66229 South Asia Development Matters More and Better Jobs in South Asia More and Better Jobs in South Asia South Asia Development Matters More and Better Jobs in South Asia Reema Nayar, Pablo Gottret, Pradeep Mitra, Gordon Betcherman, Yue Man Lee, Indhira Santos, Mahesh Dahal, and Maheshwor Shrestha © 2012 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org 1 2 3 4 14 13 12 11 This volume is a product of the staff of The World Bank with external contributions. The ï¬? ndings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judg- ment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to the work is given. For permission to reproduce any part of this work for commercial purposes, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Ofï¬?ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN (paper): 978-0-8213-8912-6 ISBN (electronic): 978-0-8213-8913-3 DOI: 10.1596/978-0-8213-8912-6 Library of Congress Cataloging-in-Publication Data More and better jobs in South Asia / World Bank. p. cm.— (South asia development matters) Includes bibliographical references. ISBN 978-0-8213-8912-6 — ISBN 978-0-8213-8913-3 (electronic) 1. Labor market—South Asia. 2. Labor policy—South Asia. 3. Employment—South Asia. 4. South Asia—Economic policy. 5. South Asia—Economic conditions. I. World Bank. HD5812.57.A6M67 2011 331.120954—dc23 Cover photo: Ray Witlin, The World Bank Cover design: Bill Pragluski, Critical Stages, LLC Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii 1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 South Asia’s track record . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Determinants of job quality and the employment challenge . . . . . . . . . . . . . . . . . . . . . . . . 9 Improving an inconducive business environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Improving workers’ skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Reforming labor market institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Creating jobs in confl ict-affected areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Annex 1A Summary statistics on South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . 42 Annex 1B Deï¬? nition of key labor market terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Annex 1C What is a “betterâ€? job, and which jobs are “betterâ€?? . . . . . . . . . . . . . . . . . . . 43 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 2 Growth and Job Quality in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Economic growth in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Sources of future growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 The track record on employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 The urgency of reform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 v vi CONTENTS Annex 2A Methodology for decomposing growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Annex 2B Sources of average annual growth in output per worker . . . . . . . . . . . . . . . . . 70 Annex 2C Shares of agriculture, industry, and services in employment and GDP . . . . . . 72 Annex 2D Methodology and data sources for labor force projections . . . . . . . . . . . . . . . 72 Annex 2E Poverty rates and the number of working poor in South Asia . . . . . . . . . . . . . 75 Annex 2F Analysis of poverty and unemployment in India . . . . . . . . . . . . . . . . . . . . . . . 79 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 3 A Proï¬? le of South Asia at Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Overview of employment and labor force participation in South Asia . . . . . . . . . . . . . . . . 86 The nature of employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Where are the better jobs? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Who holds better jobs? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Annex 3A Deï¬? nitions and criteria used in proï¬? le of South Asia at work . . . . . . . . . . . 117 Annex 3B Regional employment patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 4 What Is Preventing Firms from Creating More and Better Jobs? . . . . . . . . . . . . . . . . . . . 125 Methodological framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Constraints in the urban formal sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Constraints in the rural nonfarm and informal sectors. . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Demand-side policy options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Constraints facing potential ï¬? rm entrants: Business entry regulations . . . . . . . . . . . . . . 153 Annex 4A Business environment constraints in high- and low-income states in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Annex 4B Tax rates as a constraint to ï¬? rms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Annex 4C Constraints facing nonbenchmark ï¬? rms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Annex 4D Access to ï¬? nance as a constraint to ï¬? rms . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Annex 4E Policy options for increasing access to ï¬? nance . . . . . . . . . . . . . . . . . . . . . . . . 162 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 5 Opening the Door to Better Jobs by Improving Education and Skills . . . . . . . . . . . . . . . 171 Education and skills in South Asian labor markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Education and access to better jobs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 The education challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 The next 20 years: Can South Asian countries improve the educational attainment of their labor forces? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Addressing disadvantages before school: The role of early childhood development . . . . . 189 Primary and secondary education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Tertiary education and preemployment training systems . . . . . . . . . . . . . . . . . . . . . . . . . 199 On-the-job training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 Annex 5A Additional tables and ï¬?gures on education and skills . . . . . . . . . . . . . . . . . . 210 Annex 5B Projections of the educational attainment of South Asia’s population and labor force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 6 The Role of Labor Market Regulations, Institutions, and Programs . . . . . . . . . . . . . . . . 229 Labor market institutions, policies, and programs in the formal sector . . . . . . . . . . . . . . 229 CONTENTS vii Labor market institutions, policies, and programs in the informal sector . . . . . . . . . . . . 246 Annex 6A Additional tables and ï¬?gures on labor market regulations and institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 7 Creating Jobs in Confl ict-Affected Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 Characteristics and intensity of armed confl ict in South Asia . . . . . . . . . . . . . . . . . . . . . . 282 Constraints to job creation in conflict-affected areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 Armed confl ict and labor markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Facilitating private sector job creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 Education service delivery in confl ict situations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 Labor market policies and programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 A jobs transition path in confl ict zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Annex 7A Deï¬? nitions of high-confl ict and low-confl ict regions in selected South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Annex 7B Labor market characteristics and educational attainment in high-confl ict and low-confl ict areas of selected South Asian countries . . . . . . . . . . . . . . 312 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 Appendixes A Household surveys used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 B Methodology used to analyze labor transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326 Boxes 1.1 Increasing productivity in agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.2 Options for reforming the power sector in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . 24 1.3 The critical role of nutrition in early childhood development . . . . . . . . . . . . . . . . . . . 31 2.1 International migration in Nepal and its effects on poverty . . . . . . . . . . . . . . . . . . . . . 63 2D.1 Trends in female labor force participation in southeast and East Asian comparator countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 3.1 Child labor in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 3.2 Composition of the labor force by employment status . . . . . . . . . . . . . . . . . . . . . . . . . 98 3.3 Determinants of informality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 3.4 Trends in India’s formal manufacturing sector, 1998–2007 . . . . . . . . . . . . . . . . . . . 109 4.1 Electricity challenges facing South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 4.2 Corruption in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 4.3 Bringing light to rural consumers in Gujarat, India . . . . . . . . . . . . . . . . . . . . . . . . . . 147 4.4 Improving performance of state-owned power suppliers in Andhra Pradesh . . . . . . 148 4.5 Cutting red tape to reduce corruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 4.6 Legislative and administrative interventions to reduce corruption in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 4.7 Public-private collaboration to implement reforms in the Cambodian garment industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 4.8 Effects of easing business entry regulations on ï¬? rm entry, employment, and formalization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 4E.1 Successful approaches to small and medium-size business banking . . . . . . . . . . . . . . 164 4E.2 Scaling up microï¬? nance institutions: The case of BRAC Bank . . . . . . . . . . . . . . . . . 164 viii CONTENTS 5.1 Recruiting teachers based on merit in Sindh, Pakistan . . . . . . . . . . . . . . . . . . . . . . . . 195 5.2 Teacher incentives schemes in Andhra Pradesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 5.3 The Reaching Out-of-School Children project in Bangladesh . . . . . . . . . . . . . . . . . . 198 5.4 Vocational education provided in the public school system . . . . . . . . . . . . . . . . . . . . 200 5.5 Learning from Australia’s systemic reforms of training and tertiary education . . . . . 202 5.6 Providing scholarships for tertiary education in Nepal . . . . . . . . . . . . . . . . . . . . . . . 206 5.7 Industry-government cooperation: The Penang Skills Development Centre . . . . . . . . 208 6.1 Severance reforms in Austria and Chile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 6.2 Unemployment beneï¬?t proposals for Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 6.3 Nongovernment players in South Asia’s informal labor market . . . . . . . . . . . . . . . . . 247 6.4 Key features of India’s Mahatma Gandhi National Rural Employment Guarantee Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 6.5 Mexico’s proactive approach to supporting small and medium-size enterprises . . . . 255 6.6 Training informal workers: Kenya’s Jua Kali experience . . . . . . . . . . . . . . . . . . . . . . 256 6.7 Business training and ï¬? nancial support for self-employed women in Sri Lanka . . . . 259 7.1 Private sector solutions to the security constraint: Lessons from Afghanistan . . . . . 294 7.2 Community-led infrastructure provision: Afghanistan’s National Solidarity Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 7.3 Improving the land rights framework in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 7.4 Improving the regulatory framework in a postconflict situation: Lessons from Liberia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 7.5 Improving schooling despite armed conflict through community schools in Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 7.6 Lessons from efforts to reintegrate Ugandan youth . . . . . . . . . . . . . . . . . . . . . . . . . . 303 7.7 Training and employing displaced people: The case of Asocolflores in Colombia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 7.8 Implementing public works in a postconfl ict environment: Sri Lanka’s Northern Province Emergency Recovery Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305 Figures 1.1 Annual growth in working-age population, employment, and labor force in selected South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2 Distribution of employment by type in South Asia, by country . . . . . . . . . . . . . . . . . . . 4 1.3 Average annual increases in mean real wages in selected countries in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.4 Percentage of workers in households below the poverty line in selected South Asian countries, by employment status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.5 Average number of months without work in the past year, casual laborers in India, by sector, 1999/2000–2009/10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.6 Distribution of rural and urban workers in selected South Asian countries, by employment type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.7 Conditional probability of moving into and out of better jobs in rural India, by education and gender, 2004/05–2007/08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.8 Annual growth in GDP per capita, by region, 1960s–2000s . . . . . . . . . . . . . . . . . . . . . 9 1.9 Annual growth in GDP per capita in South Asia, by country, 1960s–2000s . . . . . . . . . 9 1.10 Sources of annual growth in labor productivity, by region, 1960–80 and 1980–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.11 Sources of annual growth in labor productivity in selected countries in South Asia, by country, 1960–80 and 1980–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 CONTENTS ix 1.12 Sources of annual growth in total factor productivity in China, India, Pakistan, and Thailand, by sector and reallocation effects . . . . . . . . . . . . . . . . . . . . . 13 1.13 Median wage and value added per manufacturing worker in India, by ï¬? rm size and type, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.14 Share of manufacturing employment in India, by ï¬? rm size and type, 1994–2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.15 Ratio of working-age to nonworking-age population in South Asia, by country, 1960–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.16 Severity of constraints reported by South Asian benchmark ï¬? rm in the urban formal sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 1.17 Cross-country comparisons of reported severity of electricity constraint and power outages for a benchmark ï¬? rm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 1.18 Percentage of ï¬? rms expected to give gifts to public ofï¬?cials, by type of interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 1.19 Severity of constraints identiï¬?ed by South Asian benchmark (nonexpanding) and expanding ï¬? rm in the urban formal sector . . . . . . . . . . . . . . . . . 22 1.20 Severity of constraints reported by micro benchmark ï¬? rm in urban and rural sectors of Bangladesh, Pakistan, and Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.21 Severity of constraints reported by micro benchmark ï¬? rm in India’s urban formal and informal sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.22 Employers’ perceptions of skills of recently graduated engineers in India . . . . . . . . . . 28 1.23 Wage premiums in selected South Asian countries, by level of education . . . . . . . . . . 29 1.24 Share of South Asian labor force with no education, with international comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 1.3.1 Percentage of children under ï¬?ve with malnutrition, by region and country . . . . . . . . 31 1.25 Employment protection indicators in selected countries . . . . . . . . . . . . . . . . . . . . . . . 33 1.26 Weeks of wages required to be paid in severance in regions, country income groups, and selected South Asian countries, by length of service . . . . . . . . . . . . . . . . . 34 1.27 Cross-country comparison of reported severity of the labor regulation constraint . . . 35 1.28 Proportion of country-years in armed confl ict, by region, 2000–08 . . . . . . . . . . . . . . 37 1.29 Severity of business environment constraints (average) reported by ï¬? rms in low-confl ict and high-confl ict areas of Afghanistan, 2008 . . . . . . . . . . . . . . . . . . . . . 38 1.30 Unemployment rates in the Northern and Eastern provinces of Sri Lanka, 1997–2001 and 2002–04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 1C.1 Percentage of workers in households below the poverty line in Bangladesh, India, and Nepal, by employment status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 1C.2 Ratio of rural nonfarm and urban wages to agricultural wages in Bangladesh, India, and Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 2.1 Annual growth in GDP per capita, by region, 1960s–2000s . . . . . . . . . . . . . . . . . . . . 50 2.2 Annual growth in GDP per capita in South Asia, by country, 1960s–2000s . . . . . . . . 50 2.3 Sources of annual growth in labor productivity, by region, 1960–80 and 1980–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 2.4 Sources of annual growth in labor productivity in selected countries in South Asia, by country, 1960–80 and 1980–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.5 Ratio of working-age to nonworking-age population in South Asia, by country, 1960–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 2.6 Sectoral shares of GDP and employment in selected countries in South Asia, 1980s–2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 2.7 Shares of agriculture and services in employment and GDP in South Asian and comparator countries, 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 x CONTENTS 2.8 Sources of annual growth in total factor productivity in India and Pakistan, by sector and reallocation effects, 1980–2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.9 Sources of annual growth in total factor productivity in China, India, Pakistan, and Thailand, by sector and reallocation effects . . . . . . . . . . . . . . . . . . . . . 58 2.10 Annual growth in working-age population, employment, and labor force in selected South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 2.11 Average annual increases in mean real wages in selected countries in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 2.12 Percentage of workers in households below the poverty line in selected South Asian countries, by employment status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 2.13 Percentage of workers in households below the poverty line in India, by employment status and gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.14 Average number of months without work in the past year, casual laborers in India, by sector, 1999–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 2.15 Distribution of per capita household expenditure in India and Nepal, by employment status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 2.16 Distribution of rural and urban workers in selected South Asian countries, by employment type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 2.17 Distribution of rural nonfarm workers in India, by employment type, 1983–2009/10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 2.18 Labor transitions in rural areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 2.19 Probability of moving into or out of better jobs in rural Bangladesh, India, and Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 3.1 Total employment in South Asia, by country, 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 3.2 Employment rates in lower- and lower-middle-income countries . . . . . . . . . . . . . . . . . 87 3.3 Male and female employment rates in South Asia, by country . . . . . . . . . . . . . . . . . . 87 3.4 Trends in employment rates in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . 88 3.1.1 Percentage of child workers attending school in South Asia, by age group and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 3.5 Female labor force participation rates in South Asia, by age group and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 3.6 Annual percentage increases in number of employed workers in South Asia, by sector and country. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 3.7 Distribution of employment in South Asia, by sector and country . . . . . . . . . . . . . . . . 95 3.8 Percentage of rural workers in the nonfarm sector in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 3.9 Percentage of rural workers in the nonfarm sector in China and India, 1983–2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 3.10 Rural nonfarm sector employment in South Asia, by economic activity and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 3.2.1 Distribution of per capita household expenditure in India and Nepal, by employment status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 3.11 Percentage of employment in South Asia classiï¬?ed as informal, by country. . . . . . . . 100 3.12 Percentage of labor force not covered by pension schemes, by region . . . . . . . . . . . . 101 3.13 Percentage of workers in households below the poverty line in Bangladesh, India, and Nepal, by employment status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 3.14 Ratio of median rural nonfarm and urban wages to agricultural wages in selected South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 3.15 Ratio of median industry and service sector wages to agricultural wages in selected South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 CONTENTS xi 3.16 Average wage, value added, and capital per manufacturing worker in India, by ï¬? rm size and type, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 3.17 Share of manufacturing employment in India, by ï¬? rm size and type, 1994–2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 3.18 Share of manufacturing employment by ï¬? rm size in India and selected East Asian economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 3.4.1 Employment in India’s formal manufacturing sector, by ï¬? rm size, type, and location, 1998–2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 3.19 Average wage, value added, and capital per service sector worker in India, by ï¬? rm size and type, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 3.20 Share of service sector employment in India, by ï¬? rm size and type, 2001 and 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 3.21 Percentage of workers with some education and percentage of workers with secondary education or above in South Asia, by employment type and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 3.22 Percentage of rural workers in the rural nonfarm sector in South Asia, by gender and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 3.23 Percentage of rural workers in the rural nonfarm sector in India, Nepal, and Pakistan, by gender and age cohort, 1999–2009/10 . . . . . . . . . . . . . . . . . . . . . . 114 3.24 Decomposition of wage gap between male and female workers in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 3.25 Decomposition of wage gaps between nonethnic minority and ethnic minority workers in India, Nepal, and Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 3B.1 Regional variations in employment rate in South Asia, by country . . . . . . . . . . . . . . 120 4.1 Severity of constraints reported by South Asian benchmark ï¬? rm in the urban formal sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 4.2 Cross-country comparisons of reported severity of the electricity constraint . . . . . . 129 4.3 Cross-country comparisons of power outages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 4.4 Cross-country comparison of reported severity of corruption constraint . . . . . . . . . 134 4.5 Cross-country comparison of bribe payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 4.2.1 Percentage of ï¬? rms expected to give gifts to public ofï¬?cials, by type of interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 4.6 Cross-country comparison of reported severity of political instability constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 4.7 Cross-country comparison of reported severity of tax administration constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 4.8 Severity of constraints reported by South Asian benchmark (nonexpanding) and expanding ï¬? rm in the urban formal sector . . . . . . . . . . . . . . . . 138 4.9 Severity of constraints reported by South Asian benchmark (manufacturing) and service ï¬? rm in the urban formal sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 4.10 Severity of constraints reported by micro benchmark ï¬? rm in urban and rural sectors of Bangladesh, Pakistan, and Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . 141 4.11 Severity of constraints reported by micro benchmark ï¬? rm in India’s urban formal and informal sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 4A.1 Severity of constraints reported by benchmark ï¬? rm in urban formal sector in high- and low-income states in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 4B.1 Cross-country comparison of reported severity of tax rate constraint . . . . . . . . . . . . 156 4B.2 Tax revenue as a percentage of GDP in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . 157 4B.3 Highest marginal corporate tax rate in South Asian countries and selected comparator countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 xii CONTENTS 4C.1 Severity of constraints reported by benchmark ï¬? rm and ï¬? rm with 60 employees in the urban formal sector in Nepal and Sri Lanka . . . . . . . . . . . . . . . 158 4C.2 Severity of constraints reported by benchmark (nonexporting) and exporting ï¬? rm in the urban formal sector in South Asia and Bangladesh . . . . . . . . . 159 4D.1 Cross-country comparison of reported severity of access to ï¬? nance constraint . . . . . 160 4D.2 Percentage of ï¬? rms with credit line or loan from ï¬? nancial institution, by ï¬? rm size and region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 5.1 Employers’ perceptions of skills of recently graduated engineers in India . . . . . . . . . 173 5.2 Evolution of skills content of urban wage workers in India, 1994–2010 . . . . . . . . . . 174 5.3 Wage premiums over elementary occupations in India, 1994–2010. . . . . . . . . . . . . . 175 5.4 Wage premiums in selected South Asian countries, by level of education . . . . . . . . . 175 5.5 Share of South Asian labor force with no education, with international comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 5.6 Predicted probability of working in rural nonfarm and urban regular wage jobs in selected South Asian countries, by level of education and gender . . . . . 178 5.7 Conditional probability of moving into and out of better jobs in rural India, by education and gender, 2005–08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 5.8 Conditional probability of moving into and out of better jobs in urban India, by education and gender, 2005–08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 5.9 Share of young cohorts with completed primary, secondary, and tertiary education in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 5.10 Mean years of education of 15–34 year olds in South Asia, by gender and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 5.11 Reading and arithmetic achievement in rural India and Pakistan, by class, 2010 . . . 186 5.12 Public expenditure on education as a share of GDP, in South Asia and other regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 5.13 Educational attainment of the South Asian labor force in 2010 (estimated) and 2030 (projected), by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 5.14 Projected shares of prime-age (35–49) and oldest (50–64) labor force participants in South Asia with no education or only primary attainment in 2030, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 5.15 Percentage of children under ï¬?ve with malnutrition, by region and country . . . . . . . 190 5.16 Share of primary and secondary enrollments in public and private institutions in selected South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 5.17 Percentage of graduates from public and private industrial training institutions employed in the organized sector in three states of India, 2002/03 . . . . . . . . . . . . . . 201 5.18 Share of preemployment training and tertiary education enrollment in public and private institutions in South Asia, by country. . . . . . . . . . . . . . . . . . . . . . 205 5.19 Percentage of ï¬? rms providing on-the-job training in regions and selected South Asian countries, by ï¬? rm size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 5A.1 Educational attainment in the labor force in South Asia, by country. . . . . . . . . . . . . 212 5A.2 Share of labor force with at least primary, upper-secondary, and tertiary education, in South Asia and international comparators . . . . . . . . . . . . . . . . . . . . . . 213 5A.3 Gross and net enrollment rates in primary and secondary education, by region and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 5A.4 Enrollment in vocational education and training as a share of secondary enrollment, by region and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 5A.5 Gross enrollment rate in tertiary education, by region and country . . . . . . . . . . . . . 216 5A.6 Share of young cohorts with completed primary and lower-secondary education in South Asia, by gender and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 CONTENTS xiii 5A.7 Share of young cohorts with completed upper-secondary and tertiary education in South Asia, by gender and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 5A.8 Mean years of education of 15–34 year olds in selected South Asian countries, by caste/ethnicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 5A.9 Share of 20–28 year olds in Afghanistan and Nepal with different years of education completed in Afghanistan (2008) and Nepal (2004), by gender and socioeconomic status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 5A.10 Percentage of children under age 5 with stunting in selected South Asian countries, by socioeconomic status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 6.1 Percentage of workforce not covered by formal pension scheme . . . . . . . . . . . . . . . . 231 6.2 Percentage of wage employees in India and Sri Lanka covered by social security, by type of worker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 6.3 The continuum of employment protection legislation . . . . . . . . . . . . . . . . . . . . . . . . 234 6.4 Weeks of wages required to be paid in severance in regions, country income groups, and selected South Asian countries, by length of service. . . . . . . . . . 236 6.5 Employment protection indicators in selected countries . . . . . . . . . . . . . . . . . . . . . . 237 6.6 Job turnover rates and labor regulations in Indian states . . . . . . . . . . . . . . . . . . . . . . 238 6.7 Cross-country comparison of reported severity of the labor regulation constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 6.8 Minimum wages as a proportion of median formal sector wages in Nepal, Pakistan, and Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240 6.9 Participation of women and members of scheduled castes and tribes in the Mahatma Gandhi National Rural Employment Guarantee Program, 2006/07–2010/11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 6A.1 Trade union membership in India, 1987–2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 6A.2 Trade union membership in Pakistan, 1999–2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 6A.3 Trade union membership in Sri Lanka, 1987–2006 . . . . . . . . . . . . . . . . . . . . . . . . . . 263 6A.4 Percentage of unionized workers in India, by employment status and sector, 1994 and 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 6A.5 Job creation and destruction flows in Sri Lanka and selected groups of countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 6A.6 Tax wedges in South Asian and international comparator countries . . . . . . . . . . . . . 266 6A.7 Ratio of minimum wage to median casual and formal sector wage, by states in India, 2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 7.1 Proportion of country-years in armed confl ict, by region, 2000–08 . . . . . . . . . . . . . 282 7.2 Top 15 countries in number of deaths from armed confl ict, 2008 . . . . . . . . . . . . . . 283 7.3 Effects of confl ict on demand for and supply of labor . . . . . . . . . . . . . . . . . . . . . . . . 284 7.4 Annual growth in GDP and number of battle deaths in India and Nepal, 2000–08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 7.5 Unemployment rates in the Northern and Eastern provinces of Sri Lanka, 1997–2001 and 2002–04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 7.6 Percentage of working-age population employed in high-confl ict and low-confl ict areas of India and Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 7.7 Percentage of working-age population employed in high-confl ict areas of Nepal, by gender, 1996–2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 7.8 Urban workers as share of all workers in India and Nepal. . . . . . . . . . . . . . . . . . . . . 289 7.9 Percentage of workforce employed in unpaid family labor in India, 2000–08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 7.10 Percentage of working-age population with completed lower-secondary education in low- and high-confl ict areas of selected South Asian countries . . . . . . . 291 xiv CONTENTS 7.11 Severity of business environment constraints (average) reported by ï¬? rms in low-confl ict and high-confl ict areas of Afghanistan, 2008 . . . . . . . . . . . . . . . . . . . . 292 7.12 Sri Lanka infrastructure accessibility index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 7.5.1 Primary school enrollment rate in Nepal among children 6–10 years old, 1996–2004. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 7.13 Central government debt as a percentage of GDP in confl ict and nonconfl ict areas of South Asia and the world, 1990s and 2000s . . . . . . . . . . . . . . . . . . . . . . . . 307 Tables 1.1.1 Returns to agricultural growth from investments in public goods and subsidies in India, 1960s–90s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.1 Top ï¬?ve constraints reported by South Asian benchmark ï¬? rm in the urban formal sector, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.2 Top ï¬?ve constraints reported by South Asian benchmark (nonexpanding) and expanding ï¬? rm in the urban formal sector, by country . . . . . . . . . . . . . . . . . . . . . 23 1.3 Top ï¬?ve constraints reported by micro benchmark ï¬? rm in the urban and rural sectors of Bangladesh, Pakistan, and Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . 25 1.4 Top ï¬?ve constraints reported by micro benchmark ï¬? rm in India’s urban formal and informal sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1A.1 Summary economic statistics of South Asian countries . . . . . . . . . . . . . . . . . . . . . . . . 42 1B.1 Deï¬? nitions of key labor market terms used in this book . . . . . . . . . . . . . . . . . . . . . . . 43 2.1 Labor productivity in South Asia and East Asia, by sector, 2008 . . . . . . . . . . . . . . . . 58 2.2 Decomposition of decline in worker poverty rates . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 2.3 Correlations of country growth rates of per capita GDP across decades . . . . . . . . . . . 69 2B.1 Sources of average annual growth in output per worker, by region, 1960–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 2B.2 Sources of average annual growth in output per worker in South Asia, by country, 1960–2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 2C.1 Regressions of shares of agriculture, industry, and services in employment and GDP, 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 2D.1.1 Female labor force participation in Bangladesh, India, and Pakistan and Asian comparator countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 2E.1 Percentage of workers in households below the poverty line in Bangladesh, by employment type, 2000–10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 2E.2 Number of working poor in Bangladesh, by employment type, 2000–10 . . . . . . . . . . 76 2E.3 Percentage of workers in households below the poverty line in India, by employment type, 1983–2004/05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 2E.4 Number of working poor in India, by employment type, 1985–2005 . . . . . . . . . . . . . 78 2E.5 Percentage of workers in households below the poverty line in Nepal, by employment type, 1995/96 and 2003/04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 2E.6 Number of working poor in Nepal, by employment type, 1995–2005 . . . . . . . . . . . . 79 2F.1 Ofï¬?cial and authors’ estimated poverty rates for urban, rural, and all workers in India, 1983–2004/05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 3.1.1 Incidence of child labor in South Asia, by age group and country . . . . . . . . . . . . . . . . 89 3.1 Male and female labor force participation, employment, and unemployment rates in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 3.2 Factors associated with participation of women in urban areas . . . . . . . . . . . . . . . . . . 92 3.3 Reasons why urban women in South Asia do not participate in the labor force, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 CONTENTS xv 3.4 Distribution of employment in South Asian countries, by location and sector . . . . . . . 94 3.5 Distribution of employment in South Asian countries, by type of employment . . . . . . 99 3.6 Distribution of formal and informal manufacturing ï¬? rms in India, by location and size, 2005. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 3.7 Distribution of service ï¬? rms in India, by location and size, 2006 . . . . . . . . . . . . . . . 109 3.8 Average years of education in South Asian countries, by sector of employment. . . . . 111 3A.1 Deï¬? nition of employment and unemployment used based on national surveys . . . . . 117 3A.2 Deï¬? nition of formal and informal workers used based on national surveys . . . . . . . 118 4.1 Top ï¬?ve constraints reported by South Asian benchmark ï¬? rm in the urban formal sector, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 4.1.1 Selected energy indicators in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . 132 4.2 Electricity constraints faced by ï¬? rms, by developing region . . . . . . . . . . . . . . . . . . . 132 4.3 Top ï¬?ve constraints reported by South Asian benchmark (nonexpanding) and expanding ï¬? rm in the urban formal sector, by country . . . . . . . . . . . . . . . . . . . . 139 4.4 Top ï¬?ve constraints reported by micro benchmark ï¬? rm in the urban and rural sectors of Bangladesh, Pakistan, and Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . 141 4.5 Top ï¬?ve constraints reported by micro benchmark ï¬? rm in India’s urban formal and informal sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 4.6 Private investment in electricity in South Asia, by country, 1999–2000 and 2001–10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 4.7 Market reforms in the power sector in selected countries in South Asia . . . . . . . . . . 145 4A.1 Top ï¬?ve constraints reported by benchmark ï¬? rm in the urban formal sector in high- and low-income states in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 5.1 Perceived demand for and deï¬?ciency in skills in programming and software engineering in Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 5.2 Routine and nonroutine skills categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 5.3 Main issues, interventions, and expected outcomes in early childhood development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 5.4 Mechanisms for funding tertiary education in South and East Asia, by economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 5A.1 Share of Indian labor force requiring high concentration of nonroutine cognitive analytical and interpersonal skills, 1994–2010 . . . . . . . . . . . . . . . . . . . . . 210 5A.2 Mean years of education of 15–34 year olds in South Asia, by gender and country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 5A.3 Summary of randomized experiments on teacher incentives . . . . . . . . . . . . . . . . . . . 211 5B.1 Assumptions underlying scenarios used to project educational attainment of population in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 5B.2 Educational attainment of the South Asian labor force in 2010 (estimated) and 2030 (projected) under various scenarios, by country . . . . . . . . . . . . . . . . . . . . . 222 6.1 Correlation coefï¬?cients among union membership, social security coverage, and employment in ï¬? rms with 10 or more employees, India, 2010 . . . . . . . . . . . . . . 232 6.2 Percentage of workers with access to formal protection instruments in India and Sri Lanka, by worker characteristic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 6.3 Year of ratiï¬?cation of International Labour Organization core conventions, Declaration on Fundamental Principles, and Rights at Work, by country . . . . . . . . . 233 6.4 Selected hiring and redundancy rules in South Asia, by country . . . . . . . . . . . . . . . . 235 6.5 Stylized characteristics of protecting workers versus protecting jobs . . . . . . . . . . . . . 242 6.6 Selected public works programs in South Asia, by country . . . . . . . . . . . . . . . . . . . . 248 6.7 Participation in the Mahatma Gandhi National Rural Employment Guarantee Program, by consumption quintile, 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 xvi CONTENTS 6.8 Skill requirements in the informal sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 6.9 Coverage of microï¬? nance in South Asia, by country, 2009/10. . . . . . . . . . . . . . . . . . 256 6.10 Selected programs supporting self-employment and microenterprises in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258 6A.1 Selected aspects of employment protection in India . . . . . . . . . . . . . . . . . . . . . . . . . . 269 6A.2 Job creation and destruction rates in large manufacturing ï¬? rms in India, by employment type, 2001–08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 6A.3 Minimum wage policies in South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . 271 7.1 Major internal armed confl icts in South Asia since 1978 . . . . . . . . . . . . . . . . . . . . . . 283 7.2 Armed confl icts in South Asia, by confl ict stage, geographic scope, and number of casualties per thousand people . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 7.3 Labor market transitions in postconfl ict zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 7.4 Policy interventions in the initial postconfl ict stages . . . . . . . . . . . . . . . . . . . . . . . . . 