation for why the culture wars have arisen in some
democracies but not others postulates the importance of political parties
(Bornschier 2010; Engeli, Green-Pedersen, and Larsen 2012; Studlar 2012;
Studlar, Cagossi, and Duval 2013). In Europe, for instance, policy on issues
such as homosexuality and abortion has moved more rapidly in a progressive
direction in countries with a secular–orthodox party cleavage, because
progressive parties have tended to triumph over confessional ones (Engeli,
Green-Pedersen, and Larsen 2012, 2013). This explanation echoes
comparisons of evangelicals in the United States and Canada. Scholarship on
the United States shows that intensifying polarization and the rise of the
Christian right have been driven by politicians and parties; citizens have
followed political elites’ leads (Fiorina, Abrams, and Pope 2005;
Hetherington 2001; Layman and Carsey 2002; Layman, Carsey, and
Horowitz 2006; Layman and Green 2006; Mason 2015, 2018). Today,
US citizens are polarized into two partisan camps relying on distinct values
and frames (Goren and Chapp 2017; Jacoby 2014; Koleva et al. 2012; Putnam
and Campbell 2011).
Clergy, Congregants, and Religious Politicians 29

While the party-centric approach may have great appeal for explaining the
wealthy and established democracies where it was developed, it likewise fails to
explain Brazil’s culture wars. One problem is the country’s relatively low levels
of party identification. Parties cannot socialize citizens who fail to identify with
them. Though Samuels and Zucco (2018) argue that party identification in
Brazil is stronger than often assumed, partisanship and anti-partisanship have
been largely limited to one party, the center-left PT (the Partido dos
Trabalhadores, or Workers’ Party). The PT is Brazil’s only party for which
levels of mass partisanship measured by the 2014 AmericasBarometer were
roughly proportional to the party’s electoral success in voting for the National
Congress. The ideologically diffuse Party of the Brazilian Democratic
Movement (the Partido do Movimento Democrático Brasileiro, or PMDB)
had party identification at 3.9 percent in 2014 but 11.2 percent of the vote in
the Chamber of Deputies, meaning the ratio of the party’s percentage of
chamber seats to mass-level partisanship was 2.9.1 Among parties of the right
and center-right, the equivalent ratio of overrepresentation ranged from 4.0 to
well over 100.0.2 Meanwhile, politicians from the “embarrassed right” (direita
envergonhada) have tended to disown their own parties’ ideological positions in
the post-1985 era, mobilizing voters on the basis of personalistic and
clientelistic appeals, but not partisanship (Ames 2001; Mainwaring,
Meneguello, and Power 2000; Power and Zucco 2012).
A second problem with applying the party-centric approach to Brazil is the
absence of strong alliances between party and religious leaders. Despite the
PT–Catholic alliance in the party’s early years (Keck 1992; Mainwaring 1986),
today Catholic leaders eschew public partisan positions or candidate support.
The UCKG (the Igreja Universal do Reino de Deus, or Universal Church of the
Kingdom of God) also tended to support PT presidential candidates throughout
the 2000s, until the party imploded in the Operation Car Wash corruption
scandal in 2016. Other evangelical leaders have supported a wide range of
centrist and rightist parties, based on personalistic rather than ideological
criteria (Dantas 2011; Freston 1993; Lisboa 2010). Evangelical leaders today
deliberately avoid putting all their eggs in any partisan basket, scattering
support across the political spectrum and using parties strategically to
improve their own religious groups’ political standing (Dantas 2011).
The next section begins to develop a clergy-centered explanation of Brazil’s
culture wars. The Brazilian case is not unique, however; this approach can

1
Party ideology is estimated following Power and Zucco (2012). Sympathy in the electorate is
estimated based on the 2014 AmericasBarometer.
2
The Brazilian Social Democratic Party (PSDB) claimed 2.4 percent of AmericasBarometer respon-
dents but got 11.4 percent of the vote (ratio of 4.1). No other party claimed over 1.0 percent of
2014 survey respondents. Levels of party identification for some parties are estimated based on an
“Other” category that in total received 0.5 percent support. Even if the PMDB is recoded as
rightist, no more than 8 percent of interviewees identified with a rightist party, yet such parties
received well over half of votes in the 2014 legislative election.
30 Introduction

illuminate other developing country contexts. For instance, scholars debate the
causes of anti-homosexuality movements in sub-Saharan Africa. One
prominent example is Uganda’s 2009 and 2012 bills that sought to create
a “gay death penalty.” These bills had very high domestic approval, and
ultimately were defeated by means of international human rights intervention.
A party-centered approach offers little purchase in Uganda, where the party
system is relatively new and ethnically based (Conroy-Krutz 2013; Conroy-
Krutz, Moehler, and Aguilar 2016). As Grossman notes, “One implication of
the weak party system prevalent in Africa is that . . . opposition parties
commonly lack . . . the . . . capacity necessary to stimulate new issue
dimensions. Instead, they are more likely to react to cues from incumbents or
organized social groups” (2015, 341). Many scholars have instead debated
whether the culprit is local tradition, religion imposed by colonists, or other
global rightist forces (J. Anderson 2011; Cheney 2012; Sadgrove et al. 2012).
However, this focus on international networks of conservative Christians
minimizes the agency of local groups to accept, reject, or reinterpret messages.
By contrast, the activism of Pentecostal leaders and congregations better
explains observed patterns (Bompani and Terreni Brown 2015; Grossman
2015). As Grossman argues, “the upward trend in the issue saliency of LGBTs
is closely related to . . . (1) a rapid growth of Pentecostal, Evangelical, and
related Renewalist or Spirit-filled churches (demand factor) and (2)
a democratization process leading to heightened political competition (supply
factor)” (2015, 338). That is, it is the alliance of religious with political actors
that has created Uganda’s distinctive manifestation of culture-war politics.

the clergy-driven culture wars

The clergy-centered approach involves three sets of actors who, through
interaction, construct democratic politics: clergy, congregants (sometimes
simply citizens), and politicians. I use the term clergy to refer to religious
leaders within congregations, including those who are unpaid and non-
ordained. Congregants are citizens who regularly attend a given
congregation – that is, a group of one or more members of the clergy and
congregants who meet regularly, among other things to discuss religious/
political ideas.3 Many congregations belong to denominations comprised of
tens, hundreds, or thousands of congregations sharing a hierarchical decision-
making structure through which they adopt shared doctrine and practices, as
well as a common identity.

3
Because this is a book about one country, I sometimes use terms, such as “church,” derived from
Christian traditions for the sake of simplicity and familiarity. However, I intend the approach to
apply broadly, including to non-Christian religious institutions. It should not necessarily apply to
local groups or clubs without religious content.
Clergy, Congregants, and Religious Politicians 31

Clergy take the political initiative first. In a seminal study, Verba, Schlozman,
and Brady (1995) noted the similarities between churches and labor unions as
sources of political socialization. Likewise, clergy play roles analogous to those
of union bosses in steering member opinion and in brokering between
politicians and voters. First, clergy take positions on many aspects of politics,
yet the secular norms of congregants sometimes lead clergy to self-censor.
Second, congregants adopt or resist clergy views. Third, congregants and
clergy vote for political representatives, who may or may not represent them
by sharing their views or responding to their policy demands. Representative
democracy and the culture wars emerge as higher-level properties of the system
of interactions among the three actors.
Observers of religion and politics often weigh the relative importance of
ideas and material concerns for explaining clergy behavior. Theology provides
the most obvious answers. Simply listening to religious leaders explain their
behavior, one might think their choices are straightforward responses to the
dictates of holy texts or core doctrines. However, this naïve view leads to
patently false assumptions: that religious traditions have fixed political
approaches (instead, temporal and geographic variation within groups is
tremendous); or that the political approaches of different traditions might be
inherently compatible or incompatible (instead, intergroup conflict is highly
contextually dependent) (Dowd 2015; Kuru 2009).
So why do religious leaders adopt different political ideas and practices in
different times and places? Scholars often note that theologies and practices are
socially constructed, but this does not imply that religious groups could adopt
any set of ideas. A loose analogy to Darwinian selection might help. Just as
individuals within a species may exhibit millions of possible trait mutations,
religious leaders could generate a very large variety of religious and political
ideas and practices. However, these new ideas only thrive if the individuals and
groups holding them thrive, or transmit them to new individuals and groups.
Some innovations spread more easily than others. Some innovations help
religious groups themselves survive and grow, while others inhibit survival of
the religious species.
Many scholars argue that material concerns affect which theological
innovations succeed. What “religious economy” scholars call “demand-side”
approaches emphasize social conditions, or the needs and demands of religious
adherents. As just one example, the Catholic Church’s leftist doctrine of
liberation theology may have been a response to intense poverty and
inequality in the developing world. However, this approach leads to the
obvious question of how, in a pluralistic world, religious leaders become
sensitized to varying potential constituencies whose interests are at odds:
slaveholders or slaves, landowners or peasants.
By contrast, the “supply-side” school focuses not on religious “consumers,”
but on the “opportunities and restrictions” facing religious leaders and
entrepreneurs (Finke and Iannaccone 1993, 27). Religious suppliers adapt
32 Introduction

their behavior to the opportunity structure created by the state, at times taking
oppositional stances and at others currying favor. Gill, for instance, argues that
the Catholic Church’s turn to the left in Latin America in the 1960s and 1970s
was not a response either to poverty or to theological innovation, but rather to
growing Protestant competition (1994, 1998). He also maintains that the Latin
American Church’s subsequent turn to the right in the 1980s was an attempt to
shore up state support for high overhead costs related to physical infrastructure
and training (Gill 1995).
The state-centric approach likely goes too far in the other direction, ignoring
clergy interest in addressing the societal conditions they face daily in their work.
Trejo brings the society-centered approach back in by contributing the insight
that religious demand shapes supply-side calculations (2009, 2014). That is,
interreligious competition for bodies in the pews made the Catholic Church
susceptible to constituents’ demands that clergy support pro-poor and pro-
indigenous activism – but only in places where the Church actually faced
competition.4 In this book, I conceive of clergy as rational actors who
maximize the perceived benefits and minimize the perceived costs of their
actions (e.g., Stark and Bainbridge 1996). Interreligious competition, and the
need to maintain and increase their group’s count of souls, critically shapes
clergy behavior. Thus, I coincide with Trejo in integrating the supply and
demand sides.
Yet a naïve interpretation of the supply-side approach leads to further
problems, promulgating a vision of clergy as cynical panderers who believe
that, as one critic put it, “the parishioner is always right” (Wilson Quarterly
2010). I do not assume that clergy adopt political beliefs opportunistically. Even
as clergy seek to maintain membership, they also hold sincere, yet evolving,
political and theological views that affect their incentive structures. Thus,
I concur with scholars who argue that political theology motivates clergy
action, which is still rational given complex and sometimes competing
objectives (Kuru 2009; Toft, Philpott, and Shah 2011).5 Hence, this book
weaves theological approaches together with integrated supply- and demand-
side explanations.

Clergy Objectives: Gathering Souls

The sine qua non objective of religious leaders, I assume, is to maintain or
increase their congregations’ counts of “souls” – the number of people in the
community. In practical terms, “souls” may be judged by the number of warm

4
Hale (2018) further argues that Catholic parishes were only able to take advantage of increased
episcopal support for leftist activism in places where parish governance had previously been
decentralized.
5
Also see Gwyneth H. McClendon and Rachel Beatty Riedl’s manuscript “From Pews to Politics in
Africa and Beyond.” Unpublished, 2018.
Clergy, Congregants, and Religious Politicians 33

bodies in the pews, or the number of names in a congregation’s registry.6 Some

congregations are nonviable at their current membership level and need to
grow, while others can comfortably thrive at the present level. Some
intensively evangelize to swell the pews, while others focus on “consciousness
raising” of a more narrowly circumscribed public (Lehmann 1996).
No religious leader, though, should be comfortable with sustained
membership loss. In most majority-Christian contexts, birth rates are around
replacement levels, and religious attendance is declining among youth; hence,
demographic trends will lead to decline in membership absent concerted efforts.
The need for souls inevitably constrains clergy behavior.
Secondarily, clergy need money – often but not always collected from their
own count of souls – to support a wide range of purposes: clergy salaries;
physical infrastructure; material and spiritual services; and reproduction of
the congregation’s membership. Very large and old denominations such as the
Roman Catholic Church have deeper pockets and greater capacity to share
resources across congregations, yet they also face the financial burden of
maintaining historical infrastructure, training facilities, and extensive
personnel. Within congregations and denominations, individual clergy may
divide responsibilities for gathering souls and money, but these tasks are
ubiquitous. Without souls and money, established religious institutions
disappear, while would-be entrepreneurs are unable to plant new religious
ventures.
Strategies for seeking souls and money depend on the extent of state
regulation of religious groups. We can think of state secularism – defined as
states’ neutrality toward religious groups – as having two dimensions: (lack of)
support/favoritism for any religious group over others; and (lack of) restriction
on groups’ activities (Fox 2015; Kuru 2009). With respect to the first dimension,
secular states do not, in the terms of the First Amendment of the United States
Constitution, “establish” official religions. Short of full establishment, though,
secular states often channel resources to religious groups. With respect to
the second dimension, both secular and non-secular states limit religious
groups’ and individuals’ abilities to practice their faith as they choose.
Restrictions are more severe in non-secular states and “assertively secular”
ones such as France and Turkey, which seek to keep religion out of the public
sphere (Kuru 2008). However, even “passively secular” states such as Brazil or
the United States inevitably impinge on religious groups, typically via laws that
are not religiously targeted. For example, nondiscrimination laws could affect
congregations’ abilities to exclude gays, while zoning codes and restrictions on
issuing radio licenses affect churches’ growth. These two dimensions are

6
I believe most clergy care about their followers’ relationships to the transcendent, not just about
a tally of warm bodies in the pews. However, the reason why clergy want to gather souls is
unimportant for this explanation.
34 Introduction

correlated; states seeking to control religious groups’ activities often also

channel financial resources toward one or more groups (Fox 2015).
The extent of state regulation of the religious market affects strategies for
gathering souls and money. The closer a state is to the secular end of each
dimension, the more leaders focus on congregants – religious consumers who
control their own religious choices and voluntarily fund congregations. The less
secular a state is, the more religious leaders focus their energies on the state
itself. With respect to the first dimension of secularism, religious groups can
lobby the state for their own budget line, or for laws requiring religious
education in public schools. Regarding secularism’s second dimension,
religious leaders might struggle against policies that inhibit their ability to
compete, or privileged groups might fight for restrictions on other groups.
States are never entirely absent from markets, including religious ones;
modern states shape the religious groups within them in countless ways.
As a result, even in relatively secular countries, religious groups always keep
one eye on the state as they compete among each other for souls and money. For
that matter, some religious groups actively push states to become less secular.
Still, the extent of state secularism affects religious leaders’ relative emphasis on
the state or society.
Because the Brazilian state is highly secular – as I will detail below – I focus on
how religious free markets shape leaders’ competitive strategies. In a religious free
market, religious groups face two challenges in collecting souls and money. The first
is competition. Given the ebullience of religious creativity, in places and times where
there are few barriers to forming new religious groups, the religious marketplace is
ever changing. New religious ventures rise, others fail, and as new groups form they
may succeed in creating demand for religious products that had never before
existed. The inherent unpredictability of an unregulated and competitive religious
market requires nimbleness from established religious groups. In Brazil, the sudden
rise and dramatic success of neo-Pentecostalism changed the competitive landscape
for both the Catholic Church and non-Pentecostal Protestant leaders, leading to the
“Pentecostalization” of both religious traditions – that is, the rise of Charismatic
Catholicism and Charismatic Protestantism. In addition, all religious competitors
have to work to create demand for any religious product at all, as societal secularism
grows and “none of the above” becomes an increasingly popular option. Thus, in
Brazil the culture wars have been fought on two fronts: between evangelicals and
Catholics, and between religious and secular groups. In some battles, evangelicals
and Catholics have been allies against a secular left; in others, evangelicals and
Catholics have been competitors.
In contexts where congregations must be largely self-supporting, the second
challenge is to induce congregants to fund services – that is, to gather money
from a collection of souls. Congregations seeking to raise money from their own
attendees face free-riding problems. Hypothetically speaking, enterprising
clergy could charge admission at the door; as a practical matter, it would be
simple enough to refuse entry to non-payers. However, the priority of collecting
Clergy, Congregants, and Religious Politicians 35

souls, even nonpaying ones, complicates the calculus. As Mancur Olson (1971)
noted decades ago, targeting “selective incentives” to donors can stimulate
voluntary contributions. Prosperity theology doctrines that portray tithing as
a financial investment operate as a supernatural selective incentive. More
commonly, congregations provide a wide range of services without charge –
from mass, prayer, or worship service, to religious education, to social
connection with others in the congregation – and invite attendees to “give
back” (Falk and Fischbacher 2006; Gouldner 1960; Sugden 1984). Church
attendees may be more susceptible to calls for reciprocity, as religiosity is
associated with sustained cooperation, fairness, trust, and trustworthiness
(L. R. Anderson and Mellor 2009; Barrios and Gandelman 2014; Henrich
et al. 2010; Migheli 2017; Tan and Vogel 2008). Discussions of politics and
public issues can become one component of a “brand” cementing loyalty.

Clergy Objectives: Theological and Political Beliefs

Though congregational leaders must seek souls and money, clergy also ascribe
to complex bundles of theological and political views. These include beliefs
about transcendent forces beyond the divining of human senses, but also about
how humans should relate to the transcendent and interact with each other in
the here and now. Thus, politics is inseparable from the larger package. Clergy
seek to impart these views to congregants.
Clergy views about political issues deal with the outcomes religious groups
should advocate, from restricting abortion, to investigating corruption, to
creating environmental regulations. Three bundles of issues have been
especially important in the culture wars: socioeconomic issues such as social
welfare and state intervention in the economy; family/sexual traditionalism,
encompassing abortion, women’s roles, homosexuality, and transgender rights;
and church–states relations, including a diverse set of policy demands regarding
what religious groups can and must do, from tax policy, to policies limiting how
and when churches can evangelize. Clergy also take positions on a very large
number of other issues, such as foreign policy and war, the environment, and
race relations. Because their various positions often fail to correspond to
consistent locations on a standard left–right spectrum, the issue space is
potentially highly multidimensional. For simplicity, I limit the discussion
primarily to the first three issue dimensions. Even on these issues, the range of
possible views is dizzying; for example, clergy range from full support to full
opposition to abortion.
Clergy views on the political process involve attitudes toward different
political actions, from voting, to marching on the capital, to terrorism.7 They
can also address broad regime types (for instance, support for democracy,
authoritarianism, or theocracy). Theologies of political process affect the way

7
This is what Philpott (2007, 2009) terms “political theology.”
36 Introduction

clergy express their policy views. For instance, in a state where abortion is legal,
clergy who oppose abortion might lobby autocrats or legislators behind closed
doors (Grzymala-Busse 2015); they might organize marches in the streets; or
they might do nothing. Political process ideals are often embedded within
overtly nonpolitical theologies dealing with the nature of human agency and
society.8
What shapes clergy beliefs? Perhaps the most important influences include
internationally disseminated doctrine, childhood socialization, and adult
training in what Sandal calls “epistemic communities” (2017; Serbin 2006).
Yet views also slowly evolve in response to professional circumstances. Clergy
in hierarchical denominations are influenced and constrained by the views of
their superiors and peers (Calfano, Michelson, and Oldmixon 2017; Calfano
and Oldmixon 2016). Those whose ambitions are blocked in a hierarchy may
switch religious establishments and adopt new views (Nielsen 2017).

How Clergy Negotiate Multiple Objectives

Scholars often debate the relative importance of ideas and material
organizational interests for clergy behavior (Levine 2012). I maintain that
both matter. Clergy hold some views that are relatively unbudging, whether
due to professional constraints or personal convictions. At the same time, the
dominant political culture and institutions – in society at large, and in the
congregation – constrain and filter what clergy are able to say. Clergy might
fail to talk about politics simply because they are focused on other things. Yet
even when clergy hold strong theological/political views, they rarely become
roving preachers seeking to push their beliefs on anyone with an available pair
of ears.
When clergy hold non-negotiable views that potentially conflict with their
responsibility for seeking souls and money, their behavior will be nuanced.
Sometimes they will de-emphasize those views, pivoting to prioritize other
issues. And sometimes ideas will shape how clergy define their interests – they
can affect clerics’ choices in pursuit of material organizational interests. To give
one example, a Catholic priest who endorses abortion will lose his job, no matter
how much a pro-abortion stance might appeal to the locals. A hypothetical priest
whose residents want him to endorse abortion might choose outreach strategies
ignoring his most fervently pro-choice constituencies.
Congregants are the most immediate constraint on clergy. In the absence of
self-restraint, clergy with views outside the congregational mainstream can lose
their own flocks – their ideas can be selected against in a Darwinian process.
Clergy may de-emphasize issues when they sense congregational disagreement,
whether between the clergy member and congregants, or among congregants.
In many democracies, moreover, citizens hold secular norms against clergy

8
See McClendon and Riedl, “From Pews to Politics.”
Clergy, Congregants, and Religious Politicians 37

publicly discussing political issues in general. Congregational constraints on

clergy speech may be especially common in Catholicism. As a result of Catholic
parishes’ social diversity, priests face the prospect of losing members on two
fronts: to both secularism and evangelicalism. The history and continued
demographic and cultural dominance of the Roman Catholic Church also
likely make some citizens particularly wary about Church engagement in
politics.
At the same time, political engagement can sometimes help clergy attract
members. In some places at some times, certain interpretations of doctrine catch
on quickly. Such forces are likely affected by the role of individual selection into
Catholic and evangelical congregations. Evangelical congregations may simply
attract more conservative individuals who are willing to receive or accept
certain kinds of clergy messages, and not others.
How do clergy know what their congregants want? In the cases of Mexican
mobilization that Trejo (2009, 2014) describes, civil-society actors told clergy
what they wanted. Their wide-ranging demands were largely related to
socioeconomic policies and ethnic recognition: for instance, Church support
for cooperatives, or for teaching in indigenous languages. In our case, however,
citizens do not typically make organized demands. Instead, clergy must surmise
the policy views of constituents through repeated interactions.
Beyond congregants, state policy also provides a system of opportunities and
threats that sometimes inhibit clergy political behavior and at other times
provoke it. As noted above, there are two dimensions of state secularism:
(lack of) state restriction of religious groups; and (lack of) state favoritism.
No state can be perfectly neutral all of the time, and religious groups are never
perfectly free to do anything they choose. Policies inevitably affect religious
groups differently, creating varying costs and benefits, though secular states
usually avoid policies that deliberately target specific groups. The most obvious
way states affect clergy political activity is by overtly restricting it. Non-secular
authoritarian regimes may actually prohibit some or all religious groups from
even meeting, much less getting involved in politics. But even relatively neutral,
secular democracies can limit what clergy are able to say about politics. In the
United States, Internal Revenue Service (IRS) rules governing tax exemption
prohibit congregations’ overt campaign activity, on penalty of losing their tax-
exempt status – a sizable disincentive.
The Brazilian state is, in Kuru’s (2009) terms, “passively secular.”
The Republican Constitution of 1891 legally separated the state from the
Catholic Church, and also established the free exercise of religion. Both
aspects of state secularism have been adopted in every subsequent
constitution. Since the transition to democracy in 1985, actual respect for civil
liberties such as freedom of association has risen, and today’s Protestant and
Catholic churches compete for souls on a relatively level playing field. In the
Religion and State (RAS2) data set, Fox (2015) codes Brazil as “supportively
secular” across the 1990 to 2008 period, meaning that the state “supports all
38 Introduction

religions more or less equally.”9 Brazil is also coded in that data set as having
one of seventy-one possible restrictions on minority religions, and three of fifty-
six possible religious regulations. And the Association of Religion Data
Archives (ARDA) rates Brazil as a 0.6 on a 0–10 scale on the Government
Regulation of Religion Index, and a 0.7 on a 0–10 scale for Government
Favoritism of Religion, following Grim and Finke’s (2006) coding.10
In the past decade, observers have feared secularism might be in danger.
In 2008, the Lula administration approved a treaty with the Holy See; the
Chamber of Deputies and Senate subsequently ratified the concordat in late
2009. The concordat’s sponsors portrayed it as an innocuous ratification of
policies long in place, yet civil-society groups ranging from the Association of
Brazilian Magistrates to evangelical groups warned about infringement on
secularism (L. A. Cunha 2009; Folha Online 2009a, 2009b; Schiavon 2009;
L. M. F. de Souza 2016). Nonetheless, over the ensuing years the concordat does
not appear to have substantially shifted the Catholic–evangelical balance of
power.
As Menchik notes, around the world, “religious organizations and secular
state authority have coevolved over the course of the twentieth century” (2016,
11). What Mayrl (2016) calls a country’s “secular settlement” – the legal
relationship between church and state that is fixed in the medium term –
simultaneously shapes and is shaped by clergy activism. The Brazilian state’s
high level of secularism permits most forms of religious engagement in politics,
while its imperfect neutrality motivates engagement. First, Brazilian secularism
allows religious groups great latitude to take part in politics. Toft, Philpott, and
Shah (2011) argue that religious groups become more politically active in
secular states, as they are unconstrained by the need to curry favor. Indeed,
the Superior Electoral Tribunal publishes an electoral code in advance of each
election that typically prohibits churches from engaging in many forms of
explicit, public campaigning, yet penalties are small to nonexistent.
Second, the Brazilian state’s control of resources and policy generates
church–state conflicts, motivating clergy activism. The state can provide
religious groups with funding, ranging from contracts for drug-treatment
programs to support for religious facilities. Religious education in public
schools has been another major issue of contention. Ostensibly
nondenominational religious studies was reinstated in the 1934 constitution,
and replicated in subsequent constitutions, including the current 1988
constitution, yet today both Catholics and evangelicals hope to tailor religious
education to suit their own objectives. Further, relatively minor restrictions

9
This quote is from the RAS2 codebook. See www.religionandstate.org for the RAS2 data set and
Brazil’s coding.
10
See www.religionandstate.org for the Grim and Finke/ARDA coding. Measures of social ten-
sions indicate problems: Brazil is scored as a 5 on the 0–10 scale for Social Regulation of
Religion – that is, the presence of social tensions and intolerance of religious groups.
Clergy, Congregants, and Religious Politicians 39

such as the need for building permits and radio licenses have become sources of
aggravation for evangelical and Pentecostal congregations seeking to grow
(Gaskill 2002). Clergy also see restrictions on public speech such as the use of
megaphones in public spaces, and on evangelization during publicly funded
services, as limitations on religious liberty. These opportunities, constraints,
and grievances created by the state provide religious leaders with strong
incentives for many forms of political activism, ranging from legislative
advocacy to support for in-group political candidacies.
One form of political message is likely to be universally well received by
congregants and political authorities. Since re-democratization in the 1980s,
Brazilian political culture has emphasized building citizenship among the
historically excluded (Baiocchi 2005; Barros, Bernardes, and Macedo 2015;
Morrison 2010). Political, media, and civil-society elites publicize their support
for democratic principles and participation. Religious leaders likewise see both
moral and strategic reasons for supporting the democratic order. In the context
of elections, religious and nonreligious civic education discusses the need for
a voto consciente, or a conscientious vote.11 Conscientious voting involves
showing up to the polls and following voting procedures – meaning, in the
context of electronic voting and very high multipartism, typing correctly one’s
chosen candidates’ electoral codes (up to five digits in some races) on
a computer touch screen. But it also entails making informed, non-clientelistic
choices. Embedded within Brazilian political culture themselves, clergy
intuitively adopt the language of democracy and participation. At the same
time, the noncontroversial, consensual nature of citizenship norms allows
religious groups espousing them to be seen as disinterested benefactors of
society.

how congregants respond to clergy

Brazil’s Father Marcelo, a popular television priest, once claimed “I have an
influence [on voters], whether I want to or not” (Novaes 2002, 72). Is he right?
National leaders of large denominations regularly seek to guide public opinion,
through Catholic missives and publications, as well as newspapers such as the
UCKG’s Folha Universal and the Assembly of God’s Mensageiro da Paz.
However, this book focuses on the influence of local religious elites. Is the
man or woman at the helm of the ship able to guide the crew?
Several aspects of political discussion in congregations may limit clergy
persuasiveness.12 Drawing on Zaller’s (1992) “R-A-S” (“Receive-Accept-
Sample”) model of opinion change, the first potential impediment to influence
involves the question of whether congregants receive and understand clergy

11
Voto consciente literally translates as “alert” or “aware vote.” However, the connotation in
Portuguese is closer to that of “conscientious vote” in English.
12
See McClendon and Riedl, “From Pews to Politics.”
40 Introduction

messages. Even when clergy talk about politics, congregants are not guaranteed
to recognize it. Congregants might be present during political sermons, yet
might fail to grasp the political content if they are uninterested or simply not
paying attention. Clergy may be intentionally or unintentionally vague if they
are concerned about annoying congregants. When parishioners and clergy
disagree at the outset, the need to keep bodies in the pews may discourage
clergy from voicing opinions forcefully (Novaes 2002). Thus, I expect some
decay in signals between the religious leader’s mouth and the congregant’s ears
and brain. This will especially be the case in Catholic parishes, where clergy may
be less inclined to express their views forcefully.
Moreover, even when congregants receive and recognize political messages
from clergy, they may not be influenced because they already agree with their
clergy at the outset. Citizens tend to self-select into congregations that are
politically compatible on core issues (Chesnut 2003b; Novaes 2002).
In addition, citizens and clergy within a congregation might often
independently favor the same politicians – ones identified with their religious
groups, or who share the same religious and political perspectives (Boas 2014;
Boas and Smith 2015; Campbell, Green, and Layman 2011; McDermott 2009).
Still, sometimes clergy have real opportunities for influence. When they do,
persuasion is likely to be partial – shaping some attitudes more strongly than
others. In his model of opinion change, Zaller (1992) notes that citizens are
more readily disposed to accept some kinds of elite messages than others; in
particular, views that elites consensually share are more likely to be persuasive.
As highly trusted, local civil-society elites, clergy may be quite influential when
they espouse democratic and participatory norms echoed by a wide range of
other elites.
Clergy also communicate information on candidates’ positions. For instance,
in church-based campaigns against Dilma Rousseff in Brazil in 2010, much
effort was spent on simply informing the religious public of her stances on
sexual and reproductive politics; directly partisan messages urging vote choice
in one direction or another were largely unnecessary. In such cases, again
I would expect clergy to be highly effective. Trust in clergy as local experts
and moral authorities is likely to lend credence to the presumably factual
information they disseminate. Furthermore, information may be conveyed
and reinforced not only through sermons, but indirectly, second-, third-, or
fourth-hand through congregant social networks.
Finally, clergy can seek to persuade congregants directly, imparting their
views on public-policy issues or candidates. Public religious speech can lead
congregants to consider issues and candidates in religious frames, and in terms
of sacred rhetoric (Bloom 2013; Marietta 2008; Ryan 2014). Topics related to
the dimensions of sexual and family traditionalism and church–state relations
might be expected to be especially susceptible to such framing effects (Layman
and Green 2006).
Clergy, Congregants, and Religious Politicians 41

Persuasion is likely to be not only partial, but also asymmetric – affecting

some individuals more strongly than others. I posit that two traits affect
congregants’ likelihood of accepting clergy’s political views. First, secular
democratic norms could foster resistance to opinion leadership by the
clergy. Second, congregants who are more doctrinally conservative may be
more persuadable (see, for instance, Campbell, Green, and Monson 2014).
Rink (2018) argues that theological conservatism among evangelical
congregants in Peru makes them more obedient to messages from authority –
but especially ones framed using theological, rather than secular, frames.
Doctrinal conservatives adhere to a strict interpretation of their religious
doctrine and texts, and are concerned about the divine consequences of
deviating from what they understand to be the religiously correct answer.
Such individuals are also likely to have higher levels of authoritarianism and
need for cognitive closure (Hetherington and Weiler 2009). Doctrinally
conservative individuals and denominations will tend to cohere more strongly
in their attitudes and electoral choices, and they are more likely to take cues
from clergy.
The democratic implications of partial and asymmetric influence are mixed.
Because clergy are relatively effective in transmitting consensual elite messages,
religious leaders play an important role in bolstering democratic participation
and citizens’ perceptions of the legitimacy of the political system. Such elite
messages can help sustain democracy in the context of Brazil’s sustained
political crises. In addition, if congregants take some cues, but ignore others,
gaps in opinion between religious groups will vary from one topic to the next.
This partial influence may ultimately dull polarization, as members of
competing religious groups share many religious and political beliefs.
However, clergy may also delegitimize both democracy and the political
system when they convey to congregants their own grievances against
a political system they perceive as unfair.
Asymmetric persuasion likely exacerbates polarization. If citizens who are
more committed to secular norms are less persuadable, and doctrinal
conservatives more so, conservatives will tend to follow clergy leads to
a greater extent than liberals. Ultimately, this could lead to growing gaps
between religious groups in ideology and policy views. When reinforced by
religious identity, these growing gaps could lead to affective polarization – that
is, growing disparities in the ways citizens feel toward religious and political in-
groups versus out-groups (Iyengar and Westwood 2015; Mason 2015; Webster
and Abramowitz 2017). This process could also lead to declining tolerance for
the civil liberties of religious out-groups.
Yet political influence in churches involves not only clergy persuasion, but
also congregants’ influence on each other (Bean 2014b; Djupe and Gilbert 2003,
2008, 2009; G. A. Smith 2008; Wald, Owen, and Hill 1988, 1990).
Congregational discussions of politics are common in Brazil. Before and after
mass or service, at social activities, luncheons, study groups, classes in life and
42 Introduction

occupational skills, or volunteer activities, Catholics and evangelical

congregants discuss the stuff of daily life, including, during an election
campaign, salient political events (Djupe and Gilbert 2006, 2009). Campaign
activists also target congregations; not only are members easier to reach, but
congregants can become persuasive agents themselves. Evangelicals are more
likely to be exposed to congregant political discussion in church than are
Catholics, in part simply as a function of differing rates of church attendance
(Gaskill 2002).
Discussions among congregants are likely to be influential. Congregational
studies find that religious communities tend to cohere over time in their political
views and voting behavior (for instance, Alex-Assensoh and Assensoh 2001;
Djupe and Gilbert 2006; Huckfeldt, Plutzer, and Sprague 1993; Jelen 1992;
Wald, Owen, and Hill 1988). Congregants’ influence on each other will
substantially magnify the influence of clergy. At the same time, asymmetric
influence is also likely to operate at this level – meaning that doctrinally
conservative congregations may be more likely to cohere than others.

how politicians respond to clergy and congregants

Beyond clergy and congregants, the third vertex in the representational triangle
is politicians and elected representatives. Just as I conceive of clergy as
motivated by both ideas and organizational interests, religiously based
politicians are motivated both to obtain office and to promote a policy
agenda. Some may also have goals of personal enrichment involving
corruption. In theory, candidates and elected officials with religious
constituencies should be responsive to both congregants and clergy, since they
rely on both for votes. Once religiously based politicians have shored up their
election prospects, however, their personal policy or corruption objectives do
not require responsiveness to any electoral base. In practice, the extent to which
religious politicians represent clergy and congregants depends on the actual
electoral power of each.
A few features of the relationships among clergy, congregants, and elected
representatives will tend to disadvantage congregants when their issue priorities
diverge from those of clergy. First, the role of clergy as brokers who select and
endorse in-group candidates privileges the agendas of clergy over those of
congregants. Clergy can control information, highlighting candidate positions
on core issues where they agree with congregants – often ones related to sexual
and family traditionalism or to religious liberty – while downplaying issues
where there is less agreement. This is consistent with work in comparative
politics showing how electoral brokers capture rents from the
representational arrangement (e.g., Stokes et al. 2013). Second, since clergy
are likely to be more influential within a congregation than other individuals,
candidates will have strong incentives to favor the views of clergy. Third, clergy
will typically be more capable of monitoring politicians with religious bases
Clergy, Congregants, and Religious Politicians 43

between election periods than will congregants. Fourth, while candidates could
provide clientelistic side-payments to either clergy or congregants, congregants
may often be cheaper, though this is not necessarily the case. Side-payments can
reduce the degree of substantive representation. Thus, congregants may end up
well represented by religiously based politicians on core issues related to sexual
and family traditionalism, yet poorly represented on noncore issues.
It is important to remember that Brazil’s permissive party system and system
of open-list proportional representation create the background conditions
leading politicians to seek clergy alliances. These rules often entail hundreds
of candidacies for local, state, and federal legislative elections. The political
incorporation and representation of Brazilian evangelicals bears intriguing
parallels to Dancygier’s (2017) description of British and Belgian parties’
incorporation of Muslim immigrants in urban ethnic enclaves. In all three
countries, electoral rules give politicians an incentive to cater to voters from
religious minorities embedded in tightly knit communities. Yet also, in all three,
leaders in those minority enclaves hold political preferences that are
substantially more conservative than those of other community members.
Ultimately, the pattern of incorporation pulls politics in a conservative
direction.

summary
What kinds of forces produce Brazil’s distinctive version of the culture wars?
And what are the consequences for Brazilian democracy? This chapter outlines
the argument of the book. I begin by considering two prominent explanations
for why the culture wars arise in some settings but not others: the intervention of
international actors, and the behavior of partisan elites. Neither approach
provides much purchase to explain the Brazilian case. By contrast, a clergy-
driven explanation not only better fits that case, but it might also prove to
provide us with leverage to understand other cases in the developing world,
such as Uganda.
In the clergy-driven culture wars, I argue that clergy hold sincere political
views, yet their behavior is also shaped by the need to keep members in the pews
and to gather the monetary resources to continue their work. On the one hand,
the danger of alienating congregants leads clergy to restrain themselves in
expressing their views; this restraint reduces the political influence of clergy,
and ultimately dampens political polarization among citizens. On the other
hand, the need to gather monetary resources sometimes stimulates clergy
activism targeting the state, since the Brazilian state both directly controls
resources and regulates churches’ other growth-oriented activities. As a result,
particularly in evangelical congregations, clergy select and campaign for in-
group politicians. The fact that clergy both select candidates and serve as
brokers between candidates and congregants privileges the interests of clergy
over those of congregants.
44 Introduction

In-Group
Politicians

ce
en

policy representation
lu

constraint (voting)
inf

clientelism or
Clergy

inf
lue
co nc
ns e
tra
int
(th
re
at
of Congregants
ex
it)

figure 2.1 Relationships of influence and representation within religious groups

Figure 2.1 provides a schematic representation of the relationships among

clergy, congregants, and politicians within one religious group. Solid,
unidirectional arrows represent lines of political influence – one actor’s ability
to impose his or her political and theological preferences on others, whether
through persuasion or by selecting the individuals in question. The dashed
unidirectional arrows represent constraint – the ability to influence another’s
actions undertaken in pursuit of their preferences, yet not to shape those
preferences. Because of the constraint congregants impose on clergy, clergy
have a stronger influence over the views of politicians than of congregants.
Finally, congregants and in-group politicians are linked in a standard model of
representation. Congregants vote for in-group candidates, in theory imposing
some constraint on their behavior. In turn, politicians are expected to
reciprocate through clientelism or through policy representation (i.e.,
substantive representation [Pitkin 1967]) when elected. However, the clergy’s
strong policy influence over in-group politicians will lead politicians to
represent the views of congregants more weakly when the interests of clergy
and congregants diverge.
Thus, in this clergy-centered approach, clergy can be considered as both vote
brokers and opinion leaders. On the one hand, evangelical clergy often serve as
intermediaries or brokers between politicians and voters, helping their chosen
politicians make electoral connections. Unlike many electoral brokers (for
instance, Stokes et al. 2013), though, clergy have their own sources of revenue
and authority independent of politicians, complicating the clergy–politician
principal–agent relationship. Even in typical models of the politician–broker
relationship, brokers capture some of the rents from politician–voter linkages.
In the politician–pastor–voter triad, however, the pastor’s independent sources
of power may enable him (or her) to capture a greater share of the “rents,”
Clergy, Congregants, and Religious Politicians 45

including not just monetary but also policy gains. Simultaneously, the pastor’s
power in the intermediary role can hamper representation of congregants’
policy and material interests when those interests deviate from those of the
pastor.
On the other hand, clergy often enhance representation by serving as opinion
leaders, orienting the opinions of both masses and in-group elites. In some
policy domains, religious socialization has come to substitute for the elite
opinion leadership typically associated with political parties. The limits of this
opinion leadership, of course, depend on the persuasive power of the clergy.
Such persuasive powers likely vary by policy domain, with clergy more
effectively socializing citizens in areas that are “logically constrained” by
religious doctrine (Layman and Green 2006).
3

Methods and Case Studies

What produced Brazil’s culture wars? And how have those culture wars affected
Brazilian democracy? The evidence I have mustered to answer these questions
comes predominantly from a case study of the 2014 presidential election.
The analysis draws heavily on several large, national-level studies: the
nationally representative 2014 Brazilian Electoral Panel Study, which
interviewed citizens across the country in seven waves, from June
to November; the nationally representative 2014 AmericasBarometer,
conducted in April; and a September 2014 Internet-based experimental survey
that recruited subjects using Facebook ads. The most novel data from that
election, though, come from the Churches North and South project:
a collection of studies of clergy and congregations conducted between
early August and the end of October 2014. The ten congregations where this
project’s research team conducted in-depth fieldwork were all located in the city
of Juiz de Fora, which will be described below. In addition, about 200 Catholic,
Pentecostal, and evangelical clergy were interviewed in the cities of Juiz de Fora
and Rio de Janeiro in the Southeast region, as were over 200 evangelical and
Pentecostal clergy attending a professional development conference in the city
of Fortaleza in the Northeast region of the country. These clergy and
congregational studies are each briefly described below.
The rich case study from 2014 is complemented with a wide range of data
from Juiz de Fora and the country as a whole, spanning the period from 2002 to
2017. First, I returned to Juiz de Fora in June 2017 to reinterview many of the
same clergy (and a few new ones) from the congregations where in-depth case
studies had been conducted in 2014. For further context, the empirical analysis
also incorporates data from the 2008 mayoral race in Juiz de Fora, from the
previous three presidential elections (those of 2002, 2006, and 2010), and from
a national-level experimental study conducted during the 2012 local elections.
Finally, the book draws on surveys conducted outside election cycles among
both citizens and legislators: the 2007–2017 AmericasBarometer and the
1990–2013 Brazilian Legislative Surveys.

46
Methods and Case Studies 47

juiz de fora and the 2008 local elections study

Since a sizable portion of the evidence in the book comes from fieldwork
conducted in one city, Juiz de Fora, evaluating the evidence will require some
knowledge of that city. Juiz de Fora has a little over half a million residents and
is about three hours by car inland from Rio de Janeiro, in the populous state of
Minas Gerais. This city’s thriving manufacturing base, its mid-ranked federal
university, and its strong healthcare sector make it a magnet for residents of
smaller cities within about a 75-kilometer radius. The city is slightly whiter than
the Brazilian average, and it has a relatively high Municipal Human
Development Index. However, it is typical of Brazilian cities in an important
way: residents very closely resemble the Brazilian population at large in terms of
their religious affiliation. The percentages identifying as both Catholic and
evangelical in 2000 and 2010 were within a percentage point of the
percentages in Brazil as a whole. The biggest difference between the religious
profile of Juiz de Fora and that of the country as a whole is that Juiz de Fora has
a slightly larger share of new, non-Christian religions, which tend to be geared
toward the middle class, and often incorporate elements of Christianity. These
include most prominently Spiritism, but also groups such as Santo Daime.
The city also resembles the country as a whole in following national political
trends, with the Workers’ Party (PT) winning recent presidential elections.
Ames and Rojo-Mendoza argue that the weak left in the city, as well as
a strong but less partisan Party of the Brazilian Democratic Movement
(PMDB), has historically led to low levels of partisan polarization and highly
personalistic politics.1 Juiz de Fora has been the site of other recent studies
focusing on the political impact of neighborhood and interpersonal social
context (Baker, Ames, and Renno 2006; A. E. Smith 2016a). This book and
those other recent studies follow a long tradition of in-depth case studies that
have examined the role of social context by focusing on social and political
dynamics within single cities (Berelson, Lazarsfeld, and McPhee 1954;
Huckfeldt, Plutzer, and Sprague 1993; Huckfeldt and Sprague 1995;
Lazarsfeld, Berelson, and Gaudet 1948).

