93858 v2 The Unfulfilled Promise Of Oil and Growth Poverty, Inclusion and Welfare in Iraq 2007–2012 The Unfulfilled Promise Of Oil and Growth Poverty, Inclusion and Welfare in Iraq 2007–2012 © 2014 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpreta- tions, and conclusions expressed in this work 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 judgment 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 this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publi- cations, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover photo courtesy of Thaiview Cover design by the Word Express, Inc. Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Chapter 1 Conflict, Growth and Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Repression, Invasion, and Sanctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Reconstruction and Insurgency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Current Policy Framework. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Legacy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Violence, Growth and Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Framework of Analysis and Report Outline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Chapter 2 Poverty, Shared Prosperity and Subjective Well-Being in Iraq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 GDP and Consumption Growth in Iraq: 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Measuring Poverty in Iraq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Poverty and Shared Prosperity in Iraq: 2007–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Spatially uneven poverty reduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Unequal consumption growth and shared prosperity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Who are the Poor? A Profile of Poverty, 2007 to 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Correlates of consumption expenditure and poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 Poverty across Space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Poverty beyond Consumption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Chapter 3 Poverty in Human Capital. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Multidimensional Poverty in Human Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Human Opportunity Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Lagging behind in Health and Nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 iv THE UNFULFILLED PROMISE OF OIL AND GROWTH Nutritional outcomes within Iraq. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Early motherhood and child nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Evidence from the Multiple Indicator Cluster surveys. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Stagnation in education among men; some improvement for young women. . . . . . . . . . . . . . . . . . . . . 63 Widespread access to basic services, but little improvement in quality. . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Chapter 4 Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare . . . . . . . . . . . . . . . . . . . . . . . 77 Violence and Insecurity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Human Development and Access to Basic Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Labor Market Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Consumption, Income and Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Inequality in consumption growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Incomes and transfers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Determinants of Consumption and Poverty across Iraq. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Explaining rural – urban welfare disparities within divisions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Explaining welfare differences between divisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Chapter 5 Understanding the Drivers of Poverty Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Potential Sources of Consumption Growth and Poverty Reduction in Iraq. . . . . . . . . . . . . . . . . . . . . 114 Decomposing Poverty Reduction: 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Chapter 6 The Growth-Employment Nexus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 The Links between Growth and Employment in Iraq. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 New Jobs in the Economy: Public Versus Private Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Youth participation in the workforce. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Female participation in the labor force. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Male participation in the labor force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 What Explains Differences in Earnings?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Chapter 7 The Labor Market for the Poor: The Rural-Urban Divide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Where do the Poor Work?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 Differences in economic activity are only one of many dimensions of differences between urban and rural households. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 For any given characteristic, rural households face higher poverty rates than urban households . . . . . . . 151 Public Sector Employment, Public Transfers and Urban Poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Explaining the differences in welfare between public and private sector. . . . . . . . . . . . . . . . . . . . . . . . . . 157 Rural employment and Agriculture in Iraq, 2007 to 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Agricultural jobs: Evolution across time and space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Rural poverty and non-farm diversification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 What explains the lower poverty rates of diversified households? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 Contents v Chapter 8 Transfers, Safety Nets, and Poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Transfers, Safety Nets and Poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Who receives transfers? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 The Public Distribution System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Simulation of the welfare impact of the PDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 Chapter 9 Policy Implications: Learning from the Past to Build a Better Future. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Policy Imperatives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Growth, redistribution and poverty reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Strengthening the Relationship between Growth and Welfare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 Economic growth, oil management and diversification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 Private sector led job creation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 Enhancing Inclusion and Social Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Intergovernmental fiscal relations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Redressing the human capital deficit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 Public works and spatially targeted programs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 Figures Figure 1: Civilian Deaths in Iraq, 2003 to Present . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Figure 2: GDP Per Capita (current US$), 1960–2012, and Timeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 3: Oil Rents (% of GDP), 1960–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 4: Log Mean Night Time Lights (district Level), 2003, 2005, 2007 and 2009 . . . . . . . . . . . . . . . . 14 Log Average Night Time Lights, 2003–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 5:  Average Civilian Casualties (Iraq Body Count): 2003 to 2010. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 6:   verage Civilian Casualties (Iraq Body Count) By Region: 2003–2010. . . . . . . . . . . . . . . . . . . . 15 Figure 7: A Figure 8: Economic Growth and Welfare: Main Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17  nnual GDP Growth in Iraq, 2007–2012 (%) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 9: A Figure 10: C umulative Distribution – Welfare Aggregate, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 11: P robability Density – Welfare Aggregate, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 12: Poverty Headcount, Gap and Severity—Iraq: 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Figure 13: Regional Poverty Headcount Rates, 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 14: Growth Incidence Curves – National, Rural-Urban, Divisional . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure 15: The Contribution of Growth and Redistribution to Poverty Reduction. . . . . . . . . . . . . . . . . . . 28 Figure 16: Shared Prosperity and the Distribution of the Bottom 40 Percent. . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 17: Population Share by Size of Household . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 18: Poverty Headcount Rates by Household Size, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 19: Poverty Headcount Rates for Female and Male Headed Households. . . . . . . . . . . . . . . . . . . . . 30 Figure 20: Poverty Headcount Rates by Education of the Head of Household . . . . . . . . . . . . . . . . . . . . . . 30 Figure 21: Share of the Poor, by Education of the Head of Household. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Figure 22: Poverty Headcount Rates by Labor Force Status of the Head of Household* . . . . . . . . . . . . . 31 Figure 23: Share of the Poor, by Labor Force Status of the Head of Household . . . . . . . . . . . . . . . . . . . . . . 31 vi THE UNFULFILLED PROMISE OF OIL AND GROWTH Poverty Headcount Rates by Employment Sector of the Head of Household. . . . . . . . . . . . . 32 Figure 24:  Share of the Poor by Employment Sector of Head of Household . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 25:  Figure 26: Public Sector Employment as a Share of Employment in Each Sector, 2007–2012. . . . . . . . 33 Figure 27: Self-Employment in the Private Sector as a Share of All Employment in the Private Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Figure 28: Division Headcount Rates (Percent) and Changes (Percentage Point), 2007–2012 . . . . . . . . 37 Figure 29: Distribution of the Poor by Divisions, 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Figure 30: Poverty in Basra and the Rest of the South, 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Figure 31: Comparing Minimum Monthly Income Needs Per Capita with Monthly per Capita Expenditures, by Consumption Deciles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 32: Dissatisfaction Across Different Dimensions, Iraq, Urban and Rural . . . . . . . . . . . . . . . . . . . . . 40 Figure 33: Rates of Dissatisfaction, Selected Dimensions, by Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 34: Headcount Rates, Different Measures of Wellbeing, Iraq, Urban and Rural, 2012 . . . . . . . . . 41 Figure 35: Headcount Rates, Different Measures of Wellbeing, Divisions, 2012 . . . . . . . . . . . . . . . . . . . . . 41 Figure 36: P overty Measures Across Consumption Deciles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Figure 37: Share of Consumption Poor Who are Poor According to other Measures. . . . . . . . . . . . . . . . .42 Figure 38: Share of Subjective, Satisfaction and Minimum Income Poor Who are Consumption Poor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Figure 39: Relative Weights of Different Dimensions of Deprivation in Determining Subjective Poverty and Dissatisfaction with Life. . . . . . . . . . . . . . . . . . . . . . . . 44 Figure 40: Weighted Incidence of Different Dimensions of Deprivation in Determining Subjective Poverty and Dissatisfaction with Life. . . . . . . . . . . . . . . . . . . . . . . . 44 Figure 41: Comparison Among Poverty Measures: Consumption Poverty and MPI in Human Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 42: Dimensions of Deprivation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 43: Overlapping Dimensions of Deprivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 44: Contribution of Each Dimension to Iraq´s MPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 45: Contribution of Circumstances to Inequality Index—Selected Opportunities. . . . . . . . . . . . . 54 Figure 46: Stunting and Wasting (0–60 Months): Iraq Versus other Countries . . . . . . . . . . . . . . . . . . . . . . 55 Figure 47: Stunting Rates and Governorate Consumption Expenditure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Figure 48: Wasting Rates and Governorate Consumption Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Figure 49: U nderweight Rates for Children and Governorate Consumption Expenditure. . . . . . . . . . . . 58 Figure 50: Stunting, Expenditures and Caloric Intake and the PDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Figure 51: Ever Married Females by Division: Percentage by Age (12–18) . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Figure 52: Ever Married Females Under 18 Across Time: Percentage by Division . . . . . . . . . . . . . . . . . . . 60 Figure 53: Ever Married Females Under 18: Percentage by Governorates . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Figure 54: Ever Given Birth to a Child: Share of Married Females Under 18. . . . . . . . . . . . . . . . . . . . . . . . 61 Figure 55: Stunting Rates and Early Marriage by Governorate, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Figure 56: Educational Level by Generation Over Time: Female and Male Population in Iraq . . . . . . . 64 Figure 57: Educational Level By Generation of Male Population and Divisions: 2012 . . . . . . . . . . . . . . . 64 Figure 58: Educational Level By Generation of Female Population and Divisions: 2012 . . . . . . . . . . . . . 64 Figure 59: G ross Enrollment by Educational Level in Iraq: Population Aged 6 and Above. . . . . . . . . . . 65 Figure 60: Net Enrollment by Educational Level in Iraq: Population Aged 6 and Above. . . . . . . . . . . . . 65 Figure 61: Gross Enrollment by Educational Level and Division: Population Aged 6 and Above. . . . . 66 Figure 62: Net Enrollment by Educational Level and Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Figure 63: A ge Distribution of THose Enrolled in Primary School . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Contents vii O verage Students Enrolled in Primary School by Age Group. . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 64:  67 Gender Gaps in Enrollment in Iraq: Female Relative to Male Figure 65:  Gross Enrollment—2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Figure 66: Enrollment by Age: Male. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Figure 67: Enrollment by Age: Female. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Percentage of Students Missing at Least 4 School Days in the Previous Figure 68:  Month by Gender and Age, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Reasons for Dropping Out of School: Male Ages 7–25. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 69:  69 Reasons for Dropping Out of School: Male Above the Age of 25. . . . . . . . . . . . . . . . . . . . . . . . Figure 70:  69 Reasons for Dropping Out of School: Female Ages 7–25. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 71:  70 Reasons for Dropping Out of School: Figure 72:  Female Above the Age of 25. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Distance in Minutes to the Nearest Establishment – 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 73:  71 Figure 74:  Access to Publicly Provided Water (Share of Households Connected to the Public Grid) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71  ercentage of Population THat have Public Network as the Main Source Figure 75: P of Water and Consider it Insufficient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Percentage of Households Declaring Public Supply of Water is Interrupted Figure 76:  More THan Once a Week. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Access to Electricity Grid (Publicly Provided) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 77:  72 Share of Households by Average Hours of Electricity from the Public Figure 78:  Network Per Day in the Past Week . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Access to Garbage Collection (Collected by Municipality or Figure 79:  Put in Designated Container) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Access to Sewage (Public Network Disposal System). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 80:  73 Percentage of Households Declaring Being Adversely Affected by Figure 81:  Sanitation Related Issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Access to Services by Consumption Quintile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 82:  74 Housing Quality Perception – 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 83:  74 Figure 84: Number of Bedrooms Per Capita. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Perceptions of Housing Quality by Quintile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 85:  75 Per Capita Number of Bedrooms by Quintile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 86:  75 Understanding Spatial Disparities in Poverty in Iraq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 87:  79 Figure 88: Number of Refugees, Selected Countries (>50,000 Persons) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Figure 89: Internally Displaced People, Selected Countries (>50,000 Persons). . . . . . . . . . . . . . . . . . . . . . . 81 Estimates of Migration and Displacement, IHSES 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 90:  81 Governorate Level Estimates of Displacement as a Share of THose Who Figure 91:  Lived Elsewhere for More THan 6 Months . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Spells of Displacement, 1970–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 92:  82 Figure 93: Fraction of Displaced in Each Division, by Year of Moving, 1970–2012 . . . . . . . . . . . . . . . . . . 82 Figure 94: Estimates of Displacement, by Year of Moving, 1970–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Figure 95: ‘Persons of Concern’, 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Figure 96:  ‘Persons of Concern’ Concentrated in Baghdad, Nineveh and Diyala . . . . . . . . . . . . . . . . . . . . . . 84 Figure 97: Civilian Casualties by Governorate (%), 2007 to 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Casualties by Division (%): 2007–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 98:  84 Share of Divisions in Total Civilian Deaths and Number of Deaths, 2007–2012 . . . . . . . . . . Figure 99:  85 viii THE UNFULFILLED PROMISE OF OIL AND GROWTH Figure 100: Adult Female Literacy, Iraq and other Countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Figure 101: Adult Male Literacy, Iraq and other Countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Figure 102: Current Educational Outcomes by Age in 2012 and Year of Birth, Ages 18–62. . . . . . . . . . 87 Figure 103: Educational Attainment by Age Cohort in 2012, Kurdistan Relative to Iraq. . . . . . . . . . . . . 88 Figure 104: Educational Attainment by Age Cohort in 2012, South Relative to Iraq. . . . . . . . . . . . . . . . . 88 Figure 105: Educational Attainment by Age Cohort in 2012, Baghdad Relative to Iraq. . . . . . . . . . . . . . 89 Figure 106: Educational Attainment by Age Cohort in 2012, North Relative to Iraq . . . . . . . . . . . . . . . . 89 Figure 107: Educational Attainment by Age Cohort in 2012, Centre Relative to Iraq. . . . . . . . . . . . . . . . 90 Figure 108: Male Life Expectancy, Iraq Versus MENA, 1960–2011 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Mortality Rate, Adult (per 1,000 Adults), Iraq, 1980–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Figure 109:  Population by Age Group, 2005–2010. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Figure 110:  Figure 111: Population Share, 2010. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92  mployment to Population Ratio (ages 15+), 2011. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Figure 112: E Figure 113: Labor Force Participation, by Age and Gender, 1990–2011. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Figure 114: Labor Market Outcomes, Share of Working Age Population (ages 15–64), 2007–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Figure 115: Unemployment, Underemployment and Joblessness*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Figure 116: Employment Status, Young Men (aged 15–29) in the Working Age, by Division . . . . . . . . 95 Figure 117: Percentage Change in Non-Employment, Employment, Working Age Population and Employment to Working Age Ratio between 2007 and 2012, by Division, for Men (15–64) Based on One Year Recall . . . . . . . . . . . . . . . . . . . 96 Figure 118: Changes in the Number of Jobs for Men by Sector of Employment between 2007 and 2012 in Kurdistan, the North and the Centre (One Year Recall). . . . . . . . . . . . . 97 Figure 119: Changes in the Number of Jobs for Men by Sector of Employment between 2007 and 2012 in Baghdad and the South (One Year Recall) . . . . . . . . . . . . . . . . . . . . . . . . . 98 Figure 120: Growth Incidence Curves—National, Divisional. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Figure 121: Changes in Per Capita Labor Income between 2007 and 2012 (‘000s of Iraqi Dinar) for the Bottom 3 Deciles of the Consumption Distribution, National and by Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Figure 122: M arginal Effects: Household Size and Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Figure 123: M arginal Effects: Household Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Figure 124: M arginal Effects: Education of the Head of Household. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Figure 125: Marginal Effects: Sector of Employment of the Head of Household. . . . . . . . . . . . . . . . . . . 106 Figure 126: Marginal Effects: Governorate Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Figure 127: Differences Within Regions – Urban Versus Rural Areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Figure 128: Welfare Differences between Divisions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Figure 129: Explaining Welfare Differences between Regions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Figure 130: Demographic Changes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Figure 131: Demographic Characteristics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Figure 132: Share of Adults Per Household. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Figure 133: Share of Working Adults Per Household. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Figure 134: Labor Income by Region – Growth Incidence Curves. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Figure 135: Public and Private Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Figure 136: Change in the Consumption-to-Income Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Changes in the Consumption-to-Income Ratio by Division. . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Figure 137:  Different Components of Consumption Per Capita. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Figure 138:  Contents ix Contributions to Poverty Reduction—Total Iraq. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Figure 139:  Growth in Labor Income is the Main Contributor Factor to Figure 140:  Changes in Moderate Poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Figure 141: Decomposition of Significant Poverty Changes by Divisions . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Figure 142: Public and Private Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Figure 143: Public Sector Employment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Figure 144: Public Jobs—Education Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129  rivate and Public Jobs: Retirement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Figure 145: P  rivate and Public Jobs: Gaps on Hours of Work Per Day. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Figure 146: P Figure 147: Youth Participation and Employment Rates, by Education Level . . . . . . . . . . . . . . . . . . . . . . 130 Figure 148: Type and Sector of Work, by Education Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Figure 149: Sectors of Employment, Youth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Figure 150: Youth Participation Rate by Public Employment Rate in the Governorate of Birth . . . . . 133 Marginal Effects of Being Active—Youth Population by Education Level. . . . . . . . . . . . . . 134 Figure 151:  Figure 152: Participation and Employment Rates, Adult Women. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Figure 153: Type and Sector of Work, Adult Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Figure 154: Economic Sector of Employment, Adult Women. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Figure 155: Participation and Employment Rates, Adult Men. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Figure 156: Type and Sector of Work, Adult Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Figure 157: Sector of Economic Activity, Adult Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Figure 158: Business Constraints. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Figure 159: Evolution of Returns to Education by Gender. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Figure 160: Evolution of Returns to Education by Gender—Public Sector. . . . . . . . . . . . . . . . . . . . . . . . . 142 Decomposing Differences in Earnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Figure 161:  Decomposing Gender Differences in Earnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Figure 162:  Evolution of Employment by Economic Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Figure 163:  Share of Employment in the Private Sector, Urban and Rural, 2007 to 2012. . . . . . . . . . . . 149 Figure 164:  Education Levels, Working Age Population, Urban and Rural Households, Figure 165:  2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Household Size and Composition Across Urban and Rural Areas, 2007 and 2012 . . . . . . 150 Figure 166:  Figure 167: Dependency and Employment Across Urban and Rural Areas, 2007 and 2012 . . . . . . . . . 150 Poverty Headcount Rates by Household Size, Urban and Rural Households, Figure 168:  2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Figure 169: Poverty Headcount Rates by Education, Urban and Rural Households, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Poverty Headcount Rates by Employment Status, Urban and Rural Figure 170:  Households, 2007 and 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Share of the Urban Poor, by Sector of Employment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Figure 171:  Headcount Rates, by Types of Urban Households. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Figure 172:  Share of the Poor, by Types of Urban Households. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Figure 173:  Headcount Rates by Employment of the Head (urban Private Sector), 2007–2012. . . . . . 156 Figure 174:  Decomposing Differences in Headcount Rates between Private and Figure 175:  Public Sector Employed Heads of Household . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Figure 176: Jobs in Agriculture for Men and Women, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Figure 177: Changes in Number of Jobs in Agriculture and Total Jobs, Men and Women, by Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 x THE UNFULFILLED PROMISE OF OIL AND GROWTH Share of Agricultural Jobs in Each Division, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Figure 178:   hare of Agriculture as a Source of Employment for Men and Women Figure 179: S within Each Division, 2007 and 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Share of Different Types of Households in Each Division, 2007 and 2012. . . . . . . . . . . . . . 163 Figure 180:  Trends in Headcount Rates of Different Types of Households, by Division, Figure 181:  2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 Figure 182: Educational Attainment of Different Types of Rural Households, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166  ecomposing Differences in Headcount Rates between Agriculture Figure 183: D and Diversified Households — 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168  DS and Pensions as a Share of GDP, 2007 to 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Figure 184: P  ources of Non-Labor Income, 2007 and 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 Figure 185: S  hare of Poor and Non-Poor Individuals Receiving Public and Figure 186: S Private Transfers, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174  hare of Calories from the PDS, by Consumption Quintile, 2007 and 2012. . . . . . . . . . . . 177 Figure 187: S PDS Expenditures, Food and Non-Food Expenditures Figure 188:  by Consumption Quintile, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177  ncidence of Shocks by Poverty Status, 2007. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Figure 189: I  ncidence of Shocks by Poverty Status, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Figure 190: I Prevalence of Shocks in Urban and Rural Areas, 2007. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Figure 191:   revalence of Shocks in Urban and Rural Areas, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 Figure 192: P  ousehold Perceptions of the Effects of Different Shocks, 2012. . . . . . . . . . . . . . . . . . . . . . . 181 Figure 193: H Figure 194: Household Perceptions of the Effects of Each Type of Shock, 2012. . . . . . . . . . . . . . . . . . . . 181 Figure 195: Main Response to Each Shock, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182  wn-Price Elasticities of Ration Items by Quintile of Per Capita Figure 196: O Consumption and Area, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 Figure 197:  O wn-Price Elasticities of Free Market Goods by Quintile of Per Capita Consumption and Area, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 O wn-Price Elasticities of Ration Items by Quintile of Per Capita Figure 198:  Consumption and Area, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Figure 199: Isopoverty Curves. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 Tables Table 1:  Mean Per Capita Consumption Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Poverty Lines (ID Per Person Per Month). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 2:  22 Table 3: Overall Poverty (Regional Poverty Lines). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23  ensitivity of Headcount Poverty Rate with Respect to the Choice of Poverty Line . . . . . . . . . Table 4: S 23 Population and Poor Population, by Region, 2007–2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 5:  25 Table 6:  Inequality in Per Capita Expenditure Distribution by Urban and Rural Areas . . . . . . . . . . . . . . . 26 Table 7: Shared Prosperity – Annual Consumption Growth of the Bottom 40 . . . . . . . . . . . . . . . . . . . . . . 28 Survey Based Estimates of Population by Division* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 8:  37 Average Minimum Per Capita Monthly Income (‘000s Iraqi Dinar) Table 9:  Required to Meet Basic Needs, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Table 10: Average Minimum Per Capita Income Required (monthly, ‘000s Iraqi Dinar), Governorates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Contents xi Consumption Poor Versus MPI Poor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Table 11:  Coverage Rates and Human Opportunity Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Table 12:  Table 13: Nutritional Indicators in the MENA Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Table 14: Nutrition Indicators in Iraq According to Different Data Sources. . . . . . . . . . . . . . . . . . . . . . . . . 56  utritional Outcomes by Wealth Quintiles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Table 15: N Table 16: Nutritional Outcomes by Age Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Table 17: Stunting and Wasting by Employment of Head. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57  verage Hours of Power Supply by Consumption Quintile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Table 18: A Completed Level of Education, Share of Iraqis Aged 15–64: Division Table 19:  Relative to National, 2007–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Anthropometrics, Children Aged 0–60 Months (share) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Table 20:  Employment Status, Men in the Working Age (aged 15–64), by Division, Table 21:  7 Day Recall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Table 22: Employment Status, Women in the Working Age (aged 15–64), by Division, 7 Day Recall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Table 23: Share of Labor Income in Total Income, and Shares of Major Sources of Non-Labor Incomes and Transfers in Total Non-Labor Income, Overall and Lowest Consumption Decile, 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Table 24: Changes in Labor Income and Main Non-Labor Income Sources between 2007 and 2012 for the Bottom 3 Consumption Deciles, National and Divisions Absolute Change (‘000s of Iraqi Dinars), 2012 Relative to 2007 . . . . . . . . . . . . . . . . . . . . . 102 Table 25: Significant Correlates of Per Capita Expenditure in 2007 and 2012 in Each Division . . . . . 103 Table 26: Probability of Being Poor: Marginal Effects of Characteristics. . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Table 27: Probability of Being in the Bottom 40 Percent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Table 28: Growth, Employment and Labor Income Relations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Returns to Education by Gender and Sector—2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Table 29:  Labor Market Outcomes for the Urban and Rural Poor, 2007 to 2012 . . . . . . . . . . . . . . . . . . . 149 Table 30:  Labor and Non-Labor Income, Households with Heads Employed Table 31:  in the Public Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Labor and Non-Labor Income, Households with Non-Employed Heads . . . . . . . . . . . . . . . . 156 Table 32:  Labor and Non-Labor Income, Households with Heads Employed in the Table 33:  Private Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Table 34: Share of Different Types of Households in Urban and Rural Areas, 2007 and 2012 . . . . . . . 163 Poverty Rates of Different Types of Households in Rural Areas, 2007 and 2012 . . . . . . . . . . 164 Table 35:  Table 36: Sectors of Employment Individual in Rural Areas, 2007 and 2012 . . . . . . . . . . . . . . . . . . . . . . . 167 Changes in Non-Agricultural Employment and Labor Earnings for Non-Poor Rural Table 37:  Diversified Households, 2007 and 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 Main Sources of Income (Share of Total), National Average and Bottom Table 38:  4 Deciles, 2007 and 2012. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Table 39: Sources of Non-Labor Income Across Iraq, Overall and Bottom Decile, 2012 . . . . . . . . . . . . 173 Table 40: Prevalence of Shocks, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180  udget Shares of Ration Items by Decile of Per-Capita Consumption, 2012 . . . . . . . . . . . . . 183 Table 41: B Table 42: Expenditure Elasticities by Quintile of Per-Capita Consumption and Area, 2012. . . . . . . . . 186 Expenditure Elasticities by Quintile of Per-Capita Consumption and Area, 2012. . . . . . . . . 188 Table 43:  O wn Price Elasticities by Subsidy Products, Egypt 1997 and Iraq 2012 . . . . . . . . . . . . . . . . . . 188 Table 44:  Average Change in Total Expenditure by Quintile in Urban Areas. . . . . . . . . . . . . . . . . . . . . . . 189 Table 45:  xii THE UNFULFILLED PROMISE OF OIL AND GROWTH Table 46: Average Change in Total Expenditure by Quintile in Kurdistan Region. . . . . . . . . . . . . . . . . . 190 Table 47: Trading Off Consumption Growth and Redistribution for Poverty Reduction. . . . . . . . . . . . 194 Boxes Box 1: A Stylized Example of the HOI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Box 2: Long Term Implications of Malnutrition in Early Childhood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Box 3: Early Motherhood and Child Outcomes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Box 4: Doing Business in Iraq –Main Messages from the Investment Climate Assessment 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Box 5: PDS Ration Items and the Valuation Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 Maps Map 1: Five Divisions of Iraq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Percentage of Land Used for Agriculture Per Square Kilometer. . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Map 2:  Share of Agricultural Land Use of Total District Land Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Map 3:  NDVI Changes in Greenness Over Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 Map 4:  Foreword I raq is a nation with a promising endowment— analyses informed the first ever Poverty Reduction in terms of its once rich human capital and an- Strategy of Iraq. cient culture, agricultural potential and natural resources—whose potential has been repeatedly This was not a one-off effort. Very soon, the team thwarted because of years of violence and insecurity. started planning the next survey, learning from the This difficult legacy has posed almost insurmount- experience of the first round, and invested in tech- able development challenges that will take sustained nical improvements to the survey instrument, field effort and years to overcome. The period covered procedures and survey scope, which has culminated by this report, between 2007 and 2012, while one in the second IHSES in 2012. This was accompa- of relative stability, still represents just the beginning nied by sustained capacity building and technical of Iraq’s path towards reconstruction and recovery. support from the World Bank; and the poverty mea- surement methodology evolved with this process. A key priority of the Government of Iraq since In a record three months after survey completion, 2005–06 has been to fill the huge knowledge gap poverty estimates were produced and agreed upon in terms of a deeper understanding of the state of with the PRSHC using best practice methodology; the economy and of a range of socio-economic in- and Iraq released its second poverty estimates in dicators of welfare with the objective of building June 2013. a strong evidence base for effective policy making. This effort has been led since 2006 by the Poverty The rich analyses presented in this report, that goes Reduction Strategy High Committee (PRSHC) of well beyond counting the poor, to give an incisive Iraq, a high-level technical working group whose understanding of the multi-layered development members represent a range of stakeholders includ- challenges faced by the nation is a testament to the ing parliamentarians, line ministries and academ- commitment of the Government of Iraq, the staff ics. The first Iraq Household and Socio-Economic of the Central Statistics Office and Kurdistan Re- Survey (IHSES) in 2006–07 represented the most gion Statistics Office. It will form the basis for a new comprehensive survey of its kind in Iraq at the time, strategy for Iraq’s development and to ensure broad- and benchmarked the welfare of the Iraqi popula- based welfare improvements for the population. tion for the first time in decades. The IHSES team demonstrated incredible courage and commitment Throughout this process, the World Bank has been at a time of insecurity to implement this nationwide a steady partner to the government, in providing exercise. The Government of Iraq drew on interna- technical assistance and capacity building, and this tional expertise to ensure the best possible standards is a collaboration we wish to continue. The new for survey design and implementation; and the poverty map for Iraq, developed with the assistance xiv THE UNFULFILLED PROMISE OF OIL AND GROWTH of the World Bank, which will provide estimates of Dr. Mehdi Al-Alaak poverty at the community level, will provide yet Chief of Staff, Prime Minister’s Office another tool to design effective poverty reduction Government of Iraq policies in Iraq. We look forward to a long and pro- (Formerly Deputy Minister of Planning and ductive collaboration with the World Bank to help Head of the Central Statistical Office, Iraq) build a more inclusive and prosperous Iraq. Acknowledgements T his report would not have been possible GMFDR), Sibel Kulaksic (Senior Economist, without the commendable efforts and re- GFMDR), Roy van der Weide (Economist, DEC- markable dedication of the staff of the Iraq PI), Brian Blankespoor (Environmental Specialist, Central Statistics Office (CSO) and the Kurdistan DECCT), Garam Dexter (ET Consultant, GFM- Region Statistics Office (KRSO). Under the able DR), Racha Ramadan (Consultant and Assistant guidance and strategic vision of Dr. Mehdi Al-Alak Professor, Cairo University) and Tara Vishwanath (Chair of the Poverty Reduction Strategy High (Lead Economist, GPVDR). A special thanks to Committee and Deputy Minister of Planning), Ms. Maurice Saade for his valuable inputs and insights. Najla Ali Murad (Executive General Manager of The team also gratefully acknowledges the sup- the Poverty Reduction Strategy), and Mr. Serwan port of Dr. Mohammed Bakir (Senior Consultant), Mohamed (Director, KRSO), the technical staff and Dr. Basil al-Hussaini (Senior Consultant) and Dr. fieldworkers involved in the two Iraq Household Giovanni Vecchi (University of Rome, Tor Verga- and Socio-Economic Surveys placed utmost em- ta). Thanks also to Ghada Haddad and Salam Al- phasis on the timeliness and accuracy of data. The maroof for their assistance at various times. breadth and depth of the analysis in this report is a testament to their commitment, including in diffi- The team acknowledges all the participants in two cult and insecure field conditions. Mr. Qusay Raoof in-country consultants to discuss the findings of the Abdulfatah (Living Conditions Statistics Director, report and various previous technical discussions CSO) deserves special acknowledgement for going on measurement and methodology. The insights well above and beyond the call of duty and work- and recommendations from these discussions have ing tirelessly with the team. We thank Mr. Ayad- helped make the analysis nuanced and deep. We Jawad and the Data Analysis Team for all their hard would like to especially thank our four peer review- work and support throughout the process. A special ers, who undertook the onerous task of providing thanks also to Dr. Jamal Ameen for his passion, cu- feedback on this large piece of work—Onno Ruhl riosity and insights. We are honored to have worked (Country Director, SACIN), Ambar Narayan (Lead with all of them, and we owe them a deep debt of Economist, GPVDR), Gabriela Inchauste (Senior gratitude. Thank you. Economist, GPVDR) and Somik Lall (Lead Ur- ban Economist, GURDR). Last and by no means This report was led by Nandini Krishnan (Senior the least, we are grateful to Marie-Helene Brick- Economist, GPVDR) with a core team compris- nell (formerly Special Representative, MNCIQ), ing Sergio Olivieri (Economist, GPVDR) and Ly- Bernard Funck (Practice Manager, GFMDR) and cia Lima (Consultant, World Bank). The extended FeridBelhaj (Country Director, MNC02) for their team comprises Kevin Carey (Lead Economist, support and encouragement. Conflict, Growth and Development 1 T he overarching context for this report, which profound marginalization. The space for market allo- focuses on the period from 2007 to 2012, is cations was constricted since so much was taking place Iraq’s status as a resource-rich, fragile and through administrative fiat, giving privileged access conflict-affected state. The country has been a nexus to state enterprises and certain cadres, especially dur- of conflict and fragility since the early 1980s, and has ing the sanctions era. Since the formal sector had to be experienced multiple types of conflict: insurgency, in- run through the government, many market activities ternational war, sectarian strife, persistent terrorism, shifted to the informal sector, with consequent effects regional fragmentation, and spillovers from conflict on productivity, investment horizon, and job quality. in other countries. What should have been a promising At the same time, the catastrophic decline of agricul- endowment of natural resources (land, oil, and gas) ture posed challenges for food supply and employment and human capital did not provide the foundation of the sector’s workforce. While some problems predated for poverty reduction and shared prosperity, because the conflict era (irrigation water supply, salinity, and the realization of potential was confounded by war desertification), throughout the 1990s and beyond, the and repression. Yet, externally imposed regime change sector was afflicted by lack of access to critical inputs in 2003 and a tortuous process of reestablishing elected and low productivity. State-owned enterprises (SOEs) civilian government had in principle set the stage for remained dominant and the state had no incentive inclusive growth in Iraq. to restructure SOEs; even aside from their value as a means of economic control, public sector jobs were one The country’s trajectory since the 1970s has been a se- of the few reliable instruments that the post-2003 gov- ries of divergences from regional and global trends, be- ernment had at its disposal. Far from being impetus ginning first with the Iran-Iraq war of the 1980s and to reform, growing oil revenues became an enabler of followed by the invasion of Kuwait, which was even the status quo. more damaging, triggering comprehensive sanctions, which didn’t end until the 2003 US-led invasion. Spatial divergence became pronounced. Whereas the During each phase, any hope of catch-up was thwarted Kurdish region had been the most victimized before by further events, and policy reform was off the table. the 1990s, with de facto independence from Baghdad, and more flexibility—and access to cash—basic needs 1991–92 forms a major rupture in Iraq’s develop- could be met more effectively. On the other hand, the ment trajectory. With the government consolidating southern provinces saw a double negative impact: the around the goal of regime survival and the private destruction of wars compounded by a failed rebellion. sector unable to function, this was the beginning of In principle, these trends should have been moderated large-scale detachment from government and the post-2003 since the Iraqi government could direct re- formal sector for many Iraqis—laying the roots of sources to areas of greatest deprivation. In practice, the 2 THE UNFULFILLED PROMISE OF OIL AND GROWTH significant variation in the quality of service delivery violence stalls social development, is associated with suggests the equalizing role of the central government large losses in productivity, as well as the destruc- was not effective. While oil continued in the enclave tion of assets and infrastructure. The WDR notes development model, other sources of economic activity that conflict is typically recurring, and it can take a such as pilgrimages, trade with Iran, and the Basra generation to recover from the setbacks to poverty port, came into play as drivers of regional growth. reduction associated with conflict. At the same time, jobs and access to services, especially security and A cross-cutting element of Iraq’s legacy is severe eco- justice, are critical to peacebuilding, catching up on nomic and social fragmentation. Civil conflict creat- deferred progress, and reducing the susceptibility to ed substantial internal displacement corresponding to conflict relapses. ethnic and sectarian divisions. Baghdad became a city of internal boundaries formed by blast walls, check- Iraq provides an apt case study for this framework. points, and no-go areas. For the country as a whole, The country has been a nexus of conflict and fra- a basic prerequisite of economic development—inter- gility since the early 1980s, and has experienced nal integration—was lost. Increasing sectarian vio- multiple types of conflict: insurgency, international lence in 2013 and militancy and armed insurgency war, sectarian strife, persistent terrorism, regional in 2014 have further fragmented the nation, leaving fragmentation, and spillovers from conflict in other swathes of the country outside government control. countries. What should have been a promising en- Achievement of the twin goals of ending extreme pov- dowment of natural resources (land, oil, and gas) erty and boosting shared prosperity is bound to be an and human capital did not provide the foundation uphill struggle in this context. Quantitative evidence for poverty reduction and shared prosperity, be- suggests that while development efforts by the govern- cause the realization of potential was confounded by ment and other actors can mitigate the negative role of war and repression. Yet externally imposed regime violence, the required effort is substantial. Thus, on a change in 2003 and a tortuous process of reestab- macro-economic level, and absent the establishment of lishing elected civilian government had in principle peace and security in the nation, maintaining growth set the stage for inclusive growth in Iraq. can in itself be a challenge in Iraq, and without these two pre-conditions in place, poverty reduction and This poverty and inclusion assessment provides the broad-based prosperity will be difficult to sustain. first in-depth analysis of Iraq’s economic and social development spanning the period 2007 to 2012, since the end of sectarian war of 2006–07, accom- Introduction panied by recovery in the oil sector, a massive scal- ing up of oil revenues, and extensive efforts by the The 2011 World Development Report (WDR 2011) government to meet the high expectations of the highlighted the interaction of conflict and fragility people. However, Iraq’s relationship with violence with poverty and exclusion. Not only do countries is not yet at an end: in 2013, sectarian violence affected by major violence experience slower pov- led to an increase in civilian mortality that has not erty reduction, the effects are cumulative. “On av- been seen since the 2007 spike in violence (Figure erage, a country experiencing major violence over 1). Moreover, a violent insurgency in 2014 has left the entire period (1981–2005) had a poverty rate parts of the country outside of government control, 21 percentage points higher than a country that saw leading to massive internal displacement in parts of no violence”. While the cost in terms of human life the country yet again. This report therefore covers falls disproportionately on men, children and wom- a period of relative stability in Iraq, following the en bear significant indirect costs; not to mention end of the sectarian violence of 2007, and ending in the consequences of large-scale displacement. Be- 2012, prior to the militancy and insurgency in the yond these human costs, large-scale and protracted northern governorates of the summer of 2014. Conflict, Growth and Development 3 FIGURE 1: Civilian Deaths in Iraq, 2003 to Present 4500 4000 3500 3000 2500 2000 1500 1000 500 0 1-Jan-03 1-Jun-03 1-Nov-03 1-Apr-04 1-Sep-04 1-Feb-05 1-July-05 1-Dec-05 1-May-06 1-Oct-06 1-Mar-07 1-Aug-07 1-Jan-08 1-Jun-08 1-Nov-08 1-Apr-09 1-Sep-09 1-Feb-10 1-July-10 1-Dec-10 1-May-11 1-Oct-11 1-Mar-12 1-Aug-12 1-Jan-13 1-Jun-13 1-Nov-13 1-Apr-14 Source: Iraq Body Count, https://www.iraqbodycount.org/database/ (Retrieved: July 29, 2014). Note: Casualties since October 2013 are estimates that have not yet been verified. This introductory chapter contains six further sec- human), leaving the country with significant chal- tions. Section 2 discusses the decades-long experi- lenges of delayed development today. ence of internal strife, war, and sanctions which left indelible marks on Iraq. Section 3 looks at the more In the 1970s, Iraq was still in the Arab mainstream. immediate context for understanding poverty and Although in policy terms this was associated with inclusion in Iraq in terms of the aftermath of the Nasserite statism and authoritarianism, the state was 2003 invasion and later civil war. Section 4 moves playing a clear developmental role, meaning provi- specifically to the economic policy context within sion of infrastructure and scaling up of access to so- which the current socio-economic outcomes are un- cial services. At the same time, the surge in oil prices folding. Section 5 summarizes in terms of the legacy in the 1970s generated substantial revenues for the of all these factors, which provides a framework for state and increased per capita income (Figure 2); thinking about the findings of subsequent chapters. there were ambitious plans for further increases in Section 6 presents quantitative estimates of the rela- production in the oil sector. Unlike the Gulf states, tionship between conflict, development efforts and Iraq had reasonable capacity for absorbing oil rev- growth in Iraq for the post-2003 period, and sec- enues at home (as opposed to overseas saving), and tion 7 concludes with the framework of analysis and prospects for oil-financed development were plau- outline of the report. sible, even allowing for some of the pitfalls of this state-centered mode of development. Repression, Invasion, and Sanctions Nevertheless, warning signs were already present by the late 1970s. Even by the standards of the region, Although the long sweep of Iraqi history is fairly the rule of Saddam was taking on an increasingly well known, it is useful to think about the coun- autocratic character, including purges within the try’s trajectory since the 1970s in terms of a series Baath party apparatus and brutal suppression of dis- of divergences from regional and global trends, dur- sent, notably among the Shia and Kurdish popula- ing which time any hope of catch-up was thwarted tions. The 1979 Islamic Revolution in Iran altered by further events. The result of these divergences the geopolitics of the region, so whereas the Sad- was to negate the beneficial effects of Iraq’s prom- dam regime was once seen as part of a modernizing ising endowment (agriculture, hydrocarbons, and republican/presidential group of states in contrast 4 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 2: GDP Per Capita (current US$), 1960–2012, and Timeline GDP per capita (current US$) 7000 Sanctions; weapons 6000 2003-May 2013: inspection, no- y Iraq invades Kuwait; 2nd Gulf war Saddam Hussein comes to power US troops complete withdrawal 113,299 violent Nationalization of petroleum US led coalition invades Iraq zones' Oil for food Iran-Iraq war; 1st Gulf war 5000 1980–88 civilian deaths Iran-Iraq program 4000 war 3000 2000 1000 0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: World Development Indicators (2012); estimates for 1990–1996 from CSO, Iraq. to the Gulf monarchies, both Iraq and Gulf states declining global demand. However, whereas in oth- were now aligned against the Islamic and Shia fun- er developing countries the loss of momentum in damentalism emanating from Iran. development was the impetus for structural reform, in Iraq, policy reform was off the table, leaving in This culminated in the costly 1980–88 war with place the 1970s structures of a large role for the Iran—initiated by Saddam—which marked the first state, ossified administrative procedures, subsidies, of Iraq’s divergences from its neighbors. While this and crowding out of the private sector. period was associated worldwide with the first wave of reducing the role of state and structural reforms, The end of the Iran-Iraq war brought little respite. in Iraq the war effort further centralized the alloca- In principle, Iraq had the opportunity to use the tive role of the state, favoring corruption and con- reconstruction effort to drive a broader catch-up nections over private sector development. In prac- with the lost years of the 1980s. However, the Sad- tical terms, the war front involved Iraq’s key oil dam regime was instead focused on its own en- producing and exporting facilities in the south, put- trenchment, and a sense of impunity around the ting an end to oil sector expansion plans and causing inner circle was pervasive—with a severe deterrent considerable destruction. While Iraq was still able to effect on anyone who might come to the attention produce and export oil during this period, and ben- of the regime. In economic terms, the war had left efitted in terms of access to global commerce from large a debt overhang, a considerable amount in suspicion about Iran, the economic base was nar- the form of war loans from the Gulf countries. The rowing as defense and food imports were prioritized. combination of deteriorating relations with the Gulf The country’s capital market access dwindled to de countries over these debts with Saddam’s increasing facto or de jure official sources, such as export credit sense of nationalistic grievance proved to be a lethal guarantees and loans from the GCC countries. combination, leading to the invasion of Kuwait in July 1990 and then a multinational military opera- As can be seen in Figure 2, Iraq’s development tra- tion to liberate Kuwait and destroy Saddam’s inva- jectory during this period was stalled growth rather sion force in January 1991. than collapse of the economy. Certainly, GDP per capita stagnated during the 1980s, but this was the This early 1990s period was Iraq’s second diver- typical experience for the region’s oil exporters due gence, and in many respects it was more damag- to low prices and cuts in production in response to ing than the first. Unlike the Iran-Iraq war, where Conflict, Growth and Development 5 Iraq had significant tacit international support, the 5 percent going off the top for Kuwait reparations, invasion of Kuwait was almost universally seen as a and other revenue released only to finance approved catastrophic miscalculation which had upended the essential imports, most notably food and medicine sovereign norms of the Arab world. The invasion re- but also basic industrial parts and equipment. Since sulted in an overwhelming alignment of Arab coun- the government’s entire policy framework—for tries against Iraq (including all the Gulf countries, jobs, subsidies, and investment—had been predi- Jordan, Egypt, and Syria), with one of the few disso- cated on access to oil revenues, the result was near- nant notes coming from the PLO and Yemen. As is total economic collapse by the middle of 1991.1 The widely known, the US-led force which liberated Ku- scale of the decline in GDP per capita in 1990–91 wait pursued Saddam’s military into southern Iraq, compared to that in the early 1980s (which reflects compounding the destruction in the southern prov- the combined effect of oil price decline and war inces from the Iran war era. However, the coalition with Iran) is striking. could not agree on an extension of the mission to remove Saddam from power and he was left in place. A critical aspect of the post-1991 sanctions environ- This lack of finality meant that Iraq would continue ment was the protracted process for putting perma- under a stringent UN sanctions regime (initiated fol- nent arrangements in place, during which time the lowing the invasion of Kuwait) until 2003. most stringent form of the sanctions applied. As a result, the 1990–91 GDP decline was locked in for The immediate aftermath of the 1991 war was a decade, in contrast to the normal pattern of a re- stark. Encouraged by the belief that Saddam’s de- construction-driven recovery. There was a wide di- mise was imminent, there were significant uprisings vergence between the Saddam regime and the UN in the Shia-dominated south and Kurdish north. Security Council over the structure of sanctions. The coalition intervention was limited to imposing Negotiations took several years, and given its secure northern and southern no-fly zones. This has lim- grip on control, the regime had little incentive to ited effect in the south because the regime could take general well-being and the impact of sanctions still move in ground forces, but the Kurdish region thereon into account. A workable sanctions system took advantage of its geography and universal dis- in terms of meeting national humanitarian needs enchantment with the central government to es- was not in place until 1996 and industrial imports tablish a de facto autonomous region based on the were essentially frozen for the entire decade. Oil three provinces that already had a nominally special output in 1995 was less than in 1960. The oil sec- status within Iraq dating from the 1970s. Backed tor could only accomplish routine maintenance— by its own Peshmerga militia, the Kurdish Region enough to keep crude export flowing at something was able to enforce its separation from the regime like post-1991 levels, but new investment, or even in Baghdad. However, one consequence of this was reinvestment to maintain existing capacity, was that Iraq now had a fragile internal border in the ruled out. As the country had defaulted on most north, with ethnic groups scattered both sides of it of its international financial obligations following and extensive hydrocarbon resources below it. This the invasion of Kuwait, its capital market access was set the stage for forced displacement as the regime gone and its domestic financial sector was insolvent. continued with a policy of “Arabizing” the areas in northern Iraq that it controlled. For ordinary Iraqis, the operational impact of the sanctions regime was through government delivery As already mentioned, Iraq came under extremely tight UN sanctions following the invasion of Ku- 1 The consequences in terms of welfare indicators is de- wait. Exports and imports were subject to a sanc- scribed in Dreze, Jean and Haris Gazdar, 1991. Hunger and tions framework in which all oil export revenues had Poverty in Iraq, 1991, World Development Vol. 20, No 7, to be paid into a designated US bank account, with pp 921–945. 6 THE UNFULFILLED PROMISE OF OIL AND GROWTH of food rations and medicine under the auspices of with each phase of isolation, while the brain drain the Oil-for-Food program, all under UN oversight.2 compounded that isolation as the country lost its The performance of the Oil-for-Food program it- intellectual connections to the outside world. Any self became a point of controversy in the run-up to commercial opening to the world required the ap- the 2003 invasion, but it seems clear that the food proval and facilitation of the regime apparatus: this component thereof—the Public Distribution Sys- was privilege in extreme form, and hardly the stage tem (PDS)—did succeed in delivering a monthly for inclusive growth. subsistence ration to most Iraqis between 1996 and 2003. Nevertheless, the sanctions did have some- what paradoxical effects: although targeting the re- Reconstruction and Insurgency gime, their effect was to increase the power of the regime since it had a role in delivering the goods Although the 2003 US-led invasion generated nu- and awarding the various export and import con- merous impacts, for the purposes of the poverty and tracts under the program. Thus, incentives for cor- inclusion assessment, some pertinent after-effects ruption (especially given the ease of trading spot are as follows. In quick succession during 2003–04, cargoes of oil) were considerable. The signal to the the Iraqi administrative and security state was dis- private sector about what kind of activities would be mantled and then reassembled. Administrative and rewarded was unmistakable. Baath party structures were abolished and the for- mer reinstated, but at the cost of an enormous loss With a broader development policy agenda off the of remaining capacity as much of the public sector table, these post-war arrangements remained domi- did not know who they worked for. By 2004, there nant throughout the 1990s. The Kurdish region be- was a rapidly deteriorating security situation which gan putting the basics of a state—in administrative transformed into the sectarian civil war of 2005–06, terms, a replica of the Iraqi state—in place. The Shia resulting in high levels of displacement across the majority provinces were already neglected in the country and within Baghdad. With a limited range pre-1991 era and now subject to the distrust of the of instruments to respond, the government placed Baghdad government. The private sector had never emphasis on security spending and public sector been given a major role in Iraq’s economy, and there jobs and pay increases to secure the loyalty (or at was no internal or external impetus for moderniza- least the non-defection) of the public. tion of private sector (e.g. WTO membership). Rela- tions with the Gulf had ruptured in 1990, and land The security situation made it impossible for the trade routes to the Gulf were effectively closed. For government to focus on reconstruction (especially practical purposes, the only open land routes were for fragile network infrastructure like electricity and through Syria and Jordan, and these routes are not water) or chronic and accumulating sector problems aligned with Iraq’s major commercial centers which such as agricultural decline. As a result, broad-based were on a north-south axis reflecting the historic in- economic development and diversification, which fluence of the rivers and the Ottoman orientation of needs such basic infrastructure, was effectively on the pre-independence administration. hold until some semblance of stability could be re- stored. At the same time, a new layer of displace- It is worth noting that by the early 2000s, Iraq had ment and internal fragmentation of cities was added already endured 20 years of disruption and isola- to previous episodes. In terms of economic growth, tion. Although the regime had little interest or the oil sector started to ramp up production, but ability in playing a developmental role, it was well- in an enclave model of development: the focus of entrenched and any potential opposition had long since been repressed or exiled. A chronic brain 2 In the Kurdish region, the food ration delivery was man- drain which had begun in the late 1970s accelerated aged directly by UN agencies. Conflict, Growth and Development 7 the companies was on oil-field rehabilitation and ensuring the system’s adequacy, affordability, and export infrastructure. Thus there were limited local fiscal sustainability. spillovers and a downgrading of priorities that usu- ally feature in emerging natural resource producers, The other, smaller, pillar of social protection is a such as local content development. cash transfer program. Along with ad hoc transfers to the vulnerable such as from religious founda- With restored oil production coming on stream tions, it can be an important source of income for at a time of high and rising global oil prices, large households in specific categories. The use of broad amounts of revenue and spending began flowing social categories rather than more effective targeting through the government, but in a context of very mechanisms (such as proxy-means testing or geo- weak public financial management (PFM) and gov- graphical targeting) to determine eligibility is likely ernance mechanisms. In the face of instability, the to result in the majority of the poor falling through government had no appetite to undertake major the cracks of the non-subsidy safety net. reform of state-owned enterprises (SOEs), even though many had been defunct since the 1980s. While oil remained the main driver of economic Despite their lack of viability, the government could and fiscal developments, there were other bright rely on the two large state-owned banks, Rafidain spots, notably the reopening of trade with Iran and and Rashid—themselves insolvent—to lend to the a revitalization of the pilgrimage cities (Najaf and SOE sector for payroll financing. This locked in Karbala), also linked to increased openness with Iran. continued state dominance of financial sector, since But there was no transformation in the structure of a financial sector restructuring would have opened GDP; it is noticeable that the share of oil rents in up the question of broader SOE reform.3 GDP remained broadly stable from the mid-1990s to the present, despite the tumultuous events dur- In terms of the social safety net, the PDS contin- ing this period (Figure 3). The crude oil sector ac- ued to function as a universal food ration and the counted for 44 percent of real GDP in 2012. backstop to subsistence consumption for the coun- try. There were also universal price subsidies for en- Overall, the reconstruction period marked the be- ergy, although these have been somewhat reduced ginning of the end of Iraq’s divergence from the over time by eliminating on-budget fuel subsidies. region. However, the reintegration was very in- Iraq continues to spend a sizable percentage of complete. Civil war acted as a further deterrent for its social protection budget on pensions. In 2010 skilled Iraqis to remain in Iraq, or for the externally more than 4 percent of Gross Domestic Product displaced population to return. Foreign investors went to pensions. This is among the highest lev- were likewise deterred. Relations with the GCC els of spending in the region. Emergency policies countries remained frozen, and unresolved legacy that were implemented after April 2003 replaced debts with the GCC countries outside the Paris regular pensions with emergency “flat” payments Club framework acted as a legal impediment to paid directly from the Ministry of Finance budget, commercial relationships. with very limited contributions from employers and employees. At the same time, only around 25 per- cent of the total labor force in Iraq is covered by 3 The public sector’s total domestic obligations have increased a mandatory pension system (most of these public significantly since 2009. Total obligations increased from sector workers). Only around two percent of the la- ID 5.2 trillion at end-2009 to ID 16.2 trillion at end-2012 (6.5 percent of GDP). The obligations are composed of bor force in the private sector is actually covered. (1) T-bills held by the banking system and bank loans total- Therefore, there are now various calls for reform of ing ID 7.5 trillion; and (2) government guaranteed loans by the pension system, including full integration of the state-owned banks to state-owned enterprises totaling ID public and private sector scheme as one fund, and 8.6 trillion. 8 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 3: Oil Rents (% of GDP), 1960–2012 120 100 80 60 40 20 0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: World Development Indicators (2012). Very little structural policy reform took place dur- revenue become an enabler of rapid growth in pub- ing the reconstruction period, and private sector lic sector employment. The size of the public sector development is probably the weakest reform area of in Iraq-measured by public spending-to-GDP ratio all. Iraq ranks 169 for starting a business and 189 is one of the highest in the region (61.1 percent of for resolving insolvency—the worst score in the GDP), and the approximately 4 million public sector world—because it is not possible to legally close a employees account for nearly half of total employ- business. The informal sector in Iraq is very large ment. Government and state-owned enterprises em- vis-à-vis the formal sector and formal sector firms ploy approximately half of the labor force. The wage are squeezed between SOEs with open-ended fund- bill in the public sector and military jobs consume a ing and informal firms that don’t have to comply large part of the recurrent budget. with the same rules as them. Yet growth in public sector employment has not ani- mated the labor market overall: participation rates re- Current Policy Framework main exceedingly low, especially for women. Multiple factors explain low participation including lack of se- The legacy factors and economic circumstances dis- curity, “wait unemployment” pending a public sector cussed above are mediated partly through the pre- job, and substantial numbers of discouraged workers vailing economic policy framework. This section out- as the labor market does not generate appealing jobs. lines this policy framework and describes how some The gains associated with public sector employment key economic outcomes are affected by it. Perhaps are substantial: one study (ERF) found based on the dominant influence on households is that in the 2007 data that the most important determinant of post-2003 era, public sector jobs and numbers be- receiving formal benefits is the sector of employment: came part of the oil redistributive mechanism. This public sector workers are 83 percent more likely than is the pattern throughout the MENA region, but it private sector workers to have formal benefits. had been somewhat constrained from operating in Iraq as oil revenues were preempted by war and sanc- International trade would normally be expected to tions, so oil revenue management had not been a sig- be a driver of integration. The potential for trade to nificant area of discretion for the government since grow would similarly be expected to be high, as the the early 1980s. However, once revenues started to constraints of sanctions were relaxed. But in addi- increase and public sector positions were one of the tion to the challenges of insecurity, trade was also few things the government could reliably deliver, oil subject to the vagaries of geopolitics. The country Conflict, Growth and Development 9 had only one consistent trade partner in the entire this policy mix favors short-term and easily reversed post-1991 era (Jordan), while its relations with oth- investments over the sustained commitments that er neighbors were in constant flux as they calculated Iraq needs to boost its growth potential. how to position themselves to their own advantage and the disadvantage of others. For example, Syria The content for intergovernmental fiscal relations is and Iran had complex stances during the 2006–07 also important. Iraq is designated a federal state in civil war, which included enabling the transit of its constitution. The common subnational layer is Sunni militants, while over time Turkey has varied composed of governorates, of which there are 18; its emphasis on relations with Baghdad and Erbil, 15 are not grouped into a region. The constitution depending in part on its own energy strategy. permits the formation of semi-autonomous regions from one or more governorates, but to date only the Iraq has trade barriers, mainly regulatory and bu- semi-autonomous Region of Kurdistan, with 3  gov- reaucratic practices, which restrict the level of trade ernorates, operates in this fashion. The boundaries and investment. These include increasingly burden- of KRG reflect a 1970s arrangement by the Saddam some import procedures, corruption at the border, regime to moderate separatism, but meaningful au- stringent requirements on certificates of origin, pre- tonomy dates from 1991. With the exception of this shipment inspection certification requirements on region, subnational powers are limited. agricultural products; significant behind-the-border barriers and inadequate mechanisms in place to per- It is important to note that a revised provincial pow- form these processes or revise them to more pro- ers law passed in 2013 which substantially increases gressive ones. Many of these issues can be traced the power of the elected provincial governments, es- to Iraq having been largely absent from the inter- pecially over public service delivery. But this law has national trading stage for more than three decades, not yet been implemented. This illustrates a more other than with regional trading partners, and hav- general issue that the assignment of powers and re- ing little experience with free trade as a result. In- sponsibilities between the layers of government is stead, trade relations are seen as an extension of po- incomplete and a large grey area is left to be deter- litical relations. mined by power politics. Iraq has an unusual macroeconomic stability con- Exclusive federal powers include (a) “drawing for- text, even allowing for the effects of oil and con- eign sovereign economic and trade policies ... set- flict. Because of the structural weaknesses of the ting up general budget of the nation and drawing financial sector, the economy is cash-based and up currency policies”; (b) drawing up financial and somewhat dollarized, given the Iraqi dinar peg to customs policies; and (c) planning policies con- the US dollar (ID1170 per dollar). In turn, the peg nected to water resources from outside Iraq. There is a product of an exchange-rate based stabilization are also a set of powers that are shared by federal during 2006–2008 which was designed to bring authorities and regional authorities which may re- down high inflation. As with many exchange rate quire some coordination between the two levels of based stabilizations, this appears to have contribut- government—for example general planning and de- ed to persistent overvaluation of the currency, com- velopment policies, administration and organization pounding the standard effects of Dutch Disease in of customs and organization and distribution of the squeezing the traded goods sector. While there are main electrical power resources. high headline GDP growth rates, these are driven by expansion in crude oil production. At the same Virtually all expenditure and revenue assignments time, there are strong indications of capital flight, are central, with policies established in Baghdad which serves to constrain domestic non-oil invest- and implementation through de-concentrated agen- ment. When combined with fragility and insecurity, cies of central ministries in the governorates and 10 THE UNFULFILLED PROMISE OF OIL AND GROWTH municipalities. The share of total public spending Some key emerging messages are as follows. First, executed through federal government including 1991–92 forms a major rupture in Iraq’s develop- de-concentrated structures is estimated to exceed ment trajectory. While the Iran-Iraq war was costly 90 percent. The two exceptions are (1) a small num- and destructive, the state had managed to maintain ber of capital projects assigned directly to gover- its core functions, and Iraq was not isolated from the norate administrations for implementation and (2) world during the 1980s. However, the invasion of projects that are included in the “Regional Develop- Kuwait triggered comprehensive sanctions months ment Program” which is a special item in the budget before Operation Desert Storm, and as noted above, (around ID4 billion) allocated to the governorates the sanctions era didn’t end until 2003. With the using a formula which is intended to capture needs. government consolidating around the goal of regime survival and the private sector unable to function, Global experience with service delivery shows that this was the beginning of large-scale detachment functioning of administrative systems can have from government and the formal sector for many a strong influence on performance. In Iraq, pub- Iraqis—laying the roots of profound marginalization. lic financial management (PFM) is a known area of weakness. In practical terms, this can mean the Second, the space for market allocations in Iraq coexistence of aggregate revenue abundance and was constricted since so much allocation was tak- line ministries, especially in the governorates, that ing place through administrative fiat. Throughout are often short of cash. The Ministry of Finance the sanctions era, state enterprises and certain cad- (MOF) has treasury operations in each governor- res (military, Ba’ath) were given favored access to ate. Line ministries have their own offices at each imports. This was privilege and connectedness in layer of government including municipality and extreme and debilitating form. Since the formal sec- they execute their local activities through these of- tor had to be run through the government, many fices rather than through the local government. The market activities shifted to the informal sector, with revised provincial powers law is supposed to bring consequent effects on productivity, investment ho- these operations under provincial control. rizon, and job quality. A closely related issue is the impact of governance shortfalls on diversification: Provincial councils have little power as most services if effort and resources are being pulled into distor- are delivered by de-concentrated offices of central tions caused by deficient procurement, arbitrary ministries. They have no hiring authority. However decision-making, and excessive procedures for com- the provincial councils do play a role (jointly with mercial transactions, they are being directed away the Ministry of Planning) in formulating and ex- from job creation and productive investment. ecuting the provincial capital projects included in the National Development Plan. In KRG, provinces Third, spatial divergence became pronounced. have more autonomy over services, but employees Whereas the Kurdish region had been the most are appointed by the regional government and paid victimized before the 1990s, with de facto inde- out of the 17 percent revenue allocation from the pendence from Baghdad, food and nutrition distri- center. Districts and municipalities engage in purely bution was directly implemented by UN agencies. local services, such as street paving and cleaning, These had more flexibility—and access to cash— and feeder roads. than the Baghdad government for the rest of Iraq, meaning that basic needs could be met more ef- fectively. On the other hand, the southern prov- Legacy inces saw a double negative impact: the destruction of wars now compounded by a failed rebellion. The purpose of the above overview has been to set the In principle, these trends should have moderated context and expectations for the poverty assessment. post-2003 since Iraqi government could direct Conflict, Growth and Development 11 resources to areas of greatest deprivation. The ad- employment of the sector’s workforce. Cereals pro- ministrative system for governorates not in a re- duction in 2000 was around one quarter of its 1990 gion is de-concentrated and therefore might be level. Some sector problems predate the conflict era insulated from capacity differentials across regions. (irrigation water supply, salinity, and desertification) In practice, there seems to be a significant variation but throughout the 1990s the sector was afflicted in the quality of delivery of common services across by lack of access to critical inputs. Severe droughts the country: the equalizing role of the central gov- in the late 1990s made things worse. Again, the ernment was not effective. There is considerable Kurdish region fared somewhat better due to avail- World Bank knowledge on approaches to establish- ability of rain-fed land and the ability to source in- ing (or re-establishing) social cohesion and basic puts through grey markets outside the sanctions services in conflict-affected and fragile environ- regime. Thus with more effective food distribution ments. However it is worth noting that three well and more diversified supply, nutritional outcomes known cases—Indonesia Kecamatan Development in northern Iraq began to diverge from the rest of Program (KDP) and local government support in country in the 1990s. Bangladesh and Sri Lanka—arguably represented cases of strong central governments seeking a di- Fifth—and related to the second point above— rect channel to local communities. Iraq’s central state-owned enterprises (SOEs) remained domi- government did not have this type of capacity, and nant. Agriculture is a good example: the agricul- thus gaps in local services across the country were tural inputs industries were concentrated around highly persistent. Baghdad both to provide sources of patronage and ensure regime control of the sector nation- With the uneven nature of security restoration wide. This would prove to be a major vulnerabil- across the country, the ingredients for a new set ity from 2003 onwards, as insecurity disrupted the of divergences between the provinces were now sector’s supply chain across the country given the in place. Since the southern provinces were finally need to get critical inputs from areas around Bagh- out of the grip of a hostile government in Bagh- dad. More generally, the state had no incentive to dad, and home to many of the country’s oil fields, restructure SOEs; even aside from their value as the pace of economic development could pick up. a means of economic control, public