310 7B.1 Labor market characteristics in high-confl ict and low-confl ict areas of Afghanistan, 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312 7B.2 Labor market characteristics in high-confl ict and low-confl ict areas of India, 2000 and 2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312 7B.3 Labor market characteristics in high-confl ict and low-confl ict areas of Nepal, 1996 and 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 7B.4 Labor market characteristics in high-confl ict and low-confl ict areas of Sri Lanka, 2004 and 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314 7B.5 Educational attainment in high-confl ict and low-confl ict areas of South Asia, by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 A.1 Labor force and living standards surveys used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324 Foreword I t is my great pleasure to introduce the institutions, development practitioners, and inaugural issue of South Asia Develop- the media. This series will serve as a vehicle ment Matters. More and Better Jobs in for voicing an in-depth synthesis of economic South Asia is timely because of its relevance and policy analysis on key development top- to the region’s 490 million young people ics for South Asia. who can make South Asia the region of Next year we will address the theme of the future. Despite having one of the low- inequality, the challenges faced by individ- est female participation rates in the devel- ual countries in South Asia, and the ways in oping world, South Asia will add at least which inequality can be addressed so that one million people to its labor force every countries can grow and develop with equity. month. The significant additions to the I hope that the knowledge gathered in the labor force could be a demographic divi- South Asia Development Matters series will dend or a curse, which is why this report beneï¬?t the entire region and that it promotes is so important: the key to future peace debate and builds consensus of all those who and poverty reduction in South Asia is the care about stimulating development and creation of enough good-quality jobs in the eradicating poverty in South Asia. decades to come. Future generations will I would like to thank the team members thank this one for using this opportunity for their high-quality product. They were to create an environment for progressively able to tackle a very difï¬?cult issue in a way better jobs, which are the only sustainable that examined technical quality, with great pathway out of poverty. openness to views from different professional The South Asia Development Matters streams. This was a tough topic, and they series will be published annually under the have produced important recommendations supervision of Kalpana Kochhar, Chief Econ- to promote policy debate. omist for the South Asia Region. I hope the series will promote dialogue and debate with Isabel Guerrero all our partners, not only policy makers but Vice President, South Asia Region also civil society organizations, academic The World Bank xvii Preface M ore and Better Jobs in South Asia and unrest. Closer to home, labor is the most launches the series South Asia important, if not the only, asset of the poor. Development Mat ters. W hen South Asia, despite impressive growth selecting the ï¬?rst topic in the series, we had no and poverty reduction over the past two doubt that it should focus on jobs. South Asian decades, remains home to more than half a countries will add 1.0 million to 1.2 million billion poor people, large numbers of whom new entrants to the labor force every month have little or no education and suffer from for the next two decades and will contribute poor health. Part of the reason is that South about 40 percent of the total new entrants to Asia has some of the worst nutrition indi- the global working-age (15–64) population. It cators in the world. Research clearly shows is not surprising that we decided to focus our that a person’s cognitive development begins attention on the changes in policies necessary in the early years of life, long before formal to create a larger number and higher quality schooling begins. Nutrition and early child- of jobs. Identification and implementation hood development have a strong positive of these policies are central to South Asia’s relationship with educational achievement employment challenge of absorbing the grow- and signiï¬?cant payoffs for lifetime learning ing number of entrants to the labor force at and labor market productivity. The chal- rising levels of productivity. lenge of generating more productive jobs is Recent global events have helped shine intensiï¬?ed because most of the countries in a brighter light on this issue. According to the region are still in confl ict or have only the International Labour Organization, recently emerged from it and many people as many as 30 million people worldwide face serious problems related to access to lost their jobs as a result of the 2008 crisis. opportunities based on gender, caste, and Youth unemployment is especially high, and socioeconomic status. inequality has increased in many countries More and Better Jobs in South Asia around the world. As recent events during attempts to answer three questions: the “Arab Springâ€? in the Middle East and North Africa demonstrate, joblessness and • Has South Asia been creating an increas- inequality can trigger political instability ing quantity and quality of jobs? xix xx PREFACE • What are the determinants of the quality South Asia’s recent track record with regard of job creation and what is the employ- to the quantity and quality of job creation. ment challenge going forward? It traces the relationship of such job cre- • What demand- and supply-side bottle- ation mostly to overall economic growth and necks need to be eased to meet South attempts to answer what needs to be done Asia’s employment challenge in the face of to meet South Asia’s employment challenge. intensifying demographic pressure? Chapter 3 discusses the key features of labor markets in South Asia, including where Although details vary by country, overall the better jobs are, who holds them, and there is reason for cautious optimism. Over the implications for the employment chal- the past two decades, the region has created lenge ahead. Chapter 4 reviews the busi- more jobs—at a rate largely comparable to ness environment constraints affecting, in growth in the working-age population—and particular, those ï¬? rms that have expanded better jobs, in terms of higher pay for wage employment and discusses policy options for workers, lower poverty for self-employed overcoming the most binding business con- workers, and reduced risk of low and uncer- straints in South Asia. Chapter 5 analyzes the tain income for the most vulnerable group of dimensions of the education and skills chal- workers. Yet, there is no room for compla- lenge in the region and discusses policy pri- cency because the challenges are big. Not only orities for improving the quality and skills of do a larger number of jobs need to be created, graduates of education and training systems. but the jobs also need to be more produc- Chapter 6 reviews the role of labor market tive and make workers less vulnerable. Our policies and institutions in encouraging job study shows that creating a larger number creation and protecting workers in the formal of more productive jobs for a growing labor and informal economy and discusses possible force calls for a multisectoral reform agenda directions for labor market policies, including that includes improving access to electricity options to increase the access of informal sec- for ï¬?rms across all sectors in urban and rural tor workers to programs that help them man- settings, dealing decisively with issues of gov- age labor market shocks and improve their ernance and corruption, improving access to future earnings potential. Finally, chapter 7 land and transport links between town and reviews the key constraints to job creation country, improving nutrition in early child- and the policy priorities for creating more hood, equipping workers with skills relevant and better jobs in conflict-affected areas. for the world of work, and reorienting labor market regulations and programs to protect workers rather than jobs. Kalpana Kochhar This book is divided into seven chapters. Chief Economist, South Asia Region Chapter 1 is an overview. Chapter 2 reviews The World Bank Acknowledgments M ore and Better Jobs in South Asia pieces by Shaghil Ahmed, Harold Alderman, is the ï¬? rst in a series of fl agship Sudeshna Ghosh Banerjee, Hai Anh Dang, reports conceived and launched Puja Vasudeva Dutta, Madhur Gautam, by Isabel Guerrero, Regional Vice President Rana Hasan, Benjamin Herzberg, Ina of the South Asia Region. It is the product of Hoxha, Kalim Hyder, Karl Jandoc, Maria a collaborative effort by many professionals Jos, Samir KC, Somik Lall, Peter Lanjouw, and institutions from both inside and out- Norman Loayza, David McKenzie, Claudio side the World Bank. Monteneg ro, M ar tin Moreno, R in ku The report was prepared by a team led by Murgai, Denis Nikitin, Sheoli Pargal, Harry Reema Nayar and Pablo Gottret, under the Patrinos, Zhiheng Png, Shumaila Rifaqat, direction of Kalpana Kochhar, Chief Econ- Hiroshi Saeki, Mehnaz Safavian, Vibhor omist of the South Asia Region. The core Saxena, Claudia Ines Vasquez, Jessica Ville- team comprised Pradeep Mitra, Yue Man gas, Tomoko Wada, and Karar Zunaid. The Lee, Indhira Santos, Gordon Betcherman, team also benefited from papers prepared Mahesh Dahal, and Maheshwor Shrestha. by colleagues at various institutions in the Wendy Carlin, Amit Dar, Lakshmi Iyer, Toby region, including Nazneen Ahmed, Rushidan Linden, and Mark Schaffer made signiï¬?cant Rahman, R. Shamsunnahar, and Mohamad contributions to speciï¬?c chapters. Yunus (Bangladesh Institute of Development Many people provided written inputs and Studies); Farzana Munshi (BRAC University, contributions on various issues. Early think Dhaka); Bibek Debroy (Center for Policy pieces and background papers were prepared Research, New Delhi); Koushik Dutta (inde- by T. N. Srinivasan, Barry Bosworth, Peter pendent consultant); Ramani Gunatilaka B. Hazell, Derek Headey, Alejandro Nin (adjunct research fellow, Faculty of Business Pratt, Derek Byerlee, David Robalino, Ash- and Economics, Monash University); Amrita ish Narain, Ernest Sergenti, Pierella Paci, Dutta, Ann George, Dev Nathan, Preet David Margolis, Mario Di Filippo, Tanja Rustagi, Alakh Sharma, and Ravi Srivas- Lohmann, Tenzin Chhoeda, Mark Dutz, tava (Institute of Human Development, New Hong Tan, Stephen O’Connell, Lucia Madri- Delhi); and Nisha Arunatilake, Roshini gal, and Meera Mahadevan. The report also Jayaweera, and Anushaka Wijesinha (Insti- draws on speciï¬?c contributions or analytical tute of Policy Studies, Colombo). xxi xxii ACKNOWLEDGMENTS The team beneï¬?ted from advice and com- Karthik Muralidharan, Somil Nagpal, Has- ments from Eliana Cardoso and Andrew san Naqvi, Claudia Nassif, Naveed Naqvi, Steer (former Chief Economists of the South John Newman, Thomas O’Brien, Robert Asia Region), Martin Rama, Michal Rut- Palacios, Dilip Parajuli, Giovanna Prennushi, kowski, and Marcelo Selowsky. Arup Banerji, Jasmine Rajbhandary, Dhushyanth Raju, Emmanuel Jimenez, and Ana Revenga were Mansoora Rashid, Susan Razzaz, Silvia peer reviewers for the report. The team is Redaelli, Francis Rowe, Deepa Sankar, Tah- grateful for the contributions of participants seen Sayed, Hisanobu Shishido, Venkatesh at various panel discussions and brainstorm- Sundararaman, and T. G. Srinivasan. Many ing sessions. They include Janamitra Devan, individuals from a variety of research, pol- Tamar Manuelyan Atinc, Shanta Devarajan, icy, and academic institutions and interna- Ariel Fiszbein, Ernesto May, John Henry tional development agencies in Bangladesh, Stein, Roberto Zagha, Rachid Benmessaoud, India, Nepal, and Pakistan participated in Ellen Goldstein, Nicholas Kraft, Susan and provided extremely useful insights at Goldmark, Asli Demirgüç-Kunt, Kaushik consultation meetings held in the countries. Basu, Nadeem Haque, Michael Walton, Unfortunately, it is not possible to name them Dilip Mukherji, Siddiqur Osmani, Manish individually here. Sabharwal, Binayak Sen, and the late Suresh The report would have not been possible Tendulkar. The team gratefully acknowl- without the able assistance of Izabela Anna edges comments and assistance from Faizud- Chmielewska, Julie-Anne Graitge, Marjorie din Ahmed, D. H. C. Aturupane, Roshan Kingston, and Elfreda Vincent. The team Darshan Bajracharya, Dan Biller, Andreas also gratefully acknowledges ï¬? nancial sup- Blom, John Blomquist, Jose Roberto Calix, port from the Multi-Donor Trust Fund for Eliana Carranza, Anthony Cholst, Maria Trade and Development, the Poverty and Correia, Halil Dundar, Simeon Ehui, Social Impact Analysis (PSIA), and the World Ejaz Syed Ghani, Sangeeta Goyal, Mary Bank Research Committee. Aziz Gökdemir, Hallward-Driemeier, Zahid Hussain, Nalin Patricia Katayama, Andrés Meneses, Santi- Jena, Dean Mitchell Jolliffe, Sanjay Kathuria, ago Pombo-Bejarano, and Janice Tuten of the Ayesha Khan, Gladys Lopez-Acevedo, Eric World Bank’s Ofï¬?ce of the Publisher coor- David Manes, Nkosinathi Mbuya, Julie dinated the editing, design, production, and McLaughlin, Cem Mete, Hanid Mukhtar, printing of this book. Abbreviations ANTA Australian National Training Authority CDC community development council CPI Competitiveness Partnership Initiative CSSP Community School Support Program DDR disarmament, demobilization, and reintegration EPZ export processing zones FATA Federally Administered Tribal Areas FPD Financial and Private Sector Development GDP gross domestic product GET Global Education Trend GHTDP Great Himalaya Trail Development Programme G–PSF Government–Private Sector Forum GW gigawatts IFC International Finance Corporation ILO International Labour Organization MCTEE Ministerial Council for Tertiary Education and Employment MGNREGA Mahatma Gandhi National Rural Employment Guarantee Act MSME micro and small and medium-size enterprise MW megawatt NATO North Atlantic Treaty Organization NCEUS National Commission for Enterprises in the Unorganised Sector NGO nongovernmental organization NSDC National Skills Development Corporation NSP National Solidarity Program NTFP nontimber forest produce NWFP-KP North West Frontier Province (Khyber Pakhtunkhwa) OECD Organisation for Economic Co-operation and Development PAF Poverty Alleviation Fund PPP public-private partnership PPP purchasing power parity xxiii xxiv ABBRE VIATIONS PRI Panchayati Raj Institutions PSDC Penang Skills Development Centre SEZ special economic zone TEWA Termination of Employment of Workmen Act TFP total factor productivity UCDP/PRIO I Uppsala Conflict Data Program and Centre for the Study of Civil War at the International Peace Research Institute UNESCO United Nations Education, Scientiï¬?c and Cultural Organization UNICEF United Nations Children’s Fund WHO World Health Organization Note: All dollar amounts are U.S. dollars (US$) unless otherwise indicated. CHAPTER 1 Overview Key Messages Message 1: South Asia has created many, years of high-performing East Asian econo- mostly better jobs. mies (excluding China). Going forward, rap- idly growing countries in South Asia need to • Job creation in South Asia averaged almost sustain and slow-growing countries to ignite 800,000 a month between 2000 and 2010. growth by easing constraints to physical and The rate of employment growth broadly human capital accumulation. Higher rates of tracked that of the working-age (15–64) pop- factor accumulation, alongside more typical ulation. Open unemployment is low. rates of TFP growth, which will vary according • Real wages rose for wage workers, and poverty to country circumstances, will allow the region declined for the self-employed as well as all to absorb new entrants to the labor force at types of wage workers. Wages and poverty are rapidly rising levels of labor productivity. the primary criteria for improved job quality • Aggregate TFP growth should also increase that guide the analysis in this book. A reduced through a faster reallocation of labor from risk of low and uncertain income for the most agriculture to industry and services, where TFP vulnerable group of workers is a secondary growth is higher. Reallocation across sectors criterion for improved job quality. It could be needs to be complemented by moving labor monitored only in India, where it is satisï¬?ed. out of lower-productivity ï¬?rms in manufac- • The improvement in job quality has been turing and services, where the overwhelming associated with accelerating economic growth majority of South Asians who are employed in Bangladesh and India since the 1980s. In in these sectors work, into higher-productivity Nepal, where growth has been slow for several ï¬?rms within those sectors. Reallocation across decades, massive out-migration in response to and within sectors will require physical capital limited opportunities at home has improved accumulation (in electricity, for example, the labor market prospects for those who remain. lack of reliable supply of which is reported Workers’ remittances have reduced poverty by job-creating ï¬? rms as an obstacle to their across a wide swath of households. operation). It will also require investment in human capital to provide workers with the skills necessary to access better jobs. Message 2: The region faces an enormous • The “demographic transitionâ€?—the period employment challenge, but its demography can during which the number of workers grows work in support of the reforms needed to meet it. more rapidly than the number of dependents— • An estimated 1.0–1.2 million new entrants will can provide a tailwind in support of policy join the labor market every month over the next reform for the next three decades in much of few decades—an increase of 25–50 percent South Asia, as the resources saved from hav- over the average number of entrants between ing fewer dependents provides a “demographic 1990 and 2010. The employment challenge for dividend.â€? This dividend can be used for high- the region is to absorb these new entrants into priority physical and human capital invest- jobs at rising levels of productivity. ments necessary to absorb the growing number • Aggregate productivity growth in South Asia of entrants into the labor force at rising wages over the last three decades was driven by an and more productive self-employment. The div- extraordinary surge in the growth of total idend can be reaped, however, only if a policy factor productivity (TFP) (a combination of framework is in place that can channel the extra changes in the efï¬?ciency with which inputs are savings into priority investments (including, for used and changes in technology). Its contribu- example, an efï¬?ciently intermediating ï¬?nancial tion was larger than in the “miracleâ€? growth sector and a business environment conducive to firms’ carrying out those investments). In equate access to land among their leading the absence of such a framework, productiv- constraints. Rural-based industry and ser- ity will grow slowly or remain stagnant, and vice ï¬?rms in Bangladesh, Pakistan, and Sri the dividend will go uncashed. The window of Lanka report as a top constraint inadequate demographic opportunity is expected to close transport, which inhibits their access to around 2040 for all countries except Sri Lanka, markets that would make them less depen- where it closed around 2005, and Afghani- dent on local demand. stan, where it will stay open beyond 2040. The • Agriculture will continue to be the largest demographic transition will eventually give employer in much of South Asia for the fore- way to old age dependency, as the share of the seeable future. Boosting TFP growth in the elderly in the population increases. sector through accelerated diversification • Continuance of high economic growth, which into cash crops and high-value activities has been an important driver of improved job will require investment in key public goods. quality, is not assured. Globally, correlations of Investment in agricultural research and country growth rates across decades are low. development has much higher returns than Thus policy reforms required to ease bottle- power, fertilizer, and credit subsidies. necks to improving job quality are needed, • Education reform and action before children irrespective of whether there is a demographic enter school are key. Poor nutrition in early dividend, in order to maintain and increase childhood, where South Asia has the weak- the pace of creation of better jobs, even in est indicators in the world, impairs cognitive lower growth environments. The window of development before children get to school, demographic opportunity lends urgency to the reducing the payoff from subsequent educa- agenda, since policies take time to bear fruit. tional investments. Policy makers must also strengthen the quality of learning at all levels to equip tomorrow’s workers, not only with Message 3: Creating more and better jobs for a academic and technical skills, but also with the growing labor force calls for a reform agenda behavioral, creative thinking, and problem- that cuts across sectors. solving skills employers increasingly demand. • Investing in reliable electricity supply is criti- • Moving away from protecting jobs to pro- cal. South Asian ï¬?rms of all types—rural and tecting workers is essential for formal sector urban, formal and informal—rate electricity as job creation in India, Nepal, and Sri Lanka. a top constraint to operations. Reported power Enterprise managers in the urban formal sec- outages are consistent with reported sever- tor report labor regulations as being a more ity: Afghanistan, Bangladesh, and Nepal have severe constraint to the operation of their busi- some of the highest reported outages in the ness than is the case for countries at their levels world. The gap between demand and supply of per capita income. High costs of dismissing of electricity is large. Reforms need to manage regular workers are, in effect, a tax on hiring the required expansion of capacity efï¬?ciently them. Reforms to encourage job creation in and improve the financial and commercial the formal economy should lower these costs, viability of the power utilities. They involve a which protect a minority of workers. These combination of investment and reform of gov- reforms must go hand in hand with reforms ernance in the sector—both are critical. that strengthen labor market institutions and • Formal urban ï¬?rms cite corruption in interac- programs that formal and informal sector tions with the state, especially in transactions workers can use to help them adjust to labor involving tax administration and utilities, as market shocks and improve their future earn- an important constraint to their operations. ings potential. Building incrementally on exist- • Informal urban ï¬?rms in India report inad- ing schemes is likely the best way forward. Overview 1 T his book investigates how more and quantity of jobs, the quality of jobs, and labor better jobs can be created in South mobility. Asia.1 It does so for two reasons. First, this region will contribute nearly 40 percent of the growth in the world’s working-age Job quantity (15–64) population over the next several Employment grew in South Asia over the decades. It is important to determine what past decade, broadly tracking growth in needs to be done to absorb them into employ- the working-age (15–64) population (ï¬?gure ment at rising levels of labor productivity. 1.1). Lack of safety nets precludes high rates Second, creating more productive jobs—with of open unemployment, which averaged jobs deï¬?ned to include all wage work and self- a little over 3 percent in the region. Thus employment—is the most reliable route out of employment growth tends to broadly mirror poverty for a region that is home to more than growth in the labor force. As the propor- 40 percent of the world’s absolute poor.2 tion of the working-age population that is in The book addresses three major questions. the labor force changes slowly, the growth • Has South Asia been creating an increas- of the labor force tends to track that of the ing number of jobs and better jobs? working-age population. Together these • What determines the quality of job cre- observations imply that employment growth ation, and what is the employment chal- can be expected to broadly reflect that of the lenge going forward? working-age population. • What demand- and supply-side bottle- Among five of the larger countries in necks need to be eased to meet South the region, employment growth since 2000 Asia’s employment challenge in the face of was highest in Pakistan, followed by Nepal intensifying demographic pressure? and Bangladesh, India, and Sri Lanka. Total employment in South Asia (excluding Afghanistan and Bhutan) rose from 473 mil- South Asia’s track record lion in 2000 to 568 million in 2010, creating This section examines South Asia’s track an average of just under 800,000 new jobs record in creating jobs. It looks at the a month. 3 4 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.1 Annual growth in working-age population, employment, and labor force in selected South Asian countries 4 3 percent 2 3.6 3.6 3.3 3.3 3.0 2.5 2.6 2.8 2.3 2.2 2.2 2.3 1 1.0 1.2 1.0 0 Sri Lanka India Bangladesh Nepal Pakistan 2000–10 1985–2010 2000–10 1995–2010 2000–10 working-age population employment labor force Sources: Authors, based on data on working-age population from UN 2010 and data on employment and labor force from national labor force surveys. FIGURE 1.2 Distribution of employment by type in South Asia, by country 100 80 34 36 43 50 62 60 77 75 9 9 80 percent 8 13 40 2 32 57 55 17 0.5 22 1 4 20 2 14 10 14 21 17 21 9 8 0 07 tan 05 sh 20 tan 09 ndia 20 nka 20 ves /0 l 08 tan 07 pa 20 lade 20 Ne 8 6 07 0 08 04 8 9 20 nis u i 20 kis La ald /0 /0 /1 /0 I Bh Pa a i ng M Sr gh 20 Ba Af regular wage or salaried casual wage all wages self-employed (high end) self-employed (low end) Source: Authors, based on data from national labor force and household surveys. Note: The data for Maldives and Sri Lanka do not allow the separation of wage employment into regular wage or salaried workers and casual laborers. In all countries except Maldives and Sri Job quality Lanka, the largest share of the employed are the low-end self-employed (ï¬?gure 1.2). 3 South Asia has created better jobs, deï¬? ned Nearly a third of workers in India and a primarily as those with higher wages for ï¬?fth of workers in Bangladesh and Pakistan wage workers and lower poverty levels for are casual laborers. Regular wage or sala- the self-employed and secondarily as jobs ried workers represent a ï¬?fth or less of total that reduce the risk of low and uncertain employment. income for the most vulnerable group of OVERVIEW 5 workers. By these measures, results have in the average number of months for been positive: which all casual laborers were without work despite looking for it (ï¬? gure 1.5). • Real wages for wage workers—both Thus, the secondary criterion for better casual and regular wage or salaried— jobs—that they should reduce the risk of grew 0.1–2.9 percent a year during vari- low and uncertain incomes for the most ous subperiods between 1983 and 2010 vulnerable—has been met in India. This for which comparisons can be made is not necessarily the case in other coun- (ï¬?gure 1.3). tries in South Asia. (For a discussion of • A higher proportion of self-employed these criteria and the way in which they workers (on whom information on earn- are used to rank jobs by quality, see ings is not available) are now in house- annex 1C.) holds above the national poverty line in Bangladesh, India, Nepal, Pakistan, and Notwithstanding the variation in wages Sri Lanka. This ï¬?gure is used as a proxy and poverty rates across employment types for improving job quality for this segment and their changes over time, there is a of the labor force (ï¬? gure 1.4).4 Increas- stable pattern of association between pov- ing proportions of casual and regular erty and the type of employment that has wage or salaried workers in Bangladesh, been maintained over time. Regular wage India, and Nepal and all wage workers or salaried workers have the highest wages in Pakistan and Sri Lanka are also now and lowest poverty rates; the self-employed in households that are above the poverty have higher poverty rates; and casual work- line. Indeed, poverty rates for all types of ers, especially in agriculture, have the low- workers during all time periods show a est wages and highest poverty rates (see decline when the data are disaggregated annex 1C). by location (rural or urban) or gender. The proportion of workers in different Thus, the primary criterion for better jobs employment types has remained largely is satisï¬?ed. unchanged over time (ï¬? gure 1.6). At this • In India over the period 1999/2000 level of aggregation, better jobs have been through 2009/10, there was a decline created mainly as a result of increasing FIGURE 1.3 Average annual increases in mean real wages in selected countries in South Asia 4 2.9 2.8 percent 1.9 2.0 2 0.1 0 Bangladesh India Nepal Pakistan Sri Lanka 2002–05 1983–2010 1999–2008 2000–09 2000–08 Source: Authors, based on data from national labor force and household surveys. 6 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.4 Percentage of workers in households below the poverty line in selected South Asian countries, by employment status a. Bangladesh, 2000–10 b. India, 1983–93/94 and 1999/2000–2004/05 c. Nepal, 1995/96–2003/04 80 80 80 70 67 70 70 61 58 60 60 60 47 47 47 48 50 50 50 44 43 42 percent percent percent 38 38 39 40 40 40 33 31 28 30 39 29 38 28 30 30 30 30 22 21 29 29 28 27 20 24 20 26 20 21 18 18 10 10 15 10 12 7 0 0 0 2000 2005 2010 1983 1993/94 1999/2000 2004/05 1995/96 2003/04 URP URP MRP MRP all regular wage or salaried all self-employed all casual labor all workers d. Pakistan, 2001/02 to 2007/08 e. Sri Lanka, 1995/96–2006/07 80 80 70 70 60 60 50 50 percent 43 percent 40 40 30 30 25 26 20 23 20 24 13 15 10 10 11 0 0 2001/02 2007/08 1995/96 2006/07 all self-employed, all wages, and all wages all self-employed all workers all workers Source: Authors, based on data from national labor force and household surveys. Note: URP = uniform recall period (the period in which respondents were asked to recall all consumption items over the same recall period [for example, 7 days]). MRP = mixed recall period (the period need not be the same for all items, [for example, 7 days for some and 365 days for others]). Figures are for workers age 15–64. FIGURE 1.5 Average number of months without work in the past year, casual laborers in India, by sector, 1999/2000–2009/10 2.0 1.8 1.9 1.6 1.5 1.5 1.4 1.4 1.4 1.1 months 1.0 0.9 0.5 0 1999/2000 2004/05 2009/10 agriculture rural nonfarm urban Source: Authors, based on data from Indian labor force and household surveys. Note: Figures are for workers age 15–64 who were available for work during at least part of the month. OVERVIEW 7 FIGURE 1.6 Distribution of rural and urban workers in selected South Asian countries, by employment type a. Bangladesh, 2000–2010 b. India, 1983–2009/10 100 100 18 18 18 15 17 90 22 21 21 90 31 33 31 38 36 39 80 35 38 80 70 70 28 45 41 35 36 60 40 42 41 60 percent percent 50 50 40 49 44 40 61 61 47 60 57 54 30 30 51 20 43 43 20 43 41 41 40 43 10 17 10 14 15 8 7 8 8 8 0 0 00 05 10 00 05 10 83 4 0 5 0 83 4 0 5 0 –9 00 –0 –1 –9 00 –0 –1 20 20 20 20 20 20 19 19 –2 –2 93 04 09 93 04 09 19 99 20 20 19 99 20 20 rural urban 19 19 rural urban casual labor self-employed regular wage or salaried c. Nepal, 1995/96–2003/04 d. Pakistan, 1999/2000–2008/09 e. Sri Lanka, 2000–2008 100 100 100 15 13 12 90 18 17 16 16 19 20 19 31 33 80 80 80 45 44 70 60 60 60 61 46 45 45 percent percent 55 percent 50 70 71 79 83 71 40 40 40 69 67 30 55 56 20 20 20 35 36 36 27 26 10 12 14 13 5 4 0 0 0 6 4 6 4 0 8 9 0 8 9 00 08 00 08 /9 /0 /9 /0 00 –0 –0 00 –0 –0 20 20 20 20 95 03 95 03 –2 –2 07 08 07 08 19 20 19 20 99 20 20 99 20 20 rural urban 19 19 rural urban rural urban self-employed casual labor self-employed regular wage or salaried wage worker Source: Authors, based on data from national labor force and household surveys. a. Data from the Bangladesh Household Income and Expenditure Surveys (HIES) were used to calculate worker poverty rates. The share of workers by employment type in the HIES differs from the share in the Bangladesh labor force surveys. The difference is likely to be partly driven by how female employment is captured, with female participation rates in the HIES less than half those reported in the labor force survey. Therefore, the changes in the share of workers by type in Bangladesh from the HIES should be interpreted with caution. For example, between 2005 and 2010 the significant increase in the share of regular wage or salaried work in urban areas was driven largely by changes in the female urban workforce reported in the HIES 2005 and HIES 2010. b. Although there is variation in the shares of casual labor and self-employment in rural areas in India, there is no persistent increase or decline in the shares throughout the whole period (for example, the increase in casual labor between 2004/05 and 2009/10 mostly reversed the decline between 1999/2000 and 2004/05); the share of regular wage or salaried workers remained constant throughout the 25-year period. quality within jobs rather than reallocation offers better jobs than agriculture. Improve- of the labor force across employment ment in job quality has been associated with categories. increasing shares of industry and services Looking across broad sectors, wages in in employment, which includes a growing industry and services are higher than in agri- share of the rural nonfarm economy in rural culture. Thus, the rural nonfarm economy employment. 8 MORE AND BETTER JOBS IN SOUTH ASIA Labor transitions Rural workers who were in agriculture in the ï¬? rst period were more likely to make The broad constancy in the share of work- the transition to a better job—to nonfarm ers across employment types masks labor work—if they had secondary or higher levels mobility at the level of individual workers. of education. This higher mobility is typically Many rural workers in Bangladesh, India, greater for workers who completed upper- and Nepal (the three countries studied in the secondary education. Conversely, workers labor transition analysis in this book) have with less education were more likely to expe- moved from agriculture to the rural non- rience a transition in the opposite direction— farm economy and vice versa. from nonfarm work to agriculture. Workers Education is closely tied to labor mobil- with lower levels of education are more likely ity. Secondary and higher levels of education to lose better jobs than they are to secure increase the ability of workers to move out them, as shown in the higher levels of tran- of agriculture, casual wage jobs, and low- sition bars for lower levels of education in end self-employment to better jobs. Although the right-hand panel compared with the left- analysis was conducted for both rural and hand panel in ï¬?gure 1.7. Workers with higher urban Bangladesh, India, and Nepal, in the levels of education are more likely to move to interest of space only the results for rural better jobs than they are to lose them. India are shown (ï¬?gure 1.7). FIGURE 1.7 Conditional probability of moving into and out of better jobs in rural India, by education and gender, 2004/05–2007/08 a. Transition by men from b. Transition by men from agriculture to nonfarm jobs nonfarm jobs to agriculture 80 80 60 60 percent percent 40 40 20 20 0 0 tio no im e y da er nd er y tio no im e y nd er nd er y ar ar ar ar pr som pr som on low co pp co low co pp im rti im rti n y ry y n y y y ar ar ar ar ar se u se u te te pr pr a a uc uc c ed ed se se c. Transition by women from d. Transition by women from 80 agriculture to nonfarm jobs nonfarm jobs to agriculture 80 60 60 percent percent 40 40 20 20 0 0 io o im e y nd er nd er y io o im e y nd er nd er y ar ar ar ar pr som pr som n n co low co pp co low co pp im rti im rti n y y y n y y y ar ar ar ar ar ar se u se u te te pr pr at at uc uc ed ed se se Source: Authors, based on data from national labor force and household surveys. Note: The probability of transition is conditional on being in a specific type of employment in the first period (e.g., panel a shows the estimated probability that a rural male worker who was in agriculture in the first period is working in a nonfarm job in the second period). The probability differs by the level of education of the worker. Upper and lower bounds of the estimated probabilities are shown. The blue lines are drawn through the midpoints of the bounds. OVERVIEW 9 Determinants of job quality and second only to that of East Asia (ï¬?gure 1.8). the employment challenge But growth experiences have varied within South Asia (figure 1.9). Growth in GDP Improving job quality for most segments per capita accelerated, particularly since of the labor force can usually occur only in the 1980s, in Bangladesh and India. It stag- a growing economy. South Asia has seen nated in Nepal and was marked by volatility an acceleration of growth in gross domes- around a broadly declining trend over the last tic product (GDP) per capita over the three four decades in Pakistan. Sri Lanka witnessed decades since 1980; its growth has been an acceleration of growth over the last ï¬?ve FIGURE 1.8 Annual growth in GDP per capita, by region, 1960s–2000s 10 8.4 8 6.7 6.0 5.9 6 4.7 4.5 percent 3.8 4.0 3.7 4 2.8 2.7 3.1 2.8 3.0 2.5 2.3 2.5 2 1.6 1.2 1.4 0.5 0.1 0 –0.2 –0.4 –0.9 –2 1961–70 1971–80 1981–90 1991–2000 2001–10 East Asia and Pacific Latin America and the Caribbean Middle East and North Africa Sub-Saharan Africa South Asia Source: Authors, based on data from World Bank 2011c. FIGURE 1.9 Annual growth in GDP per capita in South Asia, by country, 1960s–2000s 9 8 7.9 7 6.6 6.6 6.3 6.2 6 5 4.8 4.9 4.6 4.5 4.3 percent 4.3 4.2 4.3 4 3.3 3.4 3 2.9 2.8 2.8 2.2 2.4 2.2 2.1 1.9 2 1.8 1.0 1.2 1.2 1.1 1 0.4 0.5 0 0.0 –1 1961–70 1971–80 1981–90 1991–2000 2001–10 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Source: Authors, based on data from World Bank 2011c. Note: Growth in the earliest available decade for Afghanistan, Bhutan, and Maldives is not based on data for the entire decade because data for the entire decade were not available. Thus Afghanistan 2001–09 is based on 2003–09, Bhutan 1981–90 is based on 1982–90, and Maldives 1991–2000 is based on 1996–2000. 10 MORE AND BETTER JOBS IN SOUTH ASIA decades, except for a dip in the 1980s, and Growth in aggregate labor productivity it managed to avoid the slowdown or stagna- can be decomposed into two factors: tion of the 1970s that affected the other coun- • “Extensiveâ€? growth, comprising growth tries in the region. in physical capital per worker (capital deepening) and growth of human capital Sources of growth per worker (education) • “Intensiveâ€? growth, comprising growth in The marked acceleration in growth in South total factor productivity (TFP), a measure Asia has allowed better jobs to be created. of the efficiency with which inputs are Among industrial and all developing regions combined to produce output. except China, aggregate labor productivity (GDP per worker) grew fastest in South Asia, This decomposition indicates that growth at 3.7 percent a year, between 1980 and in TFP made a larger contribution to the 2008 (ï¬?gure 1.10). This performance repre- growth of aggregate labor productivity in sents a striking turnaround from the preced- South Asia during 1980 –2008 than did ing two decades (1960–80), when aggregate physical and human capital accumula- labor productivity in South Asia grew just tion (see ï¬?gure 1.10). In fact, the contribu- 1.6 percent a year—more slowly than any tion of TFP growth was higher than in the other region, including Sub-Saharan Africa. high-performing East Asian economies FIGURE 1.10 Sources of annual growth in labor productivity, by region, 1960–80 and 1980–2008 9 8 7 6 4.8 5 4 percent 1.3 1.0 3 0.4 0.6 1.1 2.0 0.4 1.3 1.4 2 0.5 1.3 1.6 0.3 0.4 0.4 0.8 1.1 0.4 0.5 2.6 2.9 0.3 0.3 0.4 0.4 2.5 0.5 1 0.3 1.8 0.2 1.0 1.3 1.3 1.2 0.5 0.9 0.8 0.9 0.9 0.5 0.9 0.1 0.4 0.5 0 –0.2 –0.7 –0.7 –1 –2 0 8 0 8 0 8 0 8 0 8 0 8 0 8 0 8 –8 00 –8 00 –8 00 –8 00 –8 00 –8 00 –8 00 –8 00 –2 –2 –2 –2 –2 –2 –2 –2 60 60 60 60 60 60 60 60 19 80 19 80 19 80 19 80 19 80 19 80 19 80 19 80 19 19 19 19 19 19 19 19 South Asia world industrial East Asia less China Latin America Middle East Sub-Saharan countries China Africa physical capital per worker education per worker total factor productivity Source: Bosworth 2010. OVERVIEW 11 FIGURE 1.11 Sources of annual growth in labor productivity in selected countries in South Asia, by country, 1960–80 and 1980–2008 5 4 2.6 3 0.5 0.2 1.2 percent 2 1.4 0.7 0.4 0.3 0.1 1.3 0.4 0.3 2.4 0.3 1 1.4 1.6 0.4 0.9 1.0 1.0 0.2 0.3 0.4 0 –0.3 –1 1960–80 1980–2008 1960–80 1980–2008 1960–80 1980–2008 1960–80 1980–2008 Bangladesh India Pakistan Sri Lanka physical capital per worker education per worker total factor productivity Source: Bosworth 2010. excluding China during their “miracleâ€? of workers—estimated at at least a third of growth years.5 all working-age men—that contributed to The sources of growth varied across the reduction in labor supply and led to ris- countries (ï¬?gure 1.11). In Bangladesh, edu- ing real wages for those left behind (World cation accounted for a fifth of the growth Bank 2010). Declining poverty, which is in aggregate labor productivity. Growth of used as a proxy for improving job quality, TFP was more important in India, reflect- owes less to growth in Nepal than to the ing its increased exposure to external and inflow of worker remittances, estimated at internal competition brought about by trade nearly a quarter of GDP. liberalization and deregulation. Capital deep- ening played a signiï¬?cant role in India and The employment challenge Pakistan. But whereas its contribution rose in India after 1980, it fell sharply in Pakistan, The pressure to create better jobs will inten- accounting for the relative importance of TFP sify very substantially over the next few growth there. Capital deepening was more decades. In its medium-fertility scenario, the important in Sri Lanka, where the share of United Nations projects that the region’s cur- investment in GDP nearly doubled following rent population of 1.65 billion will increase its “big bangâ€? opening up in 1977. 25 percent by 2030 and 40 percent by 2050. Job quality has not been associated Given the region’s generally youthful popu- with accelerated economic growth every- lation, the working-age population is pro- where in the region. In Nepal, for example, jected to increase even more (35 percent by growth in per capita GDP remained at 2030 and 50 percent by 2050). about 2 percent a year during the last three Two scenarios reveal the job creation decades. It was the massive out-migration implications of these demographic changes. 12 MORE AND BETTER JOBS IN SOUTH ASIA In the first, there is no increase in female and human capital accumulation) and less labor force participation rates from current on the extraordinary growth of TFP seen levels. In this scenario, South Asia adds nearly in the last three decades.7 As the region has 1 million entrants a month to the labor force become more open to the international econ- between 2010 and 2030. The proportion- omy, it is importing better-quality capital ate increases are largest in countries with the and intermediate goods at world prices and youngest populations (Afghanistan, Nepal, using standard technology to produce goods Pakistan) and smallest in the single aging that are either sold domestically or exported country in the region (Sri Lanka). in competitive world markets. Inasmuch as Under the second scenario, female labor the technology is widely used internationally, participation rates increase 10 percentage the increases in TFP arising from it will be points by 2030 in Bangladesh, India, and limited to what is routine in global best prac- Pakistan, which together account for 95 per- tice. For a country such as India, which has cent of the region’s working-age population a large internal market, domestic sales could and have the lowest rates of female participa- lead to temporarily larger increases in TFP as tion (31 percent in Bangladesh, 30 percent in less competitive producers exit the market. India, and 22 percent in Pakistan). (This phe- But, even with acceleration in “second-gener- nomenon would be consistent with observed ationâ€? structural reforms, TFP growth is not behavior in Indonesia, the Republic of Korea, likely to continue at the rates triggered by the Malaysia, and Thailand between 1960 and reforms of the 1990s. Hence, a key task for 2000.) Participation rates remain unchanged policy makers will be to create an improving in the rest of South Asia. Nearly 1.2 million enabling environment for factor accumula- new entrants a month join the labor force tion (physical capital deepening and human between 2010 and 2030, intensifying labor capital formation), which, alongside more market pressure in Bangladesh, India, and routine rates of TFP growth, can deliver ris- Pakistan. These projections imply a huge ing wages and declining poverty. increase over the just under 800,000 entrants Aggregate TFP growth could also be a month that joined the labor force between increased as a result of a faster reallocation of 1990 and 2010. labor out of low-productivity agriculture. The Can high economic growth, which has contribution of reallocation to TFP growth been the major driver of improving job quality has been substantially greater in East Asia in some South Asian countries, be expected than in South Asia (ï¬?gure 1.12). Reallocation to continue over the next few decades? The accounted for two-thirds of aggregate TFP historical evidence from around the world growth in Thailand between 1977 and 1996, shows on the contrary that growth rates are a period during which the share of agricul- highly unstable over time: the cross-decade ture in employment fell nearly a third. The rank correlation of growth rates per capita contribution of reallocation to TFP growth in for 94 countries across ï¬?ve decades is a mere China between 1978, when reforms started, 0.1–0.4—and correlations with time periods and 1993 was nearly one-third. During this more than two decades apart are typically period, the share of agriculture in employ- negligible.6 The rarity of sustained growth ment fell more than a ï¬?fth.8 In contrast, real- is underlined by the fact that since 1950, per location contributed 15 percent to aggregate capita GDP has grown at a rate of 7 percent TFP growth in Pakistan and 20 percent in or more—the rate required to double living India between 1980 and 2008, during which standards every 10 years—in only 13 coun- time the share of agriculture fell just under a tries, 9 of them in East Asia (World Bank ï¬?fth in both countries. 2008b). The creation of better jobs also requires Looking forward, productivity growth that labor be moved more rapidly not only in the region will need to rely more on fac- out of agriculture into industry and ser- tor accumulation (physical capital deepening vices but also out of lower-productivity into OVERVIEW 13 FIGURE 1.12 Sources of annual growth in total factor productivity in China, India, Pakistan, and Thailand, by sector and reallocation effects 4 0.2 0.3 3 1.0 0.5 percent 0.6 2 3.1 1.3 0.2 1.4 1 0.4 1.1 0.6 0.3 0.2 0.3 0.1 0.5 0.5 0.2 0.3 0 Pakistan Thailand India China China 1980–2008 1977–96 1980–2008 1978–93 1993–2004 reallocation services industry agriculture Sources: Authors, based on data from Bosworth 2005, 2010; Bosworth and Collins 2008. Note: The contribution of reallocation during a decade is calculated as aggregate TFP growth minus the sum over the three sectors of TFP growth weighted by the share of the sector in GDP at the beginning of the decade. FIGURE 1.13 Median wage and value added per manufacturing worker in India, by firm size and type, 2005 (percentage of median wages/value added per worker in formal firms with 200 or more employees) a. Median wage per worker b. Median value added per worker 100 100 90 90 80 80 70 70 70 66 64 60 60 60 55 53 percent percent 50 50 50 47 40 37 40 37 30 25 30 26 23 19 21 20 18 20 11 10 10 6 0 0 1–4 5–9 10–19 20–49 50–99 100–199 1–4 5–9 10–19 20–49 50–99 100–199 number of workers number of workers formal firms informal firms Sources: Authors, based on data on formal firms from the Annual Survey of Industries and data on informal firms from the National Sample Survey manufacturing surveys. Note: Formal firm with 200 or more employees = 100 percent. higher-productivity ï¬?rms within industry and firms employing one to four workers aver- services. Wage differentials between smaller age one-quarter the levels of ï¬?rms employing and larger ï¬? rms are particularly marked in more than 200 workers (ï¬?gure 1.13).9 India’s manufacturing sector, where both Output per worker and wages are also output and wages per worker in formal much lower in informal ï¬?rms than in formal 14 MORE AND BETTER JOBS IN SOUTH ASIA ï¬?rms within the same size class. In informal fewer than 50 people. As in manufacturing, firms with one to four workers, these mea- the size distribution of ï¬? rms did not change sures are just 25–50 percent of those of formal between 2001 and 2006. ï¬? rms the same size. The difference probably Notwithstanding its declining share of reflects both the higher capital intensity and employment, agriculture will continue to be the higher skill levels at larger firms versus the largest employer among the three broad smaller ones and at formal ï¬?rms versus infor- sectors in most of South Asia for some time. mal ï¬?rms of the same size. Firm-size produc- For this reason, it is important that agricul- tivity differentials exist in other countries, but tural productivity be increased to ensure that they are particularly high in India compared the quality of jobs be improved for workers with East Asia. in the sector (box 1.1). Although output per worker and wages at Accelerating the exit from agriculture to larger, formal ï¬? rms are higher, more than industry and services and enabling industrial 80 percent of employment in manufactur- and service sector ï¬? rms to expand, become ing in India is in micro ï¬?rms (ï¬?rms with 1–4 more productive, and thus pay higher wages workers) and small ï¬? rms (ï¬? rms with 5–49 requires urgent action on a number of fronts. workers), a situation that has persisted over The limited educational attainment of the time (ï¬?gure 1.14). In fact, half of employment labor force, inadequate infrastructure, and is in own-account manufacturing enterprises low capital intensity of most ï¬?rms imply that that do not hire any wage workers. The con- realizing higher TFP growth through the centration of employment in micro and small intersectoral and intrasectoral reallocation of firms is even higher in services, where 96 labor will require substantial investment in percent of workers are in ï¬? rms that employ human and physical capital. FIGURE 1.14 Share of manufacturing employment in India, by firm size and type, 1994–2005 100 90 80 70 60 percent 50 40 30 20 10 0 4 9 9 5 9 0– 9 9 0+ 4 9 9 50 9 10 –99 9 0+ 4 9 20 9 50 9 10 –99 9 0+ 1– 5– –1 –4 10 0–9 19 1– 5– –1 –4 19 1– 5– –1 –4 19 20 20 20 0– 0– 10 20 10 20 10 1994 2000 2005 informal directory manufacturing establishments (at least 1 hired worker and more than 6 workers in total) informal nondirectory manufacturing establishments (at least 1 hired worker but fewer than 6 workers in total) informal own-account manufacturing enterprises (no hired workers) formal Sources: Authors, based on data on formal firms from the Annual Survey of Industries and data on informal firms from the National Sample Survey manufacturing surveys. Note: The data show a small share (1 percent or less) of informal employment in the larger firms as well. OVERVIEW 15 BOX 1.1 Increasing productivity in agriculture The exit of workers from agriculture to industry well suited for small-scale production. Improved and services, whether rural or urban based, is an technology, together with a reduction of implicit important correlate of economic development. It is taxation, could boost the yields of such crops. nevertheless critical to ensure that TFP in agricul- Rising incomes, urbanization, and changing con- ture—the key driver of economic growth over the sumer preferences are creating strong demand for long haul— continues to grow. Increasing agricul- high-value commodities in most South Asian coun- tural TFP is important for two reasons. First, it can tries. The shift has increased incentives to diversify, to provide better jobs for workers who remain in the which farmers across the subcontinent are respond- sector. Second, it allows workers to transition more ing. Agricultural diversiï¬?cation has proceeded most rapidly from agriculture to industry and services, rapidly for fruits and vegetables in Bangladesh, where TFP growth is higher. Bhutan, and Nepal; horticulture, ï¬? shing, and live- Notwithstanding South Asia’s transformation stock in India; and livestock in Pakistan. These from a food deï¬?cit to a food surplus region, the pro- developments have occurred despite the disincentives ductivity of agriculture remains low. India and Paki- created by policies that favor food security crops (rice stan have improved their agricultural productivity and wheat), such as those in India, Pakistan, and Sri over the years; elsewhere in the region, improvements Lanka. A shift from cereal-based to high-value agri- began only in the 1990s, after decades of relative culture requires substantial farm-level investment, stagnation. There is some room to expand area under as well as greater exposure to risk. It is necessary to cultivation in selected rain-fed parts of the region (in widen access to ï¬? nancial and insurance services for Afghanistan, Bhutan, Nepal, and eastern Sri Lanka). many smallholders in order to enable them to partici- There are also some unexploited opportunities for pate in the high-value supply chains. expanding area through watershed development and Core public goods are particularly important in irrigation (in Afghanistan, Bhutan, India, Nepal, and agriculture. Only the public sector can invest in much Pakistan). But the bulk of future growth will have to research and development, because private inves- rely on boosting TFP growth, which has lagged inter- tors are not able to appropriate rents, except in a national best practice. few cases, such as hybrid seeds. Public investment in The key to accelerating TFP growth lies in diver- agriculture has been an important driver of growth sifying into cash crops (tea, sugarcane, cotton, and poverty reduction in India and can provide high spices, and rubber) and high-value activities. Cash returns to investment in South Asia. The highest crops have traditionally been important sources of returns to public spending during the 1970s through agricultural growth and employment in many parts the 1990s tended to be in research and development, of South Asia. They are more labor intensive than roads, education, and irrigation (box table 1.1.1). food staples (mechanization options are limited) and Although marginal returns have diminished over BOX TABLE 1.1.1 Returns to agricultural growth from investments in public goods and subsidies in India, 1960s–90s (percent) Public good 1960s 1970s 1980s 1990s Agricultural research and development 3.12 5.90 6.95 6.93 Road investment 8.79 3.80 3.03 3.17 Educational investment 5.97 7.80 3.88 1.53 Irrigation investment 2.65 2.10 3.61 1.41 Credit subsidies 3.86 1.68 5.20 0.89 Power subsidies 1.18 0.95 1.66 0.58 Fertilizer subsidies 2.41 3.03 0.88 0.53 Irrigation subsidies 2.24 1.22 2.38 — Sources: Fan, Gulati, and Thorat 2008; World Bank staff. Note: — = Not available. (continues next page) 16 MORE AND BETTER JOBS IN SOUTH ASIA BOX 1.1 Increasing productivity in agriculture (continued) time, they remain signiï¬?cant. In contrast, returns to restrictions on marketing arrangements also constrain input subsidies (fertilizer, power, and credit) are gen- productivity growth in agriculture. There is thus a erally low. substantial agenda of institutional reform. Institutional weaknesses such as thin land markets, suboptimal water-use arrangements, and regulatory Sources: World Bank staff; Hazell et al. 2011. FIGURE 1.15 Ratio of working-age to nonworking-age population in South Asia, by country, 1960–2008 2.3 2.1 ratio of working-age population to 1.9 nonworking-age population 1.7 1.5 1.3 1.1 0.9 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Source: Authors, based on data from World Bank 2011c. Cashing the demographic dividend Typically, there is an initial decline in this South Asia’s changing demographic proï¬? le ratio, reflecting a drop in the infant mortal- can help it meet the enormous employment ity rate that precedes a decline in the fertility challenge it faces.10 All South Asian coun- rate. The ratio subsequently increases as the tries are undergoing a process, known as the baby boom caused by the lagged decline in “demographic transition,â€? by which high fer- the fertility rate becomes part of the work- tility and mortality rates are replaced by low ing-age population. The resulting rise in the ones. A key indicator of where a country is share of the working-age to the nonworking- situated in the transition is the inverse depen- age population implies that there are fewer dency ratio, which is the ratio of the working- dependents to support (figure 1.15). The age population to the dependent population. resources saved as a result—the “demographic OVERVIEW 17 dividendâ€?