The 2008 Juiz de Fora Local Election Study

Juiz de Fora’s 2008 mayoral race exemplifies many of the themes to be found
throughout the book. Six candidates competed for mayor in the October 5 city-
council and first-round mayoral elections. The 2008 mayoral front-runner,
Margarida Salomão from the PT, is a linguistics professor and former rector
of the Federal University, and is lesbian. Her second-place rival, Custódio
Mattos from the Party of Brazilian Social Democracy (PSDB), is a former

1
Barry Ames and Reynaldo T. Rojo-Mendoza, “Urban Context and Political Behavior:
Partisanship and Polarization in Two Brazilian Cities.” Unpublished paper, 2014.
48 Introduction

mayor and was the choice of most of the local political establishment. Though
Margarida took first place in the first round, Custódio managed to turn his
campaign around and won with a comfortable margin three weeks later.
Margarida’s sexuality featured prominently in the campaign, and stimulated
vehement evangelical opposition. Custódio’s campaign put out a controversial ad
on the Free Electoral Hour (Horário Eleitoral Gratuito, the federally mandated free
television time allocated to each candidate) showing the candidate eating lunch with
his family. This advertisement was widely interpreted as a jab at Margarida for her
homosexuality. She responded by putting out an ad of herself talking about her
relationship with her father and sisters, but the damage was done. Evangelical
churches also took an active role in opposing Margarida (Miranda 2008a,
2008b). The Council of Pastors of Juiz de Fora issued a letter to member churches
supporting Custódio, in part because he was “married and had children,” and
would not “damage the Christian family” (Miranda 2008a). Member pastors were
instructed to discuss the mayoral elections with parishioners. Catholic churches,
meanwhile, took no public position on the mayoral race, though a group of
Catholics had approved a list of candidates for city council.
There is every reason to believe that dynamics within both Catholic and
evangelical churches in this particular case are typical of those elsewhere in
Brazil over the past decade. The electoral mobilization of religious groups,
partially based on issues related to sexuality, has become increasingly
common in Brazil. Both the 2010 and 2014 presidential election campaigns
were marked by Catholic and Protestant mobilization related to homosexuality
and abortion; the eventual winner of both elections, Dilma Rousseff, is a female,
twice-divorced former guerrilla fighter. And politicians’ sexuality has been
prominent in other mayoral elections as well. For instance, the 2008 mayoral
elections in São Paulo featured a great deal of controversy surrounding the
sexual orientation of Gilberto Kassab, the eventual victor.
Chapter 7 incorporates survey data gathered following this election.
In November 2008, a team of students from the Federal University of Juiz de
Fora interviewed 1,089 Juiz de Fora residents about the recently concluded local
elections. Interviewers asked about religious behavior and experiences as part of
a larger questionnaire focusing on mechanisms of social influence, building on
the recently concluded 2002–2006 Two City Study. Respondents were
clustered within twenty-two neighborhoods that had been randomly sampled
for the prior study, with approximately fifty respondents per neighborhood.

the 2014–2017 churches north and south project

The Clergy Survey

In August and September of 2014, a team of researchers surveyed 425 Catholic,
evangelical, and Pentecostal clergy members in the cities of Juiz de Fora and Rio
de Janeiro, and at a conference of Protestant clergy in Fortaleza (state of Ceará,
Methods and Case Studies 49

in the Northeast region of the country). The ten- to twelve-minute interviews

dealt with a range of topics: basic congregational characteristics; frequency of
preaching on various subjects; frequency of political activities within the
congregation; and clergy’s personal views on the legitimacy of democracy and
the political system. Importantly, the survey incorporated a question-order
experiment. In Version A, a battery of four questions about church
attendance and local religious competition were included at the beginning of
the questionnaire; Version B included these items at the end. Thus, the
difference in responses between the two versions indicates the ways responses
change when clergy are induced to reflect on the membership pressures facing
their congregations.
Creating a truly random sample of clergy proved to be impossible, as the
research team did not find a sampling frame – i.e., a comprehensive list of clergy
from which one might sample. For the purposes of contacting Catholic priests,
both Rio de Janeiro and Juiz de Fora were stratified into regions based on
geography and socioeconomic status, and research assistants sought to
interview clergy in neighborhood parishes within each region. For the
purposes of contacting Protestant clergy, we relied on clergy association
membership lists and clergy contacted at clergy association meetings. While
the team sought to interview both male and female clergy from congregations of
varying socioeconomic status, geographically distributed across the two cities,
obtaining clergy interviews was difficult. The sample was opt-in and relied on
substantial networking through the clergy association officials.
A conference of Protestant clergy in the Northeast region of the country
provided an opportunity to substantially improve the geographic scope of the
sample, and to interview a large number of clergy in one place. The conference,
organized by the Apostolic Discipleship Movement (Movimento do
Discipulado Apostólico, or MDA), was a professional development seminar
on church growth through discipleship and ministry by means of “cell groups”
(referring to the rapid reproduction and multiplication of small groups).
The MDA association, which parallels similar church-growth models based
on small groups in the United States, was founded by Abe Huber,
a missionary and the president/founder of Brazil’s Church of Peace. MDA
conferences are organized across Brazil at a relatively low registration cost.
During the three-day conference, seminars focused largely on ostensibly
nonpartisan topics related to evangelical theology and church growth, though
participants prayed for elected officials, and one speaker gave a talk on social
justice, hunger, and evangelical missions.
Interviews were conducted on the second and third days of the
conference. Interviewers were instructed to randomly approach every other
clergy member in line near the food court and book-sales table. They were
given a quota of a minimum of one third female clergy and were instructed
to interview only pastors with name tags that identified them as being from
the Northeast region. However, the conference setting also imposed trade-
50 Introduction

offs. After a day of interviewing, it became clear that the face-to-face survey
interview was too long for administration within a conference (it lasted
about twelve minutes). Overnight, a reduced-length questionnaire was
developed (Version C) including only the twelve most theoretically
important survey items.
Table 3.1 presents basic characteristics of the sample, as well as numbers of
interviews conducted in each sample component. The non-traditional sample
design has obvious drawbacks, in that it is not clear to which extent the clergy
interviewed are representative of all clergy in Brazil. For that matter, we lack
a sampling frame or basis of comparison. Nonetheless, this is, to the best of my
knowledge, the first study to examine the political attitudes and behavior of
Brazilian clergy. Apart from the fact that clergy attending a professional
development conference are likely somewhat more resourceful and motivated
than average, in the context of small, upstart evangelical and Pentecostal
congregations, I am unaware of reasons why this sample would deviate from
the population of Brazilian clergy. Overall, the Version A/Version B split was
slightly imbalanced by city. To address this issue, analysis controls for the
location (i.e., includes fixed effects for city of interview).
As mentioned above, the survey included several questions about competition
and membership pressures. Directly after asking clerics about their attendance
levels, interviewers asked if their number of attendees had risen, declined, or
remained the same in the past two years. Next, they asked, “How active is your
church in conducting outreach to try to invite new people to attend? Are you
very active, somewhat active, not very active, or do you not try to recruit new
members?” Finally, they asked, “Do you worry that other churches might try to
attract members of your church?” Response options to this last question were “a
lot,” “a little,” and “not at all.” While the questions on attendance levels and
changes were included in all questionnaires, the questions on outreach efforts
and worry over competition were not asked in Version C.
Figure 3.1 presents responses to these questions. Reported membership
trends were on average quite sunny – implausibly so, given that church
attendance is largely a zero-sum game. That is, though Brazil’s population is
growing (albeit fairly slowly), the percentage identifying as nonreligious is also
slowly increasing. As a result, one congregation’s increased attendance will
usually need to be matched, at least approximately, by another congregation’s
attendance decline. However, only 7 respondents (all evangelical or
Pentecostal) reported a recent membership loss; 107 reported stability; and
276 reported growth. Given the low rate of reporting attendance loss, the
leftmost set of bars presents the proportion of clergy within each religious
tradition who report that attendance is either stable or declining. While
Catholic clergy were unwilling to report membership loss, they were much
more likely to report stagnant attendance than evangelicals and Pentecostals.
To the extent that any religious group would feel insecure about membership
despite the impressive self-reported numbers, it would be the Catholic Church.
Methods and Case Studies 51

table 3.1 Sample statistics, clergy study

Roman Entire
Catholic Evangelical Pentecostal Sample

Total interviews 71 190 164 425

Full questionnaire total 71 119 98 288
Juiz de Fora 44 34 19 97
Rio de Janeiro 27 48 27 102
Fortaleza 37 52 89
Version C questionnaire
Fortaleza 71 66 137
Proportion female 0.01* 0.18 0.24 0.17
Median clergy tenure (years 7.5 8 7 8
in congregation)
Median reported weekly 600 250 300 300
attendance

* One survey was mistakenly conducted with a female Catholic lay leader who was assisting
a priest. Rather than eliminate the data, I include her interview.

Next, the middle set of bars presents the percentage of respondents within
each religious tradition who say that outreach efforts are “very active.” While
50.5 percent of evangelicals and 49.5 percent of Pentecostals chose the topmost
response, only 20 percent of Catholics did so. In all three religious traditions,
the bulk of the remaining respondents said their congregation was “somewhat
active.” Only about 20 percent of respondents in each group reported that their
congregation was “not very” or “not at all” active. Outreach efforts are
correlated with changes in weekly attendance at .38.
Finally, the last set of bars presents the percentage of respondents in each
group who said they worried “a lot” or “a little” about membership
competition. Reported worry does not correspond neatly to the more
objective measures. Overall, reported levels of worry are relatively similar
across the three groups, running from 36 percent to 47.5 percent. Very few
Catholic and evangelical respondents reported high levels of worry – just
4.9 percent and 5 percent of each group. By contrast, 21.6 percent of
Pentecostals reported high levels of worry. Worry is actually very high in
those congregations with the most active outreach and with rising attendance.
For instance, 81.5 percent of those who reported high levels of worry also said
attendance was rising, while only 47.4 percent of who worried “a little” and
63.7 percent of those who worried “not at all” said attendance was rising.
Perhaps Catholic clergy failed to report declining attendance because the
Catholics who converted to evangelicalism had never attended mass to begin
with. If Catholicism were losing adherents “in name only,” we would find that
church attenders constituted a growing proportion of those identifying as
52 Introduction

80
Percentage of Clergy Reporting
60
40
20
0

Membership Outreach Very Strong Worried about

Stagnant/Dropping Competition

Catholic Evangelical Pentecostal

figure 3.1 Membership changes, outreach, and competition (clergy reports)

Catholic, leading to rising measured levels of church attendance. However, the

2007–2017 waves of the AmericasBarometer show no change in average levels
of church attendance, with the exception of a jump upward between 2007 and
2008 that may be the result of changing sampling strategies.
What are the characteristics of congregations with high propensity to grow –
that is, ones with rising attendance and strong outreach efforts? Not
surprisingly, Pentecostal and evangelical congregations are much more likely
to have these two characteristics. More interestingly, growth-oriented
congregations are more likely to have female clergy and to be predominantly
upper-middle class. However, large congregations are just as likely to be
growth-oriented as small ones, and the location is statistically insignificant
after taking religious tradition into account.

The Congregational Study

Quantitative and qualitative fieldwork was also conducted in various
congregations in Juiz de Fora in 2014 and again in 2017 to understand what
attendees thought about religion and politics, and how influence works (or not)
within congregations. The qualitative fieldwork had three components. First,
between late July and early November 2014, I engaged in participant
observation in evangelical and Catholic congregations throughout the city.
The analysis is based on field notes – ranging from summaries of worship
Methods and Case Studies 53

services, sermons, and Sunday school messages, to descriptions of

congregations’ physical properties – from eleven evangelical and Pentecostal
congregations and five Catholic congregations or Charismatic prayer
communities.
Second, I also conducted many informal, open-ended qualitative interviews
with clergy in both 2014 and 2017. In the 2017 reinterviews, I sought to
understand how congregations had dealt with the series of political and
economic crises that the country faced since the 2014 election: the
impeachment of the president who had been elected in that contest; the
Operation Car Wash corruption scandal; the implication of President Michel
Temer in that scandal; and the economic recession.
Third, in October 2014, between the first- and second-round elections, the
research team conducted seven focus groups: two at Catholic sites
(a congregation and a prayer community); four at evangelical and
Pentecostal sites; and one in a classroom at the Federal University of Juiz de
Fora, with youth identifying as nonreligious who had been recruited through
community advertisements. One of the evangelical/Pentecostal groups was
comprised not of congregants, but of Pentecostal clergy from small,
affiliated congregations. The agenda for the focus groups involved giving
each participant a handout listing ten short fictitious scenarios about
congregational politicking in the invented city of “Bela Vista” (see Appendix
B). As participants discussed these vignettes, they revealed both their norms
and their actual experiences. For focus groups conducted in congregations, we
provided refreshments but no compensation; I paid the participants in the
nonreligious focus group a nominal fee for their attendance. I analyzed
annotated transcripts of each focus group, developed by one of the research
assistants who was present at all groups. Appendix A lists the church
observation and focus groups, with anonymized names and brief
descriptions of the respective congregations.
In addition, the research team conducted 850 quantitative interviews: 433
with congregants at nine religious sites, and 417 with people at nearby
community sites. One of the religious sites was a very small Charismatic
Catholic prayer community with fewer than twenty members, of whom we
were able to interview seven. Due to the small number of respondents from that
community, those interviews are excluded from most of the analysis, though the
results are discussed in the text of Chapter 6. The eight remaining congregations
represent varying traditions within Catholicism, evangelicalism, and
Pentecostalism, as well as neighborhoods of different socioeconomic levels.
The quantitative interview sample includes three Catholic parishes – one each
in upper-class, working-class, and low-income neighborhoods – and
a Charismatic Catholic community catering to the working class. In addition,
it includes two traditional evangelical and two Pentecostal congregations of
upper-middle-class and working-class profiles. Where possible, Catholic and
evangelical/Pentecostal congregations were selected in the same neighborhoods,
54 Introduction

though several were located downtown, drawing participants from across the
city.2 Quantitative exit interviews were conducted with approximately fifty
attendees at worship services in each church. Interviewers used gender quotas
and were told to approach every second person exiting the church. In addition,
to assess the extent to which socialization in churches imposes constraint
beyond that found in the broader population, interviews were also conducted
at five neighborhood sites near the churches: four public health clinics and, to
capture upper-income citizens who use private health providers, a shopping
mall catering to the upper-middle class.
The congregational data have two major limitations in terms of their
representativeness. First, a sample of this size cannot be assumed to be
representative of the entire population of congregations across the country, or
even within one city. Still, though random sampling of a large sample of
congregations would have been ideal, the congregations studied were
deliberately selected to represent a diverse array of religious approaches,
congregation sizes, socioeconomic levels, and – based on conversation with
local academic experts in the field – levels of political activism. That is, the
necessarily limited sample of cases was implicitly stratified based on
socioeconomic status and religious affiliation, and was selected for greater
variance on both the independent and dependent variables (Geddes 2003;
King, Keohane, and Verba 1994).
Second, people interviewed in church tend to be the kinds of people who go
to church. People encountered in focus groups in church tend to be the kinds of
people who are willing to stick around after service, or even come back on
a weeknight at the invitation of a friendly American researcher. Levels of church
attendance are high in Brazil in general, but we must assume that those levels are
somewhat overreported due to social desirability bias. However, respondents in
the congregational survey have given concrete behavioral evidence of their
willingness to attend. As a result of this second limitation, we cannot assume
that the congregational data are representative of the broader population, even
simply the population within each religious group in the city of Juiz de Fora.
To address this limitation, I use representative national-level surveys, as well as
a representative local-level study of the city of Juiz de Fora in 2008, to draw
more general inferences. I also use a pair of national-level but non-
representative Internet-based studies in which respondents were recruited
through Facebook advertisements to examine mechanisms of influence
experimentally. By contrast, the congregational data serve two purposes.
First, they enable us to examine diversity within congregations as well as
differences across congregations within a denomination. Second, the
quantitative congregational data can be merged with interviews of clergy

2
To be precise, the neighborhoods of Centro, Morro da Glória, and São Mateus, as defined by the
IBGE (Brazilian Institute of Geography and Statistics).
Methods and Case Studies 55

from seven of the eight congregations studied to examine the interplay between
clergy and their congregants.
The congregations referenced in the quantitative analysis are the following.
Note that congregation names are anonymized, and not linked to the
anonymized names used in the qualitative observations reported in Appendix
A and throughout the text, for the purpose of maintaining respondents’
confidentiality.
• “Catholic Site 1”: A parish in a low-income neighborhood, with a weekly
attendance of about 1,800–2,000.
• “Catholic Site 2”: A parish in an upper-middle-class neighborhood with
a weekly attendance of about 1,000.
• “Catholic Site 3”: A parish in a working-class neighborhood with a single
priest who serves two parishes, and a weekly attendance of about 150.
• “Catholic Site 4”: A Charismatic Catholic community with an estimated
weekly attendance over 1,000.
• “Evangelical Site 1”: A large, middle-class evangelical congregation belong-
ing to a historical Protestant denomination, with a weekly attendance of
between 2,000 and 2,500.
• “Evangelical Site 2”: A small, low-income storefront evangelical congrega-
tion belonging to a historical Protestant denomination, with a weekly atten-
dance of under 100.
• “Pentecostal Site 1”: A mixed-income Pentecostal congregation with
a weekly attendance of 350–400.
• “Pentecostal Site 2”: A low-income neo-Pentecostal congregation with
a weekly attendance of about 2,000.
Figure 3.2 further introduces the sites of the quantitative fieldwork by
presenting a few basic religious beliefs in the religious and community sites.
The top half of the figure presents views regarding the nature of the Bible:
whether it is “an ancient book recorded by men,” “the inspired word of
God,” or “the real word of God to be taken literally.” Levels of biblical
literalism varied dramatically across the religious sites, from 38.8 percent at
one evangelical Protestant site to 66.7 percent at a Pentecostal site. While it is
hard to detect patterns in this small sample of congregations, levels of biblical
literalism look to be highest at the Pentecostal sites and lowest in the middle-
class Catholic congregations. Indeed, in a multivariate analysis of both
congregational and community site interviews at the individual level, biblical
literalism is lower among upper-income respondents and higher among
evangelical and Pentecostal respondents. Negligible proportions of
respondents at any religious site said the Bible was an ancient book recorded
by men. Across the four evangelical sites, only one respondent chose this
answer, while six respondents at Catholic sites did so. By contrast, about one
in ten respondents at community sites chose this option.
56 Introduction

100 4.1 2.5 4.9 10.8

Interpretation of Bible

33.3
80

38.8 40.0
39.0 51.2
(Percentage)

55.7 56.3
60.2
47.4
60
40

66.7
57.1 56.1 60.0
48.8
20

41.8 43.8 38.8 41.8

0

Cath 1 Cath 2 Cath 3 Cath 4 Prot 1 Prot 2 Pent 1 Pent 2 Community

Sites

God’s Divinely Ancient

Literal Word Inspired Book
1
Agreement (0–1 Scale)
.75
.5
.25

Cath 1 Cath 2 Cath 3 Cath 4 Prot 1 Prot 2 Pent 1 Pent 2 Community

Sites

Feel God’s Love Fear God’s Wrath

figure 3.2 Core religious beliefs at the sites of the quantitative congregational study

The survey also asked respondents the extent to which they agreed or
disagreed on a five-point scale (that is, a scale running from “strongly
disagree” to “strongly agree”) with the statements “I frequently feel God’s
love” and “I often fear God’s wrath.” Belief in a punitive or wrathful divine
presence may affect behavior in a very different way from belief in a divinity that
is loving but not wrathful (for instance, Shariff and Rhemtulla 2012). Responses
to each of these questions are converted to a scale running from 0 to 1, so a score
of 0 corresponds to “strongly disagree,” 0.25 to “disagree,” 0.5 to “neither
agree nor disagree,” 0.75 to “agree,” and 1 to “strongly agree.” The bottom
half of Figure 3.2 presents the average level of agreement with these two
statements at each site. If a site average is statistically significantly above 0.5,
Methods and Case Studies 57

as indicated by the whiskers corresponding to confidence intervals, this

indicates that the average person at that site agreed (at least slightly) with the
statement in question. Average levels of agreement with the statement that
“I frequently feel God’s love” are extremely high across the board, at every
site. In fact, even among nonreligious respondents, the average level of
agreement with this statement was .91. Sites vary a great deal more in the
extent to which participants said they feared God’s wrath. At all four
Catholic sites, the average is statistically indistinguishable from 0.5, or
neutral. By contrast, in three of the four evangelical/Pentecostal sites, the
average is significantly above the neutral point.

summary
This chapter has begun to introduce the empirical studies that will occupy most
of the rest of this book. Much of the data in this book comes from a series of
studies conducted in the city of Juiz de Fora, a medium-sized city in the
Southeast region that is broadly representative of key religious and political
trends. The Juiz de Fora studies include a neighborhood-based survey of the
2008 local election; a congregation-based study of the 2014 presidential
campaign; and 2017 reinterviews of clergy interviewed in 2014. After
describing these studies, this chapter presented quantitative data on the
sample characteristics, setting up the future analysis.
Beyond the studies described in this chapter, the analysis will incorporate
many other national-level studies, which will be introduced in the appropriate
chapters. Variable coding and sampling strategies are also described in further
detail in the Appendix.
part ii

WHAT CLERGY THINK, SAY, AND DO

4

What Clergy Think and Say: Religious Teachings and

Political Views

What do Catholic and evangelical clergy tell their flocks about contemporary
policy issues and Brazilian democracy? Diving into the data from the 2014
clergy study, this chapter focuses on clergy teachings with respect to policy
issues such as homosexuality and abortion, and on their views of democracy
and the fairness of the political system. Chapter 5 then investigates clergy speech
with respect to elections and electoral participation.
Not surprisingly, clergy differ in their views. The study reveals very large
differences between Catholics and evangelicals in the priority given to one single
issue, homosexuality. As we will find in Chapters 6 and 9, the acceptability of
homosexuality and same-sex marriage constitutes the single policy issue tying
the views of evangelical citizens to those of their in-group elites at the pulpit and
in Congress. Beyond this one issue, Catholic and evangelical clergy also differ,
though less markedly and unwaveringly, in their positions on a broader group
of doctrinally conservative religious teachings and policy issues. Among the
policy issues on which Catholic and evangelical clergy differ are race, the
environment, and economic policy.
Second, in contrast to the religious gap in policy views, Catholic and
evangelical clergy nearly universally endorse democracy as a regime type.
The solid support for democracy on the part of clergy is particularly striking
when contrasted with the ambivalence among citizens. At the same time, clergy
profess greatly varying levels of internal (intra-congregational) and external
(societal) tolerance for diversity of opinion. While democratic attitudes are of
interest in themselves, I will show that they also shape other political behavior
of clergy, with profound downstream consequences.
Third, there are large gaps between religious groups in terms of their
grievances against the current Brazilian political system. While Catholic clergy
overwhelmingly believe the state is neutral toward their group, a substantial
minority of evangelical clergy believes that the country’s president and laws are
biased against them. Perceived fair treatment matters a great deal; clergy who

61
62 What Clergy Think, Say, and Do

think the system does not treat their group fairly tend to perceive the political
system in general as less legitimate.
What motivates clergy speech on political issues? The political teachings of
clergy, I argue in this and the next chapter, are driven by three factors. First are
democratic attitudes. In Chapter 8, we will find that Brazilian clergy’s firmly
rooted, nearly universally shared democratic commitments fortify democracy,
as pastors and priests transmit their attitudes to congregants. Nonetheless, this
chapter shows that political intolerance aggravates some clergy’s perceptions
that the political system is unfair – particularly evangelical clergy. Moreover,
intolerance also depresses support for political participation (see Chapter 5).
Second, institutional considerations matter. A novel survey experiment that
I conducted reveals that the objective of gathering souls affects what clergy say.
Likely because they actually experience dwindling attendance in real life,
Catholic clergy are uniquely sensitive to survey-experimental reminders that
they face membership pressures. Priests become less conservative about matters
of sexual morality and the family when reminded of the threat of losing
members. Evangelical clergy, by contrast, persist in talking about conservative
positions even when reminded of competitive pressures. However, as we will see
in the next chapter, institutional pressures shape both evangelical and Catholic
clergy advocacy of political participation.
Third, clergy policy views are closely tied to their theological perspectives.
Evangelicals’ conservatism is linked to deeply held, long-standing traditions of
doctrinal conservatism – traditions that are also correlated with lower internal
and external tolerance. This bundle of theological positions also motivates
clergy grievances against the political system, and, as we will see in the next
chapter, it drives politicking from the pulpit.

three scenes

1 In August 2014, about seventy clergy gather for the monthly meeting of the
Juiz de Fora Council of Pastors at the Comunidade Manancial, an evangelical
church in the Bairro Progresso. The worship space, like that of many evangelical
churches, is an unadorned, warehouse-like hall with white-painted cinderblock
walls and high ceilings, an oblong rectangle with a stage at the end farthest from the
street entrance. Two and a half hours into the monthly meeting, after we had
breakfasted and then prayed and sung together, the keynote speaker got up on
stage: Pastor Osésa Rodrigues, president of a very small national evangelical party,
the Liberal Christian Party.1 Pastor Osésa happened to be visiting from the Federal
District. In his hour-long talk, Pastor Osésa listed the threats to evangelical
churches today, threats both political and social. The LGBT movement loomed
large as a villain. The movement, he said, wanted to include “new genders” in

1
This spelling of Pastor Osésa’s name is based on media reports.
Religious Teachings and Political Views 63

public-school curriculums. They were sending gay couples to evangelical churches

in an effort to take the churches to court for discrimination when the churches
rejected the couples. The LGBT movement wanted to end the idea of God in Brazil
and to eliminate “Mother’s Day” and “Father’s Day” in favor of
“Caretaker’s Day.” Pastor Osésa exhorted the assembled clergy to work with
Parent-Teacher Councils to fight the LGBT menace [CO3; similar discussions of
the LGBT threat and “Caretaker’s Day” in CO1 and CO33].
2 A couple of months later, my research assistants and I sat in a small,
one-room apartment in a poor neighborhood and talked with several mem-
bers of the Disciples of Love Catholic prayer community. José Luiz, the
group’s leader and founder, explained that, “People on the margins of
society are our brothers.” Describing his own rough past, he told us that,
“Just from a [psychological] complex, I excluded myself . . . People call the
child of God a vagrant, a prostitute, even though no one reaches out their
hand to help that person.” Moving on to talk about legislation related to the
LGBT movement, José Luiz opined that “as far as I understand, and what
I understand from Jesus Christ, that doesn’t matter. If I were there [in
government], I wouldn’t either sign it or block it.” The community accepts
gays: “Our rule for life, our statute – we include the excluded.” Still, he
said, homosexuality is a sin – a lifestyle that should be abandoned – but the
church “can’t be moralistic” [FG2].
3 In mid-September 2014, the Comunidade Resgate, a Charismatic Catholic
religious community, brought Gloria Polo to speak to an audience of several
thousand people at the community’s rural facilities on the outskirts of Juiz de
Fora. Ms. Polo is a Colombian orthodontist who was struck by lightning in
1995 and had a near-death experience during which she says she saw both
Heaven and Hell, and was forced to confront her sins. After returning to life,
she had a religious awakening, and became world-famous for preaching
Catholic repentance. As Ms. Polo talked to us that dusty Saturday under
a large open-air awning, the menace of the Devil overshadowed us.
The Devil, she said, has his grip on youths who play video games, get tattoos,
and are disrespectful to their parents. It is the Devil who encourages young
people to believe condoms will protect against pregnancy. The Devil promotes
this lie because he needs the blood sacrifice of aborted babies – a sacrifice
attended by the abortion doctors who serve as demonic priests. “Governments
are lying demons” when they say abortion is a woman’s right, she told us
[CO12].

teachings: religious and policy issues

What do clergy say to their flocks about policy issues, ranging from same-sex
marriage and abortion to care for the poor? This study focuses not on what
clergy say they believe, but rather on how frequently they discuss various
64 What Clergy Think, Say, and Do

topics.2 In the first two examples above, Pastor Osésa and José Luiz, the founder
of the Catholic prayer community, actually agreed on the sinfulness of
homosexuality. What distinguished their two very different congregational
approaches was not so much beliefs as the way they weighted the importance
of homosexuality relative to other issues. For José Luiz, the need for inclusion
outweighed opposition to homosexuality. Despite basic and important
theological distinctions, Catholics and evangelicals in Brazil often agree with
each other on many doctrinal and policy-related questions: that God loves
everyone, for example; that homosexuality and abortion are sins; and that it
is important to care for the poor. However, they discuss these topics with
greatly varying frequency. Clergy reports of the frequency with which they
discuss different topics give a better sense of what their congregants are more
or less likely to hear within church walls.

Religious Teachings
Before addressing policy issues, let us examine how clergy explain their faith to
congregants. Western Christian religious traditions vary greatly in perspectives
on core theological issues: God’s love and wrath; the question of whether faith
brings material rewards in the here and now; the need for evangelism to spread
the faith; and biblical prophecies regarding the End Times (that is, the
prophesied end of the current era and return of Jesus Christ). Surveys revealed
that three of these topics were quite common across religious traditions. First,
Catholic and evangelical clergy all reported that discussions of “God’s love and
pardon” are ubiquitous. Equally as important as God’s love, Pentecostal and
evangelical clergy also emphasized the need for evangelism (“the importance of
helping non-believers find God”) and to avoid sin. These latter two topics are
somewhat less common in Catholic congregations than in evangelical and
Pentecostal ones, but Catholic clergy still said they discuss these topics
frequently.
By contrast, three other issues are a source of differences within and across
religious traditions. “God’s wrath over the sinful people” was the topic on
which Catholic clergy differed the most from their Protestant colleagues.
While Catholic religious leaders on average “rarely” discuss this issue, it is on
average “frequent” in evangelical congregations, and somewhere between
occasional and frequent in Pentecostal ones. The End Times are also an
infrequent topic of discussion in Catholic congregations, but somewhere
between occasionally and frequently discussed by evangelicals and Pentecostals.
Finally, there are muted differences across religious traditions in emphasis on
the idea that “God will reward the faithful with prosperity and good health” –
a doctrine often called “prosperity theology.” However, the small inter-

2
Questions on topics of preaching were not asked in the third version of the questionnaire (see
Chapter 3 for a discussion of Version C).
Religious Teachings and Political Views 65

.25
Clergy Priorities in Discussion
0
–.25
–.5

Catholic Evangelical Pentecostal

Homosexuality Abortion Chastity Hard Work/

as Sin Responsibility

Environment Traditional Racism Ministry to Poor

Family

figure 4.1 Priorities in policy-related teachings, by religious tradition

tradition differences mask great disagreement within traditions – especially

Pentecostalism and evangelicalism. Prosperity theology is the issue on which
Brazilian clergy most seriously disagree not just in emphasis or nuance, but in
terms of very basic principles. Though the approach has its roots in neo-
Pentecostalism, not only most Catholic clergy but also many evangelicals and
even some Pentecostals vehemently disagree with it. As Sister Enilda, an adult
Sunday-school teacher in an Assembly of God (Pentecostal) congregation, told
her assembled pupils, “Our prosperity is knowing that God loves us . . . It’s an
error to believe that prosperity has to do with things or riches. We shouldn’t put
our heart inside things” [CO21].

Policy-Related Teachings
And what did clergy say about policy issues? Not surprisingly, some topics are
emphasized more in Catholic congregations, and others in evangelical ones.
Figure 4.1 assesses the reported priority given to preaching on various topics, by
the clergy member’s denomination. To produce the figure, responses to each
item were first recoded to run from 0 to 1, where 0 corresponds to discussing an
issue “very rarely,” and 1 corresponds to discussing it “very frequently.” Issues
were then ranked in terms of their order of priority in their religious tradition,
where 0 represents the average frequency of discussing all issues within that
tradition. Issues significantly above the 0 line are discussed more often than
66 What Clergy Think, Say, and Do

average, and issues significantly below the 0 line are discussed less frequently
than the average issue.
Figure 4.1 presents the eight potential topics of preaching in their order of
priority for Catholic clergy. Two survey topics relate to socioeconomic issues.
Clergy preaching on “the need for hard work and responsibility” might
encourage opposition to social assistance. By contrast, preaching on “ministry
to the poor” might encourage support for social assistance – especially in Brazil,
where the distinction between states and churches as providers of charity has
not been politicized in the way it has in the United States. Four other questions
relate to sexual and family traditionalism. Clergy who preach more frequently
on chastity and the traditional family, or on homosexuality or abortion as sins,
might push their congregants to the right on this second dimension. Finally,
clergy were asked about preaching on two other issues, “the need to care for the
environment” and “combating racism.” More frequent preaching on these two
issues would encourage left-leaning stances on these topics, though not
necessarily on the other dimensions.
In general, Catholic clergy appear to prioritize preaching on issues that
would push their congregants to the left. At the top end of the scale, Catholic
leaders say they give greatest priority to ministry to the poor, with combating
racism in second place. Those priorities are nearly inverted among evangelicals
and Pentecostals, for whom combating racism is – together with protecting the
environment – among the least important topics of discussion.
On the socioeconomic dimension, Catholic clergy give very high priority to
ministry to the poor, but only moderate emphasis to the need for hard work and
responsibility. By contrast, evangelical and Pentecostal clergy say they give
approximately equal emphasis to these two topics. Still, the differences
between Catholics and evangelicals/Pentecostals on socioeconomic concerns
are not dramatic. Many evangelical clergy strongly emphasize the importance
of ministry to the poor. For instance, one pastor affiliated with the progressive
global evangelical movement known as the “Integral Mission” gave a sermon in
which he repeatedly returned to a one-sentence refrain, Jesus’ command to his
disciples to feed an assembled crowd: “You give them something to eat”
[CO11].3
Figure 4.1 also addresses four topics related to family/sexual
traditionalism. Even highly conservative clergy may exhibit reticence about
discussing sensitive, personal issues with congregants. As Pastora Denise,
who runs a small storefront Pentecostal church on the border between
a lower-class and middle-class neighborhood, remarked, “These are
choices people make. If a person wants to talk with us about choices,
we’re going to talk about God’s will and we’re going to say that those

3
Nonetheless, the Integral Mission is viewed with suspicion by other evangelicals and Pentecostals.
One pastor reported that many of his colleagues perceived the movement as overly theologically
liberal, and insufficiently committed to a strict interpretation of the Bible [CO23].
Religious Teachings and Political Views 67

choices have consequences. But if a person doesn’t want to discuss their

sexual choices with us, we’re not going to force the conversation all the
time” [CO22]. And another Pentecostal pastor did not think he needed to
talk much about homosexuality or abortion because everyone knew they
were forbidden, though he made an exception for educating the church’s
youth [CO33]. Nonetheless, evangelical and Pentecostal clergy do generally
talk about three of the four items related to family and sexual traditionalism
more frequently than do their Catholic counterparts.
The one issue on which there are no inter-tradition gaps in frequency of
preaching is abortion. Though the Catholic speaker Gloria Polo’s anti-abortion
talk stood out for its fire-and-brimstone drama, I encountered little variance
among Catholic clergy in opposition to abortion. In fact, abortion is an issue on
which some Pentecostals take more liberal stances than do Catholics or
evangelicals. As discussed in the first chapter, Bishop Edir Macedo, the
founder of the Universal Church of the Kingdom of God (UCKG), is a rare
example of an openly pro-choice religious leader. While Bishop Macedo’s
stance on abortion has apparently not prevented the denomination from
growing, his stance is not shared across his denomination. Prominent UCKG
pastors who run for office tend to take pro-life stances. Still, in the (admittedly
extremely small) sample of four UCKG pastors interviewed in the clergy survey,
the mean frequency of discussing abortion is 0.25 on the 0-to-1 scale,
corresponding to “rarely” discussing abortion. This is statistically
significantly below the mean score of .68 (corresponding to slightly less than
“frequently”) among non-UCKG Pentecostal clergy.
Inter-tradition differences on chastity and the importance of the traditional
family are muted. For evangelicals and Pentecostals, the traditional family is
listed as the single most important priority in preaching. Nonetheless, this topic
still ranks third out of eight among Catholics. Chastity is also discussed
somewhat more frequently in the two Protestant groups than among
Catholics, though differences are not statistically significant.
One single issue strongly differentiates Catholic clergy from their evangelical
counterparts: homosexuality. Catholic clergy discuss “the sin of
homosexuality” far less frequently than they discuss any other issue.
On average, Catholic clergy say they talk about the “sin of homosexuality”
a little more often than “rarely.” As Father Wilson, the priest of a small parish in
a lower middle-class neighborhood, told me, they never talk about
homosexuality because “We’re all children of God” [CO28]. Evangelical and
Pentecostal clergy, by contrast, give homosexuality about the same level of
priority as abortion.
Finally, Catholic and evangelical/Pentecostal clergy differ substantially in the
extent to which they talk with their flocks about “taking care of the
environment” and “combating racism.” Catholic priests mention racism as
their second most important priority in preaching, after ministry to the poor.
Environmental concerns take fourth place. By contrast, concern for racism and
68 What Clergy Think, Say, and Do

the environment rank last and second-to-last for both evangelicals and
Pentecostals.

What Shapes the Policy Priorities and Religious Teachings of Clergy?

Why do Catholic, evangelical, and Protestant clergy advocate for different policy
priorities? The objective of gathering souls – that is, congregations’ needs to
maintain or increase attendance within Brazil’s highly competitive religious free
market – is one major force shaping religious groups’ engagement in politics,
and by extension the Brazilian culture wars. At the same time, variations in
doctrinal traditions, historical distributions of resources, and demographic bases
lead Catholic religious leaders to respond differently to membership pressures
than their evangelical and Pentecostal counterparts.
To test the effect of membership pressures, I used a survey experiment. As we
saw in the previous chapter, the clergy survey asked religious leaders four
questions about the extent to which they felt membership pressures:
– On average, how many total people attend mass/service here each week?
– In the last two years, has the number of people attending risen, fallen, or
continued the same?
– How much effort does your church put into outreach to get new people to
attend? Is your outreach very active, somewhat active, not very active, or not
at all active?
– Are you worried that other churches would attempt to attract some of your
members? A lot, a little, or not at all?
In the treatment version of the questionnaire, “membership-threat”
questions were asked at the beginning of the survey, right before questions
about topics of preaching and political activism. In the control version, by
contrast, the treatment was administered at the very end of the questionnaire.
Differences in responses between the treatment and control group, then, can
indicate how clergy priorities change when they are reminded of the threat of
membership loss.
Figure 4.2 assesses the impact of the treatment on topics of preaching.
Evangelicals and Pentecostals are grouped together, given the similarities of
their responses. When a bar and its confidence interval are below 0, this
indicates that the experimental treatment significantly reduces preaching
about the topic. Though the number of Catholics interviewed is low, the
treatment has a large and statistically significant impact within this group,
substantially reducing emphasis on preaching related to supernatural
intervention in the material world (prosperity theology and the End Times);
and to strict behavioral codes (chastity, abortion, and God’s wrath). However,
the treatment is not associated with a change in emphasis on topics such as
God’s love or ministry to the poor – where clergy in the control condition group
are near the ceiling. Interestingly, it is also unassociated with a changing
Religious Teachings and Political Views 69

God’s Love

Ministry to Poor

Prosperity Theology

Evangelism

Hard Work/Responsibility

Chastity

Homosexuality

Abortion

God’s Wrath

End Times

–.4 –.2 0 .2 .4

Catholic Evangelical/Pentecostal

figure 4.2 Competitive threat affects core and policy-related teachings – but only
among Catholics
70 What Clergy Think, Say, and Do

emphasis on homosexuality, suggesting that the reluctance of Catholic clergy to

talk about this issue is driven by forces other than simple concern about losing
members.
In other words, being reminded of the threat of membership loss leads
Catholic clergy to distinguish themselves from evangelicals and Pentecostals.4
Why do Catholic priests not seek to emulate their most successful competitors?
Anthony Downs, a classic political theorist, famously posited that competitors
such as political parties – or, in this case, religious groups – will tend to move
closer together ideologically, in order to capture each other’s market (Downs
1957). Indeed, scholars widely perceive the Charismatic Catholic movement as
a Catholic attempt to compete with Pentecostalism by adopting popular
worship styles (Chesnut 2009; A. R. de Souza 2007). So why are Catholics
and evangelicals not moving closer together in their political positions?
We must recall that Catholic clergy are fighting battles on two fronts. Not
only must they contend with evangelicalism and Pentecostalism, but they must
also seek to keep members in church at all, particularly youth members.
As noted in the introduction, in April of 2014, about six months before this
clergy study was conducted, the LAPOP AmericasBarometer showed that
Catholicism was on the cusp of becoming a minority religion among 16-to-25-
year-olds in Brazil. While some of these young people were leaving Catholicism
for evangelicalism and Pentecostalism, many of them were instead identifying as
having no religion at all. Self-identification as atheist or agnostic remains rare
and stigmatized in Brazil, even among youth (Exame 2013) [FG7]; more young
people, though, now report that they have no religion, but believe in God and
pray. Catholic clergy may move to distinguish themselves from evangelicals and
Pentecostals, and reduce their emphasis on fire-and-brimstone topics, as an
overture to this group of “spiritual but not religious” youth, and more
generally to maintain Catholic youth in the pews.
Many Catholics perceive that evangelical and Pentecostal congregations do
a better job at the “warm and fuzzy” aspect of congregational life – that
Protestant congregations are warmer and more welcoming. One Catholic focus-
group participant remarked that, “in evangelical churches, people who come in
are treated well, everyone hugs.” She thought that Catholic congregations should
imitate this warm welcome so that “Catholics who are lost would come back”
[FG1]. And indeed, in the congregation study members of evangelical and
Pentecostal congregations reported having more friends, and a higher
proportion of their friends, at church. Catholic priests who are worried about
attracting new members may have a hard time inducing long-time attendees
suddenly to start hugging new people. However, reducing emphasis on certain
religious teachings may be another way to make their churches more inviting.

4
In fact, if we examine only respondents in the control condition group, Catholic responses in
Figure 4.1 are less distinct from their Protestant counterparts.
Religious Teachings and Political Views 71

Yet another explanation may also be at play. In the experiment presented

above, Catholic clergy failed to budge on homosexuality, the issue that most
distinguishes them from evangelicals. Catholic resistance to discussing
homosexuality may become a constraint that closes off certain channels –
namely, an attempt to reclaim members lost to evangelical and Pentecostal
congregations. Why would Catholic clergy be so resistant to changing their
positions on homosexuality? Their resistance is likely in part ideological. Still,
another factor may be at work. While it is necessarily very difficult to get solid data
on this topic, wide-ranging sources suggest that levels of homosexuality among
Catholic clergy in Brazil and worldwide are very high relative to the population at
large (Bacarji 2016; Cozzens 2000; Flock 2013; Gandini 2015; Wolf 1989). Same-
sex-attracted clergy members may be deeply uncomfortable discussing this topic
with congregants, and simply avoid it altogether. Moreover, among non-
homosexual clergy, regular interaction with colleagues known to be gay might
increase tolerance and reinforce the position that “we’re all children of God”
(Allport 1979).
Though membership pressures affect what clergy tell their flocks,
competition does not tell the whole story. Religious and policy-related
teachings are clearly based on more than just an attempt to appeal to the
median religious consumer. Moreover, membership threat has no statistically
significant impact on the behavior of evangelical clergy, with the minor
exception of preaching on chastity.
The most obvious determinant of policy-related teachings is core religious
doctrine. For instance, perhaps a strong belief in the immediacy of the End Times,
as well as a heightened fear of God’s wrath, could lead some clergy to focus more
on matters of personal morality, such as teaching about the importance of
chastity. To test this, I attempted to create two factor variables. The first was
related to three conservative “core” religious teachings: God’s wrath, the End
Times, and the importance of avoiding sin. The second dealt with the four
components of sexual-family traditionalism discussed above: the traditional
family, chastity, abortion, and homosexuality. However, these two factor
variables were so closely correlated with each other – even after controlling for
evangelical/Pentecostal religious traditions – that the unavoidable conclusion was
that these seven topics of teaching really stem from a single latent dimension.
Thus, instead I created one factor variable from these seven variables which
I called conservative teachings, or simply conservatism.
Competition for members and conservative core religious teachings
constitute the two strongest determinants of policy-related teachings. Beyond
these two variables, few characteristics of congregations and clergy shape what
clergy talk about. Nonetheless, after controlling for a host of factors, “dummy”
variables for religious traditions remain statistically significant. That is,
differences between religious traditions have not been fully explained by the
variables measured in the study. Relative to Catholics, evangelical and
72 What Clergy Think, Say, and Do

Pentecostal clergy still report higher levels of conservative religious teachings,

and lower levels of preaching on ministry to the poor.

clergy and the political process: democratic attitudes

and tolerance
The clergy survey also explored what religious leaders thought about how
people should work together to make political decisions – that is, leaders’
ideas about democracy. At the very end of the survey, clergy were asked to
what extent they agreed or disagreed that “Democracy may have problems, but
it’s the best form of government.”5 Support for democracy in the abstract was
very high across the board. In each religious tradition, the average clergy
member “agreed” or “strongly agreed” that democracy is the best system of
government, despite any problems it might have. In total, 86 percent of Catholic
priests, 88 percent of evangelical pastors, and 91 percent of Pentecostal ones
agreed or strongly agreed with the statement; only 10 percent of Catholics and
6 percent of evangelicals and Pentecostals disagreed or strongly disagreed. This
support for democracy is particularly striking when clergy are contrasted with
the general population. In the AmericasBarometer survey of the general
Brazilian population in the same year, 63 percent of citizens in all religious
groups agreed (weakly or strongly) that democracy in the abstract was the best
system of government.
The solid support for democracy among clergy is also signaled by the finding
that this attitude has few statistically significant determinants. Democratic
legitimacy was universally high in all groups, regardless of congregational
characteristics or religious traditions. In a country where the most recent
democratic period is three decades old, this is a sign that the democratic
regime – at least in the abstract – is quite consolidated in the attitudes of local
civil-society leaders. This is an important win for Brazilian democracy, and
suggests that clergy may serve as spokespeople who help to bolster the regime in
troubled times.
The project also explored clergy attitudes toward democracy as enacted
within their congregations and more immediate personal worlds. In an era of
increasing polarization, do clergy preach tolerance of fellow citizens who
disagree? Survey questions asked about internal tolerance – tolerance for
divergent opinions on important matters within the congregation – and
external tolerance – tolerance for debate and disagreement within society in
general. To measure internal tolerance, interviewers asked clergy to what extent
they agreed or disagreed, using the same five-point scale, that “churches such as
yours” should “give everyone the right to express opinions about the direction
of the church.” To measure external tolerance, clergy reported how frequently

5
Responses were on a five-point scale, ranging from “Strongly disagree” to “Strongly agree.”
Religious Teachings and Political Views 73

or infrequently they talked to congregants about “the importance of respecting

different points of view,” using a five-point scale running from “very rarely” to
“very frequently.”6
Tolerance for divergence in opinions varied across religious traditions.
The average Catholic priest agreed that everyone in his parish should have the
right to have an opinion about the church, and said that he talked about
respecting different points of view somewhere between “frequently” and
“very frequently.” This level of tolerance – particularly of internal tolerance –
might seem surprising, since the Roman Catholic Church is a very large,
hierarchical, and nondemocratic organization. Nonetheless, the responses
signal how priests on the ground feel about discussion within their parishes.
Responses differed substantially in Pentecostal and evangelical
congregations. The average Protestant religious leader gave a response
somewhere between neutrality and weak agreement that everyone in the
congregation should have the right to an opinion about the direction of the
church. External tolerance was also lower: Protestant clergy said they talked
about the need to respect other points of view a bit more often than
“occasionally.” After controlling for a host of factors, religious tradition
remains the strongest determinant of internal and external tolerance.
In addition, internal tolerance is lower in large congregations, likely because it
is simply less practical to take everyone’s opinions into account in congregations
with many members. Membership threat did not affect political-process
attitudes among either Catholics or Protestants.

clergy and the political process: perceived neutrality

or bias of the political system
As I observed evangelical congregations during the 2014 campaign, I often heard
leaders talk about being threatened – not by other religious groups, but by an
alliance between the PT, the political system, and what they called “the gay
movement.” The threat was palpable in Pastor Osésa’s sermon to the Council of
Pastors, discussed in the introduction to this chapter. A few months later, sitting in
the little storefront Hope Church of the Nazarene in a low-income neighborhood of
the city, Pastor Eduardo told those of us listening that “The gay movement – I’m
certain it’s paid by the PT. For a long time, they’ve been wanting to implant the gay
doctrine in the schools . . .. The gays are two percent of the nation – it’s a small
percentage, but they’re doing enormous damage to the country, to the whole world.
It’s a plague” [FG5]. A week earlier, I had observed a Sunday-school lesson taught
by Pastor Eduardo’s wife, Sister Lídia. Sister Lídia told the fifteen of us assembled
that morning that the drought the region was experiencing at the moment “has an
explanation. God is the Lord. God will send rain when it’s the right moment.” It was

6
The question on internal tolerance was asked of all respondents, while the questions on external
tolerance and the legitimacy of democracy were asked in Version C of the questionnaire.
74 What Clergy Think, Say, and Do

up to all of us to pray for rain, and to teach our children to pray – but she also
implied that political leaders had extra responsibility. Referring to the PT, she
argued that, “One party in the country said that it was the salvation of the
nations, but it’s not like that . . . I said years ago, if God sends a drought to Brazil,
we’ll see if [former president] Lula is going to be able to make it rain.” She compared
the PT’s spiritual perfidy to that of the Islamic State: “The Islamic religion doesn’t
preach violence. It’s the Islamic State, falsely acting in the name of Islam. And in
Brazil, too, the LGBT movement is doing this in the name of the people” [CO34].
But the threat extended beyond the “gay movement.” In a large, middle-class
Baptist church across the city, for instance, congregants reported that the pastor
preached that “today in Congress, they’re considering a lot of laws that could hurt
the preaching of the gospel and our freedom of expression” [FG4].
As a final approach to understanding the political beliefs and attitudes of
clergy, I examine their perceptions of the way the political system treats
their groups. The clergy survey contained two questions about religious
groups’ relationships to politicians and the political system. The first
asked, “Talking about your religious group’s mission and its position in
society, do you think this country’s laws (1) favor your group, (2) hurt your
group, or (3) neither favor it nor hurt it?” Immediately following the clergy
member’s response, interviewers asked, “And the current president [Dilma
Rousseff], does she (1) favor your group, (2) hurt your group, or (3) neither
favor it nor hurt it?”
Figure 4.3 presents responses to these questions. Because answers to the two
questions are highly correlated, I created three index variables to summarize the
average extent to which each cleric believes that the country’s laws and current
president favor their group, threaten their group, or are neutral. A score of 1 on
a dimension would mean that the clergy member chose that response for both
questions, while a score of 0.5 indicates that the clergy member chose the given
response for one question, and 0 indicates they chose that response for neither
question. A priest who, for instance, says that the country’s laws are neutral but the
president is prejudiced against his group would get a score of 0.5 for “perceived
state harm,” 0.5 for “perceived state neutrality,” and 0 for “perceived state
favoritism.”
Figure 4.3 shows dramatic differences between Catholics, and
evangelicals and Pentecostals. Catholics perceive a much higher level of
state neutrality and a much lower level of state harm toward their group.
While the average level of perceived harm is .33 for evangelicals and .37 for
Pentecostals, it is only .03 for Catholics. By contrast, the average level of
perceived state neutrality is .81 for Catholics, but .47 and .48 for
evangelicals and Pentecostals, respectively. In fact, in the entire survey,
only three Catholic priests thought that either the country’s laws or the
current president hurt their group, while 73 percent thought the state was
neutral on both dimensions. Interestingly, levels of perceived favoritism are
similar across all three groups, between .16 and .21.
Religious Teachings and Political Views 75

1
.75
.5
.25
0

Perceived State Harm Perceived State Neutrality Perceived State Favoritism

Catholic Evangelical Pentecostal

figure 4.3 Perceived state neutrality or bias

What Shapes Perceived State Neutrality or Bias?

