Wealth, Inclusive Growth

Wealth, Inclusive Growth
and Sustainability
The excessive pursuit of economic interests has resulted in severe environmental

and social problems, such as climate change, biodiversity loss, and inequality and
disparity. There is an urgent need for broader measures of progress to complement
gross domestic product (GDP). This book provides a wide range of economic
evaluations of environmental and societal issues including climate change,
emission problem from garbage landfills, and income inequality. The book
explains that sustainability indicators and well-being measures can be effective
guides for policymaking and how they can strike a balance between economic,
environmental, and societal interests.
This book summarizes current practices and theories of economic evaluation
for sustainability and provides understanding of emerging trends in this area. It
also stresses the importance of environmental policies and business actions in
achieving sustainable growth and puts forth why countries should take natural
capital and other conventional inputs into consideration.
Shunsuke Managi is Distinguished Professor and Director at the Urban Institute

and Department of Urban and Environmental Engineering, Kyushu University.
Routledge Studies in the Modern World Economy
Chinese Trade
Trade Deficits, State Subsidies and the Rise of China
Rich Marino
The Evolution of Economic Wellbeing

Progress-Driven Economic Policies in the Era of Globalization
Zuhayr Mikdashi
Entrepreneurship and Local Economic Development

A Comparative Perspective on Entrepreneurs, Universities and Governments
Edited by Bruno Dallago and Ermanno Tortia
The Future of the Economy

East-West Perspectives on Pathways Through Disruption
John Powers and Vikram Khanna
Regional Government Competition

Chen Yunxian and Gu Wenjing
The Dynamics of Growth in Emerging Economies

The Case of Turkey
Edited by Arzu Akkoyunlu Wigley and Selim Çağatay
Economic Integration in Asia

Key Prospects and Challenges with the Regional Comprehensive
Economic Partnership
Edited by Deeparghya Mukherjee
Wealth, Inclusive Growth and Sustainability

Edited by Shunsuke Managi
For more information about this series, please visit www.routledge.com/Routledge-

Studies-in-the-Modern-World-Economy/book-series/SE0432
Wealth, Inclusive Growth
and Sustainability
Edited by Shunsuke Managi

First published 2019
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
52 Vanderbilt Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2019 selection and editorial matter, Shunsuke Managi; individual chapters,
the contributors
The right of Shunsuke Managi to be identified as the author of the editorial
material, and of the authors for their individual chapters, has been asserted
in accordance with sections 77 and 78 of the Copyright, Designs and Patents
Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in writing from
the publishers.
Trademark notice: Product or corporate names may be trademarks or
registered trademarks, and are used only for identification and explanation
without intent to infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
Names: Managi, Shunsuke, editor.
Title: Wealth, inclusive growth and sustainability / edited by
Shunsuke Managi.
Description: Abingdon, Oxon ; New York, NY : Routledge, 2019. |
Series: Routledge studies in the modern world economy ; 185 |
Includes bibliographical references and index.
Identifiers: LCCN 2018048243 | ISBN 9780367002367 (hardback) |
ISBN 9780429400636 (ebook)
Subjects: LCSH: Economic development—Environmental aspects. |
Sustainable development. | Income distribution. | Welfare economics. |
Environmental economics.
Classification: LCC HD75.6 .W425 2019 | DDC 338.9/27—dc23
LC record available at https://lccn.loc.gov/2018048243
ISBN: 978-0-367-00236-7 (hbk)
ISBN: 978-0-429-40063-6 (ebk)
Typeset in Times New Roman

by Apex CoVantage, LLC
Contents
List of figures vii

List of tables xi
List of maps xiv
List of contributors xv
Preface xvii
Introduction 1
S H U N S U K E M ANAGI
1 An evaluation of inclusive capital stock for urban planning 5

H I D E M I C H I F U JI I AND S HUNS UKE MANAGI
2 Impact of infrastructure in India 23

I S WA RYA S A N K ARAL I NGAM, MOI NUL I S L AM,
WATA R U N O Z AWA AND S HUNS UKE MANAGI
3 Sustainability evaluation for energy infrastructure:

a hybrid simulation approach to inclusive wealth 43
E B R A H I M A LY AND S HUNS UKE MANAGI
4 Public debt as a negative stock in sustainability indicator 77

M A S AY U K I S ATO AND S HUNS UKE MANAGI
5 Global marine fisheries with economic growth 87

Y O G I S U G I AWA N, MOI NUL I S L AM AND S HUNS UK E MA N A G I
6 Inclusive wealth adjusted by total factor productivity as

a sustainability measurement: global productivity analysis 135
I S M A A D D I J U MBRI , MOI NUL I S L AM AND S HUNSU K E MA N A G I
vi Contents
7 Relative income, community attachment and subjective
well-being: evidence from Japan 167
T E T S U YA T SURUMI , KONG JOO S HI N, AT S US H I IMA U JI
A N D S H U N S UKE MANAGI
8 Environmental value of green spaces in Japan: an application

of the life satisfaction approach 195
T E T S U YA T SURUMI AND S HUNS UKE MANAGI
9 Greenery and well-being: assessing the monetary value of

greenery by type 225
T E T S U YA T SURUMI , AT S US HI I MAUJI AND S H U N SU K E MA N A G I
10 The impact of climate change and extreme events on

agriculture in Africa 261
T H I E R RY C OUL I BALY, MOI NUL I S L AM AND SH U N SU K E MA N A G I
11 Sustainable development and performance measurement:

global productivity decomposition 286
R O B I K U R N IAWAN AND S HUNS UKE MANAGI
12 Social cost-benefit analysis of decontamination after

the Fukushima Dai-ichi Nuclear Power Plant accident 310
K A Z U N O B U OKUBO, RI NTARO YAMAGUCHI AN D SH U N SU K E MA N A G I
13 National and subnational sustainability: a case study of Japan 327

K A Z U N O B U OKUBO, RI NTARO YAMAGUCHI AN D SH U N SU K E MA N A G I
14 The effect of landfill gas emission on global warming

and workers’ health in Indonesia 354
R A F I D M A H F UL , MOI NUL I S L AM, HI ROF UMI NA K AYA MA
A N D S H U N S UKE MANAGI
15 Human capital change and social impact under China’s

universal two-child policy 374
Q I U Y I C H E N , MOI NUL I S L AM AND S HUNS UKE MA N A G I
16 The impact of renewable energy on sustainability: empirical

analysis employing the Inclusive Wealth Indicator 405
M O E G I I G AWA AND S HUNS UKE MANAGI
17 Consumer demand for fully automated driving technology 418

K O N G J O O S HI N AND S HUNS UKE MANAGI
Index 440
Figures
0.1 Classification in inclusive wealth 2

1.1 Diagram outlining the approaches in this study 13
1.2 Decomposition analysis of changes in produced capital stock
(JPY 100 million) 17
1.3 Decomposition analysis of changes in human capital stock
1.4 Decomposition analysis of changes in natural capital stock
2.1 Public infrastructure investment as percentage of Indian GDP 24
2.2 Teledensity per 1000 population 27
2.3 Composition of internet subscription, 2016 29
2.4 Government share in infrastructure spending (in %) 36
2.5 Under-implementation of infrastructure projects 38
2.6 State-wise share of infrastructure spending (% of total) 39
3.1 The wealth trajectory after introducing a new project to
the economy 45
3.2 Schematic diagram for model components and the process flow 47
3.3 The state of projects under evaluation and the triggers control
transition between different states 49
3.4 Dynamics of the interaction between an energy project and
the different capital assets 51
3.5 Impacts on capital assets for BEE biomass power plant 59
3.6 Impacts on capital assets for the Norther wind farm 60
3.7 Impacts on capital assets for BEE biomass power plant
(not considering education effect) 61
3.8 Impacts on capital assets for the Norther wind farm
(not considering education effect) 62
3.9 CO2 damages from Belgian proposed power projects 63
3.10 Energy supply index for Belgian proposed power projects 63
3.11 The change in wealth caused by the proposed Belgian
energy projects 64
3.12 Impacts of the proposed CCGT power plant on the Egyptian
capital assets 66
viii Figures
3.13 Impacts of the proposed solar power project on the Egyptian
capital assets 67
3.14 Impacts of the proposed wind power project on the Egyptian
capital assets 68
3.15 Impacts of the proposed CCGT power plant on the Egyptian
capital assets (without education impact) 69
3.16 Impacts of the proposed solar power project on the Egyptian
3.17 Impacts of the proposed wind power project on the Egyptian
3.18 CO2 damages from the three proposed power projects in Egypt 72
3.19 ESI from the three proposed energy projects in Egypt 72
3.20 The changes in wealth caused by the proposed energy
projects in Egypt 73
4.1 The relationship and expected sign of public debt and GS 79
4.2 Public debt of Japan 80
4.3 Estimated relationship between GS and public debt 82
4.4 GS path of each country 83
5.1 Global fisheries catch and estimated stock trends 92
5.2 Comparison of world catch and estimated stock levels
A. Average annual catch changes from 1961 to 2010 (%)
B. Average annual stock changes from 1961 to 2010 (%) 93
5.3 Projection of total landing and stock values for 70 fishing
countries 100
5.4 Projection of landings and stocks for the examined countries 101
6.1 Mean TFP change, efficiency change and technical change
(1990–2010) 146
6.2 Mean TFP change in six regions (1990–2010) 147
6.3 Percentage of wealth composition of G7 countries (1990–2010) 155
6.4 TFP growth of the G7 countries (1990–2010) 156
6.5 Percentage of wealth composition of BRICS countries
(1990–2010) 157
6.6 TFP growth of BRICS countries (1990–2010) 159
6.7 Comparison between average TFP and IW-TFP adjusted growth
of BRICS 161
7.1 Conceptual diagram of the estimation model 175
7A.1 Distribution of life satisfaction 194
7A.2 Distribution of attachment to a community 194
8.1 Marginal willingness to pay for a 1% increase in green spaces 211
8.2 Estimated marginal willingness to pay for green spaces 215
9A.1 Changes in the extent of green areas in the Tokyo
metropolitan region 251
10.1a Temperature trends in the period 1961–2014 267
10.1b Temperature trends in the period 1961–1986 267
10.1c Temperature trends in the period 1987–2014 267
Figures ix
10.1d Rainfall trends in the period 1961–2014 267
10.1e Rainfall trends in the period 1961–1986 268
10.1f Rainfall trends in the period 1987–2014 268
10.2 PIN per level of income 275
10.3a Growth in inclusive wealth per capita (1990–2014) 276
10.3b Most occurring level of income per country (1990–2014) 276
10.4a The numbers of funds in Africa in 2010 277
10.4b Level of income per countries (2010–2014) 277
11.1 Average TFP and contribution effect of inputs/outputs for
140 countries 295
11.2 Inefficiency score change in China 304
11.3 Inefficiency score change in Canada 305
12.1 The schedule of air dose rate reduction without human
intervention for different initial levels 317
12.2 How to identify the timing of capital service recovery 317
13.1 Group 1-a 335
13.2 Group 1-b 336
13.3 Group 1-c 337
13.4 Group 1-d 338
13.5 Group 1-e 339
13.6 Group 2 341
13.7 Group 3 342
13.8 Group 4-a 346
13.9 Group 4-b 346
13.10 Group 4-c 347
13.11 Group 5-a 347
13.12 Group 5-b 348
13.13 Group 5-c 348
13.14 Tokyo’s PDP and DAU per capita wealth under different
discount rates 349
13.15 Osaka’s PDP and DAU per capita wealth under different
discount rates 349
14.1 Yearly disposed quantities of Benowo landfill 361
14.2 Yearly disposed quantities of Tamangapa landfill 362
14.3 Methane emission calculation from Tamangapa landfill 364
14.4 Methane emission calculation at Benowo landfill 364
14.5 Carbon dioxide emission calculation at Tamangapa landfill 365
14.6 Carbon dioxide emission calculation at Benowo landfill 366
14.7 Hydrogen sulfide emission calculation at Tamangapa landfill 366
14.8 Hydrogen sulfide emission calculation at Benowo landfill 367
14.9 Amount of illness suffered by informal worker 368
14.10 Amount of illness spend money 369
14.11 Amount of informal worker salaries 370
15.1 Total population forecast 381
15.2 Total population forecast by UN 381
x Figures
15.3 China’s age structure in recent years 382
15.4 Total population by broad age group forecast by UN 383
15.5 Population pyramids 384
15.6 China’s average educational level 390
15.7 China’s life expectancy at birth 391
15.8 Fertility rate of China and the world 394
15.9 Fertility rate of mainland China, Hong Kong and Macao 394
15.10 Fertility rate of Russian Federation 397
16.1 Electricity generation from renewable energy between 2001
and 2015, excluding large hydropower projects generating more
than 10 MW (GWh) 406
17.1 When would FAV be available in the market? 424
17.2 When would FAV be available in the market? (by age groups) 424
17.3 Purchase intention for FADS by consumer characteristics
(N = 246,642) 426
17.4 WTP of FADS by consumer characteristics (N = 188,089) 428
Tables
1.1 Literature review of city evaluation indexes 6

1.2 Basic data for 20 ordinance-designated cities in 2010 14
1.3 Relative evaluation score of the inclusive urban capital stock 15
1A.1 Grouping of local government budget expenditures 20
2.1 Principal component analysis test results 30
2.2 Correlation matrix 30
2.3 Summary statistics of the compiled dataset 32
2.4 Detailed description of variables used 32
2.5 Regression coefficients 35
3.1 Description of the project characteristics 49
3.2 Air and water pollutants 53
3.3 Predicted time series and max corresponding MAPE 57
3.4 Belgium projects data 58
3.5 Egypt projects data 65
4.1 Estimation result of ARCH-M 82
5.1 Panel unit root tests 96
5.2 Model selection summary 97
5.3 Long- and short-run estimates of the PMG 98
5A.1 Estimated biomass 107
6.1 Ranking and summary of estimated TFP 144
6.2 Means of TFP, efficiency change and technical change
(1990–2010) based on regions 146
6.3 Result of IW-TFP adjusted per capita and percentage of change 148
6.4 Results after TFP adjusted per capita and percentage of change 153
7.1 Studies on relative income hypothesis 168
7.2 Descriptive statistics 172
7.3 Estimation results (full sample) 177
7.4 Estimation results (subsample) 179
7.5a Direct, indirect and total effects of relative income on life
satisfaction (500 m mesh) 182
7.5b Direct, indirect and total effects of relative income on life
7.5c Direct, indirect and total effects of relative income on life
xii Tables
7.5d Direct, indirect and total effects of relative income on life
7.5e Direct, indirect and total effects of relative income on life
satisfaction (municipality level) 183
7A.1 Variables 191
8.1 Descriptive statistics (Obs. = 2158) 203
8.2 Correlation of each variables concerning green spaces 206
8.3 Estimation results (not including cross terms) 209
8.4a Estimation results (including cross term: Affection for
neighborhood greenery) 212
8.4b Estimation results (including cross term: Affection for
world greenery) 212
8.4c Estimation results (including cross term: Interaction with
greenery in the last 5 years) 213
8.4d Estimation results (including cross term: Remembered
interaction with greenery until the age of 12) 213
8.4e Estimation results (including cross term: Knowledge regarding
the multiple functions of forests) 214
8.4f Estimation results (including cross term: Quality of greenery) 214
8A.1 Definition of “green spaces” adopted for this study 217
8A.2 Main questionnaire content 218
8A.3 Forests’ multiple functions 220
8A.4 Multiple functions of agriculture 221
9.1 Survey questions for each measure of subjective well-being 228
9.2 Correlations among indices related to well-being 229
9.3 Classification of greenery in this study 230
9.4 Descriptive statistics 234
9.5 Parameter estimates (total greenery): Cantril ladder 237
9.6 Parameter estimates (total greenery): life satisfaction 239
9.7 Parameter estimates (total greenery): subjective happiness 239
9.8 Parameter estimates (total greenery): affect balance 239
9.9 Parameter estimates (total greenery): mental health 240
9.10 Parameter estimates of various types of greenery 241
9.11 Monetary values of greenery (JPY) 243
9A.1 Characteristics of respondents with comparison 252
9A.2 Indicators of Better Life Index that are used for control
variables in this study 253
9A.3 Correlations among total greenery by distance band 258
9A.4 Parameter estimates (total greenery): Cantril ladder 258
9A.5 Parameter estimates (total greenery): life satisfaction 258
9A.6 Parameter estimates (total greenery): subjective happiness 259
9A.7 Parameter estimates (total greenery): mental health 259
9A.8 Parameter estimates of various types of greenery 259
10.1a Summary statistics for the period 1967–2014 per countries 265
10.1b Summary statistics for the period 1961–1986 per countries 265
10.1c Summary statistics for the period 1987–2014 per countries 266
Tables xiii
10.2 Production model in Africa 268
10.3 The impact of climate change on the whole continent for
the periods 1961–1986 and 1987–2014 270
10.4 Repartition of countries per level of income 272
10.5 Impact of climatic variables per level of income of countries 273
10.6 Total number of countries with at least one agricultural fund 278
10.7 Number of agricultural funds for developing countries, per
country in 2010 for the sample of this study 278
11.1 Aggregate (overall) technical inefficiency 296
11.2 Technical inefficiency by income level 298
12.1 Estimated cost for decontamination in previous studies 312
12.2 Final disposal plans of soil and waste according to their
characteristics 314
12.3 Difference in the timing of resuming capital services with
and without decontamination in the case of the once under
the order area 318
12.4 Discounted government expenditures from fiscal years 2011 to
2016 (unit: JPY billions) 320
12.5 Total benefit by area (unit: JPY billions) 321
13.1 Share of manufacturing in 20 prefectures that have maintained
their level of total wealth since 2000 330
13.2 Wealth change under TU, TU pc, and DAU, 2000–2010 333
13.3 Change in DAU per capita under different discount rates 350
14.1 Health effects reported from involvement in informal recycling
(Eerd, 1996) 359
14.2 Input review (assumption) for Benowo and Tamangapa landfills
in LandGEM 3.02 software 363
14.3 Salient features of Benowo and Tamangapa landfills,
Indonesia 364
14.4 Health and safety risk associated with informal recycling 368
15.1 Average ideal number of children among people with different
characteristics 385
15.2 Human capital change with respect to the base year 2000 389
16.1 The lifecycle greenhouse gas emissions of electricity supply
technologies 407
16.2 Descriptive statistics value 412
16.3 Empirical results 414
17.1 The results of selected merits (multiple selections and top
three selections) 421
17.2 The results of selected demerits (multiple selections and top
three selections) 422
17.3 The results of factor analyses: merit and demerit factors 422
17.4 Determinants of consumer demand of FAV and FADS
(individual-level analysis) 431
17.5 Determinants of consumer demand of FAV and FADS
(municipality-level analysis) 434
Maps
6.1 Average of TFP for 140 countries (1990–2010) 143

6.2 Before IW-TFP adjusted 150
6.3 After IW-TFP adjusted 150
6.4 Before IWI adjusted 151
6.5 After IWI adjusted 152
8.1 Locations of “green spaces” and questionnaire respondents
in the Tokyo metropolitan area 199
8.2 Locations of “green spaces” and questionnaire respondents
in the Kinki region 200
9.1 Vegetation detected via satellite imagery and the NDVI 231
9.2 Vegetation data post-classification 232
11.1 Average GDP inefficiency score 299
11.2 Average carbon damages inefficiency score 301
11.3 Natural capital inefficiency score 302
14.1 The location of landfill study cases 360
15.1 Human capital of mainland China 2000 387
15.5 Fertility projection by UN 395
15.6 Urbanization rate of mainland China 2005 400
15.7 Urbanization rate of mainland China 2013 400
15.8 Population distribution (2015) and population change
(2002–2013) of mainland China 401
15.9 Population density of mainland China 2015 401
17.1 Distribution of purchase intention at municipality level 427
17.2 Distributions of WTP at municipality level 428
17.3 Distributions of WTP at municipality level: consumers
with high PI 429
Contributors
Ebrahim Aly is a doctoral student at the Department of Urban and Environmental

Engineering, Faculty of Engineering, Kyushu University.
Qiuyi Chen is formerly a master’s student at the Department of Urban and Envi-
ronmental Engineering, Faculty of Engineering, Kyushu University.
Thierry Coulibaly is a doctoral student at the Department of Urban and Environ-
mental Engineering, Faculty of Engineering, Kyushu University.
Hidemichi Fujii is Associate Professor at the Department of International Econ-
omy and Business, Faculty of Economics, Kyushu University.
Moegi Igawa is a doctoral student at the Department of Urban and Environmental
Atsushi Imauji is a doctoral student at the Faculty of Policy Studies, Nanzan
University.
Moinul Islam is a researcher at the Department of Urban and Environmental
Isma Addi Jumbri is a doctoral student at the Graduate School of Environmental
Studies, Tohoku University.
Robi Kurniawan is a doctoral student at the Graduate School of Environmental
Studies, Tohoku University.
Rafid Mahful is a master’s student at the Department of Urban and Environmen-
tal Engineering, Faculty of Engineering, Kyushu University.
Shunsuke Managi is Distinguished Professor at the Urban Institute and Depart-
ment of Urban and Environmental Engineering, Faculty of Engineering, Kyushu
University/IGES Fellow, Institute for Global Environmental Strategies.
Hirofumi Nakayama is Associate Professor at the Department of Urban and
Environmental Engineering, Faculty of Engineering, Kyushu University.
Wataru Nozawa is Assistant Professor at the Department of Urban and Environ-
xvi Contributors
Kazunobu Okubo is a research fellow at the Graduate School of Human Devel-
opment and Environment, Kobe University.
Iswarya Sankaralingam is a master’s student at the School of Public and Envi-
ronmental Affairs, Indiana University.
Masayuki Sato is Associate Professor at the Graduate School of Human Devel-
opment and Environment, Kobe University.
Kong Joo Shin is formerly Assistant Professor at the Department of Urban and
Environmental Engineering, Faculty of Engineering, Kyushu University.
Yogi Sugiawan is a doctoral student at the Department of Urban and Environ-
Tetsuya Tsurumi is Associate Professor at the Faculty of Policy Studies, Nanzan
University.
Rintaro Yamaguchi is a researcher at the Center for Social and Environmental
Systems Research, National Institute for Environmental Studies.
Preface
The last century witnessed severe environmental and social problems such as cli-
mate change, biodiversity loss, and inequality and disparity caused by excessive
pursuit of economic interests. The experience has shifted our society from eco-
nomic oriented to sustainability oriented, and global initiatives have been initi-
ated. At Rio+20 in 2012, the need for broader measures of sustainable growth was
discussed and set forth in the outcome document, indicating that new comprehen-
sive or integrated indicators are highly demanded in the global society.
This book, following the new paradigm, provides a wide range of economic
evaluations of environmental and societal issues including climate change, emis-
sion problem from garbage landfills, and income inequality using sustainability
indicators and well-being measures. Those evaluations, reemphasizing the trade-
off between economic values and environmental and societal values, suggest how
we can guide our policy for economic developments taking balance between eco-
nomic, environmental, and societal interests using those indicators and measures.
I would like to thank the contributing authors for their involvement. This
research was partially funded by Specially Promoted Research through a Grant-
in-Aid (Scientific Research 26000001) from the Japanese Ministry of Education,
Culture, Sports, Science and Technology (MEXT) and was supported by the Envi-
ronment Research and Technology Development Fund (S-14, S-15, S-16) and the
4th Environmental Economics Research Fund of the Japanese Ministry of the
Environment. Of course, all remaining errors are my responsibility alone.
Introduction
Shunsuke Managi
Blinding ignorance does mislead us. O! Wretched mortals, open your eyes!
– Leonardo da Vinci
A lesson from the last century is that innocent pursuit of economic interests can
lead to catastrophe. We have witnessed water and air contamination, ecosystem
and ozone layer destruction, and sea level and temperature rise, and they eventu-
ally can threaten our lives.
We continued to expand the list of what we need to pay attention to for sustain-
ability of our society and now it is so long. The task to achieve a balance between
achieving rapid economic growth and sustainability has become the most diffi-
cult ever and it requires systematic approaches that are supported by international
cooperation and the local community.
The task is, by its nature, quantitative. We asked to decide how much of eco-
nomic interests should be given up, for example, to conserve environmental
quality. The difficulty is in that many things, such as ecosystem, health, and air
quality, do not have any market value. The economic approach overcomes this
difficulty using a basic idea of asset valuation. For example, health is regarded as
an asset that benefits one by enabling earning income by working. Such benefits
are thought of as the dividends of the asset, and the value of the asset is computed
by discounted valuation.
The Inclusive Wealth Index is an evaluation framework based on the economic
approach. Attaching economic values to education level, water quality, and other
capitals that are not traded in markets, it attempts to serve as a sustainability indi-
cator. As its name indicates, it is inclusive: the Inclusive Wealth Index contains
virtually any valuable entities as assets to be evaluated. It is useful to have an
inclusive index because if one plans an action for his economic interests at a cost
of, say, ecosystem, the index can tell us the net benefit of the action, the economic
benefit from the action less the cost for the destroyed ecosystem. That cannot be
done by traditional economic measures, such as national economic accounts that
exclude most of the beneficial entities that do not have market values. The Inclu-
sive Wealth Index does not ignore them, and we believe this is the economic way
to open our eyes as wretched mortals.
2 Shunsuke Managi
Figure 0.1 Classification in inclusive wealth
Figure 0.1 describes the basic classification in the Inclusive Wealth Index.
It consists of three classes of assets: manufactured capital, human capital, and
natural capital. Manufactured capital is the standard capital in economics, such
as machines, factories, and so on. Human capital includes skills, education,
and health, and natural capital includes subsoil resources, the atmosphere, and
ecosystem.
This book consists of 17 chapters about practice with the Inclusive Wealth
Index. Chapter 1 applies the IWI framework for evaluating urban sustainability.
The authors propose an empirical approach for how to compute inclusive urban
capital and demonstrate it using Japanese ordinance-designated cities. They dis-
cuss the advantage of their framework with the IWI through data envelopment
analysis a decomposition analysis.
Chapter 2 examines the role of infrastructure in economic growth in India.
Using state-wise data of transportation, energy, and telecommunication, the
authors estimate the contribution of those kinds of infrastructure to GDP growth
and relate the results to major issues in infrastructure investment in the country,
such as regional and sector imbalance.
Chapter 3 provides an evaluation model of energy infrastructure projects. Such
projects accompany not only accumulation of physical capital but also destruc-
tion of environmental and health capitals. Their model gives a comprehensive
evaluation that takes impacts to those capitals into account. They apply the model
to projects with different fuel cycles in Belgium and Egypt and demonstrate the
advantage of their model to the standard project evaluations and how results from
the model can be useful for policymaking.
Chapter 4 examines the relationship between genuine savings and public debt.
A previous work in the literature suggests that public debt accumulation indi-
cates necessity of more distorting taxation in future and therefore genuine savings
should be revised downward according to marginal cost of public debt. However,
it is difficult to estimate it and thus whether public debt has a significant impact on
genuine savings. This chapter estimates the impact using panel data of countries
and confirms the large effect of public debt on genuine savings.
Introduction 3
Chapter 5 explores marine fishery stock and its relationship to economic fac-
tors. Using international data of catch and stock over almost 50 years, the authors
estimate the effect of economic growth on the catch and stock. The results sug-
gest that economic growth deteriorates fishery stock and marine ecosystem for
countries with a low income level, while beneficial impacts exists for high income
countries. That also suggests that a decline in catch and indications of stock recov-
ery over the next two decades.
It has been known that it is desirable to adjust the IWI incorporating the total
factor productivity, and many studies have worked on this issue. Chapter 6 con-
tributes to this literature by adding natural capital as an input in estimation of
productivity. The results show that incorporating natural capital in estimation of
productivity affects the estimation outcome and thus affects the IWI.
Chapter 7 explores the determinants of well-being. Previous studies in the lit-
erature discovered that relative income is an important determinant of subjective
well-being. This chapter examines the effects of relative income in greater detail,
decomposing it into a direct effect and an indirect effect through community
attachment.
Chapter 8 applies the life satisfaction approach to evaluate people’s prefer-
ence for green space. Combining survey data and geographic green coverage
data, the authors estimate people’s marginal willingness to pay for green space.
They also examine how the willingness to pay is related to distance, amount, and
knowledge.
Chapter 9 evaluates greenery in Tokyo in monetary value, based on the rela-
tionship between greenery types and well-being types. Utilizing satellite images
with a pixel resolution of 61 cm, the authors of the chapter can extract greenery
data at the tree level and examine the relationship between greenery and well-
being more precisely. They also use multiple measures of well-being and discuss
how the valuation results depend on the choice of the measure of well-being.
The central theme of Chapter 10 is the impact of climate change. It has been
known that developing countries will be affected the most by climate change. This
chapter examines the relationship between development levels and the impact of
climate change on agriculture, in particular on African countries that are classified
as developing countries. One of their main findings is that countries with a lower
income level suffer from the climate change impact more than those with a higher
income level and that some countries with a higher income level overcome the
damage caused by climate change.
Chapter 11 extends the work of Chapter 6, capturing the effect of undesirable
outputs to productivity measure. In the productivity estimation, the authors include
carbon damage as a bad output of production, and show that natural capital and
carbon damage, with GDP, are the main contributors to productivity change.
Chapter 12, using the framework of inclusive wealth, provides a cost-benefit
analysis of the decontamination which began in Fukushima after the great earth-
quake in 2011, which has made some radioactive areas habitable again. They
compare the net benefit from the contamination and that from doing nothing; that
is, waiting until natural radioactive decay and weathering lowers the radioactive
4 Shunsuke Managi
level to a safe one. The results, with much uncertainty as they discuss, indicate
that the contamination was not socially desirable.
Chapter 13 asks fundamental questions about inclusive wealth: (1) How differ-
ent are the sustainability assessments of a nation and its subnational regions?
(2) Are subnational units heterogeneous in term of economic sustainability as
defined by increasing wealth? (3) How sensitive are sustainability assessments to
different population ethics? They answer these questions using Japanese prefec-
ture data.
Chapter 14 evaluates two landfill sites in Indonesia, Tamangapa and Benowo,
assessing their impact to health capital around the landfill sites through meth-
ane (CH4), carbon dioxide (CO2), and hydrogen sulfide (H2S) emission. They
conclude that the health damage is substantial and suggest that Indonesia should
improve its waste management.
Chapter 15 evaluates population planning policy in China through the lens
of human capital. Based on population forecasts from the standard prediction
method, they evaluate the one-child policy which was introduced in 1979 and
phased out at the end of 2015 and the two-child policy which has been effective
since the beginning of 2016. Their results suggest that the two-child policy cannot
substantially improve human capital and recommend that additional policies that
give couples incentives to have more babies be taken into consideration.
Chapter 16 explores the impacts of renewable energy on sustainability across
countries worldwide. Using the IWI as a sustainability indicator, they show that
large-hydro and non-renewable energy have positive impacts on IWI, whereas
wind, traditional bioenergy, modern bioenergy, and small-hydro have negative
impacts, and that “cleaner energy” such as solar energy, wind energy, and bioen-
ergy have negative impacts on the level of natural capital.
Automated driving technology is expected to bring a paradigm shift to trans-
portation and logistics. Chapter 17 evaluates current consumer demand by pur-
chase intention and willingness to pay for fully automated vehicles. The authors
explore the relationship of purchase intention and willingness to pay to subjective
advantages and disadvantages of automated vehicles, using regression analyses.
1 An evaluation of inclusive
capital stock for urban
planning
Hidemichi Fujii and Shunsuke Managi
1 Introduction
Rapid urbanization has caused significant problems in terms of food security,
resource availability, environmental pollution, employment, and living conditions
(UNEP 2012). Sustainable city design has recently received significant academic
and policy attention due to the increasing complexity of cities considering various
aspects such as living, labor, transport, and the environment (Mori and Yamashita
2015; Managi 2016). Rapid urban development needs extensive natural resources
and major government budget allocations to maintain the regional environment,
but government budgets and natural resources are limited. Therefore, efficient
spending and resource use are important factors for achieving sustainable urban
development.
To understand the effect of urban planning policies on sustainability, it is impor-
tant to use a comprehensive target index that covers the various urban factors. An
evaluation index helps policymakers to consider the urban context. Additionally,
an urban policy that refers to only one dimension might worsen other factors.
Many city evaluation indexes have been developed in previous studies; Table 1.1
shows a list of proposed city evaluation indexes, focusing on the purpose, key
evaluation dimensions, and the aggregate of the total score.
The city evaluation indexes listed in Table 1.1 mainly apply flow data related
to economic, social, and environmental factors. Some indexes evaluate capital
stock (e.g., GCCI evaluates human capital, GPI evaluates infrastructure develop-
ment). However, none of these indexes explicitly looks at human, environmental,
and produced capital stocks. Dasgupta et al. (2015) pointed out the problem with
using flow data in evaluation frameworks: “GDP does not record the depreciation
of capital assets even though GDP can increase despite the depletion of natural
resources”. Economic theory offers advantages when evaluating urban capital
stock, as it considers the relationship between flow and capital data. Economic
theory (e.g., Dasgupta et al. 2015) justifies the use of this inclusive capital mea-
sure to assess sustainable development. Once there is a positive increase in stock
value, which is its shadow price multiplied by each stock measure, it can assist
in identifying what subjects the policy needs to support each problem. However,
the existing city evaluation indexes shown in Table 1.1 are limited in their use of
Table 1.1 Literature review of city evaluation indexes
Author Index name Purpose Key evaluation dimensions (factors Aggregation

or indicators)
A. T. Kearney Global Cities To examine a city’s current Business activities, Human capital, Weighted summation
(2014) Index (GCI) performance based on five key Information exchange, of five key evaluation
dimensions. Cultural experience, and Political factors
engagement
Institute for Global Power To evaluate the major cities of the Economy, Research and Sum of all scores of
Urban Strategies City Index world on their global potential and development, Cultural interaction, the key dimension
(2013) (GPCI) comprehensive power. Livability, Environment, and factors
Accessibility
The Economist Global City To rank cities according to their Economic strength, Physical Weighted summation
Intelligence Unit Competitiveness demonstrated ability to attract capital, capital, Financial maturity, of score for key
(2012a) Index (GCCI) businesses, talent, and visitors. Institutional effectiveness, Social dimension factors.
and cultural character, Weighting is based on
Human capital, Environmental and expert interviews.
natural hazards, and Global appeal
The Economist Green City Index To make a major contribution to CO2, Energy, Buildings, Transport, Equally weighted
Intelligence Unit (GrCI) the debate about environmentally Waste and land use, summation of eight
(2012b) sustainable cities. Water, Air quality, and key evaluation factors
Environmental governance rebased out of 100
UN-Habitat City Prosperity To measure the prosperity factors Productivity, Quality of life, Geometric average
(2012) Index (GPI) at work in individual cities, which Infrastructure development, of score for five
pinpoints areas for policy intervention. Environmental sustainability, prosperity factors
Equity and social inclusion, and
Governance and legislation
ARCADIS Sustainable To give a snapshot of sustainability in Literacy, Education, Green Addition of scores
(2015) Cities Index each city, choosing to reflect areas in spaces, Health, Dependency of sustainable cities
(SCI) which local authorities have the power ratio, Income inequality, Work- indicators
to enhance the sustainability of their life balance, Property prices,
city. Transport infrastructure, Energy
use and renewables mix,
Natural catastrophe exposure,
Air pollution, Greenhouse
gas emissions, Solid waste
management, Drinking water
and sanitation, Energy efficiency,
Importance to global networks,
GDP per capita, Ease of doing
business, and Cost of doing
business
International ISO 37120 To measure over a period of time 46 core indicators and 54 Not aggregated into
Organization for the management performance of city supporting indicators addressing one index
Standardization services and quality of life of the city. the Economy, Education, Energy,
(2014) Environment, Recreation,
Safety, Shelter, Solid waste,
Telecommunications and
innovation, Finance, Fire and
emergency response, Government,
Health, Transportation, Urban
planning, Wastewater, and Water
and sanitation
Source: Author selected some of the indices introduced in López-Ruiz et al. (2014) and added recent indexes.
Note: GCI weights each factor as Business activities (30%), Human capital (30%), Information exchange (15%), Cultural experience (15%), and Political engagement
(10%).
8 Hidemichi Fujii and Shunsuke Managi
economic theory to measure sustainability. One potential cause of this is the lim-
ited availability of capital stock data from national statistical databases.
This study proposes a tool for evaluating inclusive urban capital stock using the
Inclusive Wealth Index (IWI) and based on economic theory. The IWI is theoreti-
cally developed by Dasgupta et al. (2015) and empirically elaborated by UNU-
IHDP and UNEP (2012, 2014) (see Managi (2015, 2016) for a review). We define
inclusive urban capital as capital stock as estimated by the IWI to be used for
urban planning to achieve sustainable urban design. We define sustainability of
societal development along which (intergenerational) well-being does not decline
(UNU-IHDP and UNEP 2014).1
Our main objective is to propose an IWI application for urban planning. To
explain the proposed application, we introduce an empirical analysis using data
for 20 Japanese cities. Dasgupta et al. (2015) noted that governments need a
measurement tool that comprehensively records wealth, including reproducible
capital, human capital, and natural capital, to measure sustainable progress. The
advantage of the IWI is that it considers natural capital, human capital, and pro-
duced capital, which are considered to be the key factors in conventional city
evaluation approaches (see UNU-IHDP and UNEP 2014; Managi 2015).2
2 The Inclusive Wealth Index (IWI) for urban planning
2.1 The evaluation of inclusive urban capital stock

What is an effective way to increase inclusive urban capital, and what drives
changes in inclusive urban capital stock? Three factors characterize capital stock.
The first is the volume of capital stock, which indicates whether the capacity
or wealth of cities is maintained over time. The second is the balance of capital
stock, which reflects the urban planning vision and urban characteristics. These
are important pieces of information for selecting the reference city for urban
policy. The third is the efficiency of capital stock use, which shows the capital
productivity of cities.
2.2 Comparing capital stock between cities

To evaluate the first and second factors, the results of the IWI estimation score
can be directly applied. The volume factor can be evaluated using the capital stock
per person, and the balance of capital stock can be evaluated using a share of each
capital stock. These two approaches for country data were introduced as the inclu-
sive wealth of nations in section 3 of chapter 1 in UNU-IHDP and UNEP (2014).
However, each dimension of inclusive urban capital stock assumes a different
role in sustainable urban development, and the relative advantage of urban char-
acteristics differs across cities. Meanwhile, a direct comparison of volume factors
may not sufficiently clarify the relative superiority of urban capital. Thus, we pro-
pose the relative superiority evaluation approach using the concept of inclusive
urban capital stock.
An evaluation of inclusive capital stock 9
To evaluate the relative superiority of urban capital stock, we adopt an empiri-
cal evaluation approach to city performance that uses the DEA method and the
three inclusive capital stock per capita factors estimated by the IWI as output.
This relative evaluation framework for multiple dimension factors is developed
by Despotis (2005). The DEA approach was developed by Charnes et al. (1978)
and has the advantage of using data from multiple inputs and outputs for the
evaluation (Cook et al. 2014).).To evaluate the relative superiority of urban capi-
tal stock, we adopt an empirical evaluation approach to city performance that uses
the DEA method and the three inclusive capital stock per capita factors estimated
by the IWI as output. This relative evaluation framework for multiple dimension
factors is developed by Despotis (2005). The DEA approach was developed by
Charnes et al. (1978) and has the advantage of using data from multiple inputs
and outputs for the evaluation (Cook et al. 2014).To evaluate the relative supe-
riority of urban capital stock, we adopt an empirical evaluation approach to city
performance that uses the DEA method and the three inclusive capital stock per
capita factors estimated by the IWI as output. This relative evaluation frame-
work for multiple dimension factors is developed by Despotis (2005). The DEA
approach was developed by Charnes et al. (1978) and has the advantage of using
data from multiple inputs and outputs for the evaluation (Cook et al. 2014).Des-
potis (2005).Despotis (2005). The DEA approach was developed by Charnes et al.
(1978) and has the advantage of using data from multiple inputs and outputs for
the evaluation (Cook et al. 2014).
We set the three inclusive urban capital stock per capita factors as output.
Under this setting, the integrated evaluation score for the accumulation of inclu-
sive urban capital stock per capita of city k can be described as follows:
Objective function: Max. βk = wp Produced capitalk + wh Human capitalk

+ wn Natural capitalk (1)
s.t.
wp Produced capital j + wh Human capital j + wn Natural capital j

≤1 ( j = 1, 2,, k ,, J ) (2)
wp , wh , wn ≥ 0 (3)
where β is the integrated evaluation score defined from zero to one. β = 1 repre-
sents prolific inclusive capital accumulation per citizen, and a lower β represents
relatively less capital accumulation compared with an efficient city. wp, wh, and
wn indicate the variable weight for produced capital, human capital, and natural
capital, respectively. j is the reference city name (1 ≤ j ≤ J ), and k is the target city
name. To estimate the integrated capital accumulation score for all cities, the DEA
model needs to be applied independently to each of the J cities.
The variable weight is defined as a non-negative number and represents the
relative superiority of capital accumulation compared with other cities. Therefore,
the combination of the three variable weights characterizes a city’s capital accu-
mulation portfolio. The preceding primal DEA model can be transformed into a
dual DEA model as follows.
Objective function: Max. θk (4)
s.t.
j= J
∑ λ Produced capital
j =1
j j ≥ θk Produced capitalk (5)
j= J
∑ λ Human capital
j =1
j j ≥ θk Human capitalk (6)
j= J
∑ λ Natural capital
j =1
j j ≥ θk Natural capitalk (7)
∑λ
j =1
j =1 (8)
0 ≤ λj ≤1 j = 1,, k ,, J (9)
where θ is the integrated evaluation score and is equal to β. λ is the intensity

weight variable for frontier line construction by an efficient city. n is the reference
city name (1 ≤ n ≤ N ), and k is the target city name. To estimate the integrated
capital accumulation score for all cities, the DEA model needs to be applied inde-
pendently to each of the N cities.
Cities can use their reference city as a benchmark for comparison; if they can
catch up to their benchmark city, they will be able to increase their integrated
evaluation score in our model. The reference city can be identified from the λ
score: λj is positive if city j is evaluated as efficient and observed as the reference
city. In the dual DEA model, the reference city is selected from among cities with
a similar portfolio of capital accumulation. Therefore, it becomes relatively easy
to take action to improve capital accumulation in inefficient cities based on the
policies of the reference city, as these cities share similar characteristics.
2.3 Decomposition analysis of urban capital stock change

The two DEA models provide the relative superiority of and the references for
each city’s capital accumulation. Although this approach offers an advantage
by evaluating the volume of capital stock per capita, it is still difficult to clarify
why urban capital stock changes. The main factors affecting changes in inclu-
sive urban capital stock differ between cities, and they depend on the city’s
characteristics and urban planning strategy. To understand the factors that change
urban capital stock, we propose taking a decomposition approach, focusing on
the efficiency and the priorities of government budget expenditures. We choose
efficiency and priorities of government expenditures because government plan-
ning is expected to increase the value of the region, which needs each expendi-
ture to support its activity.
To decompose changes in urban capital stock, the following three indicators
are used: (1) the efficiency of government budget use (EFFICIENCY), (2) the
priorities in government budget allocation (PRIORITY), and (3) the scale of
the government budget (SCALE). We defined the EFFICIENCY indicator as the
change in urban capital stock divided by government budget expenditure, and
it represents the efficiency of budget usage for urban capital stock growth. This
indicator can increase for two reasons: urban capital stock growth while maintain-
ing the same budget or budget expenditure reduction while maintaining urban
capital stock. The reduction of expenditures can be achieved by effectively using
the budget to increase urban capital stock.
The PRIORITY indicator is calculated as the budget expenditure for a specific
area (e.g., education and health) divided by the total budget expenditure and rep-
resents the share of the budget allocated to that specific area. The PRIORITY indi-
cator reflects the relative priority of budget allocations in urban planning. Finally,
the SCALE indicator shows the scale of urban planning capacity, which is the
total amount of urban budget expenditures.
Here, we introduce the equation for the decomposition analysis of the IWI
change in human capital stock ∆IWItHuman , t +1 +1
(= IWItHuman − IWItHuman). Equation (10)
represents the decomposition model for change in human capital stock using three
factors.
, t +1
, t +1 ∆IWItHuman Budget tHuman
∆IWItHuman = × × ∑ Budget ti
Budget tHuman ∑ i Budget ti i
t
= EFFICIENCYHuman t
× PRIORITYHuman ×SCALE t (10)
, t +1
where ∆IWItHuman is the change in urban capital stock in terms of human capital
and i is the type of budget expenditure. The change in capital stock can be under-
stood as a capital flow, considering the depreciation of stock. Government bud-
get expenditures increase capital flows but do not directly increase capital stock.
Thus, it is consistent to evaluate the efficiency of government budget use focusing
on the capital flow per budget expenditure in each field.
Next, we use the decomposition analysis framework to explain changes in capi-
tal flows. To understand the equation more easily, we set the variables as follows.
The flow of human capital from year t to year t + 1 is FHt (= ∆IWItHuman , t +1
), budget
use efficiency for human capital in year t is E H (= EFFICIENCYHuman), the pri-
t t
ority of budget spending on human capital in year t is PHt (= PRIORITYHuman t

),
and the scale of the government budget is St (= SCALE ). In this case, we obtain
t
FHt = E tH × PHt ×St .

Next, we set the change in the flow of human capital as ∆FHt , t +1 = (= FHt +1 − FHt );
then, we obtain equation (11).
∆FHt , t +1 = FHt +1 − FHt = E tH+1 × PHt +1 ×St +1 − E tH × PHt ×St (11)
Here, we set the change of each indicator as ∆E tH,t +1 = E tH+1 − E tH, ∆PHt ,t +1 = PHt +1 − PHt
t ,t +1
∆P
H
t +1
=P
H − PHt , ∆St ,t +1 = St +1 − St. Then, we obtain equation (12) by applying the Laspeyres
index type decomposition approach.
∆FHt ,t +1 = E tH+1 × PHt +1 ×St +1 − E tH × PHt ×St
= (E tH + ∆E tH,t +1 )×(PHt + ∆PHt ,t +1 )×(St + ∆St ,t +1 ) − E tH × PHt ×St
= ∆E tH,t +1 × PHt ×St +1 + E tH ×∆PHt ,t +1 ×St + E tH × PHt ×∆St ,t +1

+ ∆E tH,t +1 ×∆PHt ,t +1 ×St +1 (12)
Following Sun (1998), we transform equation (12) into equation (13) by allo-
cating the interaction term.
1
∆FHt , t +1 = ∆E tH, t +1 × PHt ×St +1 +
2
(∆EtH, t+1 ×∆PHt , t+1 ×St+1 + ∆EtH, t+1 × PHt
1
×∆St , t +1 ) + ∆E tH, t +1 ×∆PHt , t +1 ×∆St , t +1
3

Efficiency chaange effect (ECE)
1
+E tH ×∆PHt , t +1 ×St +
2
(∆EtH, t +1 ×∆PHt , t +1 ×St +1 + EtH ×∆PHt , t +1 ×∆St , t +1 )
1
+ ∆E tH, t +1 ×∆PHt , t +1 ×∆St , t +1
3

Priority change effect (PCE)
1
+E tH × PHt ×∆St , t +1 +
2
(∆EtH, t +1 × PHt ×∆St , t +1 + EtH ×∆PHt , t +1 ×∆St , t +1 )
1 (13)
+ ∆E tH, t +1 ×∆PHt , t +1 ×∆St , t +1
3

Scale change effect (SCE)
In this transformation, we allocate the second-order interaction terms and the

third-order interaction term equally to each decomposed factor. This allocation of
interaction terms follows the complete decomposition model developed by Sun
(1998).
Therefore, the change in the human capital flow (∆FHt , t +1 ) is decomposed
into changes in budget use efficiency (ΔECE, first term), changes in the human
capital priorities in budget allocation (ΔPCE, second term), and changes to the
'DWD(QYHORSPHQW$QDO\VLV
Data variables estimated by IWI Estimation results
Produced capital per capita Integrated evaluation score
Combining two results
Human capital per capita Reference city would provide helpful
information for urban
Natural capital per capita Relative superiority planning.
(e.g., Drivers of inclusive
'HFRPSRVLWLRQ$QDO\VLV urban capital flow in
reference city)
Data variables Estimation results
Change in urban capital flow Efficiency change of
estimated by IWI Cities can use the effort
government budget use
and urban strategy of
Change in specific government Priority change of government their reference city as a
budget expenditure budget allocation benchmark.
Total government budget Scale change of government
expenditure budget expenditure
Figure 1.1 Diagram outlining the approaches in this study
government budget scale (ΔSCE, third term). Considering the three decomposed
factors within the inclusive urban capital flow, we can understand why each city’s
capital accumulation changed. Additionally, considering the relative superiority
of capital accumulation and the reference city information from the DEA model,
the preferable urban planning practice would be to increase inclusive wealth by
referring to the reference city’s activities. Figure 1.1 shows a diagram outlining
the approaches to explaining the relationship between DEA and decomposition
analysis using the IWI as inclusive urban capital.
3 Case study for a Japanese ordinance-designated city
3.1 Data description for the empirical study

In this section, we demonstrate our approach using data for 20 Japanese ordi-
nance-designated cities. A Japanese ordinance-designated city is a city that has
a population larger than 50,000 and that has been designated through a Local
Autonomy Act. We obtain the city level IWI data from Managi (2016), which
estimates the IWI following UNU-IHDP and UNEP (2014). Local government
budget expenditure data are observed from the regional statistics database in the
social and demographic statistics published by the Statistics Bureau, Ministry of
Internal Affairs and Communications, Japan. All monetary data are deflated to
2010 prices. To apply decomposition analysis to the government budget expen-
diture, we categorized it into four groups: human capital investment, produced
capital investment, natural capital investment, and others. We provide the integra-
tion of each budget expenditure group in the appendix.
Table 1.2 shows the basic information for the 20 cities. From Table 1.2, it can
be seen that there is diversity in the population and industry structures among
Table 1.2 Basic data for 20 ordinance-designated cities in 2010
Population Ratio of Day Income Labor Labor Labor

population population per share share share
over 65 ratio capita of the of the of the
years old primary secondary tertiary
industry industry industry
1000 % % 1000 % % %
persons Yen
Sapporo 1914 20.5 100.6 3017 0.13 12.27 87.61
Sendai 1046 18.3 107.3 3230 0.09 11.73 88.18
Saitama 1222 19.1 92.8 3730 0.09 15.93 83.98
Chiba 962 20.7 97.5 3618 0.15 13.71 86.14
Yokohama 3689 20.0 91.5 3883 0.09 17.11 82.79
Kawasaki 1426 16.6 89.5 3838 0.14 24.49 75.37
Sagamihara 718 19.2 87.9 3315 0.37 23.60 76.03
Niigata 812 23.1 101.8 2892 0.47 19.50 80.03
Shizuoka 716 24.6 103.3 3168 0.20 22.89 76.92
Hamamatsu 801 22.6 99.7 3079 0.47 31.11 68.42
Nagoya 2264 20.8 113.5 3705 0.03 17.91 82.06
Kyoto 1474 22.4 108.5 3317 0.08 17.58 82.33
Osaka 2665 22.5 132.8 3132 0.04 17.31 82.65
Sakai 842 22.5 94.4 3281 0.13 24.77 75.10
Kobe 1544 22.9 102.6 3516 0.09 15.85 84.06
Okayama 710 21.3 104.2 3088 0.25 17.85 81.90
Hiroshima 1174 19.7 102.1 3281 0.14 17.34 82.52
Kita-Kyushu 977 25.1 102.7 2974 0.09 21.34 78.57
Fukuoka 1464 17.4 111.9 3305 0.07 12.54 87.40
Kumamoto 734 20.8 N/A 2994 0.49 13.29 86.22
the cities. Kita-Kyushu has a high ratio of its population over 65 years old, while
Kawasaki and Fukuoka have a low ratio. Saitama, Yokohama, Sagamihara,
Kawasaki, and Sakai have a low day population ratio: these cities are located
near Tokyo and Osaka, and many citizens commute to these larger cities. Niigata,
Hamamatsu, and Kumamoto are active in the agricultural industry. Kawasaki,
Hamamatsu, and Sakai have a high labor share in secondary industries compared
to the others.
3.2 Empirical study for the relative evaluation

of inclusive capital stock
Table 1.3 shows the relative evaluation results for inclusive capital stock accumu-
lation using the DEA model. From Table 1.3, Hamamatsu, Nagoya, and Osaka are
evaluated as being efficient cities. The efficiency score represents the achieved
Table 1.3 Relative evaluation score of the inclusive urban capital stock
City name Integrated Weight portfolio (w) Intensity of reference city (λ)
evaluation
score Human Produced Natural Hamamatsu Nagoya Osaka
capital capital capital
Sapporo 0.530 0.000 0.894 0.106 0.323 0.000 0.677
Sendai 0.667 0.149 0.798 0.052 0.235 0.006 0.760
Saitama 0.657 0.993 0.000 0.007 0.076 0.924 0.000
Chiba 0.605 0.190 0.776 0.034 0.148 0.587 0.265
Yokohama 0.646 0.997 0.000 0.003 0.030 0.970 0.000
Kawasaki 0.651 0.999 0.000 0.001 0.007 0.993 0.000
Sagamihara 0.669 0.967 0.000 0.033 0.415 0.585 0.000
Niigata 0.795 0.000 0.745 0.255 0.777 0.000 0.223
Shizuoka 0.952 0.953 0.000 0.047 0.583 0.417 0.000
Hamamatsu 1.000 0.000 0.000 1.000 1.000 0.000 0.000
Nagoya 1.000 1.000 0.000 0.000 0.000 1.000 0.000
Kyoto 0.608 0.000 0.852 0.148 0.450 0.000 0.550
Osaka 1.000 0.000 1.000 0.000 0.000 0.000 1.000
Sakai 0.644 0.996 0.000 0.004 0.037 0.963 0.000
Kobe 0.659 0.153 0.835 0.012 0.051 0.198 0.751
Okayama 0.792 0.191 0.744 0.066 0.291 0.483 0.226
Hiroshima 0.762 0.000 0.865 0.135 0.409 0.000 0.591
Kita-Kyushu 0.530 0.153 0.821 0.026 0.115 0.147 0.738
Fukuoka 0.641 0.000 0.979 0.021 0.062 0.000 0.938
Kumamoto 0.578 0.000 0.860 0.140 0.425 0.000 0.575
20-city average 0.719 0.387 0.508 0.104 0.272 0.364 0.365
Note 1: The integrated evaluation score is defined from zero to 1, and a higher score represents more
urban capital stock per capita.
Note 2: The summation of the weight portfolio equals 1. The type of capital stock with the highest
weight score has relative superiority over the others.
Note 3: The summation of the intensity of the reference city equals 1. The city with the highest
intensity score shares the most similar characteristics with its objective city.
urban capital accumulation per capita compared with a reference point deter-
mined by the efficient cities. The weight portfolio and the intensity of the refer-
ence city represent the relative superiority and the reference point information,
respectively. The summation of both variables is equal to 1.
The weight portfolio shows the relative superiority of capital accumulation per
capita. Capital with a high weight score represents a strength of the city relative to
others. For example, Sapporo has a greater relative advantage in produced capital
accumulation than it has in human or natural capital. Hamamatsu has a relative
advantage in natural capital accumulation. One interpretation of this result is that
Hamamatsu has a large share in the agricultural industry. Similarly, a high weight
for natural capital is observed in Niigata and Kumamoto, which are both popular
agricultural areas. Thus, these three cities have a relative superiority in natural
capital accumulation because they have active agricultural industries.
The weight score for human capital accumulation is high in Saitama, Yoko-
hama, Kawasaki, and Nagoya. These four cities all have high income per capita
in common (see Table 1.2). A high-income family can allocate a sufficient portion
of the household budget to education and health maintenance, which contribute
to increasing human capital accumulation. Therefore, high-income cities have a
relative superiority in human capital accumulation.
The pattern for the reference city correlates with the weight portfolio. For
example, Yokohama and Kawasaki, which observed high weight scores in human
capital accumulation, tend to select Nagoya as a reference city. This high intensity
score means that the integrated evaluation score and weight portfolio are reflected
in the reference city’s characteristics. The assignment of the reference city reflects
its relative superiority to the objective city. Additionally, the characteristics of
capital accumulation for the objective city are similar to those of the reference
city. In our model, the benchmark city is selected from a set of relatively similar
cities, and therefore we compare one city to other cities with similar components
of capital. Thus, our model is valid to understanding the integrated evaluation
score and benchmarking.
We should note that the DEA approach limits the number of variables. Cook
et al. (2014) discussed the limitations of the nonparametric production frontier
approach: “it is likely that a significant portion of decision-making units (DMUs)
will be deemed as efficient if there are too many inputs and outputs given the num-
ber of DMUs”. Nonparametric frontier analysis, including the DEA approach, has
difficulty evaluating small sample datasets.
3.3 Empirical study for a decomposition analysis of

inclusive capital stock
Figures 1.2 to 1.4 show the results of the decomposition analysis for produced
capital, human capital, and natural capital, respectively. The dot in the figures
indicates the change in the inclusive urban capital flow between the first period
(2000 to 2005) and the second period (2005 to 2010). The bar chart shows the
contribution of each factor, and the summation of the bar chart equals the score
represented by the dot.
From Figure 1.2, it can be seen that the changes in produced capital flow are diverse
among cities. Produced capital flow growth is mainly achieved by increasing the effi-
ciency change effect (ΔECE). Meanwhile most cities observed a negative effect from
changes to the priorities in government budget expenditures (ΔPCE). One interpreta-
tion of the growth of ΔECE is the unit cost decline due to technological progress in
civil engineering fields (OECD 2008). Technological progress in infrastructure build-
ing contributes to cost reduction and improved convenience and endurance.
Osaka, which has relative superiority in produced capital accumulation,
increased capital flow through efficiency improvement. The types of government
30,000
ECE PCE SCE F(Produced)
25,000
20,000
15,000
10,000
5,000
–5,000
–10,000
–15,000
Sapporo
Sendai
Saitama
Chiba
Yokohama
Kawasaki
Sagamihara
Niigata
Shizuoka
Hamamatsu
Nagoya
Kyoto
Osaka
Sakai
Kobe
Okayama
Hiroshima
Kitakyushu
Fukuoka
Kumamoto
Figure 1.2 Decomposition analysis of changes in produced capital stock (JPY 100 million)
efforts and urban policies used in Osaka will be more effective when applied to
cities with similar characteristics. As we explained in Section 3.2, the reference
city tends to have characteristics similar to those of the target city. From Table 1.3,
Sapporo, Sendai, and Fukuoka selected Osaka as a reference city. The common
characteristic is that each of these four cities is the largest in its respective region.
To efficiently increase produced capital flows with a limited government budget,
we suggest that these three cities refer to the policies and government activities in
Osaka as helpful information.
Niigata increased produced capital flow due to an increased priority in the bud-
get expenditures, and the budget scale increased while the ΔECE was negatively
affected, which made it difficult to maintain capital accumulation given a lim-
ited future budget. This trend is not seen in the other cities. This trend occurred
because Niigata’s local government budget needed to allocate more to human
health and welfare for its aging population. Therefore, our results suggest that
Niigata should refer to the urban planning policies and efforts in Hamamatsu,
which is its reference city.
Next, we discuss the changes to the flows of human capital (Figure 1.3) and
natural capital (Figure 1.4). Figure 1.3 shows that many cities decreased their
human capital flow due to negative ΔECE, especially Sapporo. The main reason
for this decrease in the human capital flow is that low fertility has become more
serious in Japan. Another possibility is the growth of the aging population. The
main factor decreasing the human capital flow in Sapporo is a decline in the health
capital flow due to a lower quality of life (see Managi 2016).
From Figure 1.4, the natural capital flow increased in many cities due to ΔECE
growth. The unique point is that Sapporo increased natural capital flow as its
5,000
–5,000
–10,000
ECE PCE SCE F(Human)

–15,000
–20,000
Sapporo
Sendai
Saitama
Chiba
Yokohama
Kawasaki
Sagamihara
Niigata
Shizuoka
Hamamatsu
Nagoya
Kyoto
Osaka
Sakai
Kobe
Okayama
Hiroshima
Kitakyushu
Fukuoka
Kumamoto
Figure 1.3 Decomposition analysis of changes in human capital stock (JPY 100 million)
800
ECE PCE SCE F(Natutral)
700
600
500
400
300
200
100
–100
–200
Sapporo
Sendai
Saitama
Chiba
Yokohama
Kawasaki
Sagamihara
Niigata
Shizuoka
Hamamatsu
Nagoya
Kyoto
Osaka
Sakai
Kobe
Okayama
Hiroshima
Kitakyushu
Fukuoka
Kumamoto
Figure 1.4 Decomposition analysis of changes in natural capital stock (JPY 100 million)
ΔPCE increased. Meanwhile, Hamamatsu, which is the reference city of Sapporo,

successfully increased natural capital stock through ΔECE improvement. These
results imply that Sapporo has the potential to increase natural capital stock while
lessening the budget priority by referring to the urban planning effort in Hama-
matsu city.
As the preceding empirical study shows, a decomposition analysis of capi-
tal flows can be used to evaluate urban planning efforts from comprehensive
viewpoints using the IWI. Additionally, the results of the DEA model, in terms of
relative superiority and reference city information, are readily available. By com-
bining the results of the decomposition analysis and the DEA model, we can pro-
vide helpful information for sustainable urban planning and policy construction.
4 Discussion and conclusion

Mori and Christodoulou (2012) identified the issues with city evaluation indexes.
They reviewed 14 sustainability indexes and noted that four factors should be
included in a city evaluation: (1) considering the triple bottom line of sustainabil-
ity, (2) creating the index for the expressed purpose of assessing urban sustain-
ability, (3) capturing the effects of leakage on other areas in the environmental
dimension, and (4) assessing the world’s cities in both developed and developing
countries in an equitable manner.
Inclusive urban capital considers economic components such as produced cap-
ital; social components, including health and education as human capital; and
environmental components such as natural capital. Therefore, inclusive urban
capital clearly considers the first criterion, the triple bottom line. Additionally,
considering the second criterion, inclusive urban capital is created for the purpose
of assessing urban sustainability and urban planning. However, inclusive urban
capital does not meet the third and fourth criteria. To meet these requirements, the
following are needed.
To meet the third criterion, further research into the leakage effect in the envi-
ronmental dimension should be conducted, focusing on networks within and
between cities. These networks are highly developed due to the development of
information and transportation technologies. Thus, it is not enough to evaluate
one region or one city to understand urban sustainability because this clearly does
not consider the network of cities. Therefore, further research focusing on city
networks (e.g., trade, transportation, people-to-people exchange) is important to
develop the evaluation application for urban sustainability.
Issues in meeting the fourth criterion are related to the limited data availability
in developing countries because the IWI needs a wide range of data variables. The
development of data estimation analysis is important to promoting future research
addressing inclusive urban capital. In addition to the country-level analysis intro-
duced in UNU-IHDP and UNEP (2014), utilizing inclusive urban capital stock
provides a comprehensive evaluation of city performance.
To apply inclusive urban capital stock to city design, a database must be con-
structed to estimate the IWI. Research focusing on the regional IWI has already
been conducted in the transportation field in cities worldwide (Nakamura et al.
2015), in the disaster management field in Japanese cities (Tanikawa et al. 2014),
and on natural, human, and produced capital in the United States (UNU-IHDP
and UNEP 2012). A large-scale city dataset is available from the Organisa-
tion for Economic Co-Operation and Development (OECD) database, which
includes 281 metropolitan areas in 29 countries (https://stats.oecd.org/Index.
aspx?DataSetCode=CITIES). The data variables in this database are divided into
eight categories: demographics, land cover, urban forms, territorial organization,
economics, the environment, labor, and innovation. These variables provide help-
ful information for both estimating inclusive urban capital stock and considering
the changes in capital flows.
Additionally, corporate financial and environmental databases are published
with free access and downloadable data (Fujii and Managi 2016). The collabora-
tion between corporate data and regional data will create a new data network useful
for creating an effective business strategy that considers regional characteristics.
Finally, a limitation of the IWI evaluation method is that it has difficulty
addressing non-marketable goods such as cultural and religious factors. Accord-
ing to Arrow et al. (2012), capital stock should be accumulated via investment
created by allocating those resources with decreasing current consumption. In
this sense, social capital, including cultural activities, might not be suitable for
evaluation as capital stock in the IWI. However, religion and culture are impor-
tant factors in the creation of human capital stock. Thus, further research should
investigate a more comprehensive evaluation approach considering regional dif-
ferences in culture and religion.
Acknowledgments
This research was funded by the Grant-in-Aid for Specially Promoted Research
[26000001B]; the Ministry of Education, Culture, Sports, Science, and Technol-
ogy, and S15 of the Ministry of the Environment, Japan. The results and conclu-
sions of this chapter do not necessarily represent the views of the funding agencies.
Table 1A.1 Grouping of local government budget expenditures
Government budget expenditure category Grouping in this study

Civil engineering works (maintenance of civil engineering Expenditure for produced
works, roads and bridges, rivers, streams and coasts, ports capital accumulation
and harbors, city planning, dwellings, airports)
Welfare (social welfare, welfare for the elderly, child welfare, Expenditure for human
livelihood protection, disaster relief) capital accumulation
Hygiene (public health, tuberculosis control, health centers, Expenditure for human
refuse disposal) capital accumulation
Education (educational administration, elementary schools, Expenditure for human
lower secondary schools, upper secondary schools, capital accumulation
universities, social education, health and physical education)
Agriculture, forestry, and fisheries (farming, raising Expenditure for natural
livestock, agricultural land, forestry, fisheries) capital accumulation
Notes
1 As UNU-IHDP and UNEP (2014) point out, “Wealth accounting also internalizes sus-
tainability by tracking the changes in the value of a nation’s capital asset stocks”. Thus,
considering the change in the IWI would be a useful tool for evaluating a change in
sustainability.
2 The IWI also offers advantages by evaluating ecosystem services using economic the-
ory. The IWI evaluation framework introduces the shadow price approach to evaluate
ecosystem services. The shadow price reflects the marginal value contribution to inter-
generational well-being for a unit change in the respective capital asset (UNU-IHDP
and UNEP 2014). Therefore, an ecosystem evaluation approach using shadow prices
captures human behavior and preferences in monetary terms, allowing easy compari-
son with other monetary data. Costanza et al. (1997) pointed out that the total value
of ecosystem services can be derived from shadow prices for all of the flows between
processes as well as for the net outputs of the system. Therefore, we can understand
natural capital stock, including ecosystem services, in monetary terms by multiplying
the shadow price of exhaustible resources and that of environmental pollution by the
volume of the resource stock and emissions.
Bibliography
ARCADIS. 2015. Sustainable cities index: Balancing the economic, social and environ-
mental needs of the world’s leading cities. https://s3.amazonaws.com/arcadis-whitepaper/
arcadis-sustainable-cities-index-report.pdf
Arrow, K. J., P. Dasgupta, L. H. Goulder, K. J. Mumford, and K. Oleson. 2012. Sustain-
ability and the measurement of wealth. Environment and Development Economics
17:317–353.
Charnes, A., W. W. Cooper, and E. Rhodes. 1978. Measuring the efficiency of decision
making units. European Journal of Operational Research 2:429–444.
Cook, W. D., K. Tone, and J. Zhu. 2014. Data envelopment analysis: Prior to choosing a
model. Omega 44:1–4.
Costanza, R., R. d’Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S.
Naeem, R. V. O’Neill, J. Paruelo, R. G. Raskin, P. Sutton, and M. van den Belt. 1997.
The value of the world’s ecosystem services and natural capital. Nature 387:253–260.
Dasgupta, P., A. Duraiappah, S. Managi, E. Barbier, R. Collins, B. Fraumeni, H. Gundim-
eda, G. Liu, and K. J. Mumford. 2015. How to measure sustainable progress. Science
350:748.
Despotis, D. K. 2005. A reassessment of the Human Development Index via data envelop-
ment analysis. The Journal of the Operational Research Society 56:969–980.
The Economist Intelligence Unit. 2012a. Hot spots: Benchmarking global city competitive-
ness. www.citigroup.com/citi/citiforcities/pdfs/eiu_hotspots_2012.pdf
The Economist Intelligence Unit. 2012b. The green city index: A summary of the green city
index research series. Siemens AG, Munich, Germany.
Fujii, H., and S. Managi. 2016. The trend of corporate environmental management study
and database. Environmental Economics and Policy Studies 18:265–272.
Institute for Urban Strategies. 2013. Global power city index 2013. The Mori Memorial
Foundation, Tokyo, Japan.
International Organization for Standardization. 2014. Sustainable development of commu-
nities: Indicators for city services and quality of life, ISO 37120, International Organiza-
tion for Standardization (ISO), Geneva, Switzerland.
Kearney, A. T. 2014. Global cities index and emerging cities outlook: Global cities, present
and future. A. T. Kearney, Chicago, IL, US.
López-Ruiz, V., J. Alfaro-Navarro, and D. Nevado-Peña. 2014. Knowledge-city index con-
struction: An intellectual capital perspective. Expert Systems with Applications
41:5560–5572.
Managi, S. (Ed.). 2015. The economics of green growth: New indicators for sustainable
societies. Routledge, New York, NY.
Managi, S. (Ed.). 2016. The wealth of nations and regions. Routledge, New York, NY.
Mori, K., and A. Christodoulou. 2012. Review of sustainability indices and indicators:
Towards a new City Sustainability Index (CSI). Environmental Impact Assessment
Review 32:94–106.
Mori, K., and T. Yamashita. 2015. Methodological framework of sustainability assessment
in City Sustainability Index (CSI): A concept of constraint and maximisation indicators.
Habitat International 45:10–14.
Nakamura, H., T. Tamaki, H. Fujii, and S. Managi. 2015. Efficiency and emissions from
urban transport: Application to world city level public transportation. Working Paper,
Kyushu University, Fukuoka, Japan.
OECD [Organization for Economic Cooperation and Development]. 2008. OECD Eco-
nomic Surveys: Japan 2008. OECD Publishing, Paris, France.
Sun, J. W. 1998. Changes in energy consumption and energy intensity: A complete decom-
position model. Energy Economics 20:85–100.
Tanikawa, H., S. Managi, and C. M. Lwin. 2014. Estimates of lost material stock of build-
ings and roads due to the great East Japan earthquake and tsunami. Journal of Industrial
Ecology 18:421–431.
UNEP [United Nations Environment Program]. 2012. Sustainable, resource efficient cities:
Making it happen! United Nations Environment Program, Nairobi, Kenya.
UN-Habitat. 2012. State of the world’s cities 2012/2013: Prosperity of cities. United
Nations Human Settlements Programme (UN-Habitat), Nairobi, Kenya.
UNU-IHDP [United Nations University International Human Dimensions Programme] and
UNEP [United Nations Environment Program]. 2012. Inclusive wealth report 2012:
Measuring progress toward sustainability. Cambridge University Press, Cambridge, UK.
UNU-IHDP [United Nations University International Human Dimensions Programme] and
UNEP [United Nations Environment Program]. 2014. Inclusive wealth report 2014 mea-
suring progress toward sustainability. Cambridge University Press, Cambridge, UK.
2 Impact of infrastructure
in India
Iswarya Sankaralingam, Moinul Islam,
Wataru Nozawa and Shunsuke Managi
1 Introduction
Infrastructure plays a vital role in the economic and overall development of a
country. In India, due to basic infrastructure constraints, economic growth is
heavily strained. Limitations in infrastructure could curtail economic growth and
productivity and even hamper the overall development of a country. With India
being one of the fastest growing major economies in the world, the need for more
and improved infrastructure is growing rapidly. There is much evidence (Estache
2004, 2006; Jones 2004) that high-quality infrastructure can increase productiv-
ity and growth by (1) stimulating economic activities, (2) reducing transaction
and trade cost and thereby improving competitiveness, (3) creating markets by
improving access, and (4) creating jobs and helping reduce the income gaps in
developing countries. Therefore, for India to move towards a sustainable form of
growth, it is essential to improve and develop the current state of infrastructure
in the country.
India is the fastest growing major economy in the world, with a growth rate of
7.1%, but this economic growth is constrained due to the availability of only basic
infrastructure. The India Infrastructure Report 2006 notes that there is a wide gap
between the potential demand for infrastructure and the available supply. To be
able to sustain its economic growth, India needs to invest more in infrastructure
and develop the quality and expand the current infrastructure network across the
country. Understanding this growing demand, the Indian government has been
increasing public spending in the infrastructure sector, as is evident from Fig-
ure 2.1. Hence, it is important to study and understand the impact of infrastructure
development on the Indian economy.
In addition, growth and development must be sustainable. The Inclusive Wealth
Report 2014 recommends the estimation of impacts of a given infrastructure pol-
icy or project on the trajectories of the capital stocks that comprise wealth to
move towards a more inclusive index. The Inclusive Wealth Index is increasingly
being used in project evaluation (Collins 2016), aiming towards a more sustain-
able framework.
With sustainable development in progress, this study analyzes the impact of
infrastructure development on the Indian economy and inclusive wealth.
24 Iswarya Sankaralingam et al.
Public Infrastructure Investment as a % of GDP

9
8
7
6
5
% of GDP
4
3
2
1
0
1998 2000 2002 2004 2006 2008 2010 2012
Year
Figure 2.1 Public infrastructure investment as percentage of Indian GDP

Source: Planning Commission, Government of India.
2 Literature review
2.1 Definition of infrastructure

The term “infrastructure” is often used vaguely and takes different definitions in
many places. Generally, infrastructure is meant to include electricity, gas, tele-
com, transport, water supply, sewerage and sanitation. According to the Organ-
isation for Economic Co-Operation and Development (OECD), the definition of
infrastructure is the system of public works in a country, state or region, including
roads, utility lines and public buildings. Development economists refer to infra-
structure as social overhead capital, while some research papers consider public
capital as a proxy for infrastructure.
In this chapter, we have considered infrastructure to be the physical stock of the
transport and communication sectors. This is mainly because capital investments
do not always directly reflect the project output due to issues such as inefficient
institutions and misuse of public capital. In addition, public capital might include
other projects such as schools or hospitals. If public capital is used as a proxy
for infrastructure, we might greatly underestimate the impact of infrastructure
growth. Therefore, considering the physical stock of infrastructure directly is the
most reasonable strategy, considering the purpose of this study.
2.2 Empirical studies showing relationship between

infrastructure development and economy
In economic studies, there is a long-standing tradition concerning how infrastruc-
ture can be incorporated into the standard setting. Arrow and Kurz (1970) were
Impact of infrastructure in India 25
the first to formalize the mixture of produced capital and public infrastructure.
Barro (1990) incorporated public expenditure into infrastructure in an endoge-
nous growth model, and Futagami et al. (1993) provided an extension treating
infrastructure capital.
In terms of empirical research, seminal work was done by Aschauer (1989) on
the role of infrastructure in economic growth. Furthermore, he suggests that lack
of infrastructure can reduce productivity. More recent studies (Gramlich 1994;
Garcia-Mil’a et al. 1996; Canning and Pedroni 2004), however, question this high
impact on output and suggest that this very high productivity reported in prior
work could be due to issues such as reverse causality, data from developed nations
only and spurious correlation due to non-stationarity of data.
Sahoo and Dash (2009) conducted a study on India specifically with country-
level data and analyzed their data using a simple ordinary least squares (OLS)
regression. Their results show that physical infrastructure stocks, capital invest-
ments and labour force play an important role in the economic growth of India.
Since there are contradictory results from various studies, it is pivotal to address
these issues with reference to the context of region and society as well. Based on
the knowledge gained from existing literature on this topic, we have developed
our regression model to suit the current research interest. The model is discussed
in the following sections.
2.3 Impacts of different infrastructure sectors
2.3.1 Transportation
Transport infrastructure is one of the most extensive public sectors, and its impact
is felt differently depending on the stage of development in the region (Estache
and Garsous 2011). For developing countries, the growth impacts from transpor-
tation are not very high, as the initial stocks are already mature. However, for
developing countries, the impact is still quite high, as these regions need these
facilities to reach better standards to be on par with the rest of the world.
The impacts of a more extensive and improved transport infrastructure would
be mostly positive, aside from some environmental degradation. Direct benefits
would be the increase in mobility and safety and the reduction of travel time, and
this would translate into economic benefit through reduced costs and improved
efficiency. According to the OECD Programme of Research on Road Transport
and Intermodal Linkages, the socio-economic spillover of transport infrastruc-
ture is also very significant in developing countries. Improved accessibility would
increase the market and labour size, thereby improving competitiveness. Con-
struction and maintenance of infrastructure projects would generate employment
in the region. Transportation infrastructure projects also have a positive impact by
promoting social inclusion in the region. In contrast, the environmental impacts
include issues such as air and water quality, noise, use of natural resources and
other externalities. This needs to be carefully considered when planning projects
as well.
With India being a developing country, transport infrastructure is still expected
to have a positive impact on the economy, which is why the main focus of public
infrastructure spending is on roadways and railways. Understanding the impor-
tance of this sector, the government set a target of building 15,000 km of road-
ways for 2016 and constructed roadways amounting to more than half of this
target, which is the highest number of kilometres built in a year in India (Hindu-
stan Times, April 2017). Railways are not discussed in detail or included in our
model because of the lack of availability of state-level data.
2.3.2 Energy
Various studies have documented the importance of access to electricity for devel-
opment, and all developing countries show a positive impact of energy infra-
structure on economic growth and human development. Garsous (2012) finds that
energy infrastructure has the most robust positive impact when compared with
other infrastructure sectors. Energy infrastructure is crucial because all other sec-
tors are inter-linked and dependent on energy as an input to function.
The Asian Development Bank (2005) reports that in rural India, strong links
between rural electrification improvements and poverty reduction have been
noted. With India’s growing economy, the demand for electricity is increasing,
and inability to meet this demand is one of the obstacles to India’s development.
Hence, it is imperative that the government increase electricity production capac-
ity, as highlighted in the Twelfth Five-Year Plan. Since 2008, under the National
Action Plan for Climate Change, India has also increasingly focused on renew-
able energy sources such as solar, wind and biomass, aiming for clean and sustain-
able development.
Although the energy sector is an important and integral part of public infra-
structure, there are no energy or electricity variables included in our main regres-
sion model. This is primarily because the total electricity consumption variable
(in terms of gigawatt hours) was highly correlated with other input variables such
as capital and labour. This issue is further discussed in Section 3 on methodology
in a more detailed manner.
2.3.3 Telecommunications
Telecommunications infrastructure includes facilities for fixed telephones, mobile
phones and internet. The impact of telecom on output growth is probably the most
researched, as there is more reliable data documented and made available. Most
studies find a positive impact of telecom infrastructure on gross domestic prod-
uct (GDP) and labour productivity. With the inter-dependency between fixed and
mobile phones, the exact magnitude of this impact is still debatable. In more recent
studies (Chakraborty and Nandi 2011), access to the internet is highly important
to increase competitiveness and social returns. Studies also note that the faster the
internet is, the higher the payoff. Interestingly, the impact of telecom penetration
on total economic output for developing countries is significantly lower than that
Teledensity
80
per thousand population
60
40 Fixed
Mobile
20
0
2008 2009 2010 2011 2012
Figure 2.2 Teledensity per 1000 population

Source: Infrastructure Statistics 2014, MOSPI, Government of India.
reported for OECD countries (Sridhar and Sridhar 2007). This could be because
the payoffs are lower when they are not regulated properly, although there might
be sufficient investment.
India is the one of the largest markets in telecommunications in terms of both
mobile phone and internet users. Reforms and liberal governmental policies in this
sector, such as setting up an independent Telecom Regulatory Authority of India
(TRAI), deregulation of the upper limit on foreign direct investment (FDI) and
other such initiatives, have helped the telecommunications sector to strengthen
and grow in India. Fixed telephone lines are slowly on the decline as the mobile
market is booming (see Figure 2.2). Furthermore, a majority of the private capital
in infrastructure projects is heavily invested in the telecommunications sector.
3 Methodology
3.1 Theoretical framework

The main body of this chapter focuses on quantitative assessment of the impact of
infrastructure stock on the Indian state GDP (GSDP). To analyze this effect, it is
ideal to utilize the production function and growth model as it relates the physi-
cal output of a production process to the physical inputs or factors of production,
namely, capital, labour and infrastructure.
To briefly introduce the concept of the production function, the basic idea is
that the quantity of output can be introduced or produced as a function of the
quantity of inputs to production. There can be a number of different inputs to pro-
duction, often known as “factors of production”, but they are generally designated
either capital or labour. The particular functional form of the production function
depends on the specific technology and production processes used.
Additionally, the neo-classical growth model states that any change in technol-
ogy will cause a shift in the production function. It becomes imperative to include
physical infrastructure stock as an input to reflect technological advancements,
apart from capital investments and labour force.
Based on Sahoo and Dash (2009), to study the impact of infrastructure devel-
opment on the state GDP, we have considered a generalized production function
and a neo-classical growth model. Our production function framework studies
the impact of inputs such as capital investment, labour force and infrastructure
development on the GSDP. The production function is as follows:
Y = f(KF, L, I)
where Y is the gross output produced in an economy, KF is capital formation or

investments, L stands for labour force and I represents the various infrastructure
indicator variables as a reflection of technical advancements.
3.2 Infrastructure variables considered and used in the

regression model
In India, the road network is a major asset for infrastructure. It helps connect
every part of the country, making it the most valuable asset to Indian infrastruc-
ture. Next to roadways, railways are the second biggest infrastructure network in
India. However, due to lack of availability of state-level railway data, we do not
include the railway network in our model. Third, air transport is one of the biggest
public infrastructures and is included in our model.
While many papers do not consider telecommunications and internet facilities
as traditional infrastructure, we consider them as infrastructure variables because
of the effect these sectors have had on the infrastructure development in India.
We use the teledensity variable as a proxy for telecommunications penetration
in the population. Our teledensity data include both wireless and wired mode of
communication. Recently, India has seen a boom in the usage of mobile phones
as compared to fixed telephones. The majority of these mobile phone users access
the internet through their mobile devices as well. In 2016, therefore, the over-
all influence of the growing telecommunications sector along with access to the
internet is captured through this teledensity variable (see Figure 2.3).
Although total electricity consumption was initially considered as an input
variable in our multivariate regression model, we did not include it in our final
model. Even though the energy sector is an important sector of public infrastruc-
ture, we dropped the electricity input variable in our model to avoid the problem
of multicollinearity, as it could be a serious issue. Multicollinearity is a statis-
tical phenomenon where two or more variables are correlated with each other
(i.e., to a certain extent, one variable can be linearly predicted by another). If
multicollinearity exists, then the standard error of each variable increases. With
Composition of Internet subscription
Fixed Wireless
0.4%
Wired
6%
Mobile Wireless
94%
Wired Fixed Wireless Mobile Wireless
Figure 2.3 Composition of internet subscription, 2016

Source: Telecom Regulatory Authority of India.
an increased standard error, the regression coefficient can become statistically

insignificant, even when it should show significant results. The issue of multicol-
linearity can be addressed in two ways:
• By removing the highly correlated variable and using the model with the
highest R2
• By using principal component analysis (PCA).
When principal component analysis was conducted, many variables were left
unexplained by the component, causing loss of information. Hence, the electricity
variable was eventually dropped (see Table 2.1).
3.3 Pre-estimation test: correlation

As mentioned earlier, correlation among variables was checked, and the issue of
multicollinearity was resolved (see Table 2.2).
3.4 Post-estimation: Hausman test

For a panel dataset, one important test to be conducted before finalizing the
regression model is the Hausman test. The Hausman test helps decide between a
fixed or random effects model by checking whether unobserved effects are cor-
related with the regressors or the explanatory variables. Unobserved effects are
part of the error term that does not change over time. In a fixed effects model, the
unobserved effects are allowed to be arbitrarily correlated with the explanatory
Table 2.1 Principal component analysis test results
Principal components/correlation
Variables considered: road, air, tele, elec
Component Eigenvalue Difference Proportion Cumulative

Comp 1 2.30897 1.33202 0.5772 0.5772
Comp 2 0.976949 0.510619 0.2442 0.8215
Comp 3 0.46633 0.218574 0.1166 0.9381
Comp 4 0.247756 0.0619 1.0000
Principal components (eigenvectors)
Variable Comp 1 Unexplained

road 0.5152 0.3872
air 0.5715 0.246
tele 0.2253 0.8828
elec 0.5977 0.1751
Descriptive statistics
Number of observations = 91
Number of components = 1
Trace = 4
Rho = 0.5772
Table 2.2 Correlation matrix
Capital Labour Elec Road Air Tele

Capital 1.000
Labour 0.7978 1.000
Elec 0.7968 0.9179 1.000
Road 0.4312 0.4381 0.5787 1.000
Air 0.6262 0.6740 0.7255 0.5504 1.000
Tele 0.2190 0.3460 0.3092 0.0619 0.1369 1.000
variables in each time period. On the other hand, in a random effects model, these
unobserved effects are assumed to be uncorrelated with the explanatory variables
in each time period.
The null hypothesis of the Hausman test is that the preferred model is the ran-
dom effects model, the alternative being the fixed effects model. In the test results,
if the probability > chi2 is significant, then our null hypothesis can be rejected and
a fixed effects model can be used. If we fail to reject the null hypothesis, then the
random effects model must be used.
The test results of our Hausman test were insignificant, and hence we fail to
reject the null hypothesis. Thus, the random effects model is the preferred model
for our panel dataset. That is, because the p-value of Hausman test statistics is
0.83, the random effects model is chosen.
3.5 Regression model

In econometric analysis, a model that is linear in its parameters and in which the
dependent variable is a function of the independent variables plus an error term
is known as a multiple linear regression (MLR). In our research, to empirically
study the impact of infrastructure development on the Indian economy, we use the
following multiple linear regression model:
gsdpit = α0 + β1 capitalf + β2 labour + β3 road + β4 air + β5 tele + εit
where gsdp is the dependent variable, and capitalf, labour, road, air and tele are
the independent variables. The parameter α is the intercept (often known as the
constant), βn are the regression coefficients, and ε is the error term. t denotes
the time variable, and i represents the entities or cross-sections in the panel. In
this case, i represents each of the Indian states considered, and t is annual data.
Regression coefficients represent the mean change in the output (GSDP) for one
unit of change in the input variable while holding other predictors in the model
constant.
The description of the variables used is as follows: gsdp is real gross state
domestic product (in 10 million Indian rupees (hereinafter INR)); capitalf is real
gross fixed capital formation (in 10 million INR); labour is the total number of
persons engaged; road is total road length (km); air is average domestic cargo
handled per day (tons); and tele stands for teledensity (per 100 people).
4 Data
Panel data, also known as longitudinal data, represents a dataset in which the
behaviour of entities is observed across time. In our study, the entities are Indian
states with data collected annually for the period 2008–2012. While India has
29 states, two states were omitted because of data non-availability: Arunachal
Pradesh and Mizoram. The state of Telangana was created only in the year 2014,
which is outside the studied time period. Hence, Telangana is also omitted from
this dataset. The data of seven northeastern states are combined and averaged
because of their small size, regional coherence and similar data format made
available by the government. Eventually, our dataset has 19 states/groups, and
the total number of observations is 91. Another point to report is that our data are
“highly balanced”, meaning that observations for all entities i are available for all
t in the period of 5 years.
The summary statistics of our dataset are presented as follows (see Table 2.3):
Table 2.3 Summary statistics of the compiled dataset
(1) (2) (3) (4) (5)
Variables N mean SD min max

year 135 2010 1.419 2008 2012
road 135 133,481 114,856 1873 410,521
air 125 60.71 129.2 0 669
tele 91 51.26 28.49 3.596 120.7
elec 135 20,383 22,051 182.7 96,642
labour 133 774,880 891,067 4108 3.536e+06
gsdp 135 147,256 155,321 2178 775,610
capitalf 132 6909 8202 0.582 34,109
Table 2.4 Detailed description of variables used
Variable Description Unit Source

name
gdp Real gross state Constant Indian Directorate of Economics and Statistics
domestic product rupees (state government)
(in 10 million)
base 2005
capital Real gross fixed Constant Indian Annual Survey of Industries – MOSPI
capital formation rupees
(in 10 million)
base 2005
labour Total number of Persons Annual Survey of Industries – MOSPI
people engaged
road Total road length Kilometres Infrastructure Statistics 2014 – MOSPI
air Average domestic Tons Infrastructure Statistics 2014 – MOSPI
cargo handled per day
tele Teledensity Telephones per Infrastructure Statistics 2014 – MOSPI
100 population
where N is the number of observations, SD stands for standard deviation, min

denotes the minimum value of a given variable, and max denotes the maximum
value of the variable.
Table 2.4 presents the descriptions, units and sources of all the variables used
in our study.
The panel dataset used in the model is secondary data, as it has been compiled
from various ministries of the Indian government. The real gross state domestic
product at constant prices with base year 2005 was prepared by the Directorate of
Economics and Statistics and retrieved from the Planning Commission, Govern-
ment of India. Recently, the Planning Commission was reorganized and renamed
NITI Aayog, and state-level GDP data can be downloaded from their website.
Data on state-level nominal gross capital formation and total number of people
engaged were obtained from the Annual Survey of Industries (ASI), Ministry of
Statistics and Programme Implementation, Government of India. The nominal
capital data were converted into real capital formation at constant prices with
base year 2005 by calculating the GDP deflator with the data available. State-level
infrastructure data on road length, average number of flights per day and teleden-
sity were collected from the statistical publication series of the Ministry of Sta-
tistics and Programme Implementation (MOSPI), titled Infrastructure Statistics.
4.1 Conversion from nominal to real form

Nominal GSDP is gross state domestic product evaluated at the current market
price, GSDP being the monetary value of all the finished goods and services pro-
duced within a state’s borders in a specific time period. Nominal GSDP may vary
because the price of commodities increases or decreases from year to year. There-
fore, it becomes necessary to control for inflation (or deflation) when compar-
ing GSDP or any other monetary variables between different time periods. Real
GSDP is the estimation made at constant prices after adjusting for inflation. The
same condition applies for all variables that are in monetary values, such as gross
fixed capital formation.
Since both real and nominal GSDP data were readily sourced from government
websites, only the nominal gross fixed capital formation data needed to be con-
verted to their real form, for which the GSDP deflator was required. The GSDP
deflator is an economic metric that converts monetary values measured at current
prices into constant values. The GSDP deflator was calculated as a ratio of nomi-
nal and real GSDP data, which was already available.
Nominal GSDP
GSDP deflator = ×100
Real GSDP
Next, the nominal gross fixed capital formation data were divided by the calcu-
lated deflators and multiplied by 100. We used the obtained values as real capital
formation data.
4.2 Reasons for choosing these variables

For the production function, variables indicating capital and labour are required.
Based on Sahoo and Dash (2009), gross fixed capital formation was used as the
capital indicator. For the labour variable, total number of people was used instead
of the regular labour force definition, primarily because of data availability. Only
the total number of people engaged (both directly and indirectly) was available
at the Indian state level. Data on other labour indicators such as employment rate
or total labour force were sparse, which could lead to an unbalanced dataset and
a reduction in the number of observations. Hence, the total number of people
engaged was used to represent the labour indicator.
For the infrastructure indicators, the transport, energy and telecommunications
sectors were initially considered, with energy being dropped eventually due to
the issue of multicollinearity. In the transport sector, road denoted by total road
length and air denoted by the average amount of domestic cargo handled were
used in the model. While domestic cargo handled cannot be a direct indication,
we use it as a proxy variable to reflect the quantity aspect of the airways of each
state. Data on railways, representing a key public sector in India, were unable to
be incorporated because the data made available by the government were grouped
by region and not by state and were therefore unusable for our research. Finally,
the telecommunications sector was represented by the teledensity, which is the
number of telephones (both fixed and mobile) per 100 people.
4.3 Limitations in the model

As major conclusions are derived from the economic modelling, we must also
consider the limitations of the model. One limitation is the reverse causality that
could exist between our dependent variable GSDP and independent variables such
as the infrastructure indicators. Reverse causality is when the dependent variable
could effect a change on the independent variables considered in the model. This
can be better explained by using our model as an example. Causality is when the
infrastructure variables cause a change in the GSDP. Reverse causality is when
changes in the GSDP are expected to affect the infrastructure variables. Since this
relationship between economy and infrastructure sector is plausible, it is impor-
tant to check for reverse causality in our model.
To run the Granger causality test on our panel data, a minimum time period of
30 years is required for the test code to run efficiently. Since our dataset has only
5 years of observations, we are unable to test for reverse causality. Another way
this could be addressed is by introducing time-lagged effects, but that method
would greatly reduce the size of the dataset, which is already relatively small.
Hence, the case of reverse causality is not addressed in our study, and we confine
our study to a simple regression model.
5 Results
The main focus of the regression results is the regression coefficient (β). The
regression coefficient represents the mean change in the output (GSDP) for 1 unit
of change in the input variable while holding other predictors in the model con-
stant. When interpreting the regression coefficients, two concepts must be consid-
ered: statistical significance and magnitude. The regression coefficients appear in
Table 2.5.
The statistical significance of a regression coefficient can be checked through
the p-values reported along with the regression results. Statistical significance
is the likelihood that a relationship between two or more variables is caused by
something other than random chance. The p-value is the smallest significance
level at which the null hypothesis can be rejected. Equivalently, it also means
Table 2.5 Regression coefficients
Variables gsdp
capitalf 1.197
(0.778)
labour 0.0438***
(0.0122)
road 0.479***
(0.0680)
air 398.5***
(67.46)
tele 323.9***
(98.72)
Constant 11,246
(19,384)
Observations 91
Number of state 19
Random effects Yes
Overall R2 0.8649
the largest significance level at which the null hypothesis cannot be rejected. In a
classical hypothesis testing, we take this hypothesis as true and require the data to
provide substantial evidence against it. In our regression model, the null hypoth-
esis is that the coefficients are equal to zero or have no effect on the dependent
variable. If the null hypothesis is rejected, then the variable is a meaningful addi-
tion or statistically significant to our model.
If p < 0.01 (denoted by ***), then the regression coefficients are significant at
the 99% confidence level. If p < 0.05 (denoted by **), then the regression coef-
ficients are significant at the 95% confidence level. If p < 0.1 (denoted by *), then
the regression coefficients are significant at the 90% confidence level.
From the regression results, it is observed that capital investment has a statisti-
cally insignificant but positive impact on the GSDP, where a change of 10 million
INR in capital formation is expected to change GSDP by 11.9 million INR, hold-
ing all else constant. Labour has a significant and positive impact as expected,
with a unit change in the labour variable effecting a 0.4 million INR change in the
GSDP, holding all else constant. Another important finding from the regression
results is that infrastructure indicators such as road, air and teledensity are posi-
tive and significant. For every kilometre expansion of roadways, GSDP increases
by approximately 4.8 million INR. Changes in the average cargo handled per day
by domestic airways shows a change in GSDP of 39.8 million INR. An increase of
1 unit in the teledensity would increase the GSDP by 32.4 million INR.
The constant is the intercept for the regression line and is the default predicted
value of the dependent value if all the other variables are zero. Since the other
variables would never take the zero value in our model, the intercept or con-
stant does not hold much importance here. The R2 value in a multiple regression
model indicates the proportion of the total sample variation in the dependent vari-
able that is explained by the independent variable. The higher the R2 value is,
the higher percentage of points the regression line passes through when the data
points and line are plotted. Our model has a high overall R2 of 0.8649, meaning
that approximately 86% of the data points will fall on the regression line, which
is good for our model.
6 Discussion
6.1 Main findings and implications of regression results

The magnitude of the regression coefficient of the capitalf variable is high, with
the implication that there is still a huge need for investments for infrastructure
projects in India. Realizing this need, since 2004 the Indian government has
been consistently trying to increase the public spending on infrastructure (see
Figure 2.4).
As of 2016, public investment in the infrastructure sector is approximately 8%
of the total Indian GDP. However, this is still not sufficient for India’s growing
needs. According to Finance Minister Mr. Arun Jaitley in 2016, India needs over
$1.5 trillion in investment in the next 10 years to bridge the gap between sup-
ply and demand. Since public capital alone cannot fund the infrastructure needs,
government is increasingly seeking to boost private investments, which currently
account for about 32% of the total infrastructure investment (ASSOCHAM).
Even this 32% share of private investment is highly skewed and mostly concen-
trated in sectors such as telecommunications and airport development. One of the
major reasons why the telecommunications sector has been highly successful in
attracting private investors is due to the higher investment returns, sector-specific
Government share in infrastructure spending (in %)

90
80
70
60
50
40
30
20
10
0
Dec-97
Sep-98
Dec-00
Sep-01
Dec-03
Sep-04
Dec-06
Sep-07
Dec-09
Sep-10
Dec-12
Sep-13
Dec-15
Sep-16
Jun-99
Jun-02
Jun-05
Jun-08
Jun-11
Jun-14
Mar-00
Mar-03
Mar-06
Mar-09
Mar-12
Mar-15
Figure 2.4 Government share in infrastructure spending (in %)

Source: Centre for Monitoring Indian Economy (CMIE).
policies, independent regulatory bodies such as TRAI and transparent policies
facilitating fair competition among private players.
To bridge the gap between the demand and supply of infrastructure invest-
ments, the Indian government is increasingly considering public-private part-
nership (PPP) as an option to attract more capital investment. According to the
Ministry of Finance, Government of India, a PPP project is based on a contract or
concession agreement between the government or a statutory entity on one side
and a private company on the other side and is intended to deliver infrastructure
service with the payment of user charges. The Reserve Bank of India notes that
the PPP format is generally not very popular among private investors due to the
low and delayed profit returns, uncertainty in profits and less efficient institution
and policies. To address these issues, the government needs to review its PPP
policies to ease the process and try to attract more capital investment from the
private sector.
The labour variable has positive and significant impact, as expected. The mag-
nitude of the regression coefficient of labour is high, with a one-person increase
in labour increasing the GSDP by 0.4 million INR. This is approximately 6212
USD (in constant values with base year 2005). The main implication of this is
that human capital will have a huge impact on the Indian GSDP. As a develop-
ing country, India’s investment in human capital, such as improved education
and health facilities, will have a significant and positive impact on the Indian
economy.
Regarding the infrastructure variables, the air and tele variables in particular
have high impact, which is consistent with the higher share of private invest-
ment pouring into these sectors. This increased interest from private investors in
a few sectors alone poses an issue for the overall development of infrastructure
in India because even if the overall percentage of private investment is approxi-
mately 40%, not all sectors receive sufficient investments to meet the demand.
This leaves much responsibility to the government to manage other essential sec-
tors such as roadways.
Thus, the major findings from our regression lead to the policy recommenda-
tion that rising infrastructure investment needs can be met by promoting private
investments across all sectors and that the Indian government must invest more in
human capital as well by giving increased attention to health care and educational
attainment.
6.2 Obstacles in the infrastructure sector

Various infrastructure publications from the government ministries, the Reserve
Bank of India and industries were reviewed for a better understanding of the
obstacles faced by this sector, and policy recommendations are provided to rectify
these obstacles. The four major issues faced are as follows:
1 Gap in infrastructure investment needs

2 Private investments are highly skewed towards the telecommunications sector
3 Under-implementation and project delays
4 Regional imbalance in development.
There is a great difference between the infrastructure needs and the actual amount
of investment being spent on infrastructure, despite the increase in the Indian
government’s infrastructure investments. One way to increase the infrastructure
investments would be to attract more private investment to this sector. While
private investment in the infrastructure sector is considerable, it is mostly con-
centrated in certain subsectors, leading to skewed investments across sectors.
A policy suggestion for the government would be to increase the scope for private
investors across sectors that are currently underinvested in and to promote more
participation through increased incentives.
Another way to increase investment would be to promote PPPs in the infra-
structure sector, but unfortunately use of this concept has not yet reached its full
potential. This is due to issues such as risk sharing, transparency, project apprais-
als, centre-state disagreement and lack of an independent regulatory body that
hinder the implementation of PPPs in the infrastructure sector.
Under-implementation and project delays imply that the outstanding invest-
ments are still in process or not yet completed. Therefore, a high under-implemen-
tation rate is not good, as the actual benefits of a project can be received only after
project completion. In Figure 2.5, we notice that the under-implementation rate
for the infrastructure sector is higher than that for the non-infrastructure sector.
This leads to cost escalation of projects and is a burden on public funds. ASSO-
CHAM reports that transport infrastructure project delays lead to a 47% increase
in project cost. The government must follow better management policies to ensure
the timely delivery of projects to avoid such escalations.
60
55
50
45 Infrastructure
Investment
40 Non-
infrastructure
investment
35
30
2007 2008 2009 2010 2011 2012 2013 2014 2015
Figure 2.5 Under-implementation of infrastructure projects

Source: ASSOCHAM, India.
State-wise share of infrastructure spending (% of total)

14
12
10
8
6
4
2
0
Jammu &.. .
Uttar Pradesh
Kerala
Uttarakhand
Mizoram
Andhra Pradesh
Maharashtra
Gujarat
Tamil Nadu
Odisha
Karnataka
Telangana
Bihar
Chhattisgarh
Rajasthan
West Bengal
Haryana
Jharkhand
Assam
Punjab
NCT of Delhi
Manipur
Meghalaya
Goa
Nagaland
Sikkim
Tripura
Arunachal...
Himachal...
Madhya...
Figure 2.6 State-wise share of infrastructure spending (% of total)

Source: ASSOCHAM, India.
Considering investment trends by location, we notice an imbalance in regional

investment, as most of the investments gravitate towards highly developed areas.
The least developed states, while they require more investments for develop-
ment, are the ones that receive the least investment (see Figure 2.6). This leads
to skewed development trends across the regions. Infrastructure projects must be
allocated, keeping this mind to ensure fairness to all.
6.3 The need for inclusive wealth-based policy analysis

Until now, India’s economic growth has been driven by the poverty reduction
goal. The country is now at a juncture where the growth and development under-
taken must be sustainable as well. This gives rise to the need for a more inclusive
index to evaluate policies in addition to evaluation based on purely economic
goals. The Inclusive Wealth Index (IWI) was created in this context to include
produced, natural and human capital in consideration when evaluating sustainable
development goals.
According to the Inclusive Wealth Report 2014, the IWI is a measure of inter-
temporal social welfare to assess nations’ social progress by taking human, natural
and produced capital into account. Human capital includes the knowledge, skills,
competencies and attributes embodied in individuals who facilitate the creation of
personal, social and economic well-being. Natural capital is everything in nature
(biotic and abiotic) capable of providing human beings with well-being, either
directly or through the production process. Produced capital is traditionally con-
sidered as capital, consisting of roads, buildings, ports, machinery and equipment.
Wealth = Ppc × PC + Phc × HC + Pnc × NC

Inclusive wealth is represented as in the preceding equation, where it is calculated
through the shadow or social price, which acts as a weight. This shows that well-
being, which moves in the same direction as wealth, will be influenced not by just
economic growth, but by all assets including human and natural capital as well.
This study provides a descriptive analysis of how the increased development
of the transport sector would affect different capital assets on the macro level to
highlight the fact that policy evaluations should consider more than just the eco-
nomic benefits. First, the impact of the transport sector on produced capital can
be derived directly from our regression results, as GDP per capita is directly pro-
portional to the produced capital in IWI. As mentioned before, a 1 km expansion
of roadways will increase GSDP by approximately 4.8 million INR. For human
capital, the effect is seen clearly from our regression results as well. A unit change
in the labour variable effects a 0.4 million INR change in the GSDP, holding all
else constant. This shows that human capital has a huge influence on the pro-
duced capital, which increases the inclusive wealth and thereby increases well-
being. Other impacts on human capital include increased access to health services
and education facilities due to improved road connectivity. This would lead to a
direct increase in the life expectancy and educational attainment of the people.
Since health and education are two main parameters for the calculation of human
capital, roadways would have a positive impact on human capital, leading to an
increase in IWI and well-being. Finally, with regard to the impact of roadways on
natural capital, we hypothesize that roadways would have negative impacts such
as reduction in natural resources for construction purposes, degradation in air
quality, noise pollution and possible loss of forest cover. Thus, roadways do not
only influence the economy but have other varied effects as well, reiterating the
need for a more inclusive policy analysis.
7 Conclusion
To summarize, this study presents a quantitative analysis of the impact of infra-
structure development on the Indian economy. We find that the transport sec-
tor has the biggest impact and the telecommunications sector also has significant
impacts on GSDP. It was also observed that human capital is a very important fac-
tor of input in India, where investments to improve the labour force would result
in increased wealth and well-being.
The issues that plague the infrastructure sector and slow down its progress
have been reviewed. Issues such as an investment gap, private investment skewed
towards certain sectors, project delays and regional imbalance in development
still exist.
This study promotes the use of inclusive wealth-based policy analysis for infra-
structure projects. Descriptive analysis was conducted to study the impacts of
roadways on different capital assets, namely, produced, human and natural capital.
One impact is the influence of roadways on social welfare through increased con-
nectivity, which increases access to health services and education facilities. These
impacts affect human capital and well-being in a positive way. Thus, in addition
to economic impacts, a particular project would have a multitude of impacts on
natural and human capital. This gives rise to the need to evaluate infrastructure
policies and projects in a more inclusive manner.
Bibliography
Arrow, K. and Kurz, M., 1970. Optimal growth with irreversible investment in a Ramsey
Model. The Econometric Society, Vol. 38, No. 2, 331–344.
Aschauer, D. A., 1989. Is public expenditure productive? Journal of Monetary Economics,
Vol. 23, No. 2, 177–200.
The Associated Chambers of Commerce and Industry of India (ASSOCHAM, India)
(2016). Analysis of Infrastructure Investment in India [online]. Available at: https://
smartnet.niua.org/sites/default/files/resources/analysis-of-infrastructure-investment-in-
india-assocham-may-2016.pdf
Barro, R. J., 1990. Government spending in a simple model of endogenous growth. Journal
of Political Economy, Vol. 98, No. 5, Part 2.
Canning, D. and Pedroni, P., 2004. The effect of infrastructure on long-run economic
growth (mimeo). Cambridge, MA: Harvard University Press.
Chakraborty, C. and Nandi, B., 2011. ‘Mainline’ telecommunications infrastructure, levels
of development and economic growth: Evidence from a panel of developing countries.
Telecommunication Policy, Vol. 35, No. 5, 441–449. https://doi.org/10.1016/j.telpol.
2011.03.004.
Collins, R., etc., 2016. Using inclusive wealth for policy evaluation: Application to electric-
ity infrastructure planning in oil-exporting countries. Faculty Research Working Paper
Series, Harvard Kennedy School.
Cook, C., 2005. Assessing the impact of transport and energy infrastructure on poverty
reduction. Asian Development Bank Publication.
Estache, A., 2004. Emerging infrastructure policy issues in developing countries: A survey
of the recent economic literature. Background Paper for the October 2004, Berlin Meet-
ing of the POVNET Infrastructure Working Group (Processed). Washington, DC: The
World Bank.
Estache, A., 2006. Infrastructure: A survey of recent and upcoming issues. Washington,
DC: The World Bank.
Estache, A. and Garsous, G., 2011. Infrastructure finance in developing countries: An over-
view. Working Paper, European Investment Bank.
Futagami, K., Morita, Y., and Shibata, A., 1993. Dynamic analysis of an endogenous
growth model with public capital, Scandinavian Journal of Economics, Vol. 95,
607–625.
Garcia-Mil’a, T., McGuire, T. J., and Porter, R. H., 1996. The effect of public capital in
state-level production functions reconsidered. Review of Economics and Statistics,
Vol. 78, 177–180.
Garsous, G., 2012. How productive is infrastructure? A quantitative survey, ECARES
Working Paper, Université libre de Bruxelles.
Gramlich, E., 1994. Infrastructure investment: A review essay. Journal of Economic Litera-
ture, Vol. 32, No. 3, 1176–1196. Retrieved from http://www.jstor.org/stable/2728606
Hindustan Times, 2017. Government constructs 22 kms of roads per day (online), April.
IIR, 2006. India Infrastructure Report, 2006. India: 3i Network.
Inclusive Wealth Report, 2014. Measuring progress toward sustainability. Cambridge:
Cambridge University Press.
Jones, S., 2004. Contribution of infrastructure to growth and poverty reduction in East Asia
and the Pacific. Background Paper. Oxford, UK: Oxford Policy Management.
OECD Directorate, 2018. OECD glossary of statistical terms – infrastructure definition.
[online] Stats.oecd.org. Available at: https://stats.oecd.org/glossary/detail.asp?ID=4511
Sahoo, P. and Dash, R. K., 2009. Infrastructure development and economic growth in India.
Journal of the Asia Pacific Economy, Vol. 14, No. 4, November, 351–365.
Sridhar, K. S. and Sridhar, V., 2007. Telecommunications infrastructure and economic
growth: Evidence from developing countries. Applied Econometrics and International
Development, Vol. 7, No. 2, 37–56.
3 Sustainability evaluation
for energy infrastructure
A hybrid simulation approach
to inclusive wealth
Ebrahim Aly and Shunsuke Managi
1 Introduction
Successful infrastructure projects are a major component of a nation’s develop-
ment. Energy projects, in particular, are necessary for development in both devel-
oped and developing countries. Energy is, as described by E. F. Schumacher, “not
just another commodity, but the precondition of all commodities” (Brown and
Sovacool, 2007). However, important questions arise when policymakers plan
and evaluate the infrastructure projects. What is the real cost of these projects?
Do these projects conform to sustainable development? To answer these ques-
tions, such projects need to be evaluated against a comprehensive sustainability
measurement. In this chapter, we apply the Inclusive Wealth Index as a measure of
intergenerational well-being, along with a hybrid simulation technique for energy
policy evaluation and planning.
Inclusive wealth (IW) is one of the contemporary wealth accounting methods,
and it can be defined as the sum value of any society’s capital assets, namely,
human, natural, and produced capital, valued at social prices (or shadow prices;
UNU-IHDP and UNEP, 2014). The IW framework overcomes the “mis-measure-
ment” of development that occurs using conventional measures such as gross
domestic product (GDP; Stiglitz et al., 2010). If (ki (t )) is the stock of asset i at
time t, ( K (t )) is the vector of the stocks of capital assets, and ( Pi (t )) is the shadow
price of asset i at time t, then wealth at time (t) is the linear summation of each
asset stock at that time (t), multiplied by its associated shadow price at the same
time (UNU-IHDP and UNEP, 2012).
Prior studies that evaluated energy projects’ sustainability, such as Holley et al.
(2017), Atilgan et al. (2016) and Kveselis et al. (2017), focused on sustainabil-
ity assessments of developing and governing energy projects. Other researchers
focused on evaluating scenarios of shifting from one energy form to another (e.g.,
Naito et al. 2009). However, these efforts were either country specific, technology
specific, or both.
On the other hand, few attempts have been made to either predict the trajec-
tory of wealth or evaluate policies in a forward-looking manner based on the
inclusive wealth. Tokimatsu et al. (2011, 2014) investigated the future dynamics
of wealth under different scenarios such as population or technological changes.
44 Ebrahim Aly and Shunsuke Managi
Collins et al. (2014) and Pearson et al. (2013) introduced demonstrative mod-
els for policy evaluation using the inclusive wealth theory. While Collins et al.
(2017) applied the inclusive wealth theory to a prospective policy evaluation,
their work focused on the impacts of shifting to a non-fossil electricity policy on
produced human capital and oil reserves in oil‑exporting countries.
In this chapter, we provide a model for evaluating the impact of energy infra-
structure projects on the trajectory of wealth. The model applies hybrid simulation
and data analysis to quantitatively measure how a prospective energy generation
project, or a mix of projects, will affect the different capital assets that compose
the wealth base of the society under study. This enables us to exploit the edges
of different simulation techniques, particularly system dynamics and agent-based
modelling, to perform a more in-depth analysis with the flexibility to study sev-
eral cases and combinations and even variations within the same fuel cycle. The
model we present herein considers a more comprehensive scope of wealth, as
we included impacts on cropland along with extractable fuel depletion in natural
capital calculations and impacts on health capital and employment in human capi-
tal calculations. Moreover, it addresses air pollution and CO2 damage resulting
from different fuel cycles. One more important characteristic of our approach is
that it is built in a generalized manner that makes it suitable for analyzing projects
in developed and developing economies. The developed model enables policy
evaluators, whether they are government officials, development agencies, or deci-
sion-makers, to choose the most sustainable energy option available. Monitoring
the future changes in the capital assets after introducing a project to the economy
not only allows for policymakers to study the effects of projects from different
categorical fuel cycles, such as renewable and non-renewable, but also allows for
them to quantify the impacts of projects from the same category (e.g., wind and
photovoltaic).
The remainder of this chapter is organized as follows. In Section 2, we present
the methodology that we followed to simulate and evaluate projects. In Section 3,
we describe the model formulation, what its constituting parts are, and how are
they connected. Section 4 presents the model application and case studies. Finally,
the conclusion is presented in Section 5.
2 Methodology
Evaluating projects based on wealth is equivalent to carrying out a social cost-
benefit analysis for that project, using the same set of shadow prices used in
sustainability analysis (Dasgupta, 2009). The project under study represents a
perturbation in the wealth course (Arrow et al., 2003) which is initiated at time (t p ).
If we think about (Wo ) as the original course wealth shall follow had the project
not been introduced and (Wn ) as the new wealth course after project introduction,
then we can say that:
Wn = Wu + ∆W p (3.1)
Sustainability evaluation for infrastructure 45
Project initiation
Wealth
Wealth
Wealth projection (Wu)
Wn
tp
Time
Figure 3.1 The wealth trajectory after introducing a new project to the economy
The first component of Equation 3.1 (Wu ), is the part of the wealth change that
is unexplained by this project; this part changes according to the dynamics of
the economy that are not related to the project. On the other hand, (∆W p ) is the
change in wealth due to the impact of the project on the different capital assets
(Figure 3.1).
We calculate the change in wealth that resulted from the project introduction in
the same way we would calculate wealth using Equation 3.2:
∆W p (t ) =∑ Pi (t )∆K i (t ) (3.2)
where (∆K i ) is the change in each capital asset caused by the project. The capital
assets we are considering in this chapter are human capital (HC), natural capital
(NC), and produced capital (PC); thus, the vector of capital assets will be as is
shown in Equation 3.3:
∆K = {∆HC , ∆NC , ∆PC } (3.3)
As previously explained, ( Pi ) should be the shadow price corresponding to the

capital asset ( K i ). The shadow price of a capital asset is defined in the 2012 inclu-
sive wealth report (UNU-IHDP, and UNEP (2012) as “the contribution a mar-
ginal unit of it is forecast to make to human well-being” (UNU-IHDP and UNEP,
2014). This definition shows how shadow prices can be effective for valuing capi-
tal assets; yet, it makes estimating them problematic. Thus, instead, in the context
of this chapter we use market prices as an approximation. Although not using
shadow prices represents a limitation in applying the inclusive wealth theory, such
an approximation has been used in different studies. IWR 2012 relied on rental
prices whenever there was a lack in shadow prices, while Collins (2013) used the
market prices of natural resources.
Quantifying the impact of a single project or even a mix of projects on a nation’s
capital base imposes a key challenge, as a sustainability analysis that is based on
the inclusive wealth framework considers projects that are small relative to the size
of the economy (Dasgupta, 2009). In addition, although projects have their own
level of uncertainty in some aspects such as delivery time or project cost, they are
still considered deterministic compared to economies at large. Additionally, energy
generation technologies that projects apply are rapidly changing, and advance-
ments in energy efficiency and fuel substitution are always being introduced.
A robust model needs to perform well on both ends: energy technology and
economy. It needs to be flexible enough to allow for evaluating different fuel
cycles or different technologies within the same fuel cycle. It should also be able
to represent macroeconomic performance and feedbacks. Conventionally, mod-
ellers use two different modelling approaches, and each one is more suitable for
one of the mentioned ends, i.e., project scale and microeconomic scale. These two
approaches are referred to as (1) bottom-up (BU) and (2) top-down (TD) models.
TD models are effective in representing the interaction between economy com-
ponents and the accumulation of wealth. They allow for modellers to analyze
dynamic systems and test their performance under certain policy changes. BU
models, in contrast, can give a detailed description of energy technologies and
the phases of projects, but they fall short in regard to describing macroeconomic
features, such as economic growth or the composition of the capital base (Hour-
cade et al., 2006).
We use a hybrid simulation approach to compensate for the limitations in each
approach and to build a model that is reliable and robust to evaluate whether tech-
nology choices, represented by projects, are following a sustainable path. System
dynamics (SD) would be an appealing technique for modelling the system of capi-
tal stocks as it focuses on exploring the links between the system components and
how they develop over time according to the structure of the system (Kwakkel
and Pruyt, 2015). Several researchers adopted SD approaches for policy analysis
and decision support (e.g., Fiddaman, 2002; Benaich and Pruyt, 2015; Uehara
et al., 2016), and some of these efforts were dedicated to wealth‑based policy
analysis (e.g., Collins et al., 2014, 2017).
In regard to modelling the technology or the project(s), we chose to use agent-
based modelling (ABM). The agent(s) in our model are the project(s) under evalu-
ation. Each agent’s behaviour is determined by a set of rules as well as a set of
internal states. As the agent goes from one state to another, its attributes change
accordingly. Using ABM enables evaluators to examine different scenarios by
“plugging in” different agents to the framework and monitor the evolution of the
system under each scenario. Another advantage of using ABM is that it allows for
evaluators to capture one of the important concepts associated with infrastructure
capital, inertia, defined as the lag between an investment decision and its change
to the system (UNU-IHDP and UNEP, 2014).
Incorporating the two simulation techniques (SD and ABM) into one model
enables us to study how relatively small economic changes (infrastructure projects
in our case) affect the bigger wealth components or the inclusive wealth as a whole.
3 Framework description
The framework we introduce in this chapter tries to capture the impacts imposed
by different energy production projects on produced, natural, and human capital
(PC, NC, and HC) as the main components of a society’s capital base. As Fig-
ure 3.2 shows, the framework relies on three main categories of data used to build
a database that feeds the hybrid simulation model. Then, the simulation model
calculates the development of the capital assets over a certain period. The output
is the change in each capital asset along with the change in total wealth caused by
this project. Another important output is an index we developed to make it easy
for evaluators to compare between the different projects in terms of how much
energy they produce. As different types of technologies or “fuel cycles” differ in
energy production, in order to compare between projects representing different
fuel cycles, we developed the so-called energy supply index (ESI).
3.1 The energy supply index

The energy supply index can be simply described as how much of the energy
demand the project under study covers at a certain point in time. It can be calcu-
lated by Equation 3.4:
jt
ESI = ×100
Et (3.4)
Macro data
retrieval Data cleaning
and processing
1
0 0 01 0 0 1
0 100
01 1 0 010
0
1
Hybrid simulation
Meteorological model
& pollutants specific
data
Model
database
model output
Project and
Fuel Cycle
data
Figure 3.2 Schematic diagram for model components and the process flow
where ( jt ) is the project electricity output at any time (t) and ( Et ) is the energy
demand at any time (t ). To calculate the energy demand ( Et ), we use a linear
regression approach that links energy consumption to economic and demographic
variables. Such an approach has been used by different researchers as an attempt
to forecast future energy demand (e.g., Bianco et al. 2009; Mohamed and Bodger,
2005). The linear regression model we used is a log-log model, where energy
demand is the dependent variable, while population ( Pt ) and (GDPt ) are the inde-
pendent variables (Equation 3.5):
log ( Et ) = β0 + β1 log ( Pt ) + β2 log (GDPt ) + εt (3.5)
3.2 The hybrid simulation model

As previously described, the simulation model is composed of two parts, each
of which uses a different simulation technique to model both the technology and
the economy. We treat the project under evaluation as an exogenous event that is
connected to the system through certain links. Each project has its own charac-
teristics, which are represented by a vector of characteristics ai . It also moves
through different pre-defined states along its lifetime. At each state, the system
receives different feedbacks from the project through the links connecting them;
hence, the system changes according to the feedback it receives. The project’s
vector of characteristics is composed of (Equation 3.6):
a = fc, condt , opdt , lf , fu , j , ems , Inv (3.6)
As Table 3.1 shows, the characteristics of the project contains the project’s opera-
tional data, environmental data, and economic data. In addition to defining the
amount of cash invested in the project, we also define the financial type of this
investment. By the type of the investment, we mean if the project is fully or par-
tially funded by international loans. The payback of the project loans represents
a burden on the economy, which will negatively affect its growth. To account
for the project loans, we use the capital recovery factor (CRF) method to evenly
distribute the loan costs over the project lifetime (n), as shown in Equation 3.7:
n
i (1 + i )
CRF = n (3.7)
(1 + i ) −1
3.2.1 Modelling the project under evaluation (ABM)