—can be used for high-priority way into the investments in physical and investments. Eventually, as the baby boom human capital that have the highest returns. cohort ages, the demographic transition gives Increased factor accumulation would then way to old age dependency. increase aggregate labor productivity, which Although the inverse dependency ratio would help absorb entrants into the labor has followed the same broad pattern in all market at rising wages and more remunera- of South Asia, there are country differences. tive self-employment. Afghanistan’s ratio started increasing only in Without policy reform, the demographic 2005. Bangladesh’s ratio rose sharply, catch- dividend cannot be increased or used to ing up with India’s in 2003 and exceeding it boost growth and living standards. In this thereafter. The improvement reflected, among event, entrants into the labor market will other things, a very rapid decline in fertility be absorbed at stagnant or slowly rising that was supported inter alia by the country’s levels of productivity, and the potential reproductive health program. Maldives saw of the demographic dividend will remain the fastest increase in the ratio, thanks to its untapped. Policies to improve the environ- plunging fertility rate. Pakistan’s ratio began ment for factor accumulation and raise a gentle climb in the 1980s and Nepal’s in the the quality of physical and human capital 1990s. formation are necessary to create better In its medium-fertility scenario, the jobs whether or not there is a demographic United Nations estimates that the inverse dividend. But the fact that, for most of the dependency ratio will peak for most South region, the window of demographic oppor- Asian countries around 2040. The excep- tunity will be open for only another three tions are Sri Lanka, where it occurred decades lends urgency to the need for policy around 2005, and Afghanistan, where it reform. will still be rising in 2040. Bangladesh, Because of volatile economic growth, Bhutan, India, and Maldives are already an uneven policy framework, and armed experiencing the demographic transition confl ict in a number of countries, there is and therefore have the potential to ben- only a broad correspondence between per eï¬?t from it. Nepal and Pakistan, where the capita GDP growth and the demographic demographic transition started later, have transition. The acceleration in India’s eco- yet to see a dividend. nomic growth started in the 1970s, when its The demographic dividend grows when inverse dependency ratio started its climb the inverse dependency ratio rises more rap- (see figures 1.9 and 1.15). Bangladesh’s idly and peaks at a higher level. This will be acceleration began in the 1980s; during the case if the fertility decline occurs soon the middle of the decade, its ratio began after the decline in infant mortality and is to increase as well. With the exception of a rapid. Policies such as creating an effective slowdown in the 1980s, Sri Lanka has seen reproductive health program and expanding an acceleration of economic growth over female primary and secondary education, nearly ï¬?ve decades since the 1960s, when which reduces family size, can help bring its inverse dependency ratio started rising this about. rapidly. Pakistan, where economic growth The resources made available by the has been volatile around a broadly declining demographic dividend can be used for phys- trend across the decades, and Nepal, where ical capital deepening (electricity, transport) growth has been low and stagnant, have yet and human capital formation (education, to see a demographic dividend. With better skills) if the business environment is condu- policies, their growth performance could cive to making such investments. Policy also improve in the future as the demographic needs to help improve the quality of ï¬? nan- transition takes hold. cial intermediation, so that the increased Except in Nepal and, to a lesser extent private savings of households find their Bhutan, the female employment rate (the 18 MORE AND BETTER JOBS IN SOUTH ASIA ratio of female employment to the female demand for labor? How do the business working-age population) in South Asia is environments of individual South Asian among the lowest in the developing world— countries compare with the rest of the devel- primarily a reflection of the region’s low oping world? rates of female labor force participation. Enterprise surveys ask ï¬? rms to rate the Participation rates are particularly low in severity of inadequacies in the various ele- the three largest countries: Pakistan, where ments of the business environment for their almost four out of ï¬?ve women do not partic- ability to operate and expand their busi- ipate in the labor force, and Bangladesh and ness. These elements, which are external to India, where slightly more than two out of the ï¬? rm and resemble public goods, include every three do not participate. Nonpartici- regulation, physical infrastructure, the pation does not imply inactivity: household availability of skilled labor, macroeconomic duties were cited as the most important rea- conditions, the quality of the judiciary, and son for nonparticipation. A rising propor- crime and corruption. The question takes tion of working-age women in increasingly the form: “How much of an obstacle is X productive employment in the near future to the operation and growth of your busi- would provide a boost to growth in coun- ness?â€? The firm’s response regarding its tries such as Bangladesh and India, which severity—rated on a ï¬?ve-point scale, with 0 are going through the demographic transi- being no obstacle and 4 being a very severe tion. While the demographic transition is obstacle—is a measure of the marginal less advanced in Pakistan, the situation is reported cost imposed by the constraint on no less urgent there since the female partici- the operation and growth of its business. pation rate is the lowest in the region. (For These data can be interpreted as the differ- a comprehensive discussion of options to ence between the ï¬? rm’s proï¬?t in the hypo- improve economic opportunities for women, thetical situation in which the business see World Bank 2012.) environment poses a negligible obstacle to The employment challenge in South its operations and the ï¬? rm’s actual proï¬?t, Asia is one of improving job quality rather given the existing quality of the business than quantity, as job growth over long environment.11 periods tracks the growth of the working- age population. The challenge will be to Power, payments, and politics ï¬? nd better jobs for a workforce whose size will increase 25–50 percent in the coming The three most common binding con- decades. In the presence of policy reform, straints for medium-size urban formal the demographic transition can provide a ï¬? rms in South Asia are electricity, corrup- favorable tailwind in support of economic tion, and political instability (table 1.1 and growth and improving job quality. Policy figure 1.16).12 Although there are some will be needed, however, to address the variations, the top three constraints facing main demand- and supply-side constraints formal, urban firms are common to most to job creation, discussed in the next two countries. In every country except Bhutan sections respectively. and Maldives, electricity is one of the top two constraints; it is the top constraint in India and Sri Lanka. Except for Bhutan, political instability is among the top three Improving an inconducive constraints in all countries where it was business environment included in the survey instrument. In ï¬? ve What constrains the demand for labor in of the eight countries studied, corruption South Asia? What types of policy reform is among the four top constraints cited by would facilitate firm expansion and the urban formal sector ï¬? rms. OVERVIEW 19 TABLE 1.1 Top five constraints reported by South Asian benchmark firm in the urban formal sector, by country South Asia region Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Electricity 2 2 1 1 2 2 1 Political instability 1 1 2 n.a. n.a. 1 3 n.a. Corruption 3 3 3 2 3 4 Tax administration 4 5 5 3 1 Labor regulations 3 4 5 5 Inadequately educated labor 2 5 2 Access to land 4 4 1 Transport 1 3 Government policy uncertainty 5 4 2 Courts 4 5 Crime, theft, and disorder 5 5 Business licensing 4 Macro instability 3 Competition 4 Source: Authors, based on Carlin and Schaffer 2011b (from World Bank enterprise surveys). Note: A benchmark firm is a medium-size manufacturing firm with 30 employees that is domestically owned, does not export or import, is located in a large city, and did not expand employment in the preceding three years. n.a. = Not applicable (question was not asked). Analysis is based on pooled sample of enterprise surveys conducted between 2000 and 2010. Access to finance and tax rates constraints are excluded. Table 1.1 and ï¬?gure 1.16 show the sever- ï¬? rms within the urban formal sector across ity and ranking of constraints for a bench- countries in South Asia. mark firm in the urban formal sector. A benchmark firm is a medium-size manu- Electricity facturing firm with 30 employees that is In most South Asian countries, the cost domestically owned, does not export or imposed on firms by the electricity con- import, is located in a large city, and did straint is among the highest in the world; in not expand employment in the preceding Afghanistan, Bangladesh, and Nepal, it is three years. A comparison across coun- higher than in other countries at similar lev- tries and regions requires that these ï¬? rm els of per capita GDP (ï¬?gure 1.17). Moreover, characteristics, which are distributed dif- the severity of the constraint has increased ferently across countries, be controlled for. over time in India, Nepal, and Pakistan. For instance, if a country has a dominance The downward slope in the ï¬? gure implies of skill-intensive ï¬? rms, the answer to the that although ï¬? rms in richer countries can question on labor skills (“How much of an be expected to make greater demands on the obstacle are labor skills to the operation electricity grid, which would lead to rising and growth of your business?â€?) might be severity of complaints, richer countries can more important than it would be in coun- more than offset those demands in the pro- tries that do not have ï¬? rms requiring such vision of electricity, resulting in lower levels skills. That said, although the severity of of severity at higher incomes per capita. constraints differs between benchmark and The high frequency of power outages in nonbenchmark ï¬? rms (for example, ï¬? rms in South Asia is consistent with the reported the service sector), the ranking of the top severity of the electricity constraint. Indeed, constraints is very similar across types of Afghanistan, Bangladesh, and Nepal have 20 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.16 Severity of constraints reported by South Asian Corruption benchmark firm in the urban formal sector Corruption is among the top ï¬?ve constraints in ï¬?ve South Asian countries (see table 1.1). South Asian benchmark ï¬?rm Firms face high levels of corruption in a political instability 2.6 range of interactions with public ofï¬? cials, particularly for utilities and tax inspections electricity 2.2 (ï¬?gure 1.18). Government interactions that have the highest frequency of bribes vary by corruption 1.8 country (with the proportion of ï¬?rms report- tax administration 1.7 ing such payments in parentheses). government policy 1.4 • A fghanistan: Government contracts uncertainty (43 percent), electrical connections macro instability 1.3 (38 percent) • Bangladesh: Utilities (42–76 percent), tax competition 1.2 meetings (54 percent), import licenses crime, theft, and (51 percent) 1.2 disorder • India: Construction permits (67 percent), access to land 1.1 tax meetings (52 percent), operating licenses (52 percent), electrical connec- customs 1.1 tions (40 percent) inadequately • Pakistan: Electrical connections (71 per- 1.1 educated labor cent), water connections (62 percent), tax labor regulations 1.0 meetings (59 percent). transport 1.0 The high frequency of bribes faced in con- necting to power supply is another dimension business licensing 0.9 of the issue of access to electricity and may courts 0.9 be related to businesses having to compete to secure power (World Bank 2008a). More telecoms 0.7 than half of ï¬? rms in Bangladesh, India, and Pakistan are expected to pay bribes during tax 0 1 2 3 severity of constraint inspections. The tax systems in these countries are complex and create not only high costs of Source: Authors, based on Carlin and Schaffer 2011b (from World Bank enterprise surveys). compliance but also opportunities for corrup- Note: A benchmark firm is a medium-size manufacturing firm with 30 employees that is tion. (Chapter 4 compares the severity of cor- domestically owned, does not export or import, is located in a large city, and did not expand employment in the preceding three years. Analysis is based on pooled sample of enterprise ruption as an obstacle to doing business and surveys conducted between 2000 and 2010. The severity of constraint is rated by firms on a the prevalence of bribes in individual South 5-point scale, with 0 being no obstacle, 1 being a minor obstacle, 2 being a moderate obstacle, 3 being a major obstacle, and 4 being a very severe obstacle. Access to finance and tax rates Asian countries and countries outside the constraints are excluded. region at similar levels of per capita GDP.) Political instability The reported costs of political instabil- some of the highest reported outages in the ity are high in Afghanistan, Bangladesh, world, with virtually 100 percent of firms and Nepal; in all three countries, it is the experiencing them. Predictably, the use of most or second-most severe constraint (see generators to mitigate the effects of uncer- chapter 4 for details). These three countries tain power supply is higher in South Asia have some of the highest reported costs of than elsewhere, with 87 percent of ï¬? rms in political instability in the world. (Chapter 7 Afghanistan, 52 percent in Sri Lanka, and examines the costs imposed by armed con- 49 percent in India having generators. flict on ï¬? rms and workers.) OVERVIEW 21 FIGURE 1.17 Cross-country comparisons of reported severity of electricity constraint and power outages for a benchmark firm 4 a. Severity of electricity constraint Bangladesh Nepal 3 Afghanistan severity of constraint Pakistan 2 Sri Lanka India 1 Maldives Bhutan 0 6 7 8 9 10 11 log of per capita GDP in purchasing power parity dollars b. Power outages percent of ï¬?rms affected by at least one power outage a month Afghanistan Nepal 100 Bangladesh 80 60 India 40 Pakistan Bhutan 20 Maldives Sri Lanka 0 6 7 8 9 10 11 log of per capita GDP in purchasing power parity dollars Source: Carlin and Schaffer 2011b (based on World Bank enterprise surveys). Note: The cross-country regression line shows the relationship between the reported severity of the electricity constraint (panel a) and the percentage of firms experiencing more than one power outage per month (panel b) for a benchmark firm and the log of per capita GDP. The shaded area is the 95 percent confidence interval band around the regression line. Vertical bars show confidence intervals of 95 percent around the reported severity of the electricity constraint (panel a) and the percentage of firms experiencing more than one power outage per month (panel b) for countries in South Asia. Analysis is based on pooled sample of enterprise surveys conducted between 2000 and 2010. For further details, including why some observations in panel b are less than zero or are more than 100 percent, see notes to figures 4.2 and 4.3. 22 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.18 Percentage of firms expected to give gifts to public officials, by type of interaction 100 90 80 70 60 percent 50 40 30 20 10 0 operating import license construction electrical phone water meetings with government license permit connection connection connection tax officials contract Afghanistan Bangladesh Pakistan India South Asia world Source: Authors, based on data from World Bank enterprise surveys. Note: Figures show percent of firms in South Asian countries citing corruption as one of their top three constraints. FIGURE 1.19 Severity of constraints identified by South Asian benchmark (nonexpanding) and expanding firm in the urban formal sector Constraints facing job-creating firms in the urban formal sector telecoms 3.0 Job-creating firms, which are similar in political instability courts 2.5 all respects to the benchmark firm except electricity 2.0 business licensing that they expanded employment during 1.5 the preceding three years, report signifi- corruption transport cantly higher severity in 14 of the 16 busi- 1.0 0.5 ness constraints (ï¬?gure 1.19). The rankings tax administration 0.0 labor regulations are shown in Table 1.2. Job-creating ï¬? rms also report higher levels of mitigation activi- government policy ties, such as using generators as a response inadequately uncertainty educated labor to unreliable electricity supply and paying macro instability customs bribes to navigate a corrupt environment. Job-creating ï¬?rms in the urban formal sec- competition access to land tor perform well in other respects, too. They crime, theft, and disorder engage in research and development, intro- benchmark (nonexpanding) expanding duce new processes and products, sell to mul- tinational companies, offer in-ï¬? rm training, Source: Authors, based on Carlin and Schaffer 2011b (based on World Bank enterprise surveys). and have better-educated managers. The Note: Analysis is based on pooled sample of enterprise surveys conducted between 2000 and 2010. A point farther away from the origin indicates that the business constraint is considered more severe. increase in the cost of constraints for expand- The severity of constraint is rated by firms on a 5-point scale, with 0 being no obstacle, 1 being a ing ï¬? rms versus the benchmark is highest in minor obstacle, 2 being a moderate obstacle, 3 being a major obstacle, and 4 being a very severe obstacle. Survey does not make clear what firms mean by “competition.â€? Only statistically significant India and Pakistan. Job-creating ï¬?rms report differences in severity between the benchmark (nonexpanding) and the expanding firms are shown. lower costs in Afghanistan and Bangladesh. OVERVIEW 23 TABLE 1.2 Top five constraints reported by South Asian benchmark (nonexpanding) and expanding firm in the urban formal sector, by country South Asian region Benchmark firm Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Benchmark firm Benchmark firm Benchmark firm Benchmark firm Benchmark firm Benchmark firm Benchmark firm Benchmark firm Expanding firm Expanding firm Expanding firm Expanding firm Expanding firm Expanding firm Expanding firm Expanding firm Expanding firm Constraint Electricity 2 2 2 2 1 1 1 2 2 2 2 1 1 Political instability 1 1 1 1 2 2 n.a. n.a. 1 1 3 1 n.a. n.a. Corruption 3 3 3 4 3 3 2 1 3 4 4 4 1 Tax administration 4 4 5 5 5 5 3 3 1 4 Labor regulations 3 3 4 4 5 5 5 5 Inadequately educated labor 2 2 5 5 2 2 Access to land 4 5 4 4 1 1 Transport 1 1 3 3 Government policy uncertainty 5 5 4 3 2 2 Courts 4 5 5 Crime, theft, and disorder 5 3 5 3 Business licensing 4 4 Macro instability 5 3 3 Competition 4 4 Source: Authors, based on Carlin and Schaffer 2011b (based on World Bank enterprise surveys). Note: Analysis is based on pooled sample of enterprise surveys conducted between 2000 and 2010. n.a. = Not applicable (question was not asked). Access to finance and tax rates constraints are excluded. of workers to more productive sectors will Although the level of severity is different, require accumulation of physical capital. job-creating ï¬? rms rank constraints in much the same way that benchmark firms do, ranking electricity, corruption, and politi- Constraints facing rural firms cal stability as the top three constraints (see Improving the business environment can table 1.2).13 spur development of the rural nonfarm The severity of the electricity constraint economy, which accounts for an increasing facing urban formal ï¬?rms—as well as urban share of rural employment in many South informal ï¬? rms and rural nonfarm enterpris- Asian countries and, therefore, the creation es—prompts a discussion of the problems of better jobs within it. (See chapter 3 for a facing the sector and the policies and other discussion of the rural nonfarm economy.) initiatives being undertaken to address them Doing so requires an understanding of the (box 1.2). The need to make substantial constraints ï¬? rms in this sector face. investment in electricity is an example of the The severity of constraints reported by point made earlier that a rapid reallocation rural firms in Bangladesh, Pakistan, and 24 MORE AND BETTER JOBS IN SOUTH ASIA BOX 1.2 Options for reforming the power sector in South Asia South Asia is characterized by low levels of access, Improving the financial and commercial viability low consumption per capita, and wide demand- of the power sector supply gaps. Some 600 million people in the region Policy makers can choose from a range of options lack access to electricity—more than 40 percent of to improve the ï¬? nancial and commercial viability of the world total. Access rates range from 44 per- the power sector: cent of the population in Nepal to 77 percent in Sri Lanka. The average annual per capita consumption • Increase the level of tariffs to reflect the cost of for the region is 500 kilowatt hour, lower than any- supply, and rationalize tariffs to address cross- where else in the world except Africa. subsidization. Some countries have not revised Supply has not kept pace with demand, resulting tariffs in years. Others have made progress toward in shortages at peak times ranging from 1 gigawatt achieving cost-reflective tariffs, primarily to reduce (GW) in Bangladesh (13 percent) to 12 GW (10 per- ï¬?scal pressure. All countries offer “lifelineâ€? rates to cent) in India. The toll on the economy is enormous: residential consumers to enable the poor to access in Pakistan, the cost of industrial load shedding at least a minimum quantity of electricity as well as is 400,000 lost jobs; in India, 17 percent of total nominally priced electricity to agriculture consumers capacity is based on expensive diesel generation. to support irrigation and food security. The burden Countries have responded through massive invest- of cross-subsidization falls on industrial and com- ment in expanding generation capacity. India added mercial consumers. Any tariff increase will need to 50 GW of capacity between 2006 and 2011 and ini- ensure that adequate safety nets are in place to min- tiated a series of “ultra megaâ€? (4 GW) generation imize the impact on the poor. Innovative initiatives projects based on competitive bidding by independent such as the separate provision of heavy-duty agri- power producers. Bangladesh plans to develop 9.4 cultural feeders for agricultural needs and regular GW of new generation capacity by 2015. Bhutan has feeders for domestic and industrial purposes in the successfully established public-private partnerships Indian state of Gujarat has allowed transparency of for a large export-oriented hydropower project. agricultural consumption and, by providing reliable Signiï¬?cant institutional reforms have also taken supply to both farmers and rural domestic consum- place since the 1990s. Most countries have unbun- ers, spurred the growth of rural productivity. dled their power sectors or corporatized previously • Reduce losses by improving collection, curbing vertically integrated power utilities. The sector has theft, and improving overall efï¬?ciency. India has been opened to private entry and greater competition initiated incentive schemes such as the Restructured in generation, transmission, and distribution, and new Accelerated Power Development Reform Program regulatory frameworks and independent regulatory (R-APDRP), which aims to limit losses to 15 bodies have been established. The degree of reform percent. varies across countries and across states in India. • Improve the capacity and independence of regula- Sector ï¬? nancial losses across the region are large, tory agencies to ensure transparency and account- resulting from the misalignment of tariffs, the high ability in tariff setting, which continues to be driven cost of power procurement, and high transmission by political exigencies. New initiatives in regula- and distribution losses. In India, the combined cash tion have been put in place, such as the implemen- loss of state-owned distribution companies is more tation of multiyear tariffs in India, which provide than $20 billion a year (total investment needs for certainty regarding the costs for which utilities can 2010–15 are estimated at $300 billion). The sector be held accountable and reduce day-to-day regula- deï¬?cit in Pakistan is estimated at about $2 billion a tory interference. year (total investment needs for 2010–20 are esti- mated at $32 billion). Enhancing the business environment for private Several challenges need to be addressed to alle- investment in power sector viate power shortages and improve service delivery. Each is addressed briefly below. The generation sector has attracted substantial pri- vate interest, but obstacles remain in the form of (continues next page) OVERVIEW 25 BOX 1.2 (continued) procedural bottlenecks (for example, land acquisi- Bhutan to India and, to a limited extent, between tion, environmental and forest clearances, provision Nepal and India. of water for thermal plants); limited technical and ï¬? nancial capacity to implement large projects; and Improving the governance of utilities and the shortage of fuel (both domestic and imported) to strengthening institutional capacity ramp up capacity utilization. The region also needs Steps are being taken to develop strong boards and to improve the operating environment to attract pri- high-quality professional management. Doing so is vate players in transmission and distribution. necessary to transform an organizational culture of risk-averse top-down bureaucratic control to one Exploiting the significant potential of intraregional more suited to commercialization. Some states in energy trade India (such as Andhra Pradesh and West Bengal) have adopted technology initiatives, particularly in One of the most cost-effective options for alleviating metering, and accountability frameworks to improve shortages in the region is increasing intraregional sector performance. energy trade. Such trade has increased in recent years, particularly in the form of hydro exports from Source: World Bank staff. Sri Lanka is compared with the severity TABLE 1.3 Top five constraints reported by micro benchmark firm in reported by urban (formal) ï¬?rms in the same the urban and rural sectors of Bangladesh, Pakistan, and Sri Lanka countries.14 As rural ï¬? rms typically employ Bangladesh, Pakistan, Sri Lanka far fewer than 30 employees, the comparison is with a micro benchmark ï¬? rm—a bench- Constraint Urban, formal Rural mark ï¬?rm with 5 employees. The results are Electricity 1 2 as follows (table 1.3 and ï¬?gure 1.20): Political instability 2 3 Corruption 3 • Rural ï¬? rms report less severe constraints Macro instability 4 1 to their operations than urban firms. Access to land 5 This pattern is not unusual in develop- Transport 4 ing countries, where larger urban firms Source: Authors, based on Carlin and Schaffer 2011a (based on World Bank enterprise surveys). are typically more productive and, during Note: Analysis is based on a pooled sample of enterprise surveys conducted in Bangladesh, Pakistan, and Sri Lanka between 2000 and 2010. Access to finance and tax rates constraints are the course of their expansion, place more excluded. Crime, theft, and disorder as well as competition constraints were not asked in the rural demands on publicly provided services surveys. Macro instability was asked only in Bangladesh and political instability was asked only in Bangladesh and Pakistan. Only the top four constraints are shown for rural firms, as the remaining than rural ï¬? rms. constraints were not reported on average as obstacles. • Rural firms identify electricity as one of the most binding constraints to their operations. They report levels of power outages similar to those for urban ï¬? rms where less than 70 percent of enterprises and use generators more intensively than use electricity from the national grid. urban sector ï¬? rms. In Bangladesh, 73 per- • Unlike their urban counterparts, rural cent of nonmetropolitan nonfarm enter- firms cite transport as one of their top prises have an electricity connection, with four constraints. Firms in Bangladesh and 99 percent of them reporting power out- Sri Lanka complain about the poor con- ages. In Sri Lanka, shortages and unreli- ditions and inaccessibility of rural roads. ability of power are severe in rural areas, Poor transport limits access to larger, 26 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.20 Severity of constraints reported by micro urban markets, forcing rural enterprises benchmark firm in urban and rural sectors of Bangladesh, to sell predominantly to local customers, Pakistan, and Sri Lanka which limits the size of the market for their products and goods.15 Rehabilitat- telecoms ing and maintaining existing rural roads electricity 3.0 labor regulations as well as building new roads would open 2.5 up opportunities for rural ï¬? rms. political instability 2.0 customs 1.5 1.0 Constraints facing informal urban firms corruption transport 0.5 0.0 Enterprise surveys in the informal and for- inadequately mal urban sectors in India make it possible macro instability educated labor to compare the constraints encountered by informal firms with those facing formal access to land business licensing firms. The comparison is between formal government and informal urban benchmark firms. As tax administration policy uncertainty courts the median ï¬? rm in the informal urban sec- tor has fewer than 30 employees, the bench- micro benchmark (urban) micro benchmark (rural) mark ï¬? rm size is 5 rather than 30.16 There are both similarities and differences Source: Authors, based on Carlin and Schaffer 2011c (based on World Bank enterprise surveys). between formal and informal firms in the Note: A point farther away from the origin indicates that the business constraint is considered more ranking of business constraints (table 1.4 and severe. The severity of constraint is rated by firms on a 5-point scale, with 0 being no obstacle, 1 being a minor obstacle, 2 being a moderate obstacle, 3 being a major obstacle, and 4 being a ï¬?gure 1.21). Both types of ï¬?rms cite electric- very severe obstacle. Only statistically significant differences in reported severity between the ity as an important constraint. Informal ï¬?rms micro benchmark firms in urban and rural sectors are shown. For further details on the sample and constraints analyzed, see the note for table 1.3. are more likely to cite access to land and trans- port and less likely to cite corruption and tax administration.17 Informal ï¬?rms report access to ï¬?nance as the most severe constraint to the FIGURE 1.21 Severity of constraints reported by micro benchmark operations of their businesses. However, this firm in India’s urban formal and informal sectors ï¬?nding does not necessarily have implications courts electricity 3.0 telecoms 2.5 corruption macro instability TABLE 1.4 Top five constraints reported by micro 2.0 1.5 benchmark firm in India’s urban formal and informal tax administration 1.0 business licensing sectors 0.5 India 0.0 crime, theft, and disorder competition Formal Informal Electricity 1 1 access to land customs Corruption 2 labor regulations government policy Tax administration 3 inadequately uncertainty Crime, theft, and disorder 4 transport educated labor Access to land 5 3 micro benchmark (formal) micro benchmark (informal) Competition 2 Transport 4 Source: Authors, based on Carlin and Schaffer 2011a (based on World Bank enterprise surveys). Government policy uncertainty 5 Note: A point farther away from the origin indicates that the business constraint is considered more Source: Authors, based on Carlin and Schaffer 2011a (based on World Bank severe. The severity of constraint is rated by firms on a 5-point scale, with 0 being no obstacle, 1 enterprise surveys). being a minor obstacle, 2 being a moderate obstacle, 3 being a major obstacle, and 4 being a very Note: Analysis is based on a pooled sample of enterprise surveys severe obstacle. Only statistically significant differences in reported severity between the micro conducted in India between 2000 and 2010. Access to finance and tax benchmark firms in urban formal and urban informal sectors are shown. For further details on the rates constraints are excluded. Political instability was not asked in the sample and constraints analyzed, see the note for table 1.4. India surveys. OVERVIEW 27 for policy, as “access to ï¬?nanceâ€? is a dimen- technical skills needed for the job. Two out of sion of the business environment that, unlike three employers reported that most of these the judiciary or tax administration, is not a skills are “veryâ€? important but that they public good. The ï¬? rm’s response regarding were only somewhat satisï¬?ed (at best) with the inadequacy of ï¬?nance could simply reflect the graduates’ skills (figure 1.22). (Similar the fact that some ï¬? rms do not have bank- concerns are echoed by Sri Lankan employ- able projects. Other indicators of access (for ers in a survey of the information technology example, the proportion of ï¬?rms using exter- workforce [Sri Lanka Information Commu- nal ï¬?nance) suggest that ï¬?nance may indeed nication Technology Association 2007].) The be an issue for micro and small ï¬?rms in some foundation for many of these skills is estab- countries. lished well before graduates enter the world The concern regarding access to land of work—in primary and secondary educa- expressed by urban informal firms may tion and indeed even earlier. reflect the impact of regulations that shape The wage premium has been rising for the operation of land markets in India. Den- higher levels of education in all countries, sity regulations in India, which limit the even as the supply of educated workers has ratio of floor space to plot area, lead cities to increased. Figure 1.23 presents trends in expand outward instead of upward. Together wage premiums for different levels of educa- with limited accessibility of public transport, tion in three South Asian countries between such expansion can make it more difï¬?cult for about 2000 and 2008 and a longer period for informal manufacturing units to be located India. The premiums reflect the differential where they should be—close to buyers and between the average earnings of a worker suppliers. Relaxing density regulations, with a particular level of educational attain- improving urban transport, and increasing ment and the average earnings of a worker the supply of property might help reduce the with the level of attainment just below (for severity of the land constraint for informal details, see chapter 5). The pattern over time urban ï¬?rms (World Bank 2011a). has been that the premium to lower levels of education has been falling while the premi- ums for upper-secondary and tertiary edu- Improving workers’ skills cation have been increasing. These changes Enterprise managers in the urban formal sec- have been taking place in a context in which tors in Bhutan, India, and Maldives report the educational attainment of the labor force inadequate skills of the labor force among has been rising. The pattern is thus consistent the top ï¬?ve constraints to the operation and with a situation in which the supply of work- growth of their ï¬?rms (see table 1.1). In Bhutan ers at lower levels of education is increasing and Maldives, inadequate skills of the labor faster than demand, whereas the demand for force are among the top two constraints; in workers with secondary or tertiary education India they rank 5th. Firms in Afghanistan is outpacing the increased supply. and Pakistan rate skills constraints among The heterogeneity of the region is reflected the least problematic. in variations across countries. India and Focused employer surveys in India and Sri Nepal have seen increases in premiums to Lanka highlight concerns with the skills of both upper-secondary and tertiary educa- tertiary education graduates. Employers hir- tion. Indeed, the wage premium for ter- ing fresh engineering graduates in India eval- tiary education more than doubled in India uated the degree of importance of a broad between 1999/2000 and 2009/10, despite a range of skills and their level of satisfaction large increase in the share of the labor force with recent hires. They rated behavioral skills with tertiary education. In Nepal, the larg- (teamwork, reliability, leadership, willingness est relative increases in wage premiums were to learn), creative thinking and problem- at the upper-secondary level, suggesting that solving skills, and specific knowledge and demand increases at this level were greater 28 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.22 Employers’ perceptions of skills of recently graduated engineers in India integrity reliability teamwork willingness to learn entrepreneurship self-discipline communication in english self-motivated flexibility leadership responsibility modern tools knowledge math/science/English creativity written communication reading technical skills experiments/data analysis verbal communication basic computer problem solving empathy system design contemporary issues advanced computer customer service 0 not at all not very somewhat very extremely satisfaction importance Source: Blom and Saeki 2011. than the small increases in supply. In con- people age 15–34 increased in all countries trast, in Pakistan and Sri Lanka, the wage between the late 1990s and early 2000s, premium increased for upper-secondary edu- but it is still low in most countries, ranging cation but decreased for tertiary education, from 2.5 years in Afghanistan to 7.1 years particularly in Sri Lanka. in India. (It is higher in Maldives [7.8 years] Despite significant progress in recent and Sri Lanka [10.2 years].) In many coun- years, the contrast between increasing tries, the picture is considerably worse for demand for higher levels of education and women (see panel a of ï¬?gure 1.24). the educational attainment of the labor Success in school is affected by what hap- force remains stark. Educational attainment pens to children before they enter school. In remains low, particularly in secondary and fact, the greatest payoffs to subsequent edu- tertiary education, with well over a quarter cational investments may well come from of the labor force in all countries except Sri addressing poor nutrition and other factors Lanka lacking any education at all (ï¬?gure in early childhood, before children enter for- 1.24). The average years of education of mal schooling (box 1.3). OVERVIEW 29 FIGURE 1.23 Wage premiums in selected South Asian countries, by level of education a. India and Sri Lanka 60 50 43 42 42 43 43 42 40 37 36 36 36 percent 31 32 33 33 30 30 26 26 20 19 21 18 20 17 14 16 15 10 13 10 0 1993/94 1999/2000 2004/05 2009/10 2000 2008 India Sri Lanka 60 b. Nepal and Pakistan 40 37 35 32 34 34 percent 32 28 26 25 24 22 23 24 22 20 14 8 10 11 13 9 0 1998/99 2007/08 1999/2000 2007/08 Nepal Pakistan incomplete primary complete primary complete lower-secondary complete higher-secondary complete tertiary Source: Authors, based on data from national labor force and household surveys. Note: The first bar for each country-year pair reflects the wage premium for even some primary education relative to no education; the last bar reflects the wage premium for completing tertiary relative to completing upper-secondary education. The education challenge facing South consider the following actions in the school Asia is broad. It includes improving nutri- subsector: tion and other factors in early childhood, increasing attainment from primary to sec- • Address information gaps by developing ondary and higher levels, ensuring equal national assessment systems that provide opportunity for all groups, and equipping reliable feedback on learning. graduates with the skills necessary to suc- • Improve capacity and accountability at the ceed in the world of work. Country priori- school level by devolving greater respon- ties will vary. In Afghanistan and Pakistan, sibility to schools while increasing their achieving universal primary education accountability to local stakeholders. remains a priority, and these countries still • Improve the quality and performance have significant gender disparities in pri- of teachers by engaging in transparent mary education. India, Maldives, and Sri recruitment and development of career Lanka are focusing on expanding upper- and pay systems that build capacity and secondary school. All countries are expand- provide incentives. ing tertiary education. A key priority for all South Asian countries is to improve the As more and more students enter higher quality of learning and skills of graduates levels of education, pressure to expand ter- at all levels. tiary education will intensify. Priorities to Improving the quality of learning in ensure a focus on the quality and relevance primary and secondary schools requires of skills of graduates of both tertiary institu- strengthening incentives and capacity in the tions and preemployment training systems school system. To do so, governments could include the following: 30 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.24 Share of South Asian labor force with no education, with international comparisons a. Female labor force 100 Afghanistan 2008 Bhutan 2003 Pakistan 2000 80 Nepal 1999 Bhutan 2007 Pakistan 2008 India 1994 India 2000 Pakistan 2009 Nepal 2008 India 2005 60 India 2008 Bangladesh 2002 India 2010 percent Bangladesh 2005 Maldives 1998 40 Maldives 2004 20 Sri Lanka 2000 Sri Lanka 2004 Sri Lanka 2008 0 Sri Lanka 2006 0 1,000 2,000 3,000 4,000 5,000 per capita GDP in purchasing power parity dollars b. Male labor force 100 80 Bhutan 2003 Afghanistan 2008 60 Bhutan 2007 percent Nepal 1999 Bangladesh 2002 Maldives 1998 Pakistan 2000 40 Bangladesh 2005 Pakistan 2008 Pakistan 2009 Nepal 2008 India 1994 India 2000 Maldives 2004 India 2005 20 India 2008 India2010 Sri Lanka 2006 Sri Lanka 2000 Sri Lanka 2008 0 Sri Lanka 2004 0 1,000 2,000 3,000 4,000 5,000 per capita GDP in purchasing power parity dollars Sources: Authors, based on data from World Bank 2011c and national labor force and household surveys. Note: The dark line shows the predicted values of the share of the labor force with no education by per capita GDP, based on a cross-country regression (excluding high-income countries). The shaded area is the 95 percent confidence interval band around the regression line. GDP = gross domestic product. OVERVIEW 31 BOX 1.3 The critical role of nutrition in early childhood development The ï¬? rst years of life— long before formal schooling only modestly. This means that high malnutrition begins—are a key period for building human capi- and micronutrient deï¬?ciencies are likely important tal. The beneï¬?ts of health and nutrition early on can contributors to developmental delays in low-income have effects that persist through life; damage from groups in South Asia; they may also be important childhood disease and malnutrition in terms of lost factors for overall cognitive development of the opportunity for learning can be difï¬? cult to undo. broader population. Income growth alone will not Low levels of cognitive development in early child- eliminate malnutrition and micronutrient deï¬? cien- hood are often strongly correlated with low socio- cies; focused attention is needed. economic status and malnutrition. In addition to Most South Asian countries do not have inte- protein-energy malnutrition, micronutrient deï¬?cien- grated policy frameworks for early childhood devel- cies, which often begin before birth, can impair cog- opment. Early childhood interventions—which nitive and motor development and therefore school include nutrition, hygiene, early cognitive stimula- outcomes. tion, and preschool programs—are among the most South Asia has some of the highest rates of cost-effective investments for improving the quality malnutrition in the world, as well as high levels of and efï¬? ciency of basic education, as well as labor anemia and iodine deï¬?ciency. As measured by stunt- market success. ing, underweight, and wasting, it has the world’s India has the strongest enabling policy frame- highest prevalence of malnutrition in children under work in South Asia, with a foundation deriving from ï¬?ve (box ï¬? gure 1.3.1). a Indeed, malnutrition rates the constitution, among other sources. National are higher than in Sub-Saharan Africa. South Asia nutrition policies and programs in Bangladesh and also has high levels of anemia and iodine deï¬?ciency. some aspects of public health campaigns related to In South Asia, as globally, rates of malnutrition, nutrition in Pakistan have contributed to improve- anemia, and iodine deï¬?cits improve with wealth, but ments in nutrition in both countries, but neither BOX FIGURE 1.3.1 Percentage of children under five with malnutrition, by region and country 70 60 59 50 47 48 49 47 42 44 42 42 40 39 40 38 percent 32 31 33 32 30 26 23 25 21 20 19 21 19 20 17 15 13 14 12 13 14 12 10 11 10 9 9 78 5 4 2 6 0 ea d c Af and ld ca ka s an an h a l an sia pa ive iï¬? di bb n es or fri an ut ist ist Ea Cari ica a hA ac n a In Ne lad ald W ric nA rth st Bh an iL k dP Pa ut No e Ea ng M Sr r gh ra th Ame So an Ba ha Af dl ia Sa id As tin e M b- La st Su stunting underweight wasting Source: Authors, based on World Bank 2011c. (continues next page) 32 MORE AND BETTER JOBS IN SOUTH ASIA BOX 1.3 The critical role of nutrition in early childhood development (continued) country has a national policy on early childhood mother-child interactions through home visits. The development. Lady Health Workers Programme in Pakistan—a Ver y few prog rams at scale seek to inte- community-based government preventive care grate early childhood interventions. Several pilot program—holds promise for promoting nutrition projects can serve as a laboratory for design- and child care through scaling up of carefully eval- ing cost-effective programs, but many lack care- uated cost-effective pilot designs. ful plans for evaluation. Potentially promising efforts are under way, however. For example, Source: Authors, based on Alderman 2011. pilots run by the International Center for Diar- a. Wasting, stunting, and underweight indicators refer to the proportion of rheal Disease Research (ICDDR) in Dhaka have children under five whose weight for height, height for age, and weight for age respectively are more than two standard deviations below the medians stronger evaluation designs and have proven the of an international reference population recognized by the World Health feasibility of promoting better parenting and Organization (WHO). • Provide information on the qualit y Beyond protecting the basic rights of of graduates of institutions and their workers, labor market institutions have employability, and strengthen quality an important role to play in regulating the assurance and accreditation. employment relationship, with potentially • Increase the role of the private sector in important implications for labor market provision and that of employers in the efficiency and social protection. Employ- management of public institutions. ment protection legislation covers the kinds • Increase the autonomy of public higher of contracts permitted and the conditions education institutions and improve incen- and procedures for termination. Restric- tives for improved performance, such as tions on nonpermanent hiring and employer those provided by moving from histori- dismissal rights can increase employment cally negotiated budgets to performance- security and provide protection to work- based approaches. ers from arbitrary dismissal by employ- • Increase contributions from students while ers. If excessively restrictive, however, they protecting students less able to pay. can discourage formal job creation, limit the efï¬? cient reallocation of labor, and fail to provide real protection, as employers Reforming labor market find ways around the rules. The evidence institutions suggests that, in some countries, notably Labor market policies in South Asia need to India and Sri Lanka, the efï¬?ciency costs of strike a balance between facilitating protec- employment protection legislation outweigh tion for the vast majority of workers, primarily the beneï¬?ts in terms of worker protection in the informal sector, who are not covered by (see chapter 6). social protection instruments while enhancing Labor market policies in India, Nepal, their incentives for income generation. Efforts and Sri Lanka are oriented toward protect- are needed on two fronts: (a) reforming statu- ing jobs. India, for example, has employment tory regulations and institutions to encour- protection laws that are considerably tighter age job creation in the formal economy while than laws in Western countries and most protecting the fundamental rights of covered other major emerging economies (figure workers and (b) building on programs that 1.25). India, Nepal, and Sri Lanka require can help informal workers adjust to labor not only notiï¬?cation but prior approval by market shocks and improve their productivity the state to lay off or retrench workers, indi- and future earnings potential. vidually or collectively. In Sri Lanka, prior FIGURE 1.25 Employment protection indicators in selected countries OECD employment protection benchmark score 4 3.5 3 0.3 2.5 2.19 1.6 0.6 1.6 0.9 1.2 1.1 1.6 2 1.3 1.1 0.0 0.9 0.8 0.3 0.5 1.5 1.35 0.7 0.9 0.6 1.6 0.0 0.4 0.5 0.3 0.0 0.2 1 0.88 0.1 0.1 0.5 0.5 0.3 0.6 0.3 1.8 1.7 1.5 0.1 0.4 0.5 0.1 0.0 1.4 0.5 1.1 1.2 1.0 1.0 1.1 1.2 0.5 0.6 0.6 0.9 0.7 0.7 0.4 0.5 0.6 0 0.1 a es a m d ile . ico l ey y ain g ca n il a sia k ep ga di an az ar ad in ur tio lan do at fri rk Ch ne ex In R Sp Ch tu nm Br bo rm n ra St Tu hA a ng a, Ca do M r Ze de Po m Ge De re d Ki ut In ite xe Ko Fe w So d Ne Lu Un te ian i Un ss Ru selected OECD countries selected non-OECD countries protection of permanent workers against individual dismissal (1) speciï¬?c requirement for collective dismissals (2) regulation of temporary employment (3) OECD average individual (1) OECD average individual + collective (1+2) OECD average total (1+2+3) Source: OECD 2009. Note: OECD = Organisation for Economic Co-operation and Development. The OECD employment protection indicators cover three aspects of employment protection: individual dismissal of permanent workers, regulation of temporary employment, and specific requirements for collective dismissal. Each subindicator ranges from 0 to 6, with 0 the least restrictive and 6 the most restrictive. The overall indicator—the sum of the three subindicators, weighted at 5/12, 5/12, and 2/12—also ranges from 0 to 6. 