I argue that three sets of variables matter when clergy evaluate whether the
state favors or harms their group: religious teachings and policy views;
democratic attitudes; and religious traditions. First, as the many examples
of clergy speech related to homosexuality suggest, religious teachings and
policy views affect grievances because state policy actually affects whether
the policies clergy want are enacted. In contemporary Brazilian politics,
both leftists and rightists could have reason to feel aggrieved. Public
policy and leftist political parties have moved leftward on matters related
to gender and sexuality in recent years. Not only did the Supreme Federal
Tribunal and National Counsel of Justice legalize same-sex marriage in
a pair of decisions in 2011 and 2013, but leftist parties increasingly push
what some conservative groups call “gender ideology” – the notion that
gender roles are socialized and mutable. These changes are likely to
particularly aggrieve those with doctrinally conservative views. Yet leftists
would also have reason to feel disappointed with the state. Though the
Workers’ Party (PT) had created many new anti-poverty programs, in
2014 Brazil remained a middle-income and highly unequal country.
Moreover, de facto discrimination against gays remained widespread.
Second, democratic attitudes help to determine what clergy believe counts
as “fair” treatment; clergy who are less committed to democratic norms
76 What Clergy Think, Say, and Do

Perceived Hurt Perceived Neutrality Perceived Favoritism

Conservatism
God’s Love
Ministry to Poor
Legitimacy of Democracy
External Tolerance
Internal Tolerance
Growth-Oriented
Weekly Attendance
Clergy Years in Congregation
Clergy Female
Lower Class Congregation
Evangelical
Pentecostal
Juiz de Fora
Rio de Janeiro

–5 0 5 10 –4 –2 0 2 –4 –2 0 2 4

figure 4.4 Determinants of perceived neutrality or bias of the political system

may be more likely to want positive treatment than those who are highly
committed to maintaining state secularism. Third, religious traditions also
matter because the histories of groups’ relationships to the Brazilian state
vary dramatically. Despite the current religious free market, Catholicism
still has structural advantages. For instance, the Church has material assets
such as real estate that give it a leg up in competition, and some of the most
prominent symbols of the Brazilian nation are Catholic. As an evangelical
focus-group participant declared, “Our Lady of Aparecida [the patron saint
of Brazil] is theirs – she’s not mine!” [FG4].
Figure 4.4 presents determinants of grievances, perceived neutrality, and
favoritism; each of these dependent variables is on a three-point scale (0,
0.5, and 1). Once again, the dots and whiskers correspond to the estimated
impact of each of the independent variables (shown on the left-hand side of
the figure) on the dependent variables. The results show, first, that political
and religious beliefs matter. Clergy who say they preach more frequently
on conservative religious views, on God’s love, and on ministry to the poor
report higher levels of perceived state harm and lower levels of perceived
state neutrality, though not all of the coefficients are statistically
significantly different from zero. Second, more tolerant clergy, and those
who more strongly support democracy, are more likely to perceive the state
as neutral.7 Moreover, I find an interactive effect: clergy who are highly

7
Could believing the state favors or hurts one’s group affect the legitimacy of democracy, rather
than the reverse? It certainly seems plausible that causality works in the opposite direction.
I developed several versions of simultaneous equations models examining the mutual influence
Religious Teachings and Political Views 77

tolerant but less doctrinally conservative are also most likely to believe the
state is neutral. Third, after taking into account policy and process views,
religious traditions still matter: evangelicals and Pentecostals have higher
levels of grievances, and lower levels of perceived state neutrality.8
Determinants of perceived state favoritism are less consistent. Those who
preach more frequently on ministry to the poor are more likely to perceive the
state as favoring their group, but the other two topics of preaching are not
significantly associated with perceived favoritism. Support for democracy is
associated with higher, and external tolerance with lower, levels of perceived
favoritism. And finally, religious traditions do not differ significantly in terms of
this issue, after taking into account these other variables.

What Are the Consequences of Perceived State Neutrality or Bias?

Political grievances often motivate groups to engage in collective action; recent
work also identifies perceived state neutrality as a key factor affecting how
religious groups get involved in politics (Gurr 1970; Toft, Philpott, and Shah
2011; Toft and Zhukov 2015; but see Basedau et al. 2017). As we will see in the
next chapter, these variables matter a great deal for the Brazilian culture wars.
Clergy who think the state is biased against their group say very different things
to their congregants about political participation than do those who think the
state is neutral toward it.
First, though, perceptions of state neutrality/bias and legitimacy are also
strongly related. At the end of the survey, clergy were asked the extent to
which they agreed or disagreed on a five-point scale that “The current
Brazilian political system is fair and deserves respect.”9 The gap between
religious groups in their perceived legitimacy of the political system is almost
as large as the gaps in grievances and perceived state neutrality. Rescaling
responses to run from 0 to 1, the mean level of legitimacy among Catholics is
.67, while it is only .39 and .41 for evangelicals and Pentecostals. Perceived state
neutrality is strongly correlated with legitimacy: clergy members who perceive
the state as neutral toward their group are much more likely to say that the
political system is legitimate in general. Interestingly, perceived state favoritism
toward one’s group also lowers one’s perceived legitimacy of the political

of democratic legitimacy and grievances. In all of the models, democratic legitimacy strongly
affected perceptions of the state’s neutrality or bias, while the state’s perceived neutrality or bias
had no impact on democratic legitimacy.
8
Membership pressures have no significant impact on perceived hurt/state neutrality. In several
alternative versions of the model not shown here, the membership-threat treatment has no impact
on perceived state bias or neutrality. Likewise, none of the three variables related to attendance
and membership – change in attendance, extent of outreach efforts, or weekly attendance –
significantly affects perceived state harm or neutrality. The one exception is that it appears that
clergy from larger congregations are less likely to believe the state favors their group.
9
This question was not asked in Version C of the questionnaire.
78 What Clergy Think, Say, and Do

system. There are no consistent differences in levels of legitimacy between those

who think the state hurts their group and those who think the state helps it.
Perceived state neutrality is so important, in fact, that once we take it into
account, there are no statistically significant differences between Catholics and
Protestants in the legitimacy of the political system.

summary
This chapter examined what clergy think about, and tell their congregants
about policy issues, the political process, and the neutrality of the political
system. Clergy differ first of all in terms of the policy issues they prioritize.
Evangelical and Pentecostal religious leaders talk with their congregants much
more frequently than do Catholics about a set of religious and policy-related
issues that I call, as a group, conservative religious teachings, such as God’s
wrath, the need to avoid sin, the need for chastity, and the “sin of
homosexuality.” When Catholic religious leaders are led to think about the
threat of losing members, they further de-emphasize these topics. By contrast,
Catholic leaders talk more frequently than do evangelicals and Pentecostals
about a series of left-leaning issues, including ministry to the poor, racism, and
the environment. While membership pressures do have some impact on what
Catholic clergy talk about, competition and the fear of membership loss do not
tell the whole story, either for Catholics or Pentecostals and evangelicals.
Instead, clerics’ choice of what to talk about is driven in large part by core
religious beliefs and religious traditions.
Second, Catholics also differ from Pentecostals and evangelicals in their
attitudes toward the political process – that is, toward how political decisions
should be made. While democracy is universally held in high regard in the
abstract, Catholic leaders are more supportive than evangelicals and
Pentecostals of diversity in opinions, both within their congregations and in
society at large. Again, the lower levels of tolerance in evangelical and
Pentecostal congregations appear to be due to basic differences between
religious traditions, rather than to differences in reactions to membership
pressures.
Third, Pentecostals and evangelicals perceive much greater levels of state bias
toward their groups than do Catholics. Three sets of factors affect perceived
state bias or neutrality: (a) policy attitudes matter – those who spend more time
talking with their congregants about either conservative policy and religious
views or ministry to the poor are less satisfied with the state’s relationship to
their religious group; (b) attitudes toward the political process also shape these
perceptions – those who are more committed to democracy and more externally
tolerant are likely to perceive the state as less biased toward their religious
group; and (c) even after accounting for policy views and attitudes toward the
political process, clergy in the Pentecostal and evangelical religious traditions
perceive greater state bias toward their religious groups. This may well be
Religious Teachings and Political Views 79

a result of the Roman Catholic Church’s real historical advantages in the

Brazilian context – advantages that range from ownership of real estate to
identification with prominent symbols of the Brazilian nation. In the next
chapter, we will find that the perceived bias or neutrality of the political
system affects what Pentecostal, evangelical, and Catholic clergy say to their
congregants about political participation.
5

What Clergy Do: Encouraging Partisan and Electoral

Politics

When and why do Brazilian clergy encourage congregants to get involved in

politics? This chapter begins by examining what religious leaders say about
different forms of participation. We discover that clergy in all religious
traditions strongly support non-electoral engagement. For instance, they say
that churches should promote social movements and advocate for legislation in
line with their values. However, clergy speech about elections is more
controversial. Religious leaders talk to congregants about elections in three
ways: (a) through ostensibly nonpartisan encouragement to go to the polls;
(b) by advocating that congregants “carefully consider” candidates, without
overt candidate endorsement; and (c) by explicitly supporting candidates for
office. The great majority of clergy speech about elections is of the first
and second varieties, as many evangelical, Pentecostal, and particularly
Catholic clergy are wary of overt candidate endorsements.
What leads clergy to promote some forms of political participation, but not
others? I argue that clergy support for – or resistance to – political participation
is motivated by several factors. First, membership pressures sometimes lead
clergy to seek political allies that can help them compete against rival groups;
yet membership pressures likely more often muffle clergy speech, as clergy are
afraid of alienating attendees. Second, three political attitudes motivate
religious leaders’ speech: tolerance for democratic dissent; doctrinal
conservatism; and the belief that the political system is biased with respect to
their group.

three scenes

1 On the Sunday morning of the 2014 first-round election (Brazilian elections

are always on Sundays), I attended a worship service at the United Church of the
Kingdom of God (UCKG) of the Bela Vista neighborhood. Though the con-
gregation is located in a mixed-income neighborhood, the people in attendance

80
Encouraging Partisan and Electoral Politics 81

were poorer and darker-skinned than the average resident. Most of the hour-
and-a-half-long service emphasized a prosperity theology message. The pastor
exhorted us to have faith that financial rewards would come to those who
prayed and who tithed to the church; tithes, in fact, were financial investments
in one’s own future earnings. As the service was ending, the pastor called all of
the congregants to the front of the room to pray. When the group was assembled
in a tight mass at the front of the hall, he began to talk with us about the election.
He spent ten minutes urging us to vote for the congregation’s candidates for
federal and state deputy – men who were supported by many local UCKG
congregations. We chanted the five-digit electoral code for these two candi-
dates, which voters would need to know in the voting booth. The pastor told us
we were free to support whomever we wanted for senator, governor, and
president. [CO30]
2. Several weeks later, I sat in a circle conducting a focus group with nine
Assembly of God (Pentecostal) pastors in a cold upstairs room of one of their
churches. Most agreed that pastors should condemn homosexuality, but not
speak against the Workers’ Party (PT) for being pro-gay. Instead, congregants
should inform themselves – as one pastor said, “the partisan aspect is obvious”
if you just look. However, a few disagreed. One held that “it’s public informa-
tion. You should tell it like it is.”1 They generally thought that clergy should not
endorse candidates, and approved of their Head Pastor André’s approach. He
“tells everyone there are candidates among the brothers and sisters [i.e., con-
gregants], and we should evaluate them and consider their proposals carefully,
and vote for them if we want to, but he doesn’t ever mention names.” Another
noted that, “the pulpit is to preach the word of God,” not to talk about
candidates. Most of the participants were uncomfortable with the very public
politicking of the celebrity Assembly of God Pastor Silas Malafaia from Rio de
Janeiro. Just one participant thought Pastor Malafaia’s behavior was appro-
priate. Still, even this participant clarified that, “You have to tell people who the
church’s candidates are, but you can’t tell them who to vote for.” [FG6]
3 In April 2016, as Brazil’s Congress was considering the impeachment of
President Dilma Rousseff, a friend in Juiz de Fora told me a story about her
small, leftist Santo Daime church. She told me that social life had become highly
polarized – both at the university and on the streets – and there was constant
danger of conflict exploding between people who were pro- and anti-
impeachment. At a congregation affiliated with her own, she said, the pastor
had come out in favor of impeachment in a sermon. Congregants who were
unhappy about that sermon had left to join her congregation. Her own pastor
had called on congregants not to talk about the impeachment at church, but “as
soon as church is over, that’s all they talk about” (personal communication).

1
Literally, “chamar os bois pelos nomes” (“call the bulls by their names”).
82 What Clergy Think, Say, and Do

what clergy say about politicians and political

participation
A classic, oft-repeated evangelical saying from an earlier era held that “Believers
don’t mess with politics.” Do clergy still think that religious leaders should stay
as far away from politics as possible? As an Assembly of God pastor announced
in an interview in 2014, “I am a politician, it’s just that the politics I preach isn’t
that of men. I am the ambassador of the Reign of God on Earth. I think about
and invest daily in eternity” [FG6]. It turns out, though, that this perspective is
now only held by a small minority of clergy. Much more common is the point of
view that, as a different informant expressed it, “In the Bible, there are rulers.
Politics is in our blood here on earth. The Lord is the most important ruler, but
He isn’t going to come here to govern our country” [FG6]. Not only did clergy
widely agree that politics was necessary, but they thought members of the clergy
had wisdom that they should contribute to the political process. As another
pastor said, “In biblical times, you always had the priest walking alongside the
king, and we need that today” [CO23].
The clergy survey explored clerics’ visions of churches’ roles in politics. Some
activities were very highly endorsed by all clergy, both evangelical and Catholic,
including “participating in social movements that fight for the rights of the
poor” and “advocating for legislation that supports the values of this church.”
By contrast, political engagement around elections was more controversial.
Religious leaders typically expressed some hesitancy about pushing their
views too forcefully. A common point of view is expressed by an Assembly of
God pastor who advocated what many Brazilians call a “conscientious vote”:
“I try to stay informed . . . because that’s my role as a leader. I don’t pressure
anyone, but I tell them to get to know the candidates and do research to see if
they fit in their philosophy” [FG6]. Another pastor held that, “I preach that they
should vote for a person who obeys the reign of God. If there are ten candidates
who support the reign of God, they can choose any of them” [CO15]. And as
a third pastor reported when asked which candidate her church would support:
“We’re still praying. We ask everyone to pray so we can understand which is the
best candidate” [CO22]. I often saw clergy lead prayers for “God’s will to be
done” through the elections [CO11, CO15, CO29]. Though these prayers were
overtly nonpartisan, they framed the choice as one with religious consequences.
It is common for candidates, particularly at the local level and for state and
federal legislative office, to seek clergy support. It is also common for clergy to
refuse that support. Five days after the first-round election, a different Assembly
of God pastor recounted the following story:
The Assembly of God isn’t so involved in politics. It’s more involved in spiritual matters.
But it isn’t removed from politics either. We don’t have candidates. Votes are free. Now,
obviously, there are some candidates who are better and others who are worse. We had
two candidates who were members of [our congregation]. They asked for my help. I told
them that I supported them, and they could distribute their materials outside the church,
Encouraging Partisan and Electoral Politics 83

100
Percentage Saying Church Leaders Are Likely to:
80
60
40
20
0

Encourage Turnout Encourage Conscientious Vote Support Candidate

Catholic Evangelical Pentecostal

figure 5.1 Church-leader discussion of election campaigns

but I wouldn’t hand out materials for them. They brought a big box of campaign
materials to me, but I didn’t do anything with them. I already threw them away. [CO33]
This pastor’s resistance to allowing campaigning within church walls was not
ironclad, however. When I reinterviewed him in 2017, he described in a slightly
rueful tone how a coordinated group of supporters of one local 2016 city-
council candidate had managed to get their candidate onto the church’s events
agenda, and that the candidate had been booked to give talks and engage in
other activities during that campaign [CO39].
Figure 5.1 shows the percentage of clergy who said that it was “likely” or
“certain” (as opposed to “unlikely” or “very unlikely”) that leaders in their
congregations would “encourage members to vote”; “encourage members to
reflect faithfully on the election” (here called “conscientious voting”); and
“support a candidate to some office.” Majorities of evangelical and
Pentecostal clergy said their congregation would probably encourage both
turnout and conscientious voting. A little over 40 percent said their
congregation would likely also support a candidate.
Catholics were less likely than evangelicals and Pentecostals to agree that
their congregations would engage in any activity, though Catholic responses
varied a great deal according to the form of political engagement in question.
Fewer than one in five Catholic priests said church leaders were likely to support
any candidate for office. Catholic focus-group participants recounted a story
84 What Clergy Think, Say, and Do

they recalled of a Catholic clergy member who had tried a decade or so prior to
mobilize Catholics to support a list of Catholic candidates: “That didn’t work.
Not at all. Not everyone agreed with it” [FG1].
However, the one electoral activity with the participation of a majority of
Catholic clergy was the encouragement of conscientious voting. It is
common for Brazilian Catholic dioceses to produce handouts for clergy to
distribute to church attendees providing guidelines for voting (similarly, in
the US context, see Holman and Shockley 2017). Several versions of such
a handout circulated in Juiz de Fora during the 2014 campaign. One
developed by the Catholic communities of the Ecclesiastic Province of Juiz
de Fora for the 2014 elections began with a quote from Pope Francis:
“No one can demand that religion should be relegated to the inner
sanctum of personal life, without influence on societal and national life,
without concern for the soundness of civil institutions, without a right to
offer an opinion on events affecting society” (Pope Francis 2013). It then
stipulated twelve guidelines. The first numbered point simply described
which offices were being elected; several argued for the need to inform
oneself and to vote against corrupt candidates; and several final points
forbade both clergy candidacies and campaigning within congregational
walls, but called on Catholics to be civically engaged in other ways.
What candidates did religious leaders support? I did not broach this
sensitive topic directly with clergy. In my church observations,
congregational politicking on behalf of legislative or local-office candidates
was much more common than clergy support for presidential candidates
(see also Valle 2013). The Brazilian Electoral Panel Studies of 2010 and
2014 asked Brazilians which presidential candidates their clergy supported,
if any. Only about 15 percent of evangelicals and Pentecostals and 5 percent
of Catholics in each year reported their clergy member’s presidential
candidate. In the handful of cases where Catholic respondents knew their
priest’s candidate, Dilma Rousseff was the strong favorite in both years.
However, no single presidential candidate dominated among evangelical
clergy, and evangelical clergy did not exhibit strong preferences for
candidates of their own religious group. In 2010, only about a quarter of
evangelical support went to Marina Silva, the only major evangelical
candidate. Four years later, about half of evangelical clergy support went
to either her or Pastor Everaldo.
Though between a third and a quarter of evangelical clergy apparently
supported the Workers’ Party (PT) candidate Dilma Rousseff in each year,
I encountered a vehement strain of anti-petismo (that is, opposition to the PT)
among some Pentecostal and evangelical religious leaders (see also
A. D. Fonseca 2014). This anti-petismo was often expressed in dualistic
language. For instance, an Assembly of God pastor saw the PT as an
apocalyptic force supporting communism, totalitarianism, atheism, violence,
and the subversion of the Brazilian nation:
Encouraging Partisan and Electoral Politics 85

Brazil always followed the US in fighting communism . . . We don’t accept totalitar-

ianism or totalitarian parties. Here in Brazil we had the Revolution of 1964 [this is
the military regime’s term for its 1964 coup]. Those who fought the revolution were
affiliated with the PT, the Communist Party, the Communist Party of Brazil. Those
parties might change their name, but they stay the same. They don’t believe in God,
they don’t fight violence, they don’t support legislation that supports families. But
none of this surprises me . . . These things are signs that Jesus is returning. That’s
why I don’t preach against the PT. I know that it’s the will of God that it happens.
[CO33; also FG5, CO15, CO34]
As the last two sentences in this quotation suggest, though, this particular
pastor took a fairly hands-off approach to electoral politics. Though they
opposed the PT, anti-petista clergy were not necessarily strong supporters of
opposition candidates; 42 percent of evangelical and Pentecostal clergy who
believed the country’s current president hurt their religious group also said their
congregational leaders were not likely to support candidates. Still, anti-petismo
may have led such clergy to steer their flocks subtly or unsubtly away from
certain partisan options.
Yet partisanship was not limited to anti-petistas. Though most Catholic
clergy claimed no particular political tendency within their congregations,
some older priests described their current or former PT militancy. Padre
Miguel of the Santa Fé parish, for instance, told us that in the early days of
the working-class neighborhood’s settlement, he had supported leftist social
movements that fought for infrastructure improvements. Focus-group
participants in a Catholic parish reported that in the PT’s early years, “it
brought renewal and you could find good politicians there”; church leaders
had warned congregants to “be wary of tucanos [candidates from the
PSDB]” [FG2]. And at the national level, the leadership of the UCKG
strongly supported the PT in the presidential elections of 2002, 2006, and
2010.
Following the 2014 election, the question of what clergy can or should say
about politicians soon became salient once again, as members of most
congregations were deeply polarized over the legitimacy of the impeachment
of Dilma Rousseff. In such a context, any stray political comment from clergy
risked alienating some congregant, as exemplified by the third anecdote in the
introduction of this chapter. Nonetheless, 27 percent of respondents to the 2017
AmericasBarometer survey reported that they had heard a religious leader
discuss the impeachment: 7 percent said the clergy had been in favor of
impeachment, 6 percent against, and 14 percent said the clergy were neutral.
In interviews in June 2017, many clergy explained that they had been reluctant
to take a public position during the run-up to the impeachment, and that
likewise they were now reluctant to discuss the corruption scandals involving
President Michel Temer. As Pastor Willian at the Good News Baptist Church
explained, “the public has their own preferences and you have to guide them
without angering them.”
86 What Clergy Think, Say, and Do

why and when do clergy talk about electoral politics?

In the next two subsections, I argue that three sets of factors largely explain why
and when clergy support political participation: membership pressures,
perceived threat, and policy priorities. Before proceeding to these
explanations, though, I consider an alternative explanation. Do clergy avoid
endorsing candidates simply because doing so would be illegal? The Superior
Electoral Tribunal publishes an electoral code in advance of each election that
usually forbids certain forms of clergy campaigning – for instance, using
electronic PA systems, hanging posters inside or outside church property, and
distributing printed materials inside churches. These activities are prohibited in
all public spaces, and congregations are considered public spaces in the electoral
code. The Brazilian prohibitions might be considered analogous to the Johnson
Amendment in the United States, which threatens churches that engage in
campaign activity with the loss of their tax-exempt status.
The electoral code cannot fully explain clergy choices of when to engage in
political speech. For one thing, clergy can make their preferences known in
many ways without engaging in the kinds of overt “advertising” prohibited by
electoral regulations, even within congregational walls. Off church property,
clergy members’ endorsements are covered by free-speech protections, though
they are still subject to other electoral norms such as those related to electoral
calendars. For another thing, the penalties for violating electoral law are low or
nonexistent, as discussed in Chapter 1, especially in comparison to the loss of
tax-exempt status in the United States, and the probability of being caught is
also low. During my fieldwork, I occasionally observed violations of all of the
standards described in the electoral code. As relatively civically engaged opinion
leaders, most clergy were certainly aware of the electoral codes, which are
explained in mainstream media outlets. In all my discussions with clergy,
though, I only once heard a religious leader refer to the electoral code as
a reason to avoid candidate endorsements, and then as an aside. While it is
possible that the electoral code discourages clergy campaigning on the margins,
it is certainly not a full explanation for clergy reluctance to endorse candidates.
I suspect that the electoral codes are more likely to reflect prevailing social
norms regarding clergy campaigning than to shape those norms.

Citizens’ Norms about Clergy and Politics

To understand when and why clergy encourage political participation, or
instead hold their tongues, I turn briefly from the supply side to the demand
side. This section investigates citizens’ norms about what clergy should do
and say. As I will argue in this and the following chapters, citizens’ secular
norms are key to understanding Brazil’s culture wars, including both clergy
political speech and the sometimes limited effectiveness of clergy
endorsements.
Encouraging Partisan and Electoral Politics 87

In focus groups, congregants received a handout with a fictitious scenario

that involved a common saying in Brazil’s evangelical communities three or four
decades ago: “In [an evangelical church], Pastor João preaches to church
members, ‘Politics isn’t for believers.’ He’s waiting for the End Times and says
they should concentrate on saving souls.”2 Focus-group participants reacted
strongly and negatively to this scenario. One Catholic participant called the
expression “horrifying.” Another noted that, “you shouldn’t talk about politics
in church because people already made up their minds,” but that churchgoers
should still have political views [FG1]. A participant in an evangelical focus
group emphasized, “There is no dichotomy between being a citizen and being
a Christian” [FG3]. And a participant in the yet another group suggested that
Pastor João from the fictitious scenario “didn’t read the Bible. It’s a lack of
knowledge of the Bible, because there’s politics in the whole Bible” [FG4].
When the focus-group scenarios turned to clergy engagement in elections,
participants universally approved of some clergy behaviors. Even participants
in a group for people self-identifying as nonreligious agreed that it was right for
clergy to inform participants about the importance of voting and to deliver
nonpartisan messages encouraging “conscientious voting”: well-informed and
policy-based vote choices that are free of personalistic and clientelistic
influences [FG7]. Civically oriented messages are non-controversial, in part
because they echo other public-awareness campaigns. Within congregations,
not only clergy but also lay religious groups promote conscientious voting. For
instance, the Citizenship Committee, a local Catholic civic group that meets at
the Juiz de Fora Cathedral, runs a public-education campaign in high schools
every election. The group urges high schoolers, for whom voting is voluntary at
the age of 16 (and compulsory at the age of 18), to evaluate the candidates
carefully and not to sell their votes to clientelistic politicians [CO32].
However, focus-group participants had mixed feelings about other
behaviors. Their discomfort was particularly notable, because church
members who are willing to attend a focus group will generally tend to be
somewhat more engaged in and supportive of their group than average. In one
of the four evangelical/Pentecostal focus groups, all but one respondent
forcefully opposed clergy campaigning [FG5]; and in a second all but one
weakly opposed it [FG6]. A participant in the former group noted that, “there
are pastors in Brazil today who impose politics on their members and don’t
want to be questioned. They lead members by the leash3 to vote to promote their
own private interests.” That leaves their members “anguished,” she said,
because “it’s distressing and contradictory for church and politics to mix”
[FG5]. But even those who made some allowance for clergy engagement in

2
The saying is literally, “Politics isn’t a believer thing.” (Política não é coisa de crente.)
3
Literally, they “impose a halter vote” (“impõem um voto de cabresto”). This is a common idiom
referring to clientelistic voting with strong control and monitoring by a local boss, or
intermediary.
88 What Clergy Think, Say, and Do

politics often gave nuanced answers, along the lines of, “This is okay, but not
that.”
In a third evangelical focus group, respondents generally agreed that it was
fine for pastors to allow selected candidates to campaign within the
congregation, but the pastor should not actually officially endorse anyone
[FG3]. In the most permissive focus group, a participant drew the line at
allowing the candidate to give a sermon: “The candidate can’t get up at the
pulpit . . . If a candidate wants to be presented to [our congregation], the pastor
talks with him and presents him to us during worship . . . and prays for him, but
he doesn’t force anyone. We’re not being led around by the leash to vote.” Also,
he added, “the pastor doesn’t just tell us that he supports someone, he tells us
why he supports them . . . We’re not told to vote blindly. The pastor makes clear
his reasons.” Another participant added that “praying for the candidate is
biblical” [FG4]. Still, informal conversations with informants in other sites
indicated that the congregation of this permissive focus group had also lost
members who were unhappy with the intensity of clergy political mobilization.
Do congregants simply adjust their norms to the behaviors of their own
clergy? Some evidence suggests this does happen to a limited extent. In 2014,
two different national-level surveys included a question asking whether it was
appropriate for “a priest or pastor to support or campaign for a certain
candidate at election time.”4 In the LAPOP AmericasBarometer, conducted
in March and April of 2014, 49.8 percent of respondents gave this
hypothetical scenario the very lowest approval rating, and 85.4 percent
disapproved to some extent. At that time, surprisingly, Catholics and
evangelicals/Pentecostals reported indistinguishable levels of support for
clergy campaigning. In June 2014, though, just two to three months later,
a gap between religious groups appeared in the first wave of the 2014
Brazilian Electoral Panel Study (BEPS). Responses among Catholics were
nearly identical in the two surveys. However, evangelical approval of that
hypothetical scenario rose by eight percentage points between March/April
and June, to 23 percent. This suggests that as clergy began to form alliances in
preparation for the campaign that would soon begin, some evangelicals were
adjusting their norms to match the behavior of their clergy. Nonetheless, it is
also noteworthy that more than three quarters of evangelicals still disapproved
of clergy campaigning in June 2014, four months before the first- and second-
round elections.

Membership Pressures and Religious Competition

This section turns back to the supply side, clergy. Both evangelical and Catholic
religious leaders operate in an environment in which they need to be attentive to
their attendance levels. Perhaps a new congregation down the street –

4
In the AmericasBarometer 2014, the question is labeled BRAREL1; in BEPS 2014, it is REL1.
Encouraging Partisan and Electoral Politics 89

coreligionist or not – will poach a congregation’s members. Perhaps the youth

members will simply decide church is irrelevant to their lives and stop coming.
Perhaps a priest will stick his foot in his mouth, or perhaps a deacon will have an
affair with the head of the choir, and a quarter of the church’s members will quit
in disgust. Or perhaps a congregation will be so fortunate as to face the problem
of growing out of its current space, and suddenly needing to acquire new
property.
There are two political approaches clergy take to try to deal with these ever-
present membership concerns. Supply-side strategies involve seeking allies
within the state – in elected office or the bureaucracy – to advantage one’s
own group or limit the competition. State allies can be useful for helping
churches get building permits, radio licenses, or government contracts to
provide drug-treatment services (a strategy for evangelization). A well-placed
ally in the bureaucracy might be able to ease up on enforcing secular legislation
that prohibits evangelizing, or he or she might redouble enforcement – perhaps
preventing a rival church from being constructed, on the pretext that it would
obstruct traffic. A legislative ally could also introduce new legislation. Thus,
religious leaders support political candidates, in part, to shore up future
support.
The second set of political strategies for dealing with membership pressures
involves attention to the demand side – actual or potential congregants.
Religious leaders may simply try to behave politically in ways they think
religious “consumers” want, such as the Catholic clergy who become less
conservative in their preaching topics, as we saw in the last chapter. But this
strategy does not always involve moving toward the median religious consumer.
Religious consumers sometimes do not really know what they want, or their
preferences can change. Thus, enterprising and charismatic clergy can create
new markets for new forms of religious-political action.
If membership pressures lead clergy to seek state allies, the membership-
threat experiment discussed in the previous chapter should induce clergy to
support candidates and more actively lobby state officials. Figure 5.2 shows that
when evangelical and Catholic clergy are reminded of the threat of competition,
evangelicals become more supportive of legislative advocacy, while Catholics
become more likely to say their congregation will probably support a candidate.
In both cases, the less engaged group imitates its more engaged religious
competitor. However, the threat treatment has no effect on support for social
movements or plans to discuss turnout or conscientious voting – all relatively
indirect ways of seeking state allies.
Membership pressures likely had a much stronger impact on evangelicals’
political activities in the 1980s and 1990s. In that period, evangelical leaders
seeking to grow their congregations became known for insider politicking in
the interest of getting radio licenses and building permits (Gaskill 2002).
Denominations such as the UCKG developed formidable methods of
campaigning within their churches in order to maximize their electoral
90 What Clergy Think, Say, and Do

Social Movements

Legislative Advocacy

Turnout

Conscientious Voting

Candidate Endorsements

–.45 –.3 –.15 0 .15 .3 .45

Catholic Evangelical/Pentecostal

figure 5.2 Competitive threat affects legislative advocacy and candidate endorsements

power (A. B. Fonseca 2008; Reich and dos Santos 2013). In the process,
they created what social-movement scholars would call new “repertoires of
participation” – new ways of organizing their communities to impact
politics (Dalton and Welzel 2014; Tarrow 1998). These styles of
campaigning are reminiscent of earlier methods of clientelistic organization
in that pastors see congregations as blocs of votes, and exert personal
influence on “their” voters. However, pastors are different from the rural
bosses, or coronéis, of old. There are no clientelistic rewards for voting with
the pastor except, perhaps, spiritual ones; nor do pastors attempt to monitor
whether voters comply. Still, these methods have proven highly effective in
promoting congregational interests.
Even if clergy try to shape the competitive landscape by seeking political
allies, religious leaders must also be aware of members’ wishes. As the third
example in this chapter’s introduction indicates, a pastor who misjudges her
audience might inadvertently drive away a third of her congregants. Clergy
political engagement is likely to be particularly risky in congregations where
congregants are politically divided, or where clergy and congregants disagree
(Djupe and Gilbert 2002, 2009).
Encouraging Partisan and Electoral Politics 91

To test this proposition, I developed a measure of political diversity within

the congregation. In the clergy survey, 82.6 percent of Catholics, as well as
73.9 percent and 71.6 percent of evangelicals and Pentecostals, respectively,
said that their congregations were politically divided with respect to the
presidential election. Furthermore, 45.7 percent of Catholic clergy,
29.9 percent of evangelicals, and 31.3 percent of Pentecostals said that they
personally had “very” or “somewhat” different views from the majority of their
members. I combine these two variables into a single measure of the political
diversity within a congregation, running from 0 to 1. The mean level of political
diversity is .66 in Catholic congregations, .56 in evangelical ones, and .55 in
Pentecostal ones. The level of clergy support for the five different political
activities – social movements, legislative advocacy, turnout, conscientious
voting, and candidate endorsements – varies according to the level of political
diversity in the congregation. In evangelical congregations, congregational
diversity significantly decreases clergy support for all five activities. Among
Catholic clergy, congregational diversity appears only to depress support for
conscientious voting and endorsements.
To summarize the results in this section, membership pressures have mixed
effects on clergy support for political participation. On the one hand, being
reminded of the threat of losing members to other congregations boosts clergy
interest in supporting candidates and advocating for legislation – both ways to
find state allies who could aid one’s religious group in competition. I speculate
that this effect may have been stronger among Pentecostal clergy in the 1980s
and 1990s, when Pentecostals developed new styles of campaigning within
congregations, or new repertoires of participation, with the goal of improving
their access to goods the state controlled, such as radio licenses and building
permits. On the other hand, clergy must also keep in mind that most Brazilians,
and even most evangelical and Pentecostal congregants, believe that churches
should not get too involved in the electoral fray. Though evangelicals and
Pentecostals, in particular, appear willing to adjust their standards to some
extent, their norms are not infinitely elastic. Clergy are especially wary of
talking about politics in politically diverse congregations.

Policy Priorities and Views of the Political Process

Brazil’s culture wars are not simply the result of religious groups’ efforts to keep
and expand their memberships in the highly competitive religious market. Ideas
also matter. Three types of beliefs affect the way clergy talk with congregants
about political activism. The first involves perceptions of how the state treats
the religious group – whether it hurts the group (here called “grievances,” or
“perceived harm”); helps the group (“perceived favoritism”); or does neither
(“perceived neutrality”). Second, theological views – in particular, doctrinal
conservatism – matter. As the introductory scenes of Chapter 4 show,
discussions of conservative religious and policy views often accompany
92 What Clergy Think, Say, and Do

appeals to defend the Christian family. That chapter also revealed that
doctrinally conservative clergy are more likely to believe the state is harming
their group interests, and less likely to believe the state is neutral. Now, I argue
that doctrinal conservatism motivates clergy calls to political activism, even
after accounting for perceptions of the state’s bias or neutrality. Even doctrinal
conservatives who perceive the state as neutral feel compelled to take action to
change policy. Third, democratic attitudes also shape calls to action. Clergy
who are more supportive of democracy and who are more tolerant, both
internally and externally, support democratic participation more strongly.
In the past two decades, Brazilian politics and public policy have moved to
the left on many issues related to sexual and family traditionalism. Most
salient is the legalization of same-sex marriage in a pair of high court
decisions in 2011 and 2013. But this move leftward is notable in other ways.
For instance, the controversy surrounding the sexual orientation of Margarida
Salomão in the 2008 Juiz de Fora mayoral race was a precursor of later
controversy over the sexual orientations of other leftist politicians in local
and legislative races. In the National Congress, leftist legislators increasingly
promote what their conservative critics call “gender ideology” – the notion
that gender roles are socially constructed, and that transgender individuals
deserve protection. These changes are no doubt influenced by global trends
leftward on such issues across Western democracies. Abortion is another hot-
button issue. While abortion is legal only in a handful of special circumstances,
rightist and leftist politicians both seek to move abortion policy away from the
status quo. After a December 2016 Supreme Federal Tribunal decision, some
observers speculated that the high court might soon legalize abortion. However,
late in 2017 the lower chamber took steps toward criminalizing abortion under
all circumstances by moving a controversial bill out of committee.
Leftward policy trends likely mobilize evangelical and Pentecostal clergy
activism. Not only do evangelicals and Pentecostals take conservative
positions on the dimension of sexual and family traditionalism, but as we saw
in the last chapter, their positions do not readily move leftward in response to
social pressure. The same methods of organizing that promote particularistic
group interests can also help clergy pursue policy goals. Moreover, as
evangelicals and Pentecostals grow in number, these methods of organizing
become increasingly effective.
Table 5.1 explores how these three sets of attitudes, in combination with
membership pressures, shape clergy political activity.5 For simplicity, the cells
marked with negative signs show that a given independent variable (listed in the
leftmost column) decreases clergy support for the form of political action found

5
The dependent variables are left on their original scales, and the analysis uses ordinal logistic
regression. Catholics and evangelicals/Pentecostals are presented together in a single model
because the multivariate models are underpowered among Catholics. I discuss results for the
separate groups when they differ meaningfully.
table 5.1 Characteristics associated with clergy support for political activity (all religious traditions combined)

Social Legislative Turnout Conscientious Candidate

Influences Variables Movements Activism Message Voting Endorsements

Membership and Growth-Oriented +

Competition Congregation
Political Diversity in − − − − −
Congregation
Perceived State Bias Perceived Harm +
Perceived Favoritism + + + +
Religious Teachings Conservatism + +
Democratic Preferences Internal Tolerance +
External Tolerance + + + + +
Legitimacy of Democracy +
Religious Tradition Evangelical − − +
Pentecostal − −
Location Juiz de Fora + + +
Rio de Janeiro − + − −
Number of Observations 276 273 275 275 275
Note: Results from ordinal logistic regression models.
Negative signs denote statistically significant negative coefficients, and positive signs denote statistically significant positive coefficients. Dark gray cells are
statistically significant at p<.05; light gray cells are statistically significant at p<.10
94 What Clergy Think, Say, and Do

at the top of the column; a positive sign indicates that the given independent
variable increases clergy support for that form of political activism. The table
only shows effects that are statistically significant at p<.10 (light gray) or p<.05
(darker gray).
This analysis again demonstrates the importance of membership pressures.
Political diversity within a congregation substantially decreases clergy support
for all five forms of activism. At the same time, candidate endorsements are
more common in growth-oriented congregations.6
Interestingly, perceiving the state as biased – either in favor of or against
one’s group – also motivates clergy political speech. Clergy who think the state
is neutral are less likely to encourage activism. The impact on candidate
endorsements is quite large – the probability of endorsing candidates drops
from .47 to .26 for those who say the state is neutral.
Results from a quasi-experiment that occurred in the Fortaleza component
of the clergy survey indicate that evangelical leaders become more supportive
of political activism when reminded of the state’s treatment of their religious
groups. As discussed in Chapter 3, on the second day of the evangelical
conference a new, shorter version of the survey was fielded, to address
logistical problems. The two questions on the state’s treatment of religious
groups had been included at the bottom of the original treatment version
(“Version A”) and close to the bottom, directly above the “membership-
threat” treatment, in the original control version (“Version B”). However,
in the new “Version C” questionnaire fielded on the second day of the
conference, these questions were moved to the top of the questionnaire,
following a single question on recent changes in attendance levels, and
directly before questions on political activities. Differences between
responses to Versions A and B on the first day of the conference, and those
to Version C on the second day, are instructive. The “political-grievance”
treatment in Version C is not a true experiment, since it was administered on
a single day of interviewing, and administration did not randomize between
Versions C and Versions A and B. Differences in responses between Version
C and Versions A and B might be the result of some other shock to political
attitudes on the third day of the conference, perhaps a politically motivating
talk. Nonetheless, I am not aware of any such shock that could explain
differences in effects.
Figure 5.3 presents the impacts of the grievance treatment from the second day
of the conference, relative to the control-condition responses on the first day.7
The sample is limited to Fortaleza respondents, since they may differ in various
ways from evangelicals and Pentecostals interviewed in Juiz de Fora or Rio de
Janeiro. Respondents to Version C on the second day of the conference were

6
Results hold if we also include a control for the membership-threat treatment.
7
The list of dependent variables omits conscientious-voting messages because this question was not
asked in Version C.
Encouraging Partisan and Electoral Politics 95

Social Movements

Legislative Advocacy

Turnout

Candidate Endorsements

0 .2 .4
figure 5.3 The quasi-experimental impact of the grievance treatment on clergy
support for political activism

substantially and significantly more likely to support legislative activism and to say
that they would promote turnout and endorse candidates. The apparent impact of
the questionnaire change on support for legislative activism is very large. Further
analysis reveals that effects are only statistically significant among those reporting
that the state is biased. Among those who believe the state is neutral, the
questionnaire change is not significantly associated with any form of political
activism. While we cannot conclusively say that these differences were due to the
placement of the grievance treatment at the beginning of the questionnaire, they are
suggestive.
Religious views also matter. In Table 5.1, doctrinal conservatism predicts
support for legislative activism and conscientious voting. Breaking out the
analysis by religious group, however, we find that doctrinal conservatism
strongly affects the behavior of evangelical and Pentecostal clergy, but has
little impact on Catholic clergy.
Finally, democratic orientations shape how clergy talk to congregants about
political participation. The most important such orientation is external
tolerance. Clergy who are more willing to accommodate disagreement within
society at large are more likely to support all five forms of political
participation. In addition, those who perceive democracy as more legitimate
are more supportive of social movements, while those who are more tolerant of
disagreement within their congregations more strongly support legislative
activism.
96 What Clergy Think, Say, and Do

summary
Part II of this book shows that the current period of religious-political
polarization in Brazil has resulted from two forces. First, clergy members’
strategic efforts to maintain and grow their membership in a constantly
evolving religious marketplace affect which policy views they discuss, and
which political actions they endorse. Different religious traditions respond
differently to competition. Sometimes membership pressures induce clergy to
change their religious and policy-related teachings. For instance, when Catholic
clergy are reminded of the threat of competition from other religious groups,
they become less doctrinally conservative. Membership pressures can also
induce clergy to support political activity, as religious leaders seek political
allies in their struggles against competitors. However, membership pressures
more often lead clergy to refrain from endorsing political activity, for fear of
offending members.
Second, however, sincere, nonstrategic theological and political views also
matter. Some teachings are impervious to membership threats, including
Catholic leaders’ reluctance to discuss homosexuality, and evangelical
emphasis on doctrinal conservatism. The result has been a widening gap
between Catholics and evangelicals in policy-related teachings, as Pentecostals
and evangelicals have been less likely to evolve in a liberal direction than
Catholics. As politicians and policy have likewise moved in a liberal direction,
conservative religious groups have perceived the shifting issue space as a threat.
This has led to heightened evangelical mobilization, using the methods of
political organizing that were first developed to help Pentecostals develop
elected state allies.
part iii

HOW CONGREGANTS RESPOND

6

Church Influence on Citizens’ Policy Views and

Partisanship

Part III of this book considers how citizens respond to the policy priorities and
political positions of clergy. Do we find evidence of the culture wars on the
ground among congregants and the nonreligious? That is, do political
engagement and religious polarization among clergy lead to a widening gap in
policy attitudes or vote choice among citizens? This chapter investigates
differences (or in many cases, the lack of differences) across religious groups
in partisanship, views on policy toward Christianity, family/sexual
traditionalism, and socioeconomic and environmental attitudes.
It reveals that religious influence on policy attitudes in Brazil is partial – limited
to a relatively narrow band of issues. The evidence confirms prior research
showing that Brazilian evangelicals are fairly liberal on issues such as poverty
policy and social insurance (McAdams and Lance 2013; Nishimura 2004; Pew
Research Center 2006). In many policy domains, the views of Catholics,
evangelicals, the nonreligious, and adherents to other religions are
indistinguishable. This is the case for socioeconomic attitudes, as well as
opinions on the environment and race. However, religious differences are large
and growing on the acceptability of homosexuality and support for same-sex
marriage. Religious gaps are also widening in attitudes related to abortion, while
members of different religious groups disagree strongly on the extent to which the
Brazilian state should privilege Christianity. Finally, though there have been no
religious differences in partisanship in much of the post-2002 period, a partisan
cleavage appeared in 2017 between Catholics and adherents to other religious
groups. Thus, at the mass level, the Brazilian culture wars revolve around
a limited but highly politically salient set of issue and partisan attitudes.
What explains these religious differences? Using the congregational surveys
and national-level studies, I tested for social influence within the churches
studied, looking for congruence within congregations and between
congregations and their broader neighborhoods. The qualitative and
quantitative results indicate that communities develop distinctive political
cultures on certain groups of issues – namely, those related to family and

99
100 How Congregants Respond

sexual traditionalism and church–state relations. While there are essentially no

inter-congregational gaps on topics such as socioeconomic policy, race, or the
environment, congregations vary a great deal from each other on the
aforementioned group of culture-war issues. This suggests that religion may
lead to polarization only in some policy domains.
Religious influence on political attitudes is also asymmetric – affecting some
groups more than others. Doctrinal conservatives are more readily influenced
by both clergy and fellow congregants – likely pushing religious communities in
a conservative direction. Yet secular norms impose guardrails on clergy
influence. The previous chapter argued that citizen resistance limits what
clergy can say from the pulpit. Now we find that this same resistance also
limits congregations’ influence on members. Thus, in the context of widely
held secular norms, the clergy-driven nature of Brazil’s culture wars reduces
the scope of polarization, relative to what one often finds in party-driven culture
wars.
Before proceeding, a note on religious terminology is in order. Until now,
Pentecostals and evangelicals have often been analyzed and discussed
separately. Going forward, though, the term “evangelical” shall typically refer
to all Protestants, evangelicals, and Pentecostals. There are a few reasons. First,
Part II showed that Pentecostal and evangelical clergy behave similarly. Second,
citizen-level surveys treat Protestants, Pentecostals, and evangelicals
inconsistently within Brazil, due to widespread disagreement over how to
classify different denominations. Third, lay evangelicals and Pentecostals
themselves tend to report their own membership in these categories
inconsistently. While the Pew Forum estimates that Pentecostals have
outnumbered non-Pentecostal evangelicals in Brazil for at least the past
decade (Pew Research Center 2006), nationally representative surveys often
indicate that evangelicals outnumber Pentecostals. It appears that some
Pentecostals may not identify as such. Clergy, as religious experts, better
understand the distinctions, and more reliably report whether they are
Pentecostals or evangelicals.

four scenes
1 In late October, six attendees of the São José Catholic Parish sat talking with
me in a brightly lit, newly furnished upstairs meeting room. Church was
important in the lives of all six focus-group participants. One was an active
member of the Cathedral’s Citizenship Committee; another took part in the
School of Faith, an adult Catholic education program; a third introduced herself
by declaring that, “My interest is God, Jesus, Our Lady, and my faith.” But
when the topic of homosexuality arose, the lack of a common framework for
understanding the issue was notable. Some opined that homosexuality was an
innate biological trait, others that it was a behavior. One person declared, “It’s
in the Bible, it’s a sin!” Another came out as a strong advocate for same-sex
Church Influence on Citizens’ Policy Views 101

marriage: “We need legislation that serves everyone. The government can’t
force the church to perform gay weddings, but for civil weddings, we’ve got
to resolve this issue once and for all.” Noting the diversity in perspectives,
I asked if they talked much about this topic at church. The response was
universally in the negative. As one participant said, “those topics could upset
someone and they’d never come back.” As the group wrapped up, the partici-
pant who had expressed the most negative views wanted to make sure her fellow
participants did not think she was opposed to gay people. Everyone agreed that
gay people should be allowed to attend church and participate in activ-
ities [FG1].
2 Four days earlier, I had talked with a group of young adults in a middle-class
evangelical church. Most participants had some post-secondary education, and
they were generally somewhat more tolerant of homosexuality than other
evangelicals I encountered. They agreed with each other that, as one participant
said, “The practice of homosexuality is a sin, but we should love them without
judging them and condemning them. Their sin is different from ours. We can’t
reject those people as a minority.” Despite the relatively tolerant message, the
participants clearly considered gays to be an out-group with which they con-
trasted their in-group. They expected gay people who attended their church to
reject “the homosexual lifestyle.” Another participant added, “Just like we
respect homosexuals as people, they have to respect our principles too” [FG3].
3 One Sunday in September, 2014, the wood-paneled sanctuary of the Juiz de
Fora Assembly of God held perhaps 100 adults, clustered in four groups, all
receiving Sunday school lessons. I sat with a group of about thirty women
listening to the lay teacher, Sister Enilda. Following the curriculum guide dis-
tributed in Assembly of God congregations throughout Brazil, the day’s lesson
was from the Epistle of James, and emphasized sins of omission such as greed
and lack of charity. Forcefully rejecting the prosperity gospel, Enilda declared
that, “There will always be poor people. It’s not their fault. We need to take care
to provide social services.” She exhorted those of us who had more than we
needed to donate our best things to others. “The church is one single body.
If one person is sick, we’re all sick. In the first church, we were all together,
eating together, doing things together. We need solidarity and charity inside the
church.” She framed social service not just as the right thing to do, but as
enlightened self-interest: “Today maybe I don’t need assistance, but tomorrow
I will . . . Wealth is not forever” [CO21].
4 In October 2014, at the Good News Baptist Church, ten focus-group
participants sat on folding chairs in a circle in a large, warehouse-style worship
space. Sitting alone underneath the high ceilings, the group seemed very small.
As participants introduced themselves, one man told us that he had come to
understand how important politics is because the country’s laws were prevent-
ing religious education in schools. Working as a hospital chaplain, he had
102 How Congregants Respond

found, was a very good “strategy” for evangelizing outside the church. He
would like to have school chaplains as well, but state secularism (“o Estado
Laico”) did not allow it. Because of that, he had become “a very politically
engaged person.” Two other participants chimed in, arguing that the Brazilian
state was only “quote-unquote, secular.” “Catholicism controls everything,”
they said, and “we can only teach Catholicism” in the schools [FG4]. (Note that
Catholic education is legally prohibited in public schools, but religious instruc-
tion that “respects cultural religious diversity” is part of the curriculum
[L. A. Cunha 2009]).