Any project, regardless of the fuel cycle it adopts, passes through five different
states along its lifetime (Figure 3.3). The first state is when the project is merely an
idea under study, or a policy option; we call this state the “pre-construction” state.
During this state, the project has no impact on the system. This state is particularly
important for capturing the aforementioned concept of inertia, as it represents the
Table 3.1 Description of the project characteristics
fc The type of the fuel used in the project or which fuel cycle the project follows.
Condt The construction start date of the project.
opdt The operation start date of the project.
lf The project lifetime.
fu The fuel consumption of this project; this particularly useful in the case of non-
renewable fuel projects.
j The electricity output of the project (kWh).
ems The vector of emissions the project produce in each of its phases or states to both
air and water.
Inv The amount of cash invested in this project, and the dates of the inflows of the
invested cash.
Project state
State triggers
? Condition triggered state
Timeout triggered state
Message triggered state
Pre-construction
Construction
Operation Idle
End of service
Figure 3.3 The state of projects under evaluation and the triggers control transition
between different states
lag between when the decision is made and when the execution starts and, conse-
quently, the impacts start to occur. The second state is the construction state, and
as the project enters this state, its impacts on the system start to appear. As well as
starting the financial impacts, the emissions during construction are emitted to the
atmosphere. Additionally, the construction employment effect and its implications
on the human capital occur, as will be explained in more detail in the description
of the system part of the model. One of the obvious attributes of the construction
state is that it is relatively short compared to the other states of the project. After the
construction state, the project moves to the main and longest state of its lifetime, the
operation state. During this state, in addition to the financial impacts and the opera-
tion emissions, the withdrawal from the fuel reserves in the case of non-renewable
energy projects begins. The idle state represents any temporary stoppage of the proj-
ect due to unforeseen circumstances, such as natural disasters or major rehabilita-
tion. The final state marks the end of the project lifetime.1
The transition between the different states is initiated by one of the three “state
triggers” we defined for the model. The first trigger is the “condition” state trigger,
which means that the project will move to the next state when a certain condition
is fulfilled. This trigger is used to move from the pre-construction to the construc-
tion states, and the condition is that the calendar date is equal to or greater than
the construction start date.
The second trigger is the timeout state trigger, which makes the project move
from its current state to the following state after a certain duration of time. This
trigger is used for the transition between the construction and operation states and
the operation and end‑of‑service states.
The last trigger type is the message state trigger. The transition, in this case, is
triggered by an exogenous message that the project, or the agent, receives. The
message trigger is used to move the project from the operation to the idle state and
back to the operation state.
3.2.2 Modelling the capital assets (SD)

In this part of the hybrid simulation model, we identify the main components of
the system and how they interact and affect each other. We also identify the links
between the agent (the project under evaluation) and the system. These links are
the channels through which the project affects the system and gets feedback from
it. The project’s effect on a nation’s wealth will appear through its effect on the
three capital assets composing the wealth, as shown in Figure 3.4. It is worth not-
ing that the sign at the end of the arrows linking the elements shows how these
elements change relative to each other. A (+) sign means that the two elements
move in the same direction. An increase or decrease at the start element leads to
an increase or decrease in the end element, whilst a (−) sign means that the two
elements move in opposite directions or that they are inversely related.
The project, directly and indirectly, affects the capital assets through its charac-
teristics, which change according to the project state. A project, as an infrastructure
Mixing height Deposition

speed Distance
CO2 damages from plant
– – –
+ Wind speed
– –
Pollutants
+ Fishiries concentration
Emissions – –
+ Forests – Health
+
Energy Fuel capital
Pasture
Project consumption – Fuel land
–
reserves
+ Crops yield
Electricity output –
– Electricity
demand + + + +
+
+
+ Educational Natural Capital
Employment attainment
+ Human Capital +
+
+ Produced Capital
O&M costs – + – –
Investments
Loans payback
Depreciation
Figure 3.4 Dynamics of the interaction between an energy project and the different capital
assets
asset, is a direct addition to the produced capital, and all the investments in this
project along its lifetime – whether they are made to increase its service life or
to increase its capacity – are also additions to the produced capital. On the other
hand, the project depreciation and operation costs can be considered as deductions
from the produced capital. Another negative impact of the project on the produced
capital comes in the form of the loan’s payback in case it was funded by borrowed
investments.
The project impacts on employment appear in the additional labour hours of
those who work in the sector (construction and energy sectors) and the additional
labour hours produced with the help of this asset or the energy produced by it.
These impacts on employment lead to an increase in the human capital. Another
impact of the project on human capital comes from its contribution in increasing
educational attainment by increasing total energy production.
The additional energy that the project adds to the system secures a share of
the demanded energy, which is a major factor affecting withdrawal from the fuel
reserves. These fuel reserves are also affected by how much fuel the project con-
sumes, and the impacts on fuel reserves are directly reflected on the natural capital.
Impacts on natural capital also arise from the emissions of pollutants associ-
ated with this particular project during the construction and operation phases.
Emissions to water affect the quality of fisheries and, hence, negatively affect the
natural capital. Likewise, emissions to air increase the concentration of pollutants
in the atmosphere, causing negative impacts to cropland, forests, and pastureland –
three basic components of natural capital. The concentration of the pollutants in
the atmosphere also depends on other exogenous meteorological parameters, such
as the mixing height of the pollutants, the deposition speed of each pollutant, the
distance from the plant, and the average wind speed.
Finally, the increase in pollutant concentration in the atmosphere with the nega-
tive impacts that have on the health capital causes a decrease in the human capital.
In the following section, we will show how we quantify the impacts of projects
on the different capital assets within the developed framework.
3.3 Q
uantification of energy infrastructure projects
impacts on the capital assets
To evaluate an energy infrastructure project or compare between different energy
projects to choose the one(s) that conforms to sustainable development, we need
to quantify the impacts it has on the wealth and its composing capital assets. In
this section, we demonstrate the approaches we used to quantify project impacts.
Generally, each of the capital stocks (S) increases by amount of the inflows ( f i )
that are added to it and decreases by the amount of the outflows ( f o ) that escape it
over the period from the project introduction (t p ) until the end of project lifetime
(t p + lf ) (Equation 3.8):
(t p +lf )
∆S = ∫ ( f i − f o ) dt (3.8)
tp
3.3.1 Produced capital

As illustrated in Figure 3.4, the project under study is directly linked to produced cap-
ital. We can quantify the project impact on the PC stock ( K ) through Equation 3.9:
∆PC = Inv − Dep − loanPbk − OM (3.9)
where ( Inv ) is the initial and periodical investments in the project. This is an exog-
enous parameter and is provided as an element of the project characteristics. ( Dep )
is the project depreciation, and it is calculated using the same depreciation rate used
in the IWR 2014 (UNU-IHDP and UNEP, 2014), which is assumed to be 4%. The
loan payback (loanPbk ) is calculated using the (CRF), as shown in Equation 3.7. To
calculate the fixed and variable operation and maintenance costs (OM ) for the differ-
ent fuel cycles and technologies, we used the data provided by Tidball et al. (2010).
3.3.2 Pollutant concentrations and project carbon damages

Before proceeding to the approaches used to quantify human and natural capital,
we had better first present how pollutant concentrations in the atmosphere were
calculated, as it is one of the main requirements for measuring how emissions
from the project affect the two capital assets.
The pollutants we considered within the scope of this chapter were categorized
into two categories: (1) emissions to air and (2) emissions to water. The constitu-
ents of each category are shown in Table 3.2. The concentration of air pollutants
was calculated using a simplified version of the Gaussian plume model used in
the ExternE report (Bickel and Friedrich, n.d. 2005). Neglecting the escape of
pollutants into the troposphere and assuming a homogeneous distribution of wind
direction, if the average annual wind speed is (u [ m / s ]) , and the mixing layer
height is ( H [ m ]), the concentration of a pollutant (i ) at distance (d [ m ]) from the
plant can be calculated using Equation 3.10:
d v
Q 1 − ⋅ i
Ci = ⋅ ⋅e H u (3.10)
u H ⋅ 2π r
where: Q is the emission factor [ g / s ]

vi is the dry deposition velocity [ m / s ]
Projected carbon damage is estimated using the same criteria used in IWR
2014. First, CO2 emissions to air are estimated and then multiplied by a damage
of $50 per ton to obtain the project’s carbon damage.
3.3.3 Human capital

The IW framework states that there are three main terms driving HC: edu-
cational attainment, country population structure, and earnings per unit of
human capital. Extending that notion, we measured the project’s impact on
HC by measuring its effects on employment, education, and health capital
(Equation 3.11):
∆HC = ∆Emp + ∆Edu − ∆Health (3.11)
Table 3.2 Air and water pollutants
Emissions to air Emissions to water
CO2 Na
CO Cl
NOx SO4
SO2 Ca
O3 Mg
PM10 K
Employment effect (∆Emp ) can be used as the additional labour hours that resulted
from this project embodied in those who work in the relating sectors, and as we are
evaluating energy production projects, we considered only construction and energy
sectors. To calculate the project’s employment effect, we multiplied the man-hours
associated with each project state (mainly construction and operation) (∆Lstt ) with
the hourly employee compensation for the sector associated with each state (rt )
sc
(Equation 3.12). The hourly employee compensation is calculated using earnings

by economic activity (ILO, 2015) and hours worked (Conference Board, 2016):
∆Emp = ∆Lstt ⋅ rt sc (3.12)
Trying to link electricity production and educational attainment is a contro-

versial approach. Although a pure economist would resist such an idea, some
researchers made attempts to establish a correlation between electricity produc-
tion and educational attainment (Leung et al., 2005; Kanagawa and Nakata,
2008). Admitting that it is not an ideal way to estimate the impacts of electricity
on educational attainment and lacking an alternative, we would go on with that
approximation. As such, we assume a level-log relation between mean years of
schooling index (MYSI) and electricity consumption ( En), as in Equation 3.13:
∆MYSI = β0 + β1 log (∆En) + ε (3.13)
The mean years of schooling index is the proxy for educational attainment, and
it is the mean years of schooling divided by 15. Educational attainment data is
acquired from Barro and Lee’s (2013) dataset, while electricity consumption is
from the World Bank (2016). The HC driven by educational attainment is calcu-
t p +∆T
Edut = e(
At ⋅ρ )
⋅ P15+ ⋅ ∫ r ⋅ e−δt dt (3.14)
tp
In Equation 3.14, ( At ) is the educational attainment or mean years of school-

ing, (ρ ) is the rate of return on education and is set equal to 8.5%, and ( P15+ ) is
an approximation for the adult population and is obtained from the World Bank
(2016). The third term represents the shadow price for human capital, and it is the
present value of the average labour wage received over the remaining working
period. (∆T ) is assumed to equal to 20 years (Yamaguchi, 2014), and the discount
rate (δ ) equals 8.5%. The average labour wage (r ) is calculated as a percentage of
total GDP and is obtained from ILO (2015).
The health term in Equation 3.11 indicates the project impacts on the health
capital, an essential component of the HC. Different energy production cycles
cause an increase in the concentrations of certain pollutants that have adverse
health impacts. Such hazards are represented as either an increase in mortality
rates or a decrease in productivity represented as restricted activity days (RAD).
For the scope of this research, we are considering the change in mortalities and
RAD for the adult working population. Although accounting for the change in
mortality rates means putting a value for human life, which is a controversial sub-
ject, we chose to use the Value of a Statistical Life (VSL) as a valuation method
nevertheless. VSL is the value that people attach to the additional years of life,
and we calculated it following the same methodology used by Arrow et al. (2012,
2013). Thus, the VSL for any country will be approximately proportional to the
US VSL at the 0.6 power of the ratio of the country’s per capita GDP to the US
per capita GDP. For the US VSL, we used the EPA 2006 estimate of $7.4 million.
Consequently, the health effects can be calculated from Equation 3.15:
∆Health = (∆Mortality ⋅ VSL) + (∆RAD ⋅ r ) (3.15)
where (r ) is the average employee compensation.
Quantifying the percentage change in mortalities (∆Mortalities) and the change

in RAD is done using a dose-response function similar to the one used in the
ExternE project. The used dose-response function estimates how the change in
PM 10 concentration affects both mortality rates and RAD, as shown in Equa-
tions 3.16 and 3.17:
∆Mortalities = 0.104 ⋅ ∆PM 10 (3.16)

∆RAD = 49.9 ⋅ ∆PM 10 (3.17)
1000 adult
3.3.4 Natural capital

In the scope of this chapter, detecting the changes in NC due to the project intro-
duction to the system is carried out by measuring the project effects on extract-
able natural resources (i.e., fossil fuels and minerals) alongside its effects on
cropland. A more thorough research would include the effects on pastureland,
forests, and fisheries; yet, the same procedure used in calculating the effects
on cropland can be applied to those components of NC. The project effect on
extractable stocks is how much it consumes from a certain extractable (extti )
multiplied by the market price of that extractable ( pt ) (USD per ton). Moreover,
i
the project effect on cropland is measured through the effect of the atmospheric
pollutants induced by it on the yield of the different crops produced by the coun-
try hosting the project. Hence, the effect on cropland at any time, (t), is calcu-
n
CRP = ∑ ∆Yt k ⋅ ptk ⋅ Atk (3.18)
k =1
The first term in Equation 3.18, (∆Yt k ), is the change in yield of crop k (ton/ha);
that is multiplied by the price of that crop ( pt ) and the harvested area of that
k
crop ( At )(ha). Yield, harvested area, and crop price data are retrieved from FAO
k
(2014).
The loss in crop yield due to exposure to pollutants is approximated using the
same ExternE methodology, as shown in Equation 3.19.
% yield loss =−3.18 − 0.188(∆SO2 concentration in ppb) (3.19)
4 Model validation and case studies

The proposed framework was applied on different energy infrastructure projects from
two countries: Belgium and Egypt. These two countries were chosen in particular for
different reasons: first, different economic group representation. Belgium, being a
member of the Organisation for Economic Co-Operation and Development (OECD)
and one of the top-ranking countries in the Human Development Index (HDI), repre-
sents developed economies, while Egypt, with an economy that can be described as an
“imperfect economy”, belongs to the developing countries group. In addition to that,
different prospective projects are in the pipeline for both of the countries. Additionally,
both of the countries have very little subsoil resources. Moreover, although triggered by
different motivations, as will be explained later, the two countries seek to change their
energy mix. In Section 4.1, we will discuss model parametrization and calibration, and
then the details of the case studies will be discussed in Section 4.2. The analysis was
run until the year 2050, and for each project, we produced results with the “arguable”
impact of electricity on educational attainment denoted [ed] and without this impact.
4.1 Model parameters
4.1.1 The need for forecasting

Any attempt to evaluate the future values of the different capital assets will neces-
sarily include forecasting the economic aspects constituting them. Future human
capital evaluation, for instance, requires knowing the projections of population,
educational attainment, labour compensation, GDP, and so on. In addition, as some
of these projections might be available through reliable sources such as the World
Bank GDP and population projections, forecasting other aspects becomes inevitable.
Predicting the future values of each of the aspects may be done in several ways,
and the advances in computational abilities facilitate the introduction of more
powerful prediction techniques. Within the scope of this chapter, predictions were
done by a common technique in time-series analysis, which are ARMA or ARIMA
models (based on which gives lower AIC). Although we are not claiming that such
technique is the most suitable for predicting each of the predicted aspects, it was
able to produce reasonable mean absolute percentage errors (MAPE). Table 3.3
presents some of the predicted time series and the maximum corresponding MAPE.
4.1.2 Fuel cycle and emissions

To quantify the emissions associated with power plants falling under different
fuel cycles, we use a simplified life cycle assessment (LCA) approach. While
a conventional LCA will be divided into three phases⸺upstream, direct, and
Table 3.3 Predicted time series and max corresponding MAPE
Predicted TS MAPE
Electricity demand 0.09%
Working hours 0.05%
Crop prices 6%
Hourly wage 2%
downstream⸺the upstream contains all the stages before plant operation, includ-
ing fuel exploration and transportation. In our model, we considered construction
only from the upstream phase along with the direct (operation) phase.
4.2 Case studies

The case studies were selected based on a multi-level diversification criterion.
First, the countries hosting the projects were selected to represent different eco-
nomic groups and different energy needs. The projects themselves belong to dif-
ferent categories and align with each country’s energy needs in a certain way.
4.2.1 Projects from Belgium

The Belgian energy mix shows that nuclear power is the dominant energy source
in that market. Nuclear energy accounts for 47% of Belgium’s electricity genera-
tion, followed by natural gas 27%, biofuels 7.9%, wind 6.5%, coal 6.2%, solar 4%,
and approximately 1% for oil and hydro. Belgium nuclear power is generated from
seven reactors commissioned during the period 1975–1985; thus, it has a relatively
homogenous lifetime of 35 years as of 2015. Starting from 2012, several indepen-
dent events have led to a series of unplanned shutdowns affecting the efficiency of
the power generation network and raising safety concerns from neighbouring coun-
tries. Despite such events and concerns, and under electricity security challenges,
the Belgian government extended the operation time of its three oldest nuclear
power plants from 2015 to 2025. Although that operation time extension helped to
avoid an electricity shortage, the capacity adequacy remains an issue as all nuclear
power plants will be shut down by 2025 (IEA, 2016). To face the upcoming electric-
ity security problem, the Belgian government will have to explore new and more
efficient energy options. Two proposed energy projects were chosen for evaluation
so that we could evaluate the impacts of moving to new energy policies.
The two proposed projects are a 352 MW offshore wind farm and a 215 MW
biomass power plant. Table 3.4 summarizes the projects’ data.
Both projects showed a positive impact on HCed and HC with biomass power
having a higher impact. However, for the same project, the impact of consider-
ing educational attainment was almost fourfold higher than the results without
considering education. The two projects showed a similar pattern in PCed dete-
rioration and negative impacts on NCed, with the wind energy project having less
severe impacts. The same can be said for PC and NC.
Table 3.4 Belgium projects data
Project Project Capacity Construction dates Investment Emissions Lifetime

(fc) code (j) (Inv) data (lf )
Start End sources
(Condt ) (opdt )
Norther wind farm BeW 352 MW 3/1/2018 7/1/2019 Total $1.3 billion, of which [1], [2], [3] 30 years
$515 million financed by EIB
BEE biomass power plant BeB 215 MW 6/1/2017 1/1/2019 $358 million [1], [2], [4] 45 years
Emissions sources: [1] Weisser (2007), [2] Turconi et al. (2013), [3] Schleisner (2000), [4] Spath and Mann (1997).
Impacts on Capital assets considering edu. attainment (Belgium/Biomass)
120
HC
100
80
60
40
$2010(bn)
20
0
PC
0.002
0.000
$2010(bn)
–0.002
–0.004
0
NC
–20
–40
–60
$2010(bn)
–80
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.5 Impacts on capital assets for BEE biomass power plant
Impacts on Capital assets considering edu. attainment (Belgium/Wind)
100 HC
80
60
40
$2010(bn)
20
0
0.010 PC
0.005
0.000
$2010(bn)
–0.005
–0.010
0
–10 NC
–20
–30
–40
$2010(bn)
–50
–60
–70
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.6 Impacts on capital assets for the Norther wind farm
Impacts on Capital assets NOT considering edu. attainment (Belgium/Biomass)
0.0005 HC
0.0004
0.0003
0.0002
$2010(bn)
0.0001
0.0000
PC
0.002
0.000
$2010(bn)
–0.002
–0.004
0
NC
–20
–40
–60
$2010(bn)
–80
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.7 Impacts on capital assets for BEE biomass power plant (not considering education effect)
Impacts on Capital assets NOT considering edu. attainment (Belgium/Wind)
0.00035
HC
0.00030
0.00025
0.00020
0.00015
$2010(bn)
0.00010
0.00005
0.00000
0.010 PC
0.005
0.000
$2010(bn)
–0.005
–0.010
0
–10 NC
–20
–30
–40
$2010(bn)
–50
–60
–70
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.8 Impacts on capital assets for the Norther wind farm (not considering education effect)
The CO2 damages are higher in the case of the biomass projects, as well as the
ESI.
Finally, Figure 3.11 shows the wealth change both projects can cause with and
without the education impact. The results show that considering the education
impact can alter the impact on wealth from a negative to a positive impact. They
also show that the wind power project has better impacts on the course of wealth.
CO2 damages
2.5
2.0
$2010, millions
1.5
1.0
0.5
Belgium/Biomass
Belgium/Wind
0.0
15
20
25
30
35
40
45
50
20
20
20
20
20
20
20
20
Year
Figure 3.9 CO2 damages from Belgian proposed power projects
Energy Supply Index
Belgium/Biomass
1.75
Belgium/Wind
1.50
% of total demand
1.25
1.00
0.75
0.50
0.25
0.00
15
20
25
30
35
40
45
50
20
20
20
20
20
20
20
20
Year
Figure 3.10 Energy supply index for Belgian proposed power projects
Expected project impacts on Wealth

Considering impacts on edu. attainment
1000
Belgium/Wind
800 Belgium/Biomass
$2010(bn)
600
400
200
0
Not considering impacts on edu. attainment
0
–200
$2010(bn)
–400
–600
Belgium/Wind
–800 Belgium/Biomass
15
20
25
30
35
40
45
50
20
20
20
20
20
20
20
20
Year
Figure 3.11 The change in wealth caused by the proposed Belgian energy projects
4.2.2 Projects from Egypt

Natural gas-fuelled power plants produce approximately 70% of Egypt’s electric-
ity, with the remaining 30% coming from power plants fuelled by oil or renewable
sources, mostly hydropower. In spite of its vast solar and wind resource potential,
the renewable energy sector in Egypt is not well developed. Egypt aims at increas-
ing the share of renewable energy in its mix to 20% (12% wind, 6% hydro, and
2% solar) by 2020 (Ministry of Electricity and Renewable Energy, 2013, 2015). In
this section, we apply the evaluation model on three proposed energy projects in
Egypt. Two of them are renewable energy projects (wind and solar) and the third
is one of the biggest gas-powered stations to be built in Egypt. Table 3.5 shows a
summary of the details of the projects.
The simulation results for the Egyptian gas-powered project show how con-
sidering the education impact converts the negative impacts on human capital to
positive impacts. It also shows how loans affect the deterioration of the PC. The
same effect of loans applies for the wind power project; yet, both wind and solar
power projects have positive impacts on human capital with and without the edu-
cation effect, as shown in Figures 3.12 to 3.17.
The CO2 damage in the case of the gas-powered power plant was almost 100
times the damage from the other projects, although the capacity of the gas‑pow-
ered plant is only approximately three times the capacity of the solar power proj-
ect (Figure 3.18). The big difference in ESI between the gas-powered plant and
the solar power project (Figure 3.19) can be explained by the greater number of
load hours that gas projects usually have than do solar power plants.
Table 3.5 Egypt projects data
Project Project Capacity Construction dates Investment Emissions Lifetime

(fc) code (j) (Inv) data sources (lf)
Start End
(Condt ) (opdt )
New Capital CCGT power EgG 4.8 GW 3/1/2016 6/1/2018 Total $2.23 billion financed by a [1], [2], [4] 45 years
plant consortium of 17 international banks
Gulf of el Zayt wind farm EgW 200 MW 1/1/2018 8/1/2019 Total $370 million, with $209 million [1], [2], [3] 30 years
donation by KFW, $33 million EU
grant, and $54 million EIB loan
Benban solar power complex EgPV 1.8 GW 10/1/2016 5/1/2020 $2 billion [1], [2], [5] 30 years
Emissions sources: [1] Weisser (2007), [2] Turconi et al. (2013), [3] Schleisner (2000), [4] Spath and Mann (2000), [5] Fthenakis et al. (2008).
Impacts on Capital assets considering edu. attainment (Egypt/Gas)
300 HC
250
200
150
$2010(bn)
100
50
0
0.00
–0.05
–0.10
–0.15
$2010(bn)
–0.20 PC
0
–100 NC
–200
–300
–400
–500
$2010(bn)
–600
–700
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.12 Impacts of the proposed CCGT power plant on the Egyptian capital assets
Impacts on Capital assets considering edu. attainment (Egypt/PV)
200 HC
150
100
$2010(bn)
50
0
–0.00
PC
–0.05
–0.10
–0.15
$2010(bn)
–0.20
–100 NC
–200
–300
–400
–500
$2010(bn)
–600
–700
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.13 Impacts of the proposed solar power project on the Egyptian capital assets
Impacts on Capital assets considering edu. attainment (Egypt/Wind)
150 HC
125
100
75
$2010(bn)
50
25
0
0.004
0.002 PC
0.000
–0.002
–0.004
–0.006
$2010(bn)
–0.008
–0.010
0
–50 NC
–100
–150
–200
–250
$2010(bn)
–300
–350
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.14 Impacts of the proposed wind power project on the Egyptian capital assets
Impacts on Capital assets NOT considering edu. attainment (Egypt/Gas)
0.00
HC
–0.02
–0.04
–0.06
$2010(bn)
–0.08
–0.10
0.00
–0.05
–0.10
–0.15
$2010(bn)
–0.20 PC
0
–100 NC
–200
–300
–400
$2010(bn)
–500
–600
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.15 Impacts of the proposed CCGT power plant on the Egyptian capital assets (without education impact)
Impacts on Capital assets NOT considering edu. attainment (Egypt/PV)
0.0025
HC
0.0020
0.0015
0.0010
$2010(bn)
0.0005
0.0000
0.020
PC
0.015
0.010
0.005
$2010(bn)
0.000
–0.005
0
NC
–50
–100
–150
–200
$2010(bn)
–250
–300
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.16 Impacts of the proposed solar power project on the Egyptian capital assets (without education impact)
Impacts on Capital assets NOT considering edu. attainment (EgW)
0.00005
HC
0.00004
0.00003
0.00002
$2010(bn)
0.00001
0.00000
0.004
0.002 PC
0.000
–0.002
–0.004
–0.006
$2010(bn)
–0.008
–0.010
0
NC
–50
–100
–150
$2010(bn)
–200
–250
15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20
Year
Figure 3.17 Impacts of the proposed wind power project on the Egyptian capital assets (without education impact)
CO2 damages
600
Egypt/Gas
500
Egypt/PV
$2010, millions
400 Egypt/Wind
300
200
100
0
4.0
3.5 Egypt/PV
3.0 Egypt/Wind
$2010, millions
2.5
2.0
1.5
1.0
0.5
0.0
15
20
25
30
35
40
45
50
20
20
20
20
20
20
20
20
Year
Figure 3.18 CO2 damages from the three proposed power projects in Egypt
Energy Supply Index

12
Egypt/Gas
10
% of total demand
Egypt/PV
8 Egypt/Wind
6
4
2
0
1.4
1.2 Egypt/PV
% of total demand
1.0 Egypt/Wind
0.8
0.6
0.4
0.2
0.0
15
20
25
30
35
40
45
50
20
20
20
20
20
20
20
20
Year
Figure 3.19 ESI from the three proposed energy projects in Egypt
Finally, Figure 3.20 shows the change of wealth that resulted from the introduc-
tion of each of the projects to the system in Egypt. It also shows how considering
education has a great impact on the wealth changes.
Developing countries, such as Egypt, might have limited energy diversification
options with growing energy needs. However, the simulation results show that
Expected project impacts on Wealth

Considering impacts on edu. attainment
0
–500
$2010(bn)
–1000
–1500
Egypt/Gas
–2000 Egypt/Wind
–2500 Egypt/PV
Not considering impacts on edu. attainment

0
–1000
–2000
$2010(bn)
–3000
–4000
–5000 Egypt/Gas
Egypt/Wind
–6000
Egypt/PV
–7000
15
20
25
30
35
40
45
50
20
20
20
20
20
20
20
20
Year
Figure 3.20 The changes in wealth caused by the proposed energy projects in Egypt
better allocations of loans and energy policy reforms can have positive impacts on
keeping the country’s wealth on a sustainable track.
5 Conclusion
We used the capabilities of a hybrid simulation to perform a wealth-based eval-
uation for prospective energy infrastructure projects. The framework introduced
in this chapter enables policymakers to explore the impacts of different energy
projects and policies on the course of different capital assets, as well as to esti-
mate the CO2 damages corresponding to each project or mix of projects. We
applied the framework to different case studies. The results showed how dif-
ferent technologies and fuel cycles affect wealth. The study also shed light on
the notion introduced in IWR 2014: if energy production has a direct positive
impact on human capital through education, higher returns for IW growth would
be generated.
Another important conclusion is how important it is for developing countries
to revise their energy investment policies and introduce more efficient incentives,
such as feed-in tariff, to encourage investments in clean energy, rather than build-
ing energy projects using borrowed funds.
When there is a significant variation between projects in terms of investment,
electricity output, or even the number of load hours, it might be impractical to
directly conclude that the project with the most preferable impact on wealth is
the best choice. The model in that case will be useful in showing policymakers
how their choice will affect the wealth trajectory. However, the model can be
informative in another way when facing the dilemma of projects that are different
in capacity. In that case, policymakers can normalize projects for capacity and
then provide new emissions vectors relative to the normalized capacity of each
project. Running the model with the new emissions vector will lead to different
pollutant concentrations and, hence, different implications for each of the capital
assets and total wealth. This way, policymakers will know the impacts of gener-
ating a certain amount of electricity by each technology or fuel cycle condition.
Note
1 We chose not to include project decommissioning in the model as the decommissioning
process complexity and legal implications vary between countries.
Bibliography
Arrow, K. J., Dasgupta, P., Goulder, L. H., Mumford, K. J., and Oleson, K. (2012). Sustain-
ability and the measurement of wealth. Environment and Development Economics,
17(October), 315–316. https://doi.org/10.1017/S1355770X12000186
ability and the measurement of wealth: Further reflections. Environment and Develop-
ment Economics, 18(4), 504–516. https://doi.org/10.1017/S1355770X13000193
Arrow, K. J., Dasgupta, P., and Mäler, K. G. (2003). Evaluating projects and assessing
sustainable development in imperfect economies. Environmental and Resource Econom-
ics, 26, 647–685.
Atilgan, B., and Azapagic, A. (2016). An integrated life cycle sustainability assessment of
electricity generation in Turkey. Energy Policy, 93, 168–186. https://doi.org/10.1016/j.
enpol.2016.02.055
Barro, R. J., and Lee, J. W. (2013). A new data set of educational attainment in the world,
1950–2010. Journal of Development Economics, 104, 184–198. https://doi.org/10.1016/j.
jdeveco.2012.10.001
Benaich, H., and Pruyt, E. (2015). Exploring traffic and congestion policies an entity-based SD
approach. Retrieved from www.systemdynamics.org/conferences/2015/papers/P1312.pdf
Bianco, V., Manca, O., and Nardini, S. (2009). Electricity consumption forecasting in Italy
using linear regression models. Energy, 34(9), 1413–1421. https://doi.org/10.1016/j.
energy.2009.06.034
Bickel, P., and Friedrich, R. (n.d. 2005). Directorate-General for Research Sustainable
Energy Systems ExternE Externalities of Energy Methodology 2005 Update. Retrieved
from www.externe.info/externe_d7/sites/default/files/methup05a.pdf
Brown, M. A., and Sovacool, B. K. (2007). Developing an “energy sustainability index” to
evaluate energy policy. Interdisciplinary Science Reviews, 32(4), 335–349. https://doi.
org/10.1179/030801807X211793
Collins, R. D. (2013). Using inclusive wealth for dynamic analyses of sustainable develop-
ment: Theory, reflection and application. Systems Dynamics Society Conference (Octo-
ber), 1–21.
Collins, R. D., Sakhrani, V., Selin, N. E., Alsaati, A., and Strzepek, K. M. (2014). Using
inclusive wealth for policy evaluation: The case of infrastructure capital. Inclusive
Wealth Report 2014: Measuring progress toward sustainability, Cambridge University
Press, Cambridge, UK.
Collins, R. D., Selin, N. E., de Weck, O. L., and Clark, W. C. (2017). Using inclusive wealth for
policy evaluation: Application to electricity infrastructure planning in oil-exporting countries.
Ecological Economics, 133, 23–34. https://doi.org/10.1016/j.ecolecon.2016.11.013
Conference Board. (2016). Total Economy Database. Retrieved from www.conference-
board.org/data/
Dasgupta, P. (2009). The welfare economic theory of green national accounts. Environmen-
tal and Resource Economics, 42(1), 3–38. https://doi.org/10.1007/s10640-008-9223-y
Fiddaman, T. S. (2002). Exploring policy options with a behavioral climate-economy
model. System Dynamics Review, 18(2), 243–267. https://doi.org/10.1002/sdr.241
Fthenakis, V. M., Hyung, C. K., and Alsema, E. (2008). Emissions from photovoltaic life
cycles. Environmental Science and Technology, 42(6), 2168–2174. https://doi.
org/10.1021/es071763q
Food And Agriculture Organization Of The United Nations (FAO). (2014). FAOSTAT.
Retrieved from www.fao.org/faostat/en/#data
Holley, C., and Lecavalier, E. (2017). Energy governance, energy security and environmen-
tal sustainability: A case study from Hong Kong. Energy Policy, 108(Supplement C),
379–389. https://doi.org/10.1016/j.enpol.2017.06.010
Hourcade, J.-C., Jaccard, M., Bataille, C., and Ghersi, F. (2006). Hybrid modeling: New
answers to old challenges introduction to the special issue of The Energy Journal. The
Energy Journal, SI2006(1), 1–11. https://doi.org/10.5547/ISSN0195-6574-EJ-
VolSI2006-NoSI2-1
International Energy Agency (IEA). (2016). Energy Policies of IEA Countries: Belgium.
Retrieved from www.iea.org/publications/freepublications/publication/
International Labor Organization. (2015). Key Indicators of the Labour Market (KILM)
Database. Retrieved from www.ilo.org
Kanagawa, M., and Nakata, T. (2008). Assessment of access to electricity and the socio-
economic impacts in rural areas of developing countries. Energy Policy, 36(6), 2016–
2029. https://doi.org/10.1016/j.enpol.2008.01.041
Kveselis, V., Dzenajavičienė, E. F., and Masaitis, S. (2017). Analysis of energy develop-
ment sustainability: The example of the lithuanian district heating sector. Energy Policy,
100(Supplement C), 227–236. https://doi.org/10.1016/j.enpol.2016.10.019
Kwakkel, J. H., and Pruyt, E. (2015). Using system dynamics for grand challenges: The
ESDMA approach. Systems Research and Behavioral Science, 32(3), 358–375. https://
doi.org/10.1002/sres.2225
Leung, C. S., and Meisen, P. (2005). How electricity consumption affects social and eco-
nomic development by comparing low, medium and high human development countries.
(153), 1–12.
Ministry of Electricity and Renewable Energy, A. R. of E. (2013). New & Renewable
Energy Authority (NERA) Annual Report. Retrieved from www.nrea.gov.eg/annual
report/Annual_Report_2012_2013_eng.pdf
Ministry of Electricity and Renewable Energy, A. R. of E. (2015). Egyptian Electricity
Holding Company Annual Report. Retrieved from www.moee.gov.eg/english_new/
EEHC_Rep/2014-2015en.pdf
Mohamed, Z., and Bodger, P. (2005). Forecasting electricity consumption in New Zealand
using economic and demographic variables. Energy, 30(10), 1833–1843. https://doi.
org/10.1016/j.energy.2004.08.012
Naito, Y., Takashima, R., Kimura, H., and Madarame, H. (2009). Evaluating replacement
project of nuclear power plants under uncertainty. Energy Policy, 38, 1321–1329. https://
doi.org/10.1016/j.enpol.2009.11.010
Pearson, L. J., Biggs, R., Harris, M., and Walker, B. (2013). Measuring sustainable develop-
ment: The promise and difficulties of implementing inclusive wealth in the Goulburn-
Broken Catchment, Australia. Sustainability: Science, Practice, & Policy, 9(1), 16–27.
https://doi.org/10.1111/mono.12011
Schleisner, L. (2000). Life cycle assessment of a wind farm and related externalities.
Renewable Energy, 20(3), 279–288. https://doi.org/10.1016/S0960-1481(99)00123-8
Spath, P. L., and Mann, M. K. (2000). Life cycle assessment of a natural gas combined-
cycle power generation system, 1–56. https://doi.org/10.2172/776930
Stiglitz, J. E., Sen, A., and Fitoussi, J.-P. (2010). Mismeasuring our lives: Why GDP doesn’t
add up, 1, 136. Retrieved from www.nowforourturn.org/Reframing/StiglitzSustainabilityE.
pdf
Tidball, R., Bluestein, J., Rodriguez, N., and Knoke, S. (2010). Cost and performance
assumptions for modeling electricity generation technologies. National Renewable
Energy Lab, NREL, Subcontract Report (November), 37–41. https://doi.org/10.2172/
993653
Tokimatsu, K., Yamaguchi, R., Sato, M., Yasuoka, R., Nishio, M., and Ueta, K. (2011).
Measuring future dynamics of genuine saving with changes of population and technol-
ogy: Application of an integrated assessment model. Environment, Development and
Sustainability, 13(4), 703–725. https://doi.org/10.1007/s10668-011-9285-9
Tokimatsu, K., Yamaguchi, R., Sato, M., Yasuoka, R., Nishio, M., and Ueta, K. (2014).
Assessing future sustainability by forecast of genuine savings paths. Environmental Eco-
nomics and Policy Studies, 16(4), 359–379. https://doi.org/10.1007/s10018-013-
0056-8
Turconi, R., Boldrin, A., and Astrup, T. (2013). Life cycle assessment (LCA) of electricity
generation technologies: Overview, comparability and limitations. Renewable and Sus-
tainable Energy Reviews, 28, 555–565. https://doi.org/10.1016/j.rser.2013.08.013
Uehara, T., Nagase, Y., and Wakeland, W. (2016). Integrating economics and system
dynamics approaches for modelling an ecological: Economic system. Systems Research
and Behavioral Science, 33(4), 515–531. https://doi.org/10.1002/sres.2373
UNU-IHDP, and UNEP. (2012). Inclusive wealth report 2012: Measuring progress toward
sustainability. Cambridge University Press, 13(4), ijshe.2012.24913daa.006. https://doi.
org/10.1108/ijshe.2012.24913daa.006
UNU-IHDP, and UNEP. (2014). Inclusive wealth report 2014: Measuring progress toward
sustainability. Cambridge University Press, 370. https://doi.org/10.1108/ijshe.2012.
24913daa.006
Weisser, D. (2007). A guide to life-cycle greenhouse gas (GHG) emissions from electric
supply technologies. Energy, 32(9), 1543–1559. https://doi.org/10.1016/j.energy.
2007.01.008
World Bank. (2016). World Development Indicators|DataBank. Retrieved from http://data
bank.worldbank.org/data/reports.aspx?source=world-development-indicators
Yamaguchi, R. (2014). Inclusive wealth with a changing but aging population. Economics
Letters, 124, 21–25. https://doi.org/10.1016/j.econlet.2014.04.011
4 Public debt as a negative stock
in sustainability indicator
Masayuki Sato and Shunsuke Managi
1 Introduction
Public debt is one of the negative future assets. It is also typically a controver-
sial issue in discussions on the sustainability of each country. Some European
countries, such as Greece and Italy, and an economically large country, such as
Japan, have a serious problem of sustaining their social institutions because of
their cumulative public debt.
Because public debt is closely related to fiscal policy, it is important to inves-
tigate its effect on the path of sustainable development; in particular, it needs to
be assessed by paying careful attention to its characteristics, especially volatility
(Sato et al. 2017). In the case of simple economic growth, there are two main
effects of volatility. Mirman (1971) suggested that volatility could boost growth
through precautionary savings. On the other hand, Bernanke (1983) showed that
volatility leads to a lower growth rate over time. More recently, van der Ploeg and
Poelhekke (2009) conducted an assessment of the potential effects of volatility
and found that the financial sector plays an important role in alleviating the harm-
ful effect of volatility in resource-exporting countries. Based on these studies,
Sato et al. (2017) investigates the effect of volatility to sustainability indicators,
and they also consider the stabilization by an effective institution.
In the previous research, however, public debt is not considered as negative
capital. This is partly because public debt finances the investment of other capital,
and is canceled out as total inclusive wealth. Only when the public government
spends the funds financed by public debt inefficiently does the negative effect of
public debt become an issue. Arronsson et al. (2012) considered the dead weight
loss of public debt through the distortion of labor market by taxation in the future.
They adjusted the sustainability indicator by the marginal excess burden of public
debt, and showed that it should be revised downward.
But the estimation of marginal negative effect of public debt is controversial
(Dahlby 2008; Jacobs 2018). The calculation in Arronsson et al. (2012) assumed
that the marginal cost of public debt is 0.1, 0.3, and 0.5. We need to consider
whether the range of marginal cost of public debt assumed is applicable to the
sustainability indicator. This chapter investigates the relationship between genu-
ine savings and public debt, and confirms the large effect of public debt on the
indicator.
78 Masayuki Sato and Shunsuke Managi
2 Method
This chapter based on the concept of inclusive wealth, which is the weighted sum
of all types of capital, as a source of well-being. When summing up each capital,
the weight is the shadow price. In the economics literature, the total of the various
kinds of stocks has been called inclusive wealth. Man-made capital (economic
physical capital), human capital, and natural capital are included in the inclusive
wealth, and even social capital, knowledge, institutions, culture, and time can be
considered as a part of inclusive wealth (Dasgupta 2009) because all of them can
contribute to producing well-being for people. Formally, inclusive wealth W is
represented by Equation (1).
Wt = K Mt + K Ht + K Nt , (1)
where K Mt is the man-made capital, K Ht is the human capital, and K Nt is the natu-
ral capital; all of these are evaluated by their shadow prices.1 The time differentia-
tion for W is often called genuine savings (GS), and the sustainability condition
requires that GS is always non-negative, as follows:
dWt dKM t dKH t dKN t

GSt = + + 0, for all t. (2)
dt dt dt dt
Non-declining GS implies a non-decreasing source of well-being in future, which

assures sustainability of the development. Therefore we always pay attention to
the sign of GS of each country.
Much effort has been directed toward the better measurement of the various
stocks required for construction of an economic indicator of sustainability. For
example, the Inclusive Wealth Index (IWI) of UNU-IHDP and UNEP and the
Adjusted Net Savings (ANS) of the World Bank are widely used databases for
assessing sustainability. Such a database, however, does not have a perfect estima-
tion of invisible capital, such as natural capital. Therefore, many researchers have
been trying to improve the indicators by a more sophisticated estimation of the
different types of capital (e.g., Fenichel and Abbott 2014).
In addition to the elaboration of the shadow price estimation, other important
types of capitals need to be introduced into the calculation of inclusive wealth.
One of the most significant is the public debt of each country (see Arronsson et al.
2012).
2.1 Direct and indirect effect of public debt

on Inclusive Wealth Index
This section discusses the empirical method to investigate the relationship
between public debt and inclusive wealth. Here, we consider both direct and indi-
rect effects; the direct effect is the contribution (possibly negative contribution) of
public debt to the inclusive wealth accumulation and the indirect effect is via vol-
atility stabilization. This is because the public debt is usually a financial vehicle
Public debt as a negative stock 79
of fiscal policy, which partly aims to stabilize the volatility of the development
path. In order to do this analysis, we follow the estimation framework adopted
by Ramey and Ramey (1995), van der Ploeg and Poelhekke (2009), and Sato
et al. (2017). They applied the ARCH-M methodology developed by Engle et al.
(1987), which is a modification of Engle’s (1982) ARCH model. It incorporates
the variance in the mean equation. The mean and variance equations are then
constructed as follows:
log(GSit ) = c1 + β1σit2 + M it β + X itϕ + εit , (3)
εit ~ N (0, σi2, t ), and (4)
σit2 = exp(c 2 + X it x + Z it y ) + φ1σt2−1 + φ2σt2−2 + ηit , (5)
where GSit represents the transformed GS in country i. Since the original value of
GS can be non-positive, it is transformed as shown below so that we may convert
it to logarithmic form.
GSit = gsit +1, (6)
where gsit is the original data of GS in country I; σ2 is the variance of GS, reflecting
their volatility; and t is the time index. M is the vector of control variables, which
appear only in the mean equation, that is, Equation (3).
In the analysis of this chapter, public debt is introduced in equations (3) through
(5) of the ARCH-M model. The effect of public debt on GS is not clear because
there are two pathways. First, if public debt is used for the efficient accumulation
of man-made capital, the negative asset should be canceled out by the real asset.
In this case, the public debt may not affect GS. However, Arronsson et al. (2012)
indicate that public debt produces dead weight loss in future when the labor tax
is levied on repayment; thus, it will eventually reduce GS. Therefore, whether the
direct effect has significance or not is itself interesting. Second, public debt stabi-
lizes the development path by fiscal policy, which implies that if the coefficient of
volatility, β1, is negative, then a decrease in volatility may contribute to a higher
GS value. Figure 4.1 illustrates the effect of public debt on GS.
3< (>)0
Public Debt GS
<0 1<0
Volatility
Figure 4.1 The relationship and expected sign of public debt and GS
250
200
150
Central
100 government debt,
total (% of GDP)
50
0
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
Figure 4.2 Public debt of Japan
3 Japanese case and data

Japan is a typical country with huge amount of public debt. The discussion about
sustainability in Japan always considers fiscal sustainability (see Figure 4.2).
The sustainability indicators, such as the Inclusive Wealth Index and GS, which
are based on the capital approach, assess Japan as a sustainable country, that is,
they indicate that the annual inclusive capital change has a positive sign. How-
ever, as the previous sections of this chapter pointed out, the accumulating public
debt can be a risk factor for sustainability in the future. To investigate this risk,
this chapter analyzes the effect of public debt on GS as the sustainability indicator.
We use GS data from the World Development Indicators (WDI) released by
the World Bank.2 As stressed by van der Ploeg (2011), The World Bank’s ANS
possibly overestimates GS, especially in the resource-rich countries. However, in
this chapter, we focus on the relationships between volatility and the various types
of accumulated capital. For this aim, the ANS of WDI is employed as GS, taking
care in the interpretation of the results. WDI provides GS data calculated by three
dK Mt
types of capital: man-made capital investment (net national savings; pM );
dK Ht dt
human capital investment (education expenditures; pH ); and the degrada-
dt
tion of natural resources (energy depletion, mineral depletion, forest depletion,
dK Nt
and CO2 emissions; pN
dt
In addition to these components, we add the public debt factor (debt) in the
regression analysis. To control other factors, we introduce variables, such as per
capita GDP growth (ggdp), age dependency (age_dep), population growth (popg),
urban population rate (urban), trade openness (trade), and government expen-
diture (gov_size), which are also taken from the WDI database for econometric
analysis.
From the preceding discussion, the mean and variance estimation equations for
our ARCH-M model can be rewritten as
log (GSi ,t ) = c1 + β1σ12i ,t + β2 ggdpi ,t + β3 popgi ,t + β4 agedepi ,t
+β5urbani ,t + β6tradei ,t + β7 gov _ sizei ,t + β8 debti ,t (7)

ε ∼ N (0, σi2,t ) (8)
σi2,t = exp (c2 + γ1 ggdpi ,t + γ 2 popgi ,t + γ 3tradei ,t + γ 4 govsizei ,t + γ 5 debti ,t (9)
The main purpose of regression analysis is to confirm whether the value of GS on

the left-hand side of the mean equation (Equation [7]) is negatively affected by
its own volatility. The expected signs of the variables are as follows. Per capita
GDP growth (ggdp) contributes GS (β1 > 0) due to increasing man-made capital
investment. Although trade openness may or may not contribute to an increase
in the level of GS itself, it is certainly a source of volatility in GS, as discussed
by van der Ploeg and Poelhekke (2009). Therefore, its expected sign in the vari-
ance equation is positive (γ3 > 0). Government size also may or may not have a
positive impact on the level of GS, but it is expected to help stabilize the develop-
ment path owing to the basic “stabilizing function” of the government. Therefore,
the expected sign of its estimated coefficient in the variance equation is negative
(γ4 < 0). The signs of population growth, age dependency, and urban population
rate are not known in advance. The sign of public debt is the most important in
this chapter, but this is also not known in advance.
4 Estimation result
By using the ARCH-M model, we obtained the estimation, as shown in Table 4.1.
We find that most of the coefficients are significantly estimated. It is worth noting
that the coefficient of volatility is negative. This implies that the volatile path of
GS is not desirable for sustainable development.
The direct effect of public debt is small. This result is consistent with the argu-
ment of canceling out in GS calculation. However, the effect of public debt on
volatility is significantly positive and seems not negligible in magnitude. Interest-
ingly, public debt boosts volatility. Because fiscal policy based on issuing public
debt is usually conducted in order to stabilize the path, this result implies that its
stabilization function does not work. This suggests that Japanese public debt is
not transformed into a productive base, but currently consumed as social security,
medical assistance, and so forth. Inevitably, without enough reinvestment into
productive capital, such as man-made capital and human capital, the inclusive
wealth decreases.
Together with the result that volatility has a negative effect on GS, it is found
that public debt can have a negative effect on sustainability. From the estimation
result, we illustrate the relationship between GS and public debt in Figure 4.3.
Table 4.1 Estimation result of ARCH-M
Mean equation (7)
Coef. Std. err. z-value p-value 95% Conf. interval
ggdp 0.2745851 0.0106228 25.85 0.000 0.2537648 0.2954054

popg 1.141283 0.0550174 20.74 0.000 1.033451 1.249116
age −0.1129911 0.0052988 −21.32 0.000 1.033451 1.249116
urban −0.061621 0.0027629 −22.30 0.000 −0.1233765 −0.1026056
trade 0.0300981 0.0013353 22.54 0.000 0.0274809 0.0327152
gov_size −0.1791174 0.0113778 −15.74 0.000 −0.2014175 −0.1568174
debt 0.000027 0.0000152 1.77 0.076 −2.85E-06 0.0000568
_cons 0.1934456 0.004139 46.74 0.000 0.1853332 0.2015579
sigma2ex −0.0412461 0.0149712 −2.76 0.000 −0.0705891 −0.0119031
Variance equation (9)
Coef. Std. err. z-value p-value 95% Conf. interval
ggdp −30.4605 3.9534 −7.70 0.000 −38.2091 −22.7120

popg 39.661 13.86496 2.86 0.004 12.48618 66.83582
trade 0.9199192 0.1960941 4.69 0.000 0.5355818 1.304257
gov_size −17.65459 3.098779 −5.70 0.000 −23.72809 −11.5811
debt 0.0130571 0.001421 9.19 0.000 0.010272 0.0158421
_cons −7.83598 0.6166445 −12.71 0.000 −9.044581 −6.627379
Number of obs. = 1804

Wald chi2(8) = 4367.76
Log likelihood = 2949.549
Prob > chi2 = 0.000
0.000027
Public Debt GS
0.00130571 -0.0412461
Volatility
Figure 4.3 Estimated relationship between GS and public debt
Based on the relationship in Figure 4.3, we calculate the negative effect of

public debt on GS.
lnGS = 0.000027 Debt − 0.041σ 2

σ 2 = 0.013Debt
hence,
lnGS = −0.00051Debt
This implies that the negative effect of public debt on the level of Japanese GS is
calculated as −0.36 because Japanese public debt is around 200% of GDP.
Figure 4.4 illustrates the Japanese path of GS (real line) and the reduction in
its value by negative public debt (dotted line). The figure also shows the case of
Italy, UK, and Spain if we assume that the negative effect of public debt is similar
to that in Japan.
18
Japan
16
14
12
10
0
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Italy
14
12
10
0
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Figure 4.4 GS path of each country

United Kingdom
12
10
0
19891991199319951997199920012003200520072009201120132015
Spain
16
14
12
10
0
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Figure 4.4 (Continued)

5 Conclusion
This chapter discussed the effect of public debt on a sustainability indicator. So
far, sustainability indicators based on capital approach have measured man-made
capital, human capital, and natural capital. These types of capital potentially pro-
duce well-being when they are used. In this sense, if the public debt is efficiently
transformed into such types of productive capital, public debt is not a problem.
However, as Arronsson et al. (2012) discussed, if there is any inefficiency and
dead weight loss due to public debt, it will negatively affect sustainability.
Especially in Japan, the cumulating public debt is a big issue in the discus-
sion on sustainability. However, we cannot refer to the issue of public debt using
traditional indicators. This chapter analyzes the empirical effect of public debt
on a sustainability indicator by the ARCH-M model. The analysis considers both
direct and indirect effects. The estimation results of the direct effect shows that
the effect of issuing public debt is not critical. This implies that the public debt is
canceled out by its transformation into productive capital. The result is consistent
with the justification that traditional inclusive wealth does not consider the pub-
lic debt as negative capital. However, this chapter found that there is an indirect
effect, which is via path stabilization. One of the most important aims of the fiscal
policy is to stabilize the volatility in the economy and as pubic debt is the main
vehicle to finance the policy, it is important to consider the effect of stabilization
through a fiscal policy based on public debt. As expected, the result shows that the
volatility of the path negatively affects GS. Importantly, the result shows that pub-
lic debt boosts the volatility. This implies that, if public debt is not transformed
into productive capital, as is the case in Japan, but consumed instead, the public
debt significantly drives down the GS. This is especially important in the case of
countries with huge public debt.
Notes
1 Dasgupta (2004) also included knowledge capital in Equation (1). However, as men-
tioned in Section 3, our data do not include knowledge capital owing to the fact that
the available database for GS is based on only three types of capital (man-made capital,
human capital, and natural capital) in Equation (1).
2 Available at http://databank.worldbank.org/.
Bibliography
Arronsson, T., C. Cialani and K. G. Lofgren (2012), “Genuine saving and the social cost of
taxation,” Journal of Public Economics, Vol. 96, pp. 211–217.
Bernanke B. (1983), “Irreversibility, uncertainty and cyclical investment,” Quarterly Jour-
nal of Economics, Vol. 98, pp. 85–106.
Dahlby, B. (2008), Marginal Cost of Public Funds, The MIT Press, Cambridge.
Dasgupta, P. (2004), Human Well-Being and the Natural Environment, Oxford University
Press, New York.
Dasgupta, P. (2009), “The Welfare Economic Theory of Green National Accounts,” Envi-
ronmental and Resource Economics, Vol. 42, pp. 3–38.
Dietz, S., E. Neumayer and I. De Soysa (2007), “Corruption, the Resource Curse and Genu-
ine Saving,” Environment and Development Economics, Vol. 12, pp. 33–53.
Engle, R. (1982), “Autoregressive Conditional Heteroskedasticity with Estimates of the
Variance of United Kingdom Inflation,” Econometrica, Vol. 50, No. 4, pp. 987–1007.
Engle, R., D. Lilien and R. Robins (1987), “Estimating Time Varying Risk Premia in the
Term Structure: The Arch-M Model,” Econometrica, Vol. 55, No. 2, pp. 391–407.
Fenichel, E. P. and J. K. Abbott (2014), “Natural Capital: From Metaphor to Measure-
ment,” Journal of the Association of Environmental and Resource Economists, Vol. 1,
No. 1, pp. 1–27.
Jacobs, B. (2018), “The Marginal Cost of Public Funds is One at the Optimal Tax System,”
International Tax and Public Finance, forthcoming.
Lee, J. (2010), “The Link between Output Growth and Volatility: Evidence from a GARCH
Model with Panel Data,” Economics Letters, Vol. 106, pp. 143–145.
Mirman, L. J. (1971), “Uncertainty and Optimal Consumption Decisions,” Econometrica,
Vol. 39, pp. 179–185.
Ramey, G. and V. A. Ramey (1995), “Cross-Country Evidence on the Link between Volatil-
ity and Growth,” American Economic Review, Vol. 85, pp. 1138–1151.
Sato, M., S. Samreth, and K. Sasaki (2017), “The Impact of Institutional Factors on the
Performance of Genuine Savings,” forthcoming.
van der Ploeg, F. (2011), “Rapacious Resource Depletion, Excessive Investment and
Insecure Property Rights: A Puzzle,” Environmental and Resource Economics, Vol. 48,
pp. 105–128.
van der Ploeg, F. and S. Poelhekke (2009), “Volatility and the Natural Resource Curse,”
Oxford Economic Papers, Vol. 61, pp. 727–760.
5 Global marine fisheries
with economic growth
Yogi Sugiawan, Moinul Islam
and Shunsuke Managi
1 Introduction
The ocean provides an enormous amount of resources that are essential not only
for providing basic human needs but also for supporting human wealth. How-
ever, the ocean’s ability to provide sustainable benefits for human well-being is
limited by its regenerative capacity, which is currently deteriorating due to over-
exploitation, pollution and coastal development (Halpern et al., 2012). This has
spurred persistent debates regarding the state of global marine fisheries over the
last two decades. Several scientists believe that marine fisheries tend to be unsus-
tainable and that the stock of global marine fisheries is facing threats of serial
depletion (Hutchings, 2000; Jackson et al., 2001; Pauly et al., 2002; Srinivasan
et al., 2010; Worm et al., 2006, 2007; Zeller et al., 2009). This is indicated by the
increasing number of fish species that are classified as overfished or as collapsed
(Branch et al., 2011; Froese et al., 2012), by declining catch trends (Pontecorvo
and Schrank, 2012; Zeller and Pauly, 2005), and by the declining mean trophic
levels of catch (Myers and Worm, 2003; Pauly et al., 1998; Pauly and Palomares,
2005). Additionally, Worm et al. (2006) raised concerns even further by arguing
that if current trends of fish overexploitation continue, global marine fisheries are
projected to collapse by 2048. On the other hand, arguments against this view
contend that current fishing practices are sustainable and that concerns of the col-
lapse of global marine fisheries are slightly exaggerated and misleading (Hilborn,
2007; Murawski et al., 2007; Pauly et al., 2013). Proponents of this view argue
that assessments of stock abundance that use catch data as a proxy are not reliable,
as a declining catch does not solely denote a declining stock and vice versa. Geph-
art et al. (2017) show that in addition to cases of fishery collapse, catch levels are
also prone to a broad variety of disruptions and shocks such as natural and man-
made disasters, policy changes, increasing fuel costs, and low fish prices. Hence,
Worm et al.’s (2006) gloomy projections of the collapse of global marine fisher-
ies, which are based on the assessment of catch time-series data, are somewhat
misleading (Hilborn, 2007; Murawski et al., 2007).
Regardless of ongoing disputes between these two contradictory views, the
amount of fish stock that is being overfished and that has collapsed is rather high.
Branch et al. (2011) explain that proportions of fish stocks that are overfished
88 Yogi Sugiawan et al.
and that have collapsed have been stable in the range of 28%–33% and 7%–13%,
respectively. This denotes an occurrence of resource deterioration due to the
exploitation of fish that exceeds maximum sustainable yields and the regenera-
tive capacities of oceans. Economists explain changes in resource availability and
environmental degradation based on economic factors. In a simple case, resource
degradation is a transient consequence of economic development that is inevi-
table. However, after reaching a certain level of economic growth, the beneficial
impacts of economic growth on resource quality will be achieved, ameliorating
damages to nature. If this holds for global marine fisheries, then stock decline can
be only temporary and it need not be considered a threat to sustainability over
the long term, as further economic growth is expected to lead to stock recovery
through the institution of better management systems and policies. This is referred
to as the environmental Kuznets curve (EKC) hypothesis. Alternatively, we might
find a monotonic relationship or even complex relationship that mainly depends
on resource stock estimates and catch data.
Most previous studies due to data availability issues have focused mainly on
the impacts of economic growth on pollution levels, which act as an inversely pro-
portional proxy for environmental quality (Grossman and Krueger, 1991; Managi
et al., 2009). These studies aim to test the existence of the EKC hypothesis and
to find a turning point in the economy after which environmental damages will
be ameliorated. However, to the best of our knowledge, only a few studies have
examined income-natural resource relationships (see for instance Ewers (2006),
Nguyen Van and Azomahou (2007), Caviglia-Harris et al. (2009), and Al-mulali
et al. (2015)), and none has examined global marine fisheries within this frame-
work. Our main contributions are at least twofold. First, we attempt to estimate
the abundance of marine fisheries by relying on a method proposed by Martell
and Froese (2013). Second, we apply an economic model to assess the sustain-
ability of global marine fisheries by examining historical relationships between
global marine ecosystems and economic growth. We employ time-series catch
and estimated stock data as proxies for measuring the state of the global marine
ecosystem.
The remainder of this chapter is organized as follows. Section 2 describes the
current state of global marine fisheries and its association with economic devel-
opment. Section 3 discusses the research methodology and data used. Section 4
presents the main study findings and an analysis of the results. Section 5 presents
the study’s conclusions and its policy implications.
2 Economic development and the state of global

marine fisheries
The impacts of economic development on resource abundance can be differenti-
ated into three stages (Grossman and Krueger, 1991). The first stage is referred
to as the scale effect, which is characterized by a persistent utilization of heavy
machinery, indicating a structural change in an economy. At this stage, economic
development has negative impacts on the environment and spurs an upward trend
Global marine fisheries with economic growth 89
of environmental degradation and resource depletion (Panayotou, 1993). How-
ever, as incomes increase, the structure of the economy may change, shifting from
a resource-intensive economy to a service- and knowledge-based technology-
intensive economy (see Tsurumi and Managi (2010) for more information). This
stage is referred to as the composition effect, which is characterized by the devel-
opment of cleaner industries and by more stringent environmental regulations
that limit environmental pressures. Tamaki et al. (2017) show that better resource
management practices are beneficial not only for reducing resource exhaustion
but also for increasing production efficiency. Finally, a wealthy nation is capable
of allocating a higher share of R&D expenditures (Komen et al., 1997), leading to
the invention of new technologies that will gradually replace obsolete technolo-
gies that tend to be dirtier and less efficient. This stage is referred to as the techni-
cal effect, which also contributes to improvements in environmental quality. The
cumulative effects of these three different stages of economic development create
an inverted U-shaped relationship between economic growth and resource abun-
dance known as the EKC hypothesis.
Although the EKC hypothesis enticingly proposes the existence of a turning
point after which further economic growth may lead to environmental improve-
ments, it has some limitations that are worth mentioning. First, the estimated
turning point of the EKC can occur amid very high levels of income. Hence, the
beneficial impacts of economic growth on environmental quality are difficult or
even impossible to achieve. For instance, Jalil and Mahmud (2009), Bölük and
Mert (2015) and Sugiawan and Managi (2016) find a relatively high EKC turning
point that lies outside of the observed sample period for the case of carbon diox-
ide emissions. Second, the EKC hypothesis is not applicable to all environmental/
resource problems. For instance, Sinha and Bhattacharya (2017) show a reverse
trend of SO2 emissions, supporting the existence of the EKC hypothesis for 139
cities in India for 2001–2013. However, Nguyen Van and Azomahou (2007) find
no evidence of the EKC hypothesis for the case of deforestation in 59 develop-
ing countries for 1972–1994. In addition, Liao and Cao (2013) reject the validity
of the EKC hypothesis for global carbon dioxide emissions, although they find a
flattening trend in carbon dioxide emissions for high-income countries. Another
caveat pertains to the fact that the beneficial impacts of economic growth on envi-
ronmental quality are only temporary. De Bruyn et al. (1998) argue that over the
long term, new technologies will emerge, creating new pollutants and environ-
mental problems. Hence, although the inverted U-shaped relationship is initially
observed, a new turning point will appear, leading to a positive correlation between
income and environmental degradation. As a result, an N-shaped curve is likely
to be observed over the long term. Finally, the composition and technical effects
of the economy may also have negative effects on the environment (Tsurumi and
Managi, 2010). This might occur as a result of the poor implementation of environ-
mental regulations or due to the invention of more resource-intensive technologies.
If this occurs, then an EKC-type relationship is unlikely to be observed.
A scale effect for global marine fisheries was observed in the early nineteenth
century, which was marked by the operation of steam trawlers, power winches, and
diesel engines (Pauly et al., 2002). This industrialization process has resulted in over-
fishing and stock collapse (Branch et al., 2011; Froese et al., 2012) and in declin-
ing mean catch trophic levels (Myers and Worm, 2003; Pauly et al., 1998; Pauly
and Palomares, 2005), suggesting a decline in environmental quality and resource
abundance. Figure 5.1 shows the total catch of global marine fisheries obtained
through the Sea Around Us Project (Pauly and Zeller, 2015). Despite continuous
improvements made to fishing methods and technologies, global marine fish catches
finally reached a peak in 1996 and declined after experiencing continuous growth for
approximately four decades. Fortunately, this decline in the global catch was also fol-
lowed by a decline in global fishery discards (Zeller and Pauly, 2005), which is attrib-
uted to advancements in technology and to the use of more efficient fishing practices.
The composition effect of the economy, which reflects structural changes in the
economy, leads to the introduction of new regulatory means of supporting better
fisheries management. For instance, the United Nations Convention on the Law
of the Sea (UNCLOS), which came into force in 1994, and the individual trans-
ferable quota (ITQ) system introduced in the late 1970s act as countermeasures
against the collapse of global marine fisheries by boosting the economic ben-
efits of fisheries while maintaining their sustainability (Soliman, 2014). Under the
UNCLOS, the nations of the world are required to maintain rates of marine fishery
exploitation at a maximum sustainable yield (MSY). Similarly, the ITQ manage-
ment system regulates the total allowable catch (TAC) for a particular fish stock
and distributes quasi-ownership rights of the TAC to fishermen (Acheson et al.,
2015). Despite flaws of the ITQ system (see for instance Acheson et al. (2015)),
Costello et al. (2008) show that the ITQ management system helps not only retard
the collapse of global marine fisheries but also helps rebuild stock.
3 Methodology
3.1 Estimating biomass stock

Unlike estimation methods for other renewable natural resources, estimating the
abundance of marine fisheries is rather challenging. The most reliable means of
determining stock status is the stock assessment technique, which involves con-
ducting scientific surveys to collect data on fish age and size distributions and on
catches per unit of effort. However, this method is costly to apply, is time intensive,
and requires access to large volumes of data (Agnew et al., 2013). In addition,
Kleisner et al. (2013) argue that the technique is only applicable for a small frac-
tion of global stocks, and thus it is not a reliable method for portraying the status
of global marine fisheries. They recommend using widely available indicators that
can provide a better indication of the status of global marine fisheries, although
such indicators may be less precise than those of the stock assessment method.
Hence, rather than utilizing stock data drawn from the well-known RAM legacy
database (Ricard et al., 2012), we prefer to estimate the stock based on catch time-
series data drawn through the Sea Around Us Project (Pauly and Zeller, 2015),
which has broader coverage, accounting for more than 160 countries.
Some previous studies (e.g., Froese and Kesner-Reyes (2002), Pauly et al.
(2008), Froese et al. (2012) and Kleisner et al. (2013)) employ the stock status
plots (SSP) method, which uses widely available catch data to depict the state
of global marine fisheries. However, the SSP method only reveals the qualitative
status of fisheries, providing no estimations on the size of fish stocks. To make
quantitative estimates of the global marine fish stock, we use a simple yet power-
ful Schaefer production function (Schaefer, 1954). This model is preferred due to
its simplicity and attractive features in terms of determining returns based on fish
stocks and effort. Additionally, the model is suited to depicting the state of global
marine fisheries, as it uses catch data, which are widely available. The stock of
biomass at time t is given by the following equation:
  B 
Bt = Bt −1 + r i Bt −1 i 1− t −1  − Ct −1 (1)
  k 
where B is biomass, C is the annual catch, r is the intrinsic rate of population

growth and k is the parameter of the carrying capacity. While catch C time-series
data are widely available, other model parameters (r, k and B) are rather difficult
to obtain. However, Martell and Froese (2013) devise a simple means of estimat-
ing equation (1) that is strictly based on catch time-series data. They propose a
means of estimating sets of feasible r and k pairs from a uniform distribution
function satisfies the following model assumptions: (1) the estimated biomass is
never collapsed, (2) the estimated biomass never exceeds the carrying capacity
and (3) the final stock lies within the assumed range of depletion. The value of r
is determined based on the resilience classification of each species, which ranges
from 0.05 to 0.5 for low resilience levels, from 0.2 to 1.0 for medium resilience
levels and from 0.6 to 1.5 for high resilience levels. Meanwhile, the potential
value of k is determined based on the maximum catch volume, which ranges from
1 to 50 times the maximum catch. Additional assumptions on the potential range
of the initial and final value of biomass must also be applied. These assumptions
are made based on the ratio between respective catches and the maximum catch
(B/k). When the B0/k ratio is less than 0.5, the initial value of biomass is assumed
to account for approximately 0.5 to 0.9 of the carrying capacity. Otherwise, it
ranges from 0.3 to 0.6 of the carrying capacity. Similarly, the final biomass is
assumed to be approximately 0.3k to 0.7k when the B/k ratio is greater than 0.5.
Otherwise, the value ranges from 0.01k to 0.4k. From these pre-determined value
ranges, we randomly draw sets of r–k pairs that satisfy the aforementioned model
assumptions. Rather than estimating the MSY, our primary interest is to estimate
biomass trends. For this purpose, we take the geometric mean of r, k and the maxi-
mum value of initial biomass, which corresponds to each feasible set of r–k pairs,
and include them in equation (1).
From Sea Around Us Project catch time-series data (Pauly and Zeller, 2015),
we estimate the stock of more than 1400 species in 164 countries for 1950–2010
(see Table 5A.1 in the appendix for more information). The catch data used in
120 4000
Millions
Millions
3500
100
Landings (metric tons)
3000
80
Stock (metric tons)

2500
60 2000
1500
40
1000
20
500
0 0
1950 1960 1970 1980 1990 2000 2010
Landings (metric tons) Stock (metric tons)
Figure 5.1 Global fisheries catch and estimated stock trends
our estimation measure the volume of catches for all purposes in each respective
country’s exclusive economic zone (EEZ) based on domestic or foreign fleets.
Figure 5.1 shows that the global stock has experienced a steady rate of decline
along with an increasing catch volume. However, the rate of decline decreased
over the time period, implying beneficial impacts of better fisheries management
protocols. We carry out a further analysis of this trend by taking into account
different characteristics of each country as is shown in Figure 5.2. We can see
that some rich countries that have adopted quota-management systems such as
Japan, the United Kingdom and the United States have managed to reduce their
catch levels and to contribute significantly to declining levels of global catch. As
a result, these countries are able to maintain or even recover their stock levels.
On the other hand, declining levels of stock are observed for developing coun-
tries such as China, Indonesia and Malaysia. These countries are characterized by
increasing scales of economy and by relatively high levels of population growth,
which are likely to place escalating pressures on marine resources.
3.2 Economic modeling

Our chapter studies the relationship between economic growth and global marine
resources based on the following general parametric models:
ln Ct = β0 + β1 ln Yit + β2 ln Yit 2 + β3 ln Yit 3 + β4 ln Pit + εit (2)

ln Bt = γ 0 + γ1 ln Yit + γ 2 ln Yit + γ 3 ln Yit + γ 4 ln Pit + εit
2 3
(3)
where C is the catch value, B is the estimated biomass value, Y is the per capita
gross domestic product (GDP) and εit is the standard error term. To avoid omitted
-1.5%
-1.0%
0.01%
-1.5%
4.5%
88%
Figure 5.2 Comparison of world catch and estimated stock levels. A. Average annual catch
changes from 1961 to 2010 (%). B. Average annual stock changes from 1961
to 2010 (%)
variable bias, our models also include population density (P) as an independent
variable. Halkos et al. (2017) show strong evidence that the decline of natural
capital is associated with the increase of another type of capital, such as human
capital. Additionally, Merino et al. (2012) show that variations in fish production
are also driven by population growth. Furthermore, to account for trends in the
variables, we include time trends in our models. We prefer to use the reduced-
form model, as it allows us to study the relationship between income and resource
abundance both directly and indirectly without being distracted by other variables
(see List and Gallet (1999)).
Our first model (referred to as the catch model) examines dynamic levels of
catch, which act as an inversely proportional proxy for resource abundance,
based on variations in economic development. However, a dispute over the reli-
ability of using catch as a proxy for resource abundance might arise, as varia-
tions in catch levels are not simply caused by variations in resource abundance.
Hence, to ensure the robustness of our findings, we use the estimated size of stock
as a proxy for resource abundance in our second model (henceforth referred
to as the biomass model). Both of our models provide several possible func-
tional forms of the income-resources relationship1 (i.e., level, linear, quadratic,
or cubic), depicting how economic growth will affect resource abundance. A
level-type relationship suggests that economic growth is neither harmful nor
beneficial for resource abundance. Meanwhile, a linear-type relationship indi-
cates constant pressures of economic growth on resource abundance. The EKC
hypothesis is confirmed if there is an inverted U-shaped relationship between
per capita income and the catch value or a U-shaped relationship between per
capita income and the estimated volume of stock, suggesting the existence of
a turning point in the economy after which economic growth is beneficial for
resource abundance. Moreover, a cubic-type relationship follows either an N- or
flipped N-shaped curve, suggesting the existence of a secondary turning point
in the economy at which point the trend of the income-resource relationship is
reversed a second time.
Our models involve nonstationary heterogeneous panel data of a large number
of time-series and cross-sectional observations (50 years of observations for 70
countries). Hence, they cannot be estimated by simply pooling the data and by
using fixed or random effect estimators, which assume identical slope coefficients
across the groups. Additionally, estimating each group separately via the mean
group estimator approach is also inappropriate, as it allows intercepts, slope coef-
ficients, and error variances to differ across groups, overlooking the fact that some
parameters may be similar across groups (Pesaran et al., 1999). Therefore, we use
the pooled mean group (PMG) method, which combines pooling and averaging
methods developed by Pesaran et al. (1999). The PMG method allows for het-
erogeneity in intercepts, short-run coefficients and error variances but restrains
long-run coefficients as identical (Pesaran et al., 1999).
The PMG method requires that all variables are not integrated at an order of
higher than 1. To obtain the integration properties of our panel data, we use panel
unit root tests, which have a higher power compared to individual unit root tests
for each cross-section (see for instance Levin et al. (2002)). We employ three
panel unit root test methods: Im, Pesaran and Shin (IPS); Fisher-type Augmented
Dickey-Fuller (ADF-Fisher); and Fisher-type Phillips-Perron (PP-Fisher) tests,
as suggested by Al-mulali et al. (2015). The aforementioned panel unit root tests
have a null hypothesis of non-stationarity and an alternative hypothesis of no
panel unit root.
After confirming the stationarity of the variables, the autoregressive distributed
lag (ARDL) representation of our models is given by the following equations:
p q r 2
ln Ct = β0 + ∑β1i ln C t −i + ∑β2i ln Yt −i + ∑β3i (ln Yt −1 )
i =1 i =0 i =0
(4)
s 3 t
+ ∑β4i (ln Yt −1 ) + ∑β5i ln Pt −i + εit
i =0 i =0
p q r 2
ln Bt = γ 0 + ∑γ1i ln B t −i + ∑γ 2i ln Yt −i + ∑γ 3i (ln Yt −1 )
i =1 i =0 i =0
s t
(5)
3
+ ∑γ 4i (ln Yt −1 ) + ∑γ 5i ln Pt −i + εit
i =0 i =0
and the error-correction equations are given by
p q r 2
∆ ln Ct = β0 + ∑β1i ∆ ln C t −i + ∑β2i ∆ ln Yt −i + ∑β3i ∆ (ln Yt −1 )
i =1 i =0 i =0
(6)
s 3 t
+ ∑β4i ∆ (ln Yt −1 ) + ∑β5i ∆ ln Pt −i + π ECTt −1 + εt
i =0 i =0
p q r 2
∆ ln Bt = γ 0 + ∑γ1i ∆ ln B t −i + ∑γ 2i ∆ ln Yt −i + ∑γ 3i ∆ (ln Yt −1 )
i =1 i =0 i =0
s 3 t (7)
+ ∑γ 4i ∆ (ln Yt −1 ) + ∑γ 5i ∆ ln Pt −i + π ECTt −1 + εt
i =0 i =0
where ECTt–1 is the lagged error-correction term and π is the speed adjustment
parameter, which measures the speed of the adjustment of the endogenous vari-
able when there is a shock in the equilibrium. The coefficient of the lagged error-
correction term is expected to be negative and statistically significant. The optimal
lag order is determined based on the smallest Akaike Information Criterion (AIC)
and Schwarz’s Bayesian Criterion (SBC) values. When the AIC and SBC provide
different lag structures, we prefer to use the AIC to prevent our model from being
parsimonious.
Data used in our analysis include a balanced panel for 70 countries for 1961–
2010. The time span and selection of countries used were constrained by the avail-
ability of data. The volume of fish catches (C) and the estimated size of biomass
(B) are measured in metric tons. Per capita real GDP (Y) is measured in constant
2005 US dollars. Population density (P) is measured in people per square kilome-
ter of land area. Fish production, per capita real GDP and population density data
were obtained from the World Bank World Development Indicators of 2015. The
size of biomass was estimated from the Sea Around Us Project catch data (Pauly
and Zeller, 2015). These data measure the volume of catches for all purposes for
each respective country’s exclusive economic zone (EEZ) for domestic or foreign
fleets. Although our estimation may be less precise than that of the well-known
RAM legacy database, it has broader coverage, making it more reliable in terms
of reflecting the current state of global marine fisheries.
4 Results and discussion

Our evaluation begins with an examination of integration properties of the vari-
ables examined based on three types of panel unit root tests: IPS, ADF-Fisher,
and PP-Fisher. The lag lengths of the panel unit root tests are selected based on
the SBC value. The test results provided in Table 5.1 show that all of the variables
were confirmed as stationary in the first difference.
We continue our analysis by determining the optimal lag length to be used in
the ARDL model. Table 5.2 presents the top five models that minimize the AIC
and SBC values by setting the maximum lag order at 4. From Table 5A.1, we can
see that for both models, the AIC and SBC present different model specifications.
We prefer to use the lag structure recommended by the AIC to avoid oversimplify-
ing the model. Thus, we have ARDL (2, 1, 1, 1, 1) for the catch model and ARDL
(3, 1, 1, 1, 1) for the biomass model.
The results of the PMG estimations are provided in Table 5.3. From Table 5.3,
we can see that over the long term, the impacts of economic growth on catch
and biomass levels are significant. However, the estimated coefficients of the two
models have opposite signs, indicating contradictory effects of economic growth
on fish production and abundance. The positive and significant coefficient of the
cubic term of the catch model suggests that the relationship between income and
global levels of catch is best described by a flipped N-shaped curve. Meanwhile,
Table 5.1 Panel unit root tests
Variables IPS ADF-Fisher PP-Fisher
Individual Individual Individual Individual Individual Individual

intercept intercept intercept intercept intercept intercept
and trend and trend and trend
Levels
ln Y −1.57280c −0.85322 232.942a 174.661b 231.278a 90.3782
ln P −1.14447 −0.44002 200.793a 233.366a 745.435a 201.230a
ln B 10.5027 4.67867 93.0637 145.118 105.696 30.4934
ln C −1.81382b 1.72201 192.504a 127.999 284.791a 120.753
First
differences
ln Y −29.9222a −28.3416a 1144.49a 977.141a 1284.04a 1209.44a
ln P −5.94732a −8.69802a 294.419a 357.171a 246.980a 210.331a
ln B −4.23248a −6.63326a 259.596a 294.584a 272.804a 279.303a
ln C −43.1150a −40.4610a 1724.29a 1544.23a 2112.53a 3167.56a
Notes: a, b and c denote statistical significance at 1%, 5% and 10% levels, respectively.
Table 5.2 Model selection summary
Catch model Biomass model
AIC SBC AIC SBC
Value ARDL Value ARDL Value ARDL Value ARDL

−1.203780 2, 1, 1, 1, 1 −0.269545 1, 1, 1, 1, 1 −6.983085 3, 1, 1, 1, 1 −5.916566 2, 1, 1, 1, 1
−1.201876 1, 1, 1, 1, 1 −0.139337 2, 1, 1, 1, 1 −6.982768 4, 1, 1, 1, 1 −5.786531 3, 1, 1, 1, 1
−1.189445 3, 1, 1, 1, 1 0.007109 3, 1, 1, 1, 1 −6.981008 2, 1, 1, 1, 1 −5.654102 4, 1, 1, 1, 1
−1.178145 4, 1, 1, 1, 1 0.150521 4, 1, 1, 1, 1 −6.969135 4, 4, 4, 4, 4 −5.539182 1, 1, 1, 1, 1
−1.148131 2, 2, 2, 2, 2 0.316540 1, 2, 2, 2, 2 −6.964262 4, 3, 3, 3, 3 −5.334852 2, 2, 2, 2, 2
Table 5.3 Long- and short-run estimates of the PMG
Variables Catch model: Biomass model:

ARDL (2,1,1,1,1) ARDL (3,1,1,1,1)
Long-run equation
ln Y −12.15962 (2.098682)a 4.427423 (1.070006)a
ln Y 2 1.818432 (0.278766)a −0.594963 (0.139026)a
ln Y 3 −0.087393 (0.012222)a 0.026085 (0.005974)a
ln P 0.935060 (0.347013)a −0.354170 (0.102098)a
Short-run equation
Δln Bt-1 – 0.591931 (0.029401)a
Δln Bt-2 – 0.033832 (0.025134)
Δln Ct-1 0.008472 (0.024725) –
Δln Y −244.7927 (110.4849)b 9.874210 (9.328565)
Δln Y 2 30.88965 (15.25024)b −1.116540 (1.124477)
Δln Y 3 −1.337844 (0.737349)c 0.043495 (0.046442)
Δln P −0.145190 (1.657649) 0.121631 (0.212329)
ECTt–1 −0.238835 (0.017113)a −0.045471 (0.006564)a
trend 0.002546 (0.001120)b −0.000561(0.000122)a
cons 8.231275 (0.596153)a 0.241409 (0.035241)a
Number of countries 70 70
Number of obs. 3360 3290
Log likelihood 2604.376 12147.41
SE of regression 0.496437 0.045452
Notes:
1. a, b and c denote statistical significance at 1%, 5% and 10% levels, respectively.
2. The numbers in parentheses are standard errors.
the opposite sign of the cubic term in the biomass model suggests the presence of
an N-shaped curve. From Table 5.3, we can also see that population growth is a
significant predictor of our models, placing continuous pressure on the environ-
ment either by inducing higher catch levels or by deteriorating stock volumes.
The catch model depicts a flipped N-shaped curve with an initial turning point
as a local minimum occurring at an income level of USD 276 per capita and
with the second turning point as a local maximum occurring at an income level
of USD 3827 per capita. Our findings suggest that in early stages of economic
development, higher income levels lead to decreasing catch levels. During this
stage, rather than being driven by economic growth, increasing catch levels are
mainly caused by population growth. At this stage of economic development, the
fisheries sector is dominated by traditional small-scale fisheries. However, after
reaching the first turning point, increasing levels of income and population growth
lead to higher catch levels, placing more pressure on the environment. This stage
of economic development illustrates the scale and technological effects of global
marine fisheries, which are marked by the rapid development of industrial-scale
fisheries and by advances in technology. This industrialization process has led
to the perceptible environmental deterioration of global fisheries (e.g., growing
numbers of overfished or collapsed stocks and declining mean trophic catch lev-
els). One the second turning point is reached, the trend reverses. While population
growth places continuous pressure on catch levels, further economic growth leads
to decreasing catch levels. At this stage of economic development, composition
effects of the economy result in the creation of new environmental regulations and
cleaner industries that preserve the environment and that undo damages of previ-
ous stages of development. However, our catch model does not support the con-
ventional EKC hypothesis, as the flipped N-shaped curve suggests the existence
of a secondary turning point beyond which environmental benefits of economic
growth will be achieved.
For the biomass model, the first turning point, which is a local maximum,
is observed at an income level of USD 661 per capita, and the second turning
point, which is a local minimum, is observed at an income level of USD 6066
per capita. Our model implies that initially, the exploitation of fish will lead to
the development of stock, which conforms to Schaefer’s (1954) production func-
tion model. However, beyond the primary turning point, further economic growth
leads to stock decline due to the overexploitation of fish above its MSY. This trend
reverses again after per capita income levels exceed the secondary turning point,
suggesting beneficial impacts of economic growth on resource abundance.
For the short term, we find significant impacts of economic development on
short-run variations at the catch level. However, its impacts on biomass levels are
not significant. We also find no significant impacts of population growth on catch
and biomass levels for the short term. Furthermore, the lagged error-correction
terms (ECTt–1) for both of our models are negative and statistically significant,
confirming the presence of cointegration between variables. These coefficients
measure the speed of endogenous variable adjustment when there is a shock in the
equilibrium. For the catch model, the absolute value of the lagged error-correction
term is 0.238835, indicating a relatively high rate of adjustment in the presence
of any shock to the equilibrium. A deviation from equilibrium catch levels in the
current period will be corrected with 23.88% in the next period. On the other
hand, the absolute value of the lagged error-correction term of the biomass model
is only 0.045471, which is fairly low. In the presence of any shock to the equi-
librium, the volume of biomass will be corrected by only approximately 4% in
the next period. Our findings imply that while the impacts of scale effects of the
economy are perceivable over the short term, beneficial impacts of composition
effects of the economy on stock recovery can only be achieved over the long term.
Both of our models suggest that declines in resource abundance are an inev-
itable consequence of fisheries sector development. However, as the economy
grows, the beneficial impacts of economic growth on resource abundance will be
attained. This results from the adoption of more stringent environmental regula-
tions, from the implementation of better fisheries management systems and from
the use of more advanced technologies. Such processes will spur a decline in
80 2500
Millions
Landings (Metric tons) Millions
70
2000
60
Stock (metric tons)

50 1500
40
30 1000
20
500
10
0 0
1960 1970 1980 1990 2000 2010 2020 2030
Year
Actual Landings Projected Landings Projected Stock Estimated Stock
Figure 5.3 Projection of total landing and stock values for 70 fishing countries
catch levels over the short term and stock recovery over the long term. Our find-
ings support Hilborn’s (2007) argument that declines in abundance should not
be considered a serious problem, as they merely serve as a means of achieving
sustainable yields.
Based on PMG estimates, we obtain a 20-year forecast from our models. For
this purpose, we use the world population prospect of the United Nations to obtain
the projected global population of 2030. We also assume that the global economy
grows at a constant rate of 2.6% per annum. The forecasts of our models are
shown in Figure 5.3. From Figure 5.3, we can see that after reaching its peak
in 1996, global catch is predicted to decline until 2030. In 2030, the volume of
global catch is expected to decrease by 2.8% from the 2010 level. Similar trends
are observed for the biomass model. However, the trend reverses in 2027. In 2030,
we expect to see improvements to global marine fish stocks, although the pre-
dicted biomass value should still exist below the 2010 level.
A more detailed analysis of the top fishing countries examined (see Fig-
ure 5.4) shows that rich countries such as Japan, the United Kingdom and the
United States contribute positively to declining global catch levels, which in
turn prevent the stock from deteriorating further. This highlights the benefi-
cial impacts of better fisheries management systems used in these countries.
Interesting findings were found in the case of Malaysia. Unlike those of other
middle-income countries, Malaysia’s total catch is expected to peak in the near
future. However, such declining catch levels are not immediately followed by
stock recovery. For other developing countries such as China and Indonesia, we
expect to see an increase in catch levels over the next two decades, leading to a
steady decline in stock levels.
Figure 5.4 Projection of landings and stocks for the examined countries
5 Conclusion
The objective of this study was to estimate the state of global marine fisheries
and to study its relationship with economic factors. For this purpose, we used
both catch levels and the estimated stock of fish as proxies for marine resource
abundance. Our models employed panel datasets on 70 fishing countries for
1961–2010.
We found no evidence of the EKC hypothesis for global marine fisheries from
catch and biomass stock models. However, our models show that the beneficial
impacts of economic growth on global marine fisheries are likely to be achieved.
Our catch model reveals the occurrence of a secondary turning point at an income
level of USD 3827 per capita after which further economic growth will lead to a
decline in catch levels. In addition, our biomass model presents a secondary turn-
ing point occurring at an income level of USD 6066 per capita after which further
economic growth will lead to stock improvements. We also found that population
density places constant pressure on resource use by increasing catch levels or
reducing stock sizes.
Our models forecast that over the next two decades, global catch levels should
decline alongside economic and population growth. We also expect to find a
slight decline in stock levels followed indications of stock recovery. However,
our models do not dismiss the need for more stringent environmental regulations
and for the use of better fisheries management practices. The higher second-
ary turning point and the small value of the lagged error-correction term of the
biomass model suggest that current quota-based management approaches that
attempt to limit catch values might help mitigate pressures on the environment
while preventing stock depletion. However, stock recovery is unlikely to be
observed over the short term.
Acknowledgments
Shunsuke Managi was supported by the following Grant in Aid from the Ministry
of Education, Culture, Sports, Science and Technology in Japan (MEXT): Grant
in Aid for Specially Promoted Research [grant number 26000001]. Yogi Sugiawan
was supported by Research and Innovation in Science and Technology Project
(RISET-PRO), Ministry of Research, Technology and Higher Education of Indo-
nesia. Any opinions, findings and conclusions expressed in this material are those
of the authors and do not necessarily reflect the views of the funding agencies.
Note
1 The functional form of the income-resource relationship is determined by the significance
of the coefficients βi and γi. A level-type relationship occurs when β1 = β2 = β3 = 0 or γ1 =
γ2 = γ3 = 0, suggesting that there is no relationship between economic growth and resource
abundance. Meanwhile, a linear-type relationship exists when β2 = β3 = 0 and β1 ≠ 0; or γ2 =
γ3 = 0 and γ1 ≠ 0. Non-linear relationship between economic growth and resource abun-
dance exists when β2 and/or β3 or when γ2 and/or γ3 are significantly different from zero.
Bibliography
Acheson, J., Apollonio, S., Wilson, J., 2015. Individual transferable quotas and conserva-
tion: A critical assessment. Ecology and Society 20.
Agnew, D. J., Gutiérrez, N. L., Butterworth, D. S., 2013. Fish catch data: Less than what
meets the eye. Marine Policy 42, 268–269.
Al-mulali, U., Weng-Wai, C., Sheau-Ting, L., Mohammed, A. H., 2015. Investigating the
environmental Kuznets curve (EKC) hypothesis by utilizing the ecological footprint as
an indicator of environmental degradation. Ecological Indicators 48, 315–323.
Bölük, G., Mert, M., 2015. The renewable energy, growth and environmental Kuznets curve
in Turkey: An ARDL approach. Renewable and Sustainable Energy Reviews 52, 587–595.
Branch, T. A., Jensen, O. P., Ricard, D., Ye, Y., Hilborn, R., 2011. Contrasting global trends
in marine fishery status obtained from catches and from stock assessments. Conservation
Biology 25, 777–786.
Caviglia-Harris, J. L., Chambers, D., Kahn, J. R., 2009. Taking the “U” out of Kuznets.
Ecological Economics 68, 1149–1159.
Costello, C., Gaines, S. D., Lynham, J., 2008. Can catch shares prevent fisheries collapse?
Science 321, 1678–1681.
De Bruyn, S. M., van den Bergh, J. C., Opschoor, J. B., 1998. Economic growth and emis-
sions: Reconsidering the empirical basis of environmental Kuznets curves. Ecological
Economics 25, 161–175.
Ewers, R. M., 2006. Interaction effects between economic development and forest cover
determine deforestation rates. Global Environmental Change 16, 161–169.
Froese, R., Kesner-Reyes, K., 2002. Impact of fishing on the abundance of marine species.
ICES Council Meeting Report CM.
Froese, R., Zeller, D., Kleisner, K., Pauly, D., 2012. What catch data can tell us about the
status of global fisheries. Marine Biology 159, 1283–1292.
Gephart, J. A., Deutsch, L., Pace, M. L., Troell, M., Seekell, D. A., 2017. Shocks to fish produc-
tion: Identification, trends, and consequences. Global Environmental Change 42, 24–32.
Grossman, G. M., Krueger, A. B., 1991. Environmental impacts of a North American free
trade agreement. National Bureau of Economic Research.
Halkos, G., Managi, S., Tsilika, K., 2017. Evaluating a continent-wise situation for capital
data. Economic Analysis and Policy 55, 57–74.
Halpern, B. S., Longo, C., Hardy, D., McLeod, K. L., Samhouri, J. F., Katona, S. K., Kleis-
ner, K., Lester, S. E., O’Leary, J., Ranelletti, M., Rosenberg, A. A., Scarborough, C.,
Selig, E. R., Best, B. D., Brumbaugh, D. R., Chapin, F. S., Crowder, L. B., Daly, K. L.,
Doney, S. C., Elfes, C., Fogarty, M. J., Gaines, S. D., Jacobsen, K. I., Karrer, L. B., Les-
lie, H. M., Neeley, E., Pauly, D., Polasky, S., Ris, B., St Martin, K., Stone, G. S., Sumaila,
U. R., Zeller, D., 2012. An index to assess the health and benefits of the global ocean.
Nature 488, 615–620.
Hilborn, R., 2007. Reinterpreting the state of fisheries and their management. Ecosystems
10, 1362–1369.
Hutchings, J. A., 2000. Collapse and recovery of marine fishes. Nature 406, 882–885.
Jackson, J. B., Kirby, M. X., Berger, W. H., Bjorndal, K. A., Botsford, L. W., Bourque, B.
J., Bradbury, R. H., Cooke, R., Erlandson, J., Estes, J. A., Hughes, T. P., Kidwell, S.,
Lange, C. B., Lenihan, H. S., Pandolfi, J. M., Peterson, C. H., Steneck, R. S., Tegner, M.
J., Warner, R. R., 2001. Historical overfishing and the recent collapse of coastal ecosys-
tems. Science 293, 629–637.
Jalil, A., Mahmud, S. F., 2009. Environment Kuznets curve for CO2 emissions: A cointe-
gration analysis for China. Energy Policy 37, 5167–5172.
Kleisner, K., Zeller, D., Froese, R., Pauly, D., 2013. Using global catch data for inferences
on the world’s marine fisheries. Fish and Fisheries 14, 293–311.
Komen, M. H., Gerking, S., Folmer, H., 1997. Income and environmental R&D: Empirical
evidence from OECD countries. Environment and Development Economics 2,
505–515.
Levin, A., Lin, C.-F., Chu, C.-S. J., 2002. Unit root tests in panel data: Asymptotic and
finite-sample properties. Journal of Econometrics 108, 1–24.
Liao, H., Cao, H.-S., 2013. How does carbon dioxide emission change with the economic
development? Statistical experiences from 132 countries. Global Environmental Change
23, 1073–1082.
List, J. A., Gallet, C. A., 1999. The environmental Kuznets curve: Does one size fit all?
Managi, S., Hibiki, A., Tsurumi, T., 2009. Does trade openness improve environmental
quality? Journal of Environmental Economics and Management 58, 346–363.
Martell, S., Froese, R., 2013. A simple method for estimating MSY from catch and resil-
ience. Fish and Fisheries 14, 504–514.
Merino, G., Barange, M., Blanchard, J. L., Harle, J., Holmes, R., Allen, I., Allison, E. H.,
Badjeck, M. C., Dulvy, N. K., Holt, J., Jennings, S., Mullon, C., Rodwell, L. D., 2012.
Can marine fisheries and aquaculture meet fish demand from a growing human popula-
tion in a changing climate? Global Environmental Change 22, 795–806.
Murawski, S., Methot, R., Tromble, G., 2007. Biodiversity loss in the ocean: How bad is
it? Science 316, 1281–1284.
Myers, R. A., Worm, B., 2003. Rapid worldwide depletion of predatory fish communities.
Nature 423, 280–283.
Nguyen Van, P., Azomahou, T., 2007. Nonlinearities and heterogeneity in environmental
quality: An empirical analysis of deforestation. Journal of Development Economics 84,
291–309.
Panayotou, T., 1993. Empirical tests and policy analysis of environmental degradation at
different stages of economic development. International Labour Organization.
Pauly, D., Alder, J., Booth, S., Cheung, W., Christensen, V., Close, C., Sumaila, U., Swartz,
W., Tavakolie, A., Watson, R., 2008. Fisheries in large marine ecosystems: Descriptions
and diagnoses: The UNEP large marine ecosystem report: A perspective on changing
conditions in LMEs of the World’s Regional Seas. UNEP Regional Seas Reports and
Studies 23–40.
Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., Torres, F., 1998. Fishing down marine
food webs. Science 279, 860–863.
Pauly, D., Christensen, V., Guénette, S., Pitcher, T. J., Sumaila, U. R., Walters, C. J., Wat-
son, R., Zeller, D., 2002. Towards sustainability in world fisheries. Nature 418,
689–695.
Pauly, D., Hilborn, R., Branch, T. A., 2013. Fisheries: Does catch reflect abundance? Nature
494, 303–306.
Pauly, D., Palomares, M.-L., 2005. Fishing down marine food web: It is far more pervasive
than we thought. Bulletin of Marine Science 76, 197–212.
Pauly, D., Zeller, D., 2015. Sea around us concepts, design and data. (www.seaaroundus.org).
Pesaran, M. H., Shin, Y., Smith, R. P., 1999. Pooled mean group estimation of dynamic
heterogeneous panels. Journal of the American Statistical Association 94, 621–634.
Pontecorvo, G., Schrank, W. E., 2012. The expansion, limit and decline of the global marine
fish catch. Marine Policy 36, 1178–1181.
Ricard, D., Minto, C., Jensen, O. P., Baum, J. K., 2012. Examining the knowledge base and
status of commercially exploited marine species with the RAM Legacy Stock Assess-
ment Database. Fish and Fisheries 13, 380–398.
Schaefer, M. B., 1954. Some aspects of the dynamics of populations important to the man-
agement of the commercial marine fisheries. Inter-American Tropical Tuna Commission
Bulletin 1, 23–56.
Sinha, A., Bhattacharya, J., 2017. Estimation of environmental Kuznets curve for SO2
emission: A case of Indian cities. Ecological Indicators 72, 881–894.
Soliman, A., 2014. Individual transferable quotas in world fisheries: Addressing legal and
rights-based issues. Ocean & Coastal Management 87, 102–113.
Srinivasan, U. T., Cheung, W. W., Watson, R., Sumaila, U. R., 2010. Food security implica-
tions of global marine catch losses due to overfishing. Journal of Bioeconomics 12,
183–200.
Sugiawan, Y., Managi, S., 2016. The environmental Kuznets curve in Indonesia: Exploring
the potential of renewable energy. Energy Policy 98, 187–198.
Tamaki, T., Shin, K. J., Nakamura, H., Fujii, H., Managi, S., 2017. Shadow prices and
production inefficiency of mineral resources. Economic Analysis and Policy
(forthcoming).
Tsurumi, T., Managi, S., 2010. Decomposition of the environmental Kuznets curve: Scale,
technique, and composition effects. Environmental Economics and Policy Studies 11,
19–36.
Worm, B., Barbier, E. B., Beaumont, N., Duffy, J. E., Folke, C., Halpern, B. S., Jackson, J.
B., Lotze, H. K., Micheli, F., Palumbi, S. R., 2006. Impacts of biodiversity loss on ocean
ecosystem services. Science 314, 787–790.
Worm, B., Barbier, E. B., Beaumont, N., Duffy, J. E., Folke, C., Halpern, B. S., Jackson, J.
B., Lotze, H. K., Micheli, F., Palumbi, S. R., 2007. Response to comments on “impacts
of biodiversity loss on ocean ecosystem services”. Science 316, 1285d–1285d.
Zeller, D., Cheung, W., Close, C., Pauly, D., 2009. Trends in global marine fisheries: A
critical view: Fisheries, trade and development. Royal Swedish Academy of Agriculture
and Forestry, Stockholm 87–107.
Zeller, D., Pauly, D., 2005. Good news, bad news: Global fisheries discards are declining,
but so are total catches. Fish and Fisheries 6, 156–159.
Appendix A
Estimated stock
Table 5A.1 Estimated biomass