33 34 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 1.26 Weeks of wages required to be paid in severance in regions, country income groups, and selected South Asian countries, by length of service a. International comparators by region 98 100 80 60 54 weeks 47 50 43 43 40 29 33 29 27 25 26 25 16 16 17 21 21 21 20 14 15 12 14 4 4 3 3 6 10 4 11 6 9 2 11 4 4 5 11 0 Af ran tin aciï¬? ia be d Af and on lo at for sia an a l an h ka pa di es ib an P s an ist ist iti d tA hA In Ne a Am c an a en n ha ve per n lad ric ric ar ca rth st pm io ns De o tio an iL k an Eas Sa Pa ut No e Ea t e C eri ng tra Sr d o - sa gh So b- Ba an C i Af ic an dl Su id om Org M th La on Ec b. International comparators by country income group 98 100 80 60 50 54 weeks 47 43 43 40 32 29 25 26 21 21 21 25 20 16 20 17 20 15 12 11 10 6 9 11 5 3 5 4 4 2 4 4 0 a tri e tri e rie - nt le- sia an a l an h pa nt le nk di un m un m es ist ist ou idd ou idd hA In Ne es es s co inco co inco lad s La rie an k ec m ec m Pa ut ng Sri gh w- - m er- m r- So gh Ba co pe Af lo co w hi in up in lo 1 year 5 years 10 years Source: Holzmann and others 2011. approval is necessary only when the work- Enterprise managers in India, Nepal, and er’s written consent for layoff cannot be Sri Lanka report labor regulations to be an obtained. This is different from India, where important constraint to the operation and large ï¬?rms need prior approval from the gov- growth of their business. In contrast, formal ernment before they can dismiss an employee sector ï¬? rms in Afghanistan and Bangladesh even if the worker consents to the dismissal. rate labor regulations as among the least Severance pay, which typically requires problematic. There is some evidence that minimum tenure in the establishment and labor regulations become more costly to ï¬?rms increases with seniority, is also high in most as per capita GDP rises (ï¬?gure 1.27). India, South Asian countries, particularly in Sri Nepal, and Sri Lanka report higher levels of Lanka but also in Bangladesh, Nepal, and severity of this constraint than other countries Pakistan (ï¬?gure 1.26). at their level of development, however. When Most countries have reasonably flexible asked which labor regulations most affected contracting rules. Bangladesh, Nepal, and the operation of their businesses, nearly one Pakistan place limits on the use of fixed- in three ï¬?rms in India for whom labor regula- term contracts, however, and lack of clarity tions were perceived as a moderate or severe in India has led to widely varying interpre- constraint reported that restrictions on dis- tations, regulations, and practices across missal are a constraint to hiring. About one states. in four cited restrictions on casual work, and OVERVIEW 35 FIGURE 1.27 Cross-country comparison of reported severity of the labor regulation constraint 4 3 severity of constraint 2 Sri Lanka Nepal Maldives Bhutan India Pakistan 1 Bangladesh Afghanistan 0 0 6 7 8 9 10 11 log of per capita GDP in purchasing power parity dollars Source: Carlin and Schaffer 2011b (based on most recent World Bank enterprise surveys). Note: The cross-country regression line shows the relationship between the reported severity of the constraint for a benchmark firm and the log of per capita GDP. The shaded area is the 95 percent confidence interval band around the regression line. Vertical bars show confidence intervals of 95 percent around the reported severity of the constraint for countries in South Asia. Analysis is based on a pooled sample of enterprise surveys conducted between 2000 and 2010. one in five cited constraints on temporary lengthy and costly procedures for resolving work. disputes and grievances provide additional In addition to creating disincentives for incentives for noncompliance. expanding the formal sector, labor market South Asian countries would beneï¬?t from rules in South Asia have failed to protect reorienting their labor market policies from workers for two main reasons. First, these “protecting jobsâ€? to “protecting workers.â€? rules cover much less than 10 percent of the As the region modernizes, an approach that labor force in most countries and less than a moves away from protecting jobs through third even in Sri Lanka, which has the most strong job security laws to encouraging more formalized labor market. Second, the impact flexibility in the labor market, while provid- of labor market rules is weakened by non- ing workers with better tools to manage the compliance. As regulations become more fluctuations of the market, will lead to more costly, ï¬?rms increasingly employ strategies to job creation and offer more protection. circumvent them, reducing de facto protec- This approach requires two coordinated tion from labor market regulations. In India, strategies. The ï¬?rst involves realigning labor for example, medium-size and large ï¬?rms in market regulations and institutions to relax organized manufacturing adjust employment the procedures and costs associated with levels by hiring and terminating contract dismissals; extending the legality of nonper- workers. In Sri Lanka, where statutory sever- manent contracts; improving protection of ance rights are generous, workers often fail to fundamental worker rights; improving the beneï¬?t from them because of nonpayment or efï¬?ciency of dispute resolution and the enforce- partial payment, particularly during periods ment of employment protection legislation; of economic distress. Limited options and and streamlining and clarifying regulations. 36 MORE AND BETTER JOBS IN SOUTH ASIA The second involves strengthening the tools contribute, coverage can be expanded only available to workers in both the formal and if governments are willing to allocate signif- informal sectors to help them manage labor icant subsidies. The level of subsidies would market shocks. These tools include income need to be determined not only in light of support in the event of unemployment and the mandate of the programs but also by active labor market programs, including cost- other calls on budgetary resources. It there- effective training and employment services. fore seems likely that protection will be built Over time, this strategy will create a more incrementally on numerous existing schemes favorable environment for formal sector job and adjusted in light of lessons learned dur- creation. Less restrictive regulations, espe- ing their implementation. cially pertaining to dismissal, could create Well-targeted and well-designed pro- incentives for formal sector job creation; ben- grams can help informal workers smooth efit certain groups, including women; and consumption and enhance their income- encourage compliance with the law. generating potential. Countries in the region In the long run, as the workforce becomes have a variety of training, public works, and better educated and productivity rises, the self-employment assistance programs, oper- formalization gains from such an institu- ated by the government as well as by pri- tional framework could be considerable. In vate and nongovernment sponsors. (Details the short run, however, this strategy would are in chapter 6.) The effectiveness of many benefit only some informal sector workers of these programs is not well understood who might be able to ï¬?nd formal sector jobs because of the lack of evaluation evidence. in a more favorable regulatory environment. However, if well targeted and efficiently The vast majority of workers would continue implemented, such programs can incor- to lack social security beneï¬?ts and disability porate both a safety net perspective (help- and health coverage. ing workers manage income-related risks) One way to close the gap between pro- and an activation perspective (helping them tected and unprotected workers would be to improve their capacity to generate income). extend (statutory) social insurance to infor- A priority is to encourage evaluation and mal workers. India’s Rashtriya Swasthya expand programs that meet standards of Bima Yojna (RSBY) is designed to provide targeting and cost-effectiveness. hospitalization coverage to households below the poverty line. The new pension law in Maldives provides for matching pension con- Creating jobs in conflict-affected tributions for informal sector workers such as areas ï¬?shers, in order to encourage informal sector Measured by the proportion of country- participation in the pension scheme. years in conflict since 2000, South Asia Any such plan needs to take account of is the most conflict-affected major region ï¬? nancing, as well as the effect on the inci- in the world (ï¬?gure 1.28). Four of the top dence of informality itself. Employers may be 10 countries in terms of direct deaths from more likely to seek ways to opt out of formal armed conflict in 2008 were in South contributory systems if they know employees Asia (Afghanistan, Pakistan, India, and can access social protection in other ways; Sri Lanka). Ongoing confl icts affect all of workers themselves may prefer to remain Afghanistan and parts of Pakistan and, at a informal, depending on how programs are lower intensity, India. Nepal and Sri Lanka ï¬? nanced. Thus the potential for social insur- are in postconfl ict status. ance to be extended into the informal sector Confl ict affects the demand for and the will depend a great deal on ï¬? nancing and supply of labor. On the demand side, it affects the scope to which individuals are able to both the incentives and the ability of ï¬?rms to contribute. Given the high prevalence of invest in conflict-affected regions and create workers with very limited or no capacity to jobs. In addition to concerns about security, OVERVIEW 37 FIGURE 1.28 Proportion of country-years in armed conflict, by region, 2000–08 0.6 0.50 0.5 0.4 proportion 0.3 0.20 0.2 0.16 0.15 0.1 0.09 0.05 0.0 sia ca Af and ia ia be d ib an As As fri hA a an nA ric rth st ar ca l st ra ut No le Ea Ea e C ri nt ra th Ame So Ce ha d Sa id nd tin M b- ea La Su p ro Eu Source: UCDP/PRIO Armed Conflict Dataset version 4-2009 (http://www.prio.no/CSCW/Datasets/Armed-Conflict/UCDP-PRIO/Armed-Conflicts-Version-X-2009/). Note: Armed conflict refers to internal armed conflicts between the government of a state and one or more internal opposition groups that result in at least 25 battle-related deaths a year. disincentives to the operation of private ï¬?rms In Afghanistan, 56 percent of the working- may arise from inadequate infrastructure, age population was employed in low-conflict confusing regulations and the poor quality of and 68 percent in high-confl ict provinces. governance in areas affected by confl ict. On These numbers usually reflect higher female the supply side, confl ict affects the capacity labor force participation in confl ict-affected of the population to supply labor—because areas. of security concerns and, over time, disrup- Labor markets in conflict zones differ tions to education, increased mortality and from labor markets in nonconflict areas morbidity, and the loss of job-related skills in several other ways. First, except in Sri and training. At the same time, economic Lanka, jobs in high-confl ict areas are more need and the absence of key income earners likely to remain rural and based on agricul- could lead to an increase in labor supply, par- ture. The terrain in rural areas makes them ticularly among women. more favorable to rebellion (Collier and Armed conflict in South Asia seems to Sambanis 2005). Furthermore, the confl ict be associated with an increase (or a smaller itself delays the structural transformation of reduction) in the share of the working-age the economy. population that is economically active and Second, there is a lower likelihood of mov- employed. Between 1996 (preconfl ict) and ing to better jobs in conflict areas, as the 2004, for example, the proportion of the workforce is more likely to remain engaged working-age population in Nepal that was in unpaid family labor, partly because of the employed increased about 2 percentage higher concentration of employment in agri- points in low-confl ict areas and more than culture and related activities. The workforce 4 percentage points in high-confl ict areas. is also less likely to become regular wage or In India, employment rates in conflict-af- salaried employees because of the relative fected areas fell less than in peaceful areas. absence of employers. 38 MORE AND BETTER JOBS IN SOUTH ASIA Third, the workforce is less likely to be Governments can improve security in con- well educated—although confl ict in Nepal flict situations and help restore livelihoods by seems to have been associated with an implementing disarmament, demobilization, improvement in school access—and there- and reintegration (DDR) programs. Such fore less able to access better jobs were programs, which target excombatants, are they to become available. This outcome underway in Afghanistan and Nepal. They reflects both the negative impact of con- include three broad phases: fl ict on the demand for schooling and the • Disarmament (collecting and disposing of destruction of schools and complementary weapons) infrastructure. • Demobilization (disbanding military Even when it is already over, armed con- structures) flict remains a serious obstacle to job cre- • Reintegration (facilitating the return of ation in South Asia. Almost 60 percent of former combatants to civilian life, the ï¬?rms included in the enterprise surveys rank armed forces, or the police). political instability as a major or severe con- straint to doing business. In Afghanistan, In addition to providing a minimal level ï¬? rms located in areas where conflict is most of security, without which economic recov- violent report that they are more severely ery is virtually impossible, the public sector constrained than their counterparts in more has a potentially important role to play in peaceful areas with regard to infrastructure, creating jobs in the early stages of a post- the regulatory environment, security, and conflict situation. In addition to employ- skills (ï¬?gure 1.29). ment through DDR programs, well-designed FIGURE 1.29 Severity of business environment constraints (average) reported by firms in low-conflict and high-conflict areas of Afghanistan, 2008 political instability electricity crime, theft, and disorder corruption telecoms access to land business licensing transport tax administration courts competition inadequately educated labor customs labor regulations 0 1.0 2.0 3.0 4.0 severity of constraint high-conflict areas low-conflict areas Source: Authors, based on data from the 2008 Afghanistan enterprise survey. Note: High-conflict provinces are defined as provinces in which the number of deaths per 1,000 population caused by terrorist incidents in 2007 was greater than 0.1. By this definition, Baghlan, Ghazni, Farah, Helmand, Kandahar, Khost, Kunarha, Nimroz, Paktika, Paktya, Panjsher, Takhar, Urozgan, Wardak, and Zabul are high-conflict areas. Only differences that are statistically significant at at least the 5 percent in an ordinary least squares regression including firm size and industry fixed effects are shown. OVERVIEW 39 and implemented public works programs barriers to ï¬? rm entry and tackling at least that target rural areas and build or reha- the most blatant pockets of corruption bilitate community infrastructure can (World Bank 2011d). In addition, govern- play an important role. Training and self- ments could facilitate private sector activi- employment assistance programs are also ties by creating “safe economic zonesâ€? that important. provide the needed security, services, and Priority is often initially given to seg- infrastructure in a focused manner and pro- ments of the population that are particularly moting “resource corridorsâ€? in areas rich in vulnerable and sectors that have the poten- natural resources to better link them to the tial to absorb large numbers of workers. Ini- rest of the economy. Although much of the tially, labor market programs and policies in effort of job creation will initially be in agri- confl ict-affected areas need to target three culture and construction, the focus should types of populations with special needs: shift over time from low-skilled agricultural jobs to higher-productivity nonfarm jobs • Excombatants, who need to be integrated and from targeted programs to broad-based into the workforce and given incentives to job creation. refrain from violence The peace dividend for job creation is • At-risk youth and war victims (families potentially large. A striking example is the who lost members, people with physical reduction in unemployment in the Northern or mental disabilities, and households that and Eastern provinces of Sri Lanka during are extremely vulnerable as a result of the the 2002–04 ceaseï¬? re (ï¬?gure 1.30). Unem- lack of a steady stream of income) ployment fell from 13.0 percent to 9.2 per- • People displaced by the confl ict, who may cent in the Northern Province and from wish to return to their homes but need 15.9 percent to 10.5 percent in the Eastern to be able to find jobs and feel secure Province, at a time when the national unem- there. ployment rate decreased from 8.8 percent to Labor market programs are important, 8.3 percent. but fiscal and capacity constraints will limit the potential for direct job creation by the public sector in confl ict environments. Both constraints are likely to be most severe FIGURE 1.30 Unemployment rates in the Northern and Eastern for nationwide confl icts. Thus the govern- provinces of Sri Lanka, 1997–2001 and 2002–04 ment is in a better position to implement 16 employment programs in Sri Lanka, where confl ict was localized, than in Afghanistan 14 or Nepal. International organizations and 12 foreign governments have key roles to play 10 percent in providing funding and building capac- 8 ity, particularly in cases of nationwide 6 confl ict. 4 Given resource constraints and the risks of the politicization of extended public sec- 2 tor involvement, policy makers need to take 0 ce ce de early steps to gradually make the private vin vin i nw ro ro sector a more signiï¬?cant creator of jobs once tio nP nP na er r ste a minimum level of security is achieved. th Ea r No Improving the regulatory environment is important in this regard, but true institu- conflict years (1997–2001) ceasefire years (2002–04) tional transformation could take a genera- tion. Governments could begin by reducing Source: Annual Reports of the Central Bank of Sri Lanka. 40 MORE AND BETTER JOBS IN SOUTH ASIA Conclusion Asia—improved the well-being of workers in South Asia. Rising aggregate labor produc- This overview began by posing three ques- tivity has largely driven this growth. South tions. This section summarizes the answers Asia has seen the fastest growth in total fac- to those questions. tor productivity in the world over the last three decades. But not all countries in the Has South Asia been creating an region have enjoyed rapid growth. Some increasing number of jobs countries that had stagnant or slow growth and better jobs? experienced massive international out-migra- South Asia has created jobs at a rate that tion, which opened up job opportunities for broadly tracks the growth in the region’s those who remained behind. Together with working-age population. Indeed, the rank- a substantial inflow of workers’ remittances, ing among five of the larger countries in the tighter labor market has contributed to the region, in descending order of growth rising real wages and declining poverty. of employment (Pakistan, Nepal, Bangla- South Asia will have to accommodate desh, India, and Sri Lanka), coincides with 1.0–1.2 million new labor market entrants their ranking by growth of the working-age a month between 2010 and 2030—a 25–50 population. percent increase over the average number of But do workers have better jobs? This new entrants in 1990–2010. Finding jobs of book uses two primary measures (increases increasing quality for this massive number in real wages for casual laborers and regu- of new workers represents an enormous chal- lar wage or salaried earners and decreases lenge. With TFP growth likely to return to in poverty rates for the self-employed) and more typical rates for the region as a whole, one secondary measure (mitigation of the policy makers have to create incentives for risk of low and uncertain income arising physical capital deepening and human capital from lack of work for the most vulnerable formation. The relative contribution of these group of workers) to assess their well-being. factors to the growth of aggregate labor pro- Based on the primary criteria, South Asian ductivity will vary by country. workers are indeed better off than they were Faster reallocation of labor from agri- in the earliest period for which comparable culture to industry and services, where TFP data are available. Real wages have risen for growth is higher, can also raise aggregate wage workers, including both casual labor- TFP growth. Such a reallocation of workers ers and regular wage or salaried earners. will need physical capital accumulation (in, Among all groups of workers—rural and for example, electricity and transport) and urban workers, men and women—a higher investment in human capital. Reallocation proportion of the self-employed belong to across sectors needs to be complemented by households that are above the poverty line. moving labor out of low-productivity ï¬?rms in In India—the only country in which the manufacturing and services, where the bulk data permitted the secondary measure to of workers in these sectors are employed. be calculated—the risk of low and uncer- tain income arising from lack of work has What bottlenecks need to be eased to been mitigated over the last decade for the meet the employment challenge given most vulnerable segment of the labor force, intensifying demographic pressure? casual workers. Accommodating new entrants into the labor force at rising levels of productivity will What determines the quality of job require reforms to ease demand- and supply- creation, and what is the employment side bottlenecks to expanding employment. challenge going forward? Investing in electricity to ensure reliable Strong economic growth in the region since power supply is the most important and per- the 1980s—second only to that of East vasive reform. But the reform agenda is not OVERVIEW 41 only about investment. Improvements in the away from protecting jobs for a minority of regulatory framework and governance of the insiders toward protecting the vast majority sector must go hand in hand. Urban formal of workers in the informal sector who lack sector ï¬? rms consistently cite corruption in protection is essential. dealings with the state—in particular, utili- ties and tax administration—among the top A challenging, but feasible agenda constraints to their operations. The proposed reform agenda is challeng- Improving the quality of education is a key ing—but it is feasible, especially given the supply-side priority. Efforts to do so should resources that will be freed up over the start with interventions in early childhood to next three decades as a result of South improve nutritional status and prevent cogni- Asia’s demographic transition. Business as tive impairment before children get to school. usual will accommodate new entrants to Policies should focus on the quality of learn- the labor force, but doing so at stagnant ing at all levels once children enter the school or barely rising levels of productivity will system. Ensuring that graduates are equipped mean that the quality of employment will not only with speciï¬?c knowledge and tech- be poor. The vicious circle of poverty and nical skills but also with the behavioral, low- productivity employment would then problem-solving, and creative thinking skills be drawn around another generation. This increasingly required in the world of work is the price of failure. It is a price that need is critical. Moving labor market institutions not be paid. 42 MORE AND BETTER JOBS IN SOUTH ASIA Annex 1A Summary statistics on South Asian countries TABLE 1A.1 Summary economic statistics of South Asian countries Statistic Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka South Asia GDP per capita, 2009 (at 2005 purchasing power parity dollars) 879a 1,419 4,525 2,993 4,967 1,047 2,358 4,256 2,713 Gross capital formation (% of GDP) 25 24 54 36 53 30 19 25 33 Gross savings (% of GDP) — 39 — 35 35 38 22 24 34 Total population (millions) 30 162 1 1,155 0.3 29 170 20 1,568 Sector share of GDP (%) 1980 Agriculture — 32 44 36 — 62 30 28 35 Industry — 21 15 25 — 12 25 30 24 Manufacturing — 14 4 17 — 4 16 18 16 Services — 48 42 40 — 26 46 43 41 2008 Agriculture 32 19 19 18 6 34 20 13 18 Industry 26 29 46 29 18 17 27 29 29 Manufacturing 16 18 7 16 7 7 20 18 16 Services 42 53 35 54 76 50 53 57 54 Sector share of employment (%) Earliest year available in national labor force surveys — 2002/03 2003 1983 1998 1998/99 1999/2000 2000 Agriculture — 51 80 63 25 77 47 37 60 Industry — 14 3 16 25 10 19 25 16 Manufacturing — 10 1 12 15 6 12 17 11 Services — 35 18 21 49 13 34 38 24 Latest year available in national labor force surveys 2007/08 2005/06 2007 2007/08 2004 2007/08 2008/09 2008 Agriculture 59 47 68 53 17 73 43 31 52 Industry 13 15 7 20 27 11 21 27 20 Manufacturing 5 11 4 12 20 7 13 19 12 Services 29 38 24 26 55 16 36 42 28 Sources: Authors, based on data from Aggarwal 2010; ILO KILM and LABORSTA databases 2010; World Bank 2011c; and national labor force surveys. Note: Totals may sum to less than 100 percent because of employment in unknown or unclassifiable categories. Employment and GDP shares for 2008 are based on most recent year available. — = Not available. a. GDP per capita 2008. OVERVIEW 43 Annex 1B Definition of key labor market terms TABLE 1B.1 Definitions of key labor market terms used in this book Term Definition Employed Persons who worked during at least part of the reference period (typically the last seven days), regardless of whether employment was formal or informal, paid or unpaid. Reference period in Bangladesh, Bhutan, India, Nepal, Pakistan, and Sri Lanka was reference week (generally past seven days). Reference period in Afghanistan and Maldives was past month. Unemployed Persons who did not work in the reference period but actively sought work. Inactive Persons who were neither employed nor unemployed during the reference period. (nonparticipant in labor force) This category includes discouraged workers—people who left the labor force because they believed no jobs were available or did not know how to search. Regular wage or salaried workers Persons who receive regular wages or salary from a job in the public or private sector. These workers are usually on the payroll and usually earn leave and supplementary beneï¬?ts. A signiï¬?cant proportion of these workers are in the public sector (ranging from 27 percent in India to 67 percent in Afghanistan). Casual laborers Persons who are paid on a casual, daily, irregular, or piece-rate basis. These workers typically do not have access to formal instruments of social protection. In rural areas, this category includes landless agricultural workers as well as workers in rural-based industry and services, such as construction. Self-employed Employers, own-account workers, and unpaid family enterprise workers. Except in Maldives and Sri Lanka, this is the largest group of workers in South Asia, where the majority of people work as own-account or family enterprise workers. In rural areas, this category comprises largely farmers working their own land, although many self- employed workers also work in the rural nonfarm sector. High-end self-employed Employers in all occupations and own-account workers and unpaid family workers working as managers, professionals, technicians, and clerks. Low-end self-employed Own-account workers and unpaid family workers working as service workers, skilled agricultural workers, craftspeople, machine operators, and workers in elementary occupations. Source: Authors. Annex 1C What is a “betterâ€? job, and which jobs are “betterâ€?? Two main criteria are used to assess job qual- of stable employment and the associated vari- ity. The primary criterion is higher average ation in income and consumption are of con- earnings. For wage workers, earnings can be cern for casual wage workers, who are typi- assessed using information on average wages. cally the poorest segment of the labor force. This information is not available on the self- Data limitations in all countries except India employed, whose earnings are in the form of precluded a consistent analysis of this second- returns to both labor and capital. As these ary criterion. The primary criterion for better ï¬?gures are not available, poverty rates (the jobs thus guides most of this book. percentage of workers living in households Various additional dimensions of job qual- below the poverty line) are used as a proxy ity are often cited, including access to non- for job quality for this segment of the labor wage beneï¬?ts, access to public social protec- force. Better jobs are thus those associated tion mechanisms, the ability to upgrade skills with higher (average) wage rates and lower and receive training on the job, and the pres- poverty rates. ence of a safe working environment. These The second criterion of job quality looks factors are strongly correlated with wages, beyond average income to its variability. Lack poverty, and job security. 44 MORE AND BETTER JOBS IN SOUTH ASIA Based on the criteria of average wages or workers reported being out of work for poverty rates and the risk of low and uncer- 0.9–1.4 months the previous year. In con- tain incomes, several observations can be trast, the self-employed reported 0.2–0.7 made about where the “betterâ€? jobs are in months out of work, and regular wage South Asia (ï¬?gures 1C.1 and 1C.2): or salaried workers reported virtually no time out of work. • Poverty rates are highest among casual • Among the casual labor force, workers wage workers and lowest among regular in the agricultural sector have the lowest wage or salaried workers. Within the rural average earnings and the highest poverty non-farm sector and in urban areas, regu- rates. Casual workers in rural-based indus- lar wage workers earn 23–59 percent more try and services (the rural nonfarm sector) than casual workers in the four countries— earn 10–50 percent more than casual work- Bangladesh, India, Nepal, and Pakistan— ers in agriculture, even though the skills where these comparisons can be made. In proï¬? les are broadly similar. Urban casual Bangladesh (with the exception of its rural workers earn up to 20–30 percent more non-farm sector), India, and Nepal, regu- than casual agricultural workers in Nepal lar wage or salaried workers have poverty and Pakistan. In India, the risk of uncer- rates that are just a third or less than those tain income arising from inability to ï¬? nd of casual workers. The poverty rates for work is also highest for agricultural casual the self-employed are typically between laborers, 49 percent of whom spent at least those of casual and regular wage workers. one month without work in the previous • In India and Nepal—the two countries for year, with an average of 1.4 months spent which data were available—casual labor is without work. In contrast, 40 percent of also associated with the least stability and rural nonfarm casual labor reported being regular wage work with the most stability. without work for at least 1 month the In India in 2009/10, for example, casual previous year, with an average of 1.1 FIGURE 1C.1 Percentage of workers in households below the poverty line in Bangladesh, India, and Nepal, by employment status a. Bangladesh, 2010 b. India, 2004/05 c. Nepal, 2003/04 agricultural casual 51 urban casual labor 45 agricultural casual labor 50 labor rural nonfarm casual agricultural casual rural nonfarm casual labor 45 31 43 labor labor urban casual labor rural nonfarm casual agricultural 44 25 31 labor self-employed rural nonfarm regular wage rural nonfarm or salaried worker 26 urban self-employed 24 17 self-employed rural nonfarm rural nonfarm self-employed 26 18 urban casual labor 16 self-employed agricultural agricultural rural nonfarm regular self-employed 22 17 10 self-employed wage or salaried worker urban regular wage urban regular wage or or salaried worker 16 14 urban self-employed 9 salaried worker rural nonfarm regular urban regular wage or urban self-employed 14 8 3 wage or salaried worker salaried worker 0 50 0 50 0 50 percent percent percent Source: Authors, based on data from national labor force and household surveys. Note: Figures are for workers age 15–64. Poverty rates for India are based on official poverty lines prevailing until 2010. Using the new official poverty lines for 2004/05 (revised in 2011) would increase poverty rates in rural areas, making the poverty rates of rural workers higher than those of urban workers for the same employment type. The hierarchy in terms of employment type would remain the same. OVERVIEW 45 FIGURE 1C.2 Ratio of rural nonfarm and urban wages to agricultural wages in Bangladesh, India, and Nepal Bangladesh, 2005/06 India, 2009/10 Nepal, 2007/08 agricultural casual agricultural casual agricultural casual 1 1 1 labor labor labor urban urban urban 1.0 1.0 1.3 casual labor casual labor casual labor rural nonfarm casual rural nonfarm casual rural nonfarm casual 1.1 1.3 1.5 labor labor labor urban regular wage urban regular wage urban regular wage 1.3 1.6 1.8 or salaried worker or salaried worker or salaried worker rural nonfarm regular rural nonfarm regular rural nonfarm regular 1.6 1.8 2.0 wage or salaried worker wage or salaried worker wage or salaried worker 0 1 2 0 1 2 0 1 2 Source: Authors, based on data from national labor force and household surveys. Note: Figures are median wages for latest year available. months spent without work. This risk is lower still for urban casual workers, 31 Notes percent of whom reported being without 1. See annex table 1A.1 for summary statistics work for at least one month, with an aver- on the eight South Asian countries age of 0.9 months spent without work. 2. The absolute poor live on less than $1.25 a • Wage differentials between sectors and day in 2005 purchasing power parity dollars. 3. See annex table 1B.1 for a deï¬?nition of key employment types partly reflect educa- terms. Srinivasan (2010) provides a more tional attainment. Workers in indus- detailed account. try and services are more educated than 4. It is not possible to rule out reasons other than workers in agriculture. Regular wage or improving job quality for falling poverty rates salaried workers are more educated than in these households. Such factors include the casual workers. Almost all have some edu- flow of workers’ remittances in Nepal, which cation, and a signiï¬?cant proportion have bring in close to a quarter of gross domestic secondary education or above. The self- product and are estimated to account for half employed are more educated than casual of the decline in national poverty rates, or an workers. Even after accounting for higher increase in hours worked by household mem- skills, however, the majority of industry bers in situations of low wage growth. 5. Young (1994, 1995) notes the dominance of and service jobs still pay more than casual physical and human capital accumulation jobs in the agricultural sector. compared with TFP growth in East Asia. Rodrik and Subramanian (2005) note the In summary, less desirable jobs are found in overwhelming importance of TFP growth in casual employment, with the most precarious India during 1980–99. and lowest-paying jobs held by agricultural 6. These data update the point made in Easterly casual workers. Self-employment is in the and others (1993). middle, with high-end self-employed work- 7. Bosworth, Collins, and Virmani (2007) ers having consumption proï¬?les and poverty emphasize this point in the context of India. rates closer to regular wage workers and low- 8. Reallocation contributed only 5 percent to end self-employed having proï¬?les and poverty TFP growth in China between 1993 and 2004 because of an extraordinarily high rate rates that are closer to those of casual labor- of within-sector TFP growth (more than 6 ers. “Better jobsâ€? are held by regular wage or percent a year in industry). salaried workers in industry and services. 46 MORE AND BETTER JOBS IN SOUTH ASIA 9. The discussion of India’s manufacturing sec- India.â€? Policy Research Working Paper, World tor uses the terms establishments and ï¬?rms Bank, Washington, DC. interchangeably. Bloom, D., and D. Canning. 2008. “Global Demo- 10. The discussion draws on Bloom, Canning, graphic Change: Dimensions and Economic and Rosenberg (2011), which is based on an Signiï¬?cance.â€? Population and Development analysis of the contribution of demographic Review 33: 17–51. change to economic growth at the global Bloom, D., D. Canning, and L. Rosenberg. level in Bloom and Canning (2008). 2011. “Demographic Change and Economic 11. Enterprise surveys provide information only Growth in South Asia.â€? Working Paper 67, about the constraints facing existing enter- Harvard School of Public Health Program prises; they are not useful in understanding on t he Globa l Demog raphy of Ag i ng, constraints perceived by potential ï¬?rms that Cambridge, MA did not enter in the ï¬?rst place. Bosworth, B. 2005. Economic Growth in Thai- 12. The question on political instability was not land: The Macroeconomic Context. Washing- asked in India and Sri Lanka. Tax rates and ton, DC: Brookings Institution. access to ï¬?nance constraints are excluded. ———. 2010. “Update of Bosworth and Collins 13. Only in Pakistan do the rankings by bench- 2003.â€? Unpublished paper, World Bank, mark and expanding ï¬?rms differ: for expand- Washington, DC. ing ï¬?rms corruption and political instability Bosworth, B., and S. Collins. 2003. “The Empir- are the top constraints. ics of Growth: An update.â€? Brookings Papers 14. The samples for the investment climate on Economic Activity 2: 113–206. assessments of rural ï¬?rms are pooled because ———. 2008. “Accounting for Growth: Com- of the limited sample size (500 ï¬?rms) in each paring China and India.â€? Journal of Economic country. Perspectives 22 (1): 45–66. 15. The importance of urban-rural links for the Bosworth, B., S. Collins, and A. Virmani. 2007. nonfarm economy has been discussed in “Sources of Growth in the Indian Economy.â€? other studies. The World Bank’s India pov- NBER Working Paper W1290, National erty assessment (2011b) ï¬?nds that that the Bureau of Economic Research, Cambridge, expansion of the nonfarm sector is more MA. http://www.nber.org/papers/w12901. closely linked to urban than to agricultural Carlin, W., and M. Schaffer. 2011a. “A Com- growth. This ï¬?nding is also conï¬?rmed by a parison of Business Environment Constraints simple multivariate regression analysis using between Formal Sector Firms and Rural and census data from Nepal. Informal Sector Firms.â€? Background study 16. To facilitate comparison, ï¬?rms with more conducted for this book. than 20 employees were dropped from the ———. 2011b. “Which Elements of the Business sample of formal ï¬?rms. Environment Matter Most for Firms and How 17. Informal ï¬?rms also cite “competitionâ€? as a Do They Vary across Countries?â€? Background severe constraint. The survey does not make study conducted for this book. clear what ï¬?rms mean by competition. Central Bank of Sri Lanka. Various years. Annual Reports. Colombo. Collier, P., and N. Sambanis. 2005. Under- References standing Civil War: Evidence and Analysis. Washington, DC: World Bank. Aggarwal, S. 2010. “Labor Input and Its Com- Easterly, W., M. Kremer, L. Pritchett, and L. position: An Industry-Level Perspective.â€? Summers. 1993. “Good Policy or Good Luck? Paper presented at the Worldklems conference, Country Growth Performance and Temporary Cambridge, MA, August. Shocks.â€? Journal of Monetary Economics 32 Alderman, H. 2011. “Early Childhood Develop- (3): 459–83. ment and the Role of Pre-school.â€? Background Fan, S., A. Gulati, and S. Thorat. 2008. “Invest- paper prepared for Regional Quality of Educa- ment, Subsidies, and Pro-Poor Growth in tion Study, South Asia Region, Human Devel- Rural India.â€? Agricultural Economics 39 (2): opment Unit, World Bank, Washington, DC. 163–70. Blom, A., and H. Saeki. 2011. “Employability Gleditsch, N., P. Wallensteen, M. Eriksson, M. and Skill Set of Newly Graduated Engineers in Sollenberg and H. Strand. 2002. “Armed Con- OVERVIEW 47 fl ict 1946–2001: A New Dataset.â€? Journal of UN (United Nations). 2010. World Population Peace Research 39 (5): 615–37. Prospects: The 2010 Revision. New York. Harbom, L., and P. Wallensteen 2009. “Armed World Bank. 2008a. Bangladesh Investment Conflict, 1946 –2008.â€? Journal of Peace Climate Assessment. Washington, DC. Research 46 (4): 577–87. ———. 2008b. The Growth Report: Strategies Hazell, P., D. Headey, A. Pratt, and D. Byerlee. for Sustained Growth and Inclusive Develop- 2011. “Structural Imbalances and Farm Non- ment. Report of the Commission on Growth Farm Employment Prospects in Rural South and Development. Washington, DC. Asia.â€? Background study conducted for this ———. 2010. Large-Scale Migration and Remit- book. tances in Nepal: Issues, Challenges and Holzmann, R., Y. Pouget, M, Vodopivec, and Opportunities. Washington, DC. M. Weber. 2011. Severance Pay Programs ———. 2011a. India Urbanization Review. around the World: History, Rationale, Status, Washington, DC. and Reforms. Social Protection Discussion ———. 2011b. Perspectives on Poverty in India: Paper 62726, World Bank, Washington, DC. Stylized Facts from Survey Data. Washington, ILO (International Labour Organization). 2010. DC KILM and LABORSTA databases. Geneva. ———. 2011c. World Development Indicators. OECD (Organisation for Economic Co-operation Washington, DC. and Development). 2009. Indicators of ———. 2011d. World Development Report 2011: Employment Protection. Paris: OECD. Security and Development. Washington, Rodrik, D., and A. Subramanian. 2005. From DC. “Hindu Growthâ€? to Productivity Surge: The ———. 2012. World Development Report 2012: Mystery of the Indian Growth Transition. IMF Gender Equality and Development. Washing- Staff Papers 52 (2): 193–228. ton, DC. Sri Lanka Information Communication Technol- Young, A. 1994. “Lessons from the East Asian ogy Association. 2007. Rising Demand: The NICs: A Contrarian View.â€? European Eco- Increasing Demand for IT Workers Spells a nomic Review 38 (3–4): 964–73. Challenging Opportunity for the IT Indus- ———. 1995. “The Tyranny of Numbers: Con- try. Colombo. fronting the Statistical Realities of the East Srinivasan, T. N. 2011. “The Utilization of Labor Asian Growth Experience.â€? Quarterly Journal in South Asia.â€? Background study conducted of Economics 110 (3): 641–80. for this book. CHAPTER 2 Growth and Job Quality in South Asia Questions and Findings Questions • Labor productivity growth since 1980 owes more to growth in total factor productiv- • What is South Asia’s recent track record with ity than to accumulation of physical and regard to the quantity and quality of job human capital, reflecting the region’s open- creation? ing up to the world economy and deregula- • What needs to be done to improve the qual- tion. Going forward, creating an enabling ity of jobs in the face of intensifying demo- framework for physical and human capital graphic pressure? accumulation to occur will be important. • Reallocation of workers across sectors has Findings played a comparatively modest role in total • Rapid growth in aggregate output per worker factor productivity growth in South Asia. in much of South Asia since 1980 has been Labor will need to be reallocated more rap- associated with rising real wages for casual idly, not only from agriculture to industry and labor and regular wage or salaried workers services but also from less productive to more and an increase in the proportion of the self- productive units within industry and services. employed above the poverty line (a proxy for Doing so will require investment in physical improved job quality). Larger shares of casual and human capital. workers and regular wage or salaried earners • Much of South Asia is going through the now also belong to households above the pov- demographic transition, where the number of erty line. workers is growing more rapidly than their • Rising real wages and declining poverty are dependents. The resources saved as a result of the primary criteria used to assess job qual- there being fewer dependents to support—the ity. By these measures, jobs improved for all demographic dividend—can be channeled into three types of workers. A secondary criterion high-priority investments, which can raise the is the reduction in the risk of low and uncer- productivity of the larger number of entrants tain incomes for the most vulnerable workers. into the labor force. But only if there is an Data limitations allowed this criterion to be enabling policy framework for doing so. monitored only in India, where it is satisï¬?ed. • Continuance of high growth is not assured: • Large-scale out-migration in countries, where globally, correlations of country growth growth has been slow, has exerted upward rates across decades are low. Structural pressure on real wages, thus beneï¬?ting work- reforms to ease demand- and supply-side ers who remain. bottlenecks to expanding employment are • Employment growth has broadly tracked needed, irrespective of whether there is a the growth of the working-age population, dividend, in order to maintain and improve creating just under 800,000 new jobs a the pace of creation of better jobs even in month between 2000 and 2010. A projected lower-growth environments. But the pros- 1.0–1.2 million entrants will enter the labor pect of reaping the demographic dividend, force every month for the next two decades. which will be available only for the next The employment challenge is to absorb them three decades, lends urgency to the need for at rapidly rising levels of productivity. reform. Growth and Job Quality in South Asia 2 T his chapter looks at the growth con- Growth in per capita GDP has accelerated, text in South Asia in which labor particularly since the 1980s, in Bangladesh market outcomes are embedded. The and India. Bhutan saw generally high growth ï¬?rst section decomposes growth in aggregate starting in the 1980s, albeit with some fluc- gross domestic product (GDP) per worker (or tuations. Maldives also enjoyed high growth, aggregate labor productivity) during the past although it experienced a deceleration three decades into the contributions of phys- between the 1990s and the first decade of ical and human capital accumulation and the 21st century. Per capita growth has been changes in total factor productivity (TFP). marked by volatility around a broadly declin- The second section explores how sources of ing trend since the 1980s in Pakistan and has growth may be different in the future. The stagnated in Nepal. Sri Lanka has witnessed third section examines South Asia’s track an acceleration of growth over the last ï¬?ve record regarding the number and quality decades, except for a dip in the 1980s, avoid- of jobs created. The last section argues for ing the slowdown or stagnation of the 1970s moving ahead quickly with reforms in order that affected the rest of the region. to absorb the rapidly growing number of new entrants to the labor market at rising levels of labor productivity even in situations Aggregate labor productivity growth of lower economic growth. Growth in aggregate output per worker, or aggregate labor productivity, may be decom- Economic growth posed into two factors:2 in South Asia • “Extensiveâ€? growth, comprising growth Improving job quality for most segments of in physical capital per worker (capital the labor force can usually occur only in a deepening) and growth of human capital growing economy. South Asia has seen an per worker (education) acceleration of growth over the three decades • “Intensiveâ€? growth, comprising growth since 1980 that is second only to that of East in TFP—a combination of changes in the Asia (figure 2.1).1 But growth experiences efï¬?ciency with which inputs are used and have varied within South Asia (ï¬?gure 2.2). changes in technology. 49 50 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 2.1 Annual growth in GDP per capita, by region, 1960s–2000s 10 8.4 8 6.7 6.0 5.9 6 4.7 4.5 percent 3.8 4.0 3.7 4 2.8 2.7 3.1 2.8 3.0 2.5 2.3 2.5 2 1.6 1.2 1.4 0.5 0.1 0 –0.2 –0.4 –0.9 –2 1961–70 1971–80 1981–90 1991–2000 2001–10 East Asia and Paciï¬?c Latin America and the Caribbean Middle East and North Africa Sub-Saharan Africa South Asia Source: Authors, based on data from World Bank 2011c. FIGURE 2.2 Annual growth in GDP per capita in South Asia, by country, 1960s–2000s 9 8 7.9 7 6.6 6.6 6.3 6.2 6 5 4.8 4.9 4.6 4.5 4.3 percent 4.3 4.2 4.3 4 3.3 3.4 3 2.9 2.8 2.8 2.2 2.4 2.2 2.1 1.9 2 1.8 1.0 1.2 1.2 1.1 1 0.4 0.5 0 0.0 –1 1961–70 1971–80 1981–90 1991–2000 2001–10 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Source: Authors, based on data from World Bank 2011c. Note: Growth in the earliest available decade for Afghanistan, Bhutan, and Maldives is not based on data for the entire decade because data for the entire decade were not available. Thus, Afghanistan 2001–10 is based on 2003–09, Bhutan 1981–90 is based on 1982–90, and Maldives 1991–2000 is based on 1996–2000. Growth in aggregate labor productivity in the growth of aggregate output per worker South Asia between 1980 and 2008 beneï¬?ted in South Asia and its sources in relation to from rapid growth in TFP, in contrast to the growth in other regions. 