sector jobs While oil continued in the enclave development were one of the few reliable instruments that the model, other sources of economic activity such as post-2003 government had at its disposal. Far from pilgrimages, trade with Iran, and the Basra port, being impetus to reform, growing oil revenues be- came into play as drivers of regional growth. On came an enabler of the status quo. State dominance the other hand, the provinces north and west of of financial sector remained, the insolvency of the Baghdad were in effect contested regions: be- system became a reason to do nothing, and banks tween sects, ethnicities, tribes, and insurgents, all now had the means to expand as their government interacting in different ways with the government. business was growing. In particular, the government faced a complex cal- culation vis-à-vis the provinces: those with clear Sixth, a cross-cutting element of Iraq’s legacy is majorities of one sect could be taken for grant- severe economic and social fragmentation. Civil ed—perhaps to their detriment—while those with conflict created substantial internal displacement more finely balanced populations could also attract corresponding to ethnic and sectarian divisions. more competition for influence but also more vio- Baghdad became a city of internal boundaries lence and insecurity. formed by blast walls, check-points, and no-go ar- eas. For the country as a whole, a basic prerequisite Fourth, there was a catastrophic decline of ag- of economic development—internal integration— riculture, posing challenges for food supply and was lost. 12 THE UNFULFILLED PROMISE OF OIL AND GROWTH Previous World Bank analytical work on Iraq pro- Violence, Growth and Development vides some guidance as to the effects that could be expected from Iraq’s recent growth pattern. The It is very difficult to quantify the entire economic 2012 Country Economic Memorandum used a and social impact of violence on growth and de- modeling framework drawing from the Oxcarre velopment. In this report, we attempt to provide methodology for describing the sectoral effects quantitative evidence of the relationship between of various options for oil revenue management in conflict and development in Iraq, in the immediate Iraq. One of the stylized options considered was al- period covered by this poverty and inclusion assess- locating all revenues to public sector pay and num- ment, and in some cases, the effect of longer term bers, which is helpful in illustrating the impact of violence and deliberate neglect. While the rest of the actual dominance of this type of spending in the report primarily draws on evidence from mi- the budget. The model shows that such spending cro data (household surveys), we begin by draw- essentially eliminates the economy’s traded good ing on a district-level dataset to establish and quan- sector. Wages rise sharply, which mitigates any tify the negative relationship between conflict and beneficial impact of government spending on the economic growth on the one hand, and the posi- real economy in terms of provision of goods and tive relationship between development efforts and services. Given spending patterns of relatively well- growth on the other. In doing so, we show that off public employees, consumption needs are met while development efforts by the government and by imports (which are high cost due to logistical other actors can mitigate the negative role of vio- deficiencies), and because there is no public accu- lence, the required effort is substantial. Thus, on mulation of capital or foreign assets, the economy a macro-economic level, and absent the establish- is completely oil-dependent. In practice, Iraq does ment of peace and security in the nation, maintain- allocate a significant portion of oil revenues to in- ing growth can in itself be a challenge in Iraq, and vestment, but long-standing deficiencies in public without these two pre-conditions in place, poverty investment management limit the productive ef- reduction and broad-based prosperity will be dif- fects of this spending. ficult to sustain. In summary, the overarching context for the pov- The relationship between growth, development and erty assessment is Iraq’s status as a resource-rich conflict can run both ways. Miguel et al. (2004) use fragile and conflict state (FCS). Achievement of the cross-country data to investigate the effect of eco- twin goals of ending extreme poverty and boosting nomic development on conflict, and find that a nega- shared prosperity is bound to be an uphill struggle tive shock to economic growth is associated with a in that context. As the Global Monitoring Reports significantly higher likelihood of conflict emerging in have shown, weakness of state capacity in FCS plays the following year. In the Iraqi context, Berman at al. a disproportionate role in the failure to achieve (2011) combine detailed data on insurgent violence the MDGs. The dynamics of income growth in with information on reconstruction projects aimed at resource-rich FCS are often closely linked to fis- restoring public services in Iraq. They find that im- sures in society (e.g. limited formal or resource sec- provements in service provision tend to reduce insur- tor growth which only benefits connected groups). gent violence, most noticeably for smaller projects, How did these forces play out at the individual and and in particular after the “surge” began in 2007.4 household level in Iraq? Before we turn to evidence from micro-level data, we provide some quantitative 4 Ahrens (2013) provides further empirical support for the negative correlation between development and violent macro-level evidence on the relationship between conflict using cross-country data for Africa. Shapiro and conflict and violence in Iraq in the period follow- Weidmann (2011) further investigate whether improved ing the US-led invasion; development efforts and communication means can influence the degree of vio- economic growth. lence in the context of Iraq. They argue that if there is in- Conflict, Growth and Development 13 On the other hand, conflict can also negatively affect urbanization. Finally, we also include GSM cell- growth and development. Abadie and Gardeazabal phone coverage to measure the role of enhanced (2003) try to estimate the economic cost of violent communication. conflict, focusing on the Basque country in Spain. The economic cost of conflict is found to be quite We use night time lights high gain data from DM- substantial; after violent conflict emerged in the late SP-OLS, which are publicly available from NOAA.8 1960s, per capita GDP in the Basque country was We derive two metrics from the night time lights found to have declined by about 10 percent relative data at the district and governorate level: the mean to a control region that has not been exposed to of all night time light pixels within an administra- violent conflict. Murdoch and Sandler (2002) use tive unit and an inequality measure (the Theil In- cross-country data in their effort to estimate the ef- dex) within an administrative unit. We expect mean fect of violent conflict on economic growth.5 They night time lights to be significantly correlated with too find that violent conflict is detrimental to eco- economic growth. Figure 4 plots this measure at nomic growth, but that the damage is most pro- nounced in the short-run.6 deed a relationship, the sign of the effect can go both ways, In what follows, we try to further understand and as improvements in communications may ultimately help quantify the relationship between the level of vio- both sides (the side who wishes to initiate violent conflict and the side who seeks to curb conflict). The question then lence on the one hand and development efforts is which side is best equipped to take advantage of the new on the other on economic growth in Iraq. We do technology. Their empirical results suggest that the coun- this by estimating a set of growth regression mod- terinsurgents gain the most as an increase in cellphone els using district level panel data with annual ob- coverage is found to have reduced insurgent violence in servations between the years 2003 (which marks Iraq. the beginning of the US-led invasion of Iraq) and 5 They consider different lengths of “growth-spells” (5 years 2010. compared to 10–25 years) to investigate the importance of the time-horizon, and also explore the significance of spa- Since disaggregated income or GDP data at high tial spill-overs. frequency is not available, we rely on night-time- 6 They offer two possible explanations for this finding: (a) lights (NTL) data as a proxy for local economic violent conflicts tend to be relatively short-lived, and (b) in the longer term, the effects of conflict on growth may get output; and information on the number of US fi- diluted with the convergence effect. nanced reconstruction projects in Iraq as a proxy 7 This has been shown to work reasonably well, see e.g. Hen- for development efforts.7 Thus, the independent derson et al. (2012). [Henderson, V., Storeygard, A. and variables considered in the growth regressions are: D. Weil, Measuring economic growth from outer space, lagged NTL (to account for convergence effects American Economic Review, 102(2), pp994–1028]. across different parts of the country with differ- 8 NOAA DMPS-OLS (The National Oceanic and Atmo- ent levels of initial growth); conflict violence (as spheric Administration’s Defense Meteorological Satellite measured by the number of civilian deaths ac- Program) has another night time light data product that cording to the Iraq Body Count); reconstruction provides a radiance correction and corrects for the top-cod- projects; ethnic fractionalization and the share of ing problem, which is present in the high gain night time lights data. Currently, the low gain data are not available agricultural land. Since the latter two variables do at a consistent annual basis necessary for this time-series not vary over time, we interact these with the vio- analysis. lent conflict and reconstruction project variables. The data display luminosity in units of Digital Number from These interaction terms, if significant, will tell 0 to 63, where high values represent high luminosity. Due to us how the effect of conflict and reconstruction the lack of inter-calibration of the satellite, we perform an efforts on growth might vary depending on the adjustment to the raw data in order to calibrate the data for degree of ethnic diversity and/or the degree of time-series analysis according to Elvidge et al. (2013). 14 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 4: Log Mean Night Time Lights (District Level), 2003, 2005, 2007 and 2009 the district level on a map for 2003, 2005, 2007 appears to be in line with GDP, reflecting a steady and 2009 and shows high levels in and around increase, especially after 2007, which represented Baghdad as well as near Basra. With regards to the peak of internal violence in Iraq. the between region differences, the highest levels are in the Central region and the lowest levels are Data on violence comes from Iraq Body Count present in the North region. Figure 5 shows the (https://www.iraqbodycount.org), which is a da- annual changes in night time lights at the district tabase of violent civilian deaths in Iraq since 2003, level; there is a small gain from 2003 to 2005, then drawn from crosschecked media report and supple- there is a decline to the minimum level at 2007, mented by review of hospital, morgue, NGO and and finally there is a steeper increase in the night official figures. We use their conflict incidents da- time lights until the last year in the analysis (2010). tabase, where each incident has a start date, an end Thus, this proxy measure for economic growth date, and an estimate of minimum and maximum Conflict, Growth and Development 15 FIGURE 5: Log Average Night Time Lights, FIGURE 7: Average Civilian Casualties (Iraq 2003–2010 Body Count) By Region: 2003– 2010 Log average night-time-lights (district) 4 North 3 8 6 4 2 2 0 Monthy average of civilian casualties (governorate) 1 2003 2004 2005 2006 2007 2008 2009 2010 0 Central 2003 2004 2005 2006 2007 2008 2009 2010 8 6 Source: Authors’ calculations derived from DMSP-OLS. 4 2 0 FIGURE 6: Average Civilian Casualties (Iraq 2003 2004 2005 2006 2007 2008 2009 2010 Body Count): 2003 to 2010 South 8 8 6 civilian casualties (governorate) 4 6 2 Monthy average of 0 4 2003 2004 2005 2006 2007 2008 2009 2010 2 0 2003 2004 2005 2006 2007 2008 2009 2010 Data on reconstruction projects and ethnic fraction- alization are drawn from the ESOC database, while population estimates are drawn from the Landscan population database.9 The available data on the cellphone network is for the Zain network, which body count. Since both the minimum and maxi- is a dominant provider in the Baghdad and South mum estimate of the body count is provided, the regions; we do not use any GSM cell phone data midpoint is used in the analysis. Incidents can occur for the north of Iraq and also exclude the Central across months; the given start date is considered districts of Nineveh, Salahaddin, and Kirkuk due to the month of the incident. Then, it is possible to lack of consistent or missing data from Zain.10 In calculate the average body count per month with- order to measure agricultural land use in Iraq we in a year at the governorate level. It is expected that the body count data have a negative effect on growth. Figure 6 shows the variation by year during 9 For a detailed discussion of the ESOC dataset kindly refer period 2003 to 2010, which shows a peak of the to the ESOC-I-v3 codebook, available at: http://esoc.princ- body count in 2007. Figure 7 displays the between eton.edu/subfiles/codebook-iraq-civil-war-dataset-v3 (ac- region differences for the same time period; the cessed 2014-05-06). A more detailed description of Land- Central region experienced the highest body count scan can be found at: http://web.ornl.gov/sci/landscan/ and the North region shows the lowest counts of landscan_documentation.shtml (accessed 2014-05-12). the three regions. 10 Jake Shapiro kindly provided data on cellphone coverage. 16 THE UNFULFILLED PROMISE OF OIL AND GROWTH use the Global Hybrid dataset (0611–2012 V2) We indeed find that in Iraq, violent conflict between produced by Fritz et al. (2011) which estimates the 2003 and 2010 had a negative effect on subsequent percentage share of land used for agriculture within growth (proxied by growth in night time lights), a one square kilometer pixel. In this analysis, we while development efforts (measured by recon- also use Normalized Differentiated Vegetation In- struction projects) mitigated this effect by promot- dex (NDVI) data constructed by the U.S. National ing growth. This finding is fairly robust across the Aeronautics and Space Administration (NASA) different methods of estimation and across the dif- Global Inventory Modeling and Mapping Studies ferent samples used. Estimates based on the System (GIMMS) at a bi-monthly frequency between 2003 GMM suggest that compared to a district with no and 2009 that measure greenness over 8 square ki- civilian casualties, a district with the average number lometers pixels available for the entire area of Iraq of casualties experienced between 5.7 and 11.4 per- (Zhu et al. 2013).11 cent lower growth depending on the specification used (Annex Table 1.1). On the other hand, a dou- The correlates of growth we are indeed most inter- bling of reconstruction projects increased growth ested in are the incidence of violent conflict on the between 5.8 and 8.4 percent. This suggests that one hand, and reconstruction efforts on the other the size of the development effort required to com- hand. Do they both have an impact on growth pensate for the negative growth effect of violence is (presumably with opposing signs)? If indeed, to substantial. what extent are the reconstruction efforts able to compensate for the damage done by violent con- Moreover, the negative effect of violent conflict on flict? Finally, under what conditions are the effects growth is strongest in the central region (includ- on growth particularly strong (or diminished)? ing Baghdad), in ethnically mixed districts, and in We consider three regions: the north comprising relatively poorer districts (proxied by low intensity Kurdistan, the south, comprising Basra, Thi Qar, of NTL). On the other hand, the positive effect of Muthanna, Missan and Qadisiya and the central development efforts on growth was strongest in the region, comprising the rest of the country. We central region (including Baghdad) and the southern use two different samples for our growth regres- region; in urban areas (with low shares of agricultural sions: one with and one without the northern re- land); in high income districts (proxied by high inten- gion (Kurdistan), the reason being that we only sity of NTL) and in ethnically homogenous districts. have GSM cell phone coverage data for the rest With the inclusion of GSM coverage information and of the country. Hence the regressions with GSM for the sub-sample excluding the Kurdistan region, coverage as one of the control variables exclude the districts with improved communication are also as- northern region from the sample, while regressions sociated with higher growth (Annex Table 1.2). without GSM coverage use the full cross-section of district observations. Thus, not only is the development effort required to mitigate the negative impact of violence on To establish the robustness of our findings, each regression model is estimated using three differ- ent types of estimation methods: (i) Pooled OLS 11 Jim Tucker NASA kindly provided the data. with governorate dummies; (ii) District Fixed Ef- 12 It also avoids the finite sample bias that affects standard fects (the “within estimator”); and (iii) System OLS and FE estimators as they try to account for the district fixed effects (the so-called “Nickell bias”). We use GMM. The GMM estimator is often considered Principal Component Analysis to reduce the number of the preferred method as it allows one to account instruments in an effective manner. Too many instruments for potential problems of endogeneity.12 Below we relative to the dimension of the panel data are found to re- present the findings based on the System GMM duce the effectiveness of the System GMM estimator (see estimator. e.g. Roodman, 2009, 2012). Conflict, Growth and Development 17 growth substantial, violence disproportionately af- it has also suffered from a long and difficult legacy fects growth in poorer parts of the country, while of violence and insecurity, which continues in some development efforts are more effective in richer parts of the nation to this day. In some sense then, parts of the country. This implies that it will take an any progress in improving the welfare of its citizens even larger development effort to maintain growth is in and of itself a commendable achievement, and when violence affects poorer districts. Keeping this, has perhaps occurred despite the odds. The objec- and the larger political economy, institutional and tive of this report is to use the rich and detailed macro-economic context in mind, we now turn survey information available to learn from the ex- to the levels, trends and correlates of poverty and periences of the five year period between 2007 and shared prosperity in Iraq between the years 2007 2012; and to identify the many opportunities for and 2012. change in the future. In general, economic growth in a country can Framework of Analysis and Report translate into welfare improvements for its citizens Outline through two main channels: by enhancing the de- mand for labor, capital and other inputs and there- This report takes a broad and comprehensive ap- fore employment and earnings; and by boosting the proach to the analysis of poverty and welfare in Iraq, resources available to the government to transfer, and examines how inclusive the development pro- redistribute and provide services to the population cess has been—across space, groups, and gender. (see Figure 8 for a simple representation). Another Over and above the typical association between pov- core function of government, which is particularly erty and large families, lack of human capital, and relevant in the case of Iraq, is the establishment and adverse labor market outcomes, poverty in Iraq is a maintenance of peace and security and the rule of reflection and symptom of much larger, structural law. Each of these elements—employment, income and often long-term factors. Iraq is an upper-mid- (earned and unearned), the access to and quality of dle income country with one of the world’s larg- basic services, and the overall environment—con- est proven oil reserves and has experienced strong tribute directly or indirectly to household welfare, growth in the period considered in this report. Yet and therefore, to poverty and shared prosperity. FIGURE 8: Economic Growth and Welfare: Main Channels Economic growth Demand for labor Private Government and capital transfers budget Earnings Health, education Peace and security; Employment Public transfers Rule of law (Labor and non-labor) and service delivery and governance Household welfare 18 THE UNFULFILLED PROMISE OF OIL AND GROWTH This report began with an overview of the insti- development—health, education, and basic servic- tutional, macro-economic and political economy es—and poverty. The chapter that follows delves of the nation, which sets to context for the find- deeper into the factors driving spatial disparities in ings and analysis in the rest of the report. The welfare across the five divisions of Iraq; and in par- next chapter follows with the core poverty and ticular, the role of past and current violence and shared prosperity diagnostic, and an examination insecurity in the country. Chapter 5 identifies the of the determinants and correlates of life satis- different sources of household incomes that ac- faction and subjective well-being and how these count for poverty reduction for the country as a move with consumption poverty. We then take up whole and for different sub-divisions to get a sense an in-depth analysis of each of the important cor- of the importance of different channels. The next relates of welfare—human development, place of three chapters focus on the links between eco- residence, labor market outcomes, and the role of nomic growth, the labor market, public and pri- public transfers. Given the strong relationship be- vate income transfers and household welfare. The tween education and poverty in Iraq, Chapter 3 last chapter pulls the implications of the analysis examines the relationship between broader human together to provide directions for policy. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 2 O ver the five year period spanning 2007 and Poverty reduction has been spatially uneven. Rural 2012, Iraq’s GDP grew at a cumulative poverty fell by 8 percentage points, compared with the rate of over 40 percent, averaging 7 per- much smaller decline of 2.5 percentage points in ur- cent per year between 2008 and 2012. At the same ban areas. While there was little discernable improve- time, Iraq’s population grew by approximately 4 mil- ment in poverty in Baghdad and Kurdistan, in the lion persons, or at an annual rate of 2.5 to 3 percent. remaining 14 governorates of the country taken to- However, per capita real consumption, the basis for gether, headcount rates fell significantly. 70 percent of measuring poverty, grew at a rate of around 1.75 per- those in the bottom 40 percent of the population live cent per year, or in cumulative terms, by only 9 percent in these governorates, with Baghdad accounting for over the five year period. High rates of GDP growth another 20 percent. The pattern of poverty reduction did not translate into commensurate consumption has been accompanied by a greater spatial concentra- growth, and the latter was also unevenly distributed tion of poverty. In 2007, half of Iraq’s poor lived in across the population and across the regions of Iraq. five governorates—Baghdad, Basra, Nineveh, Baby- Consumption of the highest quintiles, the top 60 per- lon and Thi-Qar. By 2012, while Baghdad’s share of cent of the consumption distribution, grew faster than the poor remained unchanged at around 19 percent, that of the lowest; consumption growth was faster in Nineveh almost doubled its share to 15.7 percent. Three rural areas than in urban areas; and consumption southern governorates, Thi-Qar, Qadisiya and Mis- grew slower in Baghdad and Kurdistan relative to the san, now account for almost a quarter of the country’s rest of Iraq. Overall, poverty headcount poverty rates poor. In 2012, 58 percent of Iraq’s poor lived in these fell from 23.6 percent in 2007 to 19.8 percent in 2012, five governorates, compared to 40 percent in 2007. a 3.8 percentage point decline. Subjective measures of wellbeing and welfare highlight Poverty in Iraq is significantly higher among larger the different elements that the Iraqi people take into households, those with less educated heads, and varies by account when evaluating their own welfare, elements the employment sector of the head of household. House- that go beyond consumption and income. Overall, hold size and composition, the education and sector of headcount rates based on consumption are fairly simi- work (in general) of the head of household and the lo- lar to those based on the minimum income question cation of the household are all strong determinants of and on life satisfaction, while poverty as measured by consumption and poverty. While public sector jobs are subjective well-bring is higher at 26 percent. In rural in general associated with a lower probability of pover- areas, 20 percent of individuals report being dissatis- ty, households dependent on agriculture and construc- fied with their lives, 24 percent have lower per capita tion are as likely to be poor compared to households with consumption than their estimated basic income needs, heads who are unemployed or out of the labor force. while more than 30 percent are poor based on the 20 THE UNFULFILLED PROMISE OF OIL AND GROWTH consumption poverty line or assess that their household FIGURE 9: Annual GDP Growth in Iraq, is poor or very poor. In urban areas, while consump- 2007–2012 (%) tion poverty headcount rates are relatively low, other measures of poverty are significantly higher. 10 9 8.58 8.43 8 In determining subjective self-assessments of poverty 7 6.61 5.81 5.86 status, the largest weight is placed on lack of con- 6 5 sumption or income, with smaller but relatively equal 4 weight on educational deprivation or on place of resi- 3 dence. In contrast, dissatisfaction with life appears to 2 1 1.38 be driven more by where an individual lives, reflecting 0 the additional importance of location specific factors 2007 2008 2009 2010 2011 2012 including security concerns, local labor market con- GDP Growth (% annual) ditions and service delivery. Using these deprivations to construct a multidimensional index of deprivation; Source: World Development Indicators, 2012. headcount rates would be higher than as measured by consumption poverty—25.6 percent if derived from TABLE 1: Mean Per Capita Consumption the subjective poverty measure and 28.4 percent if the Expenditure life (dis)satisfaction measure were used (compared to 19.8 percent consumption poverty) in 2012. Percentage Annual 2007 2012 change change Urban/Rural GDP and Consumption Growth in Urban 191.1 207.7 8.6 1.67 Iraq: 2007–2012 Rural 133.7 152.9 14.4 2.72 Region Over the five year period spanning 2007 and 2012, 1 Kurdistan 277.9 279.0 0.4 0.08 Iraq’s GDP grew at a cumulative rate of over 40 per- 2 Baghdad 184.0 201.4 9.5 1.83 cent, and averaged an annual rate of 7 percent be- 3 Rest of Iraq 150.2 167.8 11.7 2.24 tween 2008 and 2012 (Figure 9). At the same time, Quintiles Iraq’s population grew by approximately 4 million Lowest quintile 80.9 83.7 3.5 0.69 persons, or at an annual rate of 2.5 to 3 percent.13 2 116.9 124.9 6.8 1.33 However, per capita real consumption, the basis 3 149.1 162.5 9.0 1.73 for measuring poverty, grew at a rate of around 4 193.2 214.1 10.8 2.08 1.75 percent per year, or in cumulative terms, by Highest quintile 332.7 366.5 10.2 1.95 only 9 percent over the five year period (Table 1). Total 174.6 190.4 9.0 1.75 Not only did these high rates of GDP growth not Source: Authors’ calculations, IHSES 2007 and 2012. translate into commensurate consumption growth, the latter was also unevenly distributed across the for the bottom 20 percent. On the other hand, con- population and across the regions of Iraq. Con- sumption grew faster where levels of consumption sumption of the highest quintiles, the top 60 per- were lower to start with—outside Baghdad and cent of the consumption distribution, grew faster Kurdistan, by 2.24 percent per annum; and in rural than that of the lowest (Table 1). For instance, the parts of the country, by 2.7 percent per year. top 40 percent of the consumption distribution ex- perienced annual growth in real per capita consump- tion of almost 2 percent, compared with 0.7 percent 13 Per capita GDP grew by 24 percent during 2007–2012. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 21 The distribution of the welfare or consumption ag- FIGURE 11: Probability Density – Welfare gregate over time reveals improvements in welfare Aggregate, 2007 and 2012 in line with the increase in mean per capita con- sumption expenditure. Figure 10 plots the poverty Probability Density Function 0.008 incidence curve, which is the cumulative distribu- tion of the welfare or consumption aggregate for Cumulative distribution 0.006 2007 (in red) and 2012 (in blue). For any possible and reasonable value of the consumption aggregate 0.004 chosen as the poverty line, the distribution of con- sumption in 2012 lies to the right of that of 2007, 0.002 implying lower rates of poverty in 2012. Figure 11 on the right panel plots the probability density func- 0 0 200 400 600 800 tion of the welfare aggregate in the two years, and Welfare aggregate (2012 prices) clearly shows, in line with Figure 10, the increase in median consumption over time. 2007 2012 Median 2007 Median 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Measuring Poverty in Iraq Iraq uses an “absolute” poverty line, which fixes a up a food poverty line and a non-food poverty line welfare threshold and is based on the Cost of Basic (For more details, see Annex Chapter 2). Needs approach (CBN). The CBN approach as ap- plied in Iraq defines the poverty line as the level of The food poverty line in Iraq was fixed at a level expenditure that allows the households to spend just equivalent to the expenditures needed to meet a enough on food to meet a certain caloric threshold, minimal nutritional intake of 2337 calories per and just enough to meet basic non-food needs. The person per day, or ID 50,473.26 per person per total poverty line is therefore calculated by adding month in 2012. In order to better account for the increasingly important differences in consump- tion expenditure across space in terms of non-food items—for instance, clothing and shelter—we allow FIGURE 10: Cumulative Distribution – Welfare the non-food allowances to vary by three regions in Aggregate, 2007 and 2012 Iraq—Baghdad, Kurdistan and the rest of Iraq. This Poverty Incidence Curve implies that for a given national food poverty line, 1.0 for each region, the corresponding non-food allow- ances are defined according to the distribution of 0.8 Cumulative distribution consumption within that particular region. 0.6 The official poverty line in Iraq is defined at the na- 0.4 tional level, i.e., the non-food allowance is defined 0.2 in accordance to the national non-food consump- tion patterns and distribution. In contrast to using 0 0 500 1000 a single national non-food allowance, in this report, Welfare aggregate and as agreed with the government, we allow for re- 2007 2012 gional variation in defining the non-food allowances, so that in effect, we use three regional poverty lines. Source: Authors’ calculations, IHSES 2007 and 2012. As a result, and as we show in the next section, the 22 THE UNFULFILLED PROMISE OF OIL AND GROWTH primary implication of adopting regional poverty fourth consumption decile may still be counted as lines as opposed to a single national line is that the poor if their per capita real consumption expenditure levels of poverty in Baghdad, and to a larger extent, fell below the cost of basic needs in the region where Kurdistan, are higher; while there is little difference they live. Similarly, an individual belonging to the for the 14 governorates comprising the rest of Iraq. 41st percentile may be classified as poor because they In addition, the choice of national or regional pov- live in Baghdad, whereas an individual living in the erty lines does not alter the trends in poverty at the rest of Iraq belonging to the 40th percentile may not. national, regional or governorate level; and for the In order to make appropriate comparisons across the parts of the country significant changes in poverty consumption distribution, therefore, we use, where were experienced, the magnitude of these changes is appropriate, an adjusted consumption aggregate, similar irrespective of the approach adopted. which rescales the welfare aggregate in each of the three regions, so that they are comparable under a Table 2 shows the resulting food and total poverty single poverty threshold (which is a weighted aver- lines, using the regional as well as the national ap- age of the three regional poverty lines).14 Thereafter, proach. Both regional and national poverty lines and throughout the rest of the report, the analysis include the same food poverty line. However, al- uses only regional poverty lines, rather than the of- lowing for regional variation in the cost of basic ficial lines as they better account for the important non-food items implies higher poverty lines for spatial differences in basic needs and welfare in Iraq. Baghdad and especially for Kurdistan, relative to the national non-food allowance, while there is little difference for the Rest of Iraq. Poverty and Shared Prosperity in Iraq: 2007–2012 In the next section, we explore levels and trends in poverty, and show that over and above level differ- The improvement in the welfare distribution is re- ences in headcount rates for Kurdistan and Bagh- flected in the decline in poverty over the 2007 to dad, both approaches yield similar findings in pat- 2012 period. Overall, headcount poverty rates, as terns and trends of poverty. One consequence of the measured using the regional poverty lines, fell from use of regional poverty lines is that the distribution 23.6 percent in 2007 to 19.8 percent in 2012, a 3.8 of the poor across the consumption distribution percentage point decline. A similar trend is evident is no longer equivalent to the bottom 2 deciles of using the official poverty line, which records a de- the consumption distribution, or strictly speaking, cline in headcount rates from 22.4 percent in 2007 to the bottom 19.8 percent of the consumption dis- 18.9 percent in 2012, a 3.5 percentage point decline. tribution. This is because poverty when using the regional lines assumes a different threshold for each In rural Iraq, poverty as measured by the regional region, so that, for instance, someone in the third or lines declined by 8 percentage points, as compared with a much smaller decrease of 2.5 percentage points in urban areas. Given the presence of universal TABLE 2: Poverty Lines (ID Per Person Per food subsidies delivered through the Public Distribu- Month) tion System (PDS), the low rates of food poverty are unsurprising (Table 3), although there has been little 2007 2012 change in these rates, perhaps because of a decline Food poverty line 35796.64 50473.26 Kurdistan poverty line 101000.50 142410.70 Baghdad poverty line 82223.19 115934.70 14 The adjusted consumption aggregate is used when compar- isons across the consumption distribution are being made, Rest of Iraq poverty line 72110.57 101675.90 for instance, in the analysis of inequality and consumption National (Official) poverty line 74822.98 105500.40 growth across different parts of the distribution. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 23 TABLE 3: Overall Poverty (Regional Poverty Lines) Poverty Headcount Rate Poverty Gap Squared Poverty Gap 2007 2012 Change 2007 2012 Change 2007 2012 Change Total poverty line Urban 17.4 14.8 –2.5 3.0 2.7 –0.2 0.8 0.8 0.0 Rural 38.9 30.6 –8.3 9.1 7.4 –1.7 3.1 2.6 –0.6 Total 23.6 19.8 –3.7 4.7 4.2 –0.5 1.5 1.3 –0.1 Food poverty line Urban 0.3 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Rural 2.9 2.2 –0.6 0.4 0.3 –0.1 0.1 0.1 0.0 Total 1.0 0.9 –0.1 0.1 0.1 0.0 0.0 0.0 0.0 Source: Authors’ calculations, IHSES 2007 and 2012. in the number of items delivered through the PDS changes in the poverty line can lead to large chang- between 2007 and 2012. Other measures of poverty, es in headcount rates (Table 4). In other words, a such as the poverty gap and the squared poverty gap, small increase or decrease in incomes and consump- did not change substantially over the 2007–2012 pe- tion can lead to large changes in the incidence of riod (Figure 12). The poverty gap, which measures poverty. For instance, a five percent increase in the the average shortfall between the consumption of poverty line in 2012 would raise poverty by 16 per- the poor and the poverty line, relative to the poverty cent, while a ten percent increase would raise pov- line fell by only half a percentage point. The squared erty by more than 30 percent. poverty gap, which is an average of the square of all consumption shortfalls, barely changed. Spatially Uneven Poverty Reduction Poverty is shallow in Iraq: consumption is dense- Although poverty has declined over the five year pe- ly concentrated near the poverty line, and small riod, poverty reduction has been spatially uneven. FIGURE 12: P  overty Headcount, Gap and TABLE 4: Sensitivity of Headcount Poverty Severity – Iraq: 2007–2012 Rate with Respect to the Choice of Poverty Line 60% 2007 2012 50% 40% Poverty Change Poverty Change Headcount from actual Headcount from 30% 23.57% Rate (%) Rate actual (%) 19.82% 20% Actual 23.6 0.0 19.8 0.0 10% 4.72% 4.20% 1.34% +5% 26.6 12.6 23.0 16.0 1.48% 0% +10% 30.8 30.8 26.1 31.5 –10% +20% 38.5 63.1 32.6 64.6 2007 2012 –5% 19.6 –16.9 17.1 –13.7 Poverty headcount Poverty gap Severity –10% 15.9 –32.6 14.2 –28.5 Source: Authors’ calculations, IHSES 2007 and 2012. –20% 10.0 –57.7 9.3 –53.3 24 THE UNFULFILLED PROMISE OF OIL AND GROWTH Using the official (single national) poverty line, poverty in both regions. Thus, for the most part, poverty in Baghdad shows little discernible change significant changes in poverty were experienced in (12.6 percent in 2007 and 12 percent in 2012); those governorates in Iraq where national and re- while a small decline is recorded in Kurdistan (from gional poverty lines are very similar. 4.3 percent in 2007 to 3.5 percent in 2012). In the rest of Iraq, official headcount rates fell from Within the RoI, the poverty reduction record was very 29.7 percent in 2007 to 24.4 percent in 2012. mixed. In the governorates south of Kurdistan and north of and around Baghdad—Diyala, Anbar, Baby- The same trend is evident when the regional lines lon, Kerbala, Salahaddin, Najaf—poverty declined are used. In Baghdad, by far the most populous gov- substantially, with the sole exception of Nineveh. In ernorate in the nation, poverty did not change sig- Nineveh and four of the southern governorates—Qa- nificantly; and in the Kurdistan region, poverty de- disiya, Muthanna, Thi-Qar and Missan, poverty in- clined, albeit at a small rate (Figure 13 and Table 5). creased significantly (Table A 2.1). Basra is the only In contrast, the rest of Iraq (RoI), comprising the 14 southern governorate where poverty fell. Table A 2.2 other governorates, registered a 5 percent decline in shows estimates and trends of poverty at the gover- headcount rates. Recall that the national and region- norate using the official national poverty line. al poverty lines are very similar for the rest of Iraq; and consequently the levels and trends are almost This pattern of poverty reduction has been accom- identical irrespective of the line. For Baghdad and panied by a greater spatial concentration of poverty. Kurdistan, the use of a regional non-food allowance In 2007, half of Iraq’s poor lived in five governor- takes into account higher expenditures required to ates—Baghdad, Basra, Nineveh, Babylon and Thi- fulfil basic non-food needs, and these higher pov- Qar. By 2012, while Baghdad’s share of the poor re- erty lines imply that the levels of poverty calculated mained unchanged at around 19 percent, Nineveh using the regional lines are higher. However, using almost doubled its share in the poor to 15.7 per- either approach, there is little perceptible change in cent. Three southern governorates, Thi-Qar, Qadis- iya and Missan, with 10 percent, 7 percent and 6.7 percent of the poor respectively, now account for almost a quarter of the country’s poor. In 2012, 58 FIGURE 13: Regional Poverty Headcount percent of Iraq’s poor lived in these five governor- Rates, 2007–2012 ates, compared to 40 percent in 2007. 0.30 Although the rates of extreme poverty (the share of 0.28 the population living on less than USD 1.25 a day, 0.26 0.24 2005 PPP) are low and have barely changed, hover- 0.22 ing around 4 percent in 2007 and 2012, the spatial 0.20 distribution of the extreme poor has altered con- 0.18 0.16 siderably. In 2007, Nineveh, Diyala and Salahaddin 0.14 together accounted for almost 35 percent of the ex- 0.12 treme poor, with another 9 percent living in Muthan- 0.10 2007 2012 2007 2012 2007 2012 2007 2012 na. In 2012, Nineveh’s share in the total population Iraq Kurdistan Baghdad Rest of Iraq of Iraqis living below the $1.25 a day line increased Upper Bound 0.25 0.21 0.15 0.13 0.22 0.21 0.29 0.23 to 18 percent, while Qadisiya, Thi Qar, Muthanna Lower Bound 0.22 0.19 0.12 0.11 0.16 0.15 0.26 0.21 and Missan together accounted for half of the ex- FGT 0: 0.24 0.20 0.14 0.12 0.19 0.18 0.27 0.22 treme poor. Thus, the increase in poverty in Nineveh Headcount Rates and the four southern governorates was accompanied Source: Authors’ calculations, IHSES 2007 and 2012. by an increasing concentration of extreme poverty. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 25 While some of these patterns and trends are due to Breaking down poverty rates within regions by urban increases in poverty within governorates, some part and rural areas, it becomes apparent that rural parts of the explanation also lies in changes in the distri- of Kurdistan and the RoI made significant gains in bution of the population. Despite its relatively lower poverty reduction, relative to urban areas in the two poverty headcount rates, Baghdad contributes to a regions, which saw small changes in headcount rates. large share of Iraq’s poor because it alone accounts In rural Kurdistan, which accounts for 20 percent for large share of the nation’s population (Table 5). of Kurdistan’s population, poverty fell by 9 percent- More than one in five Iraqis live in Baghdad, al- age points, from 32 percent to 23 percent. Poverty though in 2012, there appears to have been little in rural parts of the rest of Iraq, where 40 percent population growth in Baghdad, which is reflected in of the region’s population lives, fell by 10 percent- a decline in the share of the population and the poor age points. In contrast, in Baghdad, rural poverty in- in urban areas relative to rural areas. In the three creased by 15 percentage points, doubling to 33 per- governorates of the Kurdistan region—Duhouk, cent by 2012. At the same time, the share of the rural Suleimaniya and Erbil, small increases in the share population in Baghdad almost doubled to 13 percent. of the poor have been accompanied by large increas- es in population in these governorates over the last Within the RoI, in the governorates where poverty five years. These changes in population are likely a increased, by and large, rural poverty rose faster result of population growth and voluntary and in- than urban poverty. In contrast, in the governorates voluntary flows, a combination of return migration where poverty fell, rural areas witnessed larger re- in response to improved local conditions and dis- ductions in poverty. Of the five governorates where placement as a result of deteriorating security and headcount rates increased—Qadisiya, Thi Qar, Mis- economic conditions in other parts of Iraq. In the san, Muthanna and Nineveh—in four, rural poverty absence of census data, it is difficult to disentangle increased at a higher rate than urban poverty over these flows, yet it is notable that survey estimates the 2007–2012 period. The exception was Muth- suggest Kurdistan added 1 million additional per- anna, where increases in poverty came entirely from sons between 2007 and 2012. urban areas. Barring Nineveh, rural areas account TABLE 5: Population and Poor Population, by Region, 2007–2012 Population Distribution of the Population Poor population Distribution of the Poor All Iraq 2007 29,752,018 100% 7,013,294 100.0% 2012 34,043,890 100% 6,748,588 100.0% Kurdistan 2007 3,839,102 13% 528,656 7.5% 2012 4,732,818 14% 584,394 8.7% Baghdad 2007 6,971,005 23% 1,345,808 19.2% 2012 7,213,046 21% 1,301,363 19.3% Rest of Iraq 2007 18,941,911 64% 5,138,751 73.3% 2012 22,098,026 65% 4,862,825 72.1% Source: Authors’ calculations, IHSES 2007 and 2012. 26 THE UNFULFILLED PROMISE OF OIL AND GROWTH for more than half the population in these gover- quintiles over time: if the growth rates of the lower norates. Five governorates rapidly reduced poverty, quintiles are higher than those of the upper, con- at rates of 14 percentage points or more—Basra, sumption growth was pro-poor. This is not the case Salahaddin, Diyala, Babylon and Kerbala. In each of in Iraq. these, rural areas recorded faster rates of decline in headcount rates than urban areas. With the exception of Kurdistan, consumption per capita grew faster for the well-off than for the less- well off, as is evident in the growth-incidence curves Unequal Consumption Growth and Shared for Iraq as a whole and for the three regions (Figure Prosperity 14). The same pattern is also evident in urban and Between 2007 and 2012, consumption grew faster rural Iraq. In contrast, the growth-incidence curves for Iraq’s relatively better off, amongst the high- for Kurdistan are relatively flat, indicating that con- est quintiles. But it also grew where consumption sumption grew evenly across the distribution, albeit levels were lower to start with: in rural Iraq and in not at a high positive rate. the RoI. While the consumption Gini coefficient for Iraq is relatively low and has increased by almost Overall, poverty reduction in Iraq over the 2007 to 3 percent over this period, the ratio between the 2012 period was driven by lower headcount rates consumption of the 90th percentile and the 10th per- in rural areas and in the Rest of Iraq, and primarily centile increased at a higher rate (Table 6). In rural explained by the growth in consumption. While re- areas, where consumption grew the fastest, the lat- distributional effects were relatively smaller, changes ter ratio has increased by 12 percent. in inequality hampered poverty reduction. If there had been no change in the distribution of consump- Increasing inequality is also evident in growth-in- tion relative to 2007, national poverty would have cidence curves, which graph the growth rate of per declined by 6 percentage points, rural poverty by capita consumption expenditure for each percentile 12.21 percentage points and poverty in the RoI by of the population. For a given percentile, the height 8 percentage points (Figure 15). of the curve represents the growth in per capita expenditure for that percentile of the population. The preceding insights of increasing inequality and These curves assess how incomes change across slower consumption growth for the lower quintiles TABLE 6: Inequality in Per Capita Expenditure Distribution by Urban and Rural Areas Bottom Half of the Distribution Upper Half of the Distribution Interquartile Range Tails p25/p10 p50/p25 p75/p50 p90/p75 p75/p25 p90/p10 Gini Total 2007 1.28 1.33 1.36 1.36 1.81 3.15 26.49 2012 1.32 1.37 1.39 1.38 1.90 3.46 27.94 Urban 2007 1.25 1.33 1.35 1.36 1.79 3.04 26.02 2012 1.30 1.36 1.38 1.36 1.87 3.33 27.39 Rural 2007 1.28 1.33 1.31 1.33 1.75 2.97 24.23 2012 1.33 1.38 1.37 1.34 1.89 3.34 27.00 Source: Authors’ calculations, IHSES 2007 and 2012. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 27 Growth Incidence Curves – National, Rural-Urban, Divisional FIGURE 14:  Iraq: Pro-rich Kurdistan: Neutral 6 6 5 4 Annual growth rate (%) Annual growth rate (%) 4 3 2 2 0 1 –2 0 –1 –4 –2 –6 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Per capita expenditure percentiles Per capita expenditure percentiles Baghdad: Pro-rich Rest of Iraq: Pro-rich 6 6 5 5 Annual growth rate (%) Annual growth rate (%) 4 4 3 3 2 2 1 1 0 0 –1 –1 –2 –2 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Per capita expenditure percentiles Per capita expenditure percentiles Urban: Pro-rich Rural: Pro-rich 6 6 5 5 Annual growth rate (%) Annual growth rate (%) 4 4 3 3 2 2 1 1 0 0 –1 –1 –2 –2 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Per capita expenditure percentiles Per capita expenditure percentiles Upper 95% con dence bound/ Median spline Lower 95% con dence bound Source: Authors’ calculations, IHSES 2007 and 2012. can also be expressed in terms of ‘shared prosperi- The average consumption (in thousands of Iraqi ty’. This measure tracks the consumption or income dinar and in 2005 purchasing power parity adjust- growth of the bottom 40 percent of the population ed Iraqi dinar) of the bottom 40 percent of the relative to the consumption or income growth of consumption distribution and of the population as the entire population. a whole in 2007 and 2012 is shown in Table 7. 28 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 15: The Contribution of Growth and Redistribution to Poverty Reduction All Iraq Rest of Iraq 4% 4% 2% 2% 3.2% 2.3% 0% 0% –2% –2% –6.1% –4% –8.3% –4% –6% –3.7% –5.1% –6% –8% –8% –10% 2007 vs 2012 2007 vs 2012 Growth Redistribution Total change in poverty Source: Authors’ calculations, IHSES 2007 and 2012. While consumption of the population as a whole for close to 10 percent (Figure 16). However, a sig- grew at an average annual rate of 1.965 percent per nificant share of each region’s population belongs to annum, consumption for those in the bottom 4 de- the national bottom 40 percent. 43 percent of the ciles of the population grew slower, at 1.113 per- population of the rest of Iraq belongs to the bottom cent per year.15 40 percent, as does 39 percent of Baghdad’s popu- lation and 29 percent of Kurdistan’s. Treating the per capita consumption of the 40th per- centile as a relative poverty line, the population of Iraq can be divided into two categories: the bottom Who are the Poor? A Profile of 40 and the top 60. Whereas nationally, the consump- Poverty, 2007 to 2012 tion of the total population grew by 9 percent over the 2007–2012 period, the consumption of the bot- Poverty in Iraq is significantly higher among larger tom 40 grew by a cumulative 5 percent, less than half households, those with less educated heads, and var- the rate of growth of consumption of the top 60. ies by the employment sector of the head of house- hold. Table A 2.3 summarizes the mean characteris- Where do the bottom 40 live? 70 percent of the tics of poor and non-poor households in 2007 and population who are in the bottom 40 percent of dis- 2012.16 In 2012, a typical poor Iraqi household had tribution live in the rest of Iraq, with Baghdad ac- counting for 21 percent, and Kurdistan accounting 15 If the national/official poverty line were used, it is consis- tent with the unscaled welfare aggregate, and the shared prosperity indicator is very similar, with the bottom 40 TABLE 7: Shared Prosperity – Annual growing by 1.201 percent per annum, while overall con- Consumption Growth of the Bottom 40 sumption was 1.876 percent per year. These calculations as- sume a cumulative inflation rate between 2007 and 2012 of Iraqi dinars (thousands, 2012 terms) 40.1 percent, based on the official CPI series. Year Bottom 40 Overall 16 There is a significant difference between the non-poor and the poor in all characteristics except for the number 2007 101.83 171.37 of household members who lived elsewhere for at least six 2012 107.63 188.88 months in 2007 and the proportion of household heads Source: Authors’ calculations, IHSES 2007 and 2012. employed in electricity, gas and water supply in 2012. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 29 FIGURE 16: Shared Prosperity and the FIGURE 18: Poverty Headcount Rates by Distribution of the Bottom 40 Household Size, 2007 and 2012 Percent 0.6 100 0.5 90 80 0.4 70 60 0.3 Percentage 50 0.2 40 30 0.1 20 10 0 1 3 5 7 9 11 13 15 17 0 2007 2012 2007 2012 2007 2012 Contribution to (% of) % region's population national bottom 40 in national bottom 40 Source: Authors’ calculations, IHSES 2007 and 2012. Kurdistan 8.10 9.95 25.10 28.63 Baghdad 21.24 20.69 36.27 39.06 Rest of Iraq 70.66 69.36 44.39 42.74 Total 100.00 100.00 40.00 40.00 had almost 8 members, with approximately 4 de- Source: Authors’ calculations, IHSES 2007 and 2012. pendents (including children and the elderly), and likely lived in urban areas. Among non-poor house- holds with employed heads, almost half were em- 11 members, almost 6 children, and was equally ployed in commerce and retail; transport and stor- likely to live in urban or rural areas. 52 percent of age; financial, insurance and professional services; or poor households’ heads had less than primary edu- public administration, health and education. Almost cation. Agriculture, construction and transport and 40 percent of heads of non-poor households had storage constituted the three most likely sectors of intermediate education or higher. employment for the head of household, accounting for almost half of all employed heads of poor house- Poverty headcount rates are significantly higher for holds. A typical non-poor Iraqi household in 2012 larger households (Figure 18). More than 90 per- cent of Iraq’s population belongs to households of 14 or fewer members, with median household sizes of between 7 to 8 members (Figure 17). A majority FIGURE 17: Population Share by Size of of the poor belong to large households: less than Household 1 percent of the poor belonged to households with 0.16 4 or fewer members. In 2007, 78 percent of the 0.14 poor belonged to households with 8 or more mem- 0.12 bers, with about half of them being from families 0.10 with more than 11 members. A similar pattern is 0.08 apparent in 2012, where three-quarters of the poor 0.06 belong to families with 8 or more members and 0.04 almost a third belong to families with more than 0.02 11 members. Poverty in fact increases steeply with 0 1 4 7 10 13 16 19 22 25 28 household size, from 4 percent among households 2007 2012 with 4 or fewer members to around 40 percent among households with 13 or more members (Fig- Source: Authors’ calculations, IHSES 2007 and 2012. ure 18). Between 2007 and 2012, poverty among 30 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 19: Poverty Headcount Rates for FIGURE 20: Poverty Headcount Rates Female and Male Headed by Education of the Head of Households Household 0.3 0.35 0.30 0.2 0.25 0.20 0.1 0.15 0.0 0.10 2007 2012 0.05 Female head Male head 0.00 Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary very large households with 13 or more members, who made up 13 percent of the population in 2012, 2007 2012 appears to have come down. Source: Authors’ calculations, IHSES 2007 and 2012. A vast majority of Iraqi households, and 92 percent of poor Iraqi households (Table A 2.3), were headed by males in 2007 and 2012. In 2012, female headed 2007 and 2012, there has been a secular decline in households faced poverty rates of 16 percent, rela- headcount rates among all education levels of house- tive to 20 percent among male headed households hold heads (Figure 20). While this is heartening, the (Figure 19). While poverty rates have come down share of the poor who belong to households whose for both male and female headed households be- heads have primary education or less has increased: tween 2007 and 2012, the decline has been larger these households account for more than 80 percent for female headed households. of the poor in 2012 and face poverty rates upwards of 20 percent (Figure 20 and Figure 21). More than One possible explanation for the decline in poverty rates among very large households and households headed by women is the Government of Iraq’s social 17 Iraqi magazine for research on markets and social protec- protection scheme. At the beginning of the 2003 war tion 2009, University of Baghdad, Social Protection Net- in Iraq, the government of Iraq announced the inclu- works in Iraq and the effect on consumer protection, See page 116 for the numbers of families benefiting from the social sion of households with unemployed heads in the so- protection scheme based on region http://www.iasj.net/ cial protection scheme. Starting in January of 2005, iasj?func=fulltext&aId=1782. the government started distributing monthly grants 18 Ministry of Planning, 2008 http://cosit.gov.iq/ to beneficiaries, based on the number of family mem- documents%5Cstatistics_ar%5Cpoverty%5Cstrategy%5C bers. The social protection scheme also included Background%20papers/%D8%AF%D9%88%D8%B1%20 widows and divorced women, among others. Grants %D8%B4%D8%A8%D9%83%D8%A9%20%D8%A7 through the social protection are determined based %D9%84%D8%AD%D9%85%D8%A7%D9%8A%D on the size of the family, and increased with family 8%A9%20%D8%A7%D9%84%D8%A7%D8%AC% D8%AA%D9%85%D8%A7%D8%B9%D9%8A%D8 size.17 All the grants were raised by 25% in 2006.18 %A9%20%D9%81%D9%8A%20%D8%A7%D9%84 %D8%AA%D8%AE%D9%81%D9%8A%D9%81%20 The relationship between poverty and the education %D9%85%D9%86%20%D8%A7%D9%84%D9%81- of the household head is striking in Iraq. Between %D9%82%D8%B1.pdf. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 31 FIGURE 21: Share of the Poor, by Education of FIGURE 22: Poverty Headcount Rates by the Head of Household Labor Force Status of the Head of Household* 100% 90% 0.35 80% 0.30 70% 0.25 60% 50% 0.20 40% 0.15 30% 0.10 20% 0.05 10% 0% 0.00 Full time employed Part time employed Unemployed Out of the labor force 2007 2012 Tertiary Higher secondary Secondary Intermediate Complete primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2007 and 2012. 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Note: * Labor force outcomes are based on 7 day recall, ILO definitions. half of poor households in 2012 had heads with less than primary education, among whom poverty rates are more than 25 percent. In contrast, in 2012 those whose heads have tertiary or higher secondary edu- FIGURE 23: Share of the Poor, by Labor Force cation face poverty rates of between 5 to 9 percent, Status of the Head of Household and account for less than 5 percent of the poor. 100% Poverty does not vary as starkly by the labor force 90% status of the household head, but rather by the sector 80% 70% of employment. As may be expected, poverty is lower 60% in households where the head is employed, between 50% 17 and 18 percent in 2007 and 2012, as compared to 40% 20 percent among those where the household head 30% is out of the labor force (Figure 22). Headcount 20% rates are almost 30 percent among households where 10% 0% the head is unemployed, although these households 2007 2012 comprise only 1 to 2 percent of all households (Fig- Out of the labor force Unemployed ure 22 and Figure 23). Surprisingly, full time employ- Part time employed Full time employed ment is not associated with much lower incidence of poverty; and in 2012, more than 42 percent of poor Source: Authors’ calculations, IHSES 2007 and 2012. households had heads who were employed full time; an increase of 7 percentage points since 2007. poverty is 33 percent, higher than among house- Figure 24 plots headcount rates by the household holds with heads who are not employed. This is head’s sector of employment, based on an annual a worrying pattern, as these two sectors alone ac- reference period. Among households whose heads count for 24 percent of all poor households, while are employed in agriculture or in construction, non-employment accounts for almost another 30 32 THE UNFULFILLED PROMISE OF OIL AND GROWTH  overty Headcount Rates by Employment Sector of the Head of Household FIGURE 24: P 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 Not employed Agriculture & nishing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & communication Financial, insurance & professional Public administration, health & education Others services 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. percent (Figure 25). In contrast, the sector with the construction, transport,storage and communication lowest incidence of poverty, mining and quarrying, and commerce and retail. In addition, employment accounts for less than 1 percent of the poor. in agriculture and commerce predominantly take the In 2012, 5 sectors of employment of the household head had poverty headcount rates of 15 percent or FIGURE 25: Share of the Poor by Employment less—Mining, Manufacturing, Commerce, Finance Sector of Head of Household and Public Administration—and employed less than 30 percent of heads of poor households. In con- 100% trast, the 5 sectors with poverty rates of 20 percent 90% 80% or more included Agriculture, Construction, Utili- 70% ties, Transport, and non-employment, and account- 60% ed for almost 70 percent of poor households. 50% 40% It should be noted that while almost all jobs in the 30% public administration sector are public sector jobs, 20% not all public sector jobs are in the public administra- 10% 0% tion sector. The vast majority of the jobs in the min- 2007 2012 ing and quarrying sector, and in the utilities (electic- Not employed Mining & quarrying ity, gas and water) sector are public sector jobs. In Agriculture & shing Manufacturing addition, there has been a significant increase in the Utilities Construction share of public sector jobs in the financial, insurance Commerce and retail Transport, storage Financial, insurance & communication and professional services sector, from 34 percent & professional Public administration, in 2007 to 65 percent in 2012 (Figure 26). Thus Other services health & education private sector activity in terms of employment in Iraq is concentrated in agriculture, manufacturing, Source: Authors’ calculations, IHSES 2007 and 2012. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 33 FIGURE 26: Public Sector Employment as a Share of Employment in Each Sector, 2007–2012 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Agriculture & nishing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & communication Financial, insurance & professional Public administration, health & education Others services 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. form of self-employment rather than wage work, many of the employment sectors that are associated and about half of the private sector jobs in finan- with lower poverty rates and account for smaller cial, insurance and professional services and in the shares of the poor are dominated by the public sec- transport, storage and communications sector con- tor. In contrast, the private sector (and in particular, sist of self-employment rather than wage employ- agriculture and construction), the engine of growth ment (Figure 27). Taken together these suggest that and job creation in a healthy economy, seems to be FIGURE 27: Self-Employment in the Private Sector as a Share of All Employment in the Private Sector 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Agriculture & nishing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & communication Financial, insurance & professional Public administration, health & education Others services 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. 34 THE UNFULFILLED PROMISE OF OIL AND GROWTH comprised of mostly self-employment opportunities Over the five year period, the correlation between (which may indeed be subsistence activities) rather consumption and location has altered in important than salaried wage employment. ways. For one, the size of the correlation between living in an urban area and higher consumption has halved. For another, the effect of living in Nineveh Correlates of Consumption Expenditure has reversed, from a relative advantage to a signifi- and Poverty cant disadvantage, while the three governorates of In this section, we explore how different factors Kurdistan, Kirkuk, and Najaf continue to be asso- come together to explain consumption expenditure, ciated with higher consumption relative to Bagh- and attempt to measure the effect of each factor dad in both years. In addition, households living in while holding others constant. We also identify their Anbar, Babylon, Salahadin and Basra tend to have role in determining whether a household is poor higher consumption in 2012 compared to similar (consumption below the poverty line) or belongs households in Baghdad. In both years, certain gov- to the bottom 40 percent (consumption below the ernorates were associated with lower consumption consumption of the 4th consumption decile). The relative to Baghdad after controlling for the effect advantage of the first approach is that it allows us of household characteristics on consumption, in- to use information across the consumption distribu- cluding Diyala, Kerbala, Wasit, Qadisiya, Thi Qar, tion although it assumes that the influence of each Muthanna and Missan. factor is linear.19 The latter approach, where the outcome is whether the household is poor or not, In Table A 2.5, we present the results of a probit estimates the effect of each factor in determining regression of the factors that predict poverty at the whether household consumption is sufficiently low, level of the household, including location, house- i.e., below the poverty line; rather than the overall hold demographics, education and work status of relationship with consumption. the head of household, and migration status. Liv- ing in an urban area reduced the likelihood of being Table A 2.4 presents the results of a regression of poor by 11 percent in 2007 and 5 percent in 2012. log per capita real consumption expenditure on a Household demographics—household size, the range of household and household head characteris- number of children and elderly—are all correlated tics, as well as location variables. In 2007, consump- with poverty in 2007 and in 2012: larger house- tion was lower for larger households and with more holds and households with higher dependency ra- children and higher for households with elderly per- tios are more likely to be poor. sons (likely the effect of pension income). It was also positively correlated with the number of employed Measures of employment status are also correlated working age males. Education was strongly associ- with poverty. An additional employed working age ated with higher consumption. Almost all sectors of male reduces the probability of poverty by about employment were correlated with higher per capita 2 percent. In addition, certain sectors of employment consumption (relative to households with non-em- (relative to being unemployed or out of the labor ployed heads), with the exception of construction, force) are correlated with lower odds of poverty—in which was associated with lower consumption. The 2007, households with heads working in electricity, same relationship with household size and composi- gas and water supply (utilities) and public adminis- tion, education, and employment remains in 2012, tration, health and education were 8 and 4 percent except that households with heads employed in agri- culture were not significantly different from house- 19 Ravallion (1996) points out that the reason for which level holds whose heads were not employed in terms of regression should be preferred is that it depends on weaker their predicted per capita consumption, when all assumption about the error term than the binary model of other factors were controlled for. being poor or not. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 35 less likely to be poor (Table A 2.5). In 2012, almost declined to 7 percent. Household size and depen- all sectors of employment except electricity, gas and dency are strong correlates with belonging to the water supply and agriculture lowered the odds of bottom 4 deciles: in 2012, each additional house- poverty. However, households with heads working in hold member increased the probability of being in construction were 7 percent more likely to be poor. the bottom 40 by 8.6 percent, each child by ap- proximately 6 percent. Belonging to a male headed Education of the head of household is strongly cor- household increases this probability by 9 percent related with poverty: at higher levels of education, in 2012. the odds of poverty fall, at an increasing rate. For instance, relative to a household with an illiterate Measures of employment and labor force are also head of household, households with primary edu- very important. Each additional employed working cated heads face 7 to 8 percent lower probability of age male lowers the risk of being in the bottom 40 being poor, while those with higher secondary and by 4 percent. All employment sectors lower or do tertiary educated heads face on average more than not alter the odds of being in the bottom 40 relative 16 percent lower odds of being poor. to the household head being unemployed or out of the labor force, with the exception of construction, How do the poor compare to the bottom 40 per- which increased the risk of being in the bottom 40 by cent of the consumption distribution? By definition, 8 percent in 2012. Employment for the household in Iraq, all the poor belong to the bottom 40, but head in public administration, finance, and mining the reverse is not true. In line with the relative con- sectors, which are mostly public sector jobs, have centration of people around the poverty line, the large impacts, lowering the odds of being in the bot- average characteristics of the bottom 4 deciles are tom 40 by 11, 9, and 16 percent respectively in 2012. remarkably similar to those of the poor. Table A 2.6 Finally, higher education for the head of household compares the characteristics, on average of the poor, starkly lowers the likelihood of being in the bottom those in the bottom 40 percent, and the top 60 per- 4 deciles of the consumption distribution. cent in 2007 and 2012. Just like the poor, the aver- age household in the bottom 40 percent is almost equally likely to live in an urban or rural area and has Poverty Across Space more than 10 members, with almost 6 dependents. 4 out of 5 heads of bottom 40 households have pri- In order to better understand the spatial dimensions mary or less education, and two-fifths are employed of poverty, the analysis that follows will further sub- in agriculture, construction, commerce and trans- divide the rest of Iraq into three parts, yielding five port (predominantly private sector jobs). In con- divisions of Iraq (Map 1), of relatively equal popula- trast, almost 80 percent of top 60 households are tion size (Table 8): urban, with a typical household size of 7 members, almost 4 of which are dependents. 42 percent of top 1. Kurdistan comprising the three governorates of 60 households have heads with more than primary the Kurdistan Regional governorate, Duhouk, education, and almost two-fifths are employed in Erbil, and Sulaimaniya, making up around 15 public administration, finance or commerce. percent of the Iraqi population 2. North comprising of the three governorates di- Probit analysis of the characteristics that predict rectly south of Kurdistan and to the North of being in the bottom 40 confirm these findings Baghdad—Nineveh, Kirkuk, and Salahadin, ac- (Table A 2.7). In 2007, living in an urban area counting for 18 percent of the population reduced the odds of being in the bottom 40 by 3. Baghdad comprising of the single governorate approximately 13 percent, a relation that contin- of Baghdad, the capital city, making up a fifth of ues to hold in 2012, although the coefficient had the Iraqi population 36 THE UNFULFILLED PROMISE OF OIL AND GROWTH 4. Central comprising of the governorates to the This divisional break-up shows the stark differenc- east, west and immediately south of Baghdad es in welfare improvements within the 14 gover- —Anbar, Diyala, Najaf, Karbala, Wasit, and norates that make up the Rest of Iraq. Figure 28 Babylon—accounting for a quarter of Iraq’s plots poverty head count rates in each division in population 2007 and 2012, and the changes in poverty over 5. South comprising the five southern most gover- the period (in percentage points). It is evident norates of Iraq—Qadisiya, Thi Qar, Muthanna, that poverty reduction was concentrated entirely Missan, and Basra—making up almost 22 per- in the Central division, where headcount rates fell cent of the total population. by 14 percentage points between 2007 and 2012. In contrast, the South was the only division where During the period from 2007 to 2012, three di- headcount rates increased, albeit slightly, over the visions witnessed larger than average population 2007–2012 period. The Central division, as a re- growth (including natural population growth as well sult, was the only division to witness a decline in its as migration)—Kurdistan, the North and the Cen- share of the poor, by 12 percentage points, while tral division. On the other hand, Baghdad’s popu- the South’s share of the poor increased by 6 per- lation barely grew, and population growth in the centage points, and the North’s by 4 percentage South was below the national average of 14 percent. points, a rate significantly higher than the increase MAP 1: Five Divisions of Iraq 'XKRN (UELO 1DLQDZD 6XODLPDQL\DK .LUNXN 6DODKDO'HHQ 'L\DOD %DJKGDG $QEDU :DVLW .HUEHOD %DE\ORQ 4DGLVL\D 0LVVDQ 1DMDI 7KL4DU 'LYLVLRQV %DJKGDG %DVUDK 0XWKDQQD &HQWUDO .XUGLVWDQ 1RUWK 6RXWK ʹͲͳͶǦͳͲ Poverty, Shared Prosperity and Subjective Well-Being in Iraq 37 TABLE 8: Survey Based Estimates of FIGURE 29: Distribution of the Poor by Population by Division* Divisions, 2007–2012 % Change 0.4 (cumulative): 0.33 0.30 2012 relative % of 2012 0.27 0.3 2007 2012 to 2007 population 0.20 0.19 0.19 0.18 Kurdistan 3,838,437 4,728,838 23.2 14.0 0.2 0.16 Baghdad 6,961,071 7,193,415 3.3 21.2 0.09 0.08 0.1 0.06 North 5,049,876 6,128,938 21.4 18.1 0.04 0.01 0.00 Centre 7,247,272 8,515,574 17.5 25.1 0 South 6,526,511 7,300,681 11.9 21.6 Total 29,623,167 33,867,446 14.3 100.0 –0.1 –0.12 Source: Authors’ calculations, IHSES 2007 and 2012. Note: * While no recent population estimates bases on census data is available, survey –0.2 based estimates of population provide some indication of the size and share of each Kurdistan Baghdad North Centre South division, although these should be interpreted with caution. Distribution of the poor 2007 Distribution of the poor 2012 Distribution of the poor change in headcount rates, because of their large popula- (2012 relative to 2007) tions (Figure 29). Source: Authors’ calculations, IHSES 2007 and 2012. Furthermore, the overall rates and trend for the FIGURE 28: Division Headcount Rates South mask an increasingly divergent performance (Percent) and Changes between Basra and the other governorates in the (Percentage Point), 2007–2012 South (Figure 30). While poverty fell by 14 per- centage points in Basra to 13 percent in 2012, in 35% Thi Qar, Muthanna, Qadisiya and Missan, the four 30.2% 28.8% 28.7% 30% other governorates in the Southern division, head- 22.3% 21.9% 25% count rates increased by 10 percentage points to a 19.3% 18.0% 20% staggering 40 percent. Similarly, within the North, 14.5% 13.8% 12.3% 15% headcount rates increased sharply in Nineveh, from 10% 20 percent in 2007 to 32 percent in 2012, while 5% poverty fell in Salahaddin and Kirkuk, the two other 1.6% 0.3% 0% Northern governorates. –1.4% –1.4% –5% –10% –15% Poverty Beyond Consumption –14.3% –20% Kurdistan Baghdad North Centre South Poverty in Iraq, as in the developing world, goes Headcount poverty rates 2007 well beyond material deprivation—the inability to Headcount poverty rates 2012 satisfy basic needs of food, shelter, clothing and Headcount poverty rates change (2012 relative to 2007) other necessities that make up a minimum standard of living. Over and above the deprivation of many Source: Authors’ calculations, IHSES 2007 and 2012. Iraqi households in human development—health, 38 THE UNFULFILLED PROMISE OF OIL AND GROWTH Poverty in Basra and the Rest of FIGURE 30:  welfare as perceived by Iraqis themselves, and how the South, 2007–2012 these relate to consumption poverty. 0.5 The first such question asks the main respondent: “In 0.40 your view, what’s the minimum monthly income that 0.4 your household needs to cover your basic needs?” 0.3 0.30 For Iraq as a whole, the average minimum monthly 0.27 income per capita reported by households as being 0.2 0.13 adequate to cover their basic needs is ID 128900. 0.10 0.1 On average, rural households report 32 percent lower minimum income requirements compared to 0 urban households (Table 9). The highest reported –0.1 minimum income needs are in Baghdad, 28 percent –0.14 higher than the national average, with the lowest in –0.2 South Basra South excl Basra the North, 21 percent lower. The largest differences between urban and rural households is in Kurdistan, Headcount poverty rates 2007 with rural households reporting needing 38 percent Headcount poverty rates 2012 lower incomes per capita per month; while Baghdad Headcount poverty rates change (2012 relative to 2007) and the North have the lowest urban-rural differen- tial of around 20 percent. It is interesting to note Source: Authors’ calculations, IHSES 2007 and 2012. that despite the regional non-food adjustment that allows for a significantly higher consumption pov- erty line in Baghdad, perceived differences between education and living standards (described in detail minimum income requirements and the poverty line in the next chapter)—, household self-assessments are substantial. Households in Baghdad report need- of their welfare status incorporate a range of other ing a minimum income that is on average 40 percent measurable and unmeasurable dimensions of wel- higher than the Baghdad poverty line. Similarly, in fare. The 2012 IHSES surveys ask a series of ques- the Central province, where poverty has come down tions to better understand subjective and relative sharply in many governorates, households report wellbeing. This section uses these measures and minimum income needs almost 30 percent higher their correlates to further our understanding of than the regional poverty lines. TABLE 9: Average Minimum Per Capita Monthly Income (‘000s Iraqi Dinar) Required to Meet Basic Needs, 2012 % Difference % Difference % Difference (Urban relative to (Relative to Consumption (Relative to Regional Rural Urban rural) Total National) poverty line poverty line) Kurdistan 100.88 138.88 37.67 131.45 1.98 142.41 –7.69 Baghdad 140.42 168.69 20.13 165.13 28.10 115.93 42.43 North 91.29 109.23 19.65 101.81 –21.02 101.68 0.13 Central 113.95 145.46 27.65 131.16 1.75 101.68 29.00 South 96.12 119.90 24.74 111.90 –13.19 101.68 10.05 All Iraq 105.57 139.68 32.30 128.90 Poverty, Shared Prosperity and Subjective Well-Being in Iraq 39 Table 10 shows average minimum per capita The concept of minimum income needs is not pure- monthly incomes reported by households in each ly an absolute one, but also, one that appears to in- governorate (rural and urban areas), as well for the crease slowly with per capita consumption expendi- nation. The largest differences between rural and tures (Figure 31). For the bottom decile, mean per urban areas in this measure is in Erbil, with urban capita expenditures are below the reported monthly households stating minimum income needs as be- minimum income needs. Thereafter, minimum in- ing 50 percent higher than those in rural areas. come needs increase slowly, at an average rate of The smallest differential is in Basra, where urban 7 percent per decile, until the 7th decile. Thereafter, and rural households report needing very similar they increase more rapidly, by 11.6 percent for the incomes. Minimum incomes required in rural and 8th and 9th decile, and by 31 percent for the top de- urban areas are relatively low in the governorates cile (compared to the 9th decile). with high consumption poverty, and on relative- ly high in Erbil, Baghdad, Najaf, Basra, and An- The survey also solicits information on satisfaction bar. Relative to the national average, residents of along various dimensions; including food, housing, Nineveh and Muthanna reported needed more income, work, local security, and trust and accep- than 30 percent lower incomes per month to meet tance within the community, and life overall. These their basic needs, whereas those in Najaf needed are asked of all household members above the age more than 40 percent more. of 15. For each of these elements and for the overall TABLE 10: Average Minimum Per Capita Income Required (monthly, ‘000s Iraqi Dinar), Governorates % Difference (Urban % Difference (Relative Rural Urban relative to rural) Total to National) NINEVEH 74.41 86.57 16.35 81.73 –36.59 MUTHANNA 75.17 100.36 33.51 86.22 –33.11 KERBALA 82.00 96.88 18.15 91.91 –28.70 QADISIYA 85.81 101.42 18.20 94.73 –26.51 THI-QAR 81.05 103.61 27.83 94.75 –26.50 MAYSAN 100.82 119.53 18.55 113.89 –11.65 SULAIMANIYA 93.18 119.41 28.15 115.01 –10.78 DIYALA 102.54 129.69 26.47 115.39 –10.48 SALAHADDIN 106.15 127.79 20.39 116.03 –9.99 WASIT 101.54 138.47 36.36 123.06 –4.53 BABYLON 108.84 148.32 36.27 127.13 –1.38 KIRKUK 114.83 143.43 24.91 134.28 4.17 DUHOK 106.15 148.82 40.20 137.10 6.35 BASRA 140.79 138.08 –1.92 138.63 7.55 ANBAR 135.62 141.96 4.68 138.71 7.61 ERBIL 103.80 155.71 50.01 146.73 13.83 BAGHDAD 140.42 168.69 20.13 165.13 28.10 NAJAF 148.43 203.18 36.89 185.80 44.14 All Iraq 105.57 139.68 32.30 128.90 0.00 40 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 31: Comparing Minimum Monthly with food, housing, income and work are the high- Income Needs Per Capita with est in Baghdad and the South; while Baghdad also Monthly Per Capita Expenditures, has the highest rates of dissatisfaction in terms of by Consumption Deciles education and local security. 