religion and citizens’ policy attitudes

What do Brazilian citizens think about policy issues? Do congregants’ attitudes
mirror the ideological divides found among clergy? If the culture wars in the
United States can provide a guide, there is reason to suspect not. Though
polarization is widening there, ideological gaps between parties remain
greater between elites and activists than between other citizens (Abramowitz
2010; Fiorina and Abrams 2008; Fiorina, Abrams, and Pope 2006; Mason
2015; Wolfe 1999).
Does the same pattern occur in Brazil? On many issues, there are few
differences between citizens in different religious groups. Attitudes on
homosexuality and same-sex marriage – the same issues where clergy gaps are
largest – constitute one important exception. Evangelicals are also becoming
more conservative on abortion, leading to a growing evangelical–Catholic gap
on this issue. Finally, members of different religious groups disagree about
policies that would privilege Christianity over other religions. Taken together,
these findings indicate incipient ideological polarization, as citizens follow
clergy cues.
Figure 6.1 draws on six waves of the LAPOP AmericasBarometer survey to
assess how different groups’ policy views changed between 2007 and 2017. All
items are converted to 0–1 scales; higher values indicate greater support for each
policy.1 The largest and most consistent religious gaps relate to homosexuality.
In 2007, 49 percent of evangelicals objected to men who had sex with men, as
opposed to 36 percent of Catholics, 28 percent of those without a religious
affiliation, and 27 percent of those in other religious groups.2 Three years later,
AmericasBarometer began to ask respondents’ views on same-sex marriage.
Support for same-sex marriage gradually rose within every group between 2010
and 2017, mirroring global trends. Though international media and networks
have rapidly diffused new norms and understandings of homosexuality, in

1
Online Appendix C explains the wording of all of these questions (or in some cases, indices).
2
Question GEN8 was as follows: “Now I’m going to talk about a controversial topic. What do you
think about men who have sex with men? Do you think they are free to do what they want, or do
you think they’re shameless, or do you think they’re mentally ill?”
Church Influence on Citizens’ Policy Views 103

Same-Sex Marriage Abortion (Mother’s Health)

.8

.8
.6

.7
.4

.6
Level of Support
.2

.5
Economic Role of State Environment (Versus Economy)

.8
.55 .6 .65 .7 .75

.6
.4
2008 2010 2012 2014 2017 2008 2010 2012 2014 2017
Year

Catholic Evangelical None/Other

figure 6.1 Religious affiliation and policy preferences

Brazil the change has been much faster among non-evangelicals. In the seven-
year period, support for same-sex marriage rose from .42 to .58 among
Catholics, and from .54 to .68 among “others”; but the rise was only from
.28 to .34 among evangelicals.
This growing religious gap suggests that different religious communities have
been constructing varying understandings of homosexuality. Indeed, I observed
very different language across congregations. Catholics often justified tolerance
by saying, “We are all children of God.” By contrast, one participant in a Baptist
church focus group declared that, “Gay people are not children of God. A child
of God is someone who receives the word” [FG4]. This was the only time I heard
such an opinion, and I suspect that this participant was in the minority in her
own congregation. Still, evangelical churches commonly attempt to “cure”
homosexuality through counseling and prayer, and insist that they will only
accept gays who undergo therapy. In the same focus group, a participant
reported that, “Right now in Congress, they’re planning laws to extinguish
the family as a social base.” Another participant added that, “the church is
going to be forced to perform this kind of union.” A third worried that the state
of Minas Gerais would pass laws so that “even in schools they teach that it’s
normal” [FG4].
The AmericasBarometer survey reveals smaller, but likewise growing, gaps
on abortion. Religious differences in opinion on whether abortion should be
allowed when the woman’s health is in danger were not statistically significant
104 How Congregants Respond

in 2012 or 2014, but the gap became statistically significant in 2017. This was
largely due to increasing conservatism among evangelicals.
However, the single abortion question from the cross-national
AmericasBarometer poorly matches the Brazilian criminal code, which permits
abortion when the mother’s life (not health) is in danger, and in the case of rape.
The inconsistency is problematic, since opinions on abortion are highly sensitive to
question wording (Lewis 2017). The 2010 and 2014 Brazilian Electoral Panel
Studies (BEPS) asked about scenarios matching legal conditions: “Should abortion
(1) never be permitted; (2) be permitted only in exceptional circumstances such as
rape or risk to the mother’s life; (3) be permitted at will only in the initial stages of
pregnancy; or (4) be permitted without restrictions?” In BEPS 2014, only 6 percent
of Catholics and 4 percent of evangelicals favored liberalizing abortion laws,
compared to 14 percent of those without a religion and 9 percent of those in
other religions. Majorities supported the status quo: 57 percent of Catholics,
51 percent of evangelicals, 60 percent of those without a religion, and 58 percent
of those in other religions. On the conservative end, 37 percent of Catholics and
45 percent of evangelicals favored making abortion completely illegal, while
26 percent of those without a religion and 33 percent of those in other religions did
so.3
Abortion and homosexuality are the two most important issues cleaving the
electorate by religion, and ultimately driving Brazil’s culture wars. Turning to
a third aspect of family and sexual traditionalism, I find few religious differences
in attitudes with respect to women’s roles in society and the labor force.
In 2007, Catholics were slightly more conservative than other groups on
women’s status in the labor force. However, between 2008 and 2017, surveys
reveal no religious differences in attitudes toward women’s roles in public office.
Survey data reveal virtually no religious differences on most other policies.
As the third scene in the opening of this chapter indicates, many evangelicals and
Pentecostals strongly support the social safety net. From 2008 to 2017,
AmericasBarometer asked two questions on socioeconomic attitudes: whether
the Brazilian state should own important industries, and whether the state should
attempt to redress economic inequality. While the electorate as a whole became
more conservative over this period, at no point were there significant gaps
between religious groups. Nor are there religious differences in self-placement
on the left–right ideological spectrum over the 2007–2017 period, nor in
environmental and racial attitudes. In 2014 and 2017, AmericasBarometer
interviewers asked respondents if they would prefer to protect the environment
or promote economic development. In three years, support for the environment
(versus economic development) declined dramatically, likely due to a severe
economic downturn. Yet there were essentially no differences between religious
groups. And in 2012 (the only year in which the question was asked), evangelicals

3
Satisfaction with the status quo declined somewhat between the BEPS 2010 and 2014.
Church Influence on Citizens’ Policy Views 105

25
Predicted Percentage Petista
20
15
10
5

2007 2008 2010 2012 2014 2017

Year

Catholic Evangelical None/Other

figure 6.2 Religious affiliation and support for the Workers’ Party

were slightly, though not statistically significant, to the left of other citizens in
supporting university quotas for Afro-descendants.
Note that these views contrast dramatically with clergy priorities.
In Chapter 4, we saw that Catholic clergy are to the left of evangelical and
Pentecostal clergy on the economy, the environment, and racism. That is,
citizens’ attitudes fail to match local religious elites’ priorities on issues outside
the domain of family and sexual traditionalism. The lack of correspondence will
provide an important clue to understanding representation in Chapter 9.
What about partisanship? Chapter 5 found a streak of anti-petismo
(opposition to the Workers’ Party, or PT) among some – though certainly not
all – evangelical and Pentecostal clergy. Does this hostility translate to
congregants? Does the older Catholic–PT alliance register among citizens
today? Between 2007 and 2014, there were few statistically significant
differences in partisanship. However, Figure 6.2 shows that as the PT began
to decline in popularity between 2012 and 2014, evangelicals were quicker to
abandon the ship than were Catholics. By 2017, a statistically significant gap
had opened between Catholics and other religious groups.4
But one more policy issue drives religious divisions over politics. Scenes from
two focus groups illustrate debates over policies promoting secularism or
favoring Christianity. At the Good News Baptist Church, a participant who
worked in drug rehabilitation worried about suffering from “pressure from the
government” that prevented evangelization in public spaces and in publicly

4
This gap is robust to controls for demographics, region, and church attendance.
106 How Congregants Respond

funded rehab programs. Another participant was upset that Brazil’s anti-
discrimination and hate-speech laws punished some forms of speech against
homosexuality. “We can’t express ourselves at all. As soon as you say, ‘the
homosexuals,’ they want to shut us up” [FG4]. But not all evangelicals and
Pentecostals wanted restrictions loosened. At the Hope Church of the
Nazarene, one participant said, “We live in a secular state (‘um Estado
Laico’). I’m completely against evangelizing in schools. That’s dictatorship
and we don’t need it.” Another participant continued, “If the evangelical
church gets that opportunity, you have to open it up to everyone. If not, you
end up saying ‘you have to be this, you can’t be that.’ I’m not against it, but you
have to give everyone the opportunity” [FG5].
Unfortunately, nationally representative survey data on attitudes toward
state regulation of and privileges for Christianity are lacking. However, the
2014 congregational study provides a sense of religious divides. Church
members and citizens across Juiz de Fora were asked the extent to which they
agreed with “passing laws that recognize Christian values as the basis for our
nation.” Religious gaps were large: 74 percent of Catholics and 85 percent of
evangelicals agreed or strongly agreed with such policies, as opposed to
60 percent of the nonreligious and 50 percent of adherents to other religion.5
Differences among Catholics, evangelicals, and “others” are statistically
significant, even after controlling for frequency of religious participation and
the location/congregation of the interview.

do religious communities affect congregants’

attitudes on policy issues?
Do religious communities affect congregants’ policy views? Average public
opinion within religious affiliations does not tell us what happens in
congregations. A group’s average opinion on an issue could be identical to
that of the broader population, yet different congregations might take greatly
varying positions. Alternatively, the group average might be very far from the
mainstream simply because the group attracts people with extreme opinions.
This is a perennial question in scholarship on religion and politics.6 The query
always recurs precisely because it is hard to demonstrate social influence
conclusively, given that many people join congregations with which they are
already ideologically compatible. In Brazil, I argue, religious groups differ from
each other due to both selection – meaning that people choose compatible
religious communities – and socialization.

5
The agreement among adherents of other religions may seem high. However, the most common
affiliations in the “other” category, such as Spiritism and Umbanda, incorporate elements of
Catholicism, and adherents often maintain affective ties to Catholicism.
6
See Gwyneth H. McClendon and Rachel Beatty Riedl’s manuscript “From Pews to Politics in
Africa and Beyond.” Unpublished, 2018.
Church Influence on Citizens’ Policy Views 107

To begin to understand church influence, we can examine whether being

present in the pews – or on wooden benches or folding metal chairs – matters.
Does attending church affect citizens’ attitudes? On any given weeknight in any
medium-sized or larger city in Brazil, one finds dozens of storefront churches
open to meet the spiritual needs of both the faithful and passersby. Many
citizens, particularly evangelicals and Pentecostals, attend some form of
worship or prayer service several times a week, often for two to three hours
per service. One might imagine that this high volume of religious attendance
increases the likelihood of persuasion by the religious community. Indeed, in the
AmericasBarometer data, frequency of church attendance is strongly negatively
correlated with acceptance of homosexuality and support for same-sex
marriage, among both evangelicals and Catholics. However, the correlation
might simply be due to the fact that people who oppose homosexuality and
same-sex marriage tend to go to church more often.
Is it selection or socialization? The Brazilian Electoral Panel Study of 2010
contains clues. In the second wave, respondents were asked if they had begun
attending a new house of worship in the past year. The rate of reported
switching was surprisingly low: of interviewees attending church at least once
a year, only 5.2 percent reported having switched congregations. Among these
forty switchers, twenty-one currently identified as Catholics, fifteen as
evangelicals or Protestants, and four as having another religion. Almost all of
the Catholics said they had moved from other Catholic congregations, while
a little over half of the evangelical switchers had likewise come from Catholic
congregations.
Catholic switchers were distinctive from their coreligionists who stayed put –
they attended mass significantly more frequently and said that religion was more
important in their lives. Catholic switchers were equally as devout as evangelicals,
though they attended church less frequently than them. Catholic switchers differed
little from stayers in most policy attitudes, with one major exception: switchers
were significantly more conservative on abortion. In fact, 59 percent of Catholic
switchers said that abortion should not be permitted even in the case of danger to
the mother’s health, while only 33 percent of Catholics who stayed put took that
stance. Catholic switchers were substantially more conservative even than
evangelicals, among whom 43 percent of stayers and 46 percent of switchers –
including 43 percent of former Catholics – opposed abortion in all circumstances.
Within evangelicalism, by contrast, switchers were indistinguishable from stayers
in either religious devotion or policy views, with one exception: switchers were
more economically conservative.
So do conservatism and religious devotion lead people to switch congregations,
or does joining a new congregation lead switchers to become more conservative?
Given the very small number of switchers, it is probably impossible to establish
causality using statistical methods. Still, we can sketch out a story. Some switchers
may quickly adopt the attitudes prevalent at their new houses of worship.
However, Catholic-to-Catholic switchers are quite devout and very conservative
108 How Congregants Respond

Strongly
Agree

.75
Agreement

.5

.25

Strongly
Disagree
Recognize Different Same-Sex Abortion Environment Anti-Racism Anti-Poverty
Christianity Gender Marriage
Roles

Catholic Evangelical Pentecostal Community

Congregations Congregations Congregations Sites

figure 6.3 Clustering in policy views at eight congregations and community sites

on abortion – likely more so even than most people in their new congregations.
The conservatism and devoutness of both formerly Catholic evangelicals and
Catholic-to-Catholic switchers suggests that conservative, devout Catholics
leave congregations they view as permissive in search of stricter churches.

Congregational Results
Even if politics sometimes leads people to switch congregations, though,
congregations can still influence those in their midst. Both socialization and
selection likely occur simultaneously – sometimes even operating on the same
individual. The Churches North and South congregational study provides an
opportunity to peak into congregations to examine how opinions cohere and
change in different communities. Figure 6.3 examines support for various
policies in the eight congregations of the quantitative study, as well as
neighboring community sites. Responses are recoded to run from 0 to 1.
When the congregational average is above 0.5, the average respondent in that
congregation agrees with the policy (even if only slightly); when the
congregational average is below 0.5, the average respondent disagrees.
The confidence interval, represented by the whiskers surrounding each dot,
represents the variance in attitudes within each congregation. The larger the
confidence interval, the greater the diversity in attitudes within a congregation.
The figure reveals tremendous variation across and within congregations on
several issues core to Brazil’s culture wars. Though all groups supported laws
recognizing Christianity, the highest support was found in one evangelical
Church Influence on Citizens’ Policy Views 109

congregation, and the lowest in a Catholic parish. Congregations also held

diverse views on gender roles. Respondents were asked how strongly they
agreed or disagreed, on a five-point scale, that “it’s important for men and
women to have different roles in society.” Three of the four evangelical/
Pentecostal congregations on balance supported having distinct gender roles,
while majorities in three of the four Catholic congregations, as well as
respondents at the community sites, opposed them. Thus, while there may be
few differences in gender attitudes between traditions, there are major
differences within them. Turning to same-sex marriage, average opinions at
the community sites and in one Catholic congregation were neutral. At the other
end of the spectrum, respondents in the four evangelical and Pentecostal
congregations very strongly disagreed with same-sex marriage, as did those in
one Catholic congregation. Finally, respondents at all sites disagreed with
legalizing abortion. Still, there were differences across sites; respondents in one
Catholic and one Pentecostal congregation were substantially more liberal on this
issue than those in other congregations, and they aligned with respondents at the
community sites. By contrast, note three policy issues on the right side of the
figure. Intra- and inter-congregational variance on environmental protection,
combating racism and discrimination, and anti-poverty policies were strikingly
low, contrasting starkly with the variance on the first four issues.
Chapter 3 described one very small Charismatic Catholic prayer community
where the research team was able to interview only seven of the group’s fewer
than twenty members. Because of the small number of observations, this site is
excluded from Figure 6.3 and most quantitative analysis. However, results for
the seven interviewed members hint at what the liberal end of Catholicism looks
like. In this group, the mean support for same-sex marriage is .61, substantially
higher than at any site shown in Figure 6.3. In addition, the group mean on the
“separate gender roles” question is .32, putting it near the bottom. However,
this liberalism did not extend to abortion, where the group mean was only .07.
Why are some congregations within a tradition relatively progressive, and
others conservative? Are congregants’ attitudes affected by those of their
religious leaders? Clergy in seven of these eight congregations were
interviewed in the clergy study (the last refused). Merging the clergy responses
with the congregational data, we can test the correlation between congregant
and clergy attitudes. Though with only seven congregations our ability to draw
inferences is limited, hierarchical linear models nonetheless demonstrate that an
index of clergy doctrinal conservatism strongly predicts congregants’ attitudes
on gender roles, same-sex marriage, and special protections for Christianity.7
Moving from the lowest to the highest levels of doctrinal conservatism,
a congregation’s predicted level of support for same-sex marriage drops from

7
An index of the clergy member’s doctrinal conservatism turns out to be a better predictor of
congregant attitudes than the frequency with which the priest or pastor reports preaching on the
individual policy topic.
110 How Congregants Respond

0.25 to 0.01. Likewise, predicted support for different gender roles rises from
0.44 to 0.62, and the predicted probability of agreeing that Christianity should
receive special legal recognition rises from .77 to .92. Clergy conservatism also
has a smaller, marginally statistically significant association with attitudes
toward abortion and the environment. However, there is no association
between clergy attitudes and support for anti-racism or anti-poverty policy.
Thus, the congregational study reveals that religious communities develop
distinctive political cultures shared by clergy and congregants alike, yet
influence is partial – limited to some attitudes but not others.
Which individuals are most readily influenced? Influence turns out to be
asymmetric – limited to certain people. First, those who spend more time at
church, in worship service and other activities, are likely to be more strongly
influenced (Djupe and Gilbert 2009). Second, people with stronger secular
norms, as measured by opposition to enshrining “Christian values” in the
Brazilian nation, may resist religious influence. Third, doctrinal conservatism
may also matter. When the average opinion in a congregation diverges from
what doctrinal conservatives believe to be biblically correct, doctrinal
conservatives may actually resist influence. At other times, though, doctrinal
conservatives may be eager to follow religious leaders.8 Fifth, I control for
gender, since studies of socialization show that women are more sensitive to
social influence than are men (Djupe, Sokhey, and Gilbert 2007; Djupe,
McClurg, and Sokhey 2018); and for education and income, since both of
these resources may help citizens resist pressure to conform.
Figure 6.4 presents results from three variance function regression models.
Each model tries to predict which individuals are closest to the average opinion
in their congregations (Western and Bloome 2009; see also Boas and Smith in
press). The dependent variable is the individual’s deviation from the predicted
value for the congregation. When a coefficient for a given independent variable
along the left-hand side of the figure is statistically significantly below 0, this
means that that characteristic makes people more likely to agree with others in
their congregations. When a coefficient is statistically significantly above 0, by
contrast, it makes people less likely to agree with their congregations.
The analysis shows that people who spend more time in church come to agree
more strongly with their fellow congregants. However, only time spent in church
activities matters; attending worship service does not bring people closer to the
average opinion within their congregations. This suggests that sermons by
themselves are insufficient for social influence within congregations. Rather,
messages must be reinforced through repeated interaction with peers and
religious leaders. At the same time, those who more strongly adhere to secular
norms are more likely to diverge from their congregations on same-sex marriage

8
At the citizen level, by necessity I develop a different measure of doctrinal conservatism than the
one used for clergy. Here, doctrinal conservatism is the mean of citizens’ levels of biblical
literalism and fear of a wrathful God.
Church Influence on Citizens’ Policy Views 111

Different Same-Sex
Gender Roles Marriage Abortion
Frequency of Church Activities

Frequency of Worship Service

Doctrinal Conservatism

Support for Secular State

Woman

Education

Income

Catholic Congregation
–.1 0 .1 –.1 0 .1 –.1 0 .1

Coefficients represent impact of each variable on divergence from congregational

opinion. Estimates and 90% confidence intervals from hierarchical models regressing
squared residuals from first-stage hierarchical models of congregational opinion.
figure 6.4 Determinants of variance in policy views

and abortion. However, contrary to expectations, secular norms lead people to

agree with their congregations on gender roles. Doctrinal conservatism works in the
opposite direction to secular norms – it decreases adherence to the average
congregational opinion on gender roles, and increases adherence to the mean
congregational opinion on same-sex marriage. Finally, gender has a small and
marginally statistically significant impact only on support for abortion, such that
women are more likely to diverge from the congregational mean on abortion. This is
contrary to expectations, but women are significantly more supportive of abortion
rights in this sample. Perhaps women are simply more willing to resist their
(overwhelmingly male) clergy on this issue, as a result of differences in other
personal experiences as well as personal interests. Last, income fortifies resistance
to congregational opinion on just one issue, abortion, while education has no
impact on socialization on any issue.

summary
This chapter examines citizens’ policy views. Religious cleavages are pronounced
and expanding on two issues: same-sex marriage and abortion. By contrast, there
is little difference across religious groups in most other policy issues. The lack of
cleavages on issues such as racism, the environment, and the economy is
noteworthy – citizens fail to mirror the issue divides present among clergy.
Nonetheless, growing differences on the abortion and same-sex-marriage issues
may partially explain a new, though yet small, religious gap in partisanship.
112 How Congregants Respond

Where do congregants’ attitudes come from? Religious leaders’ levels of

doctrinal conservatism are correlated with members’ political views on issues
related to family and sexual traditionalism. This is likely due both to selection
and socialization. On the one hand, conservative citizens seek out conservative
congregations. On the other, once congregants arrive in the pews, clergy can
have substantial influence, and people who spend more time in the
congregational social environment are more strongly influenced. Moreover,
even after controlling for doctrinal conservatism and engagement in church,
evangelicals are more likely to cohere to the group position than are Catholics.
Thus, in the clergy-driven culture wars, the doctrinal positions clergy take
contribute to polarization in political views between congregations, and
between members of different religious traditions.
Nonetheless, citizens can resist clergy influence. As I show at the end of this
chapter and will show in the next one, secular norms condition congregants’
responses to church influence. Moreover, as the previous chapter demonstrated,
secular norms also constrain what clergy can say to their congregations without
fear of losing members.
Ultimately, the partial and asymmetric nature of clergy influence shapes
Brazil’s culture wars. The fact that clergy have influence on a fairly narrow
band of policy views may limit the scope of the culture wars at the citizen level.
Nonetheless, the fact that the citizens who are most readily influenced tend to be
conservatives, and ones within conservative congregations, ultimately polarizes
Brazilian politics.
7

Church Influence on Voting Behavior

In Chapter 5, we saw that both Catholic and evangelical clergy encourage

political participation. Sometimes, they also venture into partisan
discussions. But are clergy able to influence voting behavior – either
turnout or candidate choices? If people self-select into ideologically
compatible congregations, correlations between religious affiliation and
electoral behavior might result from that self-selection, rather than from
congregational influence.
This chapter begins by examining the many different ways religious affiliation
and vote choices have been correlated in recent elections. This exploration orients
us to the patterns we will seek to explain. The chapter then turns to the question
of what messages Catholics and evangelicals say they hear about elections in
church. Chapter 5 has already described what clergy say they say. However,
congregants might not get the messages clergy report sending (e.g., Zaller 1992).1
Perhaps clergy are unclear or deliberately vague; perhaps congregants are absent,
or simply tune out as soon as the pastor starts yammering about politics.
The results show that citizen and clergy accounts roughly align, yet leakage also
occurs in the communication pipeline. Even in the most politically engaged
congregations, some congregants fail to get the message.
Finally, the chapter turns to the thorny question of influence. Considering
multiple forms of evidence, the results support the conclusion that churches
directly influence congregants’ vote choices. Still, influence is far from
a naïve “hypodermic” model of persuasion in which clergy might inject
messages directly into the veins of congregants (on “hypodermic models of
communication,” and the model’s status as a straw man in contemporary
theory, see Lasswell 2011; Bineham 1988). Analysis of representative
surveys would indicate that clergy political messages are only occasionally
influential. To address the shortcomings of those surveys, the chapter turns

1
Also see Gwyneth H. McClendon and Rachel Beatty Riedl’s manuscript “From Pews to Politics in
Africa and Beyond.” Unpublished, 2018.

113
114 How Congregants Respond

to survey experiments conducted online in the 2012 local and 2014

presidential elections, as well as quantitative and qualitative data from the
Churches North and South congregational study. Taken together, these
forms of evidence strongly indicate that clergy and religious communities
as a whole influence voters, especially evangelical ones. Citizens who are
doctrinally conservative are more likely to respond to political influence
within church walls, while those holding highly secular norms resist
influence. Overall, and in combination with results from the prior chapter,
this suggests that at the citizen level, Brazil’s culture wars are fought most
intensively on the political right, and in evangelical congregations.
Still, it is worth remembering that evangelical clergy campaigning has often
ultimately had little impact on election outcomes, particularly in executive
races, because clergy have failed to coordinate their endorsements. Given the
fragmentation of both evangelical denominations and the Brazilian party
system, it has usually been hard to identify the candidate or party of choice
for evangelical elites and voters. Yet when clergy do coordinate, they can sway
election outcomes.

two scenes
1 On the morning of September 7, 2014, I missed several calls from my
fieldwork supervisor. By the time I saw the cell-phone notifications and
called her back, things were going badly. She and a team of research
assistants were visiting the Universal Church of the Kingdom of God
(UCKG) in the Santa Emélia neighborhood to conduct interviews with
congregants as they left the morning’s prayer service. They had been stand-
ing in the dimly lit, carpeted front entry hall, as I had previously agreed with
the congregation’s head pastor. Unexpectedly, the doorman – one of a large
number of smartly uniformed, devoted attendants who made things run
around the congregation – had asked the research team to leave.
I managed to get the doorman on the phone and smooth things over.
About an hour later, though, I received another call from my fieldwork
supervisor: The head pastor had rescinded his authorization for them to
conduct interviews within the church. They could keep the completed
surveys, he agreed, but they would have to leave. Later, I heard that shortly
before the doorman had asked them to leave the first time, a couple of
interviewers had looked into a meeting room with an open door to the front
entry hall, where they had seen church members assembling large stacks of
campaign posters for the state and federal deputy candidates that were
supported by the local UCKG congregations [CO13].
2 In October 2014, the Hope Church of the Nazarene was renting a storefront
congregation in a dusty, low-income neighborhood of Juiz de Fora. The roll-
down, metal front door spanned the front of the church, which was perhaps
Church Influence on Voting Behavior 115

six meters wide. On Sundays, the music blasting from the church surely woke
any sleeping neighbors for a block or two in each direction – perhaps one reason
for the congregation’s later move (see Chapter 9). At a focus group one Sunday
afternoon that October, the fluorescent lights overhead were turned off, and the
indirect light coming in through the sunny entryway gave the meeting a mellow
atmosphere. Congregants disagreed with each other over whether they should
support coreligionist candidates. One participant held that, “without
Christians in charge, things are only going to get worse, and I’m afraid the
church will end up accepting the values of the world.” As an example, he
mentioned that “in the Vatican, the Catholic Church is wanting to support
homosexuality within church. We should pray and watch out so that doesn’t
happen.” But another participant disagreed: “When we say that we need
evangelical legislators to defend our rights, I think that’s selfish. And the other
[religious] institutions, what will happen to them? Is it just about our rights?”
Another participant chimed in: “we shouldn’t just worry about our rights, but
about the rights of society.” She pointed out that evangelicals in office too often
get involved in corruption. “Maybe some laws are going to affect us, but the
solution isn’t to have believers [evangelicals] there [in government], because the
evangelical church is losing its identity” [FG5].

religious affiliation and voting behavior

Do Brazilian voters choose candidates along religious lines? As the two scenes
above indicate, evangelicals debate the extent to which candidate religion
matters. Analysis of surveys from the 2002, 2006, 2010, and 2014
presidential elections as well as the 2008 mayoral election in Juiz de Fora
reveals tremendous variation across time and space in the ways religion
“matters.” Sometimes evangelical candidates attract evangelical voters – but
non-evangelical candidates can also capitalize on religious cleavages. Whether
candidates polarize the electorate along religious lines depends on the partisan
context and mobilization by religious leaders.
To start, consider first-round presidential vote choice in 2002 and 2006 in
Juiz de Fora as well as Caxias do Sul, a city in the southern state of Rio Grande
do Sul.2 In 2002, religion mattered a good deal in both cities. A plurality of
Catholics, the nonreligious, and adherents to other religions preferred Lula da
Silva, who would go on to win the presidency in a second round runoff.
By contrast, a majority of evangelicals preferred the Presbyterian Anthony
Garotinho. In 2006, by contrast, religious affiliation was uncorrelated with
the vote – evangelicals voted nearly identically to Catholics in both cities.

2
The results discussed here are based on self-reports in the “Two City Study,” a six-wave, four-
year panel study focused on understanding how social context affected vote choice (for more
information on the study, and on the city of Caxias do Sul, see Online Appendix C).
116 How Congregants Respond

The major cleavage in that election was geographic. Those living in Juiz de Fora
preferred Lula, and those in Caxias do Sul preferred the candidate Geraldo
Alckmin.3
Further, consider 2010 and 2014 presidential vote choice, as reported in the
Brazilian Electoral Panel Studies. In both elections, Dilma Rousseff received
a plurality or majority of preferences in every religious group. Nonetheless,
religious affiliation still mattered, in that evangelicals were notably less
enthusiastic about the petista candidate than members of other groups. Yet,
despite the fact that most major candidates were consistent across the two
elections, the extent of in-group voting varied from one campaign to the
other. Assembly of God member Marina Silva captured greater support from
her coreligionists in 2014 than in 2010. In addition, the 2014 presidential
election included a minor candidate, Assembly of God Pastor Everaldo Dias
Pereira. As discussed in Chapter 5, Brazilian Electoral Panel Study (BEPS) data
indicate that Pastor Everaldo may have received more than a third of evangelical
clergy endorsements. However, he received less than 2 percent of evangelical
votes reported in BEPS 2014. One evangelical focus-group respondent reported
what he perceived as his congregation’s preference ordering, after Marina Silva
narrowly lost the first-round election in 2014: “Because he’s Catholic, Aécio
isn’t exactly our church’s ideal candidate . . . But if the choice is between an iron
rod and Dilma, we’ll vote for the iron rod. It’s bad for us now” [FG4].
Finally, Juiz de Fora’s 2008 local election presented yet another scenario:
religiously polarized voting despite the fact that all candidates shared the same
Catholic religious identification. In the first-round mayoral race, the
businessman Custódio Mattos was the strong preference of evangelicals and
those without a religion, while Margarida Salomão was the strong preference of
Catholics and those of other religions. Evangelical opposition to Margarida was
driven largely by her sexual orientation; the nonreligious preference for
Custódio may instead have been motivated by his business credentials.
Evangelicals’ united and fierce opposition to Margarida likely proved decisive
in that election, lending credence to a common evangelical claim: when
evangelical clergy are united, they can swing elections.
Taken together, these results urge a hard-to-resist conclusion: religious
affiliation matters in Brazil, but not due to any automatic tendency toward
religious in-group voting. Indeed, the second scene in the introduction of this
chapter suggests substantial ambivalence over the notion of voting based simply
on candidates’ religious identity. Instead, religious communities construct their
own stories of the political implications of their religious identities, beliefs, and
practices anew within the context of each election. Sometimes they interpret
elections in terms of group interests. At other times they focus on secular

3
The geographic divide relates to long-standing partisan differences in the two cities (for instance,
see Barry Ames and Reynaldo T. Rojo-Mendoza, “Urban Context and Political Behavior:
Partisanship and Polarization in Two Brazilian Cities.” Unpublished paper, 2014).
Church Influence on Voting Behavior 117

regional conditions and needs. At times they home in on the compatibility

between individual candidates’ personal lives and religious doctrine. In each
new contest, religious elites and masses debate, cajole, argue, persuade, joke,
and gossip as they collectively interpret and reinterpret political information
filtering in from the broader political environment.

receiving persuasive signals: what voting messages

do citizens hear in church?
On the Wednesday afternoon before the first-round election, I attended a prayer
service in a small, middle-class evangelical congregation in downtown Juiz de
Fora. Only about twenty-five people were in attendance, mostly retirees who
may have been downtown for errands. The service was short and subdued, and
there was little mention of politics. However, as the pastor concluded, he led the
assembled in prayer for the election, calling out, “Let us elect men and women
of God, who understand God’s will. The world is ruined, and only Jesus can
fix it.”
What political messages do citizens receive in church? Do the signals
religious leaders send break down between the mouths of clergy and the ears
of congregants? Prominent theoretical models of persuasion posit “receiving”
a message as the first step in persuasion (Petty and Cacioppo 1986; Zaller
1992). Before we attempt to parse out clergy influence, we need to assess
whether congregants actually receive the persuasive “treatment.”
To understand what citizens hear from clergy, I examined survey reports
from a local-level study of the 2008 Juiz de Fora election and the 2014 national-
level BEPS. Citizens are somewhat less likely to recount each type of election-
related message than are clergy, indicating some decay between clergy and
congregant. Nonetheless, strengthening confidence in the validity of the
measures, other patterns in the data correspond with clergy responses.
Citizens in both elections are more likely to report nonpartisan civic messages
than other kinds of messages. Between one half and three fifths of evangelicals in
both 2008 and 2014 said that clergy encouraged them to go to the polls or to
vote conscientiously. About two in five Catholics in each campaign did so.
By contrast, a little under 40 percent of evangelicals, and about one in five
Catholics, described having heard discussion of the political candidates.
Partisan messages and overt endorsements were the least commonly reported
communications in every religious group. In fact, in each campaign, only
around 5 percent of Catholics said that their clergy had endorsed candidates.
The one major difference between survey accounts from 2008 and 2014
relates to evangelical endorsements. About 35 percent of evangelicals in Juiz
de Fora in 2008 reported their clergy had endorsed a mayoral candidate; six
years later, just 15 percent of evangelicals nationwide described a similar
endorsement of a presidential candidate. This inconsistency is telling. Recall
118 How Congregants Respond

100
80
Percentage Reporting
60
40
20
0

Leaders Leaders Leaders Congregant

Encourage Encourage Support Discussion
Turnout Conscientious Candidate of Election
Voting

Catholic Evangelical Pentecostal

Congregations Congregations Congregations

figure 7.1 Campaigning and electoral discussion in eight congregations, Juiz de Fora,
2014

that the 2008 local election was marked by fierce opposition to the front-runner,
Margarida Salomão, on the basis of her sexual orientation. By contrast, though
the 2014 election featured two prominent evangelical candidates and a petista
female front-runner, neither in-group voting nor policy-related threat mobilized
evangelical clergy in the same way. The UCKG congregation discussed in the
first scene in the introduction of this chapter likely perceived my interviewers as
overly intrusive into the congregation’s campaign operations. However, the
congregation was exclusively focused on the legislative elections. As presented
in the first scene in Chapter 5, UCKG congregations in Juiz de Fora largely
ignored the executive-level races in 2014. In sum, these elections drive home the
conclusion that actual levels of evangelical mobilization vary dramatically by
race and over time, depending in large part on the nature of the candidates and
issues at stake in each election.
To what extent do congregations differ from each other? Once again, the
2014 Churches North and South congregational study provides the opportunity
to peek within religious communities. Figure 7.1 depicts dramatic variance in
clergy engagement across eight congregations. Similar percentages at all four
Catholic congregations reported hearing nonpartisan civic messages. In general,
Church Influence on Voting Behavior 119

members of evangelical and Pentecostal congregations were more likely to

report hearing turnout and conscientious-voting messages, yet the
congregation with the lowest reported level of civic messages in the sample is
Pentecostal. At this congregation’s Sunday service the week before the election,
I heard not a word about politics; the utterly nonpolitical sermon dealt with
how to have a loving relationship with one’s children.
In six of the eight congregations, very few respondents reported clergy
support for candidates. Only in two evangelical/Pentecostal congregations did
high percentages of congregants report clergy as supporting candidates. Finally,
the last set of estimates demonstrates that even in the least politicized
congregations, casual political discussions within church walls can mobilize
and influence congregants. Between 40 percent and 65 percent of respondents in
each of seven congregations reported hearing fellow church members talk
politics. Only one congregation had notably higher reported levels of political
discussion – the same one as had the highest levels of reported clergy
engagement on every other measure.
Do congregants’ reports match those of their clergy? Since the research team
interviewed clergy from seven of the eight congregations in the quantitative
congregational study, we can assess the correspondence between what clergy
and congregants say. I find that clergy and congregants generally agreed about
what was happening inside church walls. In addition, congregants who
attended worship services and activities more frequently, and who had
stronger congregational networks, reported clergy endorsements more
accurately. However, congregant reports of clergy speech regarding turnout
and conscientious voting are not correlated with clergy reports of their own
speech on those topics. Instead, congregants tended to say that clergy who
campaigned actively on behalf of candidates also encouraged turnout and
conscientious voting. Perhaps the rhetorical style of clergy who endorse
candidates from the pulpit makes a bigger impression on congregants than,
for instance, the printed guides to a conscientious vote common in Catholic
parishes.

accepting persuasive messages: do congregations affect

voting behavior?
The case of the minor presidential candidate Pastor Everaldo Dias in 2014 raises
interesting questions. If the electorate had been limited to evangelical clergy,
Pastor Everaldo might have performed quite well. In the wave of BEPS 2014
immediately following the first-round election, twice as many evangelical
respondents said Pastor Everaldo was their pastor’s choice than said he was
their own. There was zero overlap – a perfect negative correlation – between the
respondent’s pastor’s preference for Everaldo Dias Pereira and the respondent’s
own personal support for him. Not a single evangelical respondent whose
120 How Congregants Respond

pastor was said to support Pastor Everaldo followed their pastor’s direction,
and none of the people who said they had voted for Pastor Everaldo reported
clergy endorsement. Do clergy preferences have any influence on congregants?
Is the case of Pastor Everaldo an anomaly – either a real-life one or a result of
survey sampling error?
This section turns to the “A” in Zaller’s (1992) “R-A-S” (“Receive-Accept-
Sample”) model of opinion change. Do congregants adopt the political views
espoused by their clergy and fellow congregants? We begin by examining
congregation influence on turnout, civic engagement, and political knowledge.
The section then turns to vote choice. Throughout, three types of evidence
provide insight: nationally representative surveys; national-level, online
surveys with embedded survey experiments; and the Churches North and
South congregational study.

Turnout and Civic Engagement

Before examining turnout, it is useful to recall that voting is compulsory in
Brazil for those aged 18 to 65. Is it reasonable to study turnout in a compulsory-
voting country? A large body of literature shows that voting is fundamentally
a social process, often dependent on ties to others (Bello and Rolfe 2014;
McClurg 2006; Nickerson 2008). In a compulsory-voting country, however,
we might expect the impact of social influence on turnout to be smaller. Still,
turnout is far from universal in Brazil, typically hovering around 80–85 percent
in a given election, and fines are relatively minor. Scholars have found that the
same individual-level traits long known to shape voting in voluntary countries –
for instance, socioeconomic status and resources – are strongly associated with
voting in compulsory countries in general, and Brazil in particular (Castro
2007; Maldonado 2011; Power 2009; Singh 2011). Thus, it is plausible that
social processes may also be associated with turnout in Brazil, despite its
compulsory-voting laws.
In Brazil, as in many other countries, turnout rates tend to be higher among
religious than nonreligious citizens. For instance, analyzing six waves of the
AmericasBarometer survey, the average self-reported turnout rate is 85 percent
among Catholics, 83 percent among evangelicals and adherents to other
religions, and 78 percent among those without a religion. However, religious
citizens are different from the nonreligious in many ways, ranging from age and
socioeconomic status to propensity to join groups. More work is need to
uncover the role of church influence.
Figure 7.2 examines how self-reported turnout correlates with self-reports of
two kinds of congregational discussions: general discussion of elections and
candidates, and clergy speech promoting turnout. After taking into account
a large number of other characteristics that influence turnout, I did find that
those who said they heard election-related messages in church also reported
voting at higher rates. However, in the Juiz de Fora local election, only
Church Influence on Voting Behavior 121

Juiz de Fora 2008 Brazil 2014

.95
Probability of Having Voted
.9
.85
.8

Candidate Discussion Turnout Messages Candidate Discussion Turnout Messages

No Messages Messages

figure 7.2 Congregational messages and turnout in Juiz de Fora, 2008, and Brazil,
2014

congregational discussion boosted turnout significantly. By contrast, in Brazil’s

2014 national election, only turnout messages were statistically significantly
related to turnout. The reasons for these differences are not clear.
Is turnout higher in congregations where clergy say they promote it? Once
again, the Churches North and South congregational study enables us to
examine the correspondence between clergy survey responses and congregant
behavior. Unfortunately, though, the congregational survey was a pre-electoral
survey, and could not measure actual self-reported turnout. Instead,
congregants were asked about the “probability you will go to vote,” and
about the extent to which they believed that “voting is a moral obligation of
every citizen.” The clergy-based measure of pro-turnout messages is entirely
uncorrelated with congregants’ self-reported likelihood to vote. However,
clergy turnout messages are correlated with turnout norms. Respondents
interviewed at sites where clergy said they promoted voting were more likely
to believe that they had a moral duty to vote. In addition, those who attended
church more frequently and who were more engaged in church networks had
a stronger sense of duty to vote.