3 Figure 2.4 shows 1960 –80 period, when extensive growth the sources of growth for Bangladesh, India, accounted for the bulk of growth in aggre- Pakistan, and Sri Lanka. Together they help gate labor productivity. Figure 2.3 presents illustrate the following points: FIGURE 2.3 Sources of annual growth in labor productivity, by region, 1960–80 and 1980–2008 9 8 7 6 4.8 5 4 percent 1.3 1.0 3 0.4 0.6 1.1 2.0 0.4 1.3 1.4 2 0.5 1.3 1.6 0.3 0.4 0.4 0.8 1.1 0.4 0.5 2.6 2.9 0.3 0.3 0.4 0.4 2.5 0.5 1 0.3 1.8 0.2 1.0 1.3 1.3 1.2 0.5 0.9 0.8 0.9 0.9 0.5 0.9 0.1 0.4 0.5 0 –0.2 –0.7 –0.7 –1 –2 0 8 0 8 0 8 0 8 0 8 0 8 0 8 0 8 –8 00 –8 00 –8 00 –8 00 –8 00 –8 00 –8 00 –8 00 –2 –2 –2 –2 –2 –2 –2 –2 60 60 60 60 60 60 60 60 19 80 19 80 19 80 19 80 19 80 19 80 19 80 19 80 19 19 19 19 19 19 19 19 South Asia world industrial East Asia less China Latin America Middle East Sub-Saharan countries China Africa physical capital per worker education per worker total factor productivity Source: Bosworth 2010. 51 52 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 2.4 Sources of annual growth in labor productivity in selected countries in South Asia, by country, 1960–80 and 1980–2008 5 4 2.6 3 0.5 0.2 1.2 percent 2 1.4 0.7 0.4 0.3 0.1 1.3 0.4 0.3 2.4 0.3 1 1.4 1.6 0.4 0.9 1.0 1.0 0.2 0.3 0.4 0 –0.3 –1 1960–80 1980–2008 1960–80 1980–2008 1960–80 1980–2008 1960–80 1980–2008 Bangladesh India Pakistan Sri Lanka physical capital per worker education per worker total factor productivity Source: Bosworth 2010. • Growth in aggregate labor productivity lowest among all developing and industrial in South Asia during 1980–2008, which regions, including Sub-Saharan Africa. averaged nearly 3.7 percent a year, was During 1980–2008, TFP growth accounted well above the world average and second for more than half of aggregate labor pro- only to that witnessed in China. However, ductivity growth in South Asia. Its contri- excluding 1997 and 1998, the years when bution ranged from just over a third in Ban- East Asia experienced a ï¬? nancial crisis, gladesh and Sri Lanka (a range comparable growth in aggregate output per worker to that in the high-performing East Asian in East Asia less China was higher than economies during 1960–96) to 50–60 per- in South Asia during 1980–96 and lower cent in India and Pakistan. The transforma- during 1999–2008. Productivity growth tion of the role of TFP growth is consistent varied within South Asia during 1980– with the picture of a region responding to 2008, when it ranged from 2 percent in improved policies that exposed it to greater Bangladesh to nearly 4.5 percent in India. internal and external competition.5 • TFP growth in South Asia—which aver- • Capital deepening was 2.5 times higher in aged about 2 percent a year over 1980– East Asia less China than in South Asia dur- 2008, 2.5 times the world average—was ing 1960–80; it was only about 40 percent second only to China’s and nearly twice higher during 1980–2008, when capital that of East Asia less China.4 This rapid deepening accounted for about 35 percent growth represented a striking turnaround of the growth in aggregate labor productiv- from the situation during 1960–80, when ity. The contribution of capital deepening virtually all growth in aggregate output per to labor productivity growth was about worker in South Asia was the result of fac- a third in India and Pakistan, more than tor accumulation and TFP growth was the 40 percent in Bangladesh, and more than GROWTH AND JOB QUALITY IN SOUTH ASIA 53 50 percent in Sri Lanka. The magnitude accelerating TFP growth, which was second of capital deepening in South Asia was 1.5 only to China. Although factor accumulation times the world average during this period. played a less prominent role than in the years • The difference in the contribution of edu- of rapid investment-led growth of East Asia cation in South Asia and in East Asia less China, both capital deepening and edu- less China steadily narrowed between cation were increasingly important sources 1980 and 2008, falling from a factor of of growth in India decade by decade over 2 during 1960–80 to a factor of 1.5 dur- 1980–2008. ing 1980–90 and about 1.3 during 1990– 2000 and 2000– 08. The magnitude of growth in education in South Asia was Demographic transition comparable to the world average during Almost all South Asian countries are expe- 1980 –2008. Within the region, educa- riencing a demographic transition—the tion accounted for 10 percent of growth process by which high fertility and mortal- in aggregate labor productivity in South ity rates are replaced by low ones. 6 A key Asia during 1980 –2008, ranging from indicator of where a country is situated in the low double digits in India, Pakistan, the transition is captured by the inverse and Sri Lanka to more than 20 percent in dependency ratio, the ratio of the working- Bangladesh. age (15– 64) population to the dependent In summary, rising aggregate labor pro- population (people under 15 and over 65) ductivity in South Asia owed a great deal to (ï¬?gure 2.5). Initially, the inverse dependency FIGURE 2.5 Ratio of working-age to nonworking-age population in South Asia, by country, 1960–2008 2.3 2.1 ratio of working-age population to 1.9 nonworking-age population 1.7 1.5 1.3 1.1 0.9 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Source: Authors, based on data from UN 2008. 54 MORE AND BETTER JOBS IN SOUTH ASIA ratio decreases, because the infant mortal- • Potential demographic dividend coun- ity rate falls before the fertility rate. The tries: Bangladesh, Bhutan, India, and ratio subsequently rises, as the baby boom Maldives7 caused by the lagged decline in the fertility • Aging country: Sri Lanka. rate becomes part of the working-age popu- This classiï¬?cation (used later in the chapter lation. The resulting rise in the share of the to project the numbers of entrants into South working-age to the nonworking-age popu- Asia’s labor markets in the coming decades) is lation means that there are fewer depen- chosen to reflect the following considerations. dents to support. The resources saved as a First, the demographic transition is over in result—the “demographic dividendâ€?—can Sri Lanka. Second, with improved policies, be used for high-priority investments. Even- Bangladesh, Bhutan, India, and Maldives, tually, as the baby boom cohort ages, the which though growing rapidly, could beneï¬?t demographic transition gives way to old-age yet more from the demographic dividend. dependency. Third, Nepal, where growth has been stag- Sri Lanka’s inverse dependency ratio nant, and Pakistan, where growth has been reached its peak around 2005. Since then it volatile around a broadly declining trend, has been declining, making it the only aging have yet to see a demographic dividend, while country in the region. Bangladesh’s ratio the demographic transition has barely begun shows a sharp increase since the mid-1980s, in Afghanistan. catching up with India’s in 2003 (the result, The resources made available by a demo- among other factors, of a very rapid decline graphic dividend can be used to deepen in fertility, which was supported by a repro- physical capital (for example by investing in ductive health program) and exceeding it electricity or transport infrastructure) as well thereafter. India’s inverse dependency ratio as human capital (by investment in educa- began to increase in the 1970s. Maldives saw tion and skills training). the fastest increase in the ratio, as a result But the realization of the dividend requires of its plunging fertility rate. In Bhutan, the a supportive policy framework, such as a inverse dependency ratio fluctuated, rising in ï¬? nancial sector that intermediates the addi- the mid-1970s and then falling through the tional savings effectively and a business envi- mid-1990s before rising sharply again. Turn- ronment that provides ï¬? rms with the incen- ing to countries with young populations, tives to make high-priority investments. Nepal’s ratio began to rise in the 1990s. Without policy reform, the demographic Pakistan’s ratio began a gentle climb in the dividend cannot be harnessed to productive 1980s. The inverse dependency ratio started ends. increasing in Afghanistan, the region’s most youthful country, only in 2005. In the medium-fertility scenario in the Sources of future growth United Nations’ population projections, the Looking forward, productivity growth in the ratio of the working-age to the nonworking- region will ï¬? rst need to rely more on factor age population in South Asia is expected to accumulation (physical capital deepening peak around 2040, except in Afghanistan, and human capital accumulation) and less where the ratio will still be increasing, and in on the extraordinary growth of TFP seen Sri Lanka, where it has already peaked. Thus in the last three decades. As the region has the demographic window of opportunity become more open to the global economy, will close after 2040 for most South Asian it is importing better-quality capital goods countries. and intermediate goods at world prices Trends in each country suggest a classiï¬?ca- and using standard technology to pro- tion into three groups: duce goods that are sold domestically or • Young countries: Afghanistan, Nepal, and exported in competitive world markets. Inas- Pakistan much as the technology used is widely used GROWTH AND JOB QUALITY IN SOUTH ASIA 55 internationally, the increases in TFP arising employment by broad sector (agriculture, from exports will be limited to what is rou- industry, and services) have changed more tine in global best practice. For a country slowly than shares of value added in South such as India, which has a large domestic Asia. In 2008, India was an outlier in having market, domestic sales could lead to larger too large a share of workers in agriculture for increases in TFP as less competitive produc- its income level. Although the share of GDP ers exit the market. Even with acceleration provided by agriculture fell by almost half in “second-generationâ€? structural reforms, between 1983 and 2008, the proportion of TFP growth, although still an important employment fell by only 20 percent. driver of long-run economic growth, is not The comparison of the shares of employ- likely to expand at the rates triggered by ment and GDP relative to average develop- the reforms of the 1990s. Hence, a key task ment experience is captured more formally by for policy makers is to create an improving comparing South Asian countries to a bench- enabling policy framework within which mark for market economies. The benchmark physical capital deepening and human capi- is derived by regressing the shares of employ- tal formation can take place. Such a frame- ment and GDP in each sector against per cap- work is needed to absorb the growing num- ita GDP, its square, and a measure of country ber of entrants into the labor force at rising size, represented by its land area, for nearly 55 productivity levels. industrial countries and emerging economies The transfer of underutilized labor from in 2008. Figure 2.7 shows the market econ- agriculture to the rest of the economy yields omy benchmarks for agriculture and services. reallocation-driven gains in TFP. The share The evolution of the benchmark is consistent of agriculture in employment in South Asia with the stylized facts in the development has generally fallen more slowly than its literature—namely, that as per capita income share in GDP (ï¬?gure 2.6). Indeed, shares of rises, the share of employment in agriculture FIGURE 2.6 Sectoral shares of GDP and employment in selected countries in South Asia, 1980s–2008 100 90 19 24 25 27 29 80 37 41 3 35 39 47 46 44 2 52 11 54 6 53 57 70 6 9 7 60 4 12 15 7 7 percent 11 9 12 50 7 9 13 12 11 17 7 19 40 14 13 16 16 11 59 67 30 18 54 20 48 16 53 46 20 45 18 31 34 30 33 28 10 19 17 20 13 0 GDP Emp GDP Emp GDP Emp GDP Emp GDP Emp GDP Emp GDP Emp GDP Emp 1984 2008 1983 2008 1980 2008 1981 2008 Bangladesh India Pakistan Sri Lanka services nonmanufacturing industry manufacturing agriculture Sources: Authors, based on data from ILO 2010; World Bank 2011c; and India National Sample Survey. Note: Emp = employment. 56 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 2.7 Shares of agriculture and services in employment and GDP in South Asian and comparator countries, 2008 a. Share of employment in agriculture, 2008 b. Share of GDP in agriculture, 2008 100 50 80 40 IN - 1983 NP - 2008 NP - 2008 IN - 1983 60 BGD - 1984 30 BGD - 1984 PK - 1980 IN - 2008 percent percent PK - 1980 SL - 1981 BGD - 2008 PK - 2008 Thailand 40 SL - 1981 20 PK - 2008 Philippines BGD - 2008 SL - 2008 IN - 2008 Philippines Thailand 20 Brazil 10 SL - 2008 Malaysia Malaysia Brazil Korea, Rep. Korea, Rep. United States United States 0 Singapore United Kingdom 0 United Kingdom Singapore .5 5 10 15 20 25 30 40 50 .5 5 10 15 20 25 30 40 50 2008 GDP per capita in 2005 purchasing power parity dollars (thousands) 2008 GDP per capita in 2005 purchasing power parity dollars (thousands) c. Share of employment in services, 2008 d. Share of GDP in services, 2008 100 100 United States United States United Kingdom 80 80 United Kingdom Singapore Singapore Korea, Rep. Brazil 60 Philippines Korea, Rep. BGD - 2008 IN - 2008 SL - 2008 percent percent 60 Brazil Malaysia NP - 2008 PK - 2008 Philippines BGD - 1984 SL - 1981 Thailand Malaysia 40 PK - 1980 IN - 1983 40 BGD - 2008 SL - 2008 Thailand PK - 2008 BGD - 1984 SL - 1981 20 PK - 1980 IN - 2008 20 NP - 2008 IN - 1983 0 .5 5 10 15 20 25 30 40 50 .5 5 10 15 20 25 30 40 50 2008 GDP per capita in 2005 purchasing power parity dollars (thousands) 2008 GDP per capita in 2005 purchasing power parity dollars (thousands) Sources: Authors, based on data from Aggarwal 2010; ILO 2010; World Bank 2011c; and India National Sample Survey. Notes: The cross-country regression lines shown are shares of employment and GDP by sector regressed on the log of 2008 GDP per capita and the log of 2008 GDP per capita squared (in 2005 purchasing power parity dollars). Figure excludes the transition economies of Europe and Central Asia. The horizontal axis is on a log scale. declines and the share of employment in ser- share of services in employment was below vices grows and that the share of employment the benchmark by nearly 10 percentage in industry rises and subsequently declines (as points in Sri Lanka, but, unlike in India, the workers move into services). The straight lines share of GDP in services was significantly in ï¬?gure 2.7 direct attention to the change in above the benchmark, indicating much the share of employment and value added in higher output per worker in services. The South Asian countries from 1980 (or the ï¬?rst share of employment in industry was signiï¬?- available year of employment data by sector cantly above the benchmark in Sri Lanka in subsequent to it) through 2008. 2008, but the share regressions for industry The regressions show that the share of ï¬?t the data much less well than those for agri- agriculture in employment in India was 14 culture and services. The share of agriculture percentage points above the benchmark in in GDP in Bangladesh—a predominantly 2008, reflecting among other things the lim- agrarian economy at the time of its creation ited absorption of labor in the nonagriculture in 1971—was already below the benchmark economy compared with countries at India’s by 14 percentage points in 2008 (see annex level of per capita income (see column 1 in 2C for full regression results). annex 2C, table 2C.1). The share of services Reallocation across sectors has played a in employment in India was 15 percentage more limited role in boosting TFP in the two points below the benchmark in 2008. The largest countries of South Asia than it did GROWTH AND JOB QUALITY IN SOUTH ASIA 57 in some East Asian countries. 8 Figure 2.8 FIGURE 2.8 Sources of annual growth in total factor productivity in presents the results of a decomposition of India and Pakistan, by sector and reallocation effects, 1980–2008 TFP growth into within- and between-sector contributions over 1980–2008 for India and Pakistan, the two countries in South Asia for a. India 4 which data on capital stocks by broad sector allow growth accounting to be conducted 0.5 for agriculture, industry, and services. On 3 average, reallocation contributed 20 per- cent to aggregate TFP growth in India and 0.5 percent 15 percent in Pakistan during the period. The 2 0.4 0.3 2.2 share of employment in agriculture fell from 1.3 67 percent in 1983 to 54 percent in 2008 in 1.1 1.2 India and from 53 percent in 1980 to 43 per- 1 cent in 2008 in Pakistan. 0.1 0.3 0.6 0.3 Reallocation was considerably more 0.6 0.4 0.5 0.3 important in East Asian countries such as 0 China and Thailand (ï¬?gure 2.9).9 In China, 1980–90 1990–2000 2000–08 1980–2008 the contribution of reallocation to the growth reallocation services industry agriculture of total factor productivity was nearly 30 per- cent between 1978, when reforms started, and b. Pakistan 1993. The share of agriculture in employment 4 fell by more than a ï¬?fth, from 71 percent to 56 percent, over this period. (Reallocation contributed a mere 5 percent to TFP growth 3 in China between 1993 and 2004, because of the extraordinarily high rate of within-sector percent TFP growth—averaging more than 6 percent 2 a year—in industry.) The share of agriculture 0.1 in employment fell by nearly a ï¬?fth, from 56 0.5 0.2 0.2 0.4 percent in 1993 to 47 percent in 2004. Real- 1 0.2 0.3 0.9 0.3 0.4 location amounted to two-thirds of aggre- 0.6 gate TFP growth in Thailand between 1977 0.6 0.4 0.2 0.3 0 0.1 and 1996, a period during which the share 1980–90 1990–2000 2000–08 1980–2008 of agriculture in employment fell by nearly a third, from 65 percent to 45 percent. reallocation services industry agriculture A comparison between China, India, and Pakistan reveals several patterns. Whereas Source: Authors, based on data from Bosworth 2010. the declining share of reallocation in China Note: The contribution of reallocation during a decade is calculated as aggregate TFP growth minus the across the two subperiods studied reflects sum over the three sectors of TFP growth weighted by the share of the sector in GDP at the beginning of the decade. a steep rise in within-sector TFP growth in industry, the increasing share of reallocation in Pakistan since 1980 reflects an across-the- in GDP in South Asia, output per worker board decline in TFP growth. In India, the in industry and services was three to eight share of reallocation falls, but, in contrast times as large as in agriculture in 2008. to Pakistan, it does so as a result of increas- This is evident from table 2.1, which shows ing, rather than declining, within-sector TFP large differences in output per worker across growth and, in contrast to China, in services the three sectors, with agriculture the low- rather than industry. est and, except in Bangladesh, services the As a result of the slower evolution in highest. The differences in sectoral output the shares of employment than in changes per worker are particularly marked in India. 58 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 2.9 Sources of annual growth in total factor productivity in China, India, Pakistan, and Thailand, by sector and reallocation effects 4 0.2 0.3 3 1.0 0.5 percent 0.6 2 3.1 1.3 0.2 1.4 1 0.4 1.1 0.6 0.3 0.2 0.3 0.1 0.5 0.5 0.2 0.3 0 Pakistan Thailand India China China 1980–2008 1977–96 1980–2008 1978–93 1993–2004 reallocation services industry agriculture Sources: Authors, based on data from Bosworth 2005, 2010; Bosworth and Collins 2008. Note: The contribution of reallocation during a decade is calculated as aggregate TFP growth minus the sum over the three sectors of TFP growth weighted by the share of the sector in GDP at the beginning of the decade. TABLE 2.1 Labor productivity in South Asia and East Asia, by sector, 2008 (per worker, in 2005 purchasing power parity dollars) Total Agriculture Industry Services Ratio Ratio Region/country (1) (2) (3) (4) (3)/(2) (4)/(2) South Asia Bangladesh 4,116 1,754 6,876 5,635 3.9 3.2 India 7,049 2,202 10,368 14,939 4.7 6.8 Nepal 2,577 1,125 3,861 8,691 3.4 7.7 Pakistan 8,287 3,778 11,097 12,430 2.9 3.3 Sri Lanka 12,842 5,257 14,334 17,928 2.7 3.4 East Asia Korea, Rep. of 49,677 17,625 73,013 44,305 4.1 2.5 Malaysiaa 36,156 26,439 60,590 27,335 2.3 1.0 Thailand 14,744 4,324 30,747 15,906 7.1 3.7 Source: Bosworth 2010. a. The utilities industry in Malaysia is included in services rather than industry. Gaps in output per worker between agricul- growth. Reallocation of workers is also ture and the rest of the economy remain necessary to accelerate the movement of in East Asia as well, with larger gaps for resources from low-productivity to high- industry than services. productivity activities within the three Hence an acceleration in the movement broad sectors. However, the limited educa- of resources from agriculture (where TFP tional attainment of the labor force in South growth has been slowest) into industry and Asia, analyzed in chapter 5, implies that services (where growth has been brisker) realizing higher TFP growth through the has the potential to increase aggregate TFP intersectoral and intrasectoral reallocation GROWTH AND JOB QUALITY IN SOUTH ASIA 59 of labor will require substantial investment force participation—the proportion of the in human capital. Deï¬?ciencies in infrastruc- working-age population that is in the labor ture, analyzed in chapter 4, imply that real- force—have moved slowly in South Asia, location will require investment in physical implying that the growth of the labor force capital as well. Creating an enabling envi- has tracked that of the working-age popula- ronment for accelerated physical and human tion. Second, in countries where the lack of capital formation and reallocation of labor social safety nets does not allow the luxury to higher-productivity areas must go hand of open unemployment, the proportion of in hand. Doing so represents the most press- the labor force that is unemployed is low ing growth challenge facing South Asia. and does not change very much. At the mar- gin, additional entrants into the labor force are absorbed into low-productivity occu- The track record on pations. Hence the growth of employment employment moves broadly in line with that of the labor This section examines South Asia’s record force. on the quantity of jobs created, the qual- Taken together, these observations imply ity of jobs created, and the degree to which that employment growth can be expected workers move across job categories. to broadly track growth in the working-age population.10 Total employment in South Asia (excluding Afghanistan and Bhutan) Job quantity rose from 473 million in 2000 to 568 million In all South Asian countries, the number in 2010, an average annual rate of growth of of jobs created has grown broadly in line 1.8 percent, ranging from just over 1 percent with the working-age population, for two a year in Sri Lanka to nearly 4 percent a year reasons (figure 2.10). First, rates of labor in Pakistan.11, 12 FIGURE 2.10 Annual growth in working-age population, employment, and labor force in selected South Asian countries 4 3 percent 2 3.6 3.6 3.3 3.3 3.0 2.5 2.6 2.8 2.3 2.2 2.2 2.3 1 1.0 1.2 1.0 0 Sri Lanka India Bangladesh Nepal Pakistan 2000–10 1985–2010 2000–10 1995–2010 2000–10 working-age population employment labor force Sources: Authors, based on data on working-age population from UN 2010 and data on employment and labor force from national labor force surveys. 60 MORE AND BETTER JOBS IN SOUTH ASIA regular wage or salaried workers and fall- Job quality ing poverty rates for the self-employed. Real Two criteria are used to assess job quality. wages in much of South Asia grew 0.1–2.9 The primary criterion is higher average earn- percent a year during various subperiods ings. For wage or salaried workers, it can be between 1983 to 2010 for which comparisons assessed using information on average earn- can be made (ï¬?gure 2.11). A higher proportion ings. Survey data do not contain information of the self-employed belong to households on earnings of the self-employed, the larg- that are above the national poverty line in est segment of the labor force in South Asia Bangladesh, India, Nepal, Pakistan, and Sri (except Maldives). Changes in poverty rates Lanka (ï¬?gure 2.12). This proportion is used (the percentage of workers living in house- as a proxy for improving job quality for the holds below the poverty line) are used as a self-employed, although falling poverty rates proxy for job quality for this segment of the in households of the self-employed could also labor force. Based on these primary criteria, be a result of an increase in other sources of better jobs are those associated with higher income, such as workers’ remittances (which (average) wage rates for wage workers and are very important in Nepal and somewhat lower poverty rates for the self-employed. important in Bangladesh) or increased hours A secondary criterion of job quality looks worked by household members.14 beyond average income to its variability. Higher proportions of casual work- Variation in income and consumption arising ers and regular wage or salaried workers in from the lack of stable employment exposes Bangladesh, India, and Nepal and all wage workers to the risk of low and uncertain workers in Pakistan and Sri Lanka also belong income. These risks can be major for casual to households above the poverty line (ï¬?gure wage workers. Because data limitations in all 2.12).15 This trend is consistent with the evi- countries in the region except India preclude dence of improving job quality provided by a consistent application of this secondary cri- rising real wages and, in Nepal, the poverty- terion, the primary criterion for better jobs reducing impact of workers’ remittances. guides most of this book.13 Poverty rates for all types of workers dur- The creation of better jobs is reflected in ing all time periods also show a decline when rising real wages for both casual workers and the data are disaggregated by location (rural FIGURE 2.11 Average annual increases in mean real wages in selected countries in South Asia 4 2.9 2.8 percent 1.9 2.0 2 0.1 0 Bangladesh India Nepal Pakistan Sri Lanka 2002–05 1983–2010 1999–2008 2000–09 2000–08 Source: Authors, based on data from national labor force and household surveys. GROWTH AND JOB QUALITY IN SOUTH ASIA 61 FIGURE 2.12 Percentage of workers in households below the poverty line in selected South Asian countries, by employment status a. Bangladesh, 2000–10 b. India, 1983–93 and 1999–2004 c. Nepal, 1995–2003 80 80 80 70 67 70 70 61 58 60 60 60 47 47 47 48 50 50 50 44 43 42 percent percent percent 38 38 39 40 40 40 33 31 28 30 39 29 38 28 30 30 30 30 22 21 29 29 28 27 20 24 20 26 20 21 18 18 10 10 15 10 12 7 0 0 0 2000 2005 2010 1983 URP 1993 URP 1999 MRP 2004 MRP 1995/96 2003/04 all regular wage or salaried all self-employed all casual labor all workers d. Pakistan, 2001/02 to 2007/08 e. Sri Lanka, 1995/96–2006/07 80 80 70 70 60 60 50 50 percent 43 percent 40 40 30 30 25 26 20 23 20 24 13 15 10 10 11 0 0 2001/02 2007/08 1995/96 2006/07 all self-employed, all wages, and all wages all self-employed all workers all workers Source: Authors, based on data from national labor force and household surveys. Note: URP = uniform recall period (the period in which respondents were asked to recall all consumption items over the same recall period [for example, 7 days]). MRP = mixed recall period (the period need not be the same for all items, [for example, 7 days for some and 365 days for others]). Figures are for workers age 15–64. or urban) or gender, although details vary by countries, but poverty rates for female work- country. Whereas urban poverty fell some- ers remained higher than for male workers in what faster than rural poverty in Bangladesh urban Bangladesh, India, and Pakistan (ï¬?g- and Sri Lanka and considerably faster in ure 2.13 shows data on India). In sum, using Nepal, the opposite was true in India between the primary criteria of higher wages for wage 1983 and 1993 and between 1999 and workers and lower poverty rates for the self- 2004.16 In Pakistan, urban and rural worker employed, South Asia has created better jobs. poverty rates fell equally rapidly between There has also been an improvement in 2001/02 and 2007/08. Poverty rates for both job quality in India based on the secondary male and female workers declined in all ï¬?ve criterion, namely, a reduced risk of low and 62 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 2.13 Percentage of workers in households below the poverty line in India, by employment status and gender a. Male workers b. Female workers 80 80 70 70 60 58 60 59 50 47 50 47 48 43 percent percent 41 41 40 40 40 34 36 38 32 32 31 31 30 28 30 30 27 28 29 22 20 25 20 18 20 22 19 18 17 15 15 12 10 10 0 0 1983 URP 1993 URP 1999 MRP 2004 MRP 1983 URP 1993 URP 1999 MRP 2004 MRP regular wage or salaried self employed casual labor all workers Source: Authors, based on data from national labor force and household surveys. Note: URP = uniform recall period (the period in which respondents were asked to recall all consumption items over the same recall period [for example, 7 days]). MRP = mixed recall period (the period need not be the same for all items, [for example, 7 days for some and 365 days for others]). Figures are for workers age 15–64. uncertain income for casual wage workers Notwithstanding the variation in pov- (India was the only country that had sufï¬?- erty rates across employment types, there is cient data on which to conduct this analysis). a consistent association between poverty and The average number of months for which all type of employment across countries and casual laborers (farm and nonfarm, rural and over time (figure 2.12). Casual labor is the urban) were without work, despite looking for most vulnerable segment of the labor force it, declined between 1999/2000 and 2009/10 and has the highest poverty rates (more than (ï¬?gure 2.14).17 Thus, the secondary criterion 40 percent in both r u ral and u rban for better jobs is established for India. Bangladesh, for example). The poverty rates Economic growth in the region has driven of the self-employed are the second highest.18 improvements in the quality of jobs. But not Poverty rates are generally lowest among all countries have enjoyed high or accelerating regular wage or salaried workers; on average, growth. Per capita GDP was virtually stag- they are one-third or less of those for casual nant in Nepal in the 1960s and 1970s and labor. This pattern of association is evident has grown at 2 percent or less a year since the in the consumption distribution by type of 1980s (ï¬?gure 2.2). Despite sluggish growth, employment (ï¬?gure 2.15). It is also consistent real wages have risen nearly 3 percent a year with observed wage differentials between since the 1980s, and poverty among workers regular wage or salaried workers and casual fell between the mid-1990s and the 2000s. labor (see chapter 3). These patterns have These improvements reflect massive out- endured over time. Hence, better jobs could migration of workers in response to limited be created either through improvement within job opportunities, which has improved labor an employment type or through the realloca- market outcomes for those who stay behind. tion of workers from job types with higher A large inflow of remittances has contributed poverty rates and lower wages to those with to declining poverty (box 2.1). lower poverty rates and higher wages. GROWTH AND JOB QUALITY IN SOUTH ASIA 63 FIGURE 2.14 Average number of months without work in the past year, casual laborers in India, by sector, 1999–2010 2.0 1.8 1.9 1.6 1.5 1.5 1.4 1.4 1.4 1.1 months 1.0 0.9 0.5 0 1999/2000 2004/05 2009/10 agriculture rural nonfarm urban Source: Authors, based on data from Indian labor force and household surveys. Note: Figures are for workers age 15–64 who were available for work during at least part of the month. BOX 2.1 International migration in Nepal and its effects on poverty Despite slow economic growth, Nepal has enjoyed earnings were Nr 16,000. Despite higher living costs higher wages and a signiï¬?cant decline in poverty rates overseas, migrants are able to save, with a typical among workers—thanks in large part to massive out- migrant saving about Nr 8,000 a month—twice the migration and inflow of workers’ remittances. Labor amount earned in Nepal. migration has been a feature of life in Nepal for 200 Nepal has the largest remittances as a share of GDP years. The primary destination for migrants was tra- of any country in the world with more than 10 mil- ditionally India, although since the 1990s migrants lion people. Ofï¬?cial remittances totaled $2.7 billion in have increasingly headed to the Middle East and 2009, equivalent to 22 percent of GDP; including infor- Malaysia. The Maoist insurgency from 1996 to 2006 mal flows and remittances from India, total inflows are accelerated the pace of migration, especially after estimated to have exceeded 25 percent of GDP. ï¬?ghting intensiï¬?ed in 2001, as many people fled rural Remittances have increased household income sig- communities affected by the hostilities. niï¬?cantly. An estimated 39 percent of all households The total number of migrants is estimated at and 84 percent of households with recent migration about 4.2 million, equivalent to 13 percent of Nepal’s experience received remittances in 2009. Income population. Migration is widespread, occurring in from remittances accounted for 24 percent of the households of all income groups and from all parts annual income of all households and two-thirds of of the country. Almost half of all households have the income of remittance-receiving households. The had at least one migrant abroad at some time. The additional income is spent largely on consumption, vast majority of migrants (93–94 percent) are men, education, and childcare. More than half the decline most of them 20–40 years old. At least one-third of in Nepal’s poverty rate between 1996 and 2004 (from working-age men in Nepal are migrants. 42 percent to 31 percent) is estimated to have been the Why is migration so prevalent in Nepal? The result of remittances. phenomenon is viewed as a response to limited Migration has also had a signiï¬? cant impact on domestic job opportunities in a stagnant economic the labor force. The male labor supply has fallen, environment with a poor business climate and politi- especially in rural areas. Remittances have caused cal instability. The majority of migrants worked in recipient households to increase their consumption of agriculture in Nepal but moved into manufactur- leisure and reduce labor supply as well. The decline ing, construction, and services (such as hotels and in the male labor supply has reduced domestic unem- catering) after migration. More than 87 percent of ployment and underemployment and led to rising migrants are literate compared with 62 percent of real wages. nonmigrants. Before migration, migrants earned about Nr 4,000 a month; after migration their average Source: World Bank 2011a. 64 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 2.15 Distribution of per capita household expenditure in occurred over time as a result of transitions India and Nepal, by employment status across the three groups. Workers in South Asia have better jobs than they previously did mainly as a result of increasing quality within a. India, 2005 rather than across employment categories 0.003 (table 2.2). Disaggregating further by sector can reveal more across-type movements in the kernel density 0.002 labor force. Separating India’s rural economy into agriculture and nonfarm (rural-based 0.001 industry and services) sectors reveals that between 1983/84 and 2004/05, the share of casual labor increased (from 25 percent to 0 28 percent) and the share of regular wage or 0 1,000 2,000 3,000 salaried jobs decreased (from 27 percent to monthly per capita expenditure (Indian rupees) 25 percent) (ï¬?gure 2.17). The share of casual b. Nepal, 2003 labor in the rural nonfarm sector increased 0.0015 to 38 percent by 2009/10, because most rural nonfarm employment was casual work.20 The World Bank’s poverty assessment of kernel density 0.001 India (World Bank 2011b) notes the increas- ingly bimodal nature of consumption by rural nonfarm regular wage or salaried workers, 0.0005 among whom a minority earns wages that are much higher than average. These trends 0 in the nonfarm economy notwithstanding, 0 1,000 2,000 3,000 median wages of casual and regular wage monthly per capita expenditure (Nepalese rupees) or salaried workers in the rural nonfarm regular wage casual wage economy increased and poverty rates fell dur- self employed ing this period. In addition, the number of months in which rural nonfarm casual labor- ers were available for work but unemployed Source: Authors, based on data from national labor force and household surveys. declined between 2004/05 and 2009/10 (see ï¬?gure 2.14). Thus, both the primary and the secondary criteria for improved job quality were met in the rural nonfarm economy. The proportion of workers in the three The effect of a growing labor force, employment types has remained broadly declining poverty rates, and changes in the unchanged over time, in both rural and proportion of workers by different employ- urban areas (figure 2.16). In India, which ment types has generally led to a decline in has the longest time series, the decline in the number of working poor (for details, see self-employment and increase in casual labor annex 2E). In Bangladesh, the number of observed between 1983/84 and 1999/2000 working poor decreased 18 percent between was reversed by 2004/05 but observed again 2000 and 2010, with declines in most types between 2004/05 and 2009/10.19 Given the of employment except rural nonfarm casual hierarchy of poverty rates (which are high- labor, urban casual labor, and urban regu- est among casual workers and lowest among lar wage or salaried workers. In India, the regular wage and salary earners), this implies number of working poor increased 2 per- that no signiï¬?cant change in poverty status cent between 1985 and 1995 for all types of GROWTH AND JOB QUALITY IN SOUTH ASIA 65 FIGURE 2.16 Distribution of rural and urban workers in selected South Asian countries, by employment type a. Bangladesh, 2000–2010 b. India, 1983–2009/10 100 100 18 18 18 15 17 90 22 21 21 90 31 33 31 38 36 39 80 35 38 80 70 70 28 45 41 35 36 60 40 42 41 60 percent percent 50 50 40 49 44 40 61 61 47 60 57 54 30 30 51 20 43 43 20 43 41 41 40 43 10 17 10 14 15 8 7 8 8 8 0 0 00 05 10 00 05 10 83 4 0 5 0 83 4 0 5 0 –9 00 –0 –1 –9 00 –0 –1 20 20 20 20 20 20 19 19 –2 –2 93 04 09 93 04 09 19 99 20 20 19 99 20 20 rural urban 19 19 rural urban casual labor self-employed regular wage or salaried c. Nepal, 1996–2004 d. Pakistan, 2000–2009 e. Sri Lanka, 2000–2008 100 100 100 15 13 12 90 18 17 16 16 19 20 19 31 33 80 80 80 45 44 70 60 60 60 61 46 45 45 percent percent 55 percent 50 70 71 79 83 71 40 40 40 69 67 30 55 56 20 20 20 35 36 36 27 26 10 12 14 13 5 4 0 0 0 96 04 96 04 0 8 9 0 8 9 00 08 00 08 00 –0 –0 00 –0 –0 19 20 19 20 20 20 20 20 –2 –2 07 08 07 08 99 20 20 99 20 20 rural urban rural urban 19 19 rural urban self-employed casual labor self-employed regular wage or salaried wage worker Source: Authors, based on data from national labor force and household surveys. a. Data from the Bangladesh Household Income and Expenditure Surveys (HIES) were used to calculate worker poverty rates. The share of workers by employment type in the HIES differs from the share in the Bangladesh labor force surveys. The difference is likely to be partly driven by how female employment is captured, with female participation rates in the HIES less than half those reported in the labor force survey. Therefore, the changes in the share of workers by type in Bangladesh from the HIES should be interpreted carefully. For example, between 2005 and 2010 the signiï¬?cant increase in the share of regular wage or salaried work in urban areas was driven largely by changes in the female urban workforce reported in the HIES 2005 and HIES 2010. b. Although there is variation in the shares of casual labor and self-employment in rural areas in India, there is no persistent increase or decline in the shares throughout the whole period (for example, the increase in casual labor between 2004/05 and 2009/10 mostly reversed the decline between 1999/2000 and 2004/05); the share of regular wage or salaried workers remained constant throughout the 25-year period. casual labor—agricultural, rural nonfarm, number of working poor decreased 2 percent and urban—and for the urban self-employed. between 1995 and 2005, falling among rural In contrast, the number of working poor fell nonfarm regular wage or salaried workers, 18 percent between 2000 and 2005 and in rural nonfarm self-employed and urban most employment types. 21 In Nepal, the casual labor. 66 MORE AND BETTER JOBS IN SOUTH ASIA TABLE 2.2 Decomposition of decline in worker poverty rates (percent) India Nepal Bangladesh Contribution to decline in worker poverty rates 1983–93 1999–2004 1995–2003 2000–10 Changes in poverty rates of different employment types 101.0 90.7 78.6 93.3 Changes in distribution of employment type –1.9 12.9 20.0 10.0 Interaction/residual 0.9 –3.6 1.5 –3.4 Sources: Authors, based on national labor force and household surveys. Note: Changes in poverty rates holds distribution of employment status constant; changes in distribution of employment status holds poverty rates constant. The interaction term equals 100 percent – (A + B). FIGURE 2.17 Distribution of rural nonfarm workers in India, by employment type, 1983–2009/10 100 90 25 26 28 28 80 38 70 60 percent 50 48 49 47 48 40 41 30 20 27 25 25 24 10 22 0 1983 1993/94 1999/2000 2004/05 2009/10 casual wage self-employed regular wage or salaried Source: Authors, based on data from national labor force and household surveys. into each of the four possible states (the Labor mobility combinations of the two possible employ- The broad constancy in the share of work- ment types in each of the two time periods ers across employment types does not nec- shown in the ï¬?gure) are generated by adapt- essarily imply a lack of mobility at the ing a technique developed by Lanjouw, level of individual workers across employ- Luoto, and Mckenzie (2011) to study pov- ment types. As an illustration, the analysis erty transitions. 22 focuses on labor transition for rural workers A sizable share of rural workers in the from agricultural work (less desirable jobs three countries is moving in both directions on average) to rural nonfarm work (more between agriculture and the rural nonfarm desirable jobs on average) and vice versa sector (ï¬?gure 2.19). The share of rural work- (ï¬?gure 2.18). Lower- and upper-bound esti- ers moving from agriculture to rural non- mates of the shares of rural workers falling farm work was 5–17 percent in Bangladesh GROWTH AND JOB QUALITY IN SOUTH ASIA 67 FIGURE 2.18 Labor transitions in rural areas between 2002 and 2005, 10 –20 percent in India between 1999 and 2004 and 3–13 percent between 2004 and 2007, and 4–11 Second period percent in Nepal between 1996 and 2004. The movement from the rural nonfarm sec- Agriculture Rural nonfarm tor back to agriculture was 5–17 percent in Rural nonfarm Agriculture No transition: More desirable agriculture both transition: Bangladesh between 2002 and 2005, 2–12 periods agriculture to percent in India between 1999 and 2004 First period rural nonfarm and 8–18 percent between 2004 and 2007, Less desirable No transition: and 3–10 percent in Nepal between 1996 and transition: rural nonfarm rural nonfarm to both periods 2004. The data do not allow a conclusion to agriculture be drawn as to which transition was larger, as the bounds for both the more desirable (agriculture to rural nonfarm labor) and less Source: Authors. desirable (rural nonfarm labor to agriculture) FIGURE 2.19 Probability of moving into or out of better jobs in rural Bangladesh, India, and Nepal a. Bangladesh, 2002–05 b. India, 1999–2004 and 2004–07 80 80 70 70 60 60 50 50 percent percent 40 40 30 30 20 20 10 10 0 0 rio re ul to rio m rio re nf to nf to ul to rio m pe ltu pe ltu pe far pe far ric rm no lture no lture ds ric rm re ds ds m m re ds th icu th non th icu tu ar th non ar ag nfa tu ag nfa u u bo agr bo agr no ric ric no ag ag bo bo c. Nepal, 1996–2004 80 70 60 50 percent 40 30 20 10 0 ul to rio m rio re nf to pe ltu pe far ric rm no lture re ds ds m th non th icu tu ar ag nfa u bo agr no ric ag bo Source: Authors, based on national labor force and household surveys. Note: The upper -and lower-bound estimates shown by the bars indicate the share of rural workers in the states of labor transitions described in ï¬?gure 2.19. In panels a and c, the red bar represents the given years. In panel b, the red bar represents 1999–2004 and the blue bar represents 2004–07. 68 MORE AND BETTER JOBS IN SOUTH ASIA transitions overlap. In fact, the similarity of 0.1–0.4. The correlation between growth both sets of transition bounds, especially in rates in 1971–80 and growth rates thereafter Bangladesh and Nepal, suggests that the pro- is 0.37 for 1981–90, 0.28 for 1991–2000, portion of workers could be similar in both and negligible thereafter (table 2.3). East- types of transitions. These results are con- erly and others (1993) attribute this ï¬? nding sistent with the observed constancy of the to exogenous shocks, such as adverse move- share of the rural workforce engaged in rural ments in the terms of trade and armed con- nonfarm activities noted later in chapter 3. fl ict, both of which are prevalent in South Although the bounds still overlap by a small Asia. The presence of such shocks is an amount in India between 1999 and 2004, the important reason for undertaking reforms size of the more desirable transition is notice- of the business environment (chapter 4) ably larger than that of the less desirable and education systems (chapter 5) to ensure one. This ï¬?nding is consistent with the large that the creation of better jobs is not overly increase in the share of the rural workforce in dependent on continued high economic rural nonfarm activities—from 25 percent in growth. 2000 to 30 percent in 2004—noted in chap- Demographic pressures lend further ter 3. The bounds of the more desirable and urgency to the need for reform. Most coun- less desirable transitions overlap much more tries in the region have a demographic between 2004 and 2007, when the share of window of opportunity during which an the rural nonfarm workforce in India was vir- enabling policy framework can help them tually flat. reap a demographic dividend. But the The less desirable transition suggests that opportunity is time bound and will close the nature of the rural nonfarm work itself for most of South Asia around 2040. Pop- might be transient and temporary in nature. ulation projections and the age structure In fact, the variance of the wage distributions of the population are used to develop two for casual nonfarm workers is higher than scenarios (details of the projections are in that for casual agricultural workers in India annex 2D): in all labor force surveys except 2004/05. Chapter 5 takes the labor transition analysis • Scenario 1: South Asia adds 1 million farther by looking at which types of workers entrants to the labor force every month are more likely to make more desirable and between 2010 and 2030. This scenario less desirable transitions. assumes no increase in the rates of female In summary, the labor transition analysis labor force participation, which are suggests that there is labor mobility in South among the lowest in the world. Asia and that labor moves to both more • Scenario 2: Female labor participation desirable and less desirable jobs. These two- rates in Bangladesh, India, and Pakistan way transitions are masked when looking at increase 10 percentage points by 2030, cross-sectional data at the aggregate level. in line with observed behavior in labor force participation rates in Indonesia, the Republic of Korea, Malaysia, and Thai- The urgency of reform land between 1960 and 2000. In this sce- nario, 1.2 million entrants are added to the The continuation of economic growth, labor force between 2010 and 2030, fur- which has been associated with improved ther intensifying labor market pressure. job quality in South Asia during the last three decades, cannot be taken for granted. By way of comparison, an average of Growth rates are famously unstable over just under 800,000 entrants joined the time: across five decades, the correlation labor market in South Asia every month of growth rates of GDP per capita for 94 between 1990 and 2010. The two sce- countries for which data are available in narios thus represent increases of 25–50 the World Development Indicators is just percent above the average for this period. GROWTH AND JOB QUALITY IN SOUTH ASIA 69 TABLE 2.3 Correlations of country growth rates of per capita GDP across decades 1961–70 1971–80 1981–90 1991–2000 2001–10 1961–70 1 0.34*** 0.37*** 0.22** –0.18* 1971–80 1 0.37*** 0.28*** 0.14 1981–90 1 0.43*** 0.07 1991–2000 1 0.18* 2001–10 1 Source: Authors’ calculations, based on data from World Bank 2011c. Note: Findings are based on 94 countries. *** Signiï¬?cant at the 1% level; ** signiï¬?cant at the 5% level; * signiï¬?cant at the 10% level. The critical question is whether increases high economic growth may not continue, of this magnitude will be absorbed at ris- highlight the importance of proceeding ing or low levels of productivity. The two quickly with reform in order to meet the scenarios, together with the possibility that employment challenge. Annex 2A Methodology for decomposing growth A country’s output in any given year depends Given an estimate for and measures of on its factor inputs—labor and (human and Y, L, K, and H, it is straightforward to solve physical) capital—as well as the efï¬? ciency for A (or a) and construct the decomposition. with which factors are used in production. 