( x 106 metric tons)
No Country No. of Year

species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
1 Albania 9 0.20 0.20 0.19 0.19 0.19 0.18 0.18 0.18 0.17 0.17 0.17 0.17 0.17 0.17 0.17
2 Algeria 185 5.12 4.53 4.18 3.94 3.78 3.66 3.57 3.50 3.44 3.40 3.37 3.33 3.30 3.29 3.28
3 American Samoa 96 0.09 0.09 0.08 0.08 0.08 0.08 0.08 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07
4 Angola 80 13.18 12.17 11.50 11.06 10.70 10.41 10.11 9.65 9.26 9.04 8.84 8.69 8.54 8.38 8.25
5 Antigua and Barbuda 32 0.10 0.09 0.08 0.08 0.07 0.07 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06
6 Argentina 61 34.28 32.29 30.93 29.94 29.18 28.58 28.10 27.72 27.40 27.13 26.90 26.70 26.54 26.39 26.23
7 Aruba 76 0.08 0.08 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.06 0.06
8 Australia 271 10.82 10.36 10.02 9.75 9.52 9.32 9.17 9.03 8.90 8.79 8.68 8.58 8.49 8.40 8.30
9 Bahamas, The 45 0.68 0.63 0.60 0.58 0.56 0.55 0.54 0.53 0.52 0.51 0.50 0.50 0.49 0.48 0.48
10 Bahrain 36 0.85 0.73 0.67 0.62 0.59 0.57 0.55 0.53 0.52 0.51 0.50 0.49 0.48 0.48 0.47
11 Bangladesh 104 18.34 16.32 15.09 14.24 13.63 13.17 12.81 12.52 12.29 12.10 11.94 11.80 11.67 11.56 11.45
12 Barbados 44 0.23 0.21 0.20 0.19 0.18 0.17 0.17 0.16 0.16 0.15 0.15 0.14 0.14 0.14 0.13
13 Belgium 75 0.53 0.51 0.49 0.46 0.43 0.40 0.37 0.34 0.34 0.33 0.32 0.32 0.31 0.31 0.31
14 Belize 57 0.22 0.21 0.20 0.19 0.18 0.18 0.17 0.17 0.17 0.16 0.16 0.16 0.16 0.15 0.15
15 Benin 89 1.69 1.58 1.51 1.46 1.41 1.38 1.35 1.33 1.31 1.30 1.29 1.27 1.26 1.24 1.23
16 Bermuda 56 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
17 Bosnia and Herzegovina 32 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
18 Brazil 260 30.21 28.00 26.61 25.60 24.87 24.27 23.77 23.35 23.01 22.70 22.37 22.04 21.73 21.28 20.84
19 Brunei Darussalam 54 0.32 0.28 0.26 0.25 0.24 0.24 0.23 0.23 0.22 0.22 0.22 0.22 0.21 0.21 0.21
20 Bulgaria 15 0.43 0.41 0.39 0.37 0.36 0.34 0.33 0.33 0.32 0.31 0.31 0.30 0.29 0.29 0.28
21 Cabo Verde 39 0.54 0.50 0.47 0.45 0.44 0.42 0.41 0.40 0.39 0.39 0.38 0.38 0.37 0.37 0.36
22 Cambodia 105 31.72 30.13 28.98 28.12 27.44 26.91 26.48 26.13 25.84 25.61 25.41 25.25 25.10 24.98 24.88
23 Cameroon 60 2.48 2.32 2.21 2.13 2.07 2.02 1.98 1.94 1.90 1.87 1.85 1.83 1.82 1.80 1.78
(Continued)
Table 5A.1 (Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
1 Albania 9 0.17 0.16 0.16 0.16 0.16 0.15 0.15 0.14 0.14 0.14 0.13 0.13 0.12 0.12 0.12
2 Algeria 185 3.28 3.27 3.27 3.26 3.26 3.25 3.25 3.25 3.24 3.23 3.21 3.19 3.18 3.16 3.15
3 American Samoa 96 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.06 0.07 0.06 0.06 0.06 0.06
4 Angola 80 8.00 7.86 7.65 7.47 7.29 6.99 6.77 6.60 6.14 5.82 5.55 5.57 5.66 5.73 5.79
6 Argentina 61 26.04 25.84 25.61 25.43 25.14 24.95 24.84 24.72 24.60 24.38 24.18 24.13 23.97 23.64 23.07
7 Aruba 76 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
8 Australia 271 8.22 8.13 8.04 7.96 7.88 7.80 7.70 7.57 7.45 7.33 7.17 7.09 6.98 6.85 6.72
9 Bahamas, The 45 0.47 0.46 0.45 0.44 0.44 0.43 0.42 0.41 0.40 0.39 0.39 0.38 0.37 0.37 0.36
10 Bahrain 36 0.46 0.45 0.44 0.44 0.43 0.43 0.43 0.43 0.43 0.43 0.43 0.43 0.43 0.43 0.43
11 Bangladesh 104 11.36 11.28 11.21 11.14 11.07 11.01 10.95 10.89 10.82 10.76 10.68 10.62 10.56 10.49 10.42
12 Barbados 44 0.13 0.13 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11 0.11 0.10 0.10 0.10
13 Belgium 75 0.30 0.29 0.29 0.29 0.29 0.29 0.28 0.28 0.28 0.27 0.27 0.26 0.25 0.25 0.26
14 Belize 57 0.15 0.15 0.15 0.14 0.14 0.14 0.14 0.14 0.13 0.13 0.13 0.13 0.13 0.13 0.13
15 Benin 89 1.21 1.19 1.16 1.16 1.15 1.13 1.13 1.13 1.12 1.12 1.12 1.12 1.12 1.12 1.11
16 Bermuda 56 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
18 Brazil 260 20.52 20.23 19.90 19.50 19.17 18.89 18.57 18.20 17.74 17.23 16.83 16.49 16.28 15.92 15.47
20 Bulgaria 15 0.28 0.28 0.27 0.27 0.27 0.27 0.26 0.26 0.26 0.26 0.25 0.25 0.24 0.23 0.23
21 Cabo Verde 39 0.36 0.35 0.34 0.34 0.33 0.32 0.32 0.31 0.31 0.31 0.31 0.30 0.30 0.29 0.29
22 Cambodia 105 24.80 24.68 24.55 24.42 24.25 24.06 23.84 23.61 23.36 23.12 22.91 22.72 22.51 22.15 21.81
23 Cameroon 60 1.77 1.76 1.75 1.75 1.74 1.73 1.71 1.68 1.65 1.60 1.56 1.52 1.48 1.44 1.40
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
1 Albania 9 0.11 0.11 0.11 0.10 0.10 0.10 0.09 0.09 0.08 0.07 0.07 0.06 0.06 0.06 0.06
2 Algeria 185 3.13 3.11 3.08 3.04 3.00 2.97 2.92 2.89 2.83 2.76 2.70 2.66 2.64 2.59 2.54
3 American Samoa 96 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07
4 Angola 80 5.86 5.97 6.01 6.08 6.16 6.28 6.36 6.45 6.56 6.60 6.63 6.59 6.59 6.60 6.62
6 Argentina 61 22.48 22.28 22.15 21.84 21.61 21.55 21.37 21.18 20.89 20.50 20.11 19.65 19.05 18.32 17.38
7 Aruba 76 0.06 0.06 0.06 0.05 0.05 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.04
8 Australia 271 6.59 6.50 6.34 6.18 6.02 5.88 5.72 5.58 5.40 5.19 5.00 4.77 4.50 4.32 4.14
9 Bahamas, The 45 0.35 0.34 0.34 0.33 0.32 0.31 0.29 0.29 0.29 0.28 0.28 0.28 0.27 0.27 0.26
10 Bahrain 36 0.43 0.43 0.42 0.42 0.42 0.41 0.40 0.40 0.39 0.39 0.38 0.38 0.38 0.37 0.37
11 Bangladesh 104 10.35 10.28 10.22 10.15 10.08 10.01 9.92 9.82 9.70 9.59 9.48 9.37 9.26 9.14 9.01
12 Barbados 44 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.09 0.09
13 Belgium 75 0.26 0.26 0.25 0.24 0.24 0.24 0.24 0.25 0.25 0.26 0.26 0.27 0.28 0.28 0.29
14 Belize 57 0.13 0.13 0.12 0.12 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11 0.11 0.11
15 Benin 89 1.11 1.10 1.10 1.09 1.07 1.06 1.05 1.03 1.01 0.98 0.95 0.91 0.87 0.84 0.81
16 Bermuda 56 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
18 Brazil 260 14.99 14.68 14.39 14.05 13.62 13.09 12.58 12.12 11.64 11.30 11.00 10.95 10.86 10.79 10.81
20 Bulgaria 15 0.22 0.20 0.19 0.18 0.17 0.17 0.16 0.15 0.15 0.15 0.15 0.16 0.16 0.17 0.18
21 Cabo Verde 39 0.29 0.29 0.28 0.28 0.27 0.27 0.26 0.26 0.26 0.26 0.25 0.25 0.25 0.25 0.25
22 Cambodia 105 21.53 21.28 20.94 20.51 20.03 19.60 19.10 18.53 17.87 17.36 16.91 16.47 16.07 15.56 15.09
23 Cameroon 60 1.37 1.34 1.32 1.30 1.29 1.27 1.26 1.25 1.24 1.23 1.21 1.18 1.14 1.12 1.09
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
1 Albania 9 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
2 Algeria 185 2.45 2.41 2.40 2.38 2.36 2.33 2.28 2.21 2.14 2.06 2.01 1.94 1.84 1.75 1.66 1.59
3 American Samoa 96 0.06 0.06 0.06 0.06 0.06 0.05 0.06 0.05 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03
4 Angola 80 6.59 6.57 6.59 6.55 6.48 6.40 6.29 6.19 6.18 6.18 6.20 6.17 6.12 6.05 6.00 5.95
5 Antigua and Barbuda 32 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
6 Argentina 61 16.69 15.77 14.81 13.73 12.99 12.39 12.02 11.62 11.24 10.91 10.56 10.26 9.61 8.98 8.16 7.69
7 Aruba 76 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02
8 Australia 271 4.02 3.88 3.76 3.62 3.49 3.36 3.26 3.14 3.08 2.99 2.91 2.81 2.71 2.70 2.69 2.68
9 Bahamas, The 45 0.26 0.26 0.25 0.25 0.24 0.24 0.23 0.23 0.23 0.22 0.22 0.21 0.21 0.21 0.21 0.21
10 Bahrain 36 0.36 0.35 0.34 0.33 0.33 0.32 0.32 0.32 0.32 0.32 0.31 0.31 0.31 0.31 0.30 0.29
11 Bangladesh 104 8.89 8.76 8.63 8.51 8.39 8.26 8.15 8.02 7.90 7.77 7.57 7.40 7.25 7.11 6.97 6.78
12 Barbados 44 0.09 0.09 0.09 0.10 0.09 0.10 0.09 0.10 0.09 0.09 0.10 0.10 0.10 0.10 0.10 0.10
13 Belgium 75 0.29 0.30 0.31 0.32 0.33 0.33 0.34 0.35 0.36 0.36 0.37 0.38 0.39 0.39 0.40 0.41
14 Belize 57 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11
15 Benin 89 0.78 0.76 0.73 0.71 0.69 0.67 0.64 0.62 0.59 0.56 0.53 0.49 0.47 0.45 0.44 0.42
16 Bermuda 56 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
17 Bosnia and Herzegovina 32 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
18 Brazil 260 10.79 10.80 10.83 10.77 10.75 10.77 10.75 10.64 10.53 10.49 10.43 10.37 10.29 10.18 10.07 9.93
19 Brunei Darussalam 54 0.17 0.17 0.16 0.16 0.16 0.16 0.16 0.15 0.15 0.14 0.14 0.13 0.13 0.12 0.12 0.11
20 Bulgaria 15 0.18 0.19 0.19 0.19 0.20 0.20 0.20 0.21 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20
21 Cabo Verde 39 0.24 0.23 0.23 0.22 0.21 0.20 0.19 0.18 0.18 0.18 0.17 0.15 0.15 0.14 0.14 0.13
22 Cambodia 105 14.57 14.04 13.39 12.73 11.98 11.21 10.44 9.79 9.09 8.39 7.63 6.85 6.17 5.59 5.46 5.31
23 Cameroon 60 1.06 1.02 0.99 0.96 0.92 0.88 0.84 0.80 0.77 0.73 0.69 0.67 0.66 0.65 0.65 0.64
species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
24 Canada 172 199.38 192.85 187.49 182.05 177.04 171.55 166.47 161.47 156.70 151.84 146.67 141.14 134.88 129.00 122.72
25 Cayman Islands 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
26 Chile 86 112.66 109.01 106.46 104.51 102.98 101.75 100.69 99.84 99.13 98.54 98.01 97.37 96.74 96.22 94.76
27 China 185 219.92 203.34 192.49 184.67 178.89 174.18 170.35 166.92 164.20 161.89 159.88 158.17 156.71 155.36 154.20
28 Colombia 74 3.23 3.17 3.13 3.09 3.05 3.02 3.00 2.98 2.95 2.92 2.88 2.84 2.78 2.69 2.57
29 Comoros 21 0.43 0.39 0.37 0.35 0.34 0.33 0.33 0.32 0.32 0.31 0.31 0.31 0.30 0.30 0.30
30 Congo, Dem. Rep. 41 0.50 0.46 0.43 0.41 0.40 0.39 0.38 0.37 0.37 0.36 0.35 0.35 0.34 0.34 0.33
31 Congo, Rep. 56 2.02 1.96 1.92 1.88 1.85 1.83 1.80 1.78 1.76 1.73 1.70 1.68 1.66 1.63 1.61
32 Costa Rica 81 2.68 2.60 2.55 2.50 2.47 2.44 2.42 2.40 2.39 2.37 2.35 2.32 2.29 2.26 2.23
33 Côte d’Ivoire 139 3.24 3.13 3.05 2.99 2.94 2.89 2.85 2.81 2.77 2.74 2.70 2.66 2.61 2.57 2.52
34 Croatia 57 1.83 1.72 1.64 1.58 1.52 1.48 1.45 1.41 1.38 1.36 1.34 1.32 1.30 1.29 1.27
35 Cuba 55 1.70 1.64 1.60 1.56 1.53 1.51 1.48 1.46 1.44 1.42 1.40 1.38 1.36 1.34 1.32
36 Curaçao 28 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
37 Denmark 135 44.06 42.98 42.06 41.27 40.60 39.82 39.13 38.53 37.91 37.30 36.75 36.26 35.72 35.18 34.58
38 Djibouti 10 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
39 Dominica 37 0.06 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04
40 Dominican Republic 92 1.74 1.53 1.40 1.31 1.24 1.19 1.15 1.12 1.09 1.07 1.05 1.03 1.02 1.00 0.99
41 Ecuador 69 20.00 19.49 19.16 18.93 18.75 18.61 18.50 18.40 18.32 18.26 18.19 18.14 18.09 18.04 17.96
42 Egypt, Arab Rep. 73 2.84 2.62 2.48 2.37 2.30 2.23 2.19 2.14 2.10 2.05 2.00 1.96 1.92 1.88 1.83
43 El Salvador 62 2.31 2.27 2.24 2.22 2.20 2.18 2.17 2.16 2.13 2.10 2.07 1.98 1.88 1.80 1.73
44 Equatorial Guinea 80 1.34 1.28 1.24 1.21 1.19 1.17 1.15 1.14 1.12 1.11 1.10 1.09 1.08 1.08 1.07
45 Eritrea 52 0.53 0.50 0.48 0.46 0.44 0.42 0.40 0.38 0.37 0.36 0.36 0.36 0.35 0.34 0.34
46 Estonia 47 2.57 2.32 2.16 2.04 1.95 1.88 1.82 1.77 1.73 1.70 1.67 1.65 1.63 1.61 1.59
(Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
24 Canada 172 116.44 109.80 103.67 97.65 91.11 85.79 81.45 77.34 73.62 69.97 67.42 65.50 64.43 65.00 66.36
25 Cayman Islands 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
26 Chile 86 92.64 92.20 91.05 90.38 89.41 88.89 87.94 87.12 87.12 87.28 86.09 85.36 84.66 83.28 80.93
27 China 185 153.16 152.19 151.31 150.59 149.84 149.17 148.64 148.06 147.33 146.55 145.65 144.60 143.68 142.72 141.94
28 Colombia 74 2.52 2.47 2.39 2.29 2.22 2.20 2.13 2.09 2.00 1.84 1.75 1.65 1.65 1.60 1.61
29 Comoros 21 0.30 0.30 0.30 0.29 0.29 0.29 0.29 0.29 0.28 0.28 0.28 0.28 0.28 0.28 0.28
30 Congo, Dem. Rep. 41 0.33 0.32 0.32 0.31 0.31 0.30 0.29 0.29 0.28 0.27 0.26 0.25 0.24 0.24 0.23
31 Congo, Rep. 56 1.57 1.52 1.47 1.46 1.44 1.42 1.40 1.37 1.33 1.29 1.23 1.18 1.12 1.06 1.02
32 Costa Rica 81 2.18 2.14 2.11 2.08 2.03 2.01 1.98 1.91 1.84 1.74 1.64 1.62 1.60 1.60 1.59
33 Côte d’Ivoire 139 2.48 2.44 2.39 2.35 2.31 2.28 2.24 2.21 2.19 2.17 2.16 2.15 2.13 2.10 2.07
34 Croatia 57 1.25 1.24 1.22 1.21 1.19 1.18 1.18 1.16 1.15 1.14 1.13 1.12 1.11 1.10 1.08
35 Cuba 55 1.30 1.27 1.25 1.23 1.21 1.19 1.17 1.14 1.11 1.08 1.04 1.01 0.98 0.95 0.92
36 Curaçao 28 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
37 Denmark 135 33.91 33.33 32.81 32.25 31.57 30.95 30.39 29.80 29.27 28.68 28.09 27.34 26.55 25.94 25.33
38 Djibouti 10 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
39 Dominica 37 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
41 Ecuador 69 17.91 17.86 17.81 17.76 17.69 17.62 17.54 17.44 17.31 17.11 16.89 16.62 16.14 15.36 14.25
42 Egypt, Arab Rep. 73 1.78 1.75 1.74 1.73 1.72 1.72 1.72 1.71 1.70 1.71 1.70 1.69 1.68 1.66 1.65
43 El Salvador 62 1.65 1.58 1.50 1.45 1.38 1.34 1.30 1.25 1.22 1.15 1.10 1.07 1.04 1.04 0.98
45 Eritrea 52 0.34 0.32 0.31 0.30 0.29 0.28 0.27 0.25 0.25 0.25 0.25 0.26 0.26 0.27 0.28
46 Estonia 47 1.57 1.55 1.53 1.52 1.49 1.47 1.45 1.43 1.40 1.38 1.35 1.33 1.31 1.30 1.28
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
24 Canada 172 67.68 69.26 70.97 72.82 74.75 76.77 78.59 80.16 81.82 83.50 85.18 86.88 89.04 91.60 94.38
25 Cayman Islands 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
26 Chile 86 79.91 78.70 77.07 73.64 71.36 69.20 66.89 63.84 61.12 58.71 55.30 52.26 49.08 46.32 44.71
27 China 185 141.31 140.78 140.23 139.26 138.35 137.13 135.65 134.28 132.43 130.44 128.96 127.01 124.95 122.07 119.55
28 Colombia 74 1.63 1.63 1.61 1.59 1.60 1.57 1.48 1.46 1.44 1.40 1.37 1.32 1.25 1.12 1.09
29 Comoros 21 0.28 0.28 0.28 0.28 0.27 0.27 0.26 0.26 0.26 0.26 0.25 0.24 0.24 0.23 0.23
30 Congo, Dem. Rep. 41 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.23
31 Congo, Rep. 56 0.97 0.91 0.86 0.82 0.77 0.74 0.72 0.71 0.69 0.69 0.68 0.68 0.69 0.70 0.71
32 Costa Rica 81 1.57 1.55 1.51 1.48 1.48 1.44 1.34 1.23 1.17 1.13 1.09 1.07 1.07 1.07 1.05
33 Côte d’Ivoire 139 2.04 2.03 2.02 2.01 2.00 2.00 1.99 1.98 1.96 1.96 1.94 1.92 1.90 1.86 1.82
34 Croatia 57 1.07 1.05 1.03 1.00 0.97 0.94 0.91 0.88 0.84 0.81 0.79 0.77 0.77 0.76 0.74
35 Cuba 55 0.90 0.87 0.85 0.83 0.80 0.77 0.74 0.71 0.68 0.66 0.63 0.61 0.60 0.59 0.58
36 Curaçao 28 0.04 0.04 0.04 0.04 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
37 Denmark 135 24.81 24.16 23.51 22.91 22.39 21.81 21.31 20.71 20.37 19.66 19.02 18.71 18.19 17.47 17.12
38 Djibouti 10 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
39 Dominica 37 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
41 Ecuador 69 13.23 12.16 11.41 10.52 10.28 9.24 7.56 6.58 6.05 5.37 4.93 4.90 4.84 4.83 4.68
42 Egypt, Arab Rep. 73 1.62 1.60 1.59 1.58 1.57 1.55 1.53 1.52 1.51 1.48 1.45 1.41 1.38 1.36 1.31
43 El Salvador 62 0.95 0.96 0.93 0.91 0.91 0.85 0.84 0.83 0.82 0.81 0.81 0.82 0.83 0.85 0.85
45 Eritrea 52 0.29 0.29 0.30 0.31 0.32 0.33 0.33 0.34 0.34 0.35 0.36 0.36 0.37 0.38 0.39
46 Estonia 47 1.27 1.26 1.26 1.25 1.24 1.23 1.21 1.20 1.19 1.17 1.16 1.15 1.15 1.15 1.14
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
24 Canada 172 96.93 100.31 103.59 106.80 110.12 113.41 116.63 119.76 122.81 125.75 128.59 131.45 134.25 137.01 139.74 142.11
25 Cayman Islands 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
26 Chile 86 42.03 39.36 36.73 35.20 35.10 33.81 33.19 32.53 31.83 31.02 28.84 27.35 26.15 25.43 24.89 24.13
27 China 185 115.98 111.59 106.18 100.62 95.56 92.32 89.54 86.09 82.85 80.32 78.06 76.03 74.21 72.28 70.17 68.05
28 Colombia 74 1.09 1.08 1.05 1.03 1.01 1.03 1.02 1.02 1.03 1.03 1.05 1.07 1.09 1.08 1.09 1.10
29 Comoros 21 0.22 0.22 0.21 0.20 0.19 0.18 0.18 0.17 0.16 0.15 0.14 0.14 0.13 0.13 0.12 0.12
30 Congo, Dem. Rep. 41 0.23 0.23 0.23 0.23 0.22 0.22 0.22 0.22 0.22 0.21 0.21 0.21 0.21 0.21 0.21 0.21
31 Congo, Rep. 56 0.71 0.72 0.73 0.74 0.75 0.76 0.76 0.77 0.77 0.77 0.77 0.77 0.77 0.77 0.76 0.76
32 Costa Rica 81 1.02 1.00 0.97 0.96 0.95 0.94 0.92 0.92 0.88 0.84 0.82 0.80 0.78 0.78 0.79 0.79
33 Côte d’Ivoire 139 1.80 1.78 1.76 1.73 1.72 1.69 1.67 1.66 1.65 1.65 1.65 1.65 1.67 1.68 1.69 1.70
34 Croatia 57 0.74 0.74 0.75 0.76 0.76 0.78 0.79 0.80 0.82 0.82 0.82 0.82 0.81 0.81 0.81 0.79
35 Cuba 55 0.57 0.56 0.55 0.53 0.52 0.51 0.49 0.49 0.49 0.49 0.49 0.50 0.51 0.51 0.52 0.53
36 Curaçao 28 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
37 Denmark 135 16.50 15.77 15.44 15.00 14.70 14.48 14.19 13.80 13.39 13.39 13.32 13.35 13.47 13.75 14.00 14.16
38 Djibouti 10 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
39 Dominica 37 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
40 Dominican Republic 92 0.62 0.61 0.61 0.60 0.61 0.60 0.60 0.59 0.58 0.58 0.58 0.58 0.58 0.58 0.58 0.57
41 Ecuador 69 4.64 4.58 4.23 4.00 4.02 4.05 4.02 3.99 4.11 4.22 4.36 4.44 4.60 4.79 4.92 5.11
42 Egypt, Arab Rep. 73 1.29 1.27 1.24 1.21 1.16 1.09 1.06 1.03 1.01 1.00 0.99 0.97 0.95 0.91 0.86 0.83
43 El Salvador 62 0.85 0.84 0.82 0.82 0.82 0.84 0.87 0.89 0.91 0.92 0.94 0.95 0.96 0.97 1.00 1.03
44 Equatorial Guinea 80 0.54 0.54 0.53 0.53 0.52 0.52 0.52 0.52 0.52 0.53 0.54 0.55 0.55 0.56 0.56 0.57
45 Eritrea 52 0.40 0.40 0.41 0.42 0.43 0.44 0.43 0.43 0.43 0.43 0.43 0.43 0.43 0.43 0.44 0.45
46 Estonia 47 1.12 1.09 1.06 1.02 1.00 0.98 0.96 0.94 0.94 0.95 0.95 0.94 0.95 0.95 0.92 0.90
species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
47 Faroe Islands 90 14.99 14.07 13.37 12.80 12.28 11.75 11.16 10.55 10.31 10.33 10.32 10.40 10.55 10.72 10.86
48 Fiji 52 1.57 1.52 1.47 1.44 1.41 1.38 1.36 1.33 1.32 1.30 1.28 1.26 1.24 1.23 1.21
49 Finland 46 3.34 2.94 2.70 2.53 2.40 2.30 2.22 2.16 2.11 2.08 2.04 2.01 1.99 1.97 1.93
50 France 294 11.35 10.54 9.99 9.60 9.30 9.08 8.89 8.73 8.58 8.44 8.34 8.22 8.08 7.97 7.88
51 French Polynesia 33 0.64 0.60 0.58 0.56 0.55 0.54 0.53 0.53 0.52 0.51 0.49 0.46 0.43 0.43 0.43
52 Gabon 79 3.58 3.44 3.34 3.26 3.20 3.15 3.11 3.07 3.04 3.02 2.99 2.96 2.94 2.91 2.89
53 Gambia, The 199 4.88 4.63 4.46 4.34 4.25 4.18 4.12 4.07 4.03 3.99 3.97 3.95 3.93 3.91 3.88
54 Georgia 37 2.78 2.40 2.20 2.06 1.97 1.91 1.86 1.82 1.79 1.77 1.75 1.73 1.72 1.70 1.69
55 Germany 118 4.80 4.64 4.51 4.39 4.27 4.16 4.05 3.97 3.88 3.80 3.74 3.70 3.64 3.59 3.51
56 Ghana 87 13.04 12.29 11.78 11.41 11.11 10.89 10.71 10.54 10.38 10.25 10.13 10.04 9.95 9.86 9.76
57 Greece 68 5.88 5.46 5.18 4.97 4.81 4.67 4.56 4.47 4.38 4.30 4.23 4.16 4.09 4.04 4.00
58 Greenland 90 16.17 15.39 14.83 14.30 13.90 13.51 13.15 12.85 12.57 12.21 11.90 11.63 11.32 10.97 10.65
59 Grenada 50 0.15 0.15 0.14 0.14 0.13 0.13 0.12 0.12 0.11 0.11 0.11 0.10 0.10 0.10 0.09
60 Guam 73 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
61 Guatemala 84 1.17 1.09 1.04 1.00 0.98 0.95 0.94 0.92 0.91 0.90 0.89 0.88 0.85 0.84 0.83
62 Guinea 121 11.02 9.69 8.95 8.47 8.12 7.87 7.67 7.51 7.39 7.28 7.19 7.10 7.01 6.91 6.83
63 Guinea-Bissau 144 8.56 7.93 7.53 7.25 7.05 6.90 6.78 6.68 6.59 6.51 6.44 6.38 6.32 6.26 6.20
64 Guyana 57 1.43 1.35 1.29 1.24 1.21 1.18 1.16 1.14 1.12 1.11 1.10 1.08 1.07 1.06 1.05
65 Haiti 67 0.58 0.49 0.44 0.41 0.39 0.37 0.36 0.35 0.34 0.34 0.33 0.33 0.32 0.32 0.31
66 Honduras 92 0.53 0.51 0.49 0.48 0.47 0.46 0.45 0.45 0.44 0.43 0.43 0.42 0.42 0.41 0.41
67 Hong Kong SAR, China 83 0.51 0.50 0.49 0.49 0.48 0.47 0.47 0.46 0.46 0.45 0.45 0.45 0.44 0.44 0.44
68 Iceland 68 47.30 45.07 43.42 42.16 41.08 40.18 39.40 38.69 38.09 37.56 37.02 36.54 35.97 35.30 34.76
69 India 206 116.12 102.69 94.67 89.26 85.23 82.10 79.68 77.59 75.79 74.52 73.58 72.54 71.76 71.12 70.62
(Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
47 Faroe Islands 90 10.89 10.49 10.14 10.02 9.96 9.97 9.91 9.94 9.99 10.07 10.22 10.32 10.40 10.50 10.53
48 Fiji 52 1.20 1.18 1.17 1.15 1.14 1.12 1.11 1.10 1.08 1.07 1.06 1.04 1.03 1.02 1.00
49 Finland 46 1.92 1.90 1.88 1.87 1.84 1.82 1.80 1.78 1.76 1.73 1.71 1.69 1.67 1.64 1.62
50 France 294 7.80 7.72 7.64 7.54 7.43 7.34 7.28 7.22 7.09 7.03 6.90 6.76 6.56 6.39 6.21
52 Gabon 79 2.86 2.83 2.79 2.74 2.69 2.63 2.60 2.55 2.51 2.46 2.42 2.38 2.32 2.26 2.15
53 Gambia, The 199 3.86 3.82 3.79 3.76 3.74 3.71 3.62 3.43 3.26 3.16 3.11 3.05 2.99 2.95 2.91
54 Georgia 37 1.68 1.66 1.65 1.63 1.61 1.59 1.57 1.55 1.52 1.49 1.45 1.41 1.36 1.32 1.26
55 Germany 118 3.44 3.39 3.31 3.25 3.17 3.12 3.05 3.01 2.96 2.89 2.83 2.78 2.74 2.71 2.71
56 Ghana 87 9.64 9.53 9.41 9.29 9.20 9.04 8.82 8.57 8.27 8.15 7.99 7.78 7.62 7.43 7.27
57 Greece 68 3.95 3.92 3.88 3.85 3.82 3.80 3.77 3.75 3.73 3.70 3.67 3.63 3.59 3.55 3.51
58 Greenland 90 10.34 9.97 9.64 9.32 8.99 8.78 8.66 8.50 8.32 8.20 8.05 7.94 7.85 7.79 7.79
59 Grenada 50 0.09 0.09 0.09 0.08 0.08 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.05 0.05
60 Guam 73 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02
61 Guatemala 84 0.82 0.81 0.80 0.79 0.77 0.77 0.76 0.75 0.74 0.73 0.71 0.69 0.69 0.67 0.66
62 Guinea 121 6.76 6.71 6.67 6.62 6.57 6.52 6.46 6.40 6.34 6.28 6.21 6.13 6.04 5.93 5.80
63 Guinea-Bissau 144 6.13 6.08 6.03 5.98 5.92 5.86 5.80 5.74 5.67 5.60 5.52 5.43 5.36 5.24 5.13
64 Guyana 57 1.04 1.03 1.02 1.01 1.00 0.99 0.98 0.97 0.95 0.94 0.93 0.92 0.91 0.89 0.87
65 Haiti 67 0.31 0.31 0.31 0.30 0.30 0.30 0.30 0.29 0.29 0.29 0.29 0.29 0.28 0.28 0.28
66 Honduras 92 0.41 0.40 0.40 0.39 0.38 0.38 0.37 0.36 0.35 0.35 0.35 0.35 0.34 0.34 0.33
68 Iceland 68 33.97 33.15 32.56 32.28 32.08 31.78 31.48 31.26 31.04 30.71 30.37 30.07 29.75 29.04 28.22
69 India 206 69.85 69.25 68.63 68.11 67.57 67.15 66.51 65.86 65.39 64.85 64.13 63.30 62.64 62.12 61.48
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
47 Faroe Islands 90 10.59 10.73 10.86 11.04 11.15 11.17 11.16 11.08 11.04 11.01 11.01 11.03 11.14 11.23 11.31
48 Fiji 52 0.98 0.97 0.96 0.96 0.95 0.95 0.94 0.94 0.93 0.93 0.93 0.93 0.93 0.93 0.93
49 Finland 46 1.60 1.58 1.58 1.57 1.55 1.53 1.52 1.50 1.49 1.47 1.46 1.45 1.45 1.43 1.41
50 France 294 6.06 5.90 5.85 5.80 5.65 5.55 5.42 5.29 5.20 5.05 4.96 4.88 4.81 4.70 4.63
52 Gabon 79 2.05 1.96 1.87 1.80 1.73 1.69 1.63 1.58 1.55 1.50 1.47 1.44 1.41 1.39 1.37
53 Gambia, The 199 2.88 2.84 2.81 2.78 2.76 2.74 2.65 2.57 2.47 2.36 2.21 1.99 1.79 1.72 1.67
54 Georgia 37 1.21 1.15 1.10 1.05 1.01 0.96 0.92 0.89 0.85 0.81 0.85 0.92 1.00 1.06 1.13
55 Germany 118 2.71 2.71 2.71 2.70 2.69 2.68 2.67 2.68 2.69 2.71 2.73 2.77 2.81 2.84 2.87
56 Ghana 87 7.14 7.03 6.90 6.81 6.73 6.68 6.60 6.50 6.33 6.22 6.12 6.00 5.88 5.71 5.59
57 Greece 68 3.47 3.43 3.38 3.32 3.27 3.20 3.14 3.05 2.97 2.90 2.83 2.77 2.70 2.63 2.55
58 Greenland 90 7.75 7.71 7.77 7.84 7.90 7.97 7.85 7.82 7.80 7.83 7.78 7.75 7.80 7.79 7.73
59 Grenada 50 0.05 0.05 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07
60 Guam 73 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.02
61 Guatemala 84 0.65 0.64 0.63 0.62 0.62 0.61 0.60 0.60 0.60 0.59 0.58 0.58 0.57 0.56 0.55
62 Guinea 121 5.68 5.56 5.45 5.35 5.26 5.19 5.15 5.11 5.08 5.05 5.04 4.98 4.89 4.84 4.81
63 Guinea-Bissau 144 5.00 4.78 4.61 4.41 4.23 4.12 3.98 3.79 3.58 3.41 3.18 3.09 3.01 2.99 2.97
64 Guyana 57 0.86 0.85 0.84 0.83 0.82 0.81 0.80 0.79 0.79 0.78 0.77 0.76 0.75 0.75 0.73
65 Haiti 67 0.28 0.28 0.28 0.28 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.26 0.26
66 Honduras 92 0.32 0.32 0.31 0.31 0.30 0.30 0.29 0.27 0.26 0.25 0.25 0.25 0.24 0.23 0.23
68 Iceland 68 27.38 26.72 26.18 26.31 26.40 25.78 25.04 24.29 23.58 22.77 22.22 21.68 21.65 21.06 20.39
69 India 206 60.94 60.55 60.08 59.11 58.21 57.52 57.07 56.60 56.16 55.70 54.71 53.61 52.37 51.30 50.71
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
47 Faroe Islands 90 11.40 11.32 11.34 11.26 11.16 11.10 10.96 10.68 10.56 10.11 9.85 9.56 9.26 9.08 9.06 9.05
48 Fiji 52 0.93 0.93 0.94 0.95 0.95 0.96 0.96 0.95 0.95 0.95 0.96 0.97 0.98 0.98 0.99 0.98
49 Finland 46 1.37 1.33 1.30 1.25 1.21 1.17 1.14 1.12 1.12 1.14 1.15 1.14 1.14 1.12 1.10 1.08
50 France 294 4.57 4.51 4.49 4.44 4.41 4.33 4.25 4.20 4.14 4.05 3.99 3.93 3.88 3.82 3.81 3.79
51 French Polynesia 33 0.22 0.22 0.23 0.23 0.23 0.23 0.24 0.24 0.24 0.25 0.25 0.25 0.26 0.26 0.27 0.27
52 Gabon 79 1.33 1.26 1.23 1.21 1.16 1.12 1.05 1.01 0.96 0.92 0.86 0.82 0.80 0.78 0.75 0.74
53 Gambia, The 199 1.64 1.63 1.62 1.62 1.64 1.64 1.64 1.60 1.57 1.52 1.47 1.44 1.41 1.38 1.34 1.31
54 Georgia 37 1.20 1.25 1.29 1.32 1.34 1.36 1.37 1.38 1.38 1.37 1.35 1.32 1.29 1.24 1.16 1.10
55 Germany 118 2.89 2.91 2.93 2.94 2.97 2.98 3.00 3.01 3.02 3.06 3.08 3.09 3.11 3.13 3.16 3.20
56 Ghana 87 5.52 5.41 5.20 5.03 4.87 4.66 4.49 4.34 4.24 4.09 3.92 3.80 3.69 3.62 3.48 3.35
57 Greece 68 2.42 2.32 2.23 2.14 2.09 2.04 2.00 1.97 1.94 1.91 1.88 1.85 1.82 1.79 1.76 1.74
58 Greenland 90 7.68 7.79 7.91 7.94 7.99 8.03 8.06 8.03 8.00 7.97 7.99 8.02 8.04 8.08 8.12 8.20
59 Grenada 50 0.07 0.07 0.07 0.07 0.07 0.07 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08
60 Guam 73 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.03
61 Guatemala 84 0.53 0.51 0.49 0.48 0.46 0.44 0.43 0.41 0.40 0.39 0.38 0.37 0.37 0.36 0.36 0.36
62 Guinea 121 4.81 4.74 4.68 4.55 4.40 4.31 4.17 4.05 3.96 3.92 3.91 3.89 3.81 3.72 3.64 3.51
63 Guinea-Bissau 144 2.95 2.92 2.88 2.84 2.81 2.77 2.69 2.63 2.58 2.53 2.49 2.45 2.44 2.43 2.44 2.47
64 Guyana 57 0.72 0.70 0.66 0.62 0.59 0.55 0.53 0.50 0.49 0.46 0.44 0.43 0.41 0.40 0.39 0.38
65 Haiti 67 0.26 0.26 0.25 0.25 0.25 0.24 0.24 0.23 0.23 0.22 0.22 0.21 0.21 0.20 0.19 0.19
66 Honduras 92 0.22 0.21 0.20 0.19 0.19 0.18 0.17 0.17 0.16 0.15 0.14 0.14 0.14 0.14 0.14 0.14
67 Hong Kong SAR, China 83 0.14 0.14 0.13 0.12 0.12 0.13 0.13 0.13 0.14 0.14 0.14 0.15 0.15 0.15 0.16 0.16
68 Iceland 68 19.95 19.60 18.85 17.88 17.42 16.98 16.28 15.50 14.61 13.90 13.31 12.81 12.71 12.45 12.28 12.24
69 India 206 49.66 48.77 47.54 46.11 44.77 43.68 42.67 42.06 41.18 40.41 39.64 38.98 38.26 37.30 36.40 35.49
species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
70 Indonesia 143 161.01 147.77 139.35 133.49 129.23 126.00 123.52 121.57 120.03 118.79 117.80 117.00 116.25 115.59 115.03
71 Iran, Islamic Rep. 132 11.25 10.48 9.92 9.48 9.14 8.85 8.62 8.41 8.24 8.08 7.95 7.82 7.71 7.61 7.52
72 Iraq 12 0.30 0.28 0.27 0.26 0.25 0.25 0.24 0.24 0.23 0.23 0.23 0.23 0.22 0.22 0.22
73 Ireland 131 14.23 13.68 13.24 12.88 12.56 12.28 12.07 11.84 11.62 11.41 11.23 11.01 10.90 10.78 10.65
74 Israel 49 0.16 0.15 0.14 0.14 0.14 0.13 0.13 0.13 0.12 0.12 0.12 0.11 0.11 0.11 0.11
75 Italy 96 24.97 23.90 23.05 22.16 21.50 20.93 20.43 19.95 19.62 19.35 19.13 18.96 18.76 18.66 18.51
76 Jamaica 62 1.58 1.48 1.41 1.35 1.31 1.27 1.24 1.22 1.19 1.17 1.15 1.14 1.12 1.11 1.09
77 Japan 159 251.82 239.16 230.31 223.55 217.91 213.23 209.03 205.37 201.65 198.30 195.06 191.98 188.74 185.81 183.20
78 Jordan 18 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
79 Kenya 45 0.48 0.45 0.42 0.41 0.39 0.38 0.37 0.36 0.36 0.35 0.34 0.34 0.33 0.33 0.32
80 Kiribati 35 4.55 4.17 3.93 3.77 3.65 3.57 3.50 3.44 3.40 3.35 3.30 3.28 3.26 3.23 3.22
81 Korea, Dem. People’s Rep. 84 22.84 22.17 21.60 21.09 20.65 20.21 19.77 19.33 18.88 18.44 18.01 17.60 17.22 16.88 16.55
82 Korea, Rep. 182 73.62 68.30 64.87 62.43 60.62 59.24 58.10 57.07 56.15 55.37 54.72 54.20 53.66 53.16 52.72
83 Kuwait 24 0.77 0.69 0.64 0.60 0.57 0.55 0.53 0.51 0.50 0.48 0.47 0.46 0.45 0.44 0.44
84 Latvia 48 3.55 3.23 3.01 2.86 2.74 2.65 2.57 2.51 2.45 2.41 2.37 2.34 2.31 2.29 2.26
85 Lebanon 20 0.24 0.21 0.19 0.18 0.17 0.16 0.15 0.15 0.15 0.14 0.14 0.14 0.13 0.13 0.13
86 Liberia 134 1.76 1.58 1.47 1.40 1.35 1.32 1.29 1.26 1.25 1.23 1.21 1.20 1.18 1.17 1.16
87 Libya 91 1.82 1.65 1.55 1.48 1.43 1.40 1.37 1.34 1.32 1.31 1.29 1.28 1.27 1.27 1.26
88 Lithuania 44 0.87 0.83 0.80 0.78 0.76 0.74 0.72 0.71 0.70 0.68 0.67 0.67 0.66 0.65 0.64
89 Madagascar 51 3.38 2.99 2.76 2.60 2.50 2.42 2.36 2.31 2.27 2.24 2.21 2.19 2.17 2.15 2.14
90 Malaysia 171 83.64 75.34 70.35 67.00 64.63 62.87 61.54 60.51 59.71 59.05 58.51 58.05 57.66 57.30 56.95
91 Maldives 61 3.17 2.86 2.67 2.54 2.44 2.37 2.32 2.28 2.24 2.22 2.19 2.18 2.16 2.15 2.14
92 Malta 101 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03
(Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
70 Indonesia 143 114.52 113.97 113.42 112.98 112.48 111.96 111.25 110.61 109.92 109.19 108.50 107.73 106.74 105.42 104.22
71 Iran, Islamic Rep. 132 7.43 7.35 7.27 7.19 7.13 7.06 7.01 6.96 6.91 6.87 6.83 6.78 6.72 6.66 6.60
72 Iraq 12 0.22 0.22 0.22 0.22 0.21 0.21 0.21 0.21 0.21 0.20 0.19 0.18 0.17 0.17 0.16
73 Ireland 131 10.53 10.44 10.37 10.28 10.18 10.07 9.94 9.81 9.69 9.57 9.46 9.39 9.36 9.39 9.39
74 Israel 49 0.11 0.11 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.09 0.09 0.09 0.09 0.09 0.09
75 Italy 96 18.35 18.17 17.96 17.73 17.56 17.41 17.21 17.05 16.84 16.57 16.28 15.84 15.44 15.11 14.69
76 Jamaica 62 1.08 1.07 1.05 1.04 1.03 1.02 1.00 0.99 0.98 0.97 0.96 0.94 0.93 0.92 0.91
77 Japan 159 181.47 179.45 177.72 175.93 173.88 172.16 170.10 167.90 165.63 163.12 160.45 157.53 154.48 151.24 147.64
78 Jordan 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
79 Kenya 45 0.32 0.31 0.31 0.30 0.30 0.29 0.28 0.28 0.27 0.27 0.26 0.25 0.25 0.24 0.24
80 Kiribati 35 3.20 3.19 3.17 3.16 3.15 3.12 3.14 3.15 3.14 3.14 3.14 3.13 3.11 3.09 3.05
82 Korea, Rep. 182 52.25 51.80 51.40 51.02 50.60 50.17 49.73 49.24 48.57 47.87 47.02 46.24 45.47 44.64 43.79
83 Kuwait 24 0.43 0.42 0.41 0.40 0.40 0.39 0.38 0.38 0.37 0.36 0.36 0.36 0.37 0.37 0.37
84 Latvia 48 2.24 2.22 2.19 2.16 2.12 2.09 2.06 2.03 1.99 1.96 1.92 1.89 1.86 1.84 1.82
85 Lebanon 20 0.13 0.13 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11
86 Liberia 134 1.15 1.14 1.13 1.12 1.11 1.10 1.08 1.05 1.03 1.02 1.01 1.01 1.00 0.99 0.97
87 Libya 91 1.25 1.25 1.25 1.25 1.24 1.24 1.24 1.24 1.24 1.24 1.24 1.24 1.24 1.24 1.23
88 Lithuania 44 0.64 0.63 0.62 0.62 0.61 0.59 0.58 0.57 0.56 0.55 0.53 0.52 0.51 0.50 0.50
89 Madagascar 51 2.12 2.11 2.10 2.08 2.06 2.04 2.02 2.00 1.98 1.96 1.94 1.93 1.92 1.90 1.88
90 Malaysia 171 56.56 56.17 55.78 55.38 54.96 54.65 54.30 53.82 53.28 52.64 51.92 51.22 50.32 49.43 48.45
91 Maldives 61 2.14 2.12 2.11 2.10 2.09 2.08 2.05 2.03 2.01 1.98 1.95 1.93 1.93 1.93 1.94
92 Malta 101 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
70 Indonesia 143 103.15 102.42 101.64 99.75 97.73 95.77 93.56 91.30 88.82 86.47 84.15 82.21 80.62 78.95 77.40
71 Iran, Islamic Rep. 132 6.55 6.52 6.48 6.44 6.40 6.36 6.31 6.23 6.13 6.01 5.87 5.74 5.60 5.42 5.27
72 Iraq 12 0.15 0.15 0.15 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.15 0.15 0.16 0.16
73 Ireland 131 9.38 9.32 9.20 9.06 8.92 8.78 8.64 8.50 8.33 8.16 7.94 7.78 7.75 7.71 7.61
74 Israel 49 0.09 0.09 0.09 0.09 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.07 0.07 0.07 0.07
75 Italy 96 14.26 13.91 13.53 13.12 12.82 12.50 12.16 11.86 11.67 11.57 11.52 11.54 11.52 11.52 11.49
76 Jamaica 62 0.89 0.88 0.88 0.87 0.86 0.85 0.85 0.84 0.84 0.85 0.85 0.85 0.85 0.86 0.86
77 Japan 159 144.29 140.76 136.91 133.04 128.85 123.99 119.63 114.99 110.44 105.76 101.40 97.48 94.59 92.59 90.67
78 Jordan 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
79 Kenya 45 0.23 0.22 0.21 0.21 0.20 0.20 0.19 0.18 0.18 0.17 0.17 0.17 0.17 0.17 0.17
80 Kiribati 35 3.06 3.04 3.04 3.03 2.99 2.99 2.99 2.98 2.95 2.95 2.95 2.86 2.78 2.67 2.57
82 Korea, Rep. 182 42.93 42.14 41.31 40.63 39.89 39.14 38.51 37.56 36.82 36.17 35.45 34.75 34.22 33.72 32.91
83 Kuwait 24 0.38 0.39 0.39 0.38 0.37 0.36 0.35 0.35 0.33 0.31 0.30 0.30 0.30 0.29 0.28
84 Latvia 48 1.80 1.77 1.74 1.72 1.69 1.66 1.63 1.61 1.59 1.57 1.56 1.55 1.54 1.53 1.53
85 Lebanon 20 0.11 0.11 0.11 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12
86 Liberia 134 0.96 0.95 0.93 0.91 0.90 0.89 0.87 0.85 0.83 0.79 0.77 0.74 0.72 0.71 0.70
87 Libya 91 1.23 1.22 1.22 1.21 1.20 1.19 1.18 1.17 1.16 1.15 1.14 1.13 1.13 1.12 1.11
88 Lithuania 44 0.49 0.48 0.47 0.46 0.45 0.44 0.43 0.43 0.42 0.41 0.41 0.41 0.40 0.40 0.39
89 Madagascar 51 1.86 1.84 1.82 1.80 1.78 1.76 1.74 1.71 1.67 1.65 1.63 1.60 1.58 1.55 1.52
90 Malaysia 171 47.53 46.65 45.75 44.97 44.23 43.64 43.07 42.57 41.97 41.52 40.97 40.42 39.82 39.03 38.24
91 Maldives 61 1.94 1.94 1.93 1.93 1.92 1.91 1.88 1.86 1.84 1.82 1.80 1.77 1.73 1.70 1.68
92 Malta 101 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
70 Indonesia 143 75.64 73.51 71.26 69.18 66.99 64.97 62.93 60.92 58.88 56.81 54.84 53.13 51.28 49.55 48.17 46.71
71 Iran, Islamic Rep. 132 4.99 4.73 4.49 4.18 3.98 3.75 3.57 3.45 3.34 3.29 3.24 3.20 3.16 3.15 3.11 3.09
72 Iraq 12 0.16 0.16 0.15 0.14 0.13 0.12 0.11 0.09 0.08 0.08 0.08 0.08 0.07 0.07 0.07 0.06
73 Ireland 131 7.44 7.11 6.83 6.67 6.31 6.09 5.90 5.82 5.71 5.48 5.25 4.92 4.61 4.37 4.19 4.14
74 Israel 49 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
75 Italy 96 11.47 11.42 11.41 11.40 11.36 11.39 11.43 11.55 11.73 11.88 12.06 12.29 12.48 12.70 12.97 13.23
76 Jamaica 62 0.87 0.88 0.88 0.89 0.90 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1.00 1.01
77 Japan 159 89.23 88.12 86.40 85.15 84.73 84.06 83.74 83.97 84.51 84.81 85.25 85.72 86.25 86.95 87.62 88.65
78 Jordan 18 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
79 Kenya 45 0.17 0.17 0.17 0.17 0.17 0.18 0.18 0.18 0.18 0.19 0.19 0.19 0.19 0.19 0.19 0.19
80 Kiribati 35 2.50 2.56 2.56 2.42 2.26 2.20 2.14 1.99 1.72 1.76 1.78 1.72 1.68 1.65 1.60 1.42
81 Korea, Dem. People’s Rep. 84 6.72 6.87 7.12 7.39 7.69 8.00 8.31 8.63 8.94 9.26 9.49 9.79 10.09 10.41 10.73 11.08
82 Korea, Rep. 182 32.17 31.44 31.19 31.26 30.70 29.87 29.23 28.85 28.61 28.40 28.04 27.66 27.26 26.83 26.34 26.10
83 Kuwait 24 0.27 0.26 0.26 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.24 0.24 0.24
84 Latvia 48 1.51 1.48 1.42 1.36 1.34 1.33 1.32 1.32 1.31 1.31 1.30 1.30 1.30 1.30 1.29 1.27
85 Lebanon 20 0.12 0.11 0.11 0.10 0.10 0.10 0.10 0.09 0.09 0.09 0.09 0.08 0.08 0.08 0.08 0.08
86 Liberia 134 0.69 0.68 0.67 0.65 0.64 0.63 0.62 0.61 0.60 0.59 0.58 0.57 0.55 0.54 0.52 0.51
87 Libya 91 1.10 1.08 1.07 1.05 1.04 1.01 0.98 0.95 0.91 0.87 0.82 0.78 0.75 0.70 0.66 0.62
88 Lithuania 44 0.38 0.37 0.35 0.32 0.31 0.30 0.29 0.29 0.28 0.27 0.27 0.26 0.26 0.24 0.24 0.23
89 Madagascar 51 1.47 1.44 1.41 1.37 1.33 1.30 1.27 1.24 1.21 1.19 1.17 1.14 1.12 1.09 1.06 1.02
90 Malaysia 171 37.41 36.43 35.47 34.59 33.77 32.86 31.98 31.26 30.53 29.72 28.87 28.16 27.46 26.84 26.42 26.01
91 Maldives 61 1.72 1.75 1.77 1.79 1.80 1.82 1.83 1.83 1.82 1.81 1.65 1.50 1.35 1.27 1.20 1.12
92 Malta 101 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
93 Marshall Islands 60 0.81 0.76 0.71 0.69 0.67 0.65 0.64 0.63 0.62 0.61 0.60 0.59 0.59 0.58 0.57
94 Mauritania 185 22.04 20.33 19.26 18.54 18.01 17.62 17.32 17.09 16.90 16.75 16.62 16.52 16.43 16.33 16.25
95 Mauritius 46 0.46 0.44 0.43 0.42 0.41 0.41 0.40 0.40 0.39 0.39 0.38 0.38 0.38 0.37 0.37
96 Mexico 142 61.21 56.43 53.36 51.20 49.54 48.21 47.11 46.13 45.40 44.54 43.71 42.91 42.15 41.51 40.89
97 Micronesia, Fed. Sts. 65 5.32 5.02 4.81 4.66 4.53 4.44 4.35 4.29 4.23 4.17 4.11 4.08 4.03 4.00 3.98
98 Montenegro 55 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
99 Morocco 175 45.88 41.12 38.33 36.41 35.04 34.07 33.34 32.73 32.17 31.70 31.33 31.02 30.76 30.50 30.27
100 Mozambique 63 4.69 4.28 4.00 3.79 3.64 3.52 3.42 3.33 3.27 3.21 3.16 3.11 3.07 3.03 3.00
101 Myanmar 92 38.31 34.64 32.35 30.78 29.65 28.80 28.15 27.63 27.21 26.88 26.62 26.41 26.27 26.15 26.06
102 Namibia 42 61.03 60.63 60.07 59.31 58.54 57.88 57.33 56.84 56.38 55.97 55.48 55.00 54.41 53.71 52.57
103 Netherlands 146 14.13 13.59 13.11 12.71 12.29 11.95 11.69 11.46 11.19 10.95 10.72 10.51 10.25 10.01 9.71
104 New Caledonia 39 0.33 0.29 0.27 0.26 0.25 0.24 0.23 0.23 0.22 0.22 0.22 0.21 0.21 0.20 0.20
105 New Zealand 87 25.37 24.48 23.82 23.30 22.88 22.52 22.22 21.95 21.71 21.52 21.32 21.13 20.99 20.88 20.77
106 Nicaragua 32 1.12 1.06 1.02 0.99 0.95 0.93 0.91 0.89 0.87 0.85 0.84 0.82 0.80 0.78 0.77
107 Nigeria 49 10.30 9.49 8.97 8.61 8.34 8.14 7.98 7.86 7.76 7.68 7.62 7.56 7.52 7.48 7.45
108 Northern Mariana Islands 83 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.10 0.10 0.10 0.10
109 Norway 136 81.93 75.81 71.72 68.86 67.03 65.07 63.83 62.43 61.40 60.59 59.76 58.98 58.26 57.76 57.29
110 Oman 59 5.19 4.64 4.30 4.07 3.90 3.76 3.64 3.55 3.47 3.41 3.35 3.31 3.26 3.22 3.18
111 Pakistan 122 25.03 23.11 21.91 21.08 20.39 19.86 19.42 19.06 18.80 18.56 18.36 18.16 17.99 17.87 17.72
112 Palau 54 1.36 1.18 1.08 1.01 0.97 0.94 0.91 0.90 0.88 0.87 0.86 0.85 0.84 0.84 0.83
113 Panama 52 5.19 4.97 4.81 4.69 4.60 4.52 4.45 4.38 4.32 4.27 4.21 4.16 4.12 4.06 3.94
114 Papua New Guinea 52 6.75 5.71 5.16 4.82 4.59 4.41 4.29 4.19 4.11 4.04 3.98 3.95 3.91 3.86 3.84
115 Peru 49 292.92 281.47 272.59 265.55 259.88 255.24 251.37 248.05 245.07 242.07 238.14 233.07 226.50 218.51 211.57
(Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
94 Mauritania 185 16.15 16.05 15.98 15.84 15.61 15.38 15.11 14.80 14.53 14.21 13.78 13.41 12.94 12.68 12.71
95 Mauritius 46 0.37 0.36 0.36 0.35 0.33 0.33 0.33 0.33 0.32 0.32 0.32 0.32 0.31 0.30 0.30
96 Mexico 142 40.50 40.24 39.95 39.49 39.15 38.93 38.51 38.09 37.67 37.29 37.04 36.77 36.73 36.70 36.14
98 Montenegro 55 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
99 Morocco 175 30.05 29.79 29.38 29.00 28.67 28.24 27.78 27.30 26.72 25.97 25.43 24.93 24.51 23.95 23.63
100 Mozambique 63 2.97 2.94 2.91 2.88 2.85 2.81 2.78 2.74 2.69 2.65 2.59 2.53 2.47 2.42 2.37
101 Myanmar 92 26.01 25.93 25.88 25.82 25.74 25.65 25.57 25.48 25.40 25.25 25.16 25.04 24.90 24.77 24.67
102 Namibia 42 50.96 49.04 46.97 44.01 39.91 36.30 34.36 32.30 30.75 28.80 26.89 24.75 23.11 22.27 21.28
103 Netherlands 146 9.42 9.21 9.01 8.93 8.79 8.62 8.43 8.26 8.11 7.97 7.83 7.72 7.69 7.70 7.74
104 New Caledonia 39 0.20 0.20 0.20 0.19 0.19 0.19 0.19 0.18 0.18 0.17 0.17 0.17 0.16 0.16 0.16
105 New Zealand 87 20.67 20.55 20.43 20.31 20.15 20.00 19.84 19.63 19.31 18.98 18.57 18.32 17.83 17.11 16.95
106 Nicaragua 32 0.75 0.74 0.72 0.70 0.67 0.65 0.62 0.60 0.58 0.55 0.54 0.52 0.50 0.48 0.47
107 Nigeria 49 7.41 7.37 7.34 7.30 7.27 7.25 7.21 7.12 7.04 6.97 6.90 6.85 6.79 6.73 6.69
109 Norway 136 56.85 56.09 54.44 52.29 51.00 50.13 49.02 48.07 47.06 46.25 45.49 44.95 43.50 41.96 41.24
110 Oman 59 3.15 3.12 3.08 3.05 3.02 3.00 2.98 2.96 2.94 2.92 2.89 2.87 2.84 2.81 2.78
111 Pakistan 122 17.52 17.32 17.09 16.90 16.75 16.59 16.43 16.30 16.13 15.94 15.81 15.68 15.53 15.35 15.08
112 Palau 54 0.83 0.82 0.82 0.81 0.81 0.81 0.80 0.79 0.80 0.79 0.79 0.79 0.79 0.79 0.80
113 Panama 52 3.84 3.76 3.64 3.57 3.52 3.51 3.49 3.45 3.43 3.38 3.36 3.34 3.24 3.10 3.06
114 Papua New Guinea 52 3.80 3.79 3.75 3.73 3.71 3.67 3.41 3.17 3.22 3.24 3.20 3.23 3.23 3.25 3.27
115 Peru 49 202.24 196.09 188.57 179.65 170.63 163.87 152.41 143.93 143.57 146.44 146.82 148.13 148.28 150.83 152.16
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
94 Mauritania 185 12.76 12.59 12.36 12.10 11.87 11.72 11.51 11.26 10.90 10.52 10.17 9.84 9.54 9.42 9.36
95 Mauritius 46 0.30 0.30 0.30 0.30 0.30 0.30 0.30 0.30 0.29 0.29 0.29 0.28 0.28 0.27 0.27
96 Mexico 142 35.36 34.25 32.80 31.80 31.24 30.72 30.12 29.51 28.80 28.23 27.51 26.93 26.28 26.02 25.75
98 Montenegro 55 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
99 Morocco 175 23.46 23.21 23.02 22.80 22.47 22.25 22.03 21.70 21.47 20.95 20.47 20.04 19.65 19.51 19.28
100 Mozambique 63 2.31 2.24 2.18 2.12 2.07 2.03 1.99 1.94 1.90 1.86 1.82 1.79 1.75 1.72 1.69
101 Myanmar 92 24.58 24.49 24.40 24.26 24.10 23.92 23.73 23.48 23.17 22.88 22.58 22.25 21.90 21.47 20.99
102 Namibia 42 20.91 20.56 19.89 19.39 18.80 18.47 18.28 18.31 17.49 17.25 17.44 17.90 18.56 19.09 19.51
103 Netherlands 146 7.76 7.74 7.69 7.57 7.45 7.34 7.12 6.93 6.78 6.64 6.46 6.35 6.28 6.24 6.18
104 New Caledonia 39 0.15 0.15 0.15 0.14 0.14 0.14 0.14 0.14 0.14 0.13 0.13 0.13 0.12 0.12 0.12
105 New Zealand 87 16.68 16.40 16.03 15.68 15.25 14.71 14.31 13.79 12.98 12.13 11.43 10.96 10.56 10.12 9.81
106 Nicaragua 32 0.46 0.45 0.44 0.44 0.45 0.45 0.45 0.46 0.47 0.47 0.48 0.48 0.48 0.48 0.48
107 Nigeria 49 6.64 6.60 6.54 6.47 6.40 6.31 6.23 6.15 6.06 5.96 5.85 5.79 5.64 5.50 5.36
109 Norway 136 40.58 40.04 39.21 38.46 37.43 36.91 36.71 36.88 36.99 37.16 37.42 37.88 37.62 37.26 36.64
110 Oman 59 2.74 2.71 2.69 2.67 2.63 2.60 2.56 2.54 2.49 2.41 2.38 2.33 2.31 2.27 2.20
111 Pakistan 122 14.86 14.66 14.41 14.19 13.94 13.68 13.42 13.17 12.88 12.54 12.24 11.87 11.49 11.12 10.61
112 Palau 54 0.80 0.78 0.76 0.73 0.70 0.67 0.65 0.64 0.62 0.62 0.61 0.63 0.62 0.62 0.63
113 Panama 52 2.99 2.87 2.82 2.75 2.61 2.57 2.38 2.34 2.28 2.28 2.21 2.21 2.18 2.15 2.10
114 Papua New Guinea 52 3.32 3.34 3.34 3.31 3.30 3.27 3.19 3.13 3.07 3.06 2.91 2.79 2.64 2.58 2.56
115 Peru 49 153.54 156.09 158.61 159.75 163.17 164.31 164.41 162.45 162.01 158.28 153.81 150.53 147.02 142.42 135.96
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
93 Marshall Islands 60 0.38 0.39 0.39 0.39 0.34 0.33 0.32 0.27 0.23 0.23 0.22 0.21 0.20 0.20 0.20 0.19
94 Mauritania 185 9.37 9.42 9.34 9.20 9.05 8.92 8.77 8.56 8.41 8.24 7.98 7.67 7.57 7.41 7.22 7.03
95 Mauritius 46 0.26 0.25 0.25 0.23 0.22 0.22 0.20 0.19 0.17 0.16 0.16 0.15 0.15 0.15 0.15 0.14
96 Mexico 142 25.40 24.88 24.27 23.67 23.48 23.30 23.06 22.79 22.49 22.16 22.07 22.12 22.09 21.94 21.64 21.41
97 Micronesia, Fed. Sts. 65 1.99 1.86 1.81 1.82 1.83 1.76 1.67 1.65 1.65 1.49 1.38 1.22 1.04 0.82 0.84 0.79
98 Montenegro 55 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
99 Morocco 175 18.88 18.42 18.30 17.98 17.74 17.56 17.16 16.52 16.16 15.92 15.68 15.45 15.37 15.35 15.16 14.83
100 Mozambique 63 1.66 1.64 1.62 1.60 1.58 1.56 1.55 1.54 1.53 1.53 1.53 1.54 1.55 1.56 1.58 1.60
101 Myanmar 92 20.40 19.78 19.10 18.50 17.90 17.28 16.59 15.97 15.33 14.71 14.09 13.47 12.92 12.46 12.19 11.91
102 Namibia 42 20.08 20.76 21.53 22.33 23.05 23.77 24.50 25.25 25.95 26.61 27.34 28.07 28.84 29.70 30.57 31.40
103 Netherlands 146 6.15 6.10 6.16 6.22 6.21 6.20 6.22 6.27 6.34 6.41 6.49 6.60 6.70 6.83 6.97 7.07
104 New Caledonia 39 0.12 0.12 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11
105 New Zealand 87 9.51 9.22 8.96 8.63 8.26 7.96 7.73 7.50 7.26 6.98 6.74 6.58 6.40 6.25 6.17 6.11
106 Nicaragua 32 0.48 0.47 0.46 0.45 0.43 0.41 0.40 0.39 0.38 0.37 0.37 0.36 0.36 0.36 0.36 0.37
107 Nigeria 49 5.22 5.06 4.91 4.75 4.60 4.45 4.29 4.14 3.99 3.82 3.67 3.51 3.35 3.21 3.06 2.93
108 Northern Mariana Islands 83 0.03 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.05 0.06 0.06 0.06 0.06
109 Norway 136 36.11 35.47 34.92 34.11 33.67 33.41 33.09 32.84 32.51 32.43 32.50 32.74 33.07 33.17 33.22 33.13
110 Oman 59 2.17 2.10 2.06 2.03 2.01 1.99 1.96 1.92 1.88 1.84 1.77 1.72 1.67 1.62 1.57 1.52
111 Pakistan 122 10.33 10.06 9.81 9.51 9.24 8.92 8.66 8.38 8.12 7.88 7.58 7.38 7.16 6.96 6.74 6.53
112 Palau 54 0.63 0.62 0.58 0.55 0.56 0.55 0.55 0.56 0.55 0.54 0.52 0.51 0.50 0.47 0.43 0.39
113 Panama 52 2.06 2.00 1.98 1.94 1.85 1.86 1.75 1.62 1.52 1.45 1.42 1.40 1.40 1.38 1.35 1.30
114 Papua New Guinea 52 2.44 2.19 2.17 2.19 2.25 2.33 2.33 2.40 2.49 2.49 2.49 2.49 2.46 2.42 2.37 2.26
115 Peru 49 125.07 119.27 112.67 108.42 110.29 105.47 97.30 93.28 87.74 85.90 78.87 72.32 69.13 65.55 61.78 58.47
species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
116 Philippines 160 37.09 32.27 29.52 27.72 26.46 25.52 24.80 24.25 23.83 23.47 23.18 22.94 22.73 22.53 22.31
117 Poland 53 5.07 4.86 4.70 4.56 4.44 4.34 4.25 4.17 4.10 4.04 3.98 3.93 3.89 3.84 3.80
118 Portugal 239 13.18 12.79 12.45 12.13 11.81 11.51 11.28 11.02 10.77 10.52 10.29 10.05 9.80 9.56 9.30
119 Puerto Rico 37 0.17 0.17 0.16 0.16 0.16 0.16 0.15 0.15 0.15 0.15 0.15 0.14 0.14 0.14 0.14
120 Qatar 43 0.25 0.24 0.22 0.22 0.21 0.20 0.20 0.20 0.20 0.20 0.19 0.19 0.19 0.19 0.19
121 Romania 26 0.34 0.33 0.32 0.31 0.30 0.30 0.29 0.29 0.28 0.27 0.27 0.26 0.26 0.25 0.25
122 Russian Federation 193 250.38 235.10 223.97 215.69 209.18 203.85 199.11 195.06 191.65 189.14 187.19 185.73 184.33 183.17 181.78
123 Samoa 60 0.42 0.39 0.37 0.36 0.34 0.33 0.33 0.32 0.31 0.31 0.30 0.30 0.29 0.29 0.28
124 São Tomé and Principe 141 1.07 1.02 0.97 0.94 0.91 0.88 0.86 0.85 0.83 0.82 0.80 0.79 0.78 0.77 0.77
125 Saudi Arabia 75 2.26 2.15 2.07 2.01 1.97 1.94 1.91 1.88 1.86 1.84 1.82 1.81 1.79 1.77 1.75
126 Senegal 156 28.12 26.93 26.09 25.46 24.97 24.57 24.21 23.90 23.63 23.38 23.08 22.76 22.47 22.17 21.88
127 Seychelles 34 1.47 1.36 1.29 1.24 1.20 1.17 1.15 1.13 1.11 1.10 1.09 1.08 1.07 1.07 1.06
128 Sierra Leone 131 6.31 5.37 4.87 4.54 4.31 4.14 4.01 3.91 3.83 3.76 3.70 3.65 3.58 3.51 3.44
129 Singapore 61 0.24 0.23 0.22 0.22 0.21 0.21 0.20 0.20 0.20 0.19 0.19 0.19 0.18 0.18 0.18
130 Sint Maarten (Dutch part) 26 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
131 Slovenia 68 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
132 Solomon Islands 14 1.87 1.70 1.60 1.53 1.47 1.43 1.41 1.39 1.37 1.36 1.35 1.35 1.34 1.32 1.32
133 Somalia 96 3.06 2.81 2.66 2.55 2.48 2.42 2.37 2.34 2.31 2.29 2.27 2.26 2.24 2.23 2.23
134 South Africa 101 39.38 37.93 36.74 35.62 34.70 33.92 33.27 32.76 32.27 31.75 31.20 30.58 29.83 29.06 28.23
135 Spain 251 29.13 28.06 27.06 26.34 25.69 25.08 24.40 23.85 23.31 22.77 22.24 21.74 21.17 20.71 20.21
136 Sri Lanka 120 13.52 12.06 11.21 10.63 10.21 9.89 9.64 9.43 9.27 9.12 9.00 8.89 8.79 8.69 8.61
137 St. Kitts and Nevis 30 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03
138 St. Lucia 56 0.15 0.15 0.14 0.14 0.14 0.13 0.13 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11
139 St. Martin (French part) 32 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01
140 St. Vincent and the Grenadines 66 0.20 0.19 0.18 0.18 0.17 0.17 0.16 0.15 0.14 0.14 0.14 0.13 0.13 0.13 0.12
141 Sudan 23 0.08 0.07 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
(Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
116 Philippines 160 22.07 21.83 21.60 21.38 21.12 20.93 20.76 20.61 20.42 20.21 20.03 19.81 19.61 19.43 19.28
117 Poland 53 3.77 3.73 3.69 3.65 3.59 3.54 3.48 3.43 3.37 3.31 3.24 3.16 3.10 3.04 3.00
118 Portugal 239 8.98 8.66 8.43 8.20 8.02 7.92 7.73 7.57 7.41 7.23 7.14 7.08 7.01 6.96 6.93
119 Puerto Rico 37 0.13 0.13 0.13 0.13 0.13 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.11 0.11 0.11
120 Qatar 43 0.19 0.19 0.19 0.18 0.18 0.18 0.18 0.18 0.18 0.18 0.17 0.17 0.17 0.17 0.17
121 Romania 26 0.25 0.24 0.24 0.24 0.24 0.24 0.24 0.23 0.23 0.23 0.22 0.22 0.22 0.22 0.22
123 Samoa 60 0.28 0.28 0.27 0.27 0.27 0.26 0.26 0.26 0.25 0.25 0.25 0.25 0.24 0.24 0.24
125 Saudi Arabia 75 1.73 1.71 1.68 1.63 1.59 1.54 1.51 1.48 1.44 1.39 1.37 1.34 1.32 1.28 1.26
126 Senegal 156 21.58 21.32 21.06 20.76 20.40 19.95 19.46 18.90 18.31 17.76 17.19 16.60 16.08 15.64 15.28
127 Seychelles 34 1.06 1.05 1.05 1.05 1.00 0.99 0.99 0.98 0.97 0.98 0.98 0.98 0.99 0.99 0.99
128 Sierra Leone 131 3.39 3.35 3.31 3.27 3.23 3.20 3.17 3.14 3.10 3.05 3.01 2.96 2.92 2.88 2.84
129 Singapore 61 0.18 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.16
131 Slovenia 68 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
132 Solomon Islands 14 1.32 1.32 1.31 1.31 1.31 1.30 1.30 1.24 1.25 1.26 1.25 1.26 1.25 1.25 1.24
133 Somalia 96 2.22 2.21 2.20 2.20 2.19 2.18 2.17 2.16 2.16 2.16 2.15 2.15 2.14 2.13 2.13
134 South Africa 101 27.38 26.48 25.60 24.42 22.79 21.45 20.82 20.24 19.71 18.96 18.17 17.46 16.95 16.66 16.33
135 Spain 251 19.80 19.39 19.01 18.64 18.26 17.95 17.56 17.34 17.11 16.84 16.63 16.31 16.05 15.75 15.53
136 Sri Lanka 120 8.51 8.46 8.41 8.36 8.31 8.27 8.23 8.21 8.18 8.15 8.10 8.04 7.98 7.93 7.79
138 St. Lucia 56 0.11 0.10 0.10 0.10 0.10 0.10 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.08
141 Sudan 23 0.06 0.06 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
116 Philippines 160 19.15 19.05 18.95 18.84 18.75 18.67 18.61 18.53 18.43 18.35 18.16 17.95 17.74 17.46 17.14
117 Poland 53 2.95 2.88 2.82 2.76 2.69 2.62 2.56 2.51 2.47 2.44 2.41 2.39 2.37 2.34 2.30
118 Portugal 239 6.90 6.81 6.72 6.67 6.64 6.58 6.48 6.36 6.26 6.20 6.15 6.15 6.15 6.14 6.13
119 Puerto Rico 37 0.11 0.10 0.10 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.09 0.09 0.09 0.09 0.09
120 Qatar 43 0.17 0.17 0.17 0.17 0.17 0.17 0.16 0.16 0.16 0.16 0.16 0.15 0.15 0.14 0.14
121 Romania 26 0.21 0.21 0.20 0.20 0.19 0.18 0.17 0.16 0.15 0.14 0.13 0.12 0.11 0.11 0.10
123 Samoa 60 0.23 0.23 0.23 0.22 0.22 0.21 0.21 0.21 0.20 0.20 0.20 0.20 0.20 0.21 0.21
125 Saudi Arabia 75 1.26 1.26 1.25 1.23 1.21 1.18 1.14 1.06 1.03 1.00 0.97 0.95 0.93 0.90 0.88
126 Senegal 156 15.00 14.63 14.37 14.12 13.93 13.73 13.54 13.35 13.01 12.75 12.51 12.14 11.82 11.40 11.10
127 Seychelles 34 0.99 1.00 1.00 1.00 1.00 0.97 0.95 0.93 0.91 0.86 0.82 0.81 0.79 0.75 0.71
128 Sierra Leone 131 2.82 2.79 2.77 2.75 2.73 2.72 2.71 2.70 2.68 2.66 2.64 2.61 2.61 2.61 2.62
129 Singapore 61 0.16 0.16 0.16 0.15 0.14 0.13 0.13 0.12 0.12 0.13 0.13 0.13 0.12 0.11 0.11
131 Slovenia 68 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
132 Solomon Islands 14 1.24 1.24 1.24 1.24 1.23 1.19 1.20 1.21 1.23 1.23 1.22 1.21 1.20 1.20 1.19
133 Somalia 96 2.12 2.11 2.10 2.09 2.08 2.06 2.04 2.03 2.00 1.98 1.97 1.94 1.92 1.88 1.84
134 South Africa 101 16.04 15.92 15.80 15.75 15.58 15.64 15.66 15.65 14.99 14.47 14.39 14.68 15.02 15.14 15.40
135 Spain 251 15.38 15.27 15.16 15.05 15.07 15.04 15.09 15.19 15.29 15.41 15.52 15.64 15.73 15.84 15.95
136 Sri Lanka 120 7.67 7.53 7.43 7.34 7.27 7.22 7.20 7.16 7.13 7.07 7.04 7.02 6.99 6.89 6.78
138 St. Lucia 56 0.08 0.08 0.08 0.07 0.05 0.05 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
141 Sudan 23 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04 0.04
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
116 Philippines 160 16.84 16.57 16.31 16.07 15.76 15.48 15.21 14.91 14.56 14.21 13.87 13.46 13.20 12.81 12.38 11.96
117 Poland 53 2.21 2.12 1.98 1.83 1.73 1.66 1.58 1.51 1.47 1.42 1.37 1.34 1.31 1.26 1.24 1.20
118 Portugal 239 6.17 6.21 6.27 6.35 6.43 6.53 6.65 6.75 6.87 6.98 7.10 7.20 7.30 7.39 7.45 7.55
119 Puerto Rico 37 0.09 0.09 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.11 0.11 0.11 0.11 0.11
120 Qatar 43 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.13 0.13 0.12 0.11 0.11 0.10 0.10
121 Romania 26 0.10 0.10 0.09 0.09 0.09 0.09 0.09 0.10 0.10 0.10 0.11 0.11 0.11 0.12 0.12 0.13
122 Russian Federation 193 101.28 98.89 95.11 91.50 88.96 87.30 86.73 86.10 87.09 87.44 89.09 90.25 91.36 91.93 92.51 92.58
123 Samoa 60 0.21 0.20 0.20 0.20 0.19 0.18 0.18 0.17 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.15
124 São Tomé and Principe 141 0.36 0.36 0.36 0.35 0.35 0.35 0.34 0.34 0.33 0.33 0.33 0.32 0.32 0.32 0.32 0.32
125 Saudi Arabia 75 0.82 0.80 0.77 0.74 0.71 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.68 0.68 0.67
126 Senegal 156 10.87 10.70 10.34 10.06 9.83 9.63 9.50 9.35 9.25 9.20 9.11 9.00 8.95 8.96 9.02 9.08
127 Seychelles 34 0.70 0.65 0.65 0.64 0.65 0.65 0.65 0.62 0.60 0.53 0.47 0.43 0.39 0.35 0.33 0.32
128 Sierra Leone 131 2.63 2.64 2.66 2.69 2.72 2.74 2.73 2.68 2.61 2.51 2.42 2.34 2.26 2.19 2.12 2.04
129 Singapore 61 0.11 0.11 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.09 0.09 0.09 0.09
130 Sint Maarten (Dutch part) 26 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
131 Slovenia 68 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
132 Solomon Islands 14 1.19 1.19 1.18 1.17 1.12 1.13 1.17 1.17 1.19 1.17 1.09 1.00 0.92 0.85 0.74 0.72
133 Somalia 96 1.79 1.75 1.67 1.61 1.56 1.49 1.43 1.38 1.31 1.16 1.04 0.92 0.87 0.84 0.82 0.79
134 South Africa 101 15.70 15.96 16.36 16.69 16.97 17.23 17.42 17.49 17.55 17.55 17.48 17.48 17.69 17.84 18.01 18.30
135 Spain 251 16.00 16.06 16.16 16.23 16.32 16.43 16.59 16.70 16.94 17.18 17.46 17.67 17.84 18.10 18.36 18.61
136 Sri Lanka 120 6.66 6.49 6.31 6.19 6.07 5.94 5.74 5.52 5.28 4.97 4.73 4.73 4.56 4.44 4.32 4.21
137 St. Kitts and Nevis 30 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
138 St. Lucia 56 0.06 0.06 0.07 0.07 0.07 0.07 0.07 0.07 0.08 0.07 0.07 0.07 0.07 0.07 0.07 0.07
139 St. Martin (French part) 32 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
140 St. Vincent and the Grenadines 66 0.08 0.08 0.08 0.09 0.09 0.09 0.09 0.10 0.10 0.10 0.10 0.10 0.11 0.11 0.11 0.11
141 Sudan 23 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04
species
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
142 Suriname 70 1.86 1.67 1.56 1.48 1.43 1.38 1.35 1.32 1.30 1.28 1.27 1.25 1.23 1.21 1.20
143 Sweden 85 11.97 11.11 10.49 10.02 9.64 9.34 9.09 8.88 8.69 8.53 8.39 8.27 8.17 8.07 7.97
144 Syrian Arab Republic 167 13.62 13.15 12.81 12.53 12.30 12.09 11.90 11.72 11.55 11.39 11.23 11.10 10.95 10.80 10.64
145 Tanzania 91 3.12 2.81 2.62 2.49 2.38 2.31 2.24 2.19 2.15 2.12 2.09 2.06 2.04 2.02 2.00
146 Thailand 90 70.15 65.01 61.48 58.89 56.89 55.31 54.03 52.96 52.06 51.29 50.62 50.03 49.49 48.97 48.45
147 Togo 112 1.52 1.43 1.38 1.33 1.30 1.27 1.25 1.22 1.20 1.18 1.16 1.14 1.13 1.11 1.09
148 Tonga 78 0.21 0.20 0.20 0.19 0.19 0.19 0.18 0.18 0.18 0.18 0.17 0.17 0.17 0.16 0.16
149 Trinidad and Tobago 95 0.92 0.87 0.84 0.81 0.79 0.77 0.75 0.73 0.72 0.71 0.70 0.69 0.68 0.67 0.67
150 Tunisia 84 2.19 2.01 1.89 1.81 1.75 1.70 1.67 1.64 1.61 1.59 1.58 1.56 1.54 1.53 1.51
151 Turkey 112 24.20 22.10 20.72 19.75 19.06 18.48 18.05 17.66 17.37 17.14 16.94 16.78 16.66 16.59 16.41
152 Turks and Caicos Islands 29 0.35 0.33 0.32 0.30 0.29 0.28 0.27 0.26 0.25 0.25 0.24 0.24 0.24 0.24 0.24
153 Tuvalu 77 0.38 0.35 0.33 0.32 0.31 0.30 0.29 0.29 0.28 0.28 0.28 0.28 0.27 0.27 0.27
154 Ukraine 34 3.32 3.18 3.06 2.97 2.90 2.84 2.80 2.76 2.73 2.70 2.68 2.66 2.64 2.61 2.59
155 United Arab Emirates 64 2.05 1.93 1.84 1.79 1.74 1.71 1.68 1.65 1.63 1.61 1.60 1.58 1.56 1.55 1.54
156 United Kingdom 260 60.37 58.10 56.35 54.98 53.77 52.79 51.99 51.32 50.72 50.19 49.72 49.25 48.92 48.61 48.28
157 United States 557 220.20 208.40 199.93 193.14 187.63 182.83 178.68 174.86 171.74 168.94 165.99 162.98 159.82 156.97 154.72
158 Uruguay 45 3.64 3.42 3.28 3.18 3.11 3.06 3.01 2.98 2.95 2.92 2.90 2.88 2.86 2.84 2.82
159 Vanuatu 62 0.23 0.21 0.20 0.19 0.19 0.18 0.17 0.17 0.17 0.17 0.17 0.17 0.16 0.16 0.16
160 Venezuela, RB 116 10.09 9.66 9.32 9.07 8.85 8.67 8.50 8.35 8.21 8.09 7.98 7.88 7.79 7.71 7.62
161 Vietnam 121 66.52 59.08 54.67 51.77 49.75 48.18 46.91 45.92 45.12 44.45 43.84 43.18 42.64 42.04 41.40
162 Virgin Islands (US) 36 0.06 0.06 0.05 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04
163 West Bank and Gaza 26 0.09 0.08 0.08 0.07 0.07 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06
164 Yemen, Rep. 63 6.87 6.33 5.99 5.74 5.56 5.43 5.32 5.25 5.19 5.14 5.10 5.07 5.04 5.02 5.00
(Continued)

species
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
142 Suriname 70 1.18 1.17 1.15 1.14 1.12 1.10 1.07 1.05 1.02 0.99 0.97 0.94 0.93 0.91 0.90
143 Sweden 85 7.89 7.80 7.71 7.61 7.49 7.39 7.31 7.24 7.16 7.05 6.95 6.85 6.76 6.68 6.61
145 Tanzania 91 1.98 1.96 1.93 1.91 1.88 1.87 1.85 1.83 1.80 1.79 1.77 1.72 1.68 1.64 1.60
146 Thailand 90 47.88 47.11 46.30 45.46 44.53 43.57 42.56 41.52 40.50 39.64 39.15 38.76 38.37 37.50 36.86
147 Togo 112 1.08 1.06 1.05 1.03 1.02 1.00 0.99 0.97 0.96 0.94 0.92 0.91 0.90 0.89 0.89
148 Tonga 78 0.16 0.16 0.15 0.15 0.15 0.15 0.14 0.14 0.14 0.13 0.13 0.13 0.12 0.12 0.12
150 Tunisia 84 1.50 1.49 1.48 1.46 1.45 1.44 1.44 1.43 1.42 1.41 1.40 1.39 1.37 1.35 1.33
151 Turkey 112 16.28 16.10 15.98 15.72 15.60 15.44 15.28 15.16 15.05 15.00 14.99 14.99 14.95 14.89 14.71
153 Tuvalu 77 0.27 0.27 0.27 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.25
154 Ukraine 34 2.56 2.54 2.50 2.48 2.46 2.45 2.43 2.41 2.40 2.38 2.33 2.29 2.20 2.14 2.08
156 United Kingdom 260 47.94 47.39 46.80 46.21 45.68 45.22 44.39 43.73 43.02 42.19 41.55 41.13 40.58 39.84 38.86
157 United States 557 152.56 150.53 148.78 147.22 145.43 143.54 141.29 138.67 135.91 133.56 131.54 129.92 128.05 126.61 124.78
158 Uruguay 45 2.80 2.78 2.77 2.74 2.73 2.73 2.73 2.72 2.71 2.71 2.70 2.69 2.66 2.62 2.55
159 Vanuatu 62 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.15 0.15 0.15 0.15 0.15 0.14
160 Venezuela, RB 116 7.55 7.46 7.40 7.32 7.23 7.14 7.07 6.97 6.85 6.73 6.65 6.55 6.47 6.40 6.33
161 Vietnam 121 40.82 40.35 39.95 39.63 39.36 39.14 38.97 38.84 38.74 38.67 38.63 38.64 38.55 38.55 38.58
164 Yemen, Rep. 63 4.98 4.96 4.95 4.93 4.91 4.90 4.88 4.86 4.84 4.80 4.75 4.71 4.64 4.57 4.52
species
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
142 Suriname 70 0.89 0.90 0.89 0.89 0.88 0.88 0.87 0.87 0.86 0.86 0.86 0.87 0.87 0.86 0.86
143 Sweden 85 6.52 6.40 6.29 6.18 6.06 5.91 5.80 5.72 5.66 5.58 5.50 5.44 5.39 5.31 5.20
145 Tanzania 91 1.58 1.56 1.53 1.53 1.52 1.49 1.47 1.45 1.43 1.41 1.39 1.36 1.34 1.31 1.31
146 Thailand 90 36.47 36.27 36.03 35.70 35.38 35.27 35.17 34.80 34.20 33.78 33.34 32.86 32.23 31.35 30.43
147 Togo 112 0.87 0.86 0.84 0.83 0.82 0.81 0.80 0.79 0.78 0.77 0.76 0.75 0.74 0.73 0.72
148 Tonga 78 0.12 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.12 0.12 0.12
150 Tunisia 84 1.31 1.28 1.26 1.24 1.22 1.18 1.15 1.11 1.06 1.02 0.98 0.95 0.93 0.90 0.88
151 Turkey 112 14.37 14.00 13.60 13.19 12.71 12.29 11.89 11.50 11.06 10.58 10.41 10.36 10.33 10.16 9.87
153 Tuvalu 77 0.26 0.26 0.26 0.26 0.25 0.25 0.25 0.26 0.26 0.25 0.26 0.26 0.25 0.25 0.25
154 Ukraine 34 2.00 1.91 1.84 1.77 1.72 1.65 1.60 1.56 1.51 1.42 1.36 1.35 1.37 1.38 1.39
156 United Kingdom 260 38.16 37.19 36.35 35.26 34.21 33.51 32.63 31.66 30.51 29.31 28.18 27.36 26.56 25.88 25.42
157 United States 557 123.04 121.25 119.56 117.77 116.05 113.69 111.58 109.85 108.09 106.63 105.45 104.28 102.92 101.53 100.39
158 Uruguay 45 2.46 2.36 2.25 2.17 2.08 2.01 1.93 1.85 1.78 1.74 1.69 1.67 1.60 1.54 1.49
159 Vanuatu 62 0.14 0.14 0.13 0.14 0.14 0.14 0.14 0.14 0.14 0.13 0.13 0.13 0.13 0.12 0.12
160 Venezuela, RB 116 6.28 6.22 6.17 6.09 6.01 5.93 5.84 5.75 5.63 5.57 5.45 5.33 5.19 5.07 4.87
161 Vietnam 121 38.58 38.62 38.61 38.51 38.26 38.06 37.85 37.61 37.37 37.13 36.81 36.50 36.14 35.77 35.33
164 Yemen, Rep. 63 4.47 4.38 4.31 4.28 4.24 4.18 4.11 4.02 3.95 3.88 3.82 3.80 3.76 3.75 3.73
(Continued)

species
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
142 Suriname 70 0.85 0.83 0.82 0.81 0.79 0.78 0.76 0.74 0.72 0.70 0.67 0.66 0.65 0.62 0.60 0.60
143 Sweden 85 5.03 4.85 4.68 4.51 4.32 4.21 4.13 4.09 4.07 4.09 4.09 4.09 4.11 4.14 4.18 4.20
144 Syrian Arab Republic 167 5.14 5.04 4.93 4.85 4.81 4.79 4.80 4.81 4.78 4.70 4.63 4.54 4.51 4.54 4.59 4.62
145 Tanzania 91 1.30 1.30 1.26 1.25 1.23 1.20 1.18 1.15 1.12 1.10 1.07 1.04 1.03 1.01 0.99 0.97
146 Thailand 90 29.43 28.38 27.61 27.13 26.67 26.23 25.66 25.34 25.13 24.82 24.63 24.33 23.90 23.96 23.72 23.55
147 Togo 112 0.72 0.72 0.72 0.72 0.72 0.72 0.73 0.73 0.73 0.74 0.74 0.75 0.76 0.76 0.77 0.79
148 Tonga 78 0.12 0.12 0.12 0.12 0.12 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13
149 Trinidad and Tobago 95 0.35 0.35 0.36 0.36 0.36 0.37 0.37 0.37 0.37 0.37 0.37 0.37 0.38 0.38 0.38 0.38
150 Tunisia 84 0.85 0.84 0.83 0.82 0.81 0.80 0.78 0.77 0.77 0.76 0.75 0.73 0.72 0.70 0.69 0.69
151 Turkey 112 9.52 9.14 8.94 8.87 8.76 8.51 8.37 8.18 7.94 7.81 7.62 7.60 7.44 7.15 7.08 7.07
152 Turks and Caicos Islands 29 0.28 0.28 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.26 0.26 0.26 0.26
153 Tuvalu 77 0.25 0.25 0.25 0.26 0.24 0.23 0.22 0.21 0.19 0.20 0.19 0.18 0.18 0.17 0.15 0.13
154 Ukraine 34 1.39 1.39 1.40 1.40 1.40 1.40 1.38 1.33 1.30 1.28 1.26 1.23 1.21 1.21 1.20 1.17
155 United Arab Emirates 64 0.98 0.95 0.92 0.89 0.86 0.82 0.78 0.74 0.72 0.69 0.67 0.65 0.63 0.61 0.60 0.59
156 United Kingdom 260 24.94 24.23 23.76 23.19 22.65 21.80 21.20 20.76 20.45 20.17 20.01 20.03 19.97 19.57 19.27 19.24
157 United States 557 99.25 98.48 97.93 97.28 97.01 96.76 96.64 96.28 95.98 95.48 94.82 94.50 94.01 93.52 93.48 93.56
158 Uruguay 45 1.43 1.38 1.33 1.27 1.20 1.18 1.15 1.14 1.12 1.10 1.06 1.01 0.94 0.90 0.86 0.84
159 Vanuatu 62 0.12 0.12 0.12 0.12 0.11 0.11 0.11 0.11 0.11 0.10 0.10 0.10 0.10 0.09 0.09 0.08
160 Venezuela, RB 116 4.65 4.40 4.18 3.99 3.75 3.58 3.47 3.27 3.06 2.86 2.53 2.37 2.27 2.20 2.17 2.09
161 Vietnam 121 34.64 33.95 33.20 32.48 31.73 30.74 29.81 28.96 28.18 27.26 26.59 25.95 25.33 24.72 24.10 23.28
162 Virgin Islands (US) 36 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
163 West Bank and Gaza 26 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02
164 Yemen, Rep. 63 3.72 3.68 3.65 3.59 3.49 3.39 3.32 3.21 3.04 2.79 2.52 2.32 2.14 2.05 2.06 2.03
6 Inclusive wealth adjusted by
total factor productivity as a
sustainability measurement
Global productivity analysis
Isma Addi Jumbri, Moinul Islam
and Shunsuke Managi
1 Introduction
In September 2015, the United Nations General Assembly adopted the 2030
Agenda for Sustainable Development Goals (SDGs). These important goals range
from responsible consumption and production, improvements in well-being, qual-
ity education and health to the protection of global assets, including oceans and
a stable climate. As such, governments need a measure to monitor and judge
whether the development programs they undertake to meet the goals are sustain-
able (Dasgupta et al., 2015). Even though gross domestic product (GDP) is viewed
as a good measure of economic activity, GDP ignores social costs, environmental
impacts and income inequality. Oleson (2011) also argued that GDP needs to be
replaced in order to measure or as an indicator of real economic growth and well-
being. It needs the indicators that promote a truly sustainable development, that is
the development that improves the quality of human life while living within the
carrying capacity of the supporting ecosystems (Costanza et al., 2014).
Having a theoretically consistency to measure the determinants of well-being, the
inclusive wealth (IW) framework, which is also termed as genuine savings (GS),
is utilized by many stakeholders to measure sustainability progress (see Atkinson
et al., 2014; Dasgupta et al., 2015; Hamilton and Clemens, 1999; Hamilton, 2000;
Hamilton and Atkinson, 1996, 2006; Mäler, 2008; World Bank, 2006). While stud-
ies by (Arrow et al., 2004; Arrow, 2007; Arrow et al., 2012; Dasgupta, 2005; Das-
gupta, 2001, 2009; Dasgupta and Mäler, 2001; Managi, 2015; Mumford, 2016;
UNU-IHDP and UNEP, 2014, 2012; World Bank, 2010) were structured on both
market and non-market assets of all types of capitals that contribute to well-being.
Instead of cross-country, IW is also utilized to measure regional wealth and assess
sustainability development. For instance, Yamaguchi, Sato, and Ueta (2016) mea-
sured the sustainability of Japan after the tsunami disaster. In other instance, Acar
and Gultekin-Karakas (2016) used to measure Turkey’s sustainability development.
Mota, Domingos, and Martins (2010) estimated for the Portugal’s and Ollivier and
Giraud (2010) assessed the sustainability for the Madagascar. The indicator that
was used in all studies almost equivalently to IW as the index is increasingly recog-
nized as an indicator for sustainability (Yamaguchi and Managi, 2017).
136 Isma Addi Jumbri et al.
This approach together with the amount of capital represents the produc-
tive ability of well-being. They consider an intergenerational well-being as an
approach to measure sustainable development. In their formulation, a society’s
economic development is sustainable at a point of time if its wealth at constant
shadow prices is non-decreasing at that time. A non-declining IW implies the pos-
sibility of non-declining human well-being.
On the other hand, Arrow et al. (2004) argued that total factor productivity
(TFP) plays an important role to assess the sustainability development using GS
or IW. Utilizing TFP as adjustment can portray a closer approximation of the real
contribution of technological innovation and efficiency played in production, as
well as other implicit capital types that are not yet considered in developing the
countries’ IW. Utilizing capital and labour as inputs, Arrow et al. (2004) found
the contribution of TFP to IW growth rate to be between 6.33% (for China) and
−0.40% (for the Middle East and North Africa). Arrow et al. (2012) found the
contribution varied from −2.12% (in Venezuela) to 2.71% (in China).
However, previous studies on TFP as an adjustment of IW ignored the effects of
natural capital as input for productivity measures (Kurniawan and Managi, 2017).
Recent research found that natural capital has a significant contribution to cross-
country efficiency changes. For instance, Brandt, Schreyer, and Zipperer (2017)
reported that failing to account for natural capital tends to lead to an underestima-
tion of productivity growth. Tang et al. (2017) found that there is a very significant
effect on the environment’s TFP and TFP estimation. By considering the environ-
mental regulation efficiency (ERE), they find that there is significance to China’s
TFP and economic development level.
To fill the gap of incorporating natural capital to TFP as an adjustment of
IW, this study considers natural capital, including oil capital gain and other con-
ventional inputs, to productivity measures. Considering these inputs, we could
analyse how these countries differ with respect to the effective utilization of
their productive assets. We can understand that the same productive base of a
country can lead to an increase or decrease in aggregate output over time due to
productivity changes in the resource. Some countries use their endowed capi-
tal efficiently with appropriate productivity changes and future-oriented stock
consumption schemes, while others do not use their capital as efficiently as they
should.
To do so, we use a deterministic nonparametric analysis called the Malmquist
Productivity Index (MPI), based on the Data Envelopment Analysis (DEA). This
methodology is widely used in the measurement of productivity (for review, see
Coelli, Rao, and Battese, 1998; Färe et al., 1994; Kerstens and Managi, 2012;
Malik, 2015; Tanaka and Managi, 2013). According to Stern (2005), DEA is a
linear programming technique that allows both the frontier itself and the dis-
tance from the frontier of each country to vary in every time period arbitrarily.
Therefore, this index that is based on distance is suitable for assessing the relation
between multivariate inputs and outputs for 140 countries from 1990 to 2010. In
addition, the measurement takes into account the efficiency of resources used and
productivity changes.
Inclusive wealth adjusted by TFP 137
This chapter is organized with the following sections. Section 2 provides an
overview of the literature, focusing on sustainability and wealth, highlighting the
previous empirical research on the cross-country productivity and the method
evolution which reinforced the contribution of this study. Section 3 describes the
methodology and data. Section 4 presents empirical results and efficiency decom-
position. Section 5 draws conclusions and suggests possibilities to extend this
research.
2 Empirical research
There are several empirical studies that provide evidence of sustainable develop-
ment based on the historical average or current level GS. For instance, Hamilton
and Clemens (1999) calculated the GS average for the 1970s and 1980s and also
provided some single-year values for the 1990s across countries. In order to deter-
mine whether each country is sustainability or not, they based it on the GS indica-
tor, either positive or negative.
Neumayer (2000) remodified the GS calculation method of the World Bank
by introducing an alternative approach of computing resource depletion. He esti-
mated the GS rates and considered its average as a benchmark of the sustainable
development of a country. Based on his study of the World Bank database, the
GS of some countries considered as unsustainable, but in reality, it is in progress.
Scheffer et al. (2000) also illustrated the assumption of economic optimality on
the effect of common practices towards the ecosystem use on society and environ-
ment that require the sustainable development strategies. They also demonstrated
that it is possible to extend the GS model on environmental issues. A study by
Dubash (2012) also argued that sustainable development has always been a com-
promise formulation that papered over the real conflict between environment and
development. This is because the framework of sustainable development expands
not only economically and socially but also environmentally (Clémençon, 2012).
Hamilton and Atkinson (2006) posited that positive IW values referring to GS
previously depended strongly on population growth rates and Yamaguchi (2014)
considered the effect of population age distribution in IW accounting. Mumford
(2016) in his study focusing on Asian countries found that there is a strong rela-
tionship between growth in GDP per capita and IW per capita in different signs.
Nonetheless, all these studies excluded TFP in the measurement.
Arrow et al. (2004) discussed whether economic development in some of the
major countries and regions was sustainable and argued that a country would
not meet the sustainability development criterion if that country’s average of
GS is negative. Arrow et al. (2004) also suggested that TFP plays an important
role in accessing sustainability development using GS or IW. The amount of IW
is calculated based on the weighted summation of all types of capital using the
shadow prices. In their study, Arrow et al. (2004) used capital and labour as the
inputs and GDP as the output and applied the conventional method in calculating
the TFP. Data on human and produced capital is measured based on Collins and
Bosworth (1996) and Klenow and Rodriguez-Clare (1997). The results indicated
the contribution of TFP to the IW growth rate (i.e., GS per inclusive wealth)
to be between −0.40% (Middle East and North Africa) and 6.33% (China). In
their study, Arrow et al. (2012) found the contribution to be between −2.12%
(Venezuela) and 2.71% (China). However, their measures did not consider the
contribution of natural capital as an input for productivity measures.
However, the above studies did not make any adjustment or correction on
the TFP even though to weight up natural capital as an explicit factor of input.
According to Jorgenson and Griliches (1967), adjustment of capital stock data
reflects changes in capacity utilization (CU). For instance, Denison (1979) used
the variation in capital share income, and Norsworthy, Harper, and Kunze (1979)
adjusted productivity by selecting the interval from which CU is believed to be
nearly 1. Morrison (1985) attempted to adjust the productivity changes by divid-
ing the productivity growth by the cost CU measure.
Such results indicate a divergence in cross-country productivity, wealth and
adjustment of capital stock productivity. However, to date there is no study on
cross-country TFP analysis that considers natural capital as an explicit factor of
input and includes oil capital gain as an adjustment. Several studies found that
oil price plays a significant role in economic and productivity measurement. For
instance, a study by Hesary et al. (2013) found that oil-producing countries like
Iran and Russia benefit from oil price shocks. For oil-consuming economies the
effect is more diverse. Estrada and Hernández de Cos (2011) suggested that fluctu-
ation in oil prices can significantly affect the potential output. In this respect, Finn
(2016) showed theoretically how, under certain conditions, a permanent increase in
oil prices lowers the equilibrium level of the capital stock in the long term.
Therefore, in this study, we provide a deeper analysis of how each input and
output affects countries’ performance and sustainability development by using
an unconventional TFP measurement. Besides considering natural capital as an
explicit factor of input, we incorporate oil capital gain as an adjustment of TFP.
3 Methodology and data
3.1 Inclusive wealth measurement

Aggregated capital is a term of inclusive investment or IW. According to Arrow
et al. (2003), an economy enjoys sustainable development if and only if, relative
to its population, its IW is at a constant price or does not decline. IW is a measure
of intergenerational well-being, where the accumulation of wealth corresponds to
sustained development and a key to economic progress (Dasgupta, 2007). Thus,
we can define IW at the time (t) as Wt that includes produced, human and natural
capital. This also can be defined as:
W (t ) = Q (t )t + ∑Pi (t ) K i (t ) = PPC PC (t ) + PHC HC (t ) + PNC NC (t )

i
where Q(t) is the shadow price of the time asset (TFP in our case). Each P is a
shadow price of the capital asset, defined as δWt /δC; K is each capital, where
PC(t), HC(t) and NC(t) represent production capital, human capital and natural
capital at time t, respectively. Therefore, the total capital asset is its quantity
multiplied by the present value of the flow of social benefits extra unit that
would be able to be generated over time. That present value is called an asset’s
shadow price. Hence, an economy’s IW is the shadow value of its productive
base, and inclusive investment, IW(t), is the shadow value of the net change in
its productive base (Dasgupta, 2008). All capitals are evaluated based on each
of their own accounting prices in the current period. The IW should be positive
and contrasted from recorded investment; thus, the sustainability condition can
be expressed as:
IW (t ) ≥ 0 for all t (time)
3.2 TFP measurement

To measure the cross-country productivity, we adopt the Malmquist Productivity
Index (MPI) as described by Coelli, Rao, and Battese (1998) and Färe et al. (1994)
as a measurement of TFP to investigate mapping from capital assets (human capi-
tal, produced capital and natural capital) to change in wealth. This is because the
MPI is suitable for assessing the relationship between inputs and outputs under
multivariate input inefficiency. By using the MPI, technological progress and the
inefficiency of resource use also can be estimated. Therefore, MPI can be consid-
ered the most suitable index for achieving the objectives of this research.
When calculating TFP, we apply produced, human and natural capital as a
separate unit, each capital as an input and GDP as an output. Because of the mul-
tivariate inputs – that is, produced, human and natural capital – we adapt the
nonparametric frontier analysis of Data Envelopment Analysis (DEA) to measure
TFP. By using the distance function specification, our problem can be formulated
as follows:
T (t ) ≡ {( xt , yt ) : xt can produce yt } (1)
In this case, x = ( x1 ,… x M ) ∈ R+M and y = ( y1 ,… x N ) ∈ R+N be the input and out-

put vectors, respectively. The technology set is defined by the equation (2), which
consists of all feasible input vectors, xt and output vectors, yt, at time t. According
to Managi (2011), the estimation of efficiency relative to the production frontiers
relies on the theory of distance or gauge functions. In economics, distance func-
tions are related to the notion of the coefficient of resources utilization (Debreu,
1951) and to efficiency measures (Farrell, 1957). This distance function is defined
t as:
d t ( yt , xt ) = min {δ : ( xt , yt / δ ) ∈ T (t )} (2)
Where δ is the maximal proportional amount to which yt can be expanded given

technology T(t). Data Envelopment Analysis (DEA) and the output-oriented
function are used to estimate the distance function under constant returns to scale
(CRS) by solving the following optimization problem (Managi, 2003).
dT (t ) ( xt , yt ) = maxδ,λδ
yit
s.t Ytλ ≥
δ
X tλ ≤ xit
λ ≥ 0 (3)
From formula (3), δ is the measure of efficiency for country i in year t. λ is an

N x 1 vector of weights, Yt and Xt and inputs are the vectors of outputs yt and inputs
xt. To estimate productivity over time, several distance functions are used for the
input-output vector for period t + 1 and technology in period t. The MPI (M0) for
the output-oriented productivity index is as follows:
1
 d t ( y x ) d t +1 ( y x ) 2
M 0 ( yt , xt , yt +1 , xt +1 ) =  t
o t +1, t +1 o t +1, t +1 
× t +1 (4)
d
 o t t( y , x ) d ( yt , xt ) 
o 
In formula (4), d represents the geometric distance to the frontier, which can
be decomposed into efficiency change (i.e., catching up to the frontier) and the
technological change (i.e., change in the production frontier; Färe et al., 1994).
Based on this formula, a country on the frontier will perform better under the
same resource constraints than a country that is further away from the frontier.
The estimation of this is performed by using DEA, a nonparametric estimation
method. M0 can be divided into two components. Efficiency change represents
the first ratio, and technological changes represent the second ratio. Based on this
formulation, we estimate the MPI and estimated IW as follows:
f : { PPC PC , PHC HC , PNC NC } Inclusive wealth (IW)(5)
In the estimation, it is possible for two countries to have different levels of

sustainable development even if they have a similar level of capital assets. This
is due to the fact that sustainable development is dependent not only on how the
countries use their capital assets, but also on their different saving rates. The MPI
is reformulated as follows:
M 0 (GDPi ,t , PCi ,t , HCi ,t , NCi ,t , GDPi ,t +1 , PCi ,t +1 , HCi ,t +1 , NCi ,t +11 )

1
 d t (GDP , PC , HC , NC ) d t +1 (GDP , PC , HC , NC ) 2
= 
i ,t +1 
i ,t i ,t +1 i ,t +1 i ,t +1
×
i ,t +1 i ,t +1 i ,t +1
(6)

 d t
(GDP i ,t , PC i ,t , HC i ,t , NCi ,t ) d t +1
(GD P i ,t , PC i ,t , HCi ,t , NCi ,t ) 
This equation indicates the feasibility of producing goods with fewer inputs.
Where d is the geometric distance to the production frontier, it represents the best
available technology for the given inputs and output. The country under analysis
in this research refers to i, which runs from 1 to 140 countries in our sample.
GDP is the corresponding value of the gross domestic product; PC stands for
produced capital; HC represents human capital; and NC stands for natural capital.
This index is measured by the ratio between 2 years; thus, a value greater than 1
represents an improvement or otherwise in the TFP calculation. Using this index
allows us to examine TFP’s contribution to IW by country.
3.3 Inclusive wealth adjusted by TFP measurement

By using the new TFP results and based on the IW methodology explained, we esti-
mate each country’s Inclusive Wealth-TFP Adjusted (IW-TFP Adjusted) for the years
1990–2010. We do so by using the annual average of IW per capita and estimate the
percentage change of TFP on the IW as the sustainability indicator. In order to estimate
the IW-TFP Adjusted, we calculate the growth rate of IW per capita. We also follow
the same process as done by Arrow et al. (2004), with the compromise of the use of
TFP growth from Collins and Bosworth (1996), which was based on GDP output.
According to the framework of sustainable development, the IW-TFP Adjusted should
be maintained at a positive and non-declining rate. The calculation is as follows:
IW ~ TFP Adjusted = IW + TFP
IW~TFP Adjusted Inclusive Wealth Adjusted with TFP

IW Inclusive wealth growth per capita
TFP TFP growth
Based on TFP results from 1990 to 2010, we measure not only IWI-TFP
Adjusted but also how countries’ performance affected IW after three main fac-
tors were taken into consideration. Those three factors are (1) climate change,
particularly the damage suffered as the result of increased atmospheric carbon;
(2) TFP’s contribution to multiple factors missing from economic growth; and
(3) oil capital gains related to the change in oil prices, which may either increase
or decrease the country’s productivity value.
Based on the 2012 and 2014 IWR, we refer to this adjusted figure as the Inclu-
sive Wealth Index Adjusted (IWI Adjusted). We measure the change in wealth
for 140 countries over 21 years on a per capita basis. To estimate the change
in wealth, we calculate the average annual growth rates in wealth and popula-
tion. According to the framework of sustainable development, the IWI-Adjusted
should remain positive and non-declining. The calculation is as follows:
IWI ~ Adjusted = IW + C + E + TFP
IWI~Adjusted Inclusive Wealth Index Adjusted

IW Inclusive wealth per capita
C Carbon damage
E Energy depletion, i.e., oil capital gain
TFP Percentage of TFP growth
Based on this adjustment, the IWI-Adjusted can have either a positive or a neg-
ative effect on countries’ IW. According to Nordhaus and Boyer (2000), climate
change will benefit some countries, whereas others will experience a negative
impact. In terms of oil price fluctuations, normally countries that are considered
oil producers will realize advantages and positive changes in wealth compared to
countries that depend on oil imports.
3.4 Data
For the sources of these data, we refer to the 2014 Inclusive Wealth Report pub-
lished by UNU-IHDP and UNEP. This dataset provides quantitative information
on data for 140 countries from 1990 to 2010. As noted earlier, the accounting
shadow balance sheet included as inputs here are not only produced capital
(PC) and human capital (HC), but also natural capital (NC). To measure NC
wealth, fossil fuels (oil, coal and natural gas), minerals (nickel, gold, silver,
iron, lead, tin, zinc, phosphate, copper and bauxite), forest resources (timber
and non-timber) and agricultural land (cropland and pastureland) are used. For
HC wealth, the calculation uses education attainment and the additional com-
pensation over time as developed by Arrow et al. (2012). The BP Statistical
Review of World Energy (BP, 2013) is used for prices of coal, natural gas and
oil. Equipment, machinery and road data are calculated for PC. In this research,
we utilize GDP as an output. All capital wealth in this calculation is based on
million constant 2005 US dollars.
4 Empirical results
4.1 Total Factor Productivity (TFP)

Based on the estimating procedure explained in Section 3, we estimate each coun-
try’s TFP, efficiency change and technical change for the 21 years included in the
dataset. The means of each country’s TFP change, efficiency change and technical
change from 1990 to 2010 are presented in Map 6.1 and Table 6.1.
Table 6.1 and Map 6.1 shows the TFP change for the 140 countries from 1990
to 2010. The countries in the table are presented in descending order based on
the magnitude of TFP changes. The table shows Singapore and China as the two
countries with the highest TFP growth. Singapore indicates 10% average growth
in TFP (caused by steady growth in efficiency) and 10% growth in technical
change. The highest TFP growth of Singapore contributes to its rapid economic
development. According to Han (2017), the rapid economic development of Sin-
gapore has been attributed to its plan-rational technocratic elite and its environ-
mental policy resulted in its international reputation as a model green city with
a remarkable expansion of green spaces and infrastructure. Australia, the United
Kingdom, India and Japan exhibit TFP growth rates of 0.5%, 3.2%, 0.8% and
0.5%, respectively. The unweighted average growth in TFP across 140 countries
is 0.7%.
>1
0.93 - 1
0.87 - 0.93
0.8 - 0.87
<0.8
No data
Map 6.1 Average of TFP for 140 countries (1990–2010)

Table 6.1 Ranking and summary of estimated TFP
Ranking Country TFP Ranking Country TFP Ranking Country TFP
1 Singapore 1.100 2 China 1.052 3 Trinidad and Tobago 1.038

4 Qatar 1.036 5 Nigeria 1.034 6 United Republic of Tanzania 1.034
7 United Kingdom 1.032 8 Albania 1.030 9 Peru 1.025
10 Thailand 1.025 11 Cambodia 1.025 12 Mozambique 1.025
13 Argentina 1.024 14 Botswana 1.024 15 Mali 1.024
16 Uganda 1.024 17 Ghana 1.023 18 Syrian Arab Republic 1.021
19 Tunisia 1.020 20 Yemen 1.020 21 Jordan 1.019
22 Namibia 1.018 23 Nepal 1.018 24 Iran 1.017
25 Malawi 1.017 26 Niger 1.017 27 Ecuador 1.016
28 Panama 1.016 29 Swaziland 1.016 30 Benin 1.016
31 Slovakia 1.015 32 Cuba 1.015 33 Dominican Republic 1.015
34 Belgium 1.014 35 Ireland 1.014 36 Maldives 1.014
37 Belize 1.014 38 Sri Lanka 1.014 39 Sudan 1.014
40 Lesotho 1.013 41 Nicaragua 1.013 42 Papua New Guinea 1.013
43 Kuwait 1.012 44 Guyana 1.012 45 Republic of Korea 1.011
46 Malaysia 1.011 47 Uruguay 1.011 48 Indonesia 1.011
49 Iraq 1.011 50 Luxembourg 1.010 51 Norway 1.010
52 Armenia 1.010 53 Paraguay 1.010 54 Myanmar 1.010
55 Rwanda 1.010 56 Cyprus 1.009 57 Czech Republic 1.009
58 Greece 1.009 59 Brazil 1.009 60 Congo 1.009
61 Guatemala 1.009 62 Pakistan 1.009 63 Mauritania 1.009
64 Mauritius 1.008 65 (Bolivarian Republic of) Venezuela 1.008 66 Cameroon 1.008
67 India 1.008 68 Philippines 1.008 69 Austria 1.007
70 Colombia 1.007 71 Côte d’Ivoire 1.007 72 Egypt 1.007
73 Finland 1.006 74 Israel 1.006 75 Algeria 1.006
76 El Salvador 1.006 77 Togo 1.006 78 Australia 1.005
79 Japan 1.005 80 Netherlands 1.005 81 New Zealand 1.005
82 Poland 1.005 83 Bulgaria 1.005 84 Costa Rica 1.005
85 (Plurinational State of) Bolivia 1.005 86 Honduras 1.005 87 Senegal 1.005
88 Sweden 1.004 89 Morocco 1.004 90 Iceland 1.003
91 Malta 1.003 92 South Africa 1.003 93 Mongolia 1.003
94 Bangladesh 1.002 95 Bahrain 1.001 96 Italy 1.001
97 Saudi Arabia 1.001 98 Chile 1.001 99 Romania 0.993
100 Fiji 1.001 101 Barbados 1.000 102 Canada 0.990
103 Afghanistan 1.000 104 Central African Republic 1.000 105 Switzerland 0.999
106 United Arab Emirates (UAE) 0.999 107 Gabon 0.999 108 Turkey 0.999
109 Denmark 0.998 110 France 0.998 111 Germany 0.998
112 Slovenia 0.998 113 United States 0.998 114 Liberia 0.998
115 Estonia 0.997 116 Hungary 0.997 117 Burundi 0.997
118 Kenya 0.997 119 Mexico 0.996 120 Zambia 0.996
121 Spain 0.995 122 Jamaica 0.993 123 Lithuania 0.993
124 Russian Federation 0.993 125 Portugal 0.992 126 Kazakhstan 0.990
127 Lao People’s Democratic Republic 0.990 128 Sierra Leone 0.989 129 Vietnam 0.988
130 Zimbabwe 0.987 131 Kyrgyzstan 0.985 132 Latvia 0.984
133 Croatia 0.981 134 Republic of Moldova 0.981 135 Democratic Republic of Congo 0.980
136 Haiti 0.977 137 Tajikistan 0.975 138 Serbia 0.973
139 The Gambia 0.968 140 Ukraine 0.966 Mean 1.007
Figure 6.1 shows the annual mean TFP change, technical change and efficiency
change for over the entire period (1990 to 2010) for 140 countries. On average,
TFP values were increasing in many countries throughout the period of study. The
worst TFP change during the study period is from 2008 to 2009 and is attributable
to the global financial crisis. The highest technical and efficiency change are in
1992–1993 and 2002–2003, respectively.
Table 6.2 shows the mean TFP change, efficiency change and technical change
results based on the different regions from 1990 to 2010. It is clear that Asia shows
the highest TFP growth (1%), followed by Latin America and the Caribbean (0.9%).
North America shows the lowest (and decreasing) TFP, followed by Europe (0.1%).
Figure 6.2 shows the cumulative TFP indices from 1990 to 2010 based on six
regions. Asia clearly experienced the highest cumulative growth, especially in
2010, followed by Africa and Latin America and the Caribbean. North America and
TFP Change Technical Change Efficiency Change

1.06
1.04
1.02
MEAN
0.98
0.96
0.94
1990- 1992- 1994- 1996- 1999- 2000- 2002- 2004- 2006- 2008-
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009
YEAR
Figure 6.1 Mean TFP change, efficiency change and technical change (1990–2010)
Table 6.2 Means of TFP, efficiency change and technical change (1990–2010) based on regions
Regions TFP change Efficiency Technical

change change
Africa 1.008 1.004 1.005
Asia 1.010 1.000 1.010
Europe 1.001 0.990 1.011
Latin America and the Caribbean 1.009 1.001 1.008
North America 0.999 0.992 1.007
Oceania 1.006 1.000 1.006
Mean 1.007 0.999 1.008
Figure 6.2 Mean TFP change in six regions (1990–2010)
Oceania were at the bottom, whereas Asia was still experiencing the highest cumu-
lative growth than the global growth in TFP. According to Clémençon (1997), eco-
nomic integration and free trade are factors that further boost the economic growth of
Asia countries, especially in the Association of Southeast Asian Nations (ASEAN).
Sonnenfeld and Mol (2006) argued that the Asia-Pacific region is unquestionable
as one of the most economically dynamic areas in the world. For instance, coun-
tries like China, India, Malaysia, Vietnam and Thailand indicated steady economic
growth in two decades, even facing the financial crisis in the late 1990s.
4.2 Inclusive Wealth considering TFP (IW-TFP adjusted)

Using the results of the new TFP and based on the method described, we estimate
each country’s IW-TFP Adjusted for 1990–2010. In this study, we use each annual
per capita IW and estimate the percentage change in TFP on the IW as an indicator of
sustainable development. We calculate the growth rate of per capita IW to estimate
the IW-TFP Adjusted. We follow the same process as in Arrow et al. (2004), com-
promising through the use of TFP growth from Collins and Bosworth (1996), which
is based on GDP output. The result of the IW-TFP Adjusted is shown in Table 6.3.
Table 6.3 shows the post-IW-TFP adjustment for the 140 countries over the
21 years from 1990 to 2000. The table demonstrates that among those 140 coun-
tries, 100 countries, representing 71% of nations, showed a positive IW-TFP
Adjusted. Before adjustment, 85 countries demonstrated positive IW; after modi-
fication, 25 countries moved from the negative to the positive bracket, and only 10
nations experienced the reverse. The following countries moved from negative to
Table 6.3 Result of IW-TFP adjusted per capita and percentage of change
Countries After adjusted % change Countries After adjusted % change

Afghanistan −0.825 −0.8% Liberia −3.565 −3.6%
Albania 3.759 3.8% Lithuania 0.594 0.6%
Algeria −0.156 −0.2% Luxembourg 2.521 2.5%
Argentina 2.724 1.3% Malawi 0.389 0.4%
Armenia 2.027 2.0% Malaysia 1.538 1.5%
Australia 0.829 0.8% Maldives 4.652 4.6%
Austria 1.841 1.8% Mali 0.768 0.7%
Bahrain 0.943 0.9% Malta 1.954 1.9%
Bangladesh 1.629 1.6% Mauritania 1.190 1.2%
Barbados 0.623 0.6% Mauritius 2.094 2.1%
Belgium 2.439 2.4% Mexico 0.516 0.5%
Belize 0.006 0.01% Mongolia −0.910 −0.9%
Benin 0.313 0.3% Morocco 1.519 1.5%
Bolivia −1.593 −1.5% Mozambique 0.022 0.02%
Botswana 2.203 2.2% Myanmar −0.326 −0.3%
Brazil 1.013 1.0% Namibia 1.001 1.0%
Bulgaria 1.451 1.5% Nepal 0.199 0.2%
Burundi −0.413 −0.4% Netherlands 1.428 1.4%
Cambodia 1.146 1.1% New Zealand 0.980 1.0%
Cameroon −0.987 −1.0% Nicaragua 0.674 0.6%
Canada 0.282 0.3% Niger 0.611 0.6%
Central African −2.095 −2.1% Nigeria 1.565 1.6%
Republic
Chile 1.224 1.2% Norway 1.299 1.3%
China 7.163 7.2% Pakistan 1.557 1.6%
Colombia 0.605 0.6% Panama 2.219 2.2%
Congo −1.809 −1.8% Papua New −1.468 −1.5%
Guinea
Costa Rica 1.298 1.3% Paraguay −0.143 −0.1%
Côte d’Ivoire 0.160 0.2% Peru 1.928 1.9%
Croatia −0.974 −1.0% Philippines 1.307 1.3%
Cuba 1.616 1.6% Poland 1.618 1.6%
Cyprus 1.891 1.9% Portugal 0.216 0.2%
Czech Republic 2.166 2.2% Qatar −0.496 −0.5%
Democratic −4.736 −4.7% Republic of 3.467 3.4%
Republic of Congo Korea
Denmark 0.291 0.3% Republic of −2.082 −2.1%
Moldova
Dominican 2.519 2.5% Romania 1.129 1.2%
Republic
Ecuador 0.641 0.6% Russian −0.554 −0.6%
Federation
Countries After adjusted % change Countries After adjusted % change
Egypt 1.357 1.4% Rwanda 1.635 1.6%

El Salvador 2.188 2.2% Saudi Arabia −1.278 −1.3%
Estonia 1.422 1.4% Senegal −0.546 −0.5%
Fiji 0.802 0.8% Serbia −1.966 2.0%
Finland 1.294 1.3% Sierra Leone −1.990 −2.0%
France 0.888 0.8% Singapore 11.923 12.0%
Gabon −1.820 −1.8% Slovakia 2.642 2.6%
The Gambia −3.169 3.2% Slovenia 1.157 1.2%
Germany 1.276 1.3% South Africa 0.395 0.4%
Ghana 1.534 1.5% Spain 1.418 1.4%
Greece 1.889 1.9% Sri Lanka 2.588 2.5%
Guatemala 1.175 1.2% Sudan −0.201 −0.2%
(former)
Guyana 1.107 1.1% Swaziland 1.929 1.9%
Haiti −1.363 −1.4% Sweden 1.034 1.0%
Honduras −0.135 −0.1% Switzerland 0.337 0.3%
Hungary 0.992 1.0% Syrian Arab 1.783 1.8%
Republic
Iceland 0.329 0.3% Tajikistan −3.562 −3.6%
India 1.557 1.6% Thailand 3.752 3.7%
Indonesia 1.080 1.0% Togo −0.023 −0.02%
Iran 0.977 1.0% Trinidad and 3.590 3.6%
Tobago
Iraq −1.886 −1.9% Tunisia 3.137 3.1%
Ireland 2.936 2.9% Turkey 0.816 0.8%
Israel 1.488 1.5% Uganda 2.064 2.1%
Italy 0.903 0.9% Ukraine −3.002 −3.0%
Jamaica −0.027 −0.03% United Arab −3.315 −3.3%
Emirates
Japan 1.406 1.4% United 4.055 4.1%
Kingdom
Jordan 2.608 2.6% United 1.046 1.0%
Republic of
Tanzania
Kazakhstan −0.670 −0.7% United States 0.401 0.4%
Kenya −0.095 −0.1% Uruguay 1.844 1.8%
Kuwait −0.278 −0.3% Venezuela −0.312 −0.3%
Kyrgyzstan −1.341 −1.3% Vietnam 0.735 0.7%
Lao People’s −2.536 −2.5% Yemen 1.590 1.5%
Democratic
Republic
Latvia 0.111 0.1% Zambia −2.779 −2.8%
Lesotho 2.205 2.2% Zimbabwe −2.438 −2.4%
positive: Belize, Benin, Botswana, Cambodia, Colombia, Côte d’Ivoire, Ecuador,
Ghana, Guyana, Indonesia, Iran, Malawi, Mali, Mozambique, Namibia, Nepal,
Nicaragua, Niger, Nigeria, Peru, the Syrian Arab Republic, Trinidad and Tobago,
Uganda, Tanzania and Yemen. The following countries moved from positive to
negative: Croatia, The Gambia, Haiti, Jamaica, Kazakhstan, Kenya, Kyrgyzstan,
Russia, Serbia and Ukraine. The movement of several countries from negative to
positive positions proved that TFP also can be considered one of the factors that
has an impact on the result. Maps 6.2 and 6.3 show the percentage of growth rate
(per capita) on IW before and after TFP adjustment.
>1
0 to 1
<0
No data
Map 6.2 Before IW-TFP adjusted
>1
0 to 1
<0
No data
Map 6.3 After IW-TFP adjusted

4.3 Inclusive Wealth Index Adjusted (IWI-Adjusted)
We also measure how countries’ performance is affected by three main factors
(carbon damage, oil capital gains, and TFP) that are taken into consideration
for the adjustment. We refer to this new adjustment as IWI-Adjusted and mea-
sure all the countries in the sample on a per capita basis. We also calculate the
annual average growth rate in wealth and population to estimate the change in
wealth.
Before the new adjustment, 55 countries experienced negative growth in per
capita IW, representing almost 39% of the 140 countries in this study. Eighty-five
out of 140 countries show a positive growth rate in wealth, representing almost
77% of the total sample. After adjusting three main factors (carbon damage, oil
capital gains, and the latest TFP results), the number of countries showing positive
growth rates in wealth increased from 85 to 101. Twenty-seven countries moved
from the negative to the positive bracket after the wealth adjustment. The fol-
lowing 11 countries move from the positive to the negative bracket after all three
factors are taken into consideration: Croatia, The Gambia, Haiti, Jamaica, Kenya,
Kyrgyzstan, Latvia, Portugal, Russian Federation, Serbia and Ukraine. Of the 140
countries, 40 (almost 29%) remain in the negative bracket after the adjustment.
These 40 countries will be confronted by long-term issues related to sustaining
their current consumption patterns.
Maps 6.4 and 6.5 show the following 27 countries that move from the nega-
tive to the positive bracket after all three factors are adjusted: Algeria, Botswana,
Cambodia, Ecuador, Ghana, Guyana, Indonesia, Iran, Iraq, Kuwait, Malawi,
Mali, Namibia, Nicaragua, Niger, Nigeria, Peru, Qatar, Saudi Arabia, Sudan, Syr-
ian Arab Republic, Trinidad and Tobago, Uganda, United Republic of Tanzania,
Venezuela and Yemen. Maps 6.4 and 6.5 show the percentages of growth rate (per
capita) on IW before and after the three component adjustment.
>1
0 to 1
<0
No data
Map 6.4 Before IWI adjusted

>1
0 to 1
<0
No data
Map 6.5 After IWI adjusted
Among the three main factors, TFP makes the biggest contribution, moving
25 countries from the negative to the positive bracket after the energy depletion
factor, which takes oil capital gains into consideration. In the case of climate
change, most of the countries in this study would experience negative signifi-
cance. Thus, TFP can be considered one of the factors that significantly contrib-
utes to several countries’ movement from the negative to the positive bracket.
Table 6.4 shows the results after TFP is adjusted per capita and percentage of
change.
4.4 G7 countries
In this study of the years 1990–2010, we also learn something from the major
development economies of the G7 countries. The countries of the G7 are Can-
ada, France, Germany, Italy, Japan, the United Kingdom and the United States.
Figure 6.3 shows the wealth composition of the G7 countries from 1990 to
2010. Most of the G7 countries have human capital advantages that can be con-
sidered a major contributor to their TFP and economic growth. In the case of
Canada, its natural capital makes the second-highest wealth contribution after
human capital, representing 31% of Canada’s IW. The UK’s growth in produc-
tivity contributes more than its human capital. This success can be attributed to
tougher competition in the product and labour markets, an increase in higher
education and faster adoption of information, communication and innovation
policies (Aghion et al., 2013).
Table 6.4 Results after TFP adjusted per capita and percentage of change
Countries After % change Countries After % change

adjusted adjusted
Afghanistan −1.025 −1.0% Liberia −3.566 −3.7%
Albania 3.535 3.5% Lithuania 0.248 0.2%
Algeria 0.667 0.6% Luxembourg 1.964 2.0%
Argentina 2.492 2.5% Malawi 0.089 0.1%
Armenia 1.164 1.1% Malaysia 1.363 1.4%
Australia 0.751 0.7% Maldives 3.936 4.0%
Austria 1.566 1.6% Mali 0.642 0.6%
Bahrain 0.622 0.6% Malta 1.366 1.4%
Bangladesh 1.332 1.3% Mauritania 0.504 0.5%
Barbados 0.329 0.3% Mauritius 1.571 1.6%
Belgium 2.058 2.1% Mexico 0.087 0.1%
Belize −0.333 −0.3% Mongolia −1.005 −1.0%
Benin −0.021 −0.02% Morocco 1.154 1.2%
Bolivia −1.630 −1.6% Mozambique −0.076 −0.1%
Botswana 1.905 1.9% Myanmar −0.478 −0.5%
Brazil 0.810 0.8% Namibia 0.843 0.8%
Bulgaria 0.993 1.0% Nepal −0.001 −0.001%
Burundi −0.671 −0.7% Netherlands 1.034 1.0%
Cambodia −0.871 −0.9% New Zealand 0.831 0.8%
Cameroon −1.076 −1.1% Nicaragua 0.233 0.2%
Canada 1.077 1.1% Niger 0.381 0.4%
Central African Republic −2.130 −2.1% Nigeria 3.501 3.5%
Chile 0.853 0.8% Norway 1.473 1.5%
China 6.942 7.0% Pakistan 1.140 1.1%
Colombia 0.509 0.5% Panama 1.430 1.4%
Congo −1.201 −1.2% Papua New Guinea −1.521 −1.5%
Costa Rica 1.018 1.0% Paraguay −0.324 −0.3%
Côte d’Ivoire −0.077 −0.1% Peru 1.803 1.8%
Croatia −1.166 −1.2% Philippines 0.711 0.7%
Cuba 0.886 0.9% Poland 1.426 1.4%
Cyprus 1.444 1.4% Portugal −0.078 −0.1%
Czech Republic 1.964 2.0% Qatar 3.526 3.5%
Democratic Republic −4.755 −4.8% Republic of Korea 2.921 3.0%
of Congo
Denmark 0.070 0.1% Republic of −2.501 −2.5%
Moldova
Dominican Republic 2.018 2.0% Romania 0.883 1.0%
Ecuador 1.618 1.6% Russian Federation −0.400 −0.4%
(Continued)
Table 6.4 (Continued)
Countries After % change Countries After % change

adjusted adjusted
Egypt 0.904 0.9% Rwanda 1.205 1.2%

El Salvador 1.756 1.8% Saudi Arabia 2.204 2.2%
Estonia 1.171 1.1% Senegal −1.004 −1.0%
Fiji 0.307 0.3% Serbia −2.049 −2.0%
Finland 1.002 1.0% Sierra Leone −2.352 −2.4%
France 0.605 0.6% Singapore 10.804 11.0%
Gabon −1.083 −1.1% Slovakia 2.448 2.4%
The Gambia −3.709 −3.7% Slovenia 0.962 1.0%
Germany 0.991 1.0% South Africa 0.115 0.1%
Ghana 1.058 1.1% Spain 1.110 1.1%
Greece 1.442 1.4% Sri Lanka 2.226 2.2%
Guatemala 0.864 0.9% Sudan (former) −1.185 −1.2%
Guyana 1.024 1.0% Swaziland 1.714 1.7%
Haiti −1.845 −1.8% Sweden 0.737 0.7%
Honduras −0.449 −0.4% Switzerland 0.171 0.2%
Hungary 0.797 0.8% Syrian Arab 1.428 1.4%
Republic
Iceland 0.099 0.1% Tajikistan −4.670 −4.7%
India 1.169 1.2% Thailand 3.117 3.1%
Indonesia 0.822 0.8% Togo −0.414 −0.4%
Iran 3.349 3.3% Trinidad and 3.778 3.8%
Tobago
Iraq 3.946 4.0% Tunisia 2.891 2.9%
Ireland 2.591 2.6% Turkey 0.509 0.5%
Israel 1.267 1.3% Uganda 1.645 1.6%
Italy 0.577 0.6% Ukraine −3.257 −3.3%
Jamaica −0.657 −0.7% United Arab −0.030 −0.03%
Emirates
Japan 1.231 1.2% United Kingdom 3.805 3.8%
Jordan 1.629 1.6% United Republic 0.792 0.8%
of Tanzania
Kazakhstan 0.536 0.5% United States 0.182 0.2%
Kenya −0.518 −0.5% Uruguay 1.533 4.4%
Kuwait 6.694 6.7% Venezuela 2.005 2.0%
Kyrgyzstan −1.970 −2.0% Vietnam 0.400 0.4%
Lao People’s Democratic −2.564 −2.6% Yemen 1.913 1.9%
Republic
Latvia −0.190 −0.2% Zambia −2.860 −2.9%
Lesotho 2.008 2.0% Zimbabwe −2.753 −2.8%
120
100
PERCENTAGE
80
60
40
20
0
Canada Japan France Germany Italy United United
Kingdom States
COUNTRY
Produced Capital Human Capital Natural Capital
Figure 6.3 Percentage of wealth composition of G7 countries (1990–2010)
The combination of human and produced capital in Japan has supported its
growth and efficiency increase in other components throughout the study’s period.
This is due to the factor that labour productivity in Japan after 1995 has been
attributed to a compositional shift in the labour market to increase the amount of
higher-quality labour due to a higher demand for higher education (Chun et al.,
2015). The United Kingdom has the highest amount of produced capital in G7 at
78% of total wealth, followed by France at 73%. Therefore, human capital is the
foremost contributor to IW growth rates for all of the G7 countries. On average,
human capital contributed to 68% of overall in IW, whereas produced capital
contributed to 25% and natural capital contributed to 7% if we look at all G7
countries as a single group.
The TFP growth of the G7 countries has undergone a powerful revival since
1990, as depicted in Figure 6.4. The TFP growth rate, which considers natural
capital as an input, shows that the UK’s TFP growth rate pattern is increased.
The UK’s TFP growth based on IW continued to progress from 1990 to 2010 but
did not do so as rapidly as France and Germany. Although France and Germany
started the 20-year period behind the UK, the UK finished ahead until 2010. The
trend in TFP growth among the G7 countries since the mid-1990s has attracted
significant attention because labour productivity growth represents an important
factor in economic growth stability (Jablanovic, 2012).
The UK’s TFP growth exceeded that of the United States starting in 1992.
Although the entire G7 suffered during the financial crisis of 2008–2009, the
United Kingdom continues to enjoy the highest TFP growth in the G7. From
2005 to 2010, TFP growth for most of the G7 countries was slowed considerably
Canada Japan France Germany Italy UK USA
1.08
1.06
1.04
TFP GROWTH
1.02
0.98
0.96
0.94
0.92
1990- 1993- 1996- 1999- 2002- 2005- 2008-
1991 1994 1997 2000 2003 2006 2009
YEAR
Figure 6.4 TFP growth of the G7 countries (1990–2010)
by the global financial and economic crisis of 2008–2009. Recovery from the
Great Recession of 2007–2009 by the United States, Japan, Canada and the four
European economies of the G7 has been slow and fitful (Jorgenson and Khuong,
2013). The Asian currency and financial crises of 1997 and 1998 affected Japan’s
TFP growth more than that of other G7 countries because Japan is the only Asian
country in the G7. Japan recovered from the Asian financial crisis and starting
in 1999, Japan’s TFP growth was higher than that of France, Germany and the
United States. Japan’s TFP growth revived during from 2000 to 2010, and Italy
showed the weakest performance during the global crisis.
4.4.1 Countries example: case of BRICS