500 The series of questions on satisfaction are immedi- 450 400 ately followed by a subjective self-assessment of the 350 household’s poverty status, asked of the same set 300 of individuals. Based on these different measures, 250 200 we construct three additional subjective and self- 150 reported measures of welfare: 100 50 0 1. Minimum income poverty: A household is 1 2 3 4 5 6 7 8 9 10 poor according to this measure if their stated Mean Minimum Income per capita minimum per capita monthly income needed to Mean monthly per capita real expenditure cover basic needs is higher than their per capita monthly expenditure 2. Satisfaction poverty: An individual aged 15 years and above is poor by this measure if they assessment of satisfaction, Figure 32 graphs the state that they are ‘not very satisfied’ or ‘not share of respondent who reported being dissatisfied satisfied at all’ with life overall across different elements. Across rural and urban ar- 3. Subjective poverty: An individual aged 15 years eas, the highest rates of dissatisfaction are related to and above is poor by this measure if they an- housing, income, work and education, with upwards swer that their household’s situation is ‘poor’ of a third of respondents reporting dissatisfaction. or ‘very poor’. Rates of dissatisfaction across different elements Figure 34 plots consumption poverty headcount vary across space (Figure 33). Rates of dissatisfaction rates and self-assessment of wellbeing according to FIGURE 32: Dissatisfaction Across Different FIGURE 33: Rates of Dissatisfaction, Selected Dimensions, Iraq, Urban and Rural Dimensions, by Division 50% 0.6 0.5 40% 0.4 30% 0.3 20% 0.2 10% 0.1 0% 0 Food Housing Income Health Work Local security Education Freedom of choice Control over life Trust and acceptance in community Life overall Food Housing Income Health Work Local security Education Kurdistan Baghdad North Iraq Rural Urban Centre South Poverty, Shared Prosperity and Subjective Well-Being in Iraq 41 these measures for Iraq as a whole and for rural FIGURE 34: Headcount Rates, Different and urban areas. Overall, headcount rates based Measures of Wellbeing, Iraq, on consumption are fairly similar to those based Urban and Rural, 2012 on the minimum income question and on life sat- isfaction, while poverty as measured by subjective 0.35 0.30 well-bring is higher at 26 percent. In rural areas, 0.25 20 percent of individuals report being dissatisfied 0.20 with their lives, 24 percent have lower per capita 0.15 consumption than their estimated basic income 0.10 needs, while more than 30 percent are poor based 0.05 on the consumption poverty line or assess that 0.00 Total Rural Urban their household is poor or very poor. In urban ar- eas, while consumption poverty headcount rates Consumption CBN (Absolute poverty line) Subjective Satisfaction MIQ are relatively low, other measures of poverty are significantly higher. Looking across divisions (Figure 35), in line with consumption poverty rates, Kurdistan has the low- consumption expenditures that are below their re- est rates of dissatisfaction, subjective poverty and ported minimum income requirements. minimum income poverty. In fact, only 5 percent of the population reported needing more income Comparing the incidence of poverty according to to fulfil basic needs than their current expenditure. these different measures over consumption deciles In Baghdad, while consumption and subjective reveals the extent to which these subjective mea- poverty are similar, other measures suggest lower sures combine both absolute and relative measures. levels of poverty. In the Centre, where consump- We can see that minimum incomes, while referenced tion poverty rates fell the most since 2007, mini- to minimum needs, appear to have a relative dimen- mum income, subjective and satisfaction poverty sion, and increase with the wealth of households, the are all much higher, with headcount rates above share of households consuming below their self-re- 30 percent. In the South in contrast, rates of dis- ported minimum income needs declines steadily as satisfaction with life are relatively low, at 23 per- consumption increases; from 62 percent among the cent, compared to consumption poverty at 30 percent and subjective poverty at 34 percent. An- FIGURE 35: Headcount Rates, Different nex Table 8.9 reports estimates at the governor- Measures of Wellbeing, Divisions, ate level. Across all measures, Sulaimaniya has the 2012 lowest headcount rates. While Muthanna has the highest rates of consumption poverty at 48 per- 0.4 cent, Qadisiya has the highest rates of subjective poverty, with half the respondents considering 0.3 that their household’s situation was poor or very 0.2 poor. The lowest rates of life satisfaction are re- ported in Baghdad, with 36 percent stating that 0.1 they were not very or not at all satisfied with life 0 overall, likely reflecting the larger rates of dissatis- Kurdistan Baghdad North Central South faction with the security situation. In contrast, in Consumption CBN (Absolute poverty line) Subjective Najaf, where consumption poverty rates are only Satisfaction MIQ 10 percent, more than 40 percent have per capita 42 THE UNFULFILLED PROMISE OF OIL AND GROWTH  overty Measures Across Consumption Deciles FIGURE 36: P 1.0 0.9 Headcount rates (Share) 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 2 3 4 5 6 7 8 9 10 Consumption CBN (Absolute poverty line) 1.00 0.76 0.15 0.07 0.00 0.00 0.00 0.00 0.00 0.00 Subjective 0.56 0.43 0.38 0.29 0.26 0.25 0.18 0.17 0.14 0.08 Satisfaction 0.30 0.24 0.25 0.24 0.19 0.20 0.20 0.17 0.18 0.15 MIQ 0.62 0.35 0.29 0.24 0.20 0.14 0.13 0.10 0.07 0.05 bottom consumption decile, to 20 percent among the consumption poverty line, while around a fifth the 5th decile, to 5 percent among the top decile. of households who express dissatisfaction with their Life satisfaction and subjective measures on the other lives are also consumption poor (Figure 38). hand, which go beyond consumption and income, do not decline as quickly with increases in consump- The correlates of subjective poverty and life (dis) tion levels. Subjective self-assessed poverty levels, satisfaction are therefore broader than those of con- which are anchored to some notion of absolute wel- sumption poverty (Annex table 8.10). For instance, fare, fall from 56 percent among the bottom decile larger household sizes, with more children and to 26 percent among the 5th decile and to 8 percent more elderly persons are associated with lower levels among the top decile. Dissatisfaction rates are the of subjective poverty and of dissatisfaction with life; least responsive to improvements in income and con- as is being female. Other correlates are common: sumption, remaining as high as 26 percent among the 5th decile and 15 percent among the top decile. FIGURE 37: Share of Consumption Poor Who In line with the different aspects of absolute and are Poor According to other relative deprivation captured by these different Measures measures, and the elements taken into account in evaluating whether a household or an individual 60 considers themselves as deprived; almost half of the 50 consumption poor are also poor in terms of subjec- 40 tive poverty and minimum income poverty, while only 28 percent are dissatisfied with life overall 30 (Figure 37). 45 percent of those who are poor in 20 the sense that their consumption is lower than their 10 perceived minimum income needs are also con- 0 sumption poor, 32.5 percent of households who Subjective Satisfaction MIQ consider themselves poor or very poor are below Poverty, Shared Prosperity and Subjective Well-Being in Iraq 43 FIGURE 38: Share of Subjective, Satisfaction We define the following deprivations: and Minimum Income Poor Who are Consumption Poor Education: Illiterate and incomplete primary (relative to Higher secondary and 50 Tertiary) 40 Complete primary and lower 30 secondary (relative to Higher secondary and Tertiary) 20 Employment: Non employed (relative to Public 10 sector employment) 0 Private sector job (relative to Subjective Satisfaction MIQ Public sector employment) Lower than average share of working age men employed Displacement Forcibly displaced non-employment, employment in the private sector, and migration: as well as fewer working age males employed are all Head of household born associated with higher poverty according to these elsewhere measures. Lower education levels also lower these Consumption Household per capita expenditure indicators of well-being. Each division is associated and minimum is lower than minimum income with lower well-being relative to Kurdistan; and income poverty: needs in addition, individuals belonging to households Quintile 1 (poorest) (relative to whose heads were born in a different governorate, Quintile 5) or report having been forcibly displaced, are more Quintile 2 (relative to Quintile 5) likely to report lower subjective welfare and satisfac- tion. Finally, and as expected, the higher the indi- Quintile 3 (relative to Quintile 5) vidual’s consumption level, the lower the rates of Quintile 4 (relative to Quintile 5) subjective poverty and dissatisfaction. Space: Subjective poverty: Division (rela- tive to Kurdistan) In order to further understand how Iraqis (in this case, Iraqis aged 15 and above) evaluate their Satisfaction: Division (relative to household’s welfare status as well as their own sat- North) isfaction or dissatisfaction with life, we restrict at- tention to dimensions of deprivation as revealed By normalizing the coefficients of the regressions by the regressions described above—education, of these dimensions against the subjective poverty employment (or lack thereof), poverty in terms of dummy and a dummy for whether an individual low consumption expenditures, either relative to is dissatisfied with life, we can construct the rela- the consumption distribution, or relative to their tive weights of each of these dimensions (Annex own notions of minimum income needs, and the Table 8.11, 8.12 and 8.13). These dimensions are division of residence, which proxies for access to aggregated up into categories: education, employ- services, local security and rule of law, and the local ment, displacement and migration, consumption environment and labor market. We exclude house- or income poverty and place of residence; and hold size, composition and the gender of the in- are shown in Figure 39. Figure 40 shows the in- dividual, as we consider these as given rather than cidence of each of these deprivations, weighted deprivations in themselves. appropriately. 44 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 39: Relative Weights of Different driven more by where an individual lives, reflecting Dimensions of Deprivation in the additional importance of location specific factors Determining Subjective Poverty including security concerns, local labor market con- and Dissatisfaction with Life ditions and service delivery. However, the incidence of these deprivations also varies: for instance, less than 6 percent of individuals report being forcibly Life displaced; while more than 60 percent of individuals satisfaction are not employed. Figure 26 shows how the incidence of these depri- Subjective vations would change depending on which sets of poverty weights were used. Given the pervasively low levels of education, the incidence of education deprivation 0% 50% 100% is large, and is weighted more in subjective poverty Education Employment Displacement and migration assessments. Similarly, while employment outcomes Consumption and minimum income poverty Space have a relatively smaller weight; because so few indi- viduals have access to public sector jobs, the weighted incidence of employment deprivations is also relative- ly large. In line with the relative importance of dif- FIGURE 40: Weighted Incidence of Different ferent dimensions, the incidence of consumption or Dimensions of Deprivation in income poverty is the largest when weighted accord- Determining Subjective Poverty ing to subjective welfare assessments, while the inci- and Dissatisfaction with Life dence of spatial dimensions is more important when weighted according to the dimensions of satisfaction. Satisfaction poverty If we were to use these broader dimensions of depri- weighted vation, weighted according to their revealed impor- incidence tance in determining self-assessments of household Subjective welfare or life satisfaction, to construct a multidi- poverty mensional index of deprivation, headcount rates weighted incidence would be higher than as measured by consumption poverty, 25.57 percent if subjective poverty weights 0% 50% 100% and deprivations were used and 28.45 percent if life satisfaction weights and dimensions were used. Education Employment Displacement and migration Consumption and minimum income poverty Space To conclude, the 2007 to 2012 period has been characterized by low rates of per capita consumption growth relative to the increase in per capita GDP. Consumption growth has been faster, and as a con- When comparing the relative weights of different sequence, poverty reduction has been larger, in rural dimensions in Figure 39, it is apparent that in deter- areas and in the 14 governorates outside of Kurdistan mining subjective self-assessments of poverty status, and Baghdad. In fact, poverty reduction has been al- the largest weight is placed on lack of consumption most entirely focused in the Central division, and a or income, with smaller but relatively equal weight few other governorates, while poverty has increased on educational deprivation or on place of residence. from already high levels in the South. Consumption In contrast, dissatisfaction with life appears to be has also grown faster for the non-poor than the poor. Poverty, Shared Prosperity and Subjective Well-Being in Iraq 45 Household size and composition, the education and to understand the different elements that the Iraqi sector of work (in general) of the head of household people take into account when evaluating their and the location of the household are all strong de- own welfare, elements that go beyond consump- terminants of consumption and poverty. But house- tion. These include concerns about the work and holds dependent on agriculture and construction are incomes, education, the ability to fulfil basic needs, no less likely to be poor relative to households with as well as local economic and security conditions. heads who are unemployed or out of the labor force; Taken together, these findings highlight the im- while public sector jobs are in general associated with portance for putting in place a set of broad based a lower probability of poverty. policy reforms to address the multidimensional de- privations faced by Iraqi households that both shape Recognizing that poverty has many facets, we also their perceptions of the present and their aspirations use subjective measures of wellbeing and welfare for the future. Poverty in Human Capital 3 P overty in consumption is but one dimension of intake, and a smaller share of food calories from the deprivation in Iraq—many individuals ex- Public Distribution System also have lower rates of perience poor health and education outcomes stunting. The negative relationship between mean and limited access to essential services. To benchmark consumption levels and nutrition is directly related poverty in human development, we create a multidi- to food expenditures and caloric intake. Stunting is mensional index of human development deprivations. more prevalent among the children below 36 months, Poverty as measured by this index generally follows suggesting that perhaps malnourished mothers may be the same pattern across space as consumption poverty. giving birth to malnourished children. Indeed, early But there are important differences. While the high motherhood is associated with poor nutritional out- level of consumption deprivation in the South coexists comes for children, even after accounting for lower with poverty in human capital, and the low consump- food consumption associated with the household’s in- tion poverty rates in Kurdistan are accompanied by come and expenditure levels, the place of residence and limited deprivation in human development; in the parental education. Centre, while welfare as measured by consumption has rebounded, significant deprivations in human capital Education, one of the strongest correlates of poverty, remain. Unequal access to basic human development varies widely across Iraq, by division and by gender. outcomes is shaped by both gender and space, in ad- The median education level for Iraqis is primary dition to household wealth and family characteristics. schooling, and for the majority of children, schooling Inequality in access to educational attainment is pri- ends after primary education. Very little has changed marily explained by gender whereas unequal access to in terms of educational outcomes between 2007 and quality housing and other basic services is driven by 2012 in Iraq. One exception is the cohort of young the place of residence. females, who are catching up with the relatively low attainment of their male counterparts. Both gross Early childhood nutrition, the lack of which can have and net enrollments display a sharp decline after pri- irreversible long term consequences, is correlated not mary school. The sole exception is Kurdistan, where only with the household’s ability to consume adequate gross enrollment rates are above 100 percent even at food, but also with space, reflecting differential ac- the intermediate and secondary levels, the highest in cess to essential services, and with maternal nutri- the country. There exist significant gender disparities tion, proxied by early motherhood. Stunting and child in gross enrollment at each level. Despite these odds, underweight rates vary in Iraq with wealth, with the among the few girls who make it to secondary and highest prevalence rates among households belonging tertiary level, net enrollment rates are slightly higher to the poorest quintiles. Governorates with higher per among girls than among boys, suggesting that they capita consumption expenditure, higher total caloric are more successful in completing each level on time. 48 THE UNFULFILLED PROMISE OF OIL AND GROWTH The lack of access to schools, the pressure to support the Multidimensional Poverty in Human family by working and the unaffordability of school- Development ing, social norms about the value and appropriateness of girls’ education, and a significant lack of interest; Poverty in consumption is but one dimension of de- all contribute to poor educational outcomes for chil- privation in Iraq—many individuals experience poor dren. The latter may reflect poor education quality, health and education outcomes and limited access but these education outcomes are more broadly also to essential services. In part, this is a legacy of years likely reflecting limited returns to education on the of violence and instability, which have led to the labor market. neglect and destruction of infrastructure, the diver- sion of public resources away from these types of Finally, turning to measures of housing quality and investments, and the loss of qualified personnel (see access to basic services, while there have been some Chapter 4 for a more detailed analysis of the role of measurable improvements in access, there are little conflict in explaining stalled progress in key human improvements in quality, with some significant varia- development outcomes). To benchmark poverty in tions across space and by household wealth. A substan- human development, we create a multidimensional tial portion of households who use the public network measure that builds on three equally weighted di- as the main source of water consider it to be insuf- mensions of deprivation in human development: ficient. Only 3 percent of households in Baghdad, and education, health and standards of living. Each di- around a tenth of households in the Centre and the mension is composed by the following deprivation South receive power for more than 12 hours. Housing indicators (each taking the values 0 or 1):20 quality is also perceived to be low, with a large share of Iraqi households reporting inadequate space and Education inadequate utilities. In fact, on average, households have only 1 bedroom for every 3 to 4 members. Overall, 1. Illiterate household head: Deprived if house- in Iraq, three decades of violence and insecurity have hold head is illiterate stalled progress in human development and service de- 2. School attendance: Deprived if any child under livery, and the nation faces a significant deficit, with 12 in the household is not currently in school far-reaching consequences for the economy and for fu- ture generations. Health Across the world, consumption poverty is associat- 1. Nutrition: Deprived if there is a malnourished ed with poorer education and health outcomes and child in the household (suffering from either limited access to and lower quality of basic services. stunting or wasting) Iraq is no exception—education is a strong determi- nant of higher per capita consumption expenditure Standard of living and of lower poverty. More than four-fifths of the poor belong to households with heads having com- 1. Electricity: Deprived if household has less than pleted primary education or less, and poverty rates 12 hours a day of electricity from the public increase systematically with lower levels of education network of the head of household. In this chapter, we analyze 2. Assets: Deprived if the household does not own a range of human development outcomes; explore more than one TV, motorbike or fridge and their links to poverty and welfare, with a particular does not own a car focus on children. We begin with some aggregate indices of deprivation and opportunity in human de- 20 The dimensions were defined so as to be as comparable as velopment in Iraq and examine how they correlate possible to the Global MPI (2013), with a focus solely on with welfare. human development. Poverty in Human Capital 49 3. Flooring: Deprived if household has a brick, poverty in Iraq is approximately 20 percent in earth or other flooring 2012, MPI poverty is 35 percent. But the gap be- 4. Water: Deprived if household has insufficient tween consumption poverty and poverty in human supply of water from network development is the largest in the Centre, 21 per- 5. Sanitation: Deprived if household does not centage points, and in urban areas, 16 percentage have flush toilet or if the toilet is shared points. This suggests that improvements in welfare as measured by consumption, which are relatively Based on the indicators presented above, each high in the Centre and in urban areas, do not al- household is assigned a deprivation score, which ways go hand in hand. The exception is Kurdistan, is the weighted sum of the number of depriva- where low rates of consumption poverty are ac- tions (all three dimensions are equally weighted as companied by low rates of multidimensional pov- is each indicator within each dimension). In com- erty in human development. puting the Multidimensional Poverty Index for human development, the larger the number of de- Poverty in human development goes beyond con- privation indicators in each dimension, the smaller sumption poverty—this is evident in Table 11: only the weight of each indicator individually. A given 10.4 percent of the country´s population is both household is considered multidimensional (MPI) MPI poor and consumption poor, while 55.3 per- poor if their deprivation score lies above the pov- cent are neither consumption poor nor MPI poor. erty cutoff of 0.33. Among the consumption poor, more than half are also MPI poor. In contrast, poverty in human de- Figure 41 displays a comparison among the fig- velopment, as measured by this index, is more wide- ures for consumption and MPI poverty. While es- ranging in that more than 70 percent of the MPI timates of poverty in multiple human development poor are non-poor in terms of consumption. dimensions are higher than consumption poverty in 2012, both measures suggest similar spatial pat- Figure 42 shows the incidence of deprivations along terns in poverty, with the Southern division having the dimensions of living standards, health and edu- the highest MPI and consumption poverty, and cation. Lack of sanitation, inadequate electricity and Kurdistan having the lowest. While consumption poor nutrition are the most prevalent deprivations FIGURE 41: Comparison Among Poverty Measures: Consumption Poverty and MPI in Human Development 50% 47% 45% 45% 40% 38% 36% 35% 35% 33% 31% 31% 30% 30% 25% 22% 18% 20% 20% 16% 15% 15% 15% 12% 10% 5% 0% Kurdistan Baghdad North Central South Rural Urban Iraq MPI headcount Consumption poverty headcount Source: Authors’ calculations, IHSES 2012. 50 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 11: Consumption Poor Versus MPI Poor free of deprivation (8%). In contrast with the trends in consumption poverty, the Central division ap- Consumption non-poor Consumption poor pears to be performing the worst in terms of the MPI non-poor 55.3% 9.4% presence of multiple dimensions of deprivation, with 86 percent of households experiencing two or MPI poor 24.9% 10.4% more deprivations, and 21 percent having four or Source: Authors’ calculations, IHSES 2012. more. This may suggest that while welfare as mea- sured by consumption has been quicker to rebound in the Centre with the improvements in security in Iraq, followed by water and school attendance. and revival of economic activity, the quality of in- Among these, malnutrition is the deprivation with frastructure and services has not yet caught up to a the highest weight in the index, since it is the only commensurate level. indicator for the Health dimension. The MPI can also be decomposed to quantify the Moreover, the vast majority of households across contributions of each of the different dimensions Iraq suffer not one but multiple deprivations in and of each indicator (Figure 44). Looking at the human development; only a tiny fraction of house- aggregate dimensions, poverty in human develop- holds suffer no deprivation on these human devel- ment is being driven mostly by deprivation in health opment indicators. Figure 43 shows the incidence (61%), followed by living standards (27 %) and edu- of overlapping dimensions of deprivation in each cation (12%). When turning to the individual indi- division, as well as rural and urban Iraq. Overall, in cators, deprivation in nutrition is the main driver of Iraq, 63 percent of households suffer from two or MPI poverty, followed by sanitation, electricity and three simultaneous deprivations; while 11 percent school attendance. experience four or more. In line with the higher rates of MPI poverty in rural areas, 51 percent of rural households are deprived in three or more Human Opportunity Index dimensions, while 21 percent experience four or more, as compared to 33 percent and 8 percent Access to basic goods and services—such as edu- of urban households. Not surprisingly, Kurdistan cation and health—can be encompassed by the stands out with the highest incidence of households term opportunity, as described in the 2006 World FIGURE 42: Dimensions of Deprivation Sanitation 83% Living standards Water 25% Flooring 6% Assets 5% Electricity 64% Health Nutrition 32% School attendance 15% Education Illiterate household head 6% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Source: Authors’ calculations, IHSES 2012. Poverty in Human Capital 51 FIGURE 43: Overlapping Dimensions of Deprivation Iraq 3% 22% 37% 26% 9% 2% Urban 4% 24% 39% 25% 7% 1% Rural 2% 15% 32% 30% 15% 6% South 3% 23% 38% 23% 9% 4% Central 3% 11% 29% 36% 17% 4% North 3% 24% 39% 26% 7% 2% Baghdad 1% 12% 45% 32% 10% 1% Kurdistan 8% 47% 34% 9% 1% No deprivations 1 deprivation 2 deprivations 3 deprivations 4 deprivations 5 or more deprivations Source: Authors’ calculations, IHSES 2012. FIGURE 44: Contribution of Each Dimension to Iraq´s MPI 1% 4% 12% 12% 8% 1% 4% 27% 9% 61% 61% Illiterate household head School Attendance Nutrition Education Health Living Standards Electricity Asset Floor Water sanitation Source: Authors’ calculations, IHSES 2012. Development Report.21 Unlike the notion of equal- access can be allocated in many different ways. These ity in earnings or income, there is widespread con- tradeoffs can be particularly important in countries sensus that such opportunities to access basic goods such as Iraq that face a huge human capital deficit. and services should be made universal and that in- Analyzing the pattern of allocation, and whether it equality in children’s opportunities propagates de- is being influenced by the circumstances of the ben- privation and weakens overall prosperity and eco- eficiaries is therefore important and lies at the core nomic growth. While it is true that more developed of the concept of equality of opportunity. countries will have more resources to deliver basic goods and services to their citizens, developing countries will face important tradeoffs as additional 21 World Bank (2006) “Equity and Development”. 52 THE UNFULFILLED PROMISE OF OIL AND GROWTH BOX 1: A Stylized Example of the HOI Consider two countries, A and B, each with a total population equally Country A distributed in 10 western communities and 10 eastern communities. The coverage rate of school enrollment (or the average enrollment rate) for both countries are 0.6, i.e. children in 60 percent of the communities attend school in each country. In the figure, red-colored communities illustrate covered communities. It is evident that access to school is distributed differently among the population in country A and B. In country B, children in western communities have a 20% probability of access to schooling, whereas the probability is 100% for eastern communities. Equality of opportunity will hold true for each country if western Country B and eastern communities have the same rate of coverage. However, while in country A school coverage is 60% in western and eastern communities, this is not true in country B. This suggests that inequality of opportunities is higher in country B. The D-index is the share of total enrollment that is “misallocated”, namely 0/12 and 4/12 for A and B, respectively. The HOI is then calculated as the coverage rate multiplied by 1 minus the D-index (the share that is equitably allocated). In other words, the HOI penalizes country B for its inequitable coverage rate relative to country A. Therefore, HOIA = C0 (1-D) = 0.6 * (1–0) = 0.60; HOIB = C0 (1-D) = 0.6 * (.33) = 0.40. The Human Opportunity Index or HOI (devel- wealth of his/her household, or the education of oped by the World Bank and external research- his/her parents, among others. ers) is an intuitive measure of the availability of an opportunity in a society (measured by the av- We estimate results for three types of opportuni- erage coverage rate) that also takes into account ties. For education-related opportunities, we de- the extent to which opportunities are distributed fine two opportunities: attending school (for chil- inequitably among individuals of different circum- dren 10–14 years old) and having finished primary stances (through the calculation of a penalty). In school (for children ages 12–16). Housing-related particular, the HOI calculates how personal cir- opportunities include (all for children ages 0–16) cumstances influence the probability of a child to considering the water supply from the public net- access the goods and services necessary to succeed work sufficient, having a public hospital within 5 in life. In other words, the HOI is an inequity- km of the household, and having electricity avail- penalized estimate of the coverage rate of an op- able from the public network for at least 20 hours portunity in a society (see Box 1). The penalty in- per day. Finally, for health-related opportunities we creases as the coverage rate among children with include the absence of the stunting, wasting, or be- different circumstances increases. Opportunities ing underweight—all recorded for children 0–60 can be defined as any good or service that any soci- months old (Table 12). Health-related opportuni- ety would consider should be provided universally, ties seem to be relatively better. The percentage such as basic education, access to water or absence of children who are not wasted or underweight of malnutrition. Circumstances are all the charac- is above 90%. On the other hand, the percent- teristics into which a child is born and which con- age of children who are found not stunted is just trol are, by definition, outside his/her. Examples over 70%. It is notable that all three health-related of circumstances are the gender of the child, the opportunities registered relatively low levels of Poverty in Human Capital 53 TABLE 12: Coverage Rates and Human We next look at the circumstances that help explain Opportunity Index the estimated inequality among the opportunities analyzed (Figure 45). In line with previous results, Group Opportunity Coverage D-Index HOI the gender of the child seems to explain about a Education Attending School (10–14 yrs) 82.7% 6.5% 77.3% third of the inequality in the education-related op- Finished primary (12–16yrs) 80.6% 6.5% 75.4% portunities. Next in relative importance, we find Housing Water supply sufficient 58.3% 17.6% 48.0% that household wealth and the father’s educational Public Hospital <5km 56.5% 23.4% 43.3% attainment together explain slightly more than a Electricity available >20h 8.8% 62.2% 3.3% third of the inequality. In terms of housing-related Health Not Stunted (0–60 months) 72.6% 3.4% 70.1% opportunities, we see that location (i.e. urban or No Wasting (0–60 months) 92.1% 1.0% 91.2% rural) and divisional differences explain more than Not Underweight 90.0% 1.9% 88.3% 70% of the estimated inequality. (0–60 months) Source: Authors’ calculations, IHSES 2012. These results highlights the significant differential access to opportunities experienced by children due to having born in a particular region of the inequality, with no stunting being the most un- country when it comes to accessing basic services. equal (D-Index of 3%). This result indicates that, In addition, inequality in educational attainment overall, children with different circumstances22 is also explained by the gender of the child, in line face relatively small differential likelihoods of ex- with the significant gender gaps in enrollment periencing bad health outcomes.23 noted later in this chapter. Next, we take up each of these separate aspects of human development— In terms of education, we find that among 10–14 health, education, and basic services—in greater year olds almost 1 in every 5 children is missing detail. school, and among those between 12–16 years old there is 1 in every 5 who has not completed prima- ry school. Also, we find moderate levels of inequal- ity in access to these opportunities. The D-Index is estimated to be 6.5 % in both cases. A third of 22 Specifically children from different groups of the popula- that inequality is due to the gender of the child. tion according to the circumstances we use in the estima- The coverage and estimated HOI among housing- tion of the HOI: gender, household wealth, family demo- related opportunities are the worst. Around 58% graphics, education of the household head, economic sector of children have access to sufficient water supply where the household head works, region (aggregated in 5 divisions), and location (either urban or rural). Other char- from the public network, and about 57% have a acteristics missing from our analysis may affect the likeli- public hospital within 5 km of their household. hood of having an opportunity. Moreover, the level of inequality of having access 23 This result does not contradict the findings in Table 1 of to these opportunities is high: water registers a D- the Annex, where stunting levels are statistically signifi- Index of 17.6%, whereas public hospitals register cant for households with certain characteristics. The rea- an index 23.4%. This last result reflects that almost son is that the two methods offer answers to different and a quarter of public hospitals should be reallocated complementary questions. Regression estimates provide so that all children have the same probability of be- the expected differential likelihood of a determined group of the children population (i.e. those living in urban set- ing near a hospital. The most worrisome oppor- tings) to experience stunting with respect to a base group. tunity is that less than 10 percent of the popula- The HOI methodology provides a summary measure of tion has continuous access (i.e. at least 20 hours the differences in the likelihood of experiencing stunting a day) to electricity. Moreover, this access is large across all groups of the children population (based on a set unequally distributed. of circumstances). 54 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 45: Contribution of Circumstances to Inequality Index – Selected Opportunities 100 2.3 12.3 7.3 90 12.6 8.2 38.0 80 5.7 6.8 70 15.0 63.5 60 20.3 67.7 9.0 50 6.6 40 15.5 40.7 17.9 30 12.9 4.3 20 2.6 5.5 5.5 34.3 28.5 3.2 5.0 10 11.1 11.5 15.2 0 4.4 5.9 Attending Finished Water supply Public Hospital Electricity School (10–14 yrs) primary (12–16yrs) su cient <5km available >20h Location Region Head: Public sector worker Head: Education Family Characteristics Wealth Gender Source: Authors’ calculations using IHSES – 2012. Lagging Behind in Health and sample of countries with similar GDP. For instance, Nutrition wasting rates in Iraq are more than twice the rates in Mongolia, which has a similar level of per capita We begin by examining a set of indicators that mea- GDP, and substantially higher than in Vietnam, sure long-term and short-term inadequacy in nutri- which has slightly lower per capita GDP. The fact tion—stunting and wasting rates—among children. that Iraq performs relatively worse on metrics of Panels A and B of Figure 46 plot these different an- short term nutritional deprivation—wasting—may thropometric measures against GDP per capita (PPP indicate that the health status of the Iraqi popula- 2005) for a sample of countries, including Iraq, with tion has deteriorated in the recent past. available WHO data between 2007 and 2012. While stunting (which measures height-for-age) refers to Iraq once had some of best health indicators in the past inadequate malnutrition and is closely associ- region. According to a WHO (2001) briefing, “the ated with socioeconomic conditions for the popula- large investments in infrastructures and in human tion as a whole, capturing mostly long run effects, resources development carried out during the six- wasting (weight-for-height) refers to short run ef- ties and seventies had led to the development of an fects, usually related to recent and severe weight loss, efficient health system that was considered one of often associated with starvation or severe disease. the best in the Middle-East Region. Malnutrition was virtually not seen, as households had easy and For both measures, stunting and wasting rates affordable access to a balanced dietary intake.”24 To- among children aged 0–60 months, the overall re- day, Iraq underperforms compared to other MENA lationship with GDP is negative as expected: poor- countries (Table 13). At 22.6 percent, stunting rates er countries with low levels of GDP per capita are in Iraq are above the regional average of 21.1 per- more severely affected by adverse nutritional out- cent. In both stunting and wasting, Iraq ranks 7th comes. While Iraq’s stunting rates are slightly be- among a sample of 11 MENA countries. The effect low the average rates commensurate with its GDP level, the country underperforms in terms of wast- 24 Kreisel W: Health situation in Iraq. 2001. www.who.int/ ing when compared to the mean performance of the disasters/repo/6386.doc. Poverty in Human Capital 55 FIGURE 46: Stunting and Wasting (0–60 Months): Iraq Versus other Countries Panel A Panel B 50 15 GTM IDN LKA 40 IDN SYR BTN 10 MDV PHL EGY ALB 30 BOL AGO SWZ NAM AGO EGY IRQ PHL NAM COG SYR COG VNM SLV ARM ALB GUY BTN SUR 20 GUY MDV 5 IRQ LKA VNM ARM MNG BLZ MAR BLZ BIH MAR MNG JOR CHN GEO BOL GEO DOM 10 JOR CHN SUR DOM BIH 0 GTM SWZ SLV 2000 4000 6000 8000 2000 4000 6000 8000 GDP PPP 2005 international dollars GDP PPP 2005 international dollars Stunting Fitted values Wasting Fitted values Source: WHO Global Database on Child growth and Malnutrition and WDI, 2013. of inadequate nutrition on early childhood can be designed to be representative for the governorates devastating, with severe adverse long run effects, as within Iraq. National estimates from both sources shown in Box 2. are fairly consistent (although stunting estimates are higher using IHSES 2012 data), and as IHSES data While the international comparisons for anthropo- allows us to further disaggregate the analysis and metric measures were made using WHO’s global link it to individual and household characteristics, database, the IHSES 2012 survey also collected da- the following section uses survey based estimates ta on anthropometric nutritional measures and was from IHSES (Table 14). BOX 2: Long Term Implications of Malnutrition in Early Childhood Health in the earliest years—actually beginning with the future mother’s health before she becomes pregnant—lays the groundwork for a lifetime of well- being. Health is directly influenced by nutrition, beginning with the mother’s pre-conception nutritional status, extending through pregnancy to early infant feeding, and continuing with diet and activity throughout childhood and into adult life. Adequate intake of both macronutrients and micronutrients is particularly important in the early months and years of life, when body growth and brain development are more rapid than during any other period. Malnourished children score poorer on tests of cognitive function, have poorer psychomotor development and fine motor skills, have lower activity levels, interact less frequently in their environments, and fail to acquire skills at normal rates. The causal nature of such a relationship is indicated by several studies linking the improvement of diets to better motor and mental development for children. A randomized study of supplement distribution during pregnancy and early childhood in Guatemala showed that, not only were treated children observed to have improved cognitive development in preschool years (Martorell 1997), when the cohort was followed into adult years, the treatment group gained significantly higher schooling and wages (Maluccio et al. 2005). Glewwe, Jacoby, and King (2001) track children from birth through primary school and find that better nourished children start school earlier and repeat fewer grades, controlling for family characteristics. Alderman, Hoddinott, and Kinsey (2006) show that increased stunting of children younger than age 2 in Zimbabwe has a causal impact on years of schooling completed 15 years after the nutritional shock. In Colombia, infants born to families at risk of malnutrition that received nutritional supplementation performed better than those who did not, especially on subtests that were primarily motoric (Super et al, 1990). Childhood nutrition not only reduces child mortality but has major economic returns coming from reduced cost of health care and increased productivity of the population over a lifetime. Nutrition impacts not only on survival, but on child development, school retention and achievement. Source: Center on the Developing Child, Harvard University, The Lancet child development in developing countries series, World Bank´s Early Childhood Care and Development in Sub-Saharan Africa. 56 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 13: Nutritional Indicators in the MENA Region Rank GDP per capita Stunting Wasting Country Year Stunting Wasting (PPP 2005) 1 5 IRAN 2011 6.8 4.0 2 2 JORDAN 2012 7.7 2.4 5288.988 3 3 TUNISIA 2010 10.0 3.0 8441.619 4 4 PALESTINIAN 2010 10.6 3.3 TERRITORIES 5 1 MOROCCO 2011 14.9 2.3 4453.112 6 6 LIBYA 2007 21.0 6.5 15699.07 7 7 IRAQ 2011 22.6 7.4 3461.817 8 9 SYRIA 2009 27.5 11.5 4466.081 9 8 EGYPT 2008 30.7 7.9 5411.725 10 11 DJIBOUTI 2012 33.5 21.5 11 10 YEMEN 2011 46.6 13.0 2192.513 Source: WHO and WDI. TABLE 14: Nutrition Indicators in Iraq TABLE 15: Nutritional Outcomes by Wealth According to Different Data Sources Quintiles Year Stunting Wasting Share of children Underweight IHSES 2012 27.35% 7.97% 0–60 months Stunted Wasting children WHO 2011 22.6% 7.4% 1 (poorest) 0.31 0.09 0.13 2 0.29 0.07 0.10 3 0.26 0.08 0.10 Nutritional Outcomes Within Iraq 4 0.26 0.07 0.07 Stunting is a primary manifestation of malnutrition 5 (richest) 0.22 0.08 0.08 in early childhood, including malnutrition during Overall 0.27 0.08 0.10 fetal development brought on by the malnourish- Source: Authors’ calculations, IHSES 2012. ment of the mother, both of which are likely corre- lated with the income level of the household. Stunt- children, with some of the nutritional deficit being ing and child underweight rates vary in Iraq with bridged with time as infants move from breastfeed- consumption quintiles, with the highest prevalence ing to a food diet (Table 16). 