Growth With Inequality: An International Comparison On Income Distribution 9789814401708, 9789814401685

Citation preview

An International Comparison on Income Distribution

8452.9789814401685-tp.indd 1

7/6/12 10:47 AM

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-fm

An International Comparison on Income Distribution editor

Jinjun XUE (Nagoya University, Japan)

World Scientific NEW JERSEY

•

8452.9789814401685-tp.indd 2

LONDON

•

SINGAPORE

•

BEIJING

•

SHANGHAI

•

HONG KONG

•

TA I P E I

•

CHENNAI

7/6/12 10:47 AM

Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

Library of Congress Cataloging-in-Publication Data Growth with inequality : an international comparison on income distribution / edited by Jinjun Xue. p. cm. ISBN 978-9814401685 1. Income distribution. 2. Income--Regional disparities. 3. Regional economic disparities. 4. Economic development--Social aspects. I. Xue, Jinjun HC79.I5G776 2012 339.2--dc23 2012011415

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

Copyright © 2012 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

In-house Editor: DONG Lixi

Typeset by Stallion Press Email: [email protected]

Printed in Singapore.

Lixi - Growth with Inequality.pmd

1

1/11/2013, 5:20 PM

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

FOREWORD

In the era of globalization and liberalization, the world is enjoying higheconomic growth as well as suffering from the pains of unequal distribution of its economic outcomes. The activities of the Occupy the Wall Street movement and anti-government demonstration in the U.K., France, Japan, Thailand, Egypt, Libya, China and so on, reflect that inequality has become an international phenomenon that happens in countries regardless of whether they were rich or poor, socialist or capitalist, authoritarian or democratic. Consequently, the inequality issue has not only become a hurdle of one country’s potential development but also a threat to its social and political stability. The spread of the Jasmine Revolution in some areas of North Africa and the Arab Spring is a good example. In light of such events, this book raises three key questions: (1) Can high growth shrink inequality gradually? (2) Can government intervention bring about the equalization of income distribution in an efficient manner? (3) Is income disparity an engine, or an obstacle, of high growth? This book will answer the above questions and display the inconsistency between high growth and increasing inequality with case studies of 11 individual countries and some OECD countries based on the original household surveys and latest statistical data. In the case of China, India, Indonesia, Hong Kong and Singapore, our studies show that high-economic growth is accompanied with increasing income inequality. This is especially so in China, where high level of economic growth (making it the second largest economy in the world) and opportunities for some people and regions to get rich first has been paralleled by rising levels of inequality and continuing poverty. Similarly in India, the benefit of economic growth went to some groups further segmenting v

b1363-fm

June 5, 2012

11:23

vi

9in x 6in

Growth and Inequality: An International. . .

Foreword

society between the upper and lower classes. In contrast, with China, Indonesia, Thailand and India, permits farmers to migrant among regions freely but this has no significant improvements in reducing regional disparity. The same story can be witnessed in Hong Kong and Singapore, the two top ranking economies in international competition and market freedom, where there is a high-economic performance with high inequality. Contrast the cases of Japan, Germany, the U.K. and Korea demonstrate another picture of equitable income distribution with long term low economic growth or depression. Japan, one of the most equalized country in the world, has been suffering low growth or negative growth since the 2000s. In order to raise labor productivity, Japan introduced the annual salary system from the United States and gave more stimulation incentives to those who are deemed proficient. However, while raised Japan’s international competitiveness slightly they also increased the number of low wage part time workers and in turn the poverty ratio and the Gini coefficient. Even the change is slight but in addition to other factors, it caused frequent changes of Prime Minister and finally the change of cooperating party and the regime. Since 2009, the Democratic Party of Japan (DPJ) has been applying some new policies in order to reduce the income gap such as cutting the income of public servants, increasing taxes for the rich and protecting low-income groups by building a new safety net. Unfortunately, to today, it seems that there are no significant improvements in economic growth and income equalization just rapid change of prime ministers. Many countries in the world are trying to balance economic growth and inequality by designing a more effective policy framework. A newly issued study shows that in OECD countries, some policy approaches have resulted in entail a double dividend as they reduce income inequality while boosting long-run GDP per capita. By contrast, several policies may entail a trade-off between reducing income inequality and raising GDP per capita. However, the facts are that many countries have experienced economic slowdown and an increase in income inequality in the past decades. Our new research in the book shows that in some countries, top earners have captured a disproportional share of the overall income gains while for others income has raised a little. This has a growing consensus that an assessment of economic performance should not focus solely on overall income growth, but should also take into account income distribution. On the income policy, a key question is whether the type of growth-enhancing policy reforms advocated for each OECD country and the BRICS in Going for Growth might have positive or negative side effects on income inequality.

b1363-fm

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Foreword

b1363-fm

vii

In conclusion, in pursuing growth and redistribution strategies simultaneously, policy makers need to be aware of possible complementarities or trade-offs between the two objectives. There are many factors that we can draw upon to explain the rising level of inequality. The book not only indicates the seriousness of inequality but also explores the factors that cause the inequality and analyzes their economic and social consequences. Our studies indicate that in most countries, family background and educational disparity are the most important factors that cause income disparity. In the case of China, the registration system is the most important factor induced urban–rural income disparity by which even the large scale labor migration cannot reduce the disparity. In the case of India and China, the economic benefits have gone to some rich groups with high educational background and skilled labors, accompanying with their political capital (social class, position, government officer, etc.) and then it made people subdivided into different class levels and stylize its social class structure. We can summarize some of the characteristics of the book as follows: First, the book provides readers some of the basic theories and analysis methodologies for the subject of income distribution help readers. This will attain a better understanding about growth and inequality and learn how to analyze the issues using a proper method and index. Second, most empirical analyses conducted in the book are based on original household surveys and the newest public data. Especially we conducted country studies by using datasets of household income survey (China), consumer’s expenditure surveys (India, Singapore, Thailand), educational expenditure survey (Japan) and so on. For the reason, we are confident to say that this book can provide a real picture of the countries via reliable use of statistical data and sources. Third, the chapters are written by some leading scholars from China, India, Korea, Japan, the U.K., U.S.A., Germany, France and the OECD. Such scholars can deliver a comprehensive explanation and in-depth understanding from original taste. Especially, the emerging economies like China and India are main targets in the book. One aspect of the book that need to be stressed concerns our take on the Kuznets Hypothesis. Kuznets developed a hypothesis to demonstrate a trade-off relationship between economic growth and income distribution by saying that income growth will increase inequality at first and then bring on an equalized income distribution gradually as times goes by. However, in our studies, it does not seem that the evidence supports the hypothesis.

June 5, 2012

11:23

viii

9in x 6in

Growth and Inequality: An International. . .

Foreword

Though some countries had experienced equal income distribution and appeared an inverted U-shape, their Kuznets curves has been reversing to an increasing trend in recent years. Even in countries like China in which egalitarianism has been a slogan of socialism, income disparity has become a serious problem. This book displays you a general understanding of income inequality, explores main factors which are deemed to have caused the disparity, and shows the future trend of growth and income. However, frankly speaking, our desire to find an efficient way to boost economic growth while realizing an equal income distribution remains unsolved. We must therefore await further study but hopefully in the meantime we provided readers with enough hints to undertake their own observations following this introductory research. Jinjun Xue (editor) 14 February 2012 Nagoya, Japan

b1363-fm

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-fm

CONTENTS

Foreword

v

Acknowledgments

xi

List of Contributors

xiii

Part 1 Globalization, Liberalization, Growth and Income Inequality

1

Chapter 1

Growth and Inequality in China

3

Chapter 2

Growth and Inequality in Hong Kong

21

Chapter 3

Growth and Inequality in the United States

53

Chapter 4

Growth and Inequality in India

79

Chapter 5

Growth and Inequality in Germany

111

Chapter 6

Growth and Inequality in Korea

121

Chapter 7

Growth and Inequality in the UK

139

Chapter 8

Growth and Inequality in Indonesia

161

Chapter 9

Growth and Inequality in Thailand

181

Chapter 10 Growth and Inequality in Singapore

191

Chapter 11 Growth and Inequality in Japan

207

ix

June 5, 2012

11:23

x

9in x 6in

Growth and Inequality: An International. . .

b1363-fm

Contents

Part 2 Factor and Policy Analysis on Income Inequality

229

Chapter 12 Informal Employment and Income Disparity

231

Chapter 13 Educational Disparity and Income Disparity

255

Chapter 14 Housing Inequality and Underlying Factors in Urban China

277

Chapter 15 Agriculture Profitability and Income Disparity

297

Chapter 16 Labor Migration and Income Inequality

313

Chapter 17 Income Inequality, Labor Migration and the Lewis Turning Point

333

Chapter 18 Trade-offs and Complementarities Between Growth and Inequality in OECD Countries

347

Index

379

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

ACKNOWLEDGMENTS

• This is part of the studies “The Great Economic Turning Point in China and its Impact on the World Economy, a Grants-in-Aid for Scientific Research awarded by Ministry of Education, Culture, Sports, Science and Technology of Japan CKAKENHIA; Project Leader: Jinjun Xue; Project Number: 23252008; Project Period: 2011–2015. We are grateful to Economic Research Center of Graduate School of Economics, Nagoya University, Japan Kitankai (the alumni of the school), for its support to publish the book. We also thank Institute of Economic Research (CASS), Institute of Population and Labor Economics (CASS), University of International Business and Economics, School of Management and Economics of Beijing Normal University (China), Nanyang Technology University (Singapore), Nagoya University, Kyoto Sangyou University (Japan), University of Hyderabad (India), Levy Economics Institute of Bard College Annan dale-on-Hudson (U.S.A.), Thailand Development Research Institute (TDRI, Thailand), University of Paderborn (Germany), Economics Department of the OECD, for their help of conducting the research. • We give our special thanks to Professor Yuko Arayama (Nagoya University), Li Shi (Beijing Normal University) for their constructive criticism and Wang Bo (UIBE), Wang Yu (Shandong University), Ms. Dong Lixi (World Scientific), Xie Shaoguang, Li Yanling and Yun Wei (Social Science Academy Press of China) for their contribution to English editing and publishing.

xi

b1363-fm

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-fm

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

LIST OF CONTRIBUTORS (by chapter order)

Jinjun XUE Professor, Economic Research Center, Graduate School of Economics, Nagoya University, Japan. Shaoguang WANG Chair Professor, Department of Government and Public Administration, The Chinese University of Hong Kong. Yin XIA Assistant Professor, School of Government, Sun Yat-sen University, China. Edward Nathan WOLFF Professor, Department of Economics, New York University, USA. Senior Scholar, Levy Economics Institute of Bard College, USA. Ajit ZACHARIAS Senior Scholar, Levy Economics Institute of Bard College, USA. Thomas MASTERSON Research Scholar, Levy Economics Institute of Bard College, USA. Vamsi VAKULABHARANAM Associate Professor, Department of Economics, University of Hyderabad, India.

xiii

b1363-fm

June 5, 2012

11:23

xiv

9in x 6in

Growth and Inequality: An International. . .

List of Contributors

Stefan GRAVEMEYER Statistical Consultant, German Statutory Accident Insurance, Germany. Thomas GRIES Professor, Center for International Economics, University of Paderborn, Germany. Kang Kook LEE Professor, College of Economics, Ritsumeikan University, Japan. Stephen DAY Associate Professor, Faculty of Economics, Oita University, Japan. Susumu HONDAI Emeritus Professor, Kobe University, Japan. Nobuki SUGITA Professor, Economic Research Center, Graduate School of Economics, Nagoya University, Japan. Kong Weng HO Senior Lecturer, School of Business, SIM University, Singapore. Fumio MAKINO Professor, Institute of Comparative Economic Studies, Hosei University, Japan Yang DU Professor, Institute of Population and Labor Economics, Chinese Academy of Social Sciences, China. Erbiao DAI Research Professor, International Center for the Study of East Asian Development (ICSEAD), Japan.

b1363-fm

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

List of Contributors

Tadashi SONODA Associate Professor, Graduate School of Economics, Nagoya University, Japan. Fang CAI Professor, Institute of Population and Labor Economics, Chinese Academy of Social Sciences, China. Meiyan WANG Associate Professor, Institute of Population and Labor Economics, Chinese Academy of Social Sciences, China. Ryoshin MINAMI Emeritus Professor, Hitotsubashi University, Japan. Xin Xin MA Research Fellow, Keio Economic Observatory, Keio University, Japan. Peter HOELLER Head of the Public Economics Division, Economics Department, OECD, France. Isabelle JOUMARD Senior Economist, Economics Department, OECD, France. Isabelle KOSKE Senior Economist, Economics Department, OECD, France.

b1363-fm

xv

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-fm

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Part 1

Globalization, Liberalization, Growth and Income Inequality

1

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

2

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 1 GROWTH AND INEQUALITY IN CHINA Jinjun Xue∗

1.1. Introduction In 1980, the father of China’s Reform and Opening-up, Deng Xiaoping, put forward the idea that “China should ‘double its national income’ ” after referencing Japan’s “National Income Doubling Plan”. Since then, the income of Chinese citizens has doubled, or even quadrupled, every 10 years. In the 30 years from 1980 to 2010, China’s gross domestic product (GDP) grew by 9% annually and in 2010, it surpassed Japan to become the world’s second largest economy. Its export volume has also surpassed that of Germany and become the largest in the world. The International Monetary Fund (IMF) predicted (World Economic Outlook 2011 ) that if China continues its high-speed growth and the U.S. economy stagnates, China would overtake America as the world’s largest economy in 2016, much faster than economists and international organizations have predicted.1 Meanwhile, living standard of Chinese people has been better off and their GDP per capita increased from U.S. $313 in 1980 to U.S. $4428 in 2010,2 ranked as an up-middle income economy in the world. Furthermore, the 12th FiveYear Plan for the National Economic and Social Development, passed by the 11th National People’s Congress in March 2011, proposed that GDP per capita should be doubled again, namely to U.S. $8856 by 2015. However, behind these “lights” there exist some “shadows” in the Chinese economy. On the one hand, the issues of increasing income disparity, ∗ Professor, Economic Research Center, Graduate School of Economics, Nagoya University, Japan. 1 Wall Street Journal Japanese version, 26 April 2011. 2 World Bank, World Development Report 2011. http://data.worldbank.org/indicator/ NY.GDP.PCAP.CD

3

b1363-ch01

June 5, 2012

11:23

4

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

nationwide environment pollution, spreading official corruption, serious conflict between the public and the governments, and so on are some of them. As we knew from international experience that when a nation’s GDP per capita got to $3000 and above, it may enter a transition phase in which some problems such as income inequality, social conflict, and political risk will come about. These problems may stagnate the economic growth and induce the country into an unstable situation both in society and politics if the nation could not pass through economic and social transition smoothly. Consequently, the country may not be able to have a sustainable development and promote its income growth but fall in a “Middle Income Trap.” China is now on the cross-road of the above transition phase and facing to the “Middle Income Trap.” This chapter will introduce the process of China’s rapid economic growth and discuss the issue of income inequality using the data sets of China Household Income Project (CHIP) and China Urban Labor Survey (CULS) in addition to the public data of the government. Moreover, the main factors causing these problems will be examined from the aspects of the policy orientation and market factors. This chapter is composed of five sections. Section 1.2 presents income inequality in China by different measures; Section 1.3 gives an explanation to the Income Kuznets curve and predicts the future trend of income distribution in China; Section 1.4 implies some economic and the social meaning of income inequality for the future development.

1.2. Income Inequality in China During the era of Mao Zedong, for about 30 years starting at the beginning of the 1950s, China advanced a socialist planned economic system under the slogan, “Make all people equal.” Under that system, urban workers’ wages were set at the same rate and fixed for long periods. Even in rural areas, there was little income disparity, because under the people’s commune system agricultural income was distributed nearly equally irrespective of efforts. This meant that there was no serious problem with inequality, despite some degree of it. On the other hand, this injudicious equality did not give people the incentive to work, leaving production inefficiencies nationwide and preventing the people from rising out of poverty. For this reason, China was called “a poor socialist country with egalitarianism”. However, in the era of Deng Xiaoping, China espoused a “let those who

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in China

b1363-ch01

5

would become wealthy first, do so” policy to boost labor incentives, and set markedly divergent incomes in each area, among regions, among industries, and among job types. Then free market principles were introduced, leaving workers’ wages and farmers’ incomes to be determined by the market. Consequently, while China has been achieving its improvements of industry and agriculture productivity and enjoying the rapid economic growth, income gaps between regions and classes has been widening and become a serious economic and social problem. Aggravating the expanding disparity is the identification of problems such as bureaucratic corruption and distorted government policy. This section will analyze three types of disparities: Urban–rural gap, inter-regional disparities, and income disparities among residents in both urban and rural areas.

1.2.1. Urban–Rural Income Gap There is an expression, “urban area looks like Europe and rural area looks like Africa in China” that accurately conveys the polarization of China’s “thriving cities” and its “impoverished countryside.” This was the frank impression of a German ambassador who was stationed in China, expressing quite well the contrast between China’s urban and rural areas. The urban–rural gap is not only a problem of the present day, but actually started from 1949 with the establishment of the new China. However, since the reforms and opening-up of the country, this gap has continued to widen speedily. While there are a variety of contributing factors to be explained, in China’s case it does not seem an exaggeration to say that the disparities have been primarily caused by the government’s policy orientation of emphasizing heavy industry over agriculture and cities over the countryside. One policy factor is the government’s strategy to promote heavy industries based on a belief that it is the path to success. Starting in the 1950s, China began implementing a development strategy that preferentially focused on heavy industries followed the development model in the former Soviet Union, which was originated from Marxist economic theories. In 1958, the country embarked on the Great Leap Forward Campaign under the slogan, “China must surpass the United Kingdom in 10 years and the United States in 15 years.” Amid frenetic advancement of industrialization and urbanization, the government began implementing differential prices where industrial goods were set with high prices and agricultural goods

June 5, 2012

11:23

6

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

with low prices, in order to support industrialization and guarantee the food supply for urban residents. This policy induced a “scissor differences” problem (a reference to the fact that, when the price indices are graphed chronologically, the prices of agricultural goods trend down while the prices of industrial goods trend up, forming a scissor-like pattern) in the price of agricultural and industrial products, stripping rural residents of much of their profits and beginning to expand the income disparity between urban and rural areas. Another government policy factor is the urban residence registry system (named hukou in Chinese) segregating cities from the countryside. During the Great Leap Forward Campaign, to prevent excessive migration into the cities and secure food supplies for city residents, starting in 1958 the government implemented the urban residence registry system (under which rural residents not registered in the cities could not enjoy urban lifestyles, housing, education, employment, welfare, retirement pension, etc.) that managed urban residents separately from rural residents. Because of this system, until the 1990s the cities were segregated from the countryside, preventing large-scale labor migration from rural to urban areas. Up to 1978, approximately 80% of China’s population was rural statue, whereas the figure is currently approximately 60%. In addition to the existence of the urban residence registry system, the government’s policy preference to emphasizing industry, urban areas, and urban residents over agriculture, rural areas, and rural residents has made China into a dual economy and society, and has resulted in a disparity between the cities and the countryside. As shown in Fig. 1.1, urban– rural gap, measured by rural household average net income (income less administrative expenses and taxes) and urban household disposable income (earnings after tax and social insurance expenses) reached 2.6 times as of 1978. Although the gap shrunk to 1.8 times, thanks to the implementation of the farm management contract system and other rural reforms starting in 1978, the gap resumed its expansion starting in 1984 when the focus of the reforms was shifted to urban areas, and had increased to 2.9 times as of 1994. Thereafter the gap temporarily decreased because of a surge in the price of agricultural products, but the disparity factor had grown all the way to 3.2 times as of 2010 (China Statistics Yearbook, 2011). It has been pointed out that this figure is still an underestimate, with some scholars holding that China is the most unequal society in the world, because the urban–rural disparity is at least six times large when taking into account factors such as the fact that rural residents do not

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in China

b1363-ch01

7

Fig. 1.1. Urban–Rural Gap in China. Source: NBS (2011). China Statistics Yearbook.

enjoy the privileges that urban residents do, namely benefits and social guarantees such as medical insurance, household allowances, corporate pensions, unemployment insurance, guaranteed minimum wages, and financial assistance for schooling and educational investment (Li and Yue, 2004). In addition, if we analyze the urban–rural disparities, urban disparities, and rural disparities using the Theil index and calculate how much they each contribute to the overall level of inequality, we find that the figures are 43%, 19%, and 38%, respectively. Accordingly, the urban–rural disparities account for more than 40% of the total, and appear to be the largest factor explaining the whole inequality (Li and Yue, 2004). According to the Lewis Model of Labor Migration (A. Lewis, 1955), in developing country, there exists a dual economy (due to the income gap between urban area and rural area) which will induce labor migration from agricultural sector to industrial sector until all surplus labor had been used out. At final, the urban–rural income gap will be vanished and then the industrialization will be comprehended. In the case of China, however, owing to the urban residence registry system, labor migration had been restricted from the 1950s to the 1980s, and this regulation induced a continuous increasing of urban–rural income gap. The urban residence registry policy has been easing since the latter half of the 1980s due to the improvement of agricultural productivity and the labor shortage induced by rapid

June 5, 2012

11:23

8

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

growth of urban economy. However, from 1980s–1990s, labor migration was limited in certain areas or within the same region. From later of 1990s, Chinese government deregulated labor policy and allowed gradually labor migration between rural to rural, local to local and finally rural to urban areas, migrant workers still have to come back to their hometown even though they have lived in the city for years. Just because of the contradictory policies, the labor market has been segmented and the labor price has distorted inducing a phenomenon where millions and millions migrants had rushed in to urban areas while the urban–rural income gap had not been shrinking but continued to widen. Figure 1.2 shows a co-relationship between labor migration and urban– rural income gap. It indicates that accompanying an increasing urban–rural gap, a large numbers of migrant workers flowed into urban areas from the countryside, numbering 70 million at the end of the 1990s (Minami and Xue, 2000), and had reached 130 million, or 24% of the rural labor force, in 2008. (Cai, 2009) and 150 million in 2010 (MOA, 2011). However, since migrant workers are very difficult to attain urban statue even if they go to urban areas, and for that reason are unable to gain employment in formal sectors in urban areas and partake of urban education, employment insurance, residences welfare, pensions, or other benefits, a disparity with urban workers has developed in terms of wages and benefits. Many surveys place migrant worker wages at less than 70% of those of urban workers (Table 1.1) (Du and Xue, 2011).

Fig. 1.2. Urban–Rural Income Gap and Labor Migration in China. Source: NBS, China Statistics Yearbook 2011; Ministry of Agriculture, Labor Survey 2011.

b1363-ch01

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch01

Growth and Inequality in China Table 1.1.

9

Wage Disparity between Urban Workers and Migrant Workers.

Migrants (NBS)

Migrants (MOA)

Local Workers

Year

Nominal

Real

Nominal

Real

Nominal

Real

Wage Difference as Share of LW, Times

2001 2002 2003 2004 2005 2006 2007

644 659 702 780 861 946 1015a

644 665.7 702.8 755.9 821.3 889 912.8

— — 781 802 855 953 1,060

— — 774 776.4 841.5 938.9 1,014.4

903 1,031 1,164 1,327 1,517 1,738 2,078

896.7 1,041.4 1,153.6 1,284.6 1,493.1 1,712.3 1,988.5

1.39 1.56 1.64 1.70 1.82 1.93 2.18

Note: “a” is the average monthly earnings for the first three quarter in 2007. Source: Xue and Du (2010). Source: The data of urban local wages are taken from China Statistical Abstract in 2008, and the data for migrants wages are from Statistical Report of NBS and Research Center of Rural Economy, MOA.

The Chinese government noticed this issue and considers increasing farmers’ income as an important policy challenge. In the 11th five-year-plan of social and economic development, the government emphasizes that its most important issue is rural measures to arrest the widening gap between cities and the countryside. Moreover, to solve the “Three rural issues” (agriculture, rural area and farmers) and lessen the burden on rural residents, on 29 December 2005, a resolution was adopted at the 19th meeting of the Standing Committee of the 10th National People’s Congress to abolish the “agriculture tax ordinance” starting on 1st January 2006, thereby doing away with the agricultural taxation system that had been in place for 2,600 years. According to the data of the Ministry of Agriculture, in the 28 provinces that succeeded in waiving the agricultural tax in 2005 ahead of schedule, the burden on farmers was reduced by 22 billion yuan, resulting in substantial benefit for some 800 million farmers. Furthermore, viewing delays in infrastructure development as a factor in rural poverty, in 2006 the government spent 151.3 billion yuan on road construction in rural districts, to improve 325,000 kilometers of roads. Simultaneously, the government also spent 2.417 billion yuan to set up 4,646 passenger terminals in rural areas nationwide, as well as 29,300 bus stops and other vehicle pickup locations. In addition, at the National People’s Congress in March 2007, a new set of policy measures is scheduled to be hammered out to exempt all future tuition and miscellaneous expenses for compulsory education of the children of rural residents in the western part of the country. At a press conference for the National People’s Congress, Premier

June 5, 2012

11:23

9in x 6in

10

Growth and Inequality: An International. . .

J. J. Xue

Minister Wen Jiabao indicated the government’s determination to solve the unemployment, poverty, and disparity issues and espoused new measures, stating, “What is most important at this time is to promote equal opportunity in education. We will implement proactive employment policy measures, gradually shrink income disparities, and establish a social insurance system that covers cities and the countryside. Laws have already been enacted to waive the agriculture tax and special agricultural products tax, and to make the nine years of compulsory education free. Going forward, we will enact laws to create a minimum lifestyle guarantee system that covers cities and the countryside. We are also currently proposing a medical and sanitation business reform act for cities and the countryside, and this will also ultimately become part of the system.” (Xinhua News Agency, March 14th 2007). These initiatives represent acceleration in measures to protect farmers, and these efforts are expected to ameliorate the relative poverty of rural areas going forward. However, the advancement of industrialization and urbanization is increasing the incomes of urban residents faster than that of rural residents, and no matter how good a policy is implemented the urban–rural income gap is widening. It is certainly no simple matter to eliminate the urban–rural disparities.

1.2.2. Regional Disparities In China, since the era of the centralized planned economy, economic regions have been divided into three: the East, the Central, and the West. The eastern region consists mainly of 12 provinces and direct cities, including Beijing, Shanghai, Tianjin, Jiangsu, Zhejiang, Guangdong, and Fujian, and is the most developed region. The central region consists of nine provinces, cities and autonomous areas, including Henan, Shanxi, Jilin, Heilongjiang, and Hubei. The western region consists of 10 provinces, cities, and autonomous regions, including Chongqing (Municipalities like Peking and Shanghai), Sichuan, Guizhou, Yunnan, Sichuan, Shaanxi, Gansu, Qinghai, Xinjiang, and Tibet, etc. Under the planned economy, China implemented a balanced development strategy and diverted financial resources to each region in a roughly equal manner. Moreover, in the era of Mao Zedong, many state-owned enterprises and heavy industry corporations were placed inland because of the Taiwan issue, conflicts with the United States and former Soviet Union, and other military strategy and political reasons. However, in actual practice, funding, management and technical

b1363-ch01

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch01

Growth and Inequality in China

11

problems prevented China from achieving effective economic development. Starting in 1978, the Chinese government chose a strategy of unbalanced development like a terraced rice field (“gradual development”) via the philosophy of “letting those who would become wealthy first, do so,” and implemented preferential policy measures for coastal regions, directing government funding, foreign capital, technology, and human resources to the eastern region. This resulted in the eastern region developing ahead of the rest as an advanced region. However, even as the eastern region functions as the driver of the country’s rapid growth, the economic disparity between the coastal and inland regions is widening. As shown in Table 1.2, to look at the differences in average income among the three major regions of the country, if we set the eastern region at 100, the relative average incomes of the central and western regions in 1980 were 68 and 57, respectively, falling gradually thereafter to 68 and 55 in 1985, 64 and 53 in 1990, and 54 and 44 in 1995, finally reaching large disparities of 44 and 35 in 2000. Recognizing the problem of regional disparities, the government extended the preferential measures being carried out in the coastal regions to certain inland regions, and starting in 1992 began to expand the open regions to include many inland regions. Furthermore, in 2001 the government hammered out a major development project for the western region and has commenced a full-fledged effort to narrow the regional disparities. The project has entailed loans, infrastructure improvement, and spread of education, technology transfer, and assistance from the eastern and central regions to the western region. This resulted in a gradual shrinking of the gap between the western and eastern regions, and between the central and eastern regions, Table 1.2. Year

1980

1985

Actual Income (yuan) Nationwide 447 825 Eastern 569 1,058 Western 389 714 Central 322 580 Index (eastern = 100) Eastern 100 100 Central 68 67 Western 57 55

Trends in Regional Disparities. 1990

1995

2000

2005

2008

1,607 2,103 1,346 1,120

4,804 6813 3,664 2,973

8,167 13,698 6,045 4,758

15,468 25,130 11,992 9,281

25,555 40,116 19,006 16,376

100 64 53

100 54 44

100 44 35

100 48 37

100 47 41

Note: 2008 figures are calculated from real gross regional product (GRP). Source: Compiled from China Statistics Yearbook, 2009 CD-ROM version, and other yearly versions.

June 5, 2012

11:23

12

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

from 2000 to 2005. However, the differences with the eastern region in terms of geographical conditions, infrastructure, capital, human resources, and technology are so large that, even in 2008, incomes in the western and central regions were still only 41% and 47%, respectively, of those in the eastern region. It appears that it will take considerable time for the gap to shrink. Major factors that could be pointed to as contributing to regional disparities include differences in government policy, geographical conditions, and levels of infrastructure development, but other related factors include the amount of foreign investment, whether there is an export oriented strategy, the development level of local businesses, the number of state-owned enterprises, and especially the differences in the education levels of residents. According to government-published data, the central government’s rate of disbursement of education expenses to each region (the percentage of education expenses to distributed to each region out of total education expenditures) in 2005 was 56% to the eastern region and a mere 18.6% to the western region. It is clear that the central government’s policies are slanted toward the eastern region. Moreover, according to recent research on the relationship between income disparities and educational disparities, China’s educational disparity among regions appears to be an important factor of regional disparity (Xue, 2009).

1.2.3. Income Inequality of Residents The Gini coefficient is often used as an indicator to express degrees of income inequality. However, because of the problems of income statistics and income reporting systems in China preventing household income surveys from being administered to all residents nationwide, the Gini coefficient for the whole country cannot be derived directly. The Urban Social and Economic Survey Department and Rural Social and Economic Survey Department of National Bureau of Statistics of China, has been taking periodic sampling surveys on the same items for urban and rural residents every year since 1985, but the data for urban and rural areas are separate and the original source data are not released to the public, so a chronological national Gini coefficient cannot be obtained from these surveys. Although the World Bank estimated China’s national Gini coefficient using data from the National Bureau of Statistics of China, its actual figures only go to 2004 (Ravallion and Chen, 2007). Then, in 2004, the Employment and Income Distribution Department of the National Development and Reform Commission (NDRC) began to release Gini coefficients

b1363-ch01

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch01

Growth and Inequality in China

13

by urban and rural areas dating from 1978 with the cooperation of the Nation Bureau of Statistics, however, neither of these estimates included adequate source data or explanation of estimation methods. In particular, even though the above were governmental estimates, they do not necessarily match with international standards in terms of such aspects as the definition of income and the methodology of surveying, so it is difficult to say that these data are sufficiently reliable. Moreover, since there are issues of high income earners concealing their income and corruption among government bureaucrats, it is difficult to accurately measure inequalities in income distribution. Therefore, starting in the 1988, some international organizations and economists from China and other countries conducted household surveys in an effort to ascertain the actual levels of income inequality. Among these, an international research group composed of scholars from the United States, the United Kingdom, China, Japan, Australia, and other countries used Chines Social Science Academy’s (CASS) large-scale household survey (China Household Income Project, CHIP), which resulted in highly reliable numerical estimates that are often used. 1.2.3.1. How Large is the Gini Coefficient? Table 1.3 and Fig. 1.3 summarize the above estimates in tabular form. Whether we look at the estimates of the World Bank staff, the estimates of a Chinese government organization, or CASS’s own household surveys, we can see that China’s national Gini coefficient rose dramatically, from 0.25 in 1978 to 0.46 in 2002. Furthermore, the Gini coefficient in 2007 was even higher, at 0.478, recording a new all-time high in the CHIP survey (Li Shi, China Daily, 12 June 2010). Table 1.3.

Changes of the Gini Coefficient.

Urban Survey Year 1988 1995 2002 2007 2008

Rural

Entire country

Gini coefficient

Rate of Change (%)

Gini coefficient

Rate of Change (%)

Gini coefficient

Rate of Change (%)

0.233 0.286 0.319

22.7 11.7

0.338 0.416 0.366

23.1 12

0.382 0.445 0.454 0.478

16.5 2.0 5.3

0.34

0.38

Sources: The Gini coefficient for 1988 and 1995, 1995 and 2002 are calculated from CHIP data, figure of 2007 is cited from Li Shi (2010); the Gini of 2008 is taken from government-issued figures (Zhang 2010: 239).

June 5, 2012

11:23

14

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

Fig. 1.3. China’s Kuznets Curves (1978–2009). Source: National data 1980–2004 (M. Ravallion and Chen, 2007), 2005–07, rural and urban data (Zhang Dongsheng, 2009).

Urban income disparity: The Gini coefficient for urban resident was 0.23 as of 1988, but since then it has risen, reaching 0.29 in 1995 and then 0.32 in 2002 (Xue, 2008). According to government’ data, the Gini coefficient was 0.34 in 2008, clearly indicating a deteriorating income distribution in urban areas as the years go by. Rural income disparity: The Gini coefficient for rural resident rose dramatically from 1988 to 1995, then improved slightly (4%) from 1995 to 2002 owing to rising prices of agricultural goods and the effect of policy measures to combat the three rural issues, but in 2008 the figure again worsened to reach a new high of 0.38. Contrasting the disparities in urban areas with those in rural areas, we could see that the Gini coefficient is higher in rural areas than urban areas, showing that income distribution is more unequal in rural areas. Factors include the large number of extremely impoverished people in rural areas, the ultra-high incomes of the managers of rural farming businesses, and corrupt officials in rural governments. Another cause is the change in the agricultural income structure that farmers who engage in agriculture are earning less from such activities, while rural industry, management and other non-agricultural income is on the rise (Zhang, 2010).

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch01

Growth and Inequality in China

15

1.2.3.2. How Serious is China’s Inequality? The above analysis clearly shows that China’s income gap is widening. However, the questions arise of how large China’s income disparities are from an international perspective, and to what level they will expand. Table 1.4 is an international comparison of income inequality using datasets of the World Bank. Although we cannot say that these data are absolutely synchronized, due to differences in survey years, definition of income and contents of the survey (for example, the Gini coefficient for India is estimated based on consumption expenditure data), and survey methods employed, the table roughly reflects the income disparities in each country. According to the table, the fact that China’s Gini coefficient is ranked after such worst examples of income distribution as South Africa, Brazil, and the Philippines, and higher than Thailand, Indonesia, India, and Bangladesh, reveals the seriousness of China’s real state of affairs. This is one reason that China is attracting the attention of the international community. Table 1.4.

Country/Region South Africa Brazil Mexico China Philippines China-Hong Kong Singapore Thailand United States Indonesia Russia India Japan Korea Germany

International Comparison of Inequality.

Gini coefficient

Ratio of Top 20% to Bottom 20%

Year

0.58 0.55 0.48 0.47 0.44 0.43 0.43 0.42 0.41 0.39 0.38 0.37 0.32 0.32 0.28

20.5 19.4 11.5 8.3 9.0 9.6 9.7 8.1 8.4 6.6 6.9 5.6 3.4 4.7 4.3

2000 2007 2006 2007 2006 2006 2006 2004 2000 2005 2005 2005 2003 1998 2000

Sources: WDI Database, the World Bank, various years. Figures for China come from Li Shi (2010); figures for Japan come from Tetsuo Fukawa (2006).

1.3. China’s Kuznets Curve How large will China’s disparities become? When income distribution can be equalized? Is it possible for China to get equalized through a gradual

June 5, 2012

11:23

16

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

growth path as it is said in economic doctrine? These are the issues that international society is highly concerned about. Here we will examine these questions by adapting income Kuznets curve based on chronological data. Based on the Kuznets hypothesis which analyzed the relationship between economic growth and income distribution in a long term, we can identify that a “polarization” occurs in the early stages of economic growth between the people who find clever ways to increase their assets and income by riding the wave of economic growth, and those who are late to ride that wave, which causes the level of inequality to rise up. As the economy eventually matures and the average income rises, along with progressive income taxes and other policies, the blessings of growth trickle-down to the lowincome earners, thereby decreasing the level of inequality. This rule of thumb can also be verified using data by cross country analysis. With a single country’s average income on the horizontal axis and its Gini coefficient on the vertical axis, an inverted U-shape appears. This is called the Kuznets curve. According to the Kuznets hypothesis, changes of the Gini coefficient in a country can be divided into three stages: Rising stage, peak stage, and falling stage. For example, Japan and most developed countries of Europe have achieved a relative equal societies with high incomes and low Gini coefficients, so they are in the third stage. In contrast, many developing countries are in the second stage. Can the Kuznets hypothesis be applied to China? This is a matter of interest to many scholars. In contrast with the long-term analysis of the Kuznets curve, China’s development history is still young and there is a shortage of the time series data. Although some scholars claim that comprehensive research along these lines is premature for these reasons, in order to ascertain future trends, we will here attempt to form a Kuznets curves for China using data released by the World Bank and the Chinese government. In Fig. 1.3, we do not yet see a full inverted U-shaped curve, because only a little over 30 years have passed since the economic reforms and opening up of the country from 1978, and there are limits on the data. However, looking at the trend we can obtain the following results. First, since the reforms and opening-up of 1978, China’s Kuznets curve is trending upward, indisputably indicating that the income gap is widening. Second, as China enters its rapid economic growth phase its Kuznets curve is passing the “deterioration stage” and approaching the “peak stage.” However, looking at the manner in which the Gini coefficient is

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in China

b1363-ch01

17

rising, the Kuznets curves appear not to have peaked yet. In view of such trends, income disparities in China are expected to get even worse as the country approaches the peak of its Kuznets curve. 1.4. Conclusions Some economic and political implications obtained from the analyses up to this point will be indicated. 1.4.1. Economic Implications The experience of China raises a question in dilemma: Should we take income disparity as an engine of high economic growth or as an obstacle to social development? Regarding measures to shrink disparities, the Chinese government is working on urban employment measures as well as measures to raise farmers’ income and reduce their tax burden. It is also conducting the western region development strategy, the phased repeal of preferential measures for the coastal region and attracting investment to the interior regions, and shifting from a development pattern focused on foreign demand to one focused on domestic demand. Whether or not these policies will pan out must be watched with diligent attention. Moreover, there is tremendous significance in the fact that China is consulting the income redistribution schemes in Japan and other developed countries, particularly their progressive income tax schemes. China is currently gradually introducing the progressive income tax scheme, yet fears remain strong that this will diminish the motivation of high income earners, who are after wealth. But for China, which has made shared prosperity its goal, there is value in paying this short-term price in order to seek sustainable, long-term economic growth. The economic history of developed country shows us that urban industry grows rapidly during the initial stages of industrialization. But since it takes a long time for the effects from this to permeate out to rural areas, this tends to give rise to urban–rural disparities and regional disparities. Moreover, the progress of moves towards a market economy has inevitably produced both rich people and poor people. In China there are still deepseated opinions to the effect that the income disparities are the desired result of economic reforms, that they carry great significance in terms of improving the willingness to work and the introduction of foreign capital,

June 5, 2012

11:23

9in x 6in

18

Growth and Inequality: An International. . .

J. J. Xue

and that a too fast reduction of such disparities would have a negative impact on economic growth.

1.4.2. Political Risk In China there is the proverb “Inequality, rather than want, is the cause of trouble.” Long ago in the Chinese history, rebellion movements arose time and time again, including those against the corruption of the bureaucracy, oppressive taxes, and farmers’ uprisings stemming from their poverty. “Equal wealth between the poor and the rich” was the slogan for mobilizing the populace back in those days. Holding this past history up as a mirror reveals the concerns that unless such disparities can be eliminated, they will act as a contributing factor to social and political destabilization. The riots which occurred in Lhasa, the capital of the Tibet Autonomous Region, in March 2008 and the large-scale insurrection which occurred in Urumqui in the Xinjiang Uygur Autonomous Region in July 2009, the migrant workers’ demonstration in Guangdong province and some coastal areas in May 2011, are not simply ethnic conflicts independence movements and a mere request for wage rise. The regional disparities and income disparities between ethnic groups which have persisted for a long period of time are also thought to be important contributing factors behind them. Furthermore, there have been demonstrations and strikes by workers at foreign companies — mainly Japanese companies in the coastal regions — that started in the fall of 2009 and have persisted down to the present. While there are certainly problems with the management systems at and the responses by these companies, these could also be characterized as an outburst of dissatisfaction with urban–rural disparities, regional disparities, and disparities among residents. There is no denying the fact that the risk of internal problems shifting out into the international community is becoming more pronounced. A series of editorials that ran in Xinhua news, a state-run newspaper agency, in May 2010 offered the strong warning that, “China’s Gini coefficient is already over 0.5, which poses the possibility of ushering in social unrest” (Economics Report, May 21 2010). Amidst such circumstances, the feeling is that reducing the urban–rural disparities and regional disparities as well as rectifying the gap between the rich and poor within rural and urban areas is extremely important for China, which is after continuous economic development. China’s characteristics of economic growth is a growth that started from the changeover to reform and opening-up policies, in other words,

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in China

b1363-ch01

19

it is development that came from the forceful leadership of the government. Because of this, the occurrence of problems like poverty, unemployment, and income inequality are mainly closely linked to government policy. For this reason, unlike with other countries it will be difficult to eliminate income inequality and China’s other problems naturally by means of economic development alone. The thinking is that these should not be eliminated by forceful policies on the part of the government. However, the Chinese government has adopted a number of income policies thus far, but these have been severely limited in their outcomes and are nowhere near to eliminating the problems from inequality. Therefore, the suggestion is that China will perform an international comparison and propose better policies.

References Cai, F. (Ed.) (2009). Report on China’s Population and Labor No. 7. Social Sciences Academic Press (China). CIA (Central Intelligence Agency). In The World Factbook. https://www.cia. gov/library/publications/the-world-factbook/. IMF, (2010). “World Economic Outlook.” Li, S., Sicular, T. and Gustafsson, G. (Ed.) (2008). Third Study on Income Redistribution to Chinese Residents. Beijing: Normal University Press. Li, S. and Yue, X.M. (2004). “Study on Income Disparities between Urban and Rural China.” Caijing Magazine, March edition. Ravallion, M. and Chen, S. (2007). “China’s (Uneven) Progress against Poverty.” Journal of Development Economics, 82(1), 1–42. Sicular, T., Yue, X.M., Gustafsson, G. and Li, S. (2007). “The Urban–Rural Income Gap and Inequality in China.” Review of Income and Wealth, 53(1), U.K. Tetsuo, F. (2006). “Income Distribution in Japan based on IRS 1987–2002.” The Japanese Journal of Social Security Policy, 5(1). World Bank, (2006). Annual Review of Development Effectiveness, http://web. worldbank.org/. World Bank, (2009). World Development Indicators 2009, World Bank, Washington D.C. World Bank, (auth.). Shiratori, M. (tran. superv.) (1994). The East Asian Miracle. Toyo Keizai. Xue, J.J., Sonoda, T. and Arayama, Y. (2008). Inequality in China. Nippon Hyoronsha. Xue, J.J. and Wei, Z. (2003). “Unemployment, Poverty and Income Disparity in Urban China.” Asian Economic Journal, 17(4). Xue, J.J. and Minami, R. (1999).“Transforming Labor Markets.” The Tribulations of Becoming a Major Power: China’s Economy at the Turning Point. Nippon Hyoronsha.

June 5, 2012

11:23

20

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

Xue, J.J. and Knight, J. (2006). “How High is Urban Unemployment in China?” Journal of Chinese Economic and Business Studies, 4(2), 91–107. Xue, J.J. and Yang, D. (2010). “Labor Market Development in China and its Implications to Income Inequality.” Paper Presentation at Workshop of International Comparison of Income Inequality, Kyoto, Japan. Xue, J.J. and Gao, X.C. (2010). “Can Returns to Education Reduce Inequality in China?” China Population Science, No. 4. Zhang, D.S. et al. (2010). 2009 Annual Report on Income Distribution to Chinese Residents. Economic Science Press.

b1363-ch01

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 2 GROWTH AND INEQUALITY IN HONG KONG Shaoguang Wang∗ and Yin Xia†

2.1. Introduction The World Bank has appraised the eight high performing Asian economies (HPAEs) — Japan, Hong Kong, the Republic of Korea (Korea hereafter), Singapore, Taiwan, Indonesia, Malaysia, and Thailand — as being the “East Asian economic miracle.” It has been said that during the period 1960–1990, these economies maintained rapid and sustained economic growth combined with “low and declining levels of income inequality.”1 However, the case of Hong Kong seems to contradict this statement in that, throughout the decades of the post-war era, Hong Kong has persisted with a much higher income inequality level than other newly industrialized economies (NIEs) in East Asia. On the whole, it is true that Hong Kong has enjoyed the fruits of economic development since the post-war era. From 1950 to 2008, the annual gross domestic product (GDP) growth rate was 4.8% (Fig. 2.1).2 In 1950, the per capita GDP of Hong Kong was much lower than that of any Western country, while by 2009, Hong Kong had become one of the wealthiest economies in the world. The statistics from the International Monetary Fund (IMF) and the World Bank show that Hong Kong’s per capita GDP ∗ Chair Professor, at Department of Government and Public Administration, Chinese University of Hong Kong. † Assistant Professor, School of Government, Sun Yat-sen University, China. 1 World Bank (1993). Figures 1 and 3 in this report demonstrate, in a very special way, that Hong Kong is similar to Korea, Taiwan and Singapore in its “rapid economic growth with relatively low income inequality.” 2 “Historical Statistics of the World Economy”, Angus Maddison, accessed on September 3, 2010. http://www.ggdc.net/maddison/Historical Statistics/vertical-file 02-2010.xls.

21

b1363-ch02

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

22

35000 30000 25000 20000 15000 10000 5000 0

S. Korea Fig. 2.1.

Taiwan

Hong Kong

Singapore

Per Capita GDP of Hong Kong, Singapore, Korea and Taiwan, 1950–2008.

Source: “Historical Statistics of the World Economy”, Angus Maddison, Accessed on September 3 2010. http://www.ggdc.net/maddison/Historical Statistics/vertical-file 02-2010.xls.

by expenditure is now among the 10 highest in the world.3 Hong Kong’s outstanding economic growth has lifted the overall income level of households, meaning that households in Hong Kong were generally better off now as compared with the income conditions three decades ago. For instance, the monthly median household income for 1971 was just $708 HKD, but by 2006, it had risen to $17,250 HKD. By discounting the increases in consumer prices, the real median household income may have increased by as much as 240% from 1971 to 2006. In this respect, the economic development of Hong Kong could indeed be called an economic miracle.

3 “World

Economic Outlook 2010”, the International Monetary Fund, accessed on September 3, 2010. http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/ WEO Apr2010all.xls; “Gross Domestic Product 2009”, the World Bank, accessed September 3, 2010. http://siteresources.worldbank.org/DATASTATISTICS/Resources/ GDP PPP.pdf; “Population 2009”, the World Bank, accessed on September 3, 2010. http://siteresources.worldbank.org/DATASTATISTICS/Resources/POP.pdf.

b1363-ch02

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

23

2.2. Hong Kong — An Economic Miracle with Unequal However, the increasing trend in household income does not necessarily indicate an equal income distribution across households. While having enjoyed rapid economic growth, Hong Kong, paradoxically, also ranks among the most unequal of economies. In contrast to the World Bank’s conclusion mentioned above, the income distribution in Hong Kong has been quite unequal since the very beginning of the post-war era. In 1957, the Gini coefficient measured for household income inequality had already reached 0.48.4 Although we do not have reliable data before 1957, we can infer from the high unemployment rate, ranging from 15% to 17%, that the level of inequality has also been very high during this period.5 We can see from Fig. 2.2 that income distribution in Hong Kong improved only in the years between 1966 and 1976. Since the middle 1970s, the Gini coefficient 0.54 0.52 0.5 0.48 0.46 0.44 0.42 0.4 1957 1963 1966 1971 1973 1976 1981 1986 1991 1996 2001 2006 Fig. 2.2. The Gini Coefficients of Household Income Distribution in Hong Kong, 1957–2006.6 Source: Data for 1957–1981 from L. Chau, “Economic Growth and Income Distribution of Hong Kong since the early 1950’s,” Department of Economics Discussion Paper No. 38 (Hong Kong: The University of Hong Kong, 1984): pp. 1–36; Data after 1981 were compiled from the statistics released by the Census and Statistics Department of HKSAR. 4 The Gini coefficient is an indicator measuring the level of income inequality. It ranges from 0 to 1, where 0 means absolute equality (all people have the same income) and 1 refers to absolute inequality (one person possess all the wealth). The higher the value of the Gini coefficient, the worse is the income inequality. 5 Tzong-biau Lin (1985). 6 “Income” in Fig. 2.2 refers to net income or market income.

June 5, 2012

11:23

24

Fig. 2.3.

9in x 6in

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

Distribution of Total Household Income by Deciles, 1966 to 2006 (Unit: %).

Source: Data for 1966–1991 from Simon Xiao-bin Zhao and Zhang Li (2005); data for 1996–2006 from Dominic K. T. Leung (2009). http://www.ancsdaap.org/cencon2009/ Papers/Hong%20Kong/HongKong.D.Leung.pdf.

has been increasing continuously, with a stable rise before 1990 and a sudden upsurge from that year onwards. The largest increases in inequality were concentrated in the late 1990s, continuing into the early 2000s and reaching the highest level in 2006, with a ratio of 0.533. This trend can also be observed in Fig. 2.3, which shows the percentage share that households from different income groups have of the total income. Obviously, from 1966 to 2006, the income share of the low-income households increased only before 1976. After 1976, it can be observed that (1) the lower the income group, the greater the reduction of its share and (2) the higher the decile of the income group, the greater the realized increase in its share of the total income. In 2006, the richest 10% received 41.4% of total income, while the poorest 10% claimed less than 1%. Even the sum of the total income from the bottom 50% of households only constituted 16% of the overall income. In other words, from the middle of the 1970s onwards, rich households, especially very rich households, were able to generate income much faster than lower-income households. To put

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

25

it another way, the rate of income increase in the low-income households lagged far behind high-income households. During the past three decades when neoliberalism prevailed, the income inequality of many countries has increased. Still, the widening of the gap between the rich and poor in Hong Kong seems to be more serious than that of many other countries. A comparison between eight Southeast and East Asian economies, including Hong Kong, Korea, Singapore, Taiwan, Indonesia, Malaysia, Philippines and Thailand7 shows that in the last two decades of the 20th century, Hong Kong ranked first place among the four East Asian little dragons in terms of Gini coefficient, and only slightly better than among the eight NIEs.8 The serious situation of Hong Kong becomes more apparent when it is compared with some OECD countries that are at a similar or even lower level of economic development in terms of per capita GDP. Figure 2.4 shows that among the 29 advanced economies whose per capita income exceeds $20,000 USD, there are nine countries whose Gini coefficients are smaller than 0.3 (among which the north European countries are around 0.25) and 17 countries ranging from 0.3 to 0.4. Only three economies’ Gini coefficients are higher than 0.4, namely the United States of America, Singapore and Hong Kong, with Hong Kong having the highest Gini coefficient at 0.475. One thing that needs to be pointed out is that the concept of “income” in Fig. 2.2 is different from that in Fig. 2.4, where the former refers to “market income” and the latter refers to “final income” or “disposable income” — the post-tax and postsocial transfer income. Figure 2.4 indicates that, even taking into account the effects of the government’s social redistribution, the income inequality in Hong Kong is still very serious. Of the 142 economies in Fig. 2.4, which are arranged in ascending order in terms of their Gini coefficients, Hong Kong ranks 32nd. The top 31 economies are almost all African and South American countries. Thus we see two contradictory sides of Hong Kong. On the one hand, Hong Kong is among the most advanced economies in the world in terms of per capita GNP. On the other hand, it is among the worst-performing economies in terms of income inequality. What is worse, the rich–poor gap has been widening rather than narrowing in recent years, and there is no sign of any improvement in the near future. This paper tries to explain 7 These economies are all “HPAEs” described by the World Bank’s 1993 report, except Philippines. 8 Ragayah Haji Mat Zin (2005).

9in x 6in

11:23

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

26

0.8 0.7 0.6 Gini (UNDP, 2009)

June 5, 2012

0.5

HK, 0.475

0.4 0.3 0.2 0.1 0 0

10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 Per Capita GDP (PPP, IMF, 2009) Fig. 2.4.

Economic Growth and Income Distribution.

Source: Data of per capita GDP from “World Economic Outlook 2010”, the International Monetary Fund, accessed September 3, 2010. http://www.imf.org/external/pubs/ft/ weo/2010/01/weodata/WEOApr2010all.xls; Gini coefficients from “Human Development Report 2009: Gini index”, UNDP, accessed on September 3, 2010. http://hdrstats. undp.org/en/indicators/161.html.

why there is such a contradiction between high-economic growth and highincome inequality in Hong Kong. Income inequality has attracted research interest from scholars of various disciplines. Previous studies have provided ample evidence for the factors that are likely to be related to changes in inequality. Conventional wisdom suggests that increasing income inequality is attributable to changes in the characteristics of the labor market (including the structures of wages and labor force features),9 economic structural changes,10 and changes in family and demographic structures.11 Research on income inequality in Hong Kong basically falls within the broad theoretical framework of inequality studies in general, but such study 9 See Frank Levy and Richard Murnane (1992); Lawrence Katz and Kevin Murphy (1992); Mark Partridge et al. (1996). 10 See Azizur Rahman Khan et al. (1999). 11 See Sheldon Danziger and Peter Gottschalk (1995).

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

27

is underdeveloped compared with its Western counterparts. We can identify several theoretical approaches from previous studies on inequality in Hong Kong, whose deficiencies are easy to point out. Zhao and Zhang,12 following the “economic structural change” approach, purport that the relocation of manufacturing operations, coupled with ineffective social policies for income redistribution under globalization, have worsened income inequality in Hong Kong.13 Simply taking income inequality as an inevitable byproduct of the process of economic restructuring, Zhao and Zhang placed their emphasis more on the absence of proper redistributive policies from the government, which resulted in a small but extremely wealthy class of the “‘new rich”’ and simultaneously created a large population of “‘working poor”. Another recent study analyzes household income inequality in Hong Kong between 1991 and 2001 and attributes the reasons to an increase in labor force participation by wives and a higher correlation between the incomes of husbands and wives.14 Previous studies have provided a broad picture of the overall level of, trends for, and sources of inequality in Hong Kong, with abundant data and insightful analysis. However, there are several problems with these discussions. First, there is a mixed usage of household inequality and individual earnings inequality. For example, Zhao and Zhang (2005) have relied on the Gini coefficient as measured for households to explain the sources of individual earnings inequality, without specifying the relationship between these two forms of inequality. Although household inequality and earnings inequality are in fact closely related to each other, examining their sources may require a different theoretical framework and variables. With regard to the sources of household inequality, besides the macro socioeconomic and demographic dimensions that are commonly examined in order to understand individual earnings inequality, the family structure is an inevitable dimension more directly related to the changes in household income disparity. Interchangeable use of earning inequality and household inequality may lead to the neglect of important variables that are closely related to the latter form of inequality but not to the former. Second, closely related to the first shortcoming, previous studies did not clearly separate the determinants of individual earnings inequality (which 12 Simon

X. B. Zhao and Li Zhang (2005). and Zhang (2005) also pointed out that the absence of effective government’s redistributive policies has resulted in a small super rich group and a large low-income group. 14 Stephen Wing-kai Chiu (2005). 13 Zhao

June 5, 2012

11:23

28

9in x 6in

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

arises from labor market conditions) from the determinants of family or household income inequality (which arises from the household formation decisions of income producers, as well as earnings inequality among those producers). The analyses used in the studies were not able to sort out the influences of economic, social and demographic structures on the earnings distribution versus the same influences on the family or household income distribution. The dependent variable, the distributional measure, was a Gini coefficient computed from household income (rather than individual earnings), although the independent variables (such as industry and occupation of employment) were labor market characteristics that directly influence the earnings distribution and, therefore, affect the household income distribution indirectly rather than directly. Variables that influence the household income distribution, specifically those that reflect household formation patterns (such as a female head of household, household size and multiple earners or no earners in the household) were frequently excluded. There is less evidence on how earnings equality translates into household income equality, and the evidence that exists does not clearly demonstrate the connections. Third, almost all the discussions on income inequality in Hong Kong focus on the social and economic variables, seldom have any studies examined the sources of inequality from the political dimension. While the main features and processes of social and economic structural changes may be similar across different economies under the process of globalization, the particular political settings of different regions play an extremely important role in shaping the level and trend of inequality. This factor is particularly important in Hong Kong, given Hong Kong’s special circumstances as a former colony with a minimalist government and a low level of unionization. This chapter examines the main sources of household income inequality in Hong Kong during the past four decades. Section two discusses the main factors that affect the distribution of net market income. We first explore the traits and trends of several important factors that might affect the market income distribution, including the industrial structure, demographic structure, labor force structure and household composition. We then determine the most important explanatory factors among these to explain the rising household income inequality using regression analysis. In section three, we turn to investigating the impact of the government’s redistributive policies on the inequality of disposable income. By comparing Hong Kong with other economies, we conclude that the government’s insufficient redistributive measures are the most important factor leading to the unequal distribution of final disposable income.

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

29

2.3. Why Hong Kong’s Income Distribution is so Unequal? Since the start of the modern era, because of its scarcity of natural resources, as well as its special geographic location, the economy of Hong Kong has been highly reliant on international trade. Thus, Hong Kong is well-known for its service sectors related to import and export, such as shipping, portage, storage, transportation, wholesale, hospitality, telecommunications, banking and insurance. The main industry in the early era was ship-building, which is also closely related to international trade. Only in the 1950s, when Western countries forbade trade with mainland China, causing Hong Kong to stop relying on the re-export trade, did manufacturing industries develop in Hong Kong. As shown in Fig. 2.5, during the two decades before 1980, the manufacturing rose to contribute more than one fourth of the overall GDP of the economy, which has been called the golden era of manufacturing in Hong Kong. Nonetheless, the GDP contribution of the service sector was close to 60% at the end of the 1950s, much higher than for other economies in the world at that time. A large service sector has been consistently seen as the most important feature of Hong Kong economy. Starting in the 1980s, the industrial structure of Hong Kong transformed in a special way. Generally, there are two periods of industrial restructuring for a given economy. During the first phase, the industrial or manufacturing upgrade brought by a technological revolution transforms the main industries from being labor-intensive to being capital or technology-intensive; this upgrade then greatly increases productive efficiency, which stimulates the demand and supply of a new service sector and promotes its development and upgrade. In other words, the preliminary condition of industrial restructuring is essentially the upgrading of manufacturing itself. Since the end of the 1970s and the early 1980s, the industrial restructuring of Korea, Taiwan and Singapore, the other three of the “Four East Asian Tigers”, have all gone through this process. The manufacturing sectors of these regions have all gone through the process of transforming from labor-intensive to capital and technology intensive industry. To take Singapore (which is most similar to Hong Kong) as an example, in 1960, when the GDP contribution of manufacturing for Hong Kong had reached 31.7%, that of Singapore was only 11.2%.15 In the 15 Nah

Seok Ling (2006). http://unpan1.un.org/intradoc/groups/public/documents/ APCITY/UNPAN024567.pdf.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

30

0% 1959

20%

40%

60%

Services

Manufacturing

80%

100%

1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 Others

Fig. 2.5. Gross Domestic Product (GDP) by Economic Activity — Percentage Contribution to GDP at Current Factor Cost, 1959–2008 (Unit: %). Source: Data for 1959–1969 from Laurence C. Chau, “Estimates of Hong Kong’s Gross Domestic Product, 1959–1969,” Hong Kong Economic Papers, Vol. 1972, No. 7 (September 1972); data for 1970–1979 comes from “10-Sector Database, July 2007”, Groningen Growth and Development Centre, accessed on September 3, 2010. http:// www.ggdc.net; data for 1980–2008 comes from the Census and Statistics Department of the HKSAR, accessed on September 3, 2010. http://www.censtatd.gov.hk/ showtableexcel2.jsp?tableID=036&charsetID=2.

1980s and 1990s, when the manufacturing sector in Hong Kong was shrinking rapidly, Singapore managed to set up a manufacturing system coupled with electronics, oil refining and the machinery industry (mainly the ship-building industry) with foreign capital and advanced technology. Singapore ultimately realized the upgrade of its industry structure and set up a comprehensive and well-balanced industrial structure that is capital and technology-intensive, centered on the heavy chemical industry. Throughout

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

31

the whole process of industrial restructuring, Singapore’s manufacturing sector has played a very important role, and it has been the major engine of economic growth. However, during the process of industrial restructuring in Hong Kong, the manufacturing sector did not go through the process of industrial upgrading. In the 1970s and 1980s, there were signs of impending upgrades; the rise of the electronics sector is proof of this. However, because of the lack of support from the government, as well as the impact of high land costs, high wages and shortages in the labor force, the manufacturing sector in Hong Kong has never transformed itself completely. Instead, beginning in the mid-1980s, the manufacturing factories almost all moved suddenly to the neighboring Pearl River Delta in order to lower costs and maintain profit margins. From 1980 to 2008, the percentage of GDP contributed by the manufacturing sector to the overall GDP declined from 22.8% to 2.5% (Fig. 2.5). The manufacturing sector almost died out in Hong Kong, resulting in the danger of the “industrial hollowing-out” of the Hong Kong economy. The situation of Hong Kong is quite rare among the world’s economic system, and constitutes a sharp contrast with Singapore, which is also a city-state economy. Today, the manufacturing sector of Singapore still contributes around 20% to the overall GDP of the local economy.16 While the manufacturing sector of Hong Kong was shrinking, the service sector was growing at a rapid rate. The GDP contribution of the service sector has grown from 68.3% in 1980 to 92.42% in 2008 (Fig. 2.5). The shrinking manufacturing sector and the expanding service sector have changed the whole economic structure of Hong Kong. At present, it has become an economy relying almost exclusively on the service sector. Accompanying the dramatic changes in the industrial structure, the labor force structure has also undergone significant changes. First, the labor force in the service sector has increased significantly. Before the 1970s, the percentage of the total labor force in the service sector never exceeded 45% 17 and it never exceeded 50% before the 1980s.18 However, in 1983, the service sector labor force percentage exceeded 50% for the first time, and has continuously increased from then to the present. It reached 82.7% in 2006. In contrast, the percentage of the labor force engaged in the 16 See Table 5.3 in Singapore Department of Statistics, Yearbook of Statistics Singapore, 2010: 68. Accessed on September 3, 2010. http://www.singstat.gov.sg/pubn/reference/ yos10/yos2010.pdf. 17 Wing Suen (1995). 18 See “10-Sector Database, July 2007”, Groningen Growth and Development Centre, accessed on September 3, 2010. http://www.ggdc.net.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

32

manufacturing sector was always above the 40% level before the 1980s, after which, accompanying the increase of the service sector labor force, it began to decline significantly to 9.7% in 2006. Today, the service sector has almost become the only industrial sector in Hong Kong. During the past two decades, no other economy has experienced such a dramatic decline in the percentage of its labor force engaged in the manufacturing sector, nor has any other economy experienced such a significant increase in its service sector labor force.19 45 41.2

40

Manufacturing 35.8

35 30

Wholesales, retailed Trade etc.

28.3

25 20

Transportation, Communication etc.

18.9

15 12.3

10

9.7

5

Finance, insurance, real estate etc. Community, social & personal services

0 1981 Fig. 2.6.

1986

1991

1996

2001

2006

Labor Force Participation Rate by Industry, 1981–2006 (Unit: %).

Source: Compiled with data from the Census and Statistics Department of the HKSAR.

In addition to changes in the services and manufacturing sectors, more and more of the potential female labor force are engaged in various industrial sectors. In the past 20 years, there has been quite an interesting phenomenon in Hong Kong. On the one hand, the male labor force participation rate declined continuously from 81.3% in 1983 to 69% in 2009, while its female counterpart increased from 47.5% to 53% during the same period. The data indicates an equalizing trend for male and female labor force participation rates.20 Previous studies have shown that the increase in the female labor force will worsen income inequality as women generally earn less than men. In order to test this hypothesis, we included the 19 Hyunsub

Kum (2008). relevant data, see the Census and Statistics Department of HKSAR, accessed on September 3, 2010. http://www.censtatd.gov.hk/showtableexcel2.jsp?tableID=007. 20 For

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

33

female labor force participation rate as one of the explanatory variables while doing statistic analysis to figure out the most important factor leading to income inequality in Hong Kong.21 The third important trend is that the demographic structure of Hong Kong is becoming more and more complicated. One of the most significant changes in recent years is the size of the elderly population. According to the statistical data (Fig. 2.6) from the Census and Statistics Department of the HKSAR, the proportion of elderly people, aged 65 years old or older, of the total population has increased from 2.8% in 1961 to 12.4% in 2006.22 The increasing population of elderly people has also led to an increase in the number of households with elderly members. As most of the elderly people do not participate in any kind of economic activities and receive very little income, thus households containing elderly people are generally located at the bottom of the overall household income structure. Compared with people from other age groups, more elderly people are residing in lowincome households.23 In 2006, 35.9% of the elderly people were residing in low-income households.24 Therefore, the increasing proportion of elderly people as a percentage of the total population might increase the income disparity among different households. The factors mentioned above might all have effects on the income distribution among different households. What, then, is the most important factor that has led to the current state of income disparity in Hong Kong? In order to get an answer, we must take into account all of the key variables that we think are important based on previous studies and use statistical analysis to sort out the most important factors among them. In our regression model, we used the Gini coefficient, measured for the household income distribution for the period from 1961 to 2006, as the dependent variable. Following from our previous discussion, we focused on

21 In fact, the ideal indicator to investigate the effects of increasing female labor force on household income distribution is the percentage change of the households containing working females to the total households. However, the systematic and long-time series data of this indicator is unavailable, thus, in our statistical analysis, we choose the female labor force participation rate as an interchangeable indicator. 22 The Census and Statistics Department of the HKSAR, A Profile of Elder Persons (2006): 3. 23 “Poverty in Hong Kong”, the Hong Kong Council of Social Service, accessed on September 3, 2010. http://www.hkcss.org.hk/cb4/ecp/pov rate 91-05.pdf. 24 “Percentage of Persons Aged 65 and Over in Low Income Households”, Social Indicators of Hong Kong, accessed on September 3, 2010. http://www.socialindicators. org.hk/chi/indicators/elderly/31.13.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

34

four independent variables, namely, the percentage contribution to GDP of the service sector versus that of the manufacturing sector (Serv/Manu), the female labor force participation rate (Female), the ratio of the elderly people to the whole population (Elder), and household formation (Hsize).25 The first independent variable might directly influence the individual earnings inequality among the working population, which would indirectly influence the income distribution among households, whilst the latter three independent variables might all have direct effects on household income distribution. Before the regression analysis, we did a correlation analysis for the five variables. The results (Table 2.1) show that all the variables are highly correlated with each other, with the total correlation coefficients ranging from 0.55 to 0.98. An extremely high correlation coefficient means that the two correlated variables actually belong to the same explanatory variable. Taking this into consideration, we made a choice between two independent variables — Elder and Hsize — whose correlation coefficient is as high as 0.98. In order to choose a more favorable variable from these two highly correlated ones, we put all of the independent variables into the regression model on the Gini coefficient, and found that the contribution of variable Elder (β = 2.76) was higher than that of variable Hsize (β = 0.07). Based Table 2.1.

Gini Ser/Manu Elder Female Hsize Note:

Statistical Correlations Between the Key Variables. Giui

Ser/Manu

Elder

Female

1 0.87∗∗∗ 0.72∗∗∗ 0.55∗∗ −0.61∗∗∗

1 0.87∗∗∗ 0.70∗∗∗ −0.78∗∗∗

1 0.85∗∗∗ −0.98∗∗∗

1 −0.87∗∗∗

∗∗∗ Significant

at the 0.001 level.