23 It is assumed that capital’s share = 0.35. Deï¬? ne Y as real GDP, K as the physical capi- An analysis that used actual income shares tal stock, A as the level of technology, and L in each period would allow for the consid- as labor (measured as “bodies of economi- eration of a much wider range of underlying cally active personsâ€?), which is assumed to production functions. However, few coun- be “augmentedâ€? by H, an index of the aver- tries are able to allocate the incomes of the age level of labor quality, measured by aver- self-employed between capital and labor, a age years of schooling. Assume that a coun- particular problem in South Asia, where the try’s output can be expressed as a function self-employed make up the bulk of the labor of these inputs, using the speciï¬?c functional force. form shown in equation (2A.1), and that L is employment for industrial countries returns to scale are constant. plus Bangladesh, India, Pakistan, and Sri Y = AK (HL)(1 ) (2A.1) Lanka and labor force for all other countries. The capital stock measure is constructed The results are reported in a form that from investment data using the perpetual decomposes growth in output per worker into inventory method, with annual depreciation the contributions from the growth of physi- rate of 0.05 percent. The construction of H cal capital per worker, education per worker, assumes that human capital is directly related and total factor productivity, as shown in to average years of schooling (S) and that equation (2A.2) (lower-case letters denote a there is a 7 percent return to each additional variable’s average annual growth rate). year of schooling: y/l = (k/l) + (1 )h a (2A.2) H = (1.07)S (2A.3) 70 MORE AND BETTER JOBS IN SOUTH ASIA Annex 2B Sources of average annual growth in output per worker TABLE 2B.1 Sources of average annual growth in output per worker, by region, 1960–2008 (percent) Change in Contribution of Output Physical capital Education per Total factor Region/period Output per worker per worker worker productivity World (83 countries) 1960–80 4.91 3.12 1.29 0.41 1.34 1980–90 3.81 1.96 0.81 0.38 0.60 1990–2000 4.03 2.41 0.97 0.38 0.85 2000–08 4.30 2.73 0.93 0.24 0.99 1980–08 4.03 2.34 0.90 0.34 0.80 Industrial countries (22 countries) 1960–80 4.45 3.21 1.20 0.42 1.45 1980–90 3.08 1.91 0.73 0.31 0.78 1990–2000 2.43 1.61 0.84 0.33 0.57 2000–08 1.83 1.08 0.76 0.14 0.13 1980–2008 2.49 1.56 0.77 0.27 0.52 Sub-Saharan Africa (19 countries) 1960–80 4.53 2.16 0.87 0.18 1.15 1980–90 1.62 (1.47) (0.11) 0.47 (1.78) 1990–2000 1.92 (1.60) (0.76) 0.48 (1.38) 2000–08 4.72 2.24 0.25 0.41 1.58 1980–2008 2.61 (0.47) (0.24) 0.45 (0.69) China 1960–80 4.89 2.65 0.91 0.44 1.27 1980–90 9.29 6.62 2.09 0.39 4.03 1990–2000 10.42 9.11 3.28 0.50 5.12 2000–08 10.18 9.25 3.45 0.39 5.20 1980–2008 9.95 8.25 2.90 0.43 4.75 East Asia less China (7 countries) 1960–80 7.84 4.46 2.58 0.56 1.27 1980–90 7.42 4.26 2.15 0.64 1.18 1990–2000 6.13 3.98 2.35 0.50 0.78 2000–08 4.46 2.53 0.78 0.49 1.30 1980–2008 6.11 3.66 1.83 0.55 1.07 1980–96 7.57 4.68 1.50 0.59 2.31 1999–2008 4.81 2.76 1.50 0.49 0.78 Latin America and the Caribbean (21 countries) 1960–80 6.06 2.84 0.87 0.36 1.58 1980–90 1.49 (1.65) 0.11 0.52 (2.32) 1990–2000 3.15 0.47 0.08 0.48 (0.14) 2000–08 3.52 1.28 0.24 0.39 0.69 1980–2008 2.66 (0.06) 0.14 0.47 (0.69) Middle East (9 countries)           1960–80 5.68 3.40 2.45 0.43 0.95 1980–90 3.78 1.26 0.44 0.60 0.36 1990–2000 3.80 1.11 0.15 0.57 (0.02) 2000–08 4.84 2.74 0.73 0.46 1.39 1980–2008 4.09 1.63 0.42 0.55 0.52 (continues next page) GROWTH AND JOB QUALITY IN SOUTH ASIA 71 TABLE 2B.1 Sources of average annual growth in output per worker, by region, 1960–2008 (continued) Change in Contribution of Output Physical capital Education per Total factor Region/period Output per worker per worker worker productivity South Asia (4 countries) 1960–80 3.93 1.63 1.04 0.27 0.31 1980–90 5.24 3.18 1.11 0.42 1.46 1990–2000 5.37 3.47 1.21 0.37 1.92 2000–08 7.05 4.94 1.67 0.34 2.90 1980–2008 5.80 3.78 1.30 0.38 2.04 Source: Bosworth 2010. Note: Average annual growth in output per worker is the sum of the physical capital, human capital, and TFP contributions to growth in productivity. Totals may not add up to total growth in output per worker because of interaction terms. Table is based on aggregated growth accounts, except for Bangladesh, India, Pakistan, and Sri Lanka data from 1980 onward, which are based on disaggregated growth accounts. For South Asia, growth accounts use employment data from national labor force surveys (latest available in August 2010). Aggregated growth accounts for regions other than South Asia use labor force estimates from World Bank 2011c (taken from the International Labour Organization [ILO]), which include the unemployed population. ILO data include extrapolated data. TABLE 2B.2 Sources of average annual growth in output per worker in South Asia, by country, 1960–2008 (percent) Change in Contribution of Output per Physical capital Education per Total factor Country/period Output worker per worker worker productivity Bangladesh 1960–80 2.39 0.21 0.27 0.21 –0.26 1980–2008 4.66 2.01 0.87 0.44 0.69 India 1960–80 4.67 1.34 1.00 0.25 0.08 1980–2008 6.2 4.39 1.41 0.42 2.60 Pakistan 1960–80 5.94 3.1 2.38 0.19 0.53 1980–2008 5.28 2.67 1.00 0.30 1.40 Sri Lanka 1960–80 4.49 2.18 0.43 0.43 1.31 1980–2008 4.91 3.08 1.62 0.27 1.16 Source: Bosworth 2010. Note: Average annual growth in output per worker is the sum of the physical capital, human capital, and TFP contributions to growth in productivity. Totals may not add up to total growth in output per worker because of interaction terms. The growth accounts use employment data from national labor force surveys (latest available in August 2010). 72 MORE AND BETTER JOBS IN SOUTH ASIA Annex 2C Shares of agriculture, industry, and services in employment and GDP TABLE 2C.1 Regressions of shares of agriculture, industry, and services in employment and GDP, 2008 Employment GDP Agriculture Industry Services Agriculture Industry Services Variable (1) (2) (3) (4) (5) (6) ln (per capita GDP, 2008 –64.77** 28.92 29.87 –44.96*** 69.57 –24.61 in 2005 constant purchasing (25.82) (18.30) (26.93) (14.13) (56.68) (54.82) power parity dollars) ln (per capita GDP, 2008 2.783** –1.505 -0.988 2.040*** –3.780 1.740 in 2005 constant purchasing (1.365) (0.954) (1.425) (0.724) (2.990) (2.899) power parity dollars, squared) ln (land area, square kilometers) 0.316 –0.210 0.0783 –0.226 1.245 –1.019 (0.216) (0.348) (0.470) (0.378) (0.887) (0.826) Bangladesh dummy –12.28 1.169 9.185 –13.90** 11.22 2.679 (7.444) (5.946) (7.745) (5.659) (15.89) (14.92) India dummy 13.96*** 1.102 –15.26*** –3.045 –2.457 5.502 (3.497) (2.895) (3.661) (2.693) (8.026) (7.433) Nepal dummy 4.107 0.396 –6.753 –1.170 1.490 –0.320 (7.754) (6.180) (8.069) (5.842) (16.61) (15.61) Pakistan dummy 2.393 1.812 –5.220 –1.613 –2.101 3.714 (3.380) (2.748) (3.462) (2.776) (7.036) (6.454) Sri Lanka dummy 2.905 4.758*** –9.682*** –2.062 –1.629 3.691** (1.900) (1.462) (2.079) (1.954) (2.232) (1.828) Constant 373.0*** –112.7 –133.0 251.1*** –300.2 149.1   (121.0) (86.40) (125.7) (69.45) (260.1) (250.8) N 55 55 55 53 53 53 Adjusted R-squared 0.860 0.095 0.799 0.827 0.080 0.376 Sources: Authors, based on data from ILO 2010; World Bank 2011c; and India National Sample Survey. Note: Standard errors in parentheses. Sample restricted to countries for which employment data were available from both 1980 (or the earliest year in the subsequent ï¬?ve-year period) and 2008 (or the latest year in the preceding ï¬?ve-year period). Data exclude transition economies in Europe and Central Asia. *** Signiï¬?cant at the 1% level; ** signiï¬?cant at the 5% level; * signiï¬?cant at the 10% level. Annex 2D Methodology and data sources for labor force projections This annex describes the assumptions used household surveys (mostly 2008) to the 2010 to generate the labor force projections for age-gender populations from the United 2030. The scenarios are developed at the Nations (UN 2008). 24 country level and aggregated for the region. Scenario 1 (demographic case) Historical period (1990–2010) Labor force projections in 2030 under sce- Trends in the region’s labor force cover nario 1 are based purely on demographic the period 1990–2010. The source of the projections, using UN (2008). In this and labor force in 1990 for each country is the the next scenario, the labor force in 2010 for LABORSTA database of the International each country is calculated using the same Labour Organization (ILO). The labor method described in the historical descrip- force in 2010 is obtained by applying the tion. Similarly, the 2030 labor force projec- (ï¬?ve-year) age/gender labor force participa- tions are obtained by multiplying the 2030 tion rates obtained from the latest available (five-year) age/gender populations by the GROWTH AND JOB QUALITY IN SOUTH ASIA 73 same current age/gender–speciï¬?c labor force The experience of four Asian countries participation rate, calculated from the latest (Indonesia, the Republic of Korea, Malaysia, household surveys. and Thailand) was used to develop this sce- nario. These countries represent an interest- ing example for South Asia, for a variety of Scenario 2 (behavioral case) reasons. First, as neighbors they share some Scenario 1 assumes no change in labor social and cultural characteristics. Second, force participation rates over the projec- they developed earlier than the large South tion period for speciï¬? c age/gender groups. Asian countries and can therefore provide Scenario 2 assumes that participation by insights as benchmark countries. Third, at women in Bangladesh, India, and Pakistan the beginning of their “take-offs,â€? Indonesia, will increase between 2010 and 2030. These Korea, and Malaysia (though not Thailand) three countries are selected for this scenario had female participation rates comparable because each has very low female partici- to those of Bangladesh and India and a bit pation rates by international standards. higher than Pakistan. In the past 50 years, to Together with the fact that they dominate different degrees, these countries experienced South Asia’s overall population (accounting strong economic growth and modernization; for 95 percent of the working-age popu- in the process, all of them except Thailand lation), this implies that any increases in experienced increases in female participation female participation in these three large rates (box 2D.1). Tracking these historical countries could have an important effect on development indicators and roughly match- the region’s labor force. Under scenario 2, ing them with projected trends (for example, participation rates are assumed to remain per capita GDP) for the South Asian coun- at current levels for men in all countries tries to 2030 suggest that an increase in and women in the other five countries in the female labor force participation rate of the region (scenarios 1 and 2 are identical 10 percentage points would approximate a for Afghanistan, Bhutan, Maldives, Nepal, convergence with the historical experience of and Sri Lanka). these benchmark countries. 74 MORE AND BETTER JOBS IN SOUTH ASIA BOX 2D.1 Trends in female labor force participation in southeast and East Asian comparator countries Table 2D.1.1 summarizes what happened to female lower end of this range, female participation rates in labor force participation in Indonesia, Korea, Indonesia, Korea, and Malaysia were in the low to Malaysia, and Thailand during the periods when mid-30s—not too different from rates in India and per capita GDP increased through the range pro- Bangladesh today, though about 10 points higher jected for Bangladesh, India, and Pakistan to 2030. than in Pakistan. The case of Thailand demonstrates In Indonesia, Korea, and Malaysia, female participa- that increasing participation is not inevitable as tion rates increased consistently, in some cases, sub- per capita GDP increases. Although female rates in stantially, with per capita GDP (in 2005 purchasing Thailand did rise as per capita GDP grew to about power parity dollars) rising from the mid-$2,000s $5,000, they subsequently stalled and then started to almost $10,000. When per capita GDP was at the to decline. BOX TABLE 2D.1.1 Female labor force participation in Bangladesh, India, and Pakistan and Asian comparator countries Projected per capita GDP 2007–30 Projected changes in female (in 2005 purchasing power Southeast and East Asian labor force participation rate Country parity dollars) comparators (percentage points) Bangladesh Mid-$2,000s to high $3,000s Indonesia 1980–90s Almost +10 Malaysia 1960–70s More than +5 Thailand 1970–80s About +5 Republic of Korea 1960–70s About +10 India High $3,000s to low $9,000s Malaysia 1970s–90s Almost +10 Thailand 1980s onward About –10 Republic of Korea 1970s–90s About +10 Pakistan Mid $3000s to Low $6000s Indonesia 1990 onward About +5 Malaysia 1970–80 About +3 Thailand 1980–90 No change Republic of Korea 1970–80 About +2 Sources: Authors, based on GDP trends from Penn World Tables 6.3 and female labor force participation rates from ILO LABORSTA. Note: Projections of per capita GDP assume that the 1990–2007 annual growth rates will continue through 2030. The per capita GDP data used are 2005 purchasing power parity estimates from the Penn World Tables 6.3. Annex 2E Poverty rates and the number of working poor in South Asia TABLE 2E.1 Percentage of workers in households below the poverty line in Bangladesh, by employment type, 2000–10 2000 2005 2010 Type of Poverty Lower Upper Poverty Lower Upper Poverty Lower Upper employment rate bound bound rate bound bound rate bound bound Rural agricultural regular wage or salaried 39.6 29.5 49.7 58.8 48.8 68.8 54.1 42.5 65.7 Rural agricultural self-employed 39.5 37.4 41.6 29.7 28.0 31.4 22.0 20.3 23.6 Rural agricultural casual labor 72.1 69.9 74.3 66.2 63.9 68.5 51.0 48.7 53.3 Rural nonagricultural regular wage or salaried 33.8 30.7 37.0 25.7 23.2 28.2 26.2 23.9 28.5 Rural nonagricultural self-employed 41.1 38.2 44.0 32.9 30.3 35.5 26.1 24.0 28.2 Rural nonagricultural casual labor 60.9 57.6 64.1 55.7 53.0 58.4 44.8 42.4 47.1 Urban regular wage or salaried 25.7 23.4 28.0 19.2 17.3 21.1 15.6 14.1 17.1 Urban self-employed 29.3 26.8 31.9 18.6 16.9 20.4 13.6 12.1 15.2 Urban casual labor 62.2 58.8 65.6 55.5 52.8 58.2 44.1 41.6 46.6 Regular wage or salaried 30.5 28.6 32.3 23.6 22.1 25.2 20.9 19.6 22.2 All self-employed 38.3 36.9 39.7 28.2 27.0 29.3 21.6 20.5 22.6 All casual labor 67.4 65.8 69.0 60.6 59.1 62.0 47.3 45.9 48.7 All rural 50.0 48.8 51.2 41.7 40.6 42.7 33.8 32.8 34.8 All urban 34.9 33.3 36.5 26.8 25.6 28.0 21.0 20.0 22.1 All workers 46.8 45.8 47.7 37.7 36.9 38.5 30.0 29.3 30.8 Source: Authors, based on data from labor force and household surveys. Note: Lower and upper bounds in the table refer to the 95 percent conï¬?dence intervals of the poverty rate estimate by employment type. 75 76 MORE AND BETTER JOBS IN SOUTH ASIA TABLE 2E.2 Number of working poor in Bangladesh, by employment type, 2000–10 Percentage change in number of working poor Number of working poor (thousands) 2000–05 2005–10 Type of employment 2000 2005 2010 Total Annual Total Annual Rural agricultural regular wage or salaried 186 244 162 31 6 –50 –8 Rural agricultural self-employed 4,025 3,555 2,177 –12 –2 –63 –9 Rural agricultural casual labor 6,036 4,938 4,018 –18 –4 –23 –4 Rural nonagricultural regular wage or salaried 1,478 1,284 1,601 –13 –3 20 5 Rural nonagricultural self-employed 2,293 1,851 1,881 –19 –4 2 0 Rural nonagricultural casual labor 2,687 3,205 3,181 19 4 –1 0 Urban regular wage or salaried 1,017 1,092 1,267 7 1 14 3 Urban self-employed 962 877 617 –9 –2 –42 –7 Urban casual labor 1,249 1,564 1,484 25 5 –5 –1 Regular wage or salaried 2,681 2,620 3,030 –2 0 14 3 Self-employed 7,280 6,282 4,675 –14 –3 –34 –6 Casual labor 9,972 9,707 8,683 –3 –1 –12 –2 All rural 16,704 15,076 13,020 –10 –2 –16 –3 All urban 3,228 3,534 3,369 9 2 –5 –1 All workers 19,932 18,609 16,389 –7 –1 –14 –3 Source: Authors, based on data from labor force and household surveys Note: The number of working poor for each employment type is calculated by multiplying the estimated total employment (estimated in ï¬?ve yearly intervals) by the share of employment and poverty rate by employment type from the closest year of the household survey. For example, for the number of working poor in India in 1985, the share of employment and poverty rate by employment type used was estimated from the 1983 labor force survey. TABLE 2E.3 Percentage of workers in households below the poverty line in India, by employment type, 1983–2004/05 1983 1993/94 1999/2000 2004/05 Poverty Lower Upper Poverty Lower Upper Poverty Lower Upper Poverty Lower Upper Type of employment rate bound bound rate bound bound rate bound bound rate bound bound Rural agricultural regular wage or salaried 52.7 51.0 54.5 43.3 40.5 46.1 37.0 34.8 39.2 26.6 24.1 29.2 Rural agricultural self-employed 37.1 36.7 37.4 28.6 28.3 29.0 24.6 24.3 24.9 17.0 16.7 17.3 Rural agricultural casual labor 60.6 60.0 61.1 47.8 47.2 48.4 43.2 42.6 43.8 30.8 30.1 31.4 Rural nonagricultural regular wage or salaried 25.2 24.2 26.1 13.6 12.9 14.2 11.0 10.4 11.6 8.4 8.0 8.8 Rural nonagricultural self-employed 41.2 40.4 42.0 27.2 26.4 27.9 26.3 25.7 27.0 17.6 17.1 18.0 Rural nonagricultural casual labor 47.0 45.9 48.1 38.0 36.9 39.1 30.2 29.2 31.2 24.7 23.9 25.5 Urban regular wage or salaried 26.2 25.6 26.7 17.8 17.3 18.2 15.6 15.1 16.0 13.7 13.3 14.1 Urban self-employed 42.8 42.2 43.4 32.3 31.8 32.9 28.2 27.7 28.7 24.3 23.8 24.8 Urban casual labor 59.9 58.9 60.9 57.1 56.1 58.0 51.5 50.6 52.4 45.5 44.5 46.4 Regular wage or salaried 28.8 28.3 29.2 17.8 17.4 18.2 15.5 15.1 15.8 12.3 12.0 12.6 Self-employed 38.6 38.3 38.9 29.0 28.8 29.3 25.5 25.3 25.8 18.5 18.3 18.7 Casual labor 58.2 57.8 58.7 47.4 46.9 47.9 42.0 41.6 42.4 31.2 30.7 31.6 All rural 43.7 43.5 44.0 33.4 33.2 33.7 29.8 29.5 30.0 20.3 20.1 20.5 All urban 38.8 38.4 39.2 30.7 30.4 31.1 27.2 26.9 27.5 23.1 22.8 23.4 All workers 42.6 42.4 42.9 32.8 32.6 33.1 29.2 29.0 29.4 20.9 20.8 21.1 Source: Authors, based on data from labor force and household surveys. Note: Lower and upper bounds in the table refer to the 95 percent conï¬?dence intervals of the poverty rate estimate by employment type. 77 78 MORE AND BETTER JOBS IN SOUTH ASIA TABLE 2E.4 Number of working poor in India, by employment type, 1985–2005 Percentage change in number of working poor Number of working poor (thousands) 1985–95 2000–05 Type of employment 1985 1995 2000 2005 Total Annual Total Annual Rural agricultural regular wage or salaried 2,290 1,177 1,320 639 –49 –6 –52 –13 Rural agricultural self-employed 36,916 36,007 31,661 25,639 –2 0 –19 –4 Rural agricultural casual labor 30,943 33,898 34,635 22,177 10 1 –36 –9 Rural nonagricultural regular wage or salaried 2,974 2,209 1,965 1,943 –26 –3 –1 0 Rural nonagricultural self-employed 9,156 8,553 8,807 8,302 –7 –1 –6 –1 Rural nonagricultural casual labor 5,641 6,336 6,079 6,701 12 1 10 2 Urban regular wage or salaried 6,210 5,484 5,355 5,438 –12 –1 2 0 Urban self-employed 9,453 10,192 9,772 10,806 8 1 11 2 Urban casual labor 5,988 7,597 7,757 6,514 27 2 –16 –3 Regular wage or salaried 11,474 8,870 8,641 8,021 –23 –3 –7 –1 Self-employed 55,525 54,751 50,240 44,746 –1 0 –11 –2 Casual labor 42,571 47,831 48,470 35,392 12 1 –27 –6 All rural 87,919 88,181 84,467 65,401 0 0 –23 –5 All urban 21,651 23,272 22,884 22,758 7 1 –1 0 All workers 109,570 111,453 107,351 88,159 2 0 –18 –4 Source: Authors, based on data from labor force and household surveys. Note: The number of working poor for each employment type is calculated by multiplying the estimated total employment (estimated in ï¬?ve yearly intervals) by the share of employment and poverty rate by employment type from the closest year of the household survey. For example, for the number of working poor in India in 1985, the share of employment and poverty rate by employment type used was estimated from the 1983 labor force survey. TABLE 2E.5 Percentage of workers in households below the poverty line in Nepal, by employment type, 1995/96 and 2003/04 1995/96 2003/04 Poverty Lower Upper Poverty Lower Upper Type of employment rate bound bound rate bound bound Rural agricultural regular wage or salaried 50.9 41.8 59.9 30.1 4.9 55.3 Rural agricultural self-employed 39.7 38.2 41.1 30.7 29.4 32.0 Rural agricultural casual labor 50.2 46.3 54.2 50.0 45.7 54.3 Rural nonagricultural regular wage or salaried 21.2 15.4 27.0 10.0 6.3 13.7 Rural nonagricultural self-employed 30.4 25.9 34.8 16.7 13.3 20.0 Rural nonagricultural casual labor 44.8 39.5 50.0 43.4 37.8 49.1 Urban regular wage or salaried 9.7 6.6 12.8 3.3 1.9 4.8 Urban self-employed 18.4 15.6 21.2 8.9 7.4 10.4 Urban casual labor 45.2 38.2 52.1 16.1 11.7 20.6 Regular wage or salaried 27.6 24.2 31.0 7.4 5.6 9.1 Self-employed 38.1 36.8 39.3 27.3 26.2 28.3 Casual labor 48.1 45.2 51.0 43.7 40.8 46.7 All rural 40.0 38.8 41.2 31.1 29.9 32.2 All urban 21.0 18.7 23.2 8.3 7.2 9.5 All workers 39.0 37.9 40.1 28.0 27.1 29.0 Source: Authors, based on data from labor force and household surveys. Note: Lower and upper bounds in the table refer to the 95 percent conï¬?dence intervals of the poverty rate estimate by employment type. GROWTH AND JOB QUALITY IN SOUTH ASIA 79 TABLE 2E.6 Number of working poor in Nepal, by employment type, 1995–2005 Percentage change in number of working poor Number of working poor (thousands) 1995–2005 Type of employment 1995 2005 Total Annual Rural agricultural regular wage or salaried 92 9 –91 –21 Rural agricultural self-employed 2,456 2,464 0 0 Rural agricultural casual labor 432 463 7 1 Rural nonagricultural regular wage or salaried 60 41 –31 –4 Rural nonagricultural self-employed 181 136 –25 –3 Rural nonagricultural casual labor 207 215 4 0 Urban regular wage or salaried 13 14 11 1 Urban self-employed 50 89 79 6 Urban casual labor 41 33 –20 –2 Regular wage or salaried 165 65 –61 –9 Self-employed 2,687 2,689 0 0 Casual labor 680 711 5 0 All rural 3,427 3,328 –3 0 All urban 104 137 31 3 All workers 3,531 3,465 –2 0 Source: Authors, based on data from labor force and household surveys Note: The number of working poor for each employment type is calculated by multiplying the estimated total employment (estimated in ï¬?ve yearly intervals) by the share of employment and poverty rate by employment type from the closest year of the household survey. For example, for the number of working poor in India in 1985, the share of employment and poverty rate by employment type used was estimated from the 1983 labor force survey. Annex 2F Analysis of poverty and unemployment in India A more complete analysis of the links In 1983 and 1993/94, the same house- between poverty and unemployment is pos- holds were sampled for the employment and sible in India, where time series data for a unemployment surveys and the consumption longer period are available. Some data com- expenditure survey. The household consump- parability issues merit attention, however. tion expenditure aggregate in the employment and unemployment surveys is the unabridged uniform recall period measure from the con- Data used sumption expenditure survey. These two Data from the employment and unemploy- datasets are comparable. ment surveys by the India National Sample In 1999/2000 and 2004/05, the employ- Survey—which are equivalent to labor force ment and unemployment surveys and surveys—were used for 1983, 1993/04, consumption expenditure survey sampled dif- 1999/2000, and 2004/05. The old ofï¬? cial ferent households. The consumption expendi- state level poverty lines for the same time ture survey used unabridged uniform recall periods were used for the analysis, because period and mixed recall period consumption they are available for all survey years. measures. The consumption module in the Employment and unemployment surveys employment and unemployment surveys used were used instead of consumption expendi- an abridged (fewer questions) mixed recall ture surveys because they contain both con- period to measure the household consump- sumption expenditure and labor force vari- tion expenditure aggregate. ables, which allow the estimation of poverty The estimates of poverty rates using rates by employment types—something the abridged mixed recall period in these two consumption expenditure survey does not. employment and unemployment surveys are 80 MORE AND BETTER JOBS IN SOUTH ASIA comparable with each other but not compa- and 2004/05 show a significant decline in rable with the 1983 and 1993/94 unabridged poverty rates for the whole population, from uniform recall period measures. The abridged 36.0 percent to 27.5 percent. It is highly mixed recall period measures should not be unlikely that poverty increased between interpreted as accurate estimates of poverty 1993/94 and 1999/2000, as this would imply rates, because mixed recall period measures an even higher percentage decline between tend to generate lower poverty rates than 1999/2000 and 2004/05. It is also unlikely uniform recall period measures and abridge- that during a period in which poverty rates ment creates differences in the estimates declined signiï¬?cantly for the whole popula- using unabridged measures. Therefore, the tion, worker poverty rates would not also poverty estimates from the 1999/2000 and have fallen. 2004/05 employment and unemployment surveys should only be used to look at trends Categories of Employment by employment type. Given the changes in the consumption Nine categories of employment are consid- measure, the estimated worker poverty rates ered, based on the main activity of each of 1993/94 and 1999/2000 cannot be com- working member of a household. They pared to determine whether poverty declined. include regular wage or salaried, self– However, official poverty measures using employed, and casual labor in the rural agri- unabridged uniform recall period from the cultural, rural nonagricultural, and urban consumption expenditure survey in 1993/94 sectors. TABLE 2F.1 Official and authors’ estimated poverty rates for urban, rural, and all workers in India, 1983–2004/05 Year/estimate Urban workers Rural workers All workers 1983 Official estimates (based on unabridged uniform recall period from consumption expenditure survey) 40.8 45.7 44.5 Authors’ estimates (based on unabridged uniform recall period from employment and unemployment survey/consumption expenditure survey) 38.8 43.7 42.6 1993/94 Official estimates (based on unabridged uniform recall period from consumption expenditure survey) 32.4 37.3 36.0 Authors’ estimates (based on unabridged uniform recall period from employment and unemployment survey/consumption expenditure survey) 30.7 33.4 32.8 1999/2000 Official estimates (based on unabridged mixed recall period from consumption expenditure survey) 23.6 27.1 26.1 Authors’ estimates (based on abridged mixed recall period from employment and unemployment survey) 27.2 29.8 29.2 2004/05 Official estimates (based on unabridged uniform recall period from consumption expenditure survey) 25.7 28.3 27.5 Official estimates (based on unabridged mixed recall period from consumption expenditure survey) 21.7 21.8 21.8 Authors’ estimates (based on abridged mixed recall period from employment and unemployment survey) 23.1 20.3 20.9 Source: Authors, based on data from national labor force and household surveys. Note: Official estimates include the entire population living in households below the poverty line; authors’ estimates include only people age 15–64. The unabridged mixed recall period measures from the consumption expenditure surveys conducted in 1999/2000 and 2004/05 are roughly but not strictly comparable, because of differences in design. GROWTH AND JOB QUALITY IN SOUTH ASIA 81 Sample growth accounting to be conducted only for Bangladesh, India, Pakistan, and Sri Lanka. The sample includes 209,223 employed indi- 3. The ï¬?gures for South Asia are GDP at purchas- viduals in 1993/94, 213,986 in 1999/2000, ing power parity–weighted averages of the ï¬?g- and 228,244 in 2004/05, about 30 percent ures for Bangladesh, India, Pakistan, and Sri of whom are women. Rural agricultural reg- Lanka and are therefore weighted most heav- ular wage or salaried workers represent just ily by India. For further details, see annex 2B, 1.7 percent of the sample in 1983 and 0.6 which contains sources of growth by region percent in 2003/04. Given this small sample per decade between 1960 and 2008. 4. Young (1994, 1995) highlights the relatively size, estimated poverty rates for this type of limited role of TFP growth in East Asia less worker should not be treated as reliable. China during its years of rapid investment-led In 1983 and 1993/94, the headcount pov- growth. erty rates from the employment and unem- 5. Bosworth, Collins, and Virmani (2007) ployment surveys were comparable to ofï¬?cial develop this point in the context of India. poverty rates, as the consumption expendi- 6. This discussion draws on Bloom, Canning, ture survey sampled the same households and Rosenberg (2011). (table 2F.1). The differences reflect the fact 7. “Potentialâ€? is used because without enabling that the estimates are for workers and not the policies, a rising ratio of the working-age whole population. population to the nonworking-age population Between 1999/2000 and 2004/05, offi- will not necessarily boost economic growth. 8. Reallocation during a period is calculated as cial estimates using the unabridged mixed a residual by subtracting the weighted sum recall period measure from the consumption of TFP growth in each sector (agriculture, expenditure survey also show a decline in industry, services) from aggregate TFP growth poverty. The differences between the ofï¬?cial during that period, where the weights are the poverty estimates and the authors’ estimates share of each sector in GDP at the beginning reflect the fact that (a) the authors’ estimates of the period. refer to workers and not the whole popula- 9. The choice of countries is dictated by the tion and (b) the mixed recall period measure availability of sectorally disaggregated growth is unabridged in the consumption expendi- accounts. The Thailand numbers are from ture survey and abridged in the employment Bosworth (2005); data from China were and unemployment surveys. The compari- adapted by the authors from Bosworth and Collins (2008). son of the mixed recall period and uniform 10. The ratio of employment to the working- recall period measures from the consumption age population can be written as follows: expenditure survey in 2004/05 also shows (employment/labor force) (labor force/ that the mixed recall period measures esti- working-age population). The ï¬?rst term is (1 mate lower poverty rates than the uniform – the unemployment rate); the second is the recall period. Hence, it is not possible to com- labor force participation rate. pare the results from the 1983 and 1993/94 11. Afghanistan and Bhutan are excluded, because employment and unemployment surveys with 2000 data are not available. Including them the 1999/2000 and 2004/05 employment and in 2010 would bring total employment to unemployment surveys. 577 million in 2010. 12. The employment status of workers is deï¬?ned here on the basis of questions on current Notes weekly status (generally the past seven days) in the labor force surveys of Bangladesh, 1. Europe and Central Asia, (not shown in the Bhutan, India, Nepal, Pakistan, and Sri ï¬?gure) grew faster than South Asia during Lanka and on the basis of employment status the 2000s, but this reflected, in part, recovery in the last month in the labor force surveys from a transition recession following the exit of Afghanistan and Maldives (where weekly from the command economy. status questions were not included). 2. The decomposition is formally presented 13. Various additional dimensions of job qual- in annex 2A. Data constraints allowed ity are often cited, such as access to nonwage 82 MORE AND BETTER JOBS IN SOUTH ASIA beneï¬?ts and public social protection mecha- 22. The details of the methodology are described nisms, the ability to upgrade skills and receive in appendix B. training on the job, and a safe working envi- 23. This annex is based on Collins (2007). ronment. There is typically a strong correlation 24. The household survey–based participation between better jobs as deï¬?ned by the criteria rates are used to maintain consistency with used in this book and many of these additional the methodology used for the projections. dimensions. ILO estimates of the labor force in 2010 are 14. The use of this proxy for the self-employed slightly different from the estimates developed was proposed in the Indian context by here. The ILO estimate of the total labor force Sundaram (2004). in South Asia in 2010 is about 27 million 15. The household survey data for Pakistan and Sri higher, with an annual growth rate between Lanka used for the poverty-employment anal- 1990 and 2010 of 2.3 percent, compared ysis do not distinguish between casual work- with the 2.1 percent posited in this book. ers and regular wage and salaried workers. 16. This trend prompted an investigation into the cost of living in urban India and a recent revi- References sion of the ofï¬?cial poverty lines. Aggarwal, S. 2010. “Labor Input and Its Compo- 17. This question was based not on a reference sition: An Industry–Level Perspective.â€? Paper week but on a reference year. A worker’s usual presented at the Worldklems conference, Cam- principal activity is determined by the activity bridge, MA, August. the worker spent most of his or her time doing Bloom, D., D. Canning, and L. Rosenberg. 2011. in the year preceding the survey. Any activ- “Demographic Change and Economic Growth.â€? ity other than the principal status constitutes In Reshaping Tomorrow, ed. E. Ghani. New a worker’s subsidiary status. “Usual’â€? status Delhi: Oxford University Press. workers include principal status workers (who Bosworth, B. 2005. Economic Growth in Thai- spent most of their time employed or looking land: The Macroeconomic Context. Brookings for jobs) and subsidiary workers (who spent Institution, Washington, DC. part of their time working or looking for jobs ———. 2010. “Update of Bosworth and Collins, in the year preceding the survey). 2003.â€? Unpublished paper, World Bank, Wash- 18. An exception is Nepal, where the poverty rate ington, D.C. for urban casual workers is signiï¬?cantly lower Bosworth, B., and S. Collins, 2003. “The Empir- than the rate for the rural self-employed, ics of Growth: An Update.â€? Brookings Papers which is driven by the subsegment of urban on Economic Activity (2): 113–206. casual labor employed in short-term contract ———. 2008. “Accounting for Growth: Compar- work. This group is more educated than daily ing China and India.â€? Journal of Economic wage workers and has poverty rates that are Perspectives 22 (1): 45–66. closer to those of urban regular wage or sala- ———. 2010. “Update of Bosworth and Collins ried workers. 2003.â€? Unpublished notes for this book. 19. The number of self-employed workers declined Bosworth, B., S. Collins, and A. Virmani 2007. between 2004/05 and 2009/10. Casual labor “Sources of Growth in the Indian Economy.â€? accounted for nearly 80 percent of net addi- NBER Working Paper 12901, National Bureau tional employment in those sectors that of Economic Research, Cambridge, MA. expanded employment between 2004/05 and Collins, S. 2007. “Economic Growth in South 2009/10 (the remaining net additional employ- Asia: A Growth Accounting Perspective.â€? In ment was in regular wage or salaried jobs). Growth and Regional Integration in South 20. This trend toward casualization of the rural Asia, ed. S. Ahmed and E. Ghani, 45– 60. nonfarm labor force was highlighted in the Delhi: Macmillan India Ltd. World Bank’s India poverty assessment for Easterly, W., M. Kremer, L. Pritchett, and L. 1983–2004/05 (World Bank 2011b). Summers. 1993. “Good Policy or Good Luck? 21. The number of working poor in India is likely Country Growth Performance and Temporary to have declined between 1985 and 2005. Shocks.â€? Journal of Monetary Economics Annex 2F explains why poverty rates for the 32 (3): 459–83. ï¬?rst and second subperiods cannot be directly ILO (International Labour Ofï¬?ce). 2010. KILM compared. and LABORSTA databases. Geneva. GROWTH AND JOB QUALITY IN SOUTH ASIA 83 Lanjouw, P., J. Luoto, and D. Mckenzie. 2011. ———. 2011a. Large-Scale Migration and Remit- “Using Repeated Cross–Sections to Explore tances in Nepal: Issues, Challenges and Movements in and out of Poverty.â€? Policy Opportunities. Washington, DC. Research Working Paper 5550, World Bank, ———. 2011b. Perspectives on Poverty in India: Washington, DC. Stylized Facts from Survey Data. Washington, Sundaram, K. 2004. “Growth of Work Oppor- DC. tunities in India: 1983 to 1999–2000.â€? Paper ———. 2011c. World Development Indicators. presented at a conference in honor of K. N. Raj Washington, DC. on Planning, Institutions, Markets and Devel- Young, A. 1994. “Lessons from the East Asian opment, Thrissur, Kerala, India, October. NICs: A Contrarian View.â€? European Eco- UN (United Nations). 2008. World Population nomic Review 38: 964–73. Prospects: The 2008 Revision. New York: ———. 1995. “The Tyranny of Numbers: Con- United Nations. fronting the Statistical Realities of the East ———. 2010. World Population Prospects: The Asian Growth Experience.â€? Quarterly Journal 2010 Revision. New York: United Nations. of Economics 110 (August): 641–80. World Bank. 1993. The East Asian Miracle. Washington, DC. CHAPTER 3 A Proï¬?le of South Asia at Work Questions and Findings Questions agriculture. Faster intersectoral reallocation of employment into industry and services will • What are they key features of labor markets require the development of not just the urban in South Asia? industrial and services sectors but also the • Where are the better jobs, and who holds rural nonfarm sector. them? • Within industry and services, better jobs are • What are the implications for the region’s with large formal ï¬?rms. The majority of work- employment challenges? ers, however, work in informal micro ï¬?rms, where value added per worker is lower and Findings which pay lower wages. Creation of better • Although employment in South Asia has been jobs will require faster intrasectoral reallo- expanding, employment rates have remained cation of labor from lower-productivity— steady and are below those in other regions, typically micro and small informal—ï¬?rms to as a result of persistently low female employ- higher-productivity—typically medium and ment and participation rates. large formal—ï¬?rms within manufacturing and • The majority of workers in the region are services. still engaged in agriculture. Self-employment • The educated are more likely to work outside is the predominant type of employment, and agriculture and be employed in regular wage a high share of wage employment is casual or salaried work. Female workers are less labor. Thus, the vast majority of work in likely to be in better jobs than men, except at South Asia—86–95 percent of total employ- the highest levels of education; they also earn ment and 71–81 percent of nonagricultural less, even after controlling for differences in employment in most countries—is informal educational attainment. Members of ethnic in nature. This picture is unlikely to change minorities are less likely to hold better jobs; signiï¬?cantly in the short to medium term. they also earn less, although much of this dif- • Jobs that pay higher wages and are associated ferential can be explained by differences in with lower poverty rates are found outside of educational attainment. A Profile of South Asia at Work 3 T his chapter profiles employment in conduct labor force surveys (Bangladesh, South Asia. Relying on household India, Nepal, Pakistan, and Sri Lanka), mea- survey data from the region’s eight surement of labor market indicators such as countries, it describes the patterns of par- labor force participation, employment, and ticipation, employment, unemployment, and unemployment is common and generally con- earnings in the region. sistent with international standards. In coun- Describing the labor market in South tries in which other household surveys are Asia is a formidable task. The region’s eight used (Afghanistan, Bhutan, and Maldives), countries vary widely in size, ranging from deï¬?nitions of these (and other) indicators can less than 1 million people each in Bhutan differ from international norms. As a result, and Maldives to 1.2 billion people—about measurement differences explain some of the three-quarters of South Asia’s population—in variation across countries presented in this India. There is diversity in the stages of devel- chapter (Srinivasan 2010 discusses in further opment, economic structures, social and cul- detail how labor market concepts are mea- tural characteristics, and conflict. Even within sured in different surveys). (Annex table 3A.1 countries there is signiï¬?cant diversity. provides more detail on the measurement The profile of South Asia at work pre- of employment and unemployment from sented is based primarily on microlevel data the national surveys as used in this book.) collected by national statistical agencies. Second, as South Asian economies are still The analysis relies on labor force surveys in heavily rural, agricultural, and informal, the some countries and on living standards sur- productive activities of many individuals may veys in others (depending on survey avail- not be fully captured by standard labor mar- ability and data quality). The latest surveys ket indicators. were conducted between 2004 and 2009/10 This chapter is organized as follows. The (see appendix table A.1). first section provides an overview of the Two caveats should be noted regarding main labor market trends, including employ- analysis across countries. First, there are ment, unemployment, and labor force par- limits to the standardization that is pos- ticipation, for the eight countries in South sible, especially between labor force and Asia, with a focus on the employment and living standards surveys. In countries that participation patterns of women. The second 85 86 MORE AND BETTER JOBS IN SOUTH ASIA section takes a closer look at the nature of (2004) to 80 percent in Nepal (2008) (the employment in the region, including loca- other countries in the region have employment tion, sector, employment status, and infor- rates of 50–65 percent). Analysis within coun- mality. The third section examines where tries shows moderate differences in regional the better jobs are. The last section analyses employment rates within countries (see annex how gender, caste/ethnicity, and education 3B). Internationally, the average employment are correlated with access to better jobs. rate is 60–70 percent for low- and lower- middle-income countries (figure 3.2). The employment rate in the three largest countries Overview of employment and in South Asia (India, Bangladesh, and Paki- labor force participation in stan) is signiï¬?cantly below the average rate for South Asia countries at similar levels of development. This section ï¬? rst examines employment in These relatively low employment rates in the region. It then addresses labor force par- South Asia refl ect persistently low female ticipation and unemployment. employment rates in all countries except Bhutan and Nepal (figure 3.3). Employ- ment rates among men are not low by Employment international standards. The (unweighted) Total employment in South Asia is estimated national average for male employment in at 574 million in 2010, with India account- South Asia is 77 percent, which is almost ing for 75 percent, Bangladesh 10 percent, identical to the male average for compara- and Pakistan 9 percent of employment in the tor countries ( Bolivia, Cambodia, China, region (ï¬?gure 3.1). In the region as a whole, Ghana, Guatemala, Indonesia, Lao Peo- 55 percent of the 1.04 billion working-age ple’s Democratic Republic, Nigeria, and population is employed. the Philippines). In contrast, the aver- Employment rates are low by international age employment rate for women in South standards in all countries except Bhutan and Asia is 21 percentage points lower than in Nepal. Employment growth looks favorable comparator countries. The male-female because of the region’s growing working-age employment rate ratio is 2.2 in the region population, as discussed in chapter 2. The and just 1.3 in comparator countries. picture is less positive in terms of employ- There is no consistent evidence of an ment rates. Employment rates among people upward trend in employment rates in South age 15–64 range from 48 percent in Maldives Asian countries (ï¬?gure 3.4). Total employment FIGURE 3.1 Total employment in South Asia, by country, 2010 1,000,000 432,497 100,000 54,103 54,155 14,694 workers (thousands) 7,894 9,666 10,000 1,000 322 90 100 10 1 Maldives Bhutan Sri Lanka Afghanistan Nepal Pakistan Bangladesh India Sources: Authors, based on working-age population ï¬?gures from UN 2010 and employment rate data from national labor force surveys. A PROFILE OF SOUTH ASIA AT WORK 87 FIGURE 3.2 Employment rates in lower- and lower-middle-income countries 90 80 Nepal 70 Bhutan 60 Afghanistan Bangladesh India Sri Lanka percent 50 Pakistan Maldives 40 30 20 10 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 2008 gross national income per capita in purchasing power parity dollars Source: Authors, based on data from World Bank 2011b and national labor force and household surveys. Note: Employment rates are for population age 15 years and above. For all countries, gross national income per capita in 2008 is adjusted for purchasing power parity. Employment rates for countries in South Asia are for latest survey year; employment rates for other countries are for 2008. FIGURE 3.3 Male and female employment rates in South Asia, by country 100 85 86 83 81 80 79 78 80 75 67 61 60 percent 44 40 38 35 29 29 22 20 0 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka 2008 2009 2007 2010 2004 2008 2009 2008 male female Source: Authors, based on data from national labor force and household surveys. rates increased in Maldives and Pakistan, Pakistan, declined in Bhutan and India, and declined moderately in India and Nepal and changed little in the other countries. signiï¬?cantly in Bhutan, and remained fairly These employment figures are for the constant in Bangladesh and Sri Lanka.1 These working-age population (15–64). Child labor, trends mirrored those of female employ- which this book does not address, remains an ment rates, which increased in Maldives and important aspect of the overall employment 88 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.4 Trends in employment rates in South Asia, by country 100 90 85 82 80 74 70 67 64 60 56 57 58 57 55 56 percent 52 50 48 48 44 40 30 20 10 0 2008 2002 2009 2003 2007 2000 2010 1998 2004 1999 2008 2000 2009 2000 2008 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Source: Authors, based on data from national labor force and household surveys. Note: Trend analysis does not take into account cyclical factors. Although the analysis attempted to use standard, consistent deï¬?nitions of employment over time, differences may reflect differences in the questions used to deï¬?ne employment in different survey rounds (see annex table 3A.1 for details). picture in South Asia, as it is in many parts of Underemployment is conventionally deï¬? ned the developing world (box 3.1). as working fewer hours than desired in mature labor markets. This may not be an appropriate definition in developing coun- Labor Force Participation and tries, where people often work for long hours Unemployment even if earnings are very low. In addition, Employment rates in South Asia closely data limitations do not permit a consistent track labor force par ticipation rates, estimate of underemployment. because measured unemployment is very Estimates of the magnitude of underem- low in most countries in the region (table ployment in South Asia vary, based on dif- 3.1). Open unemployment rates in low- ferent definitions. Underemployment was income countries tend to be low, even if estimated at 48 percent of the total workforce labor market conditions are unattractive. in Afghanistan in 2008 (Islamic Republic For the region as a whole, 3.2 percent of the of Afghanistan and World Bank 2010) and labor force—19 million people in a labor 24.5 percent in Bangladesh in 2006 (Rahman force of 593 million—was unemployed in 2008). These ï¬?gures are based on a deï¬?nition 2010. The reported unemployment rate that classiï¬?es as underemployed workers who was high only in Maldives (15.3 percent), work 35 or fewer hours a week on average. In where it mainly reflects the methodology India one measure used by the National Sam- used to calculate unemployment. 