From this study, we also can learn from the several countries that show signifi-
cant results from the beginning to the end of the study period. In this case, we
took Brazil, Russia, India, China and South Africa (also known as the BRICS
countries) as examples of countries with high rates of economic growth. In
general, each of the BRICS countries is different in terms of history, human
capital, produced capital, natural capital and economic strategies. However, the
BRICS countries have two things in common: high growth potential compared
to other countries and high global economic growth rates. Together, the BRICS
countries have more than three billion inhabitants, which represent almost 40%
South Africa
China
COUNTRY
India
Russia
Brazil
0 10 20 30 40 50 60 70
PERCENTAGE (%)
Natural Capital Human Capital Produced Capital
Figure 6.5 Percentage of wealth composition of BRICS countries (1990–2010)
of the world’s population. Therefore, BRICS countries are not only important
resource suppliers for industrialized countries but are also significant to the
global economy.
Figure 6.5 shows the wealth composition among BRICS countries from 1990
to 2010. Natural resources and strong investment in human capital contribute to
Brazil’s TFP and economic. Brazil has human capital advantages compared to
China and Russia, and Brazil has the second-largest amount of natural capital
(after Russia) among the BRICS countries. For instance, European Union (EU)
consumes large quantities of agricultural imports from Brazil and it also supplies
more than three-quarters of the EU’s beef and one-half of its soy meal and soy-
beans (Oleson, 2011).
In 2011, South Africa joined the BRIC grouping 2 years after it was estab-
lished, thus creating BRICS. From 1990 to 2010, South Africa’s average TFP
growth was 0.3%, which was slightly higher than Russia. Historically, gold
mining and minerals have been the backbone of South Africa’s economy. How-
ever, South Africa is now shifting to industry and services. Although South
Africa had a slower pace of TFP growth among the BRICS countries, South
Africa’s pace is better than Russia’s. South Africa is often called as the “gate-
way” to the African continent, and this status probably played an important
role in South Africa’s admission to BRICS (Shubin, 2015). In the early 1990s,
by the end of the apartheid era, South Africa had an exceptionally high unem-
ployment rate and an unusually low employment rate. That notwithstanding,
the post-apartheid government pursued a broadly Porterian strategy that used
labour market and industrial policies to increase labour productivity and to
support “decent work”. Over two decades beginning during the 1994 transition
to democracy, this strategy contributed to very modest employment growth
(Nattrass and Seekings, 2015).
South Africa is also like India in that it has a population that is much younger
than the other BRICS countries. People who are below the age of 35 years rep-
resent almost 66% of South Africa’s population, thus allowing South Africa to
enjoy the benefits of future dividends, which provide substantial investment into
the human capital represented by youth (United Nations Population Fund, 2017).
Therefore, the combination of South Africa’s human and produced capital has
supported the increase in other components’ growth efficiency during the 1990 to
2010 study period.
We also found that China has made a relatively significant shift in its inputs,
significantly contributing to growth. China enjoys human capital advantages (i.e.,
its population is 1.4 billion, or one-quarter of the world’s population). Human
capital and development, such as life expectancy, infant mortality, adult literacy
and political development, have been universally recognized as key determinants
of China’s economic growth.
According to Wang and Yao (2003), China’s accumulation of human capital
was quite rapid and contributed significantly to TFP growth and welfare. After
the 1989–1990 recession, China reformed its market strategies, and in 1993,
its investment rapidly increased, reaching almost 13% of GDP growth. Starting
in the mid-1990s, consumption and export became more important and invest-
ment progressed in 2000, when China’s government increased its spending on
large-scale infrastructure projects, while both foreign and domestic manufac-
turing investment were booming (Zheng, Bigsten, and Hu, 2009). Therefore,
the combination of human and produced capital has also supported China’s
growth.
In Russia, natural capital plays an important role in economic growth. Most
of Russia’s export orientation is based on natural resources and relies on energy
revenues to drive economic growth. Russia is the most resource-rich country in
the world, holding approximately 30% of the world’s natural resources. Russia
has the world’s largest mineral resources and natural gas reserves, the second-
largest coal reserves, the eighth-largest oil reserves and the largest water reserve
on the planet (Sheng-Jun, 2011). According to Desai (2006), the Russian econ-
omy is driven by expanding oil production and investment. GDP growth was
remarkable in 2000, at 10.4%. Therefore, Russia’s economy is driven by extrac-
tive industries and energy-intensive and low-technology manufacturing (Thurner
and Roud, 2015). However, during the financial crisis and failing oil prices of
2008–2009, Russia suffered the largest decline in TFP growth among the BRICS
countries (see Figure 6.6). In 2003, Russia possessed 6% of proven world oil
reserves and 27% of natural gas reserves, along with 9% of global exports and
29% of gas exports (Kuboniwa, Tabata, and Ustinova, 2005). Russia also is the
largest country in the world and has a population of more than 142 million; these
factors also contributed to its TFP growth, especially from 1990 to 1997. The
Brazil Russian India China South Africa
1.15
1.1
1.05
1
TFP GROWTH
0.95
0.9
0.85
0.8
0.75
0.7
1990- 1992- 1996- 1999- 2002- 2005- 2008-
1991 1994 1997 2000 2003 2006 2009
YEAR
Figure 6.6 TFP growth of BRICS countries (1990–2010)
combination of natural capital and human capital in Russia has supported its
growth and efficiency.
India’s economic growth is supported by strong capital imports and its status as
a destination for foreign investors. Its exchange rate policy, which is more flexible
than that of China, makes India more attractive to foreign investors. Realloca-
tion of workers from agriculture to industry and services has contributed 1.2% to
India’s annual productivity growth. Forty-five percent of India’s growth comes
from services (Bosworth and Collins, 2007). One vital element of India’s rapid
economic growth since the early 1990s has been the improved performance of its
manufacturing sector. Output in manufacturing grew by 5.7% per year from 1993
to 2005 (Arnold et al., 2012). In term of wealth composition, India’s human capi-
tal makes the largest contribution, which is almost 65% of India’s total IW when
natural capital is considered (Figure 6.5).
Figure 6.6 shows an analysis of the TFP growth using the IW framework that
considers natural capital. Based on this study, China and India indicated a minimal
decline in TFP growth during the global financial and economic crisis of 2008–
2009. Indeed, in 2009, India’s overall TFP growth increased over 2008 and China
experienced a small decline in its TFP growth rate during the crisis. Although
South Africa’s TFP growth rate is considered stable, during the economic crisis,
South Africa’s suffering was second only to Russia among the BRICS countries.
Before the crisis, Russia experienced a better performance in its TFP growth rate;
after the crisis, it experienced a strong slump.
Among the BRICS countries, China is still leading in term of TFP growth from
1990 to 2010 followed by India, Brazil and South Africa. In 1997–1998, the Asian
financial crisis occurred and damaged the foreign exchange market. During the
global financial crisis in 2008–2009, all countries including BRICS also suffered
for a couple of years. Pati (2017) also agreed that the impact of the global finan-
cial crisis was well pervasive with no exception to the BRICS countries. How-
ever, most BRICS countries were able to recover quickly from the global financial
crisis. In general, the economic growth among the BRICS countries also plays the
significant role for the global economic stabilization.
Compare to other BRICS countries, both the Asian and global financial crises
made a huge impact on the entire Russia economy. The main cause of the drop
is because of the decline in the Asian markets for oil; and a drop in steel exports
to the United States is because of import regulations (Fukumoto, 1990). The two
financial and economic crises that hit Russia in 1998 and 2008 shows that the
country appeared to be finally entering into a sustainable development path of
growth. In the aftermath of the 2009 financial crisis, Russia has embarked on an
ambitious program of modernization and diversification focused on new technol-
ogy and institutional reforms (Malle, 2013).
Based on this study, China and India experienced a minimal decline in its
TFP growth rate during the global financial and economic crisis of 2008–2009.
In 2009, India’s overall TFP growth rate was even slightly higher compared
to 2008. China experienced a small decline in its TFP growth rate in the years
2008–2009. Brazil quickly reverted to the positive average TFP growth rate
after suffered a decline in the TFP growth rate during the years of economic
crisis. Even South Africa’s TFP growth rate was considered stable, but dur-
ing the economic crisis South Africa also suffered after Russia in the group of
BRICS countries. Before the crisis, Russia experienced a better performance
in TFP growth rate but a strong slump in its TFP growth rate after the crisis.
Throughout the study period, Brazil’s average TFP growth increased by 0.9%,
the second-highest rate (after China) among the BRICS countries. The main fac-
tors underlying Brazil’s economic growth were human capital improvements,
high levels of gross fixed capital and the economy’s increasing openness to the
international market (Andrade and Garcia, 2015). In this study period, China
shows dynamic economic progress from 1990 to 2010. This rapid growth is
related to the China’s export economy and represents a massive investment in
infrastructure and human capital due to the beginning of the Cultural Revolu-
tion (Hofman, Zhao, and Ishihara, 2007).
Adjusted by TFP, we found a large discrepancy result of IW as shown in Fig-
ure 6.7. China indicates the highest growth with a 7.163 score (7.2% of change),
while Russia moves to the negative bracket with −0.554 (−0.6% of change). Natu-
ral capital plays an important role in the economic growth. Russia is confronted
by long-term issues related to sustaining its current consumption. In the case of
Russia, a stringent environmental regulation and Russia’s pace of industrialization
should be consistent with its environmental management to achieve and maintain
sustainable development.
8.000
7.000
6.000
5.000
4.000
AVERAGE
3.000
2.000
1.000
0.000
Brazil Russia India China South Africa
−1.000
COUNTRY
Average TFP IW-TFP Adjusted
Figure 6.7 Comparison between average TFP and IW-TFP adjusted growth of BRICS
5 Conclusion
This study has combined the concept of inclusive wealth, including natural capital
as an input, with an adjustment of oil capital gain in productivity measurement.
We utilized the Malmquist Productivity Index to measure cross-country produc-
tivity accounting during the 1990–2010 study period.
The TFP results from using GDP as an output and IW (human, produce and
natural capital) as input shows significantly different results among 140 countries.
The growth in TFP also depends on the contribution of two components, namely,
countries’ technical efficiency change and technological change. Natural capital,
including oil capital gain related to the change in oil prices, also is significant to
the productivity value of the country.
In the case of the BRICS and G7 countries, we even find that the composition of
human and produced capital is higher and contributes more to productivity from
1990 to 2010. However, the loss of natural capital is not high enough to compen-
sate for both human and produced capital. Therefore, it is suggested, for example,
that countries investing in natural capital should compensate for human capital
development, non-renewable natural capital depletion and return to a sustainable
agenda. For instance, in the case of Russia, there is a significant link between
global oil price and the productivity output of the Russian economy. Therefore, it
is suggested that a country like Russia needs to reduce its economy depending on
global trade environment, especially oils or fuels, in order to achieve sustainable
development.
This study also presents several important findings for economic policy evalu-
ation and planning. For instance, as part of a country’s sustainable development,
a country must target not only GDP growth as a primary objective but should
also move towards incorporating IW into the measurement of TFP. For the coun-
try that is striving to improve the citizens well-being, this finding suggested that
IW is useful in assessing sustainable development. For future, more sophisticated
studies, an expanded dataset is also needed, for example one that includes more
countries, years and factors such as renewable energy stock and health in human
capital.
Acknowledgements
This chapter was supported by Grant-in-Aid for Specially Promoted Research
(26000001) by the Japan Society for the Promotion of Science.
Isma Addi Jumbri received the scholarship support from the Ministry of Educa-
tion Malaysia and Universiti Teknikal Malaysia Melaka (UTeM) throughout the
study period.
Bibliography
Acar, S., and Gultekin-Karakas, D. (2016). Questioning Turkeys ‘Miracle’ Growth from a
Sustainability Perspective. The Journal of Environment & Development 25(2):
131–158.
Aghion, P., Beesley, T., Browne, J., Caselli, F., Lambert, R., Lomax, R., Pissarides, C.,
Stern, N., and Van Reenen, J. (2013). Investing for Prosperity. London: LSE.
Andrade, D. C., and Garcia, J. R. (2015). Estimating the Genuine Progress Indicator (GPI)
for Brazil from 1970 to 2010. Ecological Economics 118(August): 49–56.
Arnold, J. M., Javorcik, B., Lipscomb, M., and Mattoo, A. (2012). Services Reform and
Manufacturing Performance Evidence from India. Retrieved from Centre for Economic
Policy Research (CEPR) website: www.cepr.org/pubs/dps/DP8011.asp.
Arrow, K. J. (2007). Global Climate Change: A Challenge to Policy. Economists’ Voice:
1–5.
Arrow, K. J., Dasgupta, P., Goulder, L., Daily, G., Ehrlich, P., Heal, G., Levin, S., Mäler,
K.-G., Schneider, S., Starrett, D., and Walker, B. (2004). Are We Consuming Too Much?
The Journal of Economic Perspectives 18(3): 147–172.
Arrow, K. J., Dasgupta, P., and Mäler, K.-G. (2003). Evaluating Projects and Assessing
Sustainable Development in Imperfect Economies. Environmental and Resource Eco-
nomics 26(4): 647–685.
Arrow, K. J., Goulder, L. H., Mumford, K. J., and Oleson, K. (2012). Sustainability and the
Measurement of Wealth. Environmental and Development Economics 17: 317–353.
Atkinson, G., Dietz, S., Neumayer, E., and Agarwala, M. (2014). Handbook of Sustainable
Development: Second Edition. Edward Elgar Publishing.
Bosworth, B., and Collins, S. M. (2007). Accounting for Growth: Comparing China and
India. NBER Working Paper Series, Cambridge, MA. Retrieved from: www.nber.org/
papers/w12943.pdf.
BP Statistical Review of World Energy 2013. BP Statistical Review of World Evergy, Lon-
don. [online]. Available at: http://large.stanford.edu/courses/2013/ph240/lim1/docs/
bpreview.pdf Brandt, N., Schreyer, P., and Zipperer, V. (2017). Productivity Measure-
ment with Natural Capital. Review of Income and Wealth 63(February): 1–21.
Chun, H., Tsutomu, M., Pyo, H. K., and Konomi, T. (2015). Do Intangibles Contribute to
Productivity Growth in East Asian Countries ? Evidence from Japan and Korea. RIETI.
Retrieved from Research Institute of Economy, Trade and Industry (RIETI) website:
www.rieti.go.jp/jp/publications/dp/15e055.pdf.
Clémençon, R. (1997). Economic Integration and the Environment in Southeast Asia:
Securing Gains from Open Markets While. Journal of Environment & Development 6(3):
317–333.
Clémençon, R. (2012). Welcome to the Anthropocene: Rio+20 and the Meaning of Sustain-
able Development. The Journal of Environment & Development 21(3): 311–338.
Coelli, T., Rao, D. P., and Battese, G. E. (1998). An Introduction to Efficiency and Produc-
tivity Analysis, Second Edition. Boston: Kluwer Academic Publisher.
Collins, S. M., and Bosworth, B.P.R. (1996). Economic Growth in East Asia: Accumulation
versus Assimilation. Brookings Papers on Economic Activity 2: 135–203.
Costanza, R., Giovannini, E., Lovins, H., McGlade, J., Ragnarddottir, K. V., Roberts, D.
Vogli, R. D., and Wilkinson, R. (2014). Development: Time to Leave GDP Behind.
Nature 505(7483): 283–285.
Dasgupta, P. (2001). Human Well-Being and the Natural Environment. New York: Oxford
University Press.
Dasgupta, P. (2005). Economics of Social Capital. Economic Record 81(255): 2–21.
Dasgupta, P. (2007). Measuring Sustainable Development: Theory and Application. Asian
Development Review 24(1): 1–10.
Dasgupta, P. (2008). Nature in Economics. Environmental and Resource Economics 39(1):
1–7.
Dasgupta, P. (2009). The Welfare Economic Theory of Green National Accounts. Environ-
mental and Resource Economics 42(1): 3–38.
Dasgupta, P., Duraiappah, A., Managi, S., Barbier, E., Collins, R., Fraumeni, B., Gundim-
eda, H., Liu, G., and Mumford, K. J. (2015). How to Measure Sustainable Progress.
Science 350(6262): 748–748.
Dasgupta, P., and Mäler, K.-G. (2001). Wealth as a Criterion for Sustainable Development.
World Economics 2(3): 19–44.
Debreu, G. (1951). The Coefficient of Resource Utilization. Journal of the Econometric
Society 19(3): 273–292.
Denison, E. (1979). Explanations of Declining Productivity Growth August. Survey Cur-
rent Business August(59): 1–24.
Desai, P. (2006). Why Is Russian GDP Growth Slowing? American Economic Review
96(2): 342–347.
Dubash, N. K. (2012). Toward Enabling and Inclusive Global Environmental Governance.
The Journal of Environment & Development 21(1): 48–51.
Estrada, A., and Hernández de Cos, P. (2011). Oil Prices and Their Effect on Potential
Output. Applied Economics Letters 19(3): 207–214.
Färe, R., Grosskopf, S., Norris, M., and Zhang, Z. (1994). Productivity Growth, Technical
Progress, and Efficiency Change in Industrialized Countries. American Economic Review
84(1): 66–83.
Farrell, M. J. (1957). The Measurement of Productive Efficiency. Journal of the Royal
Statistical Society. Series A (General) 120(3): 253–290.
Finn, M. G. (2016). Perfect Competition and the Effects of Energy Price Increases on Eco-
nomic. Journal of Money, Credit and Banking 32(3): 400–416.
Fukumoto, M. (1990). Russia’s Participation in APEC and Economic Development in the
Far East. APEC Study Center, Institute of Developing Economies.
Hamilton, K. (2000). Genuine Saving as a Sustainability Indicator. OECD Publishing.
Hamilton, K., and Atkinson, G. (1996). Air Pollution and Green Accounts. Energy Policy
24(7): 675–684.
Hamilton, K., and Atkinson, G. (2006). Wealth, Welfare, and Sustainability: Advances in
Measuring Sustainable Development. UK: Edward Elgar Publishing, Inc.
Hamilton, K., and Clemens, M. (1999). Genuine Savings Rates in Developing Countries.
The World Bank Economic Review 13(2): 333–356.
Han, H. (2017). Singapore, a Garden City: Authoritarian Environmentalism in a Develop-
ment State. The Journal of Environment & Development 26(1): 3–24.
Hesary, F. T., Yoshino, N., Abdoli, G., and Farzinvash, A. (2013). An Estimation of the
Impact of Oil Shocks on Crude Oil Exporting Economies and Their Trade Partners.
Frontiers of Economics in China 8(4): 571–591.
Hofman, B., Zhao, M., and Ishihara, Y. (2007). Asian Development Strategies : China and
Indonesia Compared. Bulletin of Indonesian Economic Studies 43(2): 171–201.
Jablanovic, V. D. (2012). Labour Productivity and the Chaotic Economic Growth Model:
G7. Chinese Business Review 11(5): 500–510.
Jorgenson, D. W., and Griliches, Z. (1967). The Explanation of Productivity Change. The
Review of Economic Studies 34(3): 249–283.
Jorgenson, D. W., and Khuong, M. K. (2013). The Emergence of the New Economic Order:
Growth in the G7 and the G20. Journal of Policy Modeling 35(3): 389–399.
Kerstens, K., and Managi, S. (2012). Total Factor Productivity Growth and Convergence in
the Petroleum Industry: Empirical Analysis Testing for Convexity. International Journal
of Production Economics 139(1): 196–206.
Klenow, P. J., and Rodriguez-Clare, A. (1997). The Neoclassical Revival in Growth Eco-
nomics: Has It Gone Too Far? NBER Macroeconomics Annual 1997 12: 73–114.
Kuboniwa, M., Tabata, S., and Ustinova, N. (2005). How Large Is the Oil and Gas Sector
of Russia? A Research Report. Eurasian Geography and Economics 46(1): 68–76.
Kurniawan, R., and Managi, S. (2017). Sustainable Development and Performance Mea-
surement: Global Productivity Decomposition. Sustainable Development: 1–9.
Mäler, K.-G. (2008). Sustainable Development and Resilience in Ecosystems. Environmen-
tal and Resource Economics 39(1): 17–24.
Malik, S. K. (2015). Measurement of Technological Change in India’s Textile Machinery
Industry Using the Malmquist Productivity Index. Margin: The Journal of Applied Eco-
nomic Research 9(2): 179–203.
Malle, S. (2013). Economic Modernization and Diversification in Russia: Constraints and
Challenges. Journal of Eurasian Studies 4(1): 78–99.
Managi, S. (2003). Luenberger and Malmquist Productivity Indices in Japan, 1955–1995.
Applied Economics Letters 10(9): 581–584.
Managi, S. (2011). Technology, Natural Resources, and Economic Growth: Improving the
Environment for a Greener Future. Cheltenham, UK: Edward Elgar Publishing.
Managi, S. (2015). The Economics of Green Growth: New Indicators for Sustainable Soci-
eties. New York: Routledge.
Morrison, C. J. (1985). On the Economic Interpretation and Measurement of Optimal
Capacity Utilization with Anticipatory Expectations. The Review of Economic Studies
52(2): 295–310.
Mota, R. P., Domingos, T., and Martins, V. (2010). Analysis of Genuine Saving and Potential
Green Net National Income: Portugal, 1990–2005. Ecological Economics 69(10): 1934–1942.
Mumford, K. J. (2016). Prosperity, Sustainability and the Measurement of Wealth. Asia &
the Pacific Policy Studies 3(2): 226–234.
Nattrass, N., and Seekings, J. (2015). Should and can labour-surplus, middle-income econ-
omies pursue labour-intensive growth? The South African challenge. CSSR Working
Paper No. 351.
Neumayer, E. (2000). Resource Accounting in Measures of Unsustainability. Environmen-
tal and Resource Economics 15(3): 257–278.
Nordhaus, W. D., and Boyer, J. (2000). Warming the World: Economic Models of Global
Warming. Cambridge, MA: The MIT Press.Norsworthy, J. R., Harper, M. J., and Kunze,
K. (1979). The Slowdown in Productivity Growth: Analysis of Some Contributing Fac-
tors. Brookings Papers on Economic 2(1979): 388–445.
Oleson, K.L.L. (2011). Shaky Foundations and Sustainable Exploiters: Problems with
National Weak Sustainability Measures in a Global Economy. The Journal of Environ-
ment & Development 20(3): 329–349.
Ollivier, T., and Giraud, P.-N. (2010). The Usefulness of Sustainability Indicators for Policy
Making in Developing Countries: The Case of Madagascar. The Journal of Environment
& Development 19(4): 399–423.
Pati, A. P. (2017). Credit Risk Stress Testing Practices in BRICS: Post-Global Financial
Crisis Scenario. Global Business Review. https://doi.org/10.1177/0972150917692269
Scheffer, M., Brock, W., and Westley, F. (2000). Socioeconomic Mechanisms Preventing
Optimum Use of Ecosystem Services: An Interdisciplinary Theoretical Analysis. Ecosys-
tems 3(5): 451–471.
Sheng-Jun, Y. (2011). Educational Policies and Economic Growth in BRICs: Comparative
Perspectives. Journal of US-China Public Administration 8(2): 188–197.
Shubin, V. (2015). South Africa in the BRICS : Last but Not Least. International Organisa-
tion Research Journal 10(2): 171–183.
Sonnenfeld, D. A., and Mol, A.P.J. (2006). Environmental Reform in Asia. The Journal of
Environment & Development 15(2): 107–111.
Stern, D. I. (2005). Beyond the Environmental Kuznets Curve: Diffusion of Sulfur-
Emissions-Abating Technology. The Journal of Environment & Development 14(1):
101–124.
Tanaka, K., and Managi, S. (2013). Measuring Productivity Gains from Deregulation of the
Japanese Urban Gas Industry. Energy Journal 34(4): 181–198.
Tang, D., Tang, J., Xiao, Z., Ma, T., and Bethel, B. J. (2017). Environmental Regulation
Efficiency and Total Factor Productivity-Effect Analysis Based on Chinese Data from
2003 to 2013. Ecological Indicators 73: 312–318.
Thurner, T. W., and Roud, V. (2015). Greening Strategies in Russia’s Manufacturing: From
Compliance to Opportunity. Journal of Cleaner Production: 1–10.
United Nations Population Fund (2017, May 31). Retrieved from www.unfpa.org/.
UNU-IHDP and UNEP (2012). Inclusive Wealth Report 2012. Cambridge: Cambridge Uni-
versity Press.
UNU-IHDP and UNEP (2014). Inclusive Wealth Report 2014. Cambridge: Cambridge
University Press.
Wang, Y., and Yao, Y. (2003). Sources of China’s Economic Growth, 1952–1999: Incorpo-
rating Human Capital Accumulation. China Economic Review 14(1): 32–52.
World Bank (2006). Where Is the Wealth of Nations?: Measuring Capital for the 21st
Century. Washington: World Bank.
World Bank (2010). The Changing Wealth of Nations: Measuring Sustainable Development
in the New Millennium. Washington: World Bank.
Yamaguchi, R. (2014). Inclusive Wealth with a Changing but Aging Population. Economics
Letters 124: 21–25.
Yamaguchi, R., and Managi, S. (2017). New Financing for Sustainable Development. The
Journal of Environment & Development 26(2): 214–239.
Yamaguchi, R., Sato, M., and Ueta, K. (2016). Measuring Regional Wealth and Assessing
Sustainable Development: An Application to a Disaster-Torn Region in Japan. Social
Indicators Research 129(1): 365–389.
Zheng, J., Bigsten, A., and Hu, A. (2009). Can China’s Growth Be Sustained? A Productiv-
ity Perspective. World Development 37(4): 874–888.
7 Relative income, community
attachment and subjective
well-being
Evidence from Japan
Tetsuya Tsurumi, Kong Joo Shin,
Atsushi Imauji and Shunsuke Managi
1 Introduction and background

The concept of subjective well-being (SWB) has received increasing attention
from both researchers and policymakers as a measure of utility that can be incor-
porated into the policy-making process and assessment (Diener and Seligman,
2004; Kahneman et al., 2004; Diener, 2006). Previous studies of the determi-
nants of SWB have demonstrated that people’s SWB is determined by various
factors: income, gender, age, marriage, stress and environmental quality, such as
green coverage and air pollution. In particular, the relationship between income
and SWB has been studied extensively. Substantial empirical evidence has been
accumulated since Easterlin posited the ‘Easterlin Paradox’, which asserts that
people with higher incomes are more likely to report being happy but that this
relationship does not hold at the national level (Easterlin, 1974). While the debate
on this paradox continues,1 the overall positive relationship between income and
individual well-being appears to be robust, and a diminishing marginal income
effect has been confirmed by the results of various empirical studies (Helliwell,
2003; Stevenson and Wolfers, 2013).
Recent related studies have found that people’s happiness is affected not only
by their own income but also by the income of others; this is generally referred
to as the ‘relative income (RI) hypothesis’. The RI hypothesis was proposed by
Duesenberry (1949) to explain saving behaviour in the United States, although
recently RI has been examined in the SWB framework. Clark and Oswald (1996)
provided some of the earliest empirical evidence of the RI effect on job satisfac-
tion, and numerous subsequent empirical studies have examined the RI effect on
SWB measures such as life satisfaction (LS). Table 7.1 presents an updated list
of RI studies based on a literature review by Brown et al. (2015). Most studies
use LS as a proxy of well-being. While own income is generally found to have a
positive impact on SWB, the literature is inconclusive with respect to the effects
of RI. As Brown et al. (2015) suggested, this lack of consensus regarding the
impact of RI may be due to “differences in data, definitions, model specification
and estimation”.
In particular, the characteristics used to identify reference groups vary: stud-
ies have used the average or predicted income of reference groups based on
Table 7.1 Studies on relative income hypothesis
Study Data Reference Estimation Relative

group method income effect
McBride US General Age (all Ordered probit Negative

(2001) Social Survey individuals who
(GSS) 1994 are in the age
range of 5 years
younger and 5
years older than
the individual
concerned)
Senik (2004) Russian Predicted income Ordered probit Positive
Longitudinal (education,
Monitoring occupation,
Survey (RLMS) industry, age,
1994–2000 gender, region,
etc.)
Ferrer-i- German Socio- Individual Ordered probit Negative
Carbonell economic characteristics with individual
(2005) Panel Study (age, education, random effect
(GSOEP) 1984, region)
1992–1997
Luttmer 1987/1988 and Local predicted OLS with state Negative
(2005) 1992/1994 earnings by FE
US National industry/
Survey of occupation
Families and for Public Use
Households, Microdata
1990 Census Areas (PUMAs)
and Current (PUMAs
Population correspond to
Survey about 150,000
inhabitants on
average)
Kingdon and SALDRU Others in the Ordered probit Positive (close
Knight (2007) National local residential and OLS neighbours);
Household cluster (close and negative (more
Survey of 1993 distant), race- distant others);
in South Africa based comparator positive or
groups negative (race)
Senik (2008) European Predicted income OLS and FE Positive
Community (education, (especially
Household occupation, in transition
Panel (ECHP) industry, gender, countries) or
1994–2001 experience, negative
region, full time
or part time)
Study Data Reference Estimation Relative
group method income effect
Caporale et al. European McBride’s (2001) Ordered probit Positive

(2009) Social Survey definition of the (Eastern
(ESS) 2002, reference group Europe) or
2004 (all individuals negative
who are in the age (Western
range of 5 years Europe)
younger and 5
years older than
the individual
concerned)
Oshio et al. Chinese, Individual Ordered logit Negative
(2011) Japanese, characteristics
Korean, and (children, marital
US’s General status, education
Social Surveys and occupational
(CGSS, JGSS, status)
KGSS, GSS)
2006
Knies (2012) German Socio- Living area (street FE Negative (West
economic level) Germany);
Panel Study insignificant
(GSOEP) 2004, (East Germany)
2005
Goerke and German Socio- Colleagues, OLS Negative
Pannenberg economic people in your (colleagues,
(2015) Panel Study occupation, people in own
(GSOEP) friends occupation)
2008–2010 insignificant
(friends)
Brown et al. Understanding 1. Age, gender Pooled ordered Positive* (age,
(2015) Society, the and education probit (OP), gender and
UK Household 2. Local district RE OP, RE OP education);
Longitudinal with Mundlak, negative*
Study 2009– FE ordered (local district)
2013 logit *significance
depends on
estimation
methods
Welsch and German Socio- Predicted income Ordered logit Positive (East
Kühling economic (education, Germany);
(2015) Panel Study gender, age) negative (West
(GSOEP) Germany)
1991–2009
Notes: RE denotes random effect; FE denotes fixed effects.
170 Tetsuya Tsurumi et al.
individual characteristics such as age, gender, marital status, number of chil-
dren, education, occupation and industry (McBride, 2001; Senik, 2004; Ferrer-i-
Carbonell, 2005; Luttmer, 2005; Senik, 2008; Caporale et al., 2009; Oshio et al.,
2011; Goerke and Pannenberg, 2015; Brown et al., 2015; Welsch and Kühling,
2015). Additionally, different levels of spatial categories have been employed
to construct measures of RI (Kingdon and Knight, 2007; Knies, 2012; Brown
et al., 2015). Kingdon and Knight (2007) used average household income mea-
sured at the street level,2 and Brown et al. (2015) used the average incomes of
405 local districts in the United Kingdom. Furthermore, some studies have used
predicted income to represent various spatial characteristics (Senik, 2004; Lutt-
mer, 2005; Senik, 2008).
RI’s negative impact on utility may be explained by envy or jealousy effects
(Senik, 2004, 2008). People who compare themselves with others are less happy,
although this comparison effect diminishes with income level (Kahneman and
Tversky, 1979; Clark and Senik, 2010). By contrast, according to Hirschman and
Rothschild (1973), RI may have a positive effect on utility through a type of
‘information effect’ that includes ambition and signal effects of future income and
lifestyle (Senik, 2004; 2008; Caporale et al., 2009; Welsch and Kühling, 2015).
Kingdon and Knight (2007) also found positive effects of RI on well-being and
suggested that this result may be due to altruism within a close community. Addi-
tionally, a ‘status effect’ may induce positive impacts. For example, living in a
richer neighbourhood may increase utility through a reputation effect derived
from a branded neighbourhood.
Given that RI may have opposing effects on utility, the total effects reported by
previous studies include both positive and negative effects. Additionally, the sign
of the total effect suggests that the effect in one direction is larger than the effect
in the other direction given the specific target groups within a reference group.
However, these studies do not distinguish the magnitude of the positive and nega-
tive impacts of RI on SWB.
In this study, we extensively examine the RI hypothesis by disentangling the
opposing effects that RI may have on SWB. Using community attachment as
a mediator variable in structural equation modelling (SEM), we examine the
indirect and direct effects of RI on LS. Our first-stage estimation includes RI
as a determinant of community attachment together with factors that have been
employed in previous studies: individual and socio-demographic variables such
as age, gender, education, income, house ownership and years in residence.3 We
also control for neighbourhood- or community-specific variables such as safety,
population density, logistics, scenery, pollution, green coverage, and community
ties and activities that may affect level of attachment.4 In the second stage of
the estimation, we use community attachment as an explanatory factor of SWB.
Empirical studies of the relationship between community attachment and SWB
suggest that there is a positive impact of community attachment on reported well-
being (Powdthavee, 2008).5 Moreover, evidence indicates that people who live in
richer areas have higher levels of attachment to their community (Taylor et al.,
1985; Anton and Lawrence, 2014). Richer neighbourhood areas are associated
with better living environments in terms of safety, schooling and other factors that
contribute to people’s satisfaction.
Relative income, community attachment 171
Using mediator analysis to identify indirect effects enables us to separate the
positive effect of RI due to status or reputation effects and relatively better living
environments from the negative effect caused by comparison and envy effects.
Moreover, we examine whether the comparison/envy effects or the information/
status effects of RI vary across age groups: younger people may be more sensitive
to their surrounding environment, and their comparison intensity may be higher
than that of older people.
We use data collected from an original large-scale Japanese survey conducted
in 2015 to match geo-coded individual-level data with spatial RI calculated from
500 m mesh data generated from various official statistical and geographic infor-
mation sources. Although we do not use individual characteristics as reference
characteristics, 500 m × 500 m area income data can provide relatively precise
effects for ‘people near us’ or ‘our neighbours’.
The results indicate that RI has an indirect positive effect on LS through a posi-
tive impact on community attachment, which mitigates its direct negative impact.
This result implies that the negative impact of RI has been overestimated in pre-
vious studies that have not considered the mechanism we present in the current
study. Additionally, we find a larger negative impact of RI for younger people.
Furthermore, our results support the notion of spiteful egalitarianism, given that
people with a household income that is lower than the reference income are less
satisfied than people with a household income that is higher than the reference
income. We also present and discuss the impact of varying the reference area by
comparing the estimation results of the direct and indirect RI impacts on LS using
larger reference areas than 500 m mesh data.
The remainder of this chapter is structured as follows. Section 2 describes the
data and discusses the empirical methodology. Section 3 presents the descriptive
statistics and the regression results. Section 4 discusses the main findings, and
Section 5 concludes.
2 Empirical analysis
2.1 Data and variables

In our analysis, we use data collected from an original Japanese web-based sur-
vey conducted in November 2015. We received 121,514 responses. To the best
of our knowledge, this survey is one of the most recent and largest nationwide
surveys to include questions on SWB and its possible determinants. We elimi-
nated observations that were missing information about a respondent’s household
income or geographic location,6 resulting in 96,312 observations for the regres-
sion analysis. Table 7A.1 provides descriptions of the variables used in the analy-
ses, and Table 7.2 presents the descriptive statistics for all the variables.
We use ratings for overall LS as the SWB measure. While the survey data also
include other SWB measures, such as happiness and life evaluation, we focus on
LS based on empirical evidence provided by Kahneman and Deaton (2010): an
increase in income improves LS but does not increase emotional well-being. Given
that SWB-related questions are particularly vulnerable to the context in which they
are asked, the SWB questions were placed at the beginning of our questionnaire.
Table 7.2 Descriptive statistics
Variable name Mean SD Min Max

Age 49.03 11.45 17 99
Male 0.61 0.49 0 1
Education 0.50 0.50 0 1
Health 3.48 1.06 1 5
Length of residence: 0.11 0.32 0 1
1 year or more and less than
3 years
3 years or more and less than
5 years
10 years
20 years
30 years
40 years
Length of residence: more than 0.04 0.19 0 1
40 years
Homeownership 0.71 0.46 0 1
Relationship with local 3.26 0.77 1 5
residents
Required time to closest station 12.23 9.96 2.5 60
Required time to bus stop 5.71 7.33 2.5 60
Required time to store 12.33 10.12 2.5 60
Environmental satisfaction 3.58 0.76 1 5
Safety 3.06 0.57 1 4
Extraversion 2.91 0.86 1 5
Agreeableness 3.53 0.66 1 5
Conscientiousness 3.19 0.78 1 5
Neuroticism 2.98 0.76 1 5
Openness 3.03 0.73 1 5
Household income 6,717,320 4,433,884 1,000,000 30,000,000
Area average income 4,569,811 710,677 1,680,500 9,594,202
Richer dummy 0.66 0.47 0 1
Life satisfaction 3.48 0.95 1 5
Community attachment 3.31 1.03 1 5
We asked the respondents: “Overall, how satisfied are you with your life?”;
the response options included ‘1 – completely dissatisfied’, ‘2 – slightly dis-
satisfied’, ‘3 – neither’, ‘4 – slightly satisfied’ and ‘5 – completely satisfied’.
Approximately half the respondents reported that they are slightly satisfied, and
the remainder of the sample is distributed among the other response options.
Overall, the scores are distributed across the scale, with an average score of 3.84
and a standard deviation of 0.95. Figure 7A.1 presents the distribution of the LS
scores by response option and indicates that approximately 60% of the respon-
dents chose ‘4 – slightly satisfied’.
In this study, we analyse income effects on LS using three measures: (1) self-
reported household income, (2) reference area income and (3) a dummy variable
that indicates whether a respondent’s household income is higher than the aver-
age area income. Household income is based on self-reported pre-tax household
income. The respondents were selected according to income range. We then recoded
the category options to the median value of each income range (see Table 7A.1).
The reference area income data reflect the average annual income of the 500 m
mesh area. These area income data are generated by Zenrin Geo Intelligence Co.,
Ltd., which specialises in GIS data marketing, and are estimated based on data
from various sources such as the Housing and Land Survey of Japan, the Popula-
tion Census of Japan and the Grid Square Statistics of the Population Census of
Japan. We use these 500 m mesh-level data to obtain a household distribution of
income ranges and calculate the weighted average area income per mesh. Then,
we combine the survey data with these area income data using the postal code
provided by the respondents.
These area income data are more detailed than prefecture- or municipality-
level data. Kingdon and Knight (2007) analysed the impact of RI on SWB in
South Africa by dividing South Africa into 187 and 360 reference areas. The
authors found a statistically significant negative effect of RI when they used the
more detailed RI derived from the 360 areas, but they did not find a significant
impact of RI when it was based on 187 areas. More detailed reference areas can be
expected to more effectively capture the ‘neighbour effect’, which may be closely
related to jealously and envy effects.
In addition, we construct a dummy variable using the two income variables
noted earlier. The Richer dummy is coded 1 if self-reported household income is
higher than the reference area income and is coded 0 otherwise. We use this vari-
able to examine whether the RI position within a community has a statistically
significant impact on LS.
Along with the income variables, community attachment is one of the main inde-
pendent variables and serves as the mediator variable in our analysis. The respon-
dents were asked to measure their attachment to the neighbourhood/community
in which they currently reside. They responded using the following 5-point scale:
(1) not attached at all, (2) not really attached, (3) neither attached nor detached,
(4) slightly attached and (5) very attached. The average score was approximately
3.31 with a standard deviation of 1.03. Figure 7A.2 presents the distribution of the
community attachment scores. Overall, we observe variation in the respondents’
subjective evaluation of community attachment. Nearly 40% of the respondents
indicated that they are slightly attached, approximately 30% reported that they are
neither attached nor detached, and more than 20% reported low attachment.
In addition, we control individual and household characteristics that have been
frequently analysed as determinants of SWB in previous studies: age, gender, sub-
jective health condition, education, marital status and personality. Age and age-
squared are both included in the estimation to test the U-shaped effect of age on
SWB that has been previously observed by scholars (Lange, 2010). To eliminate
possible multicollinearity between these two variables, the age variable is centred.
A university dummy is coded 1 if the respondent graduated from university and is
coded 0 otherwise. A married dummy is coded 1 if the respondent reported being
currently married, and we do not distinguish whether the respondent has never
been married or was previously married.
To generate personality proxies,7 we use the ‘Big Five’ personality dimensions –
extraversion, agreeableness, conscientiousness, neuroticism and openness – which
are calculated based on the average scores on related items from the Ten-Item Per-
sonality Inventory (TIPI) proposed by Gosling et al. (2003).
We also control for resident and residential area variables, which are particularly
important determinants of the mediator, community attachment. A homeownership
dummy is coded 1 if the respondent lives in a self-owned residence regardless
of residence type. We also asked the respondents how long they have lived in
a particular residence, and these responses are coded into categorical variables
with seven values. Additionally, we use time required to access bus stops, train/
subway stations and daily shopping to control for variations in location in terms of
convenience. Furthermore, we collect information on the respondents’ subjective
evaluation of neighbourhood safety (4-point scale) and the natural environment in
a locality (5-point scale). Finally, we control for population density at the munici-
pality level based on area and population data available from the 2010 census.
2.2 Statistical specification and methodology

We use SEM to examine whether RI has direct and indirect impacts on LS. Ordi-
nary least squares regression (OLS) and ordered logit models are commonly
used in studies of SWB determinants, but a few recent studies have used media-
tor analysis to address endogeneity problems and to analyse the multi-layered
relationship between some of the determinants and SWB. Mohanty and Ullah
(2012) used SEM to disentangle the effects of happiness and education on wages.
Powdthavee et al. (2015) used SEM to closely examine the relationship between
education and LS and found that years of education directly affect LS and have
an indirect impact on LS through the following analysed mediators: income,
employment, material status, children in the household and health condition.
Additionally, using SEM, Li et al. (2014) demonstrated that people’s recognition
of pollution levels affects SWB in China.
Using SEM, we can calculate the indirect, direct and total effects of the vari-
ables that are controlled in both equations (see Hayes and Preacher, 2014; Linden
Age (+ or –)
Male dummy (+ or –) Age (+ or –)
Personality (?) Age squared (+)
Household income (+) Male dummy (–)
Relative income (?) Married dummy (+)
Richer dummy (+) Personality (?)
Length of residence (+) Health (+)
Required time to closest station (–) Household income (+)
Required time to bus stop (–) Relative income (–)
Required time to store (–) Richer dummy (+)
Homeownership (+) Education (+)
Relationship with locaI residents (+) Safety (+)
Education (–) Environmental satisfaction (+) Exogenous
Safety (+) variables
Environmental satisfaction (+)
Endogenous
Community attachment Life satisfaction
variables
Figure 7.1 Conceptual diagram of the estimation model

Note: Plus or minus signs in parentheses denote expected signs of coefficient of each variable.
and Karlson, 2013; Preacher and Hayes, 2008). Figure 7.1 presents a conceptual
diagram of the estimation model with the expected signs of the coefficients for
each variable.
We use maximum likelihood estimation (MLE) to compute coefficients for the
preceding equations, where ε is the error term, and we assume that the dependent
variable and the mediator variable are continuous variables.
Z i = α1 + γ1V1′i + γ 2Wi ′ + γ 3 X ai + γ 4 X bi + γ 5 Richeri + γ 6Y1′i + ε1 (1)
LSi = α2 + β1V2′i + β2Wi ′ + β3 X ai + βa X bi + β5 Richeri + β6Y2′i + β7 Z 2i + ε1 (2)
We employ estimation equations (1) and (2) as follows: (1) regresses the set
of explanatory variables on the mediator variable, community attachment, while
(2) regresses community attachment along with other determinants on LS. LSi is
the life satisfaction measure, and Zi is the mediator, community attachment. Xai
is individual i’s logged value for annual household income, following Stevenson
and Wolfers (2013). Xbi is the logged average reference area income of i’s resi-
dential area. Following the findings of Oshio et al. (2011), we control for whether
respondents are relatively rich or poor in a neighbourhood using the Richeri
dummy. V′1i and V′2i are vectors of the individual socio-demographic variables,
except for the income-related variables. We control for slightly different sets of
variables for LS and Z based on previous empirical studies. V′1i includes age, the
gender dummy, the college dummy, duration of residence and the homeownership
dummy, and V′2i includes age, age (squared), material status, the gender dummy,
the college dummy and subjective health. W′i is the vector of the personality vari-
ables: extraversion, agreeableness, conscientiousness, neuroticism and openness.
These personality factors are well-discussed determinants in the SWB literature
but are rarely examined as determinants of place attachment. Nonetheless, we
expect people’s personality to have some impact on their attachment to the com-
munity in which they reside. Y′1i and Y′2i are vectors of residential area character-
istics. In Equation (1), we control access to public transportation, access to the
store, and a subjective evaluation of relationship with neighbours, surrounding
nature and safety. In Equation (2), we control surrounding nature and safety. α1
and α2 are the constants, and ε1 and ε2 are the error terms.
3 Results
Table 7.3 displays the main results of the regression analyses conducted with the
full sample. According to the results of the first-stage estimation with commu-
nity attachment as the dependent variable, we observe no statistically significant
effect of absolute household income, although there is a robust positive effect of
RI on the respondents’ community attachment. Moreover, whether the respon-
dents are richer or poorer than the average area income does not appear to affect
their community attachment. This result supports the findings of previous studies
indicating that residents share this positive community effect regardless of their
income positions within the community through gains from reputation and other
neighbourhood qualities.
Regarding the other control variables, respondents who have lived in an area
longer and those who own the house in which they live are more attached to their
community than their counterparts. Additionally, those who are satisfied with
neighbourhood safety and environmental conditions and those with better access
to public transportation are likely to have higher community attachment. These
results are somewhat expected. However, we also observe a less obvious effect:
younger people have a relatively higher level of attachment than older people.
Contrary to the results of previous studies, we find that men are more attached
to their community than women and that people with higher education levels are
more attached to their community than less-educated people.
The results of the second-stage equation indicate that absolute household
income and community attachment level have a positive effect on LS, while RI
affects LS negatively. These results support the findings of previous studies. In
addition, we find a robust positive impact of the richer dummy. This result implies
that the overall income effect differs depending on RI level and that richer people
in the neighbourhood are more satisfied than their neighbours with below average
Table 7.3 Estimation results (full sample)
Endogenous variable: Model (1) Model (2) Model (3)

Community attachment
Age −1.20E-03*** −1.27E-03*** −1.26E-03***

Male 0.12*** 0.12*** 0.12***
Education 0.04*** 0.04*** 0.04***
Length of residence: 0.10*** 0.10*** 0.10***
1 year or more and less than
3 years
3 years or more and less
than 5 years
than 10 years
than 20 years
than 30 years
than 40 years
more than 40 years
Homeownership 0.18*** 0.18*** 0.18***
Relationship with local 0.40*** 0.40*** 0.40***
residents
Required time to closest −0.01*** −4.97E-03*** −4.97E-03***
station
Required time to bus stop −3.46E-03*** −3.45E-03*** −3.45E-03***
Required time to store −1.94E-03*** −1.99E-03*** −1.99E-03***
Environmental satisfaction 0.13*** 0.13*** 0.13***
Safety 0.35*** 0.35*** 0.35***
Extraversion 0.08*** 0.08*** 0.08***
Agreeableness 0.08*** 0.08*** 0.08***
Conscientiousness 0.01 0.01 0.01
Neuroticism 0.02*** 0.02*** 0.02***
Openness 0.04*** 0.04*** 0.04***
Log household income −2.34E-03 −0.01 −0.01
Log area average income 0.14*** 0.14***
Richer dummy 0.01
Constant −0.48*** −2.53*** −2.53***
(Continued )
Endogenous variable: Model (1) Model (2) Model (3)

Community attachment
Endogenous variable: Life satisfaction

attachment 0.13*** 0.13*** 0.13***
age18 −0.03*** −0.03*** −0.03***
age18sq 3.66E-04*** 3.56E-04*** 3.54E-04***
male_dum18 −0.15*** −0.15*** −0.15***
marriage18 0.30*** 0.30*** 0.30***
college_dum 0.05*** 0.06*** 0.06***
Subjective Health 0.21*** 0.21*** 0.21***
Environmental satisfaction 0.13*** 0.13*** 0.13***
Safety 0.11*** 0.11*** 0.11***
Extraversion 0.06*** 0.06*** 0.06***
Agreeableness 2.41E-03 1.90E-03 1.88E-03
Conscientiousness 0.02*** 0.02*** 0.02***
Neuroticism −0.11*** −0.11*** −0.11***
Openness −0.01*** −0.01*** −0.01***
Log household Income 0.17*** 0.17*** 0.14***
Log area average income −0.13*** −0.09***
Richer dummy 0.06***
Constant −0.49*** 1.36*** 1.32***
Sample size 96,312 96,312 96,312
Note: ***, **, * denote 1%, 5%, 10% significance levels, respectively. All models were estimated
with robust standard errors.
area income.8 All the remaining control variables are statistically significant at
the 10% level or better. We find that females have higher LS than men and that
the married respondents have higher LS than the unmarried respondents. We also
observe a U-shaped effect of age and find that people with a university degree
or higher are more satisfied with their life than those with lower levels of educa-
tion. Additionally, subjective health evaluation, neighbourhood safety and living
environment have positive impacts on overall LS. The direction of the impacts of
personal characteristics and subjective evaluations of the community are consis-
tent with the findings of previous studies.
We repeat the analysis of the same empirical model employed in the main anal-
yses using sub-samples divided by age group, since age, in addition to variables
such as years spent in the current neighbourhood, also affects community attach-
ment and LS. Table 7.4 presents the estimation results for four sub-sample analy-
ses of different age groups. While the overall results of the sub-sample analyses
by age cohort indicate that our findings from the full sample analysis are robust,
we also observe some variance in the impact of the explanatory factors across the
age groups in both the first- and second-stage estimation results.
Table 7.4 Estimation results (subsample)
Endogenous Model (3) Model (3) Model (3) Model (3)

variable: Age (18–39) Age (40–49) Age (50–59) Age (60 or
Community above)
attachment
Male 0.07*** 0.13*** 0.17*** 0.09***

Education 0.03** 0.07*** 3.66E-04 0.01
Length of residence: 0.08*** 0.16*** 0.09** 0.03
1 year or more and
less than 3 years
Length of residence: 0.13*** 0.20*** 0.13*** 0.04
3 years or more and
less than 5 years
Length of residence: 0.19*** 0.24*** 0.22*** 0.12**
5 years or more and
less than 10 years
Length of residence: 0.25*** 0.28*** 0.28*** 0.14***
10 years or more
and less than 20
years
20 years or more
and less than 30
years
30 years or more
and less than 40
years
more than 40 years
Homeownership 0.17*** 0.18*** 0.17*** 0.17***
Relationship with 0.32*** 0.40*** 0.44*** 0.47***
local residents
Required time to −4.73E-03*** −0.01*** −4.52E-03*** −3.80E-03***
closest station
Required time to −3.73E-03*** −4.00E-03*** −3.44E-03*** −2.34E-03***
bus stop
Required time to −2.53E-03*** −2.73E-03*** −1.32E-03** −1.87E-03***
store
Environmental 0.12*** 0.13*** 0.12*** 0.14***
satisfaction
Safety 0.35*** 0.33*** 0.35*** 0.36***
Extraversion 0.08*** 0.08*** 0.09*** 0.07***
Agreeableness 0.10*** 0.06*** 0.09*** 0.07***
Conscientiousness 8.33E-04 0.01* −4.38E-03 0.02*
Neuroticism 0.02** 0.03*** 0.02*** 0.01
(Continued )
Endogenous Model (3) Model (3) Model (3) Model (3)

variable: Age (18–39) Age (40–49) Age (50–59) Age (60 or
Community above)
attachment
Openness 0.04*** 0.04*** 0.05*** 0.02**

Log household 2.30E-03 −0.04*** 0.01 0.02
income
Log area average 0.16*** 0.16*** 0.11*** 0.15***
income
Richer dummy 0.02 0.02 1.09E-03 0.01
Constant −2.73*** −2.47*** −2.59*** −3.09***
Endogenous variable: Life satisfaction
Community 0.10*** 0.14*** 0.13*** 0.14***
attachment
Male −0.18*** −0.18*** −0.16*** −0.05***
Married 0.36*** 0.28*** 0.23*** 0.19***
Education 0.09*** 0.06*** 0.03*** 0.02*
Health 0.25*** 0.22*** 0.20*** 0.15***
Environmental 0.13*** 0.14*** 0.14*** 0.11***
satisfaction
Safety 0.10*** 0.10*** 0.12*** 0.10***
Extraversion 0.08*** 0.06*** 0.06*** 0.04***
Agreeableness 0.03*** 2.60E-03 −0.02** −0.01
Conscientiousness 0.01 0.02*** 0.02*** 0.05***
Neuroticism −0.10*** −0.11*** −0.12*** −0.09***
Openness −1.68E-04 −0.01* −0.02** −4.56E-03
Log household 0.09*** 0.18*** 0.18*** 0.13***
Income
Log area average −0.13*** −0.10*** −0.07** −0.07*
income
Richer dummy 0.07*** 0.12*** 0.07*** −0.02
Constant 1.83*** 0.09 −0.15 0.86*
Sample size 20,286 30,317 27,140 18,569
Note: ***, **, * denote 1%, 5%, 10% significance level, respectively. All models were estimated
with robust standard errors.
In the first-stage analysis, there are several notable findings in terms of the
income-related variables. First, the richer dummy did not have an impact on
attachment level in any age group, and this result is consistent with the full-sample
result. Second, absolute household income affects community attachment nega-
tively among respondents in the 40–49 age group. This may be due to the trend
in the labour force in which workers in their forties are more likely to experience
relocation due to job-related reasons, and this trend appears to be prominent
among high-earning professionals working for large companies. According to the
most recent report9 by the Ministry of Health, Labour and Welfare, people in their
thirties migrate the most in terms of absolute numbers, and they tend to move
to find better residences. By contrast, people in their mid-forties to mid-fifties
move because of their jobs and are likely to have jobs that require frequent and
cyclical transfer to different locations; these individuals are known as Tenkin-zoku
in Japan. Finance and insurance companies are the major industries with a high
share of Tenkin-zoku, and an official report by the Japanese National Tax Agency
claimed that employees in these industries have relatively higher income.
Third, the positive impact of reference income is smaller in the 50–59 age
group than in the other age groups. Compared to younger generations, members
of this age group may care less about the status or information effects of living
in areas with higher RI. Additionally, even if people in this age cohort migrate
because of a job-related transfer, they may be indifferent to their new community
because they have high expectations of returning to their original neighbourhood
within a few years or after retirement.
Regarding the results of the second-stage analyses of the sub-samples by age
group, we find a larger positive effect of absolute income for people in their forties
and fifties, which suggests that income is relatively more important for people in
these age groups than for those in other age groups. Thus, people in their forties
and fifties may value the same unit of income more than people in younger or
older age groups. Additionally, reference area income has a larger negative impact
on younger age groups than on other age groups. Furthermore, we observe that
the richer dummy has the greatest positive impact on LS for the 40–49 age group
and that it has no statistically significant impact on LS in the over-60 age group.
Using the coefficients from the estimation results presented in Tables 7.3 and
7.4, we calculate the direct, indirect and total effects of RI on LS. Table 7.5a
presents the impact of a 1% increase in RI on LS for each sample group in our
analysis. Overall, the direct negative impact of RI is larger than the indirect posi-
tive impact through community attachment; thus, the total effect is negative in
all cases. Nonetheless, the indirect effect is sufficiently large, and the magnitude
of the direct impact suggests that the negative effect of RI is stronger for the
younger population. Given that we observe a statistically significant indirect posi-
tive impact of RI for all age groups, the combined effect of the positive and nega-
tive consequences of RI depend on an individual’s life stage.
We further examine the direct and indirect effects of RI on LS using larger
reference areas: 1500 m mesh, 2500 m mesh, and 3500 m mesh areas and the
municipality level. Tables 7.5b–e provide the results for various reference areas
that are comparable to the results of the 500 m mesh area presented in Table 7.5a.
Overall, there are no significant differences in income effects for the full sample
across different reference areas. However, the results for the larger reference areas
vary from the 500 m mesh results when we analyse the RI effect by age group.
The most significant difference is observed in the results for the largest reference
areas: we do not find a statistically significant direct negative effect of RI on LS
Table 7.5a Direct, indirect and total effects of relative income on life satisfaction (500 m
mesh)
Direct effect Indirect effect Total effect

Full sample −0.091 0.018 −0.073
Age (18–39) −0.133 0.016 −0.118
Age (40–49) −0.096 0.023 −0.073
Age (50–59) −0.068 0.015 −0.053
Age (60 and above) −0.068 0.02 −0.047
Note: We show only statistically significant effects.
Table 7.5b Direct, indirect and total effects of relative income on life satisfaction (1500 m
mesh)

Full sample −0.104 0.021 −0.083
Age (18–39) −0.122 0.021 −0.101
Age (40–49) −0.134 0.024 −0.110
Age (50–59) −0.074 0.019 −0.055
Age (60 and above) −0.073 0.025 −0.049
Note: See Table 7.5a.
Table 7.5c Direct, indirect and total effects of relative income on life satisfaction (2500 m
mesh)

Full sample −0.100 0.022 −0.078
Age (18–39) −0.109 0.020 −0.089
Age (40–49) −0.136 0.024 −0.112
Age (50–59) −0.071 0.021 −0.051
Age (60 and above) −0.066 0.027 −0.039
Table 7.5d Direct, indirect and total effects of relative income on life satisfaction (3500 m
mesh)

Full sample −0.095 0.022 −0.073
Age (18–39) −0.105 0.022 −0.083
Age (40–49) −0.133 0.024 −0.110
Age (50–59) −0.067 0.021 −0.046
Age (60 and above) 0.027 0.027
Table 7.5e Direct, indirect and total effects of relative income on life satisfaction (munici-
pality level)

Full sample −0.084 0.023 −0.062
Age (18–39) −0.089 0.024 −0.065
Age (40–49) −0.112 0.023 −0.089
Age (50–59) 0.020 0.020
Age (60 and above) 0.029 0.029
for respondents over 60 years of age; however, this effect disappears when we
widen the reference area to a 3500 m mesh and larger. Additionally, if we use RI at
the municipality level, we do not observe a statistically significant direct effect of
RI on LS for respondents in their fifties. However, there are robust indirect posi-
tive effects of RI in all sub-samples by age. Thus, the total effects are equivalent
to the magnitude of the indirect effects. Additionally, the magnitude of the indirect
positive effects of RI in larger references areas are larger than the indirect effect
of the 500 m mesh average RI.
4 Discussion
Overall, the results of this analysis reveal both negative and positive effects of RI
on LS. Hence, the total effects observed in the previous literature may have over-
estimated either the negative or the positive effects of RI. In the case of Japan,
the total RI effect on well-being is negative. Hence, the envy effect appears to
outweigh the benefits of living in a richer neighbourhood. This result may be
partly due to the very competitive nature of Japanese society, which places con-
siderable pressure on individuals to maintain high performance and a reputable
economic status.
Nevertheless, we observe a substantial positive effect of RI on SWB through
the positive effect of RI on community attachment, which mitigates the total nega-
tive impact of RI on LS. Given that we use spatial reference groups to calculate
the RI of a respondent, the positive effect we observe is likely to be a proxy for
a status effect, which should be distinguished from the Hirschman-type informa-
tion effects that are observed when using reference characteristics for occupation
in high mobility/uncertainty countries such as Russia, the former East Germany
and other former Soviet Union countries in Eastern Europe (Senik, 2004, 2008;
Welsch and Kühling, 2015; Caporale et al., 2009). In addition to status effects,
altruism within a close community, as suggested by Kingdon and Knight’s (2007)
study in South Africa, could also explain our finding of the positive effects of
RI through community attachment. While the signalling effect may not be very
strong in developed countries such as Japan, a positive boost from altruism may
exist regardless of a country’s development level.
We also find that people with a household income that is higher than the aver-
age area income are more satisfied with their life. In other words, people with rela-
tively low income who are surrounded by richer people are the least satisfied with
their lives. This result provides partial supportive evidence for the ‘spiteful egali-
tarianism’ concept, which is based on the notion that “it makes me worse-off to
see the richer getting richer” (Luttmer, 2005). Within the context of our analysis,
this result also indicates that being relatively rich in a neighbourhood mitigates
the negative effect of RI. This result appears to be sensible, since being richer than
neighbours can spare people from envying others and could result in such people
acquiring positive utility from being envied and respected by the others.
The sub-sample analyses by age group indicate that income effects vary sub-
stantially depending on age. In particular, the magnitude of the negative effect
for the younger population, those under the age of 40 years, is noticeably larger
than that of older age groups, which is consistent with previous findings. Felix
and Richard (2015) found the strongest income effect for individuals in their thir-
ties to fifties compared with younger and older age groups. Additionally, Hsieh
(2011) suggested that the effect of household income on happiness is significantly
smaller for older adults than for young or middle-aged adults.10 The findings of
the present study could be attributed to the combination of a high-pressure envi-
ronment with the increased financial responsibilities that the younger population
faces and the tendency for envy/jealousy effects to decline with age. Clark and
Senik (2010) demonstrated that the act of comparing oneself with others makes
people unhappy and that people are most likely to compare themselves with col-
leagues. During retirement, people may be less likely to compare themselves with
others because they have lost this familiar reference group.
We also find that using the RI of different reference areas affects the results
of the RI effect on LS. Kingdon and Knight (2007) noted that although Luttmer
claimed to observe a negative impact of neighbours’ RI on SWB in the United
States, the public micro-data areas used in Lutmmer’s study included an average
of 150,000 inhabitants, which is hardly the size of a ‘neighbourhood’. As noted
earlier, we observe very different overall results when we change the size of the
reference area. However, a comparison of the results for reference areas of vari-
ous size indicates that the direct negative effect of RI is significant for smaller
reference area units. Additionally, we find that older residents’ LS is affected by
RI in relatively smaller areas but that this negative impact disappears when the
reference area is expanded to a 3500 m mesh area or larger. Such variation in the
results depending on the size of the reference area may be partially explained by
the decline of mobility and the range of living areas as people age. Additionally,
our results indicate that the magnitude of the RI impact varies with the size of
the reference area: the indirect positive effects of RI on LS through community
attachment are larger when we use larger reference areas. This result suggests that
people base the concept of place attachment on the municipality rather than the
smaller neighbourhood surrounding their residence.
Our results not only extend the empirical findings of RI studies but also add
to the understanding of what determines place attachment and how community
attachment affects LS. Previous studies have found varying income effects on
community attachment. Taylor et al. (1985) reported that community attachment
in neighbourhoods with a high concentration of richer people was higher than that
in areas with greater diversity in terms of income. Anton and Lawrence (2014)
also observed that richer people have higher levels of community attachment
and suggested that mobility increases with income and that richer people have
relatively more freedom to choose where they want to live. By contrast, place
attachment may be strong in lower-income areas because of the residents’ relative
isolation and because the residents provide support to one another (Fried, 2000).
This study demonstrates a robust positive reference income effect on com-
munity attachment in Japan. This result may be explained by the relatively low
degree of isolation of any particular region or area given Japan’s well-developed
physical and social infrastructure. Our results also indicate that own income or
being richer or poorer than others in the neighbourhood affects people’s degree
of attachment to their neighbourhood. This result appears to suggest that regard-
less of their income or their placement in an income class, people attain similar
utility from the benefits associated with living in a rich neighbourhood, and vice
versa. Furthermore, we confirm the findings of previous studies: there is a posi-
tive impact of community attachment on reported well-being (Powdthavee, 2008;
Frey and Stutzer, 2002; Bjørnskov, 2003; Bruni and Stanca, 2008; Becchetti et al.,
2008; Binder and Freytag, 2013; Tsurumi and Managi, 2015; Helliwell and Put-
nam, 2004).11
5 Conclusion
This study examined the impact of RI on LS by using SEM to distinguish the
indirect effect of RI through community attachment. We used data collected from
an original large-scale survey conducted in Japan and matched the geo-coded
individual survey data with reference area income in 500 m mesh units using
alternative income data sources based on official statistics.
Our findings reveal a positive effect of absolute household income level on
self-reported LS, which is consistent with previous studies, as well as a positive
impact of community attachment on LS, which supports the results of previous
studies. The results of a mediator analysis indicate that there is a somewhat intri-
cate relationship between RI and LS. We observe a robust indirect positive effect
of reference area income on LS through its positive effect on community attach-
ment. Given that the direct effect of reference income level on LS is negative
and larger than the indirect effect, the total effect remains negative. However, the
magnitude of the indirect effect is sufficiently large that if this indirect positive
effect of RI is not taken into account, the total effect of RI is likely to be overesti-
mated. In addition, our sub-sample analyses indicate that older people and richer
people with above average reference area income experience a relatively smaller
impact of RI on LS than their younger and less-wealthy counterparts.
Our results suggest that overall, income inequality in a neighbourhood has
a negative impact on people’s SWB; however, the elimination or reduction of
inequality at the micro-geographic level is difficult, and its necessity is still
debated. Nevertheless, our findings suggest that income has a positive effect on
community attachment, which in turn has a positive effect on LS. Neighbour-
hood- and municipality-level efforts to maintain living standards and increase
residents’ attachment to their community would contribute to an overall increase
in residents’ welfare.
The empirical results of this study provide additional insight into the rela-
tionship between RI and SWB, and the analysis could be extended via several
avenues. We have identified an indirect positive effect of reference income on
well-being through community attachment; however, further examination and
clarification of the mechanisms of these RI effects is needed. The impact of sta-
tus or reputation effects and the impact of altruism, as suggested by Kingdon
and Knight (2007), are possible explanations for the indirect positive effect and
should be investigated further. Time-series data could provide more evidence of
the effects of age and economic mobility on the direct negative impact of RI on
people’s well-being. Although cross-country comparisons on this topic are very
limited due to data availability, a comparison of income effects across various
cultures with different socio-economic statuses, such as those of developed and
developing countries, may be possible. Finally, further studies that include an
expanded combination of reference characteristics may be able to provide empiri-
cal evidence that further distinguishes the mechanisms of the positive impact of
RI on SWB.
Notes
1 Easterlin (1974, 1995) proposed the Easterlin Paradox. However, Hagerty and Veen-
hoven (2003) and Stevenson and Wolfers (2008) found evidence that appears to dis-
prove the paradox (positive income effects at both the individual and country levels).
Easterlin (2010) reaffirmed the paradox.
2 This method divides Germany into 1.5 million street sections in which a section con-
tains approximately 25 households on average.
3 Age (Sampson, 1988; Bonaiuto et al., 1999; Lewicka, 2010; Hidalgo and Hernández,
2001; Brown et al., 2004; Ng et al., 2005; Shamai and Ilatov, 2005; Lewicka, 2010;
Belanche et al., 2016; Wirth et al., 2016), gender (Mesch and Manor, 1998; Hidalgo
and Hernández, 2001; Brown et al., 2004; Lewicka, 2005), education (Belanche et
al., 2016; Wirth et al., 2016), income (Taylor et al., 1985; Brown et al., 2004; Anton
and Lawrence, 2014), house ownership (Mesch and Manor, 1998; Brown et al., 2003;
Lewicka, 2010) and years in residence (Kasarda and Janowitz, 1974; Sampson, 1988;
Brown et al., 2003; Brown et al., 2004; Shamai and Ilatov, 2005; Lewicka, 2010; Wirth
et al., 2016).
4 Taylor et al. (1985), Sampson (1988), Mesch and Manor (1998), Bonaiuto et al. (1999),
Brown et al. (2004), Lewicka (2005), Brown and Raymond (2007), Lewicka (2010),
Anton and Lawrence (2014), Casakin et al. (2015).
5 The OECD’s Better Life Index includes community attachment as a determinant of the
aggregate happiness measure.
6 A total of 24,912 respondents refused to provide their household income and/or the
postal code for their residences. Two hundred ninety respondents provided an incorrect
postal code that could not be geo-coded.
7 Some studies adopt personality traits for analysis to control individual random effects
(see Boyce and Wood, 2011; Ambrey and Fleming, 2011, 2013; Ambrey et al., 2014;
Goerke and Pannenberg, 2015; Powdthavee et al., 2015).
8 This result is consistent with the results that Oshio et al. (2011) observed in their
analysis.
9 National Institute of Population and Social Security Research (2013). Overview of the
Results of the Seventh National Survey on Migration (p. 8), www.ipss.go.jp/ps-idou/e/
m07e/mig07e_summary.pdf.
10 These two previous studies analyzed the absolute income effect rather than the impact
of RI.
11 Bjørnskov (2003) used trust and citizen participation behaviour to construct an indi-
vidual-level social capital measure and found that the positive impact of social capital
is more influential than the income effect. Moreover, Helliwell and Putnam (2004)
found that relationships with people close to the respondents and other social capital
measures have an indirect positive effect on SWB through health.
Bibliography
Ambrey, C. L., Fleming, C. M. (2011). Valuing Scenic Amenity Using Life Satisfaction
Data. Ecological Economics. 72: 106–115.
Ambrey, C. L., Fleming, C. M. (2013). Public Greenspace and Life Satisfaction in Urban
Australia. Urban Studies. 51(6): 1290–1321.
Ambrey, C. L., Fleming, C. M., Chan, A. Y. (2014). Estimating the Cost of Air Pollution in
South East Queensland: An Application of the Life Satisfaction Non-Market Valuation
Approach. Ecological Economics. 97: 172–181.
Anton, C. E., Lawrence, C. (2014). Home Is Where the Heart Is: The Effect of Place of
Residence on Place Attachment and Community Participation. Journal of Environmental
Psychology. 40: 451–461.
Becchetti, L., Pelloni, A., Rossetti, F. (2008). Relational Goods, Sociability, and Happiness.
Kyklos. 61: 343–363.
Belanche, D., Casaló, L. V., Orús, C. (2016). City Attachment and Use of Urban Services:
Benefits for Smart Cities. Cities. 50: 75–81.
Binder, M., Freytag, A. (2013). Volunteering, Subjective Well-Being and Public Policy.
Journal of Economic Psychology. 34: 97–119.
Bjørnskov, C. (2003). The Happy Few: Cross-Country Evidence on Social Capital and Life
Satisfaction. Kyklos. 56: 3–16.
Bonaiuto, M., Aiello, A., Perugini, M., Bonnes, M., Ercolani, A. P. (1999). Multidimen-
sional Perception of Residential Environment Quality and Neighbourhood Attachment
in the Urban Environment. Journal of Environmental Psychology. 19: 331–352.
Boyce, C. J., Wood, A. M. (2011). Personality and the Marginal Utility of Income: Person-
ality Interacts with Increases in Household Income to Determine Life Satisfaction. Jour-
nal of Economic Behavior & Organization. 78: 183–191.
Brown, B. B., Perkins, D. D., Brown, G. (2003). Place Attachment in a Revitalizing Neigh-
borhood: Individual and Block Levels of Analysis. Journal of Environmental Psychol-
ogy. 23: 259–271.
Brown, B. B., Perkins, D. D., Brown, G. (2004). Incivilities, Place Attachment and Crime:
Block and Individual Effects. Journal of Environmental Psychology. 24: 359–371.
Brown, G., Raymond, C. (2007). The Relationship between Place Attachment and Land-
scape Values: Toward Mapping Place Attachment. Applied Geography. 27: 89–111.
Brown, S., Gray, D., Roberts, J. (2015). The Relative Income Hypothesis: A Comparison
of Methods. Economics Letters. 130: 47–50.
Bruni, L., Stanca, L. (2008). Watching Alone: Relational Goods, Television and Happiness.
Journal of Economic Behavior and Organization. 65: 506–528.
Caporale, G. M., Georgellis, Y., Tsitsianis, N., Yin, Y. P. (2009). Income and Happiness
across Europe: Do Reference Values Matter? Journal of Economic Psychology. 30:
42–51.
Casakin, H., Hernández, B., Ruiz, C. (2015). Place Attachment and Place Identity in Israeli
Cities: The Influence of City Size. Cities. 42: 224–230.
Clark, A. E., Oswald, A. (1996). Satisfaction and Comparison Income. Journal of Public
Economics. 61: 359–381.
Clark, A. E., Senik, C. (2010). Who Compares to Whom? The Anatomy of Income Com-
parisons in Europe. The Economic Journal. 120: 573–594.
Diener, E. (2006). Guidelines for National Indicators of Subjective Well-Being and Ill-
Being. Applied Research in Quality of Life. 1: 151–157.
Diener, E., Seligman, M.E.P. (2004). Beyond Money. Psychological Science in the Public
Interest. 5: 1–31.
Donaldson, S. I., Dollwet, M., Rao, M. A. (2015). Happiness, Excellence, and Optimal
Human Functioning Revisited: Examining the Peer-Reviewed Literature Linked to Posi-
tive Psychology. The Journal of Positive Psychology. 10(3): 185–195.
Duesenberry, J. S. (1949). Income, Saving and the Theory of Consumer Behavior. Boston:
University of Harvard Press.
Easterlin, R. A. (1974). Does Economic Growth Improve the Human Lot? Some Empirical
Evidence. In P. A. David and M. W. Reder (Eds.), Nations and Households in Economic
Growth: Essays in Honor of Moses Abramovitz (pp. 89–125). New York: Academic
Press.
Easterlin, R. A. (1995). Will Raising the Incomes of All Increase the Happiness of All.
Journal of Economic Behavior and Organization. 27: 35–47.
Easterlin, R. A. (2010). Well-Being, Front and Center: A Note on the Sarkozy Report.
Population and Development Review. 36: 119–124.
Felix, C., Richard, L. E. (2015). When Does Money Matter Most? Examining the Associa-
tion between Income and Life Satisfaction over the Life Course. Psychology and Aging.
30: 120–135.
Ferrer-i-Carbonell, A. (2005). Income and Well-Being: An Empirical Analysis of the Com-
parison Income Effect. Journal of Public Economics. 89: 997–1019.
Frey, B. S., Stutzer, A. (2002). Happiness and Economics. Princeton and Oxford: Princeton
University Press.
Fried, M. (2000). Continuities and Discontinuities of Place. Journal of Environmental Psy-
chology. 20: 193–205.
Goerke, L., Pannenberg, M. (2015). Direct Evidence for Income Comparison and Subjec-
tive Well-Being across Reference Groups. Economic Letters. 137: 95–101.
Gosling, S. D., Rentfrow, P. J., Swann Jr., W. B. (2003). A Very Brief Measure of the Big-
Five Personality Domains. Journal of Research in Personality. 37: 504–528.
Hagerty, M. R., Veenhoven, R. (2003). Wealth and Happiness Revisited: Growing National
Income Does Go with Greater Happiness. Social Indicators Research. 64(1): 1–27.
Hayes, A. F., Preacher, K. J. (2014). Satisfaction Mediation Analysis with a Multicategori-
cal Independent Variable. British Journal of Mathematical and Statistical Psychology.
67: 451–470.
Helliwell, J. F. (2003). How’s Life? Combining Individual and National Variables to
Explain Subjective Well-Being. Economic Modelling. 20: 331–360.
Helliwell, J. F., Putnam, R. R. (2004). The Social Context of Well-Being. Philosophical
Transactions of the Royal Society B. 29: 1435–1446.
Hidalgo, M. C., Hernández, B. (2001). Place Attachment: Conceptual and Empirical Ques-
tions. Journal of Environmental Psychology. 21: 273–281.
Hirschman, A. O., Rothschild, M. (1973). The Changing Tolerance for Income Inequality in the
Course of Economic Development. The Quarterly Journal of Economics. 87: 544–566.
Hsieh, C. M. (2011). Money and Happiness: Does Age Make a Difference?. Ageing and
Society. 31: 1289–1306.
Kahneman, D., Deaton, A. (2010). High income improves evaluation of life but not emo-
tional well-being. Proceedings of the National Academy of Sciences. 107 (38):
16489–16493.
Kahneman, D., Krueger, A. B., Schkade, D., Schwarz, N., Stone, A. (2004). Toward
National Well-Being Accounts. The American Economic Review. 94: 429–434.
Kahneman, D., Tversky, A. (1979). On the Interpretation of Intuitive Probability: A Reply
to Jonathan Cohen. Cognition. 7: 409–411.
Kasarda, J. D., Janowitz, M. (1974). Community Attachment in Mass Society. American
Sociological Review. 39: 328–339.
Kingdon, G. G., Knight, J. (2007). Community, Comparisons and Subjective Well-Being
in a Divided Society. Journal of Economic Behavior and Organization. 64: 69–90.
Knies, G. (2012). Income Comparisons among Neighbours and Satisfaction in East and
West Germany. Social Indicators Research. 106: 471–489.
Lange, T. (2010). Culture and Life Satisfaction in Developed and Less Developed Nations.
Applied Economics Letters. 17: 901–906.
Lewicka, M. (2005). Ways to Make People Active: The Role of Place Attachment, Cultural
Capital, and Neighborhood Ties. Journal of Environmental Psychology. 25: 381–395.
Lewicka, M. (2010). What Makes Neighborhood Different from Home and City? Effects
of Place Scale on Place Attachment. Journal of Environmental Psychology. 30: 35–51.
Li, Z., Folmer, H., Xue, J. (2014). To What Extent Does Air Pollution Affect Happiness?
The Case of the Jinchuan Mining Area, China. Ecological Economics. 99: 88–99.
Linden, A., Karlson, A. K. (2013). Using Mediation Analysis to Identify Causal Mecha-
nisms in Disease Management Interventions. Health Services Outcomes Res Method. 13:
86–108.
Luttmer, E.F.P. (2005). Neighbors as Negatives: Relative Earnings and Well-Being. Quar-
terly Journal of Economics. 120: 963–1002.
McBride, M. (2001). Relative-Income Effects on Subjective Well-Being in the Cross-
Section. Journal Economic Behavior and Organization. 45: 251–278.
Mesch, G., Manor, O. (1998). Social Ties, Environmental Perception, and Local Attach-
ment. Environment and Behavior. 30: 504–519.
Mohanty, M. S., Ullah, A. (2012). Direct and Indirect Effects of Happiness on Wage: A
Simultaneous Equations Approach. The Journal of Socio-Economics. 41: 143–152.
Ng, S. H., Kam, P. K., Pong, R.W.M. (2005). People Living in Ageing Buildings: Their
Quality of Life and Sense of Belonging. Journal of Environmental Psychology. 25:
347–360.
Oshio, T., Nozaki, K., Kobayashi, M. (2011). Relative Income and Happiness in Asia:
Evidence from Nationwide Surveys in China, Japan, and Korea. Social Indicators
Research. 104: 351–367.
Powdthavee, N. (2005). Unhappiness and Crime: Evidence from South Africa. Economica.
72: 531–547.
Powdthavee, N. (2008). Putting a Price Tag on Friends, Relatives, and Neighbours: Using
Surveys of Life Satisfaction to Value Social Relationships. The Journal of Socio-
Economics. 37: 1459–1480.
Powdthavee, N., Lekfuangfu, W. N., Wooden, M. (2015). What’s the good of education on
our overall quality of life? A simultaneous equation model of education and life satisfac-
tion for Australia. Journal of Behavioral and Experimental Economics. 54: 10–21.
Preacher, K. J., Hayes, A. F. (2008). Asymptotic and Resampling Strategies for Assessing
and Comparing Indirect Effects in Multiple Mediator Models. Behavior Research Meth-
ods. 40: 879–891.
Sampson, R. J. (1988). Local Friendship Ties and Community Attachment in Mass Society:
Multilevel Systemic Model. American Sociological Review. 53: 766–779.
Senik, C. (2004). When Information Dominates Comparison Learning from Russian Sub-
jective Panel Data. Journal of Public Economics. 88: 2099–2123.
Senik, C. (2008). Ambition and Jealousy: Income Interactions in the ‘Old’ Europe versus
the ‘New’ Europe and the United States. Economica. 75: 495–513.
Shamai, S., Ilatov, Z. (2005). Measuring Sense of Place: Methodological Aspects. Tijd-
schrift Voor Economische en Sociale Georgrafie. 96: 467–476.
Stevenson, B., Wolfers, J. (2008). Economic growth and subjective well-being: Reassessing
the easterlin paradox. NBER Working Paper No. 14282.
Stevenson, B., Wolfers, J. (2013). Subjective Well-Being and Income: Is There Any Evi-
dence of Satiation? American Economic Review. 103: 598–604.
Taylor, R. B., Gottredson, S. D., Brower, S. (1985). Attachment to Place: Discriminant
Validity, and Impacts of Disorder and Diversity. American Journal of Community Psy-
chology. 13: 525–542.
Tsurumi, T., Managi, S. (2015). Environmental Values of Green Spaces in Japan: An Appli-
cation of the Life Satisfaction Approach. Ecological Economics. 120, 1–12.
Welsch, H., Kühling, J. (2015). Income Comparison, Income Formation, and Subjective
Well-Being: New Evidence on Envy versus Signaling. Journal of Behavioral and Exper-
imental Economics. 59(C): 21–31.
Wirth, V. T., Grêt-Regamey, A., Moser, C., Stauffacher, M. (2016). Exploring the Influence
of Perceived Urban Change on Residents’ Place Attachment. Journal of Environmental
Psychology. 46: 67–82.
Appendix
Table 7A.1 Variables
Variable name Definition
(1) Life satisfaction Questionnaire: How satisfied are you with your life as a
whole?
Completely dissatisfied = 1, Completely satisfied = 5 (scale
1–5)
(2) Community Questionnaire: How attached are you to your local
attachment community?
Completely detached = 1, Completely attached = 5 (scale
1–5)
(3) Extraversion Questionnaire: Please select an item which is the most
applicable.
 Extraverted/Enthusiastic Completely disagree = 1,

Strong agree = 5 (scale 1–5)
 Reserved/Quite Completely disagree = 1, Strong agree =
5 (scale 1–5)
(4) Agreeableness Questionnaire: Please select an item which is the most

applicable.
 Critical/Quarrelsome Completely disagree = 1, Strong

agree = 5 (scale 1–5)
 Sympathetic/Warm Completely disagree = 1, Strong
(5) Conscientiousness Questionnaire: Please select an item which is the most

applicable.
 Dependable/Self-disciplined Completely disagree = 1,

 Disorganised/Careless Completely disagree = 1, Strong
(Continued )
(6) Neuroticism Questionnaire: Please select an item which is the most

applicable.
 Anxious/Easily upset Completely disagree = 1, Strong

 Calm/Emotionally stable Completely disagree = 1,
(7) Openness Questionnaire: Please select an item which is the most

applicable.
 Open to new experience, complex Completely disagree =

1, Strong agree = 5 (scale 1–5)
 Conventional/Uncreative Completely disagree = 1,
(8) Age Age of respondent

(9) Material status Respondent is married = 1. Not married = 0.
(10) Gender Respondent is male = 1. Respondent is female = 0.
(11) Subjective health Questionnaire: All in all, how would you describe your state
of health?
Very poor = 1, Very good = 5 (scale 1–5)
(12) Household income Choices: 1. Below 200 million = 100 million, 2. 200 million
or above and below 300 million = 250 million, 3. 300
million or above and below 400 million = 350 million,
4. 400 million or above and below 500 million = 450
million, 5. 500 million or above and below 600 million =
550 million, 6. 600 million or above and below 700 mil
lion = 650 million, 7. 700 million or above and below 800
million = 750 million, 8. 800 million or above and below
900 million = 850 million, 9. 900 million or above and
below 1000 million = 950 million, 10. 1000 million or
above and below 1500 million = 1250 million, 11. 1500
12. 2000 million or above and below 3000 million = 2500
million, 13. Above 3000 million = 3000 million
(13) Area average income (1) Below 300 million = 150 million, (2) 300 million or
above and below 500 million = 400 million, (3) 500 million
or above and below 700 million = 600 million, (4) 700
(5) 1000 million or above and below 1500 million = 1250
million, (6) Above 1500 million = 1500 million
Area average income = ((1) × Number of households of
(1) + (2) × Number of households of (2) + (3) × Number
of households of (3) + (4) × Number of households of
(4) + (5) × Number of households of (5) + (6) × Number of
households of (6))/ Total number of households
(14) Length of residence Choices: 1. Below 1 year, 2. 1 year or above and below 3
years, 3. 3 years or above and below 5 years, 4. 5 years or
above and below 10 years, 5. 10 years or above and below
20 years, 6. 20 years or above and below 30 years, 7. 30
years or above and below 40 years, 8. Above 40 years
These are dummy variables respectively (reference = below
1 year).
(15) Required time to Choices: 1. Below 5 minutes = 2.5, 2. 5 minutes or above
closest station and below 10 minutes = 7.5, 3. 10 minutes or above and
below 15 minutes = 12.5, 4. 15 minutes or above and below
30 minutes = 22.5, 5. 30 minutes or above and below 45
minutes = 37.5, 6. 45 minutes or above and below 60
minutes = 52.5, 7. Above 60 minutes = 60
bus stop and below 10 minutes = 7.5, 3. 10 minutes or above and
minutes = 37.5, 6. 45 minutes or above and below 60
store and below 10 minutes = 7.5, 3. 10 minutes or above and
30 minutes = 22.5, 5. 30 minutes or above and below
(18) Homeownership Respondent who living at Homeowner/House, Homeowner/
Mansion or Homeowner/Apartment = 1. The others = 0.
(19) Relationship with Degree of satisfaction of relationship with local residents.
local residents Completely dissatisfied = 1, Completely satisfied = 5 (scale
1–5)
(20) Education Respondent who complete university/college, master’s
education, doctorate education = 1. The others = 0.
(21) Environmental Degree of environmental satisfaction.
satisfaction Completely dissatisfied = 1, Completely satisfied = 5 (scale
1–5)
(22) Safety Questionnaire: Please tell us about safety of your
neighbourhood.
Very dangerous = 1, Very safe = 4 (scale 1–4)
(23) Population density Data source: Japanese National Population Census 2010
Total population/Administrative district area (Population/
km²)
Sample size 96,312
.6
.5
.2 .3 .4
Fraction
.1
0
1 2 3 4 5
Life satisfaction
Figure 7A.1 Distribution of life satisfaction

.6
.5
.2 .3 .4
Fraction
.1
0
1 2 3 4 5
Attachment
Figure 7A.2 Distribution of attachment to a community

8 Environmental value of
green spaces in Japan
An application of the life
satisfaction approach
Tetsuya Tsurumi and Shunsuke Managi
1 Introduction
Understanding the value of greenery has long been a major focus of research
in environmental economics. Various studies have been conducted on this issue,
applying the Contingent Valuation Method (CVM) and the hedonic pricing
method (for the survey: Barrio and Loureiro, 2010; Brander and Koetse, 2011;
Saphores and Li, 2012). There has been an increase in research that has applied
CVM to assess the multiple functions of green spaces because of their applicabil-
ity to both the use and passive use values. However, as Johnston (2006) suggests,
prior assessments have often presented substantial hypothetical bias, superficial
answers, or strategic responses, and most authors have found significant diver-
gence between stated and actual behaviors. Similarly, the hedonic method may
also provide biased results because land prices can be underestimated if changes
in the environment are not immediately reflected in land valuations. Bayer et al.
(2009), for example, claim that conventional hedonic models underestimate the
willingness to pay (WTP) for clean air. The primary problems of the hedonic
methods arise from their dependence on the equilibrium assumption, which is
met only if there is a sufficiently wide variety of houses, prices adjust rapidly,
households have full information, and transaction and moving costs are zero (Frey
et al., 2010).
The present study applies the Life Satisfaction Approach (LSA) to assess the
value of green spaces. The LSA estimates the marginal utility of both income and
non-market goods (such as the functions of a green space) and uses this proxy ratio
to calculate the monetary value of the latter. Because the LSA does not address
direct responses to the environment, we can avoid superficial answers or strategic
responses. Further, since it explicitly captures individual welfare in the absence of
market equilibrium, we can ascertain the full utility consequences independent of
the degree of capitalization in the housing and labor markets (Frey et al., 2010).
Thus, we can expect a different evaluation of the environment from that obtained
under the CVM and hedonic approaches. The purpose of the present study is to
understand the value of green spaces around the residence. We use the objective
indicator, the green coverage rates of the circumference of a residence, calculated
using geographic information systems (GIS) data. Specifically, we investigate
196 Tetsuya Tsurumi and Shunsuke Managi
how the effects of green spaces on the inhabitants differ by its distance from a
residence. We focus on the green spaces in the daily living area, at distances of
0–100 m, 100–300 m, 300–500 m, 500–1000 m, 1000–1500 m, and 1500–2000
m from a residence.
Literature in field experiments and psychology suggests that the interaction
with green spaces has relaxing effects (Hartig et al., 2003; Morita et al., 2007;
Park et al., 2010; Lee et al., 2011; Thompson et al., 2012; Tsunetsugu et al., 2013).
For example, Tsunetsugu et al. (2013) found that a short-term viewing of for-
ests has physiological relaxing effects, such as lowered diastolic blood pressure
and heart rate; and viewing forest landscapes caused higher parasympathetic ner-
vous activity and lower sympathetic nervous activity than urban landscapes, and
induced a positive mood. Even if the green coverage rate is the same, when inhab-
itants have more opportunities to interact with green spaces (or walk) around resi-
dential areas they may receive greater benefits from green spaces than those who
have few opportunities to be in close contact with nature. Similarly, if people have
a great affection for green spaces, or possess knowledge regarding the multiple
functions of green spaces, then they may receive more benefit from the green
spaces than those with the same green coverage rate around their residence. Fur-
thermore, the quality of green spaces may have a relationship to people’s moods.
The present study, therefore, considers people’s preference for greenery using
the following parameters: (1) affection for neighborhood greenery, (2) affection
for world greenery, (3) degree of interaction with greenery in the last 5 years,
(4) degree of remembered interaction with greenery until the age of 12 years,
(5) level of knowledge regarding the multiple functions of forests, (6) quality of
greenery within a 5-minute walk from home, and (7) quality of greenery within a
15-minute walk from home.
This research makes contributions to the existing literature as follows: (1) we
identify the economic value of green spaces around residences in Japan by apply-
ing the LSA to investigate how the effects of green spaces on well-being differ in
terms of existing coverage, distance from the residence, and people’s preference
for greenery; and (2) we contribute to the literature by incorporating endogeneity
into the discussion using instrumental variables in LSA.
The remainder of the chapter is organized as follows. Section 2 reviews recent
studies. Section 3 describes our dataset and Section 4 presents the estimation
model. Section 5 summarizes our estimation results and Section 6 concludes.
2 Literature review
Application of the LSA is growing in the literature. Primarily, previous studies
have used the LSA to estimate the amount of damage caused by air and noise
pollution,1 whereas few analyze the monetary value of green spaces. To our
knowledge, research works on the relationship between green spaces and resi-
dents’ well-being have been presented by Smyth et al. (2008) for China, Ambrey
and Fleming (2011, 2013) for Australia, MacKerron and Mourato (2013) for the
United Kingdom, and Kopmann and Rehdanz (2013) for 31 European countries.
Environmental value of green spaces in Japan 197
Smyth et al. (2008) empirically determine that parks in China’s urban areas
significantly increase well-being. Ambrey and Fleming (2011) also examine the
relationship between parks and well-being. They use regression analysis to exam-
ine the relationship between well-being and “the distance between residences and
parks” in Australia. The results of the analysis show that parks can increase well-
being; especially, parks within 50 km of one’s residence have a relatively larger
positive effect. Ambrey and Fleming (2013) use the more objective green cover-
age rate data from GIS as an index of the amount of green spaces. In particular, the
green coverage rate of a residence in a 700 m neighborhood shows a statistically
significant increase in well-being for inhabitants of urban areas in Australia. Using
this result, green spaces in major cities (parks, community parks, cemeteries, sta-
diums, national parks, and natural reserve areas) have been valued monetarily
using LSA. The result of this analysis shows that the marginal WTP (MWTP)
annually for a 1% increase in green spaces (i.e., 143 m2) is $467 per capita.
MacKerron and Mourato (2013) use unique data obtained from a smart-
phone application that signals participants at random moments and presents a
brief questionnaire while determining the respondent’s geographical coordinates
using global positioning satellites (GPS). They have collected over one million
responses from more than 20,000 participants and investigate the relationship
between momentary subjective well-being and individuals’ immediate environ-
ments within the United Kingdom. The estimation result implies that participants
are significantly and substantially happier when they are outdoors in all green or
natural habitat types than when they are in urban environments. Kopmann and
Rehdanz (2013) use the European Quality of Life Survey (EQLS) from 31 Euro-
pean countries, with a total of 35,634 observations, and the Coordination of Infor-
mation on the Environment (CORINE) Land Cover database, which provides
information for 44 land cover categories in a raster format of 100 m resolution
for European countries (2006 data). The estimation results indicate that MWTPs
tend to be higher for natural areas that are scarcer, and that a nonlinear relation-
ship between land cover and well-being is preferred to a linear relationship, which
implies decreasing benefits from individual landscape amenities.
In the current study, we contribute to the existing literature on LSA by separat-
ing data on green spaces by distance (the shape of donuts) and showing how the
effects of green coverage rates differ by the current levels of green coverage rates,
distance from people’s houses, and people’s preferences for green spaces. We also
contribute to the literature on LSA by considering endogeneity using instrumental
variables and higher-definition land cover data than previous literature.2
With regard to distance form people’s houses, Sander et al. (2010) analyze the
green coverage rates of Ramsey County and the urbanizing Dakota County to the
south in Minnesota (United States). They find that a 10% increase in green cov-
erage within 100 m and 250 m of a home increases its property value by 0.48%
($1371) and 0.29% ($836), respectively. However, a statistically significant rela-
tionship between the green coverage rates and value of properties situated more
than 250 m away is not found. Sander and Haight (2012) focus on the green
coverage rates of Dakota County, and find a 10% increase in green coverage rate
within a 100 m, 250 m, 500 m, and 750 m radius of a home increases its property
value by 0.58% ($1853), 0.32% ($1030), 0.61% ($1947), and 0.35% ($1102),
respectively. No literature applying the LSA considers the effect of green cover-
age rates within walking distance of residences.
In association with people’s preference for greenery, Neilson and Wichmann
(2014) consider the role of social networks in the non-market valuation of public
goods. They construct a model of valuing public goods and find that network
structure almost always matters both for utility and valuation. These authors sug-
gest that even if the presence of public goods (such as a park) does not generate
private utility for an individual, if that public good provides his or her friends with
utility, which he or she values, then the individual in question might have a posi-
tive WTP for that good. The authors also suggest that a certain individual might
only derive enjoyment from the park when visiting with friends rather than alone.
This may imply that memories of interacting public goods with their family or
friends affect valuation of green space around residences. In relation to memories
of interaction, attachment to public goods may affect an individual’s valuation of
them. For instance, Larson et al. (2013) suggest that place of residence, involve-
ment in community activities, country of birth, and the length of time respon-
dents have lived in the region are important determinants of how people value the
natural environment surrounding them. Moreover, understanding the benefits of
the public goods in question may also affect their valuations. No study applying
LSA investigates how the effects of green spaces differ by preferences for green
spaces, such as interaction with, affection to, and knowledge of green spaces;
thus, this study considers these factors.
3 Data
3.1 Respondent data

This study uses data derived from an internet survey conducted on October 11 and
12, 2012, which targeted people residing in Japan’s largest region (the capital area
of Kanto) and second largest region (Kansai). The survey selected respondents
based on prefectural demographics, such as population, sex, and age ratios.3 There
were 2158 valid responses that formed the basis for our subsequent analysis.4
3.2 Data on green spaces

Green spaces data are derived from the “Digital Map 5000” from the Geospa-
tial Information Authority of Japan. Data on the three major urban areas (Kanto,
Chubu, and Kinki regions) are recorded in the Digital Map 5000. Chubu region
was omitted from the analysis because the data in this study were taken from
two periods.5 The Digital Map 5000 classifies land use into 15 categories: for-
ests, wastelands among others; rice fields; fields and other farmland; land being
reclaimed; vacant lots; industrial sites; general low-rise residential areas; dense
low-rise residential areas; mid-to-high-rise residential areas; commercial and
Map 8.1 Locations of “green spaces” and questionnaire respondents in the Tokyo metro-
politan area
business land; road sites; parks and other green space; other sites for communal
and public facilities; rivers and lakes among others; and other. This study defines
“green spaces” as the four categories of forests, wastelands among others, rice
fields, fields and other farmland, and parks and other green spaces. The definition
of each category is shown in Table 8A.1 in the appendix.
Maps 8.1 and 8.2 plot the locations of our respondents in the Kanto and Kansai
regions. The shaded parts correspond to the green spaces. We calculate the green
coverage rate of residences based on respondents’ addresses. Specifically, the
residents’ addresses are plotted and distances of 100 m, 300 m, 500 m, 1000 m,
1500 m, and 2000 m are drawn using the plot as the center; then, we separate the
area by the distance from the center in the following: a circle with a radius of
100 m, a donut shape from 100 to 300 m, from 300 to 500 m, from 500 to 1000
m, from 1000 to 1500 m, and from 1500 to 2000 m. The green coverage rates for
Map 8.2 Locations of “green spaces” and questionnaire respondents in the Kinki region
respondents’ neighborhoods are calculated as the area ratio of “green spaces” in

the separated areas above.6
3.3 Quality of green spaces and diversity of the respondents’

preferences
Because high-quality green spaces may increase well-being to a greater extent,
the quality of such spaces, which cannot be completely measured by the green
coverage rates alone, must also be considered. Moreover, the diversity of prefer-
ences may give rise to conflict. Therefore, in this study, the following five compo-
nents derived from the questionnaire responses are considered:
1 Quality of green spaces

2 Affection for green spaces (neighborhood)
3 Affection for green spaces (worldwide)
4 Degree of interaction with green spaces (in the past 5 years)
5 Degree of interaction with green spaces (until the age of 12 years).
The first question is intended to ascertain the quality of green spaces in a neigh-
borhood.7 The quality of green spaces reflects the possibility that higher quality
greenery translates to a higher positive effect on well-being given the same green
coverage rate. Furthermore, questions 2–5 consider the diversity of preferences
for green spaces. Question 2 asks about the respondent’s attachment to green
spaces in his/her neighborhood; that is, whether one normally has a feeling of
affection toward the neighborhood’s green spaces. Question 3 expands on the
concept of neighborhood and questions the degree of attachment one has to green
spaces throughout the world. An increase in green coverage rate is expected to
have different effects on well-being corresponding to the levels of affection for
green spaces. Question 4 is intended to determine the degree to which the respon-
dent normally has opportunities to interact with green spaces, as it is expected
that more frequent interaction corresponds to a larger effect of the green spaces
on well-being. Question 5 inquires the degree to which the respondent remembers
interacting with green spaces, up to the age of 12 years. This is intended to take
into consideration the possibility that interaction with green spaces at a young
age leads to an attachment to it, and that green spaces may influence well-being
through such attachment. The specific questions are shown in Table 8A.2 in the
appendix.
3.4 Knowledge of green spaces

To evaluate the multiple functions of green spaces and the effect on respon-
dents, the survey ascertains respondents’ knowledge regarding these multiple
functions. According to the Science Council of Japan (2001), the multiple
functions of green spaces include biodiversity conservation, global environ-
mental protection, sediment disaster prevention and soil conservation, water-
sheds, creation of a pleasant environment, health and recreation, culture, and
material production. These functions were presented to the respondents, who
were then asked how much knowledge they had regarding each. They were
asked to rate their level of knowledge on each of the multiple functions on a
scale of 1 (completely unfamiliar with the function) to 11 (very familiar with
the function), and the average was used as an indicator of their knowledge. The
specific functions presented to respondents are shown in Tables 8A.3 and 8A.4
in the appendix.
3. Control variables

Moreover, due to differences in governmental land utilization plans, an increase
in green spaces may affect the convenience of a local area. The circumstances
of people’s lives, means of transportation, and retail availability are consid-
ered important factors for living a comfortable life. Previous studies suggest
that the surrounding environment can have a statistically significant effect on
well-being. For example, Balducci and Checchi (2009) examine the relation-
ship between well-being and living conditions in 10 major cities around the
world (Toronto, New York, London, Paris, Berlin, Stockholm, Milan, Beijing,
Seoul, and Tokyo) by assessing transportation, retail availability, parks, and
community life (e.g., meeting with friends and neighbors, volunteering, and
social activities). Their results indicate that living conditions have a significant
influence on well-being.
Thus, in this study, a convenience metric is generated from the GIS data to con-
sider diversities of local convenience when the green coverage rate is the same.
Specifically, data from the 2005 census is used to create an index of the following
five convenience indicators:8
1 Number of retail stores (within a 2 km radius of home)

2 Number of restaurants (within a 2 km radius of home)
3 Distance to the nearest public cultural facility
4 Distance to the nearest railway station
5 Distance to the nearest bus stop.
In addition, we include population density for town-level residences and a

prefecture dummy to eliminate factors that are not considered by the preceding
convenience indices. Moreover, Balducci and Checchi (2009) find that commu-
nity life is statistically, significantly related to an increase in well-being. Hence,
respondents were asked if there was someone nearby with whom they regularly
consult and their response was used as an indicator in the regression model (see
the questionnaire in Table 8A.2 in Appendix for more details regarding these
questions).9
In addition, we include some of the indicators used in previous studies to avoid
the omitted variable problem in regression analysis. Tella et al. (2001), Blanch-
flower and Oswald (2004), and Peiro (2006) examine how validating factors other
than those related to income influence subjective well-being. Tella et al. (2001)
show that unemployment statistically, significantly reduces well-being, while
Blanchflower and Oswald (2004) show that men have lower levels of well-being
compared with women and that a U-shaped trend accompanies aging, with the
lowest level of well-being found among those in their thirties and forties. Accord-
ing to Peiro (2006), health anxiety has also been shown to statistically, signifi-
cantly reduce well-being.10
In addition, as another control variable, personality indicators were introduced
to the model to remove the impact of fluctuating personalities in the responses
related to well-being. Further, the number of major life events that have occurred
in the past 5 years (e.g., divorce, unemployment, serious sickness or injury, the
death of a close relation) was included to consider individual heterogeneity.
Table 8.1 summarizes the indicators used in this study’s regression model to show
the descriptive statistics of the indicators.
Table 8.1 Descriptive statistics (Obs. = 2158)
Variable Unit Mean Std. dev. Min. Max.

Green coverage rate % 9.43 14.16 0 97.32
(0–100 m)
(0–100 m) (past)
(100–300 m)
(100–300 m) (past)
(300–500 m)
(300–500 m) (past)
(500–1000 m)
(500–1000 m) (past)
(1000–1500 m)
Green coverage % 21.36 15.34 0 81.55
rate (1000–1500 m)
(past)
(1500–2000 m)
Green coverage % 22.71 15.67 0 82.18
rate (1500–2000 m)
(past)
Affection for Index 6.70 2.51 1 11
neighborhood
greenery
Affection for world Index 7.36 2.20 1 11
greenery
Degree of interaction Index 5.41 2.82 1 11
with greenery in the
last 5 years
Degree of Index 6.80 2.75 1 11
remembered
interaction with
greenery until the
age of 12
Level of knowledge Index 6.94 2.56 1 11
regarding the
multiple functions of
forests
Quality of greenery Index 5.87 2.45 1 11
within a 5-min walk
from home
(Continued)
Variable Unit Mean Std. dev. Min. Max.
Quality of greenery Index 6.16 2.43 1 11

within a 15-min
walk from home
Happiness Index 7.21 2.05 1 11
Household income Yen 5873494 3533006 500000 2.10E+07
(2012)
Household income Yen 5912882 3578747 500000 2.10E+07
(2011)
Unemployment – 0.021 0.14 0 1
dummy
Age Year 46.61 13.15 16 85
Female dummy – 0.45 0.50 0 1
Below the high – 0.25 0.44 0 1
school dummy
College dummy – 0.15 0.36 0 1
University dummy – 0.46 0.50 0 1
Graduate dummy – 0.061 0.24 0 1
Marriage dummy – 0.71 0.45 0 1
Altruism Index 1.92 0.77 1 3
Time discount rate % 22.08 21.53 −10 100
Risk aversion Index 6.52 2.28 1 11
Competitiveness Index 2.76 0.98 1 5
Presence of someone person 2.19 0.72 1 3
to consult nearby
Health status Index 2.90 1.03 1 5
Number of major life Index 2.15 1.28 1 5
events
Population density person 12522.29 7765.43 0 85698.84
Number of retail number 7.87 2.70 3.5 17.5
stores (within a 2 km
radius of home)
Number of number 3.32 1.86 1.2 9.6
restaurants (within
a 2 km radius of
home)
Distance to the km 0.84 0.73 0.0048 8.71
nearest railway
station
nearest bus stop
nearest public
cultural facility
nearest hospital
Population density person 12522.29 7765.43 0 85698.84
4 Model
The LSA can be calculated using the following equation (Frey et al., 2010):
SWB = f ( x, y, θ ′ z ) (1)
Here, SWB is the subjective degrees of well-being obtained from the question-
naire responses, x represents the non-market goods to be assessed, y is income,
and θ'z represents those factors that affect other subjective degrees of well-being.
The MWTP in relation to the marginal variation of x can be obtained by differ-
entiating Equation (1) and defining dSWB = 0. This is expressed in Equation (2):
MWTP = −dy / dx = (δ f / δ x) / (δ f / δ y ), (2)
In this study, SWB corresponds to an indicator of individuals’ subjective degrees

of happiness, x corresponds to the green coverage rate (0–100 m, 100–300 m,
300–500 m, 500–1000 m, 1000–1500 m, and 1500–2000 m), y corresponds to
household income, and z corresponds to control variables.
The control variables are as follows: unemployment dummy (unemployed = 1,
other = 0), age (years), female dummy (female = 1, man = 0), marriage dummy
(married = 1, other = 0), state of health (on a scale of 1 to 5, with 5 representing
those with the least of concern about their health), dummy variables for final aca-
demic degree (high school, college, university, and master’s or PhD; the reference
is high school dummy), personality index (time discount, risk aversion, competi-
tiveness, and altruism), social capital (whether there is someone to consult nearby
on a scale of 1 to 3, with 3 representing those having the most access), the number
of retail stores, the number of restaurants, distance to the nearest public cultural
facility, distance to the nearest hospital, distance to the nearest railway station,
distance to the nearest bus stop, population density, and prefecture dummy. To
control other factors we cannot consider, we include the index representing the
number of major life events experienced. Finally, to take account of quality of
green space and diversity of respondents’ preferences for green spaces, we include
cross terms of these.11
We show the correlation between green coverage rates and the variables we
use as the cross term. The result, shown in Table 8.2, implies small correlations
between green coverage rates and the frequency with which respondents visit
green spaces, although the signs of the correlations are positive. This implies
that there is a possibility that with no relation to green coverage rates, we tend
to visit green spaces at our own initiative. Moreover, we find small correlations
between green coverage rates and the other variables. Meanwhile, to avoid multi-
collinearity, we separately include green coverage rates by distance in the regres-
sion models.
The two-stage least-squares method is used in this study to account for reverse
causality. We include instrumental variables (IV) of income and green spaces,
namely, the previous year’s income and past green spaces. In terms of income,
Table 8.2 Correlation of each variables concerning green spaces
Green Green Green Green Green Green Affection Affec- Degree of Degree of Level of Qual- Quality of
coverage coverage coverage coverage coverage coverage for neigh- tion for interac- remem- knowl- ity of greenery
rate (0– rate rate rate rate rate borhood world tion with bered edge greenery within a
100 m) (100– (300– (500– (1000– (1500– greenery greenery greenery interac- regard- within a 15-minute
300 m) 500 m) 1000 m) 1500 m) 2000 m) in the last tion with ing the 5-minute walk from
5 years greenery multiple walk home
until the functions from
age of 12 of forests home
Green coverage 1
rate (0–100 m)
Green coverage 0.6079 1
rate (100–300 m)
Green coverage 0.3973 0.7453 1
rate (300–500 m)
Green coverage 0.2969 0.5097 0.6606 1
rate (500–1000 m)
Green coverage 0.2522 0.4175 0.5113 0.7192 1
rate (1000–1500 m)
Green coverage 0.2361 0.3862 0.4781 0.6163 0.7795 1
rate (1500–2000 m)
Affection for 0.1186 0.1787 0.1975 0.2227 0.2132 0.1954 1
neighborhood
greenery
Affection for world 0.0696 0.0791 0.0860 0.1175 0.0912 0.0741 0.5673 1
greenery
Degree of 0.0404 0.0904 0.0693 0.0728 0.0647 0.0597 0.3073 0.3319 1
interaction with
greenery in the last
5 years
Degree of −0.0246 −0.0281 −0.0163 0.0023 0.0063 0.0049 0.2195 0.3107 0.4238 1
remembered
interaction with
greenery until the
age of 12
Level of knowledge 0.0093 0.0460 0.0360 0.0729 0.0614 0.0612 0.2366 0.3467 0.2507 0.2548 1
regarding the
multiple functions
of forests
Quality of greenery 0.1348 0.2200 0.2302 0.2305 0.2332 0.2106 0.4933 0.1679 0.2235 0.0856 0.0138 1
within a 5-minute
walk from home
Quality of greenery 0.1189 0.2018 0.2213 0.2484 0.2491 0.2210 0.5679 0.2322 0.2480 0.1412 0.0717 0.8354 1
within a 15-minute
walk from home
there is a possibility that people with a higher level of well-being will have higher
incomes (Graham, 2011). Further, depending on their level of well-being, people
may desire to live in an area surrounded by green spaces. The previous year’s
income is derived from the questionnaire responses and the past green coverage
rate is obtained from the 2000 and 2001 editions of the Digital Map 5000 for the
capital and Kansai regions, respectively.
5 Estimation results
Table 8.3 shows the estimation results of the model including the first- and sec-
ond-order green coverage rates. We obtain statistically significant coefficients
for first- and second-order green coverage rates only with regard to green within
100–300 m and 300–500 m.12 With obtained coefficients for these green coverage
ratios, we calculated MWTP for a 1% increase in these green coverage ratios by
existing green spaces. Figure 8.1 shows the estimated MWTP, which implies the
effects of green spaces on life satisfaction decline in proportion to an increase
in the existing green coverage rate. This result is consistent with the findings of
Kopmann and Rehdanz (2013), which indicate that MWTP tends to be higher
for natural areas that are scarcer. In our dataset, average green coverage rate for
100–300 m is 14.73% and for 300–500 m is 18.25%; thus, the average MWTPs
for green coverage within 100–300 m and within 300–500 m are JPY 93,714 and
JPY 160,065, respectively. Moreover, the Figure 8.1 shows the slope of the green
within 100–300 m is steeper than that within 300–500 m, and the signs of MWTP
for the green coverage rate within 100–300 m and within 300–500 m change
from positive to negative at 23.8% and 59.6%, respectively. These percentages
imply a threshold for the green coverage rate people desire in their residential
surrounding.
Tables 8.4a–f show the estimation results of the models including cross terms
of green coverage rates.13 We obtain statistically significant coefficients for both
green spaces and cross terms. Figure 8.2 shows the relationship between MWTP
for green spaces and the levels of each variable we use as a cross term by dis-
tance from the respondents’ houses, calculated using the obtained coefficients in
Tables 8.4a–f. MWTP varies according to the levels of the horizontal axis vari-
ables. These results show that stronger the levels of affection with green spaces,
higher the degree of interaction, more extensive knowledge of their multiple func-
tions, and the higher the quality of green space, the higher their MWTP. Thus,
we find that the diversity of preference does give rise to conflict. The negative
value of MWTP for green spaces implies that those people do not need more
green spaces and, if land use change occurs, then they prefer other kinds of use
other than green spaces. We also find that conflicts are relatively large with regard
to affection for neighborhood green spaces and affection for worldwide green
spaces. Moreover, we find that the absolute values of MWTPs tend to be larger
the closer the respondents’ houses are to green spaces. However, it is notable
that the absolute values of the MWTPs for the green spaces within 0–100 m with
regard to affection for neighborhood greenery, knowledge regarding the multiple
Table 8.3 Estimation results (not including cross terms)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m

Green coverage rate −0.00379 0.00171* 0.0161* 0.00672 −0.000556 0.0113
(0.50) (1.87) (1.74) (0.68) (−0.05) (1.06)
(Green coverage rate)2 −0.000136 −0.000360** −0.000270** −0.000114 −0.0000228** −0.000182
(−1.10) (−2.23) (−1.85) (−0.79) (−0.16) (−1.26)
Household income 7.04e-08*** 6.93e-08*** 6.98e-08*** 7.04e-08*** 7.07e-08*** 7.10e-08***
(6.14) (6.04) (6.08) (6.13) (6.15) (6.18)
Unemployment −1.041*** −1.050*** −1.043*** −1.048*** −1.048*** −1.036***
(−3.98) (−4.02) (−3.99) (−4.01) (−4.01) (−3.96)
Age −0.104*** −0.103*** −0.102*** −0.103*** −0.103*** −0.102***
(−5.78) (−5.77) (−5.69) (−5.73) (−5.75) (−5.65)
(Age)2 0.00111*** 0.00111*** 0.00110*** 0.00111*** 0.00111*** 0.00110***
(6.12) (6.11) (6.03) (6.07) (6.09) (6.00)
Female dummy 0.446*** 0.444*** 0.449*** 0.455*** 0.451*** 0.446***
(5.19) (5.18) (5.23) (5.28) (5.24) (5.19)
Marriage dummy 1.122*** 1.120*** 1.118*** 1.124*** 1.124*** 1.128***
(12.18) (12.16) (12.14) (12.19) (12.18) (12.21)
College dummy 0.0455 0.0523 0.0465 0.0417 0.0372 0.0507
(0.39) (0.44) (0.39) (0.35) (0.32) (0.43)
University dummy 0.157* 0.159* 0.159* 0.159* 0.153* 0.162*
(1.77) (1.80) (1.80) (1.79) (1.73) (1.82)
Graduate dummy 0.390** 0.387** 0.393** 0.396** 0.386** 0.414**
(2.34) (2.33) (2.36) (2.37) (2.31) (2.47)
Health status 0.289*** 0.293*** 0.296*** 0.294*** 0.292*** 0.293***
(7.54) (7.68) (7.73) (7.67) (7.63) (7.65)
Time discount rate −0.0000214 0.0000963 0.000904 0.0000733 −0.0000686 −0.0000686
(−0.01) (0.06) (0.05) (0.04) (−0.04) (−0.04)
(Continued)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m
Risk aversion −0.0491*** −0.0511*** −0.0506*** −0.0503*** −0.0499*** −0.0488***

(−2.97) (−3.10) (−3.07) (−3.05) (−3.02) (−2.95)
Competitiveness −0.308*** −0.308*** −0.305*** −0.309*** −0.309*** −0.306***
(−7.80) (−7.81) (−7.72) (−7.81) (−7.80) (−7.72)
Altruism 0.125** 0.130*** 0.130*** 0.126*** 0.126** 0.124***
(2.57) (2.68) (2.67) (2.59) (2.59) (2.55)
Number of major life events −0.198*** −0.199*** −0.199*** −0.197*** −0.196*** −0.196***
(−6.59) (−6.62) (−6.61) (−6.55) (−6.51) (−6.53)
Presence of someone to consult 0.625*** 0.622*** 0.614*** 0.620*** 0.623*** 0.621***
nearby (11.61) (11.58) (11.36) (11.51) (11.57) (11.53)
Number of retail stores 0.0135 0.0168 0.0134 0.0154 0.0178 0.0144
(0.27) (0.33) (0.26) (0.30) (0.35) (0.28)
Number of restaurants 0.0174 0.0207 0.0234 0.0166 0.00550 0.0182
(0.23) (0.27) (0.30) (0.21) (0.07) (0.23)
Distance to the nearest railway −0.00165 0.00315 −0.00226 0.00672 0.0126 0.0120
station (−0.03) (0.05) (−0.04) (0.11) (0.21) (0.20)
Distance to the nearest bus stop 0.0422 0.0605 0.0431 0.0325 0.0222 0.0330
(0.20) (0.29) (0.21) (0.16) (0.11) (0.16)
Distance to the nearest public −0.151 −0.148 −0.236 −0.212 −0.208 −0.218
cultural facility (−0.49) (−0.47) (−0.77) (−0.69) (−0.68) (−0.71)
Distance to the nearest hospital 0.306 0.366 0.319 0.276 0.256 0.256
(1.15) (1.33) (1.17) (1.05) (0.99) (1.00)
Population density 5.74e-06 7.31e-06 7.06e-06 6.49e-06 5.58e-06 5.68e-06
(1.06) (1.34) (1.30) (1.19) (1.03) (1.04)
Observation 2158 2158 2158 2158 2158 2158
Notes: Values in parentheses are z-values.
*, ** and *** indicate significance at the 10% level, the 5% level and the 1% level, respectively. Prefecture dummies are included. Instrumented variables are Income
and Greenspace. Instrumental variables are past values of Income and Greenspace.
300000
200000
100000
0
Marginal WTP (YEN)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
-100000
-200000
-300000
-400000
-500000
-600000
-700000
-800000
Green coverage rate (%)
100m to 300m 300m to 500m
Figure 8.1 Marginal willingness to pay for a 1% increase in green spaces

Note: Average green coverage rate for 100–300 m is 14.73% and that for 300–500 m is 18.25%.
functions of greenery, and quality of greenery are smaller than those for the green
spaces within 100–300 m. Combined with the fact that we do not obtain sta-
tistically significant coefficients for first- and second-order coefficients of green
spaces within 0–100 m while we obtain statistically significant coefficients for
that within 100–300 m and 300–500 m, relatively small MWTPs for the green
spaces within 0–100 m implies that there is a possibility that there is a suitable
distance of green spaces from the houses.
With regard to the control variables for unemployment, age, gender, marital
status, academic dummies, and health status, the expected signs are found to be
statistically significant. As for the personality indicators, the only indicator that is
not statistically significant is time discount; however, for all other indicators the
expected sign was significant. By contrast, statistically significant results for the
convenience indices are not obtained.
6 Conclusion and discussion

This study used the LSA to evaluate the values of green spaces in Japan. Our main
contribution to the literature is as follows. First, the estimation results demonstrate
Table 8.4a Estimation results (including cross term: Affection for neighborhood greenery)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m

Green coverage rate −0.0381*** −0.0422*** −0.0382*** −0.0359*** −0.0329*** −0.0303***
(−4.63) (−5.74) (−5.85) (−6.24) (−6.37) (−6.21)
Green coverage rate*Affection 0.00457*** 0.00526*** 0.00487*** 0.00436*** 0.00390*** 0.00370***
for neighborhood greenery (4.63) (6.46) (6.83) (7.17) (7.15) (7.06)
(6.17) (5.94) (5.94) (5.95) (6.06) (6.03)
Observation 2158 2158 2158 2158 2158 2158
Note: Due to space limitations, we show preliminary independent variables. Concerning the parameters of other explanatory variables not included in this table, we
obtain almost the same results as Table 8.3. Values in parentheses are z- values.
*, **, and *** indicate significance at the 10% level, the 5% level, and the 1% level, respectively. Prefecture dummies are included. Instrumented variables are Income
and Greenspace. Instrumental variables are past values of Income and Greenspace.
Table 8.4b Estimation results (including cross term: Affection for world greenery)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m

(−3.59) (−3.87) (−3.56) (−3.07) (−3.52) (−3.42)
Green coverage rate*Affection 0.00400*** 0.00378*** 0.00318*** 0.00230*** 0.00225*** 0.00213***
for world greenery (3.45) (4.10) (4.06) (3.39) (3.63) (3.57)
(6.14) (6.01) (6.04) (6.08) (6.08) (6.09)
Observation 2158 2158 2158 2158 2158 2158
Note: See notes in Table 8.4a.
Table 8.4c Estimation results (including cross term: Interaction with greenery in the last 5 years)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m

(−3.50) (−3.25) (−2.86) (−2.84) (−3.08) (−2.73)
Green coverage rate*Interaction 0.00291*** 0.00268*** 0.00230*** 0.00190*** 0.00170*** 0.00150***
with greenery in the last 5 years (3.42) (3.96) (3.98) (3.76) (3.73) (3.40)
(6.20) (6.20) (6.21) (6.16) (6.11) (6.10)
Observation 2158 2158 2158 2158 2158 2158
Table 8.4d Estimation results (including cross term: Remembered interaction with greenery until the age of 12)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m

Green coverage rate −0.0213*** −0.0170*** −0.0118** −0.0106** −0.00929** −0.00796**
(−3.45) (−2.97) (−2.34) (−2.36) (−2.25) (−1.98)
Green coverage rate 0.00270*** 0.00232*** 0.00181*** 0.00146*** 0.00108** 0.000919**
(3.33) (3.37) (3.14) (2.93) (2.36) (2.11)
(6.18) (6.14) (6.17) (6.15) (6.14) (6.16)
Observation 2158 2158 2158 2158 2158 2158
*Remembered interaction with greenery until the age of 12
Table 8.4e Estimation results (including cross term: Knowledge regarding the multiple functions of forests)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m
Green coverage rate −0.0140** −0.0192*** −0.0176*** −0.0161*** −0.0161*** −0.0132***

(−2.07) (−2.75) (−2.98) (−3.01) (−3.43) (−2.93)
Green coverage rate 0.00151* 0.00237*** 0.00244*** 0.00204*** 0.00190*** 0.00158***
(1.75) (2.96) (3.61) (3.44) (3.61) (3.15)
(6.08) (5.97) (5.98) (5.98) (5.94) (5.97)
Observation 2158 2158 2158 2158 2158 2158
* Knowledge regarding the multiple functions of forests
Table 8.4f Estimation results (including cross term: Quality of greenery)
0–100 m 100–300 m 300–500 m 500–1000 m 1000–1500 m 1500–2000 m
Green coverage rate −0.0163** −0.0210*** −0.0154** −0.0188*** −0.0191*** −0.0183***

(−2.17) (−3.01) (−2.59) (−3.49) (−3.81) (−3.85)
Green coverage rate 0.00186* 0.00269*** 0.00220*** 0.00241*** 0.00227*** 0.00223***
(1.92) (3.32) (3.21) (4.09) (4.22) (4.26)
(6.10) (5.92) (6.02) (6.05) (6.04) (6.03)
Observation 2158 2158 2158 2158 2158 2158
Note: See notes in Table 8.4a. With regard to 0–100 m, 100–300 m, and 500–100 m, Quality of greenery corresponds to that within a 5-minute walk from home, whereas
with regard to 1000–1500 m and 1500–2000 m, Quality of greenery corresponds to that within a 15-minute walk from home.
* Quality of greenery
300000 300000
200000 200000
100000 100000
MWTP FOR GREENERY (YEN)