∗∗ Significant

Hsize

1

at the 0.01 level.

25 Here, family composition refers to the average number of persons within the household, the relationship between the members of a household, and the demographic and economic characteristics of each member in a household. It is said that households with fewer members, more female members, or more poorly-paid individual members tend to have lower overall household income. Here, considering the availability of long timeseries data, we only choose household size (average number inside the household) as the independent variable. As the latter two dimensions of family composition are related to the other two independent variables — the female labor force participation rate and the labor force structure variable, and the data of these two dimensions is hard to collect, in this study we have only chosen “household size” to represent the family composition.

b1363-ch02

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

35

Table 2.2. Regression of Major Factors on Household Income Inequality in Hong Kong. Model 1

Variables

Ser/Manu Elder Female R2 ∆R2

Model 2

β1

t1

β2

t2

0.92

12.88

0.60 0.35

4.07 2.36

0.84

0.86 0.03

Model 3 β3

t3

0.51 4.99 9.83 6.39 −0.48 −5.98 0.94 0.08

on this finding, we kept the variables Serv/Manu, Elder and Female as the independent variables, which were then put into the regression model on the Gini coefficient later on. The results of stepwise regression (Table 2.2) show that the first variable that entered the regression model is variable Serv/Manu, with the R2 reaching 0.84. This means that this variable is the most important variable explaining the current state of the Gini coefficient and predicting its further development. In model 2, variable Elder is included, causing R2 to increase by 0.03. Similarly, in model 3, variable Female is included, with a corresponding 0.08 increase to R2 . In other words, for the period from 1961 to 2006, among the three independent variables, variable Serv/Manu is the most important explanatory and predictive variable for the variation of the Gini coefficient. The higher the value of this variable — that is, the higher the weighting of service sector of the whole industrial structure and the lower the weighting of the manufacturing sector —, the worse is the household income inequality. Also having a significant effect on the Gini coefficient is variable Female; however, this has a negative effect. This means that the increase in the size of the female labor force will to some extent alleviate household income inequality. This finding has somewhat contradicted that of in previous studies. For instance, a study of household income inequality in Hong Kong between 1991 and 2001 concluded that the increase in the number of working females in households was the major factors that lead to the worsening income disparity during that period.26 The aging population represented by variable Elder is positively correlated with the Gini coefficient; however, its effect on the Gini coefficient is much smaller than that of variable Serv/Manu and variable Female. 26 Stephen

Wing-kai Chiu (2005).

June 5, 2012

1991

1996

2001

2006

1986

1991

1996

2001

2006

Manufacturing Construction Wholesale and retail trade, restaurants and hotels Transport, storage, and communication Financing, insurance, real estate and business services Community, social and personal services Others

0.392 0.389 0.378

0.404 0.371 0.364

0.465 0.403 0.410

0.480 0.373 0.447

0.525 0.378 0.474

35.8 6.2 22.3

28.2 6.9 22.5

18.9 8.1 24.9

12.3 7.6 26.2

9.7 6.8 27.2

0.255

0.271

0.338

0.376

0.384

8

9.8

10.9

11.3

11.6

0.419

0.479

0.578

0.573

0.603

6.4

10.6

13.4

16.1

17

0.539

0.526

0.592

0.703

0.769

18.4

19.9

22.3

25.5

26.9

0.776

0.598

0.706

0.751

0.763

2.9

2.1

1.5

1

0.8

27 Manufacturing: Including spinning of cotton and other yarn, weaving and knitting of cotton and other fabrics; bleaching, dyeing, finishing; manufacturing of wearing apparel, knitwear and other made-up textile goods; and manufacturing of carpets, cordages, ropes and twines, food, beverage, tobacco, footwear, leather products, rubber products, plastic products, wood products, printed matters and paper products, metal products, machinery, chemicals, chemical products, glass and pottery. Construction: Including building construction, civil engineering, plumbing, electrical wiring, air-conditioning installing and repair. Wholesale, retail and import/export trades, restaurants and hotels: Including wholesale and retail trade; import and export trade; peddlers; Chinese general brokers; other commercial agents; restaurants; cafes; hotels and rooming houses. Transport, storage and communications: Including land transport, water transport and air transport; services allied to transport; storage and warehousing; and post and telecommunications. Financing, insurance, real estate and business services: Including financing; insurance; real estate; offices of lawyers, accountants, auditors, architects, surveyors and advertising agents and data processing services. Community, social and personal services: Including government services; educational service; medical, dental and other health services; sanitary services; welfare institutions; religious organizations; cinemas and theaters; radio and television broadcasting; libraries and museums; electrical repair shops; automobile repair garages and other household and personal services. Others: Including such industries as “Agriculture and fishing”; “Mining and quarrying”; “Electricity, gas and water” and industrial activities inadequately described or unclassifiable. The within-group income disparity is measured by variance of log earnings.

Growth and Inequality: An International. . .

Source: Hon-kwong Lui, “The Widening Income Dispersion in Hong Kong: 1986–2006” (paper presented at the conference on Social Inequality and Social Mobility, the Chinese University of Hong Kong, Hong Kong, March 14, 2008), accessed September 3, 2010. http://www.cpu.gov.hk/english/documents/conference/20080314%20Lui%20Hon-kwong.pps.

9in x 6in

1986

S. G. Wang and Y. Xia

Industry

11:23

Employee by Industry and Within-Group Income Disparity.27

36

Table 2.3.

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

37

In short, the regression on the Gini coefficient shows that the industrial restructuring — the expanding service sector coupled with a shrinking manufacturing sector — is the most important factor leading to the increasing income disparity.28 We can explain this finding in several ways. First, during the process of industrial restructuring, income disparity within each industrial sector was widening, with only “construction” and the “other” industrial sectors that very few people were engaged in being exceptions (Table 2.3). For the 20 years between 1986 and 2006, the “finance, insurance, land and commercial service” industry and “community, society and personal service” industry both had within-group Gini coefficients that were quite high, increasing 44% and 43% respectively. Some of sections of the services industry, such as “transport, storage and communications,” whose market income inequality rates were not very high, have also become more and more unequal during this 20 year period, with the within-group Gini coefficient increasing to more than 51%. Rising income disparity within each industrial sector has definitely led to an increase in the overall income disparity. Second, during the process of industrial restructuring, the changes within different industrial sectors were not favorable for the alleviation of income inequality in Hong Kong. Industries with comparably narrower income disparities, such as manufacturing, have been shrinking continuously, with a corresponding decreasing number of employees. From Table 2.3 we can see that in 1986, the working population engaged in the manufacturing industry constituted 35.8% of the total working population in Hong Kong. However, in 2006, this ratio had declined to 9.7%, a 26.15% decline for the 20-year period. Accompanying this change, the growth of the services sector with lower income inequality — the “wholesale, retail and import/export trades, restaurant and hotels” and “transport, storage and communications” sections — was quite limited, with the ratio of people engaged in these sections only increasing by 8.5%. This situation may be related to the special characteristics of Hong Kong’s industrial restructuring. During the process of industrial restructuring, Hong Kong did not experience the usual upgrading of its manufacturing sector. The consequence of this is that, during the process of service development, Hong Kong was unable to develop the productive service industries, especially the “transport, storage and communications” group. Within the aforementioned 20-year period, “finance, insurance, real estate 28 Some

studies conclude that the growth of the service sector will alleviate income inequality, which contradicts our findings.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch02

S. G. Wang and Y. Xia

38

and business services” and “community, social and personal services” were the two fastest growing service sectors, whose collective working population increased from 24.87% of the total working population in 1986 to 43.9% in 2006, a 19.1% increase for the 20-year period. However, these two service industries are well-known for their high within-group income disparity. Therefore, the fewer people engaged in industries with lower income disparity and the more people engaged in the more unequal industries, the larger the income gap for the working population as a whole will be.29 Third, during the process of industrial restructuring, the income gap between different industrial sectors was also increasing, as is clearly shown in Table 2.4. To take the female working population as an example, the Table 2.4.

Median Monthly Income by Industry, 1986–2007 (Unit: $ HKD). Sex 1986

1991

1996

2001

2006

2007

Changes

Manufacturing

F 1,900 4,000 6,900 8,000 7,000 7,500 M 3,000 6,000 10,000 12,000 12,000 12,000

5,600 9,000

Construction

F 2,200 5,500 10,000 9,000 M 3,000 6,000 10,000 10,000

8,300 9,000 9,500 10,000

6,800 7,000

F 2,100 5,000 8,000 8,000 8,000 8,000 M 3,000 6,000 10,000 11,000 11,500 12,000

5,900 9,000

F 2,900 6,000 9,500 10,500 11,000 10,500 M 3,500 6,000 10,000 10,000 10,000 10,300

7,600 6,800

Financing, insurance, real estate and business services F 3,200 6,500 12,000 13,000 13,000 14,000 M 3,900 8,000 13,000 15,000 14,000 15,000

10,800 11,100

Wholesale, retail and import/export trades, restaurants and hotels Transport, storage and communications

Community, social and personal services Highest–Lowest Highest–Lowest

F 2,700 4,800 M 3,600 7,600

7,500 13,00

6,000 5,500 5,500 16,00 15,000 15,000

F 1,300 2,500 M 900 2,000

4,500 3,000

7,000 6,000

7,500 5,500

8,500 5,000

2,800 11,400 8,000 4,600

Source: The Census and Statistics Department of the HKSAR, female and male in Hong Kong: Key statistics 2010: Table 5.2. 29 The Gini coefficient measured for the income distribution of working population in Hong Kong was 0.39 in 1986. This value increased to 0.432 in 2006. See Hon-kwong Lui, “The Widening Income Dispersion in Hong Kong: 1986–2006” (paper presented at Conference on Social Inequality and Social Mobility, the Chinese University of Hong Kong, Hong Kong, 14 March 2008). Accessed on September 3, 2010. http://www.cpu.gov.hk/ english/documents/conference/20080314%20Lui%20Hon-kwong.pps.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

39

lowest-paid occupation was “community, social and personal services,” whose monthly median income in 2007 was only $5,500 HKD, having only increased by $2,800 HKD between 1986 and 2007. Women received the highest pay in “wholesale, retail and import/export trades, restaurant and hotels,” whose median monthly income reached $14,000 HKD in 2007, increasing by $10,800 HKD between 1986 and 2007, an almost four-fold increase compared with the increase in the lowest-income occupations. For the working male population, the lowest income sector was “construction,” whose median monthly income only increased by $7,000 HKD during the same 21 year period. The highest income occupation for men was “community, social and personal services,” whose median monthly income increased $11,400 HKD during this same period. In summary, from 1986 to 2007, for both the female and male working population, the gaps between the highest median monthly income and the lowest median monthly income increased significantly. For women, the lowest–highest median monthly income gap increased from $1,300 HKD to $8,500 HKD, while for men it increased from $500 HKD to $5,000 HKD, a 6.5-fold increase and 5.6-fold increase respectively. Although within this 21 year period, the incomes of both female and male workers from all industrial sectors increased, it is those industries with the highest income level that experienced the greatest increases, while those at the bottom of the income pyramid experience the lowest increases. Fourth, during the process of industrial restructuring, the working population moved to the two ends of the overall income structure, rather than concentrating around the median level. As shown in Table 2.5, between 1986 and 2007 the working population employed in the manufacturing sector decreased by more than 700,000. Where did this labor force go? The data shows that the female labor force was mainly employed in “community, social and personal services” jobs, which increased by 411,800, and “wholesale, retail and import/export trades, restaurant and hotels,” which increased by 370,500. For female workers, these two groups of occupations are actually the two with the lowest average income. The male labor force was mainly employed in “finance, insurance, real estate and business services,” which increased by 221,200 workers, and “wholesale, retail and import/export trades, restaurant and hotels,” which increased by 170,300 workers. For males, these two occupational groups enjoyed the highest average income. For both males and females, much less of the labor force went to the sectors with median income levels, such as “construction” and “transport, storage and communications.” At present, these occupational groups employ very few workers. While the industrial sectors at the median income

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch02

S. G. Wang and Y. Xia

40 Table 2.5.

Working Population by Industry.

Median Monthly Employment Earnings Sex (2007) 1986

1991

1996

2001

2006

2007 Change

Manufacturing

F M

7,500 12,000

424.3 292.2 179.7 116.1 76.9 72.1 −352.2 495 424.8 310.3 208.9 140 130.3 −364.7

Construction

F M

9,000 10,000

9.5 10.9 15.7 20.5 19.3 20.4 191.6 214 257.5 268.2 250 254.3

10.9 62.7

Wholesale, retail and import/export trades, restaurants and hotels F M

8,000 12,000

212.3 300.3 404.4 481.6 558.9 582.8 390.7 431.8 503.4 499.4 545.9 561

370.5 170.3

F M

10,500 10,300

26.9 40.5 62.3 73.2 83.8 85.4 190.1 233.1 273.9 279.7 285.3 286.9

58.5 96.8

F M

14,000 15,000

64.1 101.8 147.7 198.5 217.8 229 97.8 127.2 216.1 283.5 307.9 319

164.9 221.2

F

5,500 15,000

208.3 278 392 512.8 598.3 620.1 253.2 258.1 275.9 287.2 293.8 301

411.8 47.8

Transport, storage and communications Financing, insurance, real estate and business services Community, social and personal services

Source: The Census and Statistics Department of the HKSAR, Women and Men in Hong Kong: Key Statistics 2010: Table 4.6.

level have a very important role in alleviating income inequality, the movement of the labor force from the median income occupations to the lowest and highest income occupations has definitely resulted in a shrinking “middle” group and two enlarging ends, thus exacerbating the overall income inequality. To sum up, the special industrial structure of Hong Kong is the fundamental factor that has led to the deteriorating situation of income inequality in Hong Kong during the past four decades. As it is difficult to solve this structural problem, in the near future, it will also be difficult to improve the situation of pre-existing market income inequality.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

41

2.4. Why is Disposable Income Distribution so Unequal in Hong Kong? In the last section, we considered why the market income distribution of Hong Kong continued to worsen during the past several decades given that the Gini coefficient was already very high in the early 1960s. As mentioned above, market income and disposable income are two different concepts. In a modern society, the market income people get from economic activities is not equal to personal disposable income. On the one hand, people need to pay taxes to the government. On the other hand, they might also receive certain kinds of social benefits (in cash or in kind). Only taking into account these two kinds of effects can we arrive at the final disposable income. Moreover, between the “market income” and the “disposable income,” there is another concept called the “after tax income.” The relationships between these three concepts are shown in Fig. 2.7. We can tell from Fig. 2.7 that the degree of inequality of disposable income is determined not only by the primary distribution of the market income, but also by the government’s redistribution efforts through its tax and social welfare policies. We have seen from the discussion of the last session that market income is quite unequal in Hong Kong. Under such circumstances, if the government’s redistribution efforts were strong enough, then it would be possible to create a relatively equal disposable income distribution. It follows that if the redistributive measures are weak, then the final disposable income will not be any more equal. Figure 2.8 reveals the actual effects of the Hong Kong government’s redistributive policies.

Fig. 2.7.

Market Income and Final Income After Tax and Disposable Income.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch02

S. G. Wang and Y. Xia

42

0.54 0.52

0.518

0.525

0.533 0.515

0.521

0.508

0.5 0.475

0.48 0.466

0.47

0.46 0.44 0.42 Original Monthly Household Income

Post-tax Monthly Household Income 1996

2001

Post-tax Post-social Transfer Monthly Household Income

2006

Fig. 2.8. The Effects of Government’s Redistribution on Income Distribution Measured by Gini Coefficient in Hong Kong. Source: The Census and Statistics Department of the HKSAR. Household income distribution in Hong Kong 2006: Table 6.6.

Undoubtedly, the government’ redistribution policies have contributed to narrowing the income gap between different income groups. After earning the market income from economic activities in the market, residents in Hong Kong need to pay salary tax and some estate-related taxes, such as the property tax, rates and government rent, among which the salary tax is progressive, in that the tax rate increases as income increases. At the same time, local residents also enjoy several social benefits, such as free compulsory education, free public medical service, public housing and a government-sponsored housing scheme. For example, in 2006, the market income Gini coefficient was 0.533, which was reduced by 2.25% to 0.521 with the effects of tax paymets, and a further reduction by 10.88% to 0.475 if accounting for the social transfer effects. Although the inequality of disposable income shows an improvement relative to that of market income distribution, the Gini coefficient of the final disposable income distribution is still 0.475, which is quite high for a city economy with a population of seven million people. It is even higher than the Gini coefficient for the mainland China as a whole, which is wellknown for its huge regional and urban–rural disparities. This is quite counterintuitive. Is this because the market income distribution is so unequal that mere reliance on the government’s redistributive measures is unable to alleviate the inequality? To provide a basis of comparison, we can select

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

43

0.6 0.5 0.4 0.3 0.2 0.1

Original Income

Post-tax Income

Final Income

0

Fig. 2.9. The Effects of Redistributions on Income Distribution in Great Britain Measured Using the Gini Coefficient. Source: Howard Glennerster, “Tibor Barna: The Redistributive Impact of Taxes and Social Policies in the UK, 1937–2005,” LSE STICERD Research Paper No. CASE115 (December 2006), p. 10.

Great Britain as the reference case whose market income inequality is most similar to that of Hong Kong. As shown in Fig. 2.9, under the “Thatcher Revolution” accompanying the neo-liberal economy policies from 1979 to 1990, Great Britain’s economy polarized rapidly. Before the Iron Lady came to power, the Gini coefficient measured for market income inequality was 0.43; when she resigned, it had climbed to 0.54, no lower than that of Hong Kong during the same period. Correspondingly, from the end of the 1970s to the early 1990s, the Gini coefficients measured for after-tax market income and final disposable income both increased to their highest historical levels. However, in contrast to the case of Hong Kong, the British government adopted much more aggressive redistributive measures. From 1977 to 2004/2005, compared with the original market income, the British after-tax income Gini coefficient had declined by 29.3%, with that of the final disposable income declining by 42.8%. Therefore, from the middle of the 1980s onwards, the Gini coefficient for final disposable income in Britain has been maintained at around the 0.3 level. Although this is markedly higher than at the end of the 1970s, it is much lower than that in Hong Kong. In fact, there are other cases that have quite similar market inequality to Great Britain, although not all of those economies have such

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch02

S. G. Wang and Y. Xia

44

effective redistributive measures as Britain. Then, what kinds of economies have relatively strong redistributive measures like Britain, and what kinds of economies have weak redistributive measures similar to those in Hong Kong? From Fig. 2.10 we can make several interesting observations. 0

0.1

0.2

0.3

0.4

0.5

0.6

Austria 99 Denmark 04 Luxembourg 01 Belgium 02 Netherlands 00 Sweden 03 Norway 00 Finland 04 France 01 Germany 02 Slovenia 05 Hungary 05 Switzerland 02 EU-15 01 New Zealand 98 Ireland 01 Spain05 Australia 03 Canada 00 Greece 05 Estonia 05 Poland 05 Italy 01 UK 09 Portugal 01 Costa Rica 00 US 05 Guatemala 04 Russia 00 Honduras 04 Singapore 09 El Salvador 00 Panama 03 Chile 99 Hong Kong 06 Argentina 97 Nicaragua 98 Peru 04 Mexico 02 Bolivia 04 Colombia 03 Brazil 99

Market Income Fig. 2.10. Countries.

Final Income

Gini Coefficients of Market Income and Disposable Income in Selected

Source: Data for Latin America come from Rodrigo Cubero and Ivanna Vladkova Hollar, “Equity and Fiscal Policy: The Income Distribution Effects of Taxation and Social Spending in Central America,” IMP Working Papers, WP/10/112 (May 2010), accessed on September 3, 2010. http://www.imf.org/external/pubs/ft/wp/2010/wp10112.pdf; data for developed countries come from David K. Jesuit and Vincent A. Mahler’s “Fiscal Redistribution Dataset” Version 2.0 (Feb. 2008), accessed on September 3, 2010. http://www.lisproject.org/publications/fiscalredistdata/fiscred.htm; data for Singapore comes from Singapore Department of Statistics, Key Household Income Trends, 2009 (Feb. 2010), accessed on September 3, 2010. http://www.singstat.gov.sg/pubn/papers/ economy/op-s16.pdf; data for Russia comes from Michele Giammatteo, “Inequality in Transition Countries: The Contribution of Markets and Government Taxes and Transfers,” Luxembourg Income Study (LIS) Working Paper Series No. 444 (August, 2006), accessed on September 3, 2010. http://www.lisproject.org/publications/liswps/443.pdf.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

45

In terms of market income, almost every country displays high degree of inequality. Among the 42 economies covered in Fig. 2.10, there are 13 economies whose Gini coefficients are above 0.5, which is similar to the level of Hong Kong. The Gini coefficients of the other economies are all above the 0.4 level. Even Sweden, which is well-known for its relative income equality, has a Gini coefficient measured for market income that is as high as 0.468. This indicates that under a capitalist economic system, the unequal distribution of market income is likely inevitable. In this aspect, Hong Kong is not exceptional; it is not much different from other capitalist economies in terms of its market income Gini coefficient. However, when considering the government’s redistributive efforts, the differences between these 42 economies begin to emerge. Basically, we can classify these economies into three categories. The first group contains 17 economies whose redistributive measures are the most intensive, able to reduce the Gini coefficient by more than 35%. Countries in this group are mainly west and north European countries. After redistribution, the Gini coefficients of their disposable income generally are reduced to less than 0.3. The second group contains 10 economies with less intensive, but still very strong, redistributive measures; they are able to reduce the Gini coefficients of market income by more than 15% but less than 35%. Most countries in this group are southern European and Anglo-Saxon countries. After redistribution, the Gini coefficients of the disposable income are basically all below the 0.4 level. The U.S.A. is the only exception in the second group. One reason for this is that its original market income inequality is simply too high; another reason is that its redistributive ability — reducing the Gini coefficient of the market income by 15.21% — is only slightly higher than the lower threshold for the second group. Therefore, its disposable income Gini coefficient is still as high as 0.418. In fact, the overall level of redistributive effects of the U.S.A. is in line with the third group, which group contains 15 economies whose redistributive abilities are the weakest, able at best to reduce the Gini coefficient of the market income by 15%. Countries in this group are mainly Latin American countries, though the group also includes Hong Kong, Singapore and Russia. Even after redistribution, their final disposable income Gini coefficients are still around 0.45. In Brazil and Chile, however, tax and social benefits did not bring about any improvement to income disparity. We can see that, along with Singapore, Hong Kong is among the most developed economies in terms of its per capita GNP, while it is among the worst economies in terms of its redistribution ability.

June 5, 2012

11:23

46

9in x 6in

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

Why are Hong Kong’s redistributive measures so weak? The reason is not especially complicated if we examine the relevant institutional settings in Hong Kong. First, the tax system of Hong Kong is not tailored for redistribution. The highest rate of personal income tax is actually very low, around 17% in recent years. In other words, the tax rate for people receiving a yearly income of $500,000 HKD is the same as that for people earning $5 billion HKD per year. The highest rate of corporate income tax is only 16.5%. Additionally, there are no such redistributive taxes in Hong Kong as taxes on inheritance, dividend and property appreciation, all of which are quite common in many other advanced economies. Thus, Hong Kong has garnered a reputation as a paradise for wealthy people. Finally, the main body of people paying the betting tax are people from the bottom of the society who are keen on horse-racing. What’s more, the burden of land selling is also borne by ordinary residents. With limited tax categories and low tax rates, the overall level of taxation in Hong Kong is quite low. The ratio of revenue to the overall GDP is below 20%, which is incomparable with most OECD countries, whose ratios are almost uniformly above 40%.30 It is because of the Hong Kong government’s limited intervention that the American conservative think tank, the Heritage Foundation, has elected Hong Kong as the freest economy in the world almost every year.31 Many people in Hong Kong are quite proud of this title. However, one of the direct consequences of limited government revenue is that the government is unable to increase the social welfare expenditure. Talking about the social welfare expenditure, although the government provides almost-free medical service, 12 years of compulsory education and public renting properties to 30% of the poor people, Hong Kong does not provide unemployment benefits or retirement security schemes (except for public servants). Moreover, the scope of the social security system (the Comprehensive Social Security Assistance Scheme) is much smaller than in most Western countries. As

30 Calculated with data in “Table 193: Government Revenue (General Revenue Account and Funds)”, the Census and Statistics Department of the HKSAR. Accessed on September 3, 2010. http://www.censtatd.gov.hk/hong kong statistics/statistical tables/ index tc.jsp?subjectID=9&tableID=193. 31 “Economic Freedom Component Scores, All Countries 2010,” the Heritage Foundation. Accessed on September 3, 2010. http://www.heritage.org/index/Explore.aspx? view=by-variables.

b1363-ch02

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch02

Growth and Inequality in Hong Kong

47

35 30 25 20 15 10 5 HK

0 1980 Fig. 2.11. (Unit: %).

1985

1990

1995

2000

2001

2002

2003

2004

2005

The Ratio of Social Expenditure to GDP in Selected OECD Countries

Source: Data for OECD countries comes from “Social Expenditure Database (SOCX)”, the OECD, accessed on September 3, 2010. http://www.oecd.org/document/9/0,3343,en 2649 34637 38141385 1 1 1 1,00.html; figures for Hong Kong are calculated with data from the Census and Statistics Department of the HKSAR, accessed September 3, 2010. http://www.censtatd.gov.hk/gb/?param=b5uniS&url=http://www.censtatd.gov.hk/ hong kong statistics/statistical tables/index tc.jsp?charsetID=2&tableID=193.

can be seen in Fig. 2.11, in terms of the ratio of social expenditure to GDP, Hong Kong (represented by the black bold line in the figure) deviates from most other OECD countries and is similar only to South Korea and Mexico. In 2005, even these two economies exceeded Hong Kong in provision of benefits, although their per capita GDP is much lower than that of Hong Kong. Data presented in Fig. 2.12 covers 91 economies. The figure clearly shows that the ratio of a government’s social welfare expenditure to GDP is negatively correlated with income inequality. In other words, the more the government spends on social welfare, the less unequal is the overall income distribution and vice versa. The Hong Kong government’s philosophy of a “small government” — low intervention, low revenue and low social expenditure — is at least partly to blame for the extremely unequal situation.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

48

Fig. 2.12. The Relationship Between Public Social Expenditure and Income Inequality.32 Source: Data for social expenditure comes from “Social Security Expenditure Database”, the ILO, accessed on September 3, 2010. http://www.ilo.org/dyn/sesame/ IFPSES.SocialDBExp; Gini coefficients come from “World Development Indicators 2009,” the World Bank, accessed September 3, 2010. http://hdrstats.undp.org/en/indicators/display cf xls indicator.cfm?indic byyear id=161.

2.5. Summary Previous discussions about income inequality in Hong Kong have mostly focused on the distribution of market income. Such works are important, as it is first necessary to understand the reasons why the inequality of market income distribution has changed over time. Analysis in the second section of this study showed that the worsening market income inequality in Hong Kong in the recent two to three decades was mainly due to Hong Kong’s special industrial restructuring model. However, we should not rely only on the analysis of the distribution of market income distribution. As the third section of this article showed, under a capitalist economic system, an unequal distribution of market income is normal, rather than exceptional. 32 “Income”

in Fig. 2.12 refers to final disposable household income.

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

49

What makes Hong Kong an outlier is that the government’s redistribution policies are too weak to smooth out the high earning inequality generated by the market. The redistributive policies of most other economies at similar levels of economic development as Hong Kong in terms of per capita GDP are much more aggressive than those practiced in Hong Kong. This tells us that, in order to alleviate the worsening condition of income inequality in Hong Kong, relying only on reducing earnings inequality might prove insufficient. More importantly, the government should strengthen its redistribution policies. Were the degree of redistribution in Hong Kong to reach the average level of the OECD countries, the Gini coefficient of the city’s final income inequality could be reduced to around 0.35. However, whether the government will change its “positive non-intervention” policy remains to be seen. It is beyond the scope of the present study, to determine the types of factors that could push the government to change major aspects of its long-standing economic philosophy.

References Bourguignon, F., Ferreira, F.H.G. and Lustig N. (2005) (Eds). The Microeconomics of Income Distribution Dynamics in East Asia and Latin America. New York: Oxford University Press. Bourguignon, F., Fournier, N. and Gurgand, M. (2001). “Fast Development with a Stable Income Distribution: Taiwan, 1979–1994.” Review of Income and Wealth, 47(2), 1–25. Bradley, D., Huber, E., Moller, S., Nielsen. F. and Stephens, J.D. (2003). “Distribution and Redistribution in Postindustrial Democracies.” World Politics, 55(2), 193–228. Chen, W.H. (2004). (Jiegou zhuanxing yu xianggang de chanye zhengce) “Structural Adjustment and Hong Kong’s Property Policy”, (Guoji jingji pinglun) International Economics Review, 2004(3–4), 33–37. Chiu, S.W.K. (2005). Family Changes and Income Inequality under Globalization: The Case of Hong Kong. Hong Kong: Hong Kong Institution of Asia-Pacific Studies. Chou, K.R. (1966). The Hong Kong Economy: A Miracle of Growth. Hong Kong: Academic Publications. Chow, S.C. and Papanek, G.F. (1981). “Laissez-Faire, Growth and Equality– Hong Kong.” The Economic Journal, 91(362), 466–485. Cloutier, N. (1997). “Metropolitan Income Inequality during the 1980s: The Impact of Urban Development, Industrial Mix, and Family Structure.” Journal of Regional Science, 37, 459–478. Danziger, S. and Gottschalk, P. (1995). America Unequal. New York: Russel Sage Foundation.

June 5, 2012

11:23

50

9in x 6in

Growth and Inequality: An International. . .

S. G. Wang and Y. Xia

Dicken, P. (1998). Global Shift: Transforming the Global Economy. London: Paul Chapman Publishing. Feng, B.Y. (2001). (Xianggang yu xinjiapo chanye jiegou ji jingji zhengce de bijiao yanjiu) “Comparative Study of the Industrial Structure and Economic Policies in Hong Kong and Singapore.” (Xueshu yanjiu) Academic Research, 2001(7), 42–45. Feng, X.Y. (2007). (Xianggang jingji jiegou zhuanxing chutan) “The Industrial Restructuring in Hong Kong,” (Jingji qianyan) Frontier of Economics, 2007(7), 20–25. Fields, G.S. (1984). “Employment, Income Distribution and Economic Growth in Seven Small Open Economies.” The Economic Journal, 94(373), 74–83. Giammatteo, M. (2006). “Inequality in Transition Countries: The Contribution of Markets and Government Taxes and Transfers.” Luxembourg Income Study (LIS) Working Paper Series No. 444. Glennerster, H. (2006). “Tibor Barna: The Redistributive Impact of Taxes and Social Policies in the UK, 1937–2005.” LSE STICERD Research Paper No. CASE115. Gottschalk, P. and Smeeding, T. (1997). “Cross-National Comparisons of Earnings and Income Inequality.” Journal of Economic Literature XXXV, 633–687. Gottschalk, P. and Smeeding, T. (1999). “Empirical Evidence on Income Inequality in Industrialized Countries.” Luxembourg Income Study Working Paper Series. Hollar, I.V. and Cubero, R. (2010). “Equity and Fiscal Policy: The Income Distribution Effects of Taxation and Social Spending in Central America.” IMF Working Papers No. 10/112. Katz, L. and Murphy, K. (1992). “Changes in Relative Wages, 1962–1987: Supply and Demand Factors.” The Quarterly Journal of Economics, 107(1), 35–78. Khan, A.R., Griffin, K. and Riskin, C. (1999). “Income Distribution in Urban China during the Period of Economic Reform and Globalization.” The American Economic Review, 89(2), 296–300. Krongkaew, M. (2005). “Income Distribution in East Asian Developing Countries: An Update.” Asian-Pacific Economic Literature, 8(2), 58–73. Kum, H. (2008). “Inequality and Structural Change.” UTIP Working Paper No. 54, pp. 14–15. Kuznets, S. (1954). “Economic Growth and Income Inequality.” American Economic Review, 45, 1–28. Kuznets, P.W. (1988). “An East Asian Model of Economic Development: Japan, Taiwan and South Korea.” Economic Development and Cultural Change, 36(3), S11–S43. Leung, D.K.T. (2009). “Insights from Census Data on Income Distribution in Hong Kong.” Paper presented at the 24th Population Census Conference 2009, Hong Kong, March 25–27, 2009. Levy, F. and Murnane, R. (1992). “U.S. Earnings Levels and Earnings Inequality: A Review of Recent Trends and Proposed Explanations.” Journal of Economic Literature, XXX, 1333–1381.

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Hong Kong

b1363-ch02

51

Lin, T.B. (1985). “Growth, Equity and Income Distribution Policies in Hong Kong.” The Developing Economies, XXIII(4), 397–398. Ling, N.S. (2006). “Singapore’s Manufacturing Sector 1991–2005”. Statistics Singapore Newsletter: 7–11. Lui, H.K. (2008). “The Widening Income Dispersion in Hong Kong: 1986–2006.” Paper presented at a conference entitled “Social Inequality and Social Mobility in Hong Kong.” Hong Kong: The Chinese University of Hong Kong, March 14, 2008. Mizoguchi, T. (1985). “Economic Development Policy and Income Distribution: The Experience in East and Southeast Asia.” The Developing Economies, XXIII-4, 307–324. Mo, P. H. (2003). “Income Inequality and Economic Growth.” Kyklos, 53(3), 293–315. Mo, P. H. (2004). Measures of Income Inequality and Economic Growth. Hong Kong: Hong Kong Baptist University. Motonishi, T. (2003). “Why Has Income Inequality in Thailand Increased?” An Analysis Using 1975–1998 Surveys. Asian Development Bank ERD Working Paper Series No. 43. Murphy, K.M., Shleifer, A. and Vishny, R. (1989). “Income Distribution, Market Size and Industrialization.” The Quarterly Journal of Economics, 104(3), 537–564. Okun, A.M. (1975). Equality and Efficiency: The Big Tradeoff. Washington: The Brookings Institute. Partridge, M., Rickman. D. and Levernier, W. (1996). “Trends in U.S. Income Inequality: Evidence from a Panel of States.” The Quarterly Review of Economics and Finance, 36(1), 17–37. Ram, R. (1997). “Level of Economic Development and Income Inequality: Evidence from the Postwar Developed World.” Southern Economic Journal, 64(2), 576–583. Suen, W. (1995). “Sectoral Shifts Impact on Hong Kong Workers.” The Journal of International Trade and Economic Development, 4(2), 135–152. World Bank. (1993). The East Asian Miracle: Economic Growth and Public Policy. Oxford: Oxford University Press: 8, 28–32. You, J-II. (1998). “Income Distribution and Growth in East Asia.” Journal of Development Studies 36(4), 37–65. Zhang, G.N. (2003). (Xianggang zhizaoye de jiuye jiegou yu chanye jiegou de xiaoyi fenxi) “The Efficiencies of the Employment and Industrial Structure of Manufacturing in Hong Kong.” (Tongji yu yuce) Statistics and Predictions, 2003(6), 74–76. Zhao, X.B.S. and Zhang, L. (2005). “Economic Growth and Income Inequality in Hong Kong: Trends and Explanations.” China: An International Journal, 3(1), 74–103. Zin, R.H.M. (2005). “Income Distribution in East Asian Developing Countries: Recent Trends.” Asian-Pacific Economic Literature, 19(2), 36–54.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch02

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 3 GROWTH AND INEQUALITY IN THE UNITED STATES Edward Nathan Wolff, Ajit Zacharias and Thomas Masterson∗

3.1. Introduction In our parlance, economic well-being refers to the household’s command over, and access to, the goods and services produced in a modern market economy during a given period of time. An income measure is normally used to measure its magnitude, since household income should, in principle, reflect the resources available to the household over a given period of time (typically, a year) for facilitating current consumption or acquiring assets. In the U.S. and many other advanced industrialized countries, gross money income (MI) is the standard measure used for this. However, MI is known to have many shortcomings. The landmark report by the Canberra Group (2001), a group of international experts on household income statistics, highlighted many of these deficiencies. In particular, MI does not include an estimate of in-kind social benefits, no valuation is included for household production or public consumption, property income is a limited indicator of the benefits from wealth holdings, and taxes are not netted out of the measure. Our aim in this paper is to show that a more comprehensive measure of income, which we call LIMEW (the Levy Institute Measure of Economic ∗ Edward Nathan Wolff, Professor, Department of Economics, New York University, USA; Senior Scholar, Levy Economics Institute of Bard College, USA. Ajit Zacharias, Senior Scholar, Levy Economics Institute of Bard College, USA; Thomas Masterson, Research Scholar, Levy Economics Institute of Bard College, USA. The author’s are deeply indebted to the Alfred P. Sloan Foundation for financial support for this project. We would also like to acknowledge the research assistance of Melissa Mahoney and contributions by Hyunsub Kum, and Selcuk Eren. An earlier version of the paper was presented at the International Association for Research in Income and Wealth 30th General Conference, Portoroz, Slovenia, August 24–30, 2008.

53

b1363-ch03

June 5, 2012

11:23

54

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

Well-Being) and which overcomes many of the shortcomings of MI, will provide more reliable information on trends in living standards, the level of inequality and trends in inequality than MI.1 As far as we are aware, LIMEW is the most comprehensive measure of income that has been developed to date. In particular, we believe that MI gives a distorted picture of actual living standards. Since the state plays a crucial role in the direct provisioning of the “necessaries and conveniences of life” (to use Adam Smith’s famous expression), such as public education and highways, we include estimates of public consumption in our measure. Since non-market household work, such as childcare, cooking, and cleaning, also provides the necessaries and conveniences of life, we also include household production in LIMEW. We also include estimates of long-run benefits from the ownership of wealth (other than homes) in the form of an imputed lifetime annuity, a procedure that, in our view, is superior to considering current property income from assets. Services derived from owner-occupied housing are valued by means of imputed rent in our measure. We use our measure to analyze trends in living standards and inequality in the United States. The findings in this chapter also highlight the important role played by the government sector in promoting increases in the living standards among the middle class. In fact, according to the LIMEW measure, the public sector was the leading source of growth in the standard of living of the middle class between 1959 and 2007. Changes in inequality have to a large extent been due to periodic spikes in household wealth. In fact, for the population as a whole, the share of income from wealth in LIMEW almost doubled between 1959 and 2007. This would have led to inequality rising much more quickly than it did, had not government spending and taxes greatly lowered inequality throughout the post-war period. We begin by briefly describing the construction of the LIMEW (Sec. 3.2).2 In Sec. 3.3, we report on time trends in LIMEW and MI from 1959 to 2007. Our concerns are two-fold. First, how have living standards changed over time on the basis of our extended income concept and how can we account for its movements over time? Second, how does this compare with the “conventional wisdom” based on money income? Section 3.4 reports on inequality trends. Concluding remarks are made in Sec. 3.5. 1 Wolff and Zacharias (2007a) provided an overview of the LIMEW and discussed results for the U.S. in the 1990s using MI, LIMEW and the Census Bureau’s broadest definition of disposable income. 2 The sources of data and methods used are described in Wolff, Zacharias and Masterson (2009) and Masterson (2010).

b1363-ch03

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in the United States Table 3.1.

b1363-ch03

55

Components of LIMEW and Average Values for 2007 (Per Household). 2007 Estimates

Derivation of LIMEW

Mean

Bottom 20%

Top 5%

Money income (MI) Less: Property income and government cash transfers Equals: Base income Plus: Income from wealth Annuity from non-home wealth Imputed rent on owner-occupied housing Less: Taxes Income taxes1 Payroll taxes1 Property taxes1 Plus: Cash transfers1 Plus: Non-cash transfers1,2 Equals: Comprehensive Disposable Income (CDI) Plus: Public consumption Equals: Post Fiscal Income (PFI) Plus: Household production Equals: LIMEW

67,622 8,760

23,819 5,252

228,015 22,254

58,863 24,687 17,285 7,402 16,242 9,986 2,332 3,924 6,428 6,358 80,093

18,567 1,748 −551 2,298 6,272 3,916 1,034 1,323 4,018 3,910 21,971

205,761 297,269 263,262 34,006 66,723 46,724 10,105 9,895 9,385 6,611 452,302

11,197 91,290 24,040 115,330

3,784 25,755 6,062 31,817

14,934 467,236 62,057 529,292

1 Aligned 2 The

with the NIPA estimates. government-cost approach is used.

3.2. Components of LIMEW LIMEW is constructed as the sum of the following components (see Table 3.1): Base money income; income from wealth; net government expenditures (both cash and non-cash transfers and public consumption, net of taxes); and household production. We provide here a summary of the procedures used to construct LIMEW.3 Base money income is defined as MI less the sum of property income (interest, dividends, and rents) and government cash transfers (e.g., social security benefits) that are included in MI. Earnings make up the overwhelming portion of base money income. The remainder consists of pensions, interpersonal transfers, workers’ compensation paid by the private sector, and other small items. In 2007, base income was a little over half of total LIMEW. 3 See

Wolff, Zacharias and Caner (2004) and Wolff and Zacharias (2007a) for more details on the methodology used to construct LIMEW.

June 5, 2012

11:23

56

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

The second component is imputed income from the household’s wealth holdings. MI includes interest, dividends, and rent. From our perspective, property income is an incomplete measure of the economic well-being derived from the ownership of assets. Owner-occupied housing yields services to their owners over many years, thereby freeing up resources otherwise spent on housing. Hence, benefits from owner-occupied housing are reckoned in terms of the replacement cost of the services derived from it (i.e., a rental equivalent).4 Under normal conditions, financial assets can be a source of economic security greater than that provided by property-type income. We estimate the benefits from non-home wealth using a lifetime annuity method.5 We calculate an annuity based on a given amount of wealth, an interest rate, and life expectancy. The annuity is the same for the remaining life of the wealth holder and the terminal wealth is assumed to be zero (in the case of households with multiple adults, we use the maximum of the life expectancy of the head of household and spouse in the annuity formula). Moreover, in our method, we account for differences in portfolio composition across households. Instead of using a single interest rate for all assets, we use a weighted average of asset-specific and historic real rates of return,6 where the weights are the proportions of the different assets in a household’s total wealth. In 2007, income from home and non-home wealth (primarily the latter) made up a little over a fifth of LIMEW. The third component is net government expenditures —the difference between government expenditures incurred on behalf of households and taxes paid by households. Our approach to determine expenditures and taxes is based on the social accounting approach (Hicks, 1946; Lakin, 2002, pp. 43–46). Government expenditures included in LIMEW are cash transfers, non-cash transfers, and public consumption. These expenditures, in general, are derived from the National Income and Product Accounts 4 This

is consistent with the approach adopted in the U.S. national accounts. method gives a better indication of resource availability on a sustainable basis over the expected lifetime than the standard bond-coupon method. The latter simply applies a uniform interest rate to the value of nonhome wealth. It thereby assumes away differences in overall rates of return for individual households ascribable to differences in household portfolios. It also assumes that the amount of wealth remains unchanged over the expected (conditional) lifetime of the wealth holder. 6 The rate of return used in our procedure is real total return (the sum of the change in capital value and income from the asset, adjusted for inflation). For example, for stocks, the total real return would be the inflation-adjusted sum of the change in stock prices plus dividend yields (see Wolff, Zacharias and Masterson (2009) for details). 5 This

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

57

[NIPA, Tables 3.12 and 3.15.5]. Government cash transfers are already treated as part of the money income of the recipients. In the case of government non-cash transfers, our approach is to distribute the appropriate actual cost incurred by the government among recipients of the benefit. An alternative, the “fungible-value method,” is based on the argument that the income value for the recipient of a given non-cash transfer is, on average, less than the actual cost incurred by the government in providing that benefit (see, for example, Canberra Group, 2001, p. 24 and p. 65). This valuation method involves estimating how much the household could have paid for the medical benefit, after meeting its expenditures on basic items such as food and clothing, with the maximum payment for the medical benefit set equal to the average cost incurred by the government. We do not use the fungible-value approach because of its implication that recipients with income below the minimum threshold receive no benefit from services such as health care. This implication is inconsistent with our goal of measuring the household’s access to or command over products. Further, unlike the social-accounting method, the fungible-value method would not yield the actual total government expenditure when aggregated across recipients. Such a feature is incompatible with our goal of estimating net government expenditures using a consistent methodology. The other type of government expenditure that we include in LIMEW is public consumption. We first select items from a detailed functional classification of government expenditures if they satisfy the general criterion of increasing the household’s access to goods or services. The household sector’s share in such expenditures can be estimated on the basis of information regarding its utilization (for example, for allocation of spending on transportation, we use the miles driven by households and businesses). In the second stage, the expenditures for each functional category are distributed among households, building on earlier studies employing the government cost approach (e.g., Ruggles and O’Higgins (1981)). Some expenditure such as education, highways, and water and sewerage are distributed on the basis of estimated patterns of utilization or consumption, while others such as public health, fire, and police are distributed equally among the relevant population.7 7 See Wolff and Zacharias (2007b) for more details. It should be noted that in the case of some expenditures, e.g., education, the government cost of provision need not coincide with the private or social benefit, as measured by an economic model. In that paper we also report the results of a sensitivity analysis to alternative assumptions regarding three components of public consumption: general public consumption, highways, and

June 5, 2012

11:23

58

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

The third part of net government expenditures is taxes. Our objective is to determine the actual tax payments made by households. We do not consider tax incidence in our analysis. Our approach is consistent with the government cost approach. We align the aggregate taxes in the annual demographic supplement (ADS) (imputed by the Census Bureau) with their NIPA counterparts, as we did for government expenditures. We include only taxes paid directly by households, including federal and state personal income taxes, property taxes on owner-occupied housing, and payroll taxes (employee portion).8 Taxes on corporate profits, on businessowned property, and on other businesses, as well as non-tax payments, are not allocated to the household sector because they are paid directly by the business sector. All told, net government expenditures amounted to 6% of LIMEW in 2007. The fourth component of LIMEW is the imputed value of household production. Three broad categories of unpaid activities are included in the definition of household production: (1) core production activities, such as cooking and cleaning; (2) procurement activities, such as shopping for groceries and for clothing; and (3) care activities, such as caring for babies and reading to children. These activities are considered as “production,” since they can be assigned, generally, to third parties apart from the person who performs them, although third parties are not always a substitute of the person, especially for the third activity.9 Our strategy for imputing the value of household production is to value the amount of time spent by individuals on the basis of its replacement cost as indicated by the average earnings of domestic servants or household employees (Kuznets, Epstein and Jenks 1941: 432, 433; Landefeld and McCulla 2000).10 Research suggests that there are significant differences schooling. New calculations for 1989 and 2000 showed that our initial major findings remain intact using alternative estimation procedures: there was a positive correlation between public consumption and the LIMEW, overall inequality was higher in 2000 than 1989, and public consumption reduced inequality. The results showed that our measure of economic well-being was robust under alternative assumptions of public consumption. 8 We do not include the employer portion of the payroll tax since it is paid directly by businesses and is not included in personal income. Sales taxes, on the other hand, should be included here. However, because of a lack of pertinent information for the allocation of sales taxes in 1959, we are unable to include them in this time-series comparison. 9 The third-party principle is sometimes ambiguous in the case of such personal care activities as shaving [see OECD (1995: 11)]. 10 Alternative approaches generally used are the opportunity cost and specialist wage approaches. As the name implies, the opportunity cost approach values the time spent by an individual on household production according to the wage that the individual

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

59

among households in the quality and composition of the “outputs” of household production, as well as the efficiency of housework (National Research Council 2005: Chap. 3). The differentials are correlated with household-level characteristics (such as wealth) and characteristics of household members [such as the influence of parental education on childrearing practices, e.g., Yeung and Stafford (2003)]. Therefore, we modify the replacement-cost procedure and apply to the average replacement cost of a discount or premium that depends on how the individual (whose time is being valued) ranks in terms of a performance index. Ideally, the performance index should account for all the factors relevant in determining differentials in household production and the weights of the factors should be derived from a full-fledged multivariate analysis. Given the absence of such research findings, we incorporated three key factors that affect efficiency and quality differentials —household income, educational attainment, and time availability —with equal weights attached to each.11 In 2007, household production made up about a fifth of LIMEW.

3.3. Trends in the Level and Composition of Well-being We first look at trends in LIMEW. Over the entire 1959–2007 period, median LIMEW grew at an annual rate of 0.67% (see Table 3.2).12 From 1959 to 1972, median LIMEW gained only 0.4% per year, while from 1972 to 1982 median LIMEW suffered an absolute decline. This was followed by a growth burst from 1982 to 1989 of 2.8% per year. However, growth slowed down from 1989 to 2007 when median LIMEW could muster only a 0.9% advance per year. How do these growth rates compare to the conventional measure MI? It is first of note that by construction, MI has lower average values is currently earning (or could potentially earn, in the case of unemployed individuals). In the specialist wage approach, household production tasks are categorized into a few groups (e.g., cooking, caring for children etc.) and valuation is performed according to the wages earned by workers in corresponding occupations (e.g., cooks, childcare workers etc.). 11 It should be noted that in our approach household production is a function only of the imputed time spent in these activities and the average cost of household workers (adjusted by income, education, and time). This approach ignores the effect of technological improvements on household production. 12 The choice of years included in the empirical work is dictated almost solely by data availability, particularly household wealth data (see Wolff and Zacharias (2007a) for details).

June 5, 2012 11:23

Trends in Income and Work, 1959 to 2007.

60

Table 3.2.

Median Values in 2007 Dollars 1982

1989

2000

2004

2007

62,479 36,988

65,465 44,388

61,150 42,989

74,316 48,388

82,320 50,575

85,520 48,530

86,080 50,000

78,458 55,614

97,962 64,636

108,945 68,752

112,648 65,887

114,083 68,031

Addendum A: Equivalence scale adjustment Equivalent LIMEW Equivalent MI

70,346 41,291

79,462 53,499

Addendum B: Annual hours of work (median values) 2,150 2,617

2,105 2,065

2,080 2,155

2,236 2,103

2,340 2,063

2,080 2,123

2,080 2,014

Total

5,084

4,600

4,501

4,718

4,749

4,683

4,593

1959–1972

1972–1982

1982–1989

Annual Percentage Change

LIMEW Money income (MI)

0.36 1.41

−0.68 −0.32

1989–2000

2000–2004

2004–2007

1959–2007

2.82 1.70

0.93 0.40

0.96 −1.03

0.22 1.00

0.67 0.63

−0.13 0.39

3.22 2.17

0.97 0.56

0.84 −1.06

0.42 1.07

1.01 1.05

Addendum A: Equivalence scale adjustment Equivalent LIMEW Equivalent MI

0.94 2.01

Addendum B: Annual hours of work Market work Housework

−0.16 −1.80

−0.12 0.43

1.04 −0.35

0.41 −0.18

−2.90 0.73

0.00 −1.74

−0.07 −0.54

Total

−0.77

−0.22

0.67

0.06

−0.35

−0.64

−0.21

Growth and Inequality: An International. . .

Market work Housework

E. N. Wolff, A. Zacharias and T. Masterson

1972

9in x 6in

LIMEW Money income (MI)

1959

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

61

than LIMEW. The median value of MI amounted to 59% of LIMEW in 2007. Over the entire 1959–2007 period, median MI grew at almost the same rate as median LIMEW, 0.63% per year compared to 0.67% per year. There are much larger differences by sub-periods. In the 1959–1972 period, median MI grew at an annual rate that was four times higher than that of median LIMEW. From 1972 to 1982, both LIMEW and MI fell in absolute terms, with LIMEW showing a rate of decline that was twice as high. In contrast, in the years 1982 to 1989, both measures recorded very high growth rates, but LIMEW grew almost twice as fast.13 From 1989 to 2007, median MI advanced at an annual rate of 0.2%, compared to 0.9% for median LIMEW. Indeed, from 1959 to 1982, median MI showed an annual gain of 0.7% while median LIMEW declined in absolute terms by 0.1% per year. In contrast, from 1982 to 2007, median MI grew by 0.6 percent per year while median LIMEW advanced by 1.4% per year. Thus, median MI showed higher growth than median LIMEW in the 1960s and 1970s while the opposite was true in the 1980s, 1990s, and 2000s. Addendum A shows trends in the various measures of well-being in equivalent dollars (that is, income adjusted for family size and composition).14 Both LIMEW and MI show a higher rate of growth when an equivalence scale adjustment is applied. This difference reflects the reduction in average household size over these years. Over the entire 1959 to 2007 period, median LIMEW and MI grew at almost the same rate, 1.01% and 1.05% per year, respectively. As before, median equivalent LIMEW displayed faster growth after 1982, while median equivalent MI grew faster before 1982. A. Composition of LIMEW In order to explain time trends in LIMEW, it is crucial to understand how the relative importance of the various components of LIMEW shifted over time. For the household sector as a whole, the most notable change was a pronounced increase in the share of income from wealth in total 13 Note the fact that 1982 is the bottom of a deep recession, which increases the measured growth accordingly. 14 The equivalence scale used here is the three-parameter scale employed in the U.S. Census Bureau’s experimental poverty measures (Short 2001: A-2), proposed originally by David Betson (1996). The scale equals (A + 0.8 + 0.5 ∗ (C − 1))0.7 for single-parent households and (A+0.5∗(C −1))0.7 for all other households, with A and C representing, respectively, the number of adults and children.

June 5, 2012

11:23

62

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

LIMEW, which jumped from 11% in 1959 to 17% in 1989 and then surged to 22% in 2007 (see Table 3.3). Fluctuations over time reflected the growing magnitude of wealth, as well as the stock market boom in the late 1990s, the bust in the early 2000s and its resurgence by the late 2000s.15 Another important development is the rise in the share of net government expenditures in total LIMEW, which increased from 1.8% in 1959 to 5.6% in 2007. The increase from 1959 to 1989 reflected mainly the sharp growth in transfers and, to a lesser extent, in public consumption that outstripped the growth in taxes. From 1989 to 2000, net government expenditure’s share declined as increases in taxes outpaced those of transfers and public consumption. The sharp increase after 2000 was primarily due to the increase in government transfers (from 9.2% to 11.2% of LIMEW, or by $2,600 per household between 2000 and 2007) and in public consumption (from 9.0% to 9.8% of LIMEW, or by $1,300 per household) and to a lesser extent from the decline in taxes (from 16.6% to 15.4% of LIMEW, or by $700 per household).16 In contrast, the share of base income in LIMEW, after rising from 55% in 1959 to 59% in 1982, fell off to 52% by 2007. The share of household production in LIMEW fell sharply, from 32% in 1959 to 21% in 2007. The overall change from 1959 to 2007 largely reflected the decline in hours spent on housework, particularly between 1959 and 1982 (see Table 3.2).17 There are marked differences in the relative importance of the components of LIMEW by LIMEW quintile. Income from wealth becomes progressively more important by quintile, advancing from 5% of total LIMEW for the lowest quintile to 36% for the highest quintile in 2007. The opposite is the case for net government expenditures, which plummets from 13% of LIMEW for the lowest quintile to −3% for the highest. There is much less variation across quintiles in terms of base income and household production.

15 The 2007 figure is before the collapse of the stock market in 2008 and 2009 from the “Great Recession.” On the surface, it might appear that the increase in income from nonhome wealth from 1959 to 2007 might be due, in part, to the increase of the average age of the population over this period, since in our formulation the annuity will rise if remaining life expectancy falls. However, median life expectancy rose over this period as well, so that remaining life expectancy remained relatively unchanged over these years. 16 All dollar values in the paper are in 2007 dollars, unless otherwise noted. 17 It is also of interest is that while hours of housework averaged almost exactly half of total hours worked over the 1959–2007 period, household production accounted for only about a fifth of LIMEW over the period.

b1363-ch03

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in the United States Table 3.3.

63

Composition of LIMEW by Quintile, 1959–2007. Quintiles

Lowest

Second

Third

Fourth

Highest

All

83,248 90,274 84,352 102,643 116,393 121,287 121,426

149,682 161,022 156,293 191,465 266,444 260,837 272,392

71,675 76,791 73,684 89,808 110,520 111,517 113,994

Mean LIMEW (in 2007$) 1959 1972 1982 1989 2000 2004 2007

20,124 21,997 23,586 28,376 30,552 31,117 30,819

42,837 44,942 42,852 52,045 56,670 58,534 59,093

62,482 65,701 61,329 74,507 82,535 85,769 86,224

Share of base income in LIMEW (percent) 1959 1972 1982 1989 2000 2004 2007

46.0 42.2 43.4 50.9 56.2 53.4 62.2

53.6 52.2 54.8 54.2 57.4 52.3 54.0

57.9 61.8 59.0 57.4 58.0 53.0 54.0

57.3 64.4 64.9 60.8 58.0 53.0 54.2

53.1 57.9 59.2 55.5 50.4 51.4 48.1

54.6 58.5 59.0 56.6 54.2 52.2 51.6

Share of income from wealth in LIMEW (percent) 1959 1972 1982 1989 2000 2004 2007

10.2 7.1 7.7 6.9 6.5 4.2 5.4

7.7 8.5 7.7 7.4 7.5 5.8 7.2

6.5 8.4 8.1 8.2 8.7 7.4 7.7

7.2 9.1 8.9 9.9 11.7 10.2 10.2

17.0 22.8 26.2 27.6 37.2 32.1 36.2

11.4 14.5 15.9 16.7 22.9 19.2 21.7

Share of net government expenditures in LIMEW (percent) 1959 1972 1982 1989 2000 2004 2007

11.3 27.9 30.7 21.4 18.0 22.2 13.0

7.6 17.2 18.6 16.0 13.2 19.1 17.9

3.2 7.0 12.0 10.2 9.7 15.1 14.7

1.4 2.2 3.6 4.0 5.3 11.3 10.7

−1.6 −4.2 −6.4 −6.1 −6.9 −2.7 −3.0

1.8 3.6 4.2 3.2 1.6 6.8 5.6

Share of household production in LIMEW (percent) 1959 1972 1982 1989 2000 2004 2007

32.5 22.8 18.2 20.7 19.3 20.3 19.4

31.1 22.1 18.9 22.4 21.8 22.8 20.9

32.4 22.7 20.9 24.2 23.5 24.5 23.5

Source: Authors’ calculations.

34.1 24.4 22.6 25.3 25.0 25.4 25.0

31.5 23.5 21.0 22.9 19.3 19.1 18.8

b1363-ch03

32.3 23.4 20.9 23.5 21.4 21.8 21.1

June 5, 2012

11:23

64

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

The most dramatic changes have taken place at the bottom and top of the distribution. For the bottom quintile, the share of net government expenditures, after surging from 11% in 1959 to 31% in 1982, declined to 18% in 2000 and then to 13% in 2007. In contrast, the share of base income jumped from 46% in 1959 to 62% in 2007. Likewise, the share of income from wealth dropped from 10% in 1959 to 5% in 2007,18 while the share of household production also fell during the period, from 33% to 19%. There was a sizeable increase in the share of income from wealth in LIMEW for the top quintile, which rose from 17% in 1959 to 36% in 2007. This was accompanied by a decline in the relative importance of base income and household production. Net government expenditure also fell off, from −1.6% in 1959 to −3.0% in 2007. In sum, for those at the bottom, there was greater reliance on base income (mainly labor income) over time, while for those at the top, income from wealth became significantly more important, and base income and household production less important. B. The Middle Class We now turn to a closer examination of the changes in the third quintile of the LIMEW distribution, because the trends in the mean value of LIMEW for this quintile provide a close approximation to the changes in the median LIMEW for all households. Focusing on the mean LIMEW for the third quintile allows us to assess the roles played by different components of the LIMEW in the well-being of the average household. The third quintile is sometimes considered the “middle class,” and we follow that convention here. As noted before, median LIMEW in 1982 was slightly lower than in 1959. The same pattern is also observed for mean LIMEW for the third quintile. The decline in the latter was partially due to the decline in household production from 32% to 21% or by $7,000 (see Tables 3.3 and 3.4, and Fig. 3.1). Decreases in housework hours and the unit value of housework represented 28% and 72% of the decline, respectively (estimates not shown). This decline was partially offset by the robust growth in net government expenditures, which climbed from 3% to 12% of LIMEW, or by $5,300. Another reason for sluggish growth in LIMEW over this period was 18 This

large decline can be traced to the rising share of debt in the household portfolio of the bottom quintile.

b1363-ch03

June 5, 2012 11:23

Table 3.4.

Contribution by Component to the Change in LIMEW and MI of the Middle Quintile, 1959–2007 (Percent). 1959–1972 LIMEW

1982–1989

1989–2000

LIMEW

MI

LIMEW

MI

10.5 2.2

−6.7 −0.9

−7.0 2.2

10.7 1.9

12.4 0.6

6.9 1.5

8.0

−0.5 −0.3 4.1

3.9

0.5 1.4 0.5

−0.3

−0.6 2.1 0.6

8.0

4.1 −0.6

3.9

−0.9

21.5

2.0 1.6

MI

LIMEW

MI

LIMEW

MI

LIMEW

MI

6.0 −1.9

−3.0 −1.1

−3.6 −1.3

1.3 0.4

2.5 0.4

16.6 4.1

23.4 1.7

3.1

−0.6 −0.4 6.0

−0.1 0.5 3.3 −0.3

−2.2

−0.3 4.4 17.1

18.0

3.3

−2.2

3.1

−3.0 1.8 12.7

10.8

2.6 97.4

−15.5 115.5

2.7 0.6

7.2

3.9

1.3 0.8 −2.3 −0.9

2.7 2.0

Addendum: Decomposition of the changes in the value of household production Total change −5,298 −2,099 5,185 1,369 ($2007) Contribution (percentage) to the change from: Change in hours −45.4 8.1 Change in unit −54.6 −108.1 value

1959–2007

−1.6

0.5

1,643

−738

31.2 68.8

94.8 5.2

17.9 9.5

18.0

−10.3 0.1 0.6

38.0

43.2

61

3231.1 −3331.1

65

Notes: 1. Middle class refers to the third quintile of the measure. The numbers shown in the line labeled “Total” refers to the percent change in the third quintile’s average between the two years. 2. Contributions of individual components add up to the total.

Growth and Inequality: An International. . .

−6.7

−0.3

−2.2 8.5

0.6 −3.2 20.7

0.7 2.0

LIMEW

2004–2007

9in x 6in

MI

2000–2004

Growth and Inequality in the United States

Base Income 7.1 Income from 2.3 wealth Home wealth 1.6 Non-home wealth 0.7 Net government 4.2 expenditures Transfers 5.3 Public 4.5 consumption Taxes −5.6 Household −8.5 production Total 5.2

1972–1982

b1363-ch03

June 5, 2012

11:23

66

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

Fig. 3.1. Contribution to the Percentage Change in the Third Quintile’s Mean LIMEW (Percent).

the drop in base income between 1972 and 1982, (from 62% to 59%, or by $4,400), that wiped out the $4,400 gain in the 1959–1972 period. The composition of LIMEW for the middle quintile remained relatively stable from 1982 to 1989 and the very high rate of growth of the mean LIMEW of the middle quintile (22%) was due to relatively balanced growth in all four components. In particular, average base income for the middle quintile rose by $6,600, and household production increased by $5,200. Most of the gain (98%) in household production was due to a rise in the unit value of housework. The growth of the mean LIMEW of the middle quintile slowed between 1989 and 2000. The composition of LIMEW of the middle quintile was also relatively stable over this period and the slowdown was attributable to the reduced growth of all components. However, between 2000 and 2004, the growth of mean LIMEW of the middle quintile slowed to a crawl, gaining only 3.9%. Over these years, the composition of LIMEW changed dramatically in favor of net government expenditures, which rose by $4,900, while base income and income from wealth declined by $2,500 and $900, respectively. These trends were largely reversed from 2004 to 2007, with net government spending showing negative growth and base income and

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

67

income from wealth showing positive gains. However, household production also declined over these years, and mean LIMEW grew by only 0.5%. Mean LIMEW of the middle quintile grew by 38% (as about the same as median LIMEW for all households) over the 1959–2007 period. Of this gain, 17.1 percentage points (or 45%) was due to the increase in net government expenditures (Table 3.4 and Fig. 3.1) in the form of an increase in transfers (18 percentage points) and public consumption (10 percentage points), while an increase in the tax burden subtracted 10 percentage points. The increase in base income added another 17 percentage points to the growth in LIMEW of the middle class, while gains in income from wealth contributed only four percentage points. Household production barely made any contribution toward the growth of middle class LIMEW over the period. Table 3.4 also presents a growth decomposition of the average MI for its middle quintile. For MI, 54% of its 43 percentage point gain was attributable to the growth of base income and 42% to increased cash transfers. According to the LIMEW measure, the public sector was the leading source of the growth in the standard of living of the middle class between 1959 and 2007. We already saw (Table 3.3) that the share of net government expenditures in the LIMEW of the middle quintile rose dramatically from 3% to 15% between 1959 and 2007. Government expenditures for the middle class grew much faster than their LIMEW over the period: As a percentage of LIMEW, expenditures rose by 17 percentage points from 12% to 29% between 1959 and 2007. Much of this increase was driven by the growth in transfers which, as a percentage of LIMEW, rose from 4% to 16% over the period, an increase of 12 percentage points. In turn, two-thirds of the increase in the percentage share of transfers in LIMEW occurred as a result of the expansion of transfer programs which did not exist in 1959 (Medicare, Medicaid and Earned Income Tax Credit (EITC), etc.). Public consumption, the other type of government expenditures, also increased much faster than LIMEW but slower relative to transfers. This was reflected in its percentage share of LIMEW, which rose from 8% to 13% between 1959 and 2007. Out of the 5 percentage points increase in the share of public consumption in LIMEW, 3 percentage points came from the increasing share of education expenditures in LIMEW. The increase in labor income was a close second to net government expenditures in contributing to the growth in the standard of living of the middle class between 1959 and 2007, while gains in income from wealth were a distant third. In contrast, the two leading factors accounting for the

June 5, 2012

11:23

68

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson Table 3.5. Share of Each Quintile and the Top 5% in Aggregate Income (In Percent), 1959–2007. Quintiles

1959 LIMEW MI 1972 LIMEW MI 1982 LIMEW MI 1989 LIMEW MI 2000 LIMEW MI 2004 LIMEW MI 2007 LIMEW MI

1

2

3

4

5

Top 5%

5.6 3.4

12.0 10.9

17.4 17.3

23.2 24.3

41.8 44.0

17.1 17.3

5.7 3.7

11.7 9.7

17.1 17.4

23.5 25.2

41.9 43.9

16.8 16.2

6.4 4.0

11.6 10.1

16.6 16.6

22.9 24.7

42.4 44.7

17.6 16.4

6.3 3.9

11.6 9.7

16.6 16.2

22.9 24.5

42.6 45.6

17.5 17.0

5.5 3.6

10.3 8.9

14.9 14.8

21.1 23.1

48.2 49.7

23.5 21.8

5.6 3.4

10.5 8.7

15.4 14.7

21.7 23.3

46.8 50.0

22.6 21.6

5.4 3.4

10.4 8.7

15.1 14.8

21.3 23.5

47.8 49.6

23.1 21.0

gain in mean LIMEW for all households were increases in base income and gains in income from wealth. According to MI, most of the growth in middle class living standards was due to rises in labor earnings over the period.

3.4. Economic Inequality As the final part of our analysis, we turn our attention to overall inequality. We investigate both differences in measured inequality among our various measures, trends over time, and the sources of rising inequality. It is striking that the income shares of the middle three quintiles were lower in 2007 than in 1959 in both the LIMEW and MI distributions (see Table 3.5).19 The change in the division of the economic pie favored the 19 The quintiles of each income measure are defined by ranking households according to that measure. Therefore, in general, a given quintile of LIMEW need not be made up of the same households as the same quintile of MI.