Vote Choice
Not surprisingly, highly religious Brazilians tend to see their vote choice as
a problem to be resolved using religious methods and criteria. As a Catholic
122 How Congregants Respond

participant in a focus group reported, “When it’s time, I just go in there [to the
voting booth], and I ask the Holy Spirit to illuminate me” [FG1]. Still, as an
evangelical focus-group participant put it, “Prayer can take you down various
paths” [FG3]. This section investigates whether clergy influence election
outcomes – whether they subtly or unsubtly steer congregants down one path
or another.
Are clergy endorsements effective? In popular accounts, clergy politicking is
seen as prima facie evidence that religion affects politics. Novaes (2002)
describes the evangelical style of activism as “religious clientelism.” However,
this seems to imply that voters automatically follow the direction of their
leaders, with relatively little resistance. Yet there is actually little evidence on
the extent to which parishioners follow church leaders’ guidance. Studies in the
United States and Mexico have both indicated that churches are less able to
sway the vote than often thought (Díaz Domínguez 2006; Putnam and
Campbell 2011). As discussed in the first chapter, Conrado (2001) reported
that around the turn of the millennium, UCKG officials projected that they
could expect a 20 percent fidelity rate in congregant support for church
candidates – notable, but far from the fidelity rates often found for political
partisans.4 If UCKG organizers’ estimates remain accurate, this rate of
compliance implies that even within what is arguably Brazil’s most politically
mobilized denomination, influence is far from automatic.
We asked survey respondents whom their clergy endorsed in three
representative studies: the local election study of the 2008 Juiz de Fora mayoral
race and the national-level 2010 and 2014 Brazilian Electoral Panel Studies.5
Interestingly, in 2008, reported endorsements are statistically insignificant as
predictors of the vote, after controlling for evangelicalism. This is striking, as
this election was, out of the three, the one featuring the highest levels of
evangelical mobilization. In both 2010 and 2014, by contrast, there is some
indication that endorsements might have mattered: clergy endorsement of
Dilma is a statistically significant predictor of the vote in 2010, while clergy
support for Marina functions in the same way in 2014. However, clergy
endorsements of other candidates are statistically insignificant in vote-choice
models in the two elections. Overall, based on these three elections, it would

4
For example, in the 2014 BEPS, 73 percent of people who identified as petistas (PT supporters)
in June 2014 voted for Dilma Rousseff in the October 2014 first-round election, while just
39 percent of those who did not identify as petistas in June voted for her in October.
5
The analysis in this paragraph results from three multinomial logistic regression models regres-
sing self-reported first round vote choice on reported clergy endorsements, controlling for
religious affiliation. In 2008, due to perfect multicollinearity, four individuals are excluded who
said that both they and their pastors had supported the minor candidate Tarcísio Delgado.
The models from BEPS 2010 and 2014 are based on the post-election waves, and include
a control for the lagged vote choice. The lagged dependent variables improve causal inference
(Morgan and Winship 2007).
Church Influence on Voting Behavior 123

appear that perhaps clergy have some influence on congregants, but that
Brazilians were certainly not lemmings willing to follow their pastors off cliffs.
Nonetheless, there are reasons to suspect that these observational results
fail fully to capture clergy influence. First, these particular campaigns might
not be the most appropriate ones for detecting clergy influence. The leading
candidates in these three races were from the Workers’ Party (PT) and the
Party of Brazilian Social Democracy (PSDB), two parties without strong
religious linkages. Clergy endorsements may also matter more in legislative,
rather than executive, races. When evangelical and Pentecostal churches get
involved in campaigning, they often do so in support of in-group candidates in
state or federal legislative races.
Second and more importantly, the use of congregant reports presents
potential problems. Congregants may not be particularly good at identifying
whom their clergy support. Not only do messages decay; more importantly,
clergy endorsements may often reach congregants only secondhand, filtered
through church networks in ways that obscure the original source. The diffuse
effect of candidate endorsements may explain the results from Juiz de Fora in
2008. In that highly mobilized election, even the major evangelical clergy
association officially took a stance. Evangelicals who failed to report the
treatment were certainly, in many cases, nonetheless effectively treated,
whether directly by their pastors or indirectly by their fellow congregants in
the pews, in Bible study, or standing at the bus stop the morning following
a rousing worship service.
Third, in the context of actual election campaigns, many persuasive forces
operate simultaneously within congregations. Direct messages from clergy
either endorsing or opposing candidates appear in an environment in which
congregants are likewise talking about the election. Moreover, clergy and
congregants also transmit persuasive arguments about policy issues, and
factual information about candidate positions on those issues. It may be
difficult to disentangle the effects of covarying factors such as church
attendance, exposure to clergy endorsements, information transmittal,
religious group identity, and general conservatism.
A couple of examples illustrate the complexity of electoral mobilization
within congregations. Take, for instance, the 2010 election discussed in the
introduction of this book. As Catholics in the pews shared DVDs in which
Dilma Rousseff appeared to endorse abortion, arguably the most important
aspect of the subterranean congregational campaigns involved information
transmittal. Nonetheless, the information transmitted carried an implicitly
persuasive, partisan message. Or take another example from the Good News
Baptist Church, in which focus-group participants discussed their anger at
policies that they perceived were preventing them from proselytizing in the
course of their public and charitable activities [FG4]. One participant argued
that “the people of God need to open their eyes and recognize” the importance
of getting involved in politics in order to maintain “freedom of expression in
124 How Congregants Respond

worshipping God.” In both of these cases, congregants framed the election in

terms of the potential consequences for policies of core interest to the religious
group. The invisible hand of clergy can be sensed as latent within the two
stories – preaching opposition to abortion, or raising congregant concern that
policies were hurting corporate evangelistic interests. Nonetheless, any actual
persuasive messages came from congregants themselves.
To isolate the effects of clergy campaigning more precisely, I turn to survey
experiments. During the 2012 local elections, a nationwide study tested the
impacts of clergy endorsements and candidate issue positions (see Online
Appendix C for details). In the control-condition group, respondents read
about a hypothetical city-council candidate named José Vargas dos Santos
who had been “working in the community for more than 20 years.” In three
treatment conditions, respondents received the additional information that
either (1) “He participates actively in the gay movement”; (2) “Various
evangelical pastors have spoken out against his political positions”; or (3) that
he was both a part of the gay movement and opposed by pastors. His party was
randomized across the PT, Party of the Brazilian Democratic Movement
(PMDB), and PSDB in all conditions. Respondents in both the treatment and
control-condition groups were then asked how likely they were “to vote for
a person like this,” using a seven-point scale (rescaled here to run from 0 to 1).
This experiment improves on the observational analysis discussed above.
First, it focuses on a local legislative contest prominently featuring candidate
positions on an issue core to Brazil’s culture wars, gay rights. This is precisely
the type of campaign where one might expect clergy politicking to
matter. Second, the controlled environment of the survey experiment enables
a cleaner treatment. The clergy message is unambiguous; it cannot be filtered
through or obscured by existing social networks. Third and perhaps more
importantly, we are able to disentangle more neatly the effects of clergy
recommendations and candidate issue positions. While scholars phrase
important cautions regarding the external validity of survey experiments
(Barabas and Jerit 2010), a research design combining experimental and
observational results enables us to triangulate across various forms of evidence.
Consider first the average treatment effects presented in the left pane of
Figure 7.3. Across all participants, all three treatments significantly depressed
candidate support, with slightly larger effects in the condition not including
information on pastors’ positions. The right pane of the figure presents
heterogeneous treatment effects across four religious traditions.
The experiment had no statistically significant effects among Catholics, the
nonreligious, or those from other religions. Catholics were slightly less
supportive of the candidate in the “gay-rights alone” condition, while those
with no religion and in other religions supported him slightly more in some
treatment conditions, but none of these results are statistically significant.
However, evangelicals responded strongly to the treatments. The largest
effects among evangelicals are found in the conditions mentioning the
 Catholics Evangelicals/Pentecostals

.6
All Respondents
.6

Level of Candidate Support (0–1 Scale)

.5
Support (0–1 Scale)
Level of Candidate

.4
.5

.3
.4

No Religion Other Religion

.6
.3

.5
Control Gay Rights

.4
Activism

.3
Evangelical Activism &
Clergy Rejection Clergy Rejection

figure 7.3 Experimental impact of clergy campaigning and candidate issue stances on citizens
126 How Congregants Respond

candidate’s gay-rights activism. Clergy opposition to the candidate had

a smaller, but still significant impact. When the two treatments were
combined, results among evangelicals are identical to those in the “gay-rights
alone” condition. These results confirm that evangelical voters pay attention to
cues from in-group elites, yet are as or more responsive to candidates’ actual
issue stances. This suggests that clergy stances might in part serve as a heuristic
to help evangelicals estimate candidates’ issue positions.
Clergy campaigning can also affect political outcomes even without persuading
people. Immediately after the question about candidate support, respondents were
asked about his electability. In an analysis not shown here, I found that all
respondents believed that all three treatments made the candidate less electable.
Candidates who are less electable will have a harder time building a base of
supporters and raising money. Thus, evangelical clergy activism can have an
outsized impact on a candidacy, even without persuading non-evangelicals.
Up to this point, the discussion has underscored a few conclusions. First,
what clergy say about candidates and elections matters – at least some of the
time. Second, congregations constitute complex information environments in
which various messages from various messengers can prove persuasive. Third,
not all congregants are equally susceptible to clergy influence. Only evangelicals
responded to the experimental treatments. Of course, it stands to reason that
evangelicals would be most sensitive to evangelical clergy. Still, it is telling that
the information on the candidates’ position on gay rights likewise had
a statistically significant effect only on evangelicals.
Other personal traits may also shape how congregants respond to clergy
campaigning. Congregants who oppose mixing religion and politics may be less
likely to support candidates who are identified as religious or supported by clergy.
Figure 7.4 presents results from a similar online national survey experiment
conducted during the 2014 general election campaign (see Online Appendix C).
The control-condition group again involved a candidate named José Vargas dos
Santos, this time cast as a candidate for state deputy and owner of a local bakery.
In this case, he was alternately described in varying treatments as endorsed by
Catholic leaders, endorsed by evangelical leaders, or active in his church. This
survey also measured secular norms by asking the extent to which respondents
agreed, on a seven-point scale, that “It’s right for people to base their attitudes on
religion,” and “It’s right for people to talk about politics and elections in church.”
The left pane of the figure shows that among all respondents, the treatments had
no impact on his support. The right pane, though, shows that the three treatments
(here combined for simplicity of presentation) had a positive effect among those
low in secular norms, and a negative effect among those high in secular norms.6

6
In an analysis of yet another experiment fielded in the same study, I found that only evangelicals
low in secular norms support in-group candidates, while non-evangelicals high in secular norms
oppose evangelical candidates. However, analysis of the 2014 BEPS indicates that secular norms
did not condition responses to reported clergy endorsements.
Church Influence on Voting Behavior 127

.4

.6
Candidate Support (0–1 Scale)

.5
.35

.4
.3
.3

.2
.0 .2 .4 .6 .8 1
.25

Secular Norms
Control Evan- Cath- Active
gelical olic in Control Religious Treatments
Endorse- Endorse- Church
ment ment

figure 7.4 The impact of candidate religious characteristics, by respondent secular

norms

Do congregational characteristics affect the extent of clergy or congregation

influence? To understand campaigning within religious communities, we turn
to the combined data from the Churches North and South congregational and
clergy studies. The disadvantage of returning to observational data is that one is
less able to control for factors covarying with different forms of religious
exposure – for instance, religious devotion. However, these observational
data are an improvement over other survey data in that they include
independent measures of clergy views.
In the Churches North and South studies, clergy activism apparently helped
congregants make up their minds in advance of elections. The congregational
study was a pre-electoral survey conducted between one and two months
before the election, when many respondents had not yet chosen candidates.
Though 86 percent of congregants interviewed reported that it was “very
likely” they would vote, only 63 percent reported their preferred
presidential candidate. In a compulsory-voting system, many people make
decisions immediately prior to the election, as the example of the woman
who asks the Holy Spirit to illuminate her in the voting booth suggests.
However, it turns out that, controlling for many personal factors that could
affect the timing of this decision-making, congregants were more likely to have
chosen a candidate in congregations where clergy reported higher levels of
electoral activism.
The congregational study also reveals that some influences operate within
congregations as a whole, beyond the agency of clergy in particular.
128 How Congregants Respond

In conservative religious communities, congregants are simply more likely to

vote the same way as each other. Political scientists have developed statistics
that assess the extent to which voters in a group are split across many options, or
instead coalesce on a single choice. One such measure is the Laakso-Taagepera
Index, which represents the “Effective Number of Parties,” or the “Effective
Number of Candidates” presented in a group of people (Laakso and Taagepera
1979). The index can range from a low of 1, when everyone in the group
chooses one single candidate, to as high as the number of members of the
group, if each person chooses his or her own separate candidate. I calculated
a Laakso-Taagepera Index for the diversity in the presidential vote in each
congregation. The resulting index ran from 1.36 in the least diverse
congregation, to 2.98 in the most diverse congregation. This index is
correlated at .74 with the level of congregational political diversity reported
by clergy.
What predicts a congregation’s level of political diversity? With just eight
churches, it is hard to make inferences. Still, even with only very few cases, this
index is statistically significantly correlated with an index of clergy electoral
activism (–.69) and clergy doctrinal conservatism (–.70). It is also correlated
with congregants’ own secular attitudes (–.60).Yet another factor is even more
strongly correlated with the fragmentation of the vote. Congregants’ own
average level of doctrinal conservatism is correlated at –.93 – a nearly perfect
negative correlation – with fractionalization of the vote.7 That is, congregations
where congregants are less supportive of secular norms, where clergy get
involved in elections, and where clergy and congregants are doctrinally
conservative are much less diverse in presidential vote choice.

summary
This chapter considered how clergy influence congregants’ electoral behavior.
As establishing social influence can be challenging, several forms of evidence
were brought to bear, including nationally representative surveys, survey
experimental data, and quantitative and qualitative data from congregational
studies. Across these studies, I conclude that clergy shape citizens’ political
behavior – turnout, civic norms, and vote choice. However, influence is far
from automatic. Once again, we find evidence that clergy influence is
asymmetric, affecting religious conservatives more than others. Evangelicals
may be more likely to take cues from clergy than Catholics, while members of
all religious groups who more strongly hold secular norms are less likely to
respond to clergy influence. Finally, doctrinally conservative congregations are
much more likely to coalesce in their candidate choices.

7
Congregants’ doctrinal conservatism is measured based on their levels of biblical literalism and
the extent to which they say they fear God’s wrath.
Church Influence on Voting Behavior 129

What is the upshot of this and the previous chapter for Brazil’s culture wars?
At the citizen level, religious politics in Brazil may be less polarized than
sometimes thought. Evangelicals, Catholics, and the nonreligious largely agree
on most major issues, with the important exceptions of church–state relations,
homosexuality, same-sex marriage, abortion, and perhaps gender roles. And
importantly, no party has developed strong voter loyalty within any particular
religious group. While religious groups often vote differently from each other,
no single pattern of religious cleavages has emerged in recent races. Sometimes
religious affiliation seems to affect vote choice; at other times clergy
mobilization appears to matter; and at still other times religion appears to be
trumped by other forces.
Still, we also find many signs of religious polarization. Though lay
evangelicals and Catholics feel similarly about many issues – ranging from anti-
poverty policy, to racism, to environmental protection – evangelicals are
becoming both increasingly conservative and increasingly distant from other
citizens on issues related to sexual and family traditionalism. Furthermore,
evangelical congregations exhibit a greater tendency toward homogeneity in
views on these issues. Evangelicals’ increasing conservatism may be one of the
forces driving a small, but potentially growing, religious gap in partisanship.
Furthermore, the growing conservatism of evangelicals is readily mobilized
into electoral politics. It bears repeating that the group now trending in
a conservative direction is precisely the one that has developed the most active
repertoires of political engagement in Brazil’s post-democratic period.
The results presented in this chapter indicate that religious groups, especially
evangelical ones, can and do influence “their” voters. Styles of campaigning that
clergy may have once deployed to promote in-group interests are readily
martialed in support of increasingly conservative candidates and issues.
Nonetheless, secular norms still matter. Brazilian citizens’ widespread
commitment to the separation of religion and politics constrains clergy speech
(Chapter 5), restricting the extent of politicking within churches. Moreover,
congregants with more secular views of society are more likely to resist
congregational influence on culture-war issues, potentially limiting the
ideological polarization of the campaigns that do occur within religious
communities (Chapter 6). Finally, secular norms lead some Brazilians – both
evangelical and Catholic – to resist the campaign-related messages they receive
within church walls.
Thus, we have a tension between growing ideological polarization and
religious engagement in politics on the one hand, and secular norms that
restrain the ferocity of such contests on the other. What is the end result for
Brazilian democracy? We turn to this question in the next two chapters.
8

Church Influence on Citizen Support for Democracy

How do clergy and congregant interactions around politics shape Brazilian

democracy? The previous chapter considered electoral politics. Elections
constitute the key institution of representative democracy – perhaps the most
important innovation in social organization in recent centuries. Nonetheless,
democracy can ultimately become sterile and fragile when citizens’ roles are
limited to turnout and voting. This chapter expands the view of democracy
beyond elections, arguing that Brazil’s religiously infused politics affects
citizens’ participation and trust in the broader polity, and in each other.
The chapter draws on several visions of democracy. Theorists of democratic
legitimacy argue that preserving democratic institutions requires citizens to
support the notion of democracy in the abstract, and to believe that their
democratic systems are basically legitimate (Booth and Seligson 2009).
Liberal, deliberative, and participatory democratic theorists go further,
positing that democracy is measured not only by how citizens feel about their
institutions, but by how they respond to and interact with their fellow citizens.
Deliberative democratic theory maintains that citizens build democracy by
coming together to discuss contentious issues across lines of difference, under
conditions of mutual respect and equality of voice (Gutmann and Thompson
1996; Ryfe 2005). Meanwhile, liberal visions of democracy do not require that
disagreeing citizens talk with each other, but that they simply tolerate each
other, respecting the rights inhering in each individual regardless of that
individual’s political views. Finally, participatory democratic theory
emphasizes citizens’ role in constructing democracy through an ongoing
process of engagement (Pateman 1970).
These visions of democracy are operationalized using several dependent
variables. First, the chapter considers the impact of clergy political
engagement on congregants’ abstract support for democracy and the political
system. Second, it turns to the relationship between religious participation and
citizens’ attitudes toward out-groups, as well as tolerance for the political

130
Church Influence on Citizen Support for Democracy 131

participation of gay candidates. Third, it examines the impact of congregational

life on several forms of political participation.
Does religious groups’ engagement in politics help or hurt the quality of
Brazilian democracy? The answer is “yes – it does both.” Congregations can
fortify democracy. Clergy transmit broad support for democracy and the
political system to congregants – no small feat in a country where democratic
legitimacy has eroded alarmingly in recent years. In addition, churches mobilize
citizens, helping them convey their political views to politicians. Yet divisive
clergy politics can also weaken democracy. When priests or pastors believe the
system is unfair to their own group, congregants lose faith in democracy and in
the legitimacy of the political system. In addition, religious attendance and
clergy campaigning heighten affective polarization along both religious and
partisan lines. Finally, evangelicals and frequent church attenders are less
socially and politically tolerant of atheists and gays. Thus, Brazil’s clergy-
driven culture wars stabilize some areas of democracy while simultaneously
eroding others.

five scenes
1 At the Good News Baptist Church, as voting returns rolled in on the Sunday
night of the 2014 first-round election, Pastor Willian sought to lift the spirits of
his congregants. Many people in attendance that night were disappointed that
Marina Silva, the presidential candidate most congregants had supported, had
narrowly missed going on to the second round, as the candidate Aécio Neves
had suddenly surged in the final two days of the race. Like a motivational
speaker, he called on congregants to shout “I am a winner!” several times,
each time getting louder. Even though Marina had narrowly missed second
place, he reminded us, the church had had some electoral victories at the level of
state and federal deputies and senator. Most importantly, he said, “you are a
candidate for the kingdom of God” [CO31].
2. In June 2017, I sat with Pastor Eduardo, head of the small storefront Hope
Church of the Nazarene, in his kitchen, eating pão de queijo (cheese bread) and
catching up on the changes in his church that had occurred since I last visited
(see Chapter 7). In general, Pastor Eduardo was not very interested in recent
political events, telling me: “What I want is to gather information about the
reign of God and the reign of Satan.” He did think politics sometimes affected
the reign of God. He told me that noise ordinances associated with environ-
mental laws inhibited his and other churches’ abilities to proselytize. His con-
gregation had had to move twice in the two years and eight months since I last
visited, in part due to complaints from the neighbors about noise volumes. He
recounted a story of a different church, where police had removed the sound
equipment. “Obviously you need to make sure people are following the rules,”
he said, “but the police are abusing the law . . . According to the Bible, according
132 How Congregants Respond

to the Word,” he continued, “Satan is the enemy. He always tries to inhibit the
reign of God through laws. Satan’s laws inhibit the reign of God on earth”
[CO43].
3 In June 2017, I also returned to the Disciples of Love Catholic prayer
community (see Chapter 4). Since I had last visited, the community had
grown from a humble one-room apartment into a three-story building, with a
large second-floor worship space and a third-floor chapel. Sitting on a crisp
leather couch in the newly constructed, high-ceilinged front entry room, I talked
with the eight assembled members about the community’s plans for
Charismatic Catholic evangelism throughout Brazil. As the conversation turned
to the state of the Brazilian nation, José Luiz, the group’s founder and leader,
situated recent events in apocalyptic terms. “This is not the end,” he explained,
“this is the beginning of the end . . . Democracy is already ending. They’re
removing the authority of the Constitution, but they continue to use the name
of democracy.” He agreed that the Brazilian people should fight to keep democ-
racy, but predicted that they would instead keep reelecting the same crooks as
always. Talking about the current president, he explained: “Temer is a
Satanist.” A bit taken aback, I asked, “How’s that?” José Luiz replied, “He
serves evil. Satan came to kill, steal, destroy. Temer has the attitude of Satan . . .
The people of God die in misery” [CO47; see also CO35].
4 On October 8, 2014, at 7:30 p.m., the Citizenship Committee convened at the
Catholic Cathedral downtown for what the chair formally announced was its
167th monthly meeting. Though the group is ostensibly ecumenical, all twelve
people in attendance were Catholic. The meeting began with the chair taking
attendance, noting several excused absences, and reading aloud the minutes from
the previous meeting. Attendees then reported on city-council activities over the
previous month – though one elderly member had entirely forgotten about the
city-council meeting to which she was assigned. A written report on that month’s
city-council activities would later be printed in the public newsletter the commit-
tee produces from members’ personal funds. The group also briefly discussed the
activities of the municipal Women’s Council, on which one of the members sat.
As the members discussed city-council business, they helped each other under-
stand the local legislative process and pending legislation.
5 Pastor Osésa Rodrigues, the president of the Liberal Christian Party, wrapped
up his hour-long talk at the Juiz de Fora Council of Pastors by calling on pastors
to take action (see Chapter 4). They should not just encourage congregants to
vote for evangelical candidates. Same-sex marriage was not a done deal, he said –
although the Supreme Federal Tribunal had legalized it, Congress could still
reverse the court’s decision legislatively. He exhorted us to work with the newly
formed, ecumenical Christian Alliance in Defense of the Family, which was
seeking to develop a popular initiative bill defining marriage as between a man
and a woman. A direct-democracy provision in Brazil’s 1988 Constitution
Church Influence on Citizen Support for Democracy 133

requires the legislature to consider any popular initiative that gathers 1.3 million
signatures. Pastor Osésa distributed pens printed with the new movement’s web
address, and called on each pastor to collect signatures within his or her con-
gregation. The talk concluded with his calling several members of the audience up
to the front to lay their hands on him to help him pray for the Christian family.

legitimate democracy
This section examines Brazilians’ perceptions of the legitimacy of democracy
and of the current political system. Have clergy and congregations exacerbated
or ameliorated the crisis of Brazilian democracy discussed in Chapter 1?
There are reasons to be hopeful that religious leaders can bolster congregant
support for democracy and the political system. As we saw in Chapter 4, clergy
express much higher levels of support for democracy than do citizens in general
in surveys. Even in 2017 in the midst of political crisis, clergy made clear their
conviction that Brazilian democracy should be defended. These attitudes could
lend legitimacy to a state that is losing its other sources of legitimacy. As highly
trusted, well-known civil-society elites, clergy could transmit their democratic
commitments to their flocks. Pastor Willian’s preaching at the Good News
Baptist Church, discussed in the first scene in the second section of this
chapter, provides an example of how clergy might help their congregants deal
with the inevitable disappointments of democracy.
There is also cause for concern that clergy could sometimes – perhaps
inadvertently – undermine congregants’ trust in political institutions. As we
also saw in Chapter 4, many evangelical and Pentecostal clergy perceive the
Brazilian political system as disadvantaging their own religious group. When
clergy preach about injustices against their religious in-group, congregants
might conclude that Brazilian democracy and the political system do not
work. Thus, as in the second and third scenes, citizens may extrapolate from
their personal dissatisfaction to the political system more broadly.
Brazilians’ levels of support for democracy and the political system fell
dramatically between 2012 and 2017 (see Chapter 1). What was the role of
religion? Strikingly, religious affiliation is not correlated with support for
democracy in the AmericasBarometer. That is, there are no consistent, statistically
significant gaps between religious groups in levels of support for democracy; nor
does church attendance predict support for democracy in any group.
However, religion is significantly correlated with citizens’ views on the
legitimacy of the political system. In every year of the AmericasBarometer,
legitimacy was higher among evangelicals and Catholics than among those
without a religion and in other religious groups.1 Moreover, in 2017 a
statistically significant gap opened between evangelicals and Catholics, such

1
These differences are robust to controls for education, wealth, gender, and size of place of
residence.
134 How Congregants Respond

Strongly
Agree

.75
Level of Support

.5

.25

Strongly
Disagree
Support for Democracy Legitimacy of Political System

Catholic Evangelical Pentecostal Community

Congregations Congregations Congregations Sites

figure 8.1 Attitudes toward the political system within eight congregations

that evangelicals were slightly more supportive of the legitimacy of the political
system; this difference is robust to controls for partisanship, political interest,
and demographics. Further, there is some evidence these religious differences
could be due to forces operating within congregations. In 2017 and across all
years, church attendance boosted levels of system support among both
Catholics and evangelicals, though not among those affiliated with other
religions.
To glimpse inside congregations, I turn once again to the 2014 Churches
North and South congregational study. Figure 8.1 examines variation in
legitimacy of democracy and the political system by congregation. Levels of
democratic legitimacy were fairly similar across congregations, though one
evangelical congregation reported significantly lower democratic legitimacy
than several others. The average legitimacy of the political system varied more
from congregation to congregation, with the two non-Pentecostal evangelical
sites a little lower than the other congregations.
What roles do clergy play? In Chapter 4, we found clergy who believe the
state is biased against their group’s interests tend to perceive the state as a whole
as less legitimate. These clergy attitudes may spill over into congregations, as
clergy talk about grievances with their congregations. Thus, I expect that when
clergy have group-related grievances, their congregants may become less
supportive of democracy and the political system; by contrast, clergy who
Church Influence on Citizen Support for Democracy 135

Democratic Legitimacy Legitimacy of Political System

Clergy Perception of State Neutrality

Clergy Support for Democracy

Church Engagement/Networks (Individual)

Doctrinal Conservatism (Individual)

Support for Secular State (Individual)

Evangelical

Educational Level

Income

Woman

Age Group
–.5 –.25 0 .25 .5 –.5 –.25 0 .25 .5

figure 8.2 Determinants of attitudes toward the state and democratic regime

strongly support democracy can bolster these attitudes. In addition,

congregants who are less supportive of state secularism may also perceive
democracy and the political system as less legitimate, since they perceive
restrictions on their own religious groups as unfair. For instance, a Baptist
focus-group participant announced that restrictions on evangelization had led
to a “critical situation.” He warned that, “if today’s government continues, we
could be facing a socialist dictatorship in Brazil” [FG4].
Figure 8.2 uses a multilevel model to examine determinants of democratic
support and regime legitimacy in the 2014 congregational study. Clergy views
mattered. Citizens whose clergy perceived the Brazilian state as neutral toward
their group, and those whose clergy more strongly supported democracy,
expressed higher levels of support for democracy and the political system. In
addition, citizens’ own levels of support for state secularism also affected the
legitimacy of democracy and the political system. As some of the examples in
this chapter suggest, the secular policies of the Brazilian political system can be
sources of grievance for those who are only weakly committed to secular norms.
After taking into account all these other variables, evangelicals had slightly
higher levels of support for democracy than did other religious groups, but
engagement in church networks was not statistically significant. Other
demographic variables also shaped legitimacy. Income was associated with
lower perceptions of state legitimacy, while older respondents perceived both
democracy and the current political system as more legitimate.

deliberative and liberal democracy

Democracy can be thought of as a set of institutions in which citizens resolve
policy disagreements under conditions of structural equality among all
136 How Congregants Respond

contestants. To maintain these institutions, democratic legitimacy – citizen

support for democracy in the abstract, and for the political system – is helpful
but insufficient. Abstract democratic commitments often founder when citizens
are asked to recognize the structural equality of those they dislike or even fear.
In this section, we examine citizens’ attitudes toward the out-groups with
whom they share the polity. Does political engagement within churches affect
citizens’ attitudes toward each other? There are reasons to worry that church
politicking could fuel mutual distrust.
Democracy carries within it a paradox. Free and fair elections are the core
component of contemporary definitions of representative democracy, yet the
combative nature of elections can make it harder for competing groups to get
along with each other in “peacetime,” and to trust the state. The very elections
that define democracy can gradually erode its quality. Electoral contests
exacerbate affective polarization – meaning large gaps in individuals’ attitudes
toward their in-groups and out-groups – among both winners and losers.
However, the legitimacy of the system itself takes a hit when contestants feel
wronged by the process, a feeling that is most common among election losers
(C. J. Anderson et al. 2005; Przeworski 1999). Citizens who come to see
electoral politics as a zero-sum game can also become intolerant of the
political participation of the groups they perceive as threats. Rising affective
polarization and mutual intolerance could potentially eventually even lead to
democratic breakdown. As Levitsky and Ziblatt observe in recent work,
“Democratic backsliding today begins at the ballot box” (2018, 5).
When religious groups take stances on contentious issues and races, they
could become affected by social polarization. Members of different religious
groups might come to dislike each other because of their groups’ policy views
and electoral alliances. Affective polarization might gradually widen over time,
as group differences are reinforced across multiple elections. The risk of
polarization is particularly great when religious groups engage in politics,
because religion is a powerful source of identity, with a basis in family ties,
childhood socialization, and adult social networks. Affective polarization along
religious lines may be exacerbated by the Brazilian context of religious
competition, if clergy seek to heighten perceived differences between groups
as a means of competition.
Religious doctrines involving a dualistic, good-versus-evil view of social
conflict may be particularly prone to affective polarization. Religious groups
sometimes literally demonize groups they dislike, attributing their behavior to
supernatural forces. One example of such literal demonization is the urban
legend that President Temer is a Satanist. This widespread myth may have
originated at the time of the 2010 election (Orrico 2016). Though prominent
evangelicals have issued defenses of Temer, stating that he is not a Satanist, the
message has not fully gotten through even to political elites or civil-society elites
such as clergy. In November 2016, for instance, evangelical federal deputy Cabo
Daciolo gave a speech on the floor of the Chamber asking Temer to renounce his
Church Influence on Citizen Support for Democracy 137

ties to Satanism (as well as Masonry) (Agência O Globo 2016). Yet literal
demonization is not limited to perceptions of Temer; it is also evident in the
way many evangelicals discuss the role of the LGBT movement, and the way
some Catholics discuss abortion. Believing that people whose politics one
dislikes are deliberately serving evil and destruction presumably reduces
tolerance of those out-groups’ civil liberties. There is no need to compromise
with the Devil, after all.
Brazil’s 2014 presidential campaign often involved heated, polarized
rhetoric. Toward the end of that campaign, I began to wonder if the intensity
of the election campaign had damaged trust between members of different
religious and political groups. In the post-election wave of the 2014 Brazilian
Electoral Panel Study (BEPS), we included questions on respondents’ attitudes
toward five groups: petistas (supporters of the Workers’ Party), pesedebistas
(supporters of the Party of Brazilian Social Democracy, or PSDB), and
Catholics, evangelicals, and “atheists, or people who don’t believe in God.”
The survey asked the extent to which respondents believed the actions of each
group are “very wrong” or “very right” on a 0–10 scale. Unfortunately, we do
not have data on these attitudes from early in the campaign, so we cannot track
how they changed over time. Still, the results indicate the extent of affective
polarization post-election.
Figure 8.3 shows the average rating of all five groups, by the respondent’s
religious affiliation. Responses have been recoded to run from 0 to 1, so the line
at 0.5 denotes a neutral rating. All four religious groups reported either neutral
or positive attitudes toward both Catholics and evangelicals. Catholics and
evangelicals, in fact, approved of each other fairly highly: Catholics rated
evangelicals at .69, and evangelicals rated Catholics .64. Apart from
Catholics’ and evangelicals’ ratings of their own groups (.83 for Catholics and
.87 for evangelicals), these were among the highest group ratings in the study.
Hence, there is very little evidence of social polarization between Catholics and
evangelicals. Furthermore, gaps between evangelicals and Catholics may be
closing over time. Comparing the ratings from 2014 with survey ratings of
evangelicals in 2002 and 2006, it appears that Catholics are gradually becoming
warmer toward evangelicals.2
Catholics’ and evangelicals’ mutual approval contrasts sharply with
partisans’ attitudes toward their party rivals. Returning to the 2014 data,
petistas scored their own political party at .84, and the PSDB at .41 – a gap of
nearly half the scale. Meanwhile, pesedebistas scored their own party as .75,
and the Workers’ Party (PT) at .43. Reassuringly, though, the partisan
polarization did not extend to religious groups. For instance, religious groups
basically coincided in their ratings of political parties. Every religious group

2
Data from 2002 and 2006 are from the Two City Study. Respondents rated only evangelicals in
those two years. Because the Two City Study was limited in geographic scope, we cannot draw
definite conclusions about change.
138 How Congregants Respond

Catholic Respondents

Evangelical Respondents

Nonreligious Respondents

Other Religion
0 .25 .5 .75 1
Levels of Approval of:
Atheists Evangelicals Catholics
PSDB PT

figure 8.3 Religious affiliation and intergroup attitudes

rated the PT more highly than the PSDB, though the PT–PSDB gap in ratings
was largest among the nonreligious and smallest among evangelicals. Likewise,
partisans were not highly polarized in their ratings of religious groups. Petistas
rated evangelicals a .73 and Catholics a .78, while pesedebistas rated
evangelicals a .78 and Catholics a .67. Thus, there is minor evidence of a PT-
Catholic versus PSDB-evangelical alliance in terms of in-group/out-group
attitudes, but the links are weak.
The absence of any serious Catholic–evangelical social tension is good news
for Brazilian democracy. Two aspects of Brazil’s religious demographics and
culture wars help to explain this felicitous finding. First is the lack of a religious
group–party alliance, which has prevented the affective polarization
characterizing the culture wars in the United States. That is, partisan divisions
do not consistently cleave religious groups. Second, the frequency of religious
conversion means that nearly all Brazilians have both evangelical and Catholic
friends and family members. Mixed-religion social networks have likely
reduced tension, just as scholars have found in the US case (Putnam and
Campbell 2012).
Nonetheless, we do find a different, and troublesome, form of affective
polarization. All groups fairly strongly disapproved of atheists. Atheists
received an average score of .09 from evangelicals, .14 from Catholics, and
Church Influence on Citizen Support for Democracy 139

.24 from adherents to other religions. Strikingly, even the nonreligious only
gave atheists a score of .30 – likely because the great majority of the nonreligious
say they believe in God. In a country where most people normatively value
religion, being nonreligious can be hard. An atheist in a focus group described
the social obstacles she faces: “In the case of people who already know me when
I mention that I’m an atheist, I see that they’re conflicted between the image they
have of atheists as something from the Devil, as people who are going to hell . . .
and the person they know me to be” [FG7].
Do forces within congregations foster mutual dislike between various
groups? Though the congregational study did not measure intergroup
attitudes, further data from the 2014 BEPS illuminate the sources of tensions
between groups. Based on the data presented in Figure 8.3, I calculate affective
polarization: the difference between the scores given to the highest- and the
lowest-rated groups (Iyengar and Westwood 2015). The measure ranges from 0
to 1, with higher values indicating a larger gap in group ratings. One score
corresponds to affective polarization between religious groups, and a second to
political parties.
In a multivariate analysis, religious variables affected partisan affective
polarization, and vice versa. Respondents who said their clergy supported a
candidate perceived bigger differences between the PT and the PSDB.
Meanwhile, people who sympathized with the PT were more affectively
polarized in both religious and partisan preferences. Using structural equation
modeling, I attempt to tease apart which form of polarization has the greater
influence on the other. Though the results must be taken with a bit of caution, it
appears that partisan polarization fuels religious polarization, but that religious
polarization has no impact on partisan polarization.
A critical aspect of democracy is that partisan contestants must recognize the
civil rights of groups they dislike or – perhaps with certain exceptions – perceive
as dangerous. There are both theological and social reasons to expect gaps
between Catholics and evangelicals in political tolerance. First, dualistic or
Manichaean religious doctrine identifying individuals, groups, or behaviors as
evil may encourage some citizens to deny equal participation to out-groups.
Second, recall that evangelical and Pentecostal clergy expressed substantially
lower levels of internal (intra-congregational) and external (extra-
congregational) tolerance than did Catholic clergy in the clergy study
(Chapter 4). If evangelicals take cues from clergy, we might expect the
religious gap in political tolerance to be greatest among those who attend
worship service more frequently.
To assess tolerance, we turn to the LAPOP AmericasBarometer, which has
regularly asked respondents whether homosexual candidates should be allowed
to run for office. This question is immediately relevant for real-world politics,
since religious groups in Brazil have often mobilized precisely to oppose gay
politicians – as in, for instance, the 2008 mayoral race in Juiz de Fora. Figure 8.4
examines the relationships among religious affiliation, congregational
 .7
Tolerance of the Rights of Gays

.7
to Run for Office (0–1 scale)

.6
.6

.5
.5

.4
Never Once/ Monthly Weekly >
Twice Weekly
.4

a Year
Catholic Evan- None Other Frequency of Church Attendance
gelical/ Religion
Pentecostal Other Religion No Religion
Religious Affiliation
Catholic Evangelical/
Pentecostal
figure 8.4 Religion, church attendance, and political tolerance
Church Influence on Citizen Support for Democracy 141

attendance, and tolerance for gays running for office. Consistent with prior
research in other countries, the nonreligious are the most tolerant, and there are
sizable differences among various religious groups (Beatty and Walter 1984).
Catholics are significantly less tolerant than are the nonreligious and those
within “other” religions, while evangelicals are less tolerant than Catholics.
The right pane of the figure examines how church attendance affects this form
of political tolerance. Religious attendance has a small, statistically significant
impact on political tolerance among Catholics. However, church attendance
more substantially reduces tolerance among evangelicals and “others.” Among
those who never attend church, the nonreligious and those in other religious
groups have nearly identical levels of tolerance, as do evangelicals and
Catholics. Yet levels of tolerance are substantially lower among evangelicals
and adherents of other religions who attend church regularly.
Thus, congregations have mixed impacts on Brazilians’ social and political
attitudes toward their fellow citizens. There is positive news. Although there is
some evidence of tension between religious groups in the highly charged
political climate in 2014, evangelicals and Catholics still had fairly positive
attitudes toward each other. Yet there is also reason to worry. Partisanship – in
particular, petismo – hurts intergroup trust, and clergy who get involved in
election campaigns contribute to partisan polarization. Moreover, acceptance
of atheists is extremely low. Finally, religious citizens are much less willing than
the nonreligious to extend civil rights to groups that they dislike, and attending
church erodes tolerance further, most prominently among evangelicals.

participatory democracy
Citizens’ ideas and behaviors support or undermine democracy on an ongoing
basis, this chapter argues; electoral participation alone is insufficient. The
previous sections investigated citizens’ attitudes toward the democratic regime
and each other. This section turns back to citizen behaviors, focusing on
participation outside of elections and campaigns. Eminent political theorist
Robert Dahl proposed that “a key characteristic of a democracy is the
continuing responsiveness of government to the preferences of its citizens,
considered as political equals” (1971, 1). Such “continuing responsiveness”
would require “institutions for making government policies depend on votes
and other expressions of preferences” (3).3 Implicit in these propositions is the
requirement that citizens regularly communicate their political desires to each
other or to government, beyond the limits of any electoral cycle.
Do congregations and clergy encourage political participation beyond
campaigns? In the previous chapter, we saw that religious leaders effectively
promote turnout and conscientious voting. Now we turn to several other forms

3
This is a tall order; Dahl recognized that the actually existing more-or-less-democratic regimes,
which he preferred to call “polyarchies,” had never fully fulfilled it.
142 How Congregants Respond

of political participation, ranging from highly institutionally oriented to

noninstitutional. First, citizens can lobby political representatives for policy
changes. Second, they can take advantage of the tools of direct democracy
such as popular referenda, attempting to force policy change. Third, they
make take to the streets in protest, using repertoires of contentious
participation to attempt to raise awareness of issues not on public agendas.
Fourth, they can work at the local level to solve communal problems in their
own communities, without the intermediation of politicians or recourse to the
legislative process.
There are two reasons to suspect that congregations encourage all these
forms of participation. First, clergy often talk about non-electoral
participation. Recall from Chapter 5 that clergy overwhelmingly agreed that
their congregations should both advocate for legislation promoting the church’s
values, and support social movements that advocate for the rights of the poor.
We have found anecdotes throughout the book in which clergy have encouraged
political participation: from lobbying efforts to stop a local gay parade, or to
halt a neighboring congregation’s construction (see Chapter 1); to Pastor
Osésa’s work to pass a popular referendum to force the legislature to consider
a bill overturning same-sex marriage (the second section of this chapter).
Furthermore, in Chapter 7, we saw that clergy efforts to promote one
particular form of participation, turnout, were effective.
Second, clergy activism is not the only force that makes religious
communities potent sites for mobilization. Instead, social networks built
within religious institutions foster political organizing. Religious communities
offer would-be community leaders a large pool of potentially mobilizable,
highly interconnected participants, as well as mechanisms for monitoring and
accountability. In addition, congregations themselves provide a wide variety of
opportunities to work in ministries and activities. That is, many aspects of
church life constitute forms of civil-society participation that could potentially
lead to further community engagement. The secondhand clothing charity store
run by the Disciples of Love Community and the Citizenship Committee at the
Catholic Cathedral constitute just two examples of congregational
contributions to civil society.
Though the congregational survey did not include questions about non-
electoral participation, the LAPOP AmericasBarometer examines many forms
of political participation. As expected, within each religious group, church
attendance is associated with contacting representatives and signing petitions,
after controlling for a host of demographic factors. Religious participation also
increases the likelihood of citizens saying that they had worked with others to
solve a community problem in the past year, or attended a meeting of an
association that seeks to improve the community. The effect of religious
attendance on community problem-solving is particularly pronounced among
Catholics, as the examples in the previous paragraph might suggest.
Church Influence on Citizen Support for Democracy 143

Political Protest
Since 2013, contentious forms of politics such as protest have become
increasingly common in Brazil. Like the other forms of participation, protest
constitutes an important way for citizens to express policy views. Protests, in
fact, may be distinctively effective in bringing new issues to the policy agenda
(Moseley 2015). Nonetheless, protest can have an ambivalent relationship with
democracy. Not only does protest often express grievances against the political
system, but in its extreme form it can destabilize elected governments or even
regimes. Thus, protests uniquely reflect the democratic troubles Brazil has faced
since 2013.
Do religious groups boost or discourage protest participation? Religious
communities could serve as a ready site for mobilizing protests. Catholics and
evangelicals heading to the streets might recruit a few church friends to go with
them, just as someone headed to a shift at the congregation’s charity shop might
also ask a friend from church for help. Asked what he would tell a parishioner
considering going to a protest, Father Miguel from the Santa Fé parish responded,
“Go! And take someone with you!” [CO36]. Though clergy largely ignored the
topic of protest when interviewed in 2014, three years later a number of clergy
explicitly acknowledged protests as legitimate. Pastor Eric from the Vila Bela
Methodist Church opined that citizens should each take part in the activism
corresponding to their social groups: teachers should take part in protests by
teachers’ groups, and union workers in workers’ strikes [CO37].
Nonetheless, evangelicalism could also discourage protest. Even in 2017,
most evangelical clergy expressed discomfort with protests. As Pastor Djalma
explained, “People don’t understand that when you’re a Christian, you don’t
have time for those things. We’re too busy building the kingdom of God”
[CO40]. A 2014 focus group for evangelical young adults indicates how
patterns of interaction within congregations and beyond church walls can
discourage protesting. A college student reported that she knew a lot of
evangelicals at the university who had gone out to the streets, but that many
other young people at church urged each other not to participate. Another
worried that if she participated in protests, she might “damage biblical
principles.” And a third explained that she had gone out to the streets, but
had been turned off by foul language and indecent behavior [FG3]. Thus, both
social influence in congregations and standards of behavior socialized in those
settings discouraged youth in this congregation from protest participation.
Figure 8.5 presents levels of protest participation, by religious group, across
the six waves of the AmericasBarometer. Across almost all of this period, the
nonreligious and those of “other” religious affiliations were most likely to take
part in protest. Nonetheless, we find intriguing religious dynamics. Between
2010 and 2014, religious gaps grew, as protest participation rose more quickly
among “nones” and “others” than among Catholics or evangelicals. However,
between 2014 and 2017, religious gaps closed, as protest participation grew
144 How Congregants Respond

15
Protest in Past Year (Percent)
10
5
0

2007 2008 2010 2012 2014 2017

Catholic Evangelical/ No/Other

Pentecostal Religion

figure 8.5 Protest participation by religious affiliation, over time

dramatically among evangelicals and Catholics. These dynamics reflect the

changing nature of protest in Brazil. Protests until 2014 were often organized
by leftist groups. By contrast, by 2017 protest had become a tool of both the
right and the left in the struggle to define the future of the country, and in the
context of prolonged political crisis.
To what extent did religious communities encourage or discourage protest
participation? Here, the experiences of evangelicals and Catholics differ
dramatically. Among Catholics, both in 2017 and across all years, church
attendance is associated with rising protest participation. Analyzing
AmericasBarometer data, the predicted probability of Catholics reporting in
2017 that they had participated in a protest in the past year rises from .13 for
those who never attended church, to .18 for those who attended church more
than once a week, after controlling for many forms of demographics. By
contrast, among evangelicals in the same period, church attendance
decreased, rather than increased, protest participation. These findings align
with the qualitative results from interviews and focus groups.

summary
How do clergy and citizens’ interactions around politics shape Brazil’s
prospects for deliberative, liberal, legitimate, or participatory democracy?
This chapter shows that churches have multivalent impacts on Brazilian
democracy. First, congregational life shapes citizens’ trust in the basic
Church Influence on Citizen Support for Democracy 145

institutions of the nation state. Clergy support for democracy can critically
bolster the system’s legitimacy, in a period when Brazilian citizens have
become increasingly disenchanted with their current regime. Nonetheless,
when clergy believe the political system is biased against their group, they
undermine support for democracy and the political regime among citizens.
Second, congregational life affects citizens’ attitudes toward out-groups. The
good news is that, despite the “clergy skirmishes” that often drive Brazil’s
culture wars, Catholics and evangelicals have fairly positive attitudes toward
each other. Still, there is some evidence that religious politicking fuels partisan
tensions, and that partisan tensions in turn fuel religious ones. While the
connection between affective polarization on religious and partisan lines was
weak in 2014, repeated battles across the same fault lines in subsequent years
could exacerbate social tensions. In addition, dualistic religious doctrines more
generally affect attitudes toward atheists and gays. Finally, both evangelical
religious affiliation and church attendance decrease Brazilians’ willingness to
extend civil liberties to groups they dislike.
Third, congregations boost participation in the many venues through which
citizens seek to shape public policy beyond elections. Evangelicals and Catholics
who attend religious services more frequently are more likely to contact their
representatives, to sign petitions, and to participate in local level efforts to
improve the community. Many forces within religious institutions foster
participation. Not only do clergy often explicitly encourage participation, but
congregational social networks provide the human resources necessary for
effective mobilization. Moreover, civil-society activities are often sponsored
within congregations, and provide a gateway to further mobilization. Protest
participation is a partial exception, however. While Catholic clergy often
encourage participation in contentious politics, evangelical clergy and
congregants tend to discourage it.
In Parts II and III of this book, we have seen how Brazil’s culture wars have
developed among clergy and citizens. Clergy, motivated by a combination of
ideas and organizational interests, polarize around a set of issues related to the
family, gender, sexuality, and the rights and responsibilities of churches; and
they sometimes get involved in partisan electoral politics. Citizens adopt some
of their attitudes on the same set of issues. Citizens take cues for electoral
behavior from their clergy, and more doctrinally conservative churches tend
to coalesce more strongly in their vote choices. Moreover, as we saw in this
chapter, religious polarization is not just a matter of abstract policy attitudes.
Rather, affective polarization extends to people’s attitudes toward their fellow
citizens and the political system. Nonetheless, secular norms lead citizens
partially to resist clergy influence and political activism, and they boost the
legitimacy of democracy and the political system. If ideological, electoral, and
social conflict are the symptoms of the Brazilian culture wars, secular norms –
even, or especially, the secular norms of highly religious citizens – are the
potential cure.
part iv

REPRESENTATION
9

The Representational Triangle

How do Brazil’s culture wars affect democratic representation? This chapter

brings in the third vertex of the “representational triangle” presented at the end
of Chapter 2: elected officials. It examines the relationships among citizens,
clergy, and the people they elect into office. In representative democratic
theories, citizens’ most important role is to select representatives, ideally
based on policy-related criteria, and to monitor those representatives outside
of elections. Elections should produce bodies of elected officials – that is,
legislatures and executives – who effectively represent the electorate in some
way. While representation might entail matching one’s constituents
demographically (“descriptive representation”), or helping them feel included
in the political system (“symbolic representation”), this chapter primarily
focuses on what Pitkin (1967) calls “substantive representation”:
representation of voters’ policy or ideological views.
Achieving substantive representation of the majority of voters has been
a persistent challenge facing the developing democracies of Huntington’s
(1991) “third wave.” In highly unequal contexts, elites enter the electoral
game with many advantages, and can become skilled in using institutions to
produce unrepresentative outcomes. Many forces and informal practices
facilitate elite control of the policy agenda: clientelism and the informal
distribution of other particularistic goods; low levels of education and
economic vulnerability; and the inadequacy of school-based civic education.
Do evangelical and Catholic elected officials effectively represent members of
their own religious groups? The empirical analysis begins by examining
legislators’ religious affiliations. While religious diversity is on the rise among
elected politicians, Catholics remain overrepresented, and all other religious
groups underrepresented. Thus, popular perceptions of an evangelical takeover
of politics are substantially overstated. Next, the chapter examines substantive
representation. We find that evangelical elected officials represent their
constituencies’ positions well on the dimension of family and sexual
traditionalism. Though quantitative data are lacking, evangelical elected

149
150 Representation

officials also appear to represent evangelicals’ desire for greater defense of in-
group religious prerogatives. However, electing in-group representatives
imposes policy costs on evangelical citizens as well. Outside of the two
aforementioned dimensions, evangelical legislators are not very good
representatives of their constituencies’ policy, ideological, or partisan
positions. In particular, elected politicians are substantially more
conservative, and more likely to be aligned with rightist parties, than are the
citizens who tend to vote for them.
What drives these partial representational failures? I argue that the answer
lies in large part in the ways religious communities build personal relations with
the politicians they tend to support for office. Religiously based politicians bring
new blood into office; they are much less likely to rely on oligarchic family ties to
get there, and they are more strongly tied to what they perceive as their
grassroots base. Moreover, they rely to a much lower degree on large
campaign donations (Netto 2016). Instead, personal ties appear to drive
evangelical politicians’ campaigns. Religious communities are one of the most
important places citizens come into contact with politicians, and the close
networks fostered in congregations enable closer relations between politicians
and their voters.
The personalization of evangelical campaigns has many positive benefits.
As a result of their personal ties, religious citizens are more likely to contact
politicians to address community or personal problems. Moreover, evangelical
politicians’ lower reliance on dynastic family ties and on large campaign
donations could both democratize the candidate recruitment process and keep
it cleaner. Nonetheless, this personalization also has negative effects. In church,
evangelicals and Pentecostals are more likely to be targeted with clientelistic
offers. Personal relations may also undermine citizens’ attention to politicians’
policy positions, particularly on issues of lower relevance within the religious
community. Furthermore, in the congregation–politician relationship, clergy
often have an informational advantage over citizens. This advantage leads
congregations to center their mobilization on politicians who align more
closely with the (conservative) views of religious leaders than the (relatively
progressive) views of religious followers.

two scenes
1 On April 17, 2016, the Chamber of Deputy’s roll-call vote to impeach
then-president Dilma Rousseff lasted six hours, as each deputy spoke in turn
before casting a vote. The first “yes” vote was cast by an evangelical,
Deputy Washington Reis from the Party of the Brazilian Democratic
Movement (PMDB), who hoped that God would “pour out blessings on
this nation” (Chagas 2016b). A little while later, evangelical Deputy
Eduardo Bolsonaro, wrapped in the flag of his state of São Paulo, declared
his vote: “For the people of São Paulo in the streets . . . for the military in
The Representational Triangle 151