2 Unem- ple Survey organization (which deï¬?nes under- ployment in other countries ranged from employment as the proportion of the usually 1.1 percent in Pakistan to 5.6 percent in Sri employed according to the usual status crite- Lanka. In Bangladesh, Sri Lanka, and espe- ria not employed the previous week) estimates cially Maldives, women have higher unem- underemployment at 9–17 percent for women ployment rates than men. In the rest of the and 2–4 percent for men in 2004/05 (Gov- region, there is little gender difference. ernment of India 2006). One problem with Although open unemployment is low in these measures is that they may overestimate South Asia, underemployment—the under- underemployment, because they do not take utilization of labor—may be prevalent. into account individuals who did not wish to A PROFILE OF SOUTH ASIA AT WORK 89 BOX 3.1 Child labor in South Asia According to the International Labour Organization 10 children between the ages of 5 and 14 working. (ILO 2010), 215 million children between the ages of Significant numbers of children are also working 5 and 14 were working in South Asia in 2008, with in Afghanistan, Bhutan, and Pakistan. Bangladesh, 115 million engaged in hazardous work. These ï¬?g- India, and Sri Lanka report lower incidences. ures indicate that 13 percent of all working children Additional dimensions of the statistical picture of in the world live in South Asia; among these children, child labor in South Asia include the following: 7 percent are engaged in hazardous work. According to these ï¬?gures, the incidence of child labor in Asia is • In most countries, boys are somewhat more likely the second highest of all regions, behind Africa. (The than girls to work. However, in Bhutan and Nepal, ILO statistics do not separate South Asia from the two countries with high child labor rates, employ- rest of the continent). ment is higher among girls. Child labor has long played an important role in • Child labor is much more prevalent in rural areas many traditional and agriculturally based societies. than in cities. The vast majority of working chil- It can also be a product of poverty and inequality, dren are engaged in agriculture and ï¬?shing. Other poor education, and confl ict. In South Asia, as in sectors with some child labor are commerce (retail some other regions, concerns about child labor are trade) and manufacturing. heightened by the presence of practices such as child • Although some children are employed as wage trafï¬?cking and bonded child labor (ILO 2010). workers (in Bangladesh and, to a lesser extent, in How prevalent is child labor in South Asia? Many India), most are household enterprise workers. In of the surveys used for this book include questions Nepal, for example, 96 percent of working children that provide data on the incidence of child labor. work in household enterprises. The employment rates for children can be computed • The incidence of child labor continues to decline in the same way they have been calculated for the gradually in Bangladesh, India, and Sri Lanka. working-age population. However, surveys differ in In Nepal and Pakistan, where the incidence of their age coverage: in some countries, employment child labor is higher, there is no clear evidence of rates can be computed for the 5- to 14-year age group; decreases in child labor over time. in others, surveys do not cover children under 10. Survey evidence suggests that the incidence of The nefarious effects of child labor can be miti- child labor varies across the region (box table 3.1.1). gated when working children continue their studies. In Nepal has the highest incidence, with about 3 in Nepal, for example, the country with the highest child BOX TABLE 3.1.1 Incidence of child labor in South Asia, by age group and country (percentage of age group) Age group Country Year 5–9 10–14 5–14 Afghanistan 2008 8.5 23.2 16.1 Bangladesh 2005 0.7 7.1 4.1 Bhutan 2003 — 19.7 — India 2008 0.2 3.4 1.8 Nepal 2008 10.9 47.0 29.7 Pakistan 2009 — 11.8 — Sri Lanka 2008 — 1.1 — Source: Authors, based on data from national labor force and household surveys. Note: — = Not available. No data on child labor are available for Maldives. (continues next page) 90 MORE AND BETTER JOBS IN SOUTH ASIA BOX 3.1 Child labor in South Asia (continued) labor rates in the region, almost 90 percent of chil- and other harmful forms of employment. The three dren who are working also attend school. In contrast, most important conventions are the UN Convention although a smaller percentage of children in Afghani- on the Rights of the Child and ILO Conventions 138 stan, Bangladesh, Bhutan, India, and Pakistan work, (minimum age of employment) and 182 (Elimination most are not in school (box ï¬?gure 3.1.1). of the Worst Forms of Child Labor). All eight coun- The international community has passed a num- tries in South Asia have ratiï¬?ed the UN Convention ber of conventions designed to protect the rights of on the Rights of the Child; only Afghanistan, Nepal, children in the labor market, through both mini- Pakistan, and Sri Lanka have ratiï¬?ed the two ILO mum working ages and protection from hazardous conventions (see chapter 6). BOX FIGURE 3.1.1 Percentage of child workers attending school in South Asia, by age group and country 100 14 13 90 21 80 53 70 68 70 60 84 percent 86 94 91 50 99 86 87 40 79 30 47 20 32 30 10 16 14 6 9 0 1 20 tan 20 esh 20 dia 20 pal 20 tan 20 esh 20 tan 20 dia 20 pal 20 tan 20 nka In In Ne Ne 08 05 05 08 08 05 03 05 08 09 08 is is u kis lad lad a Bh iL an an Pa ng ng Sr gh gh Ba Ba Af Af age 5–9 age 10–14 attending school not attending school Source: Authors, based on data from national labor force and household surveys. Note: Data on 5- to 9-year-olds are not available for Bhutan, Pakistan, and Sri Lanka. TABLE 3.1 Male and female labor force participation, employment, and unemployment rates in South Asia, by country Participation rate Employment rate Unemployment rate Country Year All Male Female All Male Female All Male Female Afghanistan 2008 65 85 45 64 83 44 2 3 2 Bangladesh 2009 60 89 31 57 85 29 5 4 8 Bhutan 2007 69 76 62 67 75 61 2 2 2 India 2010 57 82 30 55 80 29 3 3 4 Maldives 2004 57 72 46 48 67 35 15 8 24 Nepal 2008 84 88 80 82 86 79 2 2 1 Pakistan 2009 52 82 22 52 81 22 1 1 1 Sri Lanka 2008 60 81 41 57 78 38 6 4 8 Source: Authors, based on data from national labor force and household surveys. Note: See annex table 3A.1 for deï¬?nition of employment and unemployment used in each country. The term participation refers to the formal deï¬?nition of labor force participation according to international norms. Application of these norms can be problematic in South Asia, because they do not take into account nonmarket activities of women. A PROFILE OF SOUTH ASIA AT WORK 91 work additional hours. In India, deï¬?ning the can be important determinants of whether underemployed as people who worked at least women participate in the labor market. In three days during the week and spent at least a India, for example, where the overall female half day searching for work, results in an esti- participation rate was 30 percent in 2010, mated underemployment rate of 5 percent.3 the rate among women from scheduled tribes Low female employment rate is primar- (46 percent) was 16 points higher and the rate ily a result of low levels of labor force par- among Muslim women (18 percent) almost ticipation among women. The lowest female 12 points lower. In Sri Lanka, where aggre- labor force participation rates are in the gate female participation was 41 percent in three large South Asian countries: Pakistan, 2008, the rate for Indian Tamil women was where almost four out of every ï¬?ve women 62 percent and the rate for Sri Lankan Moors do not participate in the labor force, and just 17 percent. Bangladesh and India, where slightly more Female participation is especially low in than two out of every three women do not urban areas. Overall labor force participa- participate.4 tion is generally lower in cities than it is in Because female participation is such an rural areas, where labor-intensive, family- important factor in defining the region’s oriented agricultural production still domi- employment picture and its evolution over nates, but this gap is especially striking for time, this issue merits a more detailed look. women. Female rural participation rates are Before proceeding, a caveat about labor higher than urban participation rates in all force statistics and the concept of partici- countries except Bangladesh (where female pation is needed. All of the surveys collect participation is low everywhere); in Afghani- data that make it possible to measure labor stan and Pakistan, the participation rate force participation according to international norms, but the application of these norms to FIGURE 3.5 Female labor force participation rates in South Asia, low-income, traditional societies can be prob- by age group and country lematic. For this reason, it seems unlikely that the actual participation of women, especially 100 in the region’s large countries, is as low as the rates of participation in the surveys indicate. 90 In what follows, the term participation needs 80 to be understood as referring to the formal deï¬?nition of labor force participation; it does 70 not take into account other activities of South 60 Asian women, including reproduction and percent household labor. 50 Except in Bhutan and Nepal, South Asian 40 countries generally have low female participa- tion rates across age groups (ï¬?gure 3.5). This 30 is especially true in Pakistan and, to a lesser 20 degree, Bangladesh, where even in the prime- age groups, the large majority of women are 10 not in the labor force. In all three countries in which data on 0 15–24 25–34 35–44 45–54 55–64 65+ caste/ethnicity are collected (India, Nepal, age group and Sri Lanka), there is considerable variation Nepal, 2008 Bhutan, 2007 Maldives, 2004 in female labor force participation along this Afghanistan, 2008 Sri Lanka, 2008 Bangladesh, 2009 dimension; differences in male labor force India, 2010 Pakistan, 2009 participation are small. The ï¬?nding on caste/ ethnicity is not surprising, as cultural factors Source: Authors, based on data from national labor force and household surveys. 92 MORE AND BETTER JOBS IN SOUTH ASIA for rural women is nearly three times that in the labor force. Sociocultural explanations for urban women. The most recent female have also been put forward, based on the urban participation rates are just 10 percent possible stigma attached to educated women in Pakistan, 18 percent in Afghanistan, and who choose to work. 19 percent in India. Moreover, there is little In all countries except India, the surveys evidence of any signiï¬?cant change, with the ask women not participating in the labor force (unweighted) average female urban participa- why they were not employed or searching for tion in the region increasing from 30 percent work. In all countries, household duties were to 33 percent over the periods studied. the number one reason cited for nonpartici- What factors are associated with female pation (table 3.3). This is especially true of labor force participation? The labor force countries with very low rates of female labor participation status of working-age women force participation (Afghanistan, Bangladesh, was regressed on individual and family char- and Pakistan). This finding is consistent acteristics, using separate logit models for with the fact that women in South Asia, like rural and urban women in each country. women across the world, bear a dispropor- Table 3.2 summarizes the key determinants tionate share of household and care respon- of participation in urban areas. sibilities and therefore face high opportunity The negative relationship between edu- costs when they work in the marketplace. cation and the labor force participation of Social norms also affect these tradeoffs. women has been noted by others studying Education was the second-most fre- the region’s labor market (World Bank 2010). quently cited reason, but there are large Various explanations have been put forward cross-country differences. Substantial num- to explain this relationship. One hypothesis bers of young women in Bhutan, Nepal, and is that better-educated women may opt out Maldives report being in school instead of of the labor market because of the scarcity of the labor force. In contrast, in Afghanistan, good jobs that are available to them and that Bangladesh, Pakistan, and Sri Lanka, fewer an income effect may be at play, with rela- than 2 in 10 urban women not participating tively high family incomes reducing the incen- in the labor force cite education as the main tives for well-educated women to participate reason. TABLE 3.2 Factors associated with participation of women in urban areas Factor Effect Age • Age increases probability of female participation in all countries. Effect weakens later in age distribution. Education • More years of schooling decreases the probability of female participation in all countries except Bhutan. Participation rates tend to be lowest for women who complete secondary or lower-secondary school; they rise only at higher- secondary levels and tertiary levels. Household • Living in a larger household reduces the probability of female participation in all countries except Afghanistan, Nepal, characteristics and Maldives (where there is no effect). • The number of children under the age of six reduces the probability of female participation in all countries except Afghanistan and Bhutan. • Ethnic minority or lower caste status increases the probability of female participation in countries for which data are available (India, Nepal, and Sri Lanka). Marital status • Being married reduces the probability of female participation in all countries except Bangladesh (one of two surveys) and Nepal. Characteristics of other • More years of schooling of the best-educated male in the household reduces the probability of female participation adults in the household in all countries, presumably because it signals an income effect. • Having males in the household who are employed increases the probability of female participation in all countries. • Having a migrant away from the household increases the probability of female participation in India and Nepal but not Maldives (no data on other countries). Source: Authors, based on data from national labor force and household surveys. A PROFILE OF SOUTH ASIA AT WORK 93 TABLE 3.3 Reasons why urban women in South Asia do not participate in the labor force, by country (percent) Country/year Old age Illness Household duties Education Discouraged Other Afghanistan, 2008 0.2 0.6 81.0 12.9 2.2 3.2 Bangladesh, 2005 0.3 2.2 81.1 15.0 0.2 1.0 Bhutan, 2007 3.6 2.7 60.4 26.9 2.6 3.8 Maldives, 2004 — 11.1 46.7 22.8 — 19.5 Nepal, 2008 7.0 0.7 51.6 29.6 4.7 6.3 Pakistan, 2009 1.3 1.0 81.1 16.2 0.1 0.4 Sri Lanka, 2008 5.3 1.8 75.5 15.8 0.4 1.3 Source: Authors, based on data from national labor force and household surveys. Note: — = Not available. Improving economic opportunities for and of the region’s working-age population lives educational attainment of women could con- in rural areas and rural employment rates tribute to improved utilization and allocation are higher than urban rates in all countries of South Asia’s female labor force. Women’s except Maldives. decision to participate in market work is not In Afghanistan, Bhutan, India, and Nepal, independent of the occupational and earn- at least half of all employment remains in ings opportunities available to women in the agriculture. Only in Maldives is this sector labor market as these impact incentives to a relatively minor source of employment. participate. Consistent with global evidence Services are important in most countries, on employment segregation by gender (World representing more than 40 percent of total Bank 2012), women in South Asia are less employment in Bangladesh, Maldives, and likely to access the better jobs (see last section Sri Lanka. The industrial sector, including of this chapter). They also earn signiï¬?cantly manufacturing, utilities, and construction, is less for the same type of job, even after con- relatively small, despite the great importance trolling for differences in education. Improv- attached to industrialization since indepen- ing opportunities requires interventions that dence (Srinivasan 2010). In Bangladesh, relax time constraints, increase access to pro- India, Maldives, Pakistan, and Sri Lanka, ductive inputs, and correct institutional and 20–27 percent of the employed workforce market failures that contribute to employment works in industry, with most of them in segregation. (For a comprehensive discussion manufacturing. As expected, these sectoral of options to improve economic opportuni- patterns differ substantially between rural ties for women, see World Bank 2012.) and urban areas. Agriculture is the largest sector of employment in rural areas in all countries except Maldives and Sri Lanka. In The nature of employment urban areas, most workers are in the service sector. Manufacturing accounts for about This section begins by describing employ- a quarter of urban workers in Bangladesh, ment patterns by location and sector in India, Pakistan, and Sri Lanka. South Asia. It then looks at employment sta- Sectoral employment patterns are chang- tus and informality. ing. The share of agriculture employment in total employment has been declining by about 0.5 percentage points a year in recent Employment patterns by location decades in countries where statistics are and sector available over time. In the ï¬?ve largest coun- Most South Asians work in rural areas tries in the region, employment growth in (table 3.4). The concentration in rural areas agriculture was slower than other sectors in reflects the fact that more than 70 percent the ï¬? rst decade of this century (ï¬?gure 3.6). 94 MORE AND BETTER JOBS IN SOUTH ASIA TABLE 3.4 Distribution of employment in South Asian countries, by location and sector (percent) Total Total Rural Urban Country/ year Agriculture Industry Services Rural Urban Agriculture Industry Services Agriculture Industry Services Afghanistan, 2007 59 13 29 85 15 68 12 21 9 19 72 Bangladesh, 2009 39 21 40 76 24 47 18 35 12 30 57 Bhutan, 2007 68 7 24 78 22 85 4 11 9 19 72 India, 2010 50 23 27 74 26 65 19 16 7 34 59 Maldives, 2004 17 27 55 66 34 24 33 43 5 16 79 Nepal, 2008 73 11 16 87 13 80 9 11 31 21 48 Pakistan, 2009 43 21 36 69 31 60 16 24 6 32 62 Sri Lanka, 2008 31 27 42 90 10 34 26 39 2 31 68 Source: Authors, based on national labor force and household surveys. Note: These data pertain to the area in which the worker’s main employment is located. The classiï¬?cation of the area is based on each country’s classiï¬?cation of rural and urban. Sri Lanka’s classiï¬?cation of rural areas includes the tea estate sector, where a large number of workers are employed. Differences in classiï¬?cation may account for some of the variations across countries. FIGURE 3.6 Annual percentage increases in number of employed workers in South Asia, by sector and country 8 7 6.8 6 5.2 5.0 5 4.7 4.3 4.0 4 3.8 percent 3.0 2.8 3 2.7 2.6 2.5 2.4 2.0 2.2 2 1.5 1.1 1 0.0 0 –0.2 –0.3 –1 Bangladesh India Nepal Pakistan Sri Lanka 2000–10 2000–10 2000–10 2000–10 2000–10 agriculture industry services total employment Sources: Authors, based on data from ILO 2011and national labor force and household surveys. A PROFILE OF SOUTH ASIA AT WORK 95 Total agricultural employment increased sig- by only 1 percentage point over nine years. niï¬?cantly in Nepal and Pakistan during this In contrast, in Bhutan, Maldives, Pakistan, period; it remained constant in Bangladesh and Sri Lanka, it increased 5–11 percentage and declined in Sri Lanka and India. points over six to nine years. The major contributors to job creation In India, employment in the nonfarm sec- everywhere have been industry and services. tor increased steadily for 25 years, rising from Industrial employment has grown very rap- 20 percent of the rural workforce in 1983 to 35 idly in Bangladesh (at almost 7 percent a percent in 2009/10. The pace of diversiï¬?cation year), Pakistan (just over 5 percent a year), away from agriculture increased over time. and India (just over 4 percent a year). Ser- During 1983–1993/94, the average annual vices employment growth has been strongest growth in nonfarm jobs was just over 2 per- in Bangladesh, Nepal, and Pakistan, at 4–5 cent. During 1993/94–1998/99, it increased percent a year. to 3 percent; between 1999 and 2004/05, it The gradual decline in the shares of agri- increased to 4 percent. In the 1980s, of the cultural employment reflects not just rural- nearly 40 million additional rural jobs gener- urban migration but also the growth of the ated in India, 6 out of 10 were in the farm sec- rural nonfarm sector across the region. The tor. In contrast, of the 56 million new rural rural nonfarm sector employs 12–59 percent jobs created between 1993 and 2004, 6 out of the total workforce (15–65 percent of the of 10 were in the nonfarm sector (World rural workforce) in South Asia (ï¬?gure 3.7). Bank 2011a). This trend has continued in Countries that are still primarily rural and recent years: between 2004/05 and 2009/10, agricultural (Bhutan, Nepal) have the small- the nonfarm sector increased from 30 percent est rural nonfarm sectors. to 35 percent of the rural workforce. The pace of development of the rural non- According to data from the 2000 and farm sector varies widely across countries 2008 China National Rural Surveys, trans- and time (ï¬?gure 3.8). In Nepal, the nonfarm formation of the rural labor market has been sector share of the rural workforce increased one of the most salient trends in China’s FIGURE 3.7 Distribution of employment in South Asia, by sector and country 100 15 13 10 90 24 22 26 34 31 80 31 70 60 16 percent 57 44 50 49 69 41 66 40 30 59 50 20 27 32 28 10 26 17 12 0 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka 2007 2005 2007 2010 2004 2008 2009 2008 rural nonfarm rural agriculture urban Source: Authors, based on data from national labor force and household surveys. 96 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.8 Percentage of rural workers in the nonfarm sector in South Asia, by country 100 90 80 76 70 65 60 percent 50 41 42 40 40 35 35 30 30 24 25 20 19 20 20 15 10 8 0 2002 2005 2003 2007 1983 1994 2000 2005 2010 1998 2004 1999 2008 1999 2008 Bangladesh Bhutan India Maldives Nepal Pakistan Source: Authors, based on data from national labor force and household surveys. development since the 1980s. India’s over- • Local public rural infrastructure provision all pace of transition has been slower than was superior (as a result of higher levels of China’s. The share of rural labor force off decentralization in China). the farm in China was lower than in India in Unlike in East Asia, most nonfarm jobs the early 1980s (a few years after the start of in South Asia are in the service sector, with the major rural reforms adopted in 1978 that commerce the largest subsector, employ- abolished the commune system and intro- ing 12–33 percent of nonfarm rural work- duced the household responsibility system). ers (figure 3.10). The manufacturing sec- However, by 2000 the share had become tor—which in other developing countries, significantly greater than India’s, at 43.5 especially East Asia, was the major source percent; by 2008 the share of China’s rural of employment for workers moving out of labor force that worked off the farm was 62 agriculture—provides less than 30 percent percent—twice that of India (ï¬?gure 3.9). of nonfarm jobs. Mukherjee and Zhang (2007) offer a number of explanations for the faster pace of rural nonfarm sector development in China: Employment Status • Local governments in China were incentiv- South Asia is far from a typical, modern ized to support the town and village enter- labor market dominated by wage or salaried prises, as they generated revenue for them. employees. In all countries except Maldives • China’s rural nonfarm sector was less and Sri Lanka, most workers are self- protected than India’s. When the sector employed (box 3.2). The dominance of self- was liberalized, it was more competitive employment is most extreme in Afghanistan, than the protected small-scale sectors in Bhutan, and Nepal, where more than three India, which were not able to compete out of every four workers are self-employed. after liberalization. The scarcity of secure work forms is even • Rural literacy was higher, making it eas- more striking when wage employment is ier for workers to move into the nonfarm broken down into regular wage or salaried sector. workers and casual workers. In Afghanistan, A PROFILE OF SOUTH ASIA AT WORK 97 FIGURE 3.9 Percentage of rural workers in the nonfarm sector in China and India, 1983–2008 100 90 80 70 62 60 percent 50 44 40 30 31 30 24 25 20 20 15 10 0 1983 1994 2000 2004 2008 India China Source: Wang, Huang, and Zhang 2011. Note: Data for China are not available for 1994 and 2004. FIGURE 3.10 Rural nonfarm sector employment in South Asia, by economic activity and country 100 11 15 90 17 16 17 26 3 26 30 2 3 3 4 80 2 9 2 10 1 7 12 4 70 7 18 0 13 9 7 20 60 17 1 5 27 percent 3 26 21 50 24 12 7 33 40 29 8 17 15 10 2 30 17 0 1 23 7 9 1 0 1 20 3 33 28 0 22 27 10 21 21 12 14 0 1 0 2 2 0 1 0 2 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka 2007 2005 2007 2010 2004 2008 2009 2008 other services public administration ï¬?nancial, insurance, real estate transportation commerce construction electricity and utility manufacturing mining Source: Authors, based on data from national labor force and household surveys. 98 MORE AND BETTER JOBS IN SOUTH ASIA BOX 3.2 Composition of the labor force by employment status Regular wage or salaried workers are deï¬?ned as regu- self- employed workers are typically farmers work- larly paid wage employees in the public or private sec- ing their own land, though many self-employed tors. These workers are usually on the regular payroll workers work in the rural nonfarm sector. There of the enterprises for which they work and usually are signiï¬?cant gender differences in the type of self- earn leave and supplementary beneï¬?ts. A signiï¬?cant employment, with women much more likely than proportion of regular wage or salaried work is in the men to be classiï¬?ed as family enterprise workers. In public sector, ranging from 27 percent in India (2010) most countries in the region, men are more likely to to 66 percent in Afghanistan (2007) (the proportions work as own-account workers. in other countries were 29 percent in Bangladesh in The category of the self-employed is very hetero- 2005; 42 percent in Nepal in 2008; 40 percent in Pak- geneous. It can be split into two groups:. istan in 2009; and 52 percent in Sri Lanka in 2008). Casual laborers are deï¬? ned as wage workers who • The high-end self-employed subgroup consists of are paid on a casual, daily, irregular, or piece-rate all employers and other self-employed workers basis. These workers typically do not have access who work as ofï¬?cials, managers, professionals, to formal instruments of social protection. In rural technicians, and clerks. On average, these work- areas, casual laborers are often landless agricultural ers are more educated than other self-employed help, though a signiï¬?cant number of casual workers workers. Their consumption distribution pro- work in the rural nonfarm sector (in, for example, ï¬? le is more similar to regular wage or salaried construction). workers. Self-employed workers consist of employers, • The low-end self-employed subgroup consists own-account workers, and unpaid family enterprise of own account and unpaid family workers who workers. They represent the largest group of work- work as service workers, skilled agricultural work- ers in most South Asian countries, ranging from ers, craftspeople, machine operators, and workers 43 percent in Sri Lanka to 82 percent in Nepal. in elementary occupations. Their consumption The majority of the self-employed are own-account proï¬?les are similar to those of casual laborers (box workers or family enterprise workers. In rural areas, ï¬?gure 3.2.1). BOX FIGURE 3.2.1 Distribution of per capita household expenditure in India and Nepal, by employment status a. India, 2005 b. Nepal, 2003 0.003 0.0015 Kernel density 0.002 Kernel density 0.001 0.001 0.0005 0 0 0 1,000 2,000 3,000 0 1,000 2,000 3,000 monthly per capita expenditure (Indian rupees) monthly per capita expenditure (Nepalese rupees) regular wage casual wage self-employed (high end) self-employed (low end) Source: Authors, based on data from national labor force and household surveys. A PROFILE OF SOUTH ASIA AT WORK 99 Bangladesh, India, and Nepal, more than shift toward regular wage work in Nepal half of wage earners are casual workers. and Pakistan. In contrast, within the rural Only in Bhutan is wage employment domi- nonfarm sectors in Bangladesh (2002–05) nated by regular wage or salaried workers. and India (1983 –2010), there was an The pattern of employment differs by increase in the share of casual labor, both location. A larger share of rural workers than as a share of total rural nonfarm employ- urban workers is self-employed. The major- ment and as a share of rural nonfarm wage ity of rural wage earners are casual laborers, employment. whereas the majority of urban wage earners are regular wage or salaried workers. Informality The distribution of workers by employ- ment status changed very little, if at all, in The issue of informality is a prominent one the past decade (table 3.5). Only in Bhutan in South Asia. Since the term informal sec- and Maldives did employment shift mark- tor was coined, about 40 years ago, consid- edly toward wage work, and it is not known erable efforts have been made to deï¬? ne and if the increase there was in regular wage or measure informality. (For a comprehensive salaried work, casual labor, or both. In the discussion of measurement and statistics, other countries, self-employment continues see ILO 2002. For a deï¬? nition of informal to dominate, with the share of wage employ- employment in India, see the National Com- ment growing very little in the past decade. mission for Enterprises in the Unorganised Within wage employment, there was a slight Sector 2009.) TABLE 3.5 Distribution of employment in South Asian countries, by type of employment (percent) Wage employment Self-employment Regular wage or Country/year salaried Casual labor Employer Own account Family enterprise Latest year Afghanistan, 2008 9 14 0.5 44 33 Bangladesh, 2005 14 22 0.3 41 22 Bhutan, 2007 21 4 0.2 25 50 India, 2010 17 32 1 33 17 Maldives, 2004 55 4 31 10 Nepal, 2008 8 10 1 35 46 Pakistan, 2009 21 17 1 34 28 Sri Lanka, 2008 57 3 29 11 Earlier year Bangladesh, 2002 14 22 0.4 44 19 Bhutan, 2003 14 0.2 18 68 India, 2000 15 31 1 31 21 Maldives, 1998 48 6 37 9 Nepal, 1999 6 10 1 39 44 Pakistan, 2000 19 18 1 41 21 Sri Lanka, 2000 57 2 28 13 Source: Authors, based on data from national labor force and household surveys. Note: Labor force surveys in Bhutan, 2003; Maldives; and Sri Lanka do not allow the separation of wage employment into regular wage workers and casual laborers. Afghanistan has only one survey. 100 MORE AND BETTER JOBS IN SOUTH ASIA Information on employment status and Asia is informal (figure 3.12). Informal- sector, ï¬? rm characteristics, and worker edu- ity rates based on pension coverage in all cation from national labor force surveys is South Asian countries except Sri Lanka are used to compare levels of informal employ- higher than in other countries with simi- ment across the region. Informal workers lar levels of gross domestic product (GDP). include all workers in the informal sector as Together with Africa, South Asia has the well as workers in the formal sector perform- highest rate of informal employment in the ing informal jobs: all workers in agriculture; world (ï¬?gure 3.12). wage workers in informal enterprises; and Although informal employment has casual laborers, family enterprise workers, traditionally been seen as a labor market and self-employed workers with less than problem—because informal workers tend to senior-secondary education in the nonagri- have low earnings and little access to formal cultural sectors. (Annex table 3A.2 shows social protection systems—recent research, country details.) Based on this deï¬?nition, an especially in Latin America, suggests that in estimated 86–95 percent of employment is some situations, individuals may choose to informal in all countries except Maldives and work informally. Analysis in South Asia has Sri Lanka, and 71–81 percent of nonagricul- emphasized the vulnerability and involuntary tural employment is informal in all countries nature of informality in the region (Chen except Bhutan, Maldives, and Sri Lanka and Doane 2008; National Commission for (ï¬?gure 3.11). Enterprises in the Unorganised Sector 2009). The estimated rates are consistent with The next section of this chapter shows that other studies showing that the vast major- informal workers in South Asia are less ity of employment in South Asia is infor- skilled, earn less, and have higher poverty mal. 5 Using lack of pension coverage as a rates than formal workers and that informal proxy, Loayza and Wada (2011) estimate manufacturing and services ï¬?rms have lower that 91 percent of the labor force in South labor productivity and pay lower wages than FIGURE 3.11 Percentage of employment in South Asia classiï¬?ed as informal, by country 100 95 92 90 87 88 89 86 79 81 80 80 74 71 71 70 60 58 56 51 percent 50 43 40 30 20 10 0 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka 2008 2005 2007 2010 2004 2008 2009 2008 all sectors nonagricultural Source: Authors, based on data from national labor force surveys. A PROFILE OF SOUTH ASIA AT WORK 101 formal ï¬? rms in the same size class. Loayza FIGURE 3.12 Percentage of labor force not covered by pension and Wada (2011) also point out that labor schemes, by region informality is higher than predicted by the level of production informality, suggesting 100 that the productivity of informal workers in South Asia is relatively low compared with that of informal workers in other countries. 80 The pervasiveness of informality in South Asia is likely to remain a core feature for a 60 percent long time. Informality is a complex, multi- faceted phenomenon that is shaped by both 40 the modes of socioeconomic organization and the relationship the state establishes with private agents through regulation and moni- 20 toring. Loayza and Wada (2011) show that the actual rates of labor informality in South 0 Asia are similar to predicted levels based on om ed As d M bbe nd Af and ciï¬? d sia a ric La tral an Pa an the determinants of informality: the legal and on lop i a hA Eu s th Am a d an Ea ica c Af ie i n pe ica rth ast ia r ec eve ut As Ce ro e C er No le E regulatory framework, educational achieve- So st d ar id tin ment, the share of youth or rural popula- tion, and the sectoral production structure (box 3.3). Most of these determinants are Source: Loayza and Wada 2011. BOX 3.3 Determinants of informality Using t wo labor informalit y measures (self- structure tilted toward agriculture favors informality employment and lack of pension coverage), Loayza by making legal protection and contract enforcement and Wada (2011) show that in cross-country regres- less relevant and valuable. Third, larger shares of sions, labor informality is negatively and signiï¬?cantly youth or rural populations are likely to increase infor- related to the strength of law and order, business mality, make monitoring more difï¬?cult and expensive, freedom from regulations, and average years of sec- place greater demands on resources for training and ondary schooling and positively and significantly the acquisition of abilities, create bottlenecks in the associated with sociodemographic transformation initial school-to-work transition, and make it more factors (the share of agriculture, youth population, difï¬?cult to expand formal public services. and rural population). All correlation coefï¬?cients are Bangladesh, India, Pakistan, and Sri Lanka have highly statistically signiï¬?cant (p-values of less than larger predicted informality levels than the grow- 1 percent) and of large magnitude (0.68–0.83). The ing East Asian countries (the Republic of Korea, predicted levels of informality for South Asian coun- Malaysia, and Singapore). Sociodemographic factors, tries are similar to actual levels. in particular the region’s high ratio of rural popula- These results have several implications. First, tion, are the largest contributors to the differences informality is more prevalent when the regulatory in predicted informality levels between South Asian framework is burdensome, the quality of government and East Asian comparator countries. Lower busi- services to formal ï¬? rms is low, and the state’s moni- ness regulatory freedom (for all countries) and low toring and enforcement power is weak. Second, the levels of education (for all countries except Pakistan) structural characteristics of underdevelopment play an play a moderate but consistent role in explaining dif- important role in explaining informality. Other things ferences in informality. Law and order does not play equal, a higher level of education is likely to reduce a major role in explaining the differences. informality by increasing productivity, potentially increasing returns to formalization. A production Source: Loayza and Wada 2011. 102 MORE AND BETTER JOBS IN SOUTH ASIA structural and take time to change. For this Among South Asian countries in which reason, large informal sectors continue to headcount poverty rates by employment sec- exist even after economies have experienced tor and status of household members can be rapid growth.6 In South Asia, despite grow- analyzed, agricultural casual workers have ing labor productivity and increasing qual- the highest poverty rates in Bangladesh and ity of jobs, there is little evidence from labor Nepal and the second highest in India (after force or industrial surveys that labor infor- urban casual labor) (ï¬?gure 3.13).8 Agricul- mality is decreasing.7 tural self-employment provides a relatively As the correlates of informality are largely good source of income (as proxied by pov- structural, the book assumes that formaliza- erty rates) in Bangladesh and India but not tion will be a slow process. Easing restric- in Nepal. The jobs associated with the low- tive labor legislation and other interventions est poverty rates in all countries are regu- to improve the business regulatory envi- lar wage work outside of agriculture in the ronment can contribute to lower informal- urban or rural nonfarm sectors. Urban self- ity, but large increases in formality are not employment in Bangladesh and Nepal is also expected to occur immediately. Therefore, associated with lower poverty rates. the approach taken in this book is to aim to The wage data tell a consistent story: work- improve the quality of all types of jobs by ers in industry and service sectors are better addressing constraints to the productivity of paid than workers in agriculture (ï¬?gure 3.14). all workers and ï¬? rms, formal or informal. Rural nonfarm employment (regular wage Such an approach is likely to have a positive and casual labor) and urban regular wage effect on reducing informality, as increasing employment offer higher wages than agri- productivity may increase the returns to for- cultural casual labor in Bangladesh, India, malization. As the majority of workers will Nepal, and Pakistan (the countries for which remain informal at least in the near term, wage employment can be split into regular efforts should be made to increase their wage or salaried and casual labor in all time access to programs that help manage labor periods observed): market shocks (see chapter 6). • Rural nonfarm and urban regular wages are 30–100 percent higher than wages for Where are the better jobs? agricultural labor. • Wages for rural nonfarm casual labor This book deï¬? nes “better jobsâ€? as jobs asso- are 10–50 percent higher than wages for ciated with higher wages (for wage workers), agricultural labor. Even with the consis- lower poverty, and a lower risk of low and tent increase in the share of rural nonfarm uncertain income. This section examines jobs, nonfarm casual labor wages in India which sectors, types of employment, and have remained about 30–50 percent higher types of ï¬? rms are associated with better job than agricultural wages since 1983. quality. • Wages for urban casual labor are higher than wages for agricultural casual labor By sector in Nepal and Pakistan (20–30 percent) but the same in Bangladesh and India. This Employment in industry and services—in evidence is consistent with the high pov- urban areas or the rural nonfarm sector— erty rates observed among urban casual yields higher wages and is associated with labor in Bangladesh and India. lower poverty rates than agricultural casual labor. Chapter 2 showed that labor produc- Wage differentials between services and tivity (measured as output per worker) is agriculture are especially large (ï¬?gure 3.15). much higher in industry and services than in In India, average hourly wages in services agriculture. The higher productivity in these were 135 percent higher than in agriculture in sectors is manifested in higher earnings. 2010. In Nepal, Pakistan, and Sri Lanka, the A PROFILE OF SOUTH ASIA AT WORK 103 FIGURE 3.13 Percentage of workers in households below the poverty line in Bangladesh, India, and Nepal, by employment status a. Bangladesh, 2010 b. India, 2004/05 c. Nepal, 2003/04 agricultural casual 51 urban casual labor 45 agricultural casual labor 50 labor rural nonfarm casual agricultural casual rural nonfarm casual labor 45 31 43 labor labor urban casual labor rural nonfarm casual agricultural 44 25 31 labor self-employed rural nonfarm regular wage rural nonfarm or salaried worker 26 urban self-employed 24 17 self-employed rural nonfarm rural nonfarm self-employed 26 18 urban casual labor 16 self-employed agricultural agricultural rural nonfarm regular self-employed 22 17 10 self-employed wage or salaried worker urban regular wage urban regular wage or salaried worker 16 14 urban self-employed 9 or salaried worker rural nonfarm regular urban regular wage urban self-employed 14 8 3 wage or salaried worker or salaried worker 0 50 0 50 0 50 percent percent percent Source: Authors, based on data from national labor force and household surveys. Note: Figures are for workers age 15–64. Poverty rates for India are based on official poverty lines prevailing until 2010. Using the new official poverty lines for 2004/05 (revised in 2011) would increase poverty rates in rural areas, making the poverty rates of rural workers higher than those of urban workers for the same employment type. The hierarchy in terms of employment type would remain the same. FIGURE 3.14 Ratio of median rural nonfarm and urban wages to agricultural wages in selected South Asian countries 4.0 3.0 3.0 2.8 2.4 2.2 2.2 2.1 2.1 2.0 2.0 ratio 2.0 2.0 1.8 1.8 1.8 1.6 1.7 1.6 1.6 1.6 1.5 1.6 1.5 1.5 1.4 1.5 1.4 1.4 1.3 1.3 1.3 1.3 1.3 1.2 1.1 1.1 1.1 1.1 1.0 1.0 1.0 1.0 0.9 0 2002 2005 1994 2000 2005 2010 1999 2008 2000 2009 Bangladesh India Nepal Pakistan rural nonfarm regular wage or salaried worker rural nonfarm casual labor urban regular wage or salaried worker urban casual labor Source: Authors, based on data from national labor force and household surveys. Note: A ratio of 1 means the median wage was equivalent to the median wage of agricultural casual labor. 104 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.15 Ratio of median industry and service sector wages to agricultural casual labor without controlling agricultural wages in selected South Asian countries for other factors was 40 percent in 2000 and 50 percent in 2008; after controlling for other 2.5 2.35 factors, including educational attainment of 2.1 the workers, it was 24 percent in both years. 2.0 2.0 1.8 There is thus a reallocation gain even without 1.5 1.5 1.4 1.4 additional investment in education. 1.5 ratio 1.0 By employment status and ï¬?rm type 0.5 Within industry and services wage employ- ment, better jobs can be defined from two 0 main angles: by worker’s employment status 2010 2008 2009 2008 India Nepal Pakistan Sri Lanka and by type of ï¬?rm. Regular wage or salaried service/agriculture workers have considerably higher average industry/agriculture hourly wages than casual laborers (see ï¬?gure Source: Authors, based on data from national labor force and household surveys. 3.14). The two types of wage employment Note: A ratio of 1 means the median wage was equivalent to the median wage in agriculture. correspond roughly to formal and informal employment in the wage sector.10 In Bangla- desh, India, Nepal, and Pakistan, regular wage workers earn 20–48 percent more than ratio of wages in services to wages agricul- casual workers in the rural nonfarm sector ture was also very high (112 percent higher and 14–74 percent more than casual workers in Nepal, 84 percent higher in Pakistan, and in the urban sector. These wage differentials 99 percent higher in Sri Lanka). partly reflect differences in education levels Industry and service jobs are better paid between regular wage workers and casual partly because they are higher-skilled jobs. laborers. Casual labor has the highest pov- Nonagricultural workers have higher average erty rates; poverty rates are generally lowest levels of education than agricultural work- for regular wage or salaried workers, among ers. Within the nonagricultural sector, regu- whom poverty rates are on average one third lar wage workers are more educated than or less those for casual labor. casual workers, with a signiï¬?cant proportion In addition to lower wages, casual labor of workers having secondary education or employment offers less stability than regular above. The higher wages observed in industry wage work. In India in 2009/10 (depending and service employment partly reflect higher on the sector of employment), 31–49 percent levels of education. of casual laborers, 8–23 percent of the self- However, the majority of industry and ser- employed (own- account and family helper), vice jobs pay more than agricultural casual and 4–7 percent of regular wage or salaried labor even after accounting for higher levels workers reported spending at least one month of education (and other individual and house- looking for work in the past year. Casual hold characteristics that can affect wages, laborers spent an average 0.9–1.4 months such as age, gender, caste/ethnicity, religion, without work the past year. Self-employed and household wealth).9 The wage premium workers reported 0.2–0.7 months of unem- is highest for public sector regular wage or ployment; regular wage or salaried workers salaried workers, where the premium over reported virtually no time without work.11 the agricultural wage is 41–50 percent in It is more difï¬? cult to assess the quality Bangladesh, India, and Nepal. Private sector of self-employment because there are no nonfarm sector jobs also offer a wage pre- wage data and self-employed workers are a mium. In India, for example, the wage dif- very heterogeneous group. The small pro- ferential between nonfarm casual labor and portion of workers classiï¬?ed as “high-endâ€? A PROFILE OF SOUTH ASIA AT WORK 105 self-employed have poverty rates that are TABLE 3.6 Distribution of formal and informal similar to those of regular wage or salaried manufacturing ï¬?rms in India, by location and size, workers. Most self-employed workers—who 2005 are typically involved in small-scale enter- (percent) prises (enterprises with no more than five Firm Formal Informal workers) as own-account workers or family characteristic sector sector Total enterprise workers—have lower poverty rates Percentage than casual workers but higher rates than reg- of all firms 0.7 99.3 100.0 ular wage workers in India and Nepal. The Location India poverty assessment (World Bank 2011a) Urban 60.2 29.0 29.2 reports that only about half of nonfarm self- Rural 39.8 71.0 70.9 employed workers regard their earnings from Firm size self-employment as remunerative. 1–49 74.4 100.0 99.8 50–99 11.9 0.0 0.1 Analyzing the question from the point of 100+ 13.7 0.0 0.1 view of the ï¬?rm is more challenging because of the lack of nationally representative ï¬? rm- Sources: Authors, based on data on formal ï¬?rms from the Annual Survey of Industries and data on informal ï¬?rms from the National Sample Survey level industrial data. Only India conducts manufacturing surveys. national surveys that cover all ï¬? rms in the manufacturing and services sectors, infor- levels of firms employing more than 200 mal and formal, rural and urban. Based on workers (ï¬?gure 3.16). Informal ï¬?rms with one the estimated level of informality and self- to four employees have 25 percent of the value employment (especially own-account and added per worker and less than 50 percent of family enterprise work) indicated in national the wages per worker of formal ï¬?rms the same labor force surveys, there is reason to believe size. Although productivity is lower in rural that the statistics presented below for India firms than in urban firms of the same size are typical of other South Asian countries.12 class and type, especially in the informal sec- Among Indian manufacturing and services, tor, real wages (adjusted using consumption- workers in larger formal firms earn higher based household deflators rather than indus- wages, which reflect higher value added per try deflators) in rural and urban ï¬?rms of the worker. However, the majority of manufactur- same size and type are similar.14 ing and services employment is in micro/small Although the ï¬? rm size productivity/wage informal ï¬?rms, which have lower value added differential exists in other regions, it is par- per worker and, hence, pay lower wages. ticularly high in India. In East Asia (China; Indonesia; Korea; Malaysia; the Philippines; India’s manufacturing sector Taiwan, China; and Thailand), the produc- In 2005, 99 percent of the 17.1 million man- tivity of small enterprises (enterprises with ufacturing ï¬? rms in India were informal (the 5–49 workers) is about 20–40 percent that of term used in India is unorganized); nearly enterprises with more than 200 workers (ADB all ï¬? rms were micro/small enterprises (1–49 2009). In India, small ï¬? rms have 12 percent workers) (table 3.6). Even in the formal sec- of the productivity and pay 19 percent of the tor (referred to in India as the organized sec- wages of large ï¬?rms. tor), 74 percent of ï¬? rms were micro/small.13 A number of factors may accounts for Seventy-one percent of all ï¬? rms were rural. the ï¬? rm size productivity/wage differentials Average productivity (value added per observed everywhere: 15 worker) and average wages are lower in smaller ï¬? rms than in larger ï¬? rms and much • Smaller and informal ï¬? rms are less capital lower in informal ï¬?rms than in formal ï¬?rms intensive. In India, for example, informal the same size. Value added and wages per ï¬? rms have less than a quarter of the capi- worker in formal firms with one to four tal per worker of formal ï¬? rms in the same employees are on average one-quarter the size class. 