35 percent of Iraqi rates among the poorest quintiles, as shown in Table children below the age of 1 are stunted, compared 15. However, stunting rates are high even for the to a third of children aged 12 to 35 months, and richest portion of the population, with more than a approximately a fourth of children aged 26 to 60 fifth of children having low height-for-age, suggest- months. A similar pattern is evident for wasting and ing widespread malnutrition in the past among all low weight children, but there is a sharp decline for segments of the population. children ages 12 months and above relative to those less than a year old. Stunting is more prevalent among the children be- low 36 months, suggesting that perhaps malnour- Household incomes are also correlated with the ished mothers may be giving birth to malnourished employment status of the head of household. While Poverty in Human Capital 57 TABLE 16: Nutritional Outcomes by Age perform worse than would be expected even at their Group low levels of per capita consumption expenditure. These are among the governorates with highest Share of children Underweight poverty headcount rates (48.4, 36.8 and 38.2 per- 0–60 months Stunted Wasting children cent respectively) in 2012 and also among the five 0–11 months 0.35 0.13 0.22 governorates in which poverty increased between 12–23 months 0.31 0.07 0.08 2007 and 2012. In general, the Southern governor- 24–35 months 0.30 0.06 0.08 ates experienced the highest rates of stunting, wast- 36–47 months 0.22 0.06 0.06 ing and underweight children. 48–60 months 0.18 0.08 0.07 Source: Authors’ calculations, IHSES 2012. Kerbela, Salahadin, Basra and Baghdad also stand out, with higher stunting rates than would be pre- there is little difference in wasting and child under- dicted based on their mean consumption level. weight rates by the employment status of the head However, all except Basra perform better in terms of household, stunting rates are 4 to 5 percent- of wasting and underweight children. This pat- age points higher in households with unemployed tern is consistent with sharp poverty reduction in heads, perhaps because those actively seeking work Kerbala and Salahadin between 2007 and 2012, of have no other source of income and cannot afford 23.1 and 24.3 percentage points respectively, which to be out of the labor force (Table 17). makes these two governorates the most successful in the country in terms of poverty reduction. Since Just as poorer countries and poorer households stunting refers to past inadequate nutrition while within countries tend to have worse nutritional in- wasting refers to more recent disease and starvation, dicators, spatial disparities in welfare are also corre- governorates that have experienced recent welfare lated with differences in nutritional markers. Figure improvements perform relatively better in terms 47, Figure 48 and Figure 49 plot stunting, wasting of more the short term indicators of wasting and and child underweight rates in each governorate underweight rates but continue to lag in terms of against their average per capita expenditures. The the longer term stunting measure. Basra however, negative relationship between nutrition and con- remains somewhat of an exception to this pattern, sumption is, particularly pronounced for stunting with relatively high estimates of stunting and under- and underweight rates, while it is weaker for wast- weight rates despite significant poverty reduction. ing rates. On average, the Kurdistan region has the best per- Three southern governorates, Muthanna, Thi-Qar formance in terms of nutritional measures, with the and Missan, have the highest incidence of stunting, exception of the wasting measure for Erbil, which wasting and underweight children, and all three seems to be an outlier. The governorate of Sulai- maniya in Kurdistan, displays the lowest rates of stunting and underweight children, and the second lowest in terms of stunting in the country. Indeed, TABLE 17: Stunting and Wasting by Sulaimaniya has had the lowest poverty headcount Employment of Head rates, both in 2007 and 2012, only 7.6 and 7.4 Underweight percent respectively. The relationship between un- Stunted Wasting children derweight adult measures (not reported here) and Employed 0.27 0.08 0.10 per capita expenditure display a similar negative Unemployed 0.32 0.09 0.12 relationship with per capita consumption as under- Out of force 0.28 0.09 0.11 weight children, with the Southern division under- Source: Authors’ calculations, IHSES 2012. performing relative to the rest of the country. 58 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 47: Stunting Rates and Governorate and per capita caloric intake from rations. The first Consumption Expenditure plots, shown in panels A and B, display a negative relationship between stunting and expenditures and 0.4 KERBELA per capita caloric intake showing that, governorates MUTHANNA MAYSAN SALAH AL-DEEN with higher per capita expenditure and higher total THI-QAR BASRAH caloric intake are the ones with lower rates of stunt- QADISIYA 0.3 BAGHDAD ing. On the other hand, the last two plots (panels C WASIT KIRKUK and D) show a positive relationship between stunting NAINAWA DIYALA DUHOK and consumption based on food rations and caloric BABYLON 0.2 ANBAR NAJAF intake from rations. Thus, higher stunting rates ap- ERBIL pear to be correlated with lower levels of food ex- penditure per capita and potentially with a higher 0.1 SULAIMANIYA level of food insecurity, and a greater dependence 100 150 200 250 300 on the Public Distribution System. Not coincidently, Mean per capita consumption expenditure the governorates highly dependent on consumption based on rations and displaying high stunting fig- Stunted Fitted values ures are the governorates on the South, in particu- lar Muthanna, Thi-Qar, Qadisiya and Missan, where Source: Authors’ calculations using IHSES 2012. poverty increased the most between 2007 and 2012. The negative relationship between mean consump- tion levels and nutrition is directly related to food Early Motherhood and Child Nutrition expenditures and caloric intake. In Figure 50, we In this section, we explore the relationship be- plot governorate level stunting rates against per cap- tween child nutrition and maternal health, and in ita expenditure, per capita caloric intake, per capita particular the hypothesis that early motherhood is expenditure based on consumption of food rations FIGURE 49: Underweight Rates for Children FIGURE 48: Wasting Rates and Governorate and Governorate Consumption Consumption Expenditure Expenditure 0.15 2.0 BASRAH MAYSAN QADISIYA MUTHANNA THI-QAR ERBIL THI-QAR 1.5 0.10 BASRAH MUTHANNA MAYSAN QADISIYA WASIT KERBELA BABYLON NAINAWA KIRKUK WASIT SALAH AL-DEEN DIYALA BAGHDAD 1.0 BAGHDAD NAINAWA DIYALA KIRKUK 0.05 KERBELA NAJAF DUHOK ERBIL SALAH AL-DEEN BABYLON NAJAF SULAIMANIYA 0.5 DUHOK ANBAR ANBAR SULAIMANIYA 0 0 100 150 200 250 300 100 150 200 250 300 Mean per capita consumption expenditure Mean per capita consumption expenditure Wasting Fitted values Underweight children Fitted values Source: Authors’ calculations using IHSES 2012. Source: Authors’ calculations using IHSES 2012. Poverty in Human Capital 59 FIGURE 50: Stunting, Expenditures and Caloric Intake and the PDS Panel A Panel B 0.4 MUTHANNA KERBELA 0.4 MUTHANNA MAYSAN THI-QAR SALAH AL-DEEN KERBELA SALAH AL-DEEN BASRAH THI-QAR MAYSAN QADISIYA BASRAH QADISIYA 0.3 BAGHDAD 0.3 BAGHDAD WASIT WASIT KIRKUK KIRKUK NAINAWA NAINAWA DIYALA DUHOK DUHOK BABYLON DIYALA BABYLON 0.2 0.2 ANBAR ANBAR NAJAF ERBIL NAJAF ERBIL 0.1 SULAIMANIYA 0.1 SULAIMANIYA 100 150 200 250 300 8 9 10 11 12 13 Mean per capita consumption expenditure Mean per capita expenditure based on rations Panel C Panel D 0.4 KERBELA MUTHANNA 0.4 MUTHANNA SALAH AL-DEEN KERBELA MAYSAN MAYSAN THI-QAR BASRAH BASRAH THI-QAR SALAH QADISIYA AL-DEEN 0.3 QADISIYA BAGHDAD 0.3 BAGHDAD WASIT WASIT KIRKUK KIRKUK NAINAWA NAINAWA DUHOK DUHOK BABYLON DIYALA DIYALA BABYLON ANBAR 0.2 NAJAF ANBAR ERBIL 0.2 NAJAF ERBIL 0.1 SULAIMANIYA 0.1 SULAIMANIYA 2500 3000 3500 4000 1200 1400 1600 1800 Mean per capita total caloric intake Mean per capita caloric intake from rations Stunted Fitted values Source: Authors’ calculations, IHSES 2012. FIGURE 51: Ever Married Females by Division: Percentage by Age (12–18) associated with poor nutritional outcomes for chil- dren. The legal minimum age of marriage in Iraq is 16% 18 years for both men and women (Article 7 (1)), 14% although a judge can allow marriage at the age of 12% 15 years, if the person is deemed physically eligible. According to the National Strategy of Combating 10% Domestic Violence in Iraq of 2013, 5% of girls get 8% married before the age of 15 years old, and 23% of 6% girls get married before the age of 18 years old. In 4% both 2007 and 2012, around 12 percent of 16 year 2% old girls and 30 percent of 18 year old girls report 0% being married. In other words, by age 16, 5 percent 12 <=13 <=14 <=15 <=16 <=17 <=18 of girls are married; and by age 18, 11 percent of Kurdistan Baghdad North all girls are married. When we look at early mar- riage across the five divisions in 2012 (Figure 51), a Central South clear pattern emerges. In Kurdistan, where poverty Source: Authors’ calculations using IHSES 2012. 60 THE UNFULFILLED PROMISE OF OIL AND GROWTH and stunting and wasting are lower, the incidence Surprisingly, the highest rates of underage marriage of early marriage is also the lowest, with 4 percent in 2012 are in relatively well-off Basra, where 16 per- of girls being married by age 18 (and 12 percent cent of girls under the age of 18 are married. of 18 year old girls being married married). Early marriage rates are higher in every other division, Figure 54 shows the proportion of under-18 mar- and especially in the South, which has the worst nu- ried girls who have ever given birth to a child in tritional indicators, with 34 percent of 18 year old Iraq and in each division in 2007 and 2012. While girls being married, and 13 percent of all girls being there is some variation across divisions, on average, married by age 18. between 40 and 50% of those who get married be- fore the age of 18 also bear a child by that age. Early Overall, 11 percent of all girls below the age of 18 childbirth is not only more risky for the mother and in the 2007 and 2012 surveys report being married the child, childbirth at an early age can negative- (Figure 52). While this share has decreased in the ly affect the mother’s nutrition; and in turn, poor South, Centre and Kurdistan, it has increased in the nutrition of mothers can increase the likelihood of North, and especially in Baghdad, from 10 percent in anthropometric failure (stunting, wasting and un- 2007 to 13 percent in 2012. In 2007, Baghdad had derweight) (Box 3).25 a lower share of underage girls married when com- pared to the national average, while in 2012 its share In fact, it appears that the relatively high stunting has substantially increased, surpassing the national rates in some of the better off governorates, i.e., rate. The North division had in 2007 also a smaller governorates where stunting was high despite rela- percentage of underage married girls than the aver- tively high per capital consumption and food expen- age of the country, but in 2012 its share is coming diture, may be correlated with the relatively high closer to the country’s mean. Figure 53 shows the rates of early marriage and correspondingly, early share of girls married by the age of 18 in each gov- motherhood in those governorates. Figure 55 plots ernorate. In the North, the share of underage mar- the share of girls below the age of 18 who have ever ried girls has increased in every governorate, and in been married in each governorate with the stunting Anbar, has increased from 4 percent to 10 percent. rates in that governorate. Many southern governor- ates, which have relatively low consumption, also have relatively high early marriage rates, and both FIGURE 52: Ever Married Females Under factors are correlated with higher stunting rates. In 18 Across Time: Percentage by other cases, such as Basra and Salahadin, high stunt- Division ing is correlated not with lower consumption but with higher prevalence of early marriage. Iraq 0.11 0.11 Over and above poverty and its household (lower South 0.13 education, poorer health, and lower food intake 0.14 Central 0.12 and nutrition) and community level correlates 0.12 (limited access to quality healthcare), early moth- North 0.10 0.09 erhood is associated with a higher likelihood of ad- Baghdad 0.13 verse anthropometric outcomes for children. When 0.10 we consider the set of mothers aged 12–23 who Kurdistan 0.04 0.06 have children at 18 or younger who would be cov- 0.00 0.05 0.10 0.15 ered in the anthropometrics module (children aged 0–60 months) of the 2012 IHSES, they make up 2012 2007 25 http://www.ncbi.nlm.nih.gov/pubmed/21628349. Poverty in Human Capital 61 FIGURE 53: Ever Married Females Under 18: Percentage by Governorates 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% Nineveh Duhouk Erbil Suleimaniya Karkouk Salahuddin Babil Kerbala Wasit Diala Al-Anbar Al-Qadisiya Al-Muthanna Thi-Qar Missan Basrah Baghdad Al-Najaf Kurdistan North Central South Baghdad 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. FIGURE 54: Ever Given Birth to a Child: Share and caloric intake from PDS rations, education of of Married Females Under 18 the mother, work status of the head of household and per capita expenditure quintiles. The results 0.6 show that living in an urban area reduced the likeli- 0.5 hood of having a stunted child in the household by 5.2%. All division specific effects are significant 0.4 and all of the divisions are more likely than Kurd- 0.3 istan to have a stunted child in the household. The 0.2 effects are largest in Baghdad and the South, and 0.1 households living there are 17.6 and 20.4 percent more likely to have stunted children. Children be- 0 Kurdistan Baghdad North Central South Iraq longing to households at the bottom of the wealth distribution are more likely to be stunted. House- 2007 2012 2007 2012 holds belonging to the top 4th and 5th quintiles are 5 and 10 percent less likely to have a stunted Source: Authors’ calculations, IHSES 2007 and 2012. child. Neither per capita PDS expenditures on food nor the daily caloric intake per person from PDS 8.4% of their age group, but have a higher share of rations is a significant correlate of stunting. This children who are stunted relative to older mothers suggests that the governorate level correlations (29 percent relative to 26 to 27 percent for older mothers).26 26 We consider the 12–23 age group category to cover 12 year We examine the role of these factors in a probit olds who have just given birth and 23 year olds with 60 month old children who would have given birth when they regression model in predicting the presence of a were 18. The IHSES survey does not contain information stunted child at the household level (Annex Table on the mother’s age at first birth, and so we use the group of 3.1). The model includes location, household size, 12–23 year old mothers with 0–60 month old children as a the age group of the mother, percapita expenditures proxy for motherhood by age 18. 62 THE UNFULFILLED PROMISE OF OIL AND GROWTH BOX 3: Early Motherhood and Child Outcomes Early motherhood is often believed to cause adverse outcomes for children, such as lower birth weights, education, work experience and wages, more persistent poverty and welfare dependency, and higher rates of infant mortality. Evidence on this subject is, however, mixed. While on the one hand there are several studies arguing that the adverse effects of teenage childbearing primarily reflect unmeasured family background rather than the true consequences of a teen birth, other contributions show that controlling for family background does not fully eliminate the adverse consequences of early motherhood for children. Levine et al (2001) find that find that early motherhood’s strong negative correlation with children’s test scores and positive correlation with children’s grade repetition is almost entirely explained by pre-birth individual and family background factors of teen mothers themselves. However, controlling for maternal background, the authors show that children of teen mothers in the United States are more likely to initiate early sexual activity and to display problem behaviors such as truancy and fighting. Based on 1991 and 1999 data for the UK, Francesconi (1997) finds that being born to a teenage mother is usually associated with worse outcomes as young adults: lower chances of higher educational attainment, greater risks of inactivity and teenage childbearing, smaller probability of being in the top decile of the income distribution and greater probability of being in the bottom decile of the earnings distribution. These results hold true even after controlling for common family or maternal background factors. Geronimus et al (1992) estimates the relationship between maternal age and low birth weight, preterm birth, indicators of prenatal care utilization, smoking and alcohol use during pregnancy, breastfeeding, and well-child visits. The study controls for pre-pregnancy family background differences between teen and older mothers by comparing sisters who experienced their first births at different ages. The findings show evidence that maternal family background accounts for many of the health-related disadvantages of the firstborn infants of teenage mothers. Raj et al (2010) find that, in a nationally representative sample for India, children born to women married as minors are significantly more vulnerable to malnutrition than those born to women married at 18. The authors show that these associations were not simply a consequence of socioeconomic vulnerabilities of mothers, or of maternal malnutrition as indicated by low BMI. They hypothesize that perhaps young mothers may be unable to ensure adequate nutrition for their children, and that combined with the limited nutritional reserves stored within the bodies of adolescent mothers probably places their offspring at substantial risk for low birth weight and inadequate access to breast milk. The findings suggest that the effects of inadequate fetal nutrition and reduced breastfeeding among neonates born to adolescent mothers extend into infancy and early childhood, maintaining their ongoing risk for malnutrition related health problems and suggesting that such vulnerabilities accumulate over the lifespan. Source: Geronimus et al (1992), Levine et al (2001), Francesconi (1997), Raj et al (2010). FIGURE 55: Stunting Rates and Early Marriage Although the education of the mother and the by Governorate, 2012 employment status of the head of household head does not appear to have a significant effect, the co- 0.4 MUTHANNA efficients for the mother age groups’ variables are SALAH AL-DEEN KERBELA MAYSAN significant and confirm our hypothesis that younger THI-QAR BASRAH QADISIYA mothers are more likely to have stunted children. A 0.3 BAGHDAD WASIT mother aged 12–23 is 12 percent more likely and KIRKUK DUHOK NAINAWA a mother aged 24–30 is 7 percent more likely to DIYALA BABYLON 0.2 NAJAF have a stunted child relative to mothers aged 31 or ERBIL ANBAR above. These results suggest that low welfare levels 0.1 SULAIMANIYA (as measured by the consumption quintiles and the 0 0.5 0.1 0.15 0.20 division dummies) for certain households and early motherhood both adversely impact nutritional out- Ever married comes for children in Iraq. Stunted Fitted values Source: Authors’ calculations, IHSES 2012. Evidence from the Multiple Indicator Cluster Surveys with dependence on PDS for food, and lower food The Multiple Indicator Cluster Survey is a sur- expenditures with stunting rates reflect the role of vey program developed by the United Nations to location and stunting rather than that of the PDS. provide internationally comparable data on the Poverty in Human Capital 63 situation of children and women. In Iraq, the sur- Stagnation in Education Among vey was conducted in 2006 and 2011. Although Men; Some Improvement for Young there is some content overlap between MICS and Women IHSES surveys, the MICS covers women and child outcomes in greater detail. This section summarizes Education, one of the strongest correlates of pov- the highlights from the MICS findings based on erty, varies widely across Iraq, by division and by UNICEF´s 2012 report. gender. While there are a few signs of improve- ment, on average, very little has changed in terms Over 55,000 women in their reproductive age were of educational outcomes between 2007 and 2012 in interviewed to collect MICS 2011 data. Survey es- Iraq. The overall median educational level in Iraq is timates suggest a national population of 33.4 mil- complete primary school, which is also the median lion, of which 16.6 million are children and adoles- level for adult and youth male cohorts as well as for cents. One in every five Iraqis in 2011 was a child young females. For the female adult population, the under the age of 5. Around 800,000 children in median education is incomplete primary. Iraq have lost one or both parents. On the health dimension, the survey findings show that around This relative stagnancy in education overall and for 35,000 infants die every year, which is equivalent men, and the pattern of limited improvement for to approximately 37 deaths out of every 1000 live young women is evident in Figure 56 which looks births for children under 5. Only 1 in 10 children at the educational level for young (18–29) and adult are exclusively breastfed after three months of life, (30–64) cohorts in the 2007 to 2012 period. While and 1 in 4 children experience stunted growth. On- for the adult male population, illiteracy remained ly 5 out of 10 children aged 18–29 months received at 17%, for the younger male cohort, the distribu- all recommended immunizations at the correct tion has worsened as there is a higher share of il- time and acute respiratory infections and diarrhea literacy between 2007 and 2012 and also a smaller remain two of the biggest killer diseases of chil- share of young males are completing primary school. dren in Iraq. Taking a closer look at women, results This pattern of little movement in education among show that early marriage is pronounced; with 1 in younger male cohorts is repeated in almost every di- 5 young women aged 15–19 being married, with a vision, except for Kurdistan, where illiteracy rates are fertility rate of 4.4 on average. much smaller and the tertiary level rates are much higher for the younger cohorts, suggesting rapid im- Compared to 2006, the MICS 2011 show that in provements within the last five years (Figure 57). some areas there has been progress in Iraq: birth registration, gender parity in primary school and For the female adult cohort, outcomes have not im- registration and attendance in primary school have proved much over time. However there is a signifi- improved substantially. Slight progress was found cant difference in educational attainment across the in under 5 mortality rates, immunization coverage younger and older cohorts, with young women catch- and child labor. On the other hand, other dimen- ing up with their male counterparts. Between 2007 sions have experienced stalling or even regression. and 2012, the share of the youth cohort continuing Not much has changed between 2006 and 2011 in their studies after primary school has also registered terms of under nutrition, low weight rate at birth, a small increase. Unlike the stagnant picture among primary school completion on time, iodized salt consumption and attitude towards domestic vio- 27 Overall, 51 percent of women in Iraq feel that a husband lence.27 Worse outcomes were registered in 2011, has the right to beat his wife for at least one of five reasons: when compared to 2006, in terms of breastfeeding, (i) if she goes out without telling him; (ii) if she neglects her treatment of diarrhea, treatment of pneumonia and children; (iii) if she argues with him; (iv) if she refuses sex early marriage. with him and (v) if she burns the food. 64 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 56: Educational Level by Generation Over Time: Female and Male Population in Iraq 2012 26% 16% 23% 8% 10% 5% 12% 2012 16% 15% 25% 12% 15% 5% 12% Female youth Male youth 2007 19% 15% 33% 10% 8% 6% 9% 2007 11% 15% 30% 13% 13% 7% 11% 2012 42% 13% 23% 8% 4% 5%5% 2012 17% 13% 29% 11% 9% 9% 11% Female adult Male adult 2007 43% 10% 23% 8% 5% 6% 4% 2007 17% 10% 28% 14% 11% 9% 10% Illiterate Incomplete primary Primary Intermediate Secondary Higher Secondary Tertiary Source: Authors’ calculations, IHSES 2007 and 2012. men, improvement in education among younger Enrollments decline sharply after primary generations of women is evident in each division, but school especially in Kurdistan (Figure 58). Illiteracy rates To better understand education dynamics and gen- and incomplete primary schooling are much lower der disparities across Iraq, in in particular, to see among young women and the percentage of young why education does not go beyond primary school- females with tertiary education is quite large, particu- ing for the majority of the population, we now ex- larly in Baghdad, the Centre and Kurdistan (17 per- amine patterns and trends in enrollment. The net cent, 12 percent and 13 percent respectively). enrollment rate at any educational level is the share FIGURE 57: Educational Level By Generation FIGURE 58: Educational Level By Generation of Male Population and Divisions: of Female Population and 2012 Divisions: 2012 Adult 18% 15% 32% 11% 7% 9% 8% Adult 45% 15% 20% 8% 3% 5% 4% South South Youth 21% 16% 27% 12% 12% 6% 7% Youth 34% 19% 21% 7% 7% 4% 9% Adult 15% 12% 31% 11% 10% 10% 11% Adult 38% 14% 27% 8% 5% 5% 4% Central Central Youth 16% 16% 25% 11% 14% 4% 14% Youth 23% 17% 28% 8% 8% 4% 12% Adult 18% 14% 32% 10% 7% 8% 10% Adult 47% 13% 26% 4%3% 4%3% North North Youth 20% 16% 26% 9% 13% 4% 11% Youth 34% 17% 26% 5% 7% 4% 7% Adult 11% 10% 27% 15% 13% 9% 16% Adult Kurdistan Baghdad Kurdistan Baghdad 29% 12% 24% 12% 8% 7% 8% Youth 11% 14% 28% 16% 13% 3% 14% Youth 14% 14% 25% 12% 12% 6% 17% Adult 28% 19% 22% 9% 7% 8% 7% Adult 61% 12% 12% 4%3%5%3% Youth 12% 13% 19% 13% 24% 6% 14% Youth 24% 13% 15% 7% 21% 7% 13% Illiterate Incomplete primary Primary Illiterate Incomplete primary Primary Intermediate Secondary Higher Secondary Tertiary Intermediate Secondary Higher Secondary Tertiary Source: Authors’ calculations, IHSES 2012. Source: Authors’ calculations, IHSES 2012. Poverty in Human Capital 65 of children of official school-going age for that level FIGURE 60: Net Enrollment by Educational who are enrolled in the level, while the gross enroll- Level in Iraq: Population Aged 6 ment ratio is the share of children of any age who and Above are enrolled in that level. Thus, gross enrollment is always at least as large as (and typically larger than) 100% net enrollment as the former measure also accounts 90% 80% for all the students outside the official age groups 70% enrolled in a particular grade or level. 60% 50% 40% In Iraq, both gross and net enrollments display 30% a sharp decline after primary school. In primary 20% school, gross enrollment rates were 104 and 106 10% 0% percent in 2007 and 2012 respectively, i.e., for ev- Primary Intermediate Secondary Tertiary ery 100 students of primary school going age, 105 students were enrolled in primary school (Figure 2007 net 87% 38% 23% 8% enrollment 59 and Figure 60). Net enrollment rates in primary 2012 net 88% 43% 23% 10% school are also quite high: 88 percent of students of enrollment primary school age were in fact enrolled in primary school in 2012. However, at higher levels of educa- Source: Authors’ calculations, IHSES 2007 and 2012. tion, gross enrollment rates decline rapidly, and net enrollment rates fall even more sharply. For every secondary gross enrollment rates were 56 percent 100 students of intermediate-level school-going in 2012, this figure drops to 23% in terms of net age, 76 students of any age were enrolled in inter- enrollment. Between the 2007 and 2012 period, mediate school in 2012, but only 43 of them were while there have been some increases in intermedi- of intermediate school going age. Similarly, while ate and secondary gross enrollment, there has been little change in net enrollment rates. FIGURE 59:  Gross Enrollment by Educational The same pattern holds in each division in Iraq, Level in Iraq: Population Aged 6 with gross and net enrollment rates decreasing and Above sharply as the level of schooling goes up (Figure 61). The sole exception is Kurdistan, where gross 120% enrollment rates are above 100 percent even at the intermediate and secondary levels, the highest in 100% the country. Moreover, net enrollment rates for 80% post-primary education have increased substan- 60% tially in Kurdistan, from 48 to 61 percent at the 40% intermediate level, and from 23 to 38 percent at the secondary level (Figure 62). 20% 0% Typically, large differences between gross and net Primary Intermediate Secondary Tertiary enrollment rates signify the presence of overage chil- 2007 gross 104% 65% 46% 12% dren at that education level, because of grade rep- enrollment etition or late entry into the level. While these dif- 2012 gross 106% 76% 56% 15% ferences typically appear in other countries at higher enrollment levels of education, these are evident even in primary Source: Authors’ calculations, IHSES 2007 and 2012. school in Iraq. Figure 63 and Figure 64 show that, in 66 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 61: Gross Enrollment by Educational Level and Division: Population Aged 6 and Above 140% 120% 100% 80% 60% 40% 20% 0% Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Primary Intermediate Secondary Tertiary 2007 gross enrollment 2012 gross enrollment Source: Authors’ calculations, IHSES 2007 and 2012. 2012, 17% of those enrolled in primary school were At the same time, there exist significant gender dis- outside the official primary school age group (6–11 parities in enrollment. Figure 65 plots the difference years) and, among these, 93 percent are under the between male and female gross enrollment rates, age of 15. Thus, the large differences between pri- which are expressed as the difference that is due to mary school gross and net enrollment are primarily gender gaps in net enrollment rates and gender gaps explained by the presence of a significant number of in overage enrollment rates. Note that a negative overage students in primary school (rather than for net enrollment rate implies higher net enrollments instance, adult remedial education). These gaps get among boys relative to girls while a negative over- larger with education until secondary education. age rate implies a smaller share of overage girls at FIGURE 62: Net Enrollment by Educational Level and Division 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Primary Intermediate Secondary Tertiary 2007 net enrollment 2012 net enrollment Source: Authors’ calculations, IHSES 2007 and 2012. Poverty in Human Capital 67 FIGURE 63: Age Distribution of THose FIGURE 65: Gender Gaps in Enrollment in Enrolled in Primary School Iraq: Female Relative to Male Gross Enrollment – 2012 16% 17% Percentage point di erence between male 5% 0% and female gross enrollment –5% –10% –15% 84% 83% –20% –25% –30% –35% Primary Intermediate Secondary Tertiary 2007 2012 Net –8% –7% 1% 1% Ages 6–11 Ages 12+ enrollment Overage –11% –25% –16% –3% Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2012. that level relative to overage boys. For the country as a whole, in primary, intermediate and secondary school, gross enrollment rates are much lower for girls Despite these odds, among the few who make it to than for boys: by 19, 32 and 15 percentage points secondary and tertiary level, net enrollment rates are respectively. Disparities between boys and girls in net slightly higher among girls than among boys. This enrollment are pronounced on the primary and in- suggests that while girls are less likely than boys to termediate level; lower female net enrollment by 7–8 make it to higher education (because of gender gaps percentage points. However, the difference between that begin early in the education process), once they gross and net enrollments, of 11 and 25 percentage reach a higher level, girls are slightly better in com- points respectively, represents the significantly higher pleting each level on time and tend to lag behind less. rate at which overage boys are enrolled in primary and intermediate education grades relative to girls. Looking across divisions, females outperform males in terms of secondary and tertiary net enrollment except in the North, meaning that a higher share of FIGURE 64:  Overage Students Enrolled in girls of the correct age group are enrolled on sec- Primary School by Age Group ondary and tertiary school, particularly in Kurdistan and Baghdad. However, for intermediate and pri- mary levels, female enrollment is lower than male 1% enrollment in all divisions, particularly in the North 2012 6% and South. This indicates that a lower share of the 93% girls enter school and continue on to higher educa- tion, but among those few who enter, a larger share 1% of them achieve higher levels of education. 2007 5% 94% Why does education end with primary Ages 21+ Ages 16–20 Ages 12–15 school for so many? Median education levels among Iraqis are low be- Source: Authors’ calculations, IHSES 2007 and 2012. cause some children never enroll in school, and few 68 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 66: Enrollment by Age: Male FIGURE 67: Enrollment by Age: Female 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 6 8 10 12 14 16 18 20 22 24 6 8 10 12 14 16 18 20 22 24 2007 2012 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. children continue on beyond primary school, for a because there was no easily accessible school (19%); range of economic and social reasons. Figure 66 they had to work to support their family (13%) or and Figure 67 plot the share of boys and girls ages because the household could not afford school ex- 6–24 who are currently enrolled in school. While penses. The need to work to support the family and enrollment rates have increased at all ages between the unaffordability of education expenses are also 2007 and 2012, dropouts begin as early as 12 years important in explaining male dropouts, accounting for boys and 9 years for girls. Enrollments begin to for 63 percent of male dropouts above the age of 25 fall sharply below 90 percent by age 13 for boys and 41 percent of male dropouts ages 7–25 (Figure and by age 10 for girls. Only 60 percent of 16 year 69 and Figure 70). However, the single largest rea- old boys are enrolled in school, and only 43 per- son for boys dropping out of school is that they no cent of girls. In addition, among those who stay longer want to attend, which makes up more than enrolled, absenteeism increases at the age of 11 for 40 percent of dropouts of boys ages 7–25. boys and again from age 15 as they start to miss school systematically more than girls, although ab- senteeism among girls begins to increase from age 16 (Figure 68). Percentage of Students Missing at FIGURE 68:  Least 4 School Days in the Previous 13 percent of the Iraqi population aged 7–79 have Month by Gender and Age, 2012 never attended school. The reasons vary substantial- 5% ly across gender and generation. Although the ma- jority of the adult males who never attended school 4% state that it was because there was no easily acces- 3% sible school (43%), and because they had to work to 2% support their families (25%), the younger generation that has never attended school declare mostly they 1% are not interested (23%), or because of sickness or 0% disability (20%). Worryingly, issues of physical access 6 7 8 9 10 11 12 13 14 15 16 17 18 and the affordability of education remain relevant for Female Male the younger cohorts. More than 50 percent of 7 to 25 year old males who never attended school did so Source: Authors’ calculations, IHSES 2012. Poverty in Human Capital 69 For women, social reasons are the most important FIGURE 70: Reasons for Dropping Out of factor in never attending school and dropping out, School: Male Above the Age of 25 accounting for almost half the adult female popula- tion that never attended school or dropped out, and Other for around 40 percent of the younger cohort (Figure 16% I have to work to support my family 71 and Figure 72). While there is some improvement Does not want to 55% in access to schools, with 18 percent of females ages 21% 7–25 who never attended school reporting the lack of access as the main reason, compared to 29 per- cent of females aged 25 and above, 11 percent of the younger cohort report that their households could not afford the costs of schooling. Almost a quarter of young women who drop out of school state that HH could not a ord they did not want to continue their education; early school expenses marriage and having to help with household chores 8% account for a further 13 percent; and unaffordability Source: Authors’ calculations, IHSES 2012. and the lack of access for an additional 12 percent. When we look at across the different divisions, the main reason for female school dropout is “social rea- Broadly speaking, these self-reported reasons for sons”, which are cited by more than 44 percent of non-attendance and drop-out point to significant respondents, with the exception of Kurdistan, where differences for boys and girls, with girls being subject most women drop out of school because they do not to social pressures and the unwillingness of the fam- have the desire to continue studying. For men, in ev- ily to continue their education, and boys succumb- ery division boys drop out of school mainly because ing to economic pressures faced by the household they have to work to support their family, which var- that require them to look for work or to quit school ies from 39 percent in Kurdistan to 57 percent in because of the unaffordability of expenses. While the Baghdad. presence of an accessible school has improved be- tween older and younger cohorts, it still remains an issue. Not wanting to go to school or to continue FIGURE 69: Reasons for Dropping Out of education also accounts for a substantial share of re- School: Male Ages 7–25 sponses, which could imply poor quality of schooling or little perceived value and returns to education. Other 14% I have to work to support my family To understand the influence of these different fac- HH could not a ord 33% tors in determining whether an individual completes school expenses 9% primary school on time or not, we estimate a model that quantifies the marginal effects of wealth (as proxied by consumption quintiles), parents’ educa- tion, place of residence, and the gender of the indi- vidual. We restrict the analysis to young people be- tween the ages of 12 and 25, who are above the age by which primary education should be completed; Does not want to 44% and we also include terms to capture gender spe- cific differences in the role of household wealth and Source: Authors’ calculations, IHSES 2012. place of residence in determining primary school 70 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 71: Reasons for Dropping Out of for instance, children belonging to households in School: Female Ages 7–25 the top consumption quintile are 14 percent more likely relative to those belonging to the bottom HH could not a ord consumption quintile. Place of residence also mat- Other school expenses There is no easily 13% ters: Living in an urban area increases the odds of 7% accessible school on-time primary school completion by 8.5 percent; 5% while living in any division lowers the odds relative Household chores to living in Kurdistan. Girls are 14 percent less likely 6% to achieve this outcome relative to boys, except girls in the top quintile and those who live in Baghdad. Thus, while poverty is higher among less educated Social reasons households, poorer households are also less likely 39% to have completed the median level of education. Does not want to 23% Early marriage Self-reported reasons for dropping out and non-at- 7% tendance are also consistent with this finding. Over and above the affordability of education expenses, Source: Authors’ calculations, IHSES 2012. certain geographic areas—urban Iraq and Kurd- istan—are more favorable towards primary school completion, either because of relatively better acces- completion (Annex Table 3.2). Parental education sibility or better quality schooling. After controlling is strongly correlated with primary school comple- for household wealth, parents’ education, and lo- tion. An individual whose mother has completed in- cation, girls are still less likely to complete primary termediate or secondary school is 20 percent more school than boys, unless they belong to the wealthi- likely to complete primary school by the age of 12 est households or live in Baghdad. relative to one whose mother has primary or lower education. Children from wealthier households are also more likely to complete primary school on time: Widespread Access to Basic Services, but Little Improvement in Quality FIGURE 72: Reasons for Dropping Out of School: Basic health and education services appear to be Female Above the Age of 25 easily accessible for the households and reachable within minutes in every division of Iraq (Figure There is no easily Other 73). The average Iraqi household takes 23 minutes accessible school 7% 6% HH could not a ord to reach a public hospital (using the usual means Early marriage school expenses of transport available to them). Public hospitals 10% 6% are closer on average in Kurdistan (19 minutes away) than in the North (29 minutes away). In Household chores 8% terms of education services, both elementary and high schools are within easy reach everywhere, par- ticularly in Kurdistan, where elementary and high schools are reachable within 6 and 10 minutes re- Does not want to spectively. These distances are more pronounced in 15% Social reasons rural than in urban areas, but rural areas do not 48% appear to be completely isolated from health and education services. Poverty in Human Capital 71 FIGURE 73: Distance in Minutes to the Nearest Establishment – 2012 South 22.8 12.6 9.6 19.4 20.9 Urban 10.2 Central 13.1 9.9 7.9 29 North 12.6 9.3 23.7 31 Baghdad 11.7 8.8 Rural 16.2 19 11.2 Kurdistan 9.8 6 Public hospitals High school Elementary school Source: Authors’ calculations, IHSES 2012. However, access to water continues to be significant- Central division, the water access has increased but ly lower in rural areas. Figure 74 plots the share of a larger share of those that have the public network households with access to piped water through the as the main source of water state that it is insuffi- public grid: while access is above 90 percent in urban cient (62% in 2012 against 49% in 2007). In every areas, in Kurdistan and Baghdad, it falls to 63 per- other division, access has increased and the share of cent in rural areas. Although there is relatively broad users considering it insufficient has declined. How- access, and there have been improvements in access ever, in Kurdistan, North and Central divisions the since 2007, a substantial portion of households who percentage of households experiencing interruptions use the public network as the main source of wa- in public water supply more than once a week has ter consider it to be insufficient (Figure 75). In the increased to more than 70 percent (Figure 76). The increased frequency of interruptions appears to be FIGURE 74: Access to Publicly Provided Water (Share of Households Connected FIGURE 75: Percentage of Population THat to the Public Grid) have Public Network as the Main Source of Water and Consider it Insufficient 0.99 0.97 0.97 0.95 1.0 0.91 0.88 0.86 0.9 0.83 0.81 0.81 0.18 0.74 0.8 South 0.70 0.43 0.7 0.63 0.6 0.62 Central 0.51 0.49 0.5 0.4 0.1 North 0.42 0.3 0.2 0.2 Baghdad 0.47 0.1 0.0 0.18 Kurdistan 0.23 Kurdistan Baghdad North Central South Rural Urban 0 0.2 0.4 0.6 0.8 2007 2012 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2012. 