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

69

top quintile and the top 5% far more than the bottom quintile. The bottom quintile showed a slight drop in its share of total LIMEW but no change in its share of total MI, while the top quintile’s share of aggregate LIMEW and MI went up by 6.0 and 5.6 percentage points, respectively. The increase in the share of the top quintile and the top 5% was relatively moderate in terms of both aggregate LIMEW and MI from 1959 to 1989, followed by a big surge from 1989 to 2000, and then little change between 2000 and 2007. The bottom quintile also saw modest growth in its share till 1989, but lost ground thereafter. In all the years studied here, the top quintile fared better according to MI than LIMEW in terms of its share in the overall pie (50% vs. 48% in 2007) and the bottom quintile received a larger share in LIMEW than MI (5.4% vs. 3.4% in 2007). However, with the exception of 1959, the top 5% had a higher share of the total in LIMEW than MI. This reflects, in the main, the greater inequality at the top in LIMEW due to our treatment of income from wealth. In fact, in 2007, income from wealth comprised 56% of the LIMEW of the top 5% (see Table 3.1). The decline in the income share of the middle class (the third quintile) between 1959 and 2007 was similar in LIMEW and MI (2.3 and 2.5 percentage points). The share of the second quintile fell by 1.6 percentage points in LIMEW and 2.2 percentage points in MI, while that of the fourth quintile fell by 1.9 and 0.8 percentage points in LIMEW and MI, respectively. The most pronounced decline in the shares of the middle three quintiles happened during the 1989–2000 period, though the 2000–2004 period was almost as bad in terms of LIMEW and slightly worse in terms of MI for the second and third quintiles. Consistent with the data on quintile shares, MI shows a larger degree of inequality than LIMEW according to the Gini coefficient (see Table 3.6). The lower inequality in LIMEW compared to MI is primarily

Table 3.6. Economic Inequality by Measure, 1959 to 2007 (Gini Coefficient × 100).

LIMEW MI

1959

1972

1982

1989

2000

2004

2007

36.1 40.3

36.3 40.7

36.0 40.9

36.3 41.8

42.3 46.0

41.0 46.5

42.0 46.2

31.9 40.0

38.2 44.1

36.5 44.5

37.8 44.3

Equivalence scale adjusted measures Equivalent LIMEW Equivalent MI

32.8 40.1

31.7 38.9

30.8 39.1

June 5, 2012

11:23

70

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

due to the inclusion of public consumption and household production. Equivalence-scale adjustment lowers measured inequality in both LIMEW and MI. This is not surprising in light of the well-known correlation that exists in the data between household size and income. The bottom rungs of the income distribution tend to have more single-person households and smaller families than the higher rungs. Additionally, in the case of LIMEW, public consumption and household production display strong positive correlation with household size. Consider, for example, households with school-age children. The single largest component of public consumption is public education, for which we have imputed per-pupil expenditures as a part of LIMEW. Households with more school-age children would, in general, have larger amounts of public consumption allocated to them. Similarly, hours spent on household production also tend to increase with both the number of adults and the number of children at home, thus producing a positive correlation between household size and value of household production.20 The Gini coefficients indicate a considerably higher level of inequality in 2007 than 1959 for both LIMEW and MI. This result is also consistent with the pattern of changes in quintile income shares discussed earlier. The increase was about the same for MI (5.8 Gini points) and LIMEW (5.9 Gini points).21 Neither measure shows considerable change in inequality between 1959 and 1972. According to MI, almost all of the increase in inequality occurred from 1989 to 2000. In somewhat similar fashion, the LIMEW measure shows almost no change in inequality from 1959 to 1982, a modest rise from 1982 to 1989 (0.2 point increase) and then a large spurt of 6.0 points from 1989 to 2000, followed by little change between 2000 and 2007. Decomposition of inequality by income components (or sources) is a standard technique used to assess the amount of inequality accounted for by individual components in the total amount of inequality (Lerman, 1999). The decomposition results are not conclusive evidence on causality.

20 A separate issue concerns the applicability of standard equivalence scales to income measures that include non-market components such as public consumption and household production. This is an area that requires further research. 21 Time trends are quite similar for equivalence-scale adjusted measures. However, the reduction in measured inequality as a result of the equivalence-scale adjustment is larger for all the other years relative to 1959, perhaps reflecting the fact that the correlation between household size and income was relatively smaller in 1959. Consequently, the overall increase in measured inequality of the equivalent income measures between 1959 and 2007 was smaller than the corresponding unadjusted measures.

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch03

Growth and Inequality in the United States

71

Table 3.7. Decomposition of Inequality by Income Source and Income Measure (Gini Points × 100). A. Contribution to Inequality

LIMEW Base money income Income from wealth Imputed rent Annuities Net government expenditures Transfers Public consumption Taxes Household production Total Money Income Base money income Property income Transfers Total

B. Contribution to the Change In Inequality

1959

2000

2007

1959–2000

2000–2007

1959–2007

19.7 6.4 1.2 5.2 −1.4

20.9 17.1 1.8 15.3 −3.9

19.4 16.5 1.5 14.9 −1.8

1.2 10.7 0.6 10.1 −2.5

−1.6 −0.6 −0.2 −0.4 2.2

−0.4 10.1 0.4 9.7 −0.4

0.8 1.8 −3.9 11.4

1.0 2.4 −7.3 8.2

1.4 2.7 −5.9 8.0

0.2 0.7 −3.4 −3.2

0.4 0.3 1.4 −0.2

0.6 1.0 −2.0 −3.4

36.1

42.3

42.0

6.2

−0.2

5.9

38.6 1.5 0.2

43.6 3.4 −1.0

43.7 3.4 −0.9

5.0 2.0 −1.3

0.1 0.0 0.1

5.1 1.9 −1.2

40.3

46.0

46.2

5.7

0.1

5.8

Note: Contribution of each income source is expressed in Gini points multiplied by 100. The numbers shown in the row labeled “Total” refers to the Gini ratio of the income measure

However, they do identify the contribution of individual components to overall inequality. The degree of inequality accounted for by a component is the product of that component’s concentration coefficient and its share in income (Kakwani, 1977). The contribution of the components to the change in the Gini coefficient between two years is calculated as the difference between the contribution to inequality accounted for by that component in the later year and the earlier year (Table 3.7). We begin by looking at the contributions of the various components of LIMEW to the level of overall inequality of LIMEW. In 2007, the leading contributor was base income, which accounted for 46% of the overall Gini coefficient for LIMEW. Income from wealth was second, accounting for 39% and followed by household production (19%). Net government expenditures actually made a negative contribution of −4.2%, mainly due to taxes, −14%. A comparison with the decomposition of MI is useful. In 2007 base income, the first component of the two measures, accounted for 94% of

June 5, 2012

11:23

72

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

the overall Gini coefficient for MI. The contribution of base income to the level of inequality was thus markedly lower in LIMEW than MI. The lower contribution is mainly due to the fact that base income constitutes a smaller share of LIMEW than of MI. The reason is that, as noted earlier, MI is considerably smaller in size than LIMEW and hence the lower share of base income (the component that is common to both measures) in the latter is to be expected. The concentration coefficient of base income is also lower in LIMEW than in MI. This difference reflects the fact that the correlation between the rank in the base income distribution and rank in the total income distribution is weaker in LIMEW than in MI.22 There are three forces that weaken the correlation. Net government expenditures in LIMEW heavily favor those at the bottom of the earnings distribution and those with no earnings at all. Second, while there is some amount of positive correlation between the value of household production and earnings (partly as a result of our valuation schema), household production still reduces the gap between earners and non-earners and can change the rankings of households in the distribution.23 Third, as discussed above, income from non-home wealth is heavily skewed toward the elderly, a group with relatively low conventional income (MI). With regard to the second component, the contribution of income from non-home wealth to overall inequality in LIMEW in 2007 was substantially higher (36%) than the contribution made by property income to MI inequality (only 8%). Both the concentration coefficient and income share were higher in LIMEW than MI. The lower share of property income in

22 The

concentration coefficient of an income source j, denoted as cj can be expressed as: cj = rj gj , where rj = cov(yj , F )/cov(yj , Fj ), gj is the Gini coefficient of income source, yj is the amount of income from the income source, and Fj and F are the cumulative distributions of the income source and total income (Lerman and Yitzhaki, 1985). Since the Gini of base income is identical in both LIMEW and MI by construction, the difference in its concentration coefficient is solely due to the difference in the “Gini correlation” between the two variables, rj . 23 Consider the following example of a household with no earnings and a household with some earnings. Assuming that both have positive hours of household production, an income measure that augments earnings with the value of household production would reduce the gap between the two households. The potential “re-ranking” effect of household production can be seen by considering an example of two households that have equal earnings but only one of them engages in household production. Earnings augmented by household production would now place one household behind the other in the distribution. While these are extreme examples, they do illustrate how the incorporation of household production can reduce gaps between households and change the rankings of households in the distribution.

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

73

MI is partly a result of the well-known shortcomings of income surveys — namely, the underreporting of property income and lack of sufficient coverage of very high-income families. However, the more significant factor behind the difference in the share is the lifetime annuity method itself. The entire amount of non-home wealth is annuitized, including assets that do not yield any current income by their very nature (e.g. the cash surrender value of defined contribution pension plans) and assets that might have yielded no income or even generated losses in a given year (e.g. rental real estate). Furthermore, given the strong positive association between age and wealth, the average effective rate of return (i.e., the ratio of annuity to non-home wealth) will generally be higher than the average rate of return implied by the actual property income receipts. The concentration coefficient of income from non-home wealth in LIMEW was also higher than that of property income in MI.24 The reason appears to be the difference in the extent to which households in the different portions of the respective income distributions rely on this source of income. The richer households in the LIMEW distribution tend to have higher income from non-home wealth. In contrast, the richer households in the MI distribution tend to have higher labor earnings, rather than higher property income. As a result, the share of income from non-home wealth in total income tends to rise in a much steeper manner in the LIMEW distribution than the MI distribution. This difference explains the relatively higher concentration coefficient for income from non-home wealth in LIMEW in comparison to property income in MI. The third component, net government expenditures, had a larger inequality-reducing effect in LIMEW (−4%) than transfers had in MI (−1%). The reason is that net government expenditures incorporate taxes, which reduce inequality. By far the biggest contributor to rising LIMEW inequality over the years 1959 to 2007 was income from wealth, particularly non-home wealth. This component accounted for 10.1 Gini points out of a total increase of 5.9 Gini points, or 171%. The other components made negative contributions — base income (−6%), net government expenditure (−6%), and household production (−58%). In contrast, the contribution of base income to the inequality of MI grew between 1959 and 2007 and explained 71% of the change in the Gini coefficient for MI. The main reason was the increase 24 In

contrast, the Gini coefficients for income from nonhome wealth and property income are similar across the years studied here, around 0.91.

June 5, 2012

11:23

74

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

in the concentration coefficient of base income (from 0.42 to 0.49). This was partly offset by a decline in the share of base income in total MI from 91% to 84%. The amount of inequality contributed by income from non-home wealth to overall LIMEW inequality was 14.9 Gini points higher in 2007 than in 1959, mainly because its share of income more than doubled over the period from 8% to 19% of LIMEW. In contrast, the contribution of property income to inequality in MI was only 3.4 Gini points higher in 2007 than in 1959. The reason is that though the concentration coefficient of property income increased sharply over these years, the share of property income in MI grew only slightly between the two years. The evidence thus suggests that although LIMEW and MI show comparable increases in inequality over the 1959–2007 period, the principal source of the increase is different in the two measures: Changes in the level and distribution of income from non-home wealth account for the bulk of the growth in the inequality of LIMEW, while for MI, base income accounts for by far the largest part in the increase in MI inequality. Net government expenditures helped ameliorate the increase in inequality in LIMEW and transfers served the same function for MI. However, the moderating effect of net government expenditures appears to be more important in LIMEW in comparison to transfers in MI: A reduction of 1.8 Gini points between 1959 and 2007, compared to a reduction of 0.6 points in MI. In both cases the contribution of the component became more negative over time. In the case of LIMEW, the share of net government spending in total LIMEW advanced from 1.8% to 5.6% from 1959 to 2007 but while the concentration coefficient of transfers declined, that of public consumption increased and that of taxes fell. In the case of MI the larger effect of transfers in the later years reflects both its increased share in total MI (from 5.1% to 11.3%) and the change in its concentration coefficient (from 0.051 to −0.054). Household production was the largest single component restraining the growth of inequality of LIMEW between 1959 and 2007. The decline in its contribution (of 3.4 Gini points) stemmed entirely from the decline in its share of LIMEW, since its concentration coefficient actually showed a modest increase. The share of household production fell by 11 percentage points, from 32% in 1959 to 21% in 2007, while the concentration coefficient was 0.35 in 1959 and 0.38 in 2007. As noted before, there was a sizeable decline in the overall hours spent on household production activities and

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

75

this development is mirrored in the fall in the share of household production in LIMEW. During the 2000s (2000–2007), LIMEW showed a slight decrease in inequality, while MI inequality increased slightly. However, the reasons are different in the two cases. In the case of LIMEW, there was a 1.6 Gini point decline in the contribution to inequality made by base money income, about equally a reflection of a decline in its concentration coefficient and its share in total LIMEW. This decline was offset by a 2.2 Gini point increase in the contribution made by net government expenditures. The smaller negative contribution made by net government spending to LIMEW inequality in the later year was due to a rise in both the concentration coefficient and the income share of transfers and public consumption, and a fall in the concentration coefficient (that is, a decrease in the progressivity) of taxes. For MI, base income contributed to the slight increase in inequality. Our results also provide an important contrast with those of Piketty and Saez (2003). Their data source is the Internal Revenue Service Statistics of Income database and their income concept is adjusted gross income (AGI) less realized capital gains.25 A key argument made by Piketty and Saez is that the surge in top income shares since the early 1970s was due to the relatively sharp increase of top wages as reflected in the growing share of labor income, at the expense of capital income, in the total income of the rich. (Piketty and Saez, 2003: pp. 17, 37). We also find a sharp decline in the share of income from wealth in the total income of the top decile on the basis of MI, but no such decline occurred on the basis of LIMEW. We also find that according to MI, 87% of the increase in overall inequality was due to the rise of inequality of base income (mainly labor earnings) from 1959 to 2007. In contrast, according to LIMEW, base income actually made a negative contribution (−6%) to the increase in overall LIMEW inequality, while income from wealth accounted for 171%26 25 Piketty and Saez also exclude some other small items in AGI such as taxable social security income. The reference distribution is the distribution of income among taxpayers (tax units). However, the number of tax units in each quantile is defined relative to the total number of potential tax units (had everyone been required to file a tax return) and the share of each quantile is defined relative to the NIPA aggregate of personal income, after adjustments required for comparability with the AGI concept excluding realized capital gains. 26 See Wolff and Zacharias (2009) for more details on the comparison over the 1989 to 2000 period.

June 5, 2012

11:23

9in x 6in

76

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

3.5. Conclusion We find that median income grew sluggishly over the 1959 to 2007 period by any measure, particularly when compared to the annual growth in GDP per capita (2.3%). The annual growth rate in median LIMEW and MI were, respectively, 0.67% and 0.63%. This congruence between LIMEW and the conventional measure, MI, masks important differences by subperiod. Median LIMEW showed much slower growth from 1959 to 1982 than median MI. Subsequently, median LIMEW grew faster from 1982 to 2007. Trend differences in median well-being can be traced to differences in the composition of the measures. Household production —included in LIMEW but not in MI — contributed greatly toward explaining the slower growth of the middle quintile’s LIMEW in the 1959–1982 period. A marked increase in net government expenditure (mainly a large increase in transfers) contributed substantially to the increase in the LIMEW of the middle quintile between 1982 and 2007. For the population as a whole, the most notable compositional change in LIMEW was the growth in the share of income from wealth, particularly between 1989 and 2000. Indeed, from 1959 to 2007, its share almost doubled. This by itself would have led to rising inequality in LIMEW. However, this development was partially offset by a rise in the share of net government expenditures in LIMEW from 1.8% to 5.6%, which mitigated the rise in inequality. The compositional change differed between the top and bottom quintiles of the LIMEW. Between 1959 and 2007, households at the bottom became more reliant on base income (mainly consisting of labor income) and, somewhat, on net government expenditures. On the other hand, income from wealth more than doubled as a share of LIMEW for those at the top. According to both MI and LIMEW, there was a substantial growth of inequality over the years from 1959 to 2007. Time trends were also similar for the two measures, though for different reasons. Both measures show a modest rise in inequality from 1959 to 1989 and then a large spike from 1989 to 2000 followed by little change through 2007. Decomposition analysis shows that income from non-home wealth made by far the largest contribution to the increase in inequality between 1959 and 2007 recorded for LIMEW. In contrast, in the case of MI, the principal factor behind the increase in inequality was the rising contribution from base income. These two factors were particularly important in explaining the inequality surge of their respective measures during the 1990s. Net government expenditures

b1363-ch03

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in the United States

b1363-ch03

77

helped moderate the increase in inequality between 1959 and 2007 in the case of LIMEW. During the 2000s, the Gini coefficient for LIMEW increased slightly. The small change in inequality over these years was due to two offsetting factors — a declining contribution to overall inequality from base income and a rising contribution from net government expenditures. The latter, in turn, was traced to a rise in both the concentration coefficient and the income share of transfers, a rise in the income share of public consumption, and a fall in the concentration coefficient (that is, a decrease in the progressivity) of taxes. All in all, the government sector has made a remarkable contribution to sustaining middle class living standards and reducing inequality. Net government expenditures augmented the base income of the middle quintile by 27% (compared to 6% overall). The large relative contribution made by net government spending to the well-being of the middle class offset to some extent their relative shortfall in terms of income from wealth. Net government spending also helped restrain the growth of inequality before 2000 and helps to explain why inequality in LIMEW was uniformly less than that of MI. References Betson, D. (1996). “Is Everything Relative? The Role of Equivalence Scales in Poverty Measurement.” Poverty Measurement Working Paper, Washington, DC: U.S. Bureau of the Census. Canberra Group, (2001). Expert Group on Household Income Statistics: Final Report and Recommendations. Ottawa: Canberra Group. Hicks, U.K. (1946). “The Terminology of Tax Analysis.” The Economic Journal, 56(221), 38–50. Kakwani, N.C. (1977). “Applications of Lorenz Curves in Economic Analysis.” Econometrica, 45(3), 719–727. Kuznets, S.S., Epstein, L. and Jenks, E. (1941). National Income and Its Composition, 1919–1938. New York: National Bureau of Economic Research. Lakin, C. (2002). “The Effects of Taxes and Benefits on Household Income, 2000–2001.” Social Analysis and Reporting Division, Office for National Statistics, U.K. http://www.statistics.gov.uk/. Landefeld, J.S. and Stephanie, H.M. (2000). “Accounting for Non-market Household Production within a National Accounts Framework.” Review of Income and Wealth, 46(3), 289–307. Lerman, R.I. and Yitzhaki, S. (1985). “Income Inequality Effects by Income Source: A New Approach and Applications to the United States.” The Review of Economics and Statistics, 67(1), 151–156.

June 5, 2012

11:23

78

9in x 6in

Growth and Inequality: An International. . .

E. N. Wolff, A. Zacharias and T. Masterson

Masterson, T. (2010). “Quality of Match for Statistical Matches Used in the 1992 and 2007 LIMEW Estimates for the United States.” Working Paper No. 618, September. Annandale-on-Hudson, N.Y.: Levy Economics Institute of Bard College. National Research Council, (2005). Beyond the Market: Designing Nonmarket Accounts for the United States. Panel to Study the Design of Nonmarket Accounts, Abraham K.G. and Mackie, C. (Eds.) Washington, DC: The National Academies Press. Organisation for Economic Co-operation and Development, (1995). Household Production in OECD Countries: Data Sources and Measurement Methods. Paris: OECD. Piketty, T. and Saez, E. (2003). “Income Inequality in the United States, 1913–1998.” Quarterly Journal of Economics, 118(1), 1–39. Ruggles, P. and O’Higgins, M. (1981). “The Distribution of Public Expenditure among Households in the United States.” Review of Income and Wealth, 27(2), 137–164. Short, K. (2001). Experimental poverty measures 1999. Current Population Reports No. 216. Washington DC: U.S. Dept. of Commerce, Economic and Statistics Administration, Bureau of the Census. Wolff, E.N., Zacharias, A. and Caner, A. (2004). Levy Institute Measure of Economic Well-Being, Concept, Measurement and Findings: United States, 1989 and 2000. Annandale-on-Hudson, N.Y.: Levy Economics Institute of Bard College. Wolff, E.N. and Zacharias, A. (2007a). “The Levy Institute Measure of Economic Well-Being: United States, 1989 to 2001.” Eastern Economic Journal, 33(4), 443–470. Wolff, E.N. and Zacharias, A. (2007b). “The Distributional Consequences of Government Spending and Taxation in the U.S., 1989 and 2000.” Review of Income and Wealth, 53(4), 692–715. Wolff, E.N. Zacharias, A. and Masterson, T. (2009). “Long-Term Trends in the Levy Institute Measure of Economic Well-Being (LIMEW), United States, 1959–2004.” Working Paper No. 556, January. Annandale-on-Hudson, N.Y.: Levy Economics Institute of Bard College. Yeung, W.J. and Stafford, F. (2003). Intra-family child care time allocation. Paper presented at 2003 Annual Meeting of Population Association of America. Minneapolis, Minnesota.

b1363-ch03

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Chapter 4 GROWTH AND INEQUALITY IN INDIA Vamsi Vakulabharanam∗ In this chapter, I investigate the question of whether class structure matters in understanding the increasing inequality in India during the period of economic liberalization. I argue in the paper that almost the entire increase in the overall Indian inequality during this period is explained by the rising inequality between classes rather than within them (unlike, for instance, in China, where the entire increase in inequality during the same period is explained by the increase in the within class inequality). There is now clear evidence from the National Sample Survey quinquennial household consumer expenditure surveys conducted in 1993–1994 and 2004–2005 that increased distance between urban elites (owners, managers, and professionals), rural rentier classes (such as moneylenders and absentee landlords) who are more stratified at the top, and unskilled urban workers, marginal farmers and agricultural workers, who are increasingly more stratified at the bottom helps us understand the distributional dynamics of the Indian growth story. Also, the urban sector is both becoming more unequal among its own classes while it is also getting more stratified vis-` a-vis the rural classes. I analyze the class structures in India and decompose (using the methodology of Yitzhaki, 1994) the overall inequality into inter-class and intra-class terms. Finally, I explain these changes by analyzing the Indian policies during this period.

4.1. Why Focus on Class Structure? It is now well known that Indian economy has been experiencing relatively high annual growth rates consistently exceeding 5% in every decade since 1980 (e.g. Dutt and Rao, 2000; Economic Surveys, 2000–2009, Ministry of Finance, Government of India). While there was a major regime shift ∗Associate

Professor, Department of Economics at the University of Hyderabad, India. [email protected] 79

b1363-ch04

June 5, 2012

11:23

80

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

in 1991 when India introduced market-oriented economic reforms in most of its sectors and increased its openness to the global economy, Indian growth story continued. It is also acknowledged now that the income distribution has taken a turn for the worse since 1991 (e.g. Himanshu, 2007). This reverses a trend of declining or stagnant inequality during the 1980s, when a growth spurt first occurred in India with a strong push from the state. However, trends in overall Indian inequality have been understudied except at a fairly general level such as the inter-household, inter-sectoral (rural vs. urban) or inter-state level (for e.g. Chaudhuri and Ravallion, 2006; Himanshu, 2007). While the already published work helps us understand inequality at a broad level, it does not illuminate the dynamics of how different groups/classes have begun to witness increases in relative prosperity or relative pauperization. Understanding these dynamics is crucial both to examine whether class structure matters in explaining distributional changes as well as to show whether particular classes or occupations have been included/excluded in the growth process. This chapter sheds light precisely on these questions by analyzing in class terms, the Indian household consumer expenditure surveys conducted by the National Sample Survey Organization (NSSO). After economic reforms were introduced in India, there have virtually been no empirical large-scale economic studies on India that have worked with the construct of Class to examine its impact on economic and social change. Especially since the 1990s, this absence becomes glaring as sociologists, political scientists and other social observers have argued that the class structure in India has not only undergone a “great transformation”1 but also that it has been central to understanding Indian economic policy initiatives during the last two decades (Chatterjee, 2008). Since policy initiatives are not undertaken in vacuum, we know that different class interests pressure the state to pursue different sets of policies, whether they are rich minorities or poor majorities. State policies that support the former interests may help promote inter-class inequality, and those that are guided primarily by the latter interests tend to be regarded by some as populist (like the National Rural Employment Guarantee Scheme (NREGS) in India) but usually tend to put a lid on growing income gaps. So far, while careful social observers (e.g. Patnaik, 2009) have pointed out that the 1 Also see a forthcoming volume — Great Transformation: Understanding India’s Political Economy” edited by John Harriss, Stuart Corbridge, Sanjay Reddy and Sanjay Ruparelia in which “class” among other important axes, received serious attention.

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

81

Indian economic policies after 1991 have tended to exacerbate inequalities between classes, there has been no large empirical study that establishes this story convincingly.2 In this paper, I use the consumer expenditure household surveys from 1993–1994 and 2004–2005 to examine whether the rising Indian inequality has been due to increased inter-class or intra-class inequalities. Also, I estimate the consumption levels of different classes in terms of consumer expenditure and examine how these levels are undergoing changes during this period. Using Yitzhaki methodology (1994) to decompose the Gini coefficient into inter-class and intra-class components, I am also able to show tendencies towards stratification in the Indian economy that shed light on the nature of the rising inequality in India.3 The rest of the paper is organized as follows. Section 4.2 presents the theoretical framework employed in order to define classes and a class structure. Section 4.3 discusses the data used to describe classes and make comparisons. It also discusses the limitations of the data sources employed. Section 4.4 presents the main results using the Indian class structure. Section 4.5 describes the Yitzhaki method briefly and discusses the results of the Yitzhaki decomposition of the Gini coefficient for India. Section 4.6 explains the results of Secs. 4.4 and 4.5 and concludes.

4.2. Defining a Class Structure for India There have been many frameworks to define classes in modern societies. While in popular parlance, class is defined in income terms (rich, middle, and poor or with many gradations in between such as upper middle class, lower middle class etc.), for many social scientists, this is not a satisfactory device since it uses the outcome i.e., income, of various social dynamics to define class. Other frameworks have tended to focus on economic, political, and consciousness based notions to define class (e.g. Thompson, 1963). Usually sociologists have tended to focus on notions of authority/power and status to define and embellish class even when they are using explicitly 2 In a different work where Indian and Chinese class structures are compared, it is shown that while the increase in Indian inequality between 1993–1994 and 2004–2005 can be attributed to the increase in the between class inequality, the increase in Chinese inequality between 1995 and 2002 can be attributed to an increase in the within class inequality. See Vakulabharanam, Zhong and Xue (2009). 3 This decomposition methodology has been applied in two previous studies — Wolff and Zacharias (2009) in the U.S. context and Zacharias and Vakulabharanam (2009) in the context of caste groups and Indian wealth inequality.

June 5, 2012

11:23

82

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

economic class categories (see Wright, (1997) for a contrast of the Marxian and Weberian class schemas). In some recently popular frameworks in Indian nationalist historiography such as subaltern studies, terms such as elite/subaltern have been deployed which are primarily power-based or consciousness-based categories. In the economics literature that excludes the income-based definitions, the towering intellect around which class frameworks have been defined is that of Marx. In this context, the predominant use of class among Marxist scholars has been around the idea of ownership. Marx theorizes the transition to capitalism in England, wherein he makes a stark distinction between those that come to own property through force and plunder, and the masses that are expropriated from their property and are forced to sell mainly their ability to perform labor (labor power). Marx builds on this description of class in the pages of Capital Volume I (1867:1976), Volume II (1894:1978) and Volume III (1894:1981) where he argues that class is defined at the site of production i.e. a class of people/workers that perform labor and earn wages while they also produce economic surplus/profits for the entire capitalist society, while the other classes appropriate and live off this surplus. Marx expands on these latter classes in Capital, Volume III and shows how it is not merely the industrial or agricultural capitalists that live off this surplus but also other classes such as moneylenders (receive interest), merchants (receive trading margins), landlords (receive rent), shareholders (receive dividends), managers (receive salaries) and the state (receives taxes) amongst others. While Marx appreciates the heterogeneity of the working class in his writings, he does not overly dwell on the differences within the working class as he was more focused on teasing out the distinctions between the working class and all the other classes that live off the surplus that is produced by the working class. In this chapter, I resort to the use of two class schemas. The first schema uses broad consciousness categories (applicable to India) to arrive at simplified class categories. This schema identifies only two classes in the urban areas viz., elite (which includes owners, managers, and professionals) and the workers (other than professionals). In rural areas, four classes are identified keeping in mind the differences between agricultural and non-agricultural workers. These four classes are — rural elite (both the big farmers and the non-agricultural elites that include owners, moneylenders, professionals, and absentee landlords), rural non-agricultural workers, small farmers, and agricultural workers. The whole schema consists of six

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

83

classes and relies primarily on how society broadly views classes in terms of self-identification of status and power. The second schema uses Marx’s latter definition to define a broad distinction between the working class and the other classes in the urban sector. The working class, which is defined by its role in production, is quite heterogeneous in most modern societies, and it is no different in India. There is a widely accepted distinction between skilled and unskilled work. The skilled are to be further sub-divided into professionals and skilled workers.4 This makes for a three-tier structure within the working class — professional, skilled, and unskilled. Generally, it is also acknowledged that there are fundamental distinctions between working people in manufacturing and services.5 Once this distinction is built into the class schema, I arrive at six sub-classes within the working class. Of the non-workers, given the limitations of data and the occupational categories available in the datasets (discussed in the next section), I arrive at two composite classes viz., owners and managers in the formal sector, and owners and managers in the informal sector. Managers are included alongside owners since their role in the labor process is primarily supervisory and they live off the surplus value that productive workers produce. In the rural sector, the Indian class structure is primarily defined by the failure of a crucial state policy that distinguishes India from the successful Asian economies to its east. This is the near complete failure of land reforms as a state policy. Given this failure, the class structure in agriculture is fundamentally split between the landed and the landless at the broadest level. But among the landed, there are many differences but the most crucial distinction of these is the quantum of land ownership. I divide the landed groups into four classes — Rich, Middle, Small and Marginal/Tenant (empirical details provided in the next section). Roughly, the first two groups hire net-labor power, the third group is balanced in its labor-power selling and buying, and the last group is net-labor power 4 This framework for urban areas is somewhat similar to the one employed by Wright (1997) although a further sectoral sub-division into services and manufacturing is not done in Erik Wright’s work. Professionals are different from skilled workers in the sense that professionals usually do white-collar kind of work and possess greater autonomy/control than the skilled blue-collar workers. An example would be an engineer and a lathe worker in a factory. 5 This is not merely in the nature of labor (material vs. immaterial as Hardt and Negri (2000) point out) but also in terms of the inequalities among different workers. It is observed that service sector has a much greater inequality among its workers compared to manufacturing.

June 5, 2012

11:23

84

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

selling.6 The non-agricultural population can be divided into five classes. First, the rentier class can be further divided into the moneylender and absentee landlord classes. Second, the non-agricultural self-employed group (usually quite heterogeneous) that cannot be further sub-divided due to data difficulties described below. Third, the working class can be divided into two classes: the rural non-agricultural professionals (for instance, the government officials and those who possess formal employment) and the other workers. In both urban and rural sectors, there is also an unclassified section that defies further sub-division given the data limitations. 4.3. Data, Definitions and Limitations The data source used for this analysis is the Indian National Sample Survey, Household Consumer Expenditure Data, rounds 50(1993–1994) and 61(2004–2005). In the 50th round, 115,354 households were selected in the sample, out of which 69,206 are from rural area, and 46,148 are from the urban area. In the 61st round, 124,643 households were selected out of which 79,297 are in the rural area, while 45,346 are from the urban area. The basic unit of analysis is the household for which monthly consumer expenditure data are available. However, I use the household size to obtain the monthly per capita expenditure levels. All the results presented are expressed as per capita consumption expenditure data adjusted to annual figures and further adjusted to 2005 purchasing power parity (PPP) U.S.$. Classes are defined using occupational data obtained from the National Classification of Occupations (NCO 3-digit, 1968 scheme) codes of various occupations listed in the Indian sample. Therefore, this generates a maximum of 1,000 occupations, although less than 400 occupations are listed in the actual scheme. NSSO assigns an occupation to a household based on the principal occupation of the members of the household. For the urban data, the proprietary and managerial classes have been split into formal and informal categories based on the occupational descriptions (NCO) and nature of the industry (National Industrial Classification Codes (NIC)). These classes include the proprietary or managerial classes directly available from the NCO descriptions as well as from the category 6 Please see Utsa Patnaik’s agrarian class schema (1987), for a similar framework. Utsa Patnaik uses her framework in fieldwork based methodology to generate these categories, not large surveys as presented in the current work.

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

85

of self-employed (based on a survey question). Based on the NIC codes that are available in the surveys, I assigned the workers either to manufacturing or service sectors. Further in each of these sectors, sub-division was made among professionals, skilled workers and unskilled workers based on the description of the occupations (NCO). For the rural sample, using occupational data for the non-agricultural population, rural professionals, and rural moneylenders are identified. Of the remaining, the self-employed among the non-agricultural community are defined based on one of the NSS survey based questions and this cannot be further sub-divided. Of these and others that own more than half-acre of land, while pursuing occupations other than agriculture, absentee landlord category is defined. Using the NSS survey based question that identifies agriculture as the primary occupation of the household, based on land data, the agricultural community is divided into five classes. Those that own more than 10 acres7 are defined as rich farmers. Those that own between 5 and 10 acres are defined as middle farmers. Those that own between two and five acres are designated as small farmers. Those that own less than two acres are called marginal farmers. Those who do not own a land but depend on agriculture as their primary source of livelihood, and define themselves as workers are defined as the agricultural workers. The biases in these surveys are that the upper end consumption groups are not adequately sampled, while the formal owner and managerial class is severely underrepresented. These have a bearing both on computing the level of inequality in the Indian economy (inequality is, therefore, underestimated) as well as in determining the relative position of a class such as the owner and managerial class vis-`a-vis other classes. Moreover, consumption distribution is usually more equal than income distribution.

4.4. Results I: Indian Class Structure 4.4.1. Simplified Class Structure Table 4.1 presents the simplified class structure, from which we can make the following observations. First, in terms of the levels of consumption, at both points in time (1993–1994 and 2004–2005), the urban classes are better off than their rural counterparts. As expected, the urban elite 7 Irrigated

land is considered to be equivalent to twice the non-irrigated land. This is based on yield per acre calculations during my field research in Telangana villages.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

86 Table 4.1.

Simplified Class Structure.

1993–94

Urban Elite Urban Workers Rural Elite Non-Ag Workers Rural Small Peasants Agricultural Workers ALL

b1363-ch04

2004–05

Mean

Pop (%)

Mean Ratio

Mean

Pop (%)

Mean Ratio

Ratio Growth

984 703 480 417 412 322 509

7.44 17.35 32.18 6.67 15.62 20.74 100.00

1.94 1.38 0.94 0.82 0.81 0.63 1.00

1189 758 559 488 456 354 582

8.19 17.13 29.44 11.01 15.67 18.56 100.00

2.04 1.30 0.96 0.84 0.78 0.61 1.00

0.06 −0.06 0.02 0.02 −0.03 −0.04 0.00

Note: 1. All statistics for Tables 4.1 to 4.11 computed from the Indian National Sample Survey rounds 1993–1994 (Round: 50) and 2004–2005 (Round: 61).

(Owners, Managers, and Professionals) is the best placed among the six classes presented, and the agricultural workers are at the bottom. Otherwise, the ranks of different classes are as expected. In terms of change (defined as the rate of change in the ratio of class mean to the population mean over the period) too, it can be seen from Table 4.1 that the biggest gainers are the urban elite. Moderate gainers are the rural elite and rural non-agricultural working class. The biggest losers are the urban working class, small peasants, and agricultural workers. 4.4.2. Detailed Class Structure Table 4.2 shows that in terms of the levels of consumption, the ranks are more or less along expected lines. The only interesting thing to note is that the consumption levels of service sector professionals are the highest amongst all classes in 2004–2005. This shows how the relative position of the service sector and the service professionals in particular after reforms has begun to witness improvement at the higher end of the income ladder.8 In terms of change, the biggest gainers are the urban professionals (service and manufacturing), urban informal managers and owners, urban unclassified, rural moneylenders, rural professionals, absentee landlords, and rural unclassified. Broadly, it is the service sector that has made rapid strides in terms of improvements in the urban areas, and the rural 8 The fact that the formal sector owners and managers do not witness a gain in their relative position although all other perceptions are to the contrary indicates how undersampled this category is in the overall sample.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India Table 4.2.

b1363-ch04

87

Detailed Class Structure.

Class

Mean

1993–1994 Pop (%)

Ratio

Mean

2004–2005 Pop (%)

Ratio

Ratiogrowth

Owner/Manager (formal) Owner/Manager (informal) ManufacturingProfessional Manufacturingskilled Manufacturingunskilled Serviceprofessional Service-skilled Serviceunskilled Urbanunclassified Urban sub-total Rich farmer Middle farmer Small farmer Marginal farmer/Tenant Ag workers Rural professional Rural moneylender AbsenteeLL+ Nonagselfemp Non-ag Self-employed AbsenteeLL+ Others Non-Ag workers Ruralunclassified Rural sub-total ALL INDIA

1345

0.95

2.64

1490

1.16

2.55

−0.03

816

3.58

1.60

982

4.32

1.68

0.05

1006

1.60

1.98

1247

0.45

2.14

0.08

682

6.17

1.34

721

6.72

1.24

−0.08

527

2.14

1.04

537

1.47

0.92

−0.11

1170

1.25

2.30

1534

1.01

2.63

0.14

884 629

4.59 4.46

1.74 1.24

1055 648

3.85 5.08

1.81 1.11

0.04 −0.10

872

0.05

1.72

1331

1.25

2.28

0.33

787 524 445 406 426

24.74 7.36 8.02 10.52 5.09

1.54 1.03 0.87 0.80 0.84

897 601 497 459 451

25.20 6.23 7.54 10.07 5.59

1.54 1.03 0.85 0.79 0.77

0.00 0.00 −0.03 −0.01 −0.08

322 600

20.74 2.62

0.63 1.18

354 928

18.56 1.50

0.61 1.59

−0.04 0.35

669

0.01

1.31

999

0.01

1.71

0.30

465

3.31

0.91

563

4.94

0.97

0.06

422

6.15

0.83

482

7.40

0.83

0.00

514

1.08

1.01

690

0.98

1.18

0.17

417 483

6.67 3.64

0.82 0.95

488 650

11.01 0.87

0.84 1.11

0.02 0.17

417 509

75.26 100.00

0.82 1.00

476 583

74.80 100.00

0.82 1.00

0.00 0.00

rentier classes that have improved in rural areas. The biggest losers are the urban unskilled workers (both manufacturing and service), urban skilled manufacturing workers, rural marginal farmers and tenants, and agricultural workers.

June 5, 2012

11:23

88

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

Tables 4.1 and 4.2 give a broad picture wherein the urban elites have begun to march ahead of the rest of the classes. The glaring gaps are between the relative enrichment of the professionals and the relative decline of the agricultural workers and small peasants. These results give a preview of how the detailed inequality decomposition results would turn out.

4.5. Results II: Analyzing Indian Inequality, Does Class Matter? In terms of the overall story of changing inequality in India, the main observation is that there is an almost four points (more than 10%, which is very likely an underestimate as discussed above) increase in the Gini coefficient between 1993–1994 and 2004–2005. This is a big change after the experience of equal growth in the 1980s. After economic reforms have been introduced, the story of Indian growth has been disequalizing. By decomposing this inequality along multiple axes such as rural–urban, inter-state, simplified class structure, complex class structure and so forth, I present competing ways of breaking down the increase in inequality into comprehensible structures. Before that, a brief note is presented on the methodology of decomposition. 4.5.1. Yitzhaki Methodology of Decomposing the Gini Coefficient The method of Gini decomposition developed originally by Shlomo Yitzhaki9 offers a unified framework for addressing certain important issues such as how much of the inequality can be explained by the inter-group and intra-group components. Moreover, this method helps us examine whether particular classes are stratified at a point in time and if they are getting increasingly more stratified over time. A brief description is sketched out below. Let G be the Gini coefficient of consumption. Using the Yitzhaki decomposition methodology (Yitzhaki, 1994), we separate G into intergroup inequality (Ib ) and a remainder (Ir ) that can be interpreted as 9 As mentioned above, this framework and its description are also available in (Wolff and Zacharias 2009) and (Zacharias and Vakulabharanam, 2009) apart from the original description in Yitzhaki (1994).

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

89

intra-group inequality: G = Ib + Ir

(4.1)

The amount of inter-group inequality is: Ib =

2cov(µi , Foi (y)) , µ

(4.2)

where y is consumption, µ is mean consumption for all persons, µi is mean consumption for group i, and Foi (y) is the mean rank of group i, i.e., the average position of the members of a group in the overall distribution. Thus, the amount of inter-group inequality is twice the covariance between the mean amounts of consumption and mean ranks of groups divided by the mean consumption for all individuals.10 The remainder term is calculated as:
Ir = si Gi Oi , (4.3) i

where si is the share of group i in aggregate consumption, Gi is the Gini coefficient of the consumption distribution within group i, and Oi is the overlapping index for group i. The index of overlapping proposed by Yitzhaki is a measure of the degree to which the range of consumption in each group overlaps with the range of consumption for all population. Overlapping can thus be seen as the opposite of stratification: the higher the amount of overlap between a group and the population, the less stratified they are as a group in terms of consumption (Yitzhaki, 1994, pp. 148–149). The amount to which group i overlaps with the overall distribution is defined as: Oi =

covi (y, Foi (y)) , covi (y, Fi (y))

(4.4)

where Foi (y) is the function that assigns to the members of group i their ranks in the overall distribution, Fi is the function that assigns to the members of group i their ranks in the consumption distribution within that group, and covi indicates that the covariance is according to the distribution within group i. The minimum value of Oi is given by the share 10 In contrast, in the standard decomposition, the between-group component would be equal to twice the covariance between the consumption of each group and the rank of each group’s mean consumption divided by overall mean consumption. The Yitzhaki decomposition takes into account the ranking of each individual within each group in the overall distribution.

June 5, 2012

11:23

90

9in x 6in

Growth and Inequality: An International. . .

b1363-ch04

V. Vakulabharanam

of group i in the population and its maximum value is equal to two. When the index equals the minimum possible value, it suggests that the group in question is a perfect stratum, i.e., it occupies an exclusive segment of the overall distribution. If a particular group has a range of consumption that coincides with the range of consumption of all persons then the index will be equal to one. Finally, if the index is greater than one, the distribution of consumption within the group is much more polarized than in the overall distribution. This can happen if the members of the group constitute two strata, one that has higher and the other that has lower than the average consumption of the whole population (Milanovic and Yitzhaki, 2002, pp. 162–163). The index of overlapping defined in Eq. (4.4) is constructed from indexes that indicate the amount by which a group overlaps with each of the other groups:
pj Oji (4.5) Oi = pi + j=i

where pi is the share of group i in the total population and Oji is the index of overlapping of group j by group i. Since the overlapping of a group by itself is equal to one by definition, its contribution to Oi is equal to its relative size. The index of overlapping of the overall distribution by a group is the weighted sum of overlapping of each of the other groups by that group, with the relative size of each group serving as the weights. In turn, the group-by-group overlapping indexes are calculated as: Oji =

covi (y, Fji (y)) , covi (y, Fi (y))

(4.6)

where Fji is the function that assigns members of group i their ranks in the distribution of group j. The index Oji indicates the extent to which the consumption of individuals in group j falls in the range of consumption of individuals in group i. (Yitzhaki, 1994, pp. 150–152). Using the Yitzhaki method, various decomposition results are presented below. 4.5.2. Rural–urban Decomposition Table 4.3 presents the results of the rural urban decomposition using the Yitzhaki method. First, it can be observed that a big share of the overall inequality (more than 80%) can be explained by the intra-group component in both years, although there is a moderate increase in the inter-group

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch04

Growth and Inequality in India Table 4.3.

Decomposition I: Rural–Urban.

1993–1994 Cons. Share Overall Intragroup effect Rural 0.65 Urban 0.35 Intergroup effect

Gini

0.286 0.344

2004–2005

Overlap Contrib.

0.326 (0.0020)

%

Cons. Share

0.326 100.00

0.993 0.799

91

0.281

86.25

0.185 0.096 0.045

56.75 0.61 29.45 0.39 13.75

Gini

Overlap Contrib.

0.363 (0.0022)

0.305 0.376

%

0.363 100.00

0.993 0.727

0.291

80.15

0.185 0.107 0.072

50.96 29.48 19.85

Note: 1. The numbers in parentheses are the standard errors. 2. Cons. Share is the consumption share of this group in the total consumption. 3. Contrib. is the absolute contribution of each term to the total Gini. 4. % is the proportion of the contribution of each term to the total Gini.

inequality, i.e. between the rural and urban areas, during the period of study. Second, it is also evident that there is an increase in the intra-group inequality i.e. within-rural or within-urban inequality or both. When we investigate this further, we find out that all the increase in the intra-group inequality comes from the urban contribution to that component. The rural contribution to the intra-group component remains unchanged because the fall in the rural share of consumption offset the increase in their Gini. Third, in terms of the overlapping indexes, urban overlapping index has declined suggesting a greater stratification of the urban vis-` a-vis the rural. These results are in consonance with a lot of studies as well as the changing sectoral contributions to the GDP, while the employment structure does not change readily. 4.5.3. Inter-state Decomposition Table 4.4 presents the summary results of the inter-state decomposition in India. According to these results too, a bigger share of the overall inequality is explained by the intra-state inequality (more than 85% in both years). In terms of changes across the two points in time, it is evident that both inter-group as well as intra-group inequality rose in absolute terms, although the proportion of the overall inequality explained by

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch04

V. Vakulabharanam

92 Table 4.4. Overall decomposition Total Intra-group Inter-group

Decomposition II: Inter-State.

Gini (1993–1994)

%

Gini (2004–2005)

0.326 0.292 0.034

100 89.56 10.44

0.363 0.312 0.051

% 100 85.96 14.04

the inter-state component has risen (from 10% to about 15%). States like Gujarat, Haryana, Karnataka, Andhra Pradesh, West Bengal and Kerala have begun to grow more rapidly than other states while they have also witnessed fairly steep increases in their inside-inequality, while other highgrowth states like Maharashtra and Tamil Nadu have witnessed moderate increases in inequality (Based on own calculations from the NSS surveys; and for growth rates of State Domestic Product, see Jayadev et al., 2007; Also see Bannerjee et al., 2009). 4.5.4. Simplified Class Decomposition Table 4.5 presents the results of the simplified class decomposition. Interclass inequality rose quite significantly. This is self-evident from the discussion regarding the gainers in the simplified class structure story above, with the urban elites making significant relative gains while the urban working people and the rural working people have lagged behind. Intra-class component also rose marginally. The main contribution to the increase in the intra-class component comes from the urban classes and the nonagricultural working class. Also, Gini indexes of these two classes have begun to witness steady increase. In terms of overlapping, both urban elite and urban workers have witnessed significant declines in their indexes, suggesting that these classes have begun to get more stratified vis-`a-vis the rural classes. The increase in the intra-group component of the rural non-agricultural working class needs more careful examination and further investigation. However, an initial hypothesis is that there is increased distress migration from among the agricultural poor to the non-agricultural population. 4.5.5. Detailed Class Decomposition Table 4.6 presents the results of the detailed class decomposition. First, there is a substantial increase in the between-class inequality. The entire

June 5, 2012 11:23

9in x 6in

Table 4.5.

Decomposition III: Simplified Class.

Overlap

0.326 0.131 0.218 0.321 0.058 0.134 0.139

0.357 0.32 0.292 0.274 0.265 0.24

0.664 0.838 0.937 0.968 0.946 0.931

2004–2005 Contrib.

%

0.326 0.257 0.031 0.059 0.088 0.015 0.034 0.031 0.068

100.00 79.04 9.51 18.10 26.99 4.60 10.43 9.51 20.76

Cons. Share

Gini

Overlap

0.363 0.167 0.223 0.283 0.092 0.123 0.113

0.389 0.34 0.318 0.313 0.273 0.233

0.553 0.786 0.929 0.997 0.946 0.895

Contrib. 0.363 0.263 0.036 0.059 0.084 0.029 0.032 0.024 0.100

% 100.00 72.24 9.92 16.25 23.14 7.99 8.82 6.61 27.66

Growth and Inequality: An International. . .

Overall Gini Intra-group Urban elite Urban workers Rural elite Non-ag workers Small peasants Ag workers Inter-group

Gini

Growth and Inequality in India

1993–1994 Cons. Share

93 b1363-ch04

June 5, 2012

Decomposition IV: Detailed Class.

1993–1994 Cons. Share

2004–2005 Contribution

Prop. 100.00 75.22 1.04

0.030

0.326

Cons. Share

Gini

Overlap

Contribution

0.425

0.460

0.363 0.245 0.006

100.00 67.49 1.65

0.363

Prop.

0.352

0.418

0.052

0.335

0.755

0.013

4.03

0.073

0.358

0.635

0.017

4.68

0.029

0.363

0.682

0.007

2.20

0.009

0.343

0.420

0.001

0.55

0.075

0.304

0.827

0.019

5.78

0.083

0.330

0.784

0.021

6.06

0.020

0.291

0.978

0.006

1.75

0.014

0.300

0.942

0.004

2.48

0.026

0.322

0.449

0.004

1.15

0.027

0.341

0.295

0.003

0.55

0.073 0.050

0.317 0.302

0.661 0.859

0.015 0.012

4.69 3.98

0.070 0.056

0.340 0.292

0.556 0.785

0.013 0.013

2.75 2.20

0.001

0.312

0.699

0.000

0.07

0.028

0.416

0.543

0.006

1.65

0.080 0.074 0.089 0.045

0.277 0.275 0.249 0.297

0.855 0.933 0.928 0.979

0.019 0.019 0.020 0.013

5.81 5.82 6.31 4.01

0.064 0.064 0.079 0.043

0.301 0.268 0.266 0.283

0.852 0.904 0.929 0.976

0.016 0.016 0.020 0.012

4.41 4.41 5.51 3.31

0.139

0.240

0.931

0.031

9.53

0.113

0.233

0.895

0.024

Growth and Inequality: An International. . .

0.023

0.326 0.245 0.003

9in x 6in

Overlap

V. Vakulabharanam

Overall Gini Intra-class Owner/Manager (formal) Owner/Manager (informal) Manufacturingprofessional Manufacturingskilled Manufacturingunskilled Serviceprofessional Service-skilled Serviceunskilled Urbanunclassified Rich farmer Middle farmer Small farmer Marginal farmer/Tenant Ag workers

Gini

11:23

94

Table 4.6.

6.61 (Continued) b1363-ch04

June 5, 2012 11:23

0.033

0.329

0.898

0.010

0.001

0.281

0.781

0.032

0.287

0.054

Prop.

2004–2005 Overlap Contribution

Cons. Share

Gini

Prop.

2.99

0.024

0.392

0.669

0.006

1.65

0.001

0.07

0.001

0.192

0.283

0.001

0.001

0.938

0.009

2.64

0.048

0.328

0.936

0.015

4.13

0.273

0.964

0.014

4.36

0.061

0.3

0.979

0.018

4.96

0.012

0.266

0.850

0.003

0.83

0.012

0.298

0.770

0.003

0.83

0.058 0.036

0.274 0.315

0.968 0.980

0.015 0.011

4.72 3.41

0.092 0.01

0.313 0.358

0.997 0.904

0.029 0.003

7.99 0.01

0.081

24.78

0.118

32.51

Growth and Inequality: An International. . .

Gini

Growth and Inequality in India

Rural professional Rural moneylender AbsenteeLL+ Non-ag self-employed Non-ag self-employed AbsenteeLL+ Others Non-ag workers Rural unclassified Inter-class

1993–1994 Overlap Contribution

Cons. Share

(Continued)

9in x 6in

Table 4.6.

95 b1363-ch04

June 5, 2012

11:23

96

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

increase in the overall Gini is accounted for by the increase in the between-class inequality. This is by far the most significant decomposition result. The details of the nature of this increase in the between class inequality have been discussed in Sec. 4.4. From the overlapping indexes in Table 4.6, the urban professionals (service and manufacturing)11 on the one side, and the marginal farmers and agricultural workers on the other side have become more stratified at the top and bottom respectively. Substantial reductions have been noted for urban professionals, urban informal managers/owners, skilled workers in manufacturing, rural professionals, and rural moneylenders. All these groups have become more stratified at the top and are performing better than other members of society. For some classes such as formal owners and managers as well as non-agricultural workers, the Gini coefficients also show substantial increase. 4.5.6. Urban Class Decomposition Table 4.7 presents the decomposition of urban classes taking the urban sector as a self-contained whole. First, the urban inequality as a whole is higher than the overall Indian inequality at both points in time. Second, urban inequality registered an increase of almost 10% in terms of the Gini coefficient during the period of study. Both the intra-class and the inter-class inequality have registered an increase during this period although the inter-class inequality rose more significantly. This latter increase is primarily because of the growing gap between the professionals, owners and managers in the formal and informal sectors vis-` a-vis unskilled workers. The increase in the intra-class inequality has come about mainly because the informal sector owners and managers have grown much larger as well as become more unequal during this period. There is also evidence of a significant increase in their overlapping index implying that this class has become less stratified. This class includes some wealthy businesses (such as in retail or wholesale) as well as some poor self-employed urban residents. From the overlapping index analysis, an upper-end stratification is visible among professionals, while the unskilled workers are showing a tendency towards stratification at the bottom. 11 In

the case of the formal owners and managers, the overlapping index is in the range of 0.4, which suggests high stratification although it showed a tendency to increase suggesting lesser stratification. This tendency may be due to the sampling problem alluded to above.

b1363-ch04

June 5, 2012 11:23

Decomposition V: Detailed Urban Classes.

1993–1994 Cons. Share

2004–2005

Overlap

Contrib.

% 100.00 85.10 04.50

0.076

0.344

Cons. Share

Gini

Overlap

Contrib.

0.425

0.775

0.376 0.305 0.025

100.00 80.92 06.66

0.376

%

0.352

0.677

0.150

0.335

0.966

0.049

14.11

0.187

0.358

0.928

0.062

16.52

0.083

0.363

0.911

0.027

07.98

0.025

0.343

0.738

0.006

01.68

0.216 0.058

0.304 0.291

0.987 0.995

0.065 0.016

18.84 04.88

0.213 0.034

0.330 0.297

0.994 0.956

0.070 0.010

18.58 02.57

0.075 0.208 0.144 0.002

0.322 0.317 0.303 0.312

0.712 0.890 0.988 0.936

0.017 0.059 0.043 0.001 0.051

05.00 17.06 12.53 00.17 14.90

0.068 0.179 0.145 0.073

0.341 0.340 0.292 0.416

0.586 0.856 0.948 0.850

0.014 0.052 0.040 0.026 0.072

03.61 13.86 10.68 06.87 19.08

Growth and Inequality: An International. . .

0.065

0.344 0.293 0.015

Growth and Inequality in India

Total Intra-class Owner/Manager (formal) Owner/Manager (informal) Manufacturingprofessional Manufacturing-skilled Manufacturingunskilled Service-professional Service-skilled Service-unskilled Unclassified Inter-class

Gini

9in x 6in

Table 4.7.

97 b1363-ch04

June 5, 2012

11:23

98

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

4.5.7. Rural Class Decomposition Table 4.8 presents the results of the decomposition of the rural classes with the rural sector considered as a self-contained whole. Overall inequality within the rural sector has risen moderately (about 5%). From the results it is clear that this increase has come about mainly because of the increase in the inequality among the non-agricultural groups and the increase in the inequality between the agrarian and non-agrarian classes. Both the interclass and intra-class components have registered almost an equal increase in absolute terms, although as a percentage of the overall rural inequality, inter-class component now explains a higher proportion. In terms of the intra-class component, the largest increase has come about among the non-agricultural workers, whose overlapping index has crossed one, implying increased polarization within this group. Given also the fact that this group has witnessed a swelling in its ranks, probably an increasing number of agricultural workers and distressed marginal farmers/tenants are joining the non-agricultural workforce. In terms of overlapping index, the most significant stratification stories are for the rural professionals, rural moneylenders and rural absentee landlords. As noted above, the rentier classes have become increasingly more stratified. 4.5.8. Agrarian Class Decomposition Table 4.9 presents the decomposition results of the agrarian classes as a separate whole. It is evident that the inequality in the agrarian sector has remained stagnant. Both the intra and inter-class components of inequality have also remained largely unchanged. The overlapping index has also not underdone transformations. This suggests that the agrarian class structure itself has largely remained unchanged during the period of economic reforms even as the populous agricultural groups have witnessed relative impoverishment vis-`a-vis the non-agricultural population. 4.5.9. Rural Non-agricultural Class Decomposition Table 4.10 presents the decomposition results of the rural non-agricultural sector. The non-agricultural sector has witnessed a significant increase in inequality (of almost 10%) with both the intra and inter-class components showing equal absolute increases, although as a proportion, the inter-class component explains a higher proportion of the overall inequality at the later point in time. The inter-class component has registered an increase

b1363-ch04

June 5, 2012 11:23

Table 4.8.

1993–1994 Overlap

0.286

Cons. Share

Gini

2004–2005 Overlap

% 100.0 88.40 9.79 10.14 11.19 6.99

0.105 0.105 0.13 0.071

0.301 0.268 0.266 0.283

0.809 0.908 0.946 0.995

0.305 0.263 0.026 0.026 0.033 0.020

0.305

Contrib.

%

0.123 0.114 0.136 0.069

0.277 0.275 0.249 0.297

0.822 0.924 0.941 0.967

0.286 0.253 0.028 0.029 0.032 0.020

100.0 86.38 8.52 8.52 10.82 6.56

0.213 0.001 0.05 0.049

0.24 0.281 0.329 0.287

0.988 0.717 0.854 0.92

0.051 0.001 0.014 0.013

17.83 0.00 4.90 4.55

0.185 0.001 0.039 0.078

0.233 0.192 0.392 0.328

0.969 0.183 0.564 0.903

0.042 0.001 0.009 0.023

13.77 0.001 2.95 7.54

0.083 0.018 0.089 0.056

0.273 0.266 0.274 0.315

0.97 0.821 0.972 0.963

0.022 0.004 0.024 0.017 0.033

7.69 1.40 8.39 5.94 11.60

0.1 0.019 0.151 0.016

0.3 0.298 0.313 0.358

0.977 0.712 1.001 0.853

0.029 0.004 0.047 0.005 0.041

9.51 1.31 15.41 1.64 13.62

Growth and Inequality: An International. . .

Contrib.

Growth and Inequality in India

Total Intra-group effect Rich farmer Middle farmer Small farmer Marginal farmer/ Tenant Ag worker Rural moneylenders Rural professional AbsenteeLL+non-ag self-emp Non-ag self-emp AbsenteeLL+Others Non-ag workers Rural unclassified Inter-group effect

Gini

9in x 6in

Cons. Share

Decomposition VI: Detailed Rural Classes.

99 b1363-ch04

June 5, 2012

9in x 6in

Decomposition VII: Detailed Agrarian Classes. 1993–1994 Overlap

0.277 0.188 0.174 0.208 0.106 0.325

0.277 0.275 0.249 0.297 0.240

0.796 0.906 0.932 0.95 0.999

2004–2005 Contrib.

%

0.277 0.241 0.041 0.043 0.048 0.030 0.078 0.036

100.00 86.93 14.80 15.52 17.33 10.83 28.16 13.07

Cons. Share

Gini

Overlap

0.281 0.176 0.177 0.218 0.119 0.310

0.301 0.268 0.266 0.283 0.233

0.779 0.894 0.938 0.987 0.987

Contrib. 0.281 0.243 0.041 0.042 0.054 0.033 0.071 0.038

% 100.00 86.44 14.59 14.95 19.22 11.74 25.27 13.56

Growth and Inequality: An International. . .

Total Intra-class Rich farmer Middle farmer Small farmer Marginal farmer/Tenant Ag worker Inter-class

Gini

V. Vakulabharanam

Cons. Share

11:23

100

Table 4.9.

b1363-ch04

June 5, 2012 11:23

Decomposition VI: Detailed Non-agricultural Classes. 2004–2005

Overlap

Contrib.

%

0.297 0.283 0.001 0.044 0.039

100.00 95.28 0.001 14.81 13.13

0.065 0.012 0.07 0.052 0.014

21.89 4.04 23.57 17.51 4.72

0.297 0.001 0.146 0.142

0.281 0.329 0.287

0.801 0.916 0.966

0.24 0.051 0.257 0.162

0.273 0.266 0.274 0.315

0.997 0.876 1 1.006

Cons. Share

Gini

Overlap

Contrib. 0.333 0.303 0.001 0.024 0.06

100.00 90.96 0.001 7.21 18.02

0.075 0.011 0.12 0.013 0.030

22.52 3.30 36.04 3.90 9.04

0.333 0.001 0.097 0.194

0.192 0.392 0.328

0.24 0.641 0.949

0.248 0.047 0.374 0.039

0.3 0.298 0.313 0.358

1.005 0.771 1.024 0.909

%

Growth and Inequality: An International. . .

Total Intra Rural moneylenders Rural professional AbsenteeLL+non-ag self-employed Non-ag self-employed AbsenteeLL+Others Non-ag workers Rural unclassified Inter

Gini

Growth and Inequality in India

1993–1994 Cons. Share

9in x 6in

Table 4.10.

101 b1363-ch04

June 5, 2012

11:23

102

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

primarily because the rentier classes and the rural professionals have become more stratified at the top. The intra-class component has shown an increase primarily because the non-agricultural workers have both increased in proportion as well as have witnessed a certain degree of polarization (from the overlapping index) that suggests a distress movement from the agrarian sector as described above. This phenomenon needs to be carefully investigated through further research.

4.6. Explanations and Discussion The Indian growth strategy of the 1990s and this decade has led to increased economic inequality among its population. The emerging patterns of this inequality have been discussed above. How do we understand and explain these patterns of inequality? Several studies have pointed to increased distress in the agrarian sector in India. More than 100,000 farmer suicides have occurred from 1998 to till date. Agricultural growth has lagged far behind the growth in other sectors such as services and manufacturing. This has been attributed to policies of economic liberalization12 as well as a slowdown caused by a decline in the returns from Green revolution technologies (Vakulabharanam and Motiram, 2007). Policies of economic liberalization have tended to cause a reduction in public investment in agriculture, as well as partial withdrawal of state support to various small farming groups. Especially before 2004–2005 (this study focuses on that period), the cutback in subsidies and the slow growth of subsidized agricultural credit on the one hand, and the introduction of trade liberalization on the other, which caused agricultural output prices to fall for some key agricultural commodities, caused a “double squeeze” of the farming community. Now it is well documented that this has led to an increased dependence of the small farmers on informal moneylenders, who also frequently combine other roles (such as that of merchant) with moneylending causing an increase in their market power vis-`a-vis small peasants (Reddy and Mishra, 2009). This has led to a further deterioration of the farmers’ living standards as they now have to pay much higher interest rates (compared to the institutional rates), and 12 Since

2004, certain reversals in policy have come about such as the introduction of the National Rural Employment Guarantee Scheme (NREGS). The results of this act and other reversals would have to be evaluated when the next consumption survey is released.

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

103

also sometimes lose control over the cropping pattern decision and so forth (see Vakulabharanam, 2004). All in all, the agricultural sector has suffered the worst and this reflects in the reduced consumption shares of all the populous groups in agriculture but especially of the marginal farmers, tenants, and agricultural workers. Indian growth strategy has tended to leave agriculture and a majority of its population behind as the above analysis based on consumption data suggests. At the same time, it is evident that the rentier classes (such as moneylenders and absentee landlords) have cashed in on the state withdrawal to make steady gains. When the agrarian sector is in distress, does the rest of the economy provide opportunities for migration for these peasants? The rural non-agricultural sector has seen impressive gains at a first glance. However, when we do a careful class analysis it is mostly the nonworking groups (mentioned earlier) that have witnessed improvements. The rural non-agricultural working population has seen improvement but its ranks have almost doubled while the inequality within this group has undergone a big increase. The latter probably results primarily from the distress that the agricultural sector has undergone causing an outward migration. In the wake of insufficient employment opportunities in the urban areas, this migration shows up as intra-rural. Clearly, while there are improvements in this sector, they do not compare with the massive development that the Chinese non-agricultural sector witnessed, especially in the 1980s and 1990s. One of the main objectives13 of the relatively new special economic zone (SEZ) strategy implemented mostly in rural areas for rapid industrialization is the increase in the level of non-agricultural employment, although results are far from evident. Apart from the enormous dispossession of peasantry that this strategy has tended to generate in states such as Maharashtra, Andhra Pradesh, and West Bengal, the positive net employment effects have not appeared on the horizon. The urban sector has grown more rapidly than the rural sector during this period with the growth path skewed in favor of the organized services sector. It is now well known that the Indian government (as the rest of the world) has been seeing the Indian economy as primarily driven by service sectors such as Information Technology, Biotechnology, Finance, Insurance, Real Estate, Transport, Hotels, and so forth. These and other similar sectors have, therefore, received a lot of infrastructural support, as well as 13 The

other and more important premise for the policy–makers is that it will generate high growth along the Chinese lines.

June 5, 2012

11:23

104

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

have benefited from easier legislation. The main beneficiaries of this set of policies (they can also be termed policies towards globalization) have been the upper end of formal workers (professionals) employed in these industries as well as the capitalist and managerial cadres. These sectors have typically been export-led, skill-intensive and dependent on overseas demand for their sustenance. In contrast, the manufacturing sector14 in India has been witnessing impressive growth only in the recent years, although the professionals in this sector have witnessed impressive gains through the period as the survey results indicate. The informal sector owners and managers within urban areas have seen impressive growth in their consumption even as the inequality within this group has increased sharply. This group is operating within the highincome generating arenas such as retail and wholesale, as well as low-income generating arenas such as self-employed petty vendors. The consumption shares of the unskilled workers have declined sharply suggesting that Indian growth trajectory has not been a rising tide for most groups in the society. In sum, the distress in the agrarian sector has not been countered by a rapid growth of decent employment opportunities in other sectors. The fact that the share of informal workers (informal workers in the informal as well as the formal sectors) in the total employment rose slightly during the first decade of reforms to more than 92% of the total employment is indicative of the fact that decent employment opportunities on a large-scale have simply not been forthcoming from the formal high-growth sectors (From NSS report on the informal sector in India in 2004–2005, Report 519). When we analyze the macro-economic growth statistics for the Indian economy during the period of analysis (1994–2005), Indian growth has been led largely by investment, exports (not net exports) and private consumption initially until 2002, and after 2002, consistently by investment and exports.15 However, consumption shares of almost all the poor groups have witnessed declines implying that it has been luxury consumption that has driven the private consumption growth patterns. Similarly, it is now widely documented that investment in agriculture has not kept pace (see Reddy 14 It remains to be seen if this will result in employment generation like in the case of China. 15 For the whole period 1994–2005, if we take the national accounts figures, about 60% of the cumulative growth can be attributed to private consumption, over 30% can be attributed to investment and about 30% to exports (not net exports) making these three components the main pillars of Indian growth (Source: Calculations from CSO Data on macroeconomic aggregates, Ministry of Statistics, India).