1964 [the date of the military coup], today and always, for the police, in the
name of God and the Brazilian family, Yes! And Lula and Dilma in jail!”
(Chagas 2016b). Eduardo Cunha, a member of the Assembly of God and
president of the Chamber of Deputies, likewise voted yes. The diverse
dedications for the “yes” vote, ranging from “for Peace in Jerusalem” to
“for the Masons,” became the stuff of satire, with one humorous newspaper
observing, “After being cited by every pro-impeachment deputy, God will be
investigated by the Public Ministry” – a reference to the fact that many
deputies voting yes were themselves being interviewed by public prosecutors
in connection with the Operation Car Wash scandal (Carrapatoso 2016;
Zorzanelli 2016). With a qualified majority of 2/3 required, the impeach-
ment would have narrowly passed to the Senate if only non-evangelicals had
voted; 289 non-evangelicals voted in favor of impeachment, and 134
against. However, evangelicals gave the “yes” vote a comfortable margin.
Of the Chamber’s eighty-one evangelical members, 93 percent voted in
favor of the impeachment, following the decision of their caucus leadership
as well as the Universal Church of the Kingdom of God (UCKG)-linked
Brazilian Republican Party (PRB) (Chagas 2016a; Galindo 2016; Redação
Pragmatismo 2016). Immediately after the Chamber of Deputies had voted
to impeach then-president Dilma Rousseff, Pastor Marco Feliciano, an
Assembly of God pastor and federal deputy from the state of São Paulo,
posted to Twitter, “VICTORY FOR BRAZIL! BYE-BYE, DARLING!”1
2 In October 2016, UCKG Bishop Marcelo Crivella, nephew of founder
Edir Macedo, was elected mayor of Rio de Janeiro. This victory was widely
described as a sign of the growing political power of evangelicals. While
evangelicals have been increasingly successful in legislative races relying on
open-list proportional representation, winning executive races has been
harder. Yet in an interview, anthropologist Ronaldo Almeida cautioned
against making too much of this triumph. Crivella, he explained,
would now:
have to negotiate with the City Council, where the UCKG does not have a strong
base. Crivella may have an electoral base, but the political base depends on alliances
with other politicians, evangelicals or not . . . [I]t is an exaggeration to say that with
the election of Crivella, the UCKG will rule Rio. Society is pluralistic and the UCKG
is only a part of the entire evangelical sector. Most evangelicals are uncomfortable
with the UCKG . . . That’s why he was constantly working to distance himself from
the brand, while still keeping a discourse connected to the UCKG. It is very subtle.
(Charleaux 2016)

1
“Bye-bye, darling” became a pro-impeachment slogan during the Chamber and Senate votes.
The expression came from a quote from former president Lula in a wire-tapped, recorded phone
call with then-president Dilma.
152 Representation

the religious ties of brazilian politicians

Before considering the links between religion and representation, it will help to
understand the distribution of religious affiliations among politicians. When
observers discuss politicians’ religious ties in Brazil, they refer both to
politicians’ own personal religious identification, and to their political links to
religious groups. For instance, Federal Deputy Celso Russomano has been
associated with the UCKG – he was even accused of improperly diverting
federal funds to the church – despite being a Catholic himself. Of the eighty-
nine evangelical deputies whom M. N. Cunha (2016a) identified in the Congress
installed in 2015, only seventy-two were official signatories of the Evangelical
Parliamentary Front (EPF), known informally as the “evangelical caucus”
(bancada evangélica).2 Though most of the evangelicals absent from the
official rolls were conservatives who voted with the caucus, four of the six
leftist evangelicals did not register with that caucus.3 Furthermore, the
officially registered caucus includes many non-evangelical sympathizers, with
a strong presence on the part of the Charismatic Catholic Renovation
(M. N. Cunha 2016a). Registering with the evangelical caucus is evidently
a political statement as much as a religious one. Even more strikingly, the
recently established Parliamentary Front for Defense of Traditional Peoples of
African Origin, dedicated to defending Afro-Brazilian religions against
discrimination by evangelicals, has 202 signatories – almost the same number
as the membership of the Roman Catholic Apostolic Parliamentary Front (208
members) (M. N. Cunha 2016b). This is despite the fact that there are likely at
least a hundred times as many Roman Catholics as practitioners of Afro-
Brazilian religions in the National Congress. Clearly, a legislator’s choice of
religious caucus is not a straightforward indicator of religious affiliation. Here,
though, I focus on politicians’ own personal religious affiliations, when known.
Gathering information on legislators’ religious affiliations requires
triangulation across various sources. Using non-publicly available data from
the intake questionnaires of new legislators from 1988 through 2014,
Mucinhato and coauthors find that the proportion of representatives who are
Catholic is steadily dropping, and the proportion who are evangelical is rising;
yet Catholics remain overrepresented and evangelicals underrepresented
(Mucinhato 2014; Simoni Junior, Mucinhato, and Mingardi 2015, 2016).
A survey of 421 of the 513 deputies elected to the 2015–2019 Congress found

2
This counts only those whom Cunha reported were actually holding office at that moment, and
who were not on leave (that is, including suplentes and excluding licenciados). The list of official
signatories can be found at: http://www.camara.leg.br/internet/deputado/frenteDetalhe.asp?
id=53658.
3
The leftist evangelicals not registered with the caucus include Chico Alencar (Partido do
Socialismo e Liberade), Fabiano Horta and Rejane Dias (PT), and Humberto de Lucena
(Partido Verde). The leftist evangelicals who are registered with the EPF are Benedita da Silva
(PT) and João Derly (Rede Sustentabilidade).
The Representational Triangle 153

that 71 percent self-identified as Catholic, 16 percent evangelical, 3 percent

“Christian” (unspecified), 5 percent with no religion, and 2 percent Spiritist,
while one deputy (0.2 percent) identified as Jewish, and two (0.5 percent) with
other religions (G1 2015). The dominance of Catholicism is more pronounced
in the Senate where just 4 percent of senators elected to the 2015–2019 National
Congress are evangelical (M. N. Cunha 2016a). Not only are evangelicals
underrepresented, but they are less likely to hold leadership roles in the
Chamber (Simoni Junior, Mucinhato, and Mingardi 2015). The nonreligious
are also underrepresented. Given general suspicion of atheists (see previous
chapter), open lack of religiosity is likely to hurt campaigns. For instance,
former president Fernando Henrique Cardoso was dogged by allegations of
atheism (ones he denied) that were widely believed to have led to his loss of the
mayoral race in São Paulo in 1985 (Folha de São Paulo 1994a, 1994b).
The case of far-rightist Jair Bolsonaro – who was a federal deputy from 1991
to 2018 and elected president in October 2018 – exemplifies the complexity of
religious affiliation. Bolsonaro has long been affiliated with the evangelical
caucus, yet he publicly identifies as Catholic. In 2011, he told reporters that
“I’m a Catholic who has attended a Baptist church for the past ten years”
(Guiame 2011). In 2016, he was baptized in the River Jordan by Assembly of
God pastor and 2014 presidential candidate Everaldo Dias Pereira, which was
widely interpreted as a public conversion to evangelicalism (Aragão 2016).
However, Bolsonaro did not publicly announce a change in his religious
affiliation (M. N. Cunha 2016c), nor did he make clear which evangelical
church he held allegiance to; at least one commentator speculated that this
ambiguity was politically strategic (Stahlhoefer 2016). Given this book’s
conceptualization of religious affiliation in terms of personal identification,
Bolsonaro is coded as Catholic in the quantitative data analyzed here.

ideology, policy attitudes, and partisanship in the

representational triangle
How does religious affiliation shape elected officials’ attitudes? Figure 9.1
examines legislators’ attitudes on a range of policy issues similar to those
examined for clergy and citizens (see Chapters 4 and 6). The data come from
the Brazilian Legislative Surveys (BLS), which have surveyed 1,146 deputies
from seven legislatures since the return to democratization (Power and Zucco
2012). Legislators’ religious affiliations and demographic characteristics are
taken from a variety of sources.4 (See Online Appendix C for further detail.)
All policy attitudes are coded so that higher values represent more liberal/

4
Religious affiliation comes primarily from data supplied by Mucinhato and coauthors
(Mucinhato 2014; Simoni Junior, Mucinhato, and Mingardi 2015, 2016). Where those data are
missing, I use Agência DIAP (2010), Gonçalves (2011), and Cunha (M. N. Cunha 2016a).
154 Representation

1
Legislator Support for Policy
.75
.5
.25
0

Abortion Same-Sex State Spending Affirmative Environment Ideological

Marriage Economic for Poor Action Identification
Role (Race-Based) (Rightist)

Catholic Pentecostal/Evangelical
None Other

figure 9.1 Policy views of federal legislators, by religion

progressive positions, with the exception of left–right ideological identification,

where higher values indicate that the deputy is to the right.
Results for some issues look familiar. Similarly to their coreligionists among
citizens and clergy, evangelical legislators hold distinctively conservative views
on homosexuality and abortion, while Catholics occupy a middle position in
between evangelicals and the nonreligious. However, the gaps between
legislators of different religious groups are much larger than among citizens.
Yet on other issues, we find unfamiliar patterns. Evangelical legislators differ
from lay evangelicals in their tendency toward conservatism on several issues
beyond the dimension of sexual-family traditionalism. Evangelicals and
Catholics are both somewhat more conservative than the nonreligious in their
responses on two economic questions. More strikingly, evangelical legislators
are quite conservative on race-based affirmative action. There is little evidence of
a correlation between racial conservatism and evangelicalism among citizens;
moreover, in the 2010–2014 AmericasBarometer survey, lay evangelicals
are slightly darker-skinned than other citizens, even after controlling for various
measures of social status and region. However, anecdotes of racial conservatism
among evangelical legislators abound. As just one example, in April 2017, federal
deputy Jair Bolsonaro, the evangelical semi-convert and future president,
announced that if elected president, he would eliminate reservations for
indigenous communities and quilombolas, communities of escaped slaves
(Veja 2017). Unlike at the citizen level, legislators’ economic and racial attitudes
factor together in a single ideological dimension with attitudes related to sexual
and family traditionalism. On this dimension, evangelical legislators are
The Representational Triangle 155

significantly more conservative than Catholics, while the nonreligious are

significantly more liberal than Catholics.5
Given their issue attitudes, it is not surprising that evangelical politicians tend
to come from centrist or rightist parties. A number of small, rightist parties have
become vehicles for evangelical candidates. The most prominent are the Christian
Social Party (PSC), home of the 2014 candidate Pastor Everaldo Dias; the
Brazilian Republican Party (PRB); and, as of 2018, Jair Bolsonaro's Social
Liberal Party (PSL) (Power and Rodrigues-Silveira 2018). Evangelicals also
often run under the banner of larger parties, including the PMDB (Party of the
Brazilian Democratic Movement), the Demócratas, and the Social Democratic
Party (PSD). Yet among the eighty-nine evangelical federal deputies identified by
Cunha (2016a) who were holding office in October 2016, six were from leftist
parties. Once again, this differs from the findings at the citizen level, where there
are few religious differences in partisanship.
Evangelical legislators are likely to be strong representatives of their
constituents on one more issue: policies toward religious groups. As discussed
in Chapter 1, evangelical legislators often favor corporatist privileges: from
favorable tax policy, to laws facilitating evangelization. Likewise, we saw in
Chapter 5 that evangelical citizens are much more likely to favor legislative
privileges for Christianity than are Catholics, members of other religions, or the
nonreligious. While we lack quantitative measures of legislator views,
qualitative evidence suggests a close correspondence between evangelical
legislators’ and citizens’ preferences.
Thus, evangelical legislators are highly representative of their constituents on
abortion and same-sex marriage, and probably on church–state relations. They
are quite unrepresentative on economic and race-related policy. Where do their
divergent views come from, and how do they persist?
Accounting for the third vertex of the representational triangle – clergy –
could help to explain this asymmetry. Evangelical legislators’ attitudes match
those of the evangelical elites who serve as electoral brokers, I argue, rather than
those of lay evangelicals. Figure 9.2 presents the differences in policy attitudes
between evangelicals and Catholics on five key issues, for clergy, citizens, and
legislators. For each attitude, responses among Catholics are standardized – the
mean Catholic response is set to 0, and values of -1.0 and 1.0 represent
Catholics’ responses a standard deviation above and below the mean. Thus, if
evangelicals’ mean on a given variable is a -0.5, the average evangelical is 0.5
standard deviations below the Catholic mean. Evangelical clergy are

5
Among all legislators, attitudes toward abortion, same-sex marriage, spending for the poor, and
affirmative action load highly on a single dimension with an eigenvalue of 1.13 (unrotated factor
analysis). Among evangelical legislators alone, attitudes toward abortion, same-sex marriage, and
state economic policy load highly on a single dimension, with an eigenvalue of 1.82. However, the
evangelical-only factor analysis is based on an extremely small number of cases, because the
seventh wave of the BLS is the only wave for which all variables are present.
156 Representation

1
Evangelical Divergence from Catholic
Positions (Standard Deviations)
.5
0
–.5
–1
–1.5

Abortion Homosexuality/ Ministry to Anti-Racism/ Environ-

Same-Sex Poor/Social Affirmative mental
Marriage Spending Action Protection

Clergy Citizens Legislators

figure 9.2 Differences between Catholics and evangelicals in policy attitudes, for
clergy, citizens, and legislators

significantly more conservative than are Catholic clergy on all five issues.
By contrast, evangelical citizens are slightly to the left of Catholics, though
not statistically significantly so, on economics, race, and the environment.
Evangelical legislators resemble clergy to a greater extent than they resemble
citizens on social spending and race-related policy. Only on the environment do
evangelical legislators appear closer to citizens than to clergy. These patterns
indicate that clergy and candidates together mobilize citizens on the basis of
abortion and same-sex marriage, yet elected officials deviate from their base on
other issues.

what explains asymmetries in the representational

triangle?
Why do evangelical politicians’ attitudes match those of clergy to a greater extent
than they match those of citizens? At least four possible mechanisms could help
explain the proximity of clergy and legislator views: the first two involve one actor
imparting its views to the other, while the third and fourth involve different
selection mechanisms. One potential explanation entails direct clergy influence:
politicians might deliberately take cues from clergy, whom they recognize as
gatekeepers to evangelical votes. Alternatively, the arrow of influence might go in
the opposite direction: clergy could take policy cues from in-group politicians, at
The Representational Triangle 157

least on issues such as economic policy. Third, clergy might hand-pick candidates –
often members of their own religious communities – who they know share their
policy priorities. Fourth, evangelicals who choose to run for office – or at least those
who win elections – may simply be more reliably conservative. Adjudicating among
these four explanations using statistical methods is beyond the scope of what it is
possible to do with the present data. However, knowledge of the Brazilian case
leads to the conclusion that all four mechanisms likely sometimes occur, but the
two selection mechanisms play an outsized role.
In Brazil’s electoral and party systems, clergy have great latitude to select
compatible candidates. Under the rules of open-list proportional
representation, used for legislative elections to the lower chamber (in federal
elections) or only chamber (in local and state elections), parties field slates of
candidates, with candidates unranked within slates. Citizens can vote for either
candidates or parties, but most citizens choose to vote for candidates; candidate
votes are aggregated at the party level to determine the number of seats each
party wins. The candidate-level votes also serve to rank candidates within party
lists, determining which candidates win the seats allocated to each party. Thus,
the spare votes of a candidate who does extremely well – that is, the votes over
what would be necessary for that candidate to win by himself or herself – can
boost the chances of other candidates on his or her party list. Likewise,
moderately popular but unelected candidates help their parties win seats
because their votes are added to the party total. Hence, parties are eager to
field candidates who promise to draw votes. Moreover, in most local, state, and
federal legislative districts, dozens of parties seek to fill candidate slates, yielding
hundreds of candidates in each race. As a result, parties are more than willing to
accept candidates with credible support from clergy. This mechanism may
explain how the UCKG hierarchy or some Assembly of God pastors field
candidates, strategically choosing the parties whose lists they believe give their
candidates the best chances of attaining office (Conrado 2001; Dantas 2011).
In other circumstances, the fourth mechanism may provide a better explanation.
Though recruiting candidates from within the religious group’s own ranks is
common in some denominations, a model of candidate recruitment in which each
congregation selects its own candidate is unsustainable. Few congregations, or even
groups of affiliated congregations, would be able to muster enough voters to get
their own candidates elected, and many would lack a high quality pool of potential
candidates. In the UCKG, the denominational hierarchy solves these problems by
coordinating candidacies across congregations. Often, though, congregations
support evangelical candidates from outside their immediate religious group.
The local politician Noraldino Lúcio Dias Junior from Juiz de Fora is
prototypical. Noraldino was elected to the Minas Gerais state legislature in 2014.
Prior to that he was a member of the city council, and in 2016 he ran a strong but
ultimately unsuccessful campaign for mayor of Juiz de Fora. Noraldino is himself
a member of a Methodist congregation. Though Methodist clergy neither recruit
candidates to run nor openly campaign from the pulpit, Noraldino relies on
158 Representation

networks among congregants themselves in his own and other Methodist

congregations across the city for high levels of Methodist support. Success breeds
success. Noraldino leverages his popularity to seek the blessing of other evangelical
clergy. Thus, in his entrepreneurial quest for support, Noraldino helps clergy solve
what might otherwise constitute a coordination problem; he enables clergy to build
ties with a likely winner who will promote their positions on key issues.
If these two selection mechanisms help to explain the correspondence
between evangelical clergy and legislator views, we should see differences in
how evangelicals and non-evangelicals campaign. In the BLS, legislators were
asked to explain their perceived support base. Legislators in all groups tend to
attribute their electoral success to grassroots supporters, rather than to
traditional and family ties, or their party (see Figure 9.3). However, this
tendency is most pronounced among evangelicals. Evangelical legislators are
much less likely than Catholics to believe that family ties or traditional
oligarchic groups helped get them elected. While 49.6 percent of Catholic
legislators say they have a family member who previously held office, only
38.9 percent of evangelicals, 27.1 percent of the nonreligious, and
30.4 percent of those of other religions do so. In addition, evangelicals are the
least party-oriented of all religious groups. Moreover, Netto (2016) shows that
evangelical legislators rely to a much lesser extent on large campaign donations.
What are evangelical legislators’ grassroots groups? Congregations regularly
bring citizens into contact with politicians. In 2012, the LAPOP
AmericasBarometer asked Brazilians if they knew anyone who had run for
office, or who was an elected official. There were large religious differences in
the rate of knowing politicians: 45 percent of Catholics and those of other
religions, and 46 percent of evangelicals, said they personally knew
a politician, while only 33 percent of those without a religious affiliation did
so. However, a multivariate analysis shows that church attendance explains the
effect of religious affiliation. Though the probability of knowing a politician is
only .34 for those who never attend church, it is .54 for those who attend church
more than once a week. The impact of church attendance is similar in
magnitude to that of belonging to a community association, or of going from
the lowest to the highest level of political interest. Thus, congregational life
often provides both Catholics and evangelicals personal ties to the political
world.
Knowing a politician personally can affect citizens’ political behavior.
Citizens with political ties more readily communicate their policy views to
people in positions of power. The evangelical campaign against Juiz de Fora’s
Miss Gay parade described in the Chapter 1 is one example of how such behind-
the-scenes activism can work. People who know politicians personally also have
better access to them if they wish to request personal help – for instance, if they
need help locating a space in a public hospital, or enrolling a child in school.
Thus, it is not surprising that analysis of the AmericasBarometer shows that
citizens who attend church more regularly are also more likely to contact local,
The Representational Triangle 159

1
Priority Given Each Electoral Base
.75
.5
.25
0

Family/Tradition Party Grassroots Groups

Catholic Legislators Pentecostal/

Evangelical Legislators
Non-Religious Legislators Other Religion

figure 9.3 The perceived electoral bases of federal legislators, by religion

state, and federal officials for policy-related or personal concerns. Religious

differences in contacting are largely explained by differences in the rates of
knowing politicians.
At the same time, personal connections to politicians forged in church could
also become a vehicle for clientelistic offers. In 2010, the AmericasBarometer
asked respondents if a politician had offered them favors or goods in exchange
for their votes. For confidentiality reasons, the survey did not ask whether the
deal was closed; we do not know the level of actual clientelism. Nonetheless, the
data provide a sense of the level of potential clientelism. Evangelicals reported
slightly higher rates of exposure to clientelistic offers: 20.0 percent, as opposed
to 15.0 percent among Catholics, 16.9 percent for the nonreligious, and
9.8 percent for those of other religions. The differences across religious
groups were not statistically significant. However, only among evangelicals
was church attendance significantly associated with receiving clientelistic
offers. The probability of receiving such offers was 26 percent for evangelicals
who attended church more than once a week, compared to 14 percent for
evangelicals who never attended church. The significant effect of church
attendance among evangelicals holds even after controlling for a host of
demographic variables and other aspects of political and social participation.
160 Representation

The ramifications for Brazilian representative democracy are profound.

The data do not clarify who within evangelical churches might make
clientelistic offers – whether pastors, lay campaign workers or electoral
brokers, or candidates themselves.6 Recent research shows that brokers often
target clientelistic offers to “their own” voters, aiming not so much to persuade
swing voters as to mobilize those who have already been persuaded (Gans-
Morse, Mazzuca, and Nichter 2014; Nichter 2008; Stokes et al. 2013). This
would be consistent with the possibility that clientelistic offers in evangelical
congregations are made by or on behalf of evangelical politicians. Nonetheless,
even if clientelistic trades occur between coreligionists, they could undermine
substantive representation, as citizens who receive material payoffs are less
likely to be able to make policy demands, or to hold politicians to account for
their policy commitments (Kitschelt 2000; Stokes et al. 2013).
These patterns also affect political institutions. The evangelical style of
campaigning contributes to the much-lamented fragmentation of the Brazilian
party system. As we have seen, evangelical politicians give very low priority to
political parties as sources of electoral support. This is largely because
evangelical congregations effectively serve as alternative centers of electoral
power. Evangelical leaders can nominate candidates, or – as in the example of
Noraldino – coordinate directly with candidates to maximize those leaders’
own political leverage.
In principle, coalitions of evangelical leaders could coordinate to form
unified evangelical parties. Solidifying evangelicals as a voting bloc supporting
one or perhaps a few parties would especially help evangelical candidates win
executive races, which require larger coalitions of voters than do legislative
ones, and hence would likely advance evangelicals’ policy goals. However,
evangelical leaders are not only motivated by such goals. Heads of religious
institutions must also advance their own institutional objectives to gather souls
to their groups. Those institutional incentives undermine the incentives for
policy-related electoral coordination. In denominations such as the UCKG
and the Assembly of God, evangelical leaders use the tools of electoral politics
not just to promote their issue priorities, but also to promote their own religious
brands. Sublimating the denominational identity to a broader evangelical one
could ultimately hurt each religious leader’s long-term institutional objectives.

summary
How have Brazil’s clergy-driven culture wars shaped representative democracy?
Evangelical legislators dramatically expand the political power of evangelical
clergy and congregations. Not only do evangelical legislators respond to the

6
Unfortunately, I also lack information from the qualitative research on this issue. I have not
observed clientelistic offers in evangelical or Catholic congregations.
The Representational Triangle 161

same church leaders as lay evangelicals, but they also become a receptive ear for
evangelical social movements and legislative activism.
There are many ways in which the cultural struggles stemming from the rapid
entry of religious groups into electoral politics have improved representation,
particularly on the political right, where substantive representation has
historically been poor in Brazil. Legislators elected with the support of
evangelical congregations match their constituents closely in their views on
key issues, especially abortion, same-sex marriage, and policy toward
evangelical churches. Since non-evangelical legislators have tended to be to
the left of non-evangelical citizens on these issues, evangelical legislators
improve congruence between citizens and elites as a whole on same-sex
marriage and abortion (Boas and Smith in press). Moreover, the evangelical
style of campaigning provides many evangelicals with personal ties to the
political world that they would otherwise lack. More generally, churches
provide a venue for citizens of all religious backgrounds to meet politicians.
These political ties enable citizens to contact politicians to express their policy
views and request personal help when needed.
However, evangelical legislators diverge from the interests of evangelical
citizens on the economy and race. The racial conservatism of evangelical
legislators is particularly notable. On these issues, evangelical politicians’
views resemble those of in-group clergy more closely than those of voters.
I argue that this divergence is due to the asymmetric influence of clergy over
politicians. Clergy constitute one kind of electoral broker, but with an unusually
high level of independent power; this autonomy gives clergy unusual leverage in
capturing policy rents. Evangelical legislators’ divergence from their
constituents’ views on these issues may also be facilitated by clientelism.
The evangelical style of campaigning also exacerbates party fragmentation,
I argue. As noted in Chapter 1, this argument illuminates old debates over the
causes of party fragmentation – whether institutional and incentive-based, or
sociological and cleavage-based (Duverger 1972; Sartori 1976). The argument
presented here brings the two sides together by suggesting that sociological
cleavages can sometimes create institutional incentives that foster
fragmentation.
10

Conclusion: Mobilizing the People of God

On January 9, 2018, the Inter-American Court of Human Rights (ICHR) ruled

that, per the American Convention on Human Rights, Costa Rica must legalize
same-sex marriage – a ruling that the court held applied to all twenty-two
signatory countries in the region. Costa Rica has long been perceived as one
of the most stably democratic countries in the region, with high citizen support
for democracy and the political system. Nonetheless, the ruling was unpopular
among Costa Ricans, most of whom oppose same-sex marriage. It upended the
campaign for the February 4 first-round presidential election. Almost
instantaneously, the evangelical preacher Fabricio Alvarado from the tiny
National Restoration Party began a meteoric ascent in the polls, rising from
somewhere between 3 percent and 5 percent of vote intentions in the week of the
ICHR ruling to take first place with 25 percent of the vote three weeks later
(BBC News 2018). In the process, he crowded out the centrist candidate
Antonio Álvarez, from one of the largest and oldest parties in Costa Rica,
who had previously been polling in first place. In the April 1 second round,
Fabricio Alvarado faced off against Carlos Alvarado (no relation), from the
small, relatively new center-leftist Citizens’ Action Party – a contest the center-
leftist handily beat. Thus, the ICHR ruling was transformational: it polarized
the presidential election, boosted the profile of the religious right, and may have
dealt the death blow to Costa Rica’s ailing traditional party system. The impacts
promise to extend throughout Latin America. As gay-rights advocates prepare
to bring domestic court cases on the basis of the ICHR ruling in countries such
as Panama, Guatemala, and El Salvador, the religious right will certainly begin
to mobilize as well (The Economist 2018).
Across Latin America, religious demographics are changing, following
patterns that parallel shifts found in Brazil. Analyzing regional trends in
the AmericasBarometer survey between 2004 and 2012, Boas and Smith
find that “the Catholic Church has lost a significant share of identifiers in
every country but Argentina, El Salvador, and Venezuela” (2015, 101).
Between 2010 and 2016/2017, new analysis indicates that the percentage of

162
Conclusion: Mobilizing the People of God 163

the population identifying as Catholic in the AmericasBarometer dropped from

60 percent to 49 percent; the majority of Catholicism’s loss contributed to swell
the ranks of evangelicalism and Pentecostalism. The percentage identifying as
evangelical/Pentecostal in the 2016/2017 round of the AmericasBarometer
ranged from 43 percent in Guatemala to 9 percent in neighboring Mexico.
Meanwhile, the percentage reporting that they did not have a religion ranged
from 39 percent in Uruguay to 2 percent in Guatemala and Paraguay.
We also find many instances of evangelicals flexing political muscle.
Guatemala, with the highest percentage of evangelicals in the region, has now
had three evangelical presidents. Most prominent among them is the strongman
General Efraín Ríos Montt, whose seventeen-month administration prosecuted
a genocidal war against highland indigenous villages in 1982 and 1983. Under
democracy, Guatemala’s evangelical presidents include Jorge Serrano Elías
(1991–1993), and the entertainer Jimmy Morales (who took office in 2016).
In Colombia, evangelical mobilization is widely credited with derailing the
government’s 2016 referendum to approve a peace deal negotiated with the
FARC rebel group. Though the referendum itself did not touch on matters of
gender or sexuality, rightist political leaders were able to link the peace deal in
the public imagination with the April 2016 approval of same-sex marriage, and
with the government’s efforts to support gay and transgender students in public
schools. And across Latin America, we find a growing number of evangelical
parties: for instance, Visión con Valores (VIVA) in Guatemala or the Costa
Rican Renovation Party, National Christian Alliance, and National
Restoration Party.
It might seem obvious that the former set of changes – demographic shifts –
trigger the latter – the rising incorporation of evangelicals in politics.
Nonetheless, this connection is less clear-cut than one might assume.
Evangelical presence in the population does not necessarily translate into
evangelical might in the voting booths, much less evangelical representation in
the halls of power. Chile dramatically illustrates the sometime disconnect
between evangelical conversion and evangelical in-group voting. In 2016, the
AmericasBarometer registered that 21 percent of Chileans identified as
evangelical – a figure not far from the evangelical presence in Brazil in 2010.
Yet Boas notes no evangelical party has formed in Chile, and that since its
democratic transition in 1990, “Chile has had, on average, only two evangelical
representatives in Congress,” while few evangelicals run for office.1 At the mass
level, Chilean evangelicals are largely disengaged from the political sphere
(Fediakova 2012). Comparing Chile to Brazil, Boas notes that,

1
Taken from Taylor Boas, “Expanding the Public Square: Evangelicals and Electoral Politics in
Latin America,” 2018, 3. Prepared for The Inclusionary Turn in Contemporary Latin America,
editors Diana Kapiszewski, Steven Levitsky, and Deborah Yashar (unpublished).
164 Representation

Like their Brazilian counterparts, Chile’s evangelicals have gained inclusion . . . Yet
inclusion has taken very different forms in the two countries. In Brazil, evangelicals
have sought and achieved influence within the halls of power, whereas in Chile, they have
remained primarily on the sidelines.2

The case of Costa Rica constitutes yet another puzzle. Though evangelicals
and Pentecostals constituted 26 percent of respondents in the 2016
AmericasBarometer, evangelical candidates have historically received only
a small fraction of that vote. In fact, one observer marveled that legislative
candidates from evangelical parties had received 8 percent of the 2016 vote –
admittedly, a dramatic rise from 3 percent in the 1998 elections (Salazar 2017).
It is also worth remembering that until the surprise event of the ICHR ruling,
fewer than one in twenty likely voters in Costa Rica intended to support the
evangelical candidate.
Thus, thinking about the broader Latin American context brings the
arguments driving this book into new relief. In this chapter, we revisit the
book’s central questions. First, what explains religious groups’ entry into
politics and the polarization of politics along religious lines? Second, what are
the consequences for Latin American democracy?

toward an explanation of the culture wars in latin

america
Why the sudden success of evangelical politicians in Costa Rica? Or to turn the
question on its head, what explains evangelicals’ mediocre performance in
Costa Rica until the turning point of January 9, 2018? More generally, why
do evangelicals enter politics in some times and places but not others?
In broad terms, there are two possible answers. Evangelicals fail to enter
politics because either (a) they don’t want to; or (b) they can’t. To put it another
way, religious groups enter politics when they have both the motivation and the
opportunity and resources to do so. All three – motivation, opportunity, and
resources – are necessary; absence of any one prevents a group from entering
politics.
Two types of threats appear to motivate candidates with religious linkages to
run for office, and also to motivate voter support for religiously linked
candidates and causes (as in, for instance, the case of the Colombian
referendum). First, actual or threatened liberalization of policies related to
sexual and family traditionalism often precipitates vehement religious
mobilization. In Latin America, same-sex marriage appears to have been
a particularly powerful trigger, perhaps because public policy has moved
leftward much more quickly on this issue than on other touchy issues such as
abortion. Thus, we have seen that legalization of same-sex marriage in Brazil,

2
Boas, “Expanding the Public Square,” 2–3.
Conclusion: Mobilizing the People of God 165

Costa Rica, and Colombia provoked a conservative backlash in each country.

Corrales argues that evangelicals and Pentecostals have been the most
important group resisting LGBT rights in Latin America (2017). Yet abortion
has at times also stimulated mass mobilization among both evangelicals and
Catholics (e.g., Heumann and Duyvendak 2015 on Nicaragua).
Second, perceived threats to the religious in-group also motivate political
action. Such threats sometimes come from the state. For instance, in
“Expanding the Public Square,” Boas argues that the careful separation of
church and state in Chile largely explains the puzzle identified above – the
dramatic political underperformance of evangelicals, especially relative to
their coreligionists in Brazil. In Brazil, following separation of church and
state in the nineteenth century, the Catholic Church gradually recovered
many privileges in the early twentieth century, which became sources of
grievance for evangelical congregations seeking to expand. Brazilian
evangelical mobilization in the 1980s was oriented precisely to dismantle
those privileges, and was ultimately successful. By contrast, in “Expanding
the Public Square,” Boas argues that Chilean evangelicals historically failed to
enter politics simply because they saw no need to do so, as the state did not
threaten their rights and privileges.
However, 2017 did see a wave of evangelical activism in Chile for the first
time. Though Catholic politicians were historically highly conservative on
matters of sexual and family traditionalism, Chilean public policy began to
move leftward on abortion and same-sex marriage in the 2010s. Same-sex civil
unions were legalized in 2015, and limited abortion rights approved by congress
in 2017. In “Expanding the Public Square,” Boas argues that these policy
threats triggered evangelical activism, yet the staunch conservatism of the
Chilean Catholic Church limited evangelical success. Many evangelical voters
essentially delegated representation to Catholic conservative politicians whom
they believed represented them substantively (i.e., in policy views), if not
descriptively (i.e., demographically).
But a purely state-centered approach misses some of the conditions that can
motivate political action. Threats from religious out-groups also often motivate
religious groups to get involved in politics. For instance, as discussed in
Chapter 2, both Trejo (2009, 2014) and Gill (1998) argue that rising
competition from Protestants triggered and shaped the course of Catholic
political engagement. And in Chapters 4 and 5, we found that threats to the
membership base led Catholic clergy to pivot toward the left on many issues,
likely seeking to appeal to more secular voters, and also to increase their
engagement in elections. Meanwhile, reminders of interreligious threat led
evangelical clergy to increasingly emphasize the importance of lobbying
elected officials.
Religious groups also need the opportunity to engage in politics. Most
basically, secularism and democracy facilitate political engagement, providing
the freedoms of conscience, speech, and association necessary for mobilization.
166 Representation

Though this point may seem obvious, it is worth remembering in the Latin
American context, where evangelical and Pentecostal activism has largely been
a phenomenon of “third-wave” (i.e., post-1980) democracies. If Latin America
enters a new era in which liberal democratic freedoms erode, the recession of
civil liberties could affect religious groups’ approaches to politics.
Beyond the impact of regime type, the electoral and party systems can also
facilitate or hinder groups’ entry into politics. In Brazil, both have been
exceptionally porous, guaranteeing many opportunities for evangelical
engagement in electoral politics. The combination of extreme multipartism
and high-magnitude electoral districts (that is, districts in which many
representatives are elected at once) has led party leaders under pressure to fill
out their slates on ballots to accept new (evangelical) blood eagerly. In addition,
Brazil’s form of open-list proportional representation in high-magnitude
districts enables candidates to target their mobilization efforts on specific
electoral corrals such as congregations (Ames 2001). The significance of these
rules is evidenced by the fact that evangelicals have had a much harder time
getting candidates elected to executive office or the Senate.
Finally, religious groups need resources: perhaps most importantly, human
ones. Demographics are not destiny, but they are correlated with groups’
fortunes. In countries with larger evangelical populations, evangelicals are
more likely to succeed in electoral politics, and successfully to mobilize
resistance to policy changes such as the promotion of LGBT rights (Corrales
2017; Mora Torres 2010). High levels of religious attendance and devotion
among evangelicals and Pentecostals increase the value of these human
resources. Resources can also be ideational and psychological. When
religiously based candidates and clergy campaign, they draw on shared
language, theology, religious symbols, and in-group identities to mobilize in-
group turnout and activism (Albertson 2014; Calfano and Djupe 2009; Chapp
2012; Oro 2003b, 2006).
This discussion highlights at least three important questions that merit future
reflection, theorizing, and investigation. First, why are religious groups so
readily motivated by these particular issues: policy related to family and
sexual traditionalism, and in-group threats? What is it about abortion and
homosexuality that makes them the most common triggers for culture wars
across the globe? One possible answer could be that these issues uniquely
violate what many evangelicals and Pentecostals see as the core of their
religious tradition. In Chapter 1, we saw that moral asceticism and frequent
evangelism are defining characteristics of these religious traditions. Abortion
and homosexuality violate dictates of moral asceticism; encroachment by the
state or other religious groups threatens the dictate to evangelize. Moreover, the
tendency toward a dualistic, good-versus-evil worldview in evangelicalism and
Pentecostalism may make adherents more susceptible to mobilization when
moral asceticism and evangelism are threatened. This is just one possibility;
answers should draw on qualitative and quantitative research in a variety of
Conclusion: Mobilizing the People of God 167

disciplinary traditions, including political psychology, religious studies, and

political behavior.
Second, how do religious groups come to perceive and frame threats
either from the state or other religious groups? Scholars can measure
religious freedoms, levels of secularism, and state encroachment on
religious groups’ rights with a reasonable degree of objectivity.
Nonetheless, we know robustly from studies of identity and intergroup
conflict that groups often construct their own narratives of deprivation
and threat that differ from social scientists’ perceptions. In addition,
entrepreneurial civil society and political leaders often manipulate
perceptions instrumentally. Perhaps Chilean evangelicals might have come
to perceive their group as threatened if community leaders had chosen to
convey that message. Throughout this book, we have seen that Brazil’s
evangelical and Pentecostal clergy and citizens often feel threatened by the
state, in ways that many observers might suspect are exaggerated.
Meanwhile, supporters of Afro-Brazilian religions likewise perceive
evangelicals as threats to their group. Thus, we need more work on the
determinants of perceived religious threat.
Third, recognizing the importance of ideational and psychological resources
opens up new questions. How do religious leaders choose which symbolic
resources to mobilize? Are some types of resources likely to be more effective
in mobilizing individuals? Future research should explore these and other
questions.

the democratic consequences of religious engagement

in politics
The broad Latin American context also illuminates the second question driving
this manuscript: how do (and how will) the culture wars shape electoral politics
and democracy in Latin America? Early studies of evangelicalism and
Pentecostalism across the developing world in the 1960s, 1970s, and 1980s
initiated a yet-unresolved debate over the religious tradition’s impact on the
democratic dispositions of adherents (Robbins 2004; Steigenga 2003). Some
scholars maintain that evangelical and Pentecostal congregations promote
conservatism, authoritarianism, and deference to authority (e.g., Bastian
1993; Corten 1999; Rink 2018). Yet proponents of what is often called the neo-
Weberian thesis perceive evangelical and Pentecostal congregations as sites
where adherents – often society’s excluded – learn civic skills and attitudes
critical for democracy (e.g., Burdick 1993a; Lankina and Getachew 2012;
D. Martin 1993; Marshall 2009; Stoll 1990). One potential resolution to this
disagreement has been to point out that heterogeneous congregations have
heterogeneous impacts on heterogeneous individuals. Some congregations are
hierarchical, while others are open and democratic; some individuals learn
168 Representation

participatory skills at church, while others fail to do so (Ireland 1993, 1999;

Robbins 2004).
My approach to this disagreement is slightly different. It is not simply that
some people receive the normatively desirable effects of religion, and others the
normatively less desirable ones. When congregations get involved in politics,
I argue, they often simultaneously integrate citizens into democratic politics and
foster rightist conformity and intolerance. These competing stories are mutually
compatible. Mobilizing the people of God has multivalent impacts in the
developing world; the two partial visions present aspects of a complex reality.

How Congregational Politics Fosters Rightism and Intolerance

Ultimately, as we will see, congregational engagement in politics incorporates
citizens into democracy in ways that fundamentally strengthen political
systems, particularly ones in crisis. Nonetheless, congregational politics also
often reshapes the distribution of political preferences in the electorate, pushing
public opinion as a whole in a conservative direction. Furthermore,
congregations may exacerbate certain authoritarian impulses among citizens.
Evangelical and Pentecostal church attendance often reinforces and
intensifies right-leaning views of congregants on sexual and family
traditionalism. Clergy in these traditions are, with some exceptions, reliably
highly conservative on issues such as abortion and homosexuality. Conservative
citizens may self-select into evangelical and Pentecostal congregations;
nonetheless, nonconservatives who happen into the congregational milieu are
influenced by the dominant ideological tendency. Congregations with
doctrinally conservative stances are most likely to cohere ideologically, and in
terms of vote choice. Thus, social influence within evangelical and Pentecostal
congregations ultimately leads many participants to adopt the congregational
central tendency on core issues. Nonetheless, citizens who more strongly adhere
to secular norms are less likely to conform to the conservative views of their
congregations.
Rightist views on matters such as abortion and homosexuality are not
necessarily politically intolerant or authoritarian. For instance, many citizens
simultaneously personally disapprove of homosexuality or same-sex marriage,
and believe that gay citizens should be accorded full political and civil rights.
However, the conservatism fostered in evangelical and Pentecostal churches
may often exacerbate authoritarian impulses, with members seeking to deny
political and civil liberties to those of whom they disapprove for religious
reasons. Two aspects of evangelical and Pentecostal congregations contribute
to this tendency. Most importantly, sacred language that conveys dualistic,
good-versus-evil views of social issues can exacerbate affective polarization,
leading citizens to intensely dislike religious and political out-groups. Dualism
and notions of spiritual warfare contribute to citizens perceiving out-groups as
sinners, as guided by supernatural evil forces, and as existential threats to the
Conclusion: Mobilizing the People of God 169

religious in-group. Literal demonization of political opponents is perhaps most

vivid in the widespread urban myth that President Michel Temer is a Satanist
(see Chapter 8), and in the 2010 Universal Church of the Kingdom of God
(UCKG) publication that implied that José Serra was possessed by demons (see
Chapter 1). Ultimately, dualistic theology and affective polarization can make
citizens unwilling to tolerate the political and civil liberties of religious out-
groups. After all, there is no need to compromise with the devil. Furthermore,
beyond the impact of dualistic theology and worldviews, evangelical and
Pentecostal clergy themselves may fail to bolster, or even undermine, support
for out-groups’ civil liberties. As noted in Chapter 4, despite evangelical and
Pentecostal religious leaders’ support for democracy in the abstract, they give
low priority to preaching internal or external political tolerance.
This discussion is based on results from a case study of a single country in
a limited time period. Does religious politicking necessarily lead to either
conservatism or intolerance? Certainly not – not even within this country case
during this historical period. Gay-friendly evangelical churches are springing up
in the largest Brazilian cities, though they remain novelties that are few in
number. And the history of the Catholic Church in some times and places in
Brazil (as across Latin America more broadly) indicates that religious
communities sometimes actively mobilize leftist action. Beyond these
examples, it is easy to imagine religion pushing even more solidly in
a leftward direction. Brazil’s Catholic left has historically been very liberal on
socioeconomic issues, but centrist or even center-rightist on matters of sexual
and family traditionalism. Meanwhile, Brazil’s gay-friendly evangelical
churches may accept gays, yet foster certain other aspects of religious
conservatism.
Yet this discussion leaves important questions remaining. For instance, why
do right-leaning religious communities often seem to grow and mobilize in
politics more effectively than do left-leaning ones? And under what
circumstances do left-leaning religious communities effectively mobilize?
Again, approaches from political psychology, religious studies, and political
behavior will be needed to address these questions. The results presented in this
book do suggest one reason why the religious left may struggle: theological
liberalism may make congregants less likely to follow the direction of clergy, or
to seek to conform to the political beliefs of their congregations.

How Congregational Politics Pushes Elite-Level Politics to the Right

Politicking within churches can also push legislative bodies – indeed, elite-level
representation in general – in a conservative direction. Recall that in Brazil’s
evangelical and Pentecostal congregations, clergy are aligned with congregants on
matters of sexual and family traditionalism, yet they are more reliably
conservative on other issues. At the mass level, evangelicals and Catholics have
identical views on issues such as economic policy, race, and the environment.
170 Representation

However, there are significant gaps in the importance evangelical and Catholic
clergy place on such issues.
Evangelical and Pentecostal forms of political organizing will tend to
privilege the issue priorities of clergy over those of congregants, as discussed
in Chapters 2 and 9. That is, clergy’s role as brokers – choosing candidates and
mobilizing voters – enables them to capture policy “rents.” As a result, when the
political preferences of clergy and congregants diverge, candidates elected with
congregational support will tend to be closer to the views of clergy. Religious
leaders can control the information that reaches congregants, focusing on issues
where candidates and congregants agree, such as abortion and homosexuality.
In addition, candidates may be more likely to provide clientelistic side-payments
to congregants than to clergy. Finally, the asymmetry in information and
sophistication between clergy and congregants will make clergy more likely to
monitor politicians’ behavior.
Electing religious rightists strengthens the power of the right at the elite level
more generally. Politicians on the religious right form alliances with other
rightist politicians, including those whom Power and Rodrigues-Silveira term
the “clientelistic right,” the “law-and-order right,” and the “economic right”
(2018). The dramatic alignment of evangelicals with other rightist groups in
supporting the impeachment of Dilma Rousseff signals the importance of
evangelicals to the right at the elite level more broadly. At the level of civil
society, the Free Brazil Movement (MBL, or Movimento Brasil Livre), a large,
new libertarian social movement organization, has begun connecting the
various facets of rightism. And in Colombia, one finds various rightisms
instrumentally linked in the campaign against the peace deal.
Once again, questions remain for future research. For instance, why are
clergy more reliably conservative than citizens on so many issues? Tentatively,
it seems likely the answer has two parts. First, clergy have more fully absorbed
and internalized internationally disseminated doctrine constructed within
globalized religious communities that contains ideological assumptions about
“what goes with what.” That is, there is no single, logically necessary story
connecting, for instance, attitudes on abortion and tariffs. (Indeed, a naïve
reading of the issues could even suggest that support for abortion rights and
opposition to tariffs should go together.) However, clergy are more likely than
citizens to grasp and accept ideological programs defining “left” and “right,”
and dictating that people who oppose abortion rights should also oppose tariffs
(e.g., Converse 1964). Second, conservative clergy are likely to be more aware
than citizens that in national politics, religious rightists support right-leaning
causes more generally.