106 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.16 Average wage, value added, and capital per manufacturing worker in India, by ï¬?rm size and type, 2005 (percentage of average wages/value added/capital per worker in formal firms with 200 or more employees) a. Mean wage per worker b. Median wage per worker 100 100 90 90 80 80 70 66 70 70 64 60 60 60 60 percent percent 54 50 50 50 43 40 40 37 32 30 27 30 25 19 23 20 17 20 18 9 11 10 10 0 0 4 9 9 9 9 9 4 9 9 9 9 9 1– 5– –1 –4 –9 19 1– 5– –1 –4 –9 19 0– 0– 10 20 50 10 20 50 10 10 firm size firm size d. Median value added per worker 100 100 c. Mean value added per worker 90 90 80 80 70 70 66 60 56 60 percent 55 percent 53 50 45 50 47 40 39 40 37 30 26 30 26 19 21 19 21 20 20 10 12 10 4 10 6 0 0 4 9 9 9 9 9 4 9 9 9 9 9 1– 5– –1 –4 –9 19 1– 5– –1 –4 –9 19 0– 0– 10 20 50 10 20 50 10 10 firm size firm size e. Mean capital per worker f. Median capital per worker 100 100 90 90 80 80 71 70 70 62 60 60 percent 50 47 50 44 45 40 39 percent 40 32 40 31 31 30 26 30 20 13 20 12 10 5 10 7 7 3 0 0 4 9 19 9 9 4 9 9 9 9 99 1– 5– –4 –9 1– 5– –1 –4 –9 – 1 0– 10 20 50 10 20 50 10 firm size firm size formal firms informal firms Sources: Authors, based on data on formal ï¬?rms from the Annual Survey of Industries and data on informal ï¬?rms from the National Sample Survey manufacturing surveys. Note: Formal ï¬?rms with 200 or more employees = 100 percent. • Workers in larger formal enterprises are perhaps as a result of economies of scale, more skilled. and they may pay “efï¬?ciency wagesâ€? (higher • Large formal enterprises are more produc- than market-clearing wages paid to encour- tive with the labor and capital they have age higher output, raise worker morale, and (their total factor productivity is higher), discourage absenteeism and shirking). A PROFILE OF SOUTH ASIA AT WORK 107 Eighty-one percent of workers in India sector became more dynamic, growing at 3 work in informal firms, and 83 percent of percent a year to 8.7 million. At the same workers are in micro or small firms, with time, employment in the informal sector most working in own-account manufactur- declined by 0.3 percent a year. The opposing ing enterprises with fewer than ï¬?ve workers. dynamics of the two periods resulted in very Combined with the large productivity and similar share of employment in the informal wage differentials observed, these figures sector and in micro/small ï¬? rms in 1994 and translate into low-productivity and low-wage 2005. jobs for the majority of the 45.1 million man- The share of employment in micro and ufacturing workers in India in 2005. small ï¬?rms is much larger in India than it is in The distribution of employment across East Asia (ï¬?gure 3.18). A number of factors can ï¬? rm size groups has not shifted over time: affect the enterprise size distribution, includ- the share of employment in micro/small ï¬?rms ing industrial composition, infrastructure, and the informal sector in 2005 is almost product market segmentation, coordination the same as it was in 1994 (figure 3.17). failures, credit constraints for small firms, Between 1994 and 2000, employment in the regulations that differ according to enterprise informal sector increased 4 percent a year, size, and industrial policy (ADB 2009). Lower- from 28.8 to 37.0 million, whereas employ- income countries tend to produce less complex ment in the formal sector remained stagnant goods, which are better suited to small-scale at 7.7 million (employment in the largest production. ï¬? rms actually fell 1 percent a year). Between Part but not all of India’s concentration of 2000 and 2005, employment in the formal employment in small ï¬? rms can be explained FIGURE 3.17 Share of manufacturing employment in India, by ï¬?rm size and type, 1994–2005 100 90 80 70 60 percent 50 40 30 20 10 0 4 9 9 5 9 0– 9 9 0+ 4 9 9 50 9 10 –99 9 0+ 4 9 20 9 50 9 10 9 9 0+ 1– 5– –1 –4 10 0–9 19 1– 5– –1 –4 19 1– 5– –1 –4 –9 19 20 20 20 0– 0– 10 20 10 20 10 1994 2000 2005 informal directory manufacturing establishments (at least 1 hired worker and more than 6 workers in total) informal nondirectory manufacturing establishments (at least 1 hired worker but fewer than 6 workers in total) informal own-account manufacturing enterprises (no hired workers) formal Sources: Authors, based on data on formal ï¬?rms from the Annual Survey of Industries and data on informal ï¬?rms from the National Sample Survey manufacturing surveys. Note: The data show a small share (1 percent or less) of informal employment in the larger ï¬?rms as well. 108 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.18 Share of manufacturing employment by ï¬?rm size in India’s service sector India and selected East Asian economies India had 16.5 million services ï¬?rms in 2006, 15 percent of which were establishments and India 83 6 11 85 percent were own-account enterprises.16 Ninety-eight percent of own-account enter- Philippines 70 8 23 prises and 73 percent of establishments have Indonesia 65 6 29 one to four workers, and nearly all ï¬? rms are Korea, Rep. 47 24 30 micro or small (table 3.7). In contrast to the manufacturing sector, India’s service sector Thailand 46 13 42 has very few establishments with 50 or more Taiwan, China 39 21 40 workers. Sixty percent of all services ï¬? rms are rural. Malaysia 28 20 53 Average productivity (value added per China 25 23 52 worker) and average wages are lower in smaller ï¬?rms than in larger ï¬?rms (ï¬?gure 3.19). 0 20 40 60 80 100 percent For example, mean wages at ï¬?rms with fewer than 19 employees are 20–47 percent those micro and small (1–49) medium (50–199) large (200+) of ï¬? rms with 20–49 workers. Although the Sources: Authors, based on data from ADB 2009, India Annual Survey of Industries, and National average wage monotonically increases with Sample Survey manufacturing surveys. ï¬?rm size, average productivity and capital per Note: Data for India are for 2005. Data for East Asia data are for latest year available between 2004 and 2007. worker do not always increase among ï¬? rms with 5–9 and ï¬?rms with 10–19 employees. Productivity and wages are also much lower in own-account enterprises ï¬?rms than in establishments of the same size class. by its level of income and industrial composi- Own-account enterprises with one to four tion. For example, India’s apparel industry is employees (the vast majority of such enter- dominated by small ï¬? rms, whereas China’s prises) have less than half the value added industry is dominated by medium-size and per worker and less than 15 percent of the large ï¬? rms (ADB 2009). It is possible that mean wage per worker of establishments the past industrial policies in India, particularly same size. One reason for this differential is the policy of reserving production of certain that own-account enterprises are less capital products for small-scale industries, which intensive. started in the late 1960s, has had signiï¬?cant An estimated 27.5 million workers were effects on ï¬? rm size distribution that are still employed in the service sectors covered in being felt today. India’s 2006 National Sample Survey. Sixty Data from the Annual Survey of Indus- percent of service workers worked in own- tries since 2005 show continued dynamism account enterprises and 40 percent in estab- in the growth of employment in the for- lishments. The majority of workers worked mal sector, suggesting that a J-curve effect in lower-productivity and lower-wage micro may be at work: reforms in the 1990s led ï¬?rms (own-account enterprises and establish- to an initial decline of employment in the ments with fewer than ï¬?ve workers). The con- formal sector in medium-size and large centration of employment in micro and small ï¬? rms before increasing from 2000 onward enterprises (enterprises with 1–49 workers) is (box 3.4). (National Sample Survey data even higher in the service sector (96 percent of on the informal sector are not available workers) than in manufacturing (83 percent). after 2005, making it impossible to analyze The employment pattern in services has the growth of the informal manufacturing not changed significantly over time, with employment and determine how the overall the share of employment in micro/small and distribution of manufacturing employment own-account enterprises remaining fairly has evolved since then.) stable between 2001 and 2006 (ï¬?gure 3.20). A PROFILE OF SOUTH ASIA AT WORK 109 BOX 3.4 Trends in India’s formal manufacturing sector, 1998–2007 Total employment in India’s formal manufacturing the decade, with employment in nonprivate/state sec- sector follows a U-shaped path, declining from 1998, tor ï¬? rms (mostly large ï¬? rms) declining steadily and reaching a trough between 2001 and 2003, and growing signiï¬?cantly as the public sector downsized. Initially, rapidly thereafter. This path was driven by employment this pattern led to a decline in overall employment, as in large ï¬?rms; employment in small and medium-size private sector employment remained stable through ï¬?rms (enterprises employing fewer than 100 workers) 2003. Beginning in 2003, private sector employment was fairly stable through 2003 and grew thereafter. rapidly increased, more than offsetting the continuing The U-shaped pattern in employment can be decline in the state sector. explained by the large compositional change over BOX FIGURE 3.4.1 Employment in India’s formal manufacturing sector, by ï¬?rm size, type, and location, 1998–2007 a. By location and ï¬?rm size b. By ï¬?rm ownership (private versus nonprivate) 12 12 10 10 8 8 millions millions 6 6 4 4 2 2 0 0 98 99 00 01 02 03 04 05 06 07 98 99 00 01 02 03 04 05 06 07 19 19 20 20 20 20 20 20 20 20 19 19 20 20 20 20 20 20 20 20 Source: Authors, based on data from Annual Survey of Industries. Note: Annual data on the formal manufacturing sector are available for 1998–2007. The year refers to the reference period in the survey. For example, the reference period in the 2000/01 survey is the accounting year ending March 31, 2001. As three-quarters of the reference period fell in 2000, the survey is labeled 2000. SME = small to medium enterprise. Employment in establishments increased 1.2 TABLE 3.7 Distribution of service ï¬?rms in India, by location and size, percent a year, from 10.7 million in 2001 to 2006 11.4 million in 2006; employment in own- (percent) account enterprises increased 0.4 percent a Own-account year, from 15.9 to 16.2 million. Data after Firm characteristic Establishments enterprises Total 2006 are not available. Percentage of all firms 14.7 85.3 100.0 Although industrial surveys with the Location same coverage of the manufacturing and Urban 63.0 36.1 40.0 service sectors are not available for other Rural 37.0 63.9 60.0 South Asian countries, other enterprise Number of employees and ï¬? rm surveys suggest a similar picture 1–49 99.7 100.0 99.96 elsewhere. Rural nonfarm enterprise sur- 50–99 0.17 0.00 0.03 veys conducted in Bangladesh, Pakistan, 100 + 0.04 0.00 0.02 and Sri Lanka show large numbers of rural Source: Authors, based on data from National Sample Survey service sector surveys. 110 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.19 Average wage, value added, and capital per service sector worker in India, by ï¬?rm size and type, 2006 (percentage of average wages/value added/ capital per workers in establishments with 20–49 employees) 100 a. Mean wage per worker b. Median wage per worker 100 90 90 80 80 70 70 58 percent 60 percent 60 50 47 50 51 40 33 40 32 30 30 20 20 20 10 3 10 2 0 0 1–4 5–9 10–19 1–4 5–9 10–19 ï¬?rm size ï¬?rm size c. Mean value added per worker d. Median value added per worker 100 100 90 90 80 80 70 70 66 68 58 percent percent 60 60 50 43 50 40 34 36 40 35 30 30 20 15 20 10 10 0 0 1–4 5–9 10–19 1–4 5–9 10–19 ï¬?rm size ï¬?rm size e. Mean capital per worker f. Median capital per worker 100 100 90 90 80 80 70 68 63 70 56 percent percent 60 52 60 51 50 50 48 40 40 30 30 20 19 20 17 10 10 0 0 1–4 5–9 10–19 1–4 5–9 10–19 ï¬?rm size ï¬?rm size establishments own-account enterprises Source: Authors, based on data from the National Sample Survey service sector surveys. Note: Establishments with 20–49 employees = 100 percent. FIGURE 3.20 Share of service sector employment in India, by ï¬?rm size and type, 2001 and 2006 100 90 80 70 60 percent 50 40 30 20 10 0 0+ 0+ 9 9 9 9 9 9 4 4 9 9 9 9 5– –1 –4 5– –1 –4 1– 1– 19 19 –9 –9 20 20 10 20 10 20 0– 0– 50 50 10 10 ï¬?rm size ï¬?rm size 2001 2006 own-account enterprise establishments Source: Authors, based on data from the National Sample Survey service sector survey. Note: Employment in 2006 excludes the ï¬?nancial intermediation sector, because the sector was not covered in 2001. A PROFILE OF SOUTH ASIA AT WORK 111 nonfarm enterprises and large numbers of or salaried workers are more educated than workers employed in the sector. The major- self-employed or casual workers. Almost all ity of rural nonfarm enterprises are micro have some education, and a signiï¬?cant pro- family enterprises with few hired workers, portion have secondary education or above. low use of capital, and lower productivity The self-employed are better educated than than urban formal ï¬? rms. casual workers. The relative educational lev- els of employment types are found in both the Who holds better jobs? urban and rural nonfarm sectors and across all South Asian countries (ï¬?gure 3.21). These This section examines the factors that affect data suggest that education is an important access to better jobs in South Asia. These determinant of access to better jobs and that factors include education, gender, and caste lack of education is likely to be a barrier to and ethnicity. mobility to moving into better jobs. (For an analysis of the effect of education on work- Education ers’ ability to access and move into better jobs, see chapter 5.) Education is an important determinant of access to better jobs. It is strongly correlated with the sector of employment in South Gender Asian labor markets. Industry and services rely on workers who are considerably more Labor force participation rates for women are educated than workers in agriculture (table low in South Asia, particularly in urban areas. 3.8). The average educational attainment is Among women who do participate in the labor particularly high in service sector industries force, the majority live in rural areas, where such as utilities, ï¬? nance, and public admin- 70 percent of South Asians of working age istration. Although manufacturing and live. Most rural female workers are employed construction jobs do not usually require as in traditional agricultural activities. The more much education as these service industries, modern and better-paid rural nonfarm sector their workforces have much more education employs mostly men. In all countries except than workers in agriculture. Maldives, a higher percentage of male rural Within industry and services, education workers are in the nonfarm sector. Although is correlated with better jobs. Regular wage more female rural workers are entering the TABLE 3.8 Average years of education in South Asian countries, by sector of employment Public administration Manufacturing Other services Construction Agriculture Commerce Transport Financial Utilities Mining Country/year Afghanistan, 2008 0.9 1.9 1.4 — 2.0 3.5 3.5 — 9.8 4.2 Bhutan, 2007 0.9 3.8 3.5 8.2 4.2 4.2 4.3 9.9 5.7 7.2 India, 2010 4.1 6.1 6.4 10.1 4.4 7.7 6.9 12.9 11.2 9.7 Maldives, 2004 3.5 4.4 3.5 6.8 4.7 7.2 7.8 9.2 7.8 8.2 Nepal, 2008 3.1 2.9 4.5 6.3 3.5 6.2 5.4 9.6 9.8 9.6 Pakistan, 2008 2.7 5.5 5.7 10.4 4.0 6.8 5.6 11.7 9.9 9.9 Sri Lanka, 2008 7.1 7.6 9.4 10.9 8.6 9.3 9.9 12.2 11.6 10.9 South Asia unweighted mean 3.2 4.6 4.9 8.8 4.5 6.4 6.2 10.9 9.4 8.5 Source: Authors, based on data from national labor force and household surveys. Note: Bangladesh is not included, because its surveys do not permit comparable calculations. — = Not available. percent se percent percent 112 se lf- se 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 lf- 0 10 20 30 40 50 60 70 80 90 100 lf- ag em agri em ri c c em agri ag plo ultu pl cul ag ploy ultu ric ym ra oy tu 17 13 m ra 47 ric m ral ul en l 0 1 a en l 2 ul en tu t tu t no ra no ca gric t no ra nf lc nf su ult nf lc as ar as a a u 22 14 ar 23 or m r lab ual or rm r l lab ral 1 or m r lab ual sa eg or 0 sa eg or sa eg or lar ul lar ul lar ul ied ar ied ar ied ar wo wag wo wag 66 59 se se 88 se wo wag lf- rk e lf- rk e 37 16 rk e em n er 28 lf- em n er em n er pl on pl on pl on o fa oy fa m rm 26 o fa 37 no yme rm e 65 5 2 no yme rm nf nt 6 ar ca no nt nf nt ar m ur su n m ur ba ca b al far ur ca 21 or an r lab m 34 ba s or n r la sual sa eg or 54 2 4 or n r lab ual sa eg bo c. Bhutan, 2007 lar ul 1 e. Nepal, 2008 sa eg or lar ul r ied ar a. Afghanistan, 2008 lar ul ied ar ied ar wo wag 87 wo wag 78 se rk e se rk e 90 wo wag lf- lf- er 62 26 se rk e em er em 42 lf- er pl u em pl u oy rb oy rb m a 52 pl u 45 oy rb ur men an 7 ur en n 15 68 m an in South Asia, by employment type and country ba t ba t 12 ur ent nc nc ba as as 37 nc lab ual 48 as lab ual 4 or 9 61 lab ual or Source: Authors, based on data from national labor force and household surveys. 5 percentage with some education or MORE AND BETTER JOBS IN SOUTH ASIA percent se percent se percent se lf- lf- 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 lf- em agri em agri em agri c ag plo cult ag plo cult ag ploy ultu ric ym ura ric ym ura 63 54 37 ric m ral ul e l ul e l 9 1 3 ul en tu nt tu nt tu t no ra no ra no ra nf lc nf lc nf lc ar as ar as ar a 52 34 or m r lab ual 23 or m r lab ual 0 2 or m r la sual sa eg or 0 sa eg or sa eg bo lar ul lar ul lar ul r ied ar ied ar ied ar wo wag wo wag 92 90 se rk e se rk e 82 wo wag se lf- 40 18 rk e lf- 29 lf- em n er em n er em n er pl on pl on pl on o fa 76 o fa 62 oy fa no ym rm no yme rm 67 2 13 no me rm nf ent nf nt 8 ar ar nf nt ar ur m ur m m ba c ba c percentage with secondary education or above ur 62 51 b ca or n r l asu or n r l asu 0 sa eg ab al sa eg ab al 4 51 d. India, 2010 or an r la sual lar ul or 2 lar ul or sa eg bo f. Pakistan, 2009 ied ar b. Bangladesh, 2005 lar ul r ied ar ied ar wo wag wo wag 93 87 wo wag se rk e se rk e lf- er 85 lf- er 50 28 se rk e em em 42 lf- er pl u pl u em oy rb oy rb m a m a 85 70 pl u oy rb ur en n 7 ur en n 30 m an 75 ba t ba t ur ent nc 18 nc ba as as 69 48 nc lab ual lab ual as 5 1 or or 59 lab ual 6 or FIGURE 3.21 Percentage of workers with some education and percentage of workers with secondary education or above A PROFILE OF SOUTH ASIA AT WORK 113 sector, the gender gap has widened over time, and Nepal, but less than 30 percent of rural increasing from 10 percentage points in 1983 women in India and less than 20 percent in to 16 percentage points in 2010 in India, for Nepal worked in the nonfarm sector at the example (ï¬?gure 3.22). end of the period. A cohort analysis for India, Nepal, and A multinomial logit analysis that looks Pakistan provides further evidence that at the determinants of labor participation women are generally stuck in agriculture and occupational choice shows that among across all countries in the region (ï¬?gure 3.23). workers with less than upper-secondary edu- This analysis tracks participation in the rural cation, men are more likely than women to nonfarm sector for the same age groups of hold regular wage jobs within the rural non- male and female workers for 8–10 years farm and urban sectors. In contrast, female beginning in 1999/2000. workers with upper-secondary education Several findings emerge. First, for every are more likely than male workers to hold age cohort, women were less likely than men regular wage jobs (see chapter 5). Given the to work in the rural nonfarm sector. Second, small percentage of women who participate younger men were more likely than older in the labor force, the share of working-age men to leave agriculture for the nonfarm sec- women in better jobs is still much lower than tor and stay there. In all countries, the cohort the share of working-age men in better jobs, of men that was 15–24 in 1999/2000 had the even at the highest levels of education. largest share of nonfarm labor by 2009/10 The lower level of educational attainment (40 –55 percent), followed by the 25–34 among women is one reason why they are cohort. Among men in cohorts other than less likely than men to be in better jobs. More 15–24, and in some cases 25–34, the share than half of the female workforce in ï¬?ve of of workers in the nonfarm sector by the end the eight countries in the region has no educa- of the period declined or remained the same. tion. Occupational segregation and lower pay Third, younger women were more likely than for the same jobs and qualiï¬?cations are other older women to exit agriculture in India likely factors. FIGURE 3.22 Percentage of rural workers in the rural nonfarm sector in South Asia, by gender and country 100 90 90 80 73 70 70 66 64 61 60 58 54 percent 50 49 49 42 40 39 40 35 36 31 30 30 28 28 25 23 22 21 20 16 19 16 15 15 12 11 13 8 10 10 4 0 2003 2007 2002 2005 1998 2004 1999 2008 2000 2009 2000 2008 1983 1994 2000 2004 2010 Bhutan Bangladesh Maldives Nepal Pakistan Sri Lanka India male female Source: Authors, based on data from national labor force and household surveys. 114 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3.23 Percentage of rural workers in the rural nonfarm sector in India, Nepal, and Pakistan, by gender and age cohort, 1999–2009/10 a. Male workers in India b. Female workers in India 60 60 50 50 40 40 percent percent 30 30 20 20 10 10 0 0 2000 2005 2010 2000 2005 2010 c. Male workers in Nepal 60 d. Female workers in Nepal 60 50 50 40 40 percent percent 30 30 20 20 10 10 0 0 1999 2008 1999 2008 e. Male workers in Pakistan f. Female workers in Pakistan 60 60 50 50 40 40 percent percent 30 30 20 20 10 10 0 0 2000 2009 2000 2009 age 15–24 age 25–34 age 35–44 age 45–54 Source: Authors, based on data from national labor force and household surveys. Note: The age cohort refers to the age of the cohort in the earliest period, that is, the line of the cohort of age 15–24 in India refers to the same cohort group: the cohort that was 15–24 years old in 2000. A PROFILE OF SOUTH ASIA AT WORK 115 Women earn wages that are 20–40 per- It is important to determine how much of cent lower than those of men, even after con- the observed wage gap is a result of produc- trolling for education levels, type of employ- tive characteristics, such as education. The ment, and other individual and household technique proposed by Blinder and Oaxaca characteristics associated with productivity. (1973) has been used to decompose the wage This result holds true for all countries and gap into the part explained by differences in almost all types of wage employment. The levels of observable characteristics (includ- lower wages could be a result of unobserved ing education) and the part not explained or unmeasured characteristics in the wage by these differences. The unexplained por- regressions; they are also consistent with tion reflects in part differences in returns to some other type of discrimination in the mar- skills.17 Figure 3.24 plots the decomposition ket that is not associated with productivity for rural and urban workers for all survey differentials. If discrimination in the labor years. It indicates that part of the wage gap is market is a major factor behind the observed the result of differences in observable produc- wage gap, a policy focus on increasing the tive characteristics but that most of the wage skill levels of women will not be sufï¬?cient. gap is unexplained. FIGURE 3.24 Decomposition of wage gap between male and female workers in South Asia, by country a. Afghanistan, 2008 b. Bangladesh, 2002 c. Bangladesh, 2005 d. India, 2000 60 40 percent 20 0 –20 rural urban rural urban rural urban rural urban e. India, 2005 f. India, 2008 g. Nepal, 1999 h. Nepal, 2008 60 40 percent 20 0 –20 rural urban rural urban rural urban rural urban i. Pakistan, 2000 j. Pakistan, 2009 k. Sri Lanka, 2000 l. Sri Lanka, 2008 60 40 percent 20 0 –20 rural urban rural urban rural urban rural urban explained unexplained Source: Authors, based on data from national labor force and household surveys. 116 MORE AND BETTER JOBS IN SOUTH ASIA Caste and ethnicity Women are more likely to work, particularly in rural areas, if they belong to an ethnic Caste and ethnicity have long been impor- minority. Among women who work, women tant factors shaping employment in South from ethnic minority groups are more likely Asia. Female labor force participation and to be in casual or self-employment in urban employment rates in India, Nepal, and Sri areas and agriculture in rural areas. Lanka vary by ethnic/caste groups. Ethnic minority groups earn lower wages The multinomial logit analysis shows that than other groups, after controlling for edu- workers from ethnic minority groups, par- cation levels and employment type. The dif- ticularly women, are less likely to be in bet- ference, though smaller in size than gender ter jobs even after controlling for education differentials, appears signiï¬?cant in India but and other factors. Among men in rural areas, not in Nepal.18 members of ethnic minority groups are more The results of the decomposition of the likely to be involved in agriculture. In urban wage differential show that in India and areas, men from ethnic minority groups are Nepal, much of the wage gap is explained more likely to be involved in casual work and by differences in observable characteristics, self-employment. including education (ï¬?gure 3.25). In contrast, Similar trends appear for women, and in Sri Lanka most of the wage differential is the differences appear larger in magnitude. unexplained. FIGURE 3.25 Decomposition of wage gaps between nonethnic minority and ethnic minority workers in India, Nepal, and Sri Lanka a. India, 1983 b. India, 2000 c. India, 2005 d. India, 2008 40 30 percent 20 10 0 rural urban rural urban rural urban rural urban e. Nepal, 1999 f. Nepal, 2008 g. Sri Lanka, 2000 h. Sri Lanka, 2008 40 30 percent 20 10 0 rural urban rural urban rural urban rural urban explained unexplained Source: Authors, based on data from national labor force and household surveys. A PROFILE OF SOUTH ASIA AT WORK 117 Annex 3A Deï¬?nitions and criteria used in proï¬?le of South Asia at work Deï¬?nitions of employment and Criteria used to deï¬?ne formal and unemployment informal workers Table 3A.1 provides the speciï¬?c deï¬? nitions Table 3A.2 provides information on how the for employment and unemployment used national surveys are used to estimate infor- for each country based on the national mal employment for each country. surveys. TABLE 3A.1 Deï¬?nition of employment and unemployment used based on national surveys Definition of unemployment (for people Country Year Definition of employment who are not employed) Afghanistan 2007/08 In past 30 days, did any one of the following: In past 30 days, did any of the following: • Worked for any organization or any individual • Attempted to find a job or start a business • Performed any agricultural work, even without pay, • Found a job that has not yet started on land owned, rented, or used by household • Awaited recall from an employer • Performed any nonagricultural work, on own account, in a business enterprise • Performed occasional paid work, such as helping someone in his or her business • Has a permanent or long-term job from which he or she will be temporarily absent Bangladesh 2002/03 In past seven days, did either of the following: • Was available for work in past seven days and tried to • Worked for at least one hour on any day for pay or find a job during past two months profit, family gain, or final use or consumption • Even if did not work in past seven days, has a job, business, enterprise, or attachment to a job such as a business, farm, or shop 2005/06 Same as Bangladesh 2002 In past seven days, did any of the following: • Was prepared for and tried to find a job • Was waiting for reappointment • Was waiting to start work after appointment 2009 In past seven days, did either of the following: • In past seven days, looked for job • Engaged in any economic activity for at least one hour as paid worker, for family gain or profit, or own use or consumption • Was temporarily absent from work Bhutan 2003 In past seven days, did any of the following: In past seven days, did any of the following: • Any farming, fishing, hunting, or gathering of fruits • Actively looked for a job or tried to start a new • Worked for money or have profitable business business • Performed unpaid work on enterprise or farm of • Was waiting for a job to start friends or relatives • Was waiting for employer’s reply 2007 Same as 2003 In past seven days, did any of the following: • Actively looked for a job or tried to start a new business • Did not actively look for a job because was waiting for employer’s reply • Did not actively look for a job because was waiting for employers recall (continues next page) 118 MORE AND BETTER JOBS IN SOUTH ASIA TABLE 3A.1 Deï¬?nition of employment and unemployment used based on national surveys (continued) Definition of unemployment (for people Country Year Definition of employment who are not employed) India 1999/2000, In past seven days, did any of the following: • In past seven days, sought work and/or was available 2004/05, • Worked or helped in household enterprise for work 2007/08, and • Worked as regular or casual wage labor 2009/10 • Was involved in public or other works Maldives 1998 and In past 30 days, did either of the following: • Not currently working because unable to find a 2004 • Was involved in economic activity most of the time suitable job, but if hired, would be available to work • Engaged in any activity that generated income Nepal 1998/99 and In past 7 days, did either of the following: In past 7 days, did either of the following: 2007/08 • Was involved at least one week in any economic • Was available and looked for work in past 30 days activities • Awaited reply on earlier inquiries or start of • Did not work but have a permanent job he or she prearranged job or business will return to Pakistan 1999/2000 In past 7 days, did any of the following: In past 7 days, did any of the following: and 2008/09 • Performed any work for pay, profit, or family gain • Looked for job during for one hour on any day • Did not work but will take a job within a month • Did not work but has a job or enterprise to return to • Was temporarily laid off and awaiting recall • Worked/helped in family business or family farm for family gain Sri Lanka 2000 In past 7 days, did either of the following: In past 7 days, was available or looking for work. • Engaged in economic activity during past 7 days • Did not engage in economic activity but had an economic activity to return to 2008 In past 7 days, did either of the following: • Already obtained a job or made arrangements to • Engaged in economic activity start self-employment activity • Had an economic activity to return to OR • Took some steps to find a job or start self- employment activity within the last four weeks OR • Awaited results of interview for work Source: Authors’ compilation. TABLE 3A.2 Deï¬?nition of formal and informal workers used based on national surveys Country Year Informal workers Formal workers Afghanistan 2007/08 • Daily laborers • Workers in public sector • Workers who receive pensions Bangladesh 2002/03 • Casual workers and day laborers • Workers in public sector • Domestic workers • Workers in private or nongovernmental formal • Workers in private informal sector sector 2005/06 • Domestic workers and apprentices • Workers in public sector • Workers working in personal households • Workers in private formal sector • Casual workers and day laborers Bhutan 2003 a • Workers in public sector 2007 • Casual paid employees • Workers in public sector India 1999/2000 • Workers in firms that do not keep written accounts • Workers in public and semipublic sector • Workers in firms with fewer than 10 employees • Workers covered under a provident fund • Workers in firms that keep written accounts • Workers in firms with more than 10 employees (continues next page) A PROFILE OF SOUTH ASIA AT WORK 119 TABLE 3A.2 Deï¬?nition of formal and informal workers used based on national surveys (continued) Country Year Informal workers Formal workers 2004/05 and • Casual workers • Workers in public sector 2009/10 • Workers without written contracts • Workers receiving social security benefits • Workers working in firms with fewer than • Workers with written job contracts 10 employees • Workers in firms with more than 10 employees • Workers not receiving any social security benefits Maldives 2004 a • Workers in public sector Nepal 1998/99 • Workers in unregistered organizations • Workers in public sector • Workers in firms with fewer than 10 employees • Workers in firms with 10 or more employees 2007/08 • Workers who are not eligible for paid leave and do • Workers in public sector not receive social security contributions • Workers for whom employers pay social security • Workers in establishments with fewer than contributions 10 employees • Workers who are eligible for paid leave • Workers working in establishments with 10 or more employees Pakistan 1999/2000 • Casual and day laborers • Workers in public sector and 2008/09 • Workers in enterprises that do not keep written • Workers in enterprises that keep written accounts accounts • Workers in establishments with 10 or more • Workers in establishments with fewer than employees 10 employees Sri Lanka 2000 • Daily laborers • Workers in public sector 2008 • Casual workers • Workers in public sector • Workers whose employers do not contribute to • Workers for whom employer contributes to pension pension scheme or provident fund and do not scheme or provident fund or provides paid leave provide paid annual leave • Workers in registered enterprise • Workers in unregistered enterprise • Workers in enterprise that maintains formal written • Workers in enterprise that does not maintain formal accounts written accounts • Workers in firms with 10 or more employees • Workers in firms with fewer than 10 employees Source: Authors’ compilation. Note: All family enterprise workers are in the informal sector. All self-employed workers with less than senior-secondary education are considered informal, and all self-employed individuals with higher education are considered formal. All workers involved in agriculture are informal. Some workers could not be identiï¬?ed as formal or informal. In all countries except Maldives, the unclassiï¬?ed portion was 10 percent or less. In Maldives 32 percent could not be classiï¬?ed. For the subsample of workers that could not be identiï¬?ed as formal or informal in each country, their status was allocated according to the formal/informal split for the employment that could be classiï¬?ed. The assumption that the distribution of formality/informality is the same for classiï¬?ed and unclassiï¬?ed workers may lead to some overestimation of informality, as the classiï¬?ed portion includes some block assignments, such as agriculture to the informal sector. Because the unidentiï¬?ed portion is small in most countries, this should not be a major issue, except perhaps in Maldives. a = No additional criteria speciï¬?c to the survey. Annex 3B Regional employment patterns Figure 3B.1 shows how employment rates difference in employment rates between the vary within each of the South Asian coun- regions with the highest and lowest rates is tries. The overall picture is one of relatively only about 10 percentage points. Although moderate differences in regional rates. the disparities in employment rates may not Unweighted coefficients of variations are be large, there are major differences in the 5–10 percent for all countries except Afghan- nature and quality of employment across istan, where regional differences are much regions (for example, the share of employ- larger. For all countries except Afghanistan, ment that is in the formal wage-earning the range is fairly narrow: in most cases, the sector). 120 MORE AND BETTER JOBS IN SOUTH ASIA FIGURE 3B.1 Regional variations in employment rate in South Asia, by country 90 West Central Mid -Western 85 84 (5.9) 80 75 Central Eastern 70 69 (9.0) percent 67 (16.7) 65 North-Central Sylhet Western 60 58 (4.7) Western 58 (6.7) 55 Chittagong 61 (5.7) Central Punjab Western Central Lagging 50 50 (7.8) 48 (8.8) 45 NWFP South 40 Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka 2008 2005 2007 2005 2004 2008 2009 2008 Source: Authors, based on data from national labor force and household surveys. Note: Each diamond represents the employment rate in a region. The regions with the highest and lowest employment rates are named. Each box represents the average employment rate of all regions in the country; ï¬?gures in parentheses are coefficients of variation, calculated as the standard deviation of the regional employment rates normalized by the mean of these rates. Notes (identiï¬?ed as activity pursued by the individ- ual for a relatively minor time, but not less 1. The labor force participation rate and the than 30 days during the past year) fell sharply employment rate in India in 2009/10 were compared to previous years. Usual principal 56.6 percent and 54.8 percent, respectively. employment status (deï¬?ned on the basis of the There is a sharp decline from 2004/05, lead- activity pursued for the longest—major—time ing to only a very small increase (in absolute during the reference year) only experienced a terms) in the labor force and total employ- modest decline. The reasons for the decline ment between 2004/05 and 2009/10. The in labor force participation and employment decline has been driven by a sharp fall in rates remain to be analyzed fully, but possible female labor force participation and employ- reasons include (a) younger workers, espe- ment rates (male rates have also declined, cially women, staying longer in education; but far less sharply), which led to a signiï¬?- (b) rising family incomes reducing the need for cant decrease in rural female labor force and women to work; and (c) shocks in the agricul- employment in absolute terms. The decline in ture sector. It is important to note that over female participation was observed in all age the longer period between 1983 and 2010 groups including school age cohorts (15–19 labor force participation and employment years old and 20–24 years old) but also older rates have increased and decreased without cohorts. Rural female employment declined a consistent trend, although the employment particularly in agriculture. Self-employment and participation rates in 2010 are lower declined both in number and as a share of than in any previous period. Using current total employment, while casual wage labor weekly status, the labor force participation accounted for nearly 80 percent of net addi- rate was 59.7 percent, 62.0 percent, 60.4 per- tional employment in those sectors that cent, 62.2 percent, and 56.6 percent in 1983, expanded employment between 2004/05 and 1994, 2000, 2005, and 2010, respectively, 2009/10. Usual subsidiary employment rates and the employment rate was 57.5 percent, A PROFILE OF SOUTH ASIA AT WORK 121 60.1 percent, 58.2 percent, 59.8 percent, and 8. Poverty rates for India are based on the 54.8 percent in 1983, 1994, 2000, 2005, and ofï¬?cial poverty lines prevailing until 2010. 2010, respectively. It may thus be premature Using the new ofï¬?cial poverty lines for to see the past ï¬?ve years as necessarily being 2004/05 (revised in 2011) would increase evidence of a longer-term trend. the poverty rates in rural areas, causing 2. In the absence of better questions in the sur- rural poverty rates to exceed urban work- vey in Maldives, the criterion “available for ers for the same employment type. The hier- work, conditional on getting a jobâ€? was archy in terms of employment type would use to deï¬?ne unemployment. This criterion, remain the same, however: agricultural which is much broader than that used in casual labor would have the highest pov- other countries, yielded a much higher unem- erty rate, followed by rural nonfarm casual ployment rate. labor and urban casual labor. Agricultural 3. This measure, based on current weekly sta- and rural nonfarm self-employment would tus questions, is an adjusted version of the have higher poverty rates than urban self- deï¬?nition of underemployment used by the employment. The lowest poverty rates National Commission for Enterprises in the would be among urban regular wage or Unorganised Sector (2009). salaried workers. 4. Using household survey data, Srinivasan 9. Results are based on regressions for each (2010) reviews labor market indicators over country-survey year. Log hourly wages were the longer term for Bangladesh (from regressed on individual and household char- 1983/84), India (from 1972/73), and Pakistan acteristics, including educational attainment (from 1990). He ï¬?nds that female participa- and sector and type of employment. tion rates in Bangladesh rose from 8 percent 10. All casual labor is classiï¬?ed as informal in to 30 percent between 1983/84 and 2005/06, this book. Some regular wage work is also female participation rates in India were posi- classiï¬?ed as informal (see annex table 3A.2 tive in rural but not urban areas, and female for details of deï¬?nition used in each country’s participation in Pakistan has been essentially labor force survey). stagnant since 1990. 11. The India poverty assessment (World Bank 5. The informality shares in ï¬?gure 3.11 are simi- 2011a, p. 133) also highlights the transient lar to those in Chen and Doane (2008). nature of rural nonfarm casual labor work: 6. For example, more than a third of total “Casual workers tend not to have year-round employment in the Republic of Korea is infor- employment and make ends meet by working mal (OECD 2007). at several jobs, often combining agricultural 7. Based on the deï¬?nition of informality used in and nonfarm activities. In 2004/05, more this book, there is no clear pattern of increas- than half (55 percent) of casual nonfarm ing or decreasing informality. In Bangladesh workers report that they are without work (2002–05), India (2000–05), and Pakistan for one or more months in the year compared (2000–09), there have been no signiï¬?cant to 8 percent of salaried workers or 12 percent changes. In Maldives (1998–2004) and Sri of self-employed. Fourteen percent of casual Lanka (2000–08), informal employment nonfarm workers report that they were seek- declined signiï¬?cantly. In Bhutan (2003–07) ing or available for additional employment and Nepal (1998–2008), it increased sig- even when working.â€? niï¬?cantly. However, because the concepts 12. Estimates from labor force surveys of infor- used to measure informality changed from mality in the nonagricultural sector indicate the earlier to the later year in the countries that Afghanistan, Pakistan, and Nepal have in which changes were observed, it is not a higher proportion of informal nonagricul- clear how much of the observed increase tural employment (about 80 percent) than or decrease was real. In Bangladesh and Bangladesh (74 percent), India (72 percent), Pakistan, where there was no change in the and Sri Lanka (68 percent). way of measuring informality, no change 13. “Organizedâ€? (formal) manufacturing ï¬?rms in the degree of informality was observed. are registered ï¬?rms, typically with more than Other studies, using alternative deï¬?nitions, 10 workers. The most commonly used deï¬?ni- have found increasing informality in India tion of the unorganized (informal) sector flows (NCEUS 2009). from the Indian Factories Act (1948), which 122 MORE AND BETTER JOBS IN SOUTH ASIA requires that an enterprise register with the large, are ofï¬?cially in the informal sector. The state government if it employs 10 or more service sector survey covers service ï¬?rms in workers and uses power or employs 20 or rural and urban locations for a large number more workers and does not use power. Units of service sectors but excludes trade. Dehejia that do not come under the purview of this law and Panagariya (2010) estimate that the sur- constitute the unorganized sector. Businesses vey covers about half of output and employ- that are organized are required to comply ment in the service sector. with health, safety, and welfare requirements. 17. This interpretation becomes problematic if Firms are required to contribute toward insur- the observable characteristics are improp- ance against sickness, disability, and maternity erly measured or omitted or unobservable and to deposit linked provident funds or pen- characteristics, such as motivation, could be sion schemes. The state monitors compliance influencing wages. Such characteristics will through a system of inspections. Unorganized bias the results if the errors and omissions ï¬?rms largely fall outside this system. The unor- are systematically different for men and ganized manufacturing sector comprises own- women. account manufacturing enterprises, directory 18. Ethnic minorities are scheduled caste and tribes manufacturing enterprises, and nondirectory in India and Dalits and Janajatis in Nepal. manufacturing enterprises subcategories. Own- account manufacturing enterprises are informal ï¬?rms that operate without any hired worker References employed on a regular basis. Nondirectory manufacturing establishments are informal ADB (Asian Development Bank). 2009. Enter- ï¬?rms that employ fewer than six workers prises in Asia: Fostering Dynamism in SMEs. (with at least one hired worker); directory Manila. manufacturing establishments employ six or Chen, M., and D. Doane. 2008. “Informality in more workers (with at least one hired worker). South Asia: A Review.â€? Background working Ninety-nine percent of unorganized manufac- paper for the Swedish International Develop- turing ï¬?rms have fewer than 10 workers. ment Cooperation Agency (Sida), Sweden. 14. Among ï¬?rms in the formal sector, real wages Dehejia, R., and A. Panagariya. 2010. “Services are slightly higher in rural ï¬?rms than urban Growth in India: A Look inside the Black ï¬?rms of the same size class. Among ï¬?rms with Box.â€? Working Paper 2010-4, Columbia Uni- one to nine employees in the informal sector versity, Program on Indian Economic Policies, (the vast majority of the unorganized sector), New York. wages in urban and rural unorganized ï¬?rms Government of India. 2006. Employment and are very similar. Unemployment Situation in India 2004– 05 15. Assessing the contribution of each factor is National Sample Survey 61st Round. Minis- challenging, given the difï¬?culty in matching try of Statistics and Programme Implemen- information across workers and ï¬?rms. For tation, National Sample Survey Ofï¬? ce, New example, the industrial surveys for India, Delhi. which provide ï¬?rm productivity and wage ILO (International Labour Ofï¬?ce). 2002. Women data, do not survey the education and skills and Men in the Informal Economy: A Statis- levels of workers. See ADB (2009) for a dis- tical Picture. Geneva: International Labour cussion of ï¬?rm-size wage differentials. Ofï¬?ce. 16. The National Sample Survey service sector ———. 2010. Accelerating Action against Child survey classiï¬?es ï¬?rms as either establishments Labour: Global Report under the Follow-Up or own-account enterprises. Own-account to the ILO Declaration on Fundamental Prin- enterprises do not employ hired workers on a ciples and Rights at Work. Geneva: Interna- regular basis; establishments do. Own-account tional Labor Ofï¬?ce. enterprises belong to the informal sector; estab- ———. 2011. Key Indicators of the Labour Mar- lishments are a mix of informal and formal ket. Geneva: International Labour Ofï¬?ce. sector ï¬?rms. As services ï¬?rms are not required Islamic Republic of Afghanistan, and World Bank. to register under the Factories Act unless they 2010. Poverty Status in Afghanistan: A Proï¬?le are engaged in manufacturing, most private Based on the National Risk and Vulnerabil- sector service enterprises, whether small or ity Assessment (NRVA) 2007/08. Ministry A PROFILE OF SOUTH ASIA AT WORK 123 of Economy, Kabul, and Economic Policy and Volume 26. Bangladesh Institute of Develop- Poverty Sector, Washington, DC. ment Studies, Dhaka. Loayza, M., and T. Wada. 2011. “Informality in Srinivasan, T. N. 2010. “The Utilization of Labor South Asia.â€? Background study conducted for in South Asia.â€? Background study conducted this book. for this book. Mukherjee, A., and X. Zhang. 2007. “Rural UN (United Nations). 2010. World Population Industrialisation in China and India: Role of Prospects: The 2010 Revision. New York: Policies and Institutions.â€? World Development United Nations. 35 (10): 1621–34. Wang, X., J., Huang, and L . Zhang, 2011. National Commission for Enterprises in the “Creating the Entrepreneur Farmers Needed Unorganised Sectors. 2009. The Challenge of Yesterday, Today and Tomorrow.â€? Paper Employment in India: An Informal Economy presented at the International Fund for Agri- Perspective. New Delhi. cultural Development (IFAD) conference on OECD (Organisation for Economic Co-operation new directions for smallholder agriculture, and Development). 2007. “Labour Markets in Rome. Brazil, China, India and Russia and Recent World Bank. 2010. India’s Employment Chal- Labour Market Developments and Prospects lenge: Creating Jobs, Helping Workers. India: in OECD Countries.â€? OECD Employment Oxford University Press. Outlook. Paris. ———. 2011a. Perspectives on Poverty in India: Rahman, R. I. 2008. “A Review of Open Unem- Stylized Facts from Survey Data. Washington, ployment and Underemployment Concepts: DC: World Bank. Measurement, Methods, and Applicationâ€? ———. 2011b. World Development Indicators. (in Bengali with title “Unmukto Bekaratta O Washington, DC: World Bank. Angshik Bekaratta Dharanar Porjalochana, ———. 2012. World Development Report: Gen- Parimap Paddhati O Proiogâ€?). Bangladesh der Equality and Development. Washington, Unnayan Shamikkha, Golden Jubilee Number. DC: World Bank. CHAPTER 4 What Is Preventing Firms from Creating More and Better Jobs? Questions and Findings Questions informal ï¬?rms. However, without informa- tion on whether inadequate ï¬?nance reflects • What are the business environment con- the lack of bankable projects, one cannot straints affecting ï¬?rms in South Asia? infer directly that access to ï¬?nance is a bind- • Do these constraints vary by sector and ï¬?rm ing constraint. Other evidence suggests that characteristics? access to ï¬?nance may be an issue for micro • What are the main policy priorities for over- and small ï¬?rms in some countries. coming the identiï¬?ed business constraints? Findings Policy focus areas The chapter identiï¬?es constraints facing ï¬? rms • The severity of the electricity constraint for all and four areas of policy focus. types of ï¬?rms reflects the large gap in the region between demand for and supply of power. Clos- ing the gap requires a substantial increase in Constraints investment, which in turn requires that power • The most binding constraints facing all types sector reforms be sustained and deepened. of urban formal firms (where the highest- • South Asian ï¬?rms face high levels of corrup- productivity and highest-paid jobs are) are tion in a range of interactions with public ofï¬?- electricity, corruption, and political instability. cials, particularly for utilities and tax inspec- Job-creating ï¬?rms are more severely affected tions. Simplifying processes in, for example, than other ï¬?rms by virtually the entire range tax administration and reducing unnecessary of constraints. Labor regulations rank high in interaction with local officials could be an India, Nepal, and Sri Lanka. effective way of tackling corruption, as it has • The rural nonfarm sector is an important route the additional beneï¬?t of reducing the cost of out of low-paid agricultural work. For ï¬?rms in red tape on ï¬?rms. this sector, electricity and political instability • Key policy options for improving access to also rank high on the list of constraints. Trans- finance for micro and small firms include port (poor road quality and inaccessibility) strengthening the institutional environment ranks much higher as a constraint for rural (secured transactions registry, credit informa- ï¬?rms, as it affects access to larger markets. tion); creating a conducive environment for • Increasing the productivity of informal ï¬?rms downscaling commercial banks and upscal- is fundamental, as they employ the majority ing microï¬?nance institutions; and providing of workers. Electricity is the top constraint ï