0 0
1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11
-100000 -100000
-200000 -200000
-300000 -300000
-400000 -400000
-500000 -500000
-600000 -600000
AFFECTION FOR NEIGHBORHOOD GREENERY AFFECTION FOR WORLD GREENERY
0m to 100m 100m to 300m 300m to 500m 500n to 1000m 1000m to 1500m 1500m to 2000m 0m to 100m 100m to 300m 300m to 500m 500n to 1000m 1000m to 1500m 1500m to 2000m
300000 300000
200000 200000
100000 100000
0 0
1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11
-100000 -100000
-200000
-200000
-300000
-300000
-400000
-400000
-500000
-500000
-600000
INTERACTION WITH GREENERY IN THE LAST 5 YEARS -600000
REMEMBERED INTERACTION WITH GREENERY UNTIL
THE AGE OF 12
300000 300000
200000 200000
100000 100000
0
0
1 2 3 4 5 6 7 8 9 10 11
-100000 1 2 3 4 5 6 7 8 9 10 11
-100000
-200000
-200000
-300000
-300000
-400000
-400000
-500000
-500000
-600000
KNOWLEDGE REGARDING THE MULTIPLE FUNCTIONS OF FORESTS
-600000
QUALITY OF GREENERY
Figure 8.2 Estimated marginal willingness to pay for green spaces
to what degree people’s MWTPs for green spaces decrease as the current amount
of green space increases. Second, our results show how people’s MWTPs increase
in proportion to their affection for it, the amount of interaction they have with it,
their knowledge of its multiple functions, and the quality of the green spaces with
which they interact. Finally, we show various MWTPs for green spaces in terms
of distance from respondents’ houses.
In our regression model, including first- and second-order coefficients for green
coverage rates, we did not obtain statistically significant first- and second-order
coefficients for green coverage rates with regard to green within 0–100 m and
over 500 m, whereas we obtained statistically significant results with regard to
green within 100–300 m and 300–500 m. As shown in Figure 8.1, we also find the
slope of MWTP within 100–300 m was steeper than that within 300–500 m, and
the signs of MWTPs for the green coverage rates within 100–300 m and within
300–500 m change from positive to negative at 23.8% and 59.6%, respectively. In
addition, considering we obtain a statistically insignificant result for green cover-
age within 100 m, there is a possibility people prefer other land use from green
(e.g., residents’ area). This trend can be observed also in Figure 8.2, which implies
a relatively small absolute value of MWTP with regard to green within 100 m.
The next step to evaluate the value of green coverage using LSA is to include
more detailed green data. This study’s dataset and data in prior literature are lim-
ited in capturing greenery details, such as trees, grass, and gardens, particularly
nearer to residences. Greenery in gardens may also affect people’s life satisfac-
tion. Applying the image from the high-resolution satellite (e.g., GEOEYE-1)14
one way to obtain a detailed study of greenery is the identification of trees and
grasses using the GIS software.
We also need to consider the differences among the effects of various kinds of
green. This study defines “green spaces” within the categories of forests, waste-
lands among others, rice fields, fields and other farmland, and parks and other
green spaces. The next step is achieved when the green data are separated in terms
of their functions, such as green in the park, green in the residence (i.e., garden),
green in road, green in the river, green in the factory, and so forth.
The other next step in the literature on LSA is to evaluate the value of green
thought to be accomplished by reconsidering the definition of life satisfaction.
There are various kinds of the prominent candidate indices of subjective well-
being, such as happiness, life satisfaction, positive affect, and mental health
(OECD, 2013; Powdthavee, 2011). There is a possibility that the effects of green
spaces on human beings differ by adopted subjective well-being indices.
Appendix
Table 8A.1 Definition of “green spaces” adopted for this study
Forests, Refers to forestland, bamboo forest, shino bamboo areas,

wastelands, etc. sasa bamboo areas, native grasslands (including abandoned
farmlands), barren land, and golf courses (minimum area unit
is 400 m2, minimum short-side length is 20 m)
Rice fields Refers to paddy fields in which paddy rice, lotus, and
arrowheads are cultivated (includes short-term fallow fields);
this includes seasonal field crops (minimum area unit is 400
m2, minimum short-side length is 20 m)
Fields and other Refers to normal fields, orchards, mulberry plantations,
agricultural land tea gardens, other kinds of tree gardens, seedling fields,
farms, pasturelands, lumbering pasturelands, and barns and
greenhouses (minimum area unit is 400 m2, minimum short-
side length is 20 m)
Parks, green Refers to parks, zoos, and botanical gardens, cemeteries,
spaces, etc. temple, and shrine grounds, amusement parks and other
facilities that have the characteristics of public places,
comprehensive sports grounds, stadiums, and facilities for
athletic competitions, such as baseball fields (minimum area
unit is 200 m2, minimum short-side length is 20 m)
Table 8A.2 Main questionnaire content
Name of variable Content of question/parameters Value

(1) Subjective degree On the whole, how happy do you 0 (extremely unhappy) to
of happiness normally feel? 10 (extremely happy)
(2) Household Approximately how much will Unit: Japanese yen
income your total household income,
including bonuses, for the 2011
fiscal year be?
(3) Unemployment Are you currently looking for a 1: unemployed (searching
job? for work), 0: not searching
(4) Age How old are you? (numeric entry, years)
(5) Sex What is your gender? 1: male, 0: female
(6) Marital status Are you married? 1: Married, 0: other
(7) Academic What was the last level of school high school, college,
dummy completed? university, and master’s or
PhD; the reference is high
school dummy
(8) Health condition Do you feel any anxiety with 1 (feel a lot) to 5 (don’t
regard to your health? feel any)
(9) Time discount If you could receive JPY 10,000 ●: 9500, 10000, 10200,
at the end of one month, or 10400, 10600, 11000,
JPY _________ at the end of 12000, 14000
13 months, then which would you ※ Interest rate used in the
choose? analysis
(10) Risk aversion Which of the following proverbs 0 (completely empathize
more closely describes your with “nothing ventured”)
behavior? “Nothing ventured, to 10 (completely
nothing gained” or “a wise man empathize with “a wise
keeps away from danger.” man”)
(11) Competitiveness Are you conscious of the standard 1 (not conscious of others)
of living of others? to 5 (extremely conscious)
(12) Altruism If you pay JPY 1000, then JPY 1: Would give to neither
99,000 will be supplemented and 2: Would give to one or
JPY 100,000 will be given to an the other
unknown needy person. If you pay 3: Would give to both
JPY 1000, then JPY 99,000 will
be supplemented and JPY 100,000
will be given to a needy person
you are familiar with.
(13) Number of In the past 5 years, how many 1 (none) to 5 times (or
shocking times have you experienced a more)
experiences shocking incident that left a deep
emotional scar?
(14) Presence of Is there anyone near you with 1: none, 2: there is
someone to whom you can consult? someone nearby with
consult nearby whom light consultation
is possible, 3: there is
someone nearby with
whom serious consultation
is possible
Name of variable Content of question/parameters Value
(15) Green coverage Created using GIS %

rate
(16) Quality of Are you satisfied with the forests 0 (not satisfied at all) to 10
greenery within and greenery within a 5-minute (extremely satisfied)
a 5-minute walk walk from your home?
from home
(17) Level of affection Do you feel any affection for 0 (feel no attachment
for neighborhood forests and greenery near your at all) to 10 (feel very
greenery home? attached)
(18) Level of affection Do you feel any affection for 0 (feel no attachment
for global forests or other greenery itself? at all) to 10 (feel very
greenery attached)
(19) Degree of How frequently have you 0 (never went) to 10 (go
interaction with interacted with forests and frequently)
greenery in the greenery in the last 5 years for any
last 5 years of the following (rest, refreshment,
forest therapy, excursions, sports
such as mountain climbing,
fishing, etc.)? Interactions can be
in your neighborhood or far away.
(20) Degree of How frequently did you interact 0 (never went) to 10 (went
remembered with forests and greenery from frequently)
interaction with birth until you were 12 years old
greenery until the for any of the following (rest,
age of 12 years refreshment, forest therapy,
excursions, sports such as
mountain climbing, fishing,
etc.)? Interactions can be in your
neighborhood or far away.
(21) Amount of Refer to main text 0 (have never heard about
knowledge the functions) to 10 (knew
regarding the of the functions quite
multiple functions well)
of forests
(22) Number of retail Created using GIS Number of retail stores
stores within a 2 km radius of
home
(23) Number of Created using GIS Number of restaurants
restaurants within a 2 km radius of
home
(24) Distance to public Created using GIS Distance to the nearest
cultural facilities public cultural facility
(25) Distance to Created using GIS Distance to the nearest
railway station railway station
(26) Distance to bus Created using GIS Distance to the nearest
stop bus stop
(27) Population Created using GIS Population density of
density residence
Table 8A.3 Forests’ multiple functions
(1) Biodiversity conservation Gene conservation

(1) Species conservation
(1) Ecosystem integrity
(2) Global environmental protection Mitigation of global warming
(1) Stabilization of the global climate system
(3) Sediment disaster prevention/soil Surface erosion prevention
conservation
(1) Shallow landslide prevention
(1) Other sediment disaster prevention
(1) Sediment discharge prevention
(1) Soil conservation
(1) (Productivity of forest maintenance)
Other natural disaster prevention
(4) Watershed Flood mitigation
(1) Water resources reservoir
(1) Water regulation
(1) Water purification
(5) Creation of a pleasant environment Climate mitigation
(1) Air purification
(1) Creation of a pleasant living environment
(6) Health and recreation Medical treatment
(1) Recuperation
(1) Recreation
(7) Culture Landscape, scenic beauty
(1) Learning and education
(1) Art
(1) Religion, festivals
(1) Traditional culture
(1) Maintenance of diversity within a region
(culture formation)
(8) Material production Timber
Food
Fertilizer
Feed
Industrial raw material for pharmaceuticals
Planting materials
Ornamental plants
Craft materials
Source: Created based on the Forestry Agency Website: www.rinya.maff.go.jp/j/keikaku/tamenteki/
con_1.html (in Japanese).
Table 8A.4 Multiple functions of agriculture
(1) Peace of mind for the future because of sustainable food supplies for the public
(2) Contribution to the Formation of the material circulation system because of
environment because agriculture
of agricultural land
use that complements (1) Contribution to the local community by controlling
biogeochemical systems the water cycle by prevention of floods, landslides,
and soil erosion (outflow); stabilization of the river
flow regime; and groundwater recharge
(2) Removal and mitigation of environmental hazards
by water purification; organic waste decomposition;
atmospheric regulation (air purification, system
relief, etc.), and prevention of excessive
accumulation and exploitation of resources
Secondary (artificial) creation and maintenance of
natural environments
(1) Protection of biodiversity as new ecosystems
(biological ecosystem conservation; genetic
resources conservation; wildlife preservation)
(2) Conservation of land space (dynamic conservation
of prime farmland; provision of green space;
preservation of Japan’s original landscapes;
formation of artificial natural landscapes)
(3) Integrity of production Formation and maintenance of the local community and
and the living culture
environment; formation
and maintenance of the (1) Promotion of the local community
local community (2) Preservation of traditional culture
Mitigation of urban tension
(1) Recovery of humanity
(2) Hands-on learning and education
Source: Compiled by the author from Science Council of Japan (2001) documents.
Notes
1 There has been extensive research on air pollution, such as the works presented by
Welsch (2002, 2006), MacKerron and Mourato (2009), Luechinger (2009, 2010), Fer-
reira and Moro (2010), Menz (2011), and Tsurumi et al. (2013). In addition, there has
also been considerable research on airport noise (van Praag and Baarsma, 2005), cli-
mate (Ferreira and Moro, 2010; Maddison and Rehdanz, 2011), floods (Luechinger and
Raschky, 2009), and droughts (Carroll et al., 2009).
2 Our greenspace data’s smallest diameter is 15 m for “Parks, green spaces, etc.” and
20 m for “Forests, wastelands, etc.,” “Rice fields,” and “Fields and other agricultural
land.”
3 The survey was conducted as follows. First, an invitation was emailed from the internet
survey company to relevant cohorts from their one million plus registered monitors in
Japan. To obtain a random sample, e-mails were sent to selected monitors based on
prefectural demographics, such as population, gender, and age, and basic attributes,
such as income level, and education.) To avoid respondent bias, the research objectives
(i.e., identifying the value of green space) were not disclosed to the research monitors
before they decided to answer the questionnaire. Sixty-five percent of the individuals
who were contacted completed the survey.
4 To avoid invalid responses, we used several similar questions to check for response
consistency and exclude the inconsistent responses.
5 There are 2000 and 2005 editions of the metropolitan area (Kanto region) and 2003 and
2008 editions of the Kansai region, but only a 2003 edition of the Chubu region.
6 This study uses a radius of between 100 m and 2000 m based on the assumption of
everyday living space accessed by foot.
7 Ideally, it should be indexed in the objective data from the GIS among others as to
whether the forests and other areas are in a deteriorated condition or whether the green-
ery is being maintained. However, this study is limited to using subjective indicators
because of the issue of data availability.
8 In a robustness check, when we use different radii with respect to convenience indica-
tors, such as 500 m and 1000 m, we obtain almost the same estimated coefficients as
for green spaces.
9 Balducci and Checchi (2009) consider the following question: “How often do you vol-
unteer for environmental problems?” We also introduced this index into our regression
models and obtained statistically significant positive parameters. In addition, including
this index caused no significant difference in the parameters of green coverage rates.
The reason it is not included in the model in the current chapter is its high correlation
with altruism.
10 See MacKerron (2012) for a detailed review of happiness studies.
11 There is a possibility that the five indices concerning diversity of preference – affec-
tion for green spaces (neighborhood), affection for green spaces (worldwide), degree
of interaction with green spaces (in the past 5 years), degree of interaction with green
spaces (until the age of 12 years), and knowledge of green spaces – have high correla-
tions with income. However, the correlation coefficients with income are small (only
0.0749, 0.0787, 0.0852, 0.0459, and 0.1256, respectively); therefore, the problem of
multicollinearity is considered minimal.
12 We obtain statistically insignificant results for green spaces when we include only first-
order coefficients for green coverage ratios for all distances.
13 We do not obtain statistically significant results for the model including first- and sec-
ond-order coefficients for green coverage rates and these cross terms, so we adopt the
model including only first-order green coverage rates and their cross terms.
14 This satellite collects 41 cm panchromatic and 1.65 m four-band multispectral (i.e.,
blue, green, red, and near-infrared) imagery.
Bibliography
Ambrey, C., Fleming, C., 2011. Valuing scenic amenity using life satisfaction data. Eco-
logical Economics 72, 106–115.
Ambrey, C., Fleming, C., 2013. Public greenspace and life satisfaction in urban Australia.
Urban Studies, 1–32.
Balducci, A., Checchi, D., 2009. Happiness and quality of city life: The case of Milan, the
richest Italian city. International Planning Studies 14(1), 25–64.
Barrio, M., Loureiro, M. L., 2010. A meta-analysis of contingent valuation forest studies.
Ecological Economics 69(5), 1023–1030.
Bayer, P., Keohane, N. O., Timmins, C., 2009. Migration and hedonic valuation the case of
air quality. Journal of Environmental Economics and Management 58, 1–14.
Blanchflower, D., Oswald, A., 2004. Well-being over time in Britain and the USA. Journal
of Public Economics 88, 1359–1386.
Brander, L. M., Koetse, M. J., 2011. The value of urban open space: Meta-analyses of
contingent valuation and hedonic pricing results. Journal of Environmental Management
92, 2763–2773.
Carroll, N., Frijters, P., Shields, M., 2009. Quantifying the costs of drought new evidence
from life satisfaction data. Journal of Population Economics, 22445–22461.
Ferreira, S., Moro, M., 2010. On the use of subjective well-being data for environmental
valuation. Environmental and Resource Economics 46, 249–273.
Frey, B. S., Luechinger, S., Stutzer, A., 2010. The life satisfaction approach to environmen-
tal valuation. Annual Review of Resource Economics 2, 139–160.
Graham, C., 2011. The Pursuit of Happiness Toward an Economy of Well-Being. The
Brookings Institution Press.
Hartig, T., Evans, G. W., Jamner, L. D., Davis, D. S., Garling, T., 2003. Tracking restoration
in natural and urban field settings. Journal of Environmental Psychology 23, 109–123.
Johnston, R. J., 2006. Is hypothetical bias universal? Validating contingent valuation
responses using a binding public referendum. Journal of Environmental Economics and
Management 52(1), 469–481.
Kopmann, A., Rehdanz, K., 2013. A human well-being approach for assessing the value of
natural land areas. Ecological Economics 93, 20–33.
Larson, S., De Freitas, D. M., Hicks, C. C., 2013. Sense of place as a determinant of peo-
ple’s attitudes towards the environment Implications for natural resources management
and planning in the Great Barrier Reef, Australia. Journal of Environmental Manage-
ment 117, 226–234.
Lee, J., Park, B.J., Tsunetsugu, Y., Ohira, T., Kagawa, T., Miyazaki, Y., 2011. Effect of for-
est bathing on physiological and psychological responses in young Japanese male sub-
jects? Public Health 125(2), 93–100.
Luechinger, S., 2009. Valuing air quality using the life satisfaction approach. The Economic
Journal 119, 482–515.
Luechinger, S., 2010. Life satisfaction and transboundary air pollution. Economics Letters
107(1), 4–6.
Luechinger, S., Raschky, P., 2009. Valuing flood disasters using the life satisfaction
approach. Journal of Public Economics 93, 620–633.
MacKerron, G., 2012. Happiness economics from 35 000 FEET. Journal of Economic
Surveys 26, 705–735.
MacKerron, G., Mourato, S., 2009. Life satisfaction and air quality in London. Ecological
Economics 68, 1441–1453.
MacKerron, G., Mourato, S., 2013. Happiness is greater in natural environments. Global
Environmental Change 68, 1441–1453.
Maddison, D., Rehdanz, K., 2011. The impact of climate on life satisfaction. Ecological
Economics 70, 2437–2445.
Menz, T., 2011. Do people habituate to air pollution? Evidence from international life sat-
isfaction data. Ecological Economics 71, 211–219.
Morita, E., Fukuda, S., Nagano, J., Hamajima, N., Yamamoto, H., Iwai, Y., Nakashima, T.,
Ohira, H., Shirakawa, T., 2007. Psychological effects of forest environments on healthy
adults: Shinrin-yoku (forest-air bathing, walking) as a possible method of stress reduc-
tion. Public Health 121(1), 54–63.
Neilson, W., Wichmann, B., 2014. Social networks and non-market valuations. Journal of
Environmental Economics and Management 67(2), 155–170.
OECD, 2013. OECD guidelines on measuring subjective well-Being, OECD Publishing.
http://dx.doi.org/10.1787/9789264191655-en
Park, B.J., Tsunetsugu, Y., Kasetani, T., Kagawa, T., Miyazaki, Y., 2010. The physiological
effects of Shinrin-yoku (taking in the forest atmosphere or forest bathing): Evidence
from field experiments in 24 forests across Japan. Environmental Health and Preventive
Medicine 15(1), 18–26.
Peiro, A., 2006. Happiness, satisfaction and socio-economic condition some international
evidence. The Journal of Socio-Economics 35, 348–365.
Powdthavee, N., van den Berg, B., 2011. Putting different price tags on the same health
condition: Re-evaluating the well-being valuation approach. Journal of Health Econom-
ics, Elsevier 30(5), 1032–1043.
Sander, H., Haight, R. G., 2012. Estimating the economic value of cultural ecosystem ser-
vices in an urbanizing area using hednic pricing. Journal of Environmental Management
113, 194–205.
Sander, H., Polasky, S., Haight, R. G., 2010. The value of urban tree cover a hedonic prop-
erty price model in Ramsey and Dakota Countries, Minnesota, USA. Ecological Eco-
nomics 69, 1646–1656.
Saphoresa, J. D., Li, W., 2012. Estimating the value of urban green areas: A hedonic pricing
analysis of the single family housing market in Los Angeles, CA. Landscape and Urban
Planning 104, 373–387.
Science Council of Japan, 2001. Evaluating the multi-functions of agriculture and forests
related to the global environment and human beings. http//www.scj.go.jp/ja/info/kohyo/
pdf/shimon-18-1.pdf (in Japanese).
Smyth, R., Mishra, V., Qian, X., 2008. The environment and well-being in urban China.
Tella, R. D., MacCulloch, R. J., Oswald, A. J., 2001. Preferences over inflation and unem-
ployment evidence from surveys of happiness. The American Economic Review 91(1),
335–341.
Thompson, C. W., Roe, J., Aspinall, P., Mitchell, R., Clow, A., Miller, D., 2012. More green
space is linked to less stress in deprived communities: Evidence from salivary cortisol
patterns. Landscape and Urban Planning 105, 221–229.
Tsunetsugu, Y., Lee, Y., Park, B. J., Tyrväinen, L., Kagawa, T., Miyazaki, J., 2013. Physi-
ological and psychological effects of viewing urban forest landscapes assessed by mul-
tiple measurements. Landscape and Urban Planning 113, 90–93.
Tsurumi, T., Kuramashi, K., Managi, S., 2013. Determinants of happiness: Environmental
degradation and attachment to nature. S. Managi (Eds.), The Economics of Biodiversity
and Ecosystem Services, 62–86. New York: Routledge.
van Praag, B., Baarsma, B., 2005. Using happiness surveys to value intangibles the case of
airport noise. Economic Journal 115, 224–246.
Welsch, H., 2002. Preferences over prosperity and pollution environmental valuation based
on happiness surveys. Kyklos 55(4), 473–494.
Welsch, H., 2006. Environment and happiness valuation of air pollution using life satisfac-
tion data. Ecological Economics 58, 801–813.
9 Greenery and well-being
Assessing the monetary value of
greenery by type
Tetsuya Tsurumi, Atsushi Imauji
and Shunsuke Managi
1 Introduction
Along with economic development, there have been changes in land use, and this
is resulting in forests, green space, and agricultural land giving way to residential
and industrial/commercial areas in many cities.1 Due to this reduction in green
space, efforts have been made in urban planning to restore greenery around resi-
dential neighborhoods for many metropolitan areas.
The importance of efforts to restore greenery has also been substantiated in
research using subjective well-being. Studies showing the relationship between
question-based subjective well-being indices and the surrounding natural environ-
ment have recently emerged (e.g., Ambrey and Fleming, 2011, 2014; MacKerron
and Mourato, 2013; Tsurumi and Managi, 2015). These studies use life satisfac-
tion or subjective happiness as subjective well-being indices, and demonstrate a
positive relationship between subjective well-being and greenery.
In this study, we use the following well-being indices: the General Health
Questionnaire (GHQ-12) as a medical/health index, “affect balance” as a psy-
chological index, and life satisfaction, subjective happiness, and the Cantril lad-
der as subjective well-being indices to measure the marginal willingness to pay
(MWTP) for green spaces. The first purpose of the current study is to compare the
well-being indices in terms of their relationship with greenery, which we expect
to provide us with a greater understanding of the value of greenery. In addition,
we need to consider that there is a diversity of greenery in urban areas, and the
impact on people may vary by type. There are many types of green spaces in
urban areas, including those within residential areas, trees lining streets, greenery
in industrial/commercial areas, parks, and green spaces by rivers, etc., and each
of these represent categories of urban planning. The second purpose of this study
is to categorize urban greenery in a detailed fashion, verify it, and then evaluate
the detailed categories.
The remainder of the chapter is organized in the following way. Section 2
reviews prior research and the originality of this chapter, Section 3 covers data
and analytical methods and Section 4 contains the results of the analysis. The
discussion and conclusions are found in Section 5.
2 Literature
Studies showing the relationship between subjective well-being indices and the
surrounding natural environment include those by Ambrey and Fleming (2011,
2014), MacKerron and Mourato (2013), and Tsurumi and Managi (2015).
Ambrey and Fleming (2011) perform a regression analysis to explore the rela-
tionship between life satisfaction and the distance from people’s residences to
parks in Australia. They find a positive relationship between parks and life sat-
isfaction. Ambrey and Fleming (2014) use green coverage data obtained from
geographic information systems (GIS) as an objective measure of the amount
of green space. In urban areas of Australia, the authors find a positive relation-
ship between the percentage of public greenspace within a 750 m radius of
an individual’s residence and their self-reported life satisfaction. MacKerron
and Mourato (2013) find that, on average, subjective happiness is higher when
individuals are in locations surrounded by greenery compared to urban envi-
ronments. Tsurumi and Managi (2015) show that people’s MWTP for green
space decreases as the current amount of green space increases. Individuals’
MWTP increases in proportion with their affection for green spaces, the amount
of interaction they have with green spaces, their knowledge of green spaces’
multiple functions, and the quality of greenery with which they normally come
into contact; there are also various MWTP values for green spaces in terms of
distance from respondents’ houses.
We note that previous studies that investigate the relationship between subjec-
tive well-being and greenery have tended to rely on a general definition of green
spaces. In this study, we assume that the value of greenery varies based on green-
ery type, which enables us to provide evidence-based proposals for future land
use policy. The effects exerted on people by greenery in parks versus greenery in
farmland, for example, are assumed to differ. It is also possible that the values of
these types of greenery depend on the extent of contact that people have with each
type of greenery. Thus, in this study, we seek to produce monetary valuations of
various types of greenery using subjective well-being indices, which has not been
attempted in previous studies.
To achieve this, detailed data on greenery are required. For example, estimating
the value of roadside greenery requires street-level data on vegetation, while esti-
mating the value of greenery on private lands requires individual premise-level
data on vegetation. In short, tree-level data are required. We obtain satellite imag-
ery data that has become available in recent years. Specifically, we use QuickBird
satellite images with a pixel resolution of 61 cm. Images with this resolution allow
extraction of data at the tree level. Trees are present only in some sections of land.
Thus, this study compiles greenery data using high-resolution satellite imagery,
which includes items that cannot be extracted from conventional datasets. Previ-
ous studies have not utilized tree-level greenery data, and, to our knowledge, no
studies have estimated the monetary valuations of specific types of greenery using
various indices of well-being. Although Li et al. (2015) highlight the advantages
of using high-resolution land-cover data over moderate resolution NDVI data, our
Greenery and well-being 227
study uses both high-resolution land-cover data and high-resolution NDVI data.
The main purpose of this study is thus to reveal more detailed and reliable values
of greenery by using high-resolution satellite images. While previous studies use
relatively rough green data, such as land use maps, we use “tree-level” greenery
data. By extracting greenery from satellite images with tree-level accuracy, we
expect to obtain more reliable values of greenery.
3 Data and estimation method
3.1 Survey
We conducted an online survey between September 24, 2014 and October 6, 2014
targeting a sample of approximately 5057 individuals in the eight selected Tokyo
wards. We obtain the sample by “ward level” strata for gender and age distribu-
tions, which resulted in 3124 responses.2 We set several trap questions, which are
different questions that should obtain the same answers. To improve the qual-
ity of the data, we then excluded inconsistent responses from the sample. We
finally obtained 2758 responses.3 To minimize the respondents’ burden, we used
multiple-choice questions, and the average length of time taken to complete our
questionnaire was 21.2 minutes.
3.2 Data on subjective well-being

The concept of cost-benefit is considered central to local policy in relation to
taxes, so that monetary values of greenery are expected to be applied to policy,
particularly in budgeting for land use planning. Our estimation strategy is in line
with the “experienced preference method,” which is discussed by Welsch and Fer-
reira (2014). The “experienced preference method” corresponds to the method of
evaluating non-market values with several indices for well-being, which is related
to “experienced utility.” The concept of “experienced utility” is distinguished
from conventional utility or “decision utility” (see Kahneman et al. (1997) for the
detailed distinction between “experienced utility” and “decision utility.”) In this
study, we thus use the term “well-being” as a synonym for “utility.”4
The relationship between people’s well-being and greenery may differ depend-
ing on the well-being indices applied. Considering this difference, we include
several indices related to well-being in our survey questions. The specific ques-
tions are provided in Table 9.1.
According to the OECD Guidelines on Measuring Subjective Well-Being, two
main types of indices measure subjective well-being. These include those related
to “life evaluation,” in which the individual evaluates his or her own life experi-
ence or life as a whole (e.g., degree of life satisfaction, degree of subjective hap-
piness, and the Cantril ladder), and indices related to “affect” (affect balance),
which measure an individual’s feelings or psychological state at a certain point
in time or during a period (OECD, 2013).5 The latter is related to a psychological
Table 9.1 Survey questions for each measure of subjective well-being
Measure of subjective well-being Survey question

Cantril ladder Please imagine a ladder with steps numbered from
0 at the bottom to 10 at the top. The top of the
ladder represents the best possible life for you,
and the bottom of the ladder represents the worst
possible life for you. On which step of the ladder
would you say that you personally feel you stand
on at this time? Responses are given on an integer
scale from 0 to 10.
Life satisfaction How satisfied are you with your life as a whole?
Responses are given on an integer scale from 0
(not at all satisfied) to 10 (completely satisfied).
Subjective happiness In general, how happy are you? Responses are
given on an integer scale from 0 (very unhappy) to
10 (very happy).
Affect balance (previous day) Based on Gallup and Rojas and Veenhoven (2013)
Yesterday, how often did you feel the following
emotions?
Enjoyment, physical pain, worry, sadness, stress,
anger, depression, affection, and happiness.
Responses are given on the following scale:
frequently, sometimes, rarely, not at all.
The responses are converted to numerical scores
as follows: frequently = 4, sometimes = 3, rarely
= 2, and not at all = 1. The average score for
negative effects (physical pain, worry, sadness,
stress, anger, and depression) are subtracted from
the average score for positive affects (enjoyment,
affection, and happiness) to produce an overall
score between −3 and 3.
Mental health (GHQ-12) Based on Goldberg (1972) and Goldberg and
Hiller (1979)
The response to each of the 12 questions
concerning GHQ-12 is assigned a score based on
a 4-point Likert scale (from 0 to 3), and the total
score (between 0 and 36) is used as a measure
of the individual’s mental health. The larger the
score, the better is the respondent’s mental health.
evaluation index. In addition to life evaluation and affect, several studies have
incorporated mental health concepts, which are related to medical/health litera-
ture. Powdthavee and van den Berg (2011) include the GHQ-12 and degrees of
life satisfaction as measures of well-being. The GHQ-12 is widely used in the
medical literature as a measure of psychological stress and pain (Guthrie et al.,
1998) and measures well-being, or particular effects, at a point in time. Recent
studies that utilize the GHQ-12 include Clark and Oswald (2002), Pevalin and
Table 9.2 Correlations among indices related to well-being
Cantril Life Subjective Affect Mental health

ladder satisfaction happiness balance (GHQ-12)
Cantril ladder 1
Life satisfaction 0.7754 1
Subjective happiness 0.8129 0.8113 1
Affect balance 0.5816 0.5134 0.6449 1
Mental health (GHQ-12) 0.5326 0.457 0.5595 0.6402 1
Ermisch (2004), Robinson et al. (2004), Oswald and Powdthavee (2007), and
Powdthavee and Vignoles (2008).
The correlations among the well-being indices are shown in Table 9.2. The
table implies that each index captures a different aspect of well-being.
3.3 Data on greenery

In this study, green spaces are classified in accordance with the land classifica-
tions of the 2011 Tokyo Land Use Survey (Ward Areas). These include greenery
in residential areas, roadsides, commercial areas, industrial areas, farmland, for-
ests, and wilderness, waterfronts, parks, public facilities, and railroads. Table 9.3
shows the details of the land use classifications adopted in this study.
The greenery data used in this study are compiled using GIS software to over-
lay green coverage data extracted from satellite imagery with data from the Tokyo
Land Use Survey. The data are derived from eight metropolitan Tokyo wards
(Suginami, Setagaya, Meguro, Shinagawa, Minato, Nakano, Shibuya, and Shin-
juku). We focus on these wards, first, because we are concerned with policies
relating to greenery in urban areas, and second, because of the limitations on the
use of data from satellite imagery.
We use satellite images in summer, when greenery is at its richest. We adopt
images from August 12, 2013, a day on which the ground surface was unobscured
by clouds. In addition to visible light (red, green, and blue), the images include
a near-infrared channel that provides wavelength information. Using these four
channels, we can identify plants. Specifically, we used the Normalized Differ-
ence Vegetation Index (NDVI) to extract vegetation data. Because our satellite
image data are remarkably clear, we can precisely distinguish greenery from other
objects. Thus, we decided to set the cut-off point concerning NDVI by eyesight.
This is also because there is no established way of automatically setting the cut-
off point, particularly with high-resolution satellite images.
Map 9.1 shows green coverage based on the NDVI, prior to classification into
land types. Greenery classified by land type is shown in Map 9.2.6
In addition, using respondents’ addresses, we can classify greenery by distance
from the respondent’s residence – that is, greenery within a 100 m radius, within a
Table 9.3 Classification of greenery in this study
Residential Detached houses (single-family houses, buildings primarily for

areas residential purposes but that are attached to private schools, classrooms,
or clinics), apartment buildings (public housing, single-story and multi-
story apartment buildings, bachelor dormitories, boarding houses,
tenement housing, terrace houses, and townhouses)
Roadsides Streets, pedestrian walkways, bicycle paths, farm roads, forest roads,
and estate pathways
Commercial Office buildings (offices, commercial premises such as banks
areas and securities companies, newspaper headquarters, broadcasting
stations, NTT property, processing centers, and medical institutions),
commercial facilities (department stores, supermarkets, retail stores,
wholesale outlets, gas stations, food and beverage stores, public bathing
facilities, and saunas), combined residential-commercial facilities
(shops and offices attached to residential facilities, hairdressing and
beauty salons, commercial outlets with workshops and residential
facilities, lodging and entertainment facilities, sports and performance
facilities, theaters, entertainment halls, and cinemas)
Industrial Specialized factories and workshops, factories attached to residential
areas facilities, warehouses, and transportation facilities (parking structures
not attached to offices or residences, parking lots, bus terminals, truck
terminals with attached storage facilities or offices, warehouses, and
covered storehouses, distribution centers, and delivery centers)
Farmland Farmland (land cultivated for produce (e.g., rice, rushes, and lotus) that
requires fresh water and irrigation facilities), fields (land cultivated for
herbaceous produce (e.g., vegetables, cereals, flowers, and seedlings),
and orchards (land cultivated for fruit or woody plants such as tea and
mulberries)
Forests and Forests (timber forests, bamboo forests, creeping pine forests, bamboo
wilderness grassy areas, mountainous areas, and areas with wild, concentrated
populations of bamboo or trees) and plains (grasslands where grasses
and shrubs grow naturally, abandoned land, and bare land)
Waterfronts Water surfaces, rivers, and waterways (rivers, canals, lakes, marshes,
reservoirs, and seas)
Parks Green spaces in parks, outdoor exercise and sporting grounds,
amusement parks, fishing ponds, outdoor swimming pools, equestrian
centers, and cemeteries
Public Government facilities (government and public agencies and branch
facilities offices, police stations and terminals, fire stations, post offices, taxation
offices, courts, and diplomatic offices), educational and cultural
facilities (kindergartens, elementary schools, junior high schools,
high schools, universities, vocational schools, other professional
training schools and institutes, research institutes, driving schools,
art galleries, exhibition centers, libraries, public halls, temples, and
shrines, churches, and neighborhood halls), health and welfare facilities
(hospitals, clinics, health centers, daycare centers, nurseries, elder-care
facilities, disabled-care facilities, and child welfare facilities)
Railroads Railroads, railroad tracks, monorails, airports, ports, and harbors
Note: These classifications are based on the land classifications in the 2011 Tokyo Land Use Survey
(Ward Areas).
0 2.5 5 10 Kilometers
Map 9.1 Vegetation detected via satellite imagery and the NDVI
100 m to 500 m radius, within a 500 m to 1000 m radius, and within a 1000 m to
1500 m radius – and investigate the relationship between the individual’s subjec-
tive well-being and each classification.7
3.4 Estimation model

In this study, we use regression analysis to analyze the well-being function.
Specifically, we estimate equation (1) and analyze the relationships among
well-being, income, and environmental factors. We then calculate the mar-
ginal effects of income and environmental factors on well-being based on
equation (2). Finally, we calculate the substitution rate for the marginal effects
Map 9.2 Vegetation data post-classification
of income and environmental factors on well-being and arrive at the income-
equivalent value of environmental factors. This income-equivalent value rep-
resents people’s MWTP for those environmental factors. It is the amount of
income required to maintain constant well-being when people experience a
change in environmental factors.
WBi = βU i′ + γVi ′ + ηWi ′+ θ X i + λYi + µZ i′ + εi (1)
 ∂WBi   ∂WBi 
MWTPi = ∆X i ∆Yi =    
 ∂X  (2)
 ∂Yi  i
where WBi represents the five types of well-being (the Cantril ladder, life sat-
isfaction, subjective happiness, affect balance, and mental health based on the
GHQ-12) of individual i; U′i is a composite vector of the Big Five personality
traits (extraversion, agreeability, conscientiousness, openness to experience, and
neuroticism) of individual I; V′i is a composite vector of demographic informa-
tion about individual i including age, sex, and marital status; W′i is a composite
vector of Better Life Index (BLI) indicators8 for individual i; Xi is the annual
after-tax income of the individual’s household; Yi is the green coverage rate for
the individual’s neighborhood, including greenery in residential areas, roadsides,
commercial areas, industrial areas, farmland, forests, and wilderness, waterfronts,
railroads, and public facilities;9 Z′i is individual i’s satisfaction with accessibility
(convenience around the residence); and ɛi is the error term.10
Kopmann and Rehdanz (2013) indicate that marginal willingness to pay
(MWTP) tends to be higher for natural areas that are scarcer. We therefore
expect the absolute values of MWTP to be larger when the green coverage
rate is smaller, because we may acknowledge the rarity of greenery. An over-
all decrease in the green coverage rate caused by increasing urbanization is
observed in the Tokyo metropolitan regions, where the total land mass of city
parks increased by about 16,000 ha between 1965 and 2003; however, farm-
land and forest areas decreased by about 219,000 ha. This led to a decrease of
approximately 22% in total green coverage (Ministry of Land, Infrastructure
and Transport, Japan, 2008). We therefore expect greenery that is scarce in
urban areas (i.e., farmland or forest and wilderness) to have a strong corre-
lation with the well-being indices. Even in the presence of the same green
coverage rate, when inhabitants have more opportunities to interact with green
spaces or walk around residential areas, they may receive greater benefits from
these activities than those who have fewer opportunities to be in close contact
with nature. We thus expect greenery with which people often come into con-
tact (i.e., residential areas, roadsides, parks) to have a strong correlation with
the well-being indices. Furthermore, the quality of greenery is thought to be
related to its value.
The descriptive statistics are reported in Table 9.4.
Table 9.4 Descriptive statistics
Variable Number of Mean Standard Minimum Maximum

observations deviation
Cantril Ladder 2758 5.73 2.03 0 10
Life satisfaction 2758 5.77 2.43 0 10
Subjective 2758 6.19 2.30 0 10
happiness
Affect balance 2758 0.55 1.18 −3 3
Mental health 2758 22.05 6.25 0 36
(GHQ-12)
Age (years) 2758 45.34 13.03 20 83
Male dummy 2758 0.52 0.50 0 1
variable
Chronic illness 2758 0.46 0.50 0 1
dummy variable
Marital status 2758 0.57 0.50 0 1
dummy variable
Extroversion 2758 3.94 1.31 1 7
Agreeableness 2758 4.87 1.03 1 7
Neuroticism 2758 3.85 1.19 1 7
Conscientiousness 2758 4.13 1.20 1 7
Openness to 2758 4.22 1.15 1 7
experience
Annual after-tax 2758 6,102,594 4,343,287 500,000 20,000,000
household income
(yen)
Self-employment 2758 0.13 0.33 0 1
dummy variable
Unemployment 2758 0.03 0.17 0 1
dummy variable
Contract-worker 2758 0.12 0.33 0 1
dummy variable
Retirement dummy 2758 0.06 0.25 0 1
variable
Involuntary- 2758 0.04 0.20 0 1
homemaker dummy
variable
Voluntary- 2758 0.12 0.33 0 1
homemaker dummy
variable
Floor space per 2758 37.02 28.03 4.17 200
person
Extended labor 2758 1.18 4.13 0 40
Junior high school 2758 0.01 0.09 0 1
completion dummy
variable
Technical school/ 2758 0.20 0.40 0 1

junior college
completion dummy
variable
Bachelor’s degree 2758 0.57 0.50 0 1
dummy variable
Graduate degree 2758 0.08 0.28 0 1
dummy variable
Number of people 2758 5.54 5.54 0 70
who can be relied
upon
Community 2758 9.10 21.99 0 208
participation (days/
year)
Participation in 2758 7.27 3.50 0 10
elections
Experience of 2758 0.39 1.18 0 10
crime (number of
incidents)
Lifelong learning 2758 37.12 63.13 0 208
(days/year)
Satisfaction with 2758 7.59 1.99 0 10
accessibility
Green coverage
rates (radius of 0 m
to 100 m)
Residential areas 2758 0.05 0.05 0 0.45
Roadsides 2758 0.02 0.02 0 0.21
Commercial areas 2758 4.58E – 03 0.01 0 0.12
Industrial areas 2758 9.01E – 04 2.70E – 03 0 0.04
Farmland 2758 4.15E – 03 0.02 0 0.22
Forests and 2758 5.72E – 04 0.01 0 0.20
wilderness
Waterfronts 2758 1.23E – 03 0.01 0 0.09
Parks 2758 0.01 0.03 0 0.64
Public facilities 2758 0.01 0.03 0 0.58
Railroads 2758 9.11E – 04 0.01 0 0.11
Total 2758 0.12 0.11 0 1.22
Green coverage
rates (radius of 100
m to 500 m)
Residential areas 2758 0.05 0.04 0 0.25
Roadsides 2758 0.03 0.01 1.67E – 04 0.11
Commercial areas 2758 0.01 0.01 0 0.07
(Continued )

Industrial areas 2758 1.16E – 03 1.92E – 03 0 0.03

Farmland 2758 4.71E – 03 0.01 0 0.09
Forests and 2758 1.24E – 03 0.01 0 0.12
wilderness
Waterfronts 2758 1.86E – 03 3.73E – 03 0 0.03
Parks 2758 0.02 0.03 0 0.33
Public facilities 2758 0.02 0.03 0 0.43
Railroads 2758 1.09E – 03 2.14E – 03 0 0.02
Total 2758 0.15 0.09 8.18E – 04 0.77
Green coverage
m to 1000 m)
Residential areas 2758 0.05 0.03 1.12E – 04 0.21
Roadsides 2758 0.03 0.01 2.15E – 03 0.07
Commercial areas 2758 0.01 0.01 2.26E – 04 0.04
Industrial areas 2758 1.13E – 03 1.51E – 03 4.53E – 06 0.01
Farmland 2758 4.34E – 03 0.01 0 0.05
Forests and 2758 1.29E – 03 0.01 0 0.08
wilderness
Waterfronts 2758 2.03E – 03 2.75E – 03 0 0.02
Parks 2758 0.03 0.03 2.37E – 05 0.21
Public facilities 2758 0.03 0.03 8.84E – 04 0.23
Railroads 2758 1.07E – 03 1.38E – 03 0 0.01
Total 2758 0.16 0.09 0.01 0.53
Green coverage
m to 1500 m)
Residential areas 2758 0.05 0.03 4.70E – 04 0.18
Roadsides 2758 0.02 0.01 1.82E – 03 0.06
Commercial areas 2758 0.01 4.36E – 03 2.95E – 04 0.03
Industrial areas 2758 1.06E – 03 1.14E – 03 9.20E – 06 0.01
Farmland 2758 4.16E – 03 0.01 0 0.04
Forests and 2758 1.18E – 03 3.94E – 03 0 0.05
wilderness
Waterfronts 2758 1.93E – 03 2.15E – 03 0 0.02
Parks 2758 0.03 0.03 1.43E – 05 0.23
Roadsides 2758 0.02 0.01 1.82E – 03 0.06
Public facilities 2758 0.03 0.02 1.78E – 03 0.21
Railroads 2758 1.11E – 03 1.14E – 03 0 0.01
Total 2758 0.15 0.08 0.02 0.46
4 Estimation results
Tables 9.5 to 9.9 report the estimation results where each well-being index is
the dependent variable, and total greenery is an independent variable. Each table
reports the results for greenery in one of the distance bands (0 to 100 m, 100 to
500 m, 500 to 1000 m, and 1000 to 1500 m).11 Other than greenery, the inde-
pendent variables are the same in each model. An ordinary least squares method
(OLS) is used in the estimation. Previous studies use both OLS and ordered pro-
bit as common empirical methods (e.g., Ferrer-i-Carbonell and Gowdy, 2007;
Luechinger, 2009). Because, except for affect balance, the well-being indices are
ordinal, ordered probit theoretically seems to be more appropriate. However, Ferrer-
i-Carbonell and Frijters (2004) found that both approaches can provide robust
Table 9.5 Parameter estimates (total greenery): Cantril ladder
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

Total greenery 0–100 m 0.077*
(0.03)
Total greenery 0.068+
100–500 m (0.035)
Total greenery 0.053
500–1000 m (0.037)
Total greenery 1000– 0.036
1500 m (0.04)
Age −0.0051** −0.0052** −0.0052** −0.0053**
(0.0017) (0.0017) (0.0017) (0.0017)
Age squared 0.000062** 0.000064*** 0.000064*** 0.000064***
(0.000018) (0.000018) (0.000018) (0.000018)
Male dummy −0.04*** −0.04*** −0.04*** −0.04***
(0.0079) (0.0079) (0.0079) (0.0079)
Chronic illness dummy −0.017* −0.017* −0.017* −0.017**
(0.0066) (0.0066) (0.0066) (0.0066)
Marital status dummy 0.044*** 0.044*** 0.044*** 0.044***
(0.008) (0.008) (0.008) (0.008)
Extraversion 0.0081** 0.008** 0.0082** 0.0082**
(0.0029) (0.0029) (0.0029) (0.0029)
Agreeableness 0.0057 0.0059+ 0.006+ 0.0061+
(0.0035) (0.0035) (0.0035) (0.0035)
Neuroticism −0.026*** −0.026*** −0.026*** −0.026***
(0.0032) (0.0032) (0.0032) (0.0032)
Conscientiousness 0.002 0.002 0.0021 0.0021
(0.0031) (0.0031) (0.0031) (0.0031)
Openness to experience 0.00029 0.00019 0.000058 0.000098
(0.0032) (0.0032) (0.0032) (0.0032)
Log annual household 0.041*** 0.04*** 0.04*** 0.041***
income (0.0044) (0.0044) (0.0044) (0.0044)
(Continued )
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m
Self-employment dummy −0.027** −0.027** −0.027** −0.027**

(ref: regular employee) (0.01) (0.01) (0.01) (0.01)
Unemployment dummy −0.062** −0.062** −0.061** −0.061**
Contract-worker dummy −0.046*** −0.045*** −0.045*** −0.045***
Retirement dummy 0.051** 0.051** 0.051** 0.052**
Involuntary-homemaker −0.0055 −0.0045 −0.0041 −0.0039
dummy (ref: regular (0.018) (0.018) (0.018) (0.018)
employee)
Voluntary-homemaker 0.01 0.01 0.011 0.011
dummy (ref: regular (0.012) (0.012) (0.012) (0.012)
employee)
Floor space per person 0.000036 0.000034 0.000031 0.000032
(0.00012) (0.00012) (0.00012) (0.00012)
Extended labor −0.0026** −0.0026** −0.0026** −0.0026**
(0.00079) (0.00079) (0.00079) (0.00079)
Junior high school −0.014 −0.013 −0.015 −0.017
completion dummy (0.04) (0.04) (0.04) (0.04)
(ref: high school)
Technical school/ junior 0.036** 0.035** 0.035** 0.035**
college completion (0.011) (0.011) (0.011) (0.011)
dummy (ref: high school)
Bachelor’s degree 0.043*** 0.043*** 0.043*** 0.043***
dummy (0.0096) (0.0096) (0.0096) (0.0096)
(ref: high school)
Graduate degree dummy 0.053*** 0.054*** 0.054*** 0.054***
(ref: high school) (0.014) (0.014) (0.014) (0.014)
Number of people who 0.0027*** 0.0027*** 0.0027*** 0.0027***
can be relied upon (0.00063) (0.00063) (0.00063) (0.00063)
Community participation 0.00023 0.00022 0.00023 0.00022
(0.00015) (0.00015) (0.00015) (0.00015)
Participation in elections 0.0024* 0.0024* 0.0024* 0.0025*
(0.00098) (0.00098) (0.00098) (0.00098)
Experience of crime −0.0074** −0.0075** −0.0075** −0.0075**
(0.0026) (0.0026) (0.0026) (0.0026)
Satisfaction with 0.025*** 0.025*** 0.024*** 0.024***
accessibility (0.0018) (0.0018) (0.0018) (0.0018)
Lifelong learning 0.000055 0.000054 0.000058 0.000059
(0.000052) (0.000052) (0.000052) (0.000052)
Constant −0.18* −0.18* −0.18* −0.18*
(0.079) (0.079) (0.079) (0.079)
Observations 2758 2758 2758 2758
Note: The estimation method is ordinary least squares.
+, *, **, *** indicate significance at the 10%, 5%, 1%, and 0.1% levels, respectively. Robust standard
errors are in parentheses.
Table 9.6 Parameter estimates (total greenery): life satisfaction
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

(0.038)
(0.044)
(0.047)
(0.05)
Observations 2758 2758 2758 2758
+, *, **, *** indicate significance at the 10%, 5%, 1%, and 0.1% levels, respectively. Robust standard
errors are in parentheses. Independent variables are the same as in Table 9.5 and we obtain almost the
same parameters concerning control variables. We then omit parameter estimates of control variables,
and constant terms because of space limitations. The signs and statistical significance levels of the
results with household income and the control variables included are like those shown in Table 9.5.
Table 9.7 Parameter estimates (total greenery): subjective happiness
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

Total greenery 0–100 m 0.11**
(0.034)
(0.04)
(0.042)
(0.045)
Observations 2758 2758 2758 2758
Note: See Table 9.6.
Table 9.8 Parameter estimates (total greenery): affect balance
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

Total greenery 0–100 m 0.26
(0.18)
(0.2)
(0.22)
Total greenery 1000–1500 m 0.42+
(0.23)
Observations 2758 2758 2758 2758
Table 9.9 Parameter estimates (total greenery): mental health
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

(0.025)
(0.029)
(0.031)
(0.033)
Observations 2758 2758 2758 2758
and similar results. This claim is supported by the results of van den Berg and
Ferrer-i-Carbonell (2007), MacKerron and Mourato (2009), Levinson (2012), and
others. Moreover, using OLS also allows us to easily interpret the implications of
the estimated coefficients and make comparisons between them. Hence, we use
OLS as the main regression algorithm for the interpretation.12
As shown in Tables 9.5 to 9.9, despite a partial lack of statistically significant
results, total greenery is found to have a positive correlation with the well-being
indices. Household income parameters are statistically significant in all cases, as
are most of the greenery parameters. Concerning the parameters of the control
variables, we generally obtain statistically significant expected signs.
The results in Tables 9.5 to 9.9 suggest that total greenery tends to have positive
correlations with the well-being indices. Next, we examine the correlation of vari-
ous types of greenery with well-being. We show the estimation results in Table
9.10. Because of space limitations, we show only the coefficients of greenery. The
estimation formula is essentially the same as that of Tables 9.5 to 9.9, although
the regression is performed with individual specifications for greenery types and
distance to avoid multicollinearity. The signs and statistical significance levels
of the results with household income and the control variables included are like
those shown in Table 9.5.13
Table 9.10 implies the following trends. First, greenery both in residential areas
and on roadsides has a statistically significant positive correlation with all the
well-being indices except for Cantril’s ladder. For Cantril’s ladder, we obtain a
statistically significant correlation only for 0 to 100 m. On the other hand, green-
ery in commercial areas does not have a statistically significant positive corre-
lation with any indices of well-being, while greenery in industrial areas has a
statistically significant positive correlation in some cases, particularly greenery
close to individuals’ homes. Next, greenery in farmland tends to have a statisti-
cally significant positive relationship with the well-being indices, especially the
well-being indices closely related to emotion; that is, subjective happiness, affect
balance, and mental health. Concerning greenery in forests and wilderness, except
for life satisfaction, we find a statistically significant positive relationship with all
Table 9.10 Parameter estimates of various types of greenery
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

Cantril Residential area 1.1+ 0.46 0.85 0.82
ladder Roadsides 4.4** 3.3 2.7 0.44
Commercial area 4.3 5.5 7.7 3.2
Industrial area 23+ 40* 36+ 45
Farmland 2 3.8 5 4.2
Forests and 2.9 8.5+ 11+ 22**
wilderness
Waterfronts −3.8 2.2 14 20
Parks 0.72 0.6 0.94 0.81
Public facilities 0.53 1.4 −0.35 −1.2
Railroads 3.9 −0.35 −21 −24
Life Residential area 1.6* 1.9+ 2.5* 3*
satisfaction Roadsides 3.5+ 4.9+ 6.5* 5.6
Commercial area 5 8 6.9 14
Industrial area 20 36+ 36 42
Farmland 1.8 4.8 7.4 11+
Forests and 4.9 4.9 7.7 29**
wilderness
Waterfronts −10+ 1.2 17 28
Parks 0.41 0.69 1.1 1
Public facilities 0.32 1.9 1.2 1.8
Railroads 6 6.9 −18 −29
Subjective Residential area 1.9** 2.4* 2.5* 3.3**
happiness Roadsides 6.4*** 7.9** 6.7* 5.6+
Industrial area 38** 46* 32 36
Farmland 4.2* 9.6** 10* 12*
Forests and 2.4 8.8 12+ 24**
wilderness
Waterfronts −2.3 0.062 19 23
Parks 1.3 0.86 1.6 0.51
Public facilities −0.11 2.4+ 0.95 0.59
Railroads 0.24 1.7 −21 −19
Affect Residential area 0.61+ 1.1* 1.4* 1.4*
balance Roadsides 2.3** 4.4** 3.8* 3.2*
Commercial area 0.93 2 3.9 0.13
Industrial area 0.29 17+ 7.4 0.5
Farmland 1.5 3.5* 4.7* 5.8*
(Continued )
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m
Forests and −0.21 7.8** 8.7* 11*

wilderness
Waterfronts 1.3 −1.3 8.5 18*
Parks −0.039 0.56 0.59 0.39
Public facilities −1.1+ 0.1 −0.33 −0.63
Railroads 0.49 8.9 −2.6 −8.5
Mental Residential area 3.1+ 4.3+ 7.1* 6.9*
health Roadsides 12** 17* 18* 18*
Commercial area 3.5 8.5 −11 4.1
Industrial area 10 54 −45 30
Farmland 14* 27** 27* 34*
Forests and −7.3 34* 44* 55*
wilderness
Waterfronts 12 −8 45 110*
Parks −5+ −1.2 0.86 −0.61
Public facilities −5.4+ −0.81 −0.44 −0.42
Railroads −13 −31 −95 −34
+, *, **, *** indicate significance at the 10%, 5%, 1%, and 0.1% levels, respectively.
well-being indices except for 0 to 100 m. Greenery near bodies of water tend to
have a positive correlation with well-being indices, particularly greenery within a
1000 to 1500 m radius, although greenery within a 0 to 100 m shows a negative
correlation with life satisfaction. Greenery in parks has no statistically significant
positive relation with the well-being indices. The only statistically significant cor-
relation of greenery in parks is a negative correlation with the mental health score
of park greenery within a 100 m radius. Greenery in public facilities does not have
a statistically significant positive correlation with well-being indices except for
greenery within a 100 to 500 m radius in the case of subjective happiness, while
greenery within a 100 m radius of public facilities has a negative relationship with
affect balance and mental health. In addition, no statistically significant results are
obtained for greenery at railroad sites.
As a robustness check, we include additional explanatory variables to control for
environmental factors other than greenery as control variables for all specifications.
We include the following additional explanatory variables from our survey: satis-
faction with traffic volume in front of residence, satisfaction with noise in front of
residence, satisfaction with air pollution, and satisfaction with water pollution (all
with scores of zero to 10). The results are almost the same as our current results.
Table 9.11 presents the monetary valuations based on Tables 9.5 to 9.10. These
figures are for 1 m² areas of greenery classified by type and distance. Valuations
Table 9.11 Monetary values of greenery (JPY)
0 to 100 to 500 to 1000 to

100 m 500 m 1000 m 1500 m
Cantril Residential area 9734+
ladder Roadsides 86,103**
Commercial area
Industrial area 12,177,172+ 701,858* 204,298+
Farmland 16,301***
Forests and wilderness 140,257+ 56,538+ 71,984**
Waterfronts
Parks
Public facilities
Railroads
Total greenery 3072* 92+
Life Residential area 12,102* 614+ 271* 199*
satisfaction Roadsides 55,707+ 3223+ 1371*
Commercial area
Industrial area 509,305+
Farmland 16,773*** 8146+
Forests and wilderness 78,403**
Waterfronts −3,342,661+
Parks
Public facilities
Railroads
Total greenery 2503* 109* 37* 25*
Subjective Residential area 22,159** 1183* 405* 333**
happiness Roadsides 154,704*** 7866** 2121* 1095+
Commercial area
Industrial area 25,536,981** 1,003,132*
Farmland 595,212* 51,028*** 18,497* 13,299*
Forests and wilderness 73,428+ 98,402**
Waterfronts
Parks
Public facilities 2561+
Railroads
Total greenery 5566** 228** 60** 34*
Affect Residential area 25,417+ 1942* 801* 503*
balance Roadsides 197,757** 16,411** 4447* 2315*
Commercial area
Industrial area 1,305,191+
(Continued )
0 to 100 to 500 to 1000 to

100 m 500 m 1000 m 1500 m
Farmland 68,064*** 31,799* 24,392*

Forests and wilderness 602,706** 203,508* 166,615*
Waterfronts 159,664*
Parks
Public facilities −175,434+
Railroads
Total greenery 362** 99* 49+
Mental Residential area 35,444 +
2101+ 1141* 687*
health Roadsides 277,958** 17,683* 5660* 3552*
Commercial area
Industrial area
Farmland 1,970,710* 142,395*** 50,933* 39,386*
Forests and wilderness 717,374* 286,329* 222,286*
Waterfronts 286,585*
Parks −260,886+
Public facilities −242,520+
Railroads
Total greenery 342+ 125* 69+
Note: +, *, **, *** indicate significance at the 10%, 5%, 1%, and 0.1% levels, respectively, which are
evaluated by the delta method. We use coefficients shown in Tables 9.5 to 9.10 to calculate monetary
values.
tend to be higher the closer the greenery is to individuals’ homes. Concerning

the differences among well-being indices, those related closely to emotion (i.e.,
mental health, affect balance, subjective happiness) tend to show relatively high
monetary values of greenery. Although there are some exceptions, values tend to
be higher in the following order: mental health, affect balance, subjective happi-
ness, life satisfaction, and the Cantril ladder.
5 Conclusion and discussion

In this study, we examined the relationships between greenery (categorized by
land type and distance) and several measures of well-being indices, arguing that
the choice of well-being measure may influence how greenery is valued and
that valuations are likely to differ markedly based on greenery type. This study
revealed that greenery has different correlations with people’s well-being depend-
ing on the specific land use with which it is associated.
For monetary values of total greenery, as Table 9.11 shows, we find a maxi-
mum monetary value of 5566 m2 for subjective happiness for 0 to 100 m and
the minimum monetary value of 25 m2 for life satisfaction for 1000 to 1500 m.
Ambrey and Fleming (2014) show that the marginal WTP annually for a 1%
increase in green spaces (i.e., 143 m2) is $467 per capita within a 700 m neigh-
borhood, which is between 5566 m2 and 25 m2. This indicates that our estimation
results additionally show variation of monetary values among various distances
from residences. Our results also show that some types of greenery have higher
monetary values than greenery in general. This implies a need to evaluate green-
ery by type to show the variation of greenery.
Concerning various types of greenery, first, our results show that both resi-
dential and roadside greenery have relatively statistically significant positive and
large correlations with well-being, which implies high monetary values. This rela-
tively strong relationship with well-being might be driven by their accessibility.
In our daily lives, we observe these types of greenery on the way to school, the
railway or bus station, the shops, and so on. This finding is consistent with those
of Gilchrist et al. (2015), who show that the frequency of greenspace use, and the
total time spent in a greenspace can predict well-being. Similarly, Tsurumi and
Managi (2015) suggest the importance of considering the amount of interaction
individuals have with greenery. Similarly, statistically insignificant coefficients
of greenery in public facilities or railroads may imply relatively little interaction
with greenery in daily life. These results seem to be related to the concept of
“hedonic inertia,” which means that one event in our day, independent from its
duration, can inﬂuence our mood throughout the entire day (see D’Acci, 2013). If
we feel good all the way to the office because we see beautiful greenery, we will
feel its positive inﬂuence all day long.
Next, we do not find statistically significant coefficients concerning greenery
within a 1000 m radius of a body of water. This finding may be related to the
threat of damage from flooding. However, adding flooding inundation area data
to the set of control variables as a robustness check has no effect. In general, there
may be considerable differences between well-managed and poorly managed
greenery along waterfronts. The proper management of greenery can be assumed
to enhance the positive effect, while poor management may mitigate, eliminate,
or reverse the positive effect.
We find that greenery in industrial areas has a positive correlation with mea-
sures of well-being that relatively relate to life evaluation rather than affect, and
considering their comparatively high monetary values, increasing greenery in
industrial areas may have a major influence on how people evaluate their lives.
Industrial areas are typically expected to negatively relate to people’s well-being;
thus, it is important to improve the environmental situation of greenery to main-
tain residents’ well-being, particularly in areas within a 500 m radius of an indus-
trial area.
On the other hand, for farmland greenery, the results indicate that monetary
values are higher in the case of emotional well-being, such as mental health and
affect balance, which implies that greenery in farmland may positively relate to
people’s emotional state more than it relates to how they evaluate their lives.
For forest and wilderness greenery, although greenery close to one’s home tends
not to have a statistically significant correlation with well-being, this may be
because the sample includes few forest and wilderness areas near residences. We
find relatively higher monetary values for emotional well-being, which is in line
with farmland greenery. Greenery in farmland and forests and wilderness tends to
decrease in Tokyo, with a small distribution centered in the western part of Tokyo.
It is important for policymakers to reconsider land use plans in terms of emotional
well-being. In addition, we need to note that greenery with which people often
come into contact (i.e., residential areas and roadsides) and greenery that is scarce
in urban areas (i.e., farmland or forest and wilderness) tend to have strong correla-
tions with the well-being indices.
Concerning public facilities and railroad greenery, we do not find statistically
significant positive or negative correlations with well-being. This may be related
to noise levels. As a robustness check, we include an additional control variable
of self-reported satisfaction with noise around individuals’ own residences. How-
ever, the estimation results of the coefficients for public facilities and railroad
greenery are similar. Since public facilities are places where comparatively many
people assemble, diminished well-being can be related to noise in the midst of
daily life. In addition, fire departments, police stations, and hospitals are included
under the aegis of public facilities, institutions that people often associate with
negative images. Negative images of such land use are likely to connect with
people’s images of greenery on the same land, resulting in a negative relationship
with well-being.
A feature of this study is its definition of greenery as individual trees within a
park rather than the park itself. The results indicate the difficulty in demonstrat-
ing a statistically significant correlation of park greenery defined in this way with
well-being indices. If anything, given the negative correlation mentioned earlier,
the correlation of park greenery may also be subject to the quality of greenery
management, as in the case of waterfront greenery. Additionally, depending on
their location, parks may be exposed to noise and security problems, such as the
nocturnal presence of noisemakers and revelers. In short, these greenspaces rep-
resent a “blind spot” with regard to public safety and are likely create feelings
of unease. This may be a factor responsible for the negative correlation of park
greenery with subjective well-being. Fleming et al. (2016) suggest that greater
access to greenspaces is associated with higher levels of life satisfaction, but the
strength of this association is strongly dependent on the fear of crime (also see
Gidlöf-Gunnarsson and Öhrström, 2007; Shackleton et al., 2015). Opportunities
for the greenery of nearby parks to raise well-being (e.g., recreational features or
features that help create a comfortable environment) appear negated by the same
greenery’s negative influence of heightened unease due to security concerns and
noise. This is likely to have led to the failure to observe statistically significant
results for the many well-being measures. This anxiety-heightening influence is
particularly reflected by greenery in parks that are extremely close to respondents
(i.e., 0 to 100 m away), which could have a negative correlation with the well-
being of people as a result. Although the issue of security has been controlled
through the use of BLI measures, that control may be insufficient. Because further
controls are difficult to implement within the scope of this study, noise and secu-
rity may also be subjects for more detailed investigations in the future, and it is
important to consider the quality of greenery as indicated in Kaźmierczak (2013)
and Tsurumi and Managi (2015).
As indicated in the introduction, while the total land mass of city parks has
increased in the Tokyo metropolitan region, farmland, and forest areas have
decreased. In contrast, we find negative correlations of greenery in parks and posi-
tive correlations of greenery in farmland and forest areas with subjective well-
being. This implies that policymakers need to reconsider land use plans, because
large budget amounts have been spent for construction and maintenance of green-
ery in parks, not greenery in farmland or forest areas. To avoid wasting budgetary
resources, policymakers should consider the possibility of negative factors related
to greenery in parks, as noted earlier, and increase greenery that we often have
contact with (i.e., residential areas or roadsides) or increase or maintain greenery
in farmland or forest areas in urban areas.
Finally, this study suggests that greening policies must consider two points: the
efficacy of greening policies differs not based on the amount of greenery but on
differences in land use, and effects on people’s well-being differ depending on land
use. Moreover, our results can contribute to guiding land use planning by quantify-
ing and comparing the value of different types of greenery and different spatial con-
figurations. Particularly, our results suggest cost-benefits of land use. Municipalities
can establish budgets for greenery by considering its benefits. Our estimated mon-
etary values are thought to be relatively high compared with other factors affecting
well-being. For example, Tsurumi and Managi (2017) show monetary values of var-
ious factors affecting well-being. The authors show that self-evaluated health status
has relatively strong effects on well-being among BLI indices, which is approxi-
mately over USD 40,000 per unit increase (health status is measured on a 5-point
scale). Our current results shown in Table 9.11 are equivalent, if policymakers plan
to increase greenery land use with high WTP or in large areas.
Notes
1 For example, as shown in Figure 9A.1. in Appendix, when looking at the changes in the
extent of green areas (city parks, forest areas, and agricultural land) in the Tokyo metro-
politan region, even though the total land mass of city parks increased by about 16,000
ha between 1965 and 2005, farmland and forest areas decreased by about 219,000 ha.
This has led to a decrease of approximately 22% in total green coverage.
2 The survey was conducted as follows. First, an invitation is emailed from the internet
survey company to relevant cohorts from their one million plus registered monitors
in Japan. To obtain a random sample, emails were sent to selected monitors based on
prefectural demographics, such as population, gender, and age, and basic attributes,
such as income level, and education.) To avoid respondent bias, the research objectives
(i.e., identifying the value of green space) were not disclosed to the research monitors
before they decided to answer the questionnaire. About 62% of the individuals who
were contacted completed the survey.
3 We show the relationship between our survey sample and population in Table 9A.1 in
the appendix, which demonstrates that our sample is almost the same as the population
in terms of age, gender, and income.
4 Benjamin et al. (2014) discuss various indices of well-being and show the theoretical
background in terms of economic literature. D’Acci (2014) discusses various monetary
evaluation methods, including studies using subjective well-being indices.
5 There is some debate as to whether the three types of life evaluation – life satisfaction,
subjective happiness, and the Cantril ladder – signify the same thing (Helliwell et al.,
2012). For example, subjective happiness is easily influenced by feelings (Diener et
al., 2010), and the Cantril ladder is more strongly correlated with income compared
to other life evaluation indices. Meanwhile, the OECD (2013) has argued that affect
should be considered a complement to the principal indices of life evaluation.
6 One of the limitations of our study is that our research area is limited to urban areas.
This is because our high-resolution satellite image is expensive. We had to use limited
areas of images where population density is high, or where possible respondents per
unit of area are high. Future research should include suburban or rural areas to com-
prehensively understand the values of various types of greenery, particularly tree-level
evaluations such as the one presented in this study.
7 Because of space limitations, we cannot present the correlations among the greenery
data categorized by land type within each distance band; however, we note that these
correlations are diverse, which suggests diversity of greenery within respondents’ local
areas. This study uses a radius of between 100 m and 1500 m based on the assumption
of everyday living space accessed by foot.
8 For control variables, we consider the factors that are included in the BLI advocated by
the OECD. From a range of aspects, the BLI focuses on 11 dimensions of well-being:
(1) housing, (2) income and wealth, (3) jobs and earnings, (4) social connections, (5)
education and skills, (7) civic engagement and governance, (8) health status, (9) sub-
jective well-being, (10) personal security, and (11) work/life balance. Except for (6)
environmental quality and (9) subjective well-being, we use the dimensions as control
variables. Specific survey questions for these are shown in Table 9A.2 in the appendix.
Although the BLI topics have not obtained a common understanding for all topics in
the factors that influence well-being, they represent one of the efforts to reveal a com-
prehensive understanding of well-being. Therefore, we used them as control variables.
9 Owing to the large correlations among the greenery in various distance bands, as shown
in Table 9A.3 of the appendix, greenery is analyzed using separate equations for each
distance band to avoid multicollinearity. One of the limitations of our study is that we
do not completely solve the problem of “spurious correlation” between greenery that is
located at different distances from residences. There is a possibility that non-negligible
correlations exist between greenery at different distances.
10 To consider the endogeneity of the income measure, we include household income for
the previous year instead of the current year, but we obtain similar results.
11 When we include all distance variables in one model, we do not obtain statisti-
cally significant coefficients for greenery variables, with some exceptions, due to
multicollinearity.
12 As a robustness check, we show the estimated results by ordered probit in Tables 9A.4 to
9A.7 in the appendix. The results are almost the same as our main results shown in Tables
9.5 to 9.9. Because affect balance includes decimal values, we use only OLS for it.
13 As a robustness check, we show the estimated results by ordered probit in Table 9A.8 in
the appendix. The results are almost the same as our main results shown in Table 9.10.
Because affect balance includes decimal values, we use only OLS for it.
Bibliography
Ambrey, C. and Fleming, C. (2011). Valuing scenic amenity using life satisfaction data.
Ecological Economics, 72: 106–115.
Ambrey, C. and Fleming, C. (2014). Public greenspace and life satisfaction in urban Aus-
tralia. Urban Studies, 51(6): 1290–1321.
Benjamin, D. J., Heffetz, O., Kimball, M. S. and Szembrot, N. (2014). Beyond happiness
and satisfaction: Toward well-being indices based on stated preference. American Eco-
nomic Review, 104(9): 2698–2735.
Clark, A. E. and Oswald, A. J. (2002). A simple statistical method for measuring how life
events affect happiness. International Journal of Epidemiology, 31: 1139–1144.
D’Acci, L. (2013). Hedonic Inertia and underground happiness. Social Indicators Research,
113(3): 1237–1259.
D’Acci, L. (2014). Monetary, subjective and quantitative Approaches to assess urban qual-
ity of life and pleasantness in cities (Hedonic Price, willingness-to-pay, positional value,
life satisfaction, isobenefit lines). Social Indicators Research, 115(2): 531–559.
Diener, E., Kahneman, D., Tov, W. and Arora, R. (2010). Incomeʼs Association with Judge-
ments of Life versus Feelings. In Ed Diener, John Helliwel and Daniel Kahneman (Eds.),
International Differences in Well-Being, 3–15. Oxford, UK: Oxford University Press.
Ferrer-i-Carbonell, A. and Frijters, P. (2004). How important is methodology for the esti-
mates of the determinants of Happiness? The Economic Journal, 114(497): 641–659.
Ferrer-i-Carbonell, A. and Gowdy, J. M. (2007). Environmental degradation and happiness.
Ecological Economics, 60(3): 509–516.
Fleming, C. M., Manning, M. and Ambrey, C. L. (2016). Crime, greenspace and life satisfaction:
An evaluation of the New Zealand experience. Landscape and Urban Planning, 149: 1–10.
Frey, B. S. and Stutzer, A. (2002). Happiness and Economics: How the Economy and Insti-
tutions Affect Well-Being. Princeton, NJ: Princeton University Press.
Gidlöf-Gunnarsson, A. and Öhrström, E. (2007). Noise and well-being in urban residential
environments: The potential role of perceived availability to nearby green areas. Land-
scape and Urban Planning, 83: 115–126.
Gilchrist, K., Brown, C., and Montarzino, A. (2015). Workplace setting and well-being:
Greenspace use and views contribute to employee well-being at peri-urban business
sites. Landscape and Urban Planning, 138: 32–40.
Goldberg, D. P. (1972). The Detection of Psychiatric Illness by Questionnaire: A Technique
for the Identification and Assessment of Non-Psychotic Psychiatric Illness. Maudsley
Monograph No. 21. London: Oxford University Press.
Goldberg, D. P. and Hiller, V. F. (1979). A scaled version of the general health question-
naire. Psychological Medicine, 9: 139–145.
Guthrie, E., Black, D., Bagalkote, H., Shaw, C., Campbell, M. and Creed, F. (1998). Psy-
chological stress and burnout in medical students: A five-year prospective longitudinal
study. Journal of The Royal Society of Medicine, 91: 237–243.
Helliwell, J. F., Layard, R. and Sachs, J. eds. (2012). World Happiness Report, Commis-
sioned for the United Nations Conference on Happiness on April 2nd, 2012 (mandated
by the General Assembly of the United Nations). New York: The Earth Institute, Colum-
bia University.
Kahneman, D., Wakker, P. P. and Sarin, R. (1997). Back to Bentham? Explorations of
experienced utility. Quarterly Journal of Economics, 112: 375–405.
Kaźmierczak, A. (2013). The contribution of local parks to neighborhood social ties. Land-
scape and Urban Planning, 109: 31–44.
Kopmann, A. and Rehdanz, K. (2013). A human well-being approach for assessing the
value of natural land areas. Ecological Economics, 93: 20–33.
Levinson, A. (2012). Valuing public goods using happiness data: The case of air quality.
Journal of Public Economics, 96(9): 869–880.
Li, W., Saphores, J.D.M. and Gillespie, T. W. (2015). A comparison of the economic ben-
efits of urban green spaces estimated with NDVI and with high-resolution land cover
data. Landscape and Urban Planning, 133: 105–117.
Luechinger, S. (2009). Valuing air quality using the life satisfaction approach. The Eco-
nomic Journal, 119(536): 482–515.
MacKerron, G. and Mourato, S. (2009). Life satisfaction and air quality in London. Eco-
logical Economics, 68(5): 1441–1453.
MacKerron, G. and Mourato, S. (2013). Happiness is greater in natural environments.
Global Environmental Change, 68: 1441–1453.
Ministry of Land, Infrastructure and Transport (2008). White Paper on Land, Infrastructure,
Transport and Tourism in Japan. (in Japanese).
OECD (2013). OECD Guidelines on Measuring Subjective Well-being, OECD Publishing.
http://dx.doi.org/10.1787/9789264191655-en
Oswald, A. J. and Powdthavee, N. (2007). Obesity, unhappiness, and the challenge of afflu-
ence: Theory and evidence. Economic Journal, 117: 441–454.
Pevalin, D. J. and Ermisch, J. (2004). Cohabiting unions, repartnering and mental health.
Psychological Medicine, 34: 1553–1559.
Powdthavee, N. (2010). The Happiness Equation: The Surprising Economics of Our Most
Valuable Asset. Icon Books.
Powdthavee, N. and van den Berg, B. (2011). Putting different price tags on the same health
condition: Re-evaluating the well-being valuation approach. Journal of Health Econom-
ics, 30(5): 1032–1043.
Powdthavee, N. and Vignoles, A. (2008). Mental health of parents and life satisfaction of
children: A within–family analysis of intergenerational transmission of well-being.
Social Indicators Research, 88: 397–422.
Robinson, K. L., McBeth, J., and MacFarlane, G. J. (2004). Psychological distress and
premature mortality in the general population: A prospective study. Annals of Epidemiol-
ogy, 14: 467–472.
Rojas, M. and Veenhoven, R. (2013). Contentment and Affect in the estimation of Happi-
ness. Social Indicators Research, 110: 415–431.
Shackleton, S., Chinyimba, A., Hebinck, P., Shackleton, C. and Kaoma, H. (2015). Multiple
benefits and values of trees in urban landscapes in two towns in northern South Africa.
Landscape and Urban Planning, 136: 76–86.
Tsurumi, T. and Managi, S. (2015). Environmental value of green spaces in Japan: An
application of the life satisfaction approach. Ecological Economics, 120: 1–12.
Tsurumi, T. and Managi, S. (2017). Monetary valuations of life conditions in a consistent
framework: The life satisfaction approach. Journal of Happiness Studies, 18(5):
1275–1303.
Van den Berg, B. and Ferrer-i-Carbonell, A. (2007). Monetary valuation of informal care:
The well-being valuation method. Health Economics, 16(11): 1227–1244.
Welsch, H. and Ferreira, S. (2014). Environment, well-being, and experienced preference.
International Review of Environmental and Resource Economics, 7(3–4): 205–239.
Appendix
100
90
80
70 41.9 37.4
33 31.1 29.8 28.4 27 25.6
million ha
60 24.8
50
40
30
50.9 50.1 48.1 48.1 47.7 47.3 46.7 46.1 45.9
20
10
0 0.2 0.3 0.5 0.7 0.9 1.3 1.5 1.7 1.9
1965 1970 1975 1980 1985 1990 1995 2000 2005
Year
City parks Forest areas Agricultural land
Figure 9A.1 Changes in the extent of green areas in the Tokyo metropolitan region
Source: Ministry of Land, Infrastructure and Transport (2008), revised by the authors.
Table 9A.1 Characteristics of respondents with comparison
This survey Comparative survey
Age
Age 20–24 2.47% 7.40%
Age 25–29 5.87% 10.14%
Age 30–34 9.43% 11.02%
Age 35–39 13.63% 11.30%
Age 40–44 12.62% 9.99%
Age 45–49 14.18% 8.64%
Age 50–54 12.62% 6.95%
Age 55–59 10.95% 6.60%
Age 60–64 7.14% 7.80%
Age 65–69 6.38% 6.39%
Age 70–74 2.86% 5.41%
Age 75–79 1.16% 4.84%
Age 80–84 0.69% 3.53%
Gender
Male 52.28% 48.27%
Female 47.72% 51.73%
Household income
0–2,999,999 24.73% 26.41%
3,000,000–4,999,999 25.20% 30.65%
5,000,000–6,999,999 15.95% 15.57%
7,000,000–9,999,999 17.98% 13.41%
10,000,000–14,999,999 10.55% 9.18%
15,000,000 + 5.58% 4.78%
Notes: Age and gender distributions are obtained from the national population census of 2010
in eight metropolitan Tokyo wards (Suginami, Setagaya, Meguro, Shinagawa, Minato, Nakano,
Shibuya, and Shinjuku). The reason the maximum age category is 80–84 is that maximum age in our
analysis sample is 83. Household income is generated by Zenrin Geo Intellgence Co., Ltd., which
specializes in GIS data marketing, and is estimated based on data from various sources, such as the
Housing and Land Survey of Japan, the Population Census of Japan, and the Grid Square Statistics
of Population Census of Japan. We use this 500 m mesh-level data to obtain a household distribution
of income ranges (0–2,999,999, 3,000,000–4,999,999, 5,000,000–6,999,999, 7,000,000–9,999,999,
10,000,000–14,999,999,15,000,000+) in eight metropolitan Tokyo wards (Suginami, Setagaya,
Meguro, Shinagawa, Minato, Nakano, Shibuya, and Shinjuku).
Table 9A.2 Indicators of Better Life Index that are used for control variables in this study
Better Life Index Indicator used in the survey Area of inquiry Selected measures
(BLI) dimension *Corresponding BLI
indicator(s)
(1) Housing Floor space per person (m2) Quality of Floor space (m2)/cohabitating members of the family
* (HO 1) Number of rooms housing
per person
(2) Income and Annual after-tax household Present “Approximately how much will your household’s total after-tax income
wealth income (Yen) and future (including bonuses and pensions) be for fiscal year 2013?”
* (IW 1) Household net consumption *Taxes include income tax, asset taxes, social security costs (pension
adjusted disposable income possibilities funds, health insurance, etc.), and resident’s tax (municipal and prefectural
taxes, etc.).
* Household income includes income from national or local government
assistance, financial assets, and land and buildings (dividends, interest,
rent, etc.).
* Household income does not include profits from the sale of financial
assets, land, or buildings.
(3) Jobs and Occupation (dummy Amount and “What is your present occupation? Please select one of the following
earnings variable) quality of work options:
* (JE 1) Employment rate Self-employed/freelancer
* (JE 2) Long-term Regular employee (contract for an unspecified period of employment)
unemployment rate Temporary/contract employee or part-time employee
* (JE 1) Involuntary part- Unemployed
time employment Retired
* (JE 2) Employees working Homemaker
on temporary contracts Other”
* Dummy variable: Regular employee (contract for an unspecified period
of employment)
(Continued )
indicator(s)
(4) Social Number of people who can Personal “In the event that you were affected by illness or accident, how many
connections be relied upon relations people would you be able to rely upon from among your family, relatives,
* (SC 1) Social network friends, neighbors, and colleagues?” Responses are given as follows: 0, 1,
support 2, 3, 4, 5, 6, 7, 8, 9, 10, or more.
Community activities (days/ Community “How often do you participate in activities in your local area?”
year) relations * Local activities may include participation in the following:
* (SC 1) Frequency of social Neighborhood councils, neighborhood associations, women’s associations,
contact elder associations, youth groups, etc.
* (SC 2) Time spent Community development; care of the elderly, disabled or children; sports
volunteering coaching; local clean-up campaigns; prevention of crime and disasters;
environmental activities; international cooperation; and submission of
policy proposals.
(5) Education and Educational attainment Amount of “Which of the following is your highest level of educational attainment? If
skills (dummy variable) education you are presently in school, please indicate the level you will attain upon
* (ES 1) Educational graduation.
attainment Junior high school
* (ES 1) Education High school
expectancy Technical/vocational/other school, junior/technical college
Bachelor’s degree
Master’s degree
Doctoral degree”
* Dummy variable (High school)
Lifelong learning (days/ Amount of Lifelong learning refers to learning activities voluntarily undertaken by
year) education people at any stage in their life regardless of location. It is not limited to
* (ES 2) Lifelong learning learning opportunities offered by schools or community education such
as courses at community centers but rather refers to a variety of learning
activities that an individual can undertake: formal education, sports,
cultural activities, hobbies, and volunteerism.
indicator(s)
(Cabinet Office Public Survey on Lifelong Learning: Supplement 1, 2004)
“How frequently do you undertake lifelong learning activities?
A few time per year
Around once per month
Around 2 to 3 times per month
Around 1 day per week
Around 2 to 3 days per week
More than 4 days per week
Not at all
Note: Lifelong learning refers to the following interests and activities:
Hobbies (music, art, flower arrangement, dance, calligraphy, recreational
activities, etc.)
Educational and cultural activities (literature, history, science, languages,
etc.)
Activities concerned with social issues (social and current affairs,
international and environmental activities, etc.)
Health and sports (fitness, medicine, nutrition, jogging, swimming, etc.)
Skills used in the home (cooking, dressmaking, kimono making, knitting,
etc.)
Child rearing and education (home education, preschool education,
educational issues, etc.)
Skills, knowledge, and qualifications used in the workplace
Computing (computers and tablets) and online activities (blog editing,
website design, etc.)
Skills and knowledge used in volunteer activities
Activities in natural environments
Formal education (high school, technical, and vocational schools,
universities, and graduate schools)”
(7) Civic Participation in elections Civic “How often do you vote in local government (city, ward, town, or village)
engagement and (index) engagement elections?”
governance * (CEG 1) Voter turnout Responses are given on a scale from 0 (never) to 10 (always).
(Continued )
indicator(s)
(8) Health status Chronic illness (index) Illnesses and Chronic illness
* (HS 2) Self-reported health disorders Dummy variable equal to 1 if the respondent has any of the following
health status conditions below and equal to 0 otherwise.
* (HS 2) Self-reported long- Problems or disabilities relating to the arms, legs, hands, feet, back, or
term illness neck (including osteoarthritis and rheumatoid arthritis)
* (HS 3) self-reported Difficulty seeing (does not include the normal use of glasses or contact
limitations on daily activities lenses for reading and writing)
Difficulty hearing
Skin conditions or allergies
Chest or respiratory problems, asthma, or bronchitis
Heart problems, blood problems, or circulatory problems
Stomach, kidney, or liver problems, or other problems involving the
digestive organs
Diabetes
Alcohol-or drug-related problems
Epilepsy
Migraines or chronic headaches
*Many previous studies used self-reported health status as an indicator
of health (“On the whole, how satisfied are you with your health?”). In
this study, however, psychological health is used as a dependent variable,
therefore the presence of physical illnesses and disorders is measured
directly in order to remove psychological health connotations from this
independent variable.
indicator(s)
(10) Personal Experience of crime in Safe living “Since you began living at your current residence, how often have you
security residence neighborhood environment been a victim of a crime committed in your neighborhood?”
(number of incidents)
* (PS 1) Homicide rate
* (PS 2) Feeling of security
(11) Work and life Long working hours (hours) Distribution of “On an average weekday in the past month, how many hours did you
balance * (WL 1) Long working time between devote to each of the following tasks?
hours work and life Working
Commuting
Household tasks
Nursing someone
Caring for someone
Child rearing
Sleeping
Eating”
Long working hours = The number of hours in excess of 50 worked per
week
Table 9A.3 Correlations among total greenery by distance band
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

0 to 100 m 1
100 to 500 m 0.7213 1
500 to 1000 m 0.5232 0.7773 1
1000 to 1500 m 0.4609 0.6296 0.8418 1
Table 9A.4 Parameter estimates (total greenery): Cantril ladder
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

(0.19)
(0.22)
(0.23)
(0.25)
Observations 2758 2758 2758 2758
Note: The estimation method is ordered probit.
+
, *, **, *** indicate significance at the 10%, 5%, 1%, and 0.1% levels, respectively. Robust standard
errors are in parentheses. Independent variables are the same as in Table 9.5 and we obtain almost the
same parameters concerning control variables; we then omit parameter estimates of control variables
and cut points because of space limitations.
Table 9A.5 Parameter estimates (total greenery): life satisfaction
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

(0.19)
(0.22)
(0.23)
(0.25)
Observations 2758 2758 2758 2758
Note: See Table 9A.4.
Table 9A.6 Parameter estimates (total greenery): subjective happiness
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

(0.19)
(0.22)
(0.23)
(0.25)
Observations 2758 2758 2758 2758
Table 9A.7 Parameter estimates (total greenery): mental health
0 to 100 m 100 to 500 m 500 to 1000 m 1000 to 1500 m

(0.18)
(0.21)
(0.23)
(0.24)
Observations 2758 2758 2758 2758
Table 9A.8 Parameter estimates of various types of greenery
0 to 100 to 500 to 1000 to

100 m 500 m 1000 m 1500 m
Cantril ladder Residential area 0.66+ 0.21 0.48 0.4
Roadsides 2.6** 2 1.5 −0.019
Commercial area 3.2 4.9 5 2.6
Industrial area 15* 26* 24+ 27
Farmland 0.99 2.4 3.1 2.3
Forests and wilderness 1.8 5.5+ 8.4* 14**
Waterfronts −2.9 2 10 13
Parks 0.4 0.25 0.52 0.47
Public facilities 0.34 0.93 −0.2 −0.76
Railroads 3 2.1 −13 −18
(Continued )
0 to 100 to 500 to 1000 to

100 m 500 m 1000 m 1500 m
Life satisfaction Residential area 0.82* 0.96+ 1.3* 1.5*

Roadsides 1.4 2.4+ 3.1+ 2.9+
Commercial area 2.2 6.1+ 4.3 7.4+
Industrial area 10 19+ 17 21
Farmland 0.67 2.3 3.4 5.3+
Forests and wilderness 3.3 2.7 4.6 15**
Waterfronts −5.8+ 1.7 7.1 15+
Parks 0.079 −0.01 0.41 0.56
Public facilities 0.015 0.93 0.7 1
Railroads 3.1 3 −10 −12
Subjective happiness Residential area 1.1** 1.4** 1.5* 1.9**
Roadsides 3.3*** 4.5** 3.8* 3+
Industrial area 21** 26* 19 19
Farmland 2.3* 5.7** 5.8* 6.8*
Forests and wilderness 1.4 5.4+ 8* 14**
Waterfronts −2.3 0.087 13+ 14
Parks 0.54 0.27 0.9 0.1
Public facilities −0.1 1.2+ 0.46 0.4
Railroads 0.33 0.036 −14 −14
Mental health Residential area 0.7+ 1.2* 1.7** 1.7**
Roadsides 2.5** 3.9** 4.3** 4*
Commercial area 0.55 2.9 −2 2
Industrial area 1.3 16 0.19 16
Farmland 3** 6.2** 6.4** 7.9**
Forests and wilderness −2 7.7** 9.5** 11*
Waterfronts 1.2 0.25 12+ 25**
Parks −1.1+ −0.25 0.17 0.01
Public facilities −0.92 0.016 −0.1 −0.11
Railroads −2.5 −4.9 −17 −11
Note: The estimation method is ordered probit.
+
, *, **, *** indicate significance at the 10%, 5%, 1%, and 0.1% levels, respectively. Because affect
balance includes decimal values, we use only OLS for it.
10 The impact of climate change
and extreme events on
agriculture in Africa
Thierry Coulibaly, Moinul Islam
and Shunsuke Managi
1 Introduction
A large part of the economy in Africa depends on the agricultural sector. Accord-
ing to an estimate in 2015, agricultural activities employed nearly 60% of Africa’s
labour force in sub-Saharan Africa (SSA; Barrios, Ouattara, and Strobl, 2008). Addi-
tionally, in the last four decades, African countries have made tremendous progress
concerning their agricultural yield. Several countries doubled their production from
the period of 1981–2000 to the period of 2001–2013 while using approximately the
same agricultural area. However, this progress could decelerate because, among all
sectors, the African agriculture is the sector most highly affected by climate change
(Barrios, Ouattara, and Strobl, 2008; Frisvold and Ingram, 1995).
Indeed, according to the previous literature, developing countries will suffer the
most from climate change, particularly those in Africa. In the same line, according
to the Intergovernmental Panel on Climate Change (IPCC), “Under the assump-
tion that access to adequate financing is not provided, Africa is the continent most
vulnerable to the impacts of projected changes because the widespread poverty
limits adaptation capabilities” (Dell, Jones, and Olken, 2012; Sarker, Alam, and
Gow, 2012); therefore, it is essential to address this matter.
To address the negative effects of climate change on agriculture in Africa, sev-
eral actions have been taken at the microeconomic level. However, because of
the typically low adaptive capacity (for economic reasons) of African farmers,
tackling climate change problems at a macroeconomic (international) level should
be more effective (Loayza and Olaberri, 2012). Thus, identifying countries facing
the same economic difficulties could help continental organizations settle priori-
ties concerning climate change.
In this regard, some studies have been conducted on an economic basis (Felber-
mayr and Gröschl, 2014; Winsemius, Jongman, Ward, and Hallegatte, 2015; Blanc,
2012). Most of them have classified countries as rich and poor or as developed and
developing. However, concerning the classification of rich and poor countries, the
definition of a poor country is neither clear nor accepted by all. Moreover, concern-
ing the classification of developed and developing countries, because all the Afri-
can countries are classified as developing countries, the assertion that all African
countries will be impacted by climate change seems very strict.
Based on this observation, the goal of this chapter is to estimate the impact of
climate change by using a clear and pertinent classification of countries. To do so,
262 Thierry Coulibaly et al.
income level was used as the method of classification. The advantage of this clas-
sification method is that it considers countries with the same level of income that
have, to some extent, national similitudes.
Indeed, even if the continent is globally poor, as mentioned by the literature
review, Africa is composed of several countries with different levels of devel-
opment. When considering the classification of income level, three classes of
income can be found, and all of them concern developing countries. Therefore, it
will be interesting to identify the income level categories that reflect the African
countries that will face negative effects of climate change.
This chapter will integrate three main improvements upon the past literature
on the aforementioned topic. It is the first work to estimate the impact of climate
change on agriculture in Africa using income level classification. In addition,
among the studies including average weather changes and extreme events in the
same model (Schlenker, Hanemann, and Fisher, 2006), this chapter uses a different
definition of extreme events (data from EM-DAT). This definition might lead to
intuitive results, as the EM-DAT database is the largest concerning extreme events.
Third, unlike most of the studies investigating the impact of climate on Africa
while focusing on SSA, in this model, both sub-Saharan and non-sub-Saharan
countries are included. This procedure was done to check the difference between
these two regions.
The remainder of the chapter is structured as follows. Section 2 introduces the
methodology, which is mainly based on an econometric approach, the data used
and the descriptive statistics. Section 3 provides the graphs of some key variables
and the results of the regression model. Moreover, Section 3 includes the interpre-
tation of the results. Section 4 presents some discussions on the topic, and Section
5 provides some policy implications.

A yearly income level classification was used for the countries in Africa. Based
on this classification, a production model was applied to the agricultural outcome.
This methodology allows the impact of climatic variables in each income level
class to be estimated.
2.1 Production model

The production model used in this chapter is based on the cross-sectional models
of past literature (Barrios, Ouattara, and Strobl, 2008; Frisvold and Ingram, 1995).
The canonical form of the equation can be expressed as follows:
α α
Yit = β0 + ∑ βi Z it + ∑ βit X it + uit
i =1 i =1
where Yit is accounted as the agricultural output, Zit as the climatic variables, Xit
as the socio-economic variables and uit as the error term. In addition to the fact
that several studies have argued that micro-level data are better for agricultural
The impact on agriculture in Africa 263
studies, macro-level data seems more appropriate for the present chapter. Then,
in order to use macro-level data, which well aggregates micro-level data, George
Frisvold identified a set of variables which catch the sources of productivity in
African agriculture.
Starting from this study, some improvements have been made which resulted
in the following linear form:
log ( yit ) = β0 + βtemp *log(tempit ) + βrain * log (rainit ) + βdrought * droughtit
+ β flood * floodit + βL log ( Lit ) + β A log ( Ait ) + βF log ( Fit )
+ βK log ( K it ) + βM log ( M it ) + ηt + ui + εit
where Yit is the total agricultural production; tempit is the annual mean tempera-
ture; rainit is the total amount of precipitation; droughtit is the number of drought
occurrences; floodit is the number of flood occurrences and Lit, Ait, Fit, Kit and Mit
are labour, livestock, fertilizers, capital and agricultural area, respectively, for the
country i at the year t. Complementary to these variables, a time fixed effect ηt,
common to all countries (representing any possible technological change during
the study period) and a country fixed effect ui (representing any country-specific
characteristics) are included. The error term of the equation is represented by εit .
This function will be used for the panel data with a fixed effect at the country
level for the period of 1961–2014.
Some robustness tests were conducted and validated the pertinence of the
model (see Annexes).
2.2 Data
The dataset consisted of unbalanced panel data including 47 African countries.
The a priori objective was to include all the African countries, but due to a lack
of data, seven countries had to be dropped. Additionally, concerning the special
case of Sudan, because of its division into two countries (North Sudan and South
Sudan) during the study period, it was not included in the dataset. (The list of
countries is available in the Annexes).
The total time period is from 1961 to 2014. However, this period was divided into
two, because the World Bank income level classification was established in 1987.
2.2.1 Production index

The net production index from the Food and Agriculture Organization (FAO) was
used as the explained variables.
The FAO indices of agricultural production show the relative level of the
aggregate volume of agricultural production for each year in comparison to
the base period 2004–2006. They are based on the sum of price-weighted
quantities of different agricultural commodities produced after deductions of
quantities used as seed and feed weighted in a similar manner. The resulting
aggregate represents, therefore, disposable production for any use except as
seed and feed.
Food and Agriculture Organization of the United Nations (FAO)
The advantage of this variable is that it considers the production of all species
and values this production in a fixed and equal “international commodity price”1
(developed by the FAO) for all the countries. Then, it allows the agricultural pro-
duction of different countries to be compared.
2.2.2 Climate
Many studies have explained the impact of the mean annual rainfall and tempera-
ture on the agricultural outcome (Barrios, Ouattara, and Strobl, 2008; Dell, Jones,
and Olken, 2012; Sarker, Alam, and Gow, 2012). Others have explained the vulner-
ability of the agricultural outcome to extreme events (Loayza and Olaberri, 2012;
Felbermayr and Gröschl, 2014; Winsemius, Jongman, Ward, and Hallegatte, 2015).
Both the average climatic changes and the extreme climate events are included in
this section. This approach allows the consideration of more information about
the causes that underlie changes in the total agricultural production. The proce-
dure, which consists of bringing together average and extreme climatic events,
has already been done, using the standard precipitation index as the main tool to
determine drought and flood events (Blanc, 2012). However, the drawback of this
method is its inability to distinguish extreme events that are damaging and those that
are not. Therefore, while this model uses the same variables, the dataset is different.
The temperature and rainfall data were collected from the Centre of Environ-
mental Data Analysis (www.ceda.ac.uk/). The annual mean temperature is repre-
sented in degrees Celsius, and the precipitation is represented in mm/year.
Concerning the extreme events, the dataset was collected from the Emer-
gency Events Database (EM-DAT: The Emergency Events Database – Université
Catholique de Louvain (UCL) – CRED, D. Guha-Sapir, www.emdat.be, Brussels,
Belgium). This database is one of the most complete concerning extreme events
worldwide. According to the collection method of http://emdat.be,
for a disaster to be entered into the database at least one of the following
criteria must be fulfilled:
• Ten (10) or more people reported killed

• Hundred (100) or more people reported affected
• Declaration of a state of emergency
• Call for international assistance.
The units of floods and droughts are the occurrences.
2.2.3 Socio-economic factors

In addition to the climatic factors, others aspects determine the crop production
of countries (Frisvold and Ingram, 1995). It is therefore important to take into
account some socio-economic factors. In the case of agriculture, these factors
consisted of the total agricultural area per country; the total head of cattle, sheep
and goats as a proxy of livestock; the yearly consumption of phosphate, nitrogen
and potassium as a proxy for the amount of fertilizer used per country; the total
population as a proxy of the labour force per country; and the yearly GDP/capita
as a proxy of capital.
Others pertinent socio-economic variables according to the literature review
were not collected, particularly the irrigation variables and the educational level
(Schlenker, Hanemann, and Fisher, 2006). However, because of the generally high
costs of the irrigation methods (Faures and Santini, 2008), the GDP/capita vari-
able (a proxy of capital) could help us consider any eventual irrigation variable.
The following tables present the summary statistics of our dataset during our
three periods of study (Table 10.1a–c).
Table 10.1a Summary statistics for the period 1967–2014 per countries
Variable Unit Observations Mean Std. dev. Min Max

PIN – 2474 356.94 102.42 128.72 730.88
Livestock Head of 2474 10,500,000 15,200,000 33,608 133,000,000
livestock
Population 1000 person 2474 21,129.18 31,997.88 131 272,194
Agricultural 1000 ha 2474 20,269.67 20,662.98 86 101,335
land
Fertilizers tons 2454 69,437.39 193,547 0 1,840,399
GDP/capita USD 2010 2136 1610.30 2126.61 130.44 19,492.68
Drought unit 2420 0.10 0.30 0 2
Flood unit 2420 0.32 0.71 0 7
Temperature degree Celsius 2474 24.05 3.53 11 30
Rainfall mm/an 2474 922.89 597.81 24 3347
Table 10.1b Summary statistics for the period 1961–1986 per countries

PIN – 1170 325.89 115.40 128.72 730.88
livestock
Population 1000 person 1170 13,838.86 19,833.18 131.50 149,488
Agricultural 1000 ha 1170 19,502.37 20,268.81 95 101,335
land
Fertilizers tons 1170 46,290.67 137,596.60 0 1,232,886
GDP/capita USD 2010 882 1403.11 1969.05 130.44 19,492.68
Drought unit 1144 0.07 0.26 0 1
Flood unit 1144 0.07 0.27 0 2
Temperature degree Celsius 1170 23.71 3.57 11.30 28.90
Rainfall mm/an 1170 944.03 615.41 25.40 3346.90
Table 10.1c Summary statistics for the period 1987–2014 per countries

PIN – 1304 384.81 79.50 182.82 727.53
livestock
Population 1000 person 1304 27,670.33 38,726.92 605.57 272,194
Agricultural 1000 ha 1304 20,958.11 20,994.24 86 98,125
land
Fertilizers tons 1284 90,529.02 231,164.80 0 1,840,399
GDP/capita USD 2010 1254 1756.03 2219.99 149.76 11,933.78
Drought unit 1276 0.12 0.33 0 2
Flood unit 1276 0.53 0.89 0 7
Temperature degree Celsius 1304 24.35 3.46 11.50 29.60
Rainfall mm/an 1304 903.93 581.16 23.70 2873.20
3 Empirical results
Two approaches were important in this study: the graph of the annual climatic
features, which is important in order to observe general trends, and the application
of the model, which is important to estimate the impact of these climatic features
on agriculture.
3.1 Figures
Graphs and the mean annual temperature and rainfall trend lines were produced
on the three periods of the study.
During the period of 1961–2014, the temperature increased and the rainfall
decreased (Figures 10.1 and 10.2). This trend is problematic because African
farmers, in general, are known to significantly depend on precipitation (Cooper,
Dimes, Rao, Shapiro, Shiferaw, and Twomlow, 2008).
During the period of 1961–1986, the rainfall decreased and the temperature
increased (Figure 10.1a–f). However, during the period of 1987–2014, the rainfall
and the temperature both increased. Nevertheless, the increase in the rainfall in this
time period was smaller than its decrease in the previous period. Additionally, the
increase in the temperature in the period of 1987–2014 was approximately two times
faster than the increase in temperature in the previous periods. (Figure 10.1a–f)
3.2 The impact of climate change on African agriculture

in 1961–2014
In the past few years, the existence of climate change seems to have been com-
pletely accepted. However, according to studies, the climatic features impacting
most economic activities differ. In any case, based on the result of the model
(Table 10.2), two variables particularly impacted African agriculture.
Temperature 1961-2014
25.5
25
24.5
Temperature
24
y = 0.0217x + 23.469
23.5
23
22.5
Year
Figure 10.1a Temperature trends in the period 1961–2014
A A B B
Temperature
Temperature
1961-1986
1961-1986 Temperature
Temperature
1987-2014
1987-2014
24.4 24.4 25.2 25.2
24.2 24.2 25 25
24.8 24.8
24 24
degree celcius
degree celcius
24.6 24.6
degree celcius
degree celcius
23.8 23.8 24.4 24.4

23.6 23.6 24.2 24.2
23.4 23.4 24 24
23.8 23.8
23.2 23.2 y = 0.0246x
y = 0.0246x
+ 24.005+ 24.005
23.6 23.6
23 23 y = 0.0103x
y = 0.0103x
+ 23.606+ 23.606 23.4 23.4
22.8 22.8 23.2 23.2
1987
1989
1991
1987
1993
1989
1995
1991
1997
1993
1999
1995
2001
1997
2003
1999
2005
2001
2007
2003
2009
2005
2011
2007
2013
2009
2011
2013
1961
1963
1965
1961
1967
1963
1969
1965
1971
1967
1973
1969
1975
1971
1977
1973
1979
1975
1981
1977
1983
1979
1985
1981
1983
1985
Year Year Year Year
Figure 10.1b
Temperature trends in the Figure 10.1c Temperature trends in the
period 1961–1986 period 1987–2014
Rainfall 1961 - 2014

1200
1150
1100
1050
Rainfall
1000
950
900
y = -1.6198x + 1075.1
850
800
Year
Figure 10.1d Rainfall trends in the period 1961–2014

A A B B
Rainfall
Rainfall
1961-1986
1961-1986 Rainfall
Rainfall
1987-2014
1987-2014
1200 1200 1100 1100
1150 1150
1050 1050
1100 1100
1000 1000
1050 1050
mm/year
mm/year
mm/year
mm/year
1000 1000 950 950
950 950
900 900
900 900 y = 1.0354x
y = +1.0354x
996.59+ 996.59
850 850
850 850 y = -6.0378x
y = -6.0378x
+ 1132.4+ 1132.4
800 800 800 800
1961
1963
1965
1961
1967
1963
1969
1965
1971
1967
1973
1969
1975
1971
1977
1973
1979
1975
1981
1977
1983
1979
1985
1981
1983
1985
1987
1989
1991
1987
1993
1989
1995
1991
1997
1993
1999
1995
2001
1997
2003
1999
2005
2001
2007
2003
2009
2005
2011
2007
2013
2009
2011
2013
Year Year Year Year
Figure 10.1e Rainfall trends in the period Figure 10.1f Rainfall trends in the period
1961–1986 1987–2014
Table 10.2 Production model in Africa
All the climatic Temperature and Flood and

variables rainfall drought
Temperature −0.295 −0.338
(0.572) (0.589)
Rainfall 0.095** 0.102***
(0.036) (0.034)
Drought −0.005 −0.011
(0.014) (0.015)
Flood 0.019** 0.022***
(0.008) (0.007)
Livestock 0.196*** 0.191*** 0.201***
(0.042) (0.043) (0.042)
Population −0.497* −0.397 −0.488*
(0.253) (0.239) (0.251)
GDP/capita 0.142 0.142 0.142
(0.097) (0.096) (0.096)
Agricultural area 1.112*** 1.113*** 1.108***
(0.181) (0.188) (0.179)
Fertilizers 0.019** 0.020** 0.019**
(0.009) (0.009) (0.009)
constant −3.783 −4.548* −4.196**
(2.395) (2.374) (1.906)
R2 0.64 0.64 0.64
N 1859 1913 1859
*p < 0.1.
**p < 0.05.
***p < 0.01.
The rainfall had a positive and significant impact on agriculture. This result high-
lights the fact that, for most African countries, annual precipitation is an important
variable to farmers.
However, the positive and significant impact of floods on African agriculture is
not as expected as the positive impacts of rainfall. Two main reasons may explain
this result.
First, since floods are generally localized in precise geographical regions of
countries and occur in a precise period (affecting one of many growing seasons),
their negative effects might be smaller than expected.
Second, floods have a positive effect because they generally contribute to the
nationwide water supply, which is important for the agricultural sector. Moreover,
the water resources in Africa are structurally low. Floods might be positive because
Africa is the continent with the lowest conversion factor of precipitation to runoff
water, averaging 15% (the second lowest continent is Australia and Oceania, with
a value of 30%). The current trends in major river basins indicate a decrease in
runoff of approximately 17% over the decade 1990–2000; moreover, lake stor-
age and major dams have reached critically low levels (IPCC, 2007; Panel et al.,
2007). Because of the importance of runoff water in agriculture, floods represent
a source of punctual increases in these long-term scarce resources.
Thus, the positive impact of flooding implies that its benefits on agriculture
outweigh its drawbacks (Loayza and Olaberri, 2012).
Nevertheless, even if floods showed some positive impacts, this effect should
be well understood because it only remains positive insofar as the benefits from
flooding are superior to its drawbacks. It is important to keep in mind that the
number of floods is counted, with no regards to the individual levels of impor-
tance. Indeed, minor floods may have a positive impact on agriculture, but big
floods (considering the number of people affected and the amount of damages) do
not (Loayza and Olaberri, 2012).
In summary, the results indicated that the general trend in the African agricul-
tural sector depends on water availability.
3.3 The impact of climate change on the whole continent

for the periods 1961–1986 and 1987–2014
Many conferences and studies concerning climate change have occurred in the past
20 years; meanwhile, the impact of climate change started more than two decades
ago. This may imply that the impact of climatic factors worsens over time.
After the model was applied to Africa for the periods 1961–1986 and 1987–
2014, some noticeable changes were observable (Table 10.3).
After the study period was divided into two periods, the climatic features with
the greatest impact on the agricultural outcomes changed. Thus, during the period
1961–1986, the only significant climatic variable was rainfall, which had a posi-
tive effect (Table 10.5, column 2). Meanwhile, during the period 1987–2014, the
climatic variables playing a major role in the agricultural production of Africa
were temperature and drought.
Table 10.3 The impact of climate change on the whole continent for the periods 1961–
1986 and 1987–2014
Africa Africa SSA Africa SSA

1961–2014 1961–1986 1961–1986 1987–2014 1987–2014
Temperature −0.295 0.404 0.399 −1.680** −1.722**
(0.572) (0.313) (0.334) (0.648) (0.837)
Rainfall 0.095** 0.171*** 0.174*** 0.051 0.053
(0.036) (0.048) (0.052) (0.048) (0.055)
Drought −0.005 0.012 0.011 −0.026*** −0.026***
(0.014) (0.009) (0.010) (0.009) (0.009)
Flood 0.019** −0.007 0.002 0.011 0.012
(0.008) (0.015) (0.014) (0.007) (0.008)
Livestock 0.196*** 0.068 0.083 0.238*** 0.224***
(0.042) (0.055) (0.061) (0.051) (0.051)
Population −0.497* −0.161 −0.112 −0.236 −0.084
(0.253) (0.298) (0.300) (0.324) (0.340)
GDP/capita 0.142 0.065 0.057 0.330*** 0.344***
(0.097) (0.075) (0.074) (0.081) (0.079)
Agricultural area 1.112*** 0.608*** 0.575** 0.708*** 0.656***
(0.181) (0.204) (0.212) (0.207) (0.225)
Fertilizers 0.019** 0.004 0.004 −0.001 −0.001
(0.009) (0.012) (0.012) (0.007) (0.007)
constant −3.783 −2.333 −2.612 0.527 −0.104
(2.395) (3.641) (3.667) (4.113) (4.754)
R2 0.64 0.21 0.21 0.68 0.67
N 1859 723 702 1066 969
These differences in the impact of climatic features during the two time periods
can be explained by two main reasons (Schlenker and Lobell, 2010).
First, temperature has increased very quickly during the last few decades
worldwide (Climate Change, Drought and Agriculture, 2011; Schlenker and
Lobell, 2010), and this is the case in our two periods (Figure 10.1a–c). Therefore,
the temperature, which was initially at a sustainable level for agriculture, had a
negative effect after a certain limit. The rejection of the nonparametric Mann-
Whitney test null hypothesis for the temperature confirmed this assertion (see
Annexes). There was a significant difference in the global mean temperature in
Africa during the two periods. This explains the negative and significant impact
of temperature in 1987–2014.
Second, the mean annual rainfall in 1961–1986 decreased five to six times
faster than it increased in 1987–2014 (Figure 10.1e–f). Meanwhile, the increase in
temperature during the period of 1987–2014 was twice as fast as that in the period
of 1961–1986 (Figure 10.1b–c). Hence, because the magnitude of the change in
temperature was greater than the magnitude of the change in precipitation, the
importance of the variation in temperature may have exceeded that of the varia-
tion in rainfall.
Finally, drought had a negative and significant effect in the second period of the
study. Because of the warming in the continent, which affected the continent, the
number of droughts increased, and the ability to address them decreased (Table
10.1b–c).
In summary, changes in the annual temperature of Africa have been important
in recent years. The fast increase in the annual mean temperature (Figure 10.1a) is
an evidence of global warming and its impact in Africa (Table 10.3)
3.4 Impact of climate on agriculture per income level
3.4.1 The classification of income level

The World Bank divides economies into four classes: low, lower middle, upper
middle, and high income.
This classification scheme was created in 1987 and is used as a guideline for
granting preferences to domestic civil work contractors and for various borrowing
terms. This classification represents an improvement in the grouping of countries
into developed and developing countries. The biggest difference between these
two classifications is the division of the developing countries into low-income,
lower-middle-income and upper-middle-income levels.
To make this classification, the World Bank uses two main components: the
gross national income (GNI) per capita and a group threshold index system.
The determination of the income class level is then produced using the follow-
ing inequality:
Lower bound threshold Upper bound threshold

< GNI <
of the income group capita of the income group
In this method, the income of countries is measured using the GNI per capita
in US dollars. The conversion from the local currency is done using the World
Bank Atlas method; the advantage of this method is the reduction in the impact of
exchange rate fluctuations in the cross-country comparison of national incomes.
Afterwards, the income group thresholds are estimated using the World Devel-
opment Indicators. These indicators include not only the economic development
status but also some socio-development characteristics.
The thresholds for the income level classification in 2014 were as follows:
• Low-income level of countries: less than or equal to $1045

• Lower-middle-income level: greater than $1045 and less than or equal to $4125
• Upper-middle-income level: greater than $4125 and less than or equal to
$12,745
• High-income level: greater than $12,745.
The repartition of the countries per income level (Table 10.4) indicates that most
of the countries in the sample were classified as low-income countries. Merely
Table 10.4 Repartition of countries per level of income
Level of income Frequency Percent Cumulus

Low income 857 66.02 66.02
Lower medium income 294 22.65 88.67
Upper medium income 147 11.33 100.00
Total 1298 100.00
30% of the countries in the sample were classified as middle-income countries.