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

105

and Misra, 2008), suggesting once again that it has been investment in non-agricultural sectors that has seen impressive growth. With this clarity, we can see how luxury consumption, non-agricultural investment and export-led growth sectors have tended to become enclave-like, while the rest of the economy that contains the predominant majority in India has not significantly benefited from the growth process. While some economists argue about the “trickle-down” effects of this sort of a growth process, it has not been visible in the Indian economy so far on any large scale. It is this skewed growth process that explains the increase in the overall inequality after 1991 reversing the impressive record with inequality in the 1980s decade. The above evidence buttresses the understanding that there is a consolidation of a new class structure in India as has been noted in other analyses (Chatterjee, 2008). Until the 1980s, literature about the pan-Indian class structure points to two ideas: An intermediate class regime (originally proposed by Kalecki in 1972 and imported to India by Raj K.N. in 1973); or a loose coalition of dominant classes (Bardhan, 1984). In the intermediate regime story, the urban intermediate class (self-employed groups, small and medium enterprise owners, traders, and the bureaucrats), and the rural intermediate class (middle peasants) played a big role in determining state policies at the expense of the big bourgeoisie. In the dominant class story, rural elites, urban middle-class bureaucrats and big bourgeoisie jostled for supremacy with none achieving it fully and the state more or less acting in a domain of relative autonomy (See McCartney and Harriss-White, 2000 for a discussion of the variants of the class regimes idea applied to pre-liberalization India). The state operated with an ambivalence wherein it took part in accumulation activities, worked for the conflicting interests of the dominant classes, while also appearing to take on a broadly developmentalist and progressive pro-poor role for itself. This class structure along with the role of the state is supposed to explain both the tendencies towards stagnation in the economy as well as why certain classes tended to benefit disproportionately during the phase of relative stagnation until 1980. After 1991, as the NSS survey results indicate, the dominant classes are the urban elites. Urban elites have incorporated the professionals (some may use the term — upper income middle classes) from among the working people, as well as the state in pushing through the agenda of economic liberalization. The state has consequently reduced its own accumulation role quite significantly. The rural intermediate classes are not quite as important in this new scheme, although their interests are usually

June 5, 2012

11:23

106

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

protected, directly or indirectly.16 Members of this class have also unevenly moved on to urban occupations to become constituents of urban capitalist classes (Damodaran, 2008). The working groups (the rural poor — small and marginal farmers, agricultural workers; as well as the urban poor — unskilled urban workers) are no longer among the main foci of the state but their interests have continued to be addressed mainly through a populist mode in order to enlist their support during elections.17 The owners and managers in the informal sector in urban areas are quite heterogeneous and certain groups (e.g. wholesale and retail)18 have probably benefited (even this may not last long once liberalization takes deep roots in these occupations) while a large section (petty vendors) has probably not. However, the employment numbers suggest that the informal sector as stated above plays the key role of absorbing employment in the face of insufficient employment opportunities in the formal sector, although this does not apparently improve the consumption levels of the informal workers (in the unskilled category in our datasets) as the above analysis shows. This consolidation of a new class structure comes through largely in the analysis of the levels and changes in the consumption patterns as revealed by the NSS consumption surveys. Several policy conclusions follow if this skewed pattern of growth and rising inequality need to be counteracted. Firstly, it is obvious that the agricultural sector needs higher public investments and better support in terms of promoting institutional lending and so forth. However, it is the poor peasants and agricultural workers who really need to be supported. In this context, an effective way to counter landlessness (in which group, poverty is quite concentrated) is to implement land reforms. Better marketmediation structures need to be developed, for instance, in the creation of input-procurement and output-marketing cooperatives among farmers (Motiram and Vakulabharanam, 2007). This could potentially make the rural sector much more egalitarian and may even contribute to higher 16 Usually in the formulation of minimum support price policies or in the prevention of large-scale land reform process, the interests of rural elites continue to be protected. Also, the rural intermediate classes have considerably diversified themselves into urban activities so that they have one foot in the rural and another foot in the urban areas. See Vakulabharanam and Motiram (2008) for a further discussion of these issues. 17 Chatterjee (2008) calls this phenomenon as the reversal of the primitive accumulation wherein the state counteracts the dispossession of the poor by throwing certain benefits in their direction. 18 With the onset of corporate retail services in the more recent years, it may be hard for the informal retailers to compete and continue to make gains.

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

107

growth due to improved consumption within the rural areas due to this. Secondly, the enclave sectors in the urban areas need to be opened up in favor of labor-intensive strategies of economic development. This will begin to generate more employment in urban areas, which will once again improve consumption levels of the poorer working groups. Thirdly, the non-agricultural sector needs to be developed along cooperative lines, perhaps taking a cue from the Chinese growth strategy especially in the 1980s and 1990s. Implementing these measures will put the Indian economy on a sounder, more sustainable and a more equal growth path. Acknowledgment I thank the Economic Research Center, Graduate School of Economics, at Nagoya University for inviting me as a visiting professor when this research was conducted. I thank Xue Jinjun, Wei Zhong and the participants of the workshop on “International Comparison of Income Inequality” conducted at Nagoya University, Japan on 27th and 28th June 2009 for their helpful comments and suggestions in preparing this draft. I thank Ajit Zacharias and YV Reddy for their comments on an earlier draft. I also thank the participants in the seminar series at the department of economics at the University of Hyderabad as well as Kerala University, Trivandrum for their comments. I also thank Economic and Political Weekly for publishing an earlier version of this article. References Bannerjee, L., Deshpande, A., Ming, Y., Ruparelia, S., Vakulabharanam, V. and Zhong, W. (2009). “Comparing Indian and Chinese Inequality after Economic Reforms.” India China Institute Working Paper, New School, New York. Bardhan, P. (1984). The Political Economy of Underdevelopment in India. New Delhi: Oxford University Press. Chatterjee, P. (2008). “Democracy and Economic Transformation in India.” Economic and Political Weekly, 19th April. Chaudhuri, S. and Ravallion, M. (2006). “Partially Awakened Giants: Uneven Growth in India and China” In Dancing with Giants: China, India and the Global Economy, Winters, L.A. and Yusuf, S. (Eds.). World Bank, Washington, DC. Damodaran, H. (2008). India’s New Capitalists. Delhi: Permanent Black. Dutt, A.K. and Rao, J.M. (2000). “Globalization and its Social Discontents: The Case of India.” Working Paper No. 2000–2006, SCEPA Working Papers

June 5, 2012

11:23

108

9in x 6in

Growth and Inequality: An International. . .

V. Vakulabharanam

from Schwartz Center for Economic Policy Analysis (SCEPA), The New School of Social Research, New York. Hardt, M. and Antonio, N. (2000). Empire. Harvard University Press. Himanshu, (2007). “Recent Trends in Poverty and Inequality: Some Preliminary Results.” Economic and Political Weekly, 10 February 2007. Jayadev, A., Motiram, S. and Vakulabharanam, V. (2007). “Patterns of Wealth Disparities in India during the Era of Liberalisation.” Economic and Political Weekly, 42 (39), 3853–3863. Kalecki, M (1972). Essays on the Economic Growth of the Socialist and the Mixed Economy. London: Unwin. Marx, K. (1976). Capital Volume 1. London: Penguin. Marx, K. (1978). Capital Volume 2. London: Penguin. Marx, K. (1981). Capital Volume 3. London: Penguin. McCartney, M. and White, B.H. (2000).“The Intermediate Regime and Intermediate Classes Revisited: A Critical Political Economy of Indian Economic Development From 1980 to Hindutva.” Working Paper Series of Queen Elizabeth House, Oxford, QEHWPS34. Milanovic, B. and Yitzhaki, S. (2002). “Decomposing World Income Distribution: Does the World Have a Middle Class?” Review of Income and Wealth, 48(2), 155–178. Motiram, S. and Vakulabharanam, V. (2007). “Corporate and Cooperative Solutions for the Agrarian Crisis in Developing Countries.” Review of Radical Political Economics, 39(3), 360–367. Patnaik, P. (2009). “A Perspective on the Growth Process in India and China.” The IDEAS Working Paper Series, Paper No. 05/2009. Patnaik, U. (1987). Peasant Class Differentiation: A Study in Method with Reference to Haryana. New Delhi: Oxford University Press. Raj, K.N. (1973). “The Politics and Economics of Intermediate Regimes.” Economic and Political Weekly, 7th July. Reddy, D.N. and Srijit, M. (2009). Agrarian Crisis in India. New Delhi: Oxford University Press. Thompson, E.P. (1963). Making of the English Working Class, London: Victor Gollancz. Vakulabharanam, V. (2004). Immiserizing Growth: Globalization and Agrarian Change in Telangana, South India between 1985 and 2000, Ph.D. Dissertation, University of Massachusetts, Amherst. Vakulabharanam, V. and Motiram, S. (2007). “Political Economy of Agrarian Distress in India Since the 1990s,” In Great Transformation? Understanding India’s New Political Economy, Harriss, J., Corbridge, S., Reddy, S. and Ruparelia, S, Forthcoming. Vakulabharanam, V., Zhong, W. and Xue, J. (2009). “Does Class Count? Class Structure and Worsening Inequality in China and India.” Working Paper, Department of Economics, Nagoya University, Japan. Wolff, E.N. and Zacharias, A. (2009). “Class Structure and Economic Inequality.” Cambridge Journal of Economics, Forthcoming Issue.

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in India

b1363-ch04

109

Wright, E.O. (1997). Class Counts: Comparative Studies in Class Analysis. Cambridge University Press. Yitzhaki, S. (1994). “Economic Distance and Overlapping Distributions.” Journal of Econometrics, 61, 147–159. Zacharias, A. and Vakulabharanam, V. (2009). “Caste and Wealth Inequality in India.” Working Paper No. 566, Annandale-on-Hudson, NY: The Levy Economics Institute of Bard College, May.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch04

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 5 GROWTH AND INEQUALITY IN GERMANY A COMPARISON OF GERMANY AND CHINA

Stefan Gravemeyer∗ and Thomas Gries†

5.1. Introduction Education is an important determinant of wages and wage distribution. According to Pereira and Martins (2000), education impacts on the distribution of wages in different ways. First, the price of skills acquired through education and reflected in the returns to education impacts on the spread of wages. These inter-educational wage level differentials cause what is referred to as between group dispersion, an extensively discussed topic in the literature. Second, wage dispersion, to an extending amount, exists within educational groups. The question of the effect of education on within group inequality is important since it can predict the impact of education policy on overall inequality. Concerning these within group wage differences Prasad (2000) uses micro data from an extensive German panel data set (German socio-economic panel: GSOEP) to estimate quantile regressions for Mincer-earnings equations. According to Prasad (2000) the relationship between wage distribution and returns to education is positive for university graduates. A negative relationship is determined for the other two educational groups: Employees with vocational training and those who completed an apprenticeship. Pereira and Martins (2000) point out that, in contrast to the results of Prasad (2000), lower quantiles are generally associated with higher returns to education. According to Ammerm¨ uller and Weber (2003) wage inequality in Germany within educational groups ∗ Stefan Gravemeyer, Statistical Consultant, German Statutory Accident Insurance, Germany; † Thomas Gries, Center for International Economics, Professor at University of Paderbon, Germany.

111

b1363-ch05

June 5, 2012

11:23

112

9in x 6in

Growth and Inequality: An International. . .

S. Gravemeyer and T. Gries

decreases with higher educational attainment as the spread of wages is smaller within higher educational levels. According to Martins and Pereira (2004) this is a necessary condition for higher education to impact negatively on overall wage inequality. Pereira and Martins (2000) stress the uniqueness of this fact compared to other Western countries. They estimate quantile regressions of Mincer-earnings equations to analyze differences in returns to education across the wage distribution and across time for Austria, Denmark, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, and the U.K. According to Pereira and Martins (2000) four different patterns emerge: In Portugal education has an increasingly positive effect upon withineducational group wage dispersion, the best (worst) paid at each educational level receive higher (lower) returns from education. Moreover, this differential has risen over time. Second, for Austria, Finland, France, Ireland, the Netherlands, Norway, Spain, Sweden, Switzerland and the U.K., a positive but stable relation between education and within-educational group dispersion is estimated. The third group compromises Denmark and Italy, where the researchers find a neutral impact of education on within-educational group dispersion. There are no notable differences in the returns to education across wage distribution. Finally, for Germany and Greece a negative relationship between returns to education and wage distribution emerges. It can be summarized that in most European countries dispersion in earnings increases with educational level. Pereira and Martins (2000) conclude that concerning within group wage dispersion, education does not reduce wage inequality. A different picture emerges in developing economies. Patrinos et al. (2006) estimate the patterns of returns to education for a mix of East Asian and Latin American countries. They first examine the returns along wage distribution as well as the pattern of returns within educational groups for each country. In contrast to the results mentioned for Europe they find evidence of decreasing returns along wage distribution for low-income countries (Mongolia, Cambodia, Vietnam, Indonesia, Thailand, and the Philippines). A clearly opposite pattern emerges for the case of Singapore (a high-income country). For the Latin American countries the results are more heterogeneous. Argentina and Chile show patterns similar to high income countries, while the other Latin American countries show mixed patterns. There is a plethora of different research results that confirm a rise in wage inequality while stressing the importance of a more differentiated

b1363-ch05

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Germany

b1363-ch05

113

analysis of within educational group disparity since investment in education is widely regarded as a means to reduce inequality, so the importance of this issue is very obvious. The results for developing economies show that relying on the pattern of return for Western economies when developing policies can be dangerous. When returns are higher at the top end of the wage distribution, as reported for most of the European countries, then investing in education would increase inequality. If the pattern is reversed, as it is in many developing economies, investing in education can be an effective means of reducing income inequality. We aim to conduct a comparative analysis of the dispersion of returns to education within education groups and across quantiles between Germany and China, the transition country with the highest growth rates. We expect significant differences since these two countries are examples of a transition and a high-income country, respectively. We use the method of quantile regression to gain a deeper insight into the within group inequality component of total inequality. We also add various socio-economic variables to the standard Mincer equation to gauge their effect on income distribution. After discussing the empirical method we provide detailed estimates of quantile returns to education for Germany and urban China.

5.2. Income Generation and Estimation Techniques Most of the empirical work that relates education to earnings is still based on Mincer’s (1974) human capital earnings function. In this model the log of individual earnings (y) in a given period can be separated into an additive function of a linear education term and a squared experience term (Card (1999)): Log y = a + bS + cX + dX 2 + e with S representing the years of completed education, X representing the number of years an individual has worked since completing school, and the residual e. Where there is no information on actual working experience, Mincer proposes using “potential experience,” which is assumed to be the number of years an individual of age A could have worked if they enrolled at age six, so school ended after exactly S years. “Potential experience” is calculated as follows: X = A—S—6 (Card (1999)). In the early literature on Mincer’s approach the earnings function was usually estimated by an ordinary least squares estimation (OLS). One main deficit is that the OLS regression relies on the mean distribution of the dependent

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

S. Gravemeyer and T. Gries

114

variable. It estimates the mean effect of education for the average individual and thus disregards variations in returns within educational groups. Quantile regression, by contrast, allows the return to vary within educational groups. While OLS estimates the wage effects of education on the mean of the conditional wage distribution, quantile regression measures the effects of education at different points of the wage distribution. Thus, differences in quantile returns can be used to measure the dispersion of return within educational groups since they represent the wage differentials between individuals at the same educational level but in different quantiles of the conditional wage distribution (Budria and Pereira, (2005)). Koenker and Basset (1978) introduced the quantile regression model. The quantile regression formula according to Buchinsky (1994) is: lnwi = Xi βθ + eθi Quantθ (Inwi |Xi ) = Xi βθ where Xi is the vector of exogenous variables and βθ is the vector of parameters. Quant θ (lnw i |Xi ) = Xi βθ is the Θ conditional quantile of lnw given X. The Θ regression quantile, 0 < Θ < 1, is defined as the solution to the problem:

min

β∈Rk

   

i:yi ≥Xi β

θ|lnwi − Xi βθ | +

 i:yi |z|

0.09922

0.58

0.564

0.19023

20.85

0.000

Estimate 0.53271

z

Estimate

−0.26198 −1.29

P > |z| Estimate

−0.02687 −4.12 0.00032 5.07 −0.80375 −16.61

0.184 0.000

0.000 0.000 0.000

Variables for Household Family size Land holding

0.00358 0.36 −0.18630 −14.63

0.716 0.000

0.01297 1.33 −0.18415 −14.59

0.346

0.000 0.000 0.000

0.01373 20.53 −0.02844 −4.39 0.00028 4.42 −0.64239 −14.09

0.000 0.000 0.000 0.000

0.00624 0.63 −0.18122 −14.24

0.527 0.000

0.01517 1.54 −0.17892 −14.17

0.122 0.000

0.24969 0.23298 0.23307 0.01846 0.14842 9,006 919.5 0.0000 72.54

0.000 0.028 0.010 0.526 0.041

0.22979 0.21370 0.21575 0.01484 0.18853 9,006 945.59 0.0000 50.92

0.000 0.044 0.016 0.609 0.009

8.67 2.19 2.59 0.63 2.05

7.97 2.02 2.41 0.51 2.61

Growth and Inequality: An International. . .

0.000

0.18414

P > |z|

S. Hondai

20.12

0.000 0.000 0.000 0.000

−0.02708 −4.17 0.00032 5.22 −0.80178 −16.65

0.196

z 0.94

0.18593

9,006 834.49 0.0000 50.37

z

0.001

0.01391 21.05 −0.02877 −4.46 0.00028 4.55 −0.63501 −14.00

9,006 817.3 0.0000 79.19

P > |z|

Estimation (4)

3.24

Variables for Household Head Years of education Years of education2 Age Age2 Sex

Variables surrounding Household Paved road Road passed by 4 wv Leather industry Wood industry Metal industry Obs Wald chi2 (6) Prob > chi2 Wald test of exogenety ch2 (1)

Estimation (3)

9in x 6in

Constant

z

Estimation (2)

11:23

Estimation Results of Probability of Selecting the Non-Agricultural Sector, 1999.

Estimation (1) Estimate

174

Table 8.3.

Note: “***”, “**” and “*” indicate that the coefficients are statistically significant at 1%, 5%, and 10% levels, respectively. b1363-ch08

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Indonesia

b1363-ch08

175

move out from the agricultural sector and try to find an employment in other sectors which had higher rates of return. In this section, we will investigate how probability of leaving the sector changes as his educational level goes up. As shown in Sec. 8.2, there are 11 sectors in Susenas and, in theory, there are 10 different ways of selecting non-agricultural employment. But an estimation method becomes complicated with many sectors and each of them cannot be regarded as an independent one. To avoid this problem, we will group them into two, the agricultural and non-agricultural sectors, and investigate how probability selecting the latter sector will change with educational level. For the estimation, we will use the following formula. (8.2) pvi = c + Xi β + ui where pv is probability of selecting the non-agricultural sector, X is the vector of independent variables, u is the random error, and i the individual household. pv takes the following value depending on selection of a sector. pvi = 0 if a household head selects the agricultural sector, pvi = 1 if a household head selects the non-agricultural sector. Selecting a sector is regressed on a vector of independent variables. They consist of three sets of variables: (1) those related to a household head (a level of education, age, and sex); (2) those related to a household (family size and land holding); and (3) those related to economic conditions surrounding a household (road conditions, and existence of leather industry, wood related industry, and machine industry). Because migration from the agricultural to the non-agricultural sector occurred mainly in the rural area, our analyses will be done only for that area. Among the variables mentioned above, those related to a household head and a household are same as ones used in Sec. 8.3. As for those with respect to economic conditions surrounding households, their types and values are defined in the following. Road conditions which refer to costs of finding works in the nonagricultural sector consist of two variables. One of them is paved roads and if there is any paved road in the village, the variable takes value of “1” and “0” otherwise. The other is roads for four wheel vehicles in villages. This variable takes value of “1” if there is any road for four wheel vehicles and “0” otherwise. The variables with respect to subsector of manufacturing industry refer to availabilities of employment opportunity in the non-agricultural sector. They will take value of “1” if there are any related factories in a relevant village and “0” otherwise.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

S. Hondai

176

The data source of economic situations surrounding households is Village Potential Statistics (“Statistik Potensi Desa” in Indonesian; hereafter abbreviated as “Podes”). The two sets of data, Susenas and Podes (BPS, 2000b), were surveyed differently, but both have same identification numbers with respect to their locations. Based on them, economic situations can be extracted for each household. Because Podes is also surveyed only in sample villages, there are many Susenas households for which surrounding economic variables are not available. We could find them for 9,006 households out of 17,087 households in rural Central Java. We estimated two formulas, one with the variables of surrounding economic situations, and the other without them. As mentioned about years of schooling education in Sec. 8.3, we will use an instrumental variable for schooling years of household head to remove an estimation bias. As estimation results, the marginal effects of variables are shown in Table 8.4 instead of showing probit coefficients.9 Let us scrutinize the results of column 1, estimation without any variable of surrounding economic situations. The first group is the variables in a household head. The coefficient of educational year is positive and statistically significant. This indicates that probability of selecting nonagricultural employment will increase as years of education increases. As for age of household head, the coefficient of age is positive and statistically significant, whereas that of age square is negative and statistically significant. The former indicates that probability of selecting non-agricultural sector declines as a household head gets old, while the latter indicates that probability of selecting non-agricultural sector goes up as a household head gets old. These two results show that firstly probability of selecting nonagricultural sector declines and it will reach at the lowest level at a certain age. Then it will start to increase because some of household heads will retire and become receivers of transfer payments. Based on the estimated coefficients, probability of selecting non-agricultural sector is lowest at the of age 56 years in rural Central Java. The coefficient of sex for a household head is significantly negative and the sign indicates that, other things being equal, probability of selecting non-agricultural sector is lower when a household head is male. The second group is the variables with respect to household characteristics. Coefficient of family size is positive, but not statistically significant. 9 The

mi =

marginal effects for a household head i in the probit model are simply given by: d Pr(yi =1) = φ(xi β)β. dx i

b1363-ch08

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Indonesia

b1363-ch08

177

As for landholding, coefficient is negative and statistically significant. It indicates that a probability of selecting employment in the non-agricultural sector is lower if a household owns agricultural land. Column 2 shows the estimation results using a square of household head’s educational year in the place of education year. The results are almost same as those in Column 1. Columns 3 and 4 show the results of estimation with the variables of surrounding economic situations. In this case, the estimated coefficients concerning to a household head and household are same sign and statistically almost same as those in Columns 1 and 2. The coefficients of newly added variables are as follows. The coefficient of paved road is positive and statistically significant. This implies that probability of selecting the nonagricultural employment increases if there is any paved road in a village. The coefficient of road for four wheel vehicles is positive and statistically significant. This indicates that probability of selecting a non-agricultural employment increases if there is any road for four wheel vehicles in a village. Among three variables relating industrial development, the coefficients of the leather industry and metal industry are positive and statistically significant. If there are any factories related to these industries in a village, probability of selecting a non-agricultural employment gets higher. As for the wood and wood related industry, the coefficient is positive, but statistically insignificant. Based on the estimated coefficients in Column 3, we will calculate probability of selecting the non-agricultural employment with respect to years of education in rural Central Java. The probabilities estimated are for a household head of 20 years old who is holding 0.2 ha of agricultural land and who is living in a village without any paved road and any road for four wheel vehicle. We assume also that the village does not have any factory in the leather industry, wood and wood related industry, and machine industry. Using formula (8-2), probabilities of selecting the non-agricultural employment are calculated and shown as in Fig. 8.2. In that, the horizontal axis is the years of schooling, while the vertical axis is probability of selecting the non-agricultural employment. The upper line shows the probabilities for female, whereas the lower line shows those of male. For both cases, probability goes up as the years of schooling get longer. For female, probability of selecting the non-agricultural employment is 100% with nine years of school education. In other word, a 20 years old female who graduated from a junior high school will leave the agricultural sector at 100% of probability. For male, probability of selecting the

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch08

S. Hondai

178

(%) 150

100 Female Male 50

0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

-50

Years of education Fig. 8.2.

Probability of Selecting the Non-Agricultural Sector, 1999.

non-agricultural employment is 100% with 14 years of school education. In other words, a 20 years old male who finished Diploma I or Diploma II programs will leave the agricultural sector at 100% of probability. As a person attains higher level of education, his probability of selecting the non-agricultural employment goes up. In addition, probability of selecting the non-agricultural employment will goes up if a household does not have any agricultural land. Probability curves in Fig. 8.2 will be pushed up depending on surrounding economic situations of households. If there are any paved road and any road for four wheel vehicles in a village, the curves will be shifted upward. An existence of any factories in the above mentioned manufacturing subsectors in a village will also shift them upward. As a result, probability of selecting the non-agricultural employment increases even at same years of school education. In sum, one’s probability of selecting the non-agricultural employment will increase as his years of school education get longer. This type of job selection will create outflow of individuals from the agricultural sector and enable them to increase returns to their education.

8.5. Conclusion Characteristics of the poor households are examined in this chapter with respect to two important variables, an employment sector and educational

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Indonesia

b1363-ch08

179

level of a household head. Our analysis shows: Incidence of poverty is (1) high in rural areas, (2) high for the household whose head works in the agriculture sector, and (3) especially high for the household whose head has only the primary school education. The estimation of the rate of return to a household head’s education is very low in the agricultural sector in comparison to other sectors in the rural area. This result indicates that encouraging people to receive higher level of education alone will not solve a poverty problem in the sector. In addition, because most of the household heads are already above 20 years old, it is not practical to provide them additional formal education further. On the contrary, their rates of return to education would be improved if they could move to the non-agricultural sectors. An appropriate way to reduce poverty in the area will be to provide them non-agricultural employment opportunities there. The estimation of household head’s probability selecting nonagricultural sectors show the following situations: (1) the probability increases as his educational level goes up; (2) the probability is higher when the road conditions surrounding his household are improved; and (3) the probability is higher when any factories of the leather, wood and wood related, and machine industry are located in the vicinity of his household. The estimation of the rates of return to household head’s education also revealed the following: (1) they went down drastically in the rural agricultural sector from 1999 to 2005; and (2) on the other hand, they increased significantly in the other sectors in the same time period. These two opposite movement of rates of return would widen an income gap between the two groups of household and could be one of the important causes deteriorating income distribution which were shown in BPS statistics (BPS, 1999b; BPS, 2005). As shown in the previous sections, the poverty is primarily a rural problem. It could be originated from many factors and could not be solved just with a single policy instrument. In the long-run, we think that education of the children is the most important instrument to alleviate the poverty because education of the present generations has positive effects on themselves and also on the future generations. However, providing higher levels of education to children is not effective to solve the present poverty problems because a majority of household heads is above 20 years old. For them, more appropriate measures are to create non-agricultural employment opportunities in the rural area and to improve transportation infrastructure to nearby urban areas. Those measures will reduce moving costs

June 5, 2012

11:23

180

9in x 6in

Growth and Inequality: An International. . .

S. Hondai

to non-agricultural sectors and encourage them to find employment opportunities there. A well coordinated set of both long term and short term measures are essential to solve the rural poverty problem in Java. References Akita, T. et al. (1999). “Inequality in the Distribution of Household Expenditures in Indonesia: a Theil Decomposition Analysis.” The Developing Economies 37(2), 197–221. Badan Pusat Statistik, (1999a, 2000a, 2001, 2005/2006). Statistik Indonesia, Jakarta. Badan Pusat Statistik, (1999b, 2002, 2005). Expenditure for Consumption of Indonesia, Jakarta. Badan Pusat Statistik, (2000b). Statisitk Potensi Desa, Jakarta. Byron, R.P. and Takahashi, H. (1989). “An analysis of the Effect of Schooling, Experience and Sex on Earnings in the Government and Private Sectors of Urban Java,” Bulletin of Indonesian Economic Studies 25(1), 105–117. Cameron, L. A. (2000), “Poverty and Inequality in Java: Examining the Impact of the Changing Age, Educational and Industrial Structure,” Journal of Developing Economics, 62(1), 149–180. Daly, A. and Fane, G. (2002). “Anti-Poverty Programs in Indonesia,” Bulletin of Indonesian Economic Studies, 38(3), 309–329. Duflo, E. (2001). “Schooling and Labor Market Consequences of School Construction in Indonesia: Evidence from an Unusual Policy Experiment,” American Economic Review, 91(4), 795–813. Lanjouw, P. et al. (2001). “Poverty, Education, and Health in Indonesia: Who Benefits from Public Spending,” Policy Research Working Papers, No. 2739, The World Bank, Washington, D.C., December. Mincer, J. (1974). Schooling Experience and Earnings. New York: Columbia University Press. Psacharopoulos, G. (1994).“Return to Investment in Education: A Global Update.” The World Bank, Policy Research Working Paper, No. 1067. Suryadarma, D. et al. (2005).“A Reassessment of Inequality and its Role in Poverty Reduction in Indonesia.” SMERU Working Paper, SMERU Research Institute, Jakarta Trostel, P., Walker, I. and Woolley, P. (2002). “Estimates of the Economic Return to Schooling for 28 Countries.” Labor Economics. 9(1), 1–16. The World Bank, (2004). World Development Indicator, Washington, DC.

b1363-ch08

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 9 GROWTH AND INEQUALITY IN THAILAND Nobuki Sugita∗ Income disparity has been and still is an important issue of debate in economic development. This is especially so in the fast growing economies, such as those in Asia. In spite of its importance, there are obstacles for analysis. The most difficult part is the measurement of disparity, caused by the lack of consistent and comprehensive data. This chapter takes the example of Thailand and tries to overcome these obstacles by using distributional statistics, specifically the gross provincial products data, and analyzes the relationship between income disparity and economic development. The comparison with Japan, using the same method is also shown.

9.1. Economic Development and Income Disparity Income disparity has been and still is an important issue in economic development strategy. In addition to its political economy implications, it has been the source of academic debate for a long time. There are keywords related to this issue, such as Get Rich First Theory (Deng Xiaoping) and Trickle-down Theory. It also has geographical significance because income disparity is quite often associated with regional difference in development level and speed. This is especially important in political sense because of the pressure from the constituents.1 “Balanced development of land” was until recently the motto of regional development in Japan. In China, Great Western Development Strategy is meant to address the issue. This is also an issue of debate in academic world. The most famous manifestation is the so-called Kuznets’ inverse U-shape curve (Fig. 9.1), ∗ Professor at Economic Research Center, Graduate School of Economics, Nagoya University, Japan. 1 This can be exacerbated when political representation is distorted. This happens when demographic change is not properly reflected in the number of elected officials.

181

b1363-ch09

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

N. Sugita

182

Fig. 9.1.

Fig. 9.2.

Kuznets Curve.

Time Series and Cross-Section.

which dictates that the income inequality first rises with development but subsequently declines. Many attempts were made to verify or refute this assertion and it is still an open question. There are basically two ways to look at the problem. The one, which I call time series view, is to observe the development of a particular economy and track the relationship between income and disparity. The other, cross-section view, is to compare the development of various countries and see how the turning points move. This has much to do with the presence (or absence) of the late comer advantage (or disadvantage) (Fig. 9.2).

b1363-ch09

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Thailand

b1363-ch09

183

9.2. Measurement of Income Disparity The first, and probably the greatest, obstacle to analysis of income disparity is the availability of credible data. This is especially hard to overcome in developing economies, where even the basic statistics are often unreliable. Another issue is whether we should take wealth (stock data) or income (flow data). It is often the former that is really the problem. But the data on stock is even more difficult to obtain and less reliable than that on flow. So most analyses are based on solely the income data, which is inevitable given the technical obstacles. In this chapter, too, we consider only the income data. There are roughly three approaches to analyze the income data, each of which has pros and cons. In what follows, I try to explain and compare their strong and weak points, both in theoretical and practical aspects. The first approach is very simple. Compare the income level of the top bracket with that of the bottom one and calculate the ratio. This has few advantages: Calculation is simple (pocket calculator will do) and the result is intuitively easy to understand even for non-experts. Analogy to compare the salary of the CEO with that of newly hired employee comes to mind. On the other hand, it is hardly a robust and appropriate measure: It is very susceptible to how the brackets are divided. It may be possible to fix the number of division of income level for meaningful comparison but the information of the mid-level income is apparently thrown away. The second approach is to use micro level, individual data and calculate the Gini coefficient. This has a solid theoretical background and the comparison among different economies or over time is generally reliable. It also has an advantage that, depending on the design of the data collection, the effects of government policy on income redistribution can be analyzed.2 The main difficulty of this methodology is that it is costly. Usually, the sample size is not so large and the survey is not conducted so frequently.3 Another problem is that the data set is often unavailable to outside researchers because of its sensitivity and this precludes third-party verification.

2 The most widely used Gini coefficient in Japan is based on the survey named “Income Redistribution Survey”, conducted by the Ministry of Health, Welfare and Labor. 3 “Income Redistribution Survey” is conducted every three years; its sample size is less than 10,000 and it is made public only two years after the survey is conducted.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

N. Sugita

184

Fig. 9.3.

Calculating Quasi-Gini Coefficient.

The third way is situated in the middle of the aforementioned two methods. It uses semi-macro, or distributional, statistics and calculates a quasi-Gini coefficient, with the assumption that in each region income is equally distributed (Fig. 9.3). There are a few possibilities of division, such as social classes and income levels but the most convenient is geographical since, in most cases, countries are composed of several local entities (regions, states, prefectures and so on) with their own administrative bodies. The principal shortcoming of this approach is that, as in the first, it is susceptible to the distortion depending on the fineness of division. The finer the division, the more inequality the calculation shows. On the other hand, the calculation itself is not complicated: Spreadsheet level is enough. The combination of easy calculation and wide availability of data also means that the third-party verification can be carried out effectively, thus ensuring the robustness of the results. Due to these merits, there are several results obtained for various economies.4 In the following sections I will show the results for Thailand and compare them with those of Japan.

9.3. Regional Income Disparity in Thailand The data used to calculate the disparity is based on Gross Regional and Provincial Products, compiled by the National Economic and Social Development Board (NESDB) of the Thai government. In Thailand, 4 See

Fig. 9.2 of Sakamaki (2006) for examples.

b1363-ch09

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Thailand

Fig. 9.4.

b1363-ch09

185

Quasi Gini Coefficient of Thailand.

provincial, regional and national data are consistent in the sense that the sum of the provincial figures becomes the regional figure and the same is true for regional and the national figures. The country is classified into 75 provinces and the Bangkok Metropolitan Area. Figure 9.4 shows the results using the provincial data. We can see that during the period of boom from the early 1980s to early 1990s there was a general trend of increasing inequality but subsequently inequality declined until the Asian financial and economic crisis, of which Thailand was the first epicenter. Rapid recovery from the crisis was associated with increased inequality but the trend upward stabilized in the 2000s. The relation with economic growth is better observed when the growth pattern is superimposed on it (Fig. 9.5). We can generally see that high growth is accompanied by an increase in inequality but there is some minor difference in timing, such as decreased inequality preceding the decline of economic growth before the 1997 crisis.

9.4. Regional Income Disparity in Japan, Comparison with Thailand Prefectural accounts in Japan are compiled and published by each prefecture and the Cabinet Office assembles and publishes them in one volume. Although they are made under the same principles, they are not consistent

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

N. Sugita

186

Fig. 9.5.

Quasi Gini Coefficient and Economic Growth.

Fig. 9.6.

Quasi Gini Coefficient of Japan.

with the national accounts because the sum does not add up to the national figure. But as long as they are compiled under the same standard, it is not a problem when we calculate the inequality measure. There are 47 prefectures in Japan. The number of division is smaller than in Thailand and a caution may be needed to compare the relative level of inequality between the two countries. Figure 9.6 shows the results of the calculation. We can observe the upward trend in the 1980s (the decade of boom or bubble) with subsequent decline after the burst of the bubble.

b1363-ch09

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Thailand

Fig. 9.7.

b1363-ch09

187

Quasi Gini Coefficient and Economic Growth.

The downward move reversed in the middle of the 1990s but the following rise is not as steep as in the 1980s. There has been much debate in Japan on whether or why the income inequality increased but by looking at this figure we see that it is not so pronounced. Figure 9.7 shows the relationship between economic growth and inequality. Although we see a large difference in average growth rate before and after the burst of the bubble, the level of inequality is more or less the same. If we compare the two countries, we immediately see the difference in the level of inequality.5 In terms of its variation over time and with relation to economic growth, we see a similar trend: In the boom years inequality increases.

9.5. Usefulness of Distributional Statistics and their Limitations We have observed that the use of regional data, as an example of distributional statistics, offers important insight into the state of inequality in economy at relatively low cost for analysis. One of the merits of it is that 5 As

noted earlier, the number of division is different but the figures calculated here for the two countries are sufficiently apart so we do not have to change the conclusion.

June 5, 2012

11:23

188

9in x 6in

Growth and Inequality: An International. . .

N. Sugita

it provides us with the time series variation for a relatively long period, which gives us enough degree of freedom for more sophisticated statistical work. We have to be aware, however, the weakness of this approach, such as its sensitivity to the number of division. 9.6. Policy Implications and Future Direction of Research This chapter has illustrated the general tendency of correlation between high economic growth and increased inequality. It might be safely said that the benefit derived from good economic performance can outweigh the discontent caused by greater inequality in most cases. But when growth slows down, the political pressure built up during the boom (and larger inequality) comes to surface. This seems to be the state of affair in many economies now. We have to wait a little longer to see the outcome of the current crisis but with this method, the results can be obtained relatively quickly and at low cost. For further study, I suggest that more comprehensive survey for more countries should be conducted. With sufficient number of countries, cross sectional analysis becomes feasible. References Economic and Social Research Institute. Cabinet Office, National Accounts, Various Issues. Economic and Social Research Institute. Prefectural Accounts, Various Issues. Ministry of Health, Welfare and Labor, Income Redistribution Survey, Various Issues. National Economic and Social Development Board, Regional and Provincial Products, Various Issues. National Economic and Social Development Board, Thai Government, National Accounts, Various Issues. Nozaki, K. (2007). “Regional Disparity in Thailand — Its Actual State and Significance to China.” Economic Science, 55(3) (in Japanese). Sakamaki, T. (2006). “Regional Disparity and Land Development Policy in East Asian Countries.” Journal of Development Finance Institute, Japan Bank for International Cooperation, (in Japanese). Sakamaki, T. (2007). “Regional Disparity in East Asian Countries — Variation in Regional Disparity Measure and Regional/Industrial Structure.” Journal of Development Finance Institute, Japan Bank for International Cooperation, (in Japanese).

b1363-ch09

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Thailand

b1363-ch09

189

Sugita, N. (1997). “Thai Economy in Figures.” Monthly report of the Japanese Chamber of Commerce and Industry in Bangkok, (in Japanese). Yamashita, M. (2004). Economic Growth and Inter-/Intra- National Income Disparity. ESRI Discussion Paper Series No. 114, Economic and Social Research Institute, Cabinet Office, August 2004 (in Japanese).

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch09

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 10 GROWTH AND INEQUALITY IN SINGAPORE Kong Weng Ho∗ While the strategy of openness had earned Singapore rapid economic growth, upward social mobility, and possibly decreasing inequality in the early years of development, the more recent years saw increasing inequality and with it an underlying possibly diminished upward intergenerational mobility due to skill-biased growth processes, skill-biased parental influence, liberalization in the education industry, and structural changes in the society which hurt the human capital accumulation of children in families under economic and intra-household stresses. In particular, the paternal influence on educational aspiration and attainment is more pronounced than the mother’s. Non-Chinese and youths from disrupted families are worse off in both educational aspirations and educational attainment.

10.1. Introduction Singapore’s economic growth has been spectacular over a period close to five decades, making many developing countries envious of her economic performance. Using 2005 prices, the real Gross Domestic Product (GDP) per person in 2009 was 11.9 times that in 1960, implying an average real growth rate of 5.3% per annum. Figure 10.1 depicts the phenomenal growth of the Singapore economy from 1960 till 2009. Although there were some dips during several episodes of recessions, the Singapore economy recovered rather quickly. Would such a rapid and persistent growth of the Singapore pie translate to opportunities for all? Is the growth process an even one, benefiting the ∗ Senior Lecturer at School of Business, SIM University. The chapter was published earlier in Crisis Management and Public Policy: Singapore’s Approach to Economic Resilience, Singapore: World Scientific.

191

b1363-ch10

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch10

K. W. Ho

192

Per Capita Real GDP, S$ in 2005 Prices 60000

50000

40000

30000

20000

10000

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1970

1972

1968

1966

1964

1962

0 1960

June 5, 2012

Year

Fig. 10.1.

GDP Per Capita of Singapore.

skilled as well as the unskilled? How do the children from the less well-to-do fare? These are some of the questions we want to explore in this paper. In this chapter, we simply measure economic growth by the change in real GDP per person over time and propose to measure opportunity by the inverse of the dependence of one’s economic status on one’s parents. If one’s economic status in terms of income, educational attainment, or occupation, is highly dependent on parental economic status, then we would say that intergenerational mobility is low, which is equivalently, for our purpose in this chapter, a low level of opportunity in the society. Intergenerational mobility or opportunity is a dynamic measure of inequality. One static measure of inequality is the Gini coefficient, which has a value in the range from zero to one. The Gini coefficient has a larger value when the income distribution is more unequal. When income inequality is high, it has two effects at the ground level of the individuals: One positive and the other negative. The positive effect is an incentive for the less-to-do to upgrade via training, educational investment, and searching of opportunity for themselves and their children. The negative effect could be merely psychological such as feeling of being envious of others or of unfairness. But if a high level of inequality is also related to access to resources for investment in physical and human capital so that opportunities in the future can be improved,

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

193

then inequality will become entrenched and sticky in the sense that there is no opportunity to move from one economic class to another for oneself or for one’s children, resulting in social stratification or intergenerational immobility. How is social mobility or intergenerational mobility measured? Before answering that question, we want to understand an important concept of social mobility: Absolute vs. relative mobility. Absolute upward mobility for all is possible but it is impossible to have relative upward mobility for all as when a person or the person’s offspring climbs up the social ladder, another person or this other person’s offspring must have slipped down the social ladder. The weights of relative positions in a ranking ladder at a point in time represent the distribution of socioeconomic status. Over time and over generations, changes in the weights of each relative position in a social ladder give a measure of the dynamic changes in the distribution of people in the social ladder, and hence there is exactly why relative social mobility is related to inequality in society. Forces affecting social mobility, which is a dynamic measure across time, will influence how a static measure of inequality will evolve dynamically. Using aggregate data, we will examine the extent of social mobility in Singapore; when we have micro data linking generations, we will provide estimates of intergenerational mobility. We will examine a simple demand-supply framework of social mobility and inequality to understand how structural changes in technology, demography, and government policies in education and the labor market may affect the joint equilibrium of social mobility and inequality. In this theoretical model, we will define inequality as wage inequality, which is the ratio of skilled wage to unskilled wage. We will use this model to investigate how educational liberalization, population expansion via skill-biased immigration, and family disruption may worsen or improve inequality and social mobility in Singapore. The rest of the chapter is as follows. Section 2 will identify the main causes of income inequality in Singapore and present some empirical findings. Section 3 provides a brief survey of existing studies on social mobility in Singapore. Section 4 will present our findings on social mobility in Singapore, using latest available data, and with a focus on intergenerational mobility in education. We will present a simple demand-supply framework of upward mobility and wage inequality in Sec. 5. Section 6 will discuss the impact of educational liberalization, population expansion, and rising family disruption. Section 7 concludes the chapter.

June 5, 2012

11:23

194

9in x 6in

Growth and Inequality: An International. . .

K. W. Ho

10.2. Causes of Income Inequality in Singapore Singapore is a very open society, welcoming foreign investment, imports, foreign talents as well as foreign unskilled migrant workers. Being open to international forces could be a reason for widening inequality as documented for both advanced and developing countries in the United Nations’ Human Development Report (1999). Furthermore, Taylor (2000) showed that external economic liberalizations of eleven developing countries had surely widened the skilled–unskilled wage gap despite ambiguous impact on economic growth. On the other hand, globalization seemed to have aided Singapore in terms of economic growth. Linking herself to a world of advanced ideas, Singapore is able to grow rapidly as a technology follower. Ho and Hoon (2009) showed theoretically in a model and empirically using Singapore data that channels of technology diffusion measured by (a) ratio of stock of G5 foreign direct investment to total capital stock, (b) ratio of imports of machinery and transport equipment from G5 countries to GDP, and (c) quality of learning as in the ratio of tertiary enrolment to employment contribute to the multi-factor productivity growth rate and hence economic growth. While Ho and Hoon (2009) did not examine the impact of advanced ideas transmission on income inequality, we suspect that these channels of global technology diffusion could be skill-biased and hence may bring about a growth process accompanied by rising income inequality in Singapore. Singapore had relied on foreign workers, both skilled and unskilled, even before her independence in 1965 as she was and still is an immigrant society to a certain extent. The locals do not want to take up low-paying manual jobs and hence an import of unskilled migrant workers is necessary. An inflow of unskilled foreign workers will reduce the upward movement of unskilled wage, if any at all, even with economic growth. On the other hand, there is also a high demand for the scarce high-end skills in Singapore and the internationally mobile foreign talent will command a world competitive wage which accelerates with economic growth in Singapore and other parts of the world. Furthermore, non-wage or asset income, usually within the portfolio of the skilled but not the unskilled, grows in tandem with economic growth. Competing for international investment and highly skilled and internationally mobile talents, Singapore has revised the marginal income tax rates and the corporate tax rate over the years, resulting in a possibly less progressive tax structure compared to the past. Such a policy encourages

b1363-ch10

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

195

workers to move up the income ladder with the correct incentives but at the same time has a potential adverse impact on income inequality. In Singapore, capital gains are not taxed, and estate duty has been abolished in 2008. The Goods and Services Tax (GST) was introduced in April 1994 at 2%. It had been raised gradually to 7% in June 2008. As a consumption tax, GST is regressive in nature. Realizing the need to assist low wage workers in Singapore, the Ministerial Committee on Low Wage Workers was set up in 2005 to recommend measures to improve employability and income security for low wage workers and to help families break out of the poverty cycle. Six spokes of assistance within the Workfare framework were outlined: Rewarding work; Ssocial support to enable work; higher skills for better jobs; expand job opportunities; creating hope for the future; and sharing in the nation’s progress. The Ministerial Committee on Low Wage Workers (2009) reported on 7 June 2009 that the government had spent over S$1.1b from 2006 to 2008 helping low wage Singaporeans. As a result, “since 2006, low wage workers have made significant progress in their incomes and availed themselves to job opportunities. The wages of the 20th percentile fulltime employed resident increased from $1,200 a month in 2006 to $1,310 in 2008.” “In addition, Singapore’s income inequality has reduced, as reflected by the drop in Singapore’s Gini coefficient form 0.489 in 2007 to 0.481 in 2008, the first decline since 1998. After adjusting for government benefits and taxes, the Gini coefficient drops even further, to 0.462.” Our question: Is the drop caused primarily by the recession or the initiatives of the government? It could be a mixture of both and we will be able to answer this question more confidently with a few more years of data in the future. Based on the discussion above, Ho (2010a) explores a time series investigation on the Gini coefficient of Singapore. The Gini coefficients are obtained from the World Institute for Development Economics Research (2008) for the earlier years and from Singapore Department of Statistics (2010) for the later years. The results and interpretations are extracted from Ho (2010a) and presented in Table 10.1. Regression 1 in Table 10.1 suggests an inverse Kuznets curve, which is seemingly surprising. Regression 2 highlights the importance of technology transmission channels in influencing income inequality in Singapore, a small open economy in a world of ideas. An increase in G5 foreign ownership of capital in Singapore will lead to an increase in income inequality. Similarly, with more tertiary students relative to employment, a proxy for quality of learning, Gini will increase. Both channels are likened to skill-biased

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

K. W. Ho

196 Table 10.1.

Determinants of Income Inequality (Gini).

ln(G5 Imports of machinery and transport equipment/GDP) ln(Tertiary enrolment/ Employment) ln(G5 FDI stock/Capital stock) ln(Real GDP per worker) ln(Real GDP per worker) 2 Growth rate Growth acceleration Constant Observations R-squared

b1363-ch10

(1) Gini

(2) Gini

(3) Gini

(4) Gini

— — — — — — −1.665∗∗∗ (0.281) 0.080∗∗∗ (0.013) — — — — 9.147∗∗∗ (1.511) 37 0.753

−0.036∗∗∗ (0.012) 0.043∗∗∗ (0.008) 0.047∗∗∗ (0.011) — — — — — — — — 0.66∗∗∗ (0.016) 34 0.846

−0.084∗∗∗ (0.022) −0.025 (0.036) 0.031∗ (0.018) 1.274∗∗ (0.514) −0.054∗∗ (0.023) — — — — −7.058∗∗ (3.007) 34 0.876

−0.039∗∗∗ (0.012) 0.041∗∗∗ (0.008) 0.052∗∗∗ (0.011) — — — — 0.194∗ (0.099) −0.121∗ (0.070) 0.656∗∗∗ (0.016) 34 0.865

Notes: Standard errors in parentheses. ∗ significant at 10%; ∗∗ significant at 5%; Source: Ho (2010a).

∗∗∗ significant

at 1%.

forces which will raise income inequality. Imports of advanced machinery from G5 countries, however, will reduce income inequality. Regression 3 shows that with these channels as controls on the right-hand side, the usual Kuznets curve will be revealed. Regression 4 shows an alternative version of the “Kuznets curve” in terms of growth rates. Structural changes while raising growth lead to an increase in income inequality. Accelerated growth beyond the normal growth will, however, reduce income inequality, possibly because more unskilled work hours are demanded. The results reported in Ho (2010a) suggest the growth process for Singapore as a technology follower is skill-biased and could explain the evolution of income inequality in Singapore rather well. In particular, technology diffusion of advanced ideas via foreign direct investment is skill-biased and the quality of learning or the extent of tertiary enrolment relative to employment, which facilitates technology diffusion of advanced ideas, is also skill-biased. However, technology diffusion via imports of advanced machinery is likely not to be skill-biased, benefiting the unskilled workers relative to the skilled workers.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

197

10.3. Past Studies on Social Mobility in Singapore Studies on social mobility in Singapore were limited and often used secondary data or a one-time survey as panel data on economic status which are not available. The literature review in Ho (2010b) showed that absolute upward mobility had been high in Singapore due to sustained economic growth and rapid expansion of educational opportunities especially for the families at the lower end of the social ladder, implying also high intergenerational mobility in the past. However, parental background remained an important determinant of one’s economic status. See, for example, Chiew (1991) and Ko (1991) who used data collected in 1983. Using data based on a 2001 study, Tan (2004) also found an important role of father in transmitting occupational status. Ng, Shen, and Ho (2009) compared Singapore with U.S. data and found similar intergenerational immobilities. As both countries have similar economic realities, welfare systems, education regimes, labor structures, and high inequalities, the similar results are hence not surprising. To maintain global competitiveness, policy makers of these two countries face the daunting challenge of overcoming immobility and inequality. Using data from the same Singapore youth survey conducted in 2002 as in Ng, Shen, and Ho (2009) who studied income mobility, Ng and Ho (2006), and Ong and Ho (2006) considered intergenerational transmission of educational attainment and occupational attainment, respectively. In particular, Ng and Ho (2006) found that youths whose parents are divorced had their educational attainment lowered by 1.8 to 1.9 years worth of schooling. A Swedish study by Jonsson and Gahler (1997) has also demonstrated that children who have experienced family dissolution or reconstitution show lower educational attainment at age 16. A single parent, usually the mother, forced to work outside instead of relying on her husband, will have less time in the supervision of the child in school work. The child is burdened with psychological costs caused by the divorce of his or her parents, leading to lower Emotional Quotient (EQ) and Social Quotient (SQ) compared to children from intact families enjoying love and care from both parents. Can we use principles of economics to explain the transmission of educational status from fathers to their children? Yes, as fathers are usually the bread winners, their educational attainment will determine usually the amount of income they will earn for their families and children. A highly educated man is also likely to marry a highly educated woman through

June 5, 2012

11:23

198

9in x 6in

Growth and Inequality: An International. . .

K. W. Ho

a process called positive assortative matching. As the price of time of the well-to-do couple is high, they are likely to choose fewer children and invest more in the education of their children relative to couples who are less wellto-do. Consequently, we observe an educational- and skill-biased parental influence on the educational attainment of their children. Findings from Ong and Ho (2006) on the transmission of occupational status showed that three parental background variables, namely father’s education, mother’s education, and father’s occupation had stronger indirect effect on job attainment of the youth via influencing educational attainment of the youth than the direct effect on job attainment. The status of parental divorce had strong negative effects, both directly on job attainment and indirectly effect on the educational attainment of the youth. Ho (2010b) concluded that absolute upward mobility had been high in the past but recent studies on income mobility pointed that Singapore could have become less mobile.

10.4. New Empirics on Singapore Using Census data and household survey data, at the aggregate level, Ho (2010b) found significant absolute upward mobility in educational attainment as well as occupational status from 1990 to 2005. The driver of the absolute upward mobility is a rapidly growing Singapore economy, giving more opportunities to educational attainment and occupational upgrading, and reflecting in the distributional change in education and occupation. How about the distribution of income growth over the years? Figure 10.2 shows the average nominal growth rate of average monthly income from work per household member among employed households by deciles from 1995 to 2009.1 Although the growth rates are nominal as real values are not available, if inflation over the same period did not penalize the rich more than the poor, the information on nominal values will remain useful. Household income inequality has worsened as the lower deciles have experienced lower growth rates than the upper deciles. Economic growth in Singapore from 1995 to 2009 had been uneven, benefiting the upper end more than the lower end. Unfortunately, Fig. 10.2 does not provide 1 Computation

and (2010).

is based on data from Singapore Department of Statistics (2007), (2009),

b1363-ch10

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

199

Nominal Annual Growth Rate of Average Monthly Income from Work Per Household Member Among Employed Households by Decile from 1995 to 2009 7%

6%

5%

4%

3%

2%

1%

0% 1-10

11-20

21-30

31-40

Fig. 10.2.

41-50 51-60 Decile

61-70

71-80

81-90

91-100

Growth of Income.

Notes: This diagram has used the latest data to update a similar figure found in Ho (2010b). Nominal values are used as real values are not available from published data from the Government of Singapore.

information on the mobility of households across deciles and does not contain information across generations. However, it does suggest that opportunity for the families at the lower end of the social ladder could be diminished. Therefore, it is all the more important to gather data linking generations systematically for future analysis on income mobility. Next, we will present our analysis using data from the two surveys commissioned by the National Youth Council in 2002 and 2005. The samples were drawn from sampling frames obtained from the Department of Statistics and matched the national youth population by nationality, age, gender and ethnicity. Reports on the surveys are published in Ho and Yip (2003) and Ho and Chia (2006). Pooling the two samples for our regression analysis, we examine the determinants of the educational aspiration of youth aged 15 to 18 who are students, and the educational attainment of working youth aged 23 to 29 separately. Missing observations in parental education have been imputed. For brevity, we will only report the regressions results using log of years of schooling. The detailed findings are available in Ho

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

K. W. Ho

200 Table 10.2.

Intergenerational Education Mobility: Log of Years of Schooling. (1) (2) (3) (4) ln(Aspiration) ln(Aspiration) ln(Attainment) ln(Attainment)

ln(Paternal eductaion) ln(Materal education) ln(Parental income) Female Non-Chinese Family disruption Year-2005 Constant Observations R-squared

0.031∗∗ (0.013) 0.011 (0.012) 0.029∗∗∗ (0.007) 0.025∗∗∗ (0.009) −0.057∗∗∗ (0.011) −0.055∗∗∗ (0.019) 0.023∗∗ (0.010) 2.358∗∗∗ (0.055) 835 0.154

Notes: Standard errors in parentheses. ∗ significant at 10%; ∗∗ significant at 5%; Source: Ho (2010a).

0.046∗∗∗ (0.013) 0.019∗ (0.011) — — 0.025∗∗∗ (0.009) −0.067∗∗∗ (0.011) −0.061∗∗∗ (0.019) 0.020∗∗ (0.010) 2.541∗∗∗ (0.027) 835 0.132 ∗∗∗ significant

0.047∗∗∗ (0.015) 0.027∗ (0.015) 0.029∗∗∗ (0.011) 0.025∗ (0.015) −0.134∗∗∗ (0.017) −0.072∗∗∗ (0.021) 0.024 (0.015) 2.174∗∗∗ (0.074) 1217 0.135

0.058∗∗∗ (0.015) 0.034∗∗ (0.013) — — 0.026∗ (0.015) −0.142∗∗∗ (0.017) −0.084∗∗∗ (0.020) 0.024 (0.015) 2.361∗∗∗ (0.026) 1217 0.124

at 1%.

(2010a). Taken from Ho (2010a), Table 10.2 reports the impact of an additional per cent of parental schooling on the schooling of the youth. Regressions (1) and (2) of Table 10.2 report the determinants of the educational aspirations of students aged 15 to 18 while Regressions (3) and (4) show the results for working youths aged 23 to 29. The influence of fathers on education aspiration and attainment is more pronounced than the mother. Parental income is an important influence too. Female teenage students are more likely to have higher educational aspirations but gender has marginal or insignificant influence on educational attainment. NonChinese and youths from disrupted families are worse off in both educational aspirations and educational attainment. The coefficient of paternal or maternal education in the regressions represents an intergenerational persistence in education, or an inverse of intergenerational mobility. A lower coefficient means higher educational opportunity not linked to parental background measured by paternal or maternal education. Once parental income is added on the right-hand side, the persistence coefficient will be reduced or become insignificant, implying

b1363-ch10

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

201

a mediating role of parental income. Does that mean economic growth will then increase educational opportunity through the channel of increasing parental income? The answer is exactly opposite if economic growth is uneven, benefiting the well-to-do more than the lower income groups. 10.5. Demand and Supply of Social Mobility A simple demand-supply framework of social mobility and wage inequality incorporating insights from the findings reported above has been developed in Ho (2010b). We will outline the setup of the model here. Simply assume the price of upward mobility to be wage equality, which is defined as the ratio of unskilled wage to skilled wage. First, consider how parents decide on educational investment in their children. Altruistic parents maximize overall welfare comprising of their own welfare and the future welfare of their children, which is influenced by parental investment on education of the children. A child is born unskilled but is likely to be converted to become a skilled adult with investment in education. When wage equality is low, the opportunity cost, or price, of educating an unskilled child to become a skilled adult will be low, and the quantity demand of upward mobility will be high, implying a downward sloping demand curve. A decrease in government subsidy to public education, for example, which influences the parental decision on education based on returns to education, will shift the demand curve of upward mobility downward. Next, we examine the influence of upward mobility on wage equality. When there are more conversions of unskilled children to skilled adults (higher upward mobility), there will be a larger supply of future skilled workers relative to the supply of future unskilled workers, leading to a lower wage inequality or higher wage equality. Hence, the price of upward mobility (defined as wage equality) is positively related to upward mobility, giving an upward-sloping supply curve. Structural changes in technology or conditions in the labor market will shift the supply curve of upward mobility. The intersection of the demand and supply curves will give the equilibrium levels of upward mobility and wage equality in the society. In the next section, we will use the demand-supply framework developed in Ho (2010b) to discuss issues related to the empirical findings and future prospects on intergenerational upward mobility and wage inequality in Singapore. We will use the framework to examine the impact on educational intergenerational mobility.

June 5, 2012

11:23

202

9in x 6in

Growth and Inequality: An International. . .

K. W. Ho

10.6. Theoretical Impact of Policies and Trends In this section, we will examine the impact of the process of educational liberalization in Singapore, the consequence of population expansion via skilled immigrants, and the implication of potentially rising cases of family disruption in Singapore. 10.6.1. Liberalization of the Education Industry A liberalization process of schools in Singapore began in 1987 when three top-performing boys’ secondary schools turned independent, which brought about public criticism over the elitist nature and the high fees charged by the independent schools. Autonomous School status was first granted to six secondary schools in 1994, giving parents and students a wider choice of schools. Integrated Program (IP) was first implemented in 2004, allowing IP schools to skip the “O” level examinations. Currently, there are a total of 13 independent schools, 36 autonomous schools, and 12 IP schools out of 168 secondary schools, 22 junior colleges or centralized institutes. These “liberalized” schools are academically more selective than other schools and it would be interesting to examine the family background of students in these schools. If they have parents with higher education and income compared to students in other schools, the liberalization process would bring about a skill-biased parental influence on education. Local schools compete with one another in some ranking or benchmarking exercises. In a conference on benchmarking for performance in education organized by the World Bank, Ho (2010c) asked if local benchmarking would translate to competition among schools with parental assistance whether lower intergenerational mobility and higher inequality would also become the consequences despite gains in efficiency. In the demand-supply framework outlined earlier, given liberalization in the education industry, the demand curve will then shift down, resulting in a reduced social upward mobility and lower wage equality, which is a result similar to a process of privatization in education as private schools rely more on the contribution from parents both in terms of time and money. 10.6.2. Population Expansion via Skill-Biased Immigration Singapore is targeting a larger population of 6.5 millions within the next 20 to 30 years, from the current level of 4.9 millions despite a very low total

b1363-ch10

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

203

fertility rate of 1.28 per resident female in 2008. Hence, the expansion of the population is likely to be facilitated by skill-biased immigration, and as a result, the skill-biased parental influence will be enhanced as the skilled immigrant parents will have an advantage over the existing unskilled parents in the educational investment of their children. This will shift down the demand curve for upward mobility. However, the supply curve for upward mobility will also shift down or pivot clockwise given a skill-biased immigration for two reasons: first, skilled immigrants are more capable of producing children who will become future skilled workers; second, an influx of skilled immigrants will induce an increased demand for skilled jobs, resulting in an increased skill premium in wages. Therefore, the net effect on upward mobility is ambiguous given a downward shift in the demand curve and a downward shift or clockwise rotation of the supply curve as illustrated in Fig. 10.3. It is, however, unambiguous that wage equality will be reduced. Wage Equality Supply Demand

A

B

Upward Mobility Fig. 10.3.

Skill-biased Immigration.

Source: Ho (2010b)

10.6.3. Damage of Family Disruption Figure 10.4 shows the rising divorce rates in Singapore since 1980. How will a rising trend of family disruption influence social mobility and inequality? Children from broken families do not receive complete love from both parents. The single parent, likely to be the mother, has to struggle between

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch10

K. W. Ho

204

Gross Divorce Rate Per 1000 Married Resident Males/Females 9 8 7 6 5 4 3 2 1 0 1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

Year Males

Fig. 10.4.

Females

Rising Divorce Rates in Singapore.

household production, market activities to bring in the dough, and coaching the child in school work, and as a result, the chances of the child becoming skilled will be reduced, implying a downward shift of the demand curve. Therefore, a rising rate of divorce will bring about lower upward mobility and wage equality. This theoretical result is consistent with the empirical result reported in Table 10.1.

10.7. Conclusion This chapter has collated some empirics on the Singapore economy and found that the strategy of openness had earned Singapore rapid economic growth, upward social mobility, and possibly decreasing inequality in the early years of development. However, the more recent years saw increasing inequality and with it an underlying possibly diminished upward intergenerational mobility due to skill-biased growth processes, skill-biased parental influence, liberalization in the education industry, and structural changes in the society which hurt the human capital accumulation of children in families under economic and relational stress. While economic growth across the board has increased the economic pie and opportunity for all, whether this channel of social progress will remain an important research and policy

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Singapore

b1363-ch10

205

question as the nature of economic growth and its benefits have become more uneven distributed. Furthermore, we would like to suggest that further research has to be conducted in Singapore on how these macro trends in economic and non-economic variables may have an influence on the wellbeing of Singaporeans as well as immigrants living in Singapore.

References Chiew, S.K. (1991). “Social Mobility in Singapore.” In Social Class in Singapore, Quah, S., Chiew, S.K., Ko, Y.C. and Lee, S.M. (Eds.) Singapore: Times Academic Press. Chung, K.Y., (1991). “Status Attainment.” In Social Class in Singapore, Quah, S., Chiew, S.K., Ko, Y.C., and Lee, S.M. (Eds.) Singapore: Times Academic Press. Ho, K.C. and Yip, J. (2003). YOUTH.sg: The State of Youth in Singapore. Singapore: National Youth Council. Ho, K.C. and Chia, W. (2006). YOUTH.sg: The State of Youth in Singapore 2006. Singapore: National Youth Council. Ho, K.W. and Hoon, H.T. (2009). “Growth Accounting for a Technology Follower in a World of Ideas: The Case of Singapore.” Journal of Asian Economics, 20(2), 156–173. Ho, K.W. (2010a). “Inequality of a Small and Globalized Economy in a World of Ideas: The Case of Singapore.” In Income Inequality: A New Threat to Globalizing Economies, April 10–12, 2010, Japan: Kyoto Sangyo University. Ho, K.W. (2010b). “Social Mobility in Singapore.” In Management of Success: Singapore Reassessed, Chong, T. (Ed.). Singapore: Institute of Southeast Asian Studies. Ho, K.W. (2010c). “Educational Mobility across Generations and Inequality: The Case of Singapore”, In Benchmarking Education Systems for Results: East Asia Regional Conference, June 21–23, 2010. Singapore: World Bank. Jonsson, J.O. and Gahler, M. (1997). “Family Dissolution, Family Reconstitution, and Children’s Educational Careers: Recent Evidence of Sweden.” Demography, 34, 277–293. Ministerial Committee on Low Wage Workers, (2009). Progress Report of Ministerial Committee on Low Wage Workers. Singapore: Ministry of Manpower. Ng, R. and Ho, K.W. (2006). “Intergeneration Educational Mobility in Singapore: An Empirical Study.” YouthScope, 1, 58–73. Ng, I., Shen, X.Y. and Ho, K.W. (2009). “Intergenerational Earnings Mobility in Singapore and the United States.” Journal of Asian Economics, 20(2), 110–119. Ong, J.H. and Ho, K.W. (2006). “Intergeneration Occupational Mobility in Singapore: An Empirical Study.” YouthScope, 1, 22–39. Singapore Department of Statistics, (2007). “Key Household Income Trends, 2006.” Occasional Paper on Income Statistics, February 2007, 1–11.

June 5, 2012

11:23

206

9in x 6in

Growth and Inequality: An International. . .

K. W. Ho

Singapore Department of Statistics, (2009). “Key Household Income Trends, 2008.” Occasional Paper on Income Statistics, January 2009, 1–10. Singapore Department of Statistics, (2010). “Key Household Income Trends, 2009.” Occasional Paper on Income Statistics, February 2010, 1–10. Taylor, L. (2000). External Liberalization, Economic Performance, and Social Policy. New York: Oxford University Press. Tan, E.S., (2004). Does Class Matter? Social Stratification and Orientations in Singapore. Singapore: World Scientific Publisher. United Development Program, (1999). Human Development Report 1999. New York: Oxford University Press. World Institute for Development Economics Research, (2008). World Income Inequality Database, United Nations.

b1363-ch10

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Chapter 11 GROWTH AND INEQUALITY IN JAPAN A STUDY BASED ON THE FAMILY BUDGET SURVEY

Fumio Makino∗

11.1. Introduction The relationship between education and increasing income inequality has been getting more and more attention as income inequality has been increasing in Japan recently. If income inequality creates disparity in education somehow, it will be clear that it is a serious problem for a society as a whole since it could be a reason to hamper inter-generational social mobility and perpetuate inequality over generation. While much study about this issue has been done for decades by sociologists considering social stratification or the hierarchical arrangement of people in a society,1 economists have paid less attention about it until recently.2 In this chapter, I deal with some problems relating to income inequality and education using Family Income and Expenditure Survey (hereafter refers to FIES ) carried out by Statistical Bureau of Ministry of Internal Affairs and Communications of Japan. Relating to it, I also describe a problem about the regional income disparity and difference of the scores in annual achievement test done by Tokyo Metropolitan Government. In Sec. 2, I overview a trend on expenditure for education by income level using FIES, and also make an international comparison of household expenditure on education. In Sec. 3, I do a quantitative analysis about the inter-relationship between three factors of household spending on education, enrollment rate ∗ Professor

at Institute of Comparative Economic Studies, Hosei University, Japan. Kikuchi (1990, 2003), Kariya (2001) and Yoshikawa (2009). 2 For example, Tachibanaki (2004). 1 see

207

b1363-ch11

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch11

F. Makino

208

in higher education and income level, and show some examples of the relationship. Section 4 is a summary of this chapter.

11.2. Income Inequality and Household Education Expenditure 11.2.1. Long-term Trend of Income Disparity The upper line of Fig. 11.1 shows changes in real income before taxes for workers households3 (not including agricultural, forestry and fisheries households). Real income had been increasing every year until 1997, when (10,000 yen/month) 4.5

60

55

50 4.0 45

40

35 3.5 30

25

20 1980 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 2000 1 “Real income before tax”

Inequality of real income (right scale)

3.0 2

3

4

5

6

7

8

9

10

Inequality of disposable income (right scale)

Fig. 11.1. Real Income and Income Inequality. Note: 1. Income inequality is defined as a ratio of the income of the top decile to that of the bottom decile. Source: FIES, respective years http://www.stat.go.jp/data/kakei/longtime/zuhyou/ j1602000.xls. 3 A household refers to a group of two or more person sharing a dwelling and living expenses as well as a one person household. Family member who does not live together is not included in member of household. In case of analyzing household education expenditure, attention should be paid how to deal with remittances to child education who lives apart from his/her parent(s).