72 THE UNFULFILLED PROMISE OF OIL AND GROWTH Percentage of Households FIGURE 76:  FIGURE 77: Access to Electricity Grid (Publicly Declaring Public Supply of Water Provided) Is Interrupted More THan Once a Week 1.00 0.99 0.99 0.99 0.99 0.99 0.99 0.97 1.0 0.85 0.9 0.82 0.81 1.0 0.73 0.8 0.9 0.66 0.7 0.8 0.55 0.6 0.69 0.7 0.7 0.7 0.65 0.5 0.6 0.6 0.51 0.4 0.5 0.5 0.5 0.3 0.4 0.35 0.4 0.2 0.3 0.1 0.2 0.0 Kurdistan Baghdad North Central Rural Urban South 0.0 Kurdistan Baghdad North Central South Rural Urban 2007 2012 2007 2012 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2012. 18). While only 15 percent of the poorest quin- driven by rural areas, where the corresponding indi- tile manages to get more than half a day of power, cator has doubled. even among the richest quintile, less than 40 per- cent receive the same. Similarly, around a third of Turning to electricity, there appears to have been a all households receive less than 8 hours of electric- significant expansion in connecting households to ity per day. the public grid: in each division and in rural and urban areas, almost all households report being FIGURE 78: Share of Households by Average connected to the public electricity grid (Figure 77). Hours of Electricity from the This would be a commendable achievement but Public Network Per Day in the connection to the grid is no guarantee of electricity Past Week supply in Iraq. In fact, on average, electricity supply is rarely higher than 12 hours, with the exception of Iraq 37% 40% 23% Kurdistan (Figure 78), where 93 percent of house- holds receive electricity for more than 12 hours. In sharp contrast, only 3 percent of households in Urban 36% 40% 24% Baghdad, and around a tenth of households in the Rural 39% 41% 20% Centre and the South receive power for more than 12 hours. Baghdad is by far the worst hit in terms South 8% 81% 11% of electricity, with more than three-quarters of all Central 53% 39% 9% households in the capital receiving less than 8 hours North 33% 41% 26% of electricity a day. Baghdad 76% 21% 3% Kurdistan 1% 6% 93% Variation in electricity supply across space seems to be the relevant metric of unequal access. Irrespec- < 8 hours 8–12 hours >12 hours tive of household wealth, few receive more than 12 hours of electricity on average per day (Table Source: Authors’ calculations, IHSES 2012. Poverty in Human Capital 73 TABLE 18: Average Hours of Power Supply by FIGURE 80: Access to Sewage (Public Network Consumption Quintile Disposal System) <8 8 to 12 >12 1(poorest) 34% 52% 15% 0.7 0.7 2 39% 45% 16% 3 40% 41% 20% 0.41 0.37 0.4 0.4 4 39% 33% 28% 5(richest) 33% 28% 39% 0.2 0.2 Source: Authors’ calculations, IHSES 2012. 0.1 0.1 0.1 Rural 0.02 0.01 North 0.0 Kurdistan Baghdad Central South Urban Garbage collection and the availability of sewage services have increased over time, but are still very 2007 2012 limited overall, reaching less than 50% of the popu- lation in both years (Figure 79 and Figure 80). The Source: Authors’ calculations, IHSES 2007 and 2012. exception is garbage collection in Kurdistan, that reached 80% of households in 2012, and sewage in Baghdad, that reached 70% of households. In par- ticular, rural areas seem to have almost no garbage although these complaints are much less prevalent collection and sewage services, and the North, Cen- in Kurdistan. tre and the South are also underserved. Similar to the other services, the quality of sanitation does not The access to some services varies significantly by appear to be satisfactory (Figure 81). Across all divi- wealth quintile (Figure 82). For instance, less than sions, households highlight being adversely affect- 70 percent of the poorest 20 percent has access to ed by issues related to inadequate sanitation, such publicly supplied water, while 90 percent of the as stagnant water and outlets of sanitary systems, top 60 percent is connected to public water supply. FIGURE 79: Access to Garbage Collection Percentage of Households FIGURE 81:  (Collected by Municipality or put Declaring Being Adversely Affected in Designated Container) by Sanitation Related Issues 45% 40% 0.8 35% 0.7 30% 0.53 25% 0.39 0.4 20% 0.3 0.3 0.3 15% 0.2 0.2 0.2 0.2 10% 0.06 0.03 5% 0% Kurdistan Baghdad North Central South Rural Urban Kurdistan Baghdad North Central Rural Urban South 2007 2012 Stagnant water Outlets of sanitary systems Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. 74 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 82: Access to Services by space and inadequate utilities. Dissatisfaction is ex- Consumption Quintile pressed by more than 40 percent of households in all divisions, although it is a little more pronounced in Baghdad, South and Central divisions (Figure 93% 92% 88% 83). These perceptions are borne out in objective 82% measures of space per capita displayed in Figure 84. 69% 59% Households have, on average, between 0.25 and 46% 0.35 bedrooms per capita, or alternately 1 bedroom 43% 35% 33% for every 3 to 4 household members, evidence that 30% 29% 25% 20% on average families live in very compact spaces across 13% the country, with no improvement over time. 1(poorest) 2 3 4 5(richest) Perceptions of housing quality in terms of inadequate Water Garbage collection Sewage space and inadequate utilities deteriorate significantly as per capita consumption declines, but especially Source: Authors’ calculations, IHSES 2007 and 2012. among the bottom 60 percent (Figure 85). More than six-tenths of households in the bottom 60 percent re- Irrespective of wealth, 70 percent of households port an inadequate number of rooms; and around 55 find that the supply of water is insufficient. Access to percent inadequate utilities, as compared to less than garbage collection by the municipality does not go 40 percent of the top quintile (Figure 86). above 60 percent even for the top quintile, but for the lowest quintile, only a fifth have access to pub- Thus, while education is one of the strongest and lic garbage collection. Similarly, connection to the most direct correlates of poverty in Iraq, it is but one public sewage network is only 43 percent among dimension of a significant deficit in human capital the top 20 percent, but falls even further, to only 13 and in access to quality services. Moreover, much of percent, among the poorest 20 percent. this deficit, a legacy of the past, has been long-stand- ing in nature, and few measurable improvements are Housing quality is also perceived to be low, with a evident in the 2007 to 2012 period. An exception large share of Iraqi households reporting inadequate is in the expanded access to certain services, but FIGURE 83: Housing Quality Perception – 2012 FIGURE 84: Number of Bedrooms Per Capita 0.50 0.7 0.45 0.61 0.61 0.61 0.57 0.6 0.40 0.49 0.49 0.49 0.48 0.35 0.44 0.5 0.40 0.30 0.4 0.25 0.3 0.20 0.15 0.2 0.10 0.1 0.05 0 0 Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Inadequate space Inadequate utilities 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. Poverty in Human Capital 75 FIGURE 85: Perceptions of Housing Quality FIGURE 86: Per Capita Number of Bedrooms by Quintile by Quintile 65% 63% 0.51 60% 0.49 57% 55% 54% 50% 45% 0.37 0.35 38% 0.31 37% 0.30 0.28 0.27 0.24 0.24 1(poorest) 2 3 4 5(richest) 1(poorest) 2 3 4 5(richest) Inadequate number of rooms Inadequate utilities 2007 2012 Source: Authors’ calculations, IHSES 2012. Source: Authors’ calculations, IHSES 2012. these have not been accompanied by improvements disadvantage in terms of primary school comple- in quality, measured either through perceptions or tion, except a lucky few in certain parts of the coun- objective measures. Poor outcomes in health and try and from relatively well-off households, who service delivery, like education, are correlated with continue further. welfare and with inequality across space. More aggregate indices reflect these findings as well. Early childhood nutrition outcomes, the lack of Inequality in access to basic child achievements is which can have irreversible long term consequences, explained by gender for education outcomes and are correlated not only with the household’s abil- by spatial inequities for housing and services relat- ity to consume adequate food, but also vary across ed indicators. A multidimensional index of human space, reflecting differential access to essential ser- development deprivations follows the same pattern vices, and with maternal nutrition. An important across space as consumption poverty. But there are finding in this regard is the positive relationship be- important differences. While the high level of con- tween early motherhood and child stunting, even sumption deprivation in the South is correlated with among otherwise similar households. multidimensional poverty, and the low consump- tion poverty rates in Kurdistan are accompanied by Education in Iraq stalls at primary school, as many limited deprivation in human development, in the children drop out or do not attend for a variety of Centre, welfare as measured by consumption has reasons—the lack of access to schools, the pres- rebounded faster, but significant deprivations in sure to support the family by working and the un- human capital remain. Overall, the analysis in this affordability of schooling, social norms about the chapter suggests that human capital and service de- value and appropriateness of girls’ education, and livery are important determinants of welfare; and a significant lack of interest. The latter may reflect go well beyond the relationship with consumption poor education quality, but these education out- poverty. In Iraq, three decades of violence and in- comes are more broadly also likely reflecting lim- security have stalled progress, and the nation faces ited returns to education on the labor market. Over a significant deficit, with far-reaching consequences and above these challenges, girls face a significant for the economy and for future generations. 4 Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare I n Iraq, the legacy of violent conflict has no doubt While Kurdistan and the South were both lagging be- had a pervasive influence on welfare and relat- hind the rest of the nation in terms of the educational ed outcomes through multiple channels–directly attainment of working age adults in 2007, they are through the loss of life and livelihoods and the displace- now on opposite trajectories. In Kurdistan, outcomes ment of people, and indirectly, through the destruction are improving significantly for the young, and they of infrastructure and markets, by limiting the access are catching up to the rest of the country. On the other and quality of health, education and basic services, by hand, while there is some improvement in education- adversely affecting rule of law and governance, and al attainment over cohorts within the South, the gap by severely constraining economic activity. with the nation is widening. Relative peace and stability alone, where experienced, There is also evidence of long-term deterioration in has not been sufficient for economic revival. Displace- health outcomes. Until the mid-1970s, Iraqi males en- ment and civilian deaths during the 2007 to 2012 joyed higher life expectancy than their counterparts in period have been concentrated in Baghdad, the North, the Middle East and North Africa (MENA) region. and in some of the Central governorates. The absence Since 1980, the beginning of the Iran-Iraq war, they of peace and security has implied little change in wel- have lagged behind. In 2011, Iraq’s Infant Mortality fare in Baghdad and the North, where the post-2003 Rate was the highest in the region, barring Yemen. violence was concentrated. It is only in the Centre where peace and stability have to some extent com- Perhaps the most direct correlate of poverty is employ- bined with an improvement in economic activity, and ment and the associated ability to earn income and fi- where job growth has outpaced the growth in the male nance consumption. Iraq has one of the lowest employ- working age population. While the South and Kurd- ment—to-adult population ratios in the region, and istan were both subject to severe prosecution under the male and female rates of employment and labor force Saddam regime, they have remained relatively un- participation are low and stagnant. Male labor force touched by the post 2003-violence; yet, they appear to participation was around 74 percent and female la- be on opposite trajectories. bor force participation around 11.5 percent in 2012. Male employment has not kept up with the growth in Spatial disparities in welfare may be driven by dif- working age male population in the South, and both ferences in human capital endowments across differ- have actually declined in Baghdad. In contrast, em- ent parts of Iraq. Iraq’s historical endowment and ployment growth outpaced growth in the working age comparative advantage in human capital has been population for men in Kurdistan, the North and the steadily eroded as a consequence of 30 years of violence, Centre, with the gap closing the fastest in the Centre. and some of these trends are evident across age cohorts. In the southern governorates, and with the exception 78 THE UNFULFILLED PROMISE OF OIL AND GROWTH of Basra, the last five years appear to have compounded life and livelihoods and the displacement of people, the neglect of the past, with declining male employ- and indirect channels, through the destruction of ment and labor force participation, declining female infrastructure and markets, by limiting the access employment in agriculture, and with young people and quality of health, education and basic services, falling further behind in human capital. by adversely affecting rule of law and governance, and by severely constraining economic activity (Fig- Household size and composition, education and la- ure 87). In this chapter, we examine some of these bor market outcomes all play a role in determining channels, insofar as data allows. consumption expenditure, the basis for measuring poverty, as do other location- specific factors that can Given the focus on the period from 2007 to 2012, imply access to (or lack of) services, employment op- on the face of it, one may expect that the recent portunities, and markets. By far the most consistent decade of violence and insecurity has been the most and striking correlate of poverty is the education of proximate factor at work in explaining the spatial the head of the household. Further, there is evidence dimensions of welfare in Iraq. In other words, that that the labor market is fragmented as a result of welfare stalled in areas which were particularly af- continuing violence and insecurity. While individu- fected by the post-2003 violence, and that out- als are able to move between rural and urban areas comes improved in places where security conditions within nearby governorates so as to equalize returns improved as normalcy returned. Below we show to their characteristics; moving across the country is that in addition, long term neglect of some parts of much harder. As a result, similar people have different the country and a recent revival in economic activity welfare levels depending on where they live. The lack of concentrated in other parts of the country may also internal integration has severely limited the potential have an important role in explaining the spatial dif- for development and sustained welfare improvements. ferences in welfare observed today. This chapter attempts to explain the observed spa- Displacement and civilian deaths during the 2007 tial disparities in welfare, and in particular, to ex- to 2012 period have been concentrated in Baghdad, amine the direct and indirect implications of three the North, and in some of the Central governorates decades of violence, conflict and insecurity on wel- (Anbar and Diyala). While the revival of economic fare. By construction, poverty is determined by the activity in Baghdad and the North has been relative- level of consumption expenditures and therefore, ly sluggish, in the Centre, employment among men, unequal growth in consumption across space and who make up a huge majority of the workforce, has time can directly influence poverty rates, trends and grown significantly faster than their working age spatial patterns in welfare. In much of the develop- population, suggesting a significant improvement in ing world, the poor also tend to have limited access economic activity during the 2007 to 2012 period. to health, education and other basic services, or to Starting from a common legacy of neglect and per- those of relatively lower quality, which in turn imply secution during the previous regime under Saddam lower human capital and a limited ability to take ad- Hussein, Kurdistan and the southern governorates, vantage of economic opportunities. Access to jobs which were relatively untouched by the post-2003 and earnings also directly determine the ability to violence, now appear to be on opposite trajectories. consume, acquire assets and invest, as do transfers, While the limited improvement in head count rates from the government or from private citizens. in Kurdistan is masking significant improvements in health and education, and in economic activity, in In Iraq, the legacy (and continuation) of violent the South, which has long been a lagging region, conflict have no doubt had a pervasive influence with the exception of Basra, continued neglect has on welfare and related outcomes through multiple led to deteriorating welfare for an already vulner- channels–the direct influence through the loss of able population. Consumption and labor incomes Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 79 FIGURE 87: Understanding Spatial Disparities in Poverty in Iraq Con ict and insecurity Education, Economic growth and Human capital health and labor markets other infrastructure Labor Economic growth; income Political economy and Governance Displacement, Consumption migration and expenditure and welfare social cohesion Private transfers Public transfers and non-labor income for the poorest 10 percent have actually declined GDP fell from $2836 to $466 or to one-sixth of its and the share of working age men who are not em- level. In the intervening years, no estimates of GDP ployed in the South has increased by 15 percent. per capita were available until recently. The 2003 invasion toppled the Saddam Hussein Violence and Insecurity government and sealed autonomy for the Kurdish region, but was also the beginning of a protract- Conflict, insecurity and civilian displacement in Iraq ed violent conflict among different power-seeking predates the 2003 US-led invasion. The almost de- groups. This internal strife has taken an immense cade long First Gulf War between Iran and Iraq in toll: since 2003, more than 110,000 civilians have the 1980s resulted in thousands of civilian casual- died as a result of violent attacks and millions of ties, a slowdown in per capita GDP, stalling the de- Iraqis became internally displaced or left for other velopment process. In July 1990, Saddam Hussein countries (predominantly Jordan and Syria).28 The invaded Kuwait, sparking the Second Gulf War and peak of civilian deaths in 2006–07 coincides with then the abortive anti-Saddam uprisings in southern the period covered by the first Iraq Household So- Iraq and the Kurdish region. This formed the begin- cio-Economic Survey (IHSES-I). During the time ning of autonomy for the Kurdish region, resulting of the second IHSES survey in 2012, violence still in major economic divergence between this region accounted for more than 4000 civilian deaths a year. and the rest of the country. Iraq became subject to a series of stringent United Nations resolutions which The scale of violence over the last decade has been included economically crippling sanctions and di- among the worst in recent times and civilians have rect UN involvement in the provision of food and paid a huge price in terms of dislocation and loss of health services. Importantly, other than the use of life. Iraqi refugees are among the most numerous in air power, the Saddam regime was free to use its remaining military power within Iraq outside the Kurdish region—to the detriment of southern Iraq 28 Iraq Body Count (www.iraqbodycount.org); estimates as of in particular. Between 1990 and 1998, per capita May 2013. 80 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 88: Number of Refugees, Selected Countries (>50,000 Persons) 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 0 Turkey Burundi Russian Federation Islamic Republic of Iran Ethiopia Rwanda Côte d'Ivoire Palestinian Various South Sudan Western Sahara Sri Lanka Mali China Central African Republic Eritrea Viet Nam Colombia Iraq Myanmar DRC Sudan Somalia Syrian Arab Republic Afghanistan Source: UNHCR Population Statistics Reference Database, United Nations High Commissioner for Refugees, year 2013; retrieved 29 July 2014. the world, ranking only behind a handful of coun- ‘displaced’ people make up 35 percent of all people tries (Figure 88 plots the number of refugees and who had lived elsewhere for a period of 6 months their country of origin in 2012 for selected countries or more. In Kurdistan, more than a third of the with more than 50000 refugees, of more than 200 population had lived elsewhere for at least 6 months countries in UNHCR’s Statistical Online Population (henceforth migrants), and of these, more than a Database in 2014). Similarly, the scale of internal third or almost 15 percent of the population re- displacement has been massive. By June 2014, the ported themselves as being displaced or returning number internally displaced people (IDPs) in Iraq from forcible displacement (Figure 90). In contrast, was estimated at 2.3 million or roughly 6 percent of Baghdad and the Northern division had the lowest the population—less only than those in Syria, Nige- rates of migrant population, but almost 45 percent ria, Columbia, Sudan, and the Democratic Republic of these individuals were either forcibly displaced or of Congo (Figure 89).29 The militant insurgency in were returning from forcible displacement. the northern governorates at the time of writing this report has displaced populations further. IDMC es- These estimates mask significant variation within timates that since late December 2013, more than divisions. For example, within the North, 12 per- 1.2 million persons have fled their homes in the gov- cent of the population of Kirkuk reported being dis- ernorates of Anbar, Nineveh, Salahaddin and Diyala. placed forcibly or having returned from forced dis- placement, and these individuals accounted for more In 2012 IHSES data, 6 percent of individuals re- than 60 percent of the migrant population (Figure ported having lived elsewhere for at least six months 91). While the overall rates of displacement in the because they were forcibly displaced or were return- Centre are relatively low, almost 10 percent of the ing from forcible displacement, or had moved for population of Diyala was displaced, accounting for security reasons, because of conventional armed conflict or civil conflict.30 This estimate implies a 29 Restricted to countries with more than 50,000 IDPs in scale of internal displacement very similar to those 2012, of the 60 countries listed in the database. implied by IDMC or UNHCR estimates, and re- 30 Overall, 17 percent of Iraqis reported having lived else- fers to movements spanning many decades. These where for a period of 6 months or more. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 81 FIGURE 89: Internally Displaced People, Selected Countries (>50,000 Persons) 7,000,000 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 0 Nepal Cote d'Ivoire Burundi Libya Chad Indonesia Sri Lanka Serbia Bosnia and Herzegovina Mali Philippines Peru Mexico Georgia Cyprus Guatemala Bangladesh Ethiopia Yemen Palestine Kenya Central African Republic India Azerbaijan Myanmar Afghanistan Turkey Somalia South Sudan Pakistan Iraq DRC Sudan Nigeria Colombia Syria Source: International Displacement Monitoring Center (IDMC), Global Database, retrieved July 29 2014, http://www.internal-displacement.org/8025708F004CE90B/(httpPages)/22FB1D4 E2B196DAA802570BB005E787C?OpenDocument. more than 60 percent of migrants. The lowest rates than 20 percent of all reported displacement oc- of displacement—2 percent of the population—were curred in 1991–92; 10 percent in 2003, and 14 per- in Najaf, Babylon, Qadisiya and Salahadin. cent in 2006. These time periods coincide with Saddam Hussein’s invasion of Kuwait in 1990, the These estimates of displacement are concentrated in US-led coalition invasion of Iraq in 2003 and the two distinct time periods: around the early 1990s peak in internal violence in 2006–07. and in the 2003 to 2007 period (Figure 92). More These different periods of displacement dispropor- tionately affected certain parts of the country. Figure FIGURE 90: Estimates of Migration and 93 shows the fraction of the displaced population Displacement, IHSES 2012 within each division by the year of having moved. The scale of movement also varied significantly over 0.50 time. The number of displaced individuals peaked 0.45 in 1991, and then in 2006, accompanied by signifi- 0.40 0.35 cant displacement in 2003 and 2007 (Figure 94). 0.30 In Kurdistan, the bulk of the displaced had moved 0.25 prior to the 2003 US led invasion of Iraq while on 0.20 0.15 the South, 60 percent of the displaced moved prior 0.10 to 2003, mostly in the 1980s and early 90s. These 0.05 movements coincided with the Kurdish and Shi’a 0 Iraq Kurdistan Baghdad North Central South uprisings of 1991. The uprisings were followed in the South by an intensification of the draining of the % lived elsewhere for at least six months Tigris-Euphrates marshes and the forced relocation % individuals displaced of Marsh Arabs; and following the establishment % lived elsewhere who were displaced of no-fly zones over the northern and southern Source: Authors’ calculations, IHSES 2012. parts of Iraq, by the establishment of the Kurdistan 82 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 91: Governorate Level Estimates of Displacement as a Share of THose Who Lived Elsewhere for More THan 6 Months 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 ERBIL SULAIMANIYA DUHOK BAGHDAD SALAH AL-DEEN NAINAWA KIRKUK NAJAF BABYLON KERBELA ANBAR WASIT DIYALA QADISIYA THI-QAR BASRAH MUTHANNA MAYSAN Kurdistan Baghdad North Centre South % lived elsewhere who were displaced % population displaced Source: Authors’ calculations, IHSES 2012. FIGURE 92: Spells of Displacement, 1970–2012 0.20 0.15 0.10 0.05 0.00 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Fraction of total displaced by year of moving Source: Authors’ calculations, IHSES 2012. FIGURE 93: Fraction of Displaced in Each Division, by Year of Moving, 1970–2012 0.5 0.4 0.3 0.2 0.1 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Kurdistan Baghdad North Centre South Source: Authors’ calculations, IHSES 2012. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 83 FIGURE 94: Estimates of Displacement, by Year of Moving, 1970–2012 2000 1800 1600 1400 1200 1000 800 600 400 200 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Total Kurdistan Baghdad North Centre South Source: Authors’ calculations, IHSES 2012. regional government. As we will try and show, since FIGURE 95: ‘Persons of Concern’, 2007–2012 then, while the Kurdistan region has witnessed steady improvement across a range of indicators, 2012 relative peace and stability in the South has, on the other hand, not been accompanied by any visible 2011 improvement in outcomes. 2010 2009 In contrast, in the other three divisions,—the North, Centre and Baghdad—, almost 95 percent of those 2008 displaced moved after 2003, and these movements 2007 spiked in 2003 and 2006–07, coinciding with the 0 1,000,000 2,000,000 US-led invasion and the subsequent surge in vio- lence within Iraq. Source: UNHCR Population Statistics Reference Database, United Nations High Commissioner for Refugees. Within the period covered by this report, UNHCR estimates on persons of concern within Iraq have Nineveh, which together accounted for 70 percent steadily increased since 2007, and have remained of deaths (Figure 97). Aggregating to the division high since the 2009 peak of almost 1.8 million per- level, over the 2007–2012 period covered by the sons (Figure 95).31 The bulk of these populations, two IHSES surveys, Baghdad, the Centre and the primarily comprising internally displaced people, North together accounted for 95 percent of civilian are concentrated in Baghdad and the governorates deaths (Figure 98). Within the Central division, the of Nineveh and Diyala (Figure 96). In each of the more than three-quarters of the casualties were in years between 2007 and 2012, two of these three two governorates, Anbar and Diyala. While internal provinces together accounted for more than 40 per- cent of all persons of concern in Iraq. 31 “Persons of Concern to UNHCR” is a general term used to Data on civilian casualties for the same period from describe all people whose protection and assistance needs Iraq Body Count finds an identical concentration are of interest to UNHCR, including asylum seekers, state- of violence and insecurity in Baghdad, Diyala and less people, internally displaced people and return refugees. 84 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 96: ‘Persons of Concern’ Concentrated FIGURE 98: Casualties by Division (%): in Baghdad, Nineveh and Diyala 2007–2012 0.45 800 Kurdistan 0.40 700 Centre 1% 0.35 600 31% 0.30 500 0.25 Baghdad 400 39% 0.20 0.15 300 0.10 200 0.05 100 0.00 0 South Nainawa Baghdad Baghdad Nainawa Diyala Baghdad Nainawa Baghdad Baghdad Nainawa Nainawa Baghdad 4% 2007 2008 2009 2010 2011 2012 North 25% % of persons of concern Number of persons of concern (thousands) Source: Iraq Body Count, 2013. Source: UNHCR Population Statistics Reference Database, United Nations High Commissioner for Refugees. Undoubtedly war, sanctions and violent insecurity in Iraq have had far-reaching and significant impacts on the nation as a whole. However, the spatial distri- violence has indeed declined since 2007, the pattern bution of violence and insecurity during the period of spatial concentration of violence has remained an covered by the surveys suggests that Kurdistan and enduring feature (Figure 99). In the 2011 Arab Ba- the South were relatively unaffected, at least accord- rometer survey in Iraq, while 40 percent of respon- ing to these measures. While the immediate explana- dents expressed insecurity about their own and their tion for the levels and trends in poverty in the South family’s safety, 60 percent of those living in Bagh- may lie in more structural factors rather than in the dad expressed the same concern, reflecting these violence of the recent decade, the latter may explain relatively higher risks. the limited improvement in welfare in Baghdad and FIGURE 97: Civilian Casualties by Governorate (percent), 2007 to 2012 45 40 35 30 25 20 15 10 5 0 ANBAR BABYLON BAGHDAD BASRAH DIYALA DUHOK ERBIL KARBALA KIRKUK MISSAN MUTHANNA NAJAF NINEVEH QADISIYA SALAHADIN SULAIMANIYA THI-QAR WASIT Source: Iraq Body Count, 2013. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 85 FIGURE 99: Share of Divisions in Total Civilian different parts of Iraq. It is widely acknowledged Deaths and Number of Deaths, that in the 1980s, Iraq had one of the best educa- 2007–2012 tional systems in the region. Iraq’s 1970 Provisional Constitution guaranteed the right to free education 25 25000 at all levels for all its citizens and stated that educa- tion was compulsory. In 1978, the state launched 20 20000 a mandatory campaign for combating illiteracy, in 15 15000 which it was obligatory for all Iraqi citizens between 10 10000 the ages of 15 and 45 to join. The program sup- ported participants until they achieved fourth-grade 5 5000 level of reading, writing and mathematics. The 0 0 Iran-Iraq war, the first Gulf war and the subsequent 2007 2008 2009 2010 2011 2012 sanctions, as well as the decades of violence that fol- Kurdistan Baghdad North lowed led to large scale destruction and deteriora- South Central Total Casualties tion in infrastructure and severe shortages of quali- fied teachers. Today, adult male and female literacy Source: Iraq Body Count, 2013. rates in Iraq are below the MENA average and well below the average for similar upper middle income the North over the 2007 to 2012 period, and the countries (Figure 100 and Figure 101). increase in poverty in Nineveh in particular. While long term trends on educational outcomes are not available for Iraq, data from IHSES provides Human Development and Access to some evidence of the long term impact of conflict. Basic Services Figure 102 plots the fraction of people in 5-year age cohorts in 2012 (birth years below age in 2012) with Spatial disparities in welfare may also be driven by a certain educational attainment. Overall, the inci- differences in human capital endowments across dence of illiteracy declines as we move to younger FIGURE 100: Adult Female Literacy, Iraq and other Countries Female literacy (ages 15+) 100 90 80 70 60 50 40 30 20 10 0 Angola Iraq MENA (all income levels) Middle income Libya Botswana Gabon Mauritius Ecuador Dominican Republic Turkey Upper middle income Jamaica South Africa Seychelles Mexico Colombia Jordan Albania Macedonia, FYR Costa Rica Bosnia and Herzegovina Serbia Romania Montenegro Argentina Bulgaria Hungary Turkmenistan Cuba Source: World Development Indicators, 2012. 86 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 101: Adult Male Literacy, Iraq and other Countries Male literacy (ages 15+) 100 90 80 70 60 50 40 30 20 10 0 Jamaica Angola Botswana Iraq MENA (all income levels) Middle income Dominican Republic Mauritius Seychelles Gabon Ecuador Colombia South Africa Mexico Libya Upper middle income Costa Rica Jordan Argentina Turkey Albania Romania Macedonia, FYR Bulgaria Hungary Serbia Montenegro Bosnia and Herzegovina Turkmenistan Cuba Source: World Development Indicators, 2012. cohorts, but for those in their mid-30s or younger improvements in tertiary education only evident in the 2012, it has increased relative to those in their among the youngest cohorts. Overall these trends late 30s. These cohorts were born after 1975, were suggest a stalling of progress in education and a pos- of school-going age during the Iran-Iraq war and sible worsening of outcomes for some cohorts. thereafter, when the mandatory literacy campaign became harder to implement. 12 percent of those Are spatial differences in education correlated with born in the 1980s, who might have been in primary patterns of poverty across Iraq? Table 19 shows the school at the time of the first Gulf war, are illiterate share of the working age population of Iraq in each in 2012. For Iraqis of all ages, the most prevalent education level in 2007 and 2012, followed by level of education is primary schooling or less: among the percentage point difference in education lev- 18–27 year olds, about 60 percent of have no more els across the working age population in each divi- than primary education. Over time, one may expect sion in 2007 and 2012, relative the Iraqi average. that while illiteracy and incomplete primary school- Overall, more than 80 percent of Iraqis of working ing decline, primary school completion increases. In age have secondary education or less; and there has Iraq, the trend is somewhat in the opposite direction, been an increase in the share with less than pri- with a relative increase in incomplete primary school- mary education from 34 to 40 percent in 2012. ing compared to primary completion with age. 18 Only two regions lag behind the national average. year olds in Iraq today are as likely to have completed People in the working age (15–64 years) living in primary school as those 30 years older. the South, where poverty rates increased, were 3 to 4 percentage points more likely to have less than The incidence of intermediate, secondary and high- primary education relative to the national average er secondary school taken together is around 30 and less likely to have primary or higher educa- percent for young Iraqis ages 18–22, the same as for tion. In 2007, Kurdistan’s working age population Iraqis in their 40s. While this in itself is worrying, was much more likely than that of the South to be within this group, over time, the share of higher sec- illiterate or have less than primary education. In ondary graduates has remained stagnant with some 2007, for instance, the share of working age adults Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 87 FIGURE 102: Current Educational Outcomes by Age in 2012 and Year of Birth, Ages 18–62* 100% 90% 80% 70% 0.17 60% 0.18 0.13 0.20 0.33 0.30 50% 0.28 0.33 0.26 0.15 0.22 40% 0.12 30% 0.12 0.15 0.16 0.15 0.16 0.48 0.15 20% 0.38 0.32 10% 0.22 0.18 0.19 0.21 0.22 0.19 0% 58–62 53–57 48–52 43–47 38–42 33–37 28–32 23–27 18–22 1950–1954 1955–1959 1960–1964 1965–1969 1970–1974 1975–1979 1980–1984 1985–1989 1990–1994 Tertiary Higher secondary Secondary Intermediate Primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2012. Note: * As younger cohorts may not have completed tertiary schooling, the results on tertiary education may not be comparable across ages and should be interpreted with caution. in Kurdistan with less than primary education was While Kurdistan and the South were both lag- 23 percentage points higher than the national aver- ging behind the rest of the nation in terms of the age of 34 percent. However, outcomes have been educational attainment of working age adults in improving over time and the deficit is being gradu- 2007, they are in fact on opposite trajectories. In ally bridged. In the North and the Centre, the gap Kurdistan, outcomes are improving significantly with the average appears to be closing. Despite no for the young, and they are catching up to the rest change in headcount rates, working age adults in of the country. On the other hand, while there Baghdad have the highest educational attainment is some improvement in educational attainment relative to the nation, but appear to be slowly los- over cohorts within the South, the gap with the ing their relative advantage except at the highest nation is widening. Figure 103 and Figure 104 levels of education. plot, for each 5-year age cohort in Kurdistan and TABLE 19: Completed Level of Education, Share of Iraqis Aged 15–64: Division Relative to National, 2007–2012 Education level (Share of working age population), All Iraq 2007 2012 Illiterate or incomplete primary 0.34 0.40 Complete primary to secondary school 0.47 0.45 Higher secondary and higher 0.16 0.15 Kurdistan Baghdad North relative Centre relative South relative relative to Iraq relative to Iraq to Iraq to Iraq to Iraq 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Illiterate or incomplete primary 23.01 17.81 –13.01 –12.51 4.85 3.60 –3.51 –4.07 4.27 3.35 Complete primary to secondary school –19.03 –14.26 12.62 7.29 –2.18 –1.18 0.11 3.91 –4.17 –1.75 Higher secondary and tertiary –5.92 –3.56 1.75 5.24 –2.13 –2.40 2.62 0.12 –0.27 –1.59 Source: Authors’ calculations, IHSES 2007 and 2012. 88 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 103: Educational Attainment by Age Cohort in 2012, Kurdistan Relative to Iraq 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 58–62 53–57 48–52 43–47 38–42 33–37 28–32 23–27 18–22 Tertiary Higher secondary Secondary Intermediate Primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2012. FIGURE 104: Educational Attainment by Age Cohort in 2012, South Relative to Iraq 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 58–62 53–57 48–52 43–47 38–42 33–37 28–32 23–27 18–22 Tertiary Higher secondary Secondary Intermediate Primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2012. in the South, by how much education levels are fathers were. A 60 year old in the South was 5 per- higher or lower relative to the national level. Older cent more likely to be illiterate and 2 percent less generations in Kurdistan (those 30 and above in likely to have secondary education than an average 2012) started out with much higher levels of il- 60 year old Iraqi while a 20 year old person is 8 literacy and incomplete primary education relative percent more likely to be illiterate and 6 percent to Iraq and lower levels of complete primary edu- less likely to have secondary education.32 cation and higher education. In contrast, the pat- tern is reversed for those in their 20s. These young people are much more likely than their counter- 32 Some age groups in the South appear to have bridged the parts to have secondary and higher levels of edu- gap with the rest of Iraq, but this is likely a result of the cation. In the South, in contrast, young people in overall stalling of progress in education than improvements their 20s are further behind their peers than their in the South. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 89 FIGURE 105: Educational Attainment by Age Cohort in 2012, Baghdad Relative to Iraq 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 58–62 53–57 48–52 43–47 38–42 33–37 28–32 23–27 18–22 Tertiary Higher secondary Secondary Intermediate Primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2012. FIGURE 106: Educational Attainment by Age Cohort in 2012, North Relative to Iraq 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 58–62 53–57 48–52 43–47 38–42 33–37 28–32 23–27 18–22 Tertiary Higher secondary Secondary Intermediate Primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2012. In Baghdad, starting from a position of significant In the North, which was also affected by the advantage, where 50 year olds where less likely post-2003 violence, deterioration in education over to be illiterate and more likely to have completed generations is starker (Figure 106). The cohorts tertiary education relative to the average, younger in their 30s and older in 2012, were more likely cohorts are moving closer to the national average than their counterparts to have completed primary (Figure 105). A 58–62 year old in Baghdad was 15 and higher education, and less likely to be illiter- percent less likely than the average to be illiterate, ate. But for those below the age of 30, the situa- compared to an 18–22 year old, who is only 7 per- tion has worsened considerably. An 18–22 year old cent less likely. While individuals living in Baghdad in the North was 20 percentage points less likely to continue to be more educated than those in other have completed secondary school and 26 percentage parts of the country, the erosion of this advantage points more likely to have primary education or less over time may be related to the continuing violence compared to their Iraqi counterparts. In the Central and insecurity. division, there is no clear trend of improvement or 90 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 107: Educational Attainment by Age Cohort in 2012, Centre Relative to Iraq 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 58–62 53–57 48–52 43–47 38–42 33–37 28–32 23–27 18–22 Tertiary Higher secondary Secondary Intermediate Primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2012. deterioration across time, although it appears that While across the region, these rates have come down for all age groups, education levels have been fairly sharply, Iraq has not brought down IMR at the same similar to the national average (Figure 107). pace. In 2011, Iraq’s IMR was the highest in the re- gion, barring Yemen (and possibly Djibouti). There is also evidence of long term deterioration in health outcomes. Until the mid-1970s, Iraqi males There is also evidence of deteriorating nutritional enjoyed higher life expectancy than their counter- outcomes for young children: according to a 2007 parts in the Middle East and North Africa (MENA) World Food Program report, 22% of children aged region. Since 1980, the beginning of the Iran-Iraq 0–5 were stunted; and almost 10 percent of children war, they have lagged behind (Figure 108). As may were underweight.34 As noted in chapter 2, more be expected in periods of violent conflict, this trend recent analysis of IHSES 2012 not only reveals high is linked to the increased rates of adult male mortal- rates of stunting (low height for age, and indica- ity since 1980 (since when data is available, Figure tor of chronic malnourishment) and wasting (low 109). Despite the Iran-Iraq war, adult male mortal- weight for height), but also important spatial dis- ity fell slightly between 1980 and 1990. This trend parities in these outcomes (Table 20). 35 percent of appears to continue until the mid-1990s. However, children aged 0–5 are stunted in the South, double since 2002, Iraq has witnessed a sharp increase in the rate in Kurdistan and significantly higher than male mortality: from 167 per 1000 to 295 per 1000 that in other divisions. The prevalence of wasting in 2011. A WFP survey reports that in 2007, 4 per- and underweight children is also significantly higher cent of household members below the age of 18 in the South than in the other divisions. were orphans, with 4 in 5 having lost their father. 33 Stagnation and deterioration in human capital mat- Worryingly, the adverse impacts of the protracted in- ter especially because they are currently affecting security on health and basic services infrastructure is also reflected in rising adult female mortality since the 33 Orphans are defined here as children who have lost at least 1980s, and a deterioration in infant mortality relative one parent. World Food Program, 2007. Comprehensive to the rest of the region. In 1960, Iraq’s ranking by Food Security and Vulnerability Analysis: Iraq. infant mortality rate within the region was roughly in 34 World Food Program, 2007. Comprehensive Food Security the middle (World Development Indicators, 2012). and Vulnerability Analysis: Iraq. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 91 FIGURE 108: Male Life Expectancy, Iraq Versus MENA, 1960–2011 80 70 60 50 40 30 20 10 0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Iraq Mena Source: World Development Indicators (2012). FIGURE 109: Mortality Rate, Adult (per 1,000 TABLE 20: Anthropometrics, Children Aged Adults), Iraq, 1980–2010 0–60 Months (share) 400 Underweight Stunted Wasting children 300 Kurdistan 0.17 0.07 0.05 200 Baghdad 0.30 0.08 0.09 100 North 0.25 0.07 0.08 0 Central 0.26 0.07 0.09 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 South 0.35 0.11 0.17 Mortality rate, adult, female (per 1,000 female adults) Mortality rate, adult, male (per 1,000 male adults) FIGURE 110: Population by Age Group, Source: World Development Indicators (2012). 2005–2010 5,000,000 4,500,000 the younger cohorts of Iraq’s working age popula- 4,000,000 tion. Almost half of the population of Iraq is be- 3,500,000 low the age of 30, and the generational deficit in 3,000,000 human capital will have significant implications for 2,500,000 2,000,000 the future (Figure 110 and Figure 111). Young 1,500,000 Iraqis are entering the labor market with much the 1,000,000 same education as their fathers and mothers. In 500,000 0 the South in particular, which accounts for a fifth 0–4 10–14 20–24 30–34 40–44 50–54 60–64 70–74 80–84 90–94 100+ of the national population, there is little evidence of improvement, even for younger cohorts, who 2005 2010 continue to fall behind their peers in the rest of the country. Source: UN DESA. 92 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 111: Population Share, 2010 decades, the adult male employment-to-popula- tion ratio has remained stagnant at 58 percent, and 90+ at the same time, male labor force participation 80–89 (LFP) has declined, especially among the youth 70–79 (Figure 113). While the latter could suggest in- 60–69 50–59 creasing years of education, it is also likely that the 40–49 ten percentage point decline in male youth LFP 30–39 is because the decades of insecurity, violence and 20–29 10–19 limited job opportunities have left young people 0–9 frustrated and discouraged. As in the rest of the 0% 20% 40% MENA region, gender differences in labor market participation are striking in Iraq. Contrary to men, % population, 2010 data from the World Development Indicators sug- Source: UN DESA. gest that adult female labor force participation has been slowly increasing over time, albeit from much lower levels. Labor Market Outcomes Two rounds of IHSES data reveal a more de- Perhaps the most direct correlate of poverty is em- tailed picture of labor market outcomes for men ployment (or the lack thereof) and the associated and women in the working age population (aged ability to earn income and finance consumption. 15–64) in Iraq (Figure 114). Based on a seven Iraq has one of the lowest employment—to-adult day recall period, a staggering 90 percent of Iraqi population ratios in the region; only the Palestin- women of working age are not in the labor force, ian Territories have a lower rate (Figure 112). This and only 10 percent are employed, with the bulk of is also in contrast to many of the other oil-rich them employed in part-time jobs. For men as well, countries in the region. Moreover, in the last two labor force participation has been stagnant around FIGURE 112: Employment to Population Ratio (ages 15+), 2011 100 90 80 70 60 50 40 30 20 10 0 Qatar United Arab Emirates Kuwait Bahrain World Oman Saudi Arabia Libya Morocco Egypt, Arab Rep. Middle East & North Africa (all income levels) Yemen, Rep. Tunisia Iran, Islamic Rep. Syrian Arab Republic Algeria Jordan Iraq West Bank and Gaza Source: World Development Indicators (2012). Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 93 FIGURE 113: Labor Force Participation, by Age and Gender, 1990–2011 80 70 60 50 40 30 20 10 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Labor force participation rate for ages 15–24, female (%) Labor force participation rate for ages 15–24, male (%) Labor force participation rate, Labor force participation rate, female (% of female population ages 15–64) male (% of male population ages 15–64) Source: World Development Indicators (2012). 70 percent.35 The only positive sign has been a shift FIGURE 114: Labor Market Outcomes, Share from part-time work to full-time work for men: in of Working Age Population 2007, 54 percent of employed Iraqi men were in (ages 15–64), 2007–2012* full-time jobs; by 2012, this rate had increased to 63 percent. 1.0 0.89 0.89 population out of the labor force 0.8 Other labor market indicators suggest a worsen- Fracton of working age 0.6 ing of outcomes. While unemployment rates for men have not changed much and remain low 0.4 0.30 0.30 (similarly, very few women who are not employed 0.2 and of working age report that they are looking 0 for work), more than 40 percent of men and 15 2007 2012 2007 2012 percent of women in part-time jobs stated that Females Males they were looking for more work (Figure 115). 0.8 This measure of involuntary underemployment 0.7 0.67 0.67 has increased over time, especially for women. 0.6 And among the youth aged 15–29, 72 percent 0.5 0.42 of women and 18 percent of men were neither in 0.4 0.36 0.30 school nor employed in 2012. At the same time, 0.3 0.24 0.2 unemployment rates for young men and women in 0.1 0.090.11 0.080.11 this age group were only 5 and 1 percent respec- 0.02 0.02 0 tively. Combined with the rates of low labor force 2007 2012 2007 2012 participation, this suggests significant labor market Females Males discouragement among young Iraqis. Employed full time Employed part time Employment to working age ratio Note: * ILO definition, 7 day recall. 35 Students are not included in the labor force. 94 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 115: Unemployment, poverty increased, it is apparent that not only is Underemployment and male participation in the workforce the lowest in the Joblessness* country, it has been declining: in 2012, 35 percent of men of working age in these governorates were 0.8 0.76 out of the labor force, an increase of 5 percentage 0.72 0.7 points since 2007. 0.6 0.5 0.43 0.41 Increasing male labor force participation in Kurd- 0.4 0.3 istan and the North has been accompanied by an 0.22 0.2 0.18 increase in male employment rates, and a relative 0.1 0.05 0.07 0.03 shift towards full time work, especially in the North, 0.00 0.00 0.03 0 2007 2012 2007 2012 where 46 percent of men of working age were work- Females Males ing in full time jobs in 2012 (Table 21). In Bagh- Unemployment (% of working age population) dad, there was a significant drop in part time male Involuntary underemployment (% of part time employed) employment, by 13 percentage points and a large Out of school, out of work (% youth in working age) increase in full time employment by ten percentage points. On the other hand, in the South, while male Source: Authors’ calculations, IHSES 2007 and 2012. Note: * ILO definition, 7 day recall. employment has declined, so has the unemployment rate, suggesting that the increase in men out of the labor force is due to discouragement. These south- Across Iraq, one may expect that areas where em- ern governorates also had the lowest rates of full time ployment rates are higher than average and grow- employment in 2012, and among the lowest rates of ing and where labor force participation is relatively part-time employment. Male underemployment is high and increasing, will also be areas where poverty significant across Iraq: even in Kurdistan, which has headcount rates are declining. This is true to some the lowest rates of involuntary underemployment, a extent. Male labor force participation is the highest quarter of those in part-time jobs would like to work in the Central division, and increased between 2007 more and cannot find work. In the Centre and the and 2012 in Kurdistan and the North, to reach the South, over half of part-time workers are involun- national average of 70 percent (Table 21). In the tarily underemployed by this measure. South, in contrast, the share of men out of the labor force has actually increased. Excluding Basra, and For women, labor force participation has remained focusing on the four southern governorates where very low and has not changed much over the 2007 TABLE 21: Employment Status, Men in the Working Age (aged 15–64), by Division, 7 Day Recall Kurdistan Baghdad North Centre South South excl Basra 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Full time employed 0.37 0.40 0.31 0.41 0.34 0.46 0.44 0.41 0.40 0.39 0.37 0.36 Part time employed 0.27 0.27 0.37 0.24 0.30 0.21 0.26 0.26 0.26 0.24 0.27 0.25 Employment to working age 0.64 0.69 0.68 0.66 0.64 0.68 0.70 0.69 0.66 0.64 0.64 0.62 ratio Out of the labor force 0.33 0.30 0.28 0.30 0.33 0.30 0.27 0.29 0.29 0.33 0.30 0.35 Unemployment 0.02 0.01 0.04 0.04 0.03 0.03 0.02 0.03 0.04 0.03 0.06 0.03 Involuntary underemployment 0.35 0.24 0.33 0.41 0.37 0.38 0.53 0.52 0.54 0.53 0.54 0.52 Source: Authors’ calculations, IHSES 2007 and 2012. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 95 TABLE 22: Employment Status, Women in the Working Age (aged 15–64), by Division, 7 Day Recall Kurdistan Baghdad North Centre South South excl Basra 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Full time employed 0.02 0.02 0.02 0.03 0.01 0.02 0.02 0.02 0.01 0.02 0.01 0.01 Part time employed 0.09 0.10 0.09 0.08 0.09 0.07 0.11 0.09 0.08 0.07 0.09 0.08 Employment to working age 0.11 0.13 0.10 0.11 0.10 0.09 0.13 0.11 0.09 0.09 0.10 0.09 ratio Out of the labor force 0.89 0.87 0.89 0.88 0.89 0.90 0.86 0.89 0.90 0.91 0.89 0.90 Unemployment 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.01 0.01 0.00 0.01 0.00 Involuntary underemployment 0.14 0.08 0.04 0.06 0.03 0.03 0.06 0.07 0.03 0.09 0.03 0.11 Source: Authors’ calculations, IHSES 2007 and 2012. to 2012 period. Of the few women who partici- southern governorates except Basra, where invol- pate in the labor force, the majority are engaged untary underemployment increased from 3 percent in part-time work in each of the divisions (Table of part-time female workers in 2007 to 11 percent 22). Employment rates among women of working in 2012. age vary from 13 percent in Kurdistan to 9 percent in the North and the South. Among women who Young men in general appear to have worse labor are employed part-time, the incidence of invol- market outcomes than their older counterparts. untary underemployment—those who work less Among men aged 15–29, full time employment than 40 hours a week and are looking for more varies from 39 percent in the North to 30 percent work—has declined from 14 percent to 8 percent in the South in 2012, compared to 50 percent for of part-time employment in Kurdistan, which is in adults aged 20–64 (Figure 116). On average, 19 line with other evidence that this labor market has percent of young men have part-time jobs, and been performing relatively well. In all other divi- among these men, rates of involuntary underem- sions, the opposite is true, especially in the four ployment are high. With the exception of Kurdistan, FIGURE 116: Employment Status, Young Men (aged 15–29) in the Working Age, by Division 0.8 0.69 0.70 0.70 0.68 0.69 0.7 0.62 0.66 0.61 0.58 0.6 0.52 0.54 0.55 0.53 0.55 0.55 0.53 0.51 0.50 0.47 0.5 0.47 0.45 0.46 0.4 0.37 0.3 0.26 0.2 0.1 0 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Kurdistan Baghdad North Centre South South excl Basra Employment to working age ratio (Young men) Involuntary underemployment (Share of part time employed) Source: Authors’ calculations, IHSES 2007 and 2012. 96 THE UNFULFILLED PROMISE OF OIL AND GROWTH more than 40 percent of young men in part-time FIGURE 117: Percentage Change in Non- jobs work fewer hours than they would like to. In Employment, Employment, the South, involuntary underemployment is as high Working Age Population and as 70 percent among young men. Employment to Working Age Ratio between 2007 and 2012, In 2012, 42 percent of young Iraqi men and 93 by Division, for Men (15–64) percent of young women were out of the labor Based on One Year Recall force (neither employed nor looking for work). As is typical, this measure counts students as being 35 30 out of the labor force, and as many young people 25 tend to still be in school, it may overestimate in- 20 activity among the young. A useful measure of in- 15 activity in this context is the proportion of young 10 men and women who are neither in school nor at 5 0 work. This then includes young people who are –5 actively seeking employment, as well as those who –10 are not, for various reasons including discourage- Kurdistan Baghdad North Central South ment. Overall, 72 percent of women and 18 per- % change in non-employed men cent of men in the 15 to 29 age group fall in this % change in employed men category, with lower rates for Kurdistan and rela- % change in working age men tively higher rates in the other divisions. For young % change in male employment to working age ratio women, the main reason cited for not looking for Source: Authors’ calculations, IHSES 2007 and 2012. work is social reasons and being a housewife. In contrast, for young men, the bulk those who fall in this category say they are not actively looking for a job because they cannot find a job, and this is age population in the South grew by 8 percent, but another indicator of labor market discouragment the share of employed men increased by only 4 per- among young men. cent. As a result, the proportion of men of working age who were unemployed or out of the labor force These estimates of labor market outcomes are based increased by 15 percent, and the male employment on the standard ILO definition and use a 7 day recall to working age ratio fell. In Baghdad, there was a 4 period. The IHSES surveys also ask respondents to percent decline in the male working age population report on any work for pay over a 12 month period, and a 5 percent decline in male employment. which allows us to estimate measures of long-term or ‘usual employment’. These estimates are broadly In contrast, employment growth outpaced growth consistent with the ILO measure, with male labor in the working age population for men in Kurdistan, force participation of 74 percent and female labor the North and the Centre, with the gap closing the force participation of 11.5 percent in 2012. They do fastest in the Centre. In the Centre, there was in fact, reveal starker disparities across the five divisions, and a 4 percent increase in the male employment rate, as seem to be in line with the trends in poverty rates a result of faster job growth relative to working age across space and time. In what follows, we focus on population growth. In Kurdistan, which witnessed men who make up an overwhelming majority of the largest increase in the working age male popula- the labor force. Male employment has not kept up tion of 30 percent between 2007 and 2012, male with the growth in working age men in the South, employment also increased at a slightly higher rate, and both have actually declined in Baghdad (Figure leading to a small increase in the employment to 117). Between 2007 and 2012, the male working working age ratio for men. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 97 The bulk of this increased employment for men in accounting for almost a fifth of male employment the Centre and in Kurdistan was concentrated in the each, very few jobs were added. Consequently, in financial, insurance and professional services sector, the South, more than 80,000 additional men were which now accounts for 16 percent of male employ- not employed in 2012 relative to 2007. In Bagh- ment, where 140,000 jobs were added in each divi- dad, male employment in the public administra- sion (Figure 118). This sector accounts for a fifth tion, health and education and commerce and retail of all male employment in Kurdistan. While male sectors declined sharply, while male employment employment declined in agriculture and in public in construction and financial services increased. In administration in the Centre and the North, jobs total however, male employment fell in Baghdad, were added in manufacturing, construction, com- and it is only because of the slightly larger decline merce and retail and in transport, storage and com- in male working age population that the number of munication. The net increase in male employment non-employed men in Baghdad also decreased. was largest in magnitude in Kurdistan, but it was in the Centre where it significantly outpaced the work- ing age male population, and as a result, the Centre Consumption, Income and Transfers was the only division where the increase in non em- ployed men was very small. Inequality in Consumption Growth The spatially uneven pattern of poverty reduction is On the contrary, in the South, the increase in non also in part, a story of unequal growth in consump- employment (almost all men dropping out of the tion: across quintiles and across space. Between 2007 labor force) was accompanied by declining male and 2012, consumption grew faster for Iraq’s rela- employment in agriculture, a sector which account- tively better off, amongst the highest quintiles. But ed for 10 percent of jobs for men in the South in it also grew where consumption levels were lower to 2007 (Figure 119). In the two largest employment start with: in rural Iraq and in the RoI. Among the sectors, construction and public administration, governorates where poverty rates increased, Nineveh, Changes in the Number of Jobs for Men by Sector of Employment between 2007 FIGURE 118:  and 2012 in Kurdistan, the North and the Centre (One Year Recall) 200,000 150,000 100,000 50,000 0 –50,000 –100,000 Not employed Agriculture & nishing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & communication Financial, insurance & professional services Public administration, health & education Others services Kurdistan North Centre Source: Authors’ calculations, IHSES 2007 and 2012. 98 THE UNFULFILLED PROMISE OF OIL AND GROWTH Changes in the Number of Jobs for Men by Sector of Employment between 2007 FIGURE 119:  and 2012 in Baghdad and the South (One Year Recall) 100,000 80,000 60,000 40,000 20,000 0 –20,000 –40,000 –60,000 –80,000 –100,000 Not employed Agriculture & nishing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & communication Financial, insurance & professional services Public administration, health & education Others services Baghdad South Source: Authors’ calculations, IHSES 2007 and 2012. Thi-Qar and Missan also experienced increases in in- between 2007 and 2012 for the lowest deciles, equality as measured by the Gini coefficients. while growth in consumption among the top deciles is on par with the Centre. This declining With the exception of Kurdistan and the Central consumption among the bottom 40 percent of the division, consumption per capita grew faster for distribution is directly reflected in the increase in the well-off than for the less- well off, as is evi- headcount poverty rates. dent in the growth-incidence curves for Iraq as a whole and for the five divisions (Figure 120). In contrast, the growth-incidence curves for Kurd- Incomes and Transfers istan and the Centre are relatively flat, indicating The trends in male employment and labor force that consumption grew evenly across the distri- participation across divisions appear to be reflected bution. The major difference between Kurdistan in changes in per capita labor income, especially and the Centre is the higher rate of consumption among the lower deciles of the consumption distri- growth in the latter: almost entirely throughout bution. Figure 121 shows the changes in per capita the consumption distribution, growth was around labor income for the bottom three deciles of the 4 percent in the Centre, compared to 1 percent or consumption distribution for Iraq as a whole, and less in Kurdistan. This is reflected in the significant for each of the divisions. In Iraq as a whole, per decline in headcount rates in the Central gover- capita labor income increased for the bottom three norates and the limited improvements in poverty deciles, although the increase was smallest for the in Kurdistan. In Baghdad and the North, while bottom 10 percent of the consumption distribu- overall consumption growth was positive, there tion. In the Centre and in Kurdistan, the bottom was almost no change in consumption among 3 deciles experienced significant increases in per the lower deciles of the distribution, and this is capita labor income, exceeding the national aver- in turn, captured in the trends in poverty. In the age; whereas in the North, labor incomes appear to South, consumption growth was actually negative have increased on par with the national average. In Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 99 FIGURE 120: Growth Incidence Curves – National, Divisional Iraq: Pro-rich Kurdistan: Neutral 7 7 6 6 5 5 Annual growth rate (%) Annual growth rate (%) 4 4 3 3 2 2 1 1 0 0 –1 –1 –2 –2 –3 –3 –4 –4 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Per capita expenditure percentiles Per capita expenditure percentiles Baghdad: Pro-rich North: Pro-rich 7 7 6 6 5 5 Annual growth rate (%) Annual growth rate (%) 4 4 3 3 2 2 1 1 0 0 –1 –1 –2 –2 –3 –3 –4 –4 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Per capita expenditure percentiles Per capita expenditure percentiles Central: Pro-rich South: Pro-rich 7 7 6 6 5 5 Annual growth rate (%) Annual growth rate (%) 4 4 3 3 2 2 1 1 0 0 –1 –1 –2 –2 –3 –3 –4 –4 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Per capita expenditure percentiles Per capita expenditure percentiles Upper 95% con dence bound/ Median spline Lower 95% con dence bound Source: Authors’ calculations, IHSES 2007 and 2012. Baghdad, the increase in per capita labor income income and per capita labor income or negligible among the poorest decile was quite low. In line increases, and these may in turn be related to the with the declining male employment to working declining consumption observed at the lower end age ratio in the South, the bottom three deciles in of the consumption distribution in the southern the South either experienced declining per capita governorates. 100 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 121: Changes in Per Capita Labor Income between 2007 and 2012 (‘000s of Iraqi Dinar) for the Bottom 3 Deciles of the Consumption Distribution, National and by Division 25 20 15 10 5 0 –5 Decile 1 Decile 2 Decile 3 Decile 1 Decile 2 Decile 3 Decile 1 Decile 2 Decile 3 Decile 1 Decile 2 Decile 3 Decile 1 Decile 2 Decile 3 Decile 1 Decile 2 Decile 3 Iraq Centre South Kurdistan Baghdad North Source: Authors’ calculations, IHSES 2007 and 2012. Consumption expenditures may be financed not only because the incidence of poverty is higher among the through labor income but also through non-labor in- non-employed, and because among the employed come and transfers. In 2012, for the average Iraqi, 68 poor, earnings are lower relative to the non-poor. percent of income came from labor, with some varia- The dependence on non-labor income and transfers tion across divisions: non-labor income shares were is highest in the South, where these sources consti- the highest in Baghdad, 37 percent, and the lowest in tute 60 percent of total income among the poorest the North, 28 percent (Table 23). However, for the decile, and the least in Kurdistan, where 42 percent lowest deciles, the dependence on non-labor income of total income derives from non-labor income and and transfers is significantly higher. On average, only transfers. This is in line with other indicators of a rel- 49 percent of the income of an Iraqi belonging to the atively well functioning labor market in the latter and lowest decile stems from labor income. This is likely poor labor market outcomes in the former. TABLE 23: Share of Labor Income in Total Income, and Shares of Major Sources of Non-Labor Incomes and Transfers in Total Non-Labor Income, Overall and Lowest Consumption Decile, 2012 Iraq Kurdistan Baghdad North Centre South Share in total Labor income Overall 68.00 69.16 63.02 71.81 68.02 68.71 income, 2012 Lowest decile 49.24 58.18 52.80 49.82 49.21 41.70 Share of non- Rations Overall 38.92 19.72 41.45 42.20 38.69 48.21 labor income, Lowest decile 59.96 42.04 53.72 64.09 59.53 62.10 2012 Pensions Overall 26.25 30.40 33.33 23.87 24.39 21.78 Lowest decile 13.27 19.40 18.10 9.72 16.86 11.43 Domestic Overall 14.45 12.09 16.18 13.30 17.10 12.24 remittances Lowest decile 11.43 13.39 18.36 9.11 8.61 12.72 Capital income Overall 8.27 14.41 3.84 8.13 11.14 3.89 Lowest decile 3.39 6.31 1.81 5.66 4.30 1.62 Source: Authors’ calculations, IHSES 2007 and 2012. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 101 In Iraq, there are four major sources of non-labor important for the lowest deciles, with more than income and transfers—income from capital (includ- half of all non-labor income coming from rations ing income from assets and property ownership), everywhere except in Baghdad. Dependence on ra- public transfers (primarily implicit incomes from tions is lower on average and for the bottom decile subsidized food rations and pension income), and in Kurdistan, where it accounts for only 20 percent private transfers, especially domestic remittances. In of non-labor income on average and 42 percent of 2012, for the average Iraqi, 8 percent of non-labor non-labor income for the bottom decile. It is high- income was comprised of capital income, 39 percent est in the South, where the corresponding shares from rations, 26 percent from pensions and 14 per- are 48 and 62 percent respectively. While domestic cent from domestic remittances. Among the lowest remittances are on average, the largest (in terms of consumption deciles, while the share of domestic their share in non-labor income) in the Centre, they remittances declines somewhat to 11 percent, the constitute a relatively larger share of non-labor in- share of capital and pension income declines sharp- come for the poorest 10 percent in Baghdad. ly (halves in the case of pension income), but the dependence on subsidized rations as an implicit in- Between 2007 and 2012, while labor incomes for come source increases to 60 percent on non-labor the lower consumption deciles in Iraq have been in- income and transfers. creasing on average, non-labor income has been fall- ing, primarily due to a decline in the implicit income While a larger share of labor income and of capi- transfer through rations owing to a reduction in the tal income in non-labor income is a sign of income number of items covered by the PDS (Table 24). earning opportunities, a greater dependence on pri- To some extent, the latter was compensated by an vate and public transfers is likely a sign of greater vul- increase in domestic remittances, and in some di- nerability. The share of capital income in non-labor visions, by an increase in pensions. In all divisions income is correlated with the importance of labor in- except the South and the lowest decile in Baghdad, come across divisions. In Kurdistan, almost 15 per- labor incomes increased relatively substantially for cent of all non-labor income on average stems from the lowest 20 and lowest 30 percent of the popula- income from capital, while in Baghdad and the South tion. For these groups, little change in non-labor it is only 3.8 percent. But amongst the bottom 10 income was also accompanied by stagnation and percent of the per capita consumption distribution, even declining labor incomes. While incomes do not the share of capital as a source of non-labor income one to one translate with consumption, the higher declines significantly, falling to less than 2 percent in dependence of the poor on transfer income rather Baghdad and the South. Thus, both income from la- than income from labor or capital as well as declin- bor and from capital are significantly lower amongst ing income from labor in the South are very likely the poor, but especially so in the South. related to the increase in poverty headcount rates in four of the five southern governorates. It is no surprise then that the lowest deciles are heavily dependent on transfers, which make up more than 95 percent of their non-labor income. In Determinants of Consumption and 2012, rations from the Public Distribution System Poverty Across Iraq made up 60 percent of non-labor income for the lowest deciles, with domestic remittances making Household size and composition, education and la- up another 11 percent, and pensions accounting for bor market outcomes all play a role in determining a further 13 percent. However, while pension in- consumption expenditure, the basis for measuring come comprises a lower share of non-labor income poverty, as do other location- specific factors that for the poorest 10 percent relative to the average, can imply access to (or the lack of) services, employ- the transfers associated with rations are much more ment opportunities, and markets. In this section, we 102 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 24: Changes in Labor Income and Main Non-Labor Income Sources between 2007 and 2012 for the Bottom 3 Consumption Deciles, National and Divisions Absolute Change (‘000s of Iraqi Dinars), 2012 Relative to 2007 Domestic Main Non Labor Deciles Capital Pensions remittances Rations Income components Labor income Iraq 1 0.32 0.34 1.09 –2.64 –0.90 4.02 2 –0.12 0.46 0.32 –2.91 –2.24 7.93 3 0.10 1.13 1.31 –3.08 –0.54 8.00 Kurdistan 1 0.33 1.53 0.87 –3.82 –1.09 14.08 2 –0.03 1.81 0.13 –3.37 –1.46 23.33 3 0.19 3.15 2.25 –3.42 2.17 23.30 Baghdad 1 0.40 –0.45 1.62 –3.13 –1.56 –2.73 2 0.08 –4.28 0.32 –3.54 –7.42 7.24 3 0.11 –2.84 2.54 –3.26 –3.45 10.27 North 1 0.45 0.25 1.06 –2.80 –1.04 4.60 2 0.02 –0.54 0.91 –1.99 –1.59 10.00 3 0.23 3.30 –0.13 –2.54 0.86 11.15 Centre 1 0.21 0.70 –0.36 –3.97 –3.42 9.08 2 –0.27 0.92 0.62 –3.96 –2.69 10.77 3 0.30 1.36 0.80 –3.61 –1.16 9.04 South 1 0.04 0.42 2.27 –1.71 1.02 –0.93 2 0.06 1.44 1.01 –2.12 0.38 0.05 3 –0.73 1.63 0.54 –2.21 –0.77 3.20 Source: Authors’ calculations, IHSES 2007 and 2012. attempt to estimate the influences of these factors (Table 25 and Tables A 4.1–A 4.5 in the Annex). in determining consumption expenditures, poverty Similarly, among otherwise similar households, larger (whether consumption falls below the threshold households and those with more children systemati- poverty line) and relative poverty (defined as be- cally tend to have lower consumption, when com- longing to the bottom 40 percent of the national pared with smaller households and households with consumption distribution). This analysis examines fewer children. The presence of elderly persons in the each division in the two survey years so as to iden- household is associated with higher consumption, es- tify common and division-specific factors that may pecially in the North and the Centre, and is probably explain spatial differences in welfare outcomes. reflecting the role of pensions in increasing household incomes. The education of the head of the household There are some important factors that are consistently is strongly correlated with per capita consumption ex- correlated with per capita consumption expenditures penditures. In all divisions, and in both survey years, across all the divisions of Iraq. Controlling for other each education level above primary schooling is posi- household characteristics that may also be correlated tively and significantly correlated with higher levels of with per capita consumption, households living in per capita consumption expenditures. urban areas have on average, higher consumption than those living in rural areas, with the exception In 2012, similar households with more employed of Baghdad in 2007 where there was no difference working age males had higher consumption than Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 103 TABLE 25: Significant Correlates of Per Capita Expenditure in 2007 and 2012 in Each Division* Kurdistan Baghdad North Centre South Correlates of per capita consumption expenditure 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Living in an urban area + + + + + + + + + Household size Number of household — — — — — — — — — — and dependency members Number of children 0–6 — — + — — — — — — — Number of children 7–17 — — — — — — — — — — Number of elderly + + + + + + — Number of employed working age males + + + + + + + Sector of Agriculture and fishing + + + + — employment Mining and Quarrying + + of the head of household Manufacturing — + + + + (relative to non- Utilities — + employed head of household) Construction — — — — — — Commerce and retail + + + + + + + + + Transport, storage and + + + + + communication Finance, insurance and + + + + + + + professional services Public administration, — — + + + + + + health and education Other — — + Education level Incomplete primary + + + + + + + of the head Complete primary + + + + + + + + + + of household (relative to Intermediate + + + + + + + + + + illiterate head of Secondary + + + + + + + + + + household) Higher secondary + + + + + + + + + + Tertiary + + + + + + + + + + Source: Authors’ calculations, IHSES 2007 and 2012. Note: * This table reports the signs of the coefficients of multivariate analysis of the significant correlates of per capita log real consumption, reported in Tables A 4.1 to A 4.5 in the Annex. those with fewer. The role of the head of house- the South but with lower per capita consumption in hold’s sector of employment appears to be more nu- Kurdistan. In places where the few oil-related jobs anced. Certain sectors have a strong and consistent are concentrated (the North and the South), house- relationship with consumption whereas others do holds with heads employed in the sector are likely not. In 2012, employment of the head of house- to have higher consumption compared to otherwise hold in commerce and retail (all divisions) and fi- similar households. nance, insurance and professional services (except in Kurdistan and Baghdad) is correlated with higher The important correlates of consumption identified consumption, while employment in construction above also broadly predict the likelihood of being is correlated with lower consumption (except in poor, i.e., of a household having consumption below Baghdad). Jobs in public administration are asso- a certain level. Table 26 summarizes the partial effects ciated with higher consumption in the Centre and or marginal probabilities of various characteristics on 104 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 26: Probability of Being Poor: Marginal Effects of Characteristics* Marginal probability effects: Partial effects Kurdistan Baghdad North Centre South of each explanatory variable (evaluated at mean values) on the probability that a household is poor 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Living in an urban area –12.8 –4.7 –0.06 –0.06 –0.19 –0.14 –0.12 Household Number of household members 2.0 0.12 0.06 0.07 0.05 0.08 0.02 0.08 0.07 size and Number of children 0–6 3.6 2.7 2.8 3.9 2.7 4.5 2.8 3.1 4.9 dependency Number of children 7–17 2.5 2.7 2.5 2.6 3.8 2.9 2.9 6.2 Number of elderly –5.3 4.3 Male headed household 7.4 3.5 6.6 Head of household lived elsewhere for at least 6 months –12.9 –13.2 –8.3 Number of employed working age males –2.6 –3.9 –6.5 –2.1 Sector of Agriculture and fishing 9.9 employment Mining and Quarrying –12.9 of the head of household Manufacturing –8.4 (relative to Utilities –5.0 –10.9 non-employed head of Construction 6.4 15.8 11.4 10.2 household) Commerce and retail –5.5 –8.2 –11.3 –5.4 –10.7 Transport, storage and communication –4.1 Finance, insurance and professional –4.9 –14.6 –15.0 –7.1 services Public administration, health and –6.0 –10.3 –5.8 –9.3 education Other –3.8 17.2 Education level Incomplete primary –2.8 –3.5 –7.7 –4.6 –9.4 of the head Complete primary –4.5 –5.0 –9.0 –6.2 –7.2 –9.3 –15.2 of household (relative to Intermediate –6.3 –5.1 –10.2 –7.8 –9.4 –15.2 –8.9 –15.1 –15.9 illiterate head Secondary –5.1 –7.2 –12.5 –12.3 –14.5 –9.0 –14.1 –16.1 of household) Higher secondary –6.1 –7.9 –9.2 –13.2 –9.3 –14.3 –9.7 –15.0 –18.8 Tertiary –7.6 –8.0 –12.1 –12.4 –10.2 –13.5 –16.1 –10.1 –15.2 –24.0 Source: Authors’ calculations, IHSES 2007 and 2012. Note: * This table and the following figures summarize results of probit regressions (marginal effects) reported in Tables A 4.6 to A 4.10 in the Annex. the likelihood of a household being poor. In other consumption as a whole. For instance, in 2012, an in- words, the coefficients of the regression indicate the crease in the number of employed working age males change in probability of a household being poor in a household was correlated with higher consump- with a unit increase in the independent variable. tion in each division, but it did not significantly alter the odds of a household being poor in the South. The advantage of this type of multivariate analy- Similarly, per capita real consumption was lower sis relative to the consumption regressions above is among households with heads employed in construc- that it allows us to quantify and compare the mar- tion in each division in 2012, but in the Centre, such ginal effects of each factor (holding all other factors households were no more likely to be poor than simi- constant) as well as to isolate characteristics that are lar households with non-employed heads. In the oth- correlated with poverty in particular rather than with er divisions, the relationship between employment in Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 105 construction and consumption translates into suffi- additional aged 7 to 17 years increased the likelihood ciently low consumption as to increase the likelihood of poverty by 3 percent while in 2012