Finally, no countries were classified as high-income countries.
3.4.2 The impact of climate change on agriculture and income

level classes
To compare the difference between the impacts of the climatic features on differ-
ent income level classes, the model was run for each class (Table 10.5).
Additionally, in order to determine whether regional distribution could explain
some differences in the impact of climate change on the countries, the model
was applied to the sub-Saharan and the non-sub-Saharan countries. Indeed, it is
common to study the impact of climate change only on the sub-Saharan countries
(Barrios, Ouattara, and Strobl, 2008; Frisvold and Ingram, 1995; Blanc, 2012).
Based on the regression results, the climatic features that impacted agriculture
the most were temperature and drought. This observation was the same for the
sub-Saharan and non-SSA countries in 1990–2014. Furthermore, the amplitudes
of the climatic variables for these two African regions were approximately the
same (Table 10.5).
When focusing on the main purpose of this study, the results show, as expected,
that the lower the income level of a country is, more likely the country will be
impacted by the climatic patterns.
Therefore, the low-income and the lower-middle-income countries suffered
from a negative effect of temperature on the total agricultural production. How-
ever, the effect of temperature was not significant for the upper-middle-income
level countries. With 88.67% of the observations (Table 10.4) classified as low-
income or lower-middle-income countries, this explains why temperature is a sig-
nificant climatic feature for the whole continent.
Furthermore, drought had a negative and significant effect on agricultural pro-
duction in the low-income African countries. This highlights the vulnerability of
farmers in low-income countries to this extreme event. This result could be the
consequence of poor drought prevention measures in these countries (Wilhite,
1996).
Moreover, rainfall had a positive and significant impact on low-income coun-
tries only. This could be the result of the dependency of these countries on rainfall.
Indeed, compared to the world averages, several African countries have a very
low water use efficiency (Gebrehiwot and Gebrewahid, 2016). Thus, unable to
Table 10.5 Impact of climatic variables per level of income of countries
All countries SSA No SSA Low Lower Upper

1990–2014 1990–2014 1990–2014 1990–2014 1990–2014 1990–2014
Temperature −1.591** −1.848** −1.831* −1.526** −1.418** 1.266
(0.621) (0.851) (0.688) (0.709) (0.612) (0.781)
Rainfall 0.037 0.017 0.069* 0.099* −0.064 0.051
(0.049) (0.062) (0.027) (0.056) (0.065) (0.044)
Drought −0.022** −0.024** −0.201** −0.018** −0.058 −0.071
(0.009) (0.009) (0.065) (0.008) (0.058) (0.054)
Flood 0.009 0.013* 0.000 0.009 0.019 0.004
(0.007) (0.007) (0.015) (0.006) (0.015) (0.016)
Livestock 0.222*** 0.195*** 0.492 0.246*** 0.487* 0.024
(0.054) (0.057) (0.333) (0.071) (0.238) (0.086)
Population −0.152 0.107 −1.721* −0.163 0.002 −1.483***
(0.333) (0.368) (0.684) (0.418) (0.566) (0.403)
GDP/capita 0.369*** 0.378*** 0.009 0.346*** 0.116 −0.021
(0.081) (0.081) (0.093) (0.081) (0.210) (0.022)
Agricultural 0.541** 0.511** 0.965* 0.683*** 0.116 2.163***
area (0.204) (0.214) (0.418) (0.155) (0.542) (0.482)
Fertilizers 0.001 0.000 0.073 0.004 −0.001 −0.020
(0.007) (0.007) (0.050) (0.005) (0.020) (0.020)
Constant 1.076 0.338 10.396 −0.735 0.788 −4.979
(4.144) (5.008) (4.920) (5.588) (6.902) (6.423)
R2 0.66 0.66 0.85 0.70 0.77 0.57
N 950 837 113 646 186 118
*p < 0.1.
**p < 0.05.
***p < 0.01.
Note: The time period selected was 1990–2014 because of data availability.
effectively use the surface water resources, rainfall has a greater importance for
these countries.
In summary, up to the lower-middle-income level, the countries in Africa are
at a high risk of experiencing negative effects of climate change on agriculture,
which corresponds to 38 countries of the sample from the year 2014.
4 Discussion
The results of this study highlight three main aspects. First, since approximately
1987, temperature seems to be the most harmful climatic feature impacting the
agricultural production in Africa. Furthermore, this negative impact of tempera-
ture on the continent is supported by the negative impact of drought.
Second, it was clearly found that the relatively poorer countries are those that
suffer the most under climate change. More precisely, countries classified as low
and lower-middle income are undergoing negative effects of climate change (par-
ticularly related to temperature), while the ones classified as upper-middle income
seem to be able to manage the negative impacts of climate change. Additionally,
the negative effects of climate features worsen in low-income countries compared
to lower-middle-income countries. In addition to the negative effects of tempera-
ture, low-income countries suffer from drought and are significantly dependent
on rainfall.
Third, since 1990, the impacts of climate change in Africa have been similar in
SSA countries and the other African countries.
Consequently, due to economic matters, several African countries are facing
the same situation with no regards to their geographical localization or their agro-
ecological zones.
Surprisingly, in spite of this concern, almost all the African countries
increased their agricultural production during the period of study (Figure 10.2);
the cause of this contradiction is the related to environmental and social issues
(Wealth, 2012).
Indeed, when changing the measure of development from an economic basis,
such as the level of income or the gross domestic product (GDP), to a sustain-
able basis, such as the Inclusive Wealth Index (IW), the situation appears more
problematic.
With the IW/capita as a reference, three kinds of capital exist for each nation,
the natural capital (represented by fossil fuels, minerals, forest resources and
agricultural land), the human capital (represented by the health and education
of population) and the produced capital (represented by equipment, machinery,
roads and other factors of nations). In the case of this study, the agricultural
production was assimilated as produced capital; agricultural area and live-
stock were assimilated as natural capital; and the population was assimilated as
human capital.
For Africa, during the period of 1990–2014, the major source of wealth was
based on the natural capital (40%). However, the African natural capital per
capita has decreased by approximately 40%, and the other sources of wealth
PIN per level of income = For each year Sum of PIN for all the
countries / number of country in the income class
600
500
400
PIN
300
Equaon of the PIN of Africa
200
y = 2.4028x + 290.14
100
0
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
Year
Low income countries Lower-middle income countries
Upper middle income countries Africa
Linear (Africa)
Figure 10.2 PIN per level of income
have increased by approximately 10% each. Moreover, most of the African

countries have experienced a decrease in their agricultural area per capita
(Wealth, 2012).
As shown in Figure 10.3a–b, except Rwanda and Kenya, all the countries clas-
sified as Low-level income are witnessing a decrease of their IW index. For most
of them, this decrease in Inclusive Wealth Index is inferior to −1 during the period
1990–2014.
Interestingly, almost all the countries in Africa classified as low and lower-
middle-income countries were classified as least developed countries (LDC) by
the LDC classification. The countries that were classified as LDC but not as low-
income or lower-middle-income countries in 2017 are Angola and Equatorial
Guinea (UN/DESA et al., 2016). The present chapter highlights the pertinence
of the LDC classification in the action against acclimate change on agriculture in
Africa, because through it, the United Nations (UN) provides funding to address
the impact of climate change.
This highlights that LDC African countries (approximately low and lower-
middle-income countries) are making progress on economic and social aspects
but are worsening on an environmental basis. It might be one of the conse-
quences of climate change negatively impacting agricultural activities, which
consequently obliges farmers to use more natural or social resources to reach a
given economic standard.
5 Conclusion and political implications

The choice of productivity over sustainability may increase agricultural produc-
tion but is already worsening other facets of countries.
Figure 10.3a Growth in inclusive wealth per capita (1990–2014)
Figure 10.3b Most occurring level of income per country (1990–2014)
Nevertheless, almost half of the LDCs in the world acknowledge in their respec-
tive intended nationally determined contributions (INDC) some social, environ-
mental and economic benefits from actions (mitigating or adaptive actions related
to climate change) taken in the agricultural sector (Environment and Natural
Resources Management Working Paper the Agriculture Sectors in the Intended
Nationally Determined Contributions). Agriculture in the LDCs is still an oppor-
tunity to create the three types of capital described by the IWR. It is important
to determine whether countries will be able to find enough funding to take these
steps.
Access to finance is a challenge for developing African countries. It has been
shown that farmers who have enough credit are more able to take some steps to
adapt to climate change (Bryan, Deressa, Gbetibouo, and Ringler, 2009; Challinor,
Figure 10.4a The numbers of funds in Figure 10.4b

Level of income per
Africa in 2010 countries (2010–2014)
Wheeler, Garforth, Craufurd, and Kassam, 2007). This is a motivation for the set-
tlement of several funds specialized in the agricultural sector in developing African
countries. For example, FAO enumerated a list of funds operating in developing
countries, from which we constructed two maps summarising their work (Figure
10.4a–b; McNellis, Mhlanga, Miller, Richter, McNellis, and Mhlanga, 2010).
To make this comparison, an average classification of the income level of Afri-
can countries during the period 2010–2014 was estimated. For each country, this
average consisted of the most frequently appearing income level from 2010 to
2014. Indeed, it happens that some countries fluctuate between two income levels,
but these fluctuations generally lasted for less than 1 or 2 years. Hence, in a given
year, one may miss the long-run economic characteristic of a country. To avoid
a possible bias of these economic fluctuations, an average frequency for 5 years
was considered.
Additionally, the period 2010 to 2014 was chosen for two reasons. First, the
funds enumerated by the FAO have been listed in 2010. Second, because the
funds generally operate with a long-term perspective we choose the longest pos-
sible time frame which includes 2014 as the final year.
Finally, two criteria must be fulfilled for a fund to be considered. First, the fund
should focus on investments in agriculture. Its impact in the agricultural busi-
ness could be direct or indirect through investments in microfinance that provide
access for smallholders. Second, the fund should focus on African developing
countries.
Most of the agricultural funds in developing countries are in low-income coun-
tries. The countries that need the most funding seem to be the ones that welcome
funding the most (see Table 10.6).
However, the presence of funds in countries classified in the lowest income
level does not seem to be the result of an advanced targeting (Figure 10.4a–b).
When closely inspecting the presence of agricultural funds in Africa, it appears
Table 10.6 Total number of countries with at least one agricultural fund
Level of income 2010–2014 Total number Number of

of funds countries
Low-income countries 2010–2014 86 34
Lower-middle-income countries 2010–2014 16 8
Upper-middle-income countries 2010–2014 10 5
Total 112 47
This table accounts only countries from the sample of this study.
Table 10.7 Number of agricultural funds for developing countries, per country in 2010 for
the sample of this study
Name of country Number Name of country Number

of funds of funds
Morocco 2 Lesotho 0
Angola 2 Madagascar 0
Burundi 0 Mali 5
Benin 3 Mozambique 8
Burkina Faso 1 Mauritania 2
Botswana 3 Mauritius 0
The central African Republic 0 Malawi 5
Cote d’Ivoire 3 Namibia 0
Cameroon 0 Niger 3
The Democratic Republic of 2 Nigeria 8
the Congo
Congo 0 Rwanda 3
Comoros 0 Senegal 4
Djibouti 0 Sierra Leone 1
Algeria 0 Somalia 0
Egypt 2 Swaziland 1
Eritrea 0 Chad 0
Ethiopia 0 Togo 2
Gabon 0 Tunisia 0
Ghana 6 United Republic 10
of Tanzania
Guinea 0 Uganda 9
The Gambia 0 South Africa 5
Guinea-Bissau 0 Zambia 9
Kenya 10 Zimbabwe 3
Libya 0
This classification uses “the Agricultural investments funds for developing countries” as source. Each
fund listed in this article have been reviewed and the main area of activities have been selected and
considered in the construction of this Table and Figure 10.4a-b
that some countries like Ghana or South Africa (classified as lower middle income
and upper middle income, respectively) host several agricultural funds, while
other countries like the Central African Republic or Burundi, which are low-
income countries, don’t welcome any.
In conclusion, even though, several aspects must be considered when settling
a fund, by considering the level of income, we can see that there is not a clear
prioritization of the low-income countries. The funds may have misdirected their
efforts away from the countries suffering the most. The impact of climate change
on agriculture highlights the clear difference in the vulnerability of low-income
countries compared to that of other types of countries.
Note
1 The “international commodity prices” are used to avoid the use of exchange rates for
obtaining continental and world aggregates and to improve and facilitate the interna-
tional comparative analysis of productivity at the national level. These international
prices, expressed in so-called international dollars, are derived by using a Geary-Khamis
formula for the agricultural sector. This method assigns a single “price” to each com-
modity. For example, 1 metric ton of wheat has the same price, regardless of the country
where it was produced. The currency unit in which the prices are expressed has no influ-
ence on the indices published.
Bibliography
Barrios, S., B. Ouattara, and E. Strobl, “The Impact of Climatic Change on Agricultural
Production: Is It Different for Africa?”, Food Policy, vol. 33, no. 4, pp. 287–298, 2008.
Blanc, E., “The Impact of Climate Change on Crop Yields in Sub-Saharan Africa,” Ameri-
can Journal of Climate Change, vol. 1, no. 1, pp. 1–13, 2012.
Bryan, E., T. T. Deressa, G. A. Gbetibouo, and C. Ringler, “Adaptation to Climate Change
in Ethiopia and South Africa: Options and Constraints,” Environmental Science & Pol-
icy, vol. 12, no. 4, pp. 413–426, 2009.
Challinor, A., T. Wheeler, C. Garforth, P. Craufurd, and A. Kassam, “Assessing the vulner-
ability of food crop systems in Africa to climate change,” Climatic Change, vol. 83,
no. 3, pp. 381–399, 2007.
Saleth, R. M., A. Dinar, and J. A., Frisbie, Climate change, drought, and agriculture: The
role of effective institutions and infrastructure. In Dinar, A. and Mendelsohn, R. (Eds.),
Handbook on Climate Change and Agriculture (pp. 466–485). Edward Elgar: Chelten-
ham, UK, 2011.
Cooper, P.J.M., J. Dimes, K.P.C. Rao, B. Shapiro, B. Shiferaw, and S. Twomlow, “Coping
Better with Current Climatic Variability in the Rain-Fed Farming Systems of Sub-Saha-
ran Africa: An Essential First Step in Adapting to Future Climate Change?,” Agriculture,
Ecosystems & Environment, vol. 126, no. 1–2, pp. 24–35, 2008. https://doi.org/10.1016/j.
agee.2008.01.007
Dell, B. M., B. F. Jones, and B. A. Olken, “Temperature Shocks and Economic Growth:
Evidence from the Last Half Century,” American Economic Journal: Macroeconomics,
vol. 4, no. 3, pp. 66–95, 2012.
Strohmaier, R., J. Rioux, A. Seggel, A. Meybeck, M. Bernoux, M. Salvatore, J. Miranda,
and A. Agostini, The agriculture sectors in the Intended Nationally Determined
Contributions: Analysis. Environment and Natural Resources Management Working
Paper No. 62, 2016.
Faures, J.-M., and G. Santini, “Water and the Rural Poor: Interventions for Improving
Livelihoods in Sub-Saharan Africa,” Rome, Italy, 2008.
Felbermayr, G., and J. Gröschl, “Naturally Negative: The Growth Effects of Natural Disas-
ters,” Journal of Development Economics., vol. 111, no. November 1, pp. 92–106, 2014.
Frisvold, G., and K. Ingram, “Sources of Agricultural Productivity Growth and Stagnation
in Sub-Saharan Africa,” Agricultural Economics, vol. 13, no. 1, pp. 51–61, 1995.
Gebrehiwot, K. A., and M. G. Gebrewahid, “The Need for Agricultural Water Management
in Sub-Saharan Africa,” Journal of Water Resource and Protection., vol. 8, no. 9, pp.
835–843, 2016.
Parry, M., M. L. Parry, O. Canziani, J. Palutikof, P. van der Linden, and C. Hanson (Eds.),
Climate change 2007-impacts, adaptation and vulnerability: Working group II contribu-
tion to the fourth assessment report of the IPCC (Vol. 4). Cambridge University Press.
Loayza, N. V., and E. Olaberri, “Natural Disasters and Growth : Going Beyond the Aver-
ages,” World Development, vol. 40, no. 7, pp. 1317–1336, 2012.
McNellis, P., N. Mhlanga, C. Miller, S. Richter, P. McNellis, and N. Mhlanga, “Agricultural
Investment Funds for Developing Countries,” Fao, pp. 1–165, 2010.
Panel, T. I. et al., Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribu-
tion of Working Group II to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change (Vol. 2). Cambridge University Press: Cambridge, UK, 2007.
Sarker, M.A.R., K. Alam, and J. Gow, “Exploring the Relationship between Climate
Change and Rice Yield in Bangladesh: An Analysis of Time Series Data,” Agricultural
Systems, vol. 112, pp. 11–16, 2012.
Schlenker, W., W. M. Hanemann, and A. C. Fisher, “The Impact of Global Warming on U.S.
Agriculture: An Econometric Analysis of Optimal Conditions,” Review of Economics
and Statistics, vol. 88, no. February, pp. 113–125, 2006.
Schlenker, W., and D. B. Lobell, “Robust Negative Impacts of Climate Change on African
Agriculture,” Environmental Research Letter, vol. 5, no. 1, p. 014010, 2010.
UN/DESA, UNCTAD, ECA, ECE, ECLAC, ESCAP, ESCWA, and UNWTO, World Eco-
nomic Situation Prospects, New York, 2016.
UNU-IHDP and UNEP, “Inclusive Wealth Report 2012: Measuring Progress Toward Sus-
tainability,” International Journal of Sustainability in Higher Education, vol. 13, no. 4,
p. ijshe.2012.24913daa.006. https://doi.org/10.1108/ijshe.2012.24913daa.006, 2012.
Wilhite, D. A., “A Methodology for Drought Preparedness,” Natural Hazards, vol. 13,
no. 3, pp. 229–252, 1996.
Winsemius, H., B. Jongman, P. J. Ward, and S. Hallegatte, Disaster risk, climate change,
and poverty: Assessing the global exposure of poor people to floods and droughts, Work-
ing Paper No. WPS7480, November, 2015.
Annexes
(a) Robustness test
i. Haussmann test

To run our model, we checked for the most appropriate estimation method. By this
logic, we ran the Haussmann test to determine whether fixed or random effects
was the best fit for our model. The results of these tests clearly indicated the fixed
effect estimation was more appropriate for our study.
It must be noted that the Haussmann test cannot be used with vce(robust),
vce(cluster cvar), or p-weighted data. Therefore, in order to run the model, we ran
the Haussmann test and we did not use the robust standard error.
Test: Ho: difference in coefficients not systematic
Chi2(62) = (b-B)’[(V_b-V_B) ^ (−1)] (b-B)

= 511.51
Prob>chi2 = 0.0000
(V_b-V_B is not positive definite)
As the results of the Haussmann test show, the probability is less than 0.05, so the
use of fixed effects is recommended for our model.
ii. Wald test to determine the need for the time fixed effect
After determining the need for the fixed effect estimation in our model, we ran
a test in order to determine the need for a time fixed effect. Although we sup-
posed that the time fixed effect could represent a certain technological change
in our model, the pertinence of this interpretation had to be proven. The Wald
test clearly rejects the hypothesis, implying that the coefficients for all years are
jointly equal to zero, accounting for the necessity of the years’ fixed effects for
the model.
To determine if time fixed effects are needed when running a fixed effects
model, we used a joint test to investigate whether the dummies for all years are
equal to 0.
F (43, 43) = 29.76

Prob > F = 0.0000
Prob >F is < 0.05, so we reject the null hypothesis that the coefficients for all years
are jointly equal to zero; therefore, time fixed effects are needed in this case.
iii. Heteroscedasticity test

We checked for heteroscedasticity in our model through the Modified Wald test
for group-wise heteroscedasticity in a fixed effect regression model. The results
of this test clearly show the presence of heteroscedasticity in the model of origin.
Consequently, we used a robust standard error for our estimation, also known as
the White-Huber standard errors. This estimator is robust to some types of mis-
specification as long as the observations are independent.
iv. Wooldridge test

We checked for the presence of serial correlation in our model by the Wooldridge
test. The results of the test strongly rejected the null hypothesis, which was postu-
lating no first-order correlation.
To test for serial correlation, we ran the Wooldridge test for autocorrelation in
panel data.
H0: no first-order autocorrelation
F (1, 43) = 19.582

Prob > F = 0.0001
No serial correlation.
v. Two-sample Wilcoxon rank-sum (Mann-Whitney) test

Because in the second part of our study we divided the period of our sample
into 1961–1986 and 1987–2014, we were interested in knowing to what extent
the rainfall and temperature variables of the two periods differed from each
other. We ran a nonparametric Wilcoxon rank-sum test, which is also known as the
Mann – Whitney two-sample statistic. The null hypothesis of this test was that the
variables tested (temperature and rainfall) in the two periods follow an identical
distribution. The results of the test showed that the distributions of temperature in
the two periods were significantly different, but they failed to show a significant
difference in the distribution for rainfall.
Two-sample Wilcoxon rank-sum (Mann-Whitney) test
For temperature
Period Observation Rank sum Expected

1987–2014 1304 1709811.5 1613700
1961–1987 1170 1351763.5 1447875
combined 2474 3061575 3061575
Unadjusted variance 3.147e+08
Adjustment for ties −41603.955
Adjusted variance 3.146e+08
Ho: temperature (period = 1987–2014) = temperature (period = 1961–1987)
z = 5.418
Prob > |z| = 0.0000
For rainfall
Period Observation Rank sum Expected

1987–2014 1304 1587842 1613700
1961–1987 1170 1473733 1447875
combined 2474 3061575 3061575
Unadjusted variance 3.147e+08
Adjustment for ties −22.193697
Adjusted variance 3.146e+08
Ho: rainfall (period = 1987–2014) = rainfall (period = 1961–1987)
z = −1.458
Prob > |z| = 0.1449
vi. Collinearity test

The problem of multicollinearity may occur when two dependent variables have a
strong level of correlation. The standard value of 0.7 served as a reference for the
presence of multicollinearity in the econometric model.
We ran the collinearity test of the independent variables and observed a prob-
lem of multicollinearity in our data between the total population and the livestock
(see Annexes). However, as this problem did not concern our variables of interest,
temperature, rainfall, drought and flood, we assumed the importance of that mat-
ter to be minor.
Correlation analysis of the continuous explanatory variables included in the
analysis.
Temperature Rainfall Drought Flood Livestock Population GDP/ Agricultural Fertilizers
capita area
Temperature 1.0000
Rainfall 0.0368 1.0000
Drought 0.0148 −0.0665 1.0000
Flood 0.0294 −0.0398 0.0916 1.0000
Livestock −0.0150 −0.1748 0.0959 0.3683 1.0000
Population 0.0093 −0.1090 0.0215 0.3377 0.7812 1.0000
GDP/capita −0.1788 −0.1019 −0.0721 −0.0131 0.0722 −0.0227 1.0000
Agricultural area −0.0791 −0.2822 0.0841 0.2200 0.6460 0.4604 0.2088 1.0000
Fertilizers −0.2344 −0.4230 −0.0086 0.1101 0.4054 0.5065 0.2249 0.3226 1.0000
List of countries
Algeria, Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Central
African Republic, Chad, Comoros, Congo, Côte d’Ivoire, Democratic Republic
of the Congo, Djibouti, Egypt, Eritrea, Ethiopia, Gabon, The Gambia, Ghana,
Guinea, Guinea-Bissau, Kenya, Lesotho, Libya, Madagascar, Malawi, Mali, Mau-
ritania, Mauritius, Morocco, Mozambique, Namibia, Niger, Nigeria, Rwanda,
Senegal, Sierra Leone, Somalia, South Africa, Swaziland, Togo, Tunisia, Uganda,
United Republic of Tanzania, Zambia, Zimbabwe.
11 Sustainable development and
performance measurement
Global productivity decomposition
Robi Kurniawan and Shunsuke Managi
1 Introduction
Global political milestones of sustainable development have grown. Concerns
and goals have shifted from an emphasis on pollution control and availability of
natural resources to a more balanced position that places human development at
the center (Quental et al., 2011). In September 2015, the United Nations General
Assembly adopted the 2030 agenda for Sustainable Development Goals (SDGs).
These important goals range from responsible consumption and production,
improvements in well-being, quality education, and health to the protection of
global assets, including oceans and a stable climate. As such, governments need
a measure to monitor and judge if their development programs to meet the goals
are sustainable (Dasgupta et al., 2015). The system of national accounts (SNA)
that is in common use today records resource flows, such as consumption, invest-
ment, employment, and government expenditure, and these do not measure all
dimensions of societal progress (Schoenaker et al., 2015). Gross domestic product
(GDP) is viewed as a good indicator of economic activity, but it ignores social
costs, environmental impacts, and income inequality. It is a poor measure of the
well-being of society and its sustainability (Costanza et al., 2014b).
Addressing the need for sustainability measurement, UNU-IHDP and UNEP
(2014) used the concept of Inclusive Wealth (IW), which is the social value (not
dollar price) of all productive assets. As such, sustainability is defined as the posi-
tive change in human well-being represented by non-decreasing inclusive wealth,
which refers to human capital, produced capital, and natural capital as sources
of human well-being. This framework has a theoretically consistent measure of
wealth measurement as a key in sustainability evaluation.
Since inclusive wealth extends the economic growth model, Arrow et al.
(2004) suggested that Total Factor Productivity (TFP) plays an important role
in assessing sustainability. The basic idea behind TFP is the concept of a time
asset. The United Nations University International Human Dimension Program
(UNU-IHDP and UNEP, 2012) treat time as a source of well-being. Using this
concept, we can understand that the same productive base of a country can lead to
an increase (decrease) in aggregate output over time due to productivity changes
in the resource utilization. Enhancement of productivity is an important key to
Sustainable development 287
economic growth, as growth escalates pollution and emission as undesirable out-
puts due to massive resource utilization (Akao and Managi, 2007).
However, there have been several studies exercising cross-country TFP
analysis, such as Arrow et al. (2004), Zhou et al. (2007), Xepapadeas and Vouvaki
(2009), Brandt et al. (2013), Kumar and Managi (2010), Mahlberg and Sahoo
(2011), Arrow et al. (2012), UNU-IHDP and UNEP (2014) and Chen et al. (2015).
However, no study offers a comprehensive review of inclusive wealth by consid-
ering natural capital and undesirable output simultaneously. The contribution of
natural capital and undesirable outputs, such as emissions or pollutants, as side
effects to the production of a country’s performance remains unknown. This study
considers natural capital and other conventional inputs, such as inputs and carbon
damages, as undesirable outputs to productivity measures.
There are several specific aspects of this study that incorporate current empiri-
cal research on sustainability. First, the study considers the concept of inclusive
wealth, including natural capital as an input and incorporated carbon damages as
an undesirable output, in productivity measurement. Second, an applied method-
ological concept provides a comprehensive analysis of efficiency level of each
country’s input/output and compares them with the estimated optimal level. Mea-
suring the contribution effect of each input and output, especially natural capital
as an input and carbon damages as an undesirable output to efficiency and pro-
ductivity change, is an important addition to the current research of sustainability.
We analyzed how these countries differ with respect to the effective utilization
of their productive assets and the reduction of the negative impacts of climate
change. Some countries use their endowed capital efficiently with appropriate
productivity changes and future-oriented stock consumption schemes, while oth-
ers do not use their capital as efficiently. Third, we analyze this performance mea-
surement for 140 countries, which represents 99% of GDP and 95% of the global
population, from 1995 to 2010.
To conduct this study, we employ a directional distance model (WRDDM) as
a new methodological approach to decompose the inefficiency productivity into
each input/output source of the studied countries.
This chapter is organized with Section 2 providing an overview of the litera-
ture, focusing on sustainability and wealth and highlighting previous empirical
research on cross-country productivity. Section 3 describes the methodology and
data. Section 4 presents empirical results on productivity and efficiency decom-
position. Section 5 draws conclusions and suggests possibilities to extend this
research.
2 Background
2.1 Sustainability and wealth

Sustainable development, which is commonly traced to the ‘Bruntland Report’,
actually originated from classical utopias that expressed themes of social jus-
tice, environmental stewardship and economic growth (Harlow et al., 2013).
288 Robi Kurniawan and Shunsuke Managi
Cross-country empirical findings confirm a closer association between human
development, socio-economic, and sociopolitical factors to sustainable develop-
ment (Mukherjee and Chakraborty, 2013). Therefore, current trajectories of eco-
nomic growth are problematic for sustainability because they are accompanied by
a depletion of many natural resources and a deterioration in various environmen-
tal services (Jorgenson, 2015).
Human well-being depends on the way that ecosystems work, which provide
essential material things such as food, wood, and clean water (Haines-Young and
Potschin, 2009). On the one hand, scholars consider ecosystems as important
determinants for an economy. The main objective of natural resource economists
is to provide a better understanding of the role of natural resources in an economy
in order to develop more sustainable methods for managing those resources to
ensure their availability to future generations (Callan and Thomas, 2000). Envi-
ronmental economists generally study theoretical or empirical studies of the eco-
nomic effects of national or local environmental policies worldwide, including
the flow of residuals coming from the economy to the environment (Callan and
Thomas, 2000). On the other hand, ecological economists acknowledge that the
human economy is a subsystem within nature (Common and Stagl, 2005). Eco-
logical welfare is not considered as secondary but rather as equal to human wel-
fare (Imran et al., 2014). Ecosystem services contribute more than twice as much
to human well-being than global GDP (Costanza et al., 2014a). Ehrlich and Goul-
der (2007) confirmed that substitution elasticity between reproducible capital and
human capital is a very important form of natural capital. Frye-Levine (2012)
argued that ecological resource efficiency is an important aspect of sustainable
development.
Sneddon et al. (2005) argued that attention is needed regarding how an amal-
gam of ideas embrace a plurality of epistemological and normative perspectives
to advance the sustainability debate beyond its post-Brundtland quagmire. Having
differing paradigms, both perceive connections between nature and economics
as essential, as well as the importance of valuing natural capital and ecosystem
services. Incorporation of natural capital and ecosystem services into decision-
making is an essential component to fulfill current and future human well-being
(Guerry et al., 2015).
The world’s political and environmental landscape has changed significantly in
recent years, and substantial progress has been made by economists working with
ecologists and other natural scientists to value some ecosystem goods and ser-
vices. Accounting in monetary units is useful for showing the relative magnitude
of ecoservices. Therefore, using market prices rather than shadow prices, official
statistics on national income generates natural capital to be only a small contribu-
tion because nature’s services are underpriced in the market (Dasgupta, 2009).
Therefore, this valuation is different than commodification or privatization.
Structured on both market and non-market assets, which are also termed as
genuine savings, inclusive wealth was introduced by Pearce and Atkinson (1993)
and then adopted by Hamilton and Clemens (1999) and Arrow et al. (2004).
This approach focuses on the amounts of all types of capital that contribute to
well-being with the amount of capital representing the productive ability of well-
being. This aggregated capital is termed as Inclusive Wealth (IW) and includes
man-made capital, human capital, natural capital, and intangible capital, such
as knowledge and institutions, as sources of human well-being. They consider
intergenerational well-being as an approach to measure sustainable development.
In their formulation, a society’s economic development is sustainable at a point
in time if its wealth at constant shadow prices are non-decreasing at that time.
Non-declining inclusive wealth implies the possibility of non-declining human
well-being.
Statisticians have begun to consider shadow prices of the market and non-
market value, including several ecosystem services. The authors of the Inclusive
Wealth Report 2014 (IWR 2014), for example, used official statistics to arrive
at the value of reproducible capital, including the natural capital UNU-IHDP
and UNEP (2014). They show with empirical evidence that natural capital has
declined in 116 of 140 countries.
2.2 Empirical research on productivity measure

Using capital and labor as inputs and GDP as an output, Arrow et al. (2004) found
that the contribution of TFP to the inclusive wealth growth rate ranged between
−0.40% (Middle East/North Africa) to 6.33% (China). In addition, they adopted
TFP using traditional growth accounting techniques and found the contribution
to be between −2.12% (Venezuela) and 2.71% (China). However, their measures
ignore the effects of natural capital as input for productivity measures.
Xepapadeas and Vouvaki (2009) found that the externality of the environment
is considerable in TFP estimations, which drives traditional TFP downwards. Tak-
ing energy as an environmental factor in the aggregate production function, they
found TFP growth has been driven into the negative range. It supposes that eco-
system services are an unpriced factor of production. Brandt et al. (2013) incorpo-
rated the use of natural capital, especially oil and gas, various minerals, and wood,
as an input factor in addition to labor and produced capital. They reported that
failing to account for natural capital tends to lead to an underestimation of pro-
ductivity growth. UNU-IHDP and UNEP (2014) calculated TFP as a residual by
considering natural capital as an explicit factor of input to the production process.
Using the Malmquist Productivity Index approach, TFP shows negative growth in
91 out of the 140 studied countries with ±6% growth rates. However, these studies
did not calculate directly undesirable output, such as emissions or pollutants, as
the side effect of production.
Using a ratio-based method, the Malmquist Productivity Index has limitations,
especially when the key focus is to measure differences in environmental impact,
such as emissions (Azad and Ancev, 2014). Empirical studies also show that the
ratio-based productivity index may overestimate productivity change (Managi,
2003).
Developing the Luenberger Productivity Index, Kumar and Managi (2010)
proposed a productivity that considers undesirable outputs, especially carbon
dioxide (CO2) and sulfur dioxide (SO2) emissions. They measured this by using
data from 51 developed and developing countries over the period 1971 to 2000.
Using this index, about half of the countries exhibited productivity growth.
Their results showed a technological catch-up of low-income countries,
although overall productivities showed different results. Chen et al. (2015)
decomposed comprehensive efficiency measures into individual input/output
components. This was applied to a panel of 99 countries over 1991 and 2003.
They found that the OECD countries performed better than the non-OECD
countries in regard to goods, labor and capital efficiencies but worse in regards
to energy efficiencies.
It is evident that cross-country productivity has been extensively analyzed from
different perspectives and methodologies. However, no study offers a compre-
hensive review of inclusive wealth that considers a sustainability measure and
undesirable output simultaneously. In this study, we provide a deeper analysis of
how individual inputs and outputs affect countries’ efficiency-productivity.
3.1 Methodology
To measure cross-country productivity, we employ the Luenberger Productivity
Index. This method has advantages of productivity change estimation with the
additive distance function model over the Malmquist Productivity Index (Balk
et al., 2008). With this form, we gained the benefits of easy computation and
incorporation of undesirable outputs into the model. We could examine the rela-
tionship between carbon damages as a negative impact of utilizing input for cross-
country productivity.
In addition, the measurement takes into account the efficiency of resource
use and productivity changes. Hence, best practice countries can be consid-
ered as most sustainable under the prevailing resource constraints. Similarly,
the distance to the frontier (benchmark) represents the inefficiency of resource
use. In other words, a country on the frontier will perform better under the
same resource constraints than a country further away from the frontier. It is
possible for two countries to have the same level of capital assets, but differ-
ent levels of sustainable development because of how they use their capital
assets or saving rates.
To conduct technical inefficiency decomposition, we employed a new meth-
odological approach proposed by Chen et al. (2015) and Barros et al. (2012),
who used the WRDDM. This model has the advantages of solving technical dif-
ficulties involved in the empirical analysis of how to decompose and quantify
the share of each individual component (outputs/inputs) for total efficiency. The
linear model form will compare individual countries’ components to estimated
benchmark/optimal levels during the study period. Furthermore, we decompose
the inefficiency productivity into each input/output source for the studied country
during the study period.
3.1.1 Technical inefficiency decomposition
This type of directional vector assumes that an inefficient country can decrease
productive inefficiency while increasing desirable outputs and decreasing unde-
sirable outputs and/or inputs in proportion to the initial combination of actual
inputs and outputs. Compared to the traditional model of efficiency measurement
that assumes all good outputs and contracts all inputs and bad outputs by the same
rate, the WRDDM formula allows for the technical inefficiency among each of the
desirable outputs, so the undesirable outputs and inputs for each Decision Making
Unit (DMU) can be different. In fact, the proportional increasing or decreasing of
input, output, and undesirable output is not always same.
It is implied that we can analyze which input/output of each country can decrease
their inefficiency score, so we can find the source of the input/output component
for improvement. For instance, specific inputs or bad outputs that are common in
production are expected to influence the provision of a country’s desirable output.
Thus, these undesirable by-products can have an evident effect on the output effi-
ciency for countries at a point where joint-production appears in production.
In this model, inputs are represented by x ∈ R+N , good outputs are represented
by y ∈ R+M , and bad/undesirable outputs are represented by b ∈ R+J . The R+* repre-
sents a non-negative Euclidean. Using this model, we can include the inefficiency
of sources of inputs, desirable outputs and undesirable outputs when we evaluate
the performance of a decision-making unit (DMU), such as firm and country. The
DDF searching can increase the good outputs and decrease the bad outputs and
inputs directionally, as defined by the following model:

D ( x, y, b|g ) = sup {β : ( x + β g , y + β g , b + β g ) ∈ T } (1)
Non-zero vector g = (−g x , g y − gb ) specifies the directions in which inputs, desir-

able/good outputs, and undesirable/bad outputs are weighted. The vector deter-
mines how to elevate productivity for inefficient exponents towards the frontier
production line. The technology reference set T = {(x, y, b): x can produce
(y, b)} satisfies the assumptions of constant returns to scale, strong disposability
of desirable outputs and inputs, and weak disposability of undesirable outputs.
Assuming that there are k = 1 . . ., K, DMUs, such as country and factory,
use the inputs x = ( x1 , x2 , ., xN ) ∈ R+N to concurrently produce desirable outputs
k k k k
J
y k = ( y1k , y2k , ., y Nk ) ∈ R+M and undesirable outputs b k = (b1k , b2k , ., bJk ) ∈ R+. The
DEA piecewise reference technology (T) can be formulated as:
 K K
T = ( x, y, b) : ∑zk ymk ≥ ym , m = 1,…., M , ∑zk b jk = b j , j = 1,…., J ,
 k =1 k =1
K 

∑ zk xnk ≤ xn , n = 1,…, N , zk ≥ 0, k = 1,…., K 
 (2)
k =1



where zk represents the intensity variable to extend or dwindle the individual specific
activities of DMU k to develop convex combinations of observed inputs and outputs.
This formula is expressed as:
′ ′ ′ M   J   N 
D ( x k , y k , b k ; g ) = ρ k = maxwy ∑ωmyαmk  + wb ∑ω bj β kj  + wx ∑ωnx γ kj  s.t
′ ′ ′ ′
 m=1   j=1   n=1 
K K
∑zk ymk ≥ ymk ′ + αmk g ym , m = 1,…., M , ∑zk b jk = b jk ′ − β kj gbj ,

k =1
′
k =1
′
(3)
K
j = 1,…., J ∑zk xnk ≤ xnk ′ − γ nk g xn , n = 1,…, N , zk ≥ 0, k = 1,…., K
′
k =1
where ρ k ′ stands for technical inefficiency that measures the maximum expansion
of desirable outputs and contraction of undesirable outputs and inputs that remain
′
technically feasible. If ρ k = 0, the DMU k’ works on the frontier of T with techni-
′
cal efficiency, meaning that this DMU is fully efficient. Otherwise, ρ k > 0, then
k′
DMU k’ operates inside the frontier of T. The αm refers to the individual ineffi-
ciency score of the desirable output ym, β kj ′ represents the individual inefficiency
score of undesirable output b j, and γ nk ′ refers to the individual inefficiency score
of input xn. ω x , ω y , ωb represent that the priorities given to the inputs, and desir-
able and undesirable outputs are normalized to unity. This formula allows for the
expansion
and shrinkage of undesirable outputs and inputs all together.
If D ( x, y, b|g ) = 0, the DMU is technically efficient because it offers no extra
improvement in desirable output, and undesirable output and inputs exist; other-
wise the technical inefficiency score is shown if D ( x, y, b|g ) > 0.
3.1.2 Time series productivity measurement

Fujii et al. (2015) combined the Luenberger Productivity Index with the WRDDM
and derived the following:
1 t +1 t t t
TFPt t +1 =
2
{ ( )
D ( xk , yk ,bk , ) − D t +1 ( xkt+, 1 ykt+, 1bkt+, 1 ) + D t ( xkt , ykt ,bkt ,
(4)
}
− D t ( xkt+, 1 ykt+, 1bkt+, 1 ) .
In this model, xt stands for input for year t, xt+1 represents input for year t+1,
yt is the desirable output for year t, and yt+1 is the desirable output for year t+1.
Undesirable output for year t isrepresented by bt , and the undesirable output for
year t+1 represented by bt+1. D t +1 ( xkt+, 1 ykt+, 1bkt+, 1 ) represents the inefficiency score

of year t+1 based on a frontier curve in year t+1. Similarly, D t ( xkt , ykt ,bkt , is the ( )
inefficiency score of year t based on frontier curve in year t. The TFP score for a
particular country indicates the productivity change compared to the benchmark
year. Then, TFP decomposed by the inefficiency score of each variable’s contribu-
tion effect for inefficiency is:
1 N 1 M
1 L k  t t t
max  ∑βnk + ∑β k
+ ∑ βl  = Dx ( xk , yk ,bk , ) + Dy ( xkt , ykt ,bkt , )
 N n=1 M m=1
m
L l=1  
(5)
+ Db ( xkt , ykt ,bkt , )
t t t
where Dx ( xk , yk ,bk , ) is the contribution effect of input variable for the inefficiency

score, Dy ( xkt , ykt ,bkt , ) represents the
contribution effect of desirable output variable
for the inefficiency score, and Db ( xkt , ykt ,bkt , ) represents the contribution effect of
undesirable output variable for the inefficiency score. From the previous equa-
tions, we can derive the following formula:

D t ( xkt , ykt ,bkt , | g ) = Dkt , ( xkt , ykt ,bkt , ) + Dyt , ( xkt , ykt ,bkt , ) + Dbt , ( xkt , ykt ,bkt , ) (6)
From the previous equations, we can derive the following formulas:
1 t +1 t t t
TFPt t, x+1 =
2
{
Dx ( x , y , b ) − Dxt+1 ( x t +1 , y t +1 , bt +1 ) + Dxt ( x t , y t , bt )
(7)
− Dxt ( x t +1 , y t +1 , bt +1 ) }
1 t +1 t t t
TFPt t, +y 1 = Dy ( x , y , b ) − Dyt+1 ( x t +1 , y t +1 , bt +1 ) + Dyt ( x t , y t , bt )
{
2 (8)

− Dyt ( x t +1 , y t +1 , bt +1 ) }
1 t +1 t t t
TFPt t,b+1 =
2
{
Db ( x , y , b ) − Dbt+1 ( x t +1 , y t +1 , bt +1 ) + Dbt ( x t , y t , bt )
(9)

− Dbt ( x t +1 , y t +1 , bt +1 ) }
where TFPt t, +1
x assures the contribution effect of the input variable to TFP
change. Increasing TFPt t, +1
x
has two possible reasons; first, the share effect of
input variables is enhanced so that TFPt t, +1 x growth represents how much TFP
increases due to input optimization. In the second possibility, inefficiency of
input decreases more than desirable and undesirable output inefficiency. In
the second possibility, TFPt t, +1
x
will increase and TFPt t, +1
y or TFPt ,b will decrease.
t +1
This is because the model assesses each variable’s share effect by maximiz-
ing relative inefficiency among input, desirable output and undesirable output.
y measures the contribution effect of the desirable output variable to TFP
TFPt t, +1
change. TFPt t,b+1 measures the contribution effect of the undesirable output vari-
able to TFP change.
3.2 Data
Compared to previous studies on TFP computation, we examine wealth as a
productive asset of each country as input, gross domestic product (GDP) as
the desirable output, and carbon damages as an undesirable output. For the
wealth sources, we refer to the Inclusive Wealth Report (IWR) published in
2014 by UNU-IHDP and UNEP. This report provides quantitative information
on 140 countries of three main productive assets as a source of human well-
being, which are human, produced, and natural capital. The basket of assets
is calculated by using the marginal contribution of each capital type to social
welfare, which is represented by the social (or shadow) price of the assets under
evaluation.
Produced capital (PC), the capital type that has the most exhaustive (and reli-
able) data, consists of equipment, machinery, roads, and another physical capital.
Human capital is a function of education attainment and the additional compensa-
tion over time as developed by Arrow et al. (2012). Natural capital wealth consists
of fossil fuels (oil, natural gas, coal), minerals (bauxite, nickel, copper, phosphate,
gold, silver, iron, tin, lead, zinc), forest resources (timber and non-timber), and
agricultural land (cropland and pastureland). Capital wealth in this calculation is
based on millions of constant 2005 US dollars.
Emissions, as bad outputs that are typical in extractive natural capitals such as
forest, energy, and agriculture, are expected to influence the provision of a coun-
try’s performance. Nordhaus and Boyer (2000) have measured carbon damages
in the following areas: agriculture, sea-level rise, other vulnerable market sectors
(energy systems), health, and non-market amenity impacts. Thus, these impacts
can have a real effect on the efficiency of outputs to a point where co-production
appeared in the production. In this study, damages accruing from carbon emis-
sions were calculated based on global emission damages and then proportion-
ally distributed to the countries. We took carbon emissions stemming from fuel
consumption, cement, and deforestation. The marginal damage costs of carbon
dioxide emissions are estimated at USD 50, based on a study by Tol, 2009. Then,
by using a formula developed by Nordhaus and Boyer (2000), we calculated the
distribution of the damages that different regions of the global economy will suf-
fer as a whole.
4 Empirical result
4.1 Productivity growth
4.1.1 TFP change and contribution effect

Productivity change is based on how many economic goods and services are
produced given specific inputs and the production of negative impacts simulta-
neously. We treat each country as it produces capital, human capital and natu-
ral capital as input to produce GDP as desirable outputs and carbon damages as
undesirable output. We use 1995 as the baseline year, so TFP changes its score
equal to zero in 1995. In this method, a positive TFP score reflects productivity
improvement.
Figure 11.1 shows the accumulative changes of TFP indicators during the study
period. The line chart in Figure 11.1 represents the average score of TFP indica-
tors for 140 countries from 1995 to 2010. The bar chart, equal to those in the line
chart, indicates the accumulative contribution effect of the inputs/outputs factor
to TFP change. By interpreting Figure 11.1, we differentiate each contributor of
inputs, desirable and undesirable outputs for TFP change.
0.02
0.015
0.01
0.005
0
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
–0.005
–0.01
–0.015
–0.02
–0.025
–0.03
GDP Human Capital Produced Capital
Natural Capital Carbon Damages Productivity Index
Figure 11.1 Average TFP and contribution effect of inputs/outputs for 140 countries
Based on Figure 11.1, the countries increase TFP during 1995 to 1997 with the
main contributors of GDP and natural capital. Growing values of indicators for
natural capital productivity over time signal improvements in resource use, which
implies that countries utilized natural capital more efficiently.
In 1998, the economic crisis that hit several countries has a significant impact
on reducing TFP in the next year. After the 1998 crisis, TFP increases until the
2008 economic crisis that had an even deeper magnitude.
During this period, carbon damages as an undesirable output have a sig-
nificant effect on TFP changes, in addition to GDP and natural capital. This
represents countries’ ability to offset the negative impact of emission that has
increased along with economic growth. In contrast, the economic crisis sig-
nificantly reduced their capability to reduce emissions, and the resulting carbon
damages are the main contributor of negative TFP changes in 2009 as an impact
of the crisis.
Our results show that, on average, TFP values have been increasing in many
countries, and 109 countries have a positive TFP score. Using this method, some
emerging countries, such as Brazil, Russia, India, China, and South Africa, and
developed countries, such as the United Kingdom and Canada, have high pro-
ductivity scores. This reflects improvement in some countries in regard to utiliz-
ing their resources and reducing undesirable output. However, countries such as
Indonesia and Saudi Arabia show decreasing TFP values over time. This negative
value shows that these countries need to improve their resource use efficiency and
manage the negative impact of emissions.
4.2 Decomposition of inefficiency scores
4.2.1 Inefficiency score of each component change

Using WRDDM, we open the black box and examine the individual drivers behind
country inefficiency. The decomposition of aggregated inefficiency scores during
1995–2010 into produced capital, human capital, and natural capital as inputs,
GDP as desirable input, and carbon damages as undesirable output is shown in
Table 11.1.
A zero value of inefficiency score implies that input/output is fully efficient in
comparison to another countries. For example, in 1995, the 0.577 score is referred
to as the value of overall inefficiency of 140 countries in that particular year. This
value is the weighted sum of corresponding figures of produced, human and natu-
ral capital, GDP, and carbon damages. It implies that the GDP can be expanded
by more than 98.2% compared to its existing level while simultaneously reducing
77.2% of the carbon damages, 20.5% of produced capital, 30.9% of human capi-
tal, and 62.0% of natural capital. The decomposition of the overall inefficiency
score into each input/output inefficiency score is effective for obtaining each com-
ponent of the overall inefficiency score of the countries.
Table 11.1 Aggregate (overall) technical inefficiency
year Input Output
Produced Human Natural GDP Carbon

capital capital capital damages
1995 0.205 0.309 0.620 0.982 0.772
1996 0.204 0.286 0.603 0.927 0.766
1997 0.188 0.301 0.605 0.848 0.782
1998 0.178 0.285 0.598 0.832 0.773
1999 0.168 0.268 0.601 0.833 0.77
2000 0.163 0.279 0.600 0.805 0.779
2001 0.162 0.288 0.597 0.800 0.778
2002 0.120 0.283 0.613 0.730 0.786
2003 0.120 0.277 0.604 0.699 0.778
2004 0.121 0.28 0.615 0.644 0.776
2005 0.131 0.284 0.618 0.607 0.779
2006 0.149 0.281 0.611 0.558 0.763
2007 0.157 0.286 0.622 0.544 0.768
2008 0.134 0.273 0.602 0.548 0.762
2009 0.12 0.265 0.599 0.562 0.745
2010 0.128 0.275 0.641 0.561 0.782
Average 0.153 0.282 0.609 0.717 0.772
Note: Total values consider a current average of heterogeneous countries but do not capture
comparisons between the same development stage, which should be different years over countries.
For input, we observe that countries report particularly different inefficiency
scores for each input. Most of the countries have high efficiencies of inputs in
terms of produced and human capital and show improvement during the study
period. Furthermore, the inefficiency levels of produced capital gradually decrease
from 0.205 in 1995 to 0.128 in 2010. Human capital inefficiency score trends also
show a decreasing pattern from 0.309 in 1995 to 0.275 in 2010. For human and
produced capital, the inefficiency levels remain low, which indicates a positive
sign that countries utilize this capital in more efficient ways.
Infrastructure capital, a part of produced capital, as the fundamental facilities
and systems serving a country, such as transportation and communication systems,
water supply, agriculture production, power plants, and schools, is important for
the direct functioning of the economy and quality of life. Produced capital is easy
to understand from a country viewpoint because reliable access to these services
facilitates economic production and consumption. This might be the reason why
many countries tend to have low inefficiency scores for this input.
Meanwhile, the inefficiency score of natural capital is relatively high and
constant, at 0.620 in 1995 and 0.641 in 2010. On average, there is no increasing
efficiency in natural capital utilization. Natural capital, such as forest, land, and
minerals, are hard to measure and generally produce high inefficiency scores.
Many natural capitals have vital services but are often intangible to people.
These values are not readily captured in markets, so their contribution to the
economy and livelihoods is not recorded. These services are taken for granted,
and a country does not know what it would cost the economy if these services
were lost.
In regard to outputs, we see a different change of the inefficiency score. In the
beginning of our study period, the inefficiency scores for GDP are high for car-
bon damage. The trend of inefficiency score gradually decreases over the exam-
ined period. The trend for the total inefficiency score shows an improvement.
The inefficiency score for GDP improves during 1995–2007. After this period, the
inefficiency score for GDP is relatively stagnant. Carbon damages, as allocated
by the global damage to each country depending on the potential effect of global
warming in their economies, show a different pattern of inefficiency scores, as the
undesirable output is stagnant during the examining period.
Even though there are differences in the overall means, we notice that the
changes of inefficiency score for GDP as desirable outputs and human-produced
capital as inputs are relatively improved during the study period. The results
show that the goods and input efficiency are better than the undesirable output
efficiency.
4.2.2 Technical inefficiency by income level

The average overall inefficiency of high-income and upper middle-income coun-
tries was generally lower than low-income and lower middle-income countries
during the study period. This means that the comprehensive efficiency of the
higher-income countries is generally better than lower-income countries. It is
Table 11.2 Technical inefficiency by income level
Income level Produced Human Natural GDP Carbon Average

capital capital capital damages by income
High income 0.075 0.193 0.582 0.352 0.643 0.369
Upper middle income 0.148 0.273 0.679 0.445 0.843 0.478
Lower middle income 0.209 0.344 0.622 0.92 0.802 0.579
Lower income 0.219 0.35 0.520 1.058 0.836 0.596
appealing to check whether the overall inefficiency of each input-output has the
same pattern.
Income level group results of Table 11.2 show input, good output, and bad out-
put inefficiency, and the high-income countries have the lowest results. There is a
relationship between the average inefficiency score for human capital, produced
capital, GDP, emissions, and income level. Higher income level has a lower inef-
ficiency score. This implies that the higher income countries are generally more
efficient than other lower income levels for these components.
Natural capital input has a different pattern compared to other components.
Lower income level has the lowest inefficiency score, which may be due to their
utilization of the local natural capital. In the energy sector, for instance, local fuels
such as firewood, dung, crop residues, and other biomass-based products still con-
tribute a major share of energy requirements for rural populations in low-income
countries. Taking Pakistan as a study case, Filmer and Pritchett (2002) noted the
importance of the local natural resource base in rural life.
There is a large gap of the inefficiency score for GDP between the high- and
upper-middle-income and the lower-middle- and low-income categories. Low-
and middle-income countries can improve their productivity by increasing GDP.
Therefore, natural capital and carbon damages are still quite high compared to
other input/output variables in high-income groups. This means that these groups
have room for improvement in their efficiencies for these variables.
4.2.3 Output inefficiency by country

We divided the aggregated technical inefficiency score for each country into five
categories. In terms of GDP as a good output, as shown in Map 11.1, each country
has a different result. On average, higher income countries have a lower ineffi-
ciency score. Included in the low aggregated score lower than 0.2 are high-income
countries, such as Japan, the United States, Canada, Belgium, United Kingdom,
Italy, Luxembourg, and Singapore. Based on the contribution of each output com-
ponent, we found that the GDP in this group is almost zero for all countries. It
implies that besides the lower score for almost all inputs/outputs, these countries
obtained their good performance from improving their economy as represented
by GDP.
Inefficiency level
Very high
High
Medium high
Medium low
Low
No data
Map 11.1 Average GDP inefficiency score

However, GDP was the highest contributor for the high-aggregate inefficiency
score for countries with an inefficiency score of more than 0.6. Included in this
group are low-to low middle-income countries, such as Niger, Tajikistan, Afghan-
istan, Democratic Republic of Congo, Liberia, Ukraine, Paraguay, Bolivia, Leso-
tho, Iraq, Armenia, Republic of Moldova, Swaziland, Guyana, and many countries
in the African region.
In the African case, high inefficiency GDP scores related to their stagnant eco-
nomic growth, as reported in The Africa Competitiveness Report (2015). The
report noted that agriculture’s share of GDP is declining, so Africa continues to
be largely agrarian, with an economy underpinned by resource-driven growth.
Manufacturing is stagnating in this region. In contrast, several countries in other
regions, such as Asia, have a path out of their poverty and boost their economy by
increasing agricultural productivity along with growth in manufacturing. Haus-
mann and Hidalgo (2011) found a linkage between manufacturing, economic
complexity and prosperity, implying that more advanced manufacturing capabili-
ties and more advanced product sets lead to higher prosperity.
In terms of the carbon damages inefficiency score, as shown in Map 11.2, only
a few countries have superior performance for carbon damages. The United States
and Japan had the lowest inefficiency scores for this undesirable output. These
results reflect that both countries can be considered less susceptible to climate
change compared to many other countries. This is the result of their relatively tem-
perate climates and less dependence on climate for their economies. In contrast,
India and several low-income countries in Africa are on the other extreme. The
impact on India comes from its extreme vulnerability to climatic shifts because of
the importance of monsoons on their agriculture. For Africa, much of their expo-
sure comes from the potential health impacts of global warming.
4.2.4 Capital inefficiency by country

On the input side, on average, countries have a low score for produced and human
capital. Produced capital is the lowest inefficiency score for almost all countries;
136 of 140 countries have an inefficiency score of less than 0.2. For human capi-
tal, 71 of 140 countries scored less than 0.2.
The results indicate that natural capital remains a significant burden for most
countries’ performance, as shown in Map 11.3. Several oil producing countries,
such as United Arab Emirates, Iran, Iraq, Venezuela, Nigeria, and Bahrain, belong
to a group of countries with an aggregated score of higher than 0.6. Another type
of this group is in depleting forest areas, such as Indonesia and Brazil.
4.3 Country examples: cases of China and Canada

We could obtain lessons learned from several countries that have a large dis-
crepancy from the beginning to the end of study period. In this case, we used
China and Canada as representative cases for developing and developed coun-
tries, respectively.
Inefficiency level
Very high
High
Medium high
Medium low
Low
No data
Map 11.2 Average carbon damages inefficiency score

Inefficiency level
Very high
High
Medium high
Medium low
Low
No data
Map 11.3 Natural capital inefficiency score

China is an emerging economic superpower, with one-quarter of the world’s
population. Hofman et al. (2007) argued that China’s rapid growth was related to
massive investment in infrastructure and human capital since the Cultural Revolu-
tion began. Shi and Huang (2014) found that extensive infrastructure investment
in China after the 1997 East Asian economic shock efficiently improved infra-
structure development in most provinces in China.
Green et al. (2007) concluded that China’s successes in basic education, espe-
cially before and during the reform, have contributed significantly to economic
growth thereafter. This conclusion is based on Barro and Lee’s (2001) data that
showed the growth of junior, secondary, technical, and higher education from
1978. This human capital development filled the demand for skills in high-end
manufacturing and service sectors in China.
By incorporating a measure of human capital stock in the production function
for 1952–1999, Wang and Yao (2003) concluded that the relative share of the
accumulation of human capital in China was quite rapid, and it contributed sig-
nificantly to growth and welfare. The combination of human and produced capital
in China has supported its growth and efficient increase of other components.
Analyzing technical efficiency using the inclusive wealth framework that con-
siders natural capital and other conventional inputs, such as input and carbon
damages as undesirable outputs, we found that China shifted their input signifi-
cantly, which contributed significantly to their growth. China has high inefficiency
for carbon damages as an undesirable output, and human and natural capital as
an input in the beginning of study period; however, they reduced it significantly
by the end of the study period. Decreasing all the inputs, desirable outputs, and
undesirable outputs to zero by the end of study period means that this country has
been utilizing their input/output optimally. Urban (2015) argues that China, as the
world’s largest consumer of natural resources, must become a leader in environ-
mental innovation for sustainable development.
Based on the wealth approach, in 1995, human capital in China contributed
51% of total input, while they produced a capital and natural capital share of 12%
and 37%, respectively. At the end of study period, even though human capital was
still the highest contributor for input, the share decreased to 42%. Natural capital
also decreased to 21%. In contrast, the share of produced capital increased from
12% to 37%.
Figure 11.2 shows China’s inefficiency source decomposition for the whole
study period. In the beginning of the study period, China’s average inefficiency
score is 0.562, which was composed of human capital (0.33), natural capital
(0.461), and carbon damages (0.822). In the end, all input-output scores are zero,
showing China’s capacity to utilize its inputs and outputs in an efficient way.
In this study, energy becomes the link that allows decreasing natural capital
and carbon to damage the inefficiency score. During the 1980 to 2001 period,
China was able to limit energy demand growth to less than half of GDP growth.
Facing increasing energy intensity after this period, beginning in 2006, China cre-
ated or re-created a policy apparatus to design and implement energy efficiency
throughout the economy to keep their energy intensity on track (Zhou et al., 2010).
1
0.9
0.8
Inefficiency score
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
95
96
97
98
99
00
01
02
03
04
05
06
07
08
09
10
19
19
19
19
19
20
20
20
20
20
20
20
20
20
20
20
GDP Carbon Damages
Figure 11.2 Inefficiency score change in China
China’s ability to limit energy demand growth does not only utilize natural capital
as an input in a more efficient way but also has an impact on reducing emissions,
as studied by Richerzhagen and Scholz (2008). In their study, potential emissions
reductions are mainly a by-product of measures embedded in energy and transport
policies aimed at cutting energy costs and increasing energy security in China.
To reduce the high environmental impact and achieve a transition to sustainable
development, it is imperative that China invests in and becomes a leader in envi-
ronmental innovation.
Increasing education from 10.77 years in 1990 to 12.08 years in 2010, Cana-
da’s human capital shifted from 50% to 69% of the total input. During the study
period, natural capital shares decreased from 34% to 33%, while the produced
capital share increased from 16% to 28%. Figure 11.3 displays Canada’s inef-
ficiency source decomposition for the whole study period. At the beginning of the
study period, the human capital inefficiency score was 0.10, natural capital was
0.29, and carbon damages was 0.45.
In contrast to China, with a large gap between market and non-market out-
put, Canada’s discrepancy is lower. Even though Canada ratified the Kyoto Pro-
tocol in 2002, between the 1990 and 2008, Canada’s greenhouse gases (GHG)
increased. However, their technical inefficiency score decreased, meaning that
Canada could reduce the potential effect of global warming in their economies.
Related to this, Lantz and Feng (2006) argued that technological and/or economic
structural changes have tended to favor CO2 emissions-intensive production pro-
cesses in Canada. Vandenbussche et al. (2006) argued that for enhanced economic
growth, countries need to invest primarily in higher education to push forward the
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
03
04
05
06
07
08
09
10
95
96
97
98
99
00
01
02
20
20
20
20
20
20
20
19
19
19
19
19
20
20
20
20

GDP Carbon Damages
Figure 11.3 Inefficiency score change in Canada
technological frontier. Canada is one of the countries with the highest education
level, as their adults aged 25–64 have attained at least upper secondary education,
and they have been able to attain benefits from their human capital.
5 Conclusion
This study combined the concept of inclusive wealth, including natural capital,
as input and incorporated carbon damages as undesirable output in productiv-
ity measurements. Using the applied developed directional distance function, we
analyzed the contribution effect of each input and output to efficiency and pro-
ductivity change as an important addition to the current research of sustainability.
We found that GDP, natural capital, and carbon damages are the main contribu-
tors to our productivity change for 140 countries over 1995–2010. Each country
differs with respect to effective utilization of their productive assets and reducing
the negative impact of climate change. Natural capital contributed positively to
the productivity change from a high-income countries group. This finding showed
that these countries used capital more efficiently, with appropriate productivity
changes and future-oriented stock consumption schemes. Carbon damages con-
tributed positively to the productivity along with the economic ability to offset the
negative impact of carbon emissions.
Regarding efficiency measures, on average, countries have high efficiencies of
produced and human capital and they showed improvement in our study period.
Countries, on average, have high inefficiency scores for natural capital and car-
bon damages. Natural capital has remained largely hidden to policymakers due
to the limitations of traditional economic indices. It becomes a significant bur-
den to measure most countries’ performance, particularly for countries that are
oil producers or with depleting forests. As a result, these countries have failed
to adequately invest in rebuilding this capital, despite evidence that returns on
investment in natural capital far outweigh investment in physical capital, such as
infrastructure. Investments in natural capital, particularly agricultural land and
forest, can produce several dividends: increasing inclusive wealth and productiv-
ity, improving agricultural resiliency and food security, and mitigating climate
change, according to UNU-IHDP and UNEP (2014).
China and Canada, representing countries in different development stages,
decreased all the inputs, desirable outputs, and undesirable outputs to zero by the
end of the study period. This implies that these countries have better utilized their
input/output. China’s attainment in basic education, massive investment in infra-
structure, and effective energy policy have contributed significantly to technical
inefficiency reduction. In contrast to China, which has a big gap between market
and non-market output, Canada’s discrepancy is lower. Technological and/or eco-
nomic structural changes have tended to favor CO2 emissions-intensive produc-
tion processes in Canada.
Future studies are needed to extend the data, since this research has severe
limitations of data, and items of natural capital that were included were limited to
agricultural land, forests, fossil fuels, and minerals. Further elaboration of data,
such as more consideration of health in human capital, ecosystem services and
natural capitals, would enhance our understanding of countries’ performance and
sustainability.
Bibliography
Akao, K. I., and Managi, S. 2007. Feasibility and optimality of sustainable growth under
materials balance. Journal of Economic Dynamics and Control, 31(12), 3778–3790.
http://doi.org/10.1016/j.jedc.2007.01.013
Arrow, K. J., Dasgupta, P., Goulder, L., Daily, G., Ehrlich, P., Heal, G., and Walker, B. 2004.
Are we consuming too much? The Journal of Economic Perspectives, 18(3), 147–172.
http://doi.org/10.1111/j.1523-1739.2007.00770.x
Arrow, K. J., Dasgupta, P., Goulder, L., Mumford, K. J., and Oleson, K. 2012. Sustainabil-
ity and the measurement of wealth. Environment and Development Economics, 17(03),
317–353. doi: 10.1017/s1355770x12000137.
Azad, M. A. S., and Ancev, T. 2014. Measuring environmental efficiency of agricultural
water use: A Luenberger environmental indicator. Journal of Environmental Manage-
ment, 145, 314–320.
Balk, B. M., Färe, R., Grosskopf, S., and Margaritis, D. 2008. Exact relations between
Luenberger productivity indicators and Malmquist productivity indexes. Economic The-
ory, 35(1), 187–190. http://doi.org/10.1007/s00199-007-0228-5
Barro, R., and Lee, J. 2001. International data on educational attainment: Updates and impli-
cations. Oxford Economic Papers, 53(3), 541–563. http://doi.org/10.1093/oep/53.3.541
Barros, C. P., Managi, S., and Matousek, R. 2012. The technical efficiency of the Japanese
banks: Non-radial directional performance measurement with undesirable output.
Omega, 40(1), 1–8.
Brandt, N., Schreyer, P., and Zipperer, V. 2013. Productivity measurement with natural
capital. OECD Economics Department Working Paper, No. 1092, OECD Publishing.
http://dx.doi.org/10.1787/5k3xnhsz0vtg-en
Callan, S. J., and Thomas, J. M. 2000. Environmental Economics and Management: The-
ory, Policy, and Applications. South-Western Mason: Dryden Press.
Chen, P.-C., Yu, M.-M., Chang, C.-C., Hsu, S.-H., and Managi, S. 2015. Non-radial direc-
tional performance measurement with undesirable outputs: An application to OECD and
non-OECD countries. International Journal of Information Technology & Decision
Making, 14(3), 481–520.
Common, M., and Stagl, S. 2005. Ecological Economics: An Introduction. Cambridge:
Cambridge University Press.
Costanza, R., de Groot, R., Sutton, P., van der Ploeg, S., Anderson, S. J., Kubiszewski, I.,
and Turner, R. K. 2014a. Changes in the global value of ecosystem services. Global Envi-
ronmental Change, 26(1), 152–158. http://doi.org/10.1016/j.gloenvcha.2014.04.002
Costanza, R., Kubiszewski, I., Giovannini, E., Lovins, H., McGlade, J., Pickett, K. E., and
Wilkinson, R. 2014b. Development: Time to leave GDP behind. Nature, 505(7483),
283–285. DOI: 10.1038/505283a
Dasgupta, P. 2009. The place of nature in economic development. Working Paper, South
Asian Network for Development and Environmental Economics (SANDEE), Kath-
mandu, Nepal.
Dasgupta, P., Duraiappah, A., Managi, S., Barbier, E., Collins, R., Fraumeni, B., . . . Mum-
ford, K. 2015. How to measure sustainable progress. Science, 350(6262), 748. http://doi.
org/10.1126/science.350.6262.748
Ehrlich, P. R., and Goulder, L. H. 2007. Is current consumption excessive? A general frame-
work and some indications for the United States. Conservation Biology, 21(5), 1145–
1154. http://doi.org/10.1111/j.1523-1739.2007.00779.x
Filmer, D., and Pritchett, L. H. 2002. Environmental degradation and the demand for chil-
dren: Searching for the vicious circle in Pakistan. Environment and Development Eco-
nomics, 7(1), 123. http://doi.org/10.1017/S1355770X02000074
Frye-Levine, L. A. 2012. Sustainability through design science: Re-imagining option
spaces beyond eco-efficiency. Sustainable Development, 20(3), 166–179. http://doi.
org/10.1002/sd.1533
Fujii, H., Cao, J., and Managi, S. 2015. Decomposition of productivity considering multi-
environmental pollutants in Chinese industrial sector. Review of Development Econom-
ics, 19(1), 75–84. http://doi.org/10.1111/rode.12123
Green, A., Little, A., Kamat, S. G., Oketch, M., and Vickers, E. 2007. Education and devel-
opment in a global era: Strategies for ‘successful globalization’. Retrieved from http://
eprints.ioe.ac.uk/4473/
Guerry, A. D., Polasky, S., Lubchenco, J., Chaplin-Kramer, R., Daily, G. C., Griffin, R., and
Vira, B. 2015. Natural capital and ecosystem services informing decisions: From prom-
ise to practice. Proceedings of the National Academy of Sciences, 112(24), 7348–7355.
http://doi.org/10.1073/pnas.1503751112
Haines-Young, R. H., and Potschin, M. B. 2009. The links between biodiversity, ecosystem
services and human well-being. In Ecosystems Ecology: A New Synthesis, Raffaelli, D.
G. and C.L.J. Frid (Eds.). Cambridge: Cambridge University Press, 110–139. DOI:
10.1017/CBO9780511750458.007
Hamilton, K., and Clemens, M. 1999. Genuine savings rates in developing countries. World
Bank Economic Review, 13(2), 333–356.
Hanouz, M. D., Brown, C., Galvan, C., Di Batista, A., Speakman, J., Hoffman, B., . . .
Heinrigs, P. 2015. Africa Competitiveness Report 2015. Geneva: World Economic
Forum. Retrieved from www3.weforum.org/docs/WEF_ACR_2015/Africa_
Competitiveness_Report_2015.pdf
Harlow, J., Golub, A., and Allenby, B. 2013. A review of utopian themes in sustainable
development discourse. Sustainable Development, 21(4), 270–280. http://doi.
org/10.1002/sd.522
Hausmann, R., Hidalgo, C. A. et al. 2011. The atlas of economic complexity: Mapping
paths to prosperity. Retrieved from www.cid.harvard.edu/documents/complexityatlas.
pdf
Hofman, B., Zhao, M., and Ishihara, Y. 2007. Asian development strategies: China and
Indonesia compared. Bulletin of Indonesian Economic Studies, 43(2), 171–200. http://
doi.org/10.1080/00074910701408057
Imran, S., Alam, K., and Beaumont, N. 2014. Reinterpreting the definition of sustainable
development for a more ecocentric reorientation. Sustainable Development, 22(2),
134–144. http://doi.org/10.1002/sd.537
Jorgenson, A. K., and Dietz, T. 2015. Economic growth does not reduce the ecological
intensity of human well-being. Sustainability Science, 10(1), 149. DOI: 10.1007/
s11625-014-0264-6
Kenneth, J. A., Dasgupta, P., Goulder, L. H., Mumford, K. J., and Oleson, K. 2012. Sustain-
ability and the measurement of wealth. Environment and Development Economics, 17,
317–353. DOI: 10.1017/S1355770X12000137
Kumar, S., and Managi, S. 2010. Environment and productivities in developed and devel-
oping countries: The case of carbon dioxide and sulfur dioxide. Journal of Environmen-
tal Management, 91(7), 1580–1592. http://doi.org/10.1016/j.jenvman.2010.03.003
Lantz, V., and Feng, Q. 2006. Assessing income, population, and technology impacts on
CO2 emissions in Canada: Where’s the EKC? Ecological Economics, 57(2), 229–238.
http://doi.org/10.1016/j.ecolecon.2005.04.006
Mahlberg, B., and Sahoo, B. K. 2011. Radial and non-radial decompositions of Luenberger
productivity indicator with an illustrative application. International Journal of Produc-
tion Economics, 131(2), 721–726.
Managi, S. 2003. Luenberger and Malmquist productivity indices in Japan, 1955–1995.
Applied Economics Letters 10(9), 581–584.
Mukherjee, S., and Chakraborty, D. 2013. Is environmental sustainability influenced by
socioeconomic and sociopolitical factors? Cross-country empirical evidence. Sustain-
able Development, 21(6), 353–371. http://doi.org/10.1002/sd.502
Nordhaus, W. D., and Boyer, J. 2000. Warming the World: Economic Models of Global
Warming. Cambridge, MA: MIT Press.
Pearce, D. W., and Atkinson, G. D. 1993. Capital theory and the measurement of sustainable
development: An indicator of ‘weak’ sustainability. Ecological Economics, 8(2), 103–108.
Quental, N., Lourenço, J. M., and Da Silva, F. N. 2011. Sustainable development policy:
Goals, targets and political cycles. Sustainable Development, 19(1), 15–29. http://doi.
org/10.1002/sd.416
Richerzhagen, C., and Scholz, I. 2008. China’s capacities for mitigating climate change.
World Development, 36(2), 308–324. http://doi.org/10.1016/j.worlddev.2007.06.010
Schoenaker, N., Hoekstra, R., and Smits, J. P. 2015. Comparison of measurement systems
for sustainable development at the national level. Sustainable Development, 23(5),
285–300. http://doi.org/10.1002/sd.1585
Shi, H., and Huang, S. 2014. How much infrastructure is too much? A new approach and
evidence from China. World Development, 56, 272–286. http://doi.org/10.1016/j.
worlddev.2013.11.009
Sneddon, C., Howarth, R. B., and Norgaard, R. B. 2006. Sustainable development in a post-
Brundtland world. Ecological Economics, 57(2), 253–268. http://doi.org/10.1016/j.
ecolecon.2005.04.013
UNU-IHDP and UNEP. 2012. Inclusive Wealth Report 2012: Measuring Progress toward
Sustainability. Cambridge: Cambridge University Press.
UNU-IHDP and UNEP. 2014. Inclusive Wealth Report 2014: Measuring Progress toward
Sustainability. Cambridge: Cambridge University Press.
Urban, F. 2015. Environmental innovation for sustainable development: The role of China.
Sustainable Development, 23(4), 203–205. http://doi.org/10.1002/sd.1587
Tol, R.S.J. 2009. The economic effects of climate change. The Journal of Economic Per-
spective, 23(2), 29–51.
Vandenbussche, J., Aghion, P., and Meghir, C. 2006. Growth, distance to frontier and com-
position of human capital. Journal of Economic Growth, 11(2), 97–127. http://doi.
org/10.1007/s10887-006-9002-y
Wang, Y., and Yao, Y. 2003. Sources of China’s economic growth 1952–1999: Incorporat-
ing human capital accumulation. China Economic Review. http://doi.org/10.1016/
S1043-951X(02)00084-6
Xepapadeas, A., and Vouvaki, D. 2009. Total factor productivity growth when factors of
production generate environmental externalities. Fondazione Eni Enrico Mattei Working
Papers, 281, 1–31.
Zhou, N., Levine, M. D., and Price, L. 2010. Overview of current energy-efficiency policies
in China. Energy Policy, 38(11), 6439–6452. http://doi.org/10.1016/j.enpol.2009.08.015
Zhou, P., Poh, K. L., and Ang, B. W. 2007. A non-radial DEA approaches to measuring
environmental performance. European Journal of Operational Research, 178(1), 1–9.
http://doi.org/10.1016/j.ejor.2006.04.038
12 Social cost-benefit analysis
of decontamination after the
Fukushima Dai-ichi Nuclear
Power Plant accident
Kazunobu Okubo, Rintaro Yamaguchi and
Shunsuke Managi
1 Introduction
A major earthquake occurred in the northeast of Japan on March 11, 2011. A
tsunami hit the area in which the Fukushima Dai-ichi Nuclear Power Plant was
located, causing a loss of the electric power required to control the reactor and
consequently triggering a hydrogen explosion. The explosion dispersed radioac-
tive substances, and the wind scattered them over a broad area. The substances are
mainly iodine and two kinds of cesium (Cs-134 and Cs-137). Iodine’s half-life is
only 8 days, so it rapidly lost radioactivity. The problem is that cesium, which has
a much longer half-life, may have contaminated land, forests, roads, buildings and
many other structures especially in the coastal area of Fukushima.
Immediately after the power plant accident, the Japanese government issued
evacuation orders to 11 municipalities in the area,1 began to monitor the air dose
rate in the affected areas, and laid out decontamination plans. After some dem-
onstration projects, the government initiated full-scale decontamination in areas
with a relatively low air dose rate, where the workers’ safety could be assured.
As this chapter goes to press, 6 years have passed since the full-scale decon-
tamination began. The initially planned decontamination is complete except for
some follow-up work. This allowed some evacuation orders to be removed. How-
ever, an area remains where the air dose rate is so high that evacuation orders are
still in effect. It would therefore be useful to perform an intermediate cost-benefit
analysis for the finished work and an ex ante analysis for the unfinished work.
To conduct a cost-benefit analysis, we must identify the scope for the cost and
the benefit of the project. In this chapter, we define cost and benefit, respectively,
as the monetary expenditure by government2 and the value of recovered capital
services. The latter is broken down into those services of produced capital (roads
and housings) and natural capital (agricultural lands and forests) that are recov-
ered by decontamination. Contamination by radioactive substances deprives us of
part or all of these useful services. The recovered services of produced and natural
capital correspond to the shadow price change (Yamaguchi et al. 2017).
Social cost-benefit analysis requires comparison with a counterfactual (UNU-
IHDP and UNEP 2012). In our context, it is relevant to note that even if there is
no human intervention, the radioactivity level is reduced by natural decay and
Decontamination after Fukushima Dai-ichi 311
“weathering,” that is, natural decontamination by rains and winds. Thus, a coun-
terfactual would be to choose to just wait until the air dose rate naturally comes
to a safe level through natural decay and weathering. This “natural decontamina-
tion” would enable us to resume the use of capital services in any case. Alterna-
tively, with human intervention, we can advance the timing of this resumption.
The difference in the timing of resuming capital services results in the difference
in the net present value of total available services with and without human inter-
vention. Since the amount of physical volume of produced and natural capital has
not been affected, this difference directly translates into the difference in shadow
prices of capital assets between actual (decontamination) and counterfactual (no
decontamination) scenarios (Yamaguchi et al. 2017). These values of the recov-
ered capital services due to decontamination by humans can be regarded as the
project’s benefits.
Adopting the preceding definition of benefits, we conduct a cost-benefit analysis
of a decontamination project. We can forecast the amounts of capital services that
will be generated in each year, provided that the amounts of capital stocks are
constant over time. However, in the case of costs (government expenditures), we
only know the timing and amount of past expenditure.3 Thus, our strategy is as
follows: if the costs that have been already paid are larger than the total benefit,
which includes the values of capital services recovered through both past and
future decontamination, we can affirmatively say that the total cost is also larger
than the total benefit. This is somewhat similar to a break-even analysis (e.g.,
Leigh and Blakely 2016).
Yasutaka et al. (2013) estimated the costs of three categories as shown:
1 Removal and temporary storage of contaminated soil

2 Disposal of relatively low-contamination waste
3 Interim storage of the removed soil and the designated waste with high
radioactivity.
They estimated the costs only for the areas in which evacuation orders are issued.
Their estimates did not include the cost for final disposal, which is required after
the interim storage. To estimate the preceding three costs, they adopted some
scenarios based on the thoroughness of forest decontamination as well as the com-
bination of decontamination techniques for agricultural lands. Owing to the sensi-
tivity to the scenarios, the estimated cost has a broad range: from JPY 236 billion
to JPY 3.9 trillion (Table 12.1). Later, they extended the geographic area to include
all affected communities (Yasutaka and Naito 2013). The estimated decontamina-
tion cost is in the range of JPY 2.5 trillion to JPY 5.1 trillion.
The Japanese national government (Nuclear Emergency Response Headquar-
ters 2013) also estimated the costs of (1) and (2) to be JPY 3.6 trillion and (3) to
be JPY 1.1 trillion, which makes the aggregated cost JPY 4.7 trillion (Table 12.1).
In a revised estimate (METI 2016),4 the costs of (1) and (2) increased to JPY
4.0 trillion and (3) was also revised to JPY 1.6 trillion, resulting in a JPY 5.6 tril-
lion total. As with the estimates by Yasutaka et al. (2013) and Yasutaka and Naito
312 Kazunobu Okubo et al.
Table 12.1 Estimated cost for decontamination in previous studies
Area Aggregated cost in JPY

Yasutaka et al. (2013) Only areas for which 236 billion–3.9 trillion
evacuation orders are issued
Yasutaka and Naito (2013) All affected areas 2.5 trillion–5.1 trillion
Nuclear Emergency Response All affected areas 4.7 trillion
Headquarters (2013)
METI (2016) All affected areas 5.6 trillion
JCER (2017) All affected areas 30.0 trillion
Source: The authors
(2013), they did not include the cost for final disposal because of the considerable
uncertainty in the scheme.
Table 12.1 shows that the third estimate (JCER 2017) is much larger than the
aforementioned studies (JPY 30 trillion), owing to the inclusion of the final dis-
posal cost.
These studies have some challenges regarding uncertainty, transparency or
assumptions. For example, owing to the early initiation of their studies, Yasutaka
et al. (2013) and Yasutaka and Naito (2013) were frustrated by the lack of certain
information at the time of publication. The assumptions and processes the Japa-
nese national government used to obtain the estimates are not open to the public.5
JCER (2017) assumes that the unit cost of managing the contaminated objects is
equivalent to the cost of managing the low-level radioactive waste from nuclear
power plants. However, as mentioned, the main radioactive substances dispersed
by the accident are Cs-134 and Cs-137, whose half-lives are 2 years and 30 years,
respectively. Since the ratio of the discharged amount of the two kinds of cesium
is known to be 50:50,6 the excess air dose rate caused by the accident should
have lowered more rapidly than it would if Cs-137 had been the sole pollutant.
Furthermore, the radioactivity level of removed soil and wastes from the areas
distant from Fukushima Dai-ichi should be much lower than the nuclear power
plant wastes. Thus, JCER (2017) may overestimate the costs of managing con-
taminated objects.
We made an effort to avoid these problems. Our study depends on new infor-
mation that Yasutaka et al. (2013) and Yasutaka and Naito (2013) could not use:
actual government expenditure data. Moreover, all of our assumptions and esti-
mation processes are open. We also consider the characteristics of Cs-134 and
Cs-137. Finally, we estimate the cost in present value to compare it with corre-
sponding benefit, while previous studies estimated the cost in current value.
Compared to cost estimation studies, benefit estimation studies are scarce. As
far as we know, there are only two of these studies.7 Iida’s estimate contained
in Nakanishi (2014) assumed that man-made capital in the evacuation area is
completely depreciated due to exposure to rain and wind. What is recovered
by the decontamination project is the value of land in the area, which amounts
to JPY 1 trillion. The other is Yamaguchi et al. (2017), who adopted the same
approach as the current chapter.8 What is recovered by the decontamination project
is the value of produced and natural capital services.9 According to their estimate,
the total recovered value ranges from JPY 0.9 trillion to JPY 1.5 trillion in 2011
value. In this chapter, we present another estimate of decontamination benefit.
The remainder of this chapter is structured as follows. In Section 2, we clarify
the differences in the aim and definition of “decontamination” between the gov-
ernment’s documents and our study. This clarification is a preparation to identify
the costs and benefits. In Section 3, we show the empirical framework including
the operational definition of cost and benefit. We also introduce the estimation
method and the data sources in this section. Section 4 is dedicated to reporting the
result of our analysis. Section 5 provides concluding remarks.
Of course, the estimates of the cost and benefit in this chapter are only a first
approximation. Although we have carefully constructed our analytical framework,
our study cannot be perfect because of much uncertainty in the scheme and timing
of decontamination work. The estimates should be revised in the future with new
information. Moreover, we hasten to note that this chapter is intended to propose
a method to estimate the benefits and costs of a decontamination project, with
limited scope. To do a thorough cost-benefit analysis would require much broader
information and other socio-economic aspects that may not be measurable.
2 Preparation for identifying the costs and benefits

In this section, we define scope of our study. In particular, it is worthwhile noting
that the benefit of health risk reduction and the cost for final disposal are out of
the scope.10
Decontamination after the Fukushima Dai-ichi Power Plant accident is prin-
cipally governed by the Act on Special Measures concerning the Handling of
Radioactive Pollution, which was enacted immediately after the accident. It stipu-
lates that the aim of decontamination is to promptly reduce the negative effect to
human health and the living environment. In other words, we can redefine the dual
aims of decontamination as:
• Reducing additional cancer mortality risk from additional radiation

exposure
• Recovering capital services of roads, housing, agricultural lands and forests
contaminated by radioactive substances.
In this chapter, we focus on the latter due to the sheer uncertainty of health risks.
The Act defines decontamination as a set of several types of work. It involves
removing soil, tree leaves and branches, and contaminated sludge to prevent dif-
fusion of the radioactive substances. However, this definition of decontamination
only covers the first phase of the whole process. We instead include the costs of
temporary storage and interim storage in addition to the costs of removing soil
and waste.
Table 12.2 Final disposal plans of soil and waste according to their characteristics
Soil Waste
no more than more than 8000 more than

8000 Bq/kg to no more than 100,000 Bq/kg
100,000 Bq/kg
Fukushima After 30 years’ At existing At the special After 30 years’
prefecture interim storage facilities without facility (Fukushima interim storage
interim storage Eco-tech Clean
Center) without
interim storage
Other At new facilities At existing At new facilities Not applicable
prefectures without interim facilities without without interim
storage interim storage storage
Source: The authors
The removed soil and waste are managed in different way depending on three
factors: the contents, the location of generation and the extent of the radioactive
concentration (Table 12.2).
The procedures for each item are as follows. Basically, the national government
requests that the contaminated objects be managed in the prefecture in which they
are generated. Fukushima prefecture, in which the Fukushima Dai-ichi Nuclear
Power Plant is located, is the exception, as it has huge volumes of removed soil
and waste with high radioactive concentration. These must be kept in the interim
storage facility until the radioactivity levels are lower than the safe levels.11
Most of the contaminated objects in Fukushima that must be managed in speci-
fied ways are removed soil. The amount is estimated to be in the range of 16 mil-
lion to 22 million cubic meters.
“Designated waste,” which must be managed in a specified way, is identified as
such by the Minister of the Environment if the radioactive concentration levels are
over 8,000 Bq/kg.12 The designated waste at Fukushima is estimated to amount to
170,000 cubic meters, which is almost equivalent to the total designated waste in
the other five prefectures, and the amount may increase for full-scale decontami-
nation in the coastal area.13
Fukushima is the only exceptional prefecture that plans to have an interim
storage facility for the removed soil and waste. It is under construction in the
coastal area.14 The soil and waste are treated differently. The soil is planned to
be accommodated in the facility irrespective of radioactive concentration level.
The designated waste with a radioactive concentration level over 100,000 Bq/kg
must be also accommodated in the facility. The designated waste with a radioac-
tive concentration level more than 8,000 Bq/kg to no more than 100,000 Bq/kg is
disposed of in a special landfill designed only for this waste.15 The waste with a
radioactive concentration level no more than 8,000 Bq/kg is disposed of in exist-
ing non-special landfills.
In addition to Fukushima prefecture, the national government chose five other
prefectures (Miyagi, Ibaraki, Tochigi, Gunma and Chiba) where the total volume
of the contaminated objects exceeds the total capacities of existing landfills. The
national government plans to construct new landfills at its own expense in each
prefecture to dispose of the removed soil and the designated waste with a radioac-
tive concentration level over 8,000 Bq/kg.16
The responsibility to determine the site for construction is given to the Ministry
of the Environment (MOE). It has already determined the candidate sites in some
prefectures17 and has begun to negotiate with citizens and their municipal govern-
ments. However, citizens and municipal governments oppose the MOE’s plan. In
the five prefectures, the investigation of candidate sites for final disposal has not
been completed even after 6 years have passed since the full-scale decontamina-
tion began.
Most of the removed soil and waste are kept in temporary stockyards or the
very places that are being decontaminated. The objects are typically stuffed in
flexible container bags to be piled up on the ground or buried under ground.
As there is much uncertainty in the location for and timing of final disposal, we
have decided to exclude the costs of final disposal from our cost estimation, as
previous studies did. This does not cause a problem: if the total benefit is smaller
than part of the total cost, it means that the total benefit cannot exceed the total
cost.
3 Method and data

In this section, we introduce the empirical framework and the method to estimate
the costs and benefits we have just defined.
To estimate the present value of the costs of the decontamination project, we
set Fiscal Year 2011 as the base year. We use government expenditure data in
current money in each year.18 In the analysis, we count only the expenditures by
the MOE.19 The expenditure in current money is discounted at 5% per annum.
Remember that we only account for the costs that have been realized.
We define the benefit of the project as the sum of the present values of recov-
ered services of capitals: roads, housing, agricultural lands and forests. We com-
pare with a counterfactual: without intervention, the schedule of air dose rate
reduction would have been governed only by natural decay and weathering. Thus,
intervention enables us to use capital services earlier, which increases the shadow
price of capital assets (Yamaguchi et al. 2017). This can be regarded as capital
revaluation in an inclusive wealth framework brought about by the exogenous
accident and subsequent intervention.
The Japanese national government declared that the evacuation order could be
lifted if the air dose rate became lower than the designated level, which is 20 mSv/
year, or equivalently 4.6 μSv/hour. For example, if the air dose rate is prospected
to be less than the designated level in 2017 even without decontamination, but
to be so in 2015 with decontamination, then we regard that the decontamination
project gained 2 more years’ services compared with no human intervention case.
The preceding framework is designed for areas where evacuation was once
ordered but the order has been removed (we call these areas the “once under
the order area” henceforth). There are other two categories of area: areas where
the evacuation was not ordered because the air dose rate have been less than
the designated level from the very beginning (the “in-situ area” henceforth),
and areas where the evacuation was ordered and the order has not yet been
removed because the air dose rate remains high (the “still under the order area”
henceforth).
For the in-situ area, we set the duration of lost services as follows.
• Roads: 1 month in Fiscal Year 2011

• Housing: 3 months in Fiscal Year 2011
• Agricultural lands: 2 years from Fiscal Year 2011 to Fiscal Year 2012
• Forests: 2 years from Fiscal Year 2011 to Fiscal Year 2012.
The listed durations and timings are set based on documents from the prefectural
or municipal governments.20
For the still under the order area, we calculate the recovered services based on
the gap between 2023 and the year when the air dose rate will come to be less
than the designated level even without decontamination. The national govern-
ment declared that it would remove the evacuation order no later than 2023. Our
calculation identified that the timing when the air dose rate will naturally come to
be less than the designated level is 2035, and the gap between with and without
cases is 13 years.
To calculate the timing of capital services recovery, we adopt the following
assumptions:
1 The main isotopes are Cs-134 and Cs-137; their half-lives are 2 years and
30 years, respectively
2 The ratio of discharged amounts of these two isotopes is 50:50
3 The contribution share to air dose rate of these two isotopes is 73:27
4 Weathering contributes to reducing air dose rate in addition to natural decay.
These assumptions come from national government documents.21 Figure 12.1 is

an illustration of time schedules of air dose rate with different initial levels. It
shows that even without human intervention, as time goes by, natural decay and
weathering cause the air dose rates to decline.
There is a point in time at which the dose rates come to be less than the desig-
nated level (20 mSv/year or 4.6 μSv/hour), at which it is assumed that people can
resume capital services. As shown in Figure 12.2, the timing of capital service
recovery differs with and without decontamination, which is important to the val-
ues of recovered services subject to discounting.
We can obtain the data on air dose rates from the Nuclear Regulation Authori-
ty’s monitoring reports.22 We calculate average dose rate among monitoring points
for the seven affected prefectures other than Fukushima, respectively, all of which
60
50
40
µSv/hour
Intial level=20 µSv/hour

30
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Year
Figure 12.1 The schedule of air dose rate reduction without human intervention for dif-
ferent initial levels
µSv/hour
The time schedule of

air dose rate for
The designated without-decontamination
safe level
The time schedule of
air dose rate for
with-decontamination case
Year
The difference in timing of resuming
Figure 12.2 How to identify the timing of capital service recovery
have the in-situ areas only (Iwate, Miyagi, Ibaraki, Tochigi, Gunma, Saitama and
Chiba).23 For Fukushima, the only prefecture that has all categories of areas, we
separately calculate average air dose rates for three areas: in-situ, once under the
order and still under the order.
Table 12.3 Difference in the timing of resuming capital services
with and without decontamination in the case of the
once under the order area
Year (1) Natural decay (2) Without (3) With

and weathering decontamination decontamination
factor (μSv/hour) (μSv/hour)
2011 1.00 16.52 16.52
2012 0.78 12.85 9.62
2013 0.62 10.21 8.81
2014 0.50 8.29 7.03
2015 0.42 6.88 5.62
2016 0.35 5.85 4.36
2017 0.31 5.08 3.79
2018 0.27 4.50 3.35
2019 0.25 4.05 3.02
2020 0.22 3.70 2.76
…
…
…
Source: The authors
To construct the counterfactual time series of air dose rates, we first estimate the
time schedule of air dose rates with only natural decay and weathering. Column (1)
in Table 12.3 shows the ratio of air dose rate at each point in time to the initial level
when we consider only natural decay and weathering. This can be thought of as a
physical discount factor, setting the 2011 value to be unity. Column (2) in Table 12.3
is for the counterfactual case, setting the average value of air dose rate in 2011 as the
initial rate of the without-decontamination case. We can construct the counterfactual
time series of air dose rates after 2012 by multiplying the ratios in column (1).
Column (3) is the time series of air dose rates for the with-decontamination
case. The first six values for years 2011–2016 are the average air dose rates of the
observations. The values after 2017 are projections obtained by multiplying the
ratio of each year in column 1 and the average value of 2016 in column 3, and
dividing the products by the ratio of 2016 in column 1.
The gray cells in columns (2) and (3) indicate the points in time when the actual
or projected dose rates first come to be less than the designated level (4.6 μSv/
hour). For the once under the order area, we can regard that capital services in
2016 and in 2017 are recovered by the decontamination project. In a similar man-
ner, we can obtain the amount of recovered capital services in the still under the
order area, which results in 13 years of services, from 2023 to 2035.
3.1 The value of annual services

To estimate the values of recovered services, we need to estimate the unit value of
each capital service. The unit values of services are discounted at 5% per annum.
The present values of recovered services in each year are the products of the unit
values and the recovered capital stocks. The sum of these products is the present
value of capital services recovered by decontamination.
3.1.1 Roads
The services of roads are not marketed, so we assume that the asset value of a
road is the sum of the discounted value of constant benefit flows in the planning
horizon and is equal to the construction cost. With this simple assumption, we can
calculate the value of service of a road in each year by reversing the process to
obtain the asset value from benefit flows.
The construction cost of 1 km of road in a rural area like coastal Fukushima
is about JPY 1.7 billion in 2011 value. We assume that the lifetime of a road is
50 years.24 Depreciation is ruled by logistic function. At the half of the lifetime,
the value becomes half of the initial value. At the end of its lifetime, the value
becomes zero. Based on the information and assumptions, we can estimate the
unit value of annual service of a road in the area to be JPY 90.7 million km/year.
The data on the length of decontaminated roads is provided by the MOE.25 Mul-
tiplying the unit value by the length of the road, we can obtain the total value of
recovered service of decontaminated roads in each year.
3.1.2 Housing
Unlike the case of roads, housing services are marketed, and we can obtain the
price data.26 Although we cannot obtain the information on each house that is
actually decontaminated, we can obtain the average capacity and average rental
price in the affected area. For example, in the coastal area of Fukushima, the
annual average rental price for housing with average capacity is JPY 1.5 million
housing unit/year. The unit value is discounted in the same way as the service of
roads. The MOE provides data on the number of decontaminated housing units.27
Multiplying the discounted unit value by this number, we can obtain the value of
recovered housing services in each year.
3.1.3 Agricultural land
The services of agricultural lands are marketed. We can obtain the data of annual
rental price from a report by the Ministry of Agriculture, Forestry and Fisheries.28
It is JPY 75,000/hectare. The unit value is discounted at 5% per annum. The MOE
provides data on the total area of decontaminated lands in each municipality.29
Multiplying the discounted unit value by the total area, we can obtain the value of
recovered services in each year.
3.1.4 Forests
Forests provide several kinds of services that contribute to human well-being.
However, most of these useful services are not marketed. The Economics of
Ecosystems and Biodiversity (TEEB) provides estimates of unit value of several
forest services: provisioning, regulating, habitat and cultural (Van der Ploeg and
de Groot 2010). Following UNU-IHDP and UNEP (2014), we use the TEEB data
on temperate and boreal forests. Note that some services are maintained in spite of
radioactive contamination. Thus, we count only the deprived services: providing
food and water and recreation services (Yamaguchi et al. 2017). The unit value
of these forest services is USD 196/hectare in 2005 value. We can obtain the unit
value of these services in JPY exchanging the dollar amount at the rate of JPY
103/USD. It is JPY 20,200/hectare. The unit value in each year is discounted to
obtain the present value. The MOE provides data on the decontaminated forest
area. Multiplying the discounted unit value by the forest area, we can obtain the
value of recovered services in each year.30
4 Results
Table 12.4 shows the government expenditures in three accounting categories: (1)
removal and temporary storage of contaminated soil, (2) disposal of contaminated
waste and (3) interim storage for objects with high radioactivity. These costs are
discounted at 5% per annum depending on the timing of expenditure. The total of
the discounted expenditures is JPY 2.7 trillion. If the dead-weight loss incurred
by public finance increases the cost by 25%,31 the social cost of the finished work
would be JPY 3.3 trillion.
Table 12.5 shows the benefits of the decontamination project in three areas: the
in-situ, once under the order and still under the order areas. For the in-situ area,
the total of the present value of the recovered capital services is JPY 277 billion.
For the once under the order area, it is JPY 186 billion. The benefit that will be
gained from future work in the still under the order area is estimated to be JPY
164 billion.
Table 12.4 Discounted government expenditures from fiscal years 2011 to 2016 (unit: JPY
billions)
Fiscal year 2011 2012 2013 2014 2015 2016 2011–2016

Removal and temporary 99.4 357.3 569.3 202.2 404.6 986.5 2619.30
storage of contaminated soil
Disposal of contaminated 3.8 19.3 39.8 46.6 53.2 243.6 406.3
waste
Interim storage for objects that 0.7 1.9 0.7 9.8 13.3 109.8 136.2
have high radioactivity
Total 103.9 378.5 609.8 258.6 471.1 1339.90 3161.80
Discounted total 103.9 360.5 553.1 223.4 387.6 1049.80 2678.30
Source: The authors’ calculations
Note: Actual expenditures for 2011–2015 and budget for 2016.
Table 12.5 Total benefit by area (unit: JPY billions)
In-situ case Once under Still under the

the order case order case
Benefit gained 277 186 –
Benefit to be gained in the future – – 164
Source: The authors’ calculation
The sum of the decontamination benefit in the three areas, JPY 627 billion in total,
largely falls short of the present value of the costs that have already been realized
(JPY 2.7 trillion). We conclude that, under our definition of costs and benefits, the
net benefit is negative.32 If we compare the total benefit with the hypothetical cost
including marginal excess burdens, which amounts to JPY 3.3 trillion, the differ-
ence becomes much larger. Note that the cost should increase in the future, not
decrease, due to the unfinished work.
5 Conclusion
We estimated the costs and benefits of decontamination after the Fukushima Dai-
ichi Power Plant accident. The costs, defined as the sum of the present values of
government expenditure from 2011 to 2016, amount to JPY 2.7 trillion in 2011
value.
The benefits are defined as the present values of recovered capital services
or, alternatively, the gain in shadow prices of capital assets such as roads, hous-
ing, agricultural lands and forests in the affected area, in line with inclusive
wealth and the sustainability assessment framework (UNU-IHDP and UNEP
2012, 2014).
To identify the amount of recovered services, we construct the counterfactual
air dose rate schedules in each prefecture, and that in three categories of areas
in Fukushima prefecture, for the without-decontamination case and compare
them with the corresponding schedules of the with-decontamination case. The
gaps between the points in time when the air dose rates come to be less than the
designated level help to identify the amount of recovered capital services and to
estimate their values.
The products of the present unit value and corresponding capital stocks (the
length of decontaminated roads, the number of decontaminated housing units,
the area of decontaminated agricultural lands, and the area of decontaminated
forest) are the present value of recovered services in each year. The sum of these
present values, estimated to be JPY 627 billion, largely falls short of the realized
cost.
There is still much uncertainty, even after 6 years have passed from the ini-
tiation of the project, so new information in the future will enable us to obtain
more precise estimation. We focused only on the costs and benefits of the MOE
project,33 excluding projects conducted by the Cabinet Office and the Ministry
of Education, Culture, Sports, Science and Technology, which are responsible for
the smaller demonstration projects and the decontamination of kindergartens and
schools,34 respectively.
Moreover, some important benefits (e.g., reducing health risks of internal
and external exposure; decreasing environmental diseconomies to the genes
of wild animals and plants; reducing the anxiety of local citizens) are not
included. Further, our estimate of efficiency loss incurred by public expendi-
ture is still crude.
Finally, it is worth stressing that our study is based on monetary terms, ignoring
welfare gains and other critical criteria such as equity and fairness. The Fuku-
shima Dai-ichi accident deprived many people of lives in their hometown. If the
well-being of those people with a very high cost of waiting for the air dose rate
to decrease to the safe level is prioritized, the benefit of decontamination project
could be much larger than our estimate.
Acknowledgment
This research has been partially supported by Grant-in-Aid for Specially Promoted
Research (JP26000001) from the Japan Society for the Promotion of Science.
Notes
1 The 11 municipalities are Iitate, Kawamata, Katsurao, Namie, Minami-souma, Futaba,
Okuma, Tomioka, Tamura, Kawauchi and Naraha.
2 Some private companies decontaminated their facilities and equipment by themselves
and requested compensation from TEPCO (the Tokyo Electric Power Company).
Legally, the costs are not classified as public decontamination costs, but as private
reparation.
3 The Ministry of Economy, Trade and Industry published the estimates of future and
total costs of removing contaminated soil and wastes and of interim storage of highly
contaminated objects (METI 2016).
4 See also Japanese National Government (2016).
5 In the government document to report cost estimates, there is only a remark that the
cost estimates were obtained by consulting some professionals.
6 See Nuclear Emergency Response Headquarters (2011).
7 Munro (2013) also helps our study. He constructed a framework for assessing the
merits of management options for the Fukushima Dai-ichi accident. It is a framework
to assess whether delayed intervention is preferable and, if preferable, how long we
should wait. Optimal delay is determined by the benefit of capital service recovery and
cancer risk reduction, and the cost of decontamination.
8 There are two essential differences between Yamaguchi et al. (2017) and our study.
First, the covered geographic area has been extended from the 11 most affected munici-
palities to all affected municipalities. Second, the duration of capital service recovery
was assumed to be 10 years. Instead, we use an estimated duration based on the air dose
rate monitoring data.
9 While Yamaguchi et al. (2017) estimated the value of recovered services from all man-
made capital, we estimate that from only roads and housing.
10 In radiation protection studies, decontamination is narrowly defined as meaning only
removing radioactive substances from objects. “Remediation” is generally used to
mean the set of activities that the Japanese government defines as decontamination
( josen). The difference in usage of the terms often confuses foreign researchers. See
Nakayama (2014).
11 If the sites for final disposal are not determined until then, they may be kept for a longer
time. However, the national government promised that, in not more than 30 years, the
soil and waste would be transferred to and disposed of in facilities outside Fukushima.
12 If the level of radioactivity concentration is over 8,000 Bq/kg, then the waste can be
“designated” at the request of municipal governments.
13 The coastal area of Fukushima has a high air dose rate and the evacuation order is still
effective there. Decontamination of the area just began in 2017.
14 The interim storage facility is located in the site surrounding the Fukushima Dai-ichi
Nuclear Power Plant. http://josen.env.go.jp/chukanchozou/about/.
15 Fukushima Eco-tech Clean Center is the landfill site for Fukushima’s designated waste
with a radioactivity level more than 8,000 to no more than 100,000 Bq/kg. The landfill
is located in the coastal area of Fukushima. See MOE (2013).
16 The waste with a radioactive concentration level no more than 8,000 Bq/kg and gen-
erated in prefectures other than Fukushima is disposed of in the existing non-special
landfills in each prefecture.
17 The MOE has not determined the candidate site for Ibaraki and Gunma. These prefec-
tures are allowed to keep the removed soil and waste in temporary stockyard because
their radioactivity level are relatively low.
18 See MOE (2017).
19 The MOE is the agency designated with the responsibility to manage the decontamina-
tion project under the Act on Special Measures concerning the Handling of Radioactive
Pollution. Agencies other than the MOE manage their own projects under emergency
orders. They began work before the Act was legislated.
20 For example, the Fukushima Prefecture Government issued the plan for decontamina-
tion of a road in Koriyama city. It contains an announcement that decontamination
began in late November and ended in early March. www.pref.fukushima.lg.jp/down-
load/1/20131030_dourojosenn.pdf.
21 See Nuclear Emergency Response Headquarters (2011) for the detail.
22 “Monitoring information of environmental radioactivity level (Houshasen monitaringu
johou).” http://radioactivity.nsr.go.jp/map/ja/
23 There are many monitoring points at each observation date. For example, the monitor-
ing data for Fukushima prefecture at April 29, 2011, has approximately 100,000 points.
24 The MIC (N.D.) provides a table of the service lives of durable goods.
25 For the in-situ area, the data are obtained from http://josen.env.go.jp/zone/.
For the once under the order area, the data are obtained from http://josen.env.go.jp/area/.
For the still under the order area, it is assumed that all roads in the area are decon-
taminated. Road length in the area is obtained from “Road Statistics 2011 (Douro
Nenpo 2011)” by the Ministry of Land, Infrastructure, Transport and Tourism: www.
mlit.go.jp/road/ir/ir-data/tokei-nen/2011tokei-nen.html.
26 See “Housing and Land Statistics 2008 (Ju-taku Tochi Toukei Chosa 2008)” by the Sta-
tistics Bureau, the Ministry of Interim Affairs and Communications: www.stat.go.jp/
data/jyutaku/2008/index.htm.
For the once under the order area, the data are obtained from http://josen.env.go.jp/
area/.
For the still under the order area, it is assumed that all housing used in the area in 2011
is decontaminated. The number of housing units in use in 2011 is assumed to be equal
to the number of evacuated households in the area. The number of evacuated house-
holds is obtained from the “Survey of the mind of residents of the affected municipali-
ties by Fukushima Dai-ichi Nuclear Power Plant accident (Genshiryoku hisai jichitai
ni okeru ju-min iko chosa)”: www.reconstruction.go.jp/topics/main-cat1/sub-cat1-4/
ikoucyousa/.
28 See the “Trade and rental prices of agricultural land (No-chi no kakaku to chin-shaku-
ryo)” by the Ministry of Agriculture, Forestry and Fisheries: www.maff.go.jp/kyusyu/
toukei/database/pdf/2-8-1_noutikakakukenri_26.pdf.
For the once under the order area, the data are obtained from http://josen.env.go.jp/area/.
For the still under the order area, it is assumed that all agricultural lands in the area
are decontaminated. The data on agricultural land area in 2010 is obtained from the
“Agricultural land survey (Menseki chosa)” by the Ministry of Agriculture, Forestry
and Fisheries: www.maff.go.jp/j/tokei/kouhyou/sakumotu/menseki/.
30 The value of timber is excluded because there is much uncertainty about whether the
contaminated wood can be marketed even after decontamination.
31 Some public finance researchers argue that there is a marginal excess burden (MEB)
when government raises revenue (Aronsson et al. 2012; Yamaguchi et al. 2016). MEB
is defined as the change in dead-weight loss due to additional government revenue
(typically, tax revenue). See the literature review on MEB by Auerbach and Hines
(2002).
32 We omit sensitivity analysis for different values of social discount rate because the
alteration of discount rate within a reasonable range is unlikely to change our qualita-
tive result.
33 The MOE also managed decontamination of public facilities (halls, libraries, water
supply facilities, sewage treatment plants, gymnasiums, sports grounds and so on). The
benefit of public facility decontamination is not included in our estimate because there
are difficulties in defining benefit and obtaining data.
34 Part of the responsibility for decontamination of schools and kindergartens was
transferred to the MOE after the Act on Special Measures concerning the Handling
of Radioactive Pollution was enacted. However, we exclude its benefit because
we have little evidence of the effect of decontamination on children’s health risk
reduction.
Bibliography
Aronsson, T., Cialani, C., and Löfgren, K. G. (2012) Genuine saving and the social cost of
taxation. Journal of Public Economics, 96(1–2), 211–217.
Auerbach, A. J., and Hines, J., Jr. (2002) Taxation and economic efficiency. In: A. J. Auer-
bach and M. Feldstein (Eds.), Handbook of Public Economics, edition 1, volume 3,
chapter 21, 1347–1421. Elsevier.
Japanese National Government (2016) Cabinet decision on the principles for the accelera-
tion of Fukushima’s recovery from nuclear power plant accident (Genshiryoku saigai kara
no Fukushima fukkou no kasoku no tame no kihon shishin) revised at 20 th, December,
2016, www.meti.go.jp/earthquake/nuclear/kinkyu/pdf/2016/1220_01.pdf (in Japanese).
JCER (Japan Center for Economic Research) (2017) The total cost of decontamination,
compensation, and decommissioning is predicted to be 50 to 70 trillion yen (Jiko syori
hiyo wa 50–70 cho en ni naru osore), www.jcer.or.jp/policy/pdf/20170307_policy.pdf
(in Japanese).
Leigh, N. G., and Blakely, E. J. (2016) Planning Local Economic Development: Theory and
Practice. Thousand Oaks, CA: Sage Publications.
METI (Ministry of of Economy, Trade and Industry) (2016) Projection on the total cost of
the Fukushima Dai-ichi Nuclear Power Plant accident and the role of TEPCO (Tokyo
Electric Power Company) and national government (Fukushima jiko oyobi kore ni kan-
ren suru kakuho subeki shikin no zentaizou to Toden to kuni no yakuwari buntan), www.
meti.go.jp/committee/kenkyukai/energy_environment/touden_1f/pdf/006_s01_00.pdf
(in Japanese).
MIC (Ministry of Internal Affairs and Communications) (N.D.) The service life of durable
goods (Taiyo nen-su hyou). In: A Guidance of Asset Valuation and Preparation for an
Accounting Book of Fixed Properties (Shisan hyoka oyobi kotei shisan daicho seibi no
tebiki), www.soumu.go.jp/main_content/000335896.pdf
MOE (Ministry of the Environment) (2013) On the plan for landfills in the Fukushima
Ecotech Clean Center (Fukushima Eko-tekku Kurin Senta umetate shobun keikaku an ni
tsuite), www.env.go.jp/jishin/rmp/conf/10/ref07_2.pdf (in Japanese).
MOE (Ministry of the Environment) (2017) Expenditures for the decontamination, waste
disposal, and interim storage (Josen, osen haikubutsu shori, chu-kan chozou sisetsu
jigyou no yosan shikkou joukyou). In: The State and Prospect of Decontamination,
Waste Disposal, and Interim Storage (Josen, chu-kan chozou shisetsu, houshasei buss-
itsu osen haikibutsu syori no genjou, seika, mito-shi), http://josen.env.go.jp/material/pdf/
outcome_outlook_170303.pdf (in Japanese).
Munro, A. (2013) The economics of nuclear decontamination: Assessing policy options for
the management of land around Fukushima Dai-ichi. Environmental Science & Policy,
33, 63–75.
Nakanishi, J. (2014) Risk analysis of nuclear power plant accident and radiation (Genpatsu-
jiko to hoshasen no risuku-gaku), Tokyo: Nihon-Hyoron-Sha (in Japanese).
Nakayama, S. (2014) Environmental remediation in Fukushima: Decontamination 3 years
after (Fukushima no kankyo-kaifuku: Jiko san-nen-go no josen). Energy and Resources,
35(2), 110–114 (in Japanese).
Nuclear Emergency Response Headquarters (2011) On the framework of predicting
future air dose rate based on monitoring data (Genzai no hosha-senryo kara shorai no
ku-kan senryo-ritsu wo yosoku suru houhou ni tsuite), (in Japanese). http://www.
kankyo.metro.tokyo.jp/policy_others/radiation/about/hangenki.files/siryo1-
120110824.pdf
Nuclear Emergency Response Headquarters (2013) Toward the acceleration of Fukushi-
ma’s recovery from nuclear power plant accident (Genshi-ryoku saigai kara no Fuku-
shima fukkou no kasoku ni mukete) revised at 20th, December, 2013, www.meti.go.jp/
earthquake/nuclear/pdf/131220_hontai.pdf (in Japanese).
UNU-IHDP and UNEP (2012) Inclusive Wealth Report 2012. Cambridge: Cambridge Uni-
versity Press.
UNU-IHDP and UNEP (2014) Inclusive Wealth Report 2014. Cambridge: Cambridge Uni-
versity Press.
Van der Ploeg, S., and de Groot, R. (2010) The TEEB valuation database – a searchable
database of 1310 estimates of monetary values of ecosystem services, Foundation for
Sustainable Development. https://www.es-partnership.org/services/data-knowledge-
sharing/ecosystem-service-valuation-database/
Yamaguchi, R., Okubo, K., and Managi, S. (2017) Fukushima. In: S. Managi (ed.), Valua-
tion of Prosperity: Developing the Inclusive Wealth Index (Yutakasa no kachi hyoka:
Shin-kokufu shihyo no kochiku). Chuo-Keizai-Sha (in Japanese).
Yamaguchi, R., Sato, M., and Ueta, K. (2016) Measuring regional wealth and assessing
sustainable development: An application to a disaster-torn region in Japan. Social Indica-
tors Research, 129(1), 365–389.
Yasutaka, T., and Naito, W. (2013) An analysis of the cost of decontamination in the desig-
nated area in Fukushima prefecture (Fukushima-ken-nai no josen jisshi kuiki ni okeru
josen no hiyo ni kansuru kaiseki), www.aist-riss.jp/wp-content/uploads/2014/12/5d22c0
54334dc5ff4f6d1e14e6d636c1.pdf (in Japanese).
Yasutaka, T., Naito, W., and Nakanishi, J. (2013) Cost and effectiveness of decontamination
strategies in radiation contaminated areas in Fukushima in regard to external radiation
dose. PLoS One, 8(9), e75308. doi:10.1371/journal.pone.0075308
13 National and subnational
sustainability
A case study of Japan
Kazunobu Okubo, Rintaro Yamaguchi
and Shunsuke Managi
1 Introduction
The Inclusive Wealth Index (IWI) is a new macro-economic index intended for
use in sustainability analyses1 and is the basis of the Inclusive Wealth Report,
published in 2012 and 2014 (UNU-IHDP and UNEP 2012, 2014).
Inclusive wealth is defined as the sum of the asset value of capital owned by
society. Here, “capital” is broadly defined to include produced (e.g., infrastruc-
ture, factory, and machine), human (e.g., education and health), and natural (e.g.,
agricultural land, forest, and subsoil resources) assets. Thus, an increase in inclu-
sive wealth should mean that the potential well-being of society has increased.
Originally, the IWI was proposed and developed for the assessment of economic
sustainability at a national level. However, economic sustainability also applies to
subnational entities. Economic sustainability at a subnational level is as important
as and perhaps even more important than that at a national level because socio-
economic activities take place at a local level. Thus, applying the IWI to subna-
tional units is highly recommended when assessing local economic sustainability.
It is natural to expect that each subnational unit has its own pattern of capital
accumulation, creating spatial heterogeneity in terms of size, composition, and
patterns of change. Thus, it is possible for a subnational economy to not be sus-
tainable, while the national economy, as a whole, is sustainable. This leads to the
question, is there considerable variation in the assessment of subnational units?
Along with spatial heterogeneity, patterns of population change differ among
subnational units. When a population is constant over time, it is sufficient to mon-
itor the change in total wealth to assess the economic sustainability of society.
Society is adjudged to be on a sustainable development path in a defined interval
if and only if the change in total wealth is non-negative in the same interval. Can
this criterion be extended to population changing economies?
Arrow et al. (2004) and Ferreira et al. (2008) show that an increasing population
size negatively affects the change in wealth, based on a World Development Indicators
dataset. As a result, in addition to total wealth, the Inclusive Wealth Report (UNU-
IHDP and UNEP 2012, 2014) shows “per capita wealth,” which we call total utilitari-
anism (TU) per capita wealth. TU per capita wealth is defined as follows:
W (t )
,
N (t )
where W (t ) and N (t ) are wealth and the population size, respectively, at time t.
Dasgupta (2001), one of the main contributors to the theoretical framework of the
IWI, proposed alternative ethical basis for sustainability analysis that he called
dynamic average utilitarianism (DAU).2 When we adopt this ethical basis, per
capita wealth is defined as follows and called as DAU per capita wealth.
W (t )
.
N * (t )
In this case, the denominator N * (t ) is not the population size at time t, but the
sum of the discounted population from time t to infinity. We call this the present
discounted population:
∞
−δ( s−t )
N * (t ) = ∫ N ( s )e ds.
s=t
The rationale behind the concept of DAU per capita wealth stems from the work
of Arrow et al. (2003), who showed that national accounts with population capital
require that we specify a welfare-subsistence consumption level, at which utility
is zero. However, the DAU criterion can be used to alleviate the difficulty in esti-
mating this level. Another, perhaps more intuitive rationale is that wealth should
be shared by many generations to come, because capitals, the components of
wealth, generate valuable services well into the future (Yamaguchi 2018). A typi-
cal example is that of landscape and biodiversity. However, produced and human
capital can also be shared by future generations through depreciation, mainte-
nance, and the inheritance of knowledge. Therefore, DAU per capita, rather than
TU per capita, is more appropriate for sustainability analyses.
Arrow et al. (2012) showed that a constant rate of population change and con-
stant-returns-to-scale technology ensure that DAU per capita is equivalent to TU
per capita. The problem is that they do not coincide in general; that is, the signs
   
of  WN ((tt))  and  NW*((tt))  may not match. How serious is this gap in terms of sustain-
   
ability analyses of economies in which the population size is changing? Is this gap
so large that it can overturn sustainability assessments?
The preceding argument can be summarized in the following three research
questions, which we examine in this chapter.
Question 1. How different are the sustainability assessments of a nation and its
subnational regions?
Question 2. Are subnational units heterogeneous in term of economic sustain-

ability, as defined by increasing wealth?
National and subnational sustainability 329
Question 3. How sensitive are sustainability assessments to different population
ethics?
Using Japan as a case, our analysis is a direct extension of two lines of inclusive
wealth studies. First, Yamaguchi (2018) focused on the change of wealth at a
national level, and compared inclusive wealth per capita, as defined by TU per
capita and DAU per capita. We follow Yamaguchi (2018) in adopting DAU as an
alternative ethical basis for sustainability analyses, but we focus on the changes
in wealth at a subnational level.
Second, Mumford (2012) was the first to study the change in subnational, state-
level inclusive wealth for a country (the United States). Similarly, Ikeda et al.
(2017) estimated the wealth of all Japanese prefectures, although several previous
studies have also examined specific Japanese prefectures (Yamaguchi et al. 2016;
Okubo 2017). Ikeda et al. (2017) examined the urban – rural disparity in Japan
by dividing 47 prefectures into three groups (urban, rural, and hybrid), based on
the size of each economy. They found differences in the patterns of the change in
wealth per capita, measured as TU per capita. Our study is the first to apply the
DAU sustainability criterion to all Japanese prefectures.
The remainder of this chapter is structured as follows. In Section 2, we describe
the general trends in capital accumulation and population change in Japan since
2000. This section is intended to provide readers with the basic information nec-
essary for the remainder of the chapter. Our empirical framework is discussed in
Section 3. Here, we propose three sustainability criteria, and then show how to
calculate the present discounted population at time t . This section also describes
the data sources of population and of wealth and its components. Section 4 reports
the assessment results of sustainability based on the three aforementioned crite-
ria. Here, we show that the results may be strongly dependent on the discount
rate under DAU. To check the robustness of the qualitative results, we conduct a
sensitivity analysis in Section 5, examining how easily sustainability assessments
are overturned by different discount rates. The final section concludes the chapter.
2 General trends of wealth accumulation and

population change in Japan in 2000s
In this preliminary section, we provide information on general trends of wealth
accumulation and population change in Japan since 2000.
2.1 Wealth accumulation

According to the estimates by Ikeda et al. (2017),3 the value of produced capital in
Japan increased from JPY 1,847 to 2,094 trillion in 2000 and 2010, respectively,
as a result of both private and public investment being depressed during the study
period, as compared with the 1990s. During the same period, Japan experienced
a steady decrease in the value of human capital, from JPY 2,794 trillion in 2000 to
JPY 2,448 trillion in 2010. This was largely the result of a decrease in the number
of workers, although there were also some positive changes. For example, the aver-
age education level (measured by schooling year) increased, and the participation
rate of women and the elderly in the economy increased, among others. The value
of natural capital (defined as the sum of the value of agricultural land, forests, and
subsoil resources) increased from JPY 95 to 106 trillion between 2000 and 2010.
Aggregating the values of the three forms of capital shows that Japan’s wealth
decreased from JPY 4,736 trillion in 2000 to JPY 4,647 trillion in 2010. While the
level of aggregate value has no economic or sustainability significance, the values
of produced, human, and natural capital made up 45%, 53%, and 2%, respec-
tively, of total wealth in 2010.
The value of produced capital had increased in all prefectures since 2000.
During the same period, the value of human capital decreased monotonically in
almost all prefectures, with the exception of Okinawa. Because natural capital
accounts for a small share of wealth, the prefectures that could maintain/increase
their wealth during this period were those that could compensate for the deprecia-
tion of human capital by accumulating produced capital.
Table 13.1 shows the prefectures that were successful in at least maintaining
their level of wealth. Figures for Japan are also provided for reference purposes.
Table 13.1 Share of manufacturing in 20 prefectures that have

maintained their level of total wealth since 2000
Value added-based Employee-based

Japan 19.90% 17.80%
Miyagi 13.40% 13.20%
Fukushima 27.50% 23.20%
Tochigi 33.00% 27.10%
Ibaraki 33.30% 25.60%
Saitama 24.30% 20.60%
Tokyo 7.10% 8.30%
Niigata 23.60% 21.00%
Fukui 28.60% 23.10%
Gifu 33.50% 26.60%
Aichi 28.30% 27.00%
Mie 34.90% 28.90%
Shiga 16.50% 13.60%
Shimane 17.60% 16.50%
Okayama 30.40% 21.90%
Ehime 22.90% 17.10%
Fukuoka 13.70% 12.50%
Oita 24.20% 17.30%
Kumamoto 19.50% 15.90%
Kagoshima 16.50% 13.60%
Okinawa 5.90% 6.70%
Source: Based on the Economic Census 2012.
With some exceptions,4 the share of manufacturing across all industries (mea-
sured either by value added or by the number of employees) is higher than the
national average. This is particularly true of plants and factories of material-inten-
sive (e.g., steel, nonferrous metal, petrochemical, textile) and technology-driven
(e.g., auto, machinery, semiconductor, and electronic device) enterprises. These
firms might have contributed to the increase in produced capital value since 2000
amidst lower levels of public investment.
2.2 Population change

The population of the country increased from 126.9 million in 2000 to 128.1
million in 2010. However, only nine prefectures (19.1%) saw an increase in their
population during this period.5 It may seem that this would result in an increase
in per capita wealth, other things being equal. However, the effect of the decreas-
ing population on per capita wealth is ambiguous if it accompanies an even faster
decrease in the working-age population.6 In fact, the working-age population has
been decreasing in Japan since the middle of the 1990s. In 2000, the working-age
population was 86.2 million, which then decreased to 81.0 million in 2010. All
prefectures, other than Okinawa, experienced a reduction in their working-age
population and human capital stocks in this period.
3 Method and data

Three criteria for assessing the sustainability of economies that have a changing
population have been proposed in the literature. In discrete expressions, letting
∆W (t ) and ∆N (t ) denote W (t + 1) −W (t ) and N (t + 1) − N (t ) , respectively,
the three criteria are as follows:
1 TU criterion: non-negative change rate of total wealth, ∆W (t ) / W (t ) ≥ 0

2 TU pc criterion: non-negative change rate of TU per capita wealth,
 W (t )
∆ 
 N (t )  ∆W (t ) ∆N (t )
≈ − ≥0
W (t ) W (t ) N (t )
 
 N (t ) 
 
3 DAU criterion: non-negative change rate of DAU per capita wealth,
 W (t ) 
∆ 
 N * (t ) ∆W t
  ( ) ∆N * (t )
≈ − * ≥ 0.
 W (t )  W (t ) N (t )
 
 N * (t )
 
Here, the present discounted population is defined as
∞
N * (t ) = ∑ N ( s )ρ(
s−t )
,
s=t
1
where ρ = and δ > 0 .
1+ δ
This decomposition reveals some general facts about the change rate of wealth.
∆N (t ) ∆N * (t )
• TU, TU pc, and DAU coincide only when = * = 0.
N (t ) N (t )
∆N (t ) ∆N * (t )
• TU pc and DAU coincide only when = * . Note that TU pc
N (t ) N (t )
being met does not necessarily mean that wealth is increasing when future
population change is also taken into account.
In our data, we combine actual population estimates for the period 2000–2014,
published by the Statistics Bureau under the Ministry of Internal Affairs and Com-
munications,7 and projected population figures provided by the National Insti-
tute of Population and Social Security Research.8 The latter projection data are
recorded in 5-year intervals, from 2015 to 2040. Thus, missing data are linearly
interpolated, and we assume that the population size is constant after 2041 at the
2040 level, N‾. The latter assumption is not problematic when the discount rate is
not negligible. Therefore, N * (t ) is obtained as follows:
∞ 2040 ∞
N * (t ) = ∑ N ( s )ρ(
s−t )
= ∑ N ( s ) ρ(
s−t )
+ ∑ N ρ(
s−t )
s=t s=t s=2041

2040 ∞ 2040
( s−t ) ( s−t )
− ∑ N ρ(
s−t )
= ∑ N ( s )ρ + ∑N ρ
s=t s=t s=t
2040 N
∑ ( N ( s ) − N )ρ
( s−t )
ds + .
s=t δ
We set the benchmark discount rate at 5%. However, sustainability analysis may
be very sensitive to discount rate. Therefore, in order to examine this, we also
conduct a sensitivity analysis (see Section 5). In our analysis, the initial year is
2000 and the interval Δ is 10 years.
4 Results
4.1 Comparing TU, TU pc, and DAU

Our operational rules for the sustainability assessment are as follows:
• We focus on the trend in the interval (from 2000 to 2010), not on the annual
change.
• If the change rate of wealth is small (within 2.5%),9 we regard this as being
constant.
• If the change rate of wealth is equal to or more than 2.5%, we regard this
as having increased or decreased.
• If wealth returns to the level of the initial year at the end, we regard it as
having been constant in this interval, irrespective of interim changes. The
result in Table 13.2 shows clearly that a sustainability assessment depends
on the criterion we choose.
Table 13.2 Wealth change under TU, TU pc, and DAU, 2000–2010
W(t) W(t)/N(t) W(t)/N*(t)

Hokkaido − constant constant
Aomori − constant +
Iwate − constant constant
Miyagi constant constant constant
Akita − constant +
Yamagata − constant constant
Fukushima constant + +
Ibaraki constant constant +
Tochigi constant constant +
Gunma − constant constant
* Saitama constant − constant
* Chiba − − constant
* Tokyo + − constant
* Kanagawa − − −
Niigata constant constant +
Toyama − constant constant
Ishikawa − − −
Fukui constant constant +
Yamanashi − constant constant
Nagano − constant constant
Gifu constant constant +
Shizuoka − − constant
* Aichi + constant +
Mie + + +
Shiga + constant +
Kyoto − − −
* Osaka − − −
Hyogo − − constant
Nara − − −
Wakayama − + +
Tottori − constant constant
Shimane constant + +
Okayama constant constant +
Hiroshima − − constant
(Continued)
W(t) W(t)/N(t) W(t)/N*(t)
Yamaguchi − constant +
Tokushima − constant +
Kagawa − constant constant
Ehime constant + +
Kochi − − constant
* Fukuoka constant constant +
Saga − − constant
Nagasaki − + +
Kumamoto constant constant constant
Oita + + +
Miyazaki − constant constant
Kagoshima constant + +
* Okinawa + constant +
The Number 20 34 42
of Sustainable
Economies
Note: Prefectures denoted with “*” experienced a population increase in the
period. Those with “+” and “−” experienced an increase and a decrease,
respectively, in wealth. “Constant” means the change was within 2.5%.
Specifically, under TU ( ∆W (t ) / W (t ) ≥ 0 ), 20 prefectures are adjudged to have