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Japan

b1363-ch11

209

consumption tax rate of 3% was raised to 5% and some major financial institutions went bankrupt. Due to deep economic depression caused by financial crisis, household nominal income went down from 1998 to 2003. While it has remained constant during recent years, its level is as low as that in the 1990s. Calculating the real income after the adjustment to take inflation, its level in 2010 is almost same as that in the year of 1987. Calculating household distribution by annual income level in 1995 when income inequality began to rise, the mode value for annual income was between 10 million to 12.5 million yen and households with over 10 million yen of income, accounted for 20% of total households. The modal value of household income in 2010 drastically went down to annual income between three and a half million to four million yen, and the ratio of households over 10 million yen to total household decreased to 11.8%. Besides, the ratio of household earning annual income less than three and a half million yen, the maximum tax exemption limit for a standard family consisting of a couple with two children had increased almost 5% points from 13.2% in 1995 to 18.4% in 2010. The household distribution for “all households” has also been shifting into a lower income group as well as “workers households”. The decreasing trend of average income level has been accompanying the rise in income inequality. Whereas there are various forms of index of income inequality (see Chapter 2 of Aoki, 1979), income difference between income of the top 10% of income (decile X) and the bottom 10% (decile I) is used as the index here. Trend of income difference defined as above is also added in Fig. 11.1. Inequality had increased in the early 1980s, decreased from the middle 1980s to the middle 1990s, and increased again until now.4 However, it should be emphasized that recent increasing inequality goes along with decreasing of average income level, while increasing income inequality in the early 1980s proceeded together with the increasing average income. Education expenditure in FIES includes the following three types of expenses related directly to education: 1) tuition fees, 2) purchase of textbooks, educational materials and other technical equipment for study, and 3) schooling fees of preparatory school, called yobiko or juku in Japanese, designed to help students pass exams of prestigious universities or schools. FIES also reports education related expenditure defined as the sum of education expenditure mentioned above and other indirect 4 The

reason of drastic increase of inequality in 1991 came from the rapid increase of income in the highest income group.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch11

F. Makino

210 (yen/month)

(%) 8

25,000

20,000 6

15,000

4 10,000

5,000 1970 72

2 74

76

78

80

82

84

86

88

Real education costs

90

92

94

96

98 2000

2

4

6

8

10

“Angel coefficient (right scale)”

Fig. 11.2. Real Educational Expenditure and Angel Coefficient. Note: Angel coefficient is defined as a ratio of education expenditure to total consumer expenditure. Sources: FIES and Statistical Yearbook of Consumer Price Index.

education expenditure on school uniform, school meals, school transport, and remittance to school going children who do not live with their parents. Figure 11.2 shows the long-term trend of education expenditure at constant prices (deflated by consumer price index) of workers’ households since 1970 until present. Whereas education expenditure went up and down from the middle 1970s to recent years, Angel coefficient defined as proportion of education expenditure to total consumption expenditure has been rising for 40 years since 1970. 11.2.2. Education Expenditure by Income Class Figure 11.3 shows Angel coefficients by income decile; Angel coefficient 1 is defined as the rate of education expenditure to total consumption expenditure and Angel coefficient 2 as the rate of education related expenditure to total consumption expenditure. The two Angel coefficients rise as the income level of household increases. This means that education expenditure has income elasticity that is greater than one, and is therefore assumed to be luxury goods. It should be noted that the higher the income level, the

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch11

Growth and Inequality in Japan

211

(%)

(%)

18

13

16

12 11

14

10 12 9 10 8 8 7 6 6 4

5

2

4

0

3 I

II

III

VI

V

VI

VII

VIII

“Enrollment ratio in higher education institutions”

“Enrolment ritio in university”

“Angel coefficient 1 (right scale)”

“Angel coefficient 2 (right scale)”

IX

X

Fig. 11.3. Angel Coefficient by Income Decile and the Enrollment Rate (Average 2008–2010). Notes: 1. The enrollment ratio in higher education institutions and the enrollment ratio in university is defined as the number of students in higher education institutions (university, colleges of technology, junior college) and university per 100 households, respectively. 2. Average rate from the year 2008–2010. Source: FIES.

wider difference between two Angel coefficients. This is because the amount of indirect educational expenditure, especially remittance to children living apart, is directly related to income level. International comparison of the relationship between Angel coefficients and income quintile among Japan, U.S. and China is shown in Fig. 11.4. For the strict comparison of Angel coefficient, it will be required to adjust age of household head in each quintile since the amount of education expenditure is highly dependent on the age structure of family members. As for Japan and U.S.,5 there is little difference in average age of household head except for the lowest quintile and second quintile in which Japanese household head is a little bit younger. 5 No

information on age of household head is available for China.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch11

F. Makino

212

7

6

5

4

3

2

1

0 first quintile

second quintile

Japan

third quintile

U.S.

fourth quintile

fifth quintile

China

Fig. 11.4. Angel Coefficient by Quintile (International Comparison). Note: Definition of household differs among three countries; total workers’ household for Japan, urban households for China, total household for U.S. Sources: FIES 2010 for Japan, Department of Urban Society and Economic Statistics, National Bureau of Statistics of China, China Urban Life and Price Yearbook for China, Bureau of Labor Statistics, Department of Labor, Consumer Expenditure Survey for U.S. (http://www.bls.gov/cex/2009/Standard/quintile.xls).

It will be obvious that the pattern of three countries’ Angel coefficient by income level is much different. While the coefficient has positive relation with income level in Japan, its value declines as income level goes up in China where education expenditure is classified as essential service. However, tendency observed in U.S. household is different from that in Japan and China because j-shaped curve between Angel coefficients and income is exhibited in U.S. household. Anyway this comparison shows that education expenditure in Japan seems to be highly income elastic from an international viewpoint. As well as the amount of education expenditure, its composition differs by income levels in Japanese workers’ households. For example, while tuition fees account to 76.8% of total education expenditure, share of spending on textbooks and educational materials and schooling fees of preparatory school is 2.1% and 19.4%, respectively for the bottom 10% household of income distribution, share of each three type of education

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch11

Growth and Inequality in Japan

213

expenditure is 69.7%, 1.6% and 28.7% for the top 10% household. Especially it should be emphasized here that parents having high-income spend a higher portion of its education expenditure to schooling fees of preparatory school to help their children to succeed in entrance examination of popular and prestigious school. Gini coefficient (pseudo Gini, strictly speaking) of annual consumption expenditure among decile income class is calculated to investigate up to what extent education expenditure explains a degree of inequality of consumption expenditure by income level (Table 11.1). Although Gini coefficient for total consumption expenditure is as low as 0.147, pseudo Gini coefficient of major expenditure categories provides an interesting contrasts that both pseudo Gini coefficient of education and education related expenditure are much higher than that of other categories. This result is consistent with higher Angel coefficient of household having high-income as shown in Fig. 11.3. Moreover, pseudo Gini coefficient of schooling fees for preparatory school, 0.306, is the highest among expenditure categories as Table 11.1.

Inequality of Expenditure and Its Factors (2010). Gini (Quasi Gini) index

Composition of Expenditure

Contribution Ratio

0.147 0.100 −0.012 0.060 0.151

100.0 21.9 6.5 6.8 3.3

100.0 14.9 −0.5 2.8 3.4

0.216 0.124 0.129

4.3 3.6 15.1

6.3 3.0 13.2

0.247 0.201 0.218

5.7 10.7 22.1

9.6 14.7 32.8

Education related costs

0.302

9.1

18.7

Tuition Textbook and study books Schooling fees of preparatory school

0.230 0.196 0.306

4.2 0.1 1.3

6.6 0.2 2.8

Total consumer expenditure Food House Utilities and water Furniture and household shores’ supplies Clothing and footwear Healthcare Transportation and telecommunications Education Entertainment and culture Others

Note: Calculation for all workers’ households except for agriculture, forestry and fishery households. Source: FIES.

June 5, 2012

11:23

214

9in x 6in

Growth and Inequality: An International. . .

b1363-ch11

F. Makino

is shown in the lower row of Table 11.1. Relative contribution of inequality of an expenditure category to total consumption inequality is obtained by multiplying pseudo Gini coefficient by share of expenditure of the category in total consumption expenditure. Relative contribution of education related expenditure is 20% and the second largest among major expenditure categories, which means that amount of education related expenditure will give a much impact on a degree of inequality of total consumption expenditure. 11.2.3. Quantitative Analysis on Education Expenditure As was discussed in the previous section, it must be noted that unlike other expenditure the amount of education expenditure is much affected by the age structure of family members. First, if none of the family member goes to school in a household, the household will not spend on education. Second, the amount of education expenditure will be much higher in a household having child who goes to university than in a household where child goes to elementary school. Third, since Japanese wage system is seniority based, the age of a worker does have significant meaning on the determination of the wage of the worker. So, it will be difficult to detect whether more spending on education in household having high-income is caused by income level itself or by age of household head. Education expenditure function is estimated in order to overcome this difficulty by using statistical data compiled in FIES in which information is available about households’ attributes such as an average age of household head, educational level achieved by children living together with household head. Problem is that student children who do not live with their parents are excluded from households’ member by the definition in FIES. Education expenditure function is formulated as follows: EDEXPijk = f (DYij , Pjk , RHEij , OCUij , AGEij )

(11.1)

EDEXP ijk : Real educational expenditure per household or per student children.6 DY ij : Real disposable income per household or per student children. Pjk : Relative price (relative of education price index to general consumer price index). 6 Kindergarten

pupils and special training school students are also included.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Japan

b1363-ch11

215

RHE ij : Ratio of the number of children studying in higher education institution to total number of student children. OCU ij : Ratio of the number of household head having white-collar jobs to total household head. AGE ij : Average age of household head. Suffixes i, j, k stand for income deciles (i = 1−10), year (j = 2000−2006), type of education expenditure (k = 1, 2; 1 = direct education spending, 2 = education-related spending), respectively. Dependent variable EDEXP and independent variable DY are both adjusted by general consumer price index. First, ∂EDEXP /∂DY is expected to be positive sign as education expenditure increases with income level. Second, ∂EDEXP /∂P is assumed to be negative because household will cut down education spending if relative education price goes up. Third, variable RHE is added in the model to reflect the difference of schooling levels of children among household. So, ∂EDEXP /∂RHE is expected to be positive since the tuition fees will rise as schooling level goes up. OCU represents the difference of social class that household head belongs to. White-collar jobs in this model include any type of non-manual permanent occupations. Since white-collar workers have higher school background and will pay more attention to educational achievement of their children than blue-collar workers do. So, ∂EDEXP/∂OCU is expected to be positive. AGE stands for an average age of household head. This variable is added because education spending in a household is to increase until household head reaches a certain age, when it will decrease after the end of children’s schooling. Unlike other variables, the inter-relationship between age of household head and amount of education spending is not linear. The problem is that a correlation between four independent variables except relative price (P ) is very high. Multi-collinearity will surely occur, if parameters are estimated by using all the independent variables in formula (1). Whereas the variable, which has the largest Variance Inflation Factor (VIF) value among independent variables in formula (1) is a real disposable income (DY ), it is impossible to exclude it from this model. The second best choice to avoid multi-collinearity is to exclude a variable of OCU having the second largest VIF value. Therefore, the model (1) is reformulated to as follows, EDEXPijk = f (DYij , Pjk , RHEij , AGEij )

(11.2)

June 5, 2012

11:23

216

9in x 6in

Growth and Inequality: An International. . .

F. Makino

Combining data for income decile groups and for seven years (from 2000 to 2006) together and performing a logit transformation on variable RHE ranging from 0 to 1, pseudo panel estimation and OLS estimation is carried out. Estimation results are shown in lines (A) and (B) of Table 11.2. First, estimation result of model A in which education expenditure per household is used as a dependent variable. OLS estimation result is provided in Table 11.2 since F-test rejected panel estimation (estimated F-value 1.66 is less than critical F-value 4.44). The signs of the estimated parameters that connect the dependent variable to the explanatory variables are consistent with the predictions mentioned above except for relative price. Estimated value for disposable income (income elasticity) is 1.513, which confirms that education expenditure is luxury goods after adjusting influences of other factors on it. It is found that there is an inverse U curve relationship between education expenditure and age of household head. On other conditions being equal, it will be found easily that the amount of educational expenditure is maximized when household head is 46.5 year old. Estimated parameter of relative price is not statistically significant except for model D. This is because there are no goods or service substitute for education. Panel estimator is preferred to OLS estimator in model C, because estimated F-value is more than critical F-value and Houseman test is used to find which model (fixed-effect model or random-effect model) is to be selected in panel estimation. Random-effect model is adopted because Chi square value that a null hypothesis (random-effect model equals fixed-effect model rejected at a 5% level). The estimated income elasticity in model C is slightly bigger than model A, which suggests the possibility that education related expenditure including indirect costs is more income elastic than direct education expenditure.7 Parameter of average age of household head is also significant in model C, which shows 47.9 years of age (during which education related expenditure maximize) which is a little higher than the counterpart calculated from parameters estimated in model A. The estimated result in model B which used education spending per student and disposable income per household, is same as that in model A. The result of model D that selected education related spending per student as a dependent variable is a little different from those in other models 7 While

income elasticity decreases in model A, it increases in model C, if household size (number of household member) is added to independent variable.

b1363-ch11

June 5, 2012 11:23

Table 11.2.

Constant

B: Per Student

−49.229 (11.45)∗∗ 1.513 (20.34)∗∗ 0.324 (0.64) 0.156 (3.22)∗∗

−6.958 (1.82) 1.017 (12.01)∗∗ −0.401 (0.95) 0.181 (4.46)∗∗

1.674 (10.02)∗∗

0.307 (2.19)∗

C: Per Household

D: Per Student

Enrollment Ratio E: Higher Education Institution

F: University

−29.955 (4.37)∗∗ 1.594 (13.27)∗∗ −0.349 (0.60) 0.023 (0.48)

−3.251 (1.91) 0.942 (9.43)∗∗ −2.058 (3.60)∗∗ 0.106 (2.20)∗

−63.207 (6.24)∗∗ 1.387 (6.06)∗∗

−66.390 (5.96)∗∗ 1.420 (5.64)∗∗

0.885 (3.22)∗∗

3.188 (4.63)∗∗

1.951 (4.79)∗∗

2.045 (4.56)∗∗ (Continued)

Growth and Inequality: An International. . .

Disposable income (DY ) Relative price (P) Rate of students in higher education institutions (RHE ) Average age of household head (AGE )

A: Per Household

Education Related Expenditure

Growth and Inequality in Japan

Independent Variables

Education Expenditure

9in x 6in

Depenent Variable

Estimating Result of Educational Expenditure and Enrollment Ratio Function.

217 b1363-ch11

June 5, 2012

Independent Variables

F-Test, Chi Square Test

−0.018 (10.24)∗∗ 0.975

B: Per Student −3.256∗10−3 (2.17)∗ 0.971

C: Per Household

D: Per Student

−9.236∗10−3 (3.11)∗∗ 0.972

Enrollment Ratio E: Higher Education Institution

F: University

−0.019 (4.44)∗∗

−0.020 (4.02)∗∗

0.917

0.913

0.973

F = 1.66 < 4.44# F = 1.66 < 4.44# F = 5.17 > 4.44# , F = 3.86 < 4.52# F = 2.86 < 4.60# F = 2.14 < 4.60# χ2 = 7.36

Notes: 1. Calculation for all workers’ households except for agriculture, forestry and fishery households by income decile groups for the year 2000–2006. 2. Variable RHE is converted by logit transformation, original value of other variables except AGE are converted to natural log values. 3. Model C is random effect model and other models are OLS estimation model. 4. AdjR2 is a coefficient of determination conditioned for a degree of freedom, ( ) is t value, and 1% and 5%, respectively, # stand for a critical value of F value. Source: FIES.

∗∗∗ , ∗∗

stands for significance levels of

F. Makino

AdjR2

A: Per Household

Education Related Expenditure

Growth and Inequality: An International. . .

Square of average age of household head (AGE 2 )

Education Expenditure

(Continued)

9in x 6in

Depenent Variable

11:23

218

Table 11.2.

b1363-ch11

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Japan

b1363-ch11

219

because income elasticity is less than one and education expenditure is directly relative to age of household head. Factors of education expenditure inequality between income levels can be obtained by using parameters estimated in model A. A simple model
yˆ = a0 + bi xi is build for that purpose, yˆ stands for hypothetical education expenditure calculated by the estimated parameters, xi is independent variable, and a0 and bi represent constant term and estimators of independent variables estimated in model A, respectively. Taking an average of ith independent variable for high-income group as xih , and that for low-income group as xil . Assuming yˆh and yˆl , as hypothetical education expenditure for high-income group and low-income group respectively, which are calculated from parameters and xih and xil . Contribution ratio of difference of ith independent variable between two income groups to the difference of yh − yˆl ). dependent variable is defined as bi (xih − xil )/(ˆ If bottom 20% of the income distribution is assumed to be low-income group, and top 20% of household to be high-income group, respectively, 103% of differences of the two group’s education related expenditure come from the difference of disposable income in model A, and the corresponding figure is 73% in model B, 93% in model C and 56% in model D. This result means that income inequality is the largest factor of education expenditure inequality among independent variables in any model used in this analysis.

11.3. Income Inequality and Education Effects 11.3.1. Evidence from FIES It could be confirmed in Sec. 2 that income difference is the largest factor in determining differences in the education expenditure; the other factors being composition of school level the children go to and the difference in the age of the head of household that were included in the analysis in the previous sections. Then, is educational attainment or schooling level of children determined by household income levels? This problem is considered in this section by using the number of household with children enrolled in the institutions of higher education. First, relationship between household income level and the number of students enrolled in institutions of higher education per 100 households is determined (hereinafter called the enrollment ratio in higher

June 5, 2012

11:23

220

9in x 6in

Growth and Inequality: An International. . .

F. Makino

education institutions). 8 Whereas the enrollment ratio in higher education institutions and the enrollment rate in university are just 3.13 and 2.58 respectively for the bottom 10% of household, their counterpart for the top 10% are 15.98 and 14.84, respectively. Simple model is set to examine whether the enrollment ratio in higher education institutions or in university is determined by income level. EHEij = g (DYij , AGEij ) EHE ij stands for the enrollment rate in higher education institutions or university in ith income level injth year and DY ij for real disposable income per household and AGE ij for average age of household head. As is in previous section, combining data for income decile groups and seven years (from 2000 to 2006) together, pseudo panel estimation and OLS estimation is carried out to obtain parameters. Since panel estimation is not accepted by F-test, OLS estimation result is exhibited in models E and F of Table 11.2. All parameters estimated are statistically significant and positive relationship between income level and the enrollment rate in higher education institutions and university is surely observed. Comparing parameter of disposable income between model E and F, it is found that model F is higher than that of model E. This means the enrollment in university is more income elastic than advancement to junior college or other higher institutions. In short, household income level is more important for household to allow children to advance to university than to junior college or other higher institutions. As is in the Sec. 2, the inverse U curve relationship between the enrollment ratio in higher education institutions and age of household head is also observed in models E and F, and average age at which the enrollment ratio is the highest is 50.3 year old for the enrollment in higher education institutions and 51.1 year old for that in university. Applying same method in the preceding section, contribution ratio of difference of income inequality to the difference in the enrollment rate in university is 57% between low-income group and high-income group. Conclusions obtained from previous paragraphs are also confirmed in FIES. Comparing household distribution by annual income level between household supported by household head of age of 45 year old to 54 year old, and household with university student children, the distribution of the 8 As

was stated before, student children living apart from their parents are not included.

b1363-ch11

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Growth and Inequality in Japan

b1363-ch11

221

18 16 14 12 10 8 6 4 2 0 −

−

−

Households with a student

−

−

−

−

−

−

Households in which householders' age from 45 year old to 54 year old Annual income (10,000 yen)

Fig. 11.5. Distribution by Annual Income Class (2006). Notes: Students who do not live with parents are not included. Source: FIES, 2006 edition. (http://www.stat.go.jp/data/kakei/2006nn/zuhyou/a506. xls).

latter is apparently more skewed to the right (higher income) than the distribution of the former is (Fig. 11.5). 11.3.2. Case Study on Tokyo University The relationship between education inequality and socio-economic inequality can be seen in Tokyo University, the most prestigious university in Japan. Tokyo University has conducted regular investigation into the annual income or occupation of students’ parents. According to the investigation results reported on the University web, the distribution by annual income of main supporters of students is more skewed to higher income than that by household head aged from 45 years to 54 years whose children are generally about 20 year old (Fig. 11.6). The occupational distribution of the main family budget supporters of Tokyo University’s students is unique and quite different from that of

June 5, 2012

11:23

222

9in x 6in

Growth and Inequality: An International. . .

F. Makino

50

40

30

20

10

0

Main family budget supporters of students in Tokyo University

General family's one

Annual income (10,000 yen)

Fig. 11.6. Distribution of Annual Income of Main Family Budget Supporters in Tokyo University and General Family’s One (2005). Sources: FIES (2005 edition), Tokyo University’s Committee on Public Information “2005 (55th) Investigation into Student Life” (http://www.u-tokyo.ac.jp/gen03/ kouhou/1348/6.html).

general adult male. The occupation of the mail supporters of Tokyo University’s students extremely concentrates on white-collar jobs, especially professional and technical workers and occupational distribution is much different from that of general adult male (Table 11.3). It is easy to understand that students of Tokyo University come from household with quite different social and economic background from household in general.

11.3.3. Case Study on Compulsory Education in Tokyo Is there any relationship between academic ability and income level? If any, at what schooling stage it comes out? Tokyo Metropolitan Board of Education has executed achievement test in every January to all 5th grade pupils of public elementary school and 2nd grade students of public junior high school, and has opened the result (available on the web of Tokyo Metropolitan Board of Education). Figures 11.7 and 11.8 show the positive relationship between average scores and income level of 49 administrative

b1363-ch11

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch11

Growth and Inequality in Japan

223

Table 11.3. Comparison of Occupational Structure Between Parents of Students in Tokyo University and Adult Male in General (2005). (%)

Professional and technical workers Managers and officials Clerical and related workers Sales workers Agricultural, forestry and fishery workers Production process workers and laborers Workers in transport and communications occupations, service workers, protective service workers Unclassified

Students Parents of Tokyo University

General Adult Male

40.9 28.3 10.0 5.3 1.1 1.9 10.5

14.5 5.6 15.8 14.9 3.2 31.6 13.2

2.1

1.3

Note: General adult male aged from 45 years to 54 years. Source: General adult male: Census of Population 2005 ” (http://www.stat.go.jp/data/ kokusei/2005/sokuhou/zuhyou/a009-1.xls) Tokyo University: Same as Fig. 11.6.

districts of Tokyo for elementary school pupils and junior high school students. Figures 11.7 and 11.8 show that as average income level increases, the average achievement test scores rises except for three richest wards (Minato ward, Chiyoda ward, and Shibuya ward). It cannot be denied that close inter-relation between income level and the academic ability has already come out in late elementary grade. As discussed in the previous section, is there any inter-relation between parents’ occupations and achievement test score of children in early schooling stages? Model, Sij = a0 + b1 × ln(Ri ) + b2 × W Ci , is used to verify hypothesis that income level and occupational structure affect achievement test score.9 In this model Sij stands for average score of jth subject of the student in ith administrative unit (ward or city), Ri for taxable income per person, and WC i for ratio of white collar workers to total workers. Parameters are depicted in Table 11.4. As the coefficient of determination value is 0.5 to 0.7, it will not be erroneous to say that average scores in a district can be explained by its average income level and occupational structure. Coefficient of determination value by subjects gives a fact that influences affected by income and occupation is larger in humanity and social subjects than in mathematics and natural 9 See

Kariya (2003) for high school students.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch11

F. Makino

224

80

70

1

5 33

2

10

26 15

60

38 27 45 31 49 6 16 34 43 19 46 20 14 17 35 3240 8 36 11 7 29 22 42 24 23 21 47 30 25 48 41 28

12

18

50

40

30

3

13

37 4 9

39

44

20 1,000

2,000

3,000

4,000

5,000

6,000

Income per person (1,000 yen)

Fig. 11.7. Inter-relation Between a Result of Achievement Test in Public Junior High School in Tokyo and Income Level. Notes: 1. The number stands for administrative unit. 2. Average score is for four subjects of Japanese language, math, social studies and science for elementary school and five subjects (English added) for junior high school. 3. Average income is equal to taxable income per person. Sources: Result of achievement test: Tokyo Metropolitan Board of Education (2005, 2006). Taxable income: JPS (2006), p.74. 70 26 33 37

5

4

60 38 35 45 40 27 34 49 20 32 29 24 31 43 16 17 6 46 36 14 8 47 19 48 11 42 7 30 1823 28 25 21 22

50

40

1

10 15

13

12

3

2 9

41

30 44

39

20 1,000

2,000

3,000

4,000

5,000

6,000

Income per person (1,000 yen)

Fig. 11.8. Inter-relation Between the Result of Achievement Test in Public Junior High School in Tokyo and Income Level (5 Subjects).

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch11

Growth and Inequality in Japan Table 11.4.

Total Japanese Mathematics Social Studies Natural Science English

225

Estimation Result of Achievement Test Score Function in Tokyo. 5th Grade in Elementary School

2nd Grade in Junior High School

Income level (b1)

Ratio of WC (b2)

AdjR2

Income level (b1)

Ratio of WC (b2)

AdjR2

23.297∗∗ 22.172∗∗ 24.585∗∗ 19.547∗∗ 22.751∗∗ —

0.666∗∗ 0.789∗∗ 0.444 0.846∗∗ 0.572∗ —

0.688 0.708 0.644 0.627 0.645 —

13.191∗∗ 17.065∗∗ 13.397∗∗ 8.870∗∗ 4.266 16.416∗∗

1.133∗∗ 0.693∗ 1.082∗∗ 1.324∗ 1.288∗∗ 0.999∗∗

0.535 0.516 0.515 0.529 0.374 0.583

Notes: 1.

∗∗ , ∗

stands for significance levels of 1% and 5%, respectively.

2. 49 samples (number of wards and cities). Source: Test score and income per person are same as in Fig. 11.7. Number of workers by occupation: Population Census 2000 (http://www.e-stat.go.jp/SG1/estat/GL08020103. do? xlsDownload &fileId=000001586319&releaseCount=1).

sciences. According to parameters estimated, the influences of income level on test scores is higher in elementary school than in junior high school, and those of occupation on scores is higher in junior high school. In other words, the results of achievement in an elementary school is more sensitive to income level of parents than in junior high school, in contrast, the result of achievement test in junior high school is more sensitive to occupation of parents. It must be noted that these results are obtained not by micro-data but by aggregated data. We have to be careful to interpret facts found in this section because many factors other than income and occupations to affect academic ability are neglected here. However, it will be true that income level or occupation structure affects the average academic ability on local basis.

11.4. Conclusion This chapter analyzed the long-term trend of education expenditure and Angel coefficient including its comparison with some other countries. Interrelation between education expenditure, income inequality, enrollment in university, and academic ability was also examined by applying simple econometric models to understand the inter-relationship. The important

June 5, 2012

11:23

226

9in x 6in

Growth and Inequality: An International. . .

F. Makino

conclusions obtained here are as follows; 1. Angel coefficient (ratio of educational expenditure to total consumer expenditure) has been increasing for 30 years from the beginning of 1970 until now. 2. Angel coefficient rises as income level increases in workers household, and educational expenditure is the largest category in consumption expense that creates inequality among various income levels of households. 3. According to the result of achievement test implemented by Tokyo Metropolitan Government, it will be clear that academic ability in early school stage is much affected by income level and occupation of parents. Income level gives quite an impact to academic ability or advancement to university. The fact that more investment in human capital will be possible and done in high-income household than in poor household in Japan, gives unequal opportunity of enrollment or job to children. Children grown in rich household can have higher probability to enter prestigious university and get better job than those in poor household. Inequality of income and education reproduces and perpetuate themselves. How to reduce income and education inequality? More government support in education, lower level in Japan than in other OECD countries, is required to prevent inequality from further increasing. References Aoki, M. (1979). “Distribution Theory (Bumpai Riron).” Tokyo: Chikuma Shobo. Japan Student Services Organization (2006). “2004 Report of Survey of Student Life’ College and Student.” 31st issue, July. JPS, (2006). Personal Income Index (2007ed.), JPS. Kariya, T. (2001). “Stratifying Japan and Education Crisis (Kaisoka Nippon to Kyoiku Kiki).” Tokyo: Yushindokobunsha. Kariya, T. (2003). “Is Difference in Education between Social Class Widening in Japan? (Kyoiku niokeru Kaiso Kakusa wa Kakudai shiteiruka).” in Higuchi, Y. et al. (eds.) Income Inequality and Social Stratification in Japan (Nippon no Shotoku Kakusa to Shakai Kaiso). Tokyo: Nihon Hyoron Sha. Kikuchi, J. (ed.) (1990). “Education and Social Mobility, Vol. 3 of Structure of Social Stratification in Japan.” (Gendai Nippon no Kaiso Kozo, 3 kan, Kyoiku to Shakai Ido). Tokyo: Tokyo University Press. Kikuchi, J. (2003). “Education Opportunity and Social Stratification in Modern Japan (Kindai Nipponn no Kyoiku Kikai to Shakai Kaiso).” Tokyo: Tokyo University Press.

b1363-ch11

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Growth and Inequality in Japan

b1363-ch11

227

Minami, R., Makino, F. and Luo, H. (2008). “Education and Economic Development in China (Chugoku no Kyoiku to Keizai Hatten).” Tokyo: Toyo Keizai Shimpo Sha. Tachibanaki, T. (2004). “Japanese Economy in Household Point of View (Kakei kara miru Nippon Keizai).” Tokyo: Iwanami Shoten. Tokyo Metropolitan Board of Education, (2005). “2004 Research Report for Improving Student’s Academic Abilities .” June, (http://www.kyoiku. metro.tokyo.jp/press/pr050609s/pr050609s 2.htm). Tokyo Metropolitan Board of Education, (2006). “2005 Research Report for Improving Student’s Academic Abilities.” June (http://www.metro.tokyo. jp/INET/CHOUSA/2006/06/60g69100.htm). Tokyo Metropolitan Board of Education, (2007). “2006 Research Report for Improving Student’s Academic Abilities.” June (http://www.metro.tokyo. jp/INET/CHOUSA/2007/06/60h6e100.htm). Yoshikawa, T. (2009). “Society Divided by Economic Background (Gakureki Bundan Shakai).” Tokyo: Chikuma Shobo.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch11

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Part 2

Factor and Policy Analysis on Income Inequality

229

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

230

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 12 INFORMAL EMPLOYMENT AND INCOME DISPARITY Yang Du∗ and Jinjun Xue†

12.1. Introduction Labor market reform is one of the essential components of China’s Reform and opening-up strategy. With gradual introduction from the market mechanism to labor allocation in both rural and urban areas, the labor market has been functioning better and better than before. The labor market development is evidenced by some stylized facts. The labor markets are more integrated across regions, between rural and urban areas, and across economic sectors with various ownerships. The returns to human capital keep increasing, which implies that productive workers get better pay. The employers are more responsive to the factor prices, including wages. Institutionally, the Chinese labor market is less discriminatory than before, which is evidenced by many previous barriers that limited labor mobility and employment increase are demolished. Meanwhile, with the changes in the labor market, the income inequality has been an issue drawing more and more attention. Although how serious the inequality is remains as a question, one has to believe that

∗ Professor at Institute of Population and Labor Economics, Chinese Academy of Social Sciences. † Professor at Graduate School of Economics, Nagoya University. The authors thank Economic Research Center at Nagoya University for supporting this research and Du Yang’s visit at Nagoya University. The authors acknowledge funding support for the China Urban Labor Survey by Chinese Academy of Social Sciences, the World Bank, Michigan State University, the Ford Foundation, and the University of Michigan. Discussions with Cai Fang and Wang Meiyan and their contributions to this chapter are grateful.

231

b1363-ch12

June 5, 2012

11:23

232

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

many aspects of labor market transition are connected with the inequality changes. For such a reason, it will be helpful to understand the inequality and its determinants in contemporary China if we can disentangle those factors of labor market development and find out their directions and magnitudes to inequality changes. The recent labor market development is characterized by the following aspects: An increasing flow of rural to urban migration, informalization, and stricter regulations. Each of these changes may have effect on inequality. In this chapter, we are going to discuss these changes and then try to see their impacts on the inequality by taking advantage of household survey data. The chapter is organized as follows. In Sec. 2, we describe the main changes in the Chinese labor market recently. In Sec. 3, we try to link those changes with inequality. Section 4 provides some empirical evidences by taking advantage of household survey data. We draw some conclusions in Sec. 5.

12.2. Labor Market Development in China China has also witnessed labor market development in the past three decades. As an essential component of economic transition, China evolved to a labor market through gradual reforms, like in other areas. Labor market mechanisms were firstly introduced in rural China by allowing farmers to make labor allocation decisions, which in turn lead to labor mobility from rural to urban areas and across regions. When China started restructuring the economy of SOEs and experienced some years of labor market dislocation, the government tends to relax their direct control on labor allocation in urban economy. Instead, the market mechanism was recognized in hiring and firing decision and wage formation, as evidenced by the 1994 Labor Law. Most migrant workers and the urban unemployed from SOEs entered informal sectors in urban labor market, which leads to a trend of informalization while, symbolized by the 2007 Labor Code, a series of labor market regulations are going to achieve a more regulated labor market.

12.2.1. Rural to Urban Migration In the early reform period — namely from early 1980s through mid-1990s, the employment of rural and urban China expanded mainly through the transformation of farmers from agricultural to non-agricultural work. Job

b1363-ch12

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

233

creation by township and village enterprises (TVEs)1 and massive labor migration from rural to urban sectors are most impressive, unique and worldwide-recognized “China miracle”. In the urban area, employment allocation system started its reform — that is, an increasing part of new entrants into the labor market were not taken care of by the government labor planning, which laid a starting point for labor market development later on. In 1980s, there was only a small amount of labor migration. The composition of migration flow was dominated by craftsman who moved within rural areas. With increasing labor productivity in agriculture, rural labor forces began to move out of rural areas in growing numbers. According to estimations by MOA (2001), rural migrants were only two million in 1983 but reached 30 million by the end of 1980s. The economic booming after 1992 encouraged migration further. Fast economic growth in coastal areas attracted more and more rural labor forces from other parts of China to seek off-farm job vacancies. In the new century, National Bureau Statistics (NBS) started collecting the information of migrants in rural household survey, so one can get a continuous series of size of migration based on consistent sampling surveys. As Table 12.1 displays, the total number of migrants has kept growing Table 12.1.

Migrant Workers and Urban Employment.

Year

Migrant Workers (million)

Urban Employment (million)

2000 2001 2002 2003 2004 2005 2006 2007

78.49 83.99 104.7 113.9 118.23 125.78 132.12 136.49

212.74 239.4 247.8 256.39 264.76 273.31 283.1 293.5

Ratio (1/2, %) 36.9 35.1 42.3 44.4 44.7 46.0 46.7 46.5

Sources: The size of the migrant workforce from National Bureau Statistics (NBS), various years (a) Yearbook of Rural Household Survey, China Statistical Press. Data on urban employment are from NBS, various years (b). Yearbook of Labor Statistics in China (various years), China Statistical Press. 1 TVEs were industries owned by townships and villages. At the early stage of the reform and they were the main forces driving the rural industrialization in 1980s and 1990s. In 2007, the employment in TVEs reached 150.9 million.

June 5, 2012

11:23

234

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

and reached 136 million in 2007, suggesting that China is experiencing an unprecedented size of migration flow in history. It is obvious that migrant workers have had a substantial role in urban labor market. In 2007, migrant workers accounted for 46.5% of total urban employment. As the result of demographic transition, declining participation rates, and fast economic growth, China is facing a turning point in the labor market, suggesting the end of the era of unlimited labor supply. In coastal areas, the most developed regions in China, shortages of both skilled and unskilled workers have been widely reported in recent years. An indication of labor shortages is the rise of average wages. After being constant for decades, average wages for migration workers started rising up in 2003. According to surveys on migration workers, in 2006 the wages of migration workers increased more than 10% than the previous year (Cai, 2007). The survey conducted in 2,749 villages of rural China indicated that three out of four villages exhausted their young human resources (Cai, 2007). Demographic data in rural areas also confirm this trend. Looking at the age profile of rural migrants, it is easy to find that only a very limited number of those below 30 years old, work in agriculture. Considering that agriculture labor productivity is low in China due to the land tenure,2 agriculture still requires a large amount of labor input. In addition, older workers in rural areas have relatively lower years of schooling and are more difficult to reallocate in non-farm sectors than the youngest generations. However, the argument of a turning point is quite controversial as both academics and the public opinion find hard to believe the existence of a labor shortage based on the large amount of labor stock in China. To further defend this argument a deeper analysis on rural population is needed. When calculating the possible migration flows from agriculture, previous studies have often used aggregated data and predicted the number of migrants by deducting estimated labor for agriculture. One of the vital drawbacks of the estimation is to get the surplus number without considering the heterogeneity among individuals. In fact, given the disparities in terms of human capital, age, experience, household composition, and the local conditions in sending places, the propensity of migration varies from person to person.

2 The

rural land is collectively owned and the farmers have the rights to cultivate, which leads to the average size of each farm been very small.

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

235

The advantage of individual level data from a national representative sample survey, is to capture individual disparities by predicting each individual’s migration probability. For that purpose, a probit model has been used where the dependent variable is whether migrating more than six months and the regressive variables include education, health status, gender, experience and its square term, and dummies for provinces. Based on the predicted probability, we can get the average probability of migration for each group categorized by age or education. As Fig. 12.1 displays, the probability of migration varies among different education groups, declining with age increases for each group. It is easy to find out that the migration probability for people with low education and more than 40 years of age is particularly low.

Fig. 12.1.

Predicted Probabilities by Age with Different Level of Human Capital.

Source: Authors’ calculation from 1% population sampling survey in 2005.

As Table 12.2 shows, those remaining in agriculture and having a low probability to work off-farms are the oldest with lowest human capital. The predicted number of labor available for non-agricultural industries sums up 43.5 million workers (last column). Because rural-to-urban migration was initiated by large scale surplus labor in rural areas, it generated two effects of transition and development. The first is a resources reallocation effect — namely, the transformation of workforce from low productivity (agricultural) sector to high

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

236 Table 12.2.

Rural Labor Forces and Migration Probability.

Age and Education Group 16–20 Primary School or below Jr. High School Sr. High School or above 21–30 Primary School or below Jr. High School Sr. High School or above 31–40 Primary School or below Jr. High School Sr. High School or above 41–50 Primary School or below Jr. High School Sr. High School or above 50 and above Primary School or below Jr. High School Sr. High School or above All

b1363-ch12

Number of Labor (million)

Predicted Probability

17.16 4.44 12.03 0.69 50.08 15.39 32.24 2.46 88.96 39.45 46.69 2.82 76.48 39.86 30.52 6.10 93.7 76.3 15.51 1.88 326.39

— 0.189 0.315 0.505 — 0.142 0.248 0.410 — 0.109 0.178 0.298 — 0.078 0.123 0.235 — 0.053 0.084 0.182 —

Predicted Migrants (million) 4.97 0.84 3.78 0.35 11.18 2.18 7.99 1.01 13.44 4.31 8.29 0.84 8.29 3.10 3.76 1.43 5.69 4.04 1.30 0.34 43.57

Source: Authors’ calculation from 1% population sampling survey in 2005.

productivity (secondary and tertiary) sectors alone contributed 21% to the overall gross domestic product (GDP) growth rate during the reform period (Cai and Wang, 1999). The second is an income effect — namely, while the wage rate of migrant workers had not increased much, the enlargement in total number of migrants has enhanced the total income of rural households as a whole. As a result, labor migration has been an effective way to both poverty reduction and narrowing the income gap between rural and urban areas. 12.2.2. Urban Economic Restructuring and Informalization The reform program termed “activating the system of permanent employment” initiated in 1987 touched upon the core system of the “iron rice bowl”, began revising the legacy of traditional labor policies under the planning system. The legal basis of this reform is Temporary Regulations on Labor Contract System of State-owned Enterprises issued by the

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

237

administration in 1986,3 with which all state-owned enterprises are required to recruit new workers based on voluntary contract with them. Under the newly introduced labor management, workers currently working at an enterprise were to be re-chosen and contracted based on their work performance and efforts. With this reform enterprise workers began to learn there is a risk to be unemployed due to overstaffing, shirking and misbehavior, though the state asked enterprises not to dismiss workers from enterprises at the time. China’s state-owned enterprise reform started in early 1980s and was characterized as to decentralize power and to give up partial profits to enterprises. Each step of the reform has implied more rights, with which SOEs make decision on labor employment — that is, as the state gradually grants autonomy to enterprises, managers of SOEs are legitimated to select and dismiss workers, and to determine and adjust compensation in accordance with enterprise’s profitability and worker’s performance. With this change of institutional environment and with increasing competition pressure on enterprises, employment has become more and more marketoriented and the “iron rice bowl” gradually breaks up. At the early stage of labor policy reforms, the more or less relaxation of labor regulations and granting of employment autonomy to enterprises were only motivated by the challenges to solve the problems caused by massive increase of urban labor force — returned sent-down youth and new graduates and to improve incentives of enterprise workers. However, since there existed soft-budget constraint in urban enterprises, their managers had no sufficient motivation of utilizing labor market as a distributor of resources. Only when SOEs began to face stronger outside competition in 1990s, the relaxed state regulations of labor allocation and autonomy of employment enterprises obtained have become stimuli of labor market development. First, as competition from non-state sectors sharpens and comparative advantage changes, numerous SOEs became lossmakers and were forced to lay off their redundant workers, and the state is unable to continue

3 In

the same year, the state issued other relevant documents such as temporary regulations on dismissal of lawbreaking worker in state-owned enterprises, Temporary regulations on state-owned enterprises workers, temporary regulations on state-owned enterprises recruitment of workers, temporary regulations on laid-off workers of state-owned enterprises, and so on. The issue of these documents signaled that the reforms of urban labor policies comprehensively started.

June 5, 2012

11:23

238

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

taking care of all those layoffs and labor market starts working. Secondly, the massive flows of rural labor to the cities, on the one hand, bring a shock to urban workers, since the latter are not capable of competing with their migrant counterparts, who have advantages with low pay and disciplinary works.4 Furthermore, the urban non-state sectors, by employing low cost migrant workers, put further pressure on SOEs. In sum, the increased competition has deepened the reform of labor policy and thus has pushed forward the labor market development. Due to the downturn of macro economy and rapid industrial structural change, many SOEs, which lost their comparative advantage and competitiveness, have been unable to fully utilize their production capacity since the late 1990s. As a response to this difficult situation, SOEs managers are forced to exercise their autonomy of disposing workers and thus a hundred of thousands urban workers have left their works in recent years. In China today, unemployment takes two forms — explicit unemployment and lay-off (xiagang). With xiagang, workers lost their work but were able to keep connection with their former employees and receive a certain amount of subsidies. Reemployment Center are established in all SOEs with the requirement of government and responsible to pay laid-off workers’ pension insurance and basic living allowance, which is shared by governments at central and local levels, enterprises and part of unemployment insurance funds. The severity of unemployment problem, as a double-edged weapon, has induced two policy intentions. First, urban governments strengthen their protection for urban workers. Given the responsibility of local governments for political stability, with a strong motivation of averting any potential social tension, urban governments have enacted various policy measures deterring labor market development. For one thing, urban governments intervene enterprises’ matters of employment adjustment and sometimes directly restrain enterprises from laying off workers regardless what situation an enterprise is in. To protect urban workers from competition of migrant workers, governments issue discriminatory employment regulations against migrants working in urban sectors by restricting jobs that migrant

4 Studies

show that there is only a small overlap between migrants’ and urban workers’ jobs because of the existing institutional segmentation, therefore migrants actually are not direct competitors of their counterparts in the cities (see for example (Solinger, 1999) and (Cai, 2000), Chapter 6).

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

239

workers can take up and imposing heavy charges for migrants entering the city (see Cai et al., 2001). Secondly, being aware of failure of planned allocation of labor force, the governments cannot but take advantage of labor market to solve the problems of employment and reemployment. As a result, small sized non-state enterprises and service sector, which used to be artificially depressed by government, are encouraged both politically and financially. This helps China’s employment structure to be diversified. In 1978 when the economic reform was about to begin, nearly 80% of urban laborers were employed in state sector, and state and collective sectors almost employed all urban workers. The two-sector domination of employment had remained until 1990s when non-state sector began to enhance its share of employment in the whole economy. Since then, things have changed dramatically — in 2001, employment shares of state and collective sectors dropped to 32% and 5.4%, respectively, while that of non-state sector increased substantially. In practice, urban employment has been always growing since the reform started and it reached 283 million in 2006, 13 million more than the previous year. As shown in Fig. 12.2, residual of employment represents 98.6 million urban employees in 2006, which is more than the sum of state and collective employment and accounts for 34.8% of urban total employment. Understanding the gap and its sources helps us know the trend of informal employment in urban labor market. Statistically, the residual between total and unit employments appeared in 1990. Prior to that very year, figures of urban employment were collected through all production units with independent accounts and registered individual enterprises. Currently, official statistics on employment come from two statistical systems. The gap between total and sectoral summation of employment comes from the multiformity of statistical sources and it is more diversified and allocated through new channels, mostly through market forces. Therefore, there are two main components of informal sector. One of them is the migrant workers from rural areas, and the other is those who lost their previous jobs in SOEs. Using the 1% population sampling survey data in 2005, which is national representative, we may observe the whole picture of informality for both migrant workers and local residents. We categorize three types of workers in urban labor market, local workers, rural migrant workers, and urban migrant workers. In contrast, most rural migrant workers (65.4%) work informally while the proportion for urban

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

240

Fig. 12.2.

Changes in Employment Structure Since the Reform.

Source: NBS, 2007. Note: SOU — State-owned Units, COL — Collective-owned Units, SRH — Shareholding Cooperative Units, JNT — Joint Ownership Units, LMT — Limited Liability Corporations, SLMT — Share Holding Corporations, Ltd., PVT — Private Enterprises, HMT — Units with Funds from Hong Kong, Macao and Taiwan, FDI — Foreign Funded Units, IND — Self-employed Individuals, RSD — Residual.

migrant workers is 29.8%. Table 12.3 presents the outcomes calculated from the data by various groups of people with different characteristics. As Table 12.3 indicates, for migrant workers, the proportion of workers in informal sector increases with aging, which implies that the older the migrant workers are, the more disadvantaged they are. The informalityage profile is different for local workers: With increasing age, the share of workers in informal sector decreases first and then increases. Education plays the same role for both migrant and local workers: The educated workers are less possible to work in informal sectors. As in Table 12.3, most of migrants in urban labor market worked in the informal sectors. Considering that migration workers have already accounted for a fairly large share of employment in urban labor market evidenced by Table 12.1, China needs to pay more attention to this group of people for decent work. When comparing with those workers with urban hukou, rural migration workers are more disadvantaged in urban labor market in terms of earnings, working intensity, and social protection.

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity Table 12.3.

b1363-ch12

241

Informality in Non-agricultural Sectors (%).

All By age group 20 and below 21–30 31–40 41–50 51–60 61 and above By education Primary school and below Junior high school Senior high school College and above By gender Male Female

Rural Migrant Workers

Local Workers

65.4

52.5

52.6

59.6 60.2 69.8 74.2 76.6 78.3

79.8 53.6 51.3 47.5 51.2 71.3

68.8 51.9 52.5 48.2 51.5 70.3

80.0 65.5 50.4 26.0

82.9 69.2 35.4 7.1

81.7 67.4 36.8 8.2

66.5 64.0

52.9 51.9

53.1 51.9

All Workers

Source: Authors’ calculation from 2005, 1% sampling survey.

12.2.3. Stricter Regulations Furthermore, the labor demand and supply situation in China has changed much during recent years. With the sustainable and rapid economic development and population age structure changes, China has ended the era of unlimited labor supply. Structural labor shortage have emerged in costal areas first and then spread to inland provinces. This has created good opportunities to protect lawful rights and interests of laborers. Under this circumstance, in order to protect lawful rights and interests of laborers better, a series of regulations and laws on China’s labor market have been issued since the end of the 1990s, especially since the new century, which include wage guideline system (1999), minimum wage regulations (2004), employment contract law (2008), employment promotion law (2008) and labor disputes mediation and arbitration law (2008). These regulations and laws can basically be seen as the more detailed and revised version of 1994 Labor Law. The contents of 1994 Labor Law is very comprehensive. However, the regulations of the law on many aspects are not that detailed. The newly issued regulations and laws since the end of 1990s have been much more detailed than 1994 Labor Law. Furthermore,

June 5, 2012

11:23

242

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

many new situations have emerged, and these new regulations and laws are mainly used to resolve these new issues.

12.3. Understanding the Inequality 12.3.1. The Impact of Migration It is widely discussed that a widening income inequality has taken place in China since the reform and opening up. Unfortunately, there are few empirical evidences that are based on national representative survey and reliable statistical methodology to describe the trend in the past three decades. As an exception, Ravallion and Chen (2007) suggest that, using Rural Household Survey (RHS) and Urban Household Survey (UHS) conducted by National Bureau of Statistics, the overall inequality had kept growing in the first two decades since the reform. Although the R.H.S and U.H.S are the best data to analyze the income inequality in China, it is worth noting that some change in the labor market that is not captured by the surveys might bias the estimation. The most pronounced factor is the rural to urban migration. As noted in Table 12.2, migrant workers accounted for 46.5% urban employment in 2007 and became indispensable component in the urban labor market. Considering that more and more migrant workers move to urban areas with their families, the actual size of floating population in urban China would be even bigger. R.H.S or U.H.S, are not well adjusted to this structural transition. National Bureau of Statistics modified the definition of residents as those living in a place more than six months in the year, which implies that most rural migrant workers and their families are defined as urban residents in population data. Migrants living out of countryside more than half year are not defined as rural residents anymore. In that case, surveys based on hukou registration system have bias in two ways. On one hand, long-term migrants who earn high income in rural areas are excluded in the R.H.S; on the other, they are not effectively included in the U.H.S, which may up-bias the urban residents’ income estimation on average. When the size of migration was relatively small, the sampling strategy would not bias the estimation on income inequality very much. For example, according to estimation by Ministry of Agriculture, there were about 30 million migrant workers by the end of 1980s, which is equivalent to 3.5% of total rural population at that time. However, the ratio of rural migrant

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

243

workers to rural population went up to 18.6%. Considering that the family migration is more widespread than in the 1990s, the ratio is just a lower bound. Missing income of long-term migrant households brings insufficient information and distortion on actual income in urban areas. Based on data provided by NBS (2006), in 2005, disposable per capita income of urban households was 10,493 yuan, whereas the net income per capita of rural households was only 3255 yuan. However, according to CULS survey, per capita income of rural-to-urban migrant households was 8,368 yuan, equivalent of 2.6 times per capita income of rural households and 80% of that of urban households. Although we can hardly claim a disappearance in rural urban income gaps, the huge magnitude of the migrant population undoubtedly serves to reduce the rural–urban income gap. Obviously, income inequality could be kind of exaggerated if the substantial middle income group is ignored. Under a dual economy, wage rates for unskilled workers, like migrant workers, persist at a subsistence level until the expanding modern sector exhausts the surplus labor. This has been the case in China till the beginning of this century. As a consequence of emerging labor shortage, the competition for labor inevitably lead to a rise of wages in the modern sector and in agriculture, and the relationship between wage rate and productivity in agriculture became close to what was expected. As we have already mentioned, in 2003, a shortage of migrant workers occurred in the Pearl River Delta region. Since then, the phenomena spread to the Yangtze River Delta region, and even to provinces in central China, from which migrant laborers are generally sent out. This trend has however been suspended by the current financial and international crisis. These labor shortages resulted in growing average wage for migrant workers after being constant for almost a decade. As Table 12.4 presents, the average wages for both migrants and local workers have grown in recent years, both in nominal and real terms. Attaining the Lewisian turning point will tend to equalize individual incomes. Before urban economic restructuring in the mid-1990s, the wage rates in formal sectors, as in SOEs, were set by institutional factors rather than by market forces. Meanwhile, in the informal sectors in which most migrant workers are employed, wage formation is determined by supply of and demand for labor. With unlimited labor supply from agricultural surplus, the wage rates for migrant workers had been kept constant for a long time even in nominal term. Although the incomes migrants earned in the urban labor market were still higher than in agriculture and contributed

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch12

Y. Du and J. J. Xue

244 Table 12.4.

Wage Increase in Urban Labor Market.

Migrants (NBS)

2001 2002 2003 2004 2005 2006 2007

Migrants (MOA)

Local Workers

Nominal

Real

Nominal

Real

Nominal

Real

644 659 702 780 861 946 1015a

644.0 665.7 702.8 755.9 821.3 889.0 912.8

— — 781 802 855 953 1,060

— — 774.0 776.4 841.5 938.9 1,014.4

903 1,031 1,164 1,327 1,517 1,738 2,078

896.7 1,041.4 1,153.6 1,284.6 1,493.1 1,712.3 1,988.5

Source: Urban local wages are from China statistical abstract in 2008, and migrants wages are from Statistical Report of NBS and Research Center of Rural Economy, MOA. Note: “a” is the average monthly earnings for the first three quarter in 2007.

1200 prim and below

Jr. High School

Sr. High School

above Sr HS

1116

1100

1094

1098

1015

1037

1000 878 900

924

877 888 794

800 728

756

853

787

728

700 687 600 2003

Fig. 12.3.

2004

2005

2006

Wage Trends of Migrant Workers by Education Attainment.

Source: Author’s calculation from RCRE data.

to poverty reduction in rural China, income inequality increased between migrant workers and those employed in formal sectors. Figure 12.3 displays migrant workers’ wage changes by education attainment in the past few years. We found that in 2003 the wage gap between skilled and unskilled workers was quite substantial and converged in the following years. Because labor shortage was more pronounced among lowskilled workers with junior high school, they had the most significant and stable wage increase with an annual growth rate of 9%. Similarly, workers

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

245

whose education attainments are primary schooling or below, the annual growth rate was 8.1%. Wage rising of low-skilled workers implies that labor shortage is not a transitory or structural phenomenon but caused imbalance between aggregated demand for and supply of labor. As a result, this trend is helpful to narrow the income gaps between migrant workers and those who worked in formal sectors. 12.3.2. Impacts of Informality During the period of serious labor market dislocation, labor market informalization helped reduce poverty by increasing the size of total employment. But, the role of informal employment on income generation has been different with changing situations of labor market. During the early stage of economic restructuring, there were three groups of labor market participants: The unemployed, informal workers, and formal workers. Compared to the unemployed, informal workers generated some income which had a positive effect on poverty reduction. When the unemployment rate goes down, the income distribution curve will shift to the right, as the left part of Fig. 12.4 shows, the effect of informal employment on income generation will not be as obvious as before. Of course, because of the relatively fixed (absolute) poverty line that is determined by subsistence expenditure, earnings from informal work are still helpful for poverty reduction although they are not good income generators any more (Cai et al., 2006). Considering that migrant workers accounted for a large share of informal employment, it is of importance to make use of urban labor market to provide income sources for those migrants. Working even in informal

Fig. 12.4.

Impacts of Informal Employment on Income and Poverty.

June 5, 2012

11:23

246

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

sector, migration workers enhance their productivity compared to in primary sector. According to the statistics from NBS (2007), wage income accounted for 38.3% of average net income in rural households in 2006. In some typical migrants sending provinces the ratio is even higher (45.6% in Chongqing and 42.8% in Hunan). Therefore working in the informal sector is an essential income generator for rural residents 12.3.3. The Impact of Social Policies In addition to active labor market programs initiated in the 1990s, China has been reforming the social benefit system after the labor market dislocation. In recent years, a host of social assistance programs are implemented in the urban labor market. Since those programs are directly targeted on low-income group with income transfer, it is good to believe that they are helpful to reduce income inequality. In urban areas, the on-going reforms on social benefit system have already made great achievement in terms of the social protection for vulnerable groups. At the very beginning, the policies attached closely with employment status were mainly used to target the employees suffering from the labor market shocks. When urban poverty emerged, the Chinese government initiated a social assistance program, Minimum Living Standard Allowance (dibao), to support the urban poor people. Gradually, the dibao program has become the main instrument against urban poverty. During the urban labor market dislocations, unemployment takes two forms — explicit unemployment and lay-off (xiagang). With xiagang, workers lost their work but were able to keep connection with their former employees and receive a certain amount of subsidies. Reemployment Centers5 that are established in all SOEs with the requirement of government are responsible to pay laid-off workers’ pension insurance and basic living allowance, which was shared by governments at central and local levels, enterprises and part of unemployment insurance funds. Therefore, the xiagang subsidy was the first form of social transfers to deal with the labor market dislocations. It is obvious that subsidizing the laid-offs was a temporary program since it only covers the workers who previously worked in SOEs and the employer could not fully fire the workers through xiagang program. The number of workers covered by xiagang program peaked at 2005. 5 Shanghai

was the first province to establish Reemployment Centers in 1996.

b1363-ch12

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch12

Informal Employment and Income Disparity

247

In 1999 the Chinese government enacted Regulations on Unemployment Insurance in order to transfer the xiagang subsidy to unemployment insurance. In this century, the government tends to protect the unemployed through unemployment issuance rather then xiagang subsidy. In 2004, there were 16 provinces removed the Reemployment Centers, which means that the laid-offs would get support from unemployment issuance. In 2005, the Ministry of Labor and Social Security (MOLSS) required that all the provinces would close the reemployment centers by the end of the year and cover the unemployed through unemployment insurance program. In 2006, the MOLSS claimed that the transformation from support workers from xiagang subsidy to unemployment issuance had been done (MOLSS, 2007) by 2006. Therefore, since the mid-1990s, unemployment insurance has become the second social assistance program. Unlike dibao and unemployment insurance that are employment related, dibao program directly target the poor, no matter the employment status. In 1993, Shanghai was the first city introducing the dibao program that supports the poor whose income are below official poverty line. Central government positively evaluated Shanghai’s experiment. In the next year, Ministry of Civil Affair proposed to extend Shanghai’s practice to the other urban areas of China. All cities and the towns where county government locate were required to set up the program since 1999. In 2003, the Ministry of Civil Affair claimed that in 2002 dibao program has covered all the 25 coverage

subsidy per capita 20.6

160 140

141

102

80

78

62

58

100

59

72

65

60

52 40 4.0

5

31

2.8

20 0

0 1999

2000

2001

Fig. 12.5.

2002

2003

2004

2005

2006

2007

2008

The Coverage and Transfer of dibao: 1999–2008.

Source: Ministry of Civil Affair, Statistic Report of Civil Affair, various years.

subsidy (yuan)

120

11.7 10

23.3

22.7

22.4

22.4

20

15

22.3

22.0

coverage (mln)

June 5, 2012

June 5, 2012

11:23

248

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

urban poor whose income are below local dibao line. As Fig. 12.5 shows, the number of the poor covered by dibao increased dramatically since 1999 and has been stable since 2002. In 2008, 22.33 million urban residents were covered by dibao program and the average per capita dibao transfer is 141 yuan. However, the dibao program has been implemented based on locality of hukou, which means that migrants have been excluded from the program despite of inclusion of them in urban population statistics.

12.4. The Changes of Labor Market on Inequality 12.4.1. Data Explanation This chapter employs unique data collected by Institute of Population and Labor Economics in five big cities, Shanghai, Wuhan, Shenyang, Fuzhou, and Xian. The two round surveys are referred to China Urban Labor Survey (CULS), which investigated both the sampled households and individuals. In particular, migrants were sampled in each round of survey. CULS1 was done in 2001. For urban household sample frame, proportional population sampling approach was used to sample an average of 15 households in each of 70 neighborhood clusters, by making use of 2000 census to sample clusters and households. On average 10 households were interviewed in each community, with additional five for spares. For migrant sample frame, 2000 Census was firstly used to sample 60 communities. Once a neighborhood was selected, the administrative records of the neighborhood committee were used to constructing a sample frame of all registered migrants in the neighborhood. In each city, 700 urban households and all the individuals in the households who are aged 16 and above were surveyed, and 600 individual migrants were surveyed. CULS2 was done in 2005. A similar sampling strategy was applied; the sample size was a bit different. In each of five cities surveyed in CULS2, 500 urban households and all individuals in the household were investigated. For migrants, not only were the sampled individual investigated but their families. This change increased the individual sample size for migrants. 12.4.2. Measure the Labor Market Changes and Inequality According to the household survey data, we look at the income inequality changes after transfer and over time respectively. The labor income inequality reflect the effects of labor market outcomes, including employment,

b1363-ch12

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Informal Employment and Income Disparity Table 12.5.

249

Income Inequality of Urban Households.

CULS1

p90/p10 GE(−1) GE(0) GE(1) GE(2) Gini A(0.5) A(1) A(2)

b1363-ch12

CULS2

Labor Income

After Income Transfer

Labor Income

After Income Transfer

6.326 0.932 0.289 0.275 0.430 0.387 0.128 0.251 0.651

6.407 0.552 0.276 0.265 0.382 0.387 0.125 0.241 0.525

6.818 0.551 0.280 0.263 0.339 0.391 0.125 0.244 0.524

6.164 0.379 0.260 0.254 0.330 0.384 0.120 0.229 0.431

Income transfer includes payment for laid-off, unemployment insurance, and dibao (Minimum Living Standard Guarantee) transfer for the poor households.

wages, and informality, etc., while the income transfer point to the impacts of social policies. Table 12.5 presets inequality indicators including percentile ratio, general entropy, Gini coefficient, and Atkinson index. The changes of those indicators between 2001 and 2005 indicated that the disparity between high income group and low income group increased. For example, the ratio of top 10% to bottom 10% increased from 6.32 to 6.82. However, one cannot simply judge whether this trend is good or not before going through the labor market changes causing the trend, which is to be answered by the following decomposition. The role of social policy is obvious to reduce the inequality. By calculating the so-called equity-sensitive average income, we may observe how the social assistance programs affect the inequality. In Table 12.5, the Atkinson Index is more strongly correlated with the extent of poverty. With increasing risk aversion parameter, society attaches more weight to income transfers at the lower end of the distribution and less weight to transfers at the top. For such a reason, we may find that the Atkinson index has a significant declining when applying the risk aversion parameter to two which is the typically used value of risk aversion parameter. For CULS1, the Atkinson index with two of e was 0.65 before income transfer and declined to 0.53 after income transfer. In the case of CULS2, the index was 0.52 and 0.46 respectively. Most the changes of income transfer could be attributed to dibao transfer.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch12

Y. Du and J. J. Xue

250

12.4.3. Regression Based Decompositions To further look at the impacts of labor market changes on inequality, we decompose the inequality based on the regression on per capita income of urban households. We group the factors affecting income inequality as household demographics, human capital, sectoral effects, informality, regional factors, and others (residuals). Table 12.6 presents regression results based on the two rounds of household survey. Most variables are statistically significant. Based on the above regression results, we may decompose the income inequality index by source. The methodology proposed by Gary Fields Table 12.6.

Regression on Household Income Per Capita. CULS1

Log of Income Per Capita Household size Ratio of adult workers to household size Percentage of male workers to total family labor force Percentage of female workers to total family labor force Average age of adult members Percentage of members with college and above education Percentage of members with senior high school Agriculture Mining and manufacturing Construction Transport Financial services Other services Ratio of wage employment with contract Ratio of wage employment without contract Ratio of self-employment Average health status of family members Ratio of members with party membership Wuhan Shenyang Fuzhou Xian cons Adj-R2 No. of obs.

Coefficients

CULS2 t

Coefficients

t

−0.07 0.46 −0.21

−5.57 4.95 −4.42

−0.11 0.64 −0.27

−5.97 5.59 −4.15

−0.10

−2.3

−0.24

−3.66

0.02 0.90

13.77 17.94

0.02 0.85

17.08 18.23

0.40

9.46

0.19

5.05

0.52 0.39 0.61 0.60 0.73 0.49 0.05 −0.04

2.42 3.03 3.91 4.95 4.83 3.97 1.07 −0.57

0.22 0.16 0.26 0.19 0.29 0.11 0.08 −0.06

1.56 4.08 4.15 4.94 5.91 2.94 1.17 −0.82

−0.01 −0.24 0.08 6.11 0.37 7.81 −0.55 −10.1 −0.62 −11.23 −0.25 −4.89 −0.68 −12.57 5.51 43.86 0.41 3426

−0.16 −2.18 0.05 4.68 0.32 6.68 −0.61 −17.38 −0.66 −18.85 −0.27 −8.04 −0.66 −19.28 7.55 59 0.54 2449

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch12

Informal Employment and Income Disparity Table 12.7.

251

Decompositions Based on Income Per Capita Regression. CULS1

Demographic Household size Percentage of male workers to total family labor force Percentage of female workers to total family labor force Average age of adult members Ratio of adult workers to household size Human Capital Percentage of members with college and above education Percentage of members with senior high school Average health status of family members Ratio of members with party membership Sectors Agriculture Mining and manufacturing Construction Transport Financial services Other services Informality Ratio of wage employment with contract to labor forces Ratio of wage employment without contract to labor forces Ratio of self-employment to labor forces Regional Wuhan Shenyang Fuzhou Xian Residual Total

9.67 1.18 0.7 0.26 2.69 4.84 14.92 9.79 1.1 0.9 3.13 6.21 0.16 0.32 0.2 1.09 1.21 3.23 0.45 0.09 0.68 −0.32 9.67 3.04 4.4 −1.33 3.56 59.08 100

CULS2 16.47 2.27 0.64 0.62 2.95 10.09 14.62 11.02 0.05 1.3 2.25 5.82 0.04 1.09 0.36 0.52 1.83 1.98 1.93 1.52 0.04 0.37 15.28 7.29 3.77 −2.02 6.24 45.8 100

(2002) is applied. Table 12.7 displays the decomposition results. We are interested in both the share of certain group of factors and its changes over time. 12.4.4. Demographics Demographic variables reflect the working capacity of the households, i.e., the extent the households use labor market. During the period of labor market dislocations, income disparities between family with and without labor forces could be small because even the households with rich human resources were unable to take advantage of them. When labor market recovered, the income gaps between those two groups of households increased.

June 5, 2012

11:23

9in x 6in

252

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

Therefore, we can see an increased contribution of demographic variables to income inequality. 12.4.5. Human Capital The returns to human capital reflected how labor market functions. It is common to see the increasing returns to human capital in a well functioning labor market (Acemoglu, 2002). Although the trend could enlarge the income gaps between well educated and less educated workers, one cannot propose policies to stop the trend. As Table 12.7 presents, the returns to human capital explain about one seventh of total income inequality among urban households in both years, which is the largest contributor to income inequality in 2001 and third largest in 2005. In such a case, when discussing the increasing income inequality in China, one must bear in mind that some factor is positive during the development procedure of labor market although it might cause increasing inequality. Furthermore, the ideal policy here is to equalize the opportunities of human capital accumulation rather than to hold back the returns to human capital. 12.4.6. Informality The effects of informalization on income inequality are mixed. On the one hand, the unemployed or the laid-offs can get employed and make earnings in informal sectors. This employment effect could reduce the overall income inequality by increasing the incomes of low-income group. On the other hand, when labor market is booming, the wage growth in informal sectors could be behind the formal sectors. In such a case, the informalization might increase the gap between formal and informal sectors and the overall inequality. Due to these two opposite directions, the contribution of informality to overall inequality is relatively small. As shown in Table 12.7, when the labor market dislocation in 2001, the employment effects was obvious, the informalization contributed only 0.45% to overall income inequality. In 2005, the income effect outweighed the employment effects more significantly and the share increased to 1.93. 12.4.7. Sectoral Effects When looking at household income inequality, the inter-sector effects do not contribute as much as human capital does. In 2001, the sector differentials

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Informal Employment and Income Disparity

b1363-ch12

253

explained about 6.2% of the inequality among urban households and the share slightly dropped to 5.8 in 2005. 12.4.8. Regional Factors Regional disparities play a role in overall income inequality. The regional factors explained 9.7% of overall inequality in 2001 and increased to 15.3% in 2005. It is good to believe that a balanced development across regions could serve to reduce the income inequality at micro-level.