How Congregational Politics Integrates Citizens into Democratic Politics

Yet while religious politicking in Brazil pushes politics to the right, it has an
even more fundamental political impact, helping to root and stabilize
Conclusion: Mobilizing the People of God 171

Brazil’s post-1985 democracy by mobilizing citizens to engage more fully

with the political system. Religious communities’ styles of political activism
suggest a fruitful comparison with two other major types of institutions that
organize mass politics: labor unions and political parties. In Chapter 2, we
noted that clergy roles are analogous to those of union bosses; both types of
leaders simultaneously guide public opinion and coordinate with candidates
(Verba, Schlozman, and Brady 1995). Religious communities also often
perform similar functions as those of political parties, I will argue.
However, the appropriateness of the party metaphor varies a great deal
across time and place.
In seminal work, Aldrich notes that political parties “are designed as
attempts to solve” three problems endemic to electoral democracy (1995,
21–22). The first is the “problem of ambition and elective office seeking”
(22); an effective party coordinates candidacies and provides a vehicle for
ambition, outlining a path for elite-level political actors to ascend to higher
levels of power. Second is the “problem of making decisions for the party and
the polity” (22); parties develop policy agendas and coordinate action to pass
laws in accordance with the agenda. The third and “most pervasive” problem
parties are designed to solve is the “problem of collective action” (23). “How do
candidates get supporters to vote for them – at least in greater numbers than
vote for the opposition – as well as get them to provide the cadre of workers and
contribute the resources needed to win election? The political party has long
been the solution” (24). The task of coordinating collective action involves
many pieces. Parties parse and package disparate, complex issues that often
lack a necessary logical connection. In so doing, they explain to elites and
masses “what goes with what” and persuade citizens to adopt one package
over another (Converse 1964). Further, parties continuously evolve brands to
sell those bundles of issues and candidates to the mass public (Lupu 2013). Last,
they mobilize political participation, raise resources, and allocate those
resources to candidates.
Thinking about parties from the perspective of Latin America reveals
a fourth core problem of democracy that political parties help to solve. One
of the most fundamental challenges of electoral democracy is maintaining
legitimacy among elites and citizens who may (quite reasonably) dislike the
messy, conflictual business of electoral politics, and who often – as much as
49.99 percent of the time, on average – end up on the losing end of political
contests (C. J. Anderson et al. 2005). A long tradition of literature on transitions
to and from democracy suggests that disgruntled elites or masses might not only
stop participating, but even potentially destabilize the system. Parties can help
to address these legitimacy problems. By giving elites and citizens a team to
identify with and root for, parties help extend time horizons beyond an
immediate loss, to the possibility of a win in the next match. In addition,
party leaders convey legitimacy to the political system when they express
confidence in it, and muster followers to keep going. This leader-driven
172 Representation

process can bolster psychological engagement with and understanding of the

major issues of the day.
In contrast to parties, congregations were certainly not purpose-built to solve
any of these four democratic problems. Nonetheless, I argue that in the
Brazilian case, religious communities have evolved forms of activism that
happen partially to address all three problems Aldrich identifies, as well as the
problem of legitimacy. Beyond the Brazilian case, congregations often help to
solve the third and fourth problems.
Congregations are potentially natural fits for tackling Aldrich’s third
“problem.” As civil-society entities in which local elites with strong
theological/political views maintain frequent, sustained contact with large
numbers of citizens, congregations readily coordinate collective action. Thus,
one of the primary jobs of clergy is precisely to bundle policy issues (in
combination with theological stances) for congregants. Clergy help the
faithful understand “what goes with what,” and persuade them to adopt the
religious in-group’s theological/ideological package. Religious identity-group
labels can also be considered brands, in ways not dissimilar to party brands.
Though many religious communities choose not to use those brands to market
bundles of candidates and issues, religious communities themselves or
entrepreneurial politicians sometimes do so. The danger, of course, is that, as
de Tocqueville warned, parties that do so increase their “power over some and
[lose] the hope of reigning over all” (2010, 484). Finally, congregations are also
distinctively good at mobilizing human and financial resources, which they do
sometimes turn to political ends. Thus, in many democracies, congregations
provide a supportive infrastructure that complements the party system in
coordinating citizens to take part in politics.
Intriguingly, Chapter 8 shows that congregations can similarly bolster the
legitimacy of the political system. Clergy typically have higher levels of support
for democracy and the political system than do their congregants, and they
communicate those supportive attitudes to their flocks. When they preach on
public issues, clergy implicitly legitimize the public fora in which those issues are
debated, while bolstering psychological engagement among citizens. Perhaps
most importantly, clergy can help congregants cope psychologically with
political losses, while legitimizing the political system. Take, for example,
Pastor Willian in the first scene of Chapter 8 – encouraging his congregants to
see themselves as “winners” after receiving disappointing election results, and
to focus simultaneously on the races where they received good news, and on the
long haul. Such motivational rhetoric, with perhaps somewhat fewer biblical
references, would not have been out of place in a speech given by a party leader
facing losses on election night.
At the same time, Chapter 8 also reveals that clergy influence cuts both ways.
Clergy can legitimate the political system in numerous ways, yet they can also
erode citizens’ perceptions of the political regime’s legitimacy. Evangelical and
Pentecostal clergy are much more likely than Catholic clergy to perceive that the
Conclusion: Mobilizing the People of God 173

political system and current officeholders harm their religious groups, as

demonstrated in Chapter 4. Such perceptions rub off on congregants. Citizens
are less likely to perceive democracy and the political system as legitimate when
they attend churches where religious leaders believe the political system harms
their group.
In the particular context of Brazil, some congregations have also, as
competitive strategies, evolved methods of organizing that address the first
and second democratic problems Aldrich identifies. First, many evangelical
and Pentecostal congregations help to provide a vehicle for would-be elites’
political ambition. Within Pentecostal and neo-Pentecostal denominations such
as the Assembly of God and the UCKG, religious leaders themselves directly
coordinate candidacies. In other evangelical and Pentecostal congregations,
arrangements are more informal. Second, religious hierarchies – especially in
certain well-organized denominations – help to coordinate legislative platforms
for “their” representatives, at least on core issues. Thus have Brazil’s clergy-
driven culture wars evolved.
Religious communities have developed these functions in the Brazilian case
for a few reasons. We can follow the framework outlined in the first section of
this chapter to identify the forces that enabled the unusual insertion of
evangelical religious groups into electoral politics in Brazil. First, perceived
threats from both the state and other religious groups initially motivated
congregations to get directly involved in coordinating candidacies. Second,
Brazil’s exceptionally open electoral and party systems created the
opportunities that permitted and enabled this mobilization. Third, high levels
of religious adherence and swelling congregations provided human resources, in
addition to the doctrinal and symbolic resources provided by the religions
themselves.
Nonetheless, it likely goes without saying that even Brazil’s best organized
Pentecostal and evangelical congregations do not fully substitute for parties in
the first two dimensions. In the clergy-driven culture wars, congregations do not
effectively coordinate elite ambition to reach the most coveted positions in the
halls of power. Nor do they create full legislative platforms on all issues, or
effectively coordinate voting (despite some corralling of legislators on certain
issues). Moreover, to the extent that religious communities substitute for some
party functions in contexts where parties have little presence in the mass public,
congregations create incentives for partisan fragmentation that may ultimately
weaken Brazilian democracy.
This discussion leaves many open questions. For instance, when, how, and
why do clergy deliberately seek to boost the legitimacy of the political system?
And to pick up a theme from earlier in the chapter, what drives clergy
perceptions that the system hurts their in-group? Results from Chapter 4
indicate that intolerance triggers grievances against the political system,
suggesting that some clergy disappointment with the political system might be
driven more by their own out-group attitudes than by what social scientists
174 Representation

would perceive as objective harms. Nonetheless, much future research remains

to be done.

the present and future of religious politics in brazil

Brazilians encounter each other on the streets, at the bus stop, in supermarkets,
in work and school, and in church. In millions upon millions of interactions
throughout their daily lives, they come to understand Brazilian democracy and
the nation state, and they make choices together about how and when to take
part in democratic politics. Religious communities often enter this national
conversation, driven by membership pressures, theological and political
beliefs, and commitment to Brazilian democracy in the abstract. They play
two roles in this ongoing process of political socialization. First,
congregations congregate. That is, churches are a site where people interact
with others who they often do not know well, including politicians, in regular,
structured encounters. Congregational leaders set aside important time to talk
about serious topics; clergy often touch on not just what people should be doing
to prepare for the afterlife, but how they should be interacting with their fellow
citizens in the here and now. Second, congregations mobilize God’s people.
By integrating citizens into Brazilian democracy, religious groups play a critical
role in the context of sustained political crisis. Yet at the same time, Brazil’s
culture-war politics are on balance more conservative, and sometimes more
intolerant, as a result of the work of churches.
What does the future hold for Brazilian democracy and religion? Ever-
advancing waves of crisis have, since 2014, hollowed out Brazil’s electoral
democracy. Trends as of 2018 are alarming: dramatically declining support for
democracy; simultaneously, the bottoming out of partisanship; and at the elite
level the apparent difficulty of coalescing around new center-leftist leaders
capable of taking Lula’s place in popular loyalties and competing effectively
with far-rightist Jair Bolsonaro. Brazilian democracy may need a refounding –
not a constitutional one, but a refounding in the organizational infrastructure
that mobilizes ideas and action to solve collective problems. Brazilian churches
can serve as one cornerstone of this refounding, as clergy and congregants
organize – not only for the afterlife, but in pragmatic hope for the here and now.
But stopping here would be Pollyannaish. Despite churches’ important
democratic roles, religious politicking also contributes to a growing threat to
the very existence of Brazilian democracy. As nostalgia for the 1964–1985
military dictatorship grew on the Brazilian right, evangelical support for far-
rightist, sometime-evangelical presidential candidate Jair Bolsonaro, as well as
other right-leaning candidates, helped to tip the balance of the vote to the far
right in the presidential election of October 2018. Political engagement in
churches ended up helping to bring to power the very elites who might
destroy democracy from within.
Conclusion: Mobilizing the People of God 175

Nonetheless, there are reasons to hope that most churches will, despite their
rightist proclivities, avoid mobilizing behind the antidemocratic far right.
Clergy support for democracy in the abstract will reduce enthusiasm for
overtly authoritarian options. Perhaps even more importantly, religious
groups’ engagement in electoral politics gives clergy who might be tempted by
authoritarian options a stake in the electoral game. Moreover, incentives to
electoral disunity – the very incentives that prevent evangelical denominations
from functioning effectively as a partisan front – may also prevent a unified
evangelical coalition behind any particular politician who promises to rig the
game to in-group advantage. If so, Brazil’s clergy-driven culture wars could
ultimately help to stabilize democracy.
Afterword

In the months leading up to Brazil’s October 2018 presidential elections,

fallacious rumors circulated in evangelical circles via the messaging app
WhatsApp. One rumor alleged that the Workers’ Party (PT) candidate
Fernando Haddad had sent baby bottles shaped like penises to state-run
childcare centers to teach children about homosexuality. Another attributed a
false quote to Haddad, in which he had supposedly called for children to
become wards of the state at age six, so that the state could forcibly reassign
their genders. Such messages – a “fake news” variant of the CDs advertising
Dilma Rousseff’s stance on abortion in the 2010 campaign – likely contributed
to evangelical support for the far-rightist Jair Bolsonaro. Indeed, sizable
religious gaps in voting appeared. Alves (2018) estimates that Bolsonaro
received 50.1 percent of Catholic and 43.0 percent of nonreligious votes, but
68.3 percent of evangelical and Pentecostal votes, and would narrowly have lost
the presidency but for evangelicals and Pentecostals.
Clergy played a role. By the final weeks of the first-round campaign,
Bolsonaro had racked up endorsements from such prominent names as Edir
Macedo, Silas Malafaia, and the Evangelical Parliamentary Front. Following
his second-round win, the president-elect’s first public appearance was at a
worship service at Pastor Silas Malafaia’s church.
I sent the final draft of this manuscript to press in July 2018; I am typing out
this afterword in early November, in the weeks following Bolsonaro’s second-
round victory. What do the past four months teach us about religion and
Brazilian democracy?
The 2018 presidential election confirms much of the book’s argument. As the
party system that had dominated Brazil’s post-1985 democracy crumbled,
culture-war polarization drove voters’ choices. Among evangelicals,
conservative stances on sexuality and gender combined with anti-petismo to
yield strong support for the right. Social-media echo chambers may have
intensified both support for Bolsonaro and fear of his opponent. Meanwhile,
many social-justice-oriented Catholic commentators fervently rejected the

176
Afterword 177

candidate who campaigned on the slogan that “A good criminal is a dead

criminal” (“Bandido bom é bandido morto”).
Yet the behavior of Pentecostal elites in 2018 departs from the story of the
previous two decades. For the first time, a largely united Pentecostal front
favored one presidential candidate – probably both leading and reacting to
evangelical citizens’ solid support for Bolsonaro. The parallel to the 2008 Juiz
de Fora mayoral campaign featuring the lesbian professor Margarida Salomão
is telling. While many historical Protestant churches maintained distance from
Bolsonaro, the upshot was still substantially greater evangelical cohesion than
seen previously in presidential politics.
Will Brazil’s rightist politicians be able to maintain this base? Bolsonaro’s
win shows the path to stable future evangelical-rightist coalitions based on
anxiety over liberal notions of gender and sexuality. Yet evangelical unity will
likely only go so far. At the level of legislative races, churches’ incentives to
disunity will continue to fragment the evangelical vote across many different
candidates and parties, largely on the right.
So, whither Brazilian democracy? As I speculated in the closing pages of the
final chapter, evangelical support for Jair Bolsonaro helped to elect a president
who threatens human rights and the democratic order, perhaps more so than
any politician since democratization. Nonetheless, religious engagement on
behalf of the eventual winner likely also deepened citizens’ attachment to a
democratic system with fraying popular legitimacy. In the coming years,
Bolsonaro’s presidency may test both elite and mass commitment to
democratic values and procedures. I maintain hope that Catholics,
evangelicals, and the nonreligious will find common ground in the dream of a
tolerant and democratic Brazil.

references
Alves, José Eustáquio Diniz. 2018. “O voto evangélico garantiu a eleição de Jair
Bolsonaro.” EcoDebate (blog). October 31, 2018. www.ecodebate.com.br/2018/10/
31/o-voto-evangelico-garantiu-a-eleicao-de-jair-bolsonaro-artigo-de-jose-eustaquio-
diniz-alves/.
appendix a

List of Focus Groups and Church Observations

Unless otherwise noted below, all interviewee and church names used in the
book are changed for the purpose of anonymizing the qualitative data.
Quotation marks in the third columns below indicate that a congregation’s
name has been changed.

Focus Groups

Citation Date Location/Description

[FG1] October 22, 2014 “São José Parish”: Large Catholic

congregation in a middle-class
neighborhood
[FG2] October 24, 2014 “Disciples of Love Community”: a small
Catholic prayer community in a low-income
neighborhood
[FG3] October 18, 2014 “Vila Bela Methodist Church”: Large, middle-
class Methodist congregation
[FG4] October 19, 2014 “Good News Baptist Church”: Large, middle-
class Baptist congregation
[FG5] October 19, 2014 “Hope Church of the Nazarene”: a small,
storefront Nazarene church in a low-income
neighborhood
[FG6] October 27, 2014 “Juiz de Fora Assembly of God”: Focus group
of evangelical pastors at a mid-sized, mixed-
income Assembly of God congregation
[FG7] October 24, 2014 Focus group for nonreligious respondents
recruited through community
advertisements; held at the Federal
University of Juiz de Fora

178
List of Focus Groups and Church Observations 179

Church Observations

Citation Date Location/Description

[CO1] August 10, 2014 Low-income, Pentecostal storefront

congregation
[CO2] October 12, 2014 “Vila Bela Methodist Church”: Large, middle-
class Methodist congregation
[CO3] August 13, 2014 Comunidade Manancial (NOTE: Church and
individuals’ names unchanged because this
was a public event involving a national
politician.)
[CO4] August 17, 2014 “São Ignacio Parish”: Small, lower-middle-
class Catholic congregation
[CO5] August 17, 2014 “Good News Baptist Church”: Large, middle-
class Baptist congregation
[CO6] August 18, 2014 “Santa Rita Community”: Large Charismatic
Catholic community
[CO7] August 22, 2014 “São José Parish”: Large Catholic
congregation in middle-class neighborhood
[CO8] August 22, 2014 “UCKG of Santa Emélia”: Large, low-income
UCKG congregation
[CO9] August 24, 2014 “UCKG of Santa Emélia”
[CO10] August 24, 2014 “Hope Church of the Nazarene”
[CO11] August 29, 2014 Session at meeting of pastors
[CO12] September 4, 2014 “Santa Rita Community”
[CO13] September 7, 2014 “UCKG of Santa Emélia”
[CO14] September 7, 2014 “Hope Church of the Nazarene”
[CO15] September 10, 2014 Comunidade Evangélica Resgatando Vidas
(NOTE: Names unchanged because this was
a public event.)
[CO16] September 11, 2014 “Juiz de Fora Assembly of God”: Mid-sized,
mixed-income Assembly of God
congregation
[CO17] September 14, 2014 Comunidade Resgate, Chácara site (NOTE:
Site and speaker names unchanged because
this was a public event involving an
international speaker.)
[CO18] September 17, 2014 “Disciples of Love Community”
[CO19] September 19, 2014 “Juiz de Fora Assembly of God”
[CO20] September 20, 2014 “São José Parish”
[CO21] September 21, 2014 “Juiz de Fora Assembly of God”
[CO22] September 22, 2014 Small, low-income Pentecostal congregation
[CO23] September 23, 2014 “Vila Bela Methodist Church”
[CO24] September 28, 2014 “São Ignacio Parish”
180 Appendix A

(continued)

Citation Date Location/Description

[CO25] September 28, 2014 “Santa Catarina Parish”: large, low-income

Catholic parish
[CO26] September 28, 2014 “Santa Catarina Parish”
[CO27] September 28, 2014 “Juiz de Fora Assembly of God”
[CO28] September 29, 2014 “São Ignacio Parish”
[CO29] October 1, 2014 “Good News Baptist Church”
[CO30] October 5, 2014 “UCKG of Bela Vista”: mid-sized, mixed-
income UCKG congregation
[CO31] October 5, 2014 “Good News Baptist Church”
[CO32] October 8, 2014 Meeting of the Citizenship Committee at the
Catholic Cathedral (NOTE: Participants’
names are anonymized, but the name of the
group is reported.)
[CO33] October 10, 2014 “Juiz de Fora Assembly of God”
[CO35] June 22, 2017 Interview, deacon, “Good News Baptist
Church”
[CO36] June 22, 2017 Interview, “Padre Miguel,” “Santa Fé Parish”
[CO37] June 22, 2017 Interview, “Vila Bela Methodist Church”
[CO38] June 22, 2017 Interview, deacon, large middle-class Catholic
parish
[CO39] June 23, 2017 Interview, “Pastor André,” “Juiz de Fora
Assembly of God”
[CO40] June 23, 2017 Interview, pastor, “Faith Mission Evangelical
Community”
[CO41] June 23, 2017 Interview, priest, “São José Parish”
[CO42] June 23, 2017 Interview, pastor, “UCKG of Santa Emelia”
[CO43] June 24, 2017 Interview, “Pastor Eduardo,” “Hope Church
of the Nazarene”
[CO44] June 25, 2017 Interview, “Pastor Willian,” “Good News
Baptist Church”
[CO45] June 25, 2017 Interview, pastor, “Light of Life Quadrangular
Church”
[CO46] June 26, 2017 Group interview, “Disciples of Love
Community”
appendix b

Focus-Group Protocol

In the fictitious city of Bela Vista (RJ), in a universe not very far away, the
following things happen . . .
1. In an evangelical church, Pastor Sérgio distributes a pamphlet instructing
church members on the importance of voting.
2. In another evangelical church in the city, Pastor João preaches to church
members, “Politics isn’t for believers.” He’s waiting for the End Times
and says they should concentrate on saving souls.
3. In a Catholic church in the Santo Tomás neighborhood, Father Luiz often
talks about the sin of homosexuality and the Workers’ Party (PT) in the
Catholic formation class, saying that the PT is very misguided in its
legislative activism.
4. In an evangelical church in the same neighborhood, Pastor Eunice is
collecting signatures for a popular initiative law to define marriage as
between a man and a woman.
5. In the Catholic cathedral in the center of the city, Father Flávio has
a private meeting with some deacons he knows well. He tells them that
he supports Dilma and is opposed to the ideology of the Party of Brazilian
Social Democracy (PSDB).
6. In an evangelical church in the city, an evangelical candidate asks Pastor
Ricardo for his support. The pastor tells the candidate that he personally
supports him, but that he can’t campaign for him inside the church.
The candidate stays outside the church after service distributing
materials.
7. At the end of May, a Catholic church in the São Leopoldo neighborhood
organizes a march against crime. At the march, Father José Luiz preaches
to the participants that the mayor isn’t doing enough to combat crime.
8. In a Four-Square Baptist church in the city, Pastor William strongly
supports the presidential candidate Marina, a state deputy, and
a federal deputy. He puts up posters for them outside the church.

181
182 Appendix B

9. In the Universal Church of the Kingdom of God in downtown Bela Vista,

at worship service on the day of the election, Pastor Carlos asks all the
people in attendance to repeat the name and electoral code of his candi-
date for state deputy.
10. In a Catholic church in a neighborhood close to the university, Father
Carlos lets a gay couple participate actively in church activities. When
some church members object to the couple’s participation, the priest
responds that, “We are all children of God.”
appendix c

Variable Coding and Information on Studies

See Online Appendix at https://dataverse.harvard.edu/dataverse/aesmith.

183
References

Abramowitz, Alan. 2010. The Disappearing Center: Engaged Citizens, Polarization, and
American Democracy. New Haven, CT: Yale University Press.
Abril.com. 2010. “Dilma pede direito de resposta a Canção Nova por sermão de padre.”
Abril, October 7.
Agência DIAP. 2010. “Evangélicos crescem no Congresso; PSC tem mais
representantes.” DIAP – Departamento Intersindical de Assessoria Parlamentar.
Agência O Globo. 2016. “Deputado pede que Temer ‘abandone satanismo’ e
‘a maçonaria.’” Epoca Negocios, November 18. http://epocanegocios.globo.com/
Brasil/noticia/2016/11/deputado-pede-que-temer-abandone-satanismo-e-maconaria
.html.
Albertson, Bethany L. 2011. “Religious Appeals and Implicit Attitudes.” Political
Psychology 32 (1): 109–30.
2014. “Dog-Whistle Politics: Multivocal Communication and Religious Appeals.”
Political Behavior, January, 1–24.
Aldrich, John H. 1995. Why Parties? The Origin and Transformation of Political Parties
in America. Chicago: University of Chicago Press.
Alex-Assensoh, Yvette, and A. B. Assensoh. 2001. “Inner-City Contexts, Church
Attendance, and African-American Political Participation.” Journal of Politics
63 (3): 886–901.
Allport, Gordon Willard. 1979. The Nature of Prejudice. New York: Basic Books.
Alonso, Angela, and Ann Mische. 2016. “Changing Repertoires and Partisan Ambivalence in
the New Brazilian Protests.” Bulletin of Latin American Research, March.
Ames, Barry. 2001. The Deadlock of Democracy in Brazil. Ann Arbor, MI: University of
Michigan Press.
Ames, Barry, Alyssa Huberts, Fabiana Machado, Lucio Rennó, David Samuels, Amy
Erica Smith, and Cesar Zucco. 2016. “Brazilian Electoral Panel Study: 2014 Results.”
IDB-TN-915. Washington, DC: Inter-American Development Bank. www.iadb.org/en/
research-and-data/publication-details,3169.html?pub_id=IDB-TN-915.
Ames, Barry, Fabiana Machado, Lucio Rennó, David Samuels, Amy Erica Smith, and
Cesar Zucco. 2013. “Brazilian Electoral Panel Survey.” Technical Note IDB-TN
-508. Washington, DC: Inter-American Development Bank. www.iadb.org/en/
research-and-data/publication-details,3169.html?pub_id=IDB-DB-105.

184
References 185

Anderson, Allan. 2004. An Introduction to Pentecostalism: Global Charismatic

Christianity. Cambridge: Cambridge University Press.
Anderson, Christopher J., André Blais, Shaun Bowler, Todd Donovan, and
Ola Listhaug. 2005. Losers’ Consent: Elections and Democratic Legitimacy.
Oxford: Oxford University Press.
Anderson, John. 2011. “Conservative Christianity, the Global South and the Battle over
Sexual Orientation.” Third World Quarterly 32 (9): 1589–1605.
Anderson, Lisa R., and Jennifer M. Mellor. 2009. “Religion and Cooperation in a Public
Goods Experiment.” Economics Letters 105 (1): 58–60.
Ang, Adrian, and John R. Petrocik. 2012. “Religion, Religiosity, and the Moral Divide in
Canadian Politics.” Politics and Religion 5 (1): 103–32.
Aragão, Jarbas. 2016. “Jair Bolsonaro é batizado por pastor no rio Jordão.” Gospel
Prime, May 12. https://noticias.gospelprime.com.br/jair-bolsonaro-batizado-rio
-jordao/.
Autero, Esa. 2015. “Blessed Are the Prosperous but Woe to the Weak.” In Pentecostals
and Charismatics in Latin America and Latino Communities, edited by
Néstor Medina and Sammy Alfaro, 169–83. Christianity and Renewal –
Interdisciplinary Studies. New York: Palgrave Macmillan.
Bacarji, Arlene Denise. 2016. “A Igreja, a homossexualidade e o clero.” Revista
Encontros Teológicos 26 (3): 111–27. https://facasc.emnuvens.com.br/ret/article/
view/217.
Baiocchi, Gianpaolo. 2005. Militants and Citizens: The Politics of Participatory
Democracy in Puerto Alegre. Stanford, CA: Stanford University Press.
Baker, Andy, Barry Ames, and Lucio R. Rennó. 2006. “Social Context and Campaign
Volatility in New Democracies: Networks and Neighborhoods in Brazil’s 2002
Elections.” American Journal of Political Science 50 (2): 382–99.
Barabas, Jason, and Jennifer Jerit. 2010. “Are Survey Experiments Externally Valid?”
American Journal of Political Science 104 (2): 226–42.
Barrios, Juan José, and Nestor Gandelman. 2014. “Religious Participation, Trust and
Reciprocity: Evidence from Six Latin American Cities.” The B. E. Journal of
Economic Analysis & Policy 15 (1): 353–76.
Barros, Antonio Teixeira de, Cristiane Brum Bernardes, and Sílvia Mugnatto Macedo.
2015. “Comunicação, cultura e política nas rádios do poder legislativo no Brasil:
identidade e perfil da programação da Rádio Senado e da Rádio Câmara.” Latin
American Research Review 50 (1): 207–27.
Basedau, Matthias, Jonathan Fox, Jan H. Pierskalla, Georg Strüver, and
Johannes Vüllers. 2017. “Does Discrimination Breed Grievances – and
Do Grievances Breed Violence? New Evidence from an Analysis of Religious
Minorities in Developing Countries.” Conflict Management and Peace Science 34
(3): 217–39.
Bastian, Jean-Pierre. 1993. “The Metamorphosis of Latin American Protestant
Groups: A Sociohistorical Perspective.” Latin American Research Review 28
(2): 33–61.
BBC News. 2018. “Costa Rica Election Goes into Runoff.” BBC News, February 5.
www.bbc.com/news/world-latin-america-42938510.
Bean, Lydia. 2014a. “Compassionate Conservatives? Evangelicals, Economic
Conservatism, and National Identity.” Journal for the Scientific Study of Religion
53 (1): 164–86.
186 References

2014b. The Politics of Evangelical Identity: Local Churches and Partisan Divides in
the United States and Canada. Princeton, NJ: Princeton University Press.
Bean, Lydia, Marco Gonzalez, and Jason Kaufman. 2008. “Why Doesn’t Canada Have
an American-Style Christian Right? A Comparative Framework for Analyzing the
Political Effects of Evangelical Subcultural Identity.” The Canadian Journal of
Sociology / Cahiers Canadiens de Sociologie 33 (4): 899–943.
Beatty, Kathleen Murphy, and Oliver Walter. 1984. “Religious Preference and Practice:
Reevaluating Their Impact on Political Tolerance.” Public Opinion Quarterly 48
(1B): 318–29.
Bebbington, David William. 1989. Evangelicalism in Modern Britain: A History from
the 1730s to the 1980s. London: Unwin Hyman.
Bello, Jason, and Meredith Rolfe. 2014. “Is Influence Mightier than Selection? Forging
Agreement in Political Discussion Networks during a Campaign.” Social Networks,
Special Issue on Political Networks, 36 (January): 134–46.
Ben-Nun Bloom, Pazit. 2013. “The Public’s Compass: Moral Conviction and Political
Attitudes.” American Politics Research 41 (6): 937–64.
Ben-Nun Bloom, Pazit, Gizem Arikan, and Marie Courtemanche. 2015. “Religious
Social Identity, Religious Belief, and Anti-Immigration Sentiment.” American
Political Science Review 109 (02): 203–21.
Berelson, Bernard R., Paul F. Lazarsfeld, and William N. McPhee. 1954. Voting: A Study of
Opinion Formation in a Presidential Campaign. Chicago: University of Chicago Press.
Bineham, Jeffery L. 1988. “A Historical Account of the Hypodermic Model in Mass
Communication.” Communication Monographs 55 (3): 230–46.
Boas, Taylor C. 2014. “Pastor Paulo vs. Doctor Carlos: Professional Titles as Voting
Heuristics in Brazil.” Journal of Politics in Latin America 6 (2): 39–72.
Boas, Taylor C., and Amy Erica Smith. 2015. “Religion and the Latin American Voter.”
In The Latin American Voter, edited by Ryan E. Carlin, Matthew Singer, and
Elizabeth Zechmeister, 99–121. University of Michigan Press.
In press. “Looks Like Me, Thinks Like Me? Descriptive Representation and Opinion
Congruence in Brazil.” Latin American Research Review 54 (3).
Bob, Clifford. 2012. The Global Right Wing and the Clash of World Politics.
Cambridge, UK: Cambridge University Press.
Bohn, Simone. 2004. “Evangélicos no Brasil: Perfil sócio-econômico, afinidades
ideológicas e determinantes do comportamento eleitoral.” Opinião Pública 10
(2): 288–338.
2007. “Contexto político-eleitoral, minorias religiosas e voto em pleitos presidenciais.
(2002–2006).” Opinião Pública 13 (2): 366–87.
Bom, Klaas. 2015. “‘I Feel the Presence of God in My Tears’: On the Theological
Contribution to the Research of Latin American Pentecostalism.” Exchange 44
(2): 177–200.
Bompani, Barbara, and S. Terreni Brown. 2015. “A ‘Religious Revolution’? Print Media,
Sexuality, and Religious Discourse in Uganda.” Journal of Eastern African Studies 9
(1): 110–26.
Booth, John A., and Mitchell A. Seligson. 2009. The Legitimacy Puzzle in Latin America:
Political Support and Democracy in Eight Latin American Nations. Cambridge,
UK: Cambridge University Press.
Borges, Laryssa. 2010. “TV católica Canção Nova exibe sermão contra voto em Dilma.”
Terra.com, October 6.
References 187

Bornschier, Simon. 2010. “The New Cultural Divide and the Two-Dimensional Political
Space in Western Europe.” West European Politics 33 (3): 419–44. https://doi.org/
10.1080/01402381003654387.
Bruneau, Thomas C. 1980. “The Catholic Church and Development in Latin America:
The Role of the Basic Christian Communities.” World Development 8 (7): 535–44.
1982. The Church in Brazil: The Politics of Religion. Austin, TX: University of Texas
Press.
Burdick, John. 1993a. Looking for God in Brazil: The Progressive Catholic Church in
Urban Brazil’s Religious Arena. Berkeley, CA: University of California Press.
1993b. “Struggling against the Devil: Pentecostalism and Social Movements in Urban
Brazil.” In Rethinking Protestantism in Latin America, edited by Virginia Garrard-
Burnett and David Stoll. Philadelphia: Temple University Press.
Calegari, Luiza. 2017. “O que pensam os manifestantes contra e a favor de Judith
Butler.” Exame, November 15. https://exame.abril.com.br/brasil/o-que-pensam
-os-manifestantes-contra-e-a-favor-de-judith-butler/.
Calfano, Brian Robert, and Paul A. Djupe. 2009. “God Talk: Religious Cues and
Electoral Support.” Political Research Quarterly 62 (2): 329–39.
Calfano, Brian Robert, Melissa R. Michelson, and Elizabeth A. Oldmixon. 2017.
A Matter of Discretion: The Politics of Catholic Priests in the United States and
Ireland. London; Lanham, MD: Rowman & Littlefield.
Calfano, Brian Robert, and Elizabeth A. Oldmixon. 2016. “Remembering to Ask the
Boss: Priming and the Dynamics of Priest Reliance on Bishop Cues.” Religions 7
(3): 21.
Calfano, Brian Robert, Elizabeth A. Oldmixon, and Jane Suiter. 2014. “Who and What
Affects the First Estate? An Analysis of Clergy Attitudes on Cultural and Economic
Issues.” Politics 34 (4): 391–404.
Campbell, David E., John C. Green, and Geoffrey C. Layman. 2011. “The Party Faithful:
Partisan Images, Candidate Religion, and the Electoral Impact of Party
Identification.” American Journal of Political Science 55 (1): 42–58.
Campbell, David E., John C. Green, and J. Quin Monson. 2014. Seeking the Promised
Land: Mormons and American Politics. Cambridge, UK: Cambridge University
Press.
Carranza, Brenda. 2006. “Catolicismo midiático.” In As Religiões no Brasil, edited by
Faustino Teixeira and Renata Menezes, 69–88. Petrópolis, Brazil: Editora Vozes.
Carrapatoso, Miguel Santos. 2016. “Impeachment Dilma. As melhores pérolas dos
deputados para justificarem o voto.” Observador, April 18. http://observador.pt/
2016/04/18/impeachment-dilma-as-melhores-perolas-dos-deputados-justificarem
-voto/.
Castro, Mônica Mata Machado de. 2007. “Eleitorado brasileiro: Composição e grau de
participação.” In Sistema Político Brasileiro: Uma introdução, 25–42. Rio de
Janeiro: Fundação Konrad Adenauer.
Chagas, Tiago. 2016a. “‘Partido da universal’ anuncia que votará a favor do impeachment.”
Notícias Gospel (blog). April 14. https://noticias.gospelmais.com.br/partido-universal
-anuncia-votara-favor-impeachment-82222.html.
2016b. “Primeiro voto do impeachment foi de um evangélico.” Notícias Gospel
(blog). April 18. https://noticias.gospelmais.com.br/primeiro-voto-impeachment
-foi-evangelico-82289.html.
188 References

Chapp, Christopher B. 2012. Religious Rhetoric and American Politics: The Endurance
of Civil Religion in Electoral Campaigns. Ithaca, NY: Cornell University Press.
Charleaux, João Paulo. 2016. “Qual a influência das igrejas evangélicas na política
brasileira.” Nexo Jornal, October 28. https://www.nexojornal.com.br/entrevista/
2016/10/28/Qual-a-influ%C3%AAncia-das-igrejas-evang%C3%A9licas-na-pol%
C3%ADtica-brasileira.
Cheney, Kristen. 2012. “Locating Neocolonialism, ‘Tradition,’ and Human Rights in
Uganda’s ‘Gay Death Penalty.’” African Studies Review 55 (2): 77–95.
Chesnut, R. Andrew. 1999. “The Salvation Army or the Army’s Salvation?: Pentecostal
Politics in Amazonian Brazil, 1962–1992.” Luso-Brazilian Review 36 (2): 33–49.
2003a. “A Preferential Option for the Spirit: The Catholic Charismatic Renewal in
Latin America’s New Religious Economy.” Latin American Politics and Society 45
(1): 55–85.
2003b. Competitive Spirits: Latin America’s New Religious Economy: Latin
America’s New Religious Economy. New York: Oxford University Press.
2009. “Charismatic Competitors: Protestant Pentecostals and Catholic Charismatics
in Latin America’s New Religious Marketplace.” In Religion and Society in Latin
America, edited by Lee M. Penyak and Walter J. Petry, 207–23. Maryknoll, NY:
Orbis Books.
Conrado, Flávio Cesar. 2001. “Política e Mídia: A Igreja Universal do Reino de Deus nas
eleições.” Religião & Sociedade 21 (2): 85–111.
Conroy-Krutz, Jeffrey. 2013. “Information and Ethnic Politics in Africa.” British
Journal of Political Science 43 (02): 345–73.
Conroy-Krutz, Jeffrey, Devra C. Moehler, and Rosario Aguilar. 2016. “Partisan Cues
and Vote Choice in New Multiparty Systems.” Comparative Political Studies, 49
(1): 3–35.
Converse, Philip E. 1964. “The Nature of Belief Systems in Mass Publics.” In Ideology
and Discontent, edited by David Apter, 75–169. London: Free Press of Glencoe.
Corrales, Javier. 2017. “Understanding the Uneven Spread of LGBT Rights in Latin
America and the Caribbean, 1999–2013.” Journal of Research in Gender Studies 7
(1): 52–82.
Corten, André. 1999. Pentecostalism in Brazil: Emotion of the Poor and Theological
Romanticism. New York: St. Martin’s Press.
Coutinho, Raquel Zanatta, and André Braz Golgher. 2014. “The Changing Landscape
of Religious Affiliation in Brazil between 1980 and 2010: Age, Period, and Cohort
Perspectives.” Revista Brasileira de Estudos de População 31 (1): 73–98.
Cowie, Sam. 2018. “Violent Deaths of LGBT People in Brazil Hit All-Time High.”
The Guardian, January 22. www.theguardian.com/world/2018/jan/22/brazil-lgbt
-violence-deaths-all-time-high-new-research.
Cozzens, Donald B. 2000. The Changing Face of the Priesthood: A Reflection on the
Priest’s Crisis of Soul. Collegeville, MN: Liturgical Press.
Cunha, Luiz Antônio. 2009. “A educação na concordata Brasil-Vaticano.” Educação &
Sociedade 30 (106): 263–80. https://doi.org/10.1590/S0101-73302009000100013.
Cunha, Magali Nascimento. 2016a. “Composição da bancada evangélica.” Mídia,
Religião E Cultura (blog).
2016b. “Frentes parlamentares de cunho religioso.” Mídia, Religião e Política (blog).
2016c. “Bancada evangélica em processo de ampliação: Bolsonaro batizado em
Israel.” Mídia, Religião e Política (blog). May 17.
References 189

Dahl, Robert. 1971. Polyarchy: Participation and Opposition. New Haven, CT: Yale
University Press.
Dalton, Russell J., and Christian Welzel. 2014. The Civic Culture Transformed: From
Allegiant to Assertive Citizens. Cambridge, UK: Cambridge University Press.
Dancygier, Rafaela M. 2017. Dilemmas of Inclusion: Muslims in European Politics.
Princeton, NJ: Princeton University Press.
Dantas, Bruna Suruagy do Amaral. 2011. “Religião e política: ideologia e ação da
bancada evangélica na Câmara Federal.” Doctoral dissertation, São Paulo:
Pontifícia Universidade Católica de São Paulo.
Demartini, Maria. 2017. “Mackenzie abre núcleo de estudos que contesta teoria da
evolução.” Exame, May 11. http://exame.abril.com.br/ciencia/mackenzie-abre
-nucleo-de-estudos-que-contesta-teoria-da-evolucao/.
Díaz Domínguez, Alejandro. 2006. “¿Influyen los ministros de culto sobre la intención
de voto?” Perfiles Latinoamericanos 28: 33–57.
Djupe, Paul A., Anand E. Sokhey, and Christopher P. Gilbert. 2007. “Present but Not
Accounted For? Gender Differences in Civic Resource Acquisition.” American
Journal of Political Science 51 (4): 906–20.
Djupe, Paul A., and Brian Calfano. 2014. God Talk: Experimenting with the Religious
Causes of Public Opinion. Philadelphia: Temple University Press.
Djupe, Paul A., and Christopher P. Gilbert. 2002. “The Political Voice of Clergy.”
Journal of Politics 64 (2): 596–609.
2003. The Prophetic Pulpit: Clergy, Churches, and Communities in American
Politics. Lanham, MD: Rowman & Littlefield Publishers.
2006. “The Resourceful Believer: Generating Civic Skills in Church.” Journal of
Politics 68 (1): 116–27.
2008. “Politics and Church: Byproduct or Central Mission?” Journal for the Scientific
Study of Religion 47 (1): 45–62.
2009. The Political Influence of Churches. Cambridge, UK; New York: Cambridge
University Press.
Djupe, Paul A., Jacob R. Neiheisel, and Kimberly H. Conger. 2018. “Are the Politics of
the Christian Right Linked to State Rates of the Nonreligious? The Importance of
Salient Controversy.” Political Research Quarterly, 71 (4): 910–22.
Djupe, Paul A., Scott D. McClurg, and Anand E. Sokhey. 2018. “The Political
Consequences of Gender in Social Networks.” British Journal of Political Science
48 (3): 637–58.
Dowd, Robert A. 2015. Christianity, Islam, and Liberal Democracy: Lessons from Sub-
Saharan Africa. Oxford: Oxford University Press.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper
& Row.
Duverger, Maurice. 1972. Party Politics and Pressure Groups: A Comparative
Introduction. New York: Thomas Y. Crowell.
The Economist. 2018. “Latin America’s Human-Rights Court Moves into Touchy
Territory.” The Economist, February 1. https://www.economist.com/news/americas
/21736199-ruling-favour-gay-marriage-will-provoke-resistance-latin-americas-human
-rights-court.
Engeli, Isabelle, Christoffer Green-Pedersen, and Lars Thorup Larsen. 2012. Morality
Politics in Western Europe: Parties, Agendas and Policy Choices. London: Palgrave
Macmillan.
190 References

2013. “The Puzzle of Permissiveness: Understanding Policy Processes Concerning

Morality Issues.” Journal of European Public Policy 20 (3): 335–52.
Exame. 2013. “Os ateus no Brasil e seu medo de ‘sair Do armário.’” Exame, June 6. http://
exame.abril.com.br/brasil/os-ateus-no-brasil-e-seu-medo-de-sair-do-armario/.
Facchini, Regina. 2010. “Movimento homossexual no Brasil: Recompondo um
histórico.” Cadernos AEL 10 (18/19): 83–123.
Falk, Armin, and Urs Fischbacher. 2006. “A Theory of Reciprocity.” Games and
Economic Behavior 54 (2): 293–315. https://doi.org/10.1016/j.geb.2005.03.001.
Fediakova, Evguenia. 2012. “Evangelicals in Democratic Chile, 1990–2008: From
‘Resistance Identity’ to ‘Project Identity.’” Religion, State and Society 40 (1): 24–48.
Ferreira, Paula, and Renata Mariz. 2017. “CNE retira gênero e orientação sexual da
Base Curricular.” O Globo, December 12. https://oglobo.globo.com/sociedade/
educacao/cne-retira-genero-orientacao-sexual-da-base-curricular-22179063.
Finke, Roger, and Laurence R. Iannaccone. 1993. “Supply-Side Explanations for
Religious Change.” Annals of the American Academy of Political and Social
Science 527 (May): 27–39.
Fiorina, Morris P., and Samuel J. Abrams. 2008. “Political Polarization in the American
Public.” Annual Review of Political Science 11 (1): 563–88.
Fiorina, Morris P., Samuel J. Abrams, and Jeremy C. Pope. 2006. Culture War?
The Myth of a Polarized America. 2nd ed. New York: Pearson Education.
Flanagan, Scott C., and Aie-Rie Lee. 2003. “The New Politics, Culture Wars, and the
Authoritarian-Libertarian Value Change in Advanced Industrial Democracies.”
Comparative Political Studies 36 (3): 235–70.
Flock, Elizabeth. 2013. “Catholic Priests: It’s ‘Empirical Fact’ That Many Clergy Are
Gay.” US News & World Report, July 29. https://www.usnews.com/news/articles/
2013/07/29/catholic-priests-its-empirical-fact-that-many-clergy-are-gay.
Folha de São Paulo. 1994a. “Acredito em Deus’, diz Fernando Henrique.” Folha de São
Paulo, April 2. www1.folha.uol.com.br/fsp/1994/4/02/brasil/16.html.
1994b. “Debate em 1985 afetou candidatura.” Folha de São Paulo, April 2. www1
.folha.uol.com.br/fsp/1994/4/02/brasil/17.html.
Folha Online. 2009a. “AMB critica aprovação de acordo entre Brasil e Vaticano.” Folha
de São Paulo, August 17. www1.folha.uol.com.br/folha/brasil/ult96u610758
.shtml.
2009b. “Câmara aprova Estatuto da Igreja Católica e regulamenta o direito
à liberdade religiosa.” Folha de São Paulo, August 27. www1.folha.uol.com.br
/poder/2009/08/615507-camara-aprova-estatuto-da-igreja-catolica-e-regulamenta
-o-direito-a-liberdade-religiosa.shtml.
Folha Universal. 2010. “Na mão do clero.” Folha Universal 968, October 24.
Fonseca, Alexandre Brasil. 2008. “Religion and Democracy in Brazil: A Study of the
Leading Evangelical Politicians.” In Evangelical Christianity and Democracy in
Latin America, edited by Paul Freston, 163–206. Oxford: Oxford University Press.
Fonseca, André Dioney. 2014. “Informação, política e fé: O jornal Mensageiro da Paz no
contexto de redemocratização do Brasil (1980–1990).” Revista Brasileira de
História 34 (68): 279–302.
Fox, Jonathan. 2015. Political Secularism, Religion, and the State. Cambridge, UK:
Cambridge University Press.
References 191

Freston, Paul. 1993. “Brother Votes for Brother: The New Politics of Protestantism in
Brazil.” In Rethinking Protestantism in Latin America, edited by Virginia Garrard-
Burnett and David Stoll, 66–110. Philadelphia: Temple University Press.
2004. “Evangelical Protestantism and Democratization in Contemporary Latin
America and Asia.” Democratization 11 (4): 21–41.
G1. 2015. “71% dos futuros deputados se dizem Católicos e 16%, Evangélicos.” G1,
January 31. http://g1.globo.com/politica/noticia/2015/01/71-dos-futuros-deputados
-se-dizem-catolicos-e-16-evangelicos.html.
Galindo, Rogerio Waldrigues. 2016. “93% dos deputados da bancada evangélica votaram
pelo impeachment.” Caixa Zero, April 17. www.gazetadopovo.com.br/blogs/caixa
-zero/93-dos-deputados-da-bancada-evangelica-votaram-pelo-impeachment/.
Gandini, Arthur. 2015. “Cerca de 75% do clero é gay, estima ex-sacerdote católico.”
A Igreja Paulistana, May 30. www.igrejapaulistana.com/2015/05/cerca-de-75-do
-clero-e-gay-diz-ex.html.
Gans-Morse, Jordan, Sebastián Mazzuca, and Simeon Nichter. 2014. “Varieties of
Clientelism: Machine Politics during Elections.” American Journal of Political
Science 58 (2): 415–32.
Garrard-Burnett, Virginia. 2009. “‘Like a Mighty Rushing Wind’: The Growth of
Pentecostalism in Contemporary Latin America.” In Religion and Society in Latin
America, edited by Lee M. Penyak and Walter J. Petry, 190–202. Maryknoll,
New York: Orbis Books.
Gaskill, Newton Jeffrey. 2002. “Power from on High: The Political Mobilization of
Brazilian Evangelical Protestantism.” Ph.D. Thesis, Austin, Texas: University of
Texas at Austin. http://repositories.lib.utexas.edu/handle/2152/11613.
Geddes, Barbara. 2003. Paradigms and Sand Castles: Theory Building and Research
Design in Comparative Politics. Ann Arbor, MI: University of Michigan Press.
Gill, Anthony. 1994. “Rendering unto Caesar: Religious Competition and Catholic
Political Strategy in Latin America, 1962–79.” American Journal of Political
Science 38 (2): 403–25.
1995. “The Institutional Limitations of Catholic Progressivism: An Economic
Approach.” International Journal of Social Economics 22 (9/10/11): 135–48.
1998. Rendering unto Caesar: The Catholic Church and the State in Latin America.
Chicago: University of Chicago Press.
Glazier, Rebecca A. 2013. “Divine Direction: How Providential Religious Beliefs Shape
Foreign Policy Attitudes.” Foreign Policy Analysis 9 (2): 127–42.
Gonçalves, Juliana. 2017. “‘Queimem a bruxa!’ Visita de Judith Butler
provoca manifestações nas ruas de São Paulo.” The Intercept (blog). November 7.
https://theintercept.com/2017/11/07/judith-butler-bruxa-manifestacoes-sao-paulo
-ideologia-genero/.
Gonçalves, Rafael Bruno. 2011. “Bancada evangélica?: Uma análise do discurso
parlamentar evangélico durante a 52a Legislatura da Câmara Federal.” Master’s
Thesis, Universidade Federal de Pelotas. http://repositorio.ufpel.edu.br:8080/
handle/123456789/1591.
Goren, Paul, and Christopher Chapp. 2017. “Moral Power: How Public Opinion on
Culture War Issues Shapes Partisan Predispositions and Religious Orientations.”
American Political Science Review 111 (1): 110–28.
Gouldner, Alvin W. 1960. “The Norm of Reciprocity: A Preliminary Statement.”
American Sociological Review 25 (2): 161–78. https://doi.org/10.2307/2092623.
192 References

Green, James N. 2015. “‘Mais amor e mais tesão’: a construção de um movimento

brasileiro de gays, lésbicas e travestis.” Cadernos Pagu (15): 271–95.
Green, James N., Marisa Fernandes, Cláudia Regina, Edward MacRae,
Luiz Amorim, Júlio Simões, Peter Fry, and Marcos Rodrigues. 2010. “MESA-
REDONDA Somos – Grupo de Afirmação Homossexual: 24 anos depois.
Reflexões sobre os primeiros momentos do movimento homossexual no
Brasil.” Cadernos AEL 10 (18/19). www.ifch.unicamp.br/ojs/index.php/ael/
article/view/2509.
Grim, Brian J., and Roger Finke. 2006. “International Religion Indexes: Government
Regulation, Government Favoritism, and Social Regulation of Religion.”
Interdisciplinary Journal of Research on Religion 2 (1). www.ncbi.nlm.nih.gov/
pmc/articles/PMC4254791/.
Grossman, Guy. 2015. “Renewalist Christianity and the Political Saliency of LGBTs:
Theory and Evidence from Sub-Saharan Africa.” The Journal of Politics 77 (2):
337–51.
Grzymala-Busse, Anna. 2015. Nations under God: How Churches Use Moral Authority
to Influence Policy. Princeton, NJ: Princeton University Press.
Guiame. 2011. “Bolsonaro: Sou um católico que frequentou a Igreja Batista por 10
anos.” Guiame, July 4. http://guiame.com.br/gospel/mundo-cristao/bolsonaro-sou
-um-catolico-que-frequentou-a-igreja-batista-por-10-anos.html.
Gurr, Ted Robert. 1970. Why Men Rebel. Princeton, NJ: Princeton University Press.
Gutmann, Amy, and Dennis Thompson. 1996. Democracy and Disagreement.
Cambridge, MA: Belknap Press of Harvard University Press.
Hagopian, Frances. 2008. “Latin American Catholicism in an Age of Religious and
Political Pluralism: A Framework for Analysis.” Comparative Politics 40 (2):
149–68.
Hale, Christopher W. 2018. “Religious Institutions and Collective Action: The Catholic
Church and Political Activism in Indigenous Chiapas and Yucatán.” Politics and
Religion 11 (1): 27–54.
Henrich, Joseph, Jean Ensminger, and Richard McElreath, et al. 2010. “Markets,
Religion, Community Size, and the Evolution of Fairness and Punishment.”
Science 327 (5972): 1480–84.
Hetherington, Marc J. 2001. “Resurgent Mass Partisanship: The Role of Elite
Polarization.” American Political Science Review 95 (3): 619–31.
Hetherington, Marc J., and Jonathan Weiler. 2009. Authoritarianism and Polarization
in American Politics. Cambridge, UK: Cambridge University Press.
Heumann, Silke, and Jan Willem Duyvendak. 2015. “When and Why Religious Groups
Become Political Players.” In Players and Arenas: The Interactive Dynamics of
Protest, edited by James M. Jasper and Jan Willem Duyvendak, 251–74.
Amsterdam: Amsterdam University Press.
Holman, Mirya R., and Kristin Shockley. 2017. “Messages from Above: Conflict and
Convergence of Messages to the Catholic Voter from the Catholic Church
Hierarchy.” Politics and Religion, July, 1–22.
Huckfeldt, Robert, Eric Plutzer, and John Sprague. 1993. “Alternative Contexts of
Political Behavior: Churches, Neighborhoods, and Individuals.” The Journal of
Politics 55 (2): 365–81.
References 193