 W (t ) W (t ) 
    ≥ 0
been sustainable, but under TU pc ∆    , the number increases
  N (t )   N (t )  
to 34 prefectures. Under DAU, the figure increases further to 42 prefectures. This
result is expected, because most prefectures experienced a population decrease in
the 2000s, and the DAU criterion is the lowest bar for economies with a decreas-
ing population.
In the same interval, Japan was deemed sustainable under TU and DAU, but not
under TU pc. This is because the total wealth of Japan and its present discounted
population were constant from 2000 to 2010, while the population increased dur-
W (t )
ing the same period. Thus, W (t ) decreased while * was constant.
N (t ) N (t )
The assessments for individual prefectures do not often coincide with that of
the country as a whole, under each criterion. Only two densely populated prefec-
tures, Saitama and Tokyo, showed similar patterns to that of Japan (meeting the
TU and DAU criteria, but not the TU pc criterion).
4.2 Comparing prefectures

Table 13.2 reveals the heterogeneity in the sustainability assessments of the pre-
fectures. Six prefectures increased their total wealth (Tochigi, Aichi, Mie, Shiga,
Oita, and Okinawa), while 14 prefectures maintained their total wealth (Miyagi,
Fukushima, Ibaraki, Saitama, Tokyo, Niigata, Fukui, Gifu, Shimane, Okayama,
Ehime, Fukuoka, Kumamoto, and Kagoshima). Thus, 20 prefectures’ economies
are adjudged to be on a sustainable development path in the period.
4.2.1 Group 1
Among these 20 prefectures, eighteen prefectures also meet the other two criteria.
However, the way to fulfill each criterion differs. According to the patterns of
changes of W(t), N(t), and N*(t), the prefectures can be divided into five subgroups
(see Figures 13.1 to 13.5).
Aichi
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
0.95 N*(t) 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Shiga
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
0.95 N*(t) 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Okinawa
1.15 1.10
1.10
1.05
1.05
W(t) 1.00
1.00 W(t)/N(t)
N(t)
0.95 W(t)/N*(t)
0.95 N*(t)
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.1 Group 1-a

Note: The left panel in Figure 13.1 to 13.13 shows the relative levels of three components (W(t), N(t),
and N *(t) ) to their initial levels. The levels in each year are normalized to year-2000 values. The right
W (t )
panel shows the movements of W(t), N (t ) , and N *((t)) , with the values again normalized to year-2000
W t
values.
Tochigi
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85
0.85
2000 2005 2010
2000 2005 2010
Mie
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 W(t)/N(t)
N(t) 1.00
N*(t) W(t)/N*(t)
0.95
0.95
0.90
0.90
0.85
2000 2005 2010
0.85
2000 2005 2010
Oita
1.15 1.15
1.10 1.10
1.05
1.05
W(t)
1.00 W(t)/N(t)
N(t) 1.00
N*(t) W(t)/N*(t)
0.95
0.95
0.90
0.90
0.85
2000 2005 2010 0.85
2000 2005 2010
Figure 13.2 Group 1-b
Group 1-(a). N(t) increased more slowly than W(t) did, while N*(t) was
constant.10
Group 1-(b) N(t) was constant and N*(t) decreased, while W(t) increased.
Group 1- (c) All of the components, W (t ), N (t ) and N * (t ), were constant
in the interval.
Group 1- (d) N (t ) was constant and N * (t ) decreased, while W (t ) was constant.
Group 1- (e) N (t ) and N * (t ) decreased, while W (t ) was constant.
Fukuoka
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.3 Group 1-c

Miyagi
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Ibaraki
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Gifu
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Okayama
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.4 Group 1-d

Kumamoto
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Fukushima
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Niigata
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Fukui
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.5 Group 1-e

Shimane
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Ehime
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Kagoshima
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
4.2.2 Group 2
Saitama and Tokyo are two exceptions in the group of prefectures that meet the
TU criterion. As in the other cases, Saitama and Tokyo meet the DAU criterion.
However, they do not meet the TU pc criterion (Figure 13.6). The total wealth
W (t ) of Saitama and Tokyo were “constant”, as was N * (t ), but N (t ) increased,
W (t )
which made ∆ negative.
N (t )
Saitama
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Tokyo
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.6 Group 2
4.2.3 Group 3
Sixteen prefectures satisfy the TU pc- and DAU criteria only (Figure 13.7). In
these prefectures, N (t ) and N * (t ) decreased as fast as, or more rapidly than,
W (t ) W (t )
W (t ) did, making ∆ and ∆ * “constant” or positive.
N (t ) N (t )
4.2.4 Group 4
Some prefectures satisfy only the DAU criterion. This group contains six pre-
fectures. They can be divided into four subgroups (Figures 13.8 to 13.10). The
W (t )
change patterns of components made ∆W (t ) negative, ∆ negative, and
N (t )
W (t )
∆ * “constant.”
N (t )
Group 4-a: N (t ) increased and N * (t ) was constant, while W (t ) decreased.
Group 4-b: N * (t ) decreased as fast as, or more slowly than, W (t ) did,
while N (t ) was constant.
Group 4-c: N (t ) and N * (t ) decreased more slowly than W (t ) did.
Hokkaido
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Aomori
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Iwate
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Akita
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.7 Group 3

Yamagata
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Gunma
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Toyama
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Yamanashi
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010

Nagano
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Wakayama
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Tottori
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Yamaguchi
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010

Tokushima
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Kagawa
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Nagasaki
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Miyazaki
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010

Chiba
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Shizuoka
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Hyogo
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Hiroshima
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010

Kochi
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Saga
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Kanagawa
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
4.2.5 Group 5
Several prefectures were considered unsustainable, according to any criteria (Fig-
ures 13.11 to 13.13). This group includes five prefectures, which can be divided
into three subgroups based on the change patterns of W (t ), N (t ), and N * (t ).
Group 5-a: N (t ) increased and N * (t ) was constant, while W (t ) decreased.
Group 5-b: N (t ) was constant, while W (t ) decreased. N * (t ) decreased,
but more slowly than W (t ) did.
Group 5-c: N (t ) and N * (t ) decreased more slowly than W (t ) did.
Ishikawa
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Kyoto
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 1.00 W(t)/N(t)
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Osaka
1.15
1.10
1.05
1.00 W(t)/N(t)
W(t)/N*(t)
0.95
0.90
0.85
2000 2005 2010
Nara
1.15 1.15
1.10 1.10
1.05 1.05
W(t)
1.00 W(t)/N(t)
1.00
N(t)
W(t)/N*(t)
N*(t) 0.95
0.95
0.90
0.90
0.85
0.85
2000 2005 2010
2000 2005 2010

5 Discussion
In the previous section, we found that the DAU criterion set the lowest bar for
most prefectures. However, the results may depend strongly on the choice of dis-
count rate.
First, consider the impact of the discount rate selection on the present dis-
counted population (PDP). The left panel of Figure 13.14 shows the movements
of Tokyo’s PDP under three discount rates, normalized to year-2000 values.
For discount rates of 3%, 5%, and 7%, the PDP decreased, was constant, and
increased, respectively. As a result, the wealth of Tokyo under the DAU criterion
increased, was constant, and decreased, respectively. This shows the importance
of choosing the “right” discount rate when assessing sustainability with different
population ethics.
In contrast, Osaka is one of the 20 prefectures where different discount rates
do not affect the sustainability assessments. Table 13.3 shows that the number of
prefectures deemed to be sustainable in terms of DAU is 45 for 3%, 42 for 5%,
and only 18 for 7%. In other words, the smaller the discount rate is (i.e., the more
we care about future generations’ welfare), the more likely it is that inclusive
wealth per capita under DAU is increasing, for our specific context (decreasing
population).
1.15 1.15
1.10 1.10
1.05 1.05
PDP (3%) W(t)/N*(t) (3%)

1.00 1.00
PDP (5%) W(t)/N*(t) (5%)
PDP (7%) W(t)/N*(t) (7%)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Figure 13.14 Tokyo’s PDP and DAU per capita wealth under different discount rates
1.15 1.15
1.10 1.10
1.05 1.05
PDP (3%) W(t)/N*(t) (3%)

1.00 1.00
PDP (5%) W(t)/N*(t) (5%)
PDP (7%) W(t)/N*(t) (7%)
0.95 0.95
0.90 0.90
0.85 0.85
2000 2005 2010 2000 2005 2010
Figure 13.15 Osaka’s PDP and DAU per capita wealth under different discount rates
Table 13.3 Change in DAU per capita under different discount rates
δ = 3% δ = 5% δ = 7%
Hokkaido + constant −
Aomori + + constant
Iwate + constant −
Miyagi + constant −
Akita + + −
Yamagata + constant −
Fukushima + + +
Ibaraki + + constant
Tochigi + + +
Gunma + constant −
* Saitama + constant −
* Chiba + constant −
* Tokyo + constant −
* Kanagawa + − −
Niigata + + −
Toyama + constant −
Ishikawa + − −
Fukui + + constant
Yamanashi + constant −
Nagano + constant −
Gifu + + −
Shizuoka + constant −
* Aichi + + −
Mie + + +
Shiga + + constant
Kyoto + − −
* Osaka − − −
Hyogo + constant −
Nara − − −
Wakayama + + constant
Tottori + constant −
Shimane + + constant
Okayama + + constant
Hiroshima + constant −
Yamaguchi + + constant
Tokushima + + constant
Kagawa + constant −
Ehime + + +
Kochi + constant −
* Fukuoka + + constant
Saga + constant −
δ = 3% δ = 5% δ = 7%
Nagasaki + + constant
Kumamoto + constant −
Oita + + +
Miyazaki + constant −
Kagoshima + + constant
* Okinawa + + constant
The 45 42 18
Number of
Sustainable
Economies
Note: Prefectures denoted with “*” experienced a population increase
in the period. Those with “+” and “−” experienced an increase and a
decrease, respectively, in wealth. “Constant” means that the change was
within 2.5%.
These results imply that when using DAU as our ethical basis, selecting the
discount rate is one of the most important phases in the sustainability analysis
process.
6 Conclusion
Here, we revisit the three research questions posed in the Introduction.
Question 1. How different are the sustainability assessments of a nation and its
subnational regions?
Answer 1. In our case study, the two are quite different. In our observational
window, Japan succeeded in maintaining total wealth in the interval under the
TU criterion, even though some prefectures’ wealth decreased. Under the TU pc
criterion, the Japanese economy as a whole was not on a sustainable develop-
ment path, because the population increased faster than the rate at which wealth
accumulated. At the same time, we found the different trend in some prefec-
tures. In contrast, inclusive wealth under the DAU criterion increased as Japan’s
PDP monotonically decreased over time, while some prefectures experienced
decreasing DAU per capita wealth. In summary, national economic sustainability/
unsustainability does not mean economic sustainability/unsustainability at a
subnational level under any criteria we used.
Question 2. Are subnational units heterogeneous in term of economic sustain-

ability as defined by increasing wealth?
Answer 2. Yes. There is a considerable variation in the sustainability analysis
results under each criterion. Under the DAU criterion, the number of unsustain-
able economies was smallest, because it is the lowest bar to clear for economies
with a decreasing population. Only five prefectures (10.6%) were found to have
economies that were not sustainable. Under the TU pc criterion, this rose to 13
prefectures (27.7%). Under the TU criterion, which is the most difficult criterion
to meet for economies with a decreasing population, 27 economies were found
to be unsustainable (57.4%). Taken together, prefectures are not homogeneous in
terms of their economic sustainability under each criterion.
Question 3. How sensitive are sustainability assessments to different population

ethics?
Answer 3. They are highly sensitive to the choice of criterion. Specifically, we
found that the choice of discount rate is critical for sustainability assessments
under DAU. The more we care about future generations’ welfare, the more likely it
is that the sustainability criterion will be met in the case of a decreasing population.
With regard to Answer 1 and Answer 2, it is not sufficient to monitor the change
of wealth only at the national level, because we may overlook local unsustain-
ability. Of course, a cut-off line does need to be set, but deciding on an optimum
spatial unit for a sustainability analysis is not straightforward (Yamaguchi et al.
2016). However, given that people’s lives are usually confined to a local region,
and that most goods and services are circulated within such regions, local indica-
tors of wealth are equally as important as national indicators.
With regard to Answer 3, we have reconfirmed that choosing an ethical per-
spective that emphasizes the welfare of future generations is critical to sustain-
ability analyses. Yamaguchi (2018) drew a similar conclusion from his analysis
at the national level, but our analysis shows that the matter could be more serious
for increasingly heterogeneous regions, usually hidden under nationally aggregate
statistics.
Acknowledgment
We thank Ikeda, Tamaki, and Nakamura, who kindly provided us with their wealth
estimation data. This research is partially supported by Grant-in-Aid for Specially
Promoted Research (JP26000001) from the Japan Society for the Promotion of
Science.
Notes
1 Arrow et al. (2012, 2013) provide a theoretical framework for the IWI.
2 See also the exposition on dynamic average utilitarianism in Arrow et al. (2004).
3 Their data doesn’t contain the value of housing capital as a component of wealth. The
true change rate of wealth may be smaller because investment in housing can be more
sensitive to decreasing population.
4 The exceptions in Table 13.1 may be divided into two subgroups. The first is composed
of prefectures that have a pivotal local city and a high volume of service industries. The
second comprises those prefectures where primary industry has a large share of value
added or the number of employees.
5 The nine prefectures that experienced a population increase from 2000 to 2010 are
Saitama, Chiba, Tokyo, Kanagawa, Aichi, Shiga, Osaka, Fukuoka, and Okinawa.
6 Yamaguchi (2014) indicates the importance of change in composition of a population
as well as change in its size for sustainability analysis.
7 “Population estimates (Jinkou suikei).” http://www.stat.go.jp/data/jinsui/index.html
8 “Population projection for subnational units (Nihon no chiikibetsu shorai suikei
jinkou).” http://www.ipss.go.jp/pp-shicyoson/j/shicyoson13/t-page.asp
9 This “within 2.5%” rule is determined arbitrarily. If we change the threshold value, the
sustainability assessment of economies will change. However, this does not change the
relative ease/difficulty of meeting three criteria.
10 As in the case of changes in wealth, we regard population as having been “constant” if
the change rate is within 2.5% in the period.
Bibliography
Arrow, K. J., Dasgupta, P., Goulder, L. H. et al. (2004) Are we consuming too much?,
Journal of Economic Perspectives 18, 147–172.
Arrow, K. J., Dasgupta, P., Goulder, L. H., Mumford, K. J., and Oleson, K. (2012) Sustain-
ability and the measurement of wealth, Environment and Development Economics 17(3),
317–353.
Arrow, K. J., Dasgupta, P., Goulder, L. H., Mumford, K. J., and Oleson, K. (2013) Sustain-
ability and the measurement of wealth: Further reflections, Environment and Develop-
ment Economics 18(4), 504–516.
Arrow, K. J., Dasgupta, P., and Mäler, K.-G. (2003) The genuine savings criterion and the
value of population, Economic Theory 21(2–3), 217–225.
Dasgupta, P. (2001) Human Well-Being and the Natural Environment, Oxford: Oxford
University Press.
Ferreira, S., Hamilton, K., and Vincent, J. R. (2008) Comprehensive wealth and future
consumption: Accounting for population growth, The World Bank Economic Review
22(2), 233–248.
Ikeda, S., Tamaki, T., Nakamura, H., and Managi, S. (2017) Inclusive wealth of regions:
The case of Japan, Sustainability Science 12, 991–1006.
Mumford, K. J. (2012) Measuring inclusive wealth at the state level in the United States,
In: UNU-IHDP and UNEP (2012), 69–86.
Okubo, K. (2017) Trend of changes in Fukushima’s inclusive wealth: A preparation for
disaster recovery study of the Great East Japan earthquake, In: Managi, S. (ed.). The
Wealth of Nations and Regions, Abingdon, UK: Routledge, 186–236.
UNU-IHDP and UNEP (2012) Inclusive Wealth Report 2012. New York: Cambridge Uni-
versity Press.
UNU-IHDP and UNEP (2014) Inclusive Wealth Report 2014. New York: Cambridge Uni-
versity Press.
Yamaguchi, R. (2014) Inclusive wealth with a changing but aging population, Economics
Letters 124(1), 21–25.
Yamaguchi, R. (2018) Wealth and population growth under dynamic average utilitarianism,
Environment and Development Economics 23(1), 1–18.
Yamaguchi, R., Sato, M., and Ueta, K. (2016) Measuring regional wealth and assessing
sustainable development: An application to a disaster-torn region in Japan, Social Indica-
tors Research 129(1), 365–389.
14 The effect of landfill
gas emission on global
warming and workers’
health in Indonesia
Rafid Mahful, Moinul Islam,
Hirofumi Nakayama
and Shunsuke Managi
In developing countries, open dumping of municipal solid waste (MSW) is a

common practice. Wastes are not properly disposed of, creating environmental
problems like groundwater, surface water and soil contamination, ultimately
adversely affecting human and animal health and agriculture productivity (Staley
and Barlaz, 2009). It is in line with increasing greenhouse gases (GHGs), there-
fore favorable management is needed to decrease the effect of greenhouse gases
(Couth et al., 2011). In order to reduce GHG emissions, or to reduce the cumula-
tive global warming impact or to utilize it as a renewable energy resource, it is
desirable to manage such sites as the waste should be collected by employing suit-
able disposal methods and should recover the useful products. It is a problem also
in Indonesia, as a developing country and one of the ASEAN countries with the
highest population (about 220 million) and growth rate of 1.2% per year based on
the World Bank (2009). Indonesia has a project that is named by Clean Develop-
ment Mechanism (CDM) to reduce GHGs (Meidiana, 2012).
In this chapter, researchers use two methodologies to determine the result:
order decay model and survey. The order decay model finds the amount of meth-
ane (CH4), carbon dioxide (CO2), and hydrogen sulfide (H2S) emissions in land-
fill. The survey method goes directly to the waste pickers to find the health effect
from the emission in Tamangapa landfill.
From 2012, Tamangapa landfill has 17.39 gigagrams/year of methane emission
and it is estimated to increase to 19.17 gigagrams/year in its closure year (2031).
The methane emission at Benowo landfill began at 2.15 gigagrams/year in 2013
and it is estimated to increase to 4.25 gigagrams/year in its closure year (2028).
Carbon dioxide emission in Tamangapa landfill was 47.71 gigagrams/year in 2013
and will increase to 52.585 gigagrams/year in 2031. Carbon dioxide emission in
Benowo landfill was 5.40 gigagrams/year in 2013 and will increase to 11.39 giga-
grams/year in 2028. Hydrogen sulfide gas emission in Tamangapa landfill was
0.0026 gigagrams/year in 2013 and will increase to 0.0029 gigagrams/year in
2031. Hydrogen sulfide emission in Benowo landfill was 0.00030 gigagrams/year
in 2013, increased to 0.00063 gigagrams/year in 2018 and then will become stable
till 2030. The hydrogen sulfide gas emission is the cause of some illnesses that
Effect of landfill gas emission in Indonesia 355
occur around landfills. By surveying Tamangapa landfill, from 200 workers, 109
waste pickers complained of headache, therefore headache is the most frequent
illness experienced by waste pickers. It may be due to the time worked by the
waste pickers. Scratch is the second most frequent illness which has been experi-
enced by 88 waste pickers. This may be due to the safety tools or safety location
in Tamangapa landfill. However, the interesting trend is illness from itches and
respiratory issues, which may be due to the effect of hydrogen sulfide gas emis-
sion from Tamangapa landfill. Itches is the third most frequent illness experienced
by waste pickers in Tamangapa landfill, as 71 waste pickers have experienced this
illness. Next, respiratory illness caused from hydrogen sulfide gas emission is the
sixth most frequent illness that has occurred in the Tamangapa landfills, obtained
by 26 waste pickers. Regarding that result and surveying of expense of curing and
breaking days while waste pickers get that illness, we are able to get conclude that
waste pickers spend around USD 424.27 on itches and USD 160.09 on respiratory
illness. It really influences their income, because the majority of waste pickers (84
people) have income of USD 0.37–3.67 per day; 80 waste pickers have income
of USD 3.67–6.97 per day and 23 waste pickers have income of USD 6.97–10.27
per day. Besides, in Benowo landfill, the workers are covered by work safety, so
in this research, health damage is assumed by nothing.
Based on this result, we can conclude that Indonesia needs good waste manage-
ment with awareness for people who work or live near/on the landfill and illness
that is caused by emissions from the landfill. For instance, by using good manage-
ment system, Benowo landfill can reduce amount of methane, carbon dioxide and
hydrogen sulfide emissions and also have a good safety environmental for workers.
1 Introduction
Methane is regarded as one of the most important GHGs because its global warm-
ing potential has been estimated to be more than 20 times of that of carbon diox-
ide, and atmospheric methane concentration has been increasing in the range of
1–2% yr−1 (IPCC, 1996). Kumar et al. (2004) said that global warming potential
(GWP) of CH4 is reported at 21 times and N2O as 310 times more than CO2 over
a period of 100 years. Methane emissions from landfill are estimated to account
for 3–19% of the anthropogenic sources in the world (IPCC, 1996). Hansen et al.
(1998) showed that methane (CH4) is an important greenhouse gas, the total posi-
tive climate forcing attributed to CH4 over the last 150 years is 40% that of carbon
dioxide and in recent research showed that solid waste disposal sites are a major
anthropogenic source of methane emission (Ishigaki et al., 2008).
In developing countries, open dumping of MSW is a common practice and so
the wastes are not properly disposed of, creating environmental problems like
groundwater, surface water and soil contamination, ultimately adversely affecting
human and animal health and agriculture productivity (Staley and Barlaz, 2009).
It is in line with increasing greenhouse gases (GHGs), therefore favorable man-
agement is needed to decrease the effect of greenhouse gases (Couth et al., 2011).
In order to reduce GHG emissions, or to reduce the cumulative global warming
356 Rafid Mahful et al.
impact or to utilize it as a renewable energy resource, it is desirable to manage
such sites as the waste should be collected by employing suitable disposal meth-
ods and should recover the useful products. It is problem also in Indonesia, as
developing country and one of the ASEAN countries with the highest population
(about 220 million) and growth rate of 1.2% per year based on the World Bank
(2009). Indonesia has a project that is named by Clean Development Mechanism
(CDM) to reduce GHGs (Meidiana, 2012).
The MSW when subjected to landfills undergoes continuous degradation of
biodegradable components under anaerobic conditions resulting in the production
of recoverable biogas consisting of CH4 and CO2 as the major components (40%–
60% each) and other gases like H2S and can be used as energy source (Deonar
Pre-Feasibility Report, 2014).
LandGEM modeling software as one of software that be able to help to estimates
the volume and composition of the generated gas throughout time as a consequence
of the degradation of organic matter in the landfill (EPA, 2005). We can take a look
from other research such in Kumar (2004), Karanjekar (2015), Faour (2007), and
Kalantarifard (2012), Land GEM is used to measure CH4, CO2, and H2S in landfill
site. Zero and order decay method are the basic rules that they used to measure it.
On the other hand, in the context of municipal solid waste management
(MSWM), the informal recycling sector refers to the waste recycling activities
of scavengers and waste pickers (Wilson et al., 2006). Hydrogen sulfide gas is
one of the important gases that will influence the health of people around the
landfill. Health is a capital asset and should be seen as a component of a person’s
human capital. One way to estimate the combined benefit of improved health is by
recording people’s willingness to pay for better health (e.g., observing how much
people spend on health); health is the most significant component of the wealth of
nations (UNU-IHDP and UNEP, 2014). To understand the health damage in the
research is essential to see the effect on the environment and human surroundings.
In this chapter, the researchers show the total gas emissions from the land-
fill, such as methane (CH4), carbon dioxide (CO2) and hydrogen sulfide (H2S).
Methane and carbon dioxide emissions, which are important greenhouse gases,
and hydrogen sulfide gas cause health damage for people around the landfill. The
results are found at two landfill sites in Indonesia that have different characteris-
tics: Tamangapa landfill is an open dump and Benowo landfill is a control dump.
This study relies LandGEM 3.02 software to calculate total of methane and car-
bon dioxide which impact the environment (global warming), and also hydrogen
sulfide gas which impacts humans around the landfill site. The researchers attempt
to find out from informal sectors that are working on the sites based on the history
of their illness and the health capital will be produced from that data.
2 Municipal solid waste Indonesia

Indonesia is the largest archipelago country in the world. It has 34 provinces and
1,904,569 square kilometres for the total of area. Indonesia as developing country
has a main problem of urban waste, specifically in municipal waste of organics,
the same as other developing countries. It can be worse due to the size of society
that is causing increasing disposal every year. The problem occurs in all steps
of solid waste management (storage, collection, transferring, transporting, treat-
ment) with the tendency to rise at the end point, landfill (Meidiana, 2012). Based
on the study of waste management conducted by UNEP in 2004, it showed that
only 33% of the indicators for level of service (LoS) of MSW has been fulfilled
by the government. It means that the level of service of waste management was
still low.
The waste management in Indonesia becomes an important priority in national
policy shows when the raising number of clean development mechanism (CDM)
projects in waste sector approved by international board of CDM and proposed by
private/public sector (Hilman, 2005). A clean development mechanism (CDM) proj-
ect in the waste sector is one of the ways to reduce GHG emissions (Meidiana, 2012).
The new waste management law no.18/2008 introduced in 2008 has not been imple-
mented well causing low LoS in waste management (Meidiana and Gamse, 2010).
In this chapter, researchers attempt to take instance from two cities in Indone-
sia, Makassar and Surabaya, as the representative of waste management in Indo-
nesia. The reason of choosing these areas is that they are two of the largest cities
in Indonesia. Another reason, according to Indonesia’s Central Bureau of Statis-
tic, Surabaya is the biggest city on Java island, the capital island in Indonesia.
Surabaya had a population of 2,870,200 and 539,342.25 tons of waste in 2015.
Yet, a good thing about waste management in Surabaya is that Surabaya was
chosen as a pilot city for other cities in Indonesia. Makassar is the largest city in
East Indonesia and also the fifth largest city in Indonesia. Makassar had a popula-
tion of 1,369,606 and 1,632,045 tons of waste in 2015. Besides, Makassar has big
problem of waste management. The management of waste disposal in Makassar is
old. The country’s municipal solid waste (MSW) generation showed an increasing
trend parallel to the development of economic condition, urbanization and rapid
growth of population (Visvanathan and Trankler, 2003). The ratio between waste
and residents in Surabaya and Makassar cities are 0.19 ton/person and 1.19 ton/
person, and they increase every year.
3 Informal sector in Indonesia

The informal sector is characterized by small-scale, labor-intensive, largely unreg-
ulated and unregistered low-technology manufacturing or provision of services
(Wilson et al., 2006). In a number of countries, the informal sector also directly
provides a waste collection service in areas where there is no formal municipal
system in place (Coad, 2003; Haan et al., 1998; Scheinberg, 2001). In cities with a
formal municipal waste collection and disposal system, at least four main catego-
ries of informal waste recycling can be identified, depending on where and how
material recovery takes place (Wilson et al., 2006);
1 Itinerant waste buyers: waste collectors who often go from door to door,
collecting sorted dry recyclable materials from householders or domestic
servants, which they buy or barter and then transport to a recycling shop of
some kind.
2 Street waste picking: secondary raw materials are recovered from mixed
waste thrown on the streets or from communal bins before collection.
3 Municipal waste collection crew: secondary raw materials are recovered
from vehicles transporting MSW to disposal site.
4 Waste picking from dumps: waste pickers/scavengers sort through wastes
prior to being covered. This is often associated with communities that live
in shacks, built from waste construction materials, on or near the dump.
In a number of countries, the informal sector also directly provides a waste col-
lection service in areas where there is no formal municipal system in place (Coad,
2003; Haan et al., 1998; Scheinberg, 2001). They collect materials when they
have been discarded as waste and add value to them by sorting, cleaning, altering
the physical shape to facilitate transport or by aggregating material (Scheinberg,
2001). Commonly collected materials are plastics, paper, cardboard, aluminum,
steel, other metals, glass and textiles (Haan et al., 1998)
Informal recyclers often from discrete social groups or belong to minorities,
examples of which include the zabbaleen in Egypt; pepenadores, Catroneros and
Buscabotes in Mexico; Basariegos, Cartoneros, Traperos and Chatarresros in
Colombia; Chamberos in Ecuador; Buzos in Costa Rica; and Cirujas in Argentina
(Medina and Dows, 2000; Barthier, 2003). Many scavengers may not be able to
enter formal sector employment because of poor education or physical disability
(Wilson et al., 2006). It becomes a stronger argument when Medina (2000) and
Wilson et al. (2006) said that scavenger/waste picker income is very low, although
they are not necessarily the very poorest in society and informal recycling occurs
in developing countries because of low levels of economic development. This
condition becomes complicated when low income is due to their low position in
the trade hierarchy for recycled materials, often badly exploited and paid very low
prices for the materials (Wilson et al., 2006).
On the other perspective which is from a macroeconomic perspective, they are
well adapted to the prevailing conditions, namely abundant supply of working
force, but scarce capital: they minimize capital expenditures and maximize hand
(and animal) power (Haan et al., 1998; Scheinberg, 2001). Although alternative
employment opportunities and associated wages were higher, scavenging would
be less financially attractive (Porter, 2002). On the other hand, informal recycling
systems can bring significant economic benefits to developing countries (Wilson
et al., 2006). The informal waste recycling systems that already exist in many
developing countries reduce the cost of formal waste management systems as
they reduce the quantity of waste for collection, resulting in less money and time
spent on collection and transport (Wilson et al., 2006),
Besides, the attitude of the formal waste management sector to informal recy-
cling is often very negative, regarding it as backward, unhygienic and generally
incompatible with a modern waste management system (Wilson et al., 2006).
Scavenging in open dumps is considered to be the most detrimental to health
Table 14.1 Health effects reported from involvement in informal recycling (Eerd, 1996)
Reported outcomes of case studies

The overall respiratory illness score for children of waste-picking parents was the same
as those with non-waste-picking parents.
There was no association between below normal pulmonary function performance and
waste picking and current/past smoking.
There was no significant relation between HIV infection/HBV infection and waste
picking.
Waste picking was not associated with abnormal lung function among respondents.
More of the waste pickers reported past health problems that the control group.
Waste pickers were in a worse state of malnutrition than the control group.
In relation to the average for height and age, both groups were normal, indicating that
neither suffered from chronic malnutrition.
Many of the waste pickers suffered from chronic backache and many complained of
general weakness; coughs were a chronic problem.
Many suffered from injuries like cuts and needle stick injuries.
Eye infections and other eye problems were highly prevalent.
A few night-shift laborers from a dump complained of suffering from severe
hallucinations due to the environment they worked in.
Many of the waste pickers suffered from intestinal protozoa and helminths.
The dumps and waste bins were infested with stray dogs and rats. Bites from dogs and
rats were quite common.
Diarrhea was extremely common among all waste pickers.
Many of the waste pickers complained of having one or more attacks of jaundice in the
last year.
Many waste pickers suffered from skin disease.
(Wilson et al., 2006). Risk from manual handling of mixed waste may come, for
example, from direct contact with broken glass, human/animal fecal matter, paper
that may have become saturated with toxic materials, containers with residues of
chemicals, pesticides or solvents, and needles and bandages from hospitals. Inha-
lation of bio aerosols, and of smoke and fumes produced by open burning waste,
can cause health problems (Wilson et al., 2006). It is more relevant with the stud-
ies that come from Eerd (1996) in Table 14.1.
4 Material
In this chapter, we choose two landfills which will be investigated in Indone-
sia. One of the landfills is in a large city of West Indonesia, the capital city of
East Java which is Surabaya (Benowo landfill); the other is in the metropolitan
city of East Indonesia, the capital city of South Sulawesi which is Makassar
(Tamangapa landfill). Both islands are representative for Indonesia because they
represent two regions in Indonesia. We attempt to find out the methane (CH4),
carbon dioxide (CO2) and hydrogen sulfide (H2S) emissions from those landfills.
100˚0˚0’F 110˚0˚0’F 120˚0˚0’F 130˚0˚0’F 140˚0˚0’F

10˚0˚0’N
10˚0˚0’N
N
PETA REPUBLIK INDONESIA
L AUT CINA S E L ATAN SAMUDERA PASIFIK

0˚0˚0’N
0˚0˚0’N
I
SUMATERA KALIMANTAN SULAWESI
WES
SULA
PAPUA
L AU T
SA L AUT JAWA
MU
DE
RA JAWA L AUT FL OR E S
IN
DO Legenda
NE
10˚0˚0’N
10˚0˚0’N
SI
A Jawa Timur
Sulawesi Selatan
Kota Makassar
Kota Surabaya
100˚0˚0’F 110˚0˚0’F 120˚0˚0’F 130˚0˚0’F 140˚0˚0’F
Map 14.1 The location of landfill study cases
Therefore, we can find out amount of production from each landfill. Based on
that result, we try to find the health damage from one landfill (Tamangapa land-
fill) by hydrogen sulfide (H2S) production that affects informal workers (waste
pickers) in that landfill.
4.1 Benowo landfill

Benowo landfill in Surabaya, East Java is ±37.4 hectares in size and its final
height will be 5–12 m. The Benowo landfill began on November 2001 and it
accepts waste from the whole of Surabaya, the capital city of East Java. The land-
fill is expected to close in 2030. Figure 14.1 shows the yearly disposed quantities
of Benowo landfill.
7,00,000
598856
6,00,000 540200 531404
478816 493853 508890
5,00,000 459426 448742 463779 539342
Quantity (Tons)
467237
4,00,000
335619
3,00,000
2,00,000 204000
1,00,000
0
2003 2005 2007 2009 2011 2013 2015
Year
Figure 14.1 Yearly disposed quantities of Benowo landfill

Source: Central Bureau of Statistic, Indonesia.
Based on Figure 14.1, the graphic of yearly disposed quantities of Benowo

landfill slightly increased from 204,000 tons/year in 2003 to 598,856 tons/year in
2006. Then, it decreases to 459,426 tons/year in 2008. The unusual trend found
from this diagram is the trend to not constantly increase or decrease in 2008 to
2015; it increases till 539,342 tons/year.
Surabaya accepted the Nirwasita Tantra trophy, which is an achievement from
the Indonesia government as the achievement in management of environment
that the best value is from management of municipal solid waste, said Musdia
Ali, head of the Surabaya environment department (Detiknews, 2017). Therefore,
Surabaya landfill in this instance is the result of good working in management of
municipal solid waste in Indonesia.
4.2 Tamangapa landfill

Tamangapa landfill in Makassar, South Sulawesi, is ±14.3 hectares in size and
its final height will be 5–12 m. Most of the areas/phases have been capped or are
almost at full capacity. However, the city is currently in negotiations regarding
expansion of this landfill. Tamangapa landfill began in 1993 and it accepts waste
from the whole of Makassar. The landfill is expected to close in 2032. Table 14.2
shows the yearly disposed quantities of Tamangapa landfill.
Based on Figure 14.2, the trend of yearly disposed quantities of Tamangapa
landfill increases every year. From 2003 until 2015 the end of the data, the dis-
posed quantities increased until 604,550 tons. It is a small differences from
Benowo landfill because the trend of waste production has decreased and then
increased in the small amount.
The amount of municipal solid waste in Tamangapa landfill constantly increased
by a large amount waste, every year. The problem of waste management is explained
by Sakka, the head of waste management in Tamangapa landfill. This case is more
complicated when the landfill is over capacity and the waste is only stacked.
18,00,000 1581666
1531287
16,00,000 1432085
1279391
14,00,000 1632045
1077874 1178633 1480907
Quantity (Tons)
12,00,000 1329770 1380149

10,00,000 1027495 1128254 1229012
8,00,000
6,00,000
4,00,000
2,00,000
0
2003 2005 2007 2009 2011 2013 2015
Year
Figure 14.2 Yearly disposed quantities of Tamangapa landfill

Source: Central Bureau of Statistic, Indonesia.
Therefore, these landfills can be compared between good waste management

and the opposite. It can be a comparison to show the faced problem, specifically
in each landfill.
5 Modeling/Method
5.1 Landfill gas generation model (LandGEM software 3.02)

Landfill Gas Emission Model (LandGEM) software is used by estimating emission
rates for methane (CH4), carbon dioxide (CO2) and hydrogen sulfide (H2S) from
municipal solid waste landfills. The US Environmental Protection Agency (EPA)
is the organization that develops LandGEM software. The model determines the
mass of methane generated using the methane generation capacity and the mass
of waste deposited, and the Intergovernmental Panel on Climate Change (IPCC)
is the international body for assessing the science related to climate change. The
IPCC was set up in 1988 by the World Meteorological Organization (WMO) and
the United Nations Environment Program (UNEP) to provide policymakers with
regular assessments of the scientific basis of climate change, its impact and future
risks, and options for adaption and mitigation.
n n
QCH4 = ∑∑ KLo (Mi / 10)e

i=1 j =0.1
− KTij
Where QCH4: annual CH4 generation (mg/yr); i = 1 year time increment; n: (year
of the calculation) – (initial year of waste acceptance); j: 0.1 year time increment;
k: CH4 generation rate(year-1); Lo: potential CH4 generation capacity (m3/Mg);
Mi: mass of waste accepted in the ith year (Mg); tij: age of the Jth section of waste
mass Mi accepted in the ith year.
Table 14.2 Input review (assumption) for Benowo and Tamangapa landfills in LandGEM
3.02 software
Tamangapa landfill Benowo landfill
Landfill characteristic
Landfill open close year 1993–2032 2001–2030
Gas pollutant
Methane 16.04
Carbon dioxide 44.01
Hydrogen sulfide
Model parameter
Methane generation rate 0.7 0.7
Potential methane generation capacity 170 m3/mg (arid area) 170 m3/mg
NMOC concentration ppmv
Methane content 50% by volume
5.2 Survey
A field survey of health damage was conducted on October 1–15, 2017, at Taman-
gapa landfill, Makassar, South Sulawesi, for the informal sector around the land-
fill. Tamangapa landfill is owned and operated by the Parks and Sanitation Agency
of the municipal government of Makassar city. The purposing sampling technique
was taken from 200 waste pickers who work or pick waste and live in the Taman-
gapa landfill area. The sample is taken from different genders, ages and side jobs.
All data were obtained through the field work in the Indonesian language. The
study also made use of extensive collection, analysis and review of secondary
data to substantiate the primary data, and these were mainly derived from relevant
academic articles, books, newspapers and magazines, and previous studies.
6 Result
6.1 Environmental impact
6.1.1 Landfill gas emission (methane, carbon dioxide,

and hydrogen sulfide emission)
Tamangapa and Benowo landfills are the two landfill sites in this research. The
location and basic knowledge of both were described in Section 1. The differences
between both of them are annual precipitation and temperature. Benowo landfill
is in an arid area (based on LandGEM parameter) because it has high temperature
and less rainfall per year (less than 25 inches). Besides, Tamangapa landfill is in a
conventional area (based on LandGEM parameter) because the temperature is not
too high and the rainfall per year is more than 25 inches. Table 14.3 shows salient
features of Benowo and Tamangapa landfills.
Table 14.3 Salient features of Benowo and Tamangapa landfills, Indonesia
Features Detail
Benowo Tamangapa
Location:
• Latitude 7o13′9.70″ S 5o10′36.02″ S
• Longitude 112o37′52.18″ E 119o29′25.35″ E
Year of start 2001 1993
Year of closure 2032 2030
Area (ha) 37.4 14.3
Annual precipitation 141.1 mm 2186 mm
Temperature (ºC) 35 27
17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4 17.4
17.2 17.3
17.5 16.9
17
GIGAGRAM
16.4
16.5
15.8
16
15.3
15.5
15
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032
YEAR
Figure 14.3 Methane emission calculation from Tamangapa landfill

Source: Calculation from LandGEM 3.02.
5.74 5.75 5.75 5.76 5.76 5.76 5.76 5.76 5.76 5.76 5.76 5.76
5.80 5.69 5.72
5.70 5.62
Gigagrams
5.60 5.49
5.50
5.40 5.30
5.30
5.16
5.20
5.10
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030
Year
Figure 14.4 Methane emission calculation at Benowo landfill

In the present study, the CH4 emission calculation from Tamangapa and Ben-
owo landfills and its annual trend are given in Figures 14.3 and 14.4.
The amount of methane emission from Tamangapa landfill increases from 2013
to 2019 and then becomes stable in 2019 till the close year. In 2014, the methane
emission from Tamangapa landfill is 15.8 gigagrams/year. From the results, we
attempt to compare with recent studies (Lando et al., 2017), which indicated that
the calculation from IPCC 2006 from 2010 is 18 gigagrams/year and stable till
the closure year of Tamangapa landfill. The result of the LandGEM software 3.02
calculation is close, yet LandGEM gives more of a trend for the methane emis-
sions to fluctuate. Due to the result from LandGEM calculation, the closure year
rate is near 18 gigagrams/year, which is 17.4.
From Figure 14.4, we are able to see that the amount of methane production
from Benowo landfill is less than Tamangapa landfill. It is caused by the amount
of waste and the rainfall in Benowo landfill is small. The methane emission began
at 5.16 gigagrams/year in 2013 and it increases to 5.75 in 2020, then becomes
stable until the closure year. There is small difference in the trend that tamangapa
landfill, when Tamangapa landfill stop to add the amount of methane emission
(CH4) in 2019.
On the other hand, there are other emission that will be calculated in this chap-
ter. This chapter also discusses about the carbon dioxide (CO2) emission that pro-
duced from both Benowo and Tamangapa landfills. The amount of carbon dioxide
is calculated from LandGEM software as well. The form of carbon dioxide (CO2)
production conducts from LandGEM formula as below,
  
  −1
QCO2 = QCH 4 x  1




 ( PCH 4
/ 100 )  
 
Where QCO2 is the production of carbon dioxide, QCH 4 is production of methane,

and PCH 4 is the methane content percentage.
From that form calculation, this research finds out carbon dioxide production
from Benowo and Tamangapa landfills. The results of calculation are shown
below.
From Figure 14.5, we can see that amount of carbon dioxide emission or CO2
is more than two times methane emission. Carbon dioxide is 41.90 gigagrams/
year in 2013 and then it increases to 47.80 gigagrams/year in 2033. The amount of
carbon dioxide is stagnant at 47.80 gigagrams/year until the closure year in 2032.
It is smaller than the methane emission which is stagnant from 2019.
Based on Figure 14.6, we can see that the carbon dioxide (CO2) in Benowo
landfill is also less than the amount of carbon dioxide (CO2) in Tamangapa land-
fill. The differences are more than two times. The amount of carbon dioxide in
0
0
0
0
0
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.7
0
.6
.5
.1
0
47
47
47
47
47
47
47
47
47
47
47
47
47
47
47
.4
48.00
47
46
0
Gigagrams
.9
46.00
44
0
.4
43
0
44.00
.9
41
42.00
40.00
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032
Year
Figure 14.5 Carbon dioxide emission calculation at Tamangapa landfill

0
0
0
0
0
0
0
0
9
9
8
5
1
.8
.8
.8
.8
.8
.8
.8
.8
.7
.7
1
.7
.7
.7
.6
3
16.00
15
15
15
15
15
15
15
15
15
15
15
15
15
.4
Gigagrams
15
5
15
15.50
.0
15
3
15.00
.5
5
14
14.50 .1
14
14.00
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030
Year
Figure 14.6 Carbon dioxide emission calculation at Benowo landfill

0. 67
0. 67
0. 67
0. 67
00 7
0. 67
0. 67
0. 67
0. 67
7
0. 66
0. 66
0. 66
0. 66
0. 65
0. 63
26
26
0. 58
2
2
2
2
2
2
2
2
2
2
2
2
2
2
00
00
00
00
00
00
00
00
00
0.00280
00
00
00
00
00
0. 50
00
2
Gigagrams
00
0.
0. 42
2
00
0. 34
0.00260
2
00
2
00
0.00240
0.
0.00220
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032
Year
Figure 14.7 Hydrogen sulfide emission calculation at Tamangapa landfill

Benowo landfill held 14.15 gigagrams/year and then it will increase to 2021
become 15.79 gigagrams/year and continue stagnant till closure year in 2030.
From that data, methane emission (CH4) emission and carbon dioxide (CO2)
emission that occur in Benowo landfill, the trend is unique. In the trend of meth-
ane emission (CH4) of Benowo landfill, the trend enhancement is stable in 2020,
although for the trend of carbon dioxide (CO2) of Benowo landfill, the trend climb
is stagnant in 2021 so there is a year different for the enhancing gas emission.
This chapter also attempts to track hydrogen sulfide gas that is related to health
damage in each landfill. For those cases, the researchers took samples from people
who worked at the landfill area.
Using the landfill gas emission model (LandGEM) calculates the hydrogen sul-
fide (H2S), which is the most crucial gas in health from municipal solid waste.
This gas can causing harmful health for society and environment surrounding the
landfill. Next, the result of this calculation will compare to survey result to know
the effect of informal sector in landfills, specifically in Tamangapa landfill.
From Figure 14.7, we can see that the trend of hydrogen sulfide emission is
similar with methane emission. Hydrogen sulfide, as well, stagnated at 0.00266
gigagrams/year in 2019 to the closure year, while it is 0.00234 gigagrams/year
and then increased until 0.00263 gigagrams/year before stagnating at 0.00266
gigagrams/year in 2019.
The unique trend is shown by the result of hydrogen sulfide (H2S) emission pro-
duction in Benowo landfill. The enhancement of the hydrogen emission (H2S) is
0. 88
0. 88
0. 88
0. 88
0. 88
0. 88
0. 88
8
0. 88
0. 88
0. 88
0. 88
0. 88
0. 87
0. 86
08
0. 84
0
0
0
0
0
0
0
0
0
0
0
0
0. 81
0
0
0. 79
00
00
00
00
00
00
00
00
00
00
00
00
00
0
00
00
0
00
Gigagrams
0
00
0.00090
00
0.
0.00080
0.00070
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030
Year
Figure 14.8 Hydrogen sulfide emission calculation at Benowo landfill

stopped, increasing early than methane (CH4) and carbon dioxide (CO2) emission.
The hydrogen emission occurred 0.00079 gigagrams/year in 2013, it smaller than
hydrogen emission (H2S) occurred in Tamangapa landfill at that same time. The
hydrogen sulfide increases 0.00088 gigagrams/year in 2018 and then becomes
stable till 2030, a little different than methane emission which stabilizes in 2020
and carbon dioxide emission which stabilizes in 2021.
The amount emission of hydrogen sulfide (H2S) in Benowo landfill is less than
three times that of Tamangapa landfill. From the final result that found from cal-
culation of methane emission (CH4), carbon dioxide (CO2), and hydrogen sulfide
(H2S) that occurs in both landfills, we can reach the conclusion that the amount of
waste and the area also influence the emission that occurs in each landfill.
6.2 Human effect

From the survey that is collected from Tamangapa landfill, Makassar, South
Sulawesi, Indonesia, the research finds that there are several illnesses that have
been obtained from waste pickers from around the landfill, including swelling,
sprains, itches, scratches, dehydration, heatstroke, respiration, headache, vomit-
ing and aches. In total, there are 10 illnesses that have been obtained from waste
pickers.
From 200 workers, 109 waste pickers obtained headaches, so headache was the
most frequent illness obtained. It may be caused by the time worked by the waste
pickers. Scratch was the second most frequent illness, attributed to 88 waste
pickers. It may be caused by the safety tools or safety location in Tamangapa
landfill. The health and safety risks associated with informal recycling include
(1) occupational health risks posed to scavenger/waste pickers and (2) commu-
nity health risks posed to the related community or general public (Cointreau,
n.d.) (see Table 14.4).
However, the interesting trend occurs from itches and respiratory illness, the
illnesses that may be occurred from the effect of hydrogen sulfide emission
from Tamangapa landfill. Itches was the third frequent illness occurring among
waste pickers in Tamangapa landfill, with 71 having obtained this illness. Respi-
ratory illness that may be caused from hydrogen sulfide emission has becomes
the sixth most frequent illness that has occurred in Tamangapa landfills, with
109
105 88
85 71
People
65
45 38
26 26 28
25 19
9 9
5
n
in
ke
he
ng
s
he
he
le
tc
io
tio
ra
ro
ac
iti
ol
ra
at
itc
ira
ac
sp
m
st
sw
dr
ad
sc
vo
at
sp
hy
he
he
re
de
Illness
Figure 14.9 Amount of illness suffered by informal workers

Source: Data of survey on Tamangapa landfill in 2017.
Table 14.4 Health and safety risk associated with informal recycling
Origin of risk factor Examples of source of possible risk

Composition of waste Toxic, allergenic and infectious components including gases, dust,
leachate, sharps, broken glass
Nature of organic Gaseous emissions, bio aerosols, dust, leachate, fine particle sizes
decomposing waste and their change in ability to cause a toxic, allergenic or infectious
health response
Handling of waste Working in traffic, shoveling, lifting, equipment vibrations, accidents
Processing of waste Odor, noise, vibration, accidents, air and water emissions,
residuals, explosions, fires
Disposal of wastes Odor, noise, vibration, stability of waste piles, air and water
emissions, explosions, fires
Source: Cointreau (n.d.).
26 waste pickers having obtained this illness. The occupational health risks to
waste pickers in developing countries are high because of manual handling and
lack of protective clothing/equipment, resulting in direct contact with waste
(Cointreau, n.d.).
From this result, we can see that hydrogen sulfide emission may become
effected for people who working in Tamangapa area. Although the effect from
hydrogen sulfide is only found in two illnesses, more investigation is needed.
6.3 Economic impact

The inclusive wealth report project tracks changes in three categories of capital
assets: human capital, produces capital, and natural capital. This chapter attempts
closer with health capital which part of human capital. It directs to human well-
being rather than through the production of goods and services that are counted
in GDP
Good health brings three benefits to a person:
1 It adds directly to the person’s well-being

2 It enables the person to be productive (a healthy person works better and
can work for longer hours than an unhealthy person)
3 It contributes to her longevity (a healthy person can be expected to live
longer than an unhealthy person).
Economists have developed elaborate method for estimating the value of each
type of benefit. For this chapter, the researcher connects hydrogen sulfide for cal-
culating the health capital. Hydrogen sulfide is one landfill gas that is of concern
to health.
Based on Figure 14.11, the result show that some of illness can be influenced
by hydrogen sulfide (H2S). Maybe it is caused by working at the landfill. The
researchers attempt to find the cost of illness that occur. Therefore, we can find the
comparison between the expense for each illness from waste pickers at the landfill.
Figure 14.10 shows the results of a questionnaire that the researcher gave to waste
pickers. The questionnaire asks the cost that waste pickers spent to cure the illness
and how many days waste pickers get off from work because of their illness.
Regarding the diagram, we are able to get conclude that waste pickers spend
around USD 424.27 from itches and USD 160.09 to respiration.
Illness really influences the income of the waste pickers. The majority of waste
pickers (84 people) only have income of USD 0.37–3.67 per day, 80 waste pick-
ers have income of USD 3.67–6.97 per day and 23 waste pickers have income of
USD 6.97–10.27 per day. To conclude, the majority of income waste pickers per
day under USD 10.27 per day is 187 waste pickers, or around 90% from the waste
picker samples.
700 611.1
600
481.11
500 424.27
400
USD
297.29
300
200 135.03 164.22 160.09
100 59.57
0
s
in
ch
n
he
ke
n
he
n
le
ra
tio
io
at
ol
ro
ac
itc
sp
at
r
ra
sw
sc
st
ad
dr
pi
at
hy
he
s
he
re
de
Illness
Figure 14.10 Amount of illness spend money

Source: Data of survey on Tamangapa landfill in 2017
Note: USD 0.000074 = IDR 1 (13/12/2017)
Figure 14.11 Amount of informal worker salaries

Source: Data of survey on Tamangapa landfill in 2017
Note: USD 0.000074 = IDR 1 (13/12/2017)
7 Conclusion
On conclusion, this research finds that methane emission in Tamangapa and Ben-
owo landfill, from 2013 to the closure year each landfill is the methane emission
from Tamangapa landfill is 15.8 gigagrams/year and the amount of methane emis-
sion increase being 17.4 gigagrams/year in 2019 until closure year. However, in
Benowo landfill, the methane emission in started 5.16 gigagrams/year in 2013 and
it increases 5.75 in 2020, then become stable until the closure year. The amount
of carbon dioxide emission or CO2 is more than two times methane emission in
Tamangapa landfill. Carbon dioxide is 41.90 gigagrams/year in 2013 and then
it increases to 47.80 gigagrams/year in 2033. The amount of carbon dioxide is
stagnant in 47.80 gigagrams/year until the closure year in 2032. On the other
hand, by small different than methane emission which is stagnant from 2019.
The amount of carbon dioxide in Benowo landfill held 14.15 gigagrams/year and
then it will increase to 2021 become 15.79 gigagrams/year and continue stagnant
till closure year in 2030. Hydrogen sulfide in Tamangapa landfill, as well, stag-
nates in 0.00266 gigagrams/year in 2019 to the closure year, while it is 0.00234
gigagrams/year and then increases until 0.00263 gigagrams/year before stagnant
in 0.00266 gigagrams/year in 2019. Besides, the hydrogen emission in Benowo
landfill occurred 0.00079 gigagrams/year in 2013, it is smaller than hydrogen
emission (H2S) occurred in Tamangapa landfill at that same time. The hydrogen
sulfide increases 0.00088 gigagrams/year in 2018 and then become stable till
2030. The hydrogen emission or H2S is caused of some illness that occur to sur-
round landfill such itches therefore by surveying on Tamangapa landfill, from 200
workers, 109 waste pickers have obtained headache, it will make headache as the
most frequently illness that is obtained by waste pickers. It may be caused of the
time works of the waste pickers. Scratch becomes the second most frequently
illness which is 88 waste picker been felt this illness. It may be caused of the
safety tools or safety location in Tamangapa landfill. However, the interesting
trend is occurred from itches and respiration, the illness that may be occurred
from the effect of hydrogen sulfide emission from Tamangapa landfill. Itches ill-
ness becomes third frequently illness that has been occurred from waste pickers
in Tamangapa landfill, where 71 waste pickers have obtained this illness. Next,
respiration illness that may be caused from hydrogen sulfide emission, as well
becomes sixth frequently illness that is occurred in Tamangapa landfills. It is
obtained by 26 waste picker in Tamangapa landfill. Regarding on that result and
surveying of expense of curing and breaking days while waste pickers get that
illness, we are able to get conclude that waste pickers spend around USD 424.27
from itches and USD 160.09 to respiration. It really influences their income due
to majority of waste picker which 84 people only have income USD 0.37–3.67 per
day, 80 societies from the waste picker samples have income USD 3.67–6.97 per day
and USD 6.97–10.27 per day which is 23 people.
Based on this result, we can conclude that Indonesia needs good waste manage-
ment with aware for people who work or live near/on the landfill and illness that
is caused by emission from landfill. Need to more investigation for health risk
that occur from the landfill to people surround that, to find more result and more
conclusion based on data.
Bibliography
Barton, J. R., Issaias, I., and Stentiford, E. I. (2008). Carbon-Making the right choice for
waste management in developing countries. Waste Management, 28(4), 690–698.
Berthier, H. C. (2003). Garbage, work and society. Resources, Conservation and Recycling,
39(3), 193–210.
Bleck, D., and Wettberg, W. (2012). Waste collection in developing countries: Tackling
occupational safety and health hazards at their source. Waste Management, 32(11),
2009–2017.
Bogner, J., and Matthews, E. (2003). Global methane emissions from landfills: New meth-
odology and annual estimates 1980–1996. Global Biogeochemical Cycles, 17(2).
Chiemchaisri, C., Juanga, J. P., and Visvanathan, C. (2007). Municipal solid waste manage-
ment in Thailand and disposal emission inventory. Environmental Monitoring and
Assessment, 135(1), 13–20.
Coad, A. (2003). Solid waste collection that benefits the urban poor – Suggested guidelines
for municipal authorities under pressure. Prepared for the Collaborative Working Group
on solid waste management in low- and middle-income countries by the SKAT Founda-
tion, St Gallen, Switzerland. http://www.skat-foundation.org/
Couth, R., Trois, C., Parkin, J., Strachan, L. J., Gilder, A., and Wright, M. (2011). Delivery
and viability of landfill gas CDM projects in Africa – A South African experience.
Renewable and Sustainable Energy Reviews, 15(1), 392–403.
Detiknews, Z. E. (2017). https://news.detik.com/berita-jawa-timur/d-3583744/surabaya-
sabet-empat-penghargaan-lingkungan-tingkat-nasional
Eerd, V. (1996). The occupational health aspects of waste collection and recycling. A sur-
vey of the literature. WASTE Working Document 4, part 1, Urban Waste Expertise Pro-
gram (UWEP). http://www.waste.nl/page/558; accessed in 2004.
EPA (2005). Landfill Gas Emissions Model (LandGEM) version 3.02 user’s guide.
Faour, A. A., Reinhart, D. R., and You, H. (2007). First-order kinetic gas generation model
parameters for wet landfills. Waste Management, 27(7), 946–953.
Ferita, H. D. (2006). City Report of Surabaya. In AUICK First 2006 Workshop. Surabaya,
Indonesia.
Hansen, L. B., Finster, K., Fossing, H., and Iversen, N. (1998). Anaerobic methane oxida-
tion in sulfate depleted sediments: Effects of sulfate and molybdate additions. Aquatic
Microbial Ecology, 14(2), 195–204.
Hilman, M. (2005). “Prospect of municipal waste management through CDM Project Indo-
nesia”, Proceeding of International Workshop on Landfill Gas development and The
CDM, September 5–7, Bali, Indonesia.
Haan, H. C., Coad, A., and Lardinois, I. (1998). Municipal waste management: Involving
micro-and-small enterprises. Guidelines for municipal managers. Turin, Italy: Interna-
tional Training Centre of the ILO, SKAT, WASTE. http://www.skatfoundation.org/
publications/waste.htm
IPCC (1996). IPCC Second Assessment on Climate Change. International Panel on Cli-
mate Change. Cambridge, UK: Cambridge University Press.
Ishigaki, T., Van Chung, C., Sang, N. N., Ike, M., Otsuka, K., Yamada, M., and Inoue, Y.
(2008). Estimation and field measurement of methane emission from waste landfills in
Hanoi, Vietnam. Journal of Material Cycles and Waste Management, 10(2), 165.
Johari, A., Ahmed, S. I., Hashim, H., Alkali, H., and Ramli, M. (2012). Economic and
environmental benefits of landfill gas from municipal solid waste in Malaysia. Renew-
able and Sustainable Energy Reviews, 16(5), 2907–2912.
Kalantarifard, A., and Yang, G. S. (2012). Estimation of methane production by LAND-
GEM simulation model from Tanjung Langsat municipal solid waste landfill, Malaysia.
International Journal of Science and Technology, 1(9), 481–487.
Karanjekar, R. V., Bhatt, A., Altouqui, S., Jangikhatoonabad, N., Durai, V., Sattler, M. L., . . ., and
Chen, V. (2015). Estimating methane emissions from landfills based on rainfall, ambient tem-
perature, and waste composition: The CLEEN model. Waste Management, 46, 389–398.
Kaushal, A., and Sharma, M. P. (2016). Methane emission from panki open dump site of
Kanpur, India. Procedia Environmental Sciences, 35, 337–347.
Kumar, S., Gaikwad, S. A., Shekdar, A. V., Kshirsagar, P. S., and Singh, R. N. (2004). Esti-
mation method for national methane emission from solid waste landfills. Atmospheric
Environment, 38(21), 3481–3487.
Lando, A. T., Nakayama, H., and Shimaoka, T. (2017). Application of portable gas detector
in point and scanning method to estimate spatial distribution of methane emission in
landfill. Waste Management, 59, 255–266.
Matter, A., Dietschi, M., and Zurbrügg, C. (2013). Improving the informal recycling sector
through segregation of waste in the household: The case of Dhaka Bangladesh. Habitat
International, 38, 150–156.
Medina, M., and Dows, M. (2000). A short history of scavenging. Comparative Civiliza-
tions Review, 42, 7–17.
Medina (2000). Scavenger cooperatives in Asia and Latin America. Resources, Conserva-
tion and Recycling, 31(1) (2000), 51–69.
Meidiana, C. (2012). Scenarios for sustainable final waste treatment in developing country.
In Waste Management-An Integrated Vision. Intech Publisher: Croatia.
Meidiana, C., and Gamse, T. (2010). Development of waste management practices in Indo-
nesia. European Journal of Scientific Research, 40(2), 199–210.
Nzeadibe, T. C., and Anyadike, R. N. (2012). Social participation in city governance and
urban livelihoods: Constraints to the informal recycling economy in Aba, Nigeria. City,
Culture and Society, 3(4), 313–325.
Ojeda-Benitez, S., Armijo-de-Vega, C., and Ramı́rez-Barreto, M. E. (2002). Formal and
informal recovery of recyclables in Mexicali, Mexico: Handling alternatives. Resources,
Conservation and Recycling, 34(4), 273–288.
Oteng-Ababio, M. (2012). The role of the informal sector in solid waste management in the
GAMA, Ghana: Challenges and opportunities. Tijdschrift voor economische en sociale
geografie, 103(4), 412–425.
Paul, J. G., Arce-Jaque, J., Ravena, N., and Villamor, S. P. (2012). Integration of the infor-
mal sector into municipal solid waste management in the Philippines: What does it need?
Waste Management, 32(11), 2018–2028.
Porter, C.R. (2002) The Economics of Waste. Washington, DC: Resources for the Future
(RFF) Press.
Sasaki, S., and Araki, T. (2013). Employer-employee and buyer-seller relationships among
waste pickers at final disposal site in informal recycling: The case of Bantar Gebang in
Indonesia. Habitat International, 40, 51–57.
Scheinberg, A., Spies, S., Simpson, M. H., and Mol, A. P. (2011). Assessing urban recycling
in low-and middle-income countries: Building on modernised mixtures. Habitat Inter-
national, 35(2), 188–198.
Scheinberg, A. (2001). Financial and economic issues in integrated sustainable waste man-
agement. Tools for decision-makers. Experiences from the Urban Waste Expertise Pro-
gramme. The Netherlands: WASTE. http://www.waste.nl/page/525
Schenck, R., and Blaauw, P. F. (2011, December). The work and lives of street waste pick-
ers in Pretoria: A case study of recycling in South Africa’s urban informal economy. In
Urban Forum (Vol. 22, No. 4, p. 411). Netherlands: Springer.
Shekarchian, M., Moghavvemi, M., Mahlia, T.M.I., and Mazandarani, A. (2011). A review
on the pattern of electricity generation and emission in Malaysia from 1976 to 2008.
Renewable and Sustainable Energy Reviews, 15(6), 2629–2642.
Staley, B. F., and Barlaz, M. A. (2009). Composition of municipal solid waste in the United
States and implications for carbon sequestration and methane yield. Journal of Environ-
mental Engineering, 135(10), 901–909.
UNU-IHDP and UNEP (2014). Inclusive Wealth Report 2014: Measuring Progress Toward
Sustainability. New York: Cambridge University Press.
Visvanathan, C., and Trankler, J. (2003). Municipal solid waste management in Asia: A
comparative analysis. In Workshop on Sustainable Landfill Management, 3–5.
Wilson, D. C., Velis, C., and Cheeseman, C. (2006). Role of informal sector recycling in
waste management in developing countries. Habitat International, 30(4), 797–808.
World Bank (2009). World Development Report 2009: Reshaping Economic Geography.
World Bank Publications.
Zhao, W., Van Der Voet, E., Zhang, Y., and Huppes, G. (2009). Life cycle assessment of
municipal solid waste management with regard to greenhouse gas emissions: Case study
of Tianjin, China. Science of the Total Environment, 407(5), 1517–1526.
15 Human capital change and
social impact under China’s
universal two-child policy
Qiuyi Chen, Moinul Islam
and Shunsuke Managi
1 Introduction
At the end of 2015, the Chinese government published the universal two-child
policy; later it was officially implemented from 2016. The two-child policy sym-
bolized the end of the one-child policy which had lasted over three decades, and
was expected to bring a new atmosphere to Chinese society. A brief review of
China’s half a century of population policy history includes the “Later, longer,
fewer” policy (get married later, wait longer between births, have fewer children)
in 1971; the one-child policy in 1980; the separate two-child policy (couples can
have two children if both parents are only children) in 2011; and selective two-
child policy (couples can have two children if one of the parents is an only child)
in 2013. It was an iterative process before the first population policy was made
during the initial stage of new China. The purpose of population policies in the
past was to control population growth in the underdeveloped economy, while the
universal two-child policy today is meant to encourage births.
As we all know, the one-child policy was considered one of the most contro-
versial policies in the world. The person who came up with this policy was not
a demographer, but policymakers in China thought that the Chinese government
had to control population before it was too late. This was astonishing for most
Chinese people at first because the preference of many children in a family was
part of traditional Chinese family culture. In the primary stage of new China,
studies about demographics were inadequate, and policymakers were eager to
solve the problem in front of them. However, we do not intend to discuss how the
one-child policy was made, nor whether it was right. We want to show what the
one-child policy brought to China, which is helpful to understand why the two-
child policy came out.
From most perspectives, the one-child policy did succeed in population control,
reduced financial burdens, improved social status of women, and so forth. On
the contrary, it accelerated aging and allowed China to become an aging society
before it achieved high economic development – called “grow old before getting
rich”. Low fertility rate, lack of labor, disappearance of demographic dividend,
imbalance of gender structure, pension problems and care for elderly people – due
to all these issues China faced, the one-child policy was replaced by the two-child
China’s universal two-child policy 375
policy that aimed at alleviating these serious situations and mitigating the conse-
quences left by the one-child policy.
There are two main opinions about the two-child policy: one is that the Chinese
government should rescind the two-child policy, stop making any further popula-
tion policy, and let couples have as many children as they want; another is that
the two-child policy will lead to a population boom again and cause new social
problems. Also, some critics think the two-child policy won’t work because it is
too late for China to take action to raise the fertility rate, and the population is
not the key problem anymore. At the beginning, experts opposed to the two-child
policy forecasted that 41 million more babies would be born every year since the
implementation of the new two-child policy, but in reality it is far less than that
number. Will the two-child policy effectively change the demographic structure
of China and solve China’s problems? These are very interesting questions to
explore.
The one-child policy has drawn wild attention by a number of researchers
who have analyzed all kinds of topics about its impact. China’s fertility rate has
reduced significantly since the implementation of the one-child policy, yet the
thought of some researchers that the one-child policy was the main reason causing
the decrease in the fertility rate still remains doubtful because we cannot compare
the actual situation with and without the population policy. For most people, it is
common sense that the fertility rate will go down naturally and this reduction is
an inevitable trend of development of human society. Population policy can affect
every aspect of people’s life such as economics, society, environment, resources,
and so forth. We believe the impact of the one-child policy remains; that is why
we want to know if the two-child policy is able to offset the negative effects of
the one-child policy.
A study by John Bongaarts and Susan Greenhalgh (1985) gives proof that
China should have implemented a two-child policy instead of a one-child policy
at the beginning. Researchers at that time already showed the relatively suitable
option. Yi Zeng and Therese Hesketh (2016) summarized the background of the
one-child policy and demonstrated various effects under the two-child policy.
Our purpose in this chapter is to explore the influence of the two-child policy in
many respects, particularly on human capital and demographic changes. Since
the two-child policy is totally new, related research at the present is deficient,
and it is impossible to get empirical evidence and data in a short time. Even
if we look deep into the research about the one-child policy, they are almost
always focused on the influence on economy or society; few relate to human
capital. The consequences of the two-child policy can’t be seen evidently for at
least 20 years until the newborn babies are old enough to join the labor market
and social life. But before that, growing newborn babies still have significant
effects on the economy in the short term. We believe that doing research on
population policy in China is meaningful because China is a country with a
large population.
It seems apparent that two-child policy will accelerate population growth,
though perhaps not markedly. Population growth means much labor and
376 Qiuyi Chen et al.
productivity improvement for society, and more babies mean that more baby care
products and care for women after giving birth are needed. However, more con-
sumption of natural resources and emission can be harmful to the environment
and affect health capital. For economic development, it is important to maintain
the rate of population increase, especially in developing countries. China is the
largest country by population, so even if the Chinese government had not made
the new two-child policy, the question is how many far-reaching impacts the
policy would bring. Also, population growth could contribute to human capital,
which until now we have estimated by education capital based on the method in
Inclusive Wealth Report 2014.
Nowadays, people are always concerned about gross domestic product (GDP)
growth, how much we produce and how much we consume. But we can’t tell
how much water is available just by measuring what’s coming out of a faucet,
according to IWR 2014. What we use to produce is the foundation of our soci-
ety. For the whole country, whether human capital will decrease or increase is
an important issue. We try to include health capital in the estimation of human
capital since population policy will affect education capital and health capital,
either directly or indirectly. However, due to the lack of data and method, we
simplified the calculation to education capital, but we will discuss health capital
as well.
The two-child policy is already the hottest topic in China, so we want to explore
how the two-child policy will influence human capital and what kind of social
impact this new population policy will bring. In this chapter, we forecast how
the two-child policy affects total population based on census data and several
surveys, then analyze the change of human capital and also use the results of
demographic change to discuss the future social impact. Estimating the change in
human capital and analyzing social impact based on the demographic change in
the future are the two main parts of our work. To estimate the change in human
capital in the second part, we need some demographic forecast data such as popu-
lation that linked to the first part. And in the third part, we explore deeper into
the demographic situation, including gender and age structure, and discuss the
social impact in the short term and mid- to long term. For human capital calcula-
tions, according to Inclusive Wealth Report 2014, we follow Arrow et al. (2012)
and Klenow and Rogríguez-Clare (1997). For forecasts on demographic change,
we follow the cohort-component method. In the Inclusive Wealth Report 2014,
researchers made calculations on the national level from 1990 to 2010. In our
work, we estimate the change of human capital in China of every province except
Hong Kong, Macao, and Taiwan.
To discuss the desire of Chinese people to have two children, we analyze the
incentives from the results of different surveys in different years, both official
and unofficial. The decrease in fertility rate worldwide is also one of our topics
of interest, so we compare China with another country with a population policy
to give a broader vision on population control. The rest of this chapter includes
three parts. In Section 2, we will introduce the data resources and method we use
in detail. In Section 3, which is the most important part, we first show the results
of our forecast on population and human capital change, then discuss the Chinese
characteristics and social problems. Section 4 discusses issues and challenges we
faced and the conclusion.
2 Data and methodology
2.1 Data
Our main data sources are National Bureau of Statistics of China and World
Bank. We mainly collect data such as population, mortality, and education data
from the census and statistical yearbook. The fertility rate is from National
Bureau of Statistics of China and the national survey of childbearing desire
conducted by National Health and Family Planning Commission of the Peo-
ple’s Republic of China. As some data is incomplete, we use interpolation and
extrapolation for estimation purposes. And because the data from childbear-
ing desire survey is not new enough, we made some adjustments according to
the current trend. The adjustment refers to the Statistical Bulletin of National
Economic and Social Development of China and National Population Devel-
opment Plan.
2.2 Methodology
There are many methods of forecasting population, and each of them has mer-
its and demerits. What we choose in this chapter may not perfect, but it is
appropriate. The population forecast model is based on the cohort-component
method.1
The basic concept of the cohort-component method is:
P(t + n) = P(t) + B(t) − D(t) + I(t) − E(t)
where:
P(t) is the population at time t.
B(t) and D(t) are the numbers of births and deaths occurring between t and t + n.
I(t) and E(t) are the numbers of immigrants and of emigrants from the country
during the period t to t + n.
Because there is no data available for migrants and other issues, we made some
assumptions and follow the modified method below.
Main assumptions
1 The data we use in this chapter is true and effective, and it is valuable for
statistical analysis
2 No migration problem
3 No effects of war, plague, or other incidents
4 Same population policy for all area of mainland China
5 Multiple births can set off the groups who have no child
6 Death rate keeps the same for every age group, based on the data of
2010
7 Women aged 20 to 44 give birth; population policy makes little effect on
females under 20 or above 44
8 No considering people without residence registration.
First, calculate future surplus of every age group:
Pxt+1 = Pxt ×(1− Dx )

120
Pt = Pt
∑ x+1
x=1
Pxt+1: the population of x + 1 years old in year t.

Pxt: the population of x years old in year t.
Dx: the mortality rate of the population aged x to x + 1.
P t : total population of each age group (except 0-year-old) in year t.
Then, calculate the population of annual birth:
44
B t = ∑ ( f x ×Wx )
x=20
B : the population of new birth in year t.

t
f x : the fertility rate of each age group.

Wx: the fertile women population of each age group.
Deduce the annual total fertility rate
44 44
B t = ∑ ( f x ×Wx ) = ∑ (β × hx ×Wx ) = β × K
x=20 x=20
44
K = ∑ (hx ×Wx )
x=20
†: total fertility rate.

hx: standardized fertility rate of each age group.
K: the standardized total population of the new birth.
P0t = B t ×(1− D0 )
P0t : the population of the 0-year-old group in year t.

D0 : the mortality rate of the population aged 0 to 1.
Finally, calculate the population at the end of year t:
P = P t + P0t
Based on all of these calculations, we can calculate the aging coefficient and
dependency ratio to analyze the impact on population structure under the two-
child policy.
K 0 = P65+ / P
Rd =
(P
0 ~14 + P65+ )
P15~ 64
K 0: aging coefficient.
Rd : dependency ratio.
P0~14: the population aged 0 to 14.
P15~ 64: the population aged 15 to 59.
P65+: the population aged 6o and over.
There is a contradiction about the dependency ratio. In the statistical yearbook of

China, people aged under 15 and over 65 are defined as the dependent population.
However, in the census of population, people aged under 15 and over 60 are the
dependent population. In consideration of the available data, here in this chapter,
we follow the definition of the statistical yearbook, which is in accord with inter-
national standard.
We use this method to do the forecast on future total population. As to gender
structure, the work has been done by Bing Xu and Maxwell Pak (2015) and we
will use their results.
For estimating the change of human capital based on the methodology from
IWR-2014, We follow Arrow et al. (2012), Klenow and Rodríguez-Clare (1997),
and Barro and Lee (2010).
(Edu .ρ)
HC(t ) = e (t) . P5+Edu(t )
Edu(t ): a function of educational attainment, average years of the educational

attainment.
ρ: interest rate, which is fixed at 8.5%.
P5+Edu(t ): the population of the age of 5 (assuming individuals engage in formal
education at the age of five, although in China individuals engage in formal
education at six generally) plus the average years of the educational attainment
and older.
Considering China’s situation, we simplified the equation:
(Edu .ρ)
HC(t ) = e (t) . P15−64(t)
P15−64(t): the population aged 15 to 64 (Generally, Chinese people retire at 60,

but in the statistical yearbook, the labor population aged from 15 to 64, and
in rural area farmers over 60 still work). We can see the meaning of P5+Edu(t )
is actually the population of labor, and because we don’t have enough data
about the population of certain age, we change this part directly to the labor
population, about which we can get this data from the dependency data in the
statistical yearbook.
We calculate human capital but not human capital wealth here because of
the difficulties of calculating the shadow price of human capital per individual.
There are some problems with shadow price, and though it is a relatively useful
measure in money, it is still controversial. To avoid this issue, we calculate only
human capital change. According to the Inclusive Wealth Report 2014, human
capital comprises education capital and health capital. But the results of that
report was calculated from education capital only. Here we also use education
capital in our calculations, because the method of calculating health capital is
still very obscure. This simplified model may not able to estimate the human
capital accurately, but we can see the trend and change of human capital. The
unit of human capital is not money because we don’t cover the shadow price to
estimate human capital wealth (HCW) based on monetary measures, which also
remain controversial. In fact, it is hard to decide the unit of human capital, so
we keep the numerical results and put the emphasis on the percentage change
of human capital.
3.1 Population
3.1.1 Population forecast results

In late 2016, the Chinese government published a national population devel-
opment plan of 2016–2030, aimed at reaching 1.45 billion population in 2030.
This goal requires the fertility rate to reach 1.8. In this plan, the government
claims that the problem of population has always been the basic, overall
and strategic issue confronting human society. The plan aims to guide the
national population development in the next 15 years; it is an important base
for population and family planning work and the support for the development
of economic and social macro decisions. The current situation of population
development has both negative and positive sides. A decrease in workforce
and aging affects economic market directly, and the gap between provinces
is expanding. The positive side is that population growth still promotes the
modern urbanization.
The balance between population and resources and environment is tight. The
Chinese government wants to find a way to enhance the coordinated development
of population, social economy, resources, and environment. To keep the popu-
lation increasing sustainably, the fertility rate should be above 2, theoretically.
Because of the death rate, and because there are those who either don’t want or
cannot have a baby, some couples should have more than two children to make
sure the population of new generations will be adequate. Yet this is still hard to
achieve, especially after the one-child policy.
Total Populaon (million)
1500
1450
1400
1350
1300
1250
total populaon under the two-child policy

total populaon without the two-child policy
Figure 15.1 Total population forecast
1600
1500
1400
1300
Total population (millions)
1200
1100
1000
900 Medium variant

High variant
800 Low variant
700
600
1950 2000 2050 2100
Figure 15.2 Total population forecast by UN

Source: United Nations, Department of Economic and Social Affairs, Population Division (2017).
World Population Prospects: The 2017 Revision.
Owing to the new two-child policy, the fertility rate will temporarily rise and
keep steady. However, the total population will decrease once reaching the peak.
In our prediction, the peak year will be 2034 under the two-child policy and 2031
without it. The Chinese government claimed that the two-child policy would
bring an additional 30 million workers to the workforce by 2050, and this seems
achievable from the predicted results.
From the UN forecast, at the medium variant, the total population will reach
almost 1.45 billion around 2030, which matches the goal of the Chinese govern-
ment in the population development plan. However, from our prediction, the pop-
ulation will reach 1.47 billion in 2030. In reality, this number is a bit far from the
range of possibility since the downward trend of childbearing desire in the young
people’s group is continuing. In 2016, China’s total population was 1.383 bil-
lion, which increased 0.59% from the previous year. Based on the statistical data,
the newborn population in 2015 was 16.55 million, 17.86 million in 2016, and
17.23 million in 2017. The proportion of second children was 51% in 2017, and
although this increased 5% over 2016, the size of the childbearing women group
has decreased. Population explosion was once a heavy burden for Chinese soci-
ety under the planned economy, but it became an advantage after the economic
reforms. Apparently, it is difficult to control population growth and maintain the
demographic dividend at the same time.
Even though the two-child policy can promote population growth, aging doesn’t
stop. As the peak of growth is around 2030, the population of middle age group
will also achieve its peak. Figure 15.3 shows China’s age structure of recent years;
we can see that in 2016 the proportion of young people expanded, but this favor-
able trend won’t be sustained over time based on Figure 15.4. It means China
China’s age structure in recent years

100.00%
90.00%
80.00%
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
1982 1987 1990 1995 2000 2005 2010 2015 2016
0-14 years old 15-64 years old above 65
Figure 15.3 China’s age structure in recent years

Source: National Bureau of Statistics of the People’s Republic of China.
1000
900
800
700
Population (millions)
0−14
600 15−64
65+
500
400
300
200
100
1950 2000 2050 2100
Figure 15.4 Total population by broad age group forecast by UN

still has a shortage of newborn babies and a high dependency rate. Figure 15.5
provides a more intuitive image of age and gender structure changes from 1950
to 2017 and the anticipated changes in the foreseeable future, from which we can
see why a population policy shift is necessary.
Bing Xu and Maxwell Pak (2015) proved that gender ratio won’t change obvi-
ously under the two-child policy. Researchers think the one-child policy exacer-
bated gender imbalance, even if before the one-child policy Chinese couples had a
preference for boys over girls because of the traditional culture. And with the one-
child policy, the selective abortion of girl babies worsened. But this doesn’t imply
that once the one-child policy is removed, the gender balance will be improved.
In our evaluations, we use the adjusted fertility rate based on the statistical
yearbook 2016 and the survey of childbearing desire 2013. When we discuss a
population policy, we have to see which group of people will be affected. Of
course, under the two-child policy, it is couples who want and are capable of
1950 2017
100 100
Males Females Males Females
90 90
80 80
70 70
60 60
Age
Age
50 50
40 40
30 30
20 20
10 10
0 0
100 50 0 50 100 100 50 0 50 100
2050 2100
100 100
Males Females Males Females
90 90
80 80
70 70
60 60
Age
Age
50 50
40 40
30 30
20 20
10 10
0 0
100 50 0 50 100 100 50 0 50 100
Population (millions)
Figure 15.5 Population pyramids

having two children. If Chinese couples all want only one child, then the two-
child policy has no meaning. Through analyzing the potential women who would
like to have two children, it can be proved that the real situation of population
growth is very likely under our forecast level.
3.1.2 Incentives: survey of childbearing desire

Having a baby is a complex decision for young couples today, which is why it is
necessary to take a deep look at the incentives of childbearing. From the paper
Fertility Intention of Rural and Urban Residents in China: Results from the 2013
Table 15.1 Average ideal number of children among people with different characteristics
Grouping variable Groups Ideal number Average Number Percentage

of children of people
distribution (%)
0 1 2 3 and
above
Total 0.1 13.2 81.8 4.9 1.93 63,037 100
Age groups 20–24 0.0 13.3 83.6 3.1 1.90 4036 6.4
25–29 0.1 13.8 82.1 3.9 1.91 10,989 17.4
30–34 0.1 12.3 83.2 4.4 1.93 15,189 24.1
35–39 0.1 12.7 82.1 5.1 1.94 15,727 24.9
40–44 0.1 14.0 79.6 6.3 1.94 17,097 27.1
Gender male 0.1 13.1 81.5 5.3 1.93 29,687 47.1
female 0.1 13.2 82.0 4.6 1.92 33,350 52.9
Residence agricultural 0.0 10.9 83.1 6.0 1.96 45,655 72.4
non-agricultural 0.2 19.2 78.4 2.1 1.83 17,382 27.6
Educational level primary school 0.0 9.9 80.0 10.1 2.03 9667 15.3
and below
secondary school 0.0 12.4 82.4 5.1 1.94 30,898 49.0
high school 0.1 15.3 82.3 2.3 1.87 11,918 18.9
college and above 0.3 16.2 81.0 2.5 1.86 10,553 16.7
Number of 0 0.7 20.0 77.6 1.7 1.81 3684 5.8
existing children 1 0.1 19.2 78.9 1.8 1.83 36,851 58.5
2 0.0 2.3 91.5 6.2 2.05 19,836 31.5
3 and above 0.0 1.9 55.2 42.9 2.53 2666 4.2
Source: National Health and Family Planning Commission of China.
National Fertility Intention Survey, we gathered the data for the adjustment of fer-
tility rate. This national fertility intention survey was conducted in 29 provinces
in China in August 2013. Survey results indicate that the ideal number of children
is 1.93 in China (Zhuang et al., 2014). Given the fertility rate in 2013 was 1.555,
there is a lot of room for growth. However, there is a big gap between the ideal
number of children and the maximum number of children a family can have.
From Table 15.1, it can be seen that people with different characteristics have
quite obvious different demands of the number of children. For instance, one of
the most outstanding phenomena is that rural residents have greater needs for chil-
dren than urban residents. Furthermore, the better educated tend to have a lower
desire for childrearing. Some believe that this is because highly educated people
will spend much more money on their children’s education since they expect their
children to achieve at least the same educational level as they did, which certainly
increases childrearing costs. The concept of “fewer and better births” in China today
means not only better health condition, but also a better nurturing environment.
In 2016, one of the largest internet companies conducted an internet survey
of fertility intention. Although it was not official, the results can present the lat-
est conception of having children. However, we have to admit these results may
represent only people living in cities or towns considering some of the rural areas
have no internet availability. According to the result, only 31.1% people want
two children. In this survey, we can see some interesting phenomena in Chinese
society. For instance, in the family which already has one child, couples are likely
to have the second one if the first child is a girl. Whether to have children or how
many children is not a problem for one person or couples, it is a problem for the
whole family, and more than 30% families have different opinions about raising
more child. Comparing the preference of older and younger people, we found that
older people have a preference for boys, while the younger generation tends to
prefer girls. In comparison with the official survey in 2013, it is clear that even
though the ideal number of children per family is 1.93, the real situation is much
far from this ideal state. Many young couples now don’t want to have any chil-
dren. People today care more about their own lives, and raising the next genera-
tion is not the life theme for many Chinese women anymore.
3.1.3 The difference between urban and rural area

It is quite evident that the impact on population of the two-child policy is less in
rural areas than in urban areas since the fact that women in some rural areas can
have two children even under the one-child policy. Data from tabulation on the
2000 population census of the people’s republic of China shows that 232,176 sec-
ond babies were born in rural area, while 66,448 were born in cities and 73,671
in towns. This big gap between rural and urban indicates that the two-child pol-
icy may have little influence in rural areas. In some megacities like Beijing and
Shanghai it costs too much to raise a child, so couples usually choose to have one
child or don’t have a child until household finances allow them to.
In 2000, the fertility rates of city, town, and village are 0.86, 1.08, and 1.43,
respectively; in 2010 these rates were 0.88, 1.15, and 1.44, respectively, according
to the population census. These data are not in accordance with World Bank due
to statistical error, but the figures indicate a sharp difference between urban and
rural areas. Besides the ideal number of children, differences are mainly reflected
in reproduction purpose and gender preference. The primary reasons that caused
those phenomena are as follows. First, the cost in rural areas is much lower. When
considering the cost to raise a child, it is not just the money couples would spend
but also the time value or opportunity cost. For example, women living in cities
could lose their jobs after taking maternity leave or parental leave, while women
in villages can go back to farm work anytime. Second, urban dwellers have a large
market in which to purchase various products to meet the needs of daily life, but
the agricultural management system in rural area determines that families can
be self-sufficient. This means that one more child in urban families could reduce
consumption levels of couples for themselves, while in rural areas children are
seen as future labor to contribute to family farming. Third, the social security
system for rural residents is not well organized. We will discuss this problem
specifically later in the part on social impact.
3.2 Human capital change

The origin of the concept of human capital can be traced back to the writings
of Adam Smith in the eighteenth century. Smith emphasized the importance of
all the residents or members of the community acquiring useful ability. When a
person gains this ability, he will pay the price, and once acquired, they became
the fixed capital (Smith 1776). Human capital is both physical and nonphysical
asset that usually represents knowledge, skills, health condition, and so forth of
workers in a certain period of time that obtained from schooling or training. As
a primary input for almost production processes, human capital is central to a
country’s sustainable development. Government can invest in human capital by
increasing expenditure in educational institutions. Vocational training is also a
way for company to improve human capital of employees. In this chapter, we
use interprovincial data to calculate the human capital of each province, but put
emphasis on comparing the change of whole mainland China at different times.
As shown in the Maps 15.1–15.4 and Table 15.2, we can see that human capi-
tal increases at first and then falls when comparing 2050 to 2030. It is the same
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Xizang Shaanxi Henan
Jiangsu
Sichuan Anhui Shanghai
Legend Hubei
province boundary Chongqing Zhejiang
Jiangxi
Human Capital (2000)
< 15000000 Hunan
Guizhou Fujian
15000001 - 40000000
40000001 - 60000000 Yunnan

60000001 - 80000000 Guangxi Guangdong
80000001 - 120000000
120000001 - 150000000
0 1,000 2,000
> 150000001 km
Hainan
Map 15.1 Human capital of mainland China 2000

N
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Legend Hubei
Jiangxi
< 15000000 Hunan
Guizhou Fujian
15000001 - 40000000
40000001 - 60000000 Yunnan

80000001 - 120000000
120000001 - 150000000
0 1,000 2,000
> 150000001 km
Hainan
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Legend Hubei
Jiangxi
< 15000000 Hunan
Guizhou Fujian
15000001 - 40000000
40000001 - 60000000 Yunnan

80000001 - 120000000
120000001 - 150000000
0 1,000 2,000
> 150000001 km
Hainan

Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Legend Hubei
Jiangxi
< 15000000 Hunan
Guizhou Fujian
15000001 - 40000000
40000001 - 60000000 Yunnan

80000001 - 120000000
120000001 - 150000000
0 1,000 2,000
> 150000001 km
Hainan
Table 15.2 Human capital change with respect to the base year 2000
Population policy Percentage change with respect

to the base year 2000
2015 2030 2050

With the two-child policy total 37.2 69.3 40.8
per capita 26.5 45.7 30.4
Without the two-child policy total 37.2 65.9 36.0
per capita 26.5 44.9 29.9
situation for human capital per capita. We can explain this phenomenon by con-
sidering the change of workforce. As a result of the disappearing of the demo-
graphic dividend, China faces a labor shortage now and in the foreseen future.
Population keeps increasing under the two-child policy as more babies come into
the world. However, even though the newborn population is able to enter the labor
market in at least 15 years, the total population of labor force is still decreasing.
The continuing decline in the workforce and the high speed of aging are indicative
of the inadequacy of the two-child policy.
Here we assume that the level of average education will keep increasing based
on the forecast from the World Bank. We admit two-child policy will definitely
affect education, and the way to analyze whether this influence is negative or
positive is based on multiple conditions such as the household income, locational
educational resources, and consumption level of city or town. Since the forecast
growth trend of average educational level from the World Bank is positive, we
have reason to assume that the two-child policy won’t cause a negative impact on
education on the whole.
Li Jing (2016) shows that when the birth restriction was enforced there was
a marked increase in education investment. The more strictly bound, the more
investment in education. This counterfactual work suggests that the one-child pol-
icy and exogenous birth limit led to a surge of investment in human capital so that
more educated generations grow up and enter the labor market, so as to further
improvement of the per capita GDP growth rate. In this respect, we wonder if the
human capital investment in each child will decline when the fertility rate rises.
Measuring the change of human capital under the two-child policy means much
more for social and economic development as human capital is the most important
component of a country’s wealth. And this forecast can provide a new perspective
for policymakers to see whether we are on the path of long-term sustainability.
Although we use only the education part to estimate human capital, it is neces-
sary to discuss the impact on health capital. Health capital is surely an important
portion of human capital and it has a direct effect on well-being and happiness. In
the IWR 2014, a measure concept of health capital was introduced that is using
expected lifetime utility and longevity. Health capital is related to many aspects
Average Educational Level (years)

10
9
8
7
6
5
4
3
2
1
0
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Figure 15.6 China’s average educational level

Source: National Bureau of Statistics of the People’s Republic of China.
80
70
60
50
40
30
20
10
0
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
Figure 15.7 China’s life expectancy at birth
Source: World Bank.
of population policies. Population increase can influence the environment and

through it health capital. Under the two-child policy, the care for newborn baby
and parturient won’t be enough considering the sanitary situation. Also, care for
elder people is a notable problem especially in the rural area of China.
Given the difficulties and lack of empirical estimates, using only life expec-
tancy would be appropriate. Health capital was treated separately from other
forms of capital due to its particularity, but it should no longer be excluded when
estimating inclusive wealth. From the data of life expectancy, we can be sure
that health capital has positive effects on human capital. The life expectancy of
Chinese people has enhanced greatly, from 43.78 in 1960 to 76.12 in 2015, but
compared to developed countries it is still low.
Human capital tends to concentrate on major cities, which is in accord with
the population concentration trend. Human capital is not only what people can
require from schools, but also from family, daily life, and experiences, so using
the education condition to estimate human capital can be underestimated. But
based on educational estimate method, we can see that major cities usually have
high human capital due to high population concentration and rich educational
resources.
3.3 Social impact

Population can directly affect all various levels of human society and develop-
ment. To discuss the impact the two-child policy may bring, we should analyze
what kinds of problems can be solved or caused under the two-child policy in the
future.
First, let us see the positive side. The two-child policy can alleviate aging prob-
lem since more babies are born. In 2016, births increased 7.9% compared with the
previous year, according to government figures, and this increase was attributable
to the new two-child policy. In 2016, 17.86 million babies were born in all – an
increase of 1.31 million over the total in 2015. The new total represents the high-
est annual number of newborns since 2000, according to government data. From
this result, the two-child policy did work. But whether the baby boom will last for
years is hard to say.
In China, there are two pension systems. For employees in urban enterprises,
it is intergeneration transfer and saving payment. For people in government and
other institutions, it is only intergeneration transfer, which is using the part of
the income of young generation to pay for the old people. This dual track system
caused unfairness in society. And as the aging problem becomes more serious, the
young generation cannot afford to pay for the old generation. That is why some
Chinese couples think having more babies will yield more pension providers.
China’s pension system was established in the 1950s; before that time almost
all old people relied on their children or other family members after they retired.
Family support has a special meaning in Chinese culture: the Confucian philoso-
phy called “filial piety”.
The pension problem, the lack of institutions for old people, trained caregivers,
and subsidies are also big issues. Of course, elderly facilities are not popular in
China because of the tradition of care by family members. Many old people think
to go to a facility is a shame because it means no family member would take care
of them.
The loss of an only child or so-called bereaved parents is a Chinese character-
istic social problem. According to the estimation of demographers, the loss of an
only child will reach 10 million in the future. The number of only childre aged 15
to 30 is 0.19 billion, and the death rate in this age group is 0.04%, which means
76,000 couples lose their only child every year. And this kind of family always
means couples are over 50 years old and cannot usually have another child. Under
the two-child policy, the loss of a only child is expected to decrease.
As the disappearance of a demographic dividend, China will face a lack of
labor soon after workers born in the 1960s begin retiring. This problem is trou-
bling not only for China but even for many developed countries. The Chinese
government tries to postpone retirement to slow down the decreasing labor mar-
ket and the pressure of the pension system. Although the total number of labor
population is still increasing, the relative decline of the working age proportion is
decreasing apparently. In the long run, the two-child policy can contribute to the
future labor market which is the most crucial element to economic development.
Countries like the United States usually depend on immigration when facing the
labor shortage, but due to China’s tradition, it is hard to imagine that China will
open the door for foreigners as immigrants. However, in China’s coastal areas like
Guangzhou, more and more labor from Africa is showing up.
Although may not be significant, the two-child policy can play a small role
in adjustment for gender structure. Sex-selective abortion was a common
phenomenon in the past three decades under the one-child policy. Hence the one-
child policy is considered as the primary reason why sex ratio was more seriously
unbalanced in China. Under the one-child policy, couples would choose to have
a boy over a girl, and in some rural areas even the government allows couples to
have a second child if the first one is a girl. As mentioned before, this is because
China has a culture with a strong preference for sons. After all, in the past, men are
the main labor in rural areas for farming, so naturally people give more attention
to boys but not girls. Nonetheless, as a social development, this discrimination
between boys and girls will disappear gradually. Sex-selective abortion is not only
a social phenomenon in China but can also be found in other Asian regions.
On the other side, the two-child policy could cause some problems simultane-
ously since no policy is perfect. Many researchers think what China had cost to
implement the one-child policy and paid for the side effects that the one-child
policy brought were heavier than the achievement the government made in popu-
lation control. Thus, it is undoubted that the two-child policy will make some new
problems over the old issues the one-child policy had left.
Environment pressure is the most likely problem the two-child policy could
create. More population means more production and consumption, more needs
for all kinds of resources and more waste. China is a big country with a vast terri-
tory and a high population density, with most population concentrated in eastern
China. Population and environment have a very complicated relationship. The
growth rate, structure, distribution, and the total population affect different parts
of the environment in different ways. And the ways people produce, consume, and
live are inextricably tied up with the environment.
In recent years, China had many natural disasters. The irrational use of grass-
land led to degradation, desertification, and salinization, and the average yield of
grassland decreased by 30%–50%. In arid and semi-arid areas, precipitation is
much lower than evaporation, vegetation is sparse, ecology is fragile, and mild
human activities are likely to lead to land degradation and desertification.
Except for environment, economic burden is the immediate problem facing
China. Total GDP growth rate may increase while GDP per capita will decrease.
At the micro level, a family will face economic pressure to raise two children.
The cost of living in cities like Beijing, Shanghai, Shenzhen, and Guangzhou is
too high for migrants from small towns. They usually cannot even afford to buy
a room. At the macro level, the government has to invest more money in public
infrastructure which will decrease national coffers.
With more children and elder people to take care of for at least 15 years, there
will be a shortage of infrastructure for healthcare and education. Hospitals for
parturient women are lacking even now, and two women sharing one ward is
common in China. According to some news, after the universal two-child policy
was implemented, many elderly pregnant women were facing health risks to give
the birth to the second baby. Therefore, China will face an immediate scarcity of
obstetricians, midwives, and so forth.
Also, there are concerns that more population will cause deterioration of national
quality because of the lack of educational resources. In the past, particularly in
rural areas, if a family has more than one child, not all of them can go to school.
And when there is a boy, parents would give more resources to the son and let
daughters go to work. At the present time, low fertility rate may be a question
most countries faced. Usually, developed countries have lower fertility rate than
developing countries. One view is that this is because much manual labor is
needed in developing country.
By comparing the fertility rate of China with the whole world and other areas
of China, we can claim that low fertility rate is an inevitable trend even with-
out family planning. In East Asia, mainland China is the only area to implement
compulsory family planning, but the change of fertility rate is almost the same.
It is possible that without the one-child policy or family planning the fertility
0
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
2016
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
China World
Figure 15.8 Fertility rate of China and the world
0
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
China Hong Kong SAR, China Macao SAR, China
Figure 15.9 Fertility rate of mainland China, Hong Kong and Macao
Source: World Bank.
Total fertility
(live births per woman)
8.00 or over
7.50 to 8.00
7.00 to 7.50
6.50 to 7.00
6.00 to 6.50
5.50 to 6.00
5.00 to 5.50
4.50 to 5.00
4.00 to 4.50
3.50 to 4.00
3.00 to 3.50
2.50 to 3.00
2.25 to 2.50
2.00 to 2.25
1.75 to 2.00
1.50 to 1.75
Less than 1.50
No data
Map 15.5 Fertility projection by UN

Source: World Population Prospects: The 2017 Revision.
rate would still be as low as it is today. Many studies about fertility rate all show
evidence to prove why fertility rate falls, and the general thought is because of
the rising education level of women. Giving birth and taking care of children are
no longer the foremost duties for women. Other opinions say, as the economy
develops, people have more choice on consumption of goods but do not want to
pay for raising children. Some economists consider the opportunity cost to raising
children rather than purchasing the goods that can benefit the parents. And many
think it is not an issue need to solve because we don’t need too many people to do
physical work since economy and technology developed.
McElroy, M. and Yang, D. T. (2000) studied how family planning affects Chi-
na’s fertility rate. They control the compulsory level of population policy in the
model, and the final result showed that without family planning the fertility rate
would rise 0.33. All of these studies prove that family planning is not the princi-
pal reason for low fertility rate. Today, India still has a high fertility rate, and we
believe that it is the result of the low educational level of women and the tradi-
tional culture which is similar to China’s “more children more happiness”.
3.4 Population policy in Vietnam and Russia

Excluding China, Vietnam has a more than 50 years’ history of population policy.
Population policy in Vietnam was launched by the Vietnamese government in the
North Vietnamese region in the early 1960s and has been modified throughout the
country today. The policy emphasizes the official family should have one or two
children. In 2014, Vietnam had a population of 92.5 million people, accounting
for 1.28% of the total population of the world. At present, the total fertility rate of
Vietnam is 1.8 births per woman, which is below the replacement level of fertility
rate 2.1, the rate that can completely replace their population itself from genera-
tion to generation, according to the World Resources Institute.
Under the two-child policy, Vietnam successfully controlled total population
growth, but problems arose at the same time. This is similar to China’s situa-
tion under the one-child policy. Reduction of birthrate, increased abortion, and
sex imbalance are the key problems the Vietnamese government wants to solve
now, and these are the reasons why the government was considering canceling
the population policy to let couples decide how many children they want without
fine or punishment in 2015, although until now the fertility rate did not increase.
From Vietnam’s case, the two-child policy is not enough to deal with the aging
situation. The fertility rate of Vietnam in 2015 was 1.96 while China’s was 1.569.
We wonder if the two-child policy is also not able to solve China’s population
problems.
China can learn a lesson from Vietnam, which has seen its birth sex ratio imbal-
ance decrease sharply in the past 20 years due to the government’s broader imple-
mentation of family planning policy and forbidden sex-selective abortion. Under
the similar cultural pressure of preference for having boys, the population control
policy in Vietnam is not as strict as in China, where it led to an imbalance of gen-
der structure. The prevalent view is that it is easy to reduce fertility rate but hard
to raise it. Most countries that have population policy attempt to encourage birth
but not restrain it. However, few of them did success. Russia may be a successful
case worthy of analysis.
The Russian labor and social security minister revealed that in the first half of
2014, in Russia, families with three or more children grew by 2.4% over the same
period last year. Negative population growth was a prominent social problem that
once plagued Russia. In order to further improve the high population, Russia in
recent years has implemented a series of policies to stimulate growth, including
the implementation of 66 federal districts to give families with three or more
children monthly subsidies, and the federal government prepared a total fund of
$280 million in 2014.
Russia is one of the countries with the longest maternity leave in the world:
mothers can receive a full allowance maternity leave for 140 days, 70 days before
and after giving birth. In addition, 140 days later, women can continue their vaca-
tion when their child is one and a half and they receive 40% of the original wage
allowance (calculated according to the average income from the previous 2 years
before the vacation). Finally, according to the law that was in force in 1989,
women could continue their vacation for a year and a half until the child is 3 years
old, and the employer has to hold her post at this time and not dismiss her. In this
kind of environment, parents would love to enjoy their parenting life.
Russia has proved that fertility rate can be raised as long as the government
gives people strong incentives to have more children. But to make these subsidy
and welfare policies requires that a country has strong finances and superior social
economic conditions to back it up.
2.5
1.5
0.5
Figure 15.10 Fertility rate of Russian Federation

Source: World Bank.
With the progress of society, there will be two kinds of fertility trends. The
first fertility trend is well known: the fertility rate in developed countries will
keep decreasing as the economy develops. Economic development has made a
fundamental change in women’s social status; women have become more active
in social activities while their desire to procreate has reduced. Social develop-
ment has also made great changes in marriage: the marriage age increases and the
divorce rate rises. Raising children is becoming more and more difficult, espe-
cially in large cities. For many parents, raising children means a serious reduction
in living standards. Russia’s grim population situation had been widely recog-
nized, but if we think carefully we can find that Russia’s fertility rate is not so low
as imagined: 1.4 to 1.6. In Asia, this number is higher than many developed coun-
tries and regions, compared with 1.39 in Japan, 1.2 in Korea, 1.15 in Singapore,
1.1 in Hong Kong and Macao that have reached the lowest levels in the world.
Fertility rates are also low in some European countries, such as 1.39 in Germany
and Spain, 1.4 in Italy, and 1.44 in Greece (according to the World Bank 2010).
Recent comparative studies of populations have produced a second interesting
fertility trend: further modernization of society, rather than a return to traditional
values, will help boost fertility rates in developed countries. Social moderniza-
tion means that women gain the social status formerly monopolized by men and
that men have the same family responsibilities as women. For example, some
developed countries in Europe have relatively high fertility rates: 2.94 in Britain,
2.1 in the United States, 2 in France, 1.98 in Sweden, 1.95 in Norway, and 1.95
in Finland. After overcoming the familiar demographic and economic difficul-
ties, “the more wealth, the fewer children” may be just one phase of population
growth. Indeed, the improvement of women’s social status will lead to a sharp
decline in population growth. But when achieving the equality between men and
women, the social structure tends to be stable, the fertility rate will rise again and
keep steady at a higher level in fertility rate.
This hypothesis is supported by data from developed countries. German
researchers found that traditional countries such as Portugal, Spain, and Greece
have a higher fertility rate while the fertility rate in social modernization countries,
such as Sweden and Norway, is lower. They also found that in developed countries
women’s family status and working conditions are relatively good and the fertility
rate is relatively high. European transnational comparative analysis clearly shows
the mutual influence degree of female employment and fertility. Facts show that
the higher the fertility rate is, the better the social support and work conditions will
be. In some countries, such as Spain, Italy, and Greece, the employment rate of
women aged 25–59 is less than 60%, and the national fertility rates are lowest in
Europe; in other countries like Iceland, the employment rate of women aged 25–59
is more than 80%, and the fertility rate is very high, reaching 2.05.
Perhaps, the solution to low fertility is to give women favorable working envi-
ronment and payment, and let men take childcare responsibilities as women.
When women don’t have to give up their career and other social value to give
birth and look after children, the fertility rate will improve naturally.
3.5 Urbanization, domestic population, and human capital transfer
In China, people tend to go to major cities to find work and make a living. Although
the cost in these cities is higher, the pay is higher, and with all kinds of resources
and choices that small cites or towns don’t have. There are two main population
transitions in China: rural areas to urban areas and common cities to big cities.
The speed of urbanization in China is fast, with the economic development more
and more population concentrate on cities, changing the urban morphology.
Recently, the national health and family planning commission issued the “China
Division of Floating Population Dynamic Population Development Report 2016”.
In 2015, China’s floating population reached 247 million, accounting for 18% of
the total population, which is equivalent to one in six people is a floating popula-
tion. In the next 10 or 20 years, China is still in the stage of rapid urbanization.
In accordance with the “new national urbanization plan” process, in 2020, China
still has more than 200 million of the floating population. In the “13th Five-Year”
period, the population has been continuing to gather along the coast and the main
traffic areas and big cities. The population of these areas continues to increase,
hence the proportion of migrant workers in the central and western regions also
increases significantly.
According to the data of China’s six population census, the level of urbaniza-
tion in each census was 12.84%, 17.58%, 20.43%, 25.84%, 35.39%, and 49.68%
from 1953 to 2010. And in 2017, the newest data shows it reached 58.52%.
Although massive numbers of rural population move into cities, their rural reg-
istration is hard to change due to the strict household registration system. They
therefore cannot enjoy some public services such as medical care, social security
or housing security that urban residents have. Compared to developed countries,
the urbanization rate is still low. In the process of urbanization, several issues
would be challenges. One of them is overexpansion of urban scale with high den-
sity population. Establishing satellite cities or new towns is an effective way to
improve traffic and living conditions.
Floating population is a concept under China’s household registration system
that commonly means a group of people who leave their place of domicile to go
other places. Internationally, this is known as internal migration. With the rapid
development of industrialization and urbanization, China has entered the most
active period of the population transfer. In 2009, the average age of the floating
population was 27.3, 27.9 in 2013, and 29.3 in 2015. Obviously, the young gen-
eration is the principal component of floating population. Megacities like Beijing,
Guangzhou, Shanghai, and Shenzhen, where most of the floating population are
assembled, possessed high human capital.
Human capital between provinces has large differences. Under the two-child
policy, more population and human capital will concentrate on south China. The
differences will be expanded and the growth of human capital is more significant
in these big cities. Meanwhile, maldistribution of human capital in provinces has
caused uneven development of regional economies. Due to geography, climate,
and other factors, economic gap between regions is an inevitable phenomenon.
Despite the imbalance of human capital distribution, an increasing number of
N
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Hubei
Legend Chongqing Zhejiang

Jiangxi
province boundary
Hunan
Urbanization Rate of 2005 (%) Guizhou Fujian
20-30
Yunnan
30-40
Guangxi Guangdong
40-50
50-60
60-90
0 1,000 2,000
km Hainan
Map 15.6 Urbanization rate of mainland China 2005
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Hubei

Jiangxi
province boundary
Hunan
Urbanization Rate of 2013 (%) Guizhou Fujian
20-30
Yunnan
30-40
Guangxi Guangdong
40-50
50-60
60-90
0 1,000 2,000
km Hainan
Map 15.7 Urbanization rate of mainland China 2013

N
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Hubei

province boundary Jiangxi
Hunan
Population change (2002-2013) Guizhou Fujian
Yunnan
Guangxi Guangdong
Population of 2015 (10 thousand)
324-1547
1548-3017
3018-4796
4797-7425 Hainan
0 1,000 2,000
7426-10849 km
Map 15.8 Population distribution (2015) and population change (2002–2013) of mainland

China
N
Heilongjiang
Nei Mongol Jilin

Xinjiang Uygur
Liaoning
Gansu Beijing
Tianjin
Hebei
Ningxia Hui Shanxi
Qinghai Shandong
Jiangsu
Hubei
Chongqing Zhejiang
Legend Jiangxi
province boundary Hunan
Guizhou Fujian
Population Density (capita/km2)
0-184 Yunnan
Guangxi Guangdong
184-434
434-778
778-1321
1321-3770 0 1,000 2,000
km Hainan
Map 15.9 Population density of mainland China 2015

young people in recent years tend to choose to leave large cities and go back to
contribute to their hometown, which spontaneously alleviated the imbalance.
4 Conclusion
In this chapter, we made forecasts on human capital and population change under
the two-child policy. We found that although the two-child policy has a positive
effect on human capital and can promote population growth by raising fertility
rate, it is still insignificant and insufficient to solve China’s social problems that
caused by demographic issues.
Our results have a number of limitations. First, in terms of calculating human
capital, although we followed Arrow et al. (2012), Klenow and Rodríguez-Clare
(1997), and Barro and Lee (2010), we made a small simplification on the formula
and concept due to the lack of a particular category of data. And we omitted the
shadow price part because we have no intent to use money as the unit of human
capital. Instead, we chose the concept of efficiency units to do the explanation.
Based on the method of Inclusive Wealth Report 2014, we didn’t explore further
mathematically. Second, when collecting demographic data from the statistical
yearbook and the census report, we applied some regular but not very precise
ways to supplement the missing data.
As long as the Chinese government insists on birth control, several missions
need to be completed. To increase the incentives for couples to have two children,
investing in the public infrastructure of education and healthcare is a fundamental
task. Under the circumstance of unreasonable raising of house prices, controlling
house prices to reduce the pressure of living is an expectation of the majority of
the people. Further policies should also be established to support the two-child
policy, and the government has to provide more social security and welfare to
women and families.
We believe that we made a small contribution on researching the relationship
between population policy and human capital and offered a new perspective for
policymakers. More research in the future will improve our defects and take
another step forward.
Note
1 www.measureevaluation.org/resources/training/online-courses-and-resources/non-
certificate-courses-and-mini-tutorials/population-analysis-for-planners/lesson-8.
Bibliography
Aiura, H., and Sato, Y. (2014). A model of urban demography. Canadian Journal of Eco-
nomics/Revue canadienne d’économique, 47(3), 981–1009.
Arnold, F., and Zhaoxiang, L. (1986). Sex preference, fertility, and family planning in
China. Population and Development Review, 221–246.
ability and the measurement of wealth. Environment and Development Economics,
17(3), 317–353.
Barro, R. J., and Lee, J. W. (2010). A new data set of educational attainment in the
world, 1950–2010. NBER Working Paper No. 15902. National Bureau of Economic
Research.
Bongaarts, J., and Greenhalgh, S. (1985). An alternative to the one-child policy in China.
Population and Development Review, 585–617.
Brockwell, P. J., and Davis, R. A. (2016). Introduction to time series and forecasting.
Springer.
Den Boer, A. M., Hudson, V. M., and Russell, J. (2015). China’s mismatched bookends: A
tale of birth sex ratios in South Korea and Vietnam. Paper prepared for the annual meet-
ing of the International Studies Association, New Orleans, February.
Dettling, L. J., and Kearney, M. S. (2014). House prices and birth rates: The impact of the
real estate market on the decision to have a baby. Journal of Public Economics, 110,
82–100.
Ebenstein, A. (2010). The “missing girls” of China and the unintended consequences of the
one child policy. Journal of Human Resources, 45(1), 87–115.
Enos, O., Torre, S., and Wilson, W. T., PhD. (2016). An economic and humanitarian case for
pressing China to rescind the two-child policy. Backgrounder No. 3146 November 18.
Feng, L. (2000). The past, present, and future of China’s population policy. Population
Research, 24(4), 23–34. (In Chinese).
Feng, Z., Liu, C., Guan, X., and Mor, V. (2012). China’s rapidly aging population creates
policy challenges in shaping a viable long-term care system. Health Affairs, 31(12),
2764–2773.
Goodkind, D. M. (1995). Vietnam’s one-or-two-child policy in action. Population and
Development Review, 85–111.
Goodkind, D. M. (March 1995). Vietnam’s one-or-two-child policy in action. Population
and Development Review, 21(1), 85–111. JSTOR 2137414. doi:10.2307/2137414.
Hesketh, T., Lu, L., and Xing, Z. W. (2005). The effect of Chinas one-child family policy
after 25 years. New England Journal of Medicine, 353(11), 1171–1176.
Klenow, P. J., and Rodriguez-Clare, A. (1997). The neoclassical revival in growth econom-
ics: Has it gone too far? NBER Macroeconomics Annual, 12, 73–103.
Li, J. (2016). Population dynamics, human capital, and economic growth: The consequence
of one-child policy in China. Diss. Clemson University.
Lu, Y., and Zheng, M. (2017). Web topic analysis of the two-child policy in China. Procedia
Computer Science, 107, 97–102.
Maruyama, A., and Yamamoto, K. (2010). Variety expansion and fertility rates. Journal of
Population Economics, 23(1), 57–71.
McElroy, M., and Yang, D. T. (2000). Carrots and sticks: Fertility effects of China’s popula-
tion policies. The American Economic Review, 90(2), 389–392.
Pham, B. N., Hill, P. S., Hall, W., and Rao, C. (2015). The evolution of population policy
in Vietnam. School of Population Health, University of Queensland. Retrieved 12 March.
Rosenzweig, M. R., and Zhang, J. (2009). Do population control policies induce more
human capital investment? Twins, birth weight and China’s “one-child” policy. The
Review of Economic Studies, 76(3), 1149–1174.
Sato, Y. (2007). Economic geography, fertility and migration. Journal of Urban Economics,
61(2), 372–387.
Sato, Y., and Yamamoto, K. (2005). Population concentration, urbanization, and demo-
graphic transition. Journal of Urban Economics, 58(1), 45–61.
Schultz, T. P. (1985). Changing world prices, women’s wages, and the fertility transition:
Sweden, 1860–1910. Journal of Political Economy, 93(6), 1126–1154.
Schultz, T. P. (2007). Population policies, fertility, women’s human capital, and child qual-
ity. Handbook of Development Economics, 4, 3249–3303.
Searchinger, T., Hanson, C., Waite, R., Harper, S., Leeson, G., and Lipinski, B. (2013).
Achieving replacement level fertility. World Resources Institute Working Paper.
Simon, C. J., and Tamura, R. (2009). Do higher rents discourage fertility? Evidence from
US cities, 1940–2000. Regional Science and Urban Economics, 39(1), 33–42.
Smith, A. (1776). An Inquiry into the Nature and Causes of the Wealth of Nations: Volume
One. London: W. Strahan and T. Cadell.
UNU-IHDP and UNEP (2014). Inclusive Wealth Report 2014: Measuring Progress Toward
Sustainability. New York: Cambridge University Press.
Vietnam Population (Live). Worldometers. Retrieved 12 March 2015.
Wang, D. (2006). China’s urban and rural old age security system: Challenges and options.
China & World Economy, 14(1), 102–116.
Wang, K., Ding, J., and Wang, F. (2016). Influence of the implementation of the universal
two-child policy on demographic structure and population spatial distribution in China.
Progress in Geography, 35(11), 1305–1316.
World Bank (2010). World Development Indicators. USA: Oxford University Press.
Xu, B., and Pak, M. (2015). Gender ratio under China’s two-child policy. Journal of Eco-
nomic Behavior & Organization, 119, 289–307.
Yi, Z., and Vaupel, J. W. (1989). The impact of urbanization and delayed childbearing on
population growth and aging in China. Population and Development Review,
425–445.9.1.
Zeng, Y., and Hesketh, T. (2016). The effects of China’s universal two-child policy. The
Lancet, 388(10054), 1930–1938.
Zhang, J. (2002). Urbanization, population transition, and growth. Oxford Economic
Papers, 54(1), 91–117.
Zhu, W. X., Lu, L., and Hesketh, T. (2009). China’s excess males, sex selective abortion,
and one child policy: Analysis of data from 2005 national intercensus survey. British
Medical Journal, 338, b1211.
Zhuang, Y., Jiang, Y., Wang, Z., Li, C., Qi, J., Wang, H., . . . Tan, M. (2014). Fertility inten-
tion of rural and urban residents in China: Results from the 2013 national fertility inten-
tion survey. Population Research, 38(3). (In Chinese).
16 The impact of renewable
energy on sustainability
Empirical analysis employing
the Inclusive Wealth Indicator
Moegi Igawa and Shunsuke Managi
1 Introduction
Both the development and deployment of renewable energy1 has been rapidly
increasing around the world for the past couple of decades. According to IRENA
statistics (the International Renewable Energy Agency), electricity generation
from renewable energy sources grew at an average annual growth rate of 5%
between 2001 and 2015. Solar and wind energy sources are particularly salient in
this regard: the average annual rate of growth during the period was 45% for solar
and 22% for wind, and the highest annual growth rate during that period was 86%
for solar in 2011 and 35% for wind in 2002.
Global investment in renewable energy has also been remarkable. In 2016, it
was estimated to be USD 2.4 trillion, roughly five times as large as that in 2004
(Frankfurt School FS-UNEP Collaborating Centre, 2017). Also, 138 GW of
electricity generation capacity was added, which was roughly double that of fos-
sil fuels and seven times more than that of new nuclear plants (Frankfurt School
FS-UNEP Collaborating Centre, 2017).
Why are these numbers growing so rapidly? One of the main drivers is the
implementation of various policy supports. Among these, the prevailing ones are
feed-in tariffs, subsidies or tax deduction for investments, subsidies for research
and development (R&D), and renewable energy certificates. In 2016, 110 coun-
tries, states, or regions employed feed-in tariffs (REN21, 2017). Under the Paris
Agreement signed in 2016,2 55 countries featured renewable energy targets as one
element of the process of curtailing greenhouse gas emissions (REN21, 2017).
Although renewable energy is still said to be insufficiently cost-effective, such
targets are not limited to wealthy countries. At COP22, also in 2016, 48 develop-
ing countries made a declaration of working towards achieving 100% renewable
energy in their respective countries (REN21, 2017).
The political rationales for such commitments can be classified into four cat-
egories. The first category is energy security. Renewable energy can contribute to
eliminating or at least reducing energy security risks in two ways (International
Energy Agency, 2007). One way is that renewable energy is expected to become
an indigenous source of electric power, and a viable alternative to imported fossil
fuels. Oil and gas commodity prices are known to fluctuate. Renewable energy
406 Moegi Igawa and Shunsuke Managi
3,000,000
2,500,000
2,000,000
wind
solar
1,500,000 hydro (㹼1MW)
hydro (1㹼10MW)
geothermal
1,000,000
bioenergy
500,000
0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Figure 16.1 Electricity generation from renewable energy between 2001 and 2015, exclud-
ing large hydropower projects generating more than 10 MW (GWh)
Source: REN21 (2017).
can be expected to reinforce such energy market instabilities but also prevent
severe negative outcomes for the whole economy, especially for energy import-
ing countries, for the overall productive capacity of an economy remains largely
unaffected in times of price shocks if there is a sufficiently developed renewable
energy supply network.
Another way renewable energy can eliminate or reduce energy security risks is
that it can establish a resilient, decentralized power generation system by small to
medium electricity generation projects, which are characteristic of the alternative
energy source that are renewables generally. Because of this, they are available
for small-scale local deployment and can expand access to electricity.
The second category of political rationale for the promotion of renewable
energy is the latter’s fiscal policy aspects of expanding not only investment in
renewable energy facilities and research, but also related industries due to a spill-
over effect. As such, investment in renewables offers governments the opportunity
to increase employment figures, and this, just as significantly, with the generation
of non-temporary jobs. In 2016, the number of workers employed in the renew-
able energy industry globally was estimated to be 9.8 million (REN21, 2017).
The third category of rationale for state-led investment in renewables is that of cli-
mate change solution. The Table 16.1 shows emissions of selected electricity supply
Impact of renewable energy on sustainability 407
technologies (gCO2eq/kWh). Excluding bioenergy and hydro, the median values of
the lifecycle greenhouse gas emissions of renewable energy are quite small, rang-
ing from 11 gCO2eq/kWh for onshore wind power to 48 gCO2eq/kWh for solar
PV (utility). On the other hand, those of coal and gas (combined cycle) are 820
gCO2eq/kWh and 490 gCO2eq/kWh, respectively (IPCC Working Group, 2014).
The forth category of rationale for investment is prevention of air pollution. In
2017, China, one country facing the problem on an enormous scale, announced
that it would invest about USD 360 million in renewable energy primarily because
air pollutants from coal-fired power plants are becoming a serious problem in
main cities (REN21, 2017).
Against the background of these political rationales of investment and the resul-
tant increase of renewable energy, much research has been conducted to ascertain
how renewable energy has affected economic or environmental performance. In
the literature, the causal relationship between economic growth (typically mea-
sured by annual GDP growth rate) and renewable energy is an especially prevalent
topic.3 In the literature, four hypotheses concerning such a causal relationship have
been raised and discussed: bidirectional causality between renewable energy and
economic growth (the feedback hypothesis), unidirectional causality from renew-
able energy to economic growth (the conservation hypothesis), unidirectional
causality from economic growth to renewable energy (the growth hypothesis),
and the absence of causality between them (the neutrality hypothesis). However,
despite such an abundance of studies, the results have been inconsistent.4
Another prevalent topic in the literature is the connection between renewable
energy and environmental, natural, or human factors. The impacts of renewable
energy on CO2 in particular have often been examined with respect to economic
growth, and similarly, a plausible consensus has not been reached. Some studies
Table 16.1 The lifecycle greenhouse gas emissions of electricity supply technologies
Electricity supply technologies Lifecycle emissions
Min Median Max

Coal 740 820 910
Gas-combined cycle 410 490 650
Biomass-cofiring 620 740 890
Biomass-dedicated 130 230 420
Geothermal 6 38 79
Hydropower 1 24 2200
Nuclear 4 12 110
Concentrate solar power 9 27 63
Solar PV-rooftop 26 41 60
Solar PV-utility 18 48 180
Wind onshore 7 11 56
Wind offshore 8 12 35
have raised the possibility of negative effects of biofuels production even while
bioenergy is expected to serve as a potential solution to climate change by extract-
ing carbon from the atmosphere Jacobson (2004) and Johnston and van Kooten
(2015). For instance, Tsao et al. (2012) has examined how non-GHG air pollution
was affected by land use changes which stemmed from expansion of biofuels
production. The study found that direct and indirect land use changes were caused
by the production of sugar cane ethanol and soybean biodiesel in Brazil, and these
land use changes resulted in the emission of more air pollutants from biofuels than
from conventional fossil fuels. Lastly, the effects of renewable energy on human
health have also attracted much attention. Fossil fuels have negative effects on
human health and are the possible cause of respiratory and cardiovascular dis-
eases. Using much cleaner renewable energy in lieu of these would have positive
effects on human health (Pablo-Romero et al., 2016).
Given the preceding considerations, the previous literature has shown that renew-
able energy may bring about positive or negative effects on economic, natural, and
human resources through various causal paths. However, each study focuses only
on one or two causal paths. How, then, should we evaluate the complicated issue
of the impacts of renewable energy? Could we understand how renewable energy
influences us as a whole? Determining whether this question can be answered and
how we might go about formulating a response are the motivations of this study.
We try to understand the impacts of renewable energy in a more comprehensive
way by examining the relationship between renewable energy and the Inclusive
Wealth Index (IWI). IWI is one of the indicators showing the level of sustainabil-
ity, which has been developed recently through a United Nations project (UNEP,
2014). IWI is calculated by multiplying different types of capital by their respec-
tive shadow prices. These capitals are largely categorized into three types: pro-
duced capital (PC), natural capital (NC), and human capital (HC).
Importantly, IWI is based on a stock concept, not on a flow concept like GDP.
While GDP shows annual production amount, IWI indicates capital amount which
will contribute to current and future wealth.
In this study, we employ econometric analysis. We formulate four equations,
each of which employs IWI and its three components (PC, NC, and HC) as
explained variables respectively. For explanatory variables, we use the same eight
variables which include electricity generation capacity from different sources of
renewable energy (solar, wind, traditional bioenergy, modern bioenergy, large-
hydro, and small-hydro), non-renewable energy supply, and GDP.
Using IWI has three advantages. First, it can synthesize various renewable
energy impacts in terms of sustainability. Although we should admit the impor-
tance of clarifying the causal relationship of renewable energy with each of the
relevant factors, it is almost impossible to evaluate renewable energy’s impacts
on sustainability just by enumerating estimated coefficients given by each study,
because there are so many different factors and causal routes. Against this, using
IWI can allow us to sum up each of the different impacts as a monetary base, that
is, a monetary or economic translation of renewable energy’s influence on the
economy. This would give us a simple but plausible view of renewable energy
impacts. Also, considering the many purposes of renewable energy policies as
mentioned earlier, this is practically important and helpful for policymakers. Sec-
ond, using IWI would help to delineate important factors which the conventional
economic indicator of GDP or HDI (the Human Development Index) has failed
to capture. One of these factors is the impact of renewables on natural resources
and/or ecological conditions (UNEP, 2012), which is calculated in IWI as NC.
Third, IWI is superior to GDP due to the fact that it is based on a stock concept,
not a flow concept like GDP. While GDP is annual production amount, IWI indi-
cates current capital amount. In this sense, IWI is more suitable for evaluating the
effects of sustainability, the latter requiring that the current production does not
diminish the wealth of future generations. As yet, there is no available research
which examines the renewable energy impacts on sustainability using IWI other
than the present study. Thus, this study is expected to provide wider views on
renewable energy’s impacts on sustainability.
The remainder of this chapter is as follows. We explain the method employed
and the data in this study in Section 2, show the results in Section 3, and discuss
the results in Section 4.
2 Method and data

In this section, we explain the basic model, the reason why we think variables
employed are adequate, how to calculate each of the variables, and identify its
data source.
2.1 Basic model

In order to examine how renewable energy has affected IWI, we use an interna-
tional panel dataset which contains 117 countries, which were the countries about
which data on renewable energy deployment and the economy was available. The
period under consideration is from the year 2000 to 2014. The basic equation in
our model is as follows:
IWI it = α + β REit + γ NREit + δ GDPit + εit (1)

PCit = α + β REit + γ NREit + δGDPit + εit (2)
NCit = α + β REit + γ NREit + δ GDPit + εit (3)
HCit = α + β REit + γ NREit + δ GDPit + εit (4)
β REit = β1 ( solarit ) + β2 ( windit ) + β3 (traditional _ bioit )
+ β4 (modern _ bioit ) + β5 (large _ hydroit ) (5)
+ β6 ( small _ hydroit )
where IWI, PC, NC, and HC is each value of Inclusive Wealth Index, produced
capital, natural capital, and human capital, respectively. RE represents a vector of
renewable energy variables, which is composed of six different renewable energy
sources shown in the equation (5). NRE is non-renewable energy, and GDP is
gross domestic product. α is the constant term, β represents the impacts of renew-
able energy, which is a vector of six coefficients values of different renewable
energy variables ({β1, β2, β3, β4, β5, β6}). γ and δ are coefficients of NRE and GDP
respectively. εit is a composite error term. We transform all the variables to natural
logarithms.
2.2 Explained variables

As already mentioned, IWI is composed of three large categories of capital: pro-
duced capital (PC), natural capital (NC), and human capital (HC), and the sum of
these three capitals is defined as Inclusive Wealth Index (IWI). In order to show
how renewable energy influences each capital, we formulate four equations (1) –
(4), which have IWI and its three components as explained variables respectively.
The data was retrieved from the IW report of 2017 (UNEP, 2017).
2.3 Explanatory variables in concern: renewable energy

The most important explanatory variables we are concerned with is renewable
energy. In the related literature, units of variables frequently used for renewable
energy are electricity consumption (Wh), electricity generation (Wh), or elec-
tricity generation capacity (W). In this study, we adopt electricity generation
capacity and the figures pertaining to the unit obtained from IRENA’s database
(The International Renewable Energy Agency). The reason for this choice is that
most impacts of renewable energy are assumed to be observed at the time when
actual electricity generation facilities are implemented (for example, invest-
ment, employment or land use change would have already occurred at the time
of a given facility’s installation, without generating electricity). Although in some
respects, we can expect renewable energy to have an impact when electricity is
generated (such as reduced air pollutants), these are undoubtedly less than those
which are brought about at the time installation of facilities. Also, whereas the
amount of electricity generation capacity is constant, electricity generation itself
is not because it is affected by natural conditions and the efficiency of the technol-
ogy employed. Furthermore, considering IWI represents different types of capital
and not their outputs, it makes sense to employ electricity generation capacity as
explanatory variables, which would contribute to the change of capitals more than
electricity generation or electricity consumption.
As shown in equation (5), we use six variables of renewable energy: solar,
wind, traditional bioenergy, modern bioenergy, large-hydro, and small-hydro.
Generally, renewable energy is classified into six large categories: solar, wind,
bioenergy, hydro, geothermal, and marine. In contrast with this general classifica-
tion schema, our classification has two specifications. First, we exclude marine
and geothermal because these two types of renewable energy are utilized only
in a few countries, such that the total amount of electricity generation from that
is relatively small. Second, we classify these large categories of bioenergy and
hydro into two smaller categories respectively because such large categories can-
not easily distinguish between traditional and modern installations.
Regarding hydro, we classify it into two smaller categories depending on the
generation capacity of projects: Large-hydro includes generation projects which
have capacity more than 10 MW, and small-hydro projects have capacity less than
10 MW.
This is because it is generally acknowledged that there are mainly two types
of hydroelectric power: massive hydroelectric dams and small run-of-the-river
plants. Although there seems to be still no agreement whether large-hydro or
small-hydro is economically or environmentally beneficial or not, large-hydro
projects and small-hydro projects is thought to have different effects on sustain-
ability because of its size, necessary land size area, or technology employed.
The general view about the impacts of large-hydro is that it needs massive land
use change for flooding area, harms local ecosystem, and force local community
to relocate to make way for reservoirs. On the other hand, small-hydro is often
expected to be environmentally benign (Paish 2002; Kosnik, 2010).
With bioenergy, we regard solid biofuels as traditional bioenergy, and liquid
biofuels, biogas, bagasse, and renewable municipal waste as modern bioenergy,
which are expected to work as new, cleaner energy (REN21, 2017; these sub cat-
egories are based on IRENA’s classification).
In some countries, there are missing values in renewable energy data and there
are two trends on how these values are missed. First, some countries have no
available values until after one year. In such cases, the value of the first year is
almost close to zero, and therefrom, the values increase. Second, some countries
have a value of zero in some specific types of renewable energy throughout all
the available years featured in the data. We assume that the reason for the missing
values is that the given countries did not implement those specific types of renew-
able energy, and thus replace these missing values with zero.
2.4 Control variables – non-renewable energy (NRE) and GDP

We need to add some control variables to these estimation equations, especially
the ones which are correlated with renewable energy variables but also may affect
explained variables (IWI, PC, NC, or HC). If we omitted such variables in these
equations, these variables would cause biases in the estimators of renewable
energy variables. We assume that such variables are non-renewable energy (NRE)
and GDP. The reason for including non-renewable energy is that it apparently has
a positive correlation with renewable energy and also may affect IWI through
similar factors such as investment, energy price, air pollutants and so on. Also,
including a non-renewable energy variable would provide us with some bench-
marks for judging whether coefficients of renewable energy variables are large or
not. We make non-renewable energy variable by subtracting total primary energy
supply from renewable energy supply in the OECD database5 (OECD), The rea-
son for including GDP is that, as defined in IWI calculation, some output in one
year (GDP) is allocated to either investment or consumption, and this investment
Table 16.2 Descriptive statistics value
Variables Mean sd min max

IWI 28.15 1.667 24.11 33.77
PC 25.99 1.917 22.19 31.49
NC 25.57 2.147 18.96 30.17
HC 27.42 1.836 23.52 33.76
Solar 1.448 2.258 0 10.55
Wind 2.433 3.062 0 11.48
Traditional Bio 1.905 2.769 0 9.077
Modern Bio 2.371 2.521 0 9.218
Small Hydropower 1.646 1.951 0 10.15
Large Hydropower 6.250 3.069 0 12.46
NRE 9.470 1.937 4.865 14.81
GDP 24.90 1.871 21.10 30.32
apparently contributes to an increase of IWI, especially through produced capital

(PC). Also, we assume that renewable energy would expand when GDP increases
because generating more GDP generally requires more total energy. Renewables
would partly meet this increased energy demand. In the Table 16.2, we show
descriptive statistics value in all the variables in this study. All the variables in the
table are natural logarithms forms of the original values.
2.5 Estimation method

Besides the control variables we stated earlier, there would be lots of country-
specific factors influencing IWI, such as population, land area, culture, and the
natural environment (especially wind conditions and intensity of solar radiation
in the context of renewable energy). Importantly, these factors may correlate with
renewable energy variables, which would, in fact, cause the same problem of
omitted variables (i.e., the problem of the biased estimators). In order to avoid this
problem, we control these effects by using a fixed estimation method, under the
assumption that these effects are time constant during the period targeted in the
analysis. In this fixed method, we transform the composite error terms in equa-
tions (1)–(4) as shown in equation (6),
εit = µi + vit (6)

E(vit | X it , µi ) = 0, t = 2000, 2001, … , 2014 (7)
where μ1 is the time constant variable specific to individual country, Xit is a matrix
of all the explanatory variables. The idiosyncratic term vit is assumed to satisfy
equation (7). Now we explain the basic idea of this method by using equation
(1) as an example. By substituting the right side of the equation (6) for the equa-
tion (1), we get the following equation.
IWI it = α + β REit + γ NREit + δGDPit + µi + vit (8)
And, by taking the average of both sides in the equation, we get the following
equation.
IWI it = α + β REit + γ NREit + δ GDPit + µi + vit (9)
Then, by subtracting (9) from (8), we get the following equation.
IWI it − IWI it = β ( REit − REit ) + γ ( NREit − NREit )

(10)
+ δ (GDPit − GDPit ) + (vit − vit )
Equation (10) can be rewritten like the following equation,
= β RE
IWI + γ NRE
+ δGDP
+ v (11)
it it it it it
where each variable is defined as follows:
= IWI − IWI
IWI it it it
REit = REit − REit

= NRE − NRE
NRE it it it

GDP = GDP − GDP
it it it
vit = vit − vit
We transform the equations (2)–(4) in the same way.