12.5. Conclusions In this chapter we describe the main changes in the Chinese labor market recently and probe their possible influences on income inequality. By taking advantage of household survey data, the impacts of some changes on income inequality are analyzed empirically. The decomposition of income inequality based on regression on income per capita indicates that labor market changes did have significant impacts on income inequality, which accounted for 41% of overall inequality in 2001, and 54% in 2005. For such a reason, it is good to believe that, if China is approaching to labor market institutions with focus on equity, labor market outcomes will be helpful to hold back the trend of increasing income inequality (if it is so). But, one has to bear in mind that, not all contributors of inequality from labor market should be contained. For instance, the disparities of human capital explained one seventh of income inequality, suggesting that the labor market functions well. Policy here to reduce the income inequality is to equalize the opportunities of human capital accumulation rather than to hold back the returns to human capital. One question remains in this chapter is to what extent the rural to urban migration affects the overall inequality. For such a concern, combining the migrants sample with urban residents is necessary and will be our further work in the future.

References Acemoglu, D. (2002). “Technical Change, Inequality, and the Labor Market.” Journal of Economic Literature, American Economic Association, 40(1), 7–72.

June 5, 2012

11:23

254

9in x 6in

Growth and Inequality: An International. . .

Y. Du and J. J. Xue

Cai, F. and Wang, D. (1999). “The Sustainability of Economic Growth and the Labor Contribution.” Journal of Economic Research (Jingji Yanjiu), (10). Cai, F., Du, Y. and Wang, M. (2001). “Household Registration System and Employment Protection.” Journal of Economic Research, (12), 41–49. Cai, F., Du, Y. and Wang, M. (2003). “The Political Economy of Labor Migration.” Shanghai: Shanghai Sanlian Bookstore, Shanghai People’s Press. Cai, F., Du, Y. and Wang, M. (2005). How Far is China to a Labor Market? (zhongguo laodongli shichang zhuanxing yu fayu). Beijing: China Commerce Press. Du, Y., Gregory, R. and Meng, X. (2006). “Impacts of Guest Worker System on Poverty and Wellbeing of Migrant Workers in Urban China.” In China Update. Asia Pacific Press. Fields, G. (2002). “Accounting for Income Inequality and Its Changes: A New Method, with Application to the Distribution of Earnings in the United States.”Discussion paper at Cornell University. Ministry of Agriculture Task Force, (2001). “Demographic Change in Rural China Change and Reform of the Land System.” Topic Report, Beijing. Ravallion, M. and Chen, S. (2007). “China’s (uneven) progress against poverty.” Journal of Development Economics 82, 1–42.

b1363-ch12

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 13 EDUCATIONAL DISPARITY AND INCOME DISPARITY Jinjun Xue

13.1. Introduction With the rapid economic development in China, the country is experiencing the most serious income disparity since the establishment of socialist China in 1949. The Gini coefficient already hit 0.478 (Li, 2010), which is beyond the international warning line. There are many factors causing the inequality in China, and this chapter will focus on the function of education on income determination and income disparity. Using CHIP (China Household Income Project) data of 1988, 1995 and 2002, we found that the Gini coefficient rose from 0.235 in 1988 to 0.345 in 2002. Though the pace of the increase sped up in the first half and slowed down in the second half of the 14 years, income disparity in China had been worsening. Inquiring the factors which caused income disparity in urban China, we conducted a regressive analysis by the Mincer equation and the result showed that the schooling years of urban employees increased from 10.4 years in 1988 to 12.1 years in 2002 while the rate of returns to education rose from 4% to 11.8% in the corresponding years. By the Oaxaca-Blinder decomposition, we found that the coefficient of education (representing returns to education) and the education endowment (representing years of schooling) accounted for more than 50% of the total contribution on income growth. This means that education diffusion was the largest factor in income growth in China. However, the result also showed that the ∗ Jinjun

Xue, Professor at Economic Research Center, Graduate School of Economics, Nagoya University, Japan. 255

b1363-ch13

June 5, 2012

11:23

9in x 6in

256

Growth and Inequality: An International. . .

J. J. Xue

increase of returns to education was a main factor that caused the inequality. In conclusion, we make a hypnosis saying that education has positive influence on income growth but negative effect on income disparity. The chapter consists of six sections. Section two provides a literature survey; section three will be on data description; section four makes regressive analysis of the Mincer function to estimate the effect of education on income determination; section five applies an Oaxaca-Blinder decomposition to measure how education contributes to income disparity. Section six makes a brief conclusion. 13.2. Literature Review There are many studies on the effects of education on income determination and income disparity. Card (1999) and Harmon (2003) used regression equations to verify that, in statistics, education was positively related to income. Psacharoupoulos (1994) reviewed more than 100 countries’ research on the rate of returns to education, which all indicated that the world average rate of returns to education was around 11%. The women’s rate of returns to education was higher than men because the highly-educated women were scarcer compared to men. The rate of returns to education would decrease when the GDP per capita increase. For low-income countries (per capita income below $2,449) was about 11%, whereas, the rate of return to education for high income countries (per capita income more than $7,620) was about 7%. Buchinsky (2001) did an study using quantile regressions model based on Census data and pointed out that in the 1960s, the low-income group’s rate of returns to education was higher than the high-income group’s, however, in the 1990s, the high-income group’s rate of returns to education was higher than that of low-income group. Ashenfelter’s (1998) study shows that school education can effectively improve personal skills and income level, thus reducing the influence of the innate, family background and factors inducing income distribution disparity. These findings suggest that human capital, as a result of education, is the most important factor contributing to the growth of income. There are also a lot of studies on returns to education in China. Byron (1990) used the Nanjing market survey data and found that the rate of returns to education was 4% in urban China. Johnson et al. (1997) used CHIP data of year 1988, respectively analysis the rate of returns to education of both urban and rural areas, and the results showed that the rate of

b1363-ch13

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Educational Disparity and Income Disparity

b1363-ch13

257

returns to education was between 3% and 4%; meanwhile, they found that the rate of returns to education for rural resident was higher than that of urban resident. Zhang et al. (2005) used 1988–2001 data from China National Bureau of Statistics (NBS), found that during that period the rate of return of time education in China rose from 4% to 10.2%, and deduced that the rate of returns to education would be affected by China’s labor market reform. Fleisher et al. (2005) indicated that the rising rate of returns to education in China was closely related to the reform of the economic system and the implementation of the policies of social welfare and the diffusion of education. Shu’s (2005) study showed that, the reform of the market economy system had promoted the flow of the labor force, thereby increased returns to human capital. Knight (1993) pointed out that the rate of returns to education was higher in collective enterprises and private enterprises compared to the state-owned enterprises. Li (2008) found that the rate of returns to education in urban China showed a significant increasing trend, and the rate of returns to higher education was dramatically higher than the rate of returns to primary education, and this induced an increasing disparity between different educational backgrounds. Xue et al. (2008) conducted a study using the 2005 Shenzhen household survey data and found that the educational differences, measured by the levels of schooling years between rural migrants and urban residents, could explain 40% reasons of the total urban-rural gap. However, by now, we have not seen so many theses on the influence of education on income disparity in China. Even the study of (Xue, Sonoda and Arayama, 2008) was only based on one city’s data and limited in the analysis of the urban-rural gap. In order to make it up, we attempt to use three cross-section data-sets which cover a large span to do a systemically econometric analysis, and interpret the influence of education on income disparity. In addition to Xue’s study and an improvement of the sample’s representativeness, in this study we take urban employees as the target and use the national household survey, namely the CHIP data of 1988, 1995 and 2002, to analyze the contribution of schooling years to the growth of income and the effect of returns on education to income disparity. 13.3. Data and Variables Description It is the key period to the reform of China’s market economic system from 1988 to 2002, and during this period of time, a lot of data containing abundant information on income distribution have been recorded. Although due

June 5, 2012

11:23

258

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

to the labor market changes, the salary level of urban citizens is undergoing great changes, especially the “migrant worker shortage” which led to the rise of workers’ wage, all these new phenomena are lacking of data and empirical support like CHIP survey. Therefore, though the CHIP of 2002 is traced back to nine years ago, we can still get a more comprehensive information which will meet the requirement for analysis. In order to make up for the less time-sensitivity, we will quote and interpret some data from the China Statistical Yearbook. In terms of sample screening, although the international standard of working age population is defined as between 16 to 64 years old, the samples of workers at 16 to 19 are comparatively few in CHIP data, thus we will set the unit of analysis in workforces at age 20 to 65. Meanwhile, since we only consider the effect of education on current income, even though education do affect the unemployment compensation and the retirement pension, school students, unemployed workers, retirees, and the items with missing values will be excluded from to the target of analysis. The definition of annual income includes the dividends, bonus, price subsidy, regional subsidy, lowest life guarantee, amateur income and income in kind. Setting 2002 consumer price index as the benchmark, we adjust each year’s income, and apply the natural logarithm of annual income to the models estimation. Regarding the schooling year, we do the calculation based on the following standard: university = 16, junior college = 15, high school graduation = 12, middle school graduation = 9, the elementary school graduation = 6, below elementary school = 0. Therefore we can calculate the working experience in accordance with the standard Mincer model as follows: Working experience = Age − Years of education − 6 According to the CHIP data, the working years have a strong correlation with age, and by calculation, the correlation coefficient is 0.9645. We use the following dummy variables in our analysis: “gender dummy variables” — we set the female as 0, male as 1; 10 provinces as “regional dummy variables” — set Gansu as 0, Beijing, Shanxi, Liaoning, Jiangsu, Anhui, Henan, Hubei, Guangdong, Yunnan as 1; 7 job types as “occupational dummy variables” — set category of production staff as 0, then set categories of business owners, individual households, technicians, people in charge of the administrative organs, administrative staff and business officers as 1; 13 sectors as “industry dummy variables” — set other industry as 0, agriculture, manufacturing, mining, construction, transportation

b1363-ch13

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Educational Disparity and Income Disparity

b1363-ch13

259

industry, commercial, real estate, health care industry, education, industry, scientific research institutions and the financial sector, government agencies as 1. Table 13.1 shows the statistical description of the samples. In 1988, the average income of urban residents in China was only 4,820 Yuan (adjusted according to the 2002 consumer price index); along with China’s rapid economic growth, the annual income per capita increased to 7,462 Yuan in 1995, and reached 12,377 Yuan in 2002. Among the employees, number of men increased from 52% in 1988 to 56% in 2002. The schooling years of workers rose from 10.37 in 1988 up to 12.06 in 2002. The average age of the workers increased from 37.11 years in 1988 to 40.39 years in 2002. In terms of education, we found that while the highly educated employees increased, the low educated employment was greatly reduced. It is worth noting that, in 2002, the ratio of private enterprises’ employees already took up 23% of the total workers, while the public ownership enterprises’ (the state-owned enterprises and collective enterprises) employees were reduced from 99% in 1988 to 70% in 2002.1 In terms of the industry, the employment ratio of manufacturing sector was reduced from 43% in 1988 to 27% in 2002, whereas the ratio of real estate industry was increased from 2% in 1988 to 11% in 2002. 13.4. Estimating income Disparity by Education Background 13.4.1. Summary on income disparity In order to measure the income gap in urban China, we apply the widely used indexes of the Gini coefficient and the Theil Entropy. Table 13.2 shows the changes in income gap for recent years. Extrapolate on individual units, the Gini coefficient in urban China was 0.235 in 1988 and 0.312 in 1995, and rose to 0.345 in 2002. Theil Entropy also rose from 0.105 in 1988 to 0.207 in 2002. At the same time, the gap between high-income class and low-income class had been widened. In 1988, the ratio between the 10% highest income group and 10% lowest income group was 2.77, and in 2002,

1 The

ratio is based on the figures issued by Statistical Yearbook of China. The reason may lie in the CHIP sample scope covers the formal state-owned enterprises, collective enterprises, and private enterprises only, and excludes the migrant workers — which included in the Statistical Yearbook of China.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch13

J. J. Xue

260

Table 13.1.

Statistical Description of the Samples 1988

Variable Annual income

1995

2002

Standard Standard Standard Average deviation Average deviation Average deviation 4820

2597

7462

5103

12377

9074

8.38

0.45

8.74

0.63

9.21

0.67

Male Age 16–19 20–29 30–39 40–49 50–59 60–65

0.52 37.11 0.03 0.21 0.35 0.27 0.13 0.01

0.50 10.32 0.18 0.41 0.48 0.44 0.33 0.09

0.53 38.62 0.01 0.18 0.33 0.35 0.12 0.01

0.50 9.70 0.10 0.38 0.47 0.48 0.33 0.11

0.56 40.39 0.00 0.13 0.32 0.38 0.16 0.01

0.50 9.11 0.05 0.34 0.46 0.49 0.37 0.09

Schooling years University or above Junior college High school Middle school Elementary school Below elementary school

10.37 0.06 0.07 0.36 0.39 0.10 0.02

2.94 0.24 0.25 0.48 0.49 0.30 0.14

11.37 0.08 0.15 0.41 0.30 0.05 0.00

2.57 0.27 0.36 0.49 0.46 0.21 0.07

12.06 0.10 0.23 0.41 0.23 0.02 0.00

2.44 0.31 0.42 0.49 0.42 0.15 0.04

Experience

20.74

11.04

21.25

10.34

22.33

9.89

552

503

559

469

597

450

0.79 0.20 0.00 0.00 0.00

0.41 0.40 0.04 0.06 0.06

0.81 0.15 0.02 0.01 0.01

0.39 0.36 0.13 0.11 0.07

0.63 0.07 0.23 0.02 0.04

0.48 0.26 0.42 0.15 0.20

0.05 0.11 0.11 0.13 0.10 0.12 0.11 0.12 0.10 0.06

0.21 0.31 0.31 0.34 0.30 0.32 0.31 0.32 0.30 0.25

0.08 0.11 0.12 0.13 0.08 0.10 0.12 0.10 0.11 0.06

0.26 0.31 0.32 0.33 0.27 0.30 0.32 0.30 0.31 0.24

0.10 0.10 0.12 0.11 0.08 0.11 0.11 0.11 0.10 0.06

0.29 0.29 0.33 0.32 0.27 0.31 0.32 0.31 0.30 0.25

Logarithm of annual income

Square of experience

Ownership of the enterprises State-owned Collective-owned Private Foreign-funded Others Region Beijing Shanxi Liaoning Jiangsu Anhui Henan Hubei Guangdong Yunnan Gansu

(Continued )

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch13

Educational Disparity and Income Disparity

Table 13.1.

261

(Continued )

1988

1995

2002

Standard Standard Standard Average deviation Average deviation Average deviation

Variable Profession Business owner Individual household Techniques People in charge of the administrative organs Administrative staff Business officers Manufactory worker Industry Agriculture Manufacturing Mining Construction Transportation industry Commercial Real estate Health care industry Education industry Scientific research institutions Financial sector Government agencies Others

0.01 0.00 0.16 0.05

0.07 0.03 0.37 0.21

0.01 0.00 0.22 0.04

0.12 0.05 0.41 0.19

0.00 0.04 0.21 0.03

0.07 0.20 0.41 0.16

0.02 0.24 0.53

0.13 0.43 0.50

0.08 0.21 0.44

0.27 0.41 0.50

0.08 0.20 0.43

0.28 0.40 0.49

0.01 0.43 0.04 0.03 0.07

0.10 0.50 0.20 0.18 0.25

0.02 0.41 0.01 0.03 0.05

0.13 0.49 0.09 0.16 0.22

0.01 0.27 0.02 0.03 0.08

0.11 0.45 0.16 0.18 0.28

0.14 0.02 0.05 0.07 0.02

0.35 0.15 0.21 0.26 0.14

0.15 0.04 0.05 0.08 0.02

0.36 0.20 0.21 0.27 0.15

0.12 0.11 0.05 0.09 0.02

0.33 0.32 0.22 0.29 0.14

0.02 0.09 0.01

0.12 0.28 0.08

0.02 0.12 0.01

0.14 0.33 0.10

0.03 0.12 0.02

0.16 0.33 0.15

Volume of sample

17,385

10,071

8,775

Notes: The figures are calculated from CHIP. Table 13.2.

Summary on Income Disparity.

Gini Coefficient Theil Entropy Median (Yuan) 10% Fractal (Yuan) 90% Fractal (Yuan) Ratio between 90% and 10% Fractal Number of Sample

1988

1995

2002

0.235 0.105 4,501 2,552 7,082 2.77 17,385

0.312 0.175 6,498 3,035 12,474 4.11 10,071

0.345 0.207 10,500 4,399 21,821 4.96 8,775

Notes: The figures are calculated from CHIP.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch13

J. J. Xue

262

it rose to 4.96. Evidently, the income gap in urban China shows a tendency of expansion. According to the data from Statistical Yearbook of China, the urban Gini coefficient had risen from 0.17 in 1988 to 0.295 in 2002, slightly lower than our calculation. By our study, we found that during 1988–2002, the income disparity was expanded rapidly, although the tendency of expansion slowed down in the period of 1995–2002. The income disparity can be explained by China’s uneven growth and great social structure change, which can be summarized as “income polarizations”. Figure 13.1 and Table 13.3 show the relationship between educational backgrounds and incomes in course of time and it tells us that during 1988– 2002, the income disparity affected by the level of education continued to expand, and could be seen that education had became a more and more decisive factor on influencing the income disparity. Especially the annual income of a university graduate increases faster and higher than people at other levels of education background. Comparing the above figures of 2002, 1998, and 1995, we can see that the income disparity between different education backgrounds is expanding gradually. The educational background change of the high-income group can be identified from the top 20% high income individuals by educational 2.50

20000 18,172

18000

2.10 2.00

16000 14,168

14000 1.51 11,562

12000

1.50

10,412

1.25

9,809

10000

9,195 8,640

8000

6,880

7,238

7,700

5,391

6000 4000 2,509 2,108 2,373 2,244 2,283

1.00 4,926

0.50

2,971

2000 0.00

0 1988

1995

Below elementary High school Ratio of College to Elementary

Fig. 13.1.

Elementary school Community college

2002 Middle school College or above

Changes of Average Income by Educational Background.

Source: The figures in the table are calculated from CHIP. Unit is Chinese Yuan.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Educational Disparity and Income Disparity Table 13.3. Background.

b1363-ch13

263

Relative Income by Education

Education

1988

1995

2002

Below elementary school Elementary school Middle school High school Junior college University or above

92 104 98 100 110 130

70 89 94 100 119 135

43 75 85 100 123 157

Notes: Set high school graduates as 100. Source: The figures in the table are calculated from CHIP.

Fig. 13.2.

Change of Educational Background in High Income Class.

Source: CHIP.

backgrounds (Fig. 13.2). The education distribution of the high income groups in 1988 was that, graduates with university degree or above took up 12.86%, junior college graduates accounted for 8.1%, 33.85% of high school, 32.5% of middle school; in 2002, and the various rates were 22.85%, 29.74%, 34.13% and 12.25%. Among the high income groups, the proportion of graduates with university degree or above increased from 21%

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

J. J. Xue

264

in 1988 to over 50% in 2002, the ratio of high school and middle school graduates decreased to 46.4% from 66%. The percentage of low educational groups (elementary school or below elementary school) declined to close to zero. Therefore, we conclude from the above analysis that educational background becomes an important means to improve people’s income in China. 13.4.2. Calculation of the rate of returns to education We use the Mincer Function (formula) to calculate the rate of returns to education in China. The basic Mincer Function is as follows.
λj Xj + µ ln Y = α0 + α1 Edu + α2 Exp + α3 Exp 2 + j

In the formula, Y indicates income (including bonus); Edu for schooling years; Exp for years of working experience; Exp2 for the square of working experience years; Xj stands for other control variables (gender, region, profession, industry and so on); µ for the error term; α1 is the rate of returns to education; α2 is the return rate to constant working experience; α3 is the rate of returns to working experience’s quadratic term; λj for the rate of returns to other control variables. Among them, α1 =

∂Y /Y ∂ ln(Y ) = ∂Edu ∂Edu/Edu

We apply three models to calculate the rate of returns to education: (1) The standard Mincer formula; (2) the Mincer formula with control in gender and region; (3) the Mincer formula with control on all conditions (gender, region, industry and profession). The calculation shows that the rate of returns to education for urban workers increased rapidly from 4% in 1988 to 8.2% in 1995, as well as to 11.8% in 2002.2 The rise in rate of returns to education could be explained by China’s transformation from a planned economy to a market economy. In the period 2 In the past 20 years, the rate of returns to education was continuously rising around the world. This phenomenon can be explained by the IT revolution and other technology innovation requires the development of economic activities turn to knowledge intensive industry. The one who is equiped with high knowledge and high skill has been more and more evaluated in the labor market. According to Amartya (2000), the difference in education background enlarged the income gap, education inequality led to income disparity.

b1363-ch13

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Educational Disparity and Income Disparity

b1363-ch13

265

of planned economy, under the circumstance of state unified management, salary could not be raised even though one’s efficiency has improved. However, after the reform and opening–up, China’s labor market has been reformed, workers’ salary began to change along one’s work efficiency. The changes can be further explained from the following three aspects. First, is the reform of state-owned enterprises. Previously, the state-owned enterprise worker’s salary was set in accordance to the government regulation. After introduction of the operational autonomy, the state-owned enterprise can independently decide the salary of its employees. Second is the development and expansion of the non-state-owned enterprises. In 1988, the non-state-owned enterprises employees took up 0.5% of the urban employees, but in 2002, the percentage was increased to 25% (Table 13.4). The non-state-owned enterprises are profit-motivated, thus they would react more validly to the employees’ productivity. Third is both the state-owned companies and non-stated enterprises introduced salary evaluation system which is based on the ability (Minami et al., 2008). In 1998, although the rate of returns to education in urban China was still lower than that of the world average, it has reached the average level of the low and middleincome countries in 2002. By our study, we also found that the rate of returns to working experience dropped from 4.8% to 2.7% during 1988–2002. There were many reasons for the change and according to Minami (2008), the main reason was that with the development of market economy, the capability oriented salary system had been popularized, thus the effect of working experience on salary had been diminished. The rate of return to gender had remarkable changed, which rose from 11.6% in 1988 to 19.6% in 2002. The income of the male workers was significantly higher than that of the female and the gap was expanding. Meanwhile, the trend of decisive coefficient was descending. According to Xue (2010), ability and other uncalculated elements play an increas˙enply important role in income determination. Profession and occupation as general human capital variables, have strong connection with schooling years. Therefore, the rate of returns to education is comparatively higher under the circumstance of no control over the industry and profession. In order to further explore the influence of education on income level, we brought university degree and above, high school, junior high school and elementary school of dummy variables to join the Psaltery equation to calculate the effect (set people didn’t graduate from elementary schools

June 5, 2012

Analysis Result of the Mincer Function.

1988

Schooling

(1)

(2)

1995 (3)

(4)

(5)

2002 (6)

(9)

0.118 (39.46∗∗∗ ) 0.027 (10.37∗∗∗ ) −0.0001 (−2.57∗∗∗ )

0.110 (39.46∗∗∗ ) 0.032 (13.33∗∗∗ ) −0.0004 (−6.64∗∗∗ ) 0.196 (15.93∗∗∗ ) 7.708 (154.96∗∗∗ ) Yes No No 0.303 8,775

0.069 (20.78∗∗∗ ) 0.032 (13.91∗∗∗ ) −0.0004 (−8.61∗∗∗ ) 0.177 (14.55∗∗∗ ) 8.628 (74.30∗∗∗ ) Yes Yes Yes 0.354

7.272 (148.70∗∗∗ ) No No No 0.167

1. t-value are in parentheses. 2. ∗∗∗ , ∗∗ , and ∗ denote statistical significance at 1%, 5% and 10% levels respectively. Notes: The calculation has eliminated the estimation value of dummy variables, as region, profession, industry. Three models have been set to meet each year’s various conditions. Source: The figures in the table are calculated from CHIP.

Growth and Inequality: An International. . .

(8)

J. J. Xue

0.040 0.037 0.030 0.082 0.075 0.052 (38.01∗∗∗ ) (25.05∗∗∗ ) (31.07∗∗∗ ) (31.46∗∗∗ ) (17.39∗∗∗ ) (38.09∗∗∗ ) Experience 0.048 0.048 0.047 0.049 0.047 0.047 (55.07∗∗∗ ) (54.39∗∗∗ ) (25∗∗∗ ) (26.56∗∗∗ ) (26.22∗∗∗ ) (51.06∗∗∗ ) Exp2 −0.0006 −0.0007 −0.0007 −0.0006 −0.0006 −0.0006 (−30.28∗∗∗ ) (−34.31∗∗∗ ) (−34.06∗∗∗ ) (−13.89∗∗∗ ) (−15.84∗∗∗ ) (−17.08∗∗∗ ) Gender 0.116 0.109 0.141 0.139 (19.17∗∗∗ ) (13.99∗∗∗ ) (12.03∗∗∗ ) (21.04∗∗∗ ) Constant 6.594 6.654 6.823 7.149 7.496 7.808 (420.23∗∗∗ ) (349.46∗∗∗ ) (138.06∗∗∗ ) (184.71∗∗∗ ) (184.42∗∗∗ ) (108.12∗∗∗ ) Province No Yes Yes No Yes Yes Occupation No No Yes No No Yes Industry No No Yes No No Yes R-squared 0.278 0.398 0.408 0.188 0.351 0.373 Observation 17,385 10,071

(7)

9in x 6in

Log income

11:23

266

Table 13.4.

b1363-ch13

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Educational Disparity and Income Disparity

b1363-ch13

267

Table 13.5. Rate of Returns to Education by Academic Background (%). 1988 University and above

0.58 (23.72∗∗∗ ) Junior college 0.47 (19.42∗∗∗ ) High school 0.38 (17.22∗∗∗ ) Middle school 0.27 (12.17∗∗∗ ) Elementary school 0.18 (7.73∗∗∗ ) 10–19 years experience 0.38 (43.16∗∗∗ ) 20–29 years experience 0.56 (64.03∗∗∗ ) 30+ years experience 0.65 (66.39∗∗∗ ) cons 6.90 (302.79∗∗∗ ) cod 0.258 Number of samples 17385

1995

2002

1.03 (12.18∗∗∗ ) 0.90 (10.79∗∗∗ ) 0.69 (8.33∗∗∗ ) 0.47 (5.69∗∗∗ ) 0.26 (3.04∗∗∗ ) 0.35 (19.34∗∗∗ ) 0.63 (34.6∗∗∗ ) 0.73 (35.68∗∗∗ ) 7.67 (91.62∗∗∗ ) 0.178 10071

1.60 (9.94∗∗∗ ) 1.29 (8.07∗∗∗ ) 1.00 (6.26∗∗∗ ) 0.72 (4.51∗∗∗ ) 0.53 (3.24∗∗∗ ) 0.26 (11.37∗∗∗ ) 0.43 (18.9∗∗∗ ) 0.61 (24.89∗∗∗ ) 7.78 (48.51∗∗∗ ) 0.164 8775

Notes: ∗ , ∗∗ , ∗∗∗ respectively indicates statistical significant at the level of 1%, 5% and 10%. T value is in the bracket. The figures in the table are calculated from CHIP.

and years of experience from 0 to 9 as 0). Regression analysis results are shown in Table 13.5 which indicate that all education dummy variables and working experience of dummy variable coefficient are statistically significance. This means that academic background and working experience have significant influence on the income. It is notable in Table 13.5 that reruns to education change to a great extent by different academic background and each level of degree of the proceeds is significantly increased over time. On the other hand, the result displays that the rate of returns to each working experience group decreased. Basing on this assumption, we have done the calculation on the rate of returns to the promotion of education. Table 13.6 shows that the effect of education promotion is significant to the increase of income and the effect is more noted as time goes by. Meanwhile, we found that not schooling years but the final academic degree, such as high school, junior college or university, is the real important factor through the salary negotiations.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

J. J. Xue

268 Table 13.6.

1988 1995 2002

Rate of Returns to the Promotion of Education Level.

Grade up from elementary school — middle school

Grade up from middle school — high school

Grade up from high school — junior college

Grade up from high school — university

−1.8 1.7 4.3

0.6 2.1 5.6

3.2 6.1 7.0

6.8 7.8 12.0

Source: CHIP. Unit = %.

13.5. Decomposing the Contributions of Education to Income Disparity 13.5.1. Methodology We apply the Blinder-Oaxaca decomposition method to analyze the contribution of education to income growth. That the Blinder-Oaxaca decomposition assumes the function of logarithm income is as follows: ln wm = Σβmj · xmj ;

ln wn = Σβnj · xnj .

Where wm is the annual income of the benchmark year; wn is the annual income of the comparative year; x stands for the factors which affect income such as education, working experience, gender, profession, region as well as industry; β is the coefficient. The Blinder-Oaxaca decomposition is often used to analyze the income differences by gender, here we use it to decompose the income gap between benchmark year and comparative year, as well as each factors’ effect on the increase of income. In order to do this, we deform the above two functions as following: ln wm − ln wn = Σβmj · xmj − Σβnj · xnj = Σβmj · xmj − Σβmj · xnj + Σβmj · xnj − Σβnj · xnj = Σβmj (xmj − xnj ) + Σ(βmj − βnj )xnj Where Σβmj (xmj − xnj ) is defined as the income gap from the stock difference of the benchmark year and comparative year (xj ), which is the income gap from the increase of education stock, namely the years of schooling; Σ(βmj − βnj )xnj is known as income gap from the coefficient difference of the standard year and comparative year (βj ), which is the income gap from the rate of returns to the stock of education, namely returns to education.

b1363-ch13

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Educational Disparity and Income Disparity Table 13.7.

b1363-ch13

269

Result of the Blinder–Oaxaca Decomposition.

Log income difference Endowments (No.) cons & other variables Total of education endowments Difference of the education stock Difference of education coefficient Percentage of endowments (%) cons & other variables Total of education endowments Difference of the education stock Difference of education coefficient

1988–1995

1995–2002

1988–2002

0.357

0.475

0.832

0.083 0.274 0.052 0.222

0.229 0.245 0.048 0.198

0.312 0.519 0.117 0.403

23.3% 76.7% 14.5% 62.3%

48.3% 51.7% 10.0% 41.7%

37.6% 62.4% 14.0% 48.4%

Source: The figures are calculated from CHIP.

The result of the Blinder-Oaxaca decomposition analysis shown in Table 13.7 proves that the inverse decomposition and positive decomposition share similar results, thus we neglect the results of the inverse decomposition. Different from the traditional decomposition, here we do the decomposition over the years. The result shows that the constant and other variables contributed a little while the education coefficient and education stock contributed more than 70% to the income growth during 1988– 1995. This means that education is the biggest factor contributing to the growth in income. Regarding the contribution of education to the income growth, the ratio of stock was 14%, the coefficient was 48.4%, and the sum of the two was 62.4% during 1988–2002. In accordance to the time span, the rate decreased from 76.7% in the first seven years (1988–1995) to 51.7% in the next seven years (1995–2002). By further decomposition, we found that the difference of the coefficient and the difference of the stock had decreased during 1988–2002. All the above results imply that the difference in education coefficient is larger than the difference in education stock, and it implies that the rise in the rate of returns to education is the decisive factor in the expansion of income gap.

13.5.2. Estimating the effect of education on income distribution We apply a fractile regression equation to analyze the affect of education on the income distribution. The definition of the fractile regression equation

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

J. J. Xue

270

is as follows: ln wi = xi βθ + uθi ,

Quant θ (ln wi |xi ) = xi βθ

Where Quant θ (ln wi |xi ) is the distribution of all conditions for ln wi of xi . When θ = 0.5, the above function is a median regression. Calculating regression function in each fractile θ is quite effective to understand the change of regression function with income distribution. The estimation of the parameters of the fractile regression function is as follows:   

θ| ln wi − xi βθ | + (1 − θ)| ln wi − xi βθ | min  β∈Rk  i:ln wi ≥xi β

i:ln wi chi2

Coefficient 0.042 0.000 −0.000004 1.642 −0.690 −1.175 −0.163 0.461 −0.206 −0.179 −0.724

z-statistics

p-value

9.110 0.300 −1.430 1.450 −0.530 −0.750 −2.710 4.330 −1.330 −1.450 −4.600

0.000 0.764 0.153 0.146 0.594 0.452 0.007 0.000 0.184 0.147 0.000

5275 −1784.0426 0.067 256.180 0.000

Note: 3,644 groups (14,135 observations) dropped because of all positive outcomes. Variables of location and household head’s education are omitted because of no/little within-group variance. Source: Calculated by the authors.

estimated by STATA, using a conditional likelihood. Tables 14.4–14.6 show the estimation results for three cases, where the dependent variable is the probability of owning low-price house, market-price house, and all residentowned house, respectively. From Tables 14.4–14.6, we can elucidate the following points: (1) In the case of low-price resident-owned houses, two variables on household income have no significant effects. In contrast, some non-income variables such as PSSOE (government official or staff of State Owned Enterprises), Hage (household head age) and Pop (household population size) have significantly positive effects, while the variable Spboss (self-employed small business owner) has a significantly negative effect. (2) In the case of market-price resident-owned houses, while non-income variable Hage has a negative effect, two income-related variables show significantly positive effects. Meanwhile, among variables reflecting household member’s employment sector (occupation), instead of

b1363-ch14

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Housing Inequality and Underlying Factors in Urban China

b1363-ch14

291

Table 14.5. The Effects of Household Characteristics on the Probability of Owning Market-Price House (2004–2007). Dependent Variable: Mpmyhouse (Yes = 1, No = 0) Hage Headincome Tincome Rurban Rrural Murban Pop PSSOE COL BP Spboss Number of observation Log likelihood Pseudo R2 LR chi2(12) Prob>chi2

Coefficient −0.048 0.000 0.00002 −0.089 0.302 0.735 −0.024 −0.079 0.196 0.291 0.235

z-statistics

p-value

−9.800 3.180 4.400 −0.120 0.360 0.650 −0.350 −0.670 1.160 2.340 1.500

0.000 0.001 0.000 0.908 0.719 0.514 0.727 0.502 0.248 0.020 0.135

4652 −1522.8911 0.092 307.160 0.000

Note: 3,813 groups (14,758 observations) dropped because of all positive outcomes. Variables of location and household head’s education are omitted because of no/little within-group variance. Source: Calculated by the authors.

PSSOE (government official or SOE staff), BP (private company staff) has a significantly positive effect. (3) In the case of all resident-owned houses, some non-income variables and two income related variables have significantly positive effects, reflecting the overall effect of the two types of variables on owning resident-owned houses. 14.5.2. The Effects of Household Characteristics on House Size As shown in Table 14.3, housing inequality measured by house floor space or per capita house floor space has risen through the period from 2004 to 2007. Furthermore, if the house quality (market value) is taken into account, it is found that real housing inequality in urban China has reached a considerably high level. In Sec. 14.5.1, we examined the effects of household characteristics on the probability of owning various resident-owned houses. In this section, using the same panel data from the NSB survey

June 5, 2012

11:23

292

9in x 6in

Growth and Inequality: An International. . .

E. B. Dai and J. J. Xue Table 14.6. The Effects of Household Characteristics on the Probability of Owning All Resident-Owned House (2004–2007). Dependent Variable: Myhouse (Yes = 1, No = 0) Hage Headincome Tincome Rurban Rrural Murban Pop PSSOE COL BP Spboss Number of observation Log likelihood Pseudo R2 LR chi2(12) Prob>chi2

Coefficient

z-statistics

p-value

0.040 0.000 0.00002 2.359 0.780 −0.508 0.284 0.579 0.039 0.542 0.348 1987 −636.8074 0.125 188.670 0.000

5.270 3.250 2.870 2.120 0.610 −0.330 3.000 3.190 0.160 2.420 1.380

0.000 0.001 0.004 0.034 0.543 0.741 0.003 0.001 0.876 0.016 0.166

Note: 4,483 groups (17,423 observations) dropped because of all positive outcomes. Variables of location and household head’s education are omitted because of no/little within-group variance. Source: Calculated by the authors.

and the Panel Regression Model, we will examine the effects of household characteristics on house size (house floor space). Table 14.7 shows the estimation results for the effect of household characteristics on house floor space. The two sets of estimation results in the table are based on the fixed effect model and the random effect model, respectively. The Hausman test suggests that we should use the results of fixed effect model. The results of fixed effect model in Table 14.7 show that total household income and a few non-income variables including PSSOE (government official and SOE staff), existence of local Hukou (registered residence status) in the current city of residence, Hage (household head age), and Pop (population size), have significantly positive effects on house floor space, showing a picture that is quite similar to what we see from the results of the fixed effect logit model on the probability of owning all resident-owned houses. However, because the migrant households without local Hukou have no chance to buy low-price resident-owned houses or rent public houses, they have to rent private houses or buy market-price houses at very high

b1363-ch14

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch14

Housing Inequality and Underlying Factors in Urban China Table 14.7. Floor Space.

293

The Effects of Income and Other Household Characteristics on House

Dependent Variable: Hspace

Coefficient t-statistics p-value Coefficient z-statistics p-value

Hage Headincome Tincome Rurban Rrural Murban Pop PSSOE COL BP Spboss cons

0.1475 0.00004 0.0002 20.9768 19.2544 4.5292 5.5116 2.8518 −0.3987 0.3771 7.6332 29.0784

Fixed Effect Model

Number of observation Number of groups R2 within between overall F -test

4.740 1.360 11.190 4.080 3.400 0.690 14.640 4.100 −0.430 0.480 7.600 5.380

0.000 0.175 0.000 0.000 0.001 0.490 0.000 0.000 0.670 0.632 0.000 0.000

Random Effect Model

0.1276 0.00008 0.0002 23.8841 27.2390 8.8243 6.2050 3.0539 −0.3784 −0.7758 9.8159 25.2629

4.670 3.440 10.380 5.350 5.460 1.530 18.180 4.810 −0.420 −1.070 10.790 5.340

19,410

19,410

4,999

4,999

0.056 0.058 0.056 F(11, 14400) = 78.16 Prob>F = 0.000

0.055 0.068 0.063 Chi2(11) = 1179.91 Prob>Chi2 = 0.000

Hausman test

0.000 0.001 0.000 0.000 0.000 0.126 0.000 0.000 0.675 0.285 0.000 0.000

Chi2(11) = 103.56 Prob>Chi2 = 0.000

Source: Calculated by the authors.

prices, resulting in typically low house floor space. Therefore, the effect of local Hukou on house floor space is stronger than that on the probability of owning various resident-owned houses.

14.6. Conclusions Using panel data from China’s NBS survey conducted in 2004–2007, this chapter measured the level of housing inequality in contemporary urban China and analyzed the underlying factors of rising housing inequality. The main findings can be summarized as follows. (1) Housing inequality in urban China has risen to a considerably high level in recent years. If inequality is measured using “adjusted per capita

June 5, 2012

11:23

294

9in x 6in

Growth and Inequality: An International. . .

E. B. Dai and J. J. Xue

house floor space,” which takes a house’s market value into account, the Gini coefficient of housing inequality in urban China has reached 0.486 in 2007, rising from 0.466 in 2004, 0.479 in 2005, and 0.481 in 2006. (2) Although China has abolished the old housing allocation system, which was controlled by various levels of government and various kinds of state-owned “working units” (i.e., companies or institutes), and saw the housing market become the major supplier of urban house after late 1990s, the effects of non-income factors — including household members’ employment sectors and Hukou (registered residence status) — on the condition of housing in urban households still remain strong. The results of panel analysis show that occupation as a government official or SOE staff, and existence of local Hukou in the current city of residence have significantly positive effects on a household’s probability of owning a low-price resident-owned house, as well as on the house size (floor space). (3) With the dramatic transition of China’s urban housing system from a planned system to a market-oriented one, household income has become an important determinant of housing conditions in urban China. Like in other market economies, household income-related characteristics, including total household income and household head income, have significantly positive effects on the probability of owning market-price resident-owned house and all resident-owned house, as well as on the house size (floor space). By comparing the effects of a household’s income characteristics and non-income characteristics summarized above, we can conclude that rising housing inequality in urban China is a combined result of the increasing impact from ongoing market-oriented reform and the persisting impact of some traditional systems and institutions including the old housing allocation system and the Hukou (registered residence status) system. Particularly, compared to the effects of market-oriented reform on housing inequality, which is usually more transparent and relatively fair to all households, the old housing allocation system and the Hukou system are favorable to only some residents, such as officials and urban registered residents, so that their effects on housing inequality are more detrimental. Thus, in order to alleviate rising housing inequality, the government should first promote reforms of these institutions that are lagging behind. Meanwhile, it is also necessary to adjust the highly uneven income distribution

b1363-ch14

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Housing Inequality and Underlying Factors in Urban China

b1363-ch14

295

and to improve the governance of the current housing market — a market that lacks property tax and obviously favors high income households. References Dai, E.B. and Xue, J.J. (2007). “Housing Disparity and Income Inequality in Urban China.” The Economic Science, 55(3), 69–84. Dai, E.B. and Xue, J.J. (2002). “Income Inequality and housing inequality in Urban China: Influence of Migrant Household.” Higashi Ajia he no Shiten (A Viewpoint to East Asia), 13(3), 44–57 (in Japanese). Gao, L. (2010). “Achievements and Challenges: 30 Years of Housing Reforms in the People’s Republic of China.” ADB Economics Working Paper Series, No. 198. Hamerle, A. and Ronning, G. (1995). “Panel analysis for qualitative variables.” In Handbook of Statistical Modeling for the Social and Behavioral Sciences, Arminger, G., Clogg, C.C. and Sobel, M.E. New York: Plenum., 401–451. He, S.J., Liu, Y.T., Wu, F.L. et al. (2010). “Social Groups and Housing Differentiation in China’s Urban Villages: An Institutional Interpretation.” Housing Studies, 25(5), 671–691. Huang, Y.Q. and Jiang, L.W. (2009). “Housing Inequality in Transitional Beijing.” International Journal of Urban and Regional Research, 33(4), 936–956. Huang, Y.Q. (2005). “From Work-unit Compounds to Gated Communities: Housing Inequality and Residential Segregation in Transitional Beijing.” In Restructuring the Chinese City: Changing Economy, Society and Space, Laurence, J., Ma, C. and Wu, F (Eds.). London and New York: Routledge. pp. 192–221. Logan, J.R., Bian, Y.J. and Bian, F.Q. (1999). “Housing inequality in urban China in the 1990s.” International Journal of Urban and Regional Research, 23(1), 7–25. Logan, J.R., Fang, Y.P. and Zhang, Z.X. (2010). “The Winners in China’s Urban Housing Reform.” Housing Studies, 25(1), 101–117. Riskin, C., Zhao, R.W. and Li, S. (Eds.) (2001). China’s Retreat from Equality: Income Distribution and Economic Transition. New York: M.E. Sharpe. Sato, H. (2006). “Housing inequality in urban China in the 1990s.” China Economic Review, 17(1), 37–50. Tang, B.S.,Wong, S.W. and Liu, S.C. (2006). “Property Agents, Housing Markets and Housing Services in Transitional Urban China.” Housing Studies, 21(6), 799–823.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch14

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 15 AGRICULTURE PROFITABILITY AND INCOME DISPARITY A CASE STUDY ON THE MARKET LABOR SUPPLY OF HOUSEHOLD HEAD AND FARM EFFICIENCY IN RURAL CHINA∗

Tadashi Sonoda†

15.1. Introduction Income disparity between urban and rural areas is one of the most remarkable types of disparity in China these days. While one can attribute the urban–rural income disparity to the rapid income increase in urban areas, one can attribute it more appropriately to sluggish income increase in rural areas. Then, a natural approach to the disparity issue is to examine why income of rural households increases so sluggishly or why they do not (or cannot) change their traditional behavior. Incomes of farm households accrue mainly from farming and engagement in market work,1 and the latter comes to gain its importance as the economy develops. This is partly because the demand for food and therefore the prices of farm commodities grow slower than the rates of wage in other industries during economic development, which makes the growth rate of income from farming more sluggish than that from market work. Under these economic circumstances, it might be more natural for rural households to improve their income by allocating more of their time to market work rather than to farming. Focusing on this point, this study ∗ This work was supported by the Nitto Foundation and the Grant-in-Aid for Scientific Research ((C) 20580234 and (B) 20402025). † Associate professor at Graduate School of Economics, Nagoya University, Japan. 1 Market work in this study refers to all activities in which household members supply their time to earn wage (including wage in kind).

297

b1363-ch15

June 5, 2012

11:23

298

9in x 6in

Growth and Inequality: An International. . .

T. Sonoda

investigates relevant factors affecting market labor supply of the head of Chinese farm households and utilizes the result to examine how to improve their income. In their empirical analysis of market labor supply of the head of farm households, Goodwin and Mishra (2004) and Zeng and Goodwin (2008) focus on farm profitability2 (production revenue per unit cost of variable input) as one of the most important factors affecting his market labor supply. They focus on it because as the head increases his hours of market work, he might make less effort to collect information about farm production technology and inputs, which is expected to cause lower profitability in farming. This relation might conversely imply that higher profitability in farming causes more effort of the head in farming, which can decrease his hours of market work. Thus, relation between the head’s market hours and farm profitability could be useful information in examining how much effort to be put in farming and market work to improve the household income or utility. This study uses the Chinese Household Income Project 2002 (Li, 2002) and follows the empirical procedure of Goodwin and Mishra (2004) to find relevant factors affecting market labor supply of the head of farm households. It also tries to make an additional interpretation of the estimation result. Goodwin and Mishra (2004) assume an agricultural household model to derive their estimation equations but they overlook an important implication of the model in examining their empirical result. This study interprets the result in terms of “separability” of the model and draws some policy implications to improve income of farm households in the western region of China. Section 2 introduces an agricultural household model to describe market labor supply of the household head, and then specifies the relation between his market work hours and farm profitability. After interpreting this relation in detail, we briefly explain how to estimate the model. Section 3 introduces data and variables used in the empirical analysis, shows the regional difference in these variables among the eastern, central, and western parts of China, and then examines the estimation results. Section 4 summarizes the results and draws some implications about how to improve income of farm households in the western region. 2 The variable E to be defined below is called “efficiency” by Goodwin and Mishra (2004), but we simply call it farm profitability to avoid confusion with other efficiency indexes of production technology.

b1363-ch15

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch15

Agriculture Profitability and Income Disparity

299

15.2. Model and Estimation Method We follow Goodwin and Mishra (2004) to assume that the head of a farm household maximizes its utility function u(C, t; s) subject to the three constraints shown below, where C denotes the amount consumed by all household members; t = T − L − M denotes the head’s hours of leisure (T: endowed time, L: farm work hours, M: market work hours); and s denotes characteristics of the head and his family (including age and family composition). The three constraints are written as C = pX − qF + wM + V

(Budget constraint)

X = f(L, F; K)

(Production technology constraint)

L+M+t=T

(Time constraint)

Variables X, F, and K respectively denote the amount produced of farm commodity, the amount of other variable input than labor (fertilizers, pesticides, and seeds), and the amount of fixed inputs (farm land and farm tools); p, q, and w respectively denote the market prices of farm commodity, other variable input, and labor, with the price of consumption commodity set equal to one; and V denotes non-labor income. As assumed in most empirical studies of agricultural household models (e.g., Jacoby, 1993), the head always works on farm (L > 0) and may or may not engage in market work (M ≥ 0). If we also assume no transaction cost in engaging in market work and no restriction on market hours,3 the first order conditions for the utility maximization problem may be written p(∂X/∂F) = q p(∂X/∂L) =

(15.1a)

∂u/∂t ∂u/∂C

∂u/∂t =w ∂u/∂C

if M > 0

>w

if M = 0

C = pX − qF + wM + V

(15.1b) (15.1c)

(15.1d)

Assuming this model as a background, Goodwin and Mishra (2004) examine relation between farm profitability and the head’s hours of market 3 Transaction

cost in engaging in market work and restriction on market hours will be examined for our empirical model.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch15

T. Sonoda

300

work. They first define farm profitability E as production revenue per unit cost of variable input4 E = pX/qF They also define a latent variable M∗ which indicates the head’s incentive to work in the market. This variable is related to his hours M of market work as M = M∗ =0

if M∗ > 0

(15.2)

∗

if M ≤ 0

Farm efficiency E and the head’s incentive M∗ to work in the market are assumed to be determined by the following simultaneous equations5 : E = α1 M∗ + x 1 β1 + u1 ∗



M = α2 E + x 2 β2 + u2

(15.3) (15.4)

where x1 denotes a vector of exogenous variables on the production side; x2 denotes a vector of exogenous variables on the consumption side; u1 and u2 are error terms.6 The coefficients α1 and α2 in Eqs. (15.3) and (15.4) represent an important feature of the model. The case α1 < 0 implies that farm profitability falls as the head’s incentive to work in the market increases. This case could happen because farm profitability could worsen as the head makes more effort to work in the market and makes less effort to collect information about farm production technology and inputs. The case α2 < 0 implies that the head’s incentive to work in the market increases as farm profitability falls. This case could happen if the head’s leisure is a normal good. That is, a fall in farm profitability is expected to decrease farm profit or income of the household, which is expected to increase his total labor supply (the sum of farm and market work hours) through the associated income effect. His hours of market work are then expected to increase because the fall in farm profitability is likely to decrease the demand for farm labor and hence the reward for farm work is expected to reduce. 4 The first order condition for variable input F implies that index E is equal to the reciprocal of production elasticity ∂ ln X/∂ ln F. Therefore, index E should be constant if the production technology is of Cobb–Douglas type. 5 It is easier to use the latent variable M∗ in constructing the empirical model because reduced form for M∗ is obtained more easily. 6 Variables included in the vectors x and x will be specified in the next section. 1 2

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Agriculture Profitability and Income Disparity

b1363-ch15

301

Thus, if α1 < 0 and α2 < 0, the farm household (or the head) faces tradeoff to improve its utility: Fall (rise) in farm profitability and increase (decrease) in market hours. On the other hand, if α1 > 0 and α2 > 0, the farm household does not face this tradeoff and it can increase market hours while improving farm profitability.7 It should be noted, however, that the sign of α1 may or may not be the same as that of α2 , which is an empirical question. The coefficient α1 also has a relevant interpretation in the agricultural household model, though Goodwin and Mishra (2004) do not examine it. An important feature of agricultural household models is called “separability” (Singh, Squire, and Strauss, 1986). In a “separable” model, the household’s behavior on the production side (demand for production factors and supply of output) is determined independently of its behavior on the consumption side (demand for consumption commodities and supply of labor). In the model used here, when the head supplies positive hours to the market, the household’s behavior on the production side (represented by the optimal quantity of F, L, and X) is determined only by the production function, Eqs. (15.1a) and (15.1b) with the marginal rate of substitution being replaced by the market rate of wage, which is not affected by the utility function, Eqs. (15.1c) and (15.1d). Consequently, the farm household acts like a firm maximizing its profit in its production organization. In a “non-separable” model, on the other hand, its behavior on both the production and consumption side should be determined jointly, which makes its behavior more difficult to describe. Now, we interpret Eqs. (15.3) in terms of the separability. Farm profitability E is a variable on the production side, while the incentive M∗ to work in the market is a variable on the consumption side. If α1 = 0, then the profitability E can be explained only by x1 , a vector of exogenous variables on the production side, which implies the model is separable. On the other hand, if α1 = 0, then the incentive M∗ affects the profitability E, which in turn implies that the model is non-separable. It should also be noted that the model with α1 = 0 is non-separable even if the head participates in market work. Therefore, the case α1 = 0 suggests the farm household face high transaction cost in engaging in market work and/or restriction on market hours (see e.g., de Janvry, Fafchamps, and Sadoulet, 1991; Benjamin, 7 The

case α1 > 0 could happen because, for example, the head might get information about better farm technology or inputs in the place he works. The case α2 > 0 could happen if leisure of the head is an inferior good.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch15

T. Sonoda

302

1992). Consequently, the hypothesis α1 = 0 can be used to test separability of the agricultural household model in this study. Equations (15.2), (15.3), and (15.4) are similar to the model of Nelson and Olson (1978), which can be estimated in the following way (Amemiya, 1985). Reduced form of endogenous variables E and M∗ may be expressed as E = x δ 1 + v1

(15.5)

M∗ = x δ 2 + v2 ∗

M=M =0

(15.6) ∗

if M > 0 if M∗ ≤ 0,

where x = (x 1 , x 2 ) and v1 and v2 denote error terms. Then, we first estimate Eq. (15.5) to obtain an ordinary least squares (OLS) estimate δˆ1 ˆ = x δˆ1 . Assuming v2 |x is distributed as N(0, σ 2 ), we also and compute E ˆ ∗ = x δˆ2 . This fitted value estimate Eq. (15.6) by tobit method to obtain M ∗ ∗ ˆ can then be used to replace M and to estimate equation (15.3) by M ˆ OLS, from which we obtain α ˆ1 and βˆ1 . Furthermore, the fitted value E can be used to replace E and to estimate equation (15.4) by tobit method, ˆ1, α ˆ 2 , βˆ1 , and βˆ2 from which we obtain α ˆ 2 and βˆ2 . The standard errors of α need to be corrected to allow for two step estimation. We compute their standard errors using bootstrap with 400 replications.

15.3. Data and Empirical Results The empirical analysis uses data of rural households in the Chinese Household Income Project 2002. The original data include 9,200 rural households but we use 6,831 households satisfying the following conditions: (1) production revenue from crop (including grains and economic crops) and/or livestock production is positive, (2) the crop production cost is positive and the sum of cultivated own land and rented land is also positive, (3) the household head is male and he engages in farming, (4) farm profitability E does not take on extreme values.8 The conditions (1) and (2) restrict farm commodities produced and production structure to some extent, the condition (3) restricts our sample to heads of typical farm households, and 8 We

exclude 27 households with E > 30 in the eastern region, five households with E > 30 in the central region, and five households with E > 20 in the western region.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Agriculture Profitability and Income Disparity Table 15.1. Variable M E

b1363-ch15

303

Variables Used in the Empirical Analysis. Description

Annual hours of market work of the household head Ratio of production revenues from farming to variable costs

Variables included in x1 educ age fworker fworker share fworker male share tot land irr share ag capital ecrop share live share grainquota disaster

Years of schooling of the household head Average age of farm workers Share of farm workers in all members Share of male farm workers in all members Sum of cultivated own land and rented land (mu) Share of irrigated land in tot land Value of large and medium sized farm tools, farm machinery and equipment, and structures used for production (yuan) Share of revenues from economic crop production in those from farming (crop and livestock production) Share of revenues from livestock production in those from farming (crop and livestock production) Dummy variable taking value of 1 if the village is assigned any compulsory grain quota in 2002. Dummy variable taking value of 1 if the village suffers from natural disaster in 2002.

Variables included in x2 age numhh childlt6 childge6 nonlaborinc debt asset dist councenter dist terminal bigcitysuburb percap income

Age of the household head Number of all household members Number of children of age less than 6 Number of children of age between 6 and 15 Non-labor income of the household (yuan) Ratio of total debt to total asset of the household Distance from the village to the nearest county seat (km) Distance from the village to the nearest transportation terminal (km) Dummy variable taking value of 1 if the village is suburb of a large or middle city Net income per capita of the village (yuan)

Note: In addition to the variables shown above, x1 includes province dummies, x2 includes educ, disaster, and province dummies. The variables nonlaborinc and percap income are deflated using the provincial price index estimated by Brandt and Holtz (2006).

the condition (4) excludes outliers of the variable E. Variables used in the empirical analysis and their description are presented in Table 15.1. To allow for regional difference in farm production and labor market, we follow Gustafsson, Li and Sicular (2008) to classify 22 provinces, autonomous regions, and directly administered municipalities into the

June 5, 2012

9in x 6in

11:23

b1363-ch15

T. Sonoda

304 Table 15.2.

Growth and Inequality: An International. . .

Regional Sample Mean of Variables Used in the Empirical Analysis. West

Number of households Characteristics of the household head Rate of participation in market work (%) M age educ

2,148

Center 2,608

East 2,075

51.68

55.83

59.66

511.38 45.60 6.63

631.22 44.83 7.19

897.57 48.27 7.66

2.80 4.50 40.34 0.64 0.34 0.21 0.85 310.64 0.07 0.24 0.38 7.50 0.51 1,011.32

3.89 4.13 40.32 0.63 0.33 0.17 0.74 519.40 0.06 0.27 0.23 8.44 0.56 1,595.05

3.80 3.89 43.93 0.66 0.35 0.11 0.53 600.65 0.06 0.28 0.23 6.12 0.66 1,722.24

32.76 6.98 0.04 1,572.44 0.65 0.29

20.08 5.01 0.02 2,190.83 0.52 0.24

21.15 4.26 0.10 3,544.96 0.44 0.14

Characteristics of the household E numhh age fworker fworker share fworker male share childlt6 childge6 nonlaborinc debt asset ecrop share live share tot land irr share ag capital Characteristics of the village dist councenter dist terminal bigcitysuburb percap income disaster grainquota

eastern, central, and western regions in the following analysis.9 Regional sample mean of the variables in Table 15.1 are shown in Table 15.2. Comparison of the regional sample mean shows that households in the western region (or their head) differ in some points from those in the other regions, which is expected to appear in estimation results of the model. The heads in the western region participate less often in market work and supply shorter hours to the market. Their market participation rate is 52% and 9 The

eastern region includes Beijing, Hebei, Liaoning, Jiangsu, Zhejiang, Shandong, and Guangdong. The central region includes Shanxi, Jilin, Anhui, Jiangxi, Henan, Hubei, and Hunan. The western region includes Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, and Xinjiang.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Agriculture Profitability and Income Disparity

b1363-ch15

305

their average market hours are 511, while similar rate and hours for the eastern region are 60% and 898, respectively. This reflects relatively unfavorable economic situations in the western region, as shown by much lower net income per capita of the village (1,572 Yuan for west and 3,545 Yuan for east) and much farther distance to the nearest transportation terminal (7.0 km for west and 4.3 km for east). Looking at farm production, farm profitability E in the western region is lower by around one point than that in the other regions. This seems to reflect the lower share of irrigated land (0.51, 0.56, and 0.66 for west, center, and east, respectively), smaller values of farm tools (1,011, 1,595, and 1,722 Yuan for west, center, and east, respectively), and higher share of revenue from livestock production (0.38, 0.23, and 0.23 for west, center, and east, respectively). Using the data introduced above, the estimation procedure in the second section is applied to estimate parameters in Eqs. (15.3) and (15.4). The logarithm of E, instead of E itself, is used for this estimation partly because skewness and kurtosis of farm profitability E is large and they seem to show distortion in their distribution.10 We first examine the estimation result for Eq. (15.3), which is presented in Table 15.3. The most important coefficient α1 , which is associated with the head’s incentive M∗ to work in the market, is nearly zero for the eastern and central regions but it is significantly negative for the western region. For the western region, therefore, farm efficiency falls as the head has more incentive to work in the market partly because he tends to make less effort to collect information about farm technology and inputs. The significant coefficient α1 for the western region also implies that an appropriate household model for this region is non-separable. The examination of Table 15.2 above suggests possible reasons for the non-separability: less developed market for labor in the western region due to, for example, limited opportunities for market work and higher transaction costs in participating in the market. Contrarily, insignificant coefficient α1 for the eastern and central regions implies that a separable household model is applicable to these regions when we examine relation between farm profitability and the head’s incentive to work. Zeng and Goodwin (2008) make a similar analysis to this study for 787 farm households in Hebei and Liaoning, which are included in the eastern region of this study, to show that the coefficient α1 is not significant. 10 Skewness

and kurtosis of index E are 3.2 and 19.4 for the western region, 3.4 and 22.6 for the central region, and 3.1 and 16.9 for the eastern region, respectively.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch15

T. Sonoda

306 Table 15.3.

Estimation Result of Eq. (15.3) (Dependent Variable: lnE). West

Number of households

2,148

Center 2,608

East 2,075

M∗ /100

−0.0094 (−2.685)

−0.0004 (−0.120)

0.0001 (0.029)

age fworker

−0.0036 (−2.688)

−0.0008 (−0.493)

−0.0036 (−2.494)

educ

−0.0028 (−0.556)

−0.0004 (−0.086)

−0.0037 (−0.541)

tot land

−0.0055 (−2.174)

0.0012 (0.602)

−0.0079 (−4.122)

fworker share

−0.2369 (−2.710)

0.0311 (0.373)

−0.2539 (−2.711)

fworker male share

0.1136 (0.999)

−0.0351 (−0.280)

0.3016 (2.055)

ag capital/100

0.0001 (0.196)

−0.0013 (−4.527)

−0.0012 (−2.542)

irr share

0.0699 (1.977)

0.0678 (2.251)

−0.1292 (−2.843)

grainquota

0.0084 (0.308)

−0.0532 (−2.011)

0.0931 (2.299)

ecrop share

0.2844 (3.423)

0.3547 (5.113)

0.5924 (9.784)

live share

−0.2694 (−3.586)

−0.5282 (−8.045)

−0.3602 (−4.580)

disaster

−0.0261 (−1.017)

−0.0712 (−3.551)

0.0447 (1.477)

0.1242

0.1593

0.1786

Adjusted R2

Note: Coefficients for province dummies are omitted. Standard errors are computed using bootstrap with 400 replications, which are used to obtain t-values in parentheses.

Other significant coefficients in Table 15.3 may be interpreted as follows.11 Farm profitability rises with revenue share of economic crops, while it falls with revenue share of livestock for all the regions. This result seems to be related to lower farm profitability in the western region because 11 These

coefficients are interpreted conditional on the head’s incentive M∗ to work in the market. If we allow M∗ to change, we need to obtain reduced form for farm profitability E from Eqs. (15.3) and (15.4). Note, however, that most of our discussion in the text can also be applied to the case where M∗ changes because the factor 1/(1−α1 α2 ) is estimated to be around 1 and because the vectors x1 and x2 share only two variables (educ and disaster).

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Agriculture Profitability and Income Disparity

b1363-ch15

307

revenue share of livestock is higher in this region. Also, the value of farm tools, land area, and irrigation share have negative effects on farm profitability for the eastern region. This effect implies that if production elasticities of these inputs are positive, more use of them increases the demand F for variable input more than the supply X of farm commodity.12 The result suggests that farm households in the eastern region have better access to purchasing fertilizers and pesticides and therefore can increase these inputs more easily as they increase farm tools and land areas, but the resulting increase in farm output is less than the increase in fertilizers. Furthermore, compulsory grain quota has a negative effect on farm profitability for the central region, while it has a positive effect for the eastern region. Finally, average age of farm workers and share of farm workers have negative effects for the western and eastern region. The result seems to show that farm profitability decreases as the number of farm workers (including the elderly members) increases because of relatively small land for farming. Now, we examine the estimation result for Eq. (15.4), which is presented in Table 15.4. The most important coefficient α2 , which is associated with farm profitability lnE, is negative for all the regions and it is statistically significant for the central and western regions. The negative coefficient α2 seems to reflect evidence that leisure of the household head is a normal good: When farm efficiency drops and income from farming reduces, the head tries to increase his hours of market work to compensate the income reduction.13 Combining the results in Tables 15.3 and 15.4, we have found the tradeoff between the head’s incentive to work in the market and farm profitability for the western region. On the other hand, we have found no correlation between these variables for the eastern region. Zeng and Goodwin (2008) also found no correlation between these variables in their empirical analysis for Hebei and Liaoning. Other significant coefficients in Table 15.4 may be interpreted as follows.14 We find similar effects of many variables among the three 12 Relation

∂ ln E/∂ ln z = ∂ ln X/∂ ln z − ∂ ln F/∂ ln z (z
= p, q) implies that ∂ ln X/∂ ln z < ∂ ln F/∂ ln z if ∂ ln E/∂ ln z < 0. 13 The negative coefficient of non-labor income also suggests that the head’s leisure is a normal good. 14 Since we estimate equation (15.4) by tobit method, the marginal effect of variable x2j (
= ln E) for all household heads is β2j Φ((α2 ln E + x 2 β2 )/σ2 ), where Φ is the distribution function of a standard normal variable, σ2 is the standard deviation of u2 , and the coefficients are interpreted conditional on index E. Here, we focus only on the sign and statistical significance of the effect of changes in x2j and therefore just look at the coefficient β2j .

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch15

T. Sonoda

308 Table 15.4.

Estimation result of Eq. (15.4) (Dependent Variable: M). West

Number of households

2,148

Center 2,608

East 2,075

lnE

−7.965 (−2.331)

−4.8488 (−2.212)

−1.2507 (−0.698)

age

0.4804 (1.882)

0.2900 (1.019)

0.5590 (1.669)

−0.0083 (−2.998)

−0.0065 (−2.087)

−0.0082 (−2.442)

educ

0.4385 (3.583)

0.6515 (5.489)

0.8534 (5.939)

childlt6

0.7750 (1.116)

1.0672 (1.356)

1.5730 (1.249)

childge6

1.3979 (3.470)

2.0310 (4.860)

2.8490 (4.709)

numhh

−0.8915 (−3.064)

−1.6900 (−5.557)

−2.1137 (−5.673)

nonlaborinc/100

−0.0764 (−3.081)

−0.0186 (−1.092)

−0.0124 (−0.807)

debt asset

−0.0370 (−0.027)

−0.2530 (−0.207)

0.1850 (0.162)

dist councenter

−0.0220 (−1.846)

−0.0812 (−4.330)

−0.0507 (−1.977)

dist terminal

−0.0170 (−0.503)

0.0533 (1.245)

−0.1524 (−2.381)

4.5222 (2.416)

4.0162 (2.419)

2.8663 (2.095)

percap income/100

−0.0360 (−0.689)

0.1361 (2.607)

0.1285 (4.918)

disaster

−0.8309 (−1.122)

−0.4341 (−0.795)

0.7832 (0.938)

11.0324 (45.058)

11.7362 (56.522)

14.3159 (55.940)

age squared

bigcitysuburb

standard error of the error term u2 log of likelihood

−4879.3

−6418.0

−5683.0

Note: Coefficients for province dummies are omitted. Standard errors are computed using bootstrap with 400 replications, which are used to obtain t-values in parentheses.

regions: The variables include the head’s age, his years of schooling, number of children of age between 6 and 15, number of family members, distance from the village to the nearest county seat, and a dummy variable for suburb of large or middle city. Market hours of the head increase with his age

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Agriculture Profitability and Income Disparity

b1363-ch15

309

until the age of 34, 22, and 29, respectively for the eastern, central, and western regions. The result suggests more opportunities for market work in the eastern region, which can also be found from the strongly significant coefficient of net income per capita of the village for this region. The positive effect of years of schooling implies that the head with higher human capital tends to supply more of his time to the market. The result is consistent with other studies which estimate supply functions of market labor for members of farm households (e.g., Huffman and Lange, 1989). The positive effect of the number of children of age between 6 and 15 shows that the head tries to work more in the market to earn more money income to feed his children. An increase in family members in Eq. (15.4) means an increase in adult members because the number of children is controlled. Therefore, the negative effect of the number of family members is inferred to reflect the resulting increase in potential market workers, which enables the head to decrease his market hours. The positive effect of a dummy variable for large or middle city suburb, as well as the negative effect of distance to the nearest county seat, implies better access to work places, which allows the head to work more in the market.

15.4. Conclusion This study employs an agricultural household model to specify relation between an index of farm profitability and the household head’s incentive to work in the market, and empirically examines this relation using data of rural households in the Chinese Household Income Project 2002. The estimation results show that the head’s incentive to work in the market has a significantly negative effect on farm profitability for the western region, farm profitability has a significantly negative effect on the head’s incentive to work in the market for the central and western region, and no significant relation between the two variables is found for the eastern region. Now, we use our empirical results to draw some implications about how to improve income of farm households in the western region because household income in this region is lower than that in other regions. According to our estimation results, the behavior of farm households in the western region might be appropriately described by a non-separable household model partly because the regional labor market has not well developed and employment opportunities for market work are limited. In this case, farm households are likely to behave anomalously in their farm production

June 5, 2012

11:23

310

9in x 6in

Growth and Inequality: An International. . .