Huckfeldt, Robert, and John Sprague. 1995. Citizens, Politics, and Social
Communication: Information and Influence in an Election Campaign.
Cambridge, UK: Cambridge University Press.
Hunter, James Davison. 1992. Culture Wars: The Struggle to Control the Family, Art,
Education, Law, and Politics in America. New York: Basic Books.
Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth
Century. Norman, OK: University of Oklahoma Press.
Inglehart, Ronald. 1990. Culture Shift in Advanced Industrial Society. Princeton, NJ:
Princeton University Press.
1997. Modernization and Postmodernization. Princeton, NJ: Princeton University
Press.
Inglehart, Ronald, and Pippa Norris. 2003. Rising Tide: Gender Equality & Cultural
Change around the World. Cambridge, UK: Cambridge University Press.
Ireland, Rowan. 1993. “The Crentes of Campo Alegre and the Religious Construction of
Brazilian Politics.” In Rethinking Protestantism in Latin America, edited by
Virginia Garrard-Burnett and David Stoll, 45–65. Philadelphia: Temple University
Press.
1999. “Popular Religions and the Building of Democracy in Latin America: Saving the
Tocquevillian Parallel.” Journal of Interamerican Studies and World Affairs 41 (4):
vi–136.
Iyengar, Shanto, and Sean J. Westwood. 2015. “Fear and Loathing across Party Lines:
New Evidence on Group Polarization.” American Journal of Political Science 59
(3): 690–707.
Jacob, Cesar Romero, Dora Rodrigues Hees, and Philippe Waniez. 2013. Religião
e Território no Brasil, 1990–2010. Rio de Janeiro: Editora PUC-Rio.
Jacobs, Andrew. 2016. “Brazil Is Confronting an Epidemic of Anti-Gay Violence.”
The New York Times, July 5, sec. Americas. www.nytimes.com/2016/07/06/
world/americas/brazil-anti-gay-violence.html.
Jacoby, William G. 2014. “Is There a Culture War? Conflicting Value Structures in
American Public Opinion.” American Political Science Review 108 (04): 754–71.
Jelen, Ted G. 1992. “Political Christianity: A Contextual Analysis.” American Journal of
Political Science 36 (3): 692–714.
Jenkins, Philip. 2002. The Next Christendom: The Rise of Global Christianity.
New York: Oxford University Press.
Kaoma, Kapya. 2009. Globalizing the Culture Wars: US Conservatives, African
Churches, and Homophobia. Somerville, MA: Political Research Associates.
2014. “The Paradox and Tension of Moral Claims: Evangelical Christianity, the
Politicization and Globalization of Sexual Politics in Sub-Saharan Africa.”
Critical Research on Religion 2 (3): 227–45.
Keck, Margaret E. 1992. The Workers’ Party and Democratization in Brazil. New
Haven, CT: Yale University Press.
King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Social Inquiry:
Scientific Inference in Qualitative Research. Princeton, NJ: Princeton University
Press.
Kitschelt, Herbert. 2000. “Linkages between Citizens and Politicians in Democratic
Polities.” Comparative Political Studies 33 (6–7): 845–79.
194 References

Koleva, Spassena P., Jesse Graham, Ravi Iyer, Peter H. Ditto, and Jonathan Haidt. 2012.
“Tracing the Threads: How Five Moral Concerns (Especially Purity) Help Explain
Culture War Attitudes.” Journal of Research in Personality 46 (2): 184–94.
Kuru, Ahmet T. 2008. “Passive and Assertive Secularism: Historical Conditions,
Ideological Struggles, and State Policies toward Religion.” World Politics 59 (4):
568–94.
2009. Secularism and State Policies toward Religion: The United States, France, and
Turkey. New York: Cambridge University Press.
Laakso, Markku, and Rein Taagepera. 1979. “‘Effective’ Number of Parties: A Measure
with Application to West Europe.” Comparative Political Studies 12: 3–27.
Lankina, Tomila, and Lullit Getachew. 2012. “Mission or Empire, Word or Sword?
The Human Capital Legacy in Postcolonial Democratic Development.” American
Journal of Political Science 56 (2): 465–83.
Lasswell, Harold Dwight. 2011. Politics: Who Gets What, When, How. Whitefish, MT:
Literary Licensing, LLC.
Layman, Geoffrey C., and Thomas M. Carsey. 2002. “Party Polarization and ‘Conflict
Extension’ in the American Electorate.” American Journal of Political Science 46
(4): 786–802.
Layman, Geoffrey C., and John C. Green. 2006. “Wars and Rumours of Wars:
The Contexts of Cultural Conflict in American Political Behaviour.” British
Journal of Political Science 36 (1): 61–89.
Layman, Geoffrey C., Thomas M. Carsey, and Juliana Menasce Horowitz. 2006. “Party
Polarization in American Politics: Characteristics, Causes, and Consequences.”
Annual Review of Political Science 9 (1): 83–110.
Lazarsfeld, Paul F., Bernard R. Berelson, and Hazel Gaudet. 1948. The People’s Choice:
How the Voter Makes Up His Mind in a Presidential Campaign. New York:
Columbia University Press.
Lehmann, David. 1996. Struggle for the Spirit: Religious Transformation and Popular
Culture in Brazil and Latin America. Oxford: Polity Press.
Lemos, Christina, and Douglas Tavolaro. 2007. O Bispo: A História Revelada de Edir
Macedo. São Paulo: Editora Larousse do Brasil.
Levine, Daniel H. 2009. “The Future of Christianity in Latin America.” Journal of Latin
American Studies 41: 121–45.
2012. Politics, Religion & Society in Latin America. Boulder, CO: Lynne Rienner
Publishers.
Levitsky, Steven, and Daniel Ziblatt. 2018. How Democracies Die. New York: Crown/
Archetype.
Lewis, Andrew R. 2017. The Rights Turn in Conservative Christian Politics: How
Abortion Transformed the Culture Wars. Cambridge, UK: Cambridge University
Press.
Lisboa, Fábio Aguiar. 2010. “Qual o papel da religião nas últimas eleições?” O Jornal
Batista, November 6. www.adiberj.org/portal/2010/11/06/qual-o-papel-da-religiao
-nas-ultimas-eleicoes/.
Løland, Ole Jakob. 2015. “The Position of the Biblical Canon in Brazil: From Catholic
Rediscovery to Neo-Pentecostal Marginalisation.” Studies in World Christianity 21
(2): 98–118.
References 195

Lupu, Noam. 2013. “Party Brands and Partisanship: Theory with Evidence from
a Survey Experiment in Argentina.” American Journal of Political Science 57 (1):
49–64.
Mainwaring, Scott. 1986. The Catholic Church and Politics in Brazil, 1916–1985.
Stanford University Press.
Mainwaring, Scott, Rachel Meneguello, and Timothy J. Power. 2000. Partidos
Conservadores No Brasil Contemporâneo: Quais São, O Que Defendem, Quais
São Suas Bases. São Paulo: Paz e Terra.
Maldonado, Arturo. 2011. “Compulsory Voting and the Decision to Vote.” 63.
AmericasBarometer Insights. Latin American Public Opinion Project. Vanderbilt
University.
Mali, Tiago. 2014. “Candidatos evangélicos crescem quase 50%.” Época, September
16. http://epoca.globo.com/tempo/eleicoes/noticia/2014/09/bcandidatos-evangelicosb
-crescem-quase-50.html.
Malta, Magno. 2014. “Speech to the Brazilian Federal Senate.” Brasília, August 20.
www25.senado.leg.br/web/atividade/pronunciamentos/-/p/texto/409272.
Margolis, Michele F. 2016. “Cognitive Dissonance, Elections, and Religion: How
Partisanship and the Political Landscape Shape Religious Behaviors.” Public
Opinion Quarterly 80 (3): 717–40.
2018. From Politics to the Pews: How Partisanship and the Political Environment
Shape Religious Identity. Chicago: University of Chicago Press.
Mariano, Ricardo. 2011. “Laicidade à brasileira: católicos, pentecostais e laicos em
disputa na esfera pública.” Civitas – Revista de Ciências Sociais 11 (2): 238–58.
Mariano, Ricardo, and Antônio Flávio Pierucci. 1992. “O envolvimento dos
pentecostais na eleição de Collor.” Novos Estudos Cebrap 34: 92–106.
Marietta, Morgan. 2008. “From My Cold, Dead Hands: Democratic Consequences of
Sacred Rhetoric.” The Journal of Politics 70 (3): 767–79.
Mariz, Cecília Loreto. 2006. “Catolicismo no Brasil contemporâneo: Reavivamento
e diversidade.” In As Religiões No Brasil, edited by Faustino Teixeira and
Renata Menezes, 53–68. Petrópolis, Brazil: Editora Vozes.
Mariz, Cecília Loreto, and Maria das Dores Campos Machado. 1997. “Pentecostalism
and Women in Brazil.” In Power, Politics, and Pentecostals in Latin America, edited
by Edward L. Cleary and Hannah W. Stewart-Gambino, 41–54. Boulder, CO:
Westview Press.
Marshall, Ruth. 2009. Political Spiritualities: The Pentecostal Revolution in Nigeria. 1st
ed. Chicago and London: University of Chicago Press.
Martin, Bernice. 2006. “Pentecostal Conversion and the Limits of the Market
Metaphor.” Exchange 35 (1): 61–91.
Martin, David. 1993. Tongues of Fire: The Explosion of Protestantism in Latin America.
Oxford: Wiley.
Mason, Lilliana. 2015. “‘I Disrespectfully Agree’: The Differential Effects of Partisan
Sorting on Social and Issue Polarization.” American Journal of Political Science 59
(1): 128–45.
2018. Uncivil Agreement: How Politics Became Our Identity. Chicago: University of
Chicago Press.
Mayrl, Damon. 2016. Secular Conversions: Political Institutions and Religious
Education in the United States and Australia, 1800–2000. Cambridge, UK:
Cambridge University Press.
196 References

McAdams, Erin S., and Justin Earl Lance. 2013. “Religion’s Impact on the Divergent
Political Attitudes of Evangelical Protestants in the United States and Brazil.”
Politics and Religion 6 (03): 483–511.
McClendon, Gwyneth, and Rachel Beatty Riedl. 2015. “Religion as a Stimulant of
Political Participation: Experimental Evidence from Nairobi, Kenya.” Journal of
Politics 77 (4): 1045–57.
McClurg, Scott D. 2006. “The Electoral Relevance of Political Talk: Examining
Disagreement and Expertise Effects in Social Networks on Political
Participation.” American Journal of Political Science 50 (3): 737–54.
McCrudden, Christopher. 2015. “Transnational Culture Wars.” International Journal
of Constitutional Law 13 (2): 434–62.
McDermott, Monika L. 2009. “Religious Stereotyping and Voter Support for
Evangelical Candidates.” Political Research Quarterly 62 (2): 340–54.
Menchik, Jeremy. 2016. Islam and Democracy in Indonesia: Tolerance without
Liberalism. Cambridge, UK: Cambridge University Press.
Mendonça, Antônio Gouvea. 2006. “Evangélicos e pentecostais: Um campo religioso em
ebulição.” In As Religiões No Brasil, edited by Faustino Teixeira and
Renata Menezes, 89–110. Petrópolis, Brazil: Editora Vozes.
Migheli, Matteo. 2017. “The Gospel and Economic Behaviour: Experimental Evidence
from a Trust Game.” Annals of Public and Cooperative Economics 33 (1): 33–45.
Mir, Luís. 2007. Partido de Deus: Fé, Poder e Política. São Paulo: Alaúde Editorial.
Miranda, Ricardo. 2008a. “ELEIÇÕES 2008: Carta Divulgada.” Tribuna de Minas.
2008b. “Evangélicos querem um quarto das cadeiras da Câmara.” Tribuna de Minas.
Mora, G. Cristina. 2008. “Marketing the ‘Health and Wealth Gospel’ across National
Borders: Evidence from Brazil and the United States.” Poetics, Religion and Culture
36 (5–6): 404–20.
Mora Torres, Jose Enrique. 2010. “The Political Incorporation of Pentecostals in
Panama, Puerto Rico and Brazil: A Comparative Analysis.” Unpublished doctoral
dissertation, Storrs, CT: University of Connecticut. https://opencommons
.uconn.edu/dissertations/AAI3451309.
Moraes, Manoel Ribeiro de. 2010. “ELEIÇÕES 2010: Democracia versus iniquidade:
O purismo religioso à disposição do retrocesso democrático.” Novos Diálogos:
Jornal de Justiça, Espiritualidade e Cultura (online).
Morgan, William, and Christopher Winship. 2007. Counterfactuals and Causal
Inference. Cambridge, UK: Cambridge University Press.
Morrison, Penelope Kay. 2010. “Democracy ‘At Risk’? Governmental and Non-
Governmental Organizations, ‘At Risk’ Youth and Programming in Juiz de Fora,
Brazil.” Unpublished doctoral dissertation. Pittsburgh, PA: University of
Pittsburgh.
Moseley, Mason W. 2015. “Contentious Engagement: Understanding Protest
Participation in Latin American Democracies.” Journal of Politics in Latin
America 7 (3): 3–48. http://d-scholarship.pitt.edu/9046/.
Moseley, Mason W., and Matthew L. Layton. 2013. “Prosperity and Protest in Brazil:
The Wave of the Future for Latin America?” 93. AmericasBarometer Insights.
Mucinhato, Rafael Moreira Dardaque. 2014. “Quem são os deputados brasileiros? Um
balanço do perfil biográfico de 1986 a 2012.” In O Congresso Nacional, Os
Partidos Políticos e o Sistema de Integridade: Representação, Participação
References 197

e Controle Interinstitucional no Brasil Contemporâneo, edited by José

Álvaro Moisés, 61–88. Rio de Janeiro: Fundação Konrad Adenauer.
Neto, João. 2015. “Dilma promete a pastores isentar igrejas da CPMF e deputado critica:
‘Um tiro no pé’.” Guiame, October 2. http://guiame.com.br/gospel/mundo-cristao/
dilma-se-reune-com-pastores-para-propor-isencao-da-cpmf-e-deputado-critica-um
-tiro-no-pe.html.
Netto, Gabriela Figueiredo. 2016. “Quando o dinheiro importa menos: uma análise do
financiamento de campanhas eleitorais dos candidatos evangélicos.” Thesis, São
Paulo: Universidade de São Paulo.
Nichter, Simeon. 2008. “Vote Buying or Turnout Buying? Machine Politics and the
Secret Ballot.” American Political Science Review 102 (1): 19–31.
Nickerson, David W. 2008. “Is Voting Contagious? Evidence from Two Field
Experiments.” American Political Science Review 102 (1): 49–57.
Nielsen, Richard A. 2017. Deadly Clerics: Blocked Ambition and the Paths to Jihad.
Cambridge, UK: Cambridge University Press.
Nishimura, Katia Mika. 2004. “Conservadorismo social: Opiniões e atitudes no
contexto da eleição de 2002.” Opinião Pública 10 (2).
Norris, Pippa, and Ronald Inglehart. 2004. Sacred and Secular: Religion and Politics
Worldwide. 2nd ed. Cambridge, UK: Cambridge University Press.
Novaes, Regina Reyes. 2002. “Crenças religiosas e convicções políticas: Fronteiras
e passagens.” In Política e Cultura: Século XXI, edited by Luis Carlos Fridman.
Rio de Janeiro: Relume Dumará.
Offutt, Stephen. 2015. New Centers of Global Evangelicalism in Latin America and
Africa. Cambridge: Cambridge University Press.
Oliveira, Alandeom W., and Kristin L. Cook. 2018. “Evolution Education and the Rise
of the Creationist Movement in Brazil.” In Evolution Education around the Globe,
edited by Hasan Deniz and Lisa A. Borgerding. Cham, Switzerland: Springer
International Publishing.
Oliver, Marcia. 2012. “Transnational Sex Politics, Conservative Christianity, and
Antigay Activism in Uganda.” Studies in Social Justice 7 (1): 83–105.
Olson, Mancur. 1971. The Logic of Collective Action: Public Goods and the Theory of
Groups. Cambridge, MA: Harvard University Press.
Oro, Ari Pedro. 2003a. “A política da Igreja Universal e seus reflexos nos campos
religioso e político brasileiros.” Revista Brasileira de Ciências Sociais 18 (53):
53–69.
2003b. “Organização eclesial e eficácia política: O caso da Igreja Universal do Reino
de Deus.” Civitas – Revista de Ciências Sociais 3 (1): 97–109.
2006. “Religião e política no Brasil.” In Religião e política no Cone Sul: Argentina,
Brasil e Uruguai, edited by Ari Pedro Oro, 75–156. São Paulo: Attar Editorial.
Orrico, Alexandre. 2016. “De onde vem essa história de que Michel Temer é Satanista?”
BuzzFeed (blog). May 16. https://www.buzzfeed.com/alexandreorrico/michel
-temer-satanista.
Pateman, Carole. 1970. Participation and Democratic Theory. Cambridge, UK:
Cambridge University Press.
Petty, Richard E., and John T. Cacioppo. 1986. “The Elaboration Likelihood Model of
Persuasion.” Edited by Thomas C. Kinnear. Advances in Experimental Social
Psychology 19: 123–92.
198 References

Pew Research Center. 2006. Spirit and Power: A 10-Country Survey of Pentecostals.
Washington, DC: Pew Forum on Religion and Public Life. www.pewforum.org/
Christian/Evangelical-Protestant-Churches/Spirit-and-Power.aspx.
2011. Global Christianity: A Report on the Size and Distribution of the World’s
Christian Denominations. www.pewforum.org/2011/12/19/global-christianity
-traditions/.
Phillips, Dom. 2015. “Afro-Brazilian Religions Struggle against Evangelical Hostility.”
The Washington Post, February 5. www.washingtonpost.com/world/the_americas/
afro-brazilian-religions-struggle-against-evangelical-hostility/2015/02/05/
b6a30c6e-aaf9-11e4-8876-460b1144cbc1_story.html.
Philpott, Daniel. 2007. “Explaining the Political Ambivalence of Religion.”
The American Political Science Review 101 (3): 505–25.
2009. “Has the Study of Global Politics Found Religion?” SSRN Scholarly Paper ID
1416404. Rochester, NY: Social Science Research Network.
Pierucci, Antônio Flávio, and Reginaldo Prandi. 1995. “Religiões e voto: a eleição
presidencial de 1994.” Opinião Pública 3 (1): 32–63.
Pitkin, Hanna Fenichel. 1967. The Concept of Representation. Berkeley, CA: University
of California Press.
Polêmica Paraíba. 2016. “VEJA VÍDEO – Marcelo Crivella Diz ‘gay é fruto de aborto
malsucedido.’” Polêmica Paraíba. www.polemicaparaiba.com.br/politica/veja
-video-marcelo-crivella-diz-gay-e-fruto-de-aborto-malsucedido/.
Pope Francis. 2013. “Evangelii Gaudium: Apostolic Exhortation on the Proclamation
of the Gospel in Today’s World.” Vatican: The Holy See. http://w2.vatican.va
/content/francesco/en/apost_exhortations/documents/papa-francesco_esortazione
-ap_201311-24_evangelii-gaudium.html.
Power, Timothy J. 2009. “Compulsory for Whom? Mandatory Voting and Electoral
Participation in Brazil, 1986–2006.” Journal of Politics in Latin America 1 (1):
97–122.
Power, Timothy J., and Cesar Zucco. 2012. “Elite Preferences in a Consolidating
Democracy: The Brazilian Legislative Surveys, 1990–2009.” Latin American
Politics and Society 54 (4): 1–27.
Power, Timothy J., and Rodrigo Rodrigues-Silveira. 2018. “The Political Right and
Party Politics.” In The Routledge Handbook of Brazilian Politics, edited by
Barry Ames. New York and London: Routledge.
Przeworski, Adam. 1999. Democracy and the Market Political and Economic Reforms
in Eastern Europe and Latin America. New York: Cambridge University Press.
Putnam, Robert D., and David E. Campbell. 2012. American Grace: How Religion
Divides and Unites Us. 2nd ed. New York: Simon & Schuster.
Ranquetat Júnior, Cesar Alberto. 2016. Laicidade à brasileira: Um estudo sobre
a controvérsia em torno da presença de símbolos religiosos em espaços públicos.
Jundiaí, São Paulo: Paco Editorial.
Rayside, David, and Clyde Wilcox. 2011. Faith, Politics, and Sexual Diversity in Canada
and the United States. Vancouver: UBC Press.
Redação Pragmatismo. 2016. “Maluf e bancada evangélica declaram apoio ao
impeachment de Dilma.” Pragmatismo Político (blog). April 7. https://www
.pragmatismopolitico.com.br/2016/04/maluf-e-bancada-evangelica-declaram-apoio
-ao-impeachment-de-dilma.html.
References 199

Reich, Gary, and Pedro dos Santos. 2013. “The Rise (and Frequent Fall) of Evangelical
Politicians: Organization, Theology, and Church Politics.” Latin American Politics
and Society 55 (4): 1–22.
Rink, Anselm. 2018. “Do Protestant Missionaries Undermine Political Authority?
Evidence from Peru.” Comparative Political Studies 51 (4): 477–513.
Robbins, Joel. 2004. “The Globalization of Pentecostal and Charismatic Christianity.”
Annual Review of Anthropology 33 (1): 117–43.
Ryan, Timothy J. 2014. “Reconsidering Moral Issues in Politics.” Journal of Politics 76
(02): 380–97.
Ryfe, David M. 2005. “Does Deliberative Democracy Work?” Annual Review of
Political Science 8 (1): 49–71.
Sadgrove, Joanna, Robert M. Vanderbeck, Johan Andersson, Gill Valentine, and
Kevin Ward. 2012. “Morality Plays and Money Matters: Towards a Situated
Understanding of the Politics of Homosexuality in Uganda.” The Journal of
Modern African Studies 50 (1): 103–29.
Salazar, Camila. 2017. “Voto por diputados evangélicos se triplicó en cinco elecciones.”
La Nación, June 26. www.nacion.com/gnfactory/investigacion/2017/partidos
-evangelicos/index.html?desktop=true.
Sampaio, Tatiana de Souza. 2016. “A influência das organizações não governamentais
na política brasileira de direitos humanos LGBT+.” Paper presented at Entre
o global e o local, Conference XVII, Encontro de História da Anpuh-Rio. Rio de
Janeiro, August 8–11.
Samuels, David, and Cesar Zucco. 2014. “The Power of Partisanship in Brazil: Evidence
from Survey Experiments.” American Journal of Political Science 58 (1): 212–25.
2018. Partisans, Anti-Partisans, and Non-Partisans: Voting Behavior in Brazil.
Cambridge, UK: Cambridge University Press.
Sandal, Nukhet A. 2017. Religious Leaders and Conflict Transformation: Northern
Ireland and Beyond. Cambridge, UK: Cambridge University Press.
Sant’Anna, Lourival. 2010. “Eleição mostra influência das igrejas.” O Estado de São
Paulo, October 10. https://brasil.estadao.com.br/noticias/geral,eleicao-mostra
-influencia-das-igrejas-imp-,623104.
Santos, André Luis dos. 2009. “O processo de formação das atitudes políticas dos jovens
da Assembléia de Deus em Porto Alegre/RS.” Revista historiador 2(2): 364–98.
Sartori, Giovanni. 1976. Parties and Party Systems: A Framework for Analysis. Vol. 1.
Cambridge, UK: Cambridge University Press.
Schiavon, Fabiana. 2009. “Acordo de Brasil e Vaticano é aprovado em comissão.”
Consultor Jurídico, August 13. www.conjur.com.br/2009-ago-13/acordo-entre
-brasil-vaticano-aprovado-comissao-camara.
Schwartz, Mildred A., and Raymond Tatalovich. 2009. “Cultural and Institutional
Factors Affecting Political Contention over Moral Issues.” Comparative Sociology
8 (1): 76–104.
Serbin, Kenneth P. 2000. Secret Dialogues: Church-State Relations, Torture, and Social
Justice in Authoritarian Brazil. Pittsburgh, PA: University of Pittsburgh Press.
Serbin, Ken. 2006. Needs of the Heart: A Social and Cultural History of Brazil’s Clergy
and Seminaries. Notre Dame, IN: University of Notre Dame Press.
Shariff, Azim F., and Mijke Rhemtulla. 2012. “Divergent Effects of Beliefs in Heaven
and Hell on National Crime Rates.” PLoS ONE 7 (6): e39048.
200 References

Simoni Junior, Sergio, Rafael Moreira Dardaque Mucinhato, and Lucas

Malta Mingardi. 2015. “O perfil biográfico do ‘alto clero’ da câmara dos
deputados no Brasil (1995–2010).” In Elites e poder politico, September 9–11.
Porto Alegre, RS, Brazil.
2016. “A elite parlamentar brasileira de 1995 a 2010: Até que ponto vai
a popularização da classe política?” Colombia Internacional 87: 109–143.
Singh, Shane. 2011. “How Compelling Is Compulsory Voting? A Multilevel Analysis of
Turnout.” Political Behavior 33 (1): 95–111.
Smith, Amy Erica. 2016a. “Talking It Out: Political Conversation and Knowledge Gaps
in Unequal Urban Contexts.” British Journal of Political Science, January, 1–19.
2016b. “When Clergy Are Threatened: Catholic and Protestant Leaders and Political
Activism in Brazil.” Politics and Religion 9 (3): 431–55.
2017. “Democratic Talk in Church: Religion and Political Socialization in the Context
of Urban Inequality.” World Development 99 (November 2017): 441–51.
2018. “Religion, Politics, and the Secular State.” In The Routledge Handbook of
Brazilian Politics, edited by Barry Ames. New York and London: Routledge.
Smith, Gregory Allen. 2008. Politics in the Parish: The Political Influence of Catholic
Priests. Washington, DC: Georgetown University Press.
Smith, T. Alexander, and Raymond Tatalovich. 2003. Cultures at War: Moral Conflicts
in Western Democracies. Peterborough, Ontario: University of Toronto Press,
Higher Education Division.
Souza, André Ricardo de. 2007. “Igreja católica e mercados: A ambivalência entre
a solidariedade e a competição.” Religião & Sociedade 27 (1): 156–74.
Souza, Lidyane Maria Ferreira de. 2016. “The 2008 Concordat in Brazil: ‘Modern
Public Religion’ or Neo-Corporatism?” In The Social Equality of Religion or
Belief, edited by Alan Carling. New York: Palgrave Macmillan.
Souza, Sandra Duarte de. 2013. “Política Religiosa E Religião Política: Os Evangélicos
E O Uso Político Do Sexo.” Estudos de Religião 27 (1): 177–201.
2014. “‘Não À Ideologia de Gênero!’ A Produção Religiosa Da Violência de Gênero
Na Política Brasileira.” Estudos de Religião 28 (2): 188–204.
Stahlhoefer, Alexander. 2016. “O Batismo de Bolsonaro.” Xadrez Verbal (blog). May
16, 2016. https://xadrezverbal.com/2016/05/16/o-batismo-de-bolsonaro/.
Stark, Rodney, and William Sims Bainbridge. 1996. A Theory of Religion. New
Brunswick, NJ: Rutgers University Press.
Steigenga, Timothy J. 2003. The Politics of the Spirit: The Political Implications of
Pentecostalized Religion in Costa Rica and Guatemala. Lanham, MD: Lexington
Books.
Steigenga, Timothy J., and Edward L. Cleary. 2007. Conversion of a Continent:
Contemporary Religious Change in Latin America. New Brunswick, NJ: Rutgers
University Press.
Stockwell, Eugene L. 1995. “Open and Closed: Protestantism in Latin America.”
Christian Century 112 (10).
Stokes, Susan C., Thad Dunning, Marcelo Nazareno, and Valeria Brusco. 2013.
Brokers, Voters, and Clientelism: The Puzzle of Distributive Politics. New York:
Cambridge University Press.
Stoll, David. 1990. Is Latin America Turning Protestant?: The Politics of Evangelical
Growth. Berkeley, CA: University of California Press.
References 201

Stroop, Christopher. 2016. “A Right-Wing International? Russian Social Conservatism,

the World Congress of Families, and the Global Culture Wars in Historical
Context.” The Public Eye, Winter 2016.
Studlar, Donley T. 2012. “The USA and Western Europe Compared: How the ‘God Gap’
Led the USA to Join the Religious World of Morality Politics.” In Morality Politics
in Western Europe, edited by Isabelle Engeli, Christoffer Green-Pedersen, and Lars
Thorup Larsen, 161–84. London: Palgrave Macmillan.
Studlar, Donley T., and Gordon J. Burns. 2015. “Toward the Permissive Society?
Morality Policy Agendas and Policy Directions in Western Democracies.” Policy
Sciences 48 (3): 273–91.
Studlar, Donley T., Alessandro Cagossi, and Robert D. Duval. 2013. “Is Morality Policy
Different? Institutional Explanations for Post-War Western Europe.” Journal of
European Public Policy 20 (3): 353–71.
Sugden, Robert. 1984. “Reciprocity: The Supply of Public Goods through Voluntary
Contributions.” The Economic Journal 94 (376): 772–87.
Tan, Jonathan H. W., and Claudia Vogel. 2008. “Religion and Trust: An Experimental
Study.” Journal of Economic Psychology 29 (6): 832–48.
Tarrow, Sidney. 1998. Power in Movement: Social Movements and Contentious
Politics. New York: Cambridge University Press.
Tavares, Flavia. 2014. “O jeito evangélico de pedir votos e fazer política.” Época,
September 21. http://epoca.globo.com/tempo/eleicoes/noticia/2014/09/o-jeito
-bevangelicob-de-pedir-votos-e-fazer-politica.html.
Tocqueville, Alexis de. 2010. Democracy in America. Translated by James T. Schleifer.
Indianapolis: The Liberty Fund.
Toft, Monica Duffy, Daniel Philpott, and Timothy Samuel Shah. 2011. God’s Century:
Resurgent Religion and Global Politics. New York: W. W. Norton & Company.
Toft, Monica Duffy, and Yuri M. Zhukov. 2015. “Islamists and Nationalists: Rebel
Motivation and Counterinsurgency in Russia’s North Caucasus.” American
Political Science Review 109 (02): 222–38.
Trejo, Guillermo. 2009. “Religious Competition and Ethnic Mobilization in Latin
America: Why the Catholic Church Promotes Indigenous Movements in Mexico.”
American Political Science Review 103 (03): 323–42.
2014. Popular Movements in Autocracies: Religion, Repression, and Indigenous
Collective Action in Mexico. Reprint edition. Cambridge, UK: Cambridge
University Press.
Tribunal Superior Eleitoral. 2012. Recurso Especial Eleitoral No. 35724. https://tse
.jusbrasil.com.br/jurisprudencia/22421031/recurso-especial-eleitoral-respe-35724
-pa-tse/inteiro-teor-110697019.
2014. Instruções Do TSE : Eleições 2014: #vempraurna. Brasília: Tribunal Superior
Eleitoral of Brazil. www.tse.jus.br/hotsites/catalogo-publicacoes/pdf/instrucoes_
eleitorais/instrucoes-tse-eleicoes-2014.pdf.
Valle, Vinicius Saragiotto Magalhães do. 2013. “Pentecostalismo e lulismo na periferia
de São Paulo: estudo de caso sobre uma Assembleia de Deus na eleição municipal de
2012.” Master’s thesis, São Paulo: Universidade de São Paulo. www.teses.usp.br/
teses/disponiveis/8/8131/tde-10042014-120558/pt-br.php.
Veja. 2017. “Bolsonaro é acusado de racismo por frase em palestra na Hebraica.” Veja,
April 6. http://veja.abril.com.br/brasil/bolsonaro-e-acusado-de-racismo-por-frase
-em-palestra-na-hebraica/.
202 References

Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality:
Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.
Vianna, Cláudia Pereira. 2015. “O movimento LGBT e as políticas de educação de
gênero e diversidade sexual: Perdas, ganhos e desafios.” Educação e Pesquisa 41 (3):
791–806. https://doi.org/10.1590/s1517-97022015031914.
Wainscott, Ann Marie. 2017. Bureaucratizing Islam: Morocco and the War on Terror.
Cambridge, UK: Cambridge University Press.
Wald, Kenneth D., Dennis E. Owen, and Samuel S. Hill. 1988. “Churches as Political
Communities.” American Political Science Review 82 (2): 531–48.
1990. “Political Cohesion in Churches.” Journal of Politics 52 (1): 197–215.
Weber, Christopher, and Matthew Thornton. 2012. “Courting Christians: How
Political Candidates Prime Religious Considerations in Campaign Ads.” Journal
of Politics 74 (02): 400–13.
Webster, Steven W., and Alan I. Abramowitz. 2017. “The Ideological Foundations of
Affective Polarization in the US Electorate.” American Politics Research 45 (4):
621–47.
Western, Bruce, and Deirdre Bloome. 2009. “Variance Function Regressions for
Studying Inequality.” Sociological Methodology 39 (1): 293–326.
Wilson Quarterly. 2010. “The Parishioner Is Always Right.” Wilson Quarterly 34 (1):
76–7.
Wolf, James G. 1989. Gay Priests. 1st ed. San Francisco: HarperCollins.
Wolfe, Alan. 1999. One Nation, After All: What Middle-Class Americans Really Think
About: God, Country, Family, Racism, Welfare, Immigration, Homosexuality,
Work, the Right, the Left, and Each Other. New York: Penguin Books.
Zaller, John R. 1992. The Nature and Origins of Mass Opinion. Cambridge, UK:
Cambridge University Press.
Zorzanelli, Marcelo. 2016. “Após ser citado por todos deputados pró-impeachment,
Deus será investigado pelo Ministério Público.” Sensacionalista, April 17. https://
www.sensacionalista.com.br/2016/04/17/apos-ser-citado-por-todos-deputados-
pro-impeac-hment-deus-sera-investigado-pelo-ministerio-publico/.
Index

abortion, 18, 20, 166 clergy involvement, 42, 43, 80, 81, 82, 83,
clergy views on, 67, 117, 119
and Dilma Rousseff, 20 electoral rules, 21, 38, 86, 157
legal status in Brazil, 3, 92, 104 and evangelical support, 7, 21, 29, 88, 90,
legislator views, 154, 155, 161 114, 116, 118, 123, 131, 157, 160
and public opinion, 3, 18, 92, 99, 103, 104, Roman Catholic Church support, 20, 29, 123
109, 110, 111 Campos, Eduardo, 3
religious-elite views on, 67, 68 Canada, 28
and religious switching/conversion, 107 Cardoso, Fernando Henrique, 20, 153
and the Roman Catholic Church, 63 Catholicism
affective polarization; see polarization, affective clergy endorsements, 84
Afro-Brazilian religions, 12, 14, 19, 152 demographic trends, 11, 12, 14, 70
agnosticism: see atheism and agnosticism demographic trends in Latin America, 162
Alckmin, Geraldo, 116 in Chile, 165
Aldrich, John H., 171 party ties, 29, 84, 85, 105
Almeida, Ronaldo, 151 and transition to democracy, 17
Alvarado, President Carlos (of Costa Rica), 162 and voting guides, 84
American Convention on Human Rights, 162 caucuses in National Congress, 152; see also
Assembly of God, 16, 39, 81, 82, 84, 101, 157 Roman Catholic Apostolic Parliamentary
atheism and agnosticism, 70, 139, 153 Front; evangelical caucus
Caxias do Sul, 115
Belgium, 43 Chamber of Deputies, 4, 20, 29, 150, 153, 157
Benedict XVI, Pope, 20 leadership, 153
Bolsonaro, Eduardo, 150 Charismatic Catholicism, 13, 34, 70, 109,
Bolsonaro, Jair, 154, 174 132, 152
brokers, electoral, 42, 44, 160, 161, 170 Charismatic Protestantism, 13, 34
chastity,
campaign finance, 6, 150, 158 clergy views on, 67
campaigns Chile, 163, 164, 165
2002 general election, 115 church–state relations, 3, 35, 76, 102, 166
2006 general election, 115 in Chile, 165
2008 local election, Juiz de Fora, 47, 116, 117 legislator preferences, 155
2010 general election, 19, 85, 116, 122, 123 and noise ordinances, 131
2014 general election, 20, 21, 116, 122 perceptions of, 76, 94

203
204
church–state relations, (cont.)
and public opinion, 99, 106, 109
civic education, 39, 149, 167
civil society; see political participation
clergy
constraints on speech of, 66
influence on candidate electability, 126
influence on congregant vote choice, 122,
123, 126, 126,
motivations of, 31, 68, 71
policy views of, 66
religious teachings of, 35, 64, 71
and socialization, 36
clientelism, 90, 149, 159, 160, 161, 170
and citizens’ norms, 87
in congregations, 150
Colombia, 163, 165, 170
communism,
opposition to, 85
compulsory and voluntary voting rules, 87, 120
Concordat (Brazil and Holy See, 2009), 38
congregations
competition among, 51, 89
growth and decline of, 50, 52, 62, 68, 70
hosting politicians, 158
and outreach, 51
political diversity (internal) of, 91, 128
and social influence, 99, 108, 110, 119,
123, 128
conscientious voting, 39
and citizen norms, 87
clergy discussion of, 80, 83, 92, 117
Constitution of Brazil, 17, 37, 133
Costa Rica, 163, 164, 165
conversion; see religious switching
Crivella, Bishop Marcelo, 18, 19,
151
culture wars
causes, 7, 27, 28
definition of, 4, 27
Cunha, Eduardo, 20, 151
Dancygier, Raphaela, 43
democracy, 165; see representative democracy;
participatory democracy
breakdown of, 136, 143
clergy support for, 5, 61, 72, 133
deliberative democracy, 130, 136
indicators of level of, 22
liberal democracy, 130, 136
participatory democracy, 130, 141
Index
public support for, 22, 72, 131, 133, 134,
150, 167, 171, 174
“third wave,” 149
representative democracy, 149
transition to, 16, 17
Demócratas Party, 155
Dias, Pastor Everaldo, 119, 155
doctrinal conservatism, 41, 42, 62, 95, 111
Downs, Anthony, 70
drought, 73
dualism; see evil, perceptions of
ecclesiastical base communities, 17
economic policy
and legislator views, 154
and public opinion, 104
education in public schools
core curriculum of, 18
and evolution, 19
religious education in, 17, 38, 101
sexual education in, 3, 18
electoral system, 6, 43, 151, 157, 160, 166
End Times theology (eschatology), 64, 71, 132
endorsements
from clergy, 80, 81, 82, 83, 84, 86, 88, 89,
117, 122
and citizen norms, 86, 87, 88, 91
environmental protection
clergy views on, 67
and public opinion, 99, 104
evangelical caucus (bancada evangélica), 16,
115, 152
Evangelical Parliamentary Front (Frente
Parlamentar Evangélica); see
Parliamentary Front for Defense of
Traditional Peoples of African Origin;
evangelical caucus
evangelicalism
and Afro-Brazilian religions, 14, 19, 152
definition of, 10, 100, 166
and demographic trends, 11, 12, 14
and demographic trends in Latin
America, 162
ideological position in Brazil, 99
and National Constituent Assembly
(1987–1988), 16
and religious attendance, 13
evangelization
restrictions on, 106
evil, perceptions of, 63, 132, 136, 166, 169
evolution, 19
Index
Feliciano, Pastor Marco, 151
frames, religious, 40
France, 26, 33
Francis, Pope, 84
Free Electoral Hour (Horário Eleitoral
Gratuito), 48
Garotinho, Anthony, 115
“gender ideology,” 3, 75, 92
gender roles
public opinion, 109, 111
Gill, Anthony, 17, 32, 165
Guatemala, 163
homosexuality, 166; see also same-sex marriage
among Catholic clergy, 71
anti-hate-speech laws, 106
clergy views on, 5, 61, 63, 64, 67, 71, 81
conversion therapy, 103
discrimination and hate crimes, 18, 75
gay candidates, attitudes toward, 139
gay-rights movement, 27, 73, 162
gay-friendly churches in Brazil, 169
politician views on, 62
public opinion on, 99, 100, 101, 102, 103
Hunter, James Davison, 4, 27
hypodermic models of persuasion, 113
ideological identification
and legislator views, 155
and public opinion, 104
Integral Mission (evangelical movement), 66
Inter-American Court of Human Rights, 162
Juiz de Fora, 47, 157
Council of Pastors, 48, 62
local elections of 2008, 47
politics of, 17
Kassab, Gilberto, 48
legislators, see politicians
legitimacy of political system
public opinion on, 136
legitimacy, perceptions of state, 130
clergy views on, 61, 74, 77
public opinion on, 23, 131, 133, 134, 135
liberal democracy; see democracy
liberation theology, 17, 31
lobbying (legislative advocacy) by churches, 17,
19, 89, 142, 158
205
Lula (President Luiz Inácio Lula da Silva),
38, 115
Macedo, Bishop Edir, 18, 67
Malafaia, Pastor Silas, 81
Malta, Senator Magno, 18
Manichaean theology; see evil, perceptions of
Mattos, Custódio, 47, 116
Methodism, 157
Methodist Church, 9
Mexico, 122
military regime, 174
ministry to poor
clergy teachings on, 66
MBL (Free Brazil Movement), 170
National Conference of Bishops of Brazil,
17, 18
National Constituent Assembly
(1987–1988), 16
National Forum for Religious Education, 17
neo-Pentecostalism, 12, 34
Neves, Aécio, 21
Nicaragua, 165
Noraldino Lúcio Dias Junior, 157
open-list proportional representation, 43, 151,
157, 166
Operation Car Wash, 6, 19, 22, 29, 151
Parliamentary Front for Defense of Traditional
Peoples of African Origin, 152
participatory democracy; see democracy
partisanship, 23, 29, 137
anti-petismo (anti-PT sentiment), 73, 81,
84
Catholic, 8, 19, 84, 85, 99, 105
evangelical, 19, 84, 105
petismo (PT support), 22, 23, 129
party system, 4, 6, 8, 160, 161, 166; see also
Workers’ Party (PT)
evangelical parties, 7, 155
rightist parties, 29, 155
Pentecostalism
and prosperity theology, 65
definition of, 10, 12, 100
and demographic trends, 13
in Africa, 30
and styles of campaigning, 91
Pereira, Pastor Everaldo Dias, 21, 116
Pitkin, Hanna Fenichel, 149
206
PMDB (Party of the Brazilian Democratic
Movement), 29, 155
polarization, affective, 41, 136, 137, 139
political participation, 141, 168
in civil society, 142
clergy discussion of, 80, 82, 83, 92, 94, 95,
117, 118, 119, 141
and congregational mobilization, 120, 131,
142, 143
contacting politicians, 150, 159
and petitions, 142
rates of, 120
political tolerance; see tolerance, political
politicians; see also evangelical caucus;
representation
party ties, 155, 158
personal ties to voters, 150, 158, 161
policy views of, 150, 154
political dynasties, 150, 158
religious affiliations of, 149, 152
support base, 150, 158
ties to clergy, 150, 156, 157, 158
popular referenda; see referendums
PRB (Brazilian Republican Party), 151, 155
prosperity theology (prosperity gospel), 13, 35,
64, 81
Protestantism
definition of, 10, 11, 100
and demographic trends, 11, 12
protests, 22, 143
Catholic views on, 143
clergy views on, 143
evangelical views on, 143
PSC (Christian Social Party), 155
PSD (Social Democratic Party), 155
PSDB (Party of Brazilian Social Democracy),
20, 29
PT; see Workers’ Party (PT)
public-school education; see education in public
schools
quilombolas, 154
racial issues
clergy views on, 66, 67
legislator views on, 154
public opinion on, 99, 104
referendums, 133
Reis, Washington, 150
religious economy, 8, 16, 31, 32, 165
free markets (religious), 34
religious frames; see frames, religious
Index
religious non-identification in Latin
America, 163
religious non-identifiers, 14, 34, 70, 143
religious switching or conversion, 107,
representation, 42, 43, 155; see also politicians
descriptive representation, 149
substantive representation, 149, 160
symbolic representation, 149
representative democracy; see democracy
Rio de Janeiro, municipal government of,
19, 151
Ríos Montt, General Efraín, 163
Roman Catholic Apostolic Parliamentary
Front, 152
Rossi, Father Marcelo, 13, 39
Rousseff, President Dilma, 18, 20, 21, 48
impeachment of, 19, 22, 81, 85, 150
voter support for, 116
Russomano, Celso, 152
Salomão, Margarida, 47, 92, 116
same-sex marriage; see also homosexuality
Costa Rican views on, 162
Latin American views on, 164
legal status of, 5, 27, 75, 92, 132
legislator views on, 154, 161
public opinion on, 99, 101, 102, 109, 111
social movements opposing, 132
Santo Daime religion, 81
São Paulo, 153
secularism, 165
and Brazilian state policy, 16, 37, 38
definition of, 33, 34, 37
development of, 38
in Chile, 165
and legislator preferences, 155
norms, 41, 99, 100, 102, 106, 109, 126,
129
Senate, 153, 166
Serra, José, 20, 169
Silva, Marina, 20, 21, 116, 131
social movements
clergy views, 82, 89
social policy (anti-poverty policy)
legislator views, 154
public opinion, 104
Spiritism (religion), 14, 47
spiritual warfare, 13, 169
state–church relations, 74
Stokes, Susan C., 42, 44
Superior Electoral Tribunal (Tribunal Superior
Eleitoral), 21, 38, 86
Index
Supreme Federal Tribunal (Supremo Tribunal
Federal), 27, 75, 92, 132
survey experiments, 24
tax policy in Brazil, 18
Temer, President Michel, 22, 85, 132
incorrect belief Temer is a Satanist, 132,
136, 169
theological liberalism, 169
Tocqueville, Alexis de, 26, 172
tolerance, political, 6, 136, 137, 168
clergy views, 61, 72, 73
external tolerance, 72, 139
internal tolerance, 72, 139,
traditional family, clergy views of, 67
Trejo, Guillermo, 32, 37, 165
Turkey, 33
turnout; see political participation
207
Uganda, 30
United Kingdom, 43
United States, 4, 26, 27, 28, 33, 37, 86, 102,
122, 138
Universal Church of the Kingdom of God
(UCKG), 80, 114
abortion views of, 18, 67
campaign practices of, 16, 20, 157, 169
candidate support from, 29, 85, 118,
157
establishment and growth of, 12
influence on congregants, 6, 122
and politicians, 151, 152
Workers’ Party (PT), 8, 17, 18, 19, 23, 29, 47,
75, 84, 85, 105
Zaller, John R., 40, 113, 117, 120
Books in the Series (continued from p. ii)

Karrie J. Koesel, Religion and Authoritarianism: Cooperation, Conflict, and the

Consequences
Ahmet T. Kuru, Secularism and State Policies toward Religion: The United States,
France, and Turkey
Andrew R. Lewis, The Rights Turn in Conservative Christian Politics: How Abortion
Transformed the Culture Wars
Damon Maryl, Secular Conversions: Political Institutions and Religious Education in
the United States and Australia, 1800–2000
Jeremy Menchik, Islam and Democracy in Indonesia: Tolerance without Liberalism
Pippa Norris and Ronald Inglehart, Sacred and Secular: Religion and Politics Worldwide
Amy Reynolds, Free Trade and Faithful Globalization: Saving the Market
Sadia Saeed, Politics of Desecularization: Law and the Minority Question in Pakistan
David T. Smith, Religious Persecution and Political Order in the United States
J. Christopher Soper and Joel S. Fetzer, Religion and Nationalism in Global Perspective
Peter Stamatov, The Origins of Global Humanitarianism: Religion, Empires, and
Advocacy
Kenneth D. Wald, The Foundations of American Jewish Liberalism