In this study, we use panel data with a large N (N = 117). As is often said, with
such data, we face the heteroscedasticity problem, which might make it diffi-
cult for us to get efficient estimators by using the ordinary least squares method
(OLS). Thus, we adopt the feasible generalized least squares method (FGLS).
3 Empirical results
In Table 16.3, we show our empirical results.
In most variables, the coefficients are statistically different from zero at 1% lev-
els. The coefficient values also vary, ranging from −0.08 to 0.91. Comparing the
coefficients of each variable in equation (1), the variables of solar, large-hydro,
and non-renewable energy (NRE) have positive coefficients, while those of wind,
traditional bioenergy, modern bioenergy, and small-hydro have negative coeffi-
cients. Next, we indicate how renewable energy affects each type of capital (PC,
NC, and HC) by pointing out these coefficients. First, with respect to PC, all vari-
ables except modern bioenergy have positive coefficients. This result is apparent
because PC is calculated based on manufactured capitals, and renewable energy
Table 16.3 Empirical results
Variables (1) (2) (3) (4)
IWI PC NC HC
Solar 0.0531*** 0.0187*** −0.0804*** 0.0886***
(0.00391) (0.00152) (0.00901) (0.00564)
Wind −0.0590*** 0.0327*** −0.131*** −0.0438***
(0.00442) (0.00136) (0.00611) (0.00681)
traditional bio −0.0207*** 0.0197*** −0.0302*** −0.00428
(0.00388) (0.00138) (0.00644) (0.00545)
modern bio −0.0208*** −0.0205*** −0.0447*** −0.00811**
(0.00410) (0.00150) (0.00585) (0.00377)
small-hydro −0.0684*** 0.00979*** −0.0830*** −0.0743***
(0.00449) (0.00187) (0.0109) (0.00642)
large-hydro 0.110*** 0.00453*** 0.267*** 0.130***
(0.00388) (0.000847) (0.00796) (0.00416)
NRE 0.115*** 0.0323*** 0.393*** 0.401***
(0.0163) (0.00479) (0.0196) (0.0227)
GDP 0.526*** 0.910*** 0.366*** 0.133***
(0.0193) (0.00602) (0.0260) (0.0230)
Constant 13.46*** 2.891*** 11.78*** 19.47***
(0.342) (0.106) (0.478) (0.364)
Observations 1755 1755 1755 1755
Number of cou_id 117 117 117 117
Standard errors in parentheses
***p < 0.01.
**p < 0.05.
investment directly induces an increase of those capitals. A possible explanation

for the negative coefficient of modern bioenergy is that investment for bioen-
ergy leads to less investment in other industries, which could actually contrib-
ute more to PC than bioenergy itself.6 Second, regarding NC, the coefficients of
large-hydro and non-renewable energy are positive, and those of other renewable
energy variables are negative. This result is opposite to a generally accepted view
that conventional hydropower and non-renewable energy would have negative
effects on the natural environment. The negative coefficients of other types of
renewable energy might be caused, as indicated in previous literature, by changes
in land use, for example. Third, regarding HC, the coefficients are positive in
solar, large-hydro and non-renewable energy (NRE), while wind, modern bio-
energy and small-hydro have negative coefficients. The possible reason for the
HC increase is that the introduction of renewable energy stimulates labor mar-
kets and leads to an improvement of average wages or employment rates. Among
variables which have positive coefficients, large-hydro and non-renewable energy
have some common features. They have been used for a much longer time and so
their prices are much lower than modern renewable energy sources. Especially for
industrial sectors, lower energy costs can bring about positive effects on produc-
tivity and therefore also wages and/or employment rates.
As for the size of the coefficients, we can ascertain the following by comparing
renewable energy’s variables with non-renewable energy and GDP variables. The
1% increase of solar, for instance, can contribute to the improvement of IW by
0.05%, which is one-tenth more than that of GDP and two-fifths more than that of
non-renewable energy.
4 Discussion
According to the preceding results, we find that the increases in renewable energy
generation capacity have statistically significant effects on IWI and its three com-
ponents of PC, NC, and HC. Moreover, we find that these effects vary depend-
ing on the types of renewable energy and explained variables. In particular, it is
important to note that some renewable energy, which has been generally acknowl-
edged as “cleaner energy”, would have negative effects on the natural environ-
ment, while non-renewable energy and large-hydro have at least some positive
effects. Also, we should pay attention to which of the three types of capital have
positive or negative effects. For instance, in the case of solar power, although IWI
itself can be increased by the promotion of solar power, this is mainly because of
the increase of PC and HC, and in fact may cause a reduction of NC.
Finally, we report two limitations of our study. One is that our estimation
model based on the assumption that there is no serial correlation in composite
error terms. In fact, one economic shock captured in a composite error term at any
one time may indeed exist at any subsequent time. For example, it is shown that
large-scale natural disasters deplete NC (Rajapaksa et al., 2017), and it would take
a long time to recover such losses. In this example, the negative shock captured in
the composite error term would continue for a while, which might cause positive
correlations between composite error terms in adjacent time periods. Second, as
stated earlier, there may be the possibility of biased coefficients caused by omit-
ting the energy price variable. In this study, we could not use the panel data of
energy price for each type of renewable energy because there is no available data.
These are subjects for future analysis.
Notes
1 In this chapter, when we refer simply to “renewable energy”, it includes solar, wind,
bioenergy, geothermal, hydro, and marine energy.
2 A global agreement within the United Nations Framework Convention on Climate
Change (UNFCC) with a goal of reducing greenhouse gas emission.
3 For instance, see Tugcu et al. (2012), Asafu-Adjaye et al. (2016), Wong et al. (2013),
Apergis and Payne (2014), Apergis and Payne (2012), and Menegaki (2011). These stud-
ies often use time-series and panel data econometric techniques such as autoregressive
distributed lag (ARDL), a panel error correction model, and the granger causality test.
The period of data is from as early as 1980 to as late as 2013. The size of sample coun-
tries ranges from 1 to 80. In these analyses, sample countries are often categorized
according to their country-specific characteristics or region, such as OECD, EU coun-
tries, developed countries, or African countries.
4 Sebri (2015) has conducted meta-analysis of empirical literature on the connection
between renewable energy consumption and economic growth, finding that the variation
in the indicated results is due to a number of characteristics including model specifica-
tion, data characteristics, estimation techniques, and developmental level of sampled
countries.
5 In the OECD database, total primary energy supply is calculated as energy production
plus energy imports, minus energy exports, minus international bunkers, then plus or
minus stock changes. Renewable energy is defined or calculated as the contribution of
renewables to total primary energy supply.
6 In this case, correlation between renewable energy and such other industry included in
idiosyncratic term might cause the biased estimator problem.
Bibliography
Apergis, N., and Payne, J. E., “Renewable energy consumption and growth in Eurasia”,
Energy Economics, vol. 32, pp. 1392–1397, 2010.
Apergis, N., and Payne, J. E., “Renewable and non-renewable energy consumption-growth
nexus: Evidence from a panel error correction model”, Energy Economics, vol. 34,
pp. 733–738, 2012.
Apergis, N., and Payne, J. E., “The causal dynamics between renewable energy, real GDP,
emissions and oil prices: Evidence from OECD countries,” Applied Economics, vol. 46,
no. 36, pp. 4519–4525, 2014.
Apergis, N., Payne, J. E., Menyah, K., and Wolde-Rufael, W., “On the causal dynamics
between emissions, nuclear energy, renewable energy, and economic growth,” Ecologi-
cal Economics, vol. 69, pp. 2255–2260, 2010.
Asafu-Adjaye, J., Byrne, D., and Alvarez, M., “Economic growth, fossil fuel and non-fossil
consumption: A Pooled Mean Group analysis using proxies for capital,” Energy Econom-
ics, vol. 60, pp. 345–356, 2016.
Boysen, L. R., Lucht, W., and Gerten, D., “Trade-offs for food production, nature conserva-
tion and climate limit the terrestrial carbon dioxide removal potential,” Global Change
Biology, vol. 23, no. 10, pp. 4303–4317, 2017.
Bringezu, S., O’Brien, M., and Schütz, H., “Beyond biofuels: Assessing global land use for
domestic consumption of biomass: A conceptual and empirical contribution to sustainable
management of global resources,” Land Use Policy, vol. 29, no. 1, pp. 224–232, 2012.
Frankfurt School FS-UNEP Collaborating Centre, “Global trends in renewable energy
investment 2017,” 2017.
International Energy Agency, “Contibution of renewables to energy security,” 2007.
The International Renewable Energy Agency, http://resourceirena.irena.org/gateway/dash-
board/, [Online].
IPCC Working Group, “Mitigation of climate change, annex III: Technology-specific cost
and performance parameters,” 2014.
Johnston, C. M. T., and van Kooten, G. C., “Back to the past: Burning wood to save the
globe,” Ecological Economics, vol. 120, pp. 185–193, 2015.
Jacobson, M. Z., “The short-term cooling but long-term global warming due to biomass
burning,” Journal of Climate, vol. 17, no. 15, pp. 2909–2926, 2004.
Kosnik, L., “The potential for small scale hydropower development in the US,” Energy
Policy, vol. 38, no. 10, pp. 5512–5519, 2010.
Menegaki, A. N., “Growth and renewable energy in Europe: A random effect model with
evidence for neutrality hypothesis,” Energy Economics, vol. 33, pp. 257–263, 2011.
OECD, https://data.oecd.org/energy/renewable-energy.htm#indicator-chart, [Online].
Ohler, A., and Fetters, I., “The causal relationship between renewable electricity generation
and GDP growth: A study of energy sources,” Energy Economics, vol. 43, pp. 125–139,
2014.
Pablo-Romero, M. P., Román, R., Sánchez-Braza, R., and Yñiguez, R., “Renewable Energy,
Emissions, and Health,” In Cao, W. and Hu, Y., Renewable Energy (pp.173–198), Inte-
chOpen, DOI: 10.5772/61717, from: https://www.intechopen.com/books/renewable-
energy-utilisation-and-system-integration/renewable-energy-emissions-and-health.
Paish, O., “Small hydro power: Technology and current status,” Renewable and Sustain-
able Energy Reviews, vol. 6, no. 6, pp. 537–556, 2002.
Rajapaksa, D., Islam, M. and Managi, S., “Natural capital depletion: The impact of natural
disasters on inclusive growth,” Munich Personal RePEc Archive, 2017.
REN21, “Renewables 2017 global status report,” 2017.
Sebri, M., “Use renewables to be cleaner: Meta-analysis of the renewable energy
consumption-economic growth nexus,” Renewable and Sustainable Energy Reviews,
vol. 42, pp. 657–665, 2015.
Tsao, C. C., Campbell, M. Mena-Carrasco, S. N. Spak, S. N., Carmichael, G. R., and Chen,
Y., “Biofuels that cause land-use change may have much larger non-GHG air quality
emissions than fossil fuels,” Environmental Science & Technology, vol. 46, pp. 10835–
10841, 2012.
Tugcu, C. T., Ozturk, I., and Aslan, A., “Renewable and non-renewable energy consump-
tion and economic growth relationship revisited: Evidence from G7 countries,” Energy
Economics, vol. 34, pp. 1942–1950, 2012.
UNEP, “Inclusive Wealth Report,” 2012.
Wong, S. L., Chang, Y., and Chia, W., “Energy consumption, energy R&D and real GDP in
OECD countries with and without oil reserves,” Energy Economics, vol. 40, pp. 51–60,
2013.
17 Consumer demand for fully
automated driving technology1
Kong Joo Shin and Shunsuke Managi
1 Introduction
Artificial intelligence (AI) technology has advanced exponentially in recent years,
and technologies such as ‘Siri’ and ‘Pepper’2 have been installed in the machines
that consumers use daily. AI technology is also used in automotive industries to
develop fully automated vehicles (FAV), which enable people to use cars without
driving. Automated driving technology has already been tested or used in public
transportation systems and on freeways in different countries. For example, in
2012, Spain conducted a successful platooning experiment with FAV on public
roads. The UK government and several relevant industries are jointly planning to
introduce FAV for use in public transportation, which will connect Heathrow Air-
port with Bristol, London, Milton Keynes, and Coventry by 2017.3 In Japan, glob-
ally recognized vehicle companies such as Toyota and Nissan have been testing
their automated driving technologies on freeways and local roads. Additionally,
the Japanese government plans to introduce FAV on selected roads by 2020. It is
likely that drivers will begin to see FAV on ordinary roads sooner than expected.
Automated driving (AD) offers a variety of benefits. According to an official
report on the AD trend by the Ministry of Land, Infrastructure, Transport and
Tourism (MLIT) (2015),4 accident reduction and traffic mitigation are two of its
major advantages. For example, 96% of traffic accidents on freeways in Japan
are due to human errors such as mishandling, carelessness, and misjudgments
by drivers; it is expected that automated driving technology will eliminate these
accidents. Additionally, approximately 60% of Japan’s traffic congestion occurs
at sag sections of roads, and another 20% occurs at tunnel entrances. Financial
loss from traffic congestion in Japan is estimated at approximately JPY 12 trillion
per year (USD 10.4 billion). FAV are expected to contribute significantly to reduc-
tions in road congestion. Furthermore, almost 40% of Japanese drivers are elderly,
and their mishandling of vehicles is a frequent cause of fatal accidents. In recent
years, these incidents have been reported regularly in the media and have raised
concerns among the public.
Despite the promise of FAV, their introduction also raises concerns about addi-
tional purchasing and maintenance costs and possible information leakages from
their software, as the recording of private information may contribute to various
Fully automated driving technology 419
crimes. Moreover, there are ongoing debates about policy issues related to road
regulation. The introduction of AD technology will largely be determined by the
decisions made by policymakers. Additionally, insurance regulation is an impor-
tant issue, as the definition of accidents would change with FAV on the road.
There is widespread interest in FAV from consumers, policymakers and related
businesses. Surveys have been used to study consumer demand for AD, focusing
mainly on their largest potential markets: the United States, the EU, China, India,
Australia, and Japan. Bekiaris et al. (1997) provided one of the earliest studies on
FAV demand using data on 407 respondents from nine countries. The study found
that people were in favor of driving assist systems, which urge drivers to pay
attention to their driving, but the respondents showed concern and disapproval
regarding FAV. Other surveys have been conducted more recently. Google tested
automated driving in 2012, by which time the HAVE it project and SARTRE
project had also been conducted; thus, later surveys were conducted under the
circumstances where fully AD seemed more realistic to consumers, and accept-
ability was also significantly increased compared to a few decades earlier. J. D.
Power (2012) used data collected from 17,400 car owners in the United States
and found that approximately 37% responded that they ‘will purchase’ or ‘will
probably purchase’ FAV. However, when respondents were asked the same ques-
tion about purchase intention (PI) assuming an additional cost of USD 3000, only
20% had a positive PI. Hence, consumers’ PI is conditional on the expected cost
of AD technology.
There are several surveys that contain Japanese samples: Continental (2013),
Aucnet (2014), and BCG (2015). Continental (2013) collected approximately
1200 consumer samples in Germany, China, Japan, and the United States. Accord-
ing to the data, the recognition rates for AD were 67%, 64%, 29%, and 50% in
Germany, China, Japan, and the United States, respectively. On the other hand,
the shares of people who wanted AD to be available were 19%, 44%, 39%, and
23% in Germany, China, Japan, and the United States, respectively. Additionally,
the shares of respondents with positive expectations of using automated vehicles
on a freeway were 17%, 36%, 39%, and 28% in Germany, China, Japan, and the
United States, respectively. These results showed significantly low recognition of
AD in Japan. The survey also asked about expectations regarding the locations
where AD would be used. Among sampled countries, the Japanese have shown
relatively high acceptance of AD. Moreover, while 61% of Japanese respondents
answered that they were ‘more inclined to agree that automated driving is a useful
advancement’, 43% of Japanese respondents answered that they ‘don’t believe
that automated driving will function reliably’. Thus, Japanese consumers do think
AD would enhance their daily lives, but they also have concerns about the reli-
ability of the related technology.
Aucnet (2014) and BCG (2015) have conducted consumer surveys about AD
and FAV in Japan. Aucnet collected 1119 samples through the internet, and the
results show that 16% of respondents ‘would like to purchase’, 34% ‘would prob-
ably like to purchase’, 7% ‘would not want to purchase’, 6.9% ‘will not purchase’,
and 36.6% ‘do not know’ about purchasing FAV. In sum, approximately 70% of
420 Kong Joo Shin and Shunsuke Managi
respondents had a positive attitude toward automated driving. BCG (2015) also
conducted an internet survey and received 1583 responses from Japanese con-
sumers who intended to purchase a car or had recently bought one at the time of
the survey. The data suggested that between 40% and 50% of respondents were
inclined to purchase FAV. The report suggested that consumers’ main reasons for
being interested in purchasing FAV were ‘utility from automated driving on a
freeway and during traffic congestion’, ‘increased safety for elders who drive’
and being ‘attracted to newly state-of-the-art technology’. Furthermore, the sur-
vey asked about willingness to pay (WTP) for partial and fully automated driving
systems (FADS) and showed that WTP for each system was approximately JPY
100,000 (USD 870) and JPY 200,000 (USD 1740), respectively.
Previous surveys and reports on AD and FAV provide basic information on the
demand for these technologies. However, the results and analyses of previous
studies do not provide details about consumers’ anticipated benefits and concerns
with regard to AD and FAV becoming a standard presence on the road. Moreover,
there are limited studies examining the determinants of consumers’ demand for
AD and FAV. The study by Payre et al. (2014) is one of the few to have examined
the factors affecting perceptions of AD; it found that consumers who owned cars
with driving assistance systems, such as adaptive cruise control (ACC) or a lane
keeping system (LKS), were more likely to be positive about purchasing FAV.
Our study provides deeper insight into consumer demand by surveying the
subjective advantages and disadvantages of AD. We analyze the determinants of
purchase intention (PI) and WTP for FADS using originally collected Japanese
household survey data obtained in 2015. To our knowledge, our survey provides
the most recent and largest consumer dataset related to AD. Our analysis com-
bines objective area data such as population density and traffic accident data with
survey data. In addition, we provide municipality-level analysis, combining the
aggregated individual-level survey data and city-level objective data.
The rest of this chapter is organized as follows. Section 2 provides details on
the data and describes the variables. Section 3 provides descriptive statistics. Sec-
tion 4 presents the estimation model and Section 5 provides estimation results and
discusses the results. Section 6 concludes.
2 Data and variables

This study uses an original internet survey conducted in November and December
2015, which received 246,642 responses. Our survey builds on previous related
surveys and offers significantly expanded coverage of questions on consumers’
perceptions of AD and FAV; it also includes individual and household characteris-
tics. An internet survey has the advantage of avoiding the interviewer bias caused
by arbitrary factors – such as the appearance or gender of interviewers – when
responding to sensitive questions such as household income, employment status
and WTP (Welsch and Kühling, 2009). Moreover, given the extensive accessibil-
ity of the internet in Japan, it is a relatively time-and cost-efficient method of col-
lecting data from a large sample compared to a face-to-face survey.
The survey contained several main questions regarding consumer demand for
AD and FAV: (1) the expected time frame in which FAV would be available for
purchase, (2) the purchase intention of FAV when they become available, and (3)
the additional amount the respondent would be willing to pay for FADS. We asked
‘when do you think FAV will be available in the market?’ and respondents were
given several time range options.
To assess PI of FAV, respondents were given four options: (1) will purchase,
(2) will consider purchasing, (3) will not purchase, and (4) do not know. We con-
structed three dummy variables for purchase intention (PI) using responses to
(1)–(3) and dropped those respondents who had chosen (4) from the regression
analyses. Additionally, respondents were asked to choose a range of WTP for
FADS from 21 options, which ranged up to JPY 3 million (approximately USD
26,000). Then, we eliminated the respondents who replied that they did not know
their WTP for FADS.
We use several categories of variables that could be determinants of consumer
demand for FAV or FADS. As main explanatory variables, we use dummy vari-
ables constructed from the merits and demerits associated with AD. We provide a
full list of merit and demerit options in Tables 17.1 and 17.2. These tables provide
the share of respondents who selected each option, and the results are further
described in the next section. We also use factor analysis to create combined indi-
ces of the merits and demerits of FAV. As shown in Table 17.3, we found three
factors per merit and demerit with eigenvalues larger than 1.
Table 17.1 The results of selected merits (multiple selections and top three selections)
Rank Merit options Multiple Top 3

1 Mitigating transportation problems involving elders 45.44% 29.92%
2 Getting off at designated places, and automatic parking 37.25% 21.01%
3 Overall reduction of the burden from driving 36.42% 18.92%
4 Automatically braking in the cases of danger 35.54% 14.66%
5 Reducing traffic accidents due to human errors 32.43% 18.38%
6 Burden reduction from long trips 32.32% 13.39%
7 Getting on at designated places 31.67% 14.31%
8 Able to switch between automatic and manual drive 28.60% 7.91%
9 Can effectively use traveling time 23.95% 8.20%
10 Automatic starting according to signals 22.19% 2.55%
11 Automatic lane change, overtaking, and merging 20.85% 3.14%
12 Automated transportation of goods 18.12% 5.44%
13 Will not need the driver’s license 12.18% 4.36%
14 Accident would not be the driver’s responsibility 11.72% 4.31%
15 Extended accessibility 10.72% 2.08%
16 Children can ride without supervision 4.01% 0.57%
17 The status of having automated vehicles 2.02% 0.24%
Table 17.2 The results of selected demerits (multiple selections and top three selections)
Rank Demerit options Multiple Top 3

1 Possible traffic accidents by technical malfunctions 53.76% 43.48%
2 The obscurity of responsibility in the accidents 48.63% 31.43%
3 Increase in initial costs and maintenance costs 42.37% 25.26%
4 Children may misuse the car without supervision 40.08% 22.56%
5 The third party can miuse the car 35.37% 18.58%
6 Increased traffic quantities 27.23% 9.36%
7 Reaching wrong destinations due to system malfunctions 25.63% 7.85%
8 Needs to learn new operative system 20.40% 8.97%
9 Recording of all routes and destinations 13.24% 4.38%
10 Possible leakage of private information 11.83% 3.09%
11 Cannot drive over speed limit 9.50% 2.99%
12 Difficult to remodel the cars 3.88% 0.75%
Table 17.3 The results of factor analyses: merit and demerit factors
Factors Options with high weight loading for a given factor Factor description
Merit 1 Mitigating transportation problems involving elders Burden reduction
Burden reduction from long trips from driving
Overall reduction of the burden from driving
Switching between automatic and manual drive
Reducing traffic accidents due to human errors
Automatic braking in the cases of danger
Automatic starting according to signals
Automatic lane changes, overtaking, and merging
Merit 2 1: Getting off at designated places, automatic parking Automatic arrival
2: Getting on at designated places and parking
3: Automated transportation of goods at designated
destinations
Merit 3 7: Will not need the driver’s license Do not need
9: Children can ride without supervision driver’s license
11: Accident would not be the driver’s responsibility
Demerit 1 3: Children may misuse the car without supervision Uncertainties and
6: The third party can miuse the car risk of AD and
7: Possible traffic accidents by technical malfunctions FAVS
10: The obscurity of responsibility in the accidents
11: Increased traffic quantities
12: Reaching wrong destinations due to system
malfunctions
Demerit 2 4: Recording of all routes and destinations Concerns regarding
5: Possible leakage of private information information
security
Demerit 3 1: Needs to learn new operative system Restrictions
8: Cannot drive over speed limit
9: Difficult to remodeling the cars
We also construct mobility-related variables that may be related to the demand
for FAV or FADS: the number of car trips per day and average driving time per
trip, purpose of car trips, commuting time, driver’s license dummy, reasons for car
ownership, and reasons for not owning a car. We also ask about dissatisfactions
with the traffic environment and about where the respondents expect to use FAV,
with the options of general roads, freeways, and inner cities. In addition, we use
individual and household characteristics such as respondents’ age, gender, marital
status, number of household members and number of children in a household,
education by university graduation, household income, occupations, and subjec-
tive health evaluation.
Furthermore, we use citizen identification information from the ‘2015 Basic
Resident Register’ to construct municipality-level variables, including population
density, number of annual traffic accidents, and injuries /deaths related to traffic
accidents. We also use another data source, CASBEE (Comprehensive Assess-
ment System for Built Environment Efficiency), to obtain municipality averages
of taxable household income and the share of households with elders. Finally, we
use prefecture dummies to control for unobservable regional characteristics that
may affect respondents’ demand for FAV.
Out of 246,642 respondents, we eliminated 6085 respondents without valid
postal codes and for which we could not merge objective municipality-level data
from alternative sources. We also dropped 98 observations from municipalities
that lacked traffic accident data and population density data, and we dropped 412
respondents who resided in municipalities without data on taxable household
income. We were left with 240,054 responses. Furthermore, we dropped 49,717
respondents who did not provide household income. Finally, we eliminated
respondents who answered that they ‘do not know’ their WTP (51,965 samples)
or PI (51,701 samples) of FADS; 188,089 and 136,388 samples were used for
WTP and PI analyses, respectively.
3 Descriptive analysis
3.1 Time frame of FAV availability in the consumer market

Figure 17.1 shows the results of consumers’ perceptions regarding the timing of
FAV becoming available in the market. On average, respondents anticipated that
FAV would be available for purchase in approximately 13 years. This is rather
consistent with their predicted actual availability according to the Automated
Vehicle Symposium (2014), which indicated that FAV would be available for
purchase in 2030. Hence, our average time frame matches the general expecta-
tion. In our survey, 53% of respondents answered that they expect AD to be on
the market within 15 years. This is somewhat higher than the 37% of respondents
in the survey of Mobility (2013) who stated that AD technology would be avail-
able on the consumer market within 15 years. While the mean is approximately
13 years, the most popular time range was 6–10 years, with almost 40% of
respondents choosing this option. On the other hand, approximately one-quarter
50~ 3.49%
41~50 0.47%
31~40 1.36%
21~30 5.74%
Years later
16~20 9.00%
11~15 12.89%
6~10 26.38%
1~5 12.37%
~1 0.94%
I don't know 27.35%
0% 5% 10% 15% 20% 25% 30%
Figure 17.1 When would FAV be available in the market?
2030
2029
2028
2027
year
2026
2025
2024
2023
~29 30~39 40~49 50~59 60~69 70~
age
all male female
Figure 17.2 When would FAV be available in the market? (by age groups)
of respondents answered that they ‘don’t know’ the time frame in which AD
would be available.
Figure 17.2 shows the distribution of expected time ranges of FAV availabil-
ity in the market by age group. The result implies that older generations expect
that FAV will be available sooner compared to younger generations. In particular,
there is a visible difference between respondents above and below the age of 50.
This trend may reflect elders’ higher anticipation that FAV will be available soon;
it differs from the results of Aucnet (2014), which found that elders had negative
views of AD. This difference may be because Aucnet’s survey had only slightly
more than 60 samples in total, and given the sample size of our data, our result
seems more reliable.
3.2 Perceptions about the merits and demerits of automatic driving

Tables 17.1 and 17.2 show the results of the merit and demerit options selected
by the respondents.5 The respondents were given 17 merit options and 12 demerit
options. Then, they were allowed to choose unlimited options, and the results are
provided in the multiple columns. If they had chosen more than three options, the
respondents were also asked to choose the top three merits or demerits, and the
results are provided in the ‘top 3’ columns. While the average share per option is
lower in the ‘top 3’ columns, we see very little change in the rankings between
multiple and the ‘top 3’ columns.
The results of the merit options in Table 17.1 indicate that consumers have high
expectations of FAV being a useful tool in the mitigation of mobility problems
and accidents related to elderly drivers. Almost half of the respondents chose this
option, and it was also ranked first in the question on the top three merits. Addi-
tionally, reduced traffic accidents and options related to improving the comfort
and convenience of driving were also popular options. On the other hand, options
such as ‘children can ride on their own’ and ‘having FAV to raise status and repu-
tation’ were rarely chosen. Additionally, people do not seem to regard FAV as a
mobility tool that can expand their current mobility.
We have compared the selection rate of children’s independent mobility as a
merit of FAV in the full sample and the sub-sample of respondents with children
(N = 30,774) and have found little difference in the rates. However, we found a
significant difference between full samples and sub-samples of respondents with-
out a drivers’ license (N = 23,150) in the selection rate of the option ‘will not need
driver’s license’ with FAV; approximately 12% of the full sample and 29% of
the sub-sample chose that option. Moreover, approximately 19% of respondents
without licenses chose this option as one of their top three merit options, while
only 4% of the entire sample chose this as one of their top three merit options.
This result clearly reflects the different expectations between drivers and people
without licenses.
Table 17.2 shows the respondents’ selection rates of the 12 demerits. Over-
all, the results indicate consumers’ strong concerns regarding the technological
dependability and safety of FAV, as well as their concerns about the additional cost
of this new and not yet available technology. These concerns are in accordance
with the findings of BCG (2015). Although the MLIT considers the possibility of
information leakage a serious issue, cost and the robustness of technology seem to
be consumers’ main concerns. Nevertheless, information security is an important
issue, and, as the introduction of FAV on public roads becomes more realistic,
there will be increased scrutiny of software issues, including the possible mal-
functioning and mishandling of stored information.
We also use the merit and demerit categories identified by factor analyses. As
shown in Table 17.3, the merit and demerit options each had three factors with
eigenvalues larger than 1. Three merit factors are (1) reduced driving burden;
(2) automatic driving to designated destinations and automatic parking; and (3)
non-requirement of driver’s license. The first category of burden reduction had
high weight loadings among the merit options with high selection rates, as shown
in Table 17.1. A study by BCG (2015) showed that US consumers highly valued
‘increased safety with FAV’, ‘reduction of insurance cost’, and ‘improved produc-
tivity due to efficient usage of time’ as major benefits of FAV. Japanese consumers
also value the safety and increased productivity associated with AD. On the other
hand, the three factors from the demerit options are as follows: (1) uncertainties of
AD, (2) concerns about information security, and (3) restrictions on driving speed.
3.3 Purchase intention and WTP for FAV or FADS

Figure 17.3 shows that approximately 12% of respondents answered that they
‘will purchase’ and 35% of respondents answered that they ‘will consider pur-
chasing’ FAV or FADS. A near-majority of respondents are inclined to purchase
FAV or FADS. Approximately 20% of respondents answered that they ‘will not
purchase’ these technologies. This share is similar to the figure in BCG’s (2015)
report. Additionally, 32% respondents answered that they ‘do not know’. One of
the reasons for this relatively large size of the agnostic group may be lack of inter-
est and information about AD. PI per sub-sample groups indicate that men are
somewhat more inclined to purchase, but PI does not seem to significantly vary
with age. Additionally, the result shows that respondents who do not own a car
The whole 12% 35% 20% 32%
Women 9% 31% 19% 41%
Men 15% 38% 22% 26%
Not holding driver's license 12% 22% 11% 55%
Not possessing vehicles 13% 28% 14% 46%
Possessing vehicles 12% 37% 22% 29%
Over the age of 60 13% 36% 20% 31%
they will purchase they will consider purchasing they will NOT purchase they don't know
Figure 17.3 Purchase intention for FADS by consumer characteristics (N = 246,642)

Map 17.1 Distribution of purchase intention at municipality level
or do not have a drivers’ license have lower PI compared to car owners but more
often respond that they ‘do not know’ their PI. As shown in Map 17.1, we observe,
overall, that the municipalities with the highest PI are located in the Hokkaido
region and in the non-coastal inner areas of Japan, where cars play a relatively
more important role in daily mobility.
Figure 17.4 shows the result of WTP for FADS by consumer characteristics.
The sub-sample of respondents without a driver’s license seems to have a rela-
tively lower PI, but they are willing to pay relatively more for FAV. This result
seems to imply that respondents without a driver’s license are polarized: either
they are uninterested in technology and unwilling to pay for it or very interested in
purchasing FADS and willing to pay much more than average drivers. Addition-
ally, similarly to PI, men have a higher WTP than women. However, in contrast to
the results of PI, elders’ WTP is significantly higher than average. This result may
be partly due to elders’ high expectations of the benefits of AD and their relatively
high household incomes.
The whole
Men
Women
Having children
Not holding driver's license
Not possessing vehicles
Possessing vehicles
Over the age of 60
0 5 10 15 20 25
The willingness to pay (10⁴yen)
Fully automated driving When parking In traffic congestion On a freeway
Figure 17.4 WTP of FADS by consumer characteristics (N = 188,089)
0-10
10-20
20-30
30-40
40-50
50-100
100 +
Map 17.2 Distributions of WTP at municipality level

*Numerical ranges in JPY 10,000.
We present two maps of WTP distribution: Map 17.2 is the map of WTP dis-
tribution of all respondents except for those who answered that they do not know
their WTP. Map 17.3 maps the WTP of respondents who answered that they will
purchase FADS. The comparison of these maps indicates a strong correlation
between the WTP and PI of consumers.
0-10
10-20
20-30
30-40
40-50
50-100
100+
Map 17.3 Distributions of WTP at municipality level: consumers with high PI

*Numerical ranges in JPY 10,000.
4 Estimation model
To examine the determinants of consumer demand for FAV and FADS, we use
OLS regression analysis for WTP and an ordered logistic regression analysis for
PI. The estimation equation is as follows:
WTPi = Constant + [ Αi ]⋅[ Merits ] + [Βi ]⋅[ Demerits ] + [ Γi ]⋅[transfer preferences ]

+ ∑(δi ⋅ Individual attributions ) + Errori
Αi , Βi , Γi , and δi are coefficients and [∙] are a matrix of variables.

We run two estimation equations for WTP, one with separate merit and demerit
dummies and the other with merit and demerit factors.
The estimation equation using an ordered logit model for PI is as follows:
 PI k 
ln   = Constant + [ Αi ]⋅[ Merits ] + [Βi ]⋅[ Demerits ]
1− PI k 
i
+ [ Γi ]⋅[Transfer preferences ]
+∑(δi ⋅ Indevidual attributions) + Errori
We also run two estimations for PI with separate sets of merit and demerit factors,
similar to the WTP analyses.
Furthermore, we use aggregate data at the municipality-level combined with
data. We use the following estimation models for WTP and PI.
WTPi = Constant + [ Αi ]⋅[ Merit rates ] + [Βi ]⋅[ Demerit rates ]

+ [ Γi ]⋅[Transfer preference rates ] + ∑(δi ⋅ Regional attributions) + Errori
 PI on averagek 
ln   = Constant + [ Αi ]⋅[ Merit rates ] + [Βi ]⋅[ Demerit rates ]
1− PI on averagek 
i
+ [ Γi ]⋅[Transfer preference rates ]

+ ∑(δi ⋅ egional attributions) + Errori
5.1 Individual-level analysis

Table 17.4 shows four regression results. For each WTP and PI, we provide the
results with separate merit and demerit options, and with the factors of the merits
and demerits calculated using factor analysis. The results suggest that the fol-
lowing merit options positively affect consumer demand for FADS: supporting
the elderly, reducing traffic accidents, and reducing the burden of driving. On the
other hand, demerit options such as an increase in initial and maintenance costs,
information leakage to third parties, and possible malfunctions were found to
negatively affect consumer demand. In particular, the merit of FAV in improving
elders’ driving has a positive and relatively large impact on WTP. The expected
improvement in mobility with FAVS, as well as the expected reduction in traffic
accidents, also had relatively large positive coefficients.
Two demerits, ‘the possibility of malfunction or accident’ and ‘unclear respon-
sibility when a traffic accident happens’, had negative effects on PI but positive
effects on WTP. This result may reflect consumers’ perception that taking mea-
sures against malfunctioning and setting clear regulations about responsibility for
traffic accidents would increase their willingness to purchase and make FADS
Table 17.4 Determinants of consumer demand of FAV and FADS (individual-level
analysis)
Variables WTP (1) WTP (2) PI (1) PI (2)

Merit options
Getting off at designated 0.306 1.754***
places, and automatic parking
Mitigating transportation 1.823*** 1.174***
problems involving elders
Burden reduction from long 0.417* 1.379***
trips
Overall reduction of the 1.668*** 1.279***
burden from driving
Reducing traffic accidents due 1.557*** 1.182***
to human errors
Extended accessibility 1.472*** 1.389***
Automatic braking in the 1.016*** 1.101***
cases of danger
Demerit options
Increase in initial and −0.915*** 0.865***
maintenance costs
Recording of all routes and −0.976*** 0.948***
destinations
The possibility of traffic 0.858*** 0.704***
accidents by malfunctions
The obscurity of responsibility 0.747*** 0.772***
in the accidents
Merit factors
Burden reduction from driving 2.285*** 1.382***
Automatic arrival and parking 0.927*** 1.586***
at designated destinations
Will not need the driver’s license 0.198* 1.051***
Demerit factors
Uncertainties and risk of AD 0.251* 0.663***
and FAVS
Concerns regarding −0.426*** 0.953***
information security
Restrictions of AD −0.917*** 1.027***
Preference on traveling and
mobility
Likes traveling −0.136 1.016 −0.150 1.025**
Prefer to make a plan when 0.316 0.935*** 0.309 0.923***
going out
Prefer to go out alone 0.367 0.869*** 0.397* 0.863***
Don’t mind paying extra cost 3.791*** 1.197*** 3.817*** 1.223***
for safety
(Continued)
Prefer to follow plans when 0.208 0.983 0.234 0.994

going out
Want to shorten traveling time −0.748*** 1.000 −0.699*** 0.997
Don’t mind paying extra cost 2.541*** 1.111*** 2.598*** 1.097***
to avoid congestions
Prefer to go out without plans 0.287 0.917*** 0.325 0.908***
Automatic driving
On general roads −1.042*** 1.610*** −1.193*** 1.629***
On a freeway −0.376 1.426*** −0.432* 1.375***
Inner-city streets 1.169*** 0.951*** 1.223*** 0.938***
Individual attributes
Age −0.952*** 1.023*** −0.951*** 1.020***
Household income 0.657*** 1.012*** 0.657*** 1.012***
Childr(en) in household 2.151*** 1.056*** 2.127*** 1.054***
Car owner 0.644 0.458*** 0.692 0.449***
Diver’s license −3.340*** 0.702*** −3.478*** 0.674***
Average driving time 0.0168*** 1.001*** 0.0169*** 1.001***
Population density of a −4.64e−05 1.000 −4.82e−05 1.000
municipality
Accidents per capita of a −372.4 2.777e+29** −400.6 1.519e+30**
municipality
Difference in municipality 1.345*** 1.106*** 1.325*** 1.114***
evaluation
Life satisfaction 0.792*** 0.988* 0.817*** 0.988*
Observations 188,089 136,388 188,089 136,388
R2 0.044 . 0.044 .
pseudo- R2 . 0.0989 . 0.0984
Note: *** p < 0.01.
**p < 0.05.
*p < 0.1.
more valuable. According to the calculation with coefficients in Table 17.4, the
monetary benefit per capita of fully removing technical malfunctions is approxi-
mately JPY 8500 (USD 74), and the benefit of clarifying regulations and under-
standing responsibility for traffic accidents is approximately JPY 7500 (USD 65).
Additionally, per capita loss from uncertainty regarding information leakage to
third parties is approximately JPY 4000 (USD 35).
The results of other mobility-related variables indicate that consumers are will-
ing to pay more for safety and the avoidance of congestion by purchasing FADS.
Additionally, consumers who regularly drive long-distances have relatively higher
PI and WTP. On the other hand, people whom ‘prefer to go out alone’ or ‘prefer
to go out without specific purposes’ had relatively lower PI and WTP. Moreover,
people who ‘make a plan beforehand when going out’, ‘want to follow plans when
going out’, and ‘want to shorten the time spent on the road’ also had relatively
lower PI and WTP. The results also indicate that a 1-minute reduction in driving
time would increase respondents’ WTP by an average of JPY 170 (USD 1.5),
which would add up to JPY 10,000 (USD 87) per hour.
In addition, we find that expectations of AD on general roads and freeways both
have positive effects on PI; however, they do not affect WTP. On the other hand,
people in urban areas have lower PI but relatively higher WTP. Considering con-
sumer characteristics, ‘having children’ and ‘not having a driver’s license’ posi-
tively affect both PI and WTP. Meanwhile, ‘not owning cars’ positively affects PI;
however, it hardly affects WTP. Respondents who expect that automated vehicles
will enter the market earlier are inclined to purchase FAV but to pay lower fees
for them. As for household income, an increase of one million yen leads to an
increase in WTP of JPY 6500 (USD 57). Given that WTP for FAD is, on average,
approximately JPY 200,000 (USD 1740), WTP is equal to JPY 240,000 (USD
2087) when household income is JPY 10 million (USD 8700). Household income
affects WTP positively, but the impact per unit is so small that it is almost irrel-
evant. Overall life satisfaction has a positive effect on WTP but a negative effect
on PI.
Regarding objective data variables, population density does not affect WTP,
but accidents-per-capita by municipality has a positive effect on PI. In addition,
the variable of the difference between personal assessment of municipal admin-
istration services and municipality average also had a positive impact on PI. This
result implies that people who are more satisfied with their municipality’s services
have a higher PI of FADS. Given that the introduction of FAV would require
the coordination of regulators and local policymakers, consumers who trust their
local policymakers may expect a smoother transition to FAV-based road and traf-
fic regulations.
5.2 Municipality-level analysis

The municipality analysis indicates that the rate of car ownership has a positive
effect on PI. However, it does not have a significant effect on PI. Additionally, the
convenience of dropping off and parking cars has the same effect. Also, munici-
palities with a higher share of people who regard expanded accessibility as a
benefit of FADS seem to have a higher average PI. In addition, concern about
the malfunctioning of FADS negatively affects both the average PI and average
WTP. This result indicates that public relations activities concerning the possible
safety issues of FADS need to be managed to improve the acceptability of FAV
(see Table 17.5).
Consumers seem to expect major accident reductions from the introduction
of FAV, and we found a positive effect of this expectation on their PI. However,
the positive impact of the expectation of accident reduction is smaller than the
negative impact of people’s concerns regarding the malfunctioning of FADS.
Other city-specific variables, such as average taxed income, the ratio of elders,
number of traffic accidents, and average subjective assessment of municipal
Table 17.5 Determinants of consumer demand of FAV and FADS (municipality-level analysis)
Merit options
Getting off at designated places, and −4.727 1.359***
automatic parking
Automated transportation of goods −0.480 1.116
Burden reduction from long trips −4.777 1.224*
Will not need the driver’s license −4.898 0.952
Accident would not be the driver’s 4.103 1.334*
responsibility
Reducing traffic accidents due to human 11.70** 0.868
errors
Extended accessibility 16.79** 1.222
Demerit options
Needs to learn new operative system 5.887 1.128
Increase initial and maintenance costs 5.412 0.941
Making a record of all tracks 9.307 1.114
The third party can miuse the car 1.801 0.899
The possibility of traffic accidents by −14.33*** 0.777***
malfunctions
Increased traffic quantities −1.994 0.965
Reaching wrong destinations due to 2.146 0.777**
system malfunctions
Merit factors
Burden reduction from driving 7.284** 1.012
Automatic arrival and parking at −2.832 1.271***
designated destinations
Will not need the driver’s license 3.748 1.139***
Demerit factors
Uncertainties and risk of AD and FAVS −3.557 0.754***
Concerns regarding information security −0.233 0.981
Restrictions of AD 0.162 1.039
Regional attributes
Average age 0.167 0.0944 1.003 1.002
The share of car owners −0.330 −0.0608 0.825*** 0.839***
Car accidents per capita 75.26 104.9 5.858 4.779
Taxable income per capita 0.438** 0.461** 1.001 1.002
Average municipal evaluation −0.745 0.166 1.022 1.004
Average life satisfaction measures −2.062 −2.330 1.003 1.003
Hokkaido −1.210* −1.298* 0.952*** 0.953***
Tohoku −0.792 −0.602 0.986 0.987
Chubu −0.380 −0.332 0.987 0.990
Kinki −0.510 −0.293 0.994 0.998
Chugoku −0.227 −0.0496 0.971** 0.966***
Shikoku 1.239 1.332 0.986 0.988
Kyushu −1.270** −1.162* 0.985 0.988
Observations 405 405 405 405
R-squared 0.150 0.118 0.456 0.459
***p < 0.01.
**p < 0.05.
*p < 0.1.
administration services, and average subjective life satisfaction did not have sta-
tistically significant impacts on the average WTP for FAV. Thus, the acceptability
of FAV would be affected by individual characteristics rather than by city-specific
characteristics.
6 Conclusion
As AD technology advances rapidly and receives more media coverage, public
consumers are increasingly exposed to this new technology and are developing
expectations. Policy makers are assisting with the investment in both hard and
soft technologies related to AD and are preparing for the full introduction of FAV
in the consumer market and on public roads by debating optimal regulatory poli-
cies. Nonetheless, as with any technology, understanding consumer demand and
perception is essential to predicting the near-future market and to grasping barri-
ers to the introduction of FADS as common consumer goods.
Our survey data indicates that Japanese consumers expect that FAV will enter the
market in approximately 13 years, and 47% of consumers are inclined to purchase
FAV. The majority of consumers expect to use AD technology on freeways and also
on general roads. Also according to the survey data, WTP for FADS is approximately
JPY 0.19 million (USD 1650). We see a strong correlation between PI and WTP. The
respondents who answered that they ‘will purchase’ FADS when available have an
average WTP of JPY 285,000 (USD 2480), and respondents who would ‘consider
purchasing’ would pay JPY 225,000 (USD 1960) on average. On the other hand,
the WTP of respondents who ‘will NOT purchase’ FADS was JPY 134,000 (USD
1165). Thus, consumers with the strongest PI have a WTP double that of consumers
with no interest in purchasing FADS. As for these WTPs, given the price of partial
ADS and FADS, BCG (2015) estimated prices for partial AD vehicles and FADS at
approximately USD 2000–5700 and approximately USD 9800, respectively. Hence,
there seems to be a significant gap between the WTP and predicted prices.
The major merits of FAD are as follows: eliminating concerns of elders, auto-
matic destination arrival and parking, reducing the burden of driving, and reduc-
tion in traffic accidents. These merits can be classified into the following three
categories: (1) reducing the burden of driving, (2) getting in and out of the car at
designated locations and automatic parking, and (3) not needing a driver’s license.
On the other hand, accidents due to malfunctions, unclear responsibility for acci-
dents, and initial costs and maintenance costs are the majority demerits of FAD.
These demerits can be classified into the following three categories: (1) anxieties
about the unfamiliarity of automated driving, (2) leakage of information, and (3)
restrictions on cars’ availability. We find that elderly people are highly interested
in FAD, as they selected more merit and demerit options than other groups.
The results of regressions indicated that merits mainly positively affect accept-
ability and demerits mainly negatively affect acceptability. Considering that peo-
ple who do not have cars or drivers’ licenses are more willing to accept automated
driving than people who do have cars and drivers’ licenses, we anticipate the
expansion of usage for people without cars.
Our municipality-level analysis found that municipality-level variables such as
averaged tax income, the share of elders, and average ratings of municipal ser-
vices as well as average life satisfaction ratings do not affect consumer demand
for FADS. The variables that show significant effects on both individual analysis
and regional analysis are related to traffic accidents. Furthermore, using individ-
ual data, we performed a multiple regression analysis of the expansion of acces-
sible areas, which is one of the major benefits of automated driving. As a result,
it is clear that inner-city areas, particularly in Tokyo, show higher acceptability of
AD than other areas.
There are several issues that can be addressed in future studies to further
our understanding of consumer demand for AD: (1) closing the gap between
the price that firms expect and the price that consumers are willing to pay; (2)
addressing the issue of possible technical malfunction, and (3) information
security. The first problem may be resolved through subsidies either to the
industry or to consumers, and also by an increase in investment to lower pro-
duction costs. As for issues (2) and (3), investment in the field of technologi-
cal development would be useful, but sharing information about technological
advancements and the merits of AD with consumers may also be an effective
strategy.
Notes
1 This research is supported by the grants from RIETI and Ministry of Education, Culture,
Sports, Science and Technology in Japan (MEXT) under a Specially Promoted Research
grant. Any opinions, findings, and conclusions expressed in this material are those of the
authors and do not necessarily reflect the views of the MEXT. We also thank Naoto Tada
for providing extensive assistance.
2 Siri is a computer program developed by Apple that serves as an intelligent assistant.
Pepper is a humanoid robot with a dialog system, which was developed by Softbank, a
Japanese mobile phone company.
3 ‘Pods’ are driverless vehicles that move on tracks and have already been tested at Heath-
row Airport. They are used as a prototype of an automated vehicle that would eventually
be used on normal roads without tracks.
4 ‘Kentou kadai no seiri’; www.mlit.go.jp/road/ir/ir-council/autopilot/pdf/05/2.pdf.
5 Out of 246,642 samples, there were 47,406 and 39,883 respondents who did not choose
merits and demerits, respectively. Also, there were 33,159 respondents who chose nei-
ther a merit nor a demerit.
Bibliography
Bansal, P. (Graduate Research Assistant), K. Kockelman (E. P. Schoch Professor in Engi-
neering) (2017) Forecasting Americans’ long-term adoption of connected and autono-
mous vehicle technologies. Transportation Research Part A. vol. 95, 49–63.
Bekiaris, E., S. Petica, and K. Brookhuis. Driver Needs and Public Acceptance Regarding
Telematic In-Vehicle Emergency Control Aids. Proc., Mobility for Everyone: 4th World
Congress on Intelligent Transport Systems, Berlin, 1997.
Boston Consulting Group (2015) The most innovative companies 2015. https://media-
publications.bcg.com/MIC/BCG-Most-Innovative-Companies-2015-Nov-2015.pdf
Cabinet Office, Government of Japan (2016) SIP (senryakuteki innovation souzou pro-
gram) jidou soukou system kenkyu kaihatsu keikaku. www8.cao.go.jp/cstp/gaiyo/sip/
keikaku/6_jidousoukou.pdf.
Casley, S. V., A. S. Jardim, and A. M. Quartulli (2013) A study of public acceptance of
autonomous cars (Bachelor of Science). Worcester Polytechnic Institute, Worcester, MA.
www.wpi.edu/Pubs/E-project/Available/E-project-043013-155601/unrestricted/A_
Study_of_Public_Acceptance_of_Autonomous_Cars.pdf.
Correia, G., Milakis, D., van Arem, B., and Hoogendoorn, R. (2016) Vehicle automation
and transport system performance. In Handbook on Transport and Urban Planning in
the Developed World (pp. 498–516). Edward Elgar Publishing.
European transport Safety Council (2016) Prioritising the safety potential of automated
driving in Europe. https://etsc.eu/wpcontent/uploads/2016_automated_driving_briefing_
final.pdf
Goodall, N. J. (2014) Ethical decision making during automated vehicle crashes. Transpor-
tation Research Record: Journal of the Transportation Research Board. no. 2424, 58–65.
Green, P. (1996) Customer needs, new technology, human factors, and driver science
research for future automobiles (in Japanese). Journal of the Japan Society of Mechani-
cal Engineers. vol. 99. no. 926, 15–18.
Hohenberger, C., S. Matthias, and W. Isabell. (2016) How and why do men and women
differ in their willingness to use automated cars? The influence of emotions across dif-
ferent age groups. Transportation Research Part A: Policy and Practice, 374–385.
Howard, D., and D. Dai. (2014) Public perceptions of self-driving cars: The case of
Berkeley, California. Transportation Research Board 93rd Annual Meeting. vol. 14.
no. 4502.
Fagnant, D., and K. Kockelman (2015) Preparing a nation for autonomous vehicles: Oppor-
tunities, barriers and policy recommendations. Transportation Research Part A. vol. 77,
167–181.
Kyriakidis, M., et al. (2017) A human factors perspective on automated driving. Theoretical
Issues in Ergonomics Science. 1–27. www.researchgate.net/publication/313367679_A_
Human_Factors_Perspective_on_Automated_Driving.
Kyriakidis, M., R. Happee, and J.C.F. de Winter (2015) Public opinion on automated driv-
ing: Results of an international questionnaire among 5000 respondents. Transportation
Research Part F: Traffic Psychology and Behaviour. vol. 32, 127–140.
Leech, J., G. Whelan, and M. Bhaiji (2015) Connected and autonomous vehicles: The UK
Economic Opportunity. KPMG. www.smmt.co.uk/wp-content/uploads/sites/2/
CRT036586F-Connectedand-Autonomous-Vehicles-%E2%80%93-The-UK-Economic-
Opportu...1.pdf.
Ministry of Economy, Trade and Industry (2014) heisei 25 nendo green jidousha gijutsu
chousa kenkyu jigyou (kongo no gijutsu kakushin kanousei wo humaeta aratana jidousha
shakai ni kansuru itaku chousa jigyou) houkokusho. www.meti.go.jp/meti_lib/
report/2014fy/E004399.pdf.
Ministry of Land, Infrastructure, Transport and Tourism (2013) auto-pilot system no jitsu-
gen ni mukete chukan torimamome. www.mlit.go.jp/road/ir/ir-council/autopilot/pdf/
torimatome/honbun.pdf.
Ministry of Land, Infrastructure, Transport and Tourism (2014a) kounaigai ni okeru saikin
no jidou unten no jitsugen ni muketa torikumi gaiyou. www.mlit.go.jp/road/ir/ir-council/
autopilot/pdf/04/8.pdf.
Ministry of Land, Infrastructure, Transport and Tourism (2014b) kentou kadai no seiri.
www.mlit.go.jp/road/ir/ir-council/autopilot/pdf/06/2.pdf.
Ministry of Land, Infrastructure, Transport and Tourism (2015) Promotion of effective
congestion control. https://www.mlit.go.jp/road/ir/ir-perform/h18/07.pdf
Missel, J. (2014) Ipsos MORI Loyalty automotive survey. www.ipsos-mori.com/research
publications/researcharchive/3427/Only-18-per-cent-of-Britons-believe-driverless-cars-
to-be-an-important-development-for-the-car-industry-to-focus-on.aspx?utm_
campaign=cmp_325684&utm_source=getanewsletter.
Lavasani, M., Jin, X., and Du, Y. (2016) Market penetration model for autonomous vehicles
on the basis of earlier technology adoption experience. Transportation Research Record:
Journal of the Transportation Research Board, vol. 2597, 67–74.
Mosquet, X., T. Dauner, N. Lang, M. Rübmann, A. Mei-Pochtler, R. Agrawal, and F.
Schmieg (2015) Revolution in the driver’s seat: The road to autonomous vehicles. The
Boston Consulting Group.
Obst, M., A. Marjovi, M. Vasic, I. Navarro, A. Martinoli, A. Amditis, P. Pantazopoulos, I.
Llatser, A.D.L. Fortelle, and X. Qian (2016) Challenge for automated cooperative driv-
ing: The AutoNet2030 approach. Automated Driving. 561–570.
Okunetto sougou kenkyujyo (2014) jidou unten gijyutsu heno kitai to ni-zu. www.aucnet.
co.jp/wp-content/uploads/78579d0abb2745e06814236458c494bd.pdf.
Payre, W., J. Cestac, and P. Delhomme (2014) Intention to use a fully automated car: Atti-
tudes and a priori acceptability. Transportation Research Part F: Traffic Psychology and
Behaviour. vol. 27, 252–263.
Petit, J., and S. E. Shladover (2015) Potential cyberattacks on automated vehicles. IEEE
Transactions on Intelligent Transportation Systems. vol. 16, 546–556.
Power, J. D. (2012) 2012 U.S. automotive emerging technologies study results. http://autos.
jdpower.com/content/press-release/gGOwCnW/2012-u-sautomotive-emerging-technol-
ogies-study.htm.
Prateek, B., K. Kockelman, and A. Singh (2016) Assessing public opinions of and interest
in new vehicle technologies: An Austin perspective. Transportation Research Part C.
vol. 67, 1–14.
Schoettle, B., and M. Sivak (2014a) A survey of public opinion about autonomous and self-
driving vehicles in the U.S., the U.K., and Australia. Michigan. http://deepblue.lib.
umich.edu/bitstream/handle/2027.42/108384/103024.pdf.
Schoettle, B., and M. Sivak (2014b) Public opinion about self-driving vehicles in China,
India, Japan, the U.S., the U.K., and Australia, Michigan. http://deepblue.lib.umich.edu/
bitstream/handle/2027.42/109433/103139.pdf?sequence=1.
Sivak, M., and Schoettle, B. (2015) Potential Impact of Self-Driving Vehicles on House-
hold Vehicle Demand and Usage. Sustainable Worldwide Transportation Program (www.
umich. edu/~ umtriswt), University of Michigan.
Sels, P., T. Dewilde, D. Cattrysse, and P. Vansteenwegen (2016) Reducing the passenger
travel time in practice by the automated construction of a robust railway timetable.
Transportation Research Part B: Methodological, 124–156.
Silberg, G., M. Manassa, K. Everhart, D. Subramanian, M. Corley, H. Fraser, and V. Sinha
(2013) Self-driving cars: Are we ready? KPMG. www.kpmg.com/US/en/IssuesAndIn-
sights/ArticlesPublications/Documents/self-driving-cars-arewe-ready.pdf.
Sommer, K. (2013) Continental mobility study 2011. Continental AG http://www. continental-
corporation.com/www/download/pressportal_com_en/themes/initiatives/channel_mobil-
ity_study_en/ov_mobility_study2013_en/download_channel/pres_mobility_study_en. pdf.
Suda, Y., and K. Aoki (2014) jidou unten gijyutsu no kaihatsu doukou to gijyutsu kadai.
Jyouhoukanri. vol. 57. no. 11, 809–817.
Underwood, S. E. (2014) Automated vehicles forecast vehicle symposium opinion survey.
Presentation at the 2014 Automated Vehicles Symposium, San Francisco, CA.
Welsch, H., and Kühling, J. (2009). Determinants of pro-environmental consumption: The
role of reference groups and routine behavior. Ecological Economics, vol. 69, no. 1,
166–176.
Wadud, Z. (2017) Fully automated vehicles: A cost of ownership analysis to inform early
adoption. Transportation Research Part A: Policy and Practice, 163–176.
Yap, M. D., C. Goncalo, and V. A. Bart (2016) Preference of travelers for using automated
vehicles as last mile public transport of multimodal train trips. Transportation Research
Part A: Policy and Practice, 1–16.
Index
Note: Boldface page references indicate tables. Italic references indicate maps and figures.
Acar, S. 135 Association of Southeast Asian Nations

Act on Special Measures concerning the (ASEAN) 147
Handling of Radioactive Pollution 313 Atilgan, B. 43
ADF-Fisher method 94, 96, 96 Atkinson, G. 137
Adjusted Net Savings (ANS) 78, 80 Atkinson, G. D. 288
affect balance 227–228, 228, 229, 239 Aucnet survey 419, 425
Africa 261, 275, 276, 277, 277; see also Australia 196–197
climate change on African agriculture; automated driving 418–419; see also fully
South Africa automated vehicle (FAV)
Africa Competitiveness Report (2015) 300 Automated Vehicle Symposium (2014) 423
Agenda for Sustainable Development Azomahou, T. 88–89
Goals (SDGs) 135
agent-based modelling (ABM) 46, 48, Balducci, A. 202
49, 50 Barro, R. J. 25, 303, 379, 402
aggregated capital 138 Barros, C. P. 290
agriculture 216, 221, 245–246, 319; Battese, G. E. 139
see also climate change on African Bayer, P. 195
agriculture BCG survey 419–420, 425
air pollution 52–53, 53, 407; see also Bekiaris, E. S. 419
specific emission Belgium energy project case study 57, 58,
Akaike Information Criterion (AIC) 59–64
value 95 Benowo landfill 356, 359–361, 360, 361,
Al-mulali, U. 94 363–367, 363, 364, 367, 370
Ambrey, C. 196–197, 226, 245 Berg, B. van den 228, 240
Annual Survey of Industries (ASI) 33 Bernanke, B. 77
ARCH-M method 79, 81, 82, 85 Better Life Index (BLI) indicators 233,
ARIMA model 56 246–247, 253–256
ARMA model 56 Bhattacharya, J. 89
Arronsson, T. C. 77, 79, 85 bioenergy 413–415, 414; see also
Arrow, K. J. 20, 24–25, 55, 136–138, renewable energy on sustainability
141–142, 286–287, 289, 294, 327–328, birth control see population planning
376, 379, 402 policy in China; Russia; Vietnam
artificial intelligence (AI) technology 418; Blanchflower, D. 202
see also fully automated vehicle (FAV) Bölük, G. 89
Aschauer, D. A. 25 Bongaarts, John 375
Asian Development Bank report (2005) 26 Bosworth, B. P. R. 137, 141
Asian financial crisis (1997–1998) 156 bottom-up (BU) modelling 46
Index 441
Boyer, J. 142, 294 systems 392; productivity measurement
BP Statistical Review of World Energy 142 and sustainable development case study
Branch, T. A. 87–88 300, 303–306, 304; sex ratio 392–393;
Brandt, N. 136, 287, 289 total factor productivity 157–158,
Brazil 157, 157, 159, 160 157, 159, 161; two-child policy
BRICS countries’ total factor productivity 374–376, 382, 389–390, 392–393,
156–160, 157, 159, 161, 161; see also 402; urbanization 399, 400; see also
specific country population planning policy in China
Brown, S. 167, 170 China Division of Floating Population
Bruntland Report 287 Dynamic Population Development
Report (2016) 399
Canada 298, 300, 304–306, 305 Christodoulou, A. 19
Cantrill’s ladder 229, 237–238, 240, 258 city evaluation indexes 5, 6–7, 8, 19–20
Cao, H.-S. 89 Clark, A. E. 167, 184, 228
capacity utilization (CU) 138 Clean Development Mechanism (CDM)
capital: defining 327; inefficiency scores 300, project 354, 356–357
302; renewable energy on sustainability Clémençon, R. 147
and 413–414, 414; well-being and Clemens, M. 137, 288
288–289; see also specific type climate change on African agriculture:
capital assets/stocks 5, 8, 45, 50–56, 51, 73; actions taken against 261; background
see also inclusive urban capital stock information 261–262; conclusions
capital recovery factor (CRF) method 48, 52 276–277, 278, 279; countries’
carbon damage inefficiency scores 297, income 272, 273; countries list 285;
300, 301, 303 data 262–265, 265, 266; discussion
carbon dioxide emissions: energy 266–274, 275, 276; empirical results
infrastructure project evaluation and 266–274; extreme events 262, 264;
52–53, 64, 72; landfill gas emission in floods 269; funds in Africa 277, 277,
Indonesia 354, 356, 359, 362, 365–367, 278, 279; Inclusive Wealth Index and
365, 366, 370; productivity measurement 274–275, 276; income 271–272, 272,
and sustainable development 287, 273, 274–275, 276, 277; international
289–290, 294–295, 303 commodity prices 264, 279n1;
CASBEE (Comprehensive Assessment methodology 262–265; overview 3, 262,
System for Built Environment 276–277, 279; period of 1961–1986
Efficiency) 423 269–270, 270; period of 1961–2014
causality 34 266, 267, 268, 268, 269; period of
Centre of Environmental Data Analysis 264 1987–2014 269–270, 270; political
cesium contamination 12, 310 implications 276–277, 278, 279;
Charnes, A. 9 production index 263–264, 274–275,
Cheechi, D. 202 275; production model 262–263, 266,
Chen, P.-C. 290 268; productivity over sustainability
childbearing desire survey 384–386, 385 276; rainfall 264, 266, 267, 268, 269;
China: age structure 382–383, 382; robustness test 281–283, 284; socio-
domestic population 399, 401; education economic factors 264–265, 265–266;
level 158, 390; family support 392; summary statistics 265, 265–266;
fertility rate 375, 394, 394, 395, 396; temperature 264, 266, 267, 274
fish catches, forecasting 100, 102; Coelli, T. 139
green spaces and well-being 196–197; Collins, R. D. 44–45
human capital 158, 387, 387, 388, Collins, S. M. 137, 141
389–391, 389, 389; human capital community attachment 170–171, 176,
change 375–377, 379–380, 387, 387, 184–185, 194
388, 389–391, 389, 389; human capital composition effect 89
transfer 399, 401, 402; immigration 392; condition state trigger 49, 50
life expectancy 158, 391, 391; one-child construction state of project 49, 50
policy 374–375, 390, 392–393; pension Continental survey 419
442 Index
Contingent Valuation Method (CVM) 195 electricity consumption and generation 54,
contribution effect 287, 292–295, 295 406, 407
Cook, W. D. 16 Emergency Events Database (EM-DAT) 264
Coordination of Information on the employment effect 53–54
Environment (CORINE) Land Cover endowed capital 287
database 197 energy infrastructure project evaluation:
Costanza, R. 20n1 agent-based modelling 46, 48,
Costello, C. 90 49, 50; background information
43–44; Belgium case study 57, 58,
Dasgupta, P. 5, 8 59–64; bottom-up modelling 46;
Dash, R. K. 25, 28, 33 capital assets 45, 50–56, 51, 73;
Data Envelopment Analysis (DEA) capital recovery factor method 48, 52;
method: inclusive urban capital stock carbon dioxide emissions 52–53, 64,
and 9–16, 13, 15; productivity and 72; characteristics of project 48, 49;
natural capital and 136, 139–140; urban conclusions 73–74; education effect
sustainability 9–16, 13, 15 53–54; Egypt case study 64, 65, 66–73,
da Vinci, L. 1 72–73; employment effect 53–54;
Deaton, A. 171 energy supply index 47–48; ExternE
De Bruyn, S. M. 89 methodology 53, 55–56; forecasting
Decision Making Unit (DMU) 291–293 and 56; fuel cycle and emissions
decision utility 227 56–57, 73–74; health effect 53–55;
decomposition model of inclusive urban human capital 53–55, 64, 73; hybrid
capital stock 10–13, 13, 16–19, 17, 18 simulation model 48–56; Inclusive
Denison, E. 138 Wealth Index and 43; inertia and 46, 48,
Desai, P. 158 50; investment policies 73; methodology
Despotis, D. K. 9 44–46, 45; model parameters 56–57,
Digital Map 5000 198 57; mortalities effect 55; natural capital
directional distance model (WRDDM) 55–56; overview 2, 44, 73–74; pollutant
287, 290–292, 296 concentrations 52–53, 53; quantification
distance functions 139–140 52–56; state of projects 48, 49; state
Domingos, T. 135 triggers 49, 50; system dynamics
dynamic average utilitarianism (DAU) modelling 46, 50–52, 51; theoretical
criterion 328–329, 331–334, 333–334, framework 47–56, 47; time-series
349, 349, 350–351 analysis 56, 57; top-down modelling 46;
Value of a Statistical Life and 55; wealth
Easterlin Paradox 167 trajectory 44–45, 45, 72–73, 73
Easterlin, R. A. 167 energy supply index (ESI) 47–48
economic development and marine Engle, R. 7
fisheries 88–90 environmental impact of landfill gas
economic growth 1, 77; see also India’s emission in Indonesia 363–367, 364,
infrastructure and economic growth; 364, 365, 366
marine fisheries with economic growth Environmental Protection Agency (EPA) 362
economic impact of landfill gas emission environmental value of green spaces in
in Indonesia 368–369, 369, 370 Japan: agriculture’s functions 216,
economic interests, innocent pursuit of 1 221; background information 195–196;
economic modelling 93–96 conclusions 215–216; content of
Economics of Ecosystems and survey 218–219; control variables
Biodiversity (TEEB) 319–320 201–202, 203–204; data 198–204;
ecosystem evaluation approach 21n2 defining greens spaces 216, 217;
education effect 53–54 descriptive statistics 202, 203–204;
Eerd, V. 359 Digital Map 5000 classifications of land
Egypt energy project case study 64, 65, use and 198; diversity of respondents’
66–73, 72–73 preferences 200–201; estimation results
Ehrlich, P. R. 288 208, 209–210, 211, 212–214; forests’
Index 443
functions 216, 220; green spaces data Fujii, H. 292
198–200, 199, 200; knowledge of green Fukushima Dai-ichi Nuclear Power Plant
spaces 201; Life Satisfaction Approach accident: Act on Special Measures
model 195–196, 205, 206–207, 208, concerning the Handling of Radioactive
215–216; literature review 196–198; Pollution and 313; background
locations of green spaces 198–200, information 310; cesium contamination
199, 200; marginal willingness to pay 310, 312; designated waste 314–315;
197, 205, 208, 211, 211, 215–216, 215; iodine contamination 310; Ministry of
overview 3, 196, 211–216; parameters the Environment and 315
196; quality of green spaces 200–201; Fukushima Dai-ichi Nuclear Power Plant
respondent data 198; subjective well- accident; see also social cost-benefit
being 205; willingness to pay 195 analysis of Fukushima Dai-ichi Nuclear
Ermisch, J. 229 Power Plant decontamination
estimation model: fully automated vehicle fully automated driving systems (FADs) 420
429–430; greenery and well-being 231, fully automated vehicle (FAV): Aucnet
233, 234–236; renewable energy on survey 419, 425; availability time frame
sustainability 412–413 423–425, 424; background information
Estrada, A. 138 418–420; BCG survey 419–420, 425;
European Quality of Life Survey (EQLS) conclusions 435–436; consumer demand
197 419–420; Continental survey 419;
Ewers, R. M. 88 data 420–422; demerits 418–419, 421,
experienced utility 227 422, 425–426; descriptive analysis
ExternE methodology 53, 55–56 423–429; discussion 430–435; dummy
extreme events 262, 264 variables 421; estimation model
429–430; explanatory variables 421;
Faour, A. A. 356 future research 436; HAVE it project
farmland greenery 245–246 419; individual-level analysis 430,
Felix, C. 184 431–432, 432–433, 436; in Japan 418;
Feng, Q. 304 markets for, largest potential 419; merits
Ferreira, S. 227, 327 418, 421, 421, 422, 423, 425–426,
Ferrer-i-Carbonell, A. 237, 240 436; mobility-related variables 423;
Fertility Intention of Rural and Urban municipality-level analysis 433, 434,
Residents in China 384–385 435, 436; overview 4, 420, 435–436;
fertility rate 375, 378, 394, 394, 395, 396, perceptions of merits/demerits 425–426;
398 purchase intention 420–421, 426–428,
Filmer, D. 298 426, 427, 428; results 430–435;
fiscal policy and public debt effect 77, 85 SARTRE project 419; surveys 419–421,
fisheries see marine fisheries with 423, 425; in United Kingdom 418;
economic growth variables 421, 423; willingness to pay
Fisher-type Augmented Dickey-Fuller 420–421, 427, 428, 429, 435
(ADF-Fisher) method 94, 96, 96 Futagami, K. 25
Fisher-type Phillips-Perron (PP-Fisher)
method 94, 96, 96 G7 countries’ total factor productivity
Fleming, C. 196–197, 226, 245–246 152, 153–154, 155–157, 155, 156, 161;
Food and Agriculture Organization (FAO) see also specific country
263, 277 Garcia-Mil’a, T. 25
forecasting 56, 100, 100 Garsous, G. 26
forests 216, 220, 319–320 gas emissions see landfill gas emission in
fossil fuels 408 Indonesia; specific type
France 155, 155 Gaussian plume model 53
Frey-Levine, L. A. 288 General Health Questionnaire (GHQ-12)
Frijters, P. 237, 240 225, 228
Froese, R. 88, 91 genuine savings (GS): gross domestic
fuel cycle and emissions 56–57, 73–74 product and 81; productivity
444 Index
measurement and natural capital and greenhouse gas (GHG) emissions 354,
135, 137–138; public debt effect and 355–357, 407
78–81, 79, 80, 82, 83, 83, 83–84, 85 greening policies 247
geographic information systems (GIS) data green spaces: in Australia 196–197; in
173, 195, 197, 226 China 196–197; data 198–200, 199, 200;
Geospatial Information Authority 198 defining 216, 217; knowledge of 201;
Gephart, J. A. 87 locations 198–200, 199, 200; quality of
Germany 155, 155 200–201; in United Kingdom 196–197;
Gilchrist, K. 245 in United States 197–198; see also
Giraud, P.-N. 135 environmental value of green spaces in
global warming 297, 300, 304, 354–356; Japan; greenery and well-being
see also climate change on African Griliches, Z. 138
agriculture gross domestic product (GDP):
Gosling, S. D. 174 depreciation of capital assets and 5;
Goulder, L. H. 288 economic activity measurement and
Granger causality test 34 286; genuine savings and 81; of India
Great Recession (2007–2009) 156 27–28, 33–34, 40; inefficiency scores
Greece 77, 252, 398 298, 299, 300; infrastructure and
greenery and well-being: affect balance economic growth in India 27–28, 33–34,
227–228, 228, 229, 239; background 40; measurements 135; renewable
information 225; Better Life Index energy on sustainability and 411–412; in
indicators 233, 246–247, 253–256; Russia 158
Cantrill’s ladder and 229, 237–238, 240, Gultekin-Karakas, D. 135
258; changes in green areas in Tokyo
251; classification of greenery 229, Haight, R. G. 197–198
230, 231, 232; conclusions 244–247; Halkos, G. 93
correlations among indices 229, 229; Hamilton, K. 137, 288
data 227–236; data on greenery 229, Harper, M. J. 138
230, 231, 231, 232; decision utility Hausmann, R. 300
and 227; description statistics 233, Hausman test 29–31
234–236; distance band correlations HAVE it project 419
258; estimation model 231, 233, health capital 390–391
234–236; estimation results 237–249; health effect 53–55
experienced utility and 227; farmland hedonic inertia 245
245–246; General Health Questionnaire hedonic pricing method 195
and 225, 228; greening policies and 247; Hernández de Cos, P. 138
hedonic inertia and 245; industrial area Hesary, F. T. 138
245; life satisfaction 226–227, 228, 229, Hesketh, Therese 375
245–246; literature review 226–227; Hidalgo, C. A. 300
management of greenery 245; marginal Hilborn, R. 100
willingness to pay and 225–226, 233; Hirschman, A. O. 170
mental health 228, 228, 229, 239; Hofman, B. 303
methodology 227–236; monetary Holley, C. 43
valuations 242, 243–244, 244–246; Hsieh, C. M. 184
Normalized Difference Vegetation Index Huang, S. 303
and 229; overview 3, 225, 244–247; human capital (HC): Canada 304; change
parameter estimates 237, 237–238, 239, 375–377, 379–380, 387, 387, 388,
240, 241–242, 258–260; public facility 389–391, 389, 389; China 158, 387,
246; railroad greenery 246; subjective 387, 388, 389–391, 389, 389; energy
well-being 225–229, 228, 229, 239; infrastructure project evaluation 53–55,
survey 227, 228, 252; water-front 64, 73; inclusive urban capital stock
245–246 and 16–19, 18; India 159; national and
Greenhalgh, Susan 375 subnational sustainability and 327, 330;
Index 445
productivity measurement and natural India: economic growth 23; gross domestic
capital 142, 155, 157, 159; productivity product 27–28, 33–34, 40; human
measurement and sustainable capital 159; manufacturing sector 159;
development and 294, 297, 303; public total factor productivity 157, 159–161,
debt effect 78, 80; renewable energy on 159, 161; see also infrastructure and
sustainability 413, 414; South Africa 158 economic growth in India
human capital wealth (HCW) 380 India Infrastructure Report (2006) 23
Human Development Index (HDI) 56 individual transferable quota (ITQ) 90
human impact of landfill gas emission in Indonesia 100, 102; see also landfill gas
Indonesia 367–368, 367, 368, 370–371 emission in Indonesia
hybrid simulation model 48–56 industrial area greenery 245
hydrogen sulfide emissions 354–356, inertia 46, 48, 50, 245
359–360, 362, 366–367, 366, 367, 370 informal sector recycling 357–359, 359,
367, 368
idle state of project 49, 50 information effects 170, 183
Ikeda, S. 329 infrastructure capital 297
Im, Pesarian and Shin (IPS) method 94, infrastructure, defining 24; see also
96, 96 energy infrastructure project evaluation;
inclusive urban capital stock: accumulation infrastructure and economic growth in
9; comparing between cities 8–10; Data India
Envelopment Analysis method and infrastructure and economic growth in
9–16, 13, 15; decomposition model of India: background information 23;
10–13, 13, 16–19, 17, 18; evaluation of conclusions 40–41; correlation matrix
8; factors in city evaluation indexes and 29, 30; data 31–34, 32; energy sector
19–20; human capital and 16–19, 18; 26; findings 36–40; gross domestic
natural capital and 16–19, 18; produced product 27–28, 33–34, 40; growth rate
capital and 16–19, 16 23; Hausman test 29–31; implications
Inclusive Wealth Index (IWI): 36–40; Inclusive Wealth Index and
classification 2, 2; climate change on 39–40; Inclusive Wealth Report
African agriculture and 274–275, 276; (2014) and 23; India Infrastructure
defining 1, 43; ecosystem evaluation Report (2006) and 23; Infrastructure
and 21n2; energy infrastructure project Statistics and 33; literature review
evaluation and 43; genuine savings 24–27, 27; methodology 27–31, 29,
and 78; infrastructure and economic 30; multicollinearity and 28–29, 34;
growth in India 39–40; productivity National Action Plan for Climate
measurement and natural capital Change and 26; obstacles 37–39, 38;
and 135, 138–139; productivity overview 2, 24, 40–41; panel dataset
measurement and sustainable 29, 31–33, 32; post-estimation test
development and 286, 288–289, 293; 29–30, 30; pre-estimation test 29, 30;
public debt effect and 77–79, 79; principle component analysis 29, 30;
renewable energy on sustainability and private spending 36–37; productivity
408–409, 414; sustainability and 327; and 23; public-private partnership 37;
urban sustainability and 8–13, 19–20 public spending 23, 24, 36–37, 36, 39,
Inclusive Wealth Index Adjusted (IWI 39; regression coefficients 34–36, 35;
Adjusted) 141–142, 151–152, 151, 152 regression model 25, 28–29, 29, 31,
Inclusive Wealth Report (2012) 45, 141, 327 34; relationship between 24–25; results
Inclusive Wealth Report (2014) 23, 39, 73, 34–36, 35; reverse causality and 34;
141–142, 289, 327, 376, 380 telecommunications sector 26–27, 27,
income: climate change on African 28, 29; Telecom Regulatory Authority of
agriculture 271–272, 272, 273, 274–275, India and 27, 37; theoretical framework
276, 277; level classes 271, 272, 273, 27–28; transport sector 25–26, 34;
274, 277; well-being and 173, 184–185 Twelfth Five-Year Plan and 26;
income effects 173, 184–185 variables 28–29, 32, 33–34, 37
446 Index
Intergovernmental Panel on Climate per capita wealth criterion 327–328,
Change (IPCC) 261, 362 331–332, 333–334; total utilitarianism
international commodity prices 264, 279n1 criterion 332–334, 333–334; wealth
iodine contamination 310 accumulation 329–331, 330
IRENA statistics 405, 410 J. D. Power 419
IS method 94, 96, 96 Jing, Li 390
Italy 77, 83, 398 Johnston, C. M. T. 408
Johnston, R. J. 195
Jacobson, M. Z. 408 Jorgenson, D. W. 138
Jaitley, Arun 36
Jalil, A. 89 Kahneman, D. 171
Japan: community attachment 185; Kalantarifard, A. 356
forecasting fish catches 100, 101; Karanjekar, R. V. 356
fully automated vehicles in 418; Kaźmierczak, A. 247
genuine savings in 83; Geospatial Kingdon, G. G. 170, 173, 183–184, 186
Information Authority 198; Ministry Kleisner, K. 90–91
of the Environment 315; public debt Klenow, P. J. 137, 376, 379, 402
effect case study 77, 80–81, 80; quota- Knight, J. 170, 173, 183–184, 186
management system for fish catches Kooten, G. C. van 408
92; Science Council 201; sustainability Kopmann, A. 196–197, 233
after tsunami disaster 135; Tenkin-zoku Kumar, S. 287, 289–290, 355–356
in 181; total factor productivity and Kunze, K. 138
155–156, 155; see also environmental Kurz, M. 24–25
value of green spaces in Japan; Japanese Kuznets curve (EKC) hypothesis 88–89, 103
ordinance-designated city case study; Kveselis, V. 43
Japanese wealth accumulation and Kyoto Protocol (2002) 304
population change (2000s)
Japan Center for Economic Research landfill gas emission in Indonesia:
(JCER) 312 background information 354–356;
Japanese ordinance-designated city Benowo landfill 356, 359–361,
case study: conclusions 19–20; data 360, 361, 363–367, 363, 364, 367,
description 13–14, 14; decomposition 370; carbon dioxide 354, 356, 359,
analysis of inclusive capital stock 362, 365–367, 365, 366, 370; Clean
16–19, 17, 18; government budget Development Mechanism project and
expenditures 11–13, 16–17, 20; Local 354, 356–357; conclusions 370–371;
Autonomy Act and 13 economic impact 368–369, 369, 370;
Japanese wealth accumulation and environmental impact 363–367, 364,
population change (2000s): comparing 364, 365, 366; global warming and
prefectures 333–334, 334–336, 335, 355; human impact 367–368, 368, 368,
336, 337–340, 340–341, 341, 342–345, 370–371; hydrogen sulfide 354–356,
346, 347, 347, 348; comparing total 359–360, 362, 366–367, 366, 367, 370;
utilitarianism, total utilitarianism per informal sector recycling 357–359,
capita wealth, and dynamic average 359, 367, 368; Landfill Gas Generation
utilitarianism criteria 332–334, Model software 356, 362–363, 363, 366;
333–334; discount rates 332, 349, 349, methane 354–356, 359, 362, 364–367,
350–351, 351; discussion 349, 350, 364, 370; methodology 362–363;
351; dynamic average utilitarianism municipal solid waste 354, 356–357;
criterion 328–329, 331–332, 332–333, order decay model 354, 356; overview
349, 349, 350–351; methodology 4, 354, 370–371; results 363–370;
331–332; population change 327, survey of health damages 354, 363;
331; present discounted population Tamangapa landfill 354–356, 359, 360,
349, 3349; previous studies 329; 361–367, 362, 363, 364, 365, 366,
results 332–348; total utilitarianism 370–371
Index 447
Landfill Gas Generation Model (LandGEM) PP-Fisher method 94, 96, 96; RAM
software 356, 362–363, 363, 366 legacy database and 90, 96; results 96,
Lantz, V. 304 96, 97, 98–100, 98, 100, 101–102; SBC
Larson, S. 198 value 95–96, 97; scale effect 88–89; Sea
least developed countries (LDCs) 275–276 Around Us Project and 90–91, 95; short-
Lee, J. 303, 379, 402 term 99; stock status plots method 91;
Levinson, A. 240 technical effect 89; total allowable catch
Liao, H. 89 and 90; United Nations Convention on
life cycle assessment (LCA) approach 56–57 the Law of the Sea and 90; U-shaped
life satisfaction (LS): greenery and well- curve and 89, 94
being 226–227, 228, 229, 239, 245–246; Martell, S. 88, 91
as well-being determinant 167, 174, Martins, V. 135
176, 178, 181, 182, 183–186, 194 maximum likelihood estimation (MLE) 175
Life Satisfaction Approach (LSA) model maximum sustainable yield (MSY) 90–91
195–196, 205, 206–207, 208, 215–216 McElroy, M. 396
Li, W. 226 mean years of schooling index (MYSI) 54
Li, Z. 174 Medina, C. 358
longitudinal data 29, 31–33, 32 men absolute percentage errors (MAPE)
Luenberger Productivity Index 289–290, 292 56, 57
Luttmer, E. F. P. 184 mental health and greenery 228, 228, 229,
239
MacKerron, G. 196–197, 226, 249 Merino, G. 93
Madagascar 135 Mert, M. 89
Mahmud, S. F. 89 message state trigger 49, 50
Malaysia 100, 102 methane emissions 354, 356, 359, 362,
Malmquist Productivity Index (MPI) 136, 364–367, 364, 370
139, 161, 289–290 Ministry of the Environment (MOE) 315
Managi, S. 89, 139, 226, 245, 247, 287, Ministry of Health, Labour and Welfare 181
289–290 Ministry of Land, Infrastructure, Transport
man-made capital 78, 80 and Tourism (MLIT) 418, 425
marginal willingness to pay (MWTP): Minnesota, green spaces and well-being in
environmental value of green spaces 197–198
in Japan and 197, 205, 208, 211, 211, Mirman, L. J. 77
215–216, 215; greenery and well-being Mol, A. P. J. 147
and 225–226, 233 Mori, K. 19
marine fisheries with economic growth: Morrison, C. J. 138
ADF-Fisher method 94, 96, 96; AIC mortalities effect 55
value 95–96, 97; autoregressive Mota, R. P. 135
distributed lag representation 95; Mourato, S. 196–197, 226, 240
background information 87–88; biomass multicollinearity 28–29, 34
stock estimates 90–92, 92, 93, 99, multiple linear regression (MLR) 31
107–134; composition effect 89; Mumford, K. J. 137, 329
conclusions 103; economic development municipal solid waste (MSW) 354, 356–357
and 88–90; economic modelling Musdia Alli 361
93–96; error-correction equation 95,
103; forecasting 100, 100; individual Nakanishi, J. 312
transferable quota and 90; IS method National Action Plan for Climate Change 26
94, 96, 96; Kuznets curve hypothesis National Health and Family Planning
and 88–89, 103; linear-type relationship Commission of the People’s Republic of
94; maximum sustainable yield 90–91; China 377; see also population planning
methodology 90–96; N-shaped curve policy in China
and 89, 96, 98; overview 3, 88, 103; national and subnational sustainability:
pooled mean group method 94, 96, 98; background information 327–329;
448 Index
capital and, defining 327; dynamic Pati, A. P. 160
average utilitarianism criterion 328–329, Payre, W. J. 420
331–334, 333–334, 349, 350–351; Pearce, D. W. 288
human capital and 327, 330; natural Pearson, L. J. 44
capital and 330; overview 4, 329; Peiro, A. 202
present discounted population 328–329; Pesaran, M. H. 94
produced capital and 327–328; research Pevalin, D. J. 228–229
questions 328–329, 351–352; total Ploeg, F. van der 77, 79–81
utilitarianism per capita wealth criterion Poelhekke, S. 77, 79, 81
327–328, 331–334, 333–334; see also pollutant concentrations 52–53, 53
Japanese wealth accumulation and pooled mean group (PMG) method 94, 96, 98
population change (2000s) population planning policy in China:
natural capital (NC): energy infrastructure age structure 378, 382–383, 382;
project evaluation 55–56; inclusive assumptions, main 377–380;
urban capital stock and 16–19, 18; background information 374–377;
national and subnational sustainability childbearing desire survey 384–386,
327; productivity measurement and 385; conclusions 402; data 377;
sustainable development 287, 294, domestic population 399, 401; fertility
297–298, 303, 305–306; public debt rate 375, 378, 394, 394, 395, 396, 398;
effect 78, 80; renewable energy health capital 390–391; human capital
on sustainability 413, 414; Russia change 375–377, 379–380, 387, 387,
158; social cost-benefit analysis of 388, 389–391, 389, 389; human capital
Fukushima Dai-ichi Nuclear Power transfer 399, 401, 402; human capital
Plant decontamination 311, 313; total wealth 380; life expectancy 391, 391;
factor productivity 136, 142, 161; methodology 377; one-child policy
see also productivity measurement and 374–375, 390, 392–393; overview
natural capital 4, 376–377, 402; population forecast
Neilson, W. 198 results 379–380, 381, 382–384, 382,
Neumayer, E. 137 383, 384; results 380–387; sex ratio
Nguyen Van, P. 88–89 392–393; social impact 391–394,
Nordhaus, W. D. 142, 294 394, 395, 396; two-child policy
Normalized Difference Vegetation Index 374–376, 382, 389–390, 392–393, 402;
(NDVI) 229 urbanization 399, 400; urban versus
Norsworthy, J. R. 138 rural areas 386–387
Norway 398 Portugal 135, 398
Nuclear Regulation Authority monitoring Powdthavee, N. 174, 228–229
reports 316 PP-Fisher method 94, 96, 96
pre-construction state of project 48, 49, 50
OECD Guidelines on Measuring present discounted population (PDP)
Subjective Well-Being 227 328–329, 349, 349
Oleson, K. L. L. 135 principle component analysis 29, 30
Ollivier, T. 135 Pritchett, L. H. 298
operation state of project 49, 50 produced capital (PC): inclusive urban
order decay model 354, 356 capital stock and 16–19, 16; national
ordinary least squares (OLS) regression and subnational sustainability 327–328;
25, 175, 237, 240 productivity measurement and natural
Organisation for Economic Co-Operation capital 142; productivity measurement
and Development (OECD) 19, 24, 25, 56 and sustainability development 294,
Oshio, T. 175 297; renewable energy on sustainability
Oswald, A. 167, 202, 228, 229 413, 414; social cost-benefit analysis
of Fukushima Dai-ichi Nuclear Power
Pak, Maxwell 379, 383 Plant decontamination 311, 313; South
panel dataset 29, 31–33, 32 Africa 158
Index 449
production index 263–264, 274–275, 275 299, 300; overview 3, 287, 305–306;
production model 262–263, 266, 268 produced capital 294, 297; productivity
productivity: change 292–297, 295, 296, change 292–297, 295, 296, 301, 302,
301, 302, 305–306; growth 294–295, 305–306; productivity growth 294–295,
295; infrastructure and economic 295; sulfur dioxide emissions 290;
growth in India and 23 technical inefficiency decomposition
productivity measurement and natural 291–292, 296–298, 296, 298, 301,
capital: background information 302; time series 292–293; total factor
135–137; capacity utilization 138; productivity change 294–295, 295
conclusions 161–162; cross-country Programme of Research on Road
161; data 142; Data Envelopment Transport and Intermodal Linkages
Analysis method 136, 139–140; distance (OECD report) 25
functions and 139–140; empirical results public debt effect: ARCH-M method and
142–161; empirical studies 137–138; 79, 81, 82, 85; background information
genuine savings 135, 137–138; human 77; conclusions 85; controversy of 77;
capital 142, 155, 157, 159; inclusive estimation result 81, 82, 83, 83, 84;
wealth adjusted by total factor fiscal policy and 77, 85; genuine savings
productivity 141–142, 147, 148–149, and 78–81, 79, 80, 82, 83, 83, 83–84,
150, 150; Inclusive Wealth Index 85; human capital and 78, 80; Inclusive
and 135, 138–139; inclusive wealth Wealth Index and 77–79, 79; Japanese
measurement 138–139; Malmquist case study 80–81, 80; man-made capital
Productivity Index and 136, 139, 161; and 78, 80; methodology 78–79; natural
methodology 138–142; overview 3, capital and 78, 80; overview 2, 85;
137, 161–162; produced capital 142; stabilization and 78, 81, 85; volatility
results 161–162; see also total factor and 77, 79, 81, 82, 85
productivity (TFP) public facility greenery 246
productivity measurement and sustainable public goods 198
development: aggregate technical
inefficiency scores 296–297, 296; QuickBird satellite images 226
background information 286–290;
Canada case study 300, 304–306, 305; railroad greenery 246
capital inefficiency scores 300, 302; rainfall 264, 266, 267, 268, 269
carbon damage inefficiency scores Ramey, G. 79
297, 300, 301, 303; carbon dioxide Ramey, V. A. 79
emissions 287, 289–290, 294–295, RAM legacy database 90, 96
303; China case study 300, 303–306, Rao, D. P. 139
304; conclusions 305–306; contribution reference area income data 173
effect 287, 292–295, 295; cross-country regression model 25, 28–29, 29, 31, 34
290; data 293–294; Decision Making Rehdanz, K. 196–197, 233
Unit 291–293; directional distance relative income (RI) hypothesis 167,
model 287, 290–292, 296; empirical 168–169, 170–171, 173, 181, 182,
results 294–305; empirical studies 287, 183–186
289–290; endowed capital 287; future renewable energy on sustainability: air
studies 306; gross domestic product pollution prevention 407; background
inefficiency scores 298, 299, 300; information 405–409, 406, 407; basic
human capital 294, 297, 303; Inclusive model 409–410; bioenergy 413–415,
Wealth Index and 286, 288–289, 293; 414; capital and 413–414, 414;
infrastructure capital 297; Luenberger climate change solution 406–407;
Productivity Index and 289–290, control variables 411–412; descriptive
292; Malmquist Productivity Index statistics 412, 412; discussion 415;
and 289–290; methodology 290–293; economic growth 407; electricity
natural capital 287, 294, 297–298, 303, 406, 407; empirical results 413–415,
305–306; output inefficiency score 298, 414; estimation model 412–413;
450 Index
explanatory variables 410–411; fiscal services recovery 311, 316, 317,
policy 406; global investment 405; 318; categories 311, 316, 320, 320;
gross domestic product and 411–412; conclusions 321–322; costs 310–312,
human capital 413, 414; impact on 312, 321; final disposal plans 314–315,
environmental, natural, or human factors 314; forests 319–320; housing 319;
407–408; Inclusive Wealth Index and in-situ area 316–317, 320–321, 321;
408–409, 414; limitations of study 415; limitations of study 321–322; man-made
methodology 409–413; natural capital capital 312; methodology 315–320;
413, 414; overview 4, 409, 415; political natural capital and 311, 313; natural
rationales 405–407; produced capital decontamination 310–311; Nuclear
413, 414; security of energy 405–406 Regulation Authority monitoring reports
Reserve Bank of India 37 316; once under order area 316–317,
reverse causality 34 320–321, 321; overview 3–4, 313, 321;
Richard, L. E. 184 preparation for 313–315; present value
Richerzhagen, C. 304 of costs 315, 318–320; previous studies
Robinson, K. L. 229 311–313, 312; produced capital and 311,
Rodriguez-Clare, A. 137, 376, 379, 402 313; results 320–321, 320, 321; roads
Rothschild, M. 170 319; scope of 310, 313–315; still under
Russia: economic crises 160; gross order area 310, 316–317, 320–321, 321
domestic product 158; natural capital social impact of China’s two-child policy
158; population planning policy 391–394, 394, 395, 396
397–398, 397; total factor productivity social networks 198
157–161, 157, 161 Sonnenfeld, D. A. 147
South Africa: human capital 158; produced
Sahoo, P. 25, 28, 33 capital 158; subjective well-being 173;
Sander, H. 197–198 total factor productivity 157–158, 157,
SARTRE project 419 159, 160
Sato, M. 77, 135 Spain 83, 84, 398, 418
scale effect 88–89 Statistical Bulletin of National Economic
Schaefer, M. B. 99 and Social Development and National
Scheffer, M. 137 Population Development Plan 377
Scholz, I. 304 status effect 170
Schreyer, P. 136 Stevenson, B. 175
Schumacher, E. F. 43 stock status plots (SSP) method 91
Schwarz’s Bayesian Criterion (SBC) structural equation modelling (SEM) 170,
value 95 174, 174–175, 175
Science Council of Japan 201 subjective well-being (SWB): environmental
Sea Around Us Project 90–91, 95 value of green spaces in Japan 205;
self-reported household income data 173 greenery and well-being 225–229, 228,
Senik, C. 184 229, 239; well-being determinants and
Shi, H. 303 167, 170, 173–174, 184–186, 205;
Sinha, A. 89 see also well-being determinants
Smith, A. 387 Sugiawan, Y. 89
Smyth, R. 196–197 sulfur dioxide emissions 290
Sneddon, C. 288 Sun, J. W. 12
social capital 78 survey of health damages 354, 363
social cost-benefit analysis of Fukushima sustainability: economic growth and,
Dai-ichi Nuclear Power Plant balancing 1; empirical studies 287;
decontamination: agriculture 319; air Inclusive Wealth Index and 327; total
dose rate reduction 316–318, 317, factor productivity and 136, 286;
321; annual services value 318–320; wealth and 286–289; see also genuine
background information 310–311, 312; savings (GS); national and subnational
benefits 310–311, 321, 432; capital sustainability; urban sustainability
Index 451
sustainable development 135; see also asset and 286; United Kingdom 155–156,
productivity measurement and 155; United States 155–156, 155;
sustainable development various countries 142, 143, 144–145;
Sustainable Development Goals (SDGs) see also productivity and natural capital
135, 286 total utilitarianism (TU) criterion 332–334,
Sweden 398 333–334
system dynamics (SD) modelling 46, total utilitarianism (TU) per capita wealth
50–52, 51 criterion 327–328, 331–334, 333–334
system of national accounts (SNA) 286 Tsao, C. C. 408
Tsunetsugu, Y. 196
Tamaki, T. 89 Tsurumi, T. 226, 245, 247
Tamangapa landfill 354–356, 359, 360, Turkey 135
361–367, 362, 363, 364, 365, 366, Twelfth Five-Year Plan 26
370–371 two-stage least-squares method 205, 208
Tang, D. 136
Taylor, R. B. 185 Ueta, K. 135
technical effect 89 United Kingdom: fish catches, forecasting
technical inefficiency decomposition 100, 101; fully automated vehicles in
291–292, 296–298, 296, 301, 302 418; genuine savings in 84; green spaces
Telecom Regulatory Authority of India and well-being in 196–197; output
(TRAI) 27, 37 inefficiency 298, 299; quota-management
Tella, R. D. 202 system for fish catches 92; total factor
temperature 264, 266, 267, 274 productivity and 155–156, 155
Ten-Item Personality Inventory (TIPI) 174 United Nations Convention on the Law of
Tidball, R. 52 the Sea (UNCLOS) 90
timeout state trigger 49, 50 United Nations Environment Program
time-series analysis 56, 57 (UNEP) 286–287, 289, 293, 320,
time series productivity measurement 357, 362
292–293 United Nations General Assembly 135, 286
Tokimatsu, K. 43 United Nations University International
top-down (TD) modelling 46 Human Dimension Program (UNU-
total allowable catch (TAC) 90 IHDP) 286–287, 289, 293, 320
total factor productivity (TFP): Brazil United States: Environmental Protection
157, 157, 159, 161, 161; BRICS Agency 362; fertility rate 398; fish
countries 156–160, 157, 159, 161; catches, forecasting 100, 101; fully
change 146–147, 146, 147, 292–295, automated vehicles 419; green
295; China 157–158, 157, 159, 161; spaces and well-being in 197–198;
contribution effect and 294–295, immigration 392; output inefficiency
295; cumulative indices 146–147, 298, 299; quota-management system
147; France 155, 155; G7 countries for fish catches 92; relative income
152, 153–154, 155–157, 155, 156; hypothesis and saving behaviour 167;
Germany 155, 155; inclusive wealth subjective well-being in 184; total factor
considering 147, 148–149, 150, 150; productivity 155–156, 155
Inclusive Wealth Index Adjusted and Urban, F. 303
141–142, 151–152, 151, 152; India 157, urban sustainability: background
159–161, 159, 161; Japan 155–156, 155; information 5, 8; capital stock between
Malmquist Productivity Index and 289; cities and, comparing 8–10; city
measurement 138–141; natural capital databases and 19–20; city evaluation
and 136, 142, 161; Russia 157–161, indexes and 5, 6–7, 8, 19–20; Data
157, 161; South Africa 157–158, 157, Envelopment Analysis method and
159, 160; sustainability and 136, 286; urban capital stock 9–16, 13, 15;
technical efficiency change and 161; decomposition model of urban capital
technological change and 161; time stock 10–13, 13, 16–19, 17, 18;
452 Index
Inclusive Wealth Index and 8–13, 167, 168–169, 170–171, 173, 181, 182,
19–20; Japanese ordinance-designated 183–186; results 176, 177–178, 178,
city case study and 13–19; overview 2, 179–180, 180–181, 182–183, 183;
8, 19–20; urban planning policies and 5; Richer dummy 173, 180; statistical
see also inclusive urban capital stock specification 174–176, 175; status effect
170; structural equation modelling 170,
Value of a Statistical Life (VSL) 55 174–175, 175; subjective well-being
Vandenbussche, J. 303 167, 170, 173–174, 184–186; variables
Vietnam 396–397 172, 173–174, 191–193; see also
Vignoles, A. 229 greenery and well-being
volatility and public debt effect 77, 79, 81, Welsch, H. 227
82, 85 Wichmann, B. 198
Vouvaki, D. 287, 289 willingness to pay (WTP) 195, 420–421,
427, 428, 429, 435
Wang, Y. 158, 303 Wilson, D. C. 358
water-front greenery 245–246 Wolfers, J. 175
wealth accumulation 329–331, 330; women’s status and fertility rates 398
see also Japanese wealth accumulation World Bank 78, 80, 271, 354, 386, 390
and population change (2000s) World Development Indicators (WDI)
wealth and sustainability 286–289; 80, 95
see also Inclusive Wealth Index (IWI) World Development indicators dataset 327
well-being determinants: background World Meteorological Organization
information 167, 168–169, 170–171; (WMO) 362
capital 288–289; community Worm, B. 87
attachment 170–171, 176, 184–185,
194; conclusions 185–186; data 171, Xepapadeas, A. 287, 289
172, 173; descriptive statistics 171, Xu, Bing 379, 383
172; discussion 183–185; ecosystem
288; income effects 173, 184–185; Yamaguchi, R. 135, 137, 313, 329
information effects 170, 183; life Yang, D. T. 396
satisfaction 167, 174, 176, 178, 181, Yao, Y. 158, 303
182, 183–186, 194; maximum likelihood
estimation 175; methodology 174–176, Zeng, Yi 375
175; ordinary least squares regression Zenrin Geo Intelligence Co., Ltd. 173
(OLS) model 174; overview 3, 170, Zhou, P. 287
185–186; relative income hypothesis Zipperer, V. 136

Wealth, Inclusive Growth

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wealth, Inclusive Growth

Uploaded by

Copyright:

Available Formats

Wealth, Inclusive Growth

The excessive pursuit of economic interests has resulted in severe environmental

Shunsuke Managi is Distinguished Professor and Director at the Urban Institute

The Evolution of Economic Wellbeing

Entrepreneurship and Local Economic Development

The Future of the Economy

Regional Government Competition

The Dynamics of Growth in Emerging Economies

Economic Integration in Asia

Wealth, Inclusive Growth and Sustainability

For more information about this series, please visit www.routledge.com/Routledge-

Edited by Shunsuke Managi

Typeset in Times New Roman

List of figures vii

1 An evaluation of inclusive capital stock for urban planning 5

2 Impact of infrastructure in India 23

3 Sustainability evaluation for energy infrastructure:

4 Public debt as a negative stock in sustainability indicator 77

5 Global marine fisheries with economic growth 87

6 Inclusive wealth adjusted by total factor productivity as

8 Environmental value of green spaces in Japan: an application

9 Greenery and well-being: assessing the monetary value of

10 The impact of climate change and extreme events on

11 Sustainable development and performance measurement:

12 Social cost-benefit analysis of decontamination after

13 National and subnational sustainability: a case study of Japan 327

14 The effect of landfill gas emission on global warming

15 Human capital change and social impact under China’s

16 The impact of renewable energy on sustainability: empirical

17 Consumer demand for fully automated driving technology 418

0.1 Classification in inclusive wealth 2

1.1 Literature review of city evaluation indexes 6

6.1 Average of TFP for 140 countries (1990–2010) 143

Ebrahim Aly is a doctoral student at the Department of Urban and Environmental

Figure 0.1 Classification in inclusive wealth

Author Index name Purpose Key evaluation dimensions (factors Aggregation

2 The Inclusive Wealth Index (IWI) for urban planning

2.1 The evaluation of inclusive urban capital stock

2.2 Comparing capital stock between cities

Objective function: Max. βk = wp Produced capitalk + wh Human capitalk

wp Produced capital j + wh Human capital j + wn Natural capital j

Objective function: Max. θk  (4)

0 ≤ λj ≤1 j = 1,, k ,, J  (9)

where θ is the integrated evaluation score and is equal to β. λ is the intensity

2.3 Decomposition analysis of urban capital stock change

ority of budget spending on human capital in year t is PHt (= PRIORITYHuman t

FHt = E tH × PHt ×St .

∆FHt , t +1 = FHt +1 − FHt = E tH+1 × PHt +1 ×St +1 − E tH × PHt ×St  (11)

∆FHt ,t +1 = E tH+1 × PHt +1 ×St +1 − E tH × PHt ×St

= (E tH + ∆E tH,t +1 )×(PHt + ∆PHt ,t +1 )×(St + ∆St ,t +1 ) − E tH × PHt ×St

= ∆E tH,t +1 × PHt ×St +1 + E tH ×∆PHt ,t +1 ×St + E tH × PHt ×∆St ,t +1

In this transformation, we allocate the second-order interaction terms and the

Figure 1.1 Diagram outlining the approaches in this study

3 Case study for a Japanese ordinance-designated city

3.1 Data description for the empirical study

Population Ratio of Day Income Labor Labor Labor

3.2 Empirical study for the relative evaluation

3.3 Empirical study for a decomposition analysis of

ECE PCE SCE F(Human)

ΔPCE increased. Meanwhile, Hamamatsu, which is the reference city of Sapporo,

4 Discussion and conclusion

Table 1A.1 Grouping of local government budget expenditures

Government budget expenditure category Grouping in this study

Public Infrastructure Investment as a % of GDP

List of figures vii

1 An evaluation of inclusive capital stock for urban planning 5

2 Impact of infrastructure in India 23

4 Public debt as a negative stock in sustainability indicator 77

5 Global marine fisheries with economic growth 87

13 National and subnational sustainability: a case study of Japan 327

17 Consumer demand for fully automated driving technology 418

0.1 Classification in inclusive wealth 2

1.1 Literature review of city evaluation indexes 6

6.1 Average of TFP for 140 countries (1990–2010) 143

Objective function: Max. θk (4)

0 ≤ λj ≤1 j = 1,, k ,, J (9)

∆FHt , t +1 = FHt +1 − FHt = E tH+1 × PHt +1 ×St +1 − E tH × PHt ×St (11)

3.2 Empirical study for the relative evaluation

3.3 Empirical study for a decomposition analysis of

2.2 Empirical studies showing relationship between

3.2 Infrastructure variables considered and used in the

∆K = {∆HC , ∆NC , ∆PC } (3.3)

log ( Et ) = β0 + β1 log ( Pt ) + β2 log (GDPt ) + εt (3.5)

a = fc, condt , opdt , lf , fu , j , ems , Inv (3.6)

∆PC = Inv − Dep − loanPbk − OM (3.9)

∆HC = ∆Emp + ∆Edu − ∆Health (3.11)

∆Emp = ∆Lstt ⋅ rt sc (3.12)

∆MYSI = β0 + β1 log (∆En) + ε (3.13)

∆Health = (∆Mortality ⋅ VSL) + (∆RAD ⋅ r ) (3.15)

∆Mortalities = 0.104 ⋅ ∆PM 10 (3.16)

% yield loss =−3.18 − 0.188(∆SO2 concentration in ppb) (3.19)