T. Sonoda

because their choice of consumption affects their organization of farm production. In an extreme case, they can reduce their supply of farm commodity in response to a rise in the price of farm commodity. In other words, a portion of their supply function of farm commodity can be downward sloping (e.g., Sonoda and Maruyama, 1999). Our estimation results also show that farm profitability is inferred to fall as employment opportunities for market work increase. Therefore, with other things being equal, income from farming could be stagnant or reduced as the regional economy develops in the western region. Given these results, one way to improve income of farm households in the western region is to create more employment opportunities for market work and also offer them opportunities to acquire enough skills for the job so that more of their members can engage in market work. The second result given above suggests that this policy might deteriorate their income from farming. But their income from market work is expected to increase more than income reduction in farming because their farm productivity seems to be low enough, as other empirical studies show (e.g., Chen et al., 2009). Moreover, their response to agricultural policies is expected to be more predictable than before if the policy stated above is sufficiently effective. We can expect more predictable response of farm households because their behavior can be described more appropriately by a separable household model as more of their members work in the market and as restriction on their work hours ceases to be binding. Therefore, the deteriorated income from farming can now be supported by effective agricultural policy such as price support for grains. Thus, combination of appropriate agricultural policies with more opportunities for market work and the related skill is expected to improve income of farm households in the western region.

References Amemiya, T. (1985). Advanced Econometrics. Harvard University Press. Benjamin, D. (1992). “Household Consumption, Labor Markets, and Labor Demand: Testing for Separation in Agricultural Household Models.” Econometrica, 60, 287–322. Brandt, L. and Holtz, C. (2006). “Spatial Price Differences in China: Estimates and Implications.” Economic Development and Cultural Change, 55, 43–86. Chen, Z., Huffman, W.E., and Rozelle, S. (2009). “Farm Technology and Technical Efficiency: Evidence from Four Regions in China.” China Economic Review, 20, 153–161.

b1363-ch15

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Agriculture Profitability and Income Disparity

b1363-ch15

311

De Janvry, A., Fafchamps, M., and Sadoulet, E. (1991). “Peasant Household Behavior with Missing Markets: Some Paradoxes Explained.” Economic Journal, 101, 1400–1417. Goodwin, B.K. and Mishra, A.K. (2004). “Farming Efficiency and the Determinants of Multiple Job Holding by Farm Operators.” American Journal of Agricultural Economics, 86, 722–729. Gustafsson, B.A., Li, S. and Sicular, T. (2008). Inequality and Public Policy in China. Cambridge University Press. Huffman, W.E. and Lange, M.D. (1989). “Off-Farm Work Decisions of Husbands and Wives: Joint Decision Making.” Review of Economics and Statistics, 71, 471–480. Jacoby, H. (1993). “Shadow Wages and Peasant Family Labour Supply: An Econometric Application to the Peruvian Sierra.” Review of Economic Studies, 60, 903–922. Li, S. (2002). Chinese Household Income Project, 2002 [Computer file]. ICPSR21741-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2009-08-14. doi:10.3886/ICPSR21741. Nelson, F.D. and Olson, L. (1978). “Specification and Estimation of a Simultaneous Equation Model with Limited Dependent Variables.” International Economic Review, 19, 695–709. Singh, I.J., Squire, L. and Strauss, J. (1986). Agricultural Household Models: Extensions, Applications and Policy. Johns Hopkins University Press. Sonoda, T. and Maruyama, Y. (1999). “Effects of the Internal Wage on Output Supply: A Structural Estimation for Japanese Rice Farmers.” American Journal of Agricultural Economics, 81, 131–143. Zeng, T. and Goodwin, B.K. (2008). Chinese Agricultural Household: Farming Efficiency and Off-Farm Labor supply. VDM Verlag Dr. M¨ uller Aktiengesellschaft & Co. KG.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch15

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Chapter 16 LABOR MIGRATION AND INCOME INEQUALITY HAS LABOR MIGRATION REALLY NOT NARROWED THE RURAL–URBAN INCOME GAP?

Fang Cai∗ and Meiyan Wang†

16.1. Introduction It is widely observed that since the economic reform and opening-up policy initiated in the late 1970s, while China has achieved unprecedented fast economic growth, its income distribution has been deteriorated (for example, Quah, 2002). No matter what measurements are employed by researchers, they all indicate that inequalities in both income and consumption among rural and urban residents have increased in most of the time of reform period, and the myth that stated relatively smaller income inequality in China by international standard has been proven mistaken (Gibson et al., 2001). The nationwide Gini coefficient enlarged from 0.31 in 1981 to 0.447 in 2001 (Martin and Chen, 2004) and/or 0.46 in 2002 (Khan and Riskin, 2004). This degree of income inequality is indeed very high, no matter comparing to China’s own past or to international general level (CDRF, 2005). In a more detailed examination (CDRF, 2005), Gini coefficient of per capita income among rural residents increased from 0.21 in 1978 to 0.34 in 1997 and 0.38 in 2002, while Gini coefficient of per capita income among urban residents increased from 0.16 in 1978 to 0.29 in 1997 and 0.34 in 2002. In each of the periods, overall Gini coefficient of income — namely 0.30 in 1978, 0.38 in 1988 and 0.45 in 2002, were bigger than that in ∗Professor

at Institute of Population and Labor Economics, Chinese Academy of Social Sciences, China. † Associate Professor at Institute of Population and Labor Economics, Chinese Academy of Social Sciences, China. 313

b1363-ch16

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

314 3.5

Real

3.0

Nominal

2.5 2.0 1.5 1.0

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1984

1982

1980

0.0

1978

0.5 1986

11:23

Income ratio (rural=1)

June 5, 2012

Fig. 16.1. Real and Nominal Income Gap between Rural and Urban Areas. Source: NBS (Various Years) China Statistical Yearbook, China Statistics Press.

either rural or urban areas, respectively, implying the existence of large and increasing income gap between rural and urban areas. Direct observation based on officially published statistics in troth indicates such a widening gap (Fig. 16.1). Scholars have been intensively probing into the factors causing the income inequality, especially the income gap between rural and urban areas. Of the abundant researches, quite many focus on chasing down the contributions of various factors to the overall inequality by employing decomposition methods. One of the commonly agreed conclusions is that rural–urban income gap contributes 40% to 60% of overall inequality of per capita income among Chinese residents (Gustafsson and Li, 2001; Kanbur and Zhang, 2004; Wan, 2007). This conclusion implies that one will to a large extent understand China’s inequality if he can explain well the income gap between rural and urban areas, on the one hand, and one will overthrow the conclusion that China’s inequality is increasing over time if he can break the myth of increasing rural urban gap, on the other. One experience can be drawn from economic histories of early-developed countries, which shows that the significant and systematic income gap between rural and urban residents eventually disappeared because of full migration of rural residents to cities. Since China has witnessed the largest flows of rural laborers and their families to urban areas, motivated by seeking higher income and better life, one can hardly explain the fact that the income gap has not even been reduced in a significant manner. Many researchers have indistinctly sensed such a paradox and therefore tried to relate the widening income inequality with the institutional restriction of labor mobility. For example, by combining Lewis model with a price

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

315

scissors model, Knight and Song (1999) depicted the distorted rural urban relationship, which is characterized by large income gap and conditioned by unlimited supply of labor. Yang and Cai (2003) examined the changes of institutions influencing labor mobility and their impacts on rural urban income gap. From the point of view of economics, labor migration means that people move from low productivity agriculture to higher productivity non-agricultural sectors, and through the convergence of marginal productivity of labor between agricultural and non-agricultural sectors, rural and urban income gap can be expected to narrow. Logically, China economists tend to believe that with the elimination of institutional barriers deterring migration, the insistent income gap can be eventually reduced by full labor mobility across regions and sectors. For example, based on a simulation, Whalley and Zhang (2004) suggest that once the obstacles imposed by household registration or hukou system to prevent labor from fully migrating is abolished, all the existing income inequality will disappear. Here, their relatively strong assumption is that hukou system is the single important obstacle of labor migration. Or in other words, with the physical existence of hukou system, migration cannot be actually happen. By building a regional agglomeration model, Hu (2002) concludes that the spatial advantages held by the coastal areas and the concentration of migration within those areas are the major attributes to the regional income gap. However, this finding is based on an incorrect observation about the spatial distribution of migration. The notable labor migration having happened in the reform period in fact is not only for its incomparable scale but also for its regional pattern — namely, the float of migrants from central and western regions to eastern regions (Cai and Wang, 2003). Lin et al. (2004) estimate a response elasticity of migrants to income gap and conclude that migration is indeed a mechanism of narrowing the income gap. In the mean time, they also suggest that the present scale of migration is not adequate to reduce the existing income gap because of the presence of hukou system and fast growth in coastal areas. Again, all above empirical results cannot be consistent with theoretical logic. While the previous studies enhance our understanding of the relationship between migration and income inequality, the paradox of why mass migration has not narrowed down the income gap between rural and urban areas in China remains unanswered. That challenges researchers and requires further empirical evidence and alternative methods to advance the studies in the topic. This chapter tries to look into the issue from

June 5, 2012

11:23

316

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

alternate angle — that is, we bring up a counterfactual question: What if the observation about increasing income gap is incorrect. The rest of the chapter is organized as follows. Through reviewing the reforms of labor migration policies and their effects on labor market development, Sec. 16.2 finds that if viewing hukou system as a whole bunch of institutional arrangements but not only a piece of paper, one can conclude that the hukou reform has been progressing throughout the reform period, which is notably embodied in the improved conditions under which migrants find a off-farm job, reside in the urban areas, and get access to some of the public services in the destination places. Section 16.3 uses population sampling survey data to incorporate migrant workers and their accompanied family members into comparison of income between rural and urban residents, so that we can get a counterfactual view on income inequality in general and rural–urban income gap in particular, which might be overslaughed in previous studies. Based on the same population sampling survey data, Sec. 16.4 decomposes the inequality index in order to provide further evidence to support the conclusion of overestimation of rural–urban gap. Section 16.5 concludes and draws some policy implications from the chapter’s findings.

16.2. Hukou System Reforms and Labor Mobility One of the unique institutional arrangements in China’s pre-reform period is its adoption of hukou system. The core of this system was to restrict population migration and employment transfers of agricultural labor, to exclude the rural population from urban welfare systems, and rural workers from possible jobs in non-agricultural industries. Lewis (1954) was later to show that a dual economic structure existed in developing countries, formed under conditions similar to the segregated labor market that the registration system and related institutions had led to in China. On the one hand, the expansion of the “price scissors” between agricultural and industrial products caused an income gap between urban and rural sectors; on the other hand, the household registration system also led to urban and rural residents having unequal access to the fruits of industrialization. At the same time it led to tremendous losses of efficiency in resource allocation. Under such an institutional arrangement, agricultural system was to strictly control all factors of production, restrict farmers’ activities to a designated sector and area, and deter sectoral and spatial migration of

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

317

labor force. This induced policy related to labor incentive and allocation seriously damaged technical and allocative efficiencies. Therefore, the introduction of Household Responsibility System (HRS) became the first step of the Chinese economic reform. Initiated in the early 1980s, HRS rapidly spread throughout the country. In early 1980 still only 1.1% of production brigades were practicing the household contract system, reaching 20% by the end of the year. The figure was 100% by the end of 1984, and 97.9% of peasants were by then practicing it. This reform is proven to have substantially improved labor incentives and spurred work efforts in agricultural production. According to a scholarly study, about 46.9% of incremental agricultural output between 1978 and 1984 can be attributed to the national inception of this change (Lin, 1992). With this manifold increase of the work effort, there was a significant decline in the labor time needed for agriculture, and a labor surplus emerged. In order to absorb it, peasant households, which now had operational autonomy, firstly switched their work from being used solely for grain production to other agricultural sectors, and then from cultivation to the overall development of agriculture, forestry, animal husbandry, fisheries and household sideline production, greatly changing the agricultural production structure, and improving labor utilization and income levels. With the improvement of agricultural productivity, however, the capacity of either cultivation or “big agriculture” (including farming, forestry, animal husbandry, sideline production and fisheries) to absorb labor was ultimately limited. However, in the early 1980s, the government did not encourage labor to leave the countryside. Noting the necessity of transferring agricultural labor, and the development potential for small rural industries located in the countryside, the government promoted a mode of agricultural labor transfer of “leaving the land but not the hometown.” i.e., encouraging peasants to shift out of agricultural production to employment in township and village enterprises (TVEs). The labor force employed in commune and brigade enterprises (then analogy to TVEs) numbered 28.27 million in 1978, rising sharply to 69.79 million in 1985. However, in 1985 the rural people who had transferred to jobs in TVEs accounted for only 18.8% of total 370 million rural laborers, and some 300 million workers remained in agriculture. According to the prevailing estimates, about 30–40% of the agricultural workforce — in absolute terms is about 100 million to 150 million — was surplus at the time (Taylor, 1993, Chapter 8). Faced with the surplus labor seeking jobs, the TVEs stagnating, and the acceleration

June 5, 2012

11:23

318

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

of urban reforms in the mid-1980s, peasants began to transfer to large, medium and small cities in search of non-agricultural jobs. Since the 1980s, gradual abolition of institutional obstacles has been the key to increased labor mobility. Observing the narrowing capacity to absorb surplus labor in the rural sectors, the government began in 1983 to allow farmers to engage in long distance transport and marketing of their products beyond local market places. This was the first time that Chinese farmers had obtained legitimate rights to do business outside their hometowns. In 1984, regulations were further relaxed, and farmers were encouraged by the state to work in nearby small towns. A major policy reform took place in 1988, when the central government allowed farmers to work in enterprises or run their own business in cities under the condition of self-sufficient staples.1 In the 1990s, the central government and local governments all took a series of measures, suitably relaxing policies restricting migration, implying a certain degree of reform in the household registration system at each stage of the reform process. In this period, policies towards labor mobility across regions diverged. The divergence can be found first between central and local governments. Concerning employment, income and social security of the country as a whole, central government encouraged labor migration between rural and urban areas and across provincial borders, while attitudes towards migration of local governments differ between sending and receiving places. In relatively poor regions whose proportion of surplus laborers and share of agriculture are high, policies towards labor migration are encouraging and supportive, and the governments take measures to help farmers move out to seek better jobs and pays. The more advanced regions, however, are more concerned with security of local employment, and welcome migrant workers only when local economies need extra labor force. In those regions, policies towards inflows of migrants change cyclically as employment pressures facing local governments change over time — that is, each time when the local unemployment problems become severe, they tend to take measures to supplant migrant workers (Cai, Du, and Wang, 2001). By the 21st century, the decision-making power for timely reform of the household registration system was actually devolved to local, in 1 At the time, rationing system of food and necessities had not been abolished and people without local hukou were not entitled to coupons for buying food and other necessities on the local market.

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

319

particular urban governments. At the time, the policy orientations of migrants-receiving regions diverged as well. Recognizing that while legitimacy of policies on migration is based on hukou system, in reality, institutional factors restricting the flow of labor are not only the hukou system itself. Concerning policy costs and benefits related to provisions of social security, social welfare and other public services behind the hukou system, two kinds of local government have changed their rejective attitude towards migration, because they do not see any expected loss of local welfare. The first kind is those fast-growing cities that have significantly benefited from the ample supply of migrant workers. The second is those whose local budgetary ability is uptight so that the provisions of public services have been too limited to worry about inclusion of migrants. However, in large and super sized cities, where variety of social welfare provisions is still related to and based on whether or not the recipients have local hukou identity, expecting and not willing to see a spill effect of local welfare, the governments are reluctant to adopt substantive reform of hukou system. Even now, the active role played by labor mobility in promoting labor market development is inarguable, even if in those receiving regions. Although there were dissimilar orientations and cyclical swings of policy towards migration, both central and local governments have taken a host of policy measures to release strict restriction of migration in the entire reform period — namely, the hukou system has been gradually abolishing. Such a process can be best seen through the following institutional adjustments. First, beginning from hukou reform in small towns in 2000, each level of governments has implemented various reforms of hukou system. The reform of the hukou system in small towns was characterized by “minimum conditions and comprehensive liberalization.” In more than 20,000 small towns nationwide, the basic condition for registration was reduced just to “having stable sources of life and legal domicile in the city”, any outside individuals or families meeting these conditions could apply to obtain an urban hukou. The governments in many medium-sized cities and even some provincial capitals made major efforts to reform their hukou systems, which was characterized by “abolishing quotas, granting conditional access.” Their approach was to relax the eligibility criteria, substantially lowering the threshold for settling in the city. Whereas, policies towards migrants in super-sized cities like Beijing and Shanghai, which were characterized by “building a high threshold and opening city gates”, gave green light for the introduction of special skills only, while on the other hand raising the entry thresholds for ordinary workers.

June 5, 2012

11:23

320

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

Second, urban welfare reform thus creates an institutional environment for the flow of rural workers to the cities. The urban economic reforms that gradually came into force in the late 1980s, such as the development of the non–state–owned economy that created the great demand for workers, reforms of the food rationing, housing distribution, health care and employment systems, have all reduced the costs to peasants of moving to the city, settling down and finding work. At the same time, due to the intervention of and role played by policy, migrant workers’ wage arrears have decreased significantly, and their working and living conditions improved (Wang, 2009). As the outcome of a series of institutional changes and policy adjustments, the scale of rural labor mobility has grown, forming the “tide migrant workers” which has caught the attention of the world. There has never been a consistent official figure regarding the migration of rural workers, and scholars often roughly estimate on the basis of partial surveys. In what follows, using an induction that has been made before, we may give a broad account from numerical changes in labor mobility beyond township boundary prior to 2000 (Ministry of Agriculture Research Group, 2001). That is, in 1983 there were only 2 million, increasing to 30 million by 1989, and reached 62 million in 1993 and 75.5 million in 2000. After 2000, National Bureau of Statistics annual survey has its own estimates on the numbers of rural migrant laborers, which we show in Table 16.1. First column denotes total amounts of rural migrant workers, from which one can perceive both the fast growth and diminishing rates of growth in recent years. Second column shows the amounts of migrant workers of permanent rural households. Namely, the difference between columns 1 and 2 is the migrants with entire family moving out. Columns 3 through 5 indicate the regional distribution of migrant workers as defined in column 2. The changes in migrants’ distribution among three regions show very similar trends in the period covered. This world’s largest internal migration in terms of scale and scope eventually led to substantial reduction of surplus labor force in rural sectors and has been bringing China into the first Lewis turning point.2

2 See Cai and Wang (2007) for a latest estimation of rural surplus labor force, and Cai (2008) for discussion on coming Lewis turning point. As for “first Lewis turning point”, a point differentiated from the point where marginal productivities of labor in agricultural and modern sectors become equal, it is referred to the period of time, which was no longer characterized by unlimited supply of labor, while the dual economy feature

b1363-ch16

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch16

Labor Migration and Income Inequality Table 16.1.

321

Magnitudes and Regional Distribution of Migrant Workers. Distribution of Migrant Workers from Permanent Rural Households (%)

Years

Migrant Workers (million)

Migrant Workers from Permanent Rural Households (million)

East

Central

West

2000 2001 2002 2003 2004 2005 2006 2007 2008

78.49 83.99 104.70 113.90 118.23 125.78 132.12 136.97 140.41

n.a. n.a. n.a. 89.60 93.53 100.38 105.68 108.75 111.82

n.a. n.a. n.a. 69.9 69.6 70.3 70.1 69.6 71.0

n.a. n.a. n.a. 14.9 14.4 14.4 14.8 15.2 13.2

n.a. n.a. n.a. 15.2 15.7 15.0 14.9 15.0 15.4

Source: (1) Data of 2000–2007 are from Department of Rural Surveys, National Bureau of Statistics, China Yearbook of Rural Household Survey (Various Years), China Statistics Press; (2) Data of 2008 is from website of National Bureau of Statistics.

Closely examining the whole picture of reforms of hukou system and related institutions, especially noticing the thirsty demand for migrant workers in urban sectors, one cannot conclude a stagnation of the reform process. With no such broad sight of China’s economic development, and with a more focus on officialese instead of actuality, Chan and Buckingham (2008) suggest that the cumulative effect of the new round of hukou reform has nothing to do with abolition of the hukou, but makes migration harder than before.

16.3. Insufficient Statistics on Income in Covering Comparison Groups Since it is agreed that labor mobility functions as a reducer of the rural urban income gap and the China’s labor migration has expanded with gradual relaxation of institutional restriction over the past decades, how can then one explain the revealed huge income gap between rural and urban areas today? In their excellent paper, while Srinivasan and Bhagwati (1999) criticized the RHS warriors in cross country regression practice — namely,

still exists — that is, without an increase in wage rate in modern sector, labor supply can be embarrassed (Lewis, 1972).

June 5, 2012

11:23

322

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

those researchers who tend to add new explanatory variables on the right-hand side of the equation to support their own hypothesis and argue with other researchers, they also called for attention that even the lefthand side, the dependent, variable needs to be handled with empirical and conceptual care. In the present case, while struggling for consistent explanation about why rural urban income gap insistently remains large and is even increasing, most researchers have never had suspicion that whether the income gap as dependent variable has really increased all the time. Ravallion and Chen (1999) implies that the fact of economic reform undergoing faster than statistical system reform makes China’s statistical data extremely complicated and requires more sophisticated investigation. In present official statistical programs, exclusion of long-term migrant workers and their accompanied family members is one of such cases and leads to underreport of migrants’ income and exaggeration of overall income inequality. The major attribute that gives rise to overstatement of rural–urban income gap comes from the sampling of rural and urban households in household income survey. When the National Bureau of Statistics (NBS) conducts the household surveys in rural and urban areas, only two types of households are concerned conceptually and sampled practically, the urban residence households and rural residence households, but the rural-to-urban migrant households are to a large extent omitted. On the one hand, migrant households are excluded from the chosen samples of urban households that are required to keep account for daily income and expenditure for the NBS survey program. The simplest reason for doing so is that the migrant households usually do not have stable housing in destination cities and therefore they are not considered as practical survey samples. On the other hand, the sampled households in rural areas cannot fully record their income from remittance. As is explained in the notes on main statistical indicators of China Statistical Yearbook (for example, NBS, 2008, p. 355), first, those households that register as local rural hukou but all family members left their registered places for one year or more are not considered as usual rural households and thus excluded from sampling framework. Second, those family members who left home for six months or more are not considered as usual rural residents, unless they keep close economic relation with the households by sending the majority of income to the household. While the exception as described as “close economic relation” is practically hard to define, the local survey teams of NBS usually follow a thumb rule, namely they view marriage couples as those having

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

323

close economic relation but do not view the relationship of parents and children as those having close economic relation. The definitions on usual rural household and usual rural residents certainly lead to the pretermission of migrant income from rural side. Because the statistically omitted migrants are expanding in size, they as a group of income earners have significant impact on total magnitude and distribution of the Chinese residents’ income. In other words, we now need to compare three groups of population when examining income levels and distribution — namely, urban native population, rural usual population, and rural-to-urban migrant population. With the existing information officially released by NBS, however, we cannot accurately divide total population into such three parts without overlapping. According to a survey conducted by NBS, an overwhelming part of the total migrant workers out of rural areas work in towns and cities (Sheng and Peng, 2006). In 2005, NBS surveyed average disposable income of urban households was 10,493 Yuan and the comparable average net income of rural households was 3,255 Yuan. Based on data from a scholarly survey conducted in the same year,3 average per capita income of rural-to-urban migrant households was 8,368 Yuan, 2.6 times that of rural households and 80% of urban households. Though we are not yet confident to allege the decline in income gap between rural and urban areas, the large amount of migrants and their income can no doubt change income distributive pattern of the country as a whole if it is included in comparison. To see how the income amount and income share of migrant workers and their accompanied families are meaningful to income gap between rural and urban residents, we use 1% population sampling survey conducted by NBS in 2005 to categorize the three groups of population: Urban native residents, rural usual residents, and rural migrants for more than six months,4 and calculate the income level of each group (Table 16.2). The categorization follows following definitions: (1) urban native residents are those who

3 This survey was conducted in 12 cities in 2005 by Institute of Population and Labor Economics, Chinese Academy of Social Sciences. In Shanghai, Shenyang, Wuhan, Xi’an and Fuzhou, 500 urban households and 500 migrant households were interviewed in each city respectively. In Wuxi, Yichang, Benxi, Zhuhai, Shenzhen, Baoji, and Daqing, 400 migrant households were interviewed in each city. Using data of migrant households in 12 cities, we calculated the average income of migrant households. 4 Unique advantage of this round of population sampling survey is its first trial to include question about individual income by asking how much did laborer earn in previous month, or how much his previous year’s income can be divided for previous month.

June 5, 2012

11:23

9in x 6in

b1363-ch16

F. Cai and M. Y. Wang

324 Table 16.2.

Growth and Inequality: An International. . .

Regrouping of Population and Income Distribution (%, Yuan/Month). Share in Total Income

Share in Total Population

Share in Labor Force

Income Per Worker

Income Per Capita

At present price Urban natives Rural usual Rural migrant Average/Total

53.14 38.76 8.10 100

39.67 56.04 4.30 100

35.24 59.84 4.92 100

846.70 363.68 924.64 561.49

408.80 211.07 575.29 305.15

At constant price Urban natives Rural usual Rural migrant Average/Total

53.03 39.29 7.68 100

39.67 56.04 4.30 100

35.24 59.84 4.92 100

794.44 346.60 824.67 527.94

383.56 201.16 513.09 286.92

Source: Authors’ calculation based on micro data of 1% population sampling survey in 2005.

hold an urban hukou and have resided in registered places or migrated between and within cities, (2) rural usual residents are those who hold a rural hukou and have not left home for more than six months, and (3) rural migrants are those who hold rural hukou and live places other than home village for more than six months.5 From the calculated income levels of the three groups listed in Table 16.1, one can see that given the hard work and high labor participation of migrant workers, their monthly income is not only higher than rural usual residents but also higher than urban native residents. Though they have relatively smaller share in total population in the samples, when we take into account this omitted income in the comparison between rural and urban income, it would revise the distribution pattern to certain extent. For example, if we only consider the comparison of income between urban native residents and rural usual residents, which is the present practice adopted by NBS and most of researchers while claiming the increasing income gap, the ratio of urban income to rural income is 1.94 at present price and 1.91 at constant price, respectively. If we incorporate rural migrants’ income into rural income, the ratio of urban to rural income declined to 1.72 at both present and constant price. The inclusion of migrant group generates a 11.86% decrease in rural–urban income gap. 5 Such

a categorization may lead to exclusion of some people. However, due to their small size, the exclusion does not influence the results of this study.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

325

There are at least two reasons showing that the foregoing results still underestimate the actual income level of migrant workers and their families and therefore the revealed bias in rural and urban income statistics is not yet sufficiently modified. First, total number of migrants has dramatically enlarged since 2005. For example, the number of migrant workers for more than one month increased by 14.63 million with a growth rate of 11.6% between 2005 and 2008. Secondly, there are more and more cases of entire family migration so that more households cut off economic relation with their home village and thus they are out of sight of statistics. For example, the difference between total migrants and those who migrated from rural usual households, namely the entire family migration, as shown in Table 16.1, was 28.6 million. 16.4. Decomposing Income Inequality Based on New Data Scholars have tried various methods and data sources to estimate and decompose income inequality between rural and urban areas. There are, however, several shortcomings in the studies aiming to understand the actual income gap. First, many are based on sampling surveys, which are not well representative to the country as a whole due to the limited sample size. Secondly, most of the studies did not well combine together the urban native residents, rural usual residents, and rural migrants, so that they cannot compare income distribution properly. Thirdly, because household income is a very complicated concept in China, the results obtained by different research teams are incomparable. Sicular et al. (2007) recognize the impact of the absence of migrants’ income on income distribution pattern and try to fill up the gap in their decomposition analysis. However, given the limitation of migrants’ information, particularly their size and allocation among regions and sectors, the determination of migrants’ weight in econometric model — or their proportion in urban population, can be problematic. Before we use data of 1% population sampling survey to analyze national income inequality, several points need to be made. First, while this dataset is advantageous in sample coverage and representation, its information on income is not sufficient. Comparing with regular components of income in official statistics, income in the survey is no more than labor income, whereas income from properties, income from transfers and the like are not included. Second, in conducting a comparison of income between rural and urban areas and among provinces, we need to

June 5, 2012

11:23

326

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

deflate the income data with proper price index. By utilizing and updating a set of such price indexes from Brandt and Holz (2006), which provide both price differences of initial and consequent years between rural and urban areas and among provinces, we calculate each group’s income in real term. In this study, we intend to analyze overall income inequality of the country as a whole, decompose income inequality into income inequality within rural and urban areas, respectively, and income inequality between rural and urban areas, and then estimate each component’s contribution to the overall inequality. Theil entropy index is the most frequently used indicator that meets the goal.6 If we use T to denote Theil entropy index, the equation for calculating it is the following:   N yi 1
yi ln T = , N i=1 y y here, y is average income, yi is the income of individual i, and N is summation of total number of persons. Equation for decomposing Theil entropy index is:   
 N N
yi N yi N yi yi ln ln = T = Ny yN Y Y i=1 i=1      
Yj
Yj Yj /Y = Tj + ln Y Y N j /N j here, Y is total income, Yj is income of group j, N is total number of persons, and Nj is the number of persons of group j. The first term of the decomposition equation is intra-group inequality and the second term is inter-group inequality. The 2005 population sampling survey, containing various types of population, people’s residence and status of registration, enables us to calculate and decompose income inequality by different categorization and compare 6 A set of Generalized Entropy Family (GE) is in common use to measure inequality of income distribution. While its value is ranged from zero to infinite, the bigger its value the less equal the distribution is. In calculation, this set of indexes can adopt different parameters, which imply to endue income in different locations of the distribution with different weights. Commonly used values of parameter are zero, one and two. When unity is adopted as the parameter value, the calculated index is called Theil entropy index. One important feature of Theil index is its capability of decomposing the index into intra- and inter-group indexes and their contributions to overall index. For the method of decomposition, see Shorrocks (1984) and Shorrocks and Wan (2005).

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

327

the results between the categorizations. In our particular categorizations, usual residents are referred to those who had resided in places for more than six months when the survey carried out, and therefore long-term migrants are included into the urban group.7 In the case of categorization by hukou, urban native residents are classified as urban residents, rural usual residents and rural-to-urban migrants as rural residents. And the other categorization considers urban natives as urban population and rural usual residents as rural population, and rural-to-urban migrants are missing from this grouping. Based on three categorizations of population, we calculated Theil entropy index both at constant price and present prices, decomposed it into inequality within rural and urban residents — namely, the summation of intra-rural inequality and intra-urban inequality, and inequality between rural and urban residents (Table 16.3). Because with or without price adjustment, the results show similar trends, we will mainly analyze the results based on price–adjusted indexes and their decomposition. Most of the existing studies on income inequality do not contain samples of migrants, thus their analysis is based on incomplete information. This present study tries to avoid such a defect. In order to get a clear cut comparison, we first take a reference that is based on the categorization, which is believed to exclude migrants. In this case, nationwide Theil entropy index calculated in our study is 0.4301 and Theil index of income between rural–urban areas is 0.1035, which contributes 24.06% to overall income inequality. Comparing with the cases with other two categorizations, this practice of excluding migrants from overall income earners generates the largest overall inequality and rural–urban gap, and thus the biggest contribution of rural–urban gap to the overall inequality. When we, however, categorize the population in accordance with hukou identity — namely, treat migrants as rural residents, total Theil index reduces to 0.4191 and Theil index of rural–urban income to 0.0752, and the contribution of rural–urban income gap to overall inequality dramatically declined to 17.94%. The conclusion is that when rural-to-urban migrants are not covered to calculate residents’ income, the income inequality is overestimated. 7 Even though some rural migrants move across town or township borders but within rural areas, they became better off through the mobility, therefore in our case, they share similar features with those migrating to cities.

June 5, 2012

9in x 6in

Theil Entropy Index and its Decompositions.

By Usual Residents

By hukou Residents

Residents Excluding Migrants

0.3405 76.29

0.1058 23.71

0.4466 100

0.3684 82.49

0.0782 17.51

0.4578 100

0.3471 75.82

0.1107 24.18

0.4189 100

0.3208 76.58

0.0980 23.39

0.4191 100

0.3439 82.06

0.0752 17.94

0.4301 100

0.3266 75.94

0.1035 24.06

Source: Authors’ calculation based on 1% population sampling survey in 2005.

Growth and Inequality: An International. . .

0.4463 100

F. Cai and M. Y. Wang

Intra Rural Between Intra Rural Between Intra Rural Between and Urban Rural and and Urban Rural and and Urban Rural and National Areas Urban Areas National Areas Urban Areas National Areas Urban Areas At present prices Theil index Contribution rate At constant price Theil index Contribution rate

11:23

328

Table 16.3.

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

329

16.5. Conclusion and Discussion This chapter questions the commonly believed notion that general inequality and rural–urban income gap have been increasing over time. We take notice of the exclusion of long-term migrants from rural to urban areas from official income statistics and try to find out those omitted people in order to examine their income in comparison with both rural and urban household incomes. Our findings show that it is not only a theoretical hypothesis that labor mobility has important role in narrowing down the rural–urban income gap, but it has reduced such income gap in reality. Although the empirical efforts are still preliminary and insufficient in proving the theoretical expectation, it indicates a new direction for further studies on the subject. Then, what theoretical and policy implications does this alternative analysis and conclusion have? Many critics might believe the authors’ alternate view ingratiating because of its underrating of the income inequality. As Basu (2008, p. 218) put it, those who complain about policies deserve more applause, because some changes and improvements usually come from those complaints. However, exact factual exploration is the precondition for finding a solution for the problem. With a critical attitude towards the presence, though one can stress severity of problems and draw greater attention, he is not necessarily advantageous in finding effective solution solving the problems if he does not obtain the accurate fact about the phenomenon. When people overestimate the income inequality, they tend to more or less forget the following logic. First of all, labor market development spurs expansion of rural and urban employment, notably the labor migration across rural– urban boundaries. Secondly, as more people shifted from unemployment and underemployment to fuller employment through mobility between regions and sectors, which mostly benefit rural laborers and their families, the income gap between rural and urban areas inevitably narrows. Thirdly, since the rural urban gap contributes the major part of overall inequality, the narrowed gap will in turn reduce the country’s income inequality. When we keep the consistence and find out the causality, a logical solution will appear naturally. Otherwise we can do nothing but complain policies, which does not help much since it is not capable of telling the policy– makers how it should farthest create employment opportunities in order to reduce inequality. Furthermore, correct understanding of facts helps determine and modify policy priority. If it is found that institutional barriers deterring labor

June 5, 2012

11:23

330

9in x 6in

Growth and Inequality: An International. . .

F. Cai and M. Y. Wang

mobility are being eliminated, that labor migration has indeed reduced the rural–urban income gap, and that, therefore, the primary distribution of income through labor market is not the major cause leading to inequality, the policy suggestions should be to encourage further relaxation of labor mobility policies and push forward complete urbanization, instead of setting up more regulations baffling the labor migration. Accordingly and more importantly, the government should focus on building social welfare system and through it eliminate the rural–urban gap in provisions of and access to social welfare, instead of trying to redistribute through labor market intervention. There is a potential problem raised by the population reallocation revealed in forgoing part of the chapter. That is, the large flows of most productive population from rural to urban sectors have brought about and will further lead to the relative fall in rural usual households’ income level. Whereas this chapter mainly aims to look into the positive effect of labor migration on reduction of income inequality, the relative status change of the left-behind is also a crucial challenge facing the policymaking. That brings forward pressing task for hukou system reform, with which the permanent migrant laborers and their dependent family members can be gradually transformed to permanent urbanites and into urbanization process be completely accomplished. This reform will balance population structure between rural and urban areas and normalize the mechanism adjusting labor supply and demand among regions and sectors.

References Basu, K. (2008). Comments on Cai Fang’s Paper “How Labor Market Develops in Transition China.” In From Authoritarian Developmentalism to Democratic Developmentalism, Aoki, M. and Wu J. (Eds.). Beijing: China CITIC Press. Brandt, L. and Holz, C.A. (2006). “Spatial Price Differences in China: Estimates and Implications.” Economic Development and Cultural Change, 55, 43–86. Cai, F. and Wang, M. (2008). “A Counterfactual Analysis on Unlimited Surplus Labor in Rural China.” China & World Economy, 16(1), 51–65. Chan, W.K. and Buckingham, W. (2008). “Is China Abolishing the Hukou System?” The China Quarterly, 195(September), 582–606. China Development Research Foundation, (CDRF) (2005). China Human Development Report 2005 Towards Human Development with Equity. Beijing: China Translation and Publishing Corporation. Gibson, J., Huang, J. and Rozelle, S. (2001). “Why is Income Inequality so Low in China Compared to Other Countries? The Effect of Household Survey Method.” Economics Letters, 71, 329–333.

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Labor Migration and Income Inequality

b1363-ch16

331

Gustafsson, B. and Li, S. (2001). A More Unequal China? Aspects of Inequality in the Distribution of Equivalent Income, China’s Retreat from Equality: Income Distribution and Economic Transition, Sharpe, M.E., Riskin, C., Zhaoch, R., Shi, L. (Eds.). Kanbur, R. and Zhang, X.B. (2004). Fifty Years of Regional Inequality in China: A Journey Through Central Planning, Reform, and Openness, United Nations University WIDER Discussion Paper No. 2004/50. Knight, J. and Song, L. (1999). “The Rural–Urban Divide, Economic Disparities and Interactions in China.” Oxford, New York: Oxford University Press. Lewis, A. (1972). “Reflections on Unlimited Labour.” In International Economics and Development, Di Marco, L. (Ed.). New York: Academic Press, 75–96. Lin, J.Y.F. (1992). “Rural Reforms and Agricultural Productivity Growth in China.” American Economic Review, 82, 34–51. Lin, J.Y.F., Wang, G.W. and Zhao, Y.H. (2004). “Regional Inequality and Labor Transfers in China.” Economic Development and Cultural Change, 52(3), 587–603. Quah, D. (2002). One Third of the World’s Growth and Inequality, Discussion Paper No. CEPDP0535, Center for Economic Performance, London School of Economics and Political Science. Ravallion, M. and Chen, S.H. (1999). “When Economic Reform Is Faster Than Statistical Reform: Measuring and Explaining Income Inequality in Rural China.” Oxford Bulletin of Economics and Statistics, 61(1), 33–56. Ravallion, M. and Chen, S.H. (2004). China’s (Uneven) Progress Against Poverty, World Bank Policy Research Paper 3408, Development Research Group, World Bank, Washington, DC. Sheng, L.P. and Peng, L.Q. (2006). “Current Situation of Rural Migrant Laborers: Quantity, Components and Individual Characteristics.” In Green Book of Population and Labor: Demographic Transition and Its Social and Economic Consequences, Fang, C. (Ed.). Beijing: Social Sciences Academic Press (China). Shorrocks, A. (1984). “Inequality Decomposition by Population Subgroups.” Econometrica, 52(6), 1369–1385. Shorrocks, A. and Wan, G.H. (2005). “Spatial Decomposition of Inequality.” Journal of Economic Geography, 5(1), 59–81. Sicular, T., Yue, X., Gustafsson, B. and Li, S. (2007). “The Urban–Rural Income Gap and Inequality in China.” Review of Income and Wealth, 53(1), 93–126. Srinivasan, T.N. and Bhagwati, J. (1999). Outward-Orientation and Development: Are Revisionists Right? Economic Growth Center Discussion Papers, No. 806, Yale University. Wan, G.H. (2007). “Understanding Regional Poverty and Inequality Trends in China: Methodological Issues and Empirical Findings.” Review of Income and Wealth, 53(1). Whalley, J. and Zhang, S.M. (2004). Inequality Change in China and (Hukou) Labour Mobility Restrictions, NBER Working Paper 10683, National Bureau of Economic Research.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch16

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 17 INCOME INEQUALITY, LABOR MIGRATION AND THE LEWIS TURNING POINT A COMPARISON OF JAPAN AND CHINA

Ryoshin Minami and Xin Xin Ma∗

17.1. Introduction In 2004 it was reported in China that migrant workers became scarce to push up their wages considerably in coastal urban areas. This event, which is called as Mingong Huang (shortage of migrant workers), caused a debate among economists in and out of this country on whether the Chinese economy has passed the Lewisian turning point (TP) or not1 (Cai, 2007, 2010; Marukawa, 2010; Meng and Bai, 2007; Tajima, 2008; Wang, 2008; Yan, 2008). Passing TP in China, if it is true, is the fourth event in East Asia following Japan in around 1960 (Minami, 1968, 1973, 2002, ch. 9),2 Taiwan at the end of 1960s and Korea at the beginning of 1970s. However it should be a puzzle if China already passed TP or not: The shortage of migrant workers has not been verified and there are no exact studies which show that the phenomenon of Mingong Huang signifies that ∗ This

paper is an abridged and English version of (Minami and Ma, 2009). For details see the original. Ryoshin Minami is a Emeritus Professor at Hitotsubashi University, Japan. Xin Xin Ma is a Research Fellow at the Keio Economic Observatory, Keio University, Japan. 1 Lewisian dual economy model assumes a coexistence of “capitalist sector” and “subsistence sector”. The former is characterized with profit maximizing behavior of capitalists, while in the latter, the marginal productivity of labor (MPL) is smaller than wages, which are determined by the subsistence level (SL) dominant in the society. In general they are represented by urban industries and agriculture respectively. Labor force of subsistence sector is supplied to capitalist sector at constant SL (unlimited supplies of labor). When MPL increases and reaches to SL, profit maximizing behavior starts to operate and labor force of subsistence sector is now available only with increasing wages (limited supplies of labor). This point in time is the “turning point” (TP). 2 For the other demarcation see (Minami, 2002, pp. 213–214). 333

b1363-ch17

June 5, 2012

11:23

334

9in x 6in

Growth and Inequality: An International. . .

R. Minami and X. X. Ma

China has passed the Lewisian TP. As TP is a point in time when “surplus labor” in agriculture, which is an index for “subsistence sector,” disappears completely, demarcation of TP cannot be made without an empirical analysis about employment status in agriculture. In this chapter, we examine a change in the urban labor market by using appropriate statistics, and estimate surplus labor in agriculture by production function approach. In these discussions, contemporary China is compared with Japanese historical experience. In Sec. 17.2, we will survey the recent change in the labor market by using unemployment and wage statistics. In Sec. 17.3, we will estimate agricultural production function and calculate the marginal productivity of labor and the size of surplus labor in order to demarcate TP. In Sec. 17.4, we will reveal factors for a change in agricultural labor. In Sec. 17.5, we summarize conclusions in this paper and refer to their significance to income distribution.

17.2. Recent Change in the Labor Market 17.2.1. Unemployment Rate The most appropriate index to express the balance of demand and supply of labor should be the unemployment rate. It is well known that there are some problems in the unemployment statistics compiled by the National Bureau of Statistics of China: It does not include unemployment of migrant workers and laid-off urban workers, who are in fact in unemployment status. Considering these problems, we estimated a more appropriate series of the unemployment rate (Minami and Xue, forthcoming). Figures from Chinese population censuses in 1990 and 2000 and 1% sample surveys in 1995 and 2005 were used as benchmarks. The unemployment rate, defined as a ratio of unemployed urban labors to the total number of urban labor force, increased considerably from 2.8% of 1985 to 12% of the first half of 2000s. The existence of large unemployed labors in urban China should be one of the counter evidences to the phenomenon of Mingong Huang. On the contrary, Japanese unemployment rate (based on census figures) decreased from 2% in the 1950s to 1% in 1960. As is shown later, labor demand increased considerably and labor market became tight by the rapid growth of urban industries or the high-pitched economic growth in the 1950s and 1960s.

b1363-ch17

9in x 6in

11:23

Growth and Inequality: An International. . .

Income Inequality, Labor Migration and the Lewis Turning Point

b1363-ch17

335

17.2.2. Real Wages and Wage Differentials Here we are arguing on the wage changes in agriculture, considered as a pool of surplus labor. According to the Lewis theory, agricultural wages are determined by SL before TP and by MPL after TP. If SL is constant over time and MPL increases, therefore, agricultural wages should demonstrate a change from a constant to an increase, which demarcates TP. In Fig. 17.1 per capita net income and per capita consumption expenditure of rural households deflated by the consumer price index are drawn as indices for SL in China. They show a steady increase since the 1980s. Growth rates for 1988–2007 are respectively 5.3% and 4.9%. However, it should be noted that a steady increase in SL does not necessarily demonstrate that the economy has passed TP, because SL itself tends to increase due to a development of culture and society even before TP. Therefore we are concerned here about an “acceleration” in the increasing trend of real wages (or the increase of the growth rate), the growth rates of these indices do not show any increase in around 2004, the year of Mingong Huang. Another point which should be studied is a change in the wage differentials between unskilled and skilled workers. Concept of unlimited supplies of labor is only applicable to unskilled workers (supplied mainly from agriculture) and skilled workers in urban industries are limited in supply even in the labor surplus society. As unskilled worker wages tend to increase 10000

Per capita net income of rural households

(Wong, log)

June 5, 2012

1000

100 1986

Per capita consumption of rural households

1988

1990

Fig. 17.1.

1992

1994

1996

1998

2000

2002

2004

2006

2008

Indices for Living Standard of Farmers in China.

Sources: Per capita net income and consumption expenditure of rural households: Statistical Yearbook of China, every issue. Consumer price index (1995 = 100): Statistical Yearbook of China 2008, Table 8.2.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch17

R. Minami and X. X. Ma

336

only after TP, it can be supposed that wage differentials between the two groups of workers may decrease after TP. Figure 17.2B depicts the ratios of SL (here per capita net income of rural households) to the three groups of urban industries, manufacturing, financial, and infra industries (electric, gas, and water supply) in China. Manufacturing is the most typical urban industry, which employs a large number of migrant workers, while the other two industries3 employ white (A) Ratio of Agriculture to Machinery: Japan 0.4

0.35

0.3

0.25

0.2 1945

1950

1955

1960

1965

1970

1975

(B) Ratio of Agriculture to Urban Industries: China 0.4 0.35 0.3

manufacturing

0.25 0.2 0.15

electric, gas and water supply

financial

0.1 0.05 1986

Fig. 17.2.

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

Wage Differentials of Agriculture to Other Industries: Japan and China.

Sources: Japan: Minami 1973, pp. 298, 307. China: Statistical Yearbook of China 2008, Table 4.27 Notes: Japan: Ratio of wages of annual contract agricultural workers (male only) to other industries. China: Ratio of per capita net income of rural households to other industries. 3 They are the industries with the highest wages among 19 industry groups in Statistical Yearbook of China 2008, Table 4.27.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Income Inequality, Labor Migration and the Lewis Turning Point

b1363-ch17

337

color and skilled workers with high wages, but not many migrant workers. Such a difference among these industry groups in employing migrant workers causes a difference in the pattern of change in wage differentials; the ratio to manufacturing decreased only slowly, while the ratios to the other industries demonstrated a rapid decrease since the 1990s even after 2004. They show that there was not a decrease in the wage differentials between skilled and unskilled workers.4 The change of real wages in Japanese agriculture (annual contract workers) shows quite different pattern compared with China. They did not show a significantly increasing trend for the pre-war period. The growth rate was only 1.2% for 1898–1938. For the post-war period it was 4.4% before TP (1954–61), and 7.1% after TP (1961–69) (Minami, 1973, pp. 147–154). In Fig. 17.2A, the ratio of agricultural wages to machinery industry (male only), which is an index for wage differential between unskilled and skilled workers, was almost constant in the 1950s and decreased considerably in the 1960s. This signifies that unskilled workers, supplied mainly from agriculture, became scarce to push up their wages considerably. This is one of the evidences for demarcating TP in around 1960. Such changes in the labor market are not seen in China.

17.3. Estimation of Agricultural Production Function and Surplus Labor 17.3.1. Agricultural Production Function Production function is estimated for primary industry by sub-periods (1990– 95, 1996–2000 and 2001–05).5 Time-series (annual) and cross-sectional statistics (by 31 provinces) are combined together in all estimation. Production function to be estimated is as follows: LnYit /Lit = A + αLnNit /Lit + βLnKit /Lit +




δj Dijt + λt + uit

Here Y, N, K, and L signify gross value added (1995 prices), labor force, gross capital stock (1995 prices) and land area. A is a constant. Suffix 4 Meng and Bai (2007), based on wage statistics in seven factories in Guangdong Province, revealed that there was not a significant increase in the real wages of unskilled workers and criticized the view that China passed TP. However we should be modest in arguing about the whole country based on such micro-statistics. 5 For details see (Minami and Ma, 2009, pp. 6–8).

June 5, 2012

11:23

338

9in x 6in

Growth and Inequality: An International. . .

R. Minami and X. X. Ma

i and j denote 31 provinces and three districts (Eastern, Central and Western)6 respectively. Parameters α and β are production elasticity of labor and capital, and the elasticity of land γ is calculated as 1 − α − β. Dij is a dummy variable for the regions. They are Eastern district dummy (one for Eastern district and zero for other districts, Central district dummy (one for central district and zero for other districts) and Western district dummy (one for Western district and zero for other districts).7 Suffix t denotes year, “t” denotes a time trend (assumed as one for 1990) and λ is its parameter. “u” is an error term. Table 17.1 gives the estimates of α: They are 0.215, 0.259, and 0.379 in the three sub-periods respectively.8 An interesting finding is an increasing trend in α, which is similar to the case of Japan. In Japan, α increased from 0.125 for 1916–30 to 0.254 for 1931–40, and to 0.562 for 1953–66 (Minami 1981, Tables 1 and 5). These results may signify that the same type of technological change of agriculture (labor-using type) was dominant both in pre-war Japan and contemporary China. Also it should be noticed that α is almost equivalent between the two countries. 17.3.2. Agricultural Surplus Labor In Table 17.1 MPL, which is obtained as a product of α and APL, is compared with the two indices for SL respectively in Estimation (1) and (2). The ratio MPL/SL demonstrates an increasing trend in all estimations, which shows a transformation in the agricultural labor market. For instance in Estimation (1), the ratio increased from 35.6% (1990–95) to 39.1% (1996–2000) and 56.6% (2001–05). In Table 17.2 surplus labor is estimated as a difference between total labor force and “equilibrium labor”, which makes MPL = SL. In the last column, the “rate of surplus labor” (a ratio of surplus labors to total labor force) is shown. For instance in Estimation (1), they are 75.7%, 71.5%, and 6 Eastern district includes Liaoning, Beijing, Hebei, Tianjin, Shandong, Jiangsu, Shanghai, Zhejiang, Fujian, Guangdong, and Hainan. Central district includes Heilongjiang, Inner Mongolia, Jilin, Shanxi, Henan, Hubei, Hunan, Anhui, Jiangxi, and Guangxi. Western district includes Xinjiang, Ningxia, Gansu, Shanxi, Tibet, Sichuan, Chongqing (since 1997), Qinghai, Guizhou, and Yunnan. Chongqing City, which was a part of Sichuan Province in 1996, was estimated from the data in 1997 and separated from Sichuan. 7 In estimation, Eastern district is a reference in these three districts dummy variables. 8 Estimation of production function was made also for three districts respectively for 1990–2007. See (Minami and Ma, 2009, Table 1).

b1363-ch17

June 5, 2012 11:23

Production Elasticity of Labor α

Marginal Productivity of Labor MPL = α APL

Subsistence Level SL

MPL/SL (%)

184 342

0.245 0.562

45 192

139 183

32.4 104.9

2,380 2,979 3,486

0.215 0.259 0.379

512 772 1,321

1,438 1,974 2,333

35.6 39.1 56.6

Estimation (1)

Estimation Subsistence Level SL

MPL/SL (%)

1,213 1,497 1,749

42.2 51.6 75.5

Sources: Japan: Minami 1973, p. 200. SL is wages for annual contract agricultural workers. China: APL is calculated as a ratio of GDP to employment of primary industry. GDP is from Statsitical Yearbook of China 2008, Table 2-1, Table 2-5. Employment is our estimates based on census. SL(Estimation 1: per capita net income of rural households, Estimation 2: per capita comsumption expenditure of rural households) is same to Figure 1. Deflator is obtained as a ratio of nominal GDP to real GDP of primary industry. GDP is from Statistical Yearbook of China 2008, Table 2-1. Notes: Japan: 1934–36 prices. China: 1995 prices.

Growth and Inequality: An International. . .

Average Productivity of Labor APL

9in x 6in

Japan (Yen) 1920–1937 1955–1968 China (Yuan) 1990–1995 1996–2000 2001–2005

Comparison of Marginal Productivity of Labor and Wages in Agriculture: Japan and China.

Income Inequality, Labor Migration and the Lewis Turning Point

Table 17.1.

339 b1363-ch17

June 5, 2012 11:23

340

9in x 6in

Table 17.2.

Estimation of Surplus Labor in Agriculture: Japan and China. (10 thousands persons)

Equilibrium Labor Force

Surplus Labor

Estimation (2)

Ratio of Surplus Labor (%)

1,455

662

793

54.5

45,907 45,671 45,803

11,129 13,077 16,112

34,778 32,761 29,691

75.7 71.5 64.8

Sources: Calculated from Table 1. Notes: Equilibirium labor force is the size of employment with MPL = SL.

Equilibrium Labor Force

Surplus Labor

Ratio of Surplus Labor (%)

19,316 24,585 29,913

26,591 21,253 15,890

57.9 46.4 34.6

Growth and Inequality: An International. . .

Japan 1906–1940 China 1990–1995 1996–2000 2001–2005

Estimation (1)

R. Minami and X. X. Ma

Total Labor Force

b1363-ch17

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Income Inequality, Labor Migration and the Lewis Turning Point

b1363-ch17

341

64.8% respectively in the three sub-periods. This demonstrates an existence of a large number of surplus labors in cotemporary China and a decreasing trend of the surplus labors which signifies that the Chinese economy is approaching to TP. The rate of surplus labor for the first sub-period (75.7% and 57.9% respectively in Estimation (1) and (2)) is comparable to the estimates of pre-war Japan by one of the present authors; 54.5% for 1916–40 in the same table.

17.4. Factors for a Change in Agricultural Labor Force 17.4.1. Outflow of Agricultural Labor Force It is MPL (therefore APL) of agricultural labor force that decides the level and its change in surplus labor. APL depends on the level of agricultural technology and the size of employment. This is the reason why we are arguing on the change of agricultural labor force. In China, according to Table 17.3, primary industry labor force increased significantly in the 1970s and 1980s, and has been decreasing since the 1990s. For all periods (1981– 2007) it shows an increasing trend: An increase per year was 1,310 thousand persons.9 In Japan, on the contrary, it showed a slightly decreasing trend in the pre-war period (since the early 1900s up to the end of 1930s) and a rapid decrease in the post-war period; the decrease was 465 thousands persons per year since 1951 through 1960, the year of TP. Such a difference in the pattern of change in the agricultural labor force between the two countries is a main reason for a difference in the change of surplus labor between the two countries. An increase of agricultural labor force is a difference between natural increase and net outflow to non-agriculture. Table 17.3 shows that smaller net outflow than natural increase was a reason for an increase in agricultural labor force in China. In the total period, net outflow per year was 5,510 thousands, which was smaller than the natural increase which was 9 Annual statistics of the number of employment by industry groups in the post-war period is published by the Bureau of Statistics in every issue of Statistical Yearbook of China. Bureau of Statistics also publishes more reliable figures from the population censuses in 1982, 1990, and 2000. One of the present authors estimate and will publish annual series, which is based on census figures and estimated for other years by linking with the annual statistics in the above (Minami and Xue, forthcoming). These figures are utilized in this chapter.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

R. Minami and X. X. Ma

342 Table 17.3.

b1363-ch17

Factors for the Changes in Primary Labor Force: Japan and China. (10 thousands persons) Increase in Primary Labor Force

Total Increase Japan 1901–1940 −4.3 1951–1970 −46.0 China 1981–1990 1,019 1991–2000 −77 2001–2007 −840 (1981–2007) 131

Natural Increase

Rate of Net Net Outflow Total Outflow (%) Increase

10.4 24.7 1,112 435 418 682

Increase in Non-primary Labor Force

14.7 70.7 93 512 1,258 551

0.96 4.93 0.27 1.13 2.83 1.25

24.2 123.4 531 734 1,282 801

Ratio of Net Inflow in Total Increase (%)

60.7 56.8 17.5 69.8 98.1 68.8

Sources: Japan: Minami 1973, p. 106. China: Estimated from the number of labor force by industry groups. The number of labor force up to 2000 is from (Minami and Xue, forthcoming). Figures from 2001 is extrapolated by linking with the figures in Statistical Yearbook of China 2008, Table 4-3. Notes: Annual figures.

6,820 thousands. In Japan, to the contrast, net outflow was much larger than natural increase. A difference of the speed of outflow in two countries is evident in the net outflow rate; 5.16% per year in the 1950s in Japan, while only 1.25% in China for 1981–2007. In China, employment absorption capacity of urban industries was smaller than in Japan. This should be a main reason for a delay in approaching to TP in China. Percentage of non-primary industries in total employment increased since the 1970s, but this change was not significant compared with Japan in the 1950s and 1960s (Minami and Ma, 2009, Fig. 4). 17.4.2. Evaluation of Demographic Factors Here we are arguing on the view, which emphasizes a role of demographic factor in changing labor market. Cai and others emphasize the fact that a decrease in the growth rate of labor supply coming from a decrease in the natural increase of population is contributing to the change of labor market from surplus to shortage in the first half of 2000s or in near future (Cai, 2007; 2010). The growth rate of total labor force in China, an index of labor supply, is smaller than Japan. For instance, it was only 1.2% for

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Income Inequality, Labor Migration and the Lewis Turning Point

b1363-ch17

343

1991–2000 and 0.9% for 2001–07, almost equivalent to pre-war Japan and half of Japan in the 1950s and 1960s (Minami and Ma, 2009, Table 6). It was due to such a rapid increase in labor demand caused by the rapid economic growth that surplus labor disappeared in Japan in around 1960. Although it cannot be denied that slow increase in labor supply is a favorable factor in approaching to and passing TP, Japanese experience shows that we should pay more attention to the labor demand than the labor supply when we argue about the change of the labor market in China. 17.5. Conclusions and Their Significance 17.5.1. Conclusions Most important conclusion of our research is that China has not yet passed the turning point (TP). We estimated agricultural production functions and calculated marginal productivity of labor (MPL) for three sub-periods: 1990–95, 1996–2000 and 2001–05. Firstly, even in the most recent subperiod, MPL is only 56.6% of per capita net income and 75.5% of per capita consumption expenditure of rural household. The rate of surplus labor (labor force whose MPL is smaller than SL in total employment) was 64.8% (Estimation (1)) and 34.6% (Estimation (2)). Secondly, the rate of surplus labor demonstrates a declining trend among three periods, which suggest that China is approaching to TP.10 The most important reason for a delay in approaching to TP was that the outflow of labor force from the rural to the urban was not enough, which disturbed an increase in MPL in agriculture. On the contrary, in Japan for the 1950s, with the rapidly growing labor demand in the urban industries, a large number of agricultural labors were absorbed in these industries, which caused a drastic decline in agricultural labor force. As a result surplus labor disappeared in around 1960. There are two reasons for insufficient outflow of agricultural labor in China. The first is a labor market segmentation between urban and rural by the registration system (Cai et al., 2005; Ma, 2008). The second is a deficient of the demand for labor in urban industries. Policies to promote migration and develop labor intensive industries should be considered. 10 Only

in two provinces (Hainan and Xinjiang) we find negative rate of surplus labor, which shows that TP was already passed. We can say that the significant number of surplus labor is still existent in Chinese agriculture and that TP has not yet been passed, but it cannot be denied that China is on a process to TP. See (Minami and Ma, 2009).

June 5, 2012

11:23

344

9in x 6in

Growth and Inequality: An International. . .

R. Minami and X. X. Ma

17.5.2. Income Distribution in the Transformation of the Labor Market One of the important influences of the change in the labor market is the impact of surplus labor to the relative income share of labor. Surplus labor in the rural areas tends to disturb wage increase in the urban industries and to cause a decline in the relative income share of labor. One of the present authors, who estimated the relative income share of labor in nonprimary industries in Japan, found that it decreased from 70% in 1896 to 46% in 1940 and explained this phenomenon by referring to the availability of agricultural surplus labor. In the postwar period, the relative share of labor continued a declining trend for some years, but turned to be stable since the beginning of 1960s (Minami and Ono 1978). A study on China, on the other hand, revealed a declining trend in the relative share of labor in the urban industries (Marukawa 2002, pp. 173–180), which may signify that China has not yet passed TP. Much more elaborate study is needed. Another influence of the change in the labor market comes from the relation between surplus labor and income distribution. Aggregate income distribution becomes more unequal when surplus labor is existent, because, as was pointed out in the above, wage increase of unskilled workers in low productivity sector with surplus labor tends to lag behind the wage increase of skilled workers in other sectors. This signifies a possibility of a coincidence of the two turning points: The Lewisian TP and the Kuznets TP (the point when inequality of income distribution turns from increasing to decreasing trend) (Minami, 1998, 2008).11 The coincidence was observed typically in Japan. During the pre-war and the early post-war periods various income differentials widened and consequently income distribution became more unequal. Since around 1960 we can observe both the narrowing of income differentials and the equalization of income distribution, which means that the two TPs were passed in the same time (Minami, 1998, 2008). In China, on the contrary, we observe both the increasing income differentials between the rural and the urban, and the worsening income distribution (Xue, Arayama, and Sonoda 2008). This signifies that China is still in the period before these two TPs. But we should be careful in this issue, because there is a possibility that the 11 Impacts

of passing Lewisian TP to income distribution are not simple, because income distribution in the total economy is composed of three components: Income distribution within rural areas, income distribution in urban areas, and rural–urban income differentials. Careful studies should be needed for the impacts of TP on these components.

b1363-ch17

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Income Inequality, Labor Migration and the Lewis Turning Point

b1363-ch17

345

income inequality problem does not disappear even if the economy passes the Lewisian TP. In many developed countries including Japan, income distribution turned to a worsening trend since the 1980s. This new phenomenon can only be explored in a different theoretical framework. References Cai, F. (Ed.) (2007). Reports on China’s Population and Labor No. 8: The Lewisian Turning Point and the Challenge of Policies. Social Sciences Academic Press (in Chinese). Cai, F. (Ed.) (2008). Reports on China’s Population and Labor No. 9: Linking up Lewis and Kuznets Turning Points. Social Sciences Academic Press (in Chinese). Cai, F. (2010). “Demographic Transition, Demographic Dividend, and Lewis Turning Point in China.” Economic Research Journal, 4, 4–13 (in Chinese). Cai, F., Du, Y. and Wang, M. (2005). “The Segmentation of Labor Market” in Cai, F., Du, Y. and Wang, M. (Eds.). Transition and Development of the Labor Market in China. Commercial Press, 181–204 (in Chinese). Du, Y. (2008). “Wage Level, Wage Differentials and Labor Structure.” in Fang C. (Ed.), Reports on China’s Population and Labor No. 9, 122–137 (in Chinese). Lewis, W.A. (1954). “Economic Development with Unlimited Supplies of Labor.” Manchester School of Economic and Social Studies, 22(2), 139–191. Ma, X.X. (2008). “Rural to Urban Migration and Wage Differentials in Urban China (1) and (2).” Journal of OHARA Institute for Social Research, 591, 39–51, 592, 62–72 (in Japanese). Marukawa, T. (2002). Crustal Movement of the Labor Market in China, Nagoya University Press (in Japanese). Marukawa, T. (2010). “Has the Chinese Economy Pass the Turning Point? Implications of the Survey on Sichuan Province.” Journal of OHARA Institute for Social Research, 616, 1–13 (in Japanese). Meng, X. and Bai, N.S. (2007). “How Much Have the Wages of Unskilled Workers in China Increased?” In China: Linking Markets for Growth, Ross, G. and Song, L. (Eds.). Asia Pacific Press at the Australian National University, 151–175. Minami, R. (1968). “The Turning Point in the Japanese Economy.” Quarterly Journal of Economics, 82(3), 380–402. Minami, R. (1973). The Turning Point in Economic Development: Japan’s Experience, Kinokuniya. Minami, R. (1981). “Long-term Changes in the Output Elasticity of Labor in Agriculture: Estimation and Analysis.” Economic Review, 32(4), 358–366 (in Japanese). Minami, R. (1998). “Eonomic Development and Income Distribution in Japan: An Assessment of the Kuznets Hypothesis.” Cambridge Journal of Economics, 22(1), 39–58.

June 5, 2012

11:23

346

9in x 6in

Growth and Inequality: An International. . .

R. Minami and X. X. Ma

Minami, R. (2002). (Cooperation by Fumio Makino) Economic Development in Japan (3rd.), Toyo Keizai Shinposha (in Japanese). Minami, R. (2008). “Income Distribution of Japan: Historical Perspective and Its Implications.” Japan Labor Review, 5(4), 5–20. Minami, R. and Ma, X.X. (2009). “The Turning Point of Chinese Economy: Compared with Japanese Experience. ” Asian Economies, 50(12), 2–20 (in Japanese). Minami, R. and Ono A. (1978). “Estimation of Factor Income and Factor Shares: Non-primary Industry.” Economic Review, 29(2), 143–169 (in Japanese). Minami, R. and Xue, J.J. (forthcoming). “Labor Force in Postwar Period,” In Long-term Economic Statistics in China, Minami, R. and Makino, F. (Eds.) Toyo Keizai Shinposha (in Japanese). Tajima, T. (2008). “Unlimited Labor Supply and the Lewisian Turning Point.” Monthly Journal of Chinese Affairs, 62(2), 40–45 (in Japanese). Wang, D. (2008). “Lewisian Turning Point: Chinese Experience,” in Fang C. (Ed.), 88–103 (in Chinese). Xue, J.J., Hiroyuki, A. and Sonoda, T. (Eds.) (2008). Inequality in China. Nihon Hyoronsha (in Japanese). Yan, S.P. (2005). Population Mobility and Migrants in China: Quantitative Analysis Based on Micro-data and Macro-data. Keiso Shobo (in Japanese). Yan, S.P. (2008).“Has China Passed the Lewisian Turning Point: Around the Social and Economic Background of Mingong Huang.” East Asia, 30–42 (in Japanese).

b1363-ch17

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Chapter 18 TRADE-OFFS AND COMPLEMENTARITIES BETWEEN GROWTH AND INEQUALITY IN OECD COUNTRIES Peter Hoeller, Isabelle Joumard and Isabell Koske∗

18.1. Introduction In many OECD countries, income inequality has increased in past decades. In some countries, top earners have captured a large share of the overall income gains, while for others income has risen only a little. There is growing consensus that assessments of economic performance should not focus solely on overall income growth, but also take into account income distribution. Some consider poverty as the relevant concern while others are concerned with income inequality more generally. A key question is whether growth-enhancing policy reforms often advocated for OECD countries should have positive or negative side effects on income inequality. More broadly, in pursuing growth and redistribution strategies simultaneously, policy makers need to be aware of possible complementarities or trade-offs between the two objectives. This chapter sheds new light on this issue, following up on recent OECD work (OECD, 2011). It first highlights differences in income inequality across the OECD and the factors driving them, such as cross-country differences in wage and non-wage income inequality, as well as in hours worked ∗ The

authors are members of the Economics Department of the OECD. Peter Hoeller is head of the Public Economics Division; Isabelle Joumard and Isabelle Koske are senior economists. This chapter relies heavily on OECD (2012). Part of this chapter is originally published by the OECD in English under the title: “Reducing Income Inequality While Boosting Economic Growth: Can It be Done?” in OECD, Economic Policy Reform 2012: Going for Growth, OECD Publishing, 2012. 347

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch18

P. Hoeller, I. Joumard and I. Koske

348

and inactivity. The chapter then provides new analysis of the policy and non-policy determinants of overall income inequality, assessing separately the drivers of labor income inequality and the redistributive role of tax and transfer systems. In each case, the analysis identifies “win-win” policies that can both reduce inequality and promote economic growth, and also highlights policies that may entail trade-offs between the two policy goals. 18.2. Understanding Inequality How does one measure income inequality? According to a report by the Stiglitz–Sen–Fitoussi Commission (Stiglitz et al., 2009), the most comprehensive income concept is household disposable income that has been adjusted for publicly-provided in-kind transfers, such as public spending on education and health care. This measure, referred to here as “adjusted household disposable income” is shaped by various factors illustrated in Fig. 18.1. All these factors can vary and shape inequality as follows1 : • Individual labor income. The dispersion of individual labor income amongst the working-age population reflects both the wage dispersion among full-time employees and the labor income dispersion among other Family formation and composition

Income concept

Relevant policy instrument

Fig. 18.1. Income.

Individual labor income

Labor, education, migration and gender policies

Household labor income

Family policies (child and elderly care)

Capital income

Household market income

Tax policies (wealth, capital income)

Taxes & cash transfers

Household disposable income

Cash transfers and taxpolicies

Individual consumption of public goods Household adjusted disposable income

Education, health and housing policies

From Individual Labor Earnings to Adjusted Household Disposable

1 OECD (2011) provides more detail on the five main income concepts shown in Fig. 18.1, and also discusses changes over time.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

• • • •

b1363-ch18

349

groups who make up the working-age population (part-time workers and the self-employed, as well as the unemployed and people not looking actively for a job).2 Household labor income. Working-age families differ in size and composition, affecting the total labor income of households. Household market income. It includes both household labor and capital income.3 Household disposable income. Household disposable income covers all households and income sources, after taxes and cash transfers. Household adjusted disposable income. It adjusts household disposable income for in-kind transfers (e.g. public spending on health, education and social housing).

The rest of this chapter covers three of these five income concepts — household labor income, household market income and household disposable income — since these are the most relevant for the build-up of inequality and the most responsive to structural reforms, while the measurement of the redistributive impact of in-kind benefits is difficult.4 Due to data availability constraints, the chapter focuses on inequality at a given point in time, while the issue should ideally also be looked at from a life-time perspective, taking into account the role of social mobility. 18.2.1. The Dispersion of Household Labor and Market Income Differs Across Countries The dispersion of household labor income is driven by four factors: (i) the dispersion of hourly earnings among those who have a full-time job; (ii) the share of part-time workers; (iii) the non-employment rate and (iv) household formation. Countries differ widely in the dispersion of earnings among full-time workers, with Chile, the United States and Portugal being the most unequal countries and Belgium, Denmark and Switzerland being the most equal ones (Table 18.1). Inequality is higher in all countries when 2 When

examining inequality in individual labor income, the unemployed and people not looking actively for a job are assigned zero income. 3 As the focus of the first three income concepts is on market income, the population covered is the working-age population. 4 About the determinants of inequality for each of the five income concepts refer to a series of OECD Economics Department Working Papers, in particular Hoeller et al. (2012), Koske et al. (2012) and Joumard et al. (2012).

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch18

P. Hoeller, I. Joumard and I. Koske

350 Table 18.1.

Switzerland Belgium Denmark Sweden Finland Slovak Republic Czech Republic Norway France Iceland Slovenia Australia Germany Italy Netherlands Spain Estonia Hungary Austria Japan OECD Luxembourg Ireland Israel Greece Poland Canada Portugal Chile United States

Labor Income Inequality and Population Group. Full-time employed

Full-time and part-time employed

Working age population

0.29 0.29 0.29 0.30 0.30 0.31 0.32 0.32 0.33 0.33 0.33 0.33 0.34 0.34 0.35 0.35 0.35 0.35 0.35 0.36 0.36 0.36 0.37 0.38 0.42 0.42 0.43 0.44 0.45 0.46

0.35 0.33 0.31 0.33 0.33 0.31 0.32 0.34 0.36 0.35 0.34 0.42 0.40 0.36 0.40 0.37 0.36 0.36 0.39 0.42 0.39 0.39 0.43 0.43 0.44 0.44 0.49 0.46 0.47 0.48

0.52 0.55 0.47 0.46 0.50 0.52 0.52 0.48 0.54 0.44 0.51 0.56 0.58 0.57 0.56 0.56 0.51 0.58 0.54 0.62 0.56 0.59 0.64 0.67 0.61 0.63 0.58 0.61 0.73 0.57

Note: Labor income inequality is measured by the Gini coefficient in 2008. The group of employed individuals includes both dependent and self-employed individuals. The working age population includes all persons aged 15 to 64 except for students and people above the country’s statutory retirement age. The Gini coefficients take into account labour income only. Data refer to 2007 for France, Korea and the United States, 2009 for Australia and Japan. The value for the OECD is calculated as an unweighted average across all OECD countries for which data are available. Source: Panel Study of Income Dynamics (PSID) for the United States; Household Income and Labour Dynamics in Australia Survey (HILDA) for Australia; National Socioeconomic Characterization Survey (CASEN) for Chile; Korean Labour and Income Panel Study (KLIPS) for Korea; Luxembourg Income Study (LIS) for Israel; Japan Household Panel Survey (JHPS) for Japan; Swiss Household Panel (SHP) for Switzerland; and European Union Statistics on Income and Living Conditions (EU-SILC) for the other countries. The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD in this and other figures is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

351

extending the analysis to part-time workers or the entire working age population (i.e. also including the unemployed and the inactive), reflecting the large income differentials between these groups and full-time workers. This effect is particularly large for countries where part-time workers make up a sizable share of total employment (e.g. Australia, Germany, Japan, the United Kingdom) and where unemployment and inactivity rates are high (e.g. Belgium, Chile, Hungary, Italy). Accounting for household size and composition reveals a more complex picture (OECD, 2008a). Working household members often combine their income, which narrows the dispersion of income because of the ensuing economies of scale in consumption, whereas the inclusion of dependents in households widens it. Incorporating capital income, which is more concentrated than labor income, increases inequality among households. Even so, given its smaller overall size, capital income is not a major determinant of total household market income dispersion (Table 18.2). Labor market income accounts for around 75% of the dispersion on average in the OECD, versus just 25% for self-employment and capital income combined. OECD-wide, inequality in income after taxes and transfers, as measured by the Gini index, was about 25% lower than for income before taxes and transfers in the late 2000s, while poverty measured after taxes and transfers was 55% lower than before taxes and transfers.5 That said, the distribution of household disposable income still varies widely across countries (Table 18.3). Indeed even after taxes and transfers, the Gini index ranged from below 0.25 in Slovenia (little inequality) to 0.5 in Chile (high inequality). Percentile ratios provide a measure of income inequality at specific points of the income distribution and are an intuitive way to gauge the width of the income distribution. In around 2008, the income of the 90th (i.e. richest) centile of households was three times higher than the income of the 10th (i.e. poorest) centile of households in several Eastern European and Nordic countries. But this ratio stood above six for Chile, Israel, Mexico and Turkey. Also, cross-country differences in the share of top income earners (99th centile) in total income are very wide, ranging from 4.5% for Sweden to 18.1% for the United States.

5 The

poverty rate is defined as the share of the population whose equivalised household disposable income is below 50% of the median income.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch18

P. Hoeller, I. Joumard and I. Koske

352 Table 18.2.

Switzerland Greece Korea Slovak Republic Slovenia Iceland France Austria Norway Czech Republic Sweden Belgium Denmark Spain Luxembourg Estonia Japan Netherlands Finland OECD Poland Germany Hungary New Zealand Canada Italy Australia Portugal Turkey Ireland United Kingdom United States Israel Mexico Chile

Labor Income Inequality by Income Source.

Wages and salaries

Self-employment income

Capital income

0.25 0.17 0.25 0.27 0.28 0.23 0.27 0.27 0.27 0.23 0.29 0.30 0.28 0.32 0.31 0.35 0.31 0.26 0.30 0.29 0.33 0.28 0.27 0.24 0.32 0.18 0.34 0.33 0.23 0.30 0.35 0.37 0.35 0.35 0.27

0.03 0.11 0.06 0.04 0.03 0.00 0.05 0.05 0.03 0.10 0.01 0.03 0.03 0.02 0.04 0.00 0.03 0.06 0.03 0.06 0.04 0.08 0.06 0.07 0.03 0.18 0.02 0.06 0.12 0.09 0.05 0.03 0.05 0.09 0.19

0.02 0.03 0.01 0.00 0.01 0.09 0.01 0.01 0.04 0.01 0.04 0.02 0.04 0.01 0.01 0.01 0.02 0.05 0.03 0.03 0.00 0.03 0.05 0.07 0.04 0.03 0.04 0.01 0.06 0.03 0.03 0.04 0.04 0.02 0.05

Note: Contributions to overall household market income inequality are derived by multiplying the concentration coefficients of each income source by their weight in total market income. The data for Greece, Hungary, Mexico and Turkey are net of taxes. Data for France and Ireland refer to the mid-2000s. Source: OECD Income Distribution and Poverty, OECD Social Expenditure Statistics (Database).

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality Table 18.3.

Denmark Czech Republic Norway Slovenia Slovak Republic Hungary Finland Sweden Austria Iceland Belgium Netherlands France Luxembourg Germany Switzerland Greece Poland New Zealand Canada Estonia OECD Italy Ireland Australia Spain United Kingdom Korea Portugal Japan United States Turkey Israel Chile Mexico

b1363-ch18

353

The Division Between the Rich and the Poor. Centile ratio

Gini index

2.79 2.91 2.98 3.02 3.12 3.15 3.16 3.17 3.20 3.21 3.32 3.34 3.39 3.44 3.54 3.66 4.02 4.05 4.17 4.18 4.28 4.28 4.30 4.41 4.55 4.55 4.56 4.82 4.86 5.02 5.90 6.15 6.20 8.52 9.68

0.24 0.25 0.24 0.23 0.25 0.27 0.25 0.25 0.26 0.29 0.25 0.29 0.27 0.28 0.29 0.30 0.30 0.30 0.32 0.32 0.31 0.31 0.33 0.31 0.33 0.31 0.33 0.31 0.34 0.32 0.37 0.40 0.37 0.48 0.46

Note: Household income inequality in the late 2000s is measured by the gap between the 10th and 90th centile and Gini index. Data for France and Ireland refer to the mid-2000s instead of the late 2000s. Source: OECD Income Distribution and Poverty, OECD Social Expenditure Statistics (database).

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

354

18.2.2. Why are top earners getting a growing share of the cake? Rising income inequality is often shaped by the increasing concentration of income at the top end of the income distribution (Hoeller, 2012). In the United States, for example, the top 1% of the population received 18% of pre-tax income in 2008, up from 8% in 1980. While the share in total income of the top earners has also risen in most other OECD countries (Table 18.4), countries vary considerably both in the extent of this increase and in when it started. Despite a growing interest in the rise in top incomes, there is still substantial disagreement about the causes and their relative importance. Some of the more prominent explanations include the following: Changes in taxation • Tax rates for high earners have come down considerably over time — this may have boosted the income that top earners declare to the tax Table 18.4.

Share of the Top 1% of Earners in Total Taxable Income (1980–2008).

Sweden Netherlands Denmark Finland Norway Belgium Spain France New Zealand Japan Italy Portugal Ireland Switzerland Germany Australia Canada United Kingdom United States

1980

2008 or latest available year

4.00 5.80 5.20 3.30 4.70 7.00 7.50 7.60 5.60 7.20 6.90 3.60 6.60 8.80 10.80 4.80 8.10 6.70 8.10

4.50 5.60 6.50 7.10 7.50 7.70 8.80 8.90 9.00 9.20 9.40 9.80 10.30 10.50 11.10 11.20 13.30 14.20 18.10

Note: The pre-tax income data exclude capital gains for all countries except Australia and Finland. The data are based on tax returns. Source: Alvaredo, F., et al. (2011). “The Top Incomes Database.” www.parisschoolofeconomics.eu/en/news/the-top-incomes-database-new-website/; Matthews, S. (2011). “Trends in Top Incomes and their Tax Policy Implications.” OECD Taxation Working Papers, No. 4, OECD Publishing.

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

355

authorities. Studies suggest that in a country with a top marginal tax rate of 50%, a cut in the marginal tax rate by 1% would boost taxable income by 1%. • Tax regimes may influence the mix of compensation, tilting it towards lower taxed forms of compensation, particularly at the top (Goolsbee, 2000; Piketty and Saez, 2003; Roine et al., 2009). For example, capital gains are often taxed at a lower rate than other income and in a few countries they are not taxed at all. Stock options also benefit from preferential tax treatment in many OECD countries (OECD, 2006a) and the same is likely to hold for carried interest arrangements. Globalization, technological change and the market for talent • New information technologies, together with globalization, have widened the market for “stars,” boosting top incomes in the sports and entertainment industries (Rosen, 1981; Gordon and Dew-Becker, 2008). • The skill requirements and responsibilities of top managers have become more complex, largely owing to stronger competition associated with deregulation and globalization (e.g. Murphy and Zabojnik, 2004; Garicano and Rossi-Hansberg, 2006; Cu˜ nat and Guadalupe, 2009). Moreover, the stability of top management positions has declined while the outside options of top managers have improved, raising their bargaining power. Outside options which include jobs overseas may explain why the top income shares of some countries influence those of others. For example, the top income share in the United States has been found to have a considerable influence on the share in Canada, while those in the United Kingdom and Australia influence the one in New Zealand (Saez and Veall, 2005; Atkinson and Leigh, 2008). • Globalization has also led to a sharp increase in the market capitalization of large multi-national companies, with the rise in executive pay closely following the rise in company size (Gabaix and Landier, 2008).

18.2.3. Country Profiles Trace the Various Inequality Dimensions Country profiles have been assembled in diamonds (Fig. 18.2). These allow comparing 24 inequality dimensions for each country with the OECD average and identifying how these inequality dimensions map into inequality of household disposable income (HDI). Country profiles for all OECD countries are provided in Hoeller et al. (2012). The country profiles reveal that

June 5, 2012

11:23

356

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

Fig. 18.2. Inequality Indicators for Italy, Sweden, the United Kingdom and the United States. Note: The dotted line represents the OECD average, the solid line represents the country shown. Where the solid line falls inside the OECD average, this implies that inequality is below the OECD average. Inversely, where the solid line is outside of the OECD average, inequality is greater. The indicators are presented in units of standard deviation. Note: Legend for Fig. 18.2. See Appendix.

inequality of household disposable income, whether adjusted for in-kind transfers or not, does not have the same origin. In some countries, wage dispersion among those working is an important factor (e.g. the United States) while in others, the non-employment rate and/or inequality in capital income are driving inequality in HDI (e.g. Italy). The country profiles also show the extent to which tax and transfer systems as well as publicly-provided services (in particular education and health) reduce

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

357

income inequality. Some countries (e.g. Sweden) are characterized by relatively low inequality in household market income (HMI) but still redistribute considerably via large tax and cash transfer systems, which brings inequality in HDI well below the OECD average. The United Kingdom and the United States display a similar inequality in HMI, which is clearly above the OECD average. Still, taxes and cash transfers redistribute more in the United Kingdom than in the United States. 18.2.4. Classifying Countries by their Inequality Patterns Five groups of OECD countries with similar inequality patterns were identified using a cluster analysis (Fig. 18.3).6 The five groups are listed below, starting with those that have the lowest dispersion of household disposable income: (i) A group — which includes four Nordic countries plus Switzerland — is characterized by below-average inequality thanks to little wage dispersion, in particular at the upper end, combined with a high employment rate. However, the share of part-time employment is above average in all these countries (except Sweden), contributing to inequality in labor income. Cash transfers are often universal and household taxes tend to be largely proportional to household income, implying only moderate redistribution through the tax and transfer system. Overall, both the dispersion in disposable income and the poverty rate are well below the OECD average. (ii) In a group of eight European countries (Belgium, the Czech Republic, Estonia, Finland, France, Italy, the Slovak Republic and Slovenia), inequality originating from the labor market is slightly below the OECD average. Wages are little dispersed in international comparison but inequality in labor income is driven by a low employment rate (in particular for Belgium, France, Italy and the Slovak Republic). 6 The

cluster analysis is performed on a set of 12 variables: The Gini index for individual labor earnings for the working age population, the ratio of the 9th to 5th deciles for wage earnings of full-time employees, the ratio of the 5th to 1st deciles for wage earnings of full-time employees, the share of part-time employment in total employment, the non-employment rate, the Gini index for household labor earnings (working age population), the Gini index for household market income for the working age population, the concentration ratio for transfers, the concentration ratio for taxes, the Gini index for household disposable income for the whole population, the income ratio of the 5th to the 1st quintile for household disposable income adjusted for in-kind public services and the poverty rate.

June 5, 2012

Higher inequality in household disposable income

358

Low dispersion in labour income (high employment rate and little wage dispersion). Cash transfers tend to be universal and taxes are not highly progressive.

Average dispersion in labour income (little wage variation but low employment or high part-time rate). Highly concentrated capital and selfemployment income. Cash transfers (largely insurance-based) and taxes are not highly progressive.

Individual labour income is concentrated, reflecting above average dispersion in wages and a low employment or high part-time rate. Taxes and transfers are not highly progressive.

Australia Canada Ireland2 Netherlands New Zealand United Kingdom

Above average wage dispersion coupled with a high part-time rate. Cash transfers are targeted and taxes are progressive.

Chile Israel Mexico Portugal Turkey United States

High concentration of labour, capital and self-employment income. The poverty rate is high.

Fig. 18.3. Country Groups with Similar Patterns of Inequality.1 1. Country groups are derived from a cluster analysis of a set of 12 core income inequality indicators, with standardized values and unsquared Euclidean distance to measure differences between groups. Various alternative scenarios have been run. They suggest that the two groups to the right are very stable. The dividing lines between the three groups to the left are less sharp.

Growth and Inequality: An International. . .

Austria Germany Greece Hungary Japan Korea Luxembourg Poland Spain

9in x 6in

Belgium Czech Republic Estonia Finland France2 Italy Slovak Republic Slovenia

P. Hoeller, I. Joumard and I. Koske

Denmark Iceland Norway Sweden Switzerland

11:23

Lower inequality in household disposable income

2. For France and Ireland, mid-2000s (instead of end-2000s) data have been used for the cluster analysis.

b1363-ch18

Source: Hoeller, P. et al. (2012). “Less Income Inequality and More Growth — Are they Compatible? Part 1: Mapping Income Inequality across the OECD.” OECD Economics Department Working Paper, No. 924, OECD Publishing.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

359

The high concentration of self-employment or capital income brings inequality in household market income close to the OECD average (except for the Slovak Republic and Slovenia). However, the size of tax and cash transfer systems as a share of GDP is large, reducing household disposable income inequality to or below the OECD average. (iii) In a group of seven other continental European countries (Austria, Germany, Greece, Hungary, Luxembourg, Poland and Spain) plus Japan and Korea, inequality originating from the labor market is at or above the OECD average. However, the underlying causes vary. The wage dispersion is wide in all these countries but in Germany it is mainly at the lower end of the wage distribution, while in Hungary and Poland, wage dispersion arises more at the upper end of the income distribution. The employment rate is also low in Greece, Hungary, Korea, Luxembourg, Poland and Spain, while the share of part-time employment is high in Austria and Japan. In some of these countries (in particular Greece and Korea), an important redistribution of labor income occurs within families. Cash transfers tend to have little redistributive impact since they are small in size (Korea) or largely insurance-based and thus not highly progressive (Austria, Germany, Greece, Hungary, Japan, Poland and Spain). Overall, both the dispersion in household disposable income and the poverty rate are close to the OECD average. (iv) Five English-speaking countries (Australia, Canada, Ireland, New Zealand and the United Kingdom) and the Netherlands all have a large share of part-time employment, driving inequality in labor income. On the other hand, the employment rate is above the OECD average in all these countries except Ireland. While small in size (for all countries except the Netherlands), cash transfers tend to be more targeted and taxes more progressive than in the other OECD countries, and therefore have a sizable redistributive impact. Household disposable income inequality is, however, above the OECD average in all these countries except for the Netherlands. (v) Chile, Israel, Mexico, Portugal, Turkey and the United States are characterized by above average inequality originating from the labor market. This reflects a wide wage dispersion coupled with a low employment rate (though here the United States is an exception). Capital and self-employment income also tend to benefit a small group of households. Cash transfers have little redistributive impact because they are small in size and often largely insurance-based. The size of

June 5, 2012

11:23

360

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

tax systems is also small in most of these countries, although some embody more progressivity than on average in the OECD. Overall, both inequality in household disposable income and the poverty rate are well above the OECD average. Income inequality in non-OECD countries is higher than in most OECD countries. Drawing cross-country comparisons is, however, difficult because income distribution data and methods often differ across these and the OECD countries. As an example, income dispersion in India is measured via a consumption survey (Gini coefficient of 0.38) while an income survey is used for South Africa (Gini coefficient of 0.70). Also, poverty in emerging countries is most often measured in absolute terms, rather than in relative terms as in most OECD countries, with the poverty line varying across countries. 18.3. What Drives Inequality? 18.3.1. Technological Change and Globalization Partly Explain Recent Trends in Labor Income Inequality Technological advances may affect labor income inequality as they can benefit higher-skilled workers more than others. For example, to the extent that medium-skilled workers focus on routine tasks that can also be accomplished by computers, technological change will reduce the demand for such workers. The opposite effect can be expected for highly-skilled and low-skilled workers who tend to focus respectively on abstract and manual non-routine tasks, both of which are harder to replace by machines. If the demand shifts are not offset by equal shifts in the composition of labor supply (e.g. by a large enough rise in tertiary education attainment), technological progress may reduce the earnings or employment of mediumskilled workers relative to both the low- and high-skilled ones. Indeed the data point to a polarization of employment by skill level (e.g. Autor et al., 2006, Goos et al., 2009). Globalization may also widen inequality. A first channel through which this may happen is offshoring. The tasks that are relocated from richer to poorer countries are typically not skill intensive from the perspective of the skill-rich country, but they are from the perspective of the skill-poor country. As a result, offshoring makes labor demand more skill intensive in both poorer and richer countries, thus increasing inequality in both groups of countries (Feenstra and Hanson, 1996). Second, if firms differ in their profitability and low-income workers are disproportionately employed

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

361

by low-productivity firms that are battered by import competition, trade may increase labor income inequality by lowering employment or the relative earnings of low-income workers (e.g. Egger and Kreickemeier, 2009; Helpman et al., 2010). The implied positive link between globalization and inequality is supported by a growing body of studies of individual firms, but it is more difficult to establish a robust link at the aggregate level. Globalization and technological change may also reinforce each other, thereby further raising inequality. On the one hand, technology may underpin globalization and on the other, the increased competition that comes with globalization may force firms to innovate. Innovation may raise labor income inequality both temporarily — since R&D is skill intensive (Dinopoulos and Segerstrom, 1999; Neary, 2003) — and permanently, provided it results in skill-biased technological change as discussed above (Acemoglu, 2002). 18.3.2. Labor Income Inequality is Also Influenced by Structural Policies Structural policies in the areas of education, labor and product markets influence labor income inequality by affecting (i) the employment rate and (ii) the dispersion of earnings among those that have a job (see Koske et al., 2012 for a detailed discussion). Policies that foster equity in education lower income inequality by reducing the dispersion of earnings. The same applies to policies that promote upper-secondary or tertiary education, at least in countries with an already high share of upper-secondary or tertiary graduates, respectively, among the working-age population. For many labor market policies, by contrast, the impact is less clear cut as they affect both the dispersion of earnings and the level of employment in sometimes conflicting ways, at least for some types of workers. Examples include increasing the minimum wage relative to the median wage, increasing the level of employment protection and increasing the generosity of unemployment benefits. One labor market reform that stands out as having a positive effect on both employment and earnings equality is lowering the gap between employment protection of temporary and permanent work. The impact of product market liberalization on income inequality is ambiguous. While boosting employment, some types of product market reforms may widen the distribution of earnings.7 7A

rough quantification of the average size of the effects of selected structural policy reforms on the dispersion of earnings is provided in Koske et al. (2012).

June 5, 2012

11:23

362

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

18.3.3. Some Countries Rely Heavily on Taxes and Transfers to Influence Distributional Outcomes Tax and transfer systems play a key role in lowering overall income inequality. Cash transfers — such as pensions, unemployment and child benefits — account for more than three quarters of the overall redistributive impact, and taxes for one quarter. However, there are large differences across the OECD in the size, composition and progressivity of taxes and cash transfers (Joumard et al., 2012). On the transfer side, pensions account for the bulk of total transfers in most but not all countries (Table 18.5). They primarily aim at redistributing income over the lifetime of individuals — those with higher income contribute more but will also receive higher pensions. Thus, pensions often redistribute comparatively less across different individuals. Other transfers are usually more progressive, although how much depends on their design, e.g. the relative portion of flat versus income-related benefits. In most countries, family and housing benefits are either universal or means-tested, thus involving more redistribution across individuals. The redistributive impact of taxes varies less across countries than the large differences in tax-to-GDP ratios would suggest. Indeed some high-tax countries show little progressivity, either because: (i) the tax mix favors consumption taxes and social security contributions over more progressive personal income and wealth and inheritance taxes8 ; (ii) the progressivity of tax schedules is limited (e.g. in the Nordic countries); or (iii) statutory progressivity is weakened by tax expenditures that benefit high-income groups most. 18.3.4. Labor Income Tax Schedules have Become More Progressive but Tax Expenditures Hamper Redistribution Whether the tax and transfer system has become more or less redistributive over time across the OECD is unclear. The progressivity of statutory labor tax schedules (including social security contributions) has increased in the majority of countries since 2000 (Table 18.6). Though there has been a 8 Consumption taxes tend to be regressive because lower-income households consume a larger share of their income. To mitigate this regressive impact, many OECD countries apply reduced rates and exemptions for goods and services that account for a large share of poorer households’ consumption basket. The evidence, however, suggests that such tax reliefs benefit high-income groups most and may thus not be an effective redistributive tool (Dalsgaard, 2000; OECD, 2010a).

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

b1363-ch18

Trade-offs and Complementarities between Growth and Inequality Table 18.5.

Mexico Korea Iceland Chile Turkey Canada Australia United States Estonia Israel Ireland New Zealand Slovak Republic United Kingdom Netherlands Norway Japan Switzerland OECD Czech Republic Luxembourg Denmark Sweden Spain Slovenia Greece Poland Finland Hungary Germany Portugal Belgium Italy France Austria

363

Cash Transfer by Accounts.

Old age

Incapacity

Family

Unemployment

Other social policy areas

1.14 1.44 1.88 4.49 4.92 3.80 3.18 5.27 5.13 4.10 2.76 4.18 5.03 5.22 4.47 4.45 7.54 6.08 6.01 6.69 4.85 5.56 6.64 6.12 8.12 9.95 8.70 7.43 7.81 8.64 9.17 6.95 11.65 10.74 10.32

0.05 0.43 1.52 0.50 0.13 0.90 1.73 1.31 1.66 2.05 1.65 2.47 1.31 2.06 2.58 3.56 0.59 2.24 1.77 2.28 1.87 3.07 3.11 2.33 1.84 0.83 2.35 2.68 2.38 1.35 2.09 1.92 1.68 1.57 2.03

0.32 0.02 1.41 0.37 0.01 0.80 1.80 0.10 1.34 1.02 2.32 2.26 1.40 2.13 0.61 1.36 0.43 0.95 1.19 1.49 2.66 1.49 1.49 0.52 1.29 0.70 0.80 1.49 2.24 1.09 0.71 1.61 0.65 1.33 2.15

0.25 0.19 0.01 0.00 0.56 0.41 0.33 0.09 0.29 0.98 0.23 0.37 0.20 1.14 0.22 0.31 0.62 0.73 0.59 0.85 1.92 0.67 2.12 0.39 0.46 0.31 1.55 0.67 1.38 1.00 3.12 0.44 1.36 0.92

0.77 0.38 0.60 0.66 1.21 0.59 0.26 0.98 0.15 1.40 1.16 0.24 1.27 0.28 1.27 0.49 1.54 0.86 1.20 0.75 2.01 0.55 0.85 2.03 1.88 1.96 2.02 1.12 1.36 2.19 1.82 2.42 2.41 2.08 2.13

Notes: 1. The data shown here exclude private mandatory spending which accounts for an important share of total social spending in some countries (in particular Chile, Germany and Switzerland). In addition, public cash transfers shown here may not fully account for those programmes and services provided, or co-financed, by local governments. Measurement gaps may be high, notably in federal countries such as Canada. 2. Incapacity-related spending covers expenditure on disability pensions and sick leave schemes (occupational injury and other sickness daily allowances). Source: OECD Social Expenditure Statistics (Database).

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

364 Table 18.6.

The Progressivity of Statutory Labor Tax Schedules.

Ireland Belgium Hungary Luxembourg Mexico Finland Spain Australia Iceland France Netherlands Canada Italy Portugal Germany OECD Greece Sweden Norway Czech Republic United States Slovak Republic Denmark Switzerland United Kingdom Turkey New Zealand Korea Japan Poland

2000

2009

0.23 0.19 0.16 0.18 0.17 0.16 0.13 0.17 0.20 0.10 0.10 0.13 0.12 0.13 0.16 0.11 0.10 0.10 0.12 0.06 0.10 0.07 0.11 0.08 0.08 0.01 0.05 0.05 0.05 0.04

0.23 0.22 0.20 0.18 0.17 0.16 0.16 0.16 0.15 0.14 0.14 0.14 0.14 0.13 0.13 0.13 0.13 0.12 0.12 0.12 0.11 0.11 0.10 0.09 0.08 0.07 0.07 0.07 0.05 0.03

Note: The progressive indicator is based on personal income tax schedules for single tax payer without children in 2000 and 2008. Net personal income tax is defined as the sum of personal income tax and employee social security contributions net of standard cash transfers. Standard tax relief measures — including those linked to marital and family status and income level — are accounted for. Non-standard tax relief measures, i.e. those determined by reference to actual expenses incurred (such as the amount of interest paid on loans), are not included. The indicator for net personal tax progressivity is calculated as the difference between the average net personal tax rate at two income levels based on the assumption of a similar income dispersion across OECD countries. This difference is then divided by the difference between the two income levels. Source: OECD (2009), Taxing Wages 2008, OECD Publishing; OECD estimates.

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

365

steep decline in top marginal income tax rates, a number of countries have cut social security contributions and introduced or strengthened in-work tax benefits, targeted at lower incomes, thus increasing the progressivity of labor taxes. By contrast, the use of tax expenditures which often benefit high-income groups most — such as tax breaks for health and child care, tertiary education, owner-occupied housing and retirement savings — has been growing (OECD, 2010b). The taxation of capital income, wealth and inheritance has also been reduced in many countries, which has clearly reduced the redistributive impact of tax systems. Indeed, capital income tends to be increasingly concentrated in the upper income brackets, as do wealth and inheritance (Piketty, 2010; Fredriksen, 2012). Property taxes vary widely across countries. They largely consist of recurrent taxes on immovable property. These taxes, however, often absorb a larger share of the income of poorer households because they are often set as a payment for the benefits of local public services (e.g. waste collection) which do not increase fully in line with income.9 18.4. Policy Trade-offs and Complementarities Between Growth and Income Equality Objectives Despite a vast theoretical literature on the link between inequality and growth, no general consensus has emerged and the empirical evidence is rather inconclusive. A simple scatter plot of inequality and growth also shows no link (Fig. 18.4). Still, specific structural reforms that aim at raising average living standards also influence the distribution of income. The findings of new research on the GDP per capita and inequality effects of various structural reforms suggest that growth-enhancing policies can be divided into three broad categories (OECD, 2012): (i) those that are likely to reduce labor income inequality; (ii) those that are likely to raise it; and (iii) those that seem to have an ambiguous effect.

9 The regressive nature of recurrent taxes on immovable property may partly fade in a lifetime perspective. Indeed, as the elderly are often income-poor but wealth-rich, property taxes based on real estate values absorb a large share of their income. In contrast, working-age households tend to have higher income and lower wealth and property taxes absorb a lower share of their income.

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

b1363-ch18

P. Hoeller, I. Joumard and I. Koske

366

Inequality in household disposable income 0.5 MEX CHL

TUR

0.4 USA

ISR

PRT JPN

GBR

ITA

AUS CAN

DEU

0.3

CHE

FRA

ESP GRC

NLD OECD-33 HUN AUT SWE

EST LUX

KOR SVK

FIN

CZE

IRL

POL

ISL

SVN

BEL NOR

DNK

0.2 0

1

2

3

4

5

6

Growth of real GDP per capita:1994–2009 average

Fig. 18.4. Growth and Inequality in OECD Countries. Note: Inequality in household disposable income is measured by the Gini index. The inequality measures refer to the late 2000s, except for France and Ireland for which they refer to the mid-2000s. Source: OECD Income Distribution and Poverty, OECD Social Expenditure Statistics (Database); OECD Economic Outlook: Statistics and Projections (Database).

18.4.1. Growth-Enhancing Policy Reforms that are Likely to Reduce Income Inequality 18.4.1.1. Improving the Quality and Reach of Education Reforms to increase human capital are important for improving living standards, and are also likely to reduce labor income inequality. New analysis shows that a rise in the share of workers with upper-secondary education is associated with a decline in labor earnings inequality (Fournier and Koske, 2012). Examples of policy initiatives to raise upper-secondary education attainment include inter alia enhanced accountability for schools, better teacher recruitment and training, and special support for pupils at risk of dropping out. Encouraging more students to pursue tertiary studies may have a more ambiguous effect on earning inequality. Such reforms tend to widen income dispersion by increasing the share of high-wage earners (the composition effect). On the other hand, new research suggests that this effect may be more than offset by a decline in the returns to tertiary education relative

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

367

to the returns to lower levels of education (Koske et al., 2012). Tuition fees that make students share at least part of the cost of tertiary education could lower disposable income inequality (as the current financing of education is regressive), provided they are accompanied by flanking measures so that the poor are not excluded from tertiary education.10

18.4.1.2. Promoting Equity in Education Raising social mobility by making educational outcomes less dependent on personal and social circumstances should boost GDP per capita by enhancing entrepreneurship, the overall quality and allocation of human capital and, ultimately, productivity. At the same time, a more equitable distribution of educational opportunities has been shown to result in a more equitable distribution of labor income (e.g. De Gregorio and Lee, 2002). Examples of reforms include postponing early tracking, strengthening links between school and home to help disadvantaged children learn, and providing early childhood care and basic schooling for all. The latter may yield large positive returns over an individual’s entire lifetime, particularly for the most disadvantaged (Chetty et al., 2011; OECD, 2006b).

18.4.1.3. Reducing the Gap Between Employment Protection on Temporary and Permanent Work If employment protection11 is much stricter for regular than for temporary contracts, workers at the margin of the labor market — such as young people — risk getting trapped in a situation where they move between temporary work and unemployment without getting into permanent work. This can have adverse implications for human capital and career progression (OECD, 2004) and, ultimately, income equality and economic growth. New OECD analysis finds that low-income workers on temporary contracts

10 For example, this could be achieved by combining tuition fees with student loans and linking repayment to income. Empirical evidence suggests that any negative effect of tuition fees on participation rates can be fully offset through improvements in the financial support for students (OECD, 2008b; Heller, 1999). 11 Employment protection refers both to regulations concerning hiring (e.g. rules favoring disadvantaged groups, conditions for using temporary or fixed-term contracts, training requirements) and firing (e.g. redundancy procedures, mandatory notification periods and severance payments, special requirements for collective dismissals and short-time work schemes).

June 5, 2012

11:23

368

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

earn less than workers with similar characteristics on permanent contracts (Fournier and Koske, 2012). This is not the case for higher-income workers. More even job protection for temporary and permanent contracts is also likely to reduce the income gap between immigrants and non-immigrants, as previous studies have shown that immigrants suffer disproportionately from contract-related labor market dualism (Causa and Jean, 2007). 18.4.1.4. Increasing Spending on Active Labor Market Policies High social benefits can reduce the incentives for work and employment. Active labor market policies may limit these adverse effects by better matching jobs with skills and enhancing job search support and monitoring. Existing empirical evidence suggests indeed that active labor market policies raise employment (Bassanini and Duval, 2006). This should entail positive effects for both GDP per capita and labor income equality. Programme design is key to reaping such gains (Martin and Grubb, 2001). 18.4.1.5. Promoting the Integration of Immigrants Better integration of immigrants in the labor market can both reduce inequality and raise GDP per capita through higher labor force participation. Targeted policies, such as language courses, and transparent systems of recognizing foreign qualifications should help to close the gap between immigrants and non-immigrants’ labor market performance. 18.4.1.6. Improving Labor Market Outcomes of Women Women tend to take on more caring responsibilities than men, meaning they work fewer hours and thus take home less pay. Arguably, their higher labor supply elasticity should lead women to be taxed at a lower rate than men. Since this is not feasible in practice, policies to improve the availability of formal care for children and the elderly can serve as an alternative solution. Such policies should help to reduce gender differences in working hours and — at least to the extent that hourly wages are little affected — pay, and at the same time improve long-run living standards through higher participation rates. 18.4.1.7. Fighting Discrimination Since at least part of the earnings gap between immigrants and nonimmigrants and between men and women is likely to be due to discrimination

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

369

(Koske et al., 2012), more effective legal rules (e.g. legal action against those who engage in discriminatory practices) could also help. 18.4.1.8. Taxing in a Way that Allows Equitable Growth Taxes do not only affect the distribution of income; they also affect GDP per capita by influencing labor use and productivity, or both (Johansson et al., 2008). Some tax reforms appear to be win-win options — improving growth prospects while narrowing the distribution of income. Many, however, may imply trade-offs between these objectives. The findings discussed in Joumard et al. (2012) and in the literature suggest some policy options that could promote growth and reduce inequality: • Re-assess tax expenditures that benefit mainly high-income groups (e.g. tax relief on mortgage interest). Cutting back such tax expenditures is likely to be beneficial both for long-term GDP per capita, allowing a reduction in marginal tax rates and for a more equitable distribution of income. Lowering tax expenditures would also reduce the complexity of the tax system, and thus tax compliance and collection costs. • Reduce distortions in taxing capital income. Tax relief — such as reduced taxation of capital gains from the sale of a principal or secondary residence — often distorts resource allocation without boosting aggregate savings and growth, and benefits mainly high-income groups. Specific tax relief may also provide tax avoidance instruments for top-income earners. In particular, there is little justification for tax breaks for stock options and carried interest. Raising such taxes would increase equity and allow a growth-enhancing cut in marginal labor income tax rates. 18.4.2. Growth-Enhancing Policy Reforms that are Likely to Raise Income Inequality 18.4.2.1. Increasing the Flexibility of Wage Determination Extending collective wage agreements to firms that are not party to the original settlement may make labor costs too high for some employers. This can hamper productivity through lower competitive pressures from the entry of new firms, and can also reduce employment (Murtin et al., 2012). However, new OECD evidence suggests that unions compress the distribution of labor earnings. To the extent that administrative extensions have

June 5, 2012

11:23

370

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

a similar effect, their overall impact on income inequality is ambiguous, reflecting offsetting effects on employment and the dispersion of labor earnings. 18.4.2.2. Shifting the Tax Mix from Personal and Corporate Income Taxes Towards Real Estate and Consumption Taxes Personal and corporate income taxes, as well as social security contributions, are the most distortive taxes as they have sizable adverse effects on labor use, productivity and capital accumulation. Shifting the tax mix away from such taxes and towards recurrent taxes on immovable property (the least distortive) and consumption taxes should thus raise living standards (Johansson et al., 2008). However, there is likely to be a trade-off with the income distribution objective since personal income taxes are progressive while real estate and consumption taxes are at best neutral in a lifetime perspective and in most cases tend to be regressive. Targeted transfers can reduce the severity of this trade-off. 18.4.3. Growth-Enhancing Policy Reforms that have an Ambiguous Effect on Income Inequality 18.4.3.1. Avoiding Too High and Long-Lasting Unemployment Benefits If unemployment benefits are too high or long-lasting, they risk reducing job-search incentives and raising wages above market-clearing levels. This lowers employment with negative effects on GDP per capita and labor income equality. In the short-run, these adverse income distribution effects are likely to be dominated by the direct inequality-reducing impact of the income support for the unemployed.12 18.4.3.2. Liberalizing Product Markets A wide range of studies illustrate the large beneficial effects of product market liberalization on productivity (e.g. Bourl`es et al., 2010; Conway 12 In addition, the adverse effects on labor income inequality that stem from lower employment may potentially be offset — at least partially — by a more compressed income distribution (if unemployment benefits are progressive or lower-income workers are more likely to receive them).

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

371

et al., 2006), but the impact on labor income inequality is uncertain. Product market liberalization generally raises employment (e.g. Bassanini and Duval, 2006; Griffith et al., 2007), but this inequality-reducing effect could potentially be offset by a wider dispersion of earnings, though the evidence on the latter link is far from conclusive (e.g. Guadalupe, 2007; Koske et al., 2012). 18.4.3.3. Lowering Minimum Labor Costs Minimum wages that are set too high can limit the job market opportunities for young and low-skilled workers. Under such circumstances, lowering relative labor costs may boost the employment of these marginal groups in the labor market (Neumark and Wascher, 2007). Greater employment in turn raises GDP per capita and reduces labor income inequality. However, existing studies, including new OECD analysis (Koske et al., 2012), suggest that a fall in the minimum wage risks widening the dispersion of wages at the bottom of the distribution among those who are already employed, so that the impact on labor income inequality among the working age population is ambiguous. The employment effect of a lower minimum wage is likely to be smaller when the initial level of minimum labor costs is already low, which increases the likelihood that labor income inequality will rise. 18.4.3.4. Moving from Income to Wealth or Inheritance Taxes Shifting taxes from income to wealth or inheritance would raise GDP per capita, since property taxes are among the least distortive taxes. As personal income, wealth and inheritance taxes all tend to be progressive, the distributional impact would depend on the relative progressivity of each tax but may be broadly neutral.

18.5. Conclusions OECD countries can be divided into five groups according to their patterns of inequality. For example, in five English-speaking countries (Australia, Canada, Ireland, New Zealand and the United Kingdom) and the Netherlands, wages are rather dispersed and the share of parttime employment is high, driving inequality in labor earnings above the OECD average. Means-tested public cash transfers and progressive household taxes reduce overall income inequality, but it remains above the

June 5, 2012

11:23

372

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

OECD average. At the other end of the scale, four Nordic countries and Switzerland all have comparatively low labor income inequality because wage dispersion is narrow and employment rates are high. Cash transfers tend to be universal and are thus less redistributive. Income inequality for this group is considerably below the OECD average. This chapter also presents new empirical analysis which shows that although technological change and globalization have played a role in widening the distribution of labor income, the marked cross-country variation is likely due to differences in policies and institutions. This leads to the following conclusions about policies and institutions: • Education policies matter. Policies that increase graduation rates from upper secondary and tertiary education and that also promote equal access to education help reduce inequality. • Well-designed labor market policies and institutions can reduce inequality. A relatively high minimum wage narrows the distribution of labor income, but if set too high it may reduce employment, which dampens its inequality-reducing effect. Institutional arrangements that strengthen trade unions also tend to reduce labor earnings inequality by ensuring a more equal distribution of earnings. Job protection reforms that make permanent and temporary contracts more even in their provisions lower income inequality through smaller wage dispersion and also possibly via higher employment. • Removing product market regulations that stifle competition can reduce labor income inequality by boosting employment. The empirical evidence for the link between product market reform and the dispersion of earnings is rather mixed. • Policies that foster the integration of immigrants and fight all forms of discrimination reduce inequality. • Tax and transfer systems play a key role in lowering overall income inequality. Three quarters of the average reduction in inequality they achieve across the OECD is due to transfers. However, the redistributive impact of cash transfers varies widely across countries, reflecting both the size and progressivity of these transfers. In some countries (e.g. Australia, the United Kingdom to a lesser extent), cash transfers are small in size but highly targeted on those in need. In some others (e.g. France or Germany), large transfers redistribute income mainly over the life-cycle rather than across individuals, and their progressivity is often low. • Of the various types of taxes, the personal income tax tends to be progressive, while social security contributions, consumption taxes and

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

373

real estate taxes tend to be regressive. But progressivity could be strengthened by cutting back tax expenditures that benefit mainly high-income groups (e.g. tax relief on mortgage interest). In addition, removing other tax reliefs such as reduced taxation of capital gains from the sale of a principal or secondary residence, stock options and carried interest — would increase equity and allow a growth-enhancing cut in marginal labor income tax rates. It would also reduce tax avoidance instruments for top-income earners. These findings, combined with past OECD and other work on GDP per capita effects of policies and institutions highlight the existence of both complementarities and trade-offs between reducing inequality and promoting economic growth: • Many policies entail a double dividend as they reduce income inequality while at the same time boosting long-run GDP per capita. Examples include facilitating the accumulation of human capital, making educational potential less dependent on personal and social circumstances, reducing labor market dualism or promoting the integration of immigrants and fostering female labor market participation. Concerning taxation, reducing tax expenditures, for instance for investing in housing, contributes to equity objectives while also allowing a growth-friendly cut in marginal tax rates. • By contrast, several policies may entail a trade-off between reducing income inequality and raising GDP per capita. For instance, administrative extensions of collective wage agreements may reduce wage earnings dispersion among workers, but if they set labor costs at too-high levels for some employers they may harm competition and productivity and possibly reduce employment. Shifting the tax mix to less-distorting taxes — in particular away from labor and corporate income taxes towards consumption and real estate taxes — would improve incentives to work, save and invest, but could undermine equity. Cash transfers targeted to lower incomes can be used to ease this trade-off. • Finally, some policies aimed at boosting GDP per capita have an uncertain impact on income inequality. For instance, avoiding too-high and long-lasting unemployment benefits may raise employment over the long run but also widen the distribution of income among workers, with an ambiguous net effect on inequality. The same holds as regards keeping minimum wages at moderate levels.

June 5, 2012

11:23

374

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

Appendix: Legend for Fig. 18.2 Individual labour earnings (ILE) ILE Gini 18-65 = ILE Gini index for working age population, including wage earners, self-employed, unemployed and non-employed ILE P9/P1 = 9th to 1st decile wage earnings for full-time employees ILE P9/P5 = 9th to 5th decile wage earnings for full-time employees ILE P5/P1 = 5th to 1st decile wage earnings for full-time employees Men/Women = Median wage earnings of men to women Part-time = Ratio of part-time workers to total employment NER = Non-employment rate Household labour earnings (HLE) HLE Gini 18-65 = HLE Gini index for working age population Gini head = Gini index for heads of household Gini spouse = Gini index for spouses HMI = Household market income CC capital = Concentration coefficient for capital income CC self-employed = Concentration coefficient for self-employment income HMI GINI 18-65 = HMI Gini index for working age population HMI Gini all = HMI Gini index for total population

Household disposable income (HDI) CC transfers TR = Concentration coefficient for cash transfers CC taxes = Concentration coefficient for household taxes HDI Gini 18-65 = HDI Gini index for working age population HDI Gini 65+ = HDI Gini index for population aged 66 and over HDI Gini all = HDI Gini index for total population Adjustment to household disposable income for public spending on: Health = Public spending on health, 5th to 1st quintile Education = Public spending on education and early childhood education and care, 5th to 1st quintile HADI = HDI adjusted for in-kind health and education public spending, 5th to 1st quintile Others Poverty rate = Relative poverty rate Gini regional = Gini index for regional GDP

b1363-ch18

June 5, 2012

9in x 6in

11:23

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

375

References Acemoglu, D. (2002). “Technical Change, Inequality, and the Labor Market.” Journal of Economic Literature, l. 40(1). Atkinson, A.B. and Leigh, A. (2008). “Top Incomes in New Zealand 1921–2005: Understanding the Effects of Marginal Tax Rates, Migration Threat, and the Macroeconomy.” Review of Income and Wealth, 54(2), 149–165. Autor, D.H., Katz L.F. and Kearney M.S. (2006). “The Polarization of the US Labour Market.” American Economic Review, 96(2), 189–194. Bassanini, A. and Duval, R. (2006). “Employment Patterns in OECD Countries: Reassessing the Role of Policies and Institutions.” OECD Economics Department Working Paper, No. 486, OECD Publishing. Bourl`es, R. et al. (2010). “Do Product Market Regulations in Upstream Sectors Curb Productivity Growth: Panel Data Evidence for OECD Countries.” OECD Economics Department Working Paper, No. 791, OECD Publishing. Causa, O. and Jean, S. (2007). “Integration of Immigrants in OECD Countries: Do Policies Matter?” OECD Economics Department Working Paper, No. 564, OECD Publishing. Chetty, R. et al. (2011). “How Does Your Kindergarten Classroom Affect Your Earnings? Evidence from Project Star.” Quarterly Journal of Economics, 126(4). Conway, P. et al. (2006). “Regulation, Competition and Productivity Convergence.” OECD Economics Department Working Paper, No. 509, OECD Publishing. Cu˜ nat, V. and Guadalupe M. (2009). “Globalization and the Provision of Incentives inside the Firm: The Effect of Foreign Competition.” Journal of Labor Economics, 27(2), 179–212. Dalsgaard, T. (2000). “The Tax System in Mexico — A Need for Strengthening the Revenue-Raising Capacity.” OECD Economics Department Working Paper, No. 233, OECD Publishing. Dinopoulos, E. and Segerstrom, P. (1999). “A Schumpeterian Model of Protection and Relative Wages.” American Economic Review, 89(3), 450–472. Egger, H. and Kreickemeier, U. (2009). “Firm Heterogeneity and the Labor Market Effects of Trade Liberalization.” International Economic Review, 50(1), 187–216. Feenstra, R.C. and Hanson, G.H. (1996). “Foreign Investment, Outsourcing and Relative Wages.” NBER Working Paper, No. 5121, National Bureau of Economic Research. Fournier, J.-M. and Koske, I. (2012). “Less Income Inequality and More Growth — Are They Compatible? Part 7: The Drivers of Labour Earnings Inequality — An Analysis Based on Conditional and Unconditional Quantile Regressions.” OECD Economics Department Working Paper, No. 930, OECD Publishing. Fredriksen, K. (2012). “Less Income Inequality and More Growth — Are They Compatible? Part 6: The Distribution of Wealth.” OECD Economics Department Working Paper, No. 929, OECD Publishing.

June 5, 2012

11:23

376

9in x 6in

Growth and Inequality: An International. . .

P. Hoeller, I. Joumard and I. Koske

Gabaix, X. and Landier, A. (2008). “Why has CEO Pay Increased So Much?” The Quarterly Journal of Economics, 123(1), 49–100. Garicano, L. and Rossi-Hansberg, E. (2006). “Organization and Inequality in a Knowledge Economy.” Quarterly Journal of Economics, 121(4), 1383–1435. Goolsbee, A. (2000). “Taxes, High-Income Executives, and the Perils of Revenue Estimation in the New Economy.” American Economic Review, 90(2), 271–275. Goos, M., Manning, A. and Salomons, A. (2009). “The Polarization of the European Labor Market.” American Economic Review Papers and Proceedings, 99(2), 58–63. Gordon, R. and Dew-Becker, I. (2008). “Controversies about the Rise of American Inequality: A Survey.” NBER Working Papers, No. 13982, National Bureau of Economic Research. Gregorio, J.De and Lee, J. (2002). “Education and Income Inequality: New Evidence from Cross-Country Data.” Review of Income and Wealth, 48(3), 395–416. Griffith, R., Harrison, R. and Macartney, G. (2007). “Product Market Reforms, Labour Market Institutions and Unemployment.” Economic Journal, 117. Guadalupe, M. (2007). “Product Market Competition, Returns to Skill, and Wage Inequality.” Journal of Labor Economics, 25(3), 439–474. Heller, D.E. (1999). “The Effects of Tuition and State Financial Aid on Public College Enrollment.” Review of Higher Education, 23(1), 65–89. Helpman, E., Itskhoki, O. and Redding, S.J. (2010). “Inequality and Unemployment in a Global Economy.” Econometrica, 78(4), 1239–1283. Hoeller, P. (2012). “Less Income Inequality and More Growth — Are They Compatible? Part 4: Top Incomes.” OECD Economics Department Working Paper, No. 927, OECD Publishing. Hoeller, P. et al. (2012). “Less Income Inequality and More Growth — Are They Compatible? Part 1: Mapping Income Inequality across the OECD.” OECD Economics Department Working Paper, No. 924, OECD Publishing. Johansson, ˚ A. et al. (2008). “Taxation and Economic Growth.” OECD Economics Department Working Paper, No. 620, OECD Publishing. Joumard, I., Pisu, M. and Bloch, D. (2012). “Less Income Inequality and More Growth — Are They Compatible? Part 3: Income Redistribution via Taxes and Transfers across OECD Countries.” OECD Economics Department Working Paper, No. 926, OECD Publishing. Koske, I., Fournier, J.M. and Wanner, I. (2012). “Less Income Inequality and More Growth — Are They Compatible? Part 2: The Distribution of Labour Income.” OECD Economics Department Working Paper, No. 925, OECD Publishing. Martin, J.P. and Grubb, D. (2001). “What Works and for Whom: A Review of OECD Countries’ Experiences with Active Labour Market Policies.” Swedish Economic Policy Review, 8(2), 9–56. Murphy, K.J. and Z´ abojn´ık, J. (2004). “CEO Pay and Appointments: A MarketBased Explanation for Recent Trends.” American Economic Review, 94,(2), 192–196.

b1363-ch18

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

Trade-offs and Complementarities between Growth and Inequality

b1363-ch18

377

Murtin, F., de Serres, A. and Hijzen, A. (2012). “The Ins and Outs of Unemployment: The Role of Labour Market Institutions.” OECD Economics Department Working Paper, OECD Publishing, forthcoming. Neary, J.P. (2003). “Presidential Address: Globalization and Market Structure.” Journal of the European Economic Association, 1(2&3), 245–271. Neumark, D. and Wascher, W. (2007). “Minimum Wages and Employment.” IZA Discussion Papers, No. 2570. OECD (2004). “OECD Employment Outlook 2004.” OECD Publishing. OECD (2006a). “The Taxation of Employee Stock Options.” OECD Tax Policy Studies, No. 11, OECD Publishing. OECD (2006b). “Starting Strong II: Early Childhood Education and Care.” OECD Publishing. OECD (2008a). “Growing Unequal? Income Distribution and Poverty in OECD Countries.” OECD Publishing. OECD (2008b). “Tertiary Education for the Knowledge Society.” Vols. 1 and 2, OECD Publishing. OECD (2010a). “OECD Economic Surveys: Czech Republic 2010.” OECD Publishing. OECD (2010b). “Tax Expenditures in OECD Countries.” OECD Publishing. OECD (2011). “Divided We Stand: Why Inequality Keeps Rising.” OECD Publishing. OECD (2012). “Reducing Income Inequality while Boosting Economic Growth: Can it be Done?” in Going for Growth, 2012 edition. Piketty, T. (2010). “On the Long-Run Evolution of Inheritance: France 1820–2050.” Paris School of Economics Working Paper, May. Piketty, T. and Saez, E. (2003). “The Evolution of Top Incomes: A Historical and International Perspective.” American Economic Review, 96(2), 200–205. Roine, J., Vlachos, J. and Waldenstr¨ om, D. (2009). “The Long-Run Determinants of Inequality: What Can We Learn from Top Income Data?” Journal of Public Economics, 93(7&8), 974–988. Rosen, S. (1981). “The Economics of Superstars.” American Economic Review, 71(5), 845–858. Saez, E. and Veall, M.R. (2005). “The Evolution of High Incomes in Northern America: Lessons from Canadian Evidence.” American Economic Review, 95(3), 831–849. Stiglitz, J., Sen, A. and Fitoussi, J.P. (2009). Report by the Commission on the Measurement of Economic Performance and Social Progress, Commission on the Measurement of Economic Performance and Social Progress.

June 5, 2012

11:23

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-ch18

June 5, 2012

13:15

9in x 6in

Growth and Inequality: An International. . .

INDEX

268, 269, 270, 271, 272, 273, 298, 302, 309, 311 Class Decomposition 92, 96, 98 Class Structure 79–81, 83, 85–88, 92, 98, 105, 106, 108 Cluster analysis 357 Compulsory Education 222, 273 Conservative Party (UK) 142, 146, 147, 148, 152 Cooperatives 106 Cornwall 141, 142, 151, 152–159 Cornwall Council 141, 142, 151, 154–157, 159 Country profiles 355, 356

Academic ability 222, 223, 226 Achievement Test Score Function 225 Adjusted per capita house floor space 286, 287, 288, 294 Agricultural household model 298, 299, 301, 302, 309, 310, 311 Agricultural policy 310 Agricultural production function 334, 337, 343 Agricultural sector 163, 165, 167, 173, 174, 175, 176, 177, 178, 179, 180 Agricultural wages 335, 337 All resident-owned house 289, 291, 292, 294 Angel Coefficient 210, 211, 212, 213, 225, 226 Austerity 148, 159

Demographic structure 26, 28, 33 Deng Xiaoping 3, 4, 181 Dibao 246, 247, 248, 249 Disposable income 25, 28, 41, 42, 43, 44, 45 Distributional statistics 181, 184, 187 Dominant Classes 105 Duflo E. 162, 168, 173

Badan Pusat Statistik (BPS) 164, 179 Bank for International Settlements (BIS) 127 Base Income 55, 62–68, 71–77 Big Society 148, 149, 151

Economic conditions surrounding households 174 Economic Liberalization 79, 102, 106 Economic Surplus 82 Economic Well-being 53, 56, 58, 78 Education 111–115, 117, 118, 120 Education Expenditure 208, 209, 210, 212, 213, 214, 215, 216, 217, 218, 219, 225 Education expenditure inequality 219 Education policies 372

Capital Account Liberalization 125, 135 Capital Controls 121, 122, 124, 136, 137 China Kuznets curve 14, 15, 16 China Urban Labor Survey (CULS) 231, 243, 248 Chinese Household Income Project (CHIP) 4, 13, 114, 255, 256, 257, 258, 259, 261, 262, 263, 266, 267, 379

b1363-Index

June 5, 2012

13:15

9in x 6in

Growth and Inequality: An International. . .

380

Educational aspiration 191, 199, 200 Educational Disparity 255 Educational endowment 255, 269 Educational groups 111, 112, 114 Educational system 164 Egalitarian Growth Model 121, 135 enrollment rate 207, 211, 220 Equality Trust 144, 148 Equilibrium labor 338, 340 Experience 258, 260, 264, 265, 266, 267, 268 Family budget 207, 221, 222 Family Income and Expenditure Survey 207 Farm productivity 310 Farm profitability 298, 299, 300, 301, 302, 305, 306, 307, 309, 310 Final income 25, 41, 43, 44, 49 Financial Crisis 121, 125, 126, 129, 133 Financial Opening 121, 122, 125, 126, 128, 133, 135 Fixed effects logit model 285, 289 Fixed effects panel regression model 284 Foreign Direct Investment (FDI) 124, 128, 134 Foreign workers 194 Gender differences 368 German socio-economic panel 111, 114 Get Rich First Theory 181 Gini (UK) 139, 140, 141, 149 Gini coefficient 12, 13, 14, 15, 16, 69–73, 77, 81, 88, 89, 96, 183, 184, 185, 186, 187, 192, 195, 255, 259, 261, 262, 313 Gini coefficient of housing inequality 280, 282, 294 Globalization 355, 360, 361, 372, 374, 376 Gordon Brown 143, 147, 148 Great Leap Forward Campaign 5, 6 Great Transformation 80, 108

Index

Green Revolution 102 Gross divorce rate 204 Group inequality 111, 113, 116–120 Growth 53–55, 57, 59, 61–69, 71, 73–77 House size (floor space) 279, 291, 292, 294 Household Characteristics 277, 279, 283, 284, 289, 290, 291, 292, 293 Household consumption expenditure 164, 167 Household Education Expenditure 207, 208 Household Income 53, 59, 77 Household Production 53–55, 58, 59, 62–67, 70–78 Household Responsibility System 317 Housing Affordability 142, 155, 156 Housing market 278, 282, 289, 294, 295 Hukou 6, 115–117, 277, 278, 280, 282, 288, 292, 293, 294, 315, 316, 318, 319, 321, 322, 324, 327, 328, 330, 331 Human capital 256, 257, 265, 275 Immigrants 368, 372, 373, 374 Income Class 210, 213, 221 Income determination 255, 256, 265, 272 Income distribution 334, 344, 345, 346 Income from Wealth 54, 55, 61–69, 71, 73, 75–77 Income gap 259, 262, 264, 268, 269, 274, 313, 314, 315, 316, 321, 322, 323, 324, 325, 327, 329, 330, 331 Industrial structure 28, 29, 30, 31, 35, 40, 50, 51 Inequality 21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 47, 48, 49, 50, 51, 53–55, 57–59, 61, 63, 65, 67, 68–78 Informal employment 231, 239, 245

b1363-Index

June 5, 2012

13:15

9in x 6in

Growth and Inequality: An International. . .

Index

Informal sector 232, 233, 239, 240, 243, 246, 252 Informalization 232, 236, 245, 252 Instrumental variable 168, 175 Inter-generational social mobility 207, 226 Inter-State Decomposition 91 Intergenerational mobility 191, 192, 193, 197, 200, 201, 202, 204 Intermediate Class Regime 105 International Monetary Fund (IMF) 127 Irregular Workers 127, 130, 131, 135 Kuznets curve (hypothesis) 4, 14, 15, 16, 17, 182 Kuznets TP 344 Labor force 26, 27, 28, 31, 32, 33, 34, 35, 39, 40 Labor income inequality 348, 350, 352, 360, 361, 365, 366, 370, 371, 372 Labor law 232, 241 Labor market 231, 232, 233, 237, 238, 239, 240, 241, 242, 243, 245, 246, 248, 249, 250, 251, 252, 253, 254, 303, 309, 310, 316, 319, 329, 330 Labor market policies 361, 368, 372 Labor Market Reform 126, 127, 130 Labor migration 313, 315, 316, 318, 321, 329, 330 Labor mobility 231, 232, 314, 315, 316, 318, 319, 320, 321, 329, 330 Labor supply 297, 298, 300, 311 Labour Party (UK) 142, 145, 146, 149, 150, 152, 159 Land Reforms 83, 106, 123 Latin Americanization 135 Layoff 126, 129 Levy Institute Measure of Economic Well-being (LIMEW), 53–54, 55–59, 60, 61–75, 76–77 Lewis Model of Labor Migration 7 Liberal Party (UK) 140, 142, 146, 147, 148, 150, 152

b1363-Index

381

Liberalization in the education industry 191, 202, 204 Low-income group 219, 220 Low-price resident-owned house 289, 290, 292, 294 Managed Openness 124, 136, 137 Manufacturing sector 29, 30, 31, 32, 34, 35, 37, 39, 51 Mao Zedong 4, 10 Marginal productivity of labor (MPL) 333, 334, 335, 339, 343 Market income 23, 25, 28, 37, 40, 41, 42, 43, 44, 45, 48 Market work 297, 298, 299, 300, 301, 303, 304, 305, 307, 309, 310 Market-price resident-owned house 289, 290, 294 Middle Class 64–68 Middle income trap 4 Migrant households 243 Migrant worker, 232, 233, 234, 236, 238, 239, 240, 242, 243, 244, 245, 254, 316, 318, 319, 320, 321, 322, 323, 324, 325, 333, 334, 336, 337 Mincer equation 113, 255 Mincer J. 167 Minimum Living Standard Allowance 246 Ministerial Committee on Low Wage Workers 195, 205 Money Income (MI) 53–55, 57, 60, 71, 75 Mutualism 142, 146, 158 National Economic and Social Development Board (NESDB), Thailand 184, 188 National Sample Survey 79, 80, 84, 86 National Youth Council 199, 205 NBS panel data 285, 288 Neo-liberalism 142, 143 Neoliberal Economic Restructuring 121, 122, 130, 135 Net Government Expenditures 55–58, 62–67, 71–75, 76–77

June 5, 2012

13:15

9in x 6in

Growth and Inequality: An International. . .

382

Non-agricultural sectors 173, 174, 179, 180 Oaxaca–Blinder Decomposition 255, 256 Old housing allocation system 280, 294 Party manifestoes (UK) 147 Party membership 115, 117, 120 Policy trade-offs and complementarities 365 Poor households 163, 165, 166, 178 1% population sampling survey 323, 324, 325, 328 Poverty lines 161, 164, 167 Primary industry 337, 339, 341 Probability of owning various types of house 279, 289 Probit 163, 175 Product market regulation 372, 374 Public Consumption 53, 54, 55–58, 62, 65, 67, 70, 71, 74, 75, 77 Quantile regression 111–115, 120 Quintile, 111, 113, 114, 119, 211, 212 Redistribution 25, 27, 41, 42, 43, 44, 45, 46, 49 Reforms and opening-up policy 3, 5, 16, 18 Regional difference 298, 303 Regional disparity 12, 188 Regression analysis 28, 34 Relative income share of labor 344 Returns to education 111–113, 117, 118, 120 Returns to schooling 275, 276 Rural disparity 6 Rural household 297, 302, 309 Rural usual residents 323, 324, 327 Rural-Urban Decomposition 90 Rural-urban income gap 313, 314, 316, 322, 324, 327, 329, 330

Index

Schooling years 257, 259, 260, 261, 264, 265, 267, 272, 273 Selecting employment 173, 177 Self-employment 115, 117–120 Separability 298, 301, 302 Service sector 29, 31, 32, 34, 35, 37, 38 Singapore 21, 22, 25, 29, 30, 31, 44, 45, 50, 51 Singapore youth survey 197 Skill-biased immigration 193, 202, 203 Skill-biased parental influence 198, 202, 203, 204 Social Expenditure 47, 48, 50 Social mobility 140, 147, 148, 149, 150, 154, 158, 193, 197, 201, 203, 204, 205 Standardized coefficients 116 Stylized facts 231 Surplus labor 334, 335, 336, 337, 338, 339, 340, 341, 343, 344 Susenas 164, 167, 174, 175 Taiwan 21, 22, 25, 29, 49, 50 Tax 25, 41, 42, 43, 44, 45, 46, 50 Tax and transfer systems 348, 356, 357, 362, 372 Taxes 53, 54, 55–56, 58, 62, 71, 73, 74, 75 Theil index 7 Tokyo 207, 221, 222, 223, 224, 225, 226, 227 Top incomes 354, 355, 374, 375, 376 Township and village enterprises (TVEs) 317 Transaction cost 299, 301, 305 Transfers 55, 56–57, 62, 67, 73, 74–75, 76 Turning point (TP) 333, 334, 335, 336, 337, 342, 341, 343, 344, 345 Underlying factors 277, 279, 281, 288, 289, 293 Unemployment Insurance 238, 246, 247, 249

b1363-Index

June 5, 2012

13:15

9in x 6in

Growth and Inequality: An International. . .

383

Index

Unemployment rate 334 Unlimited supplies of labor 333, 335, 345 Unskilled worker 335, 337, 344 Upward mobility 193, 197, 198, 201, 202, 203, 204 Urban disparity 6 Urban native residents 323, 324, 325, 327

b1363-Index

Urban-rural income gap 5, 7, 8, 10 Wage differentials 335, 336, 337, 345 Wage distribution 111–115, 117, 118, 120 Working Class 82–84, 92, 108 Xiagang 238, 246, 247

June 5, 2012

13:15

9in x 6in

Growth and Inequality: An International. . .

This page intentionally left blank

b1363-Index

June 5, 2012

13:17

9in x 6in

Growth and Inequality: An International. . .

ABOUT THE AUTHOR

Dr. Jinjun XUE is a professor in Economics at the Economic Research Center, Graduate School of Economics of Nagoya University, Japan. He has working experiences as a professor at the Economics School of Wuhan University, China; Fulbright senior scholar at the Economic Growth Center, Yale University, U.S.A.; associate professor at the Institute of Economic Research, Hitotsubashi University, Japan; visiting professor at the Department of Economics, Oxford University, U.K. Currently, he is a visiting researcher at the Energy Research Institute of NDRC China; visiting professor at Beijing University of Technology; Xi’an Transportation University. Dr. Xue has many publications in English, Japanese and Chinese including: Low-carbon Economics (by Social Science Academy Press of China (in Chinese), 2011, Minerva Shoubo (in Japanese) and the World Scientific (in English), 2012; Inequality in China (main author and editor) by Nippon Hyoron-Sha (in Japanese), Social Science Academy Press of China (in Chinese), 2008; “Income Determination and Income Discrimination in Shenzhen,” (co-author) Urban Studies, Vol. 48, No. 7, May 2011; “Rethinking the Educational Disparity and Income Disparity,” (in Chinese), China Population Science, No. 2, April 2011; “How High is Urban Unemployment in China,” Journal of Chinese Economy and Business Studies, co-author, Vol. 4-2, 2006; “Unemployment, Poverty and Income Disparity in Urban China,” (co-author), Asian Economic Journal, Vol. 17-4, 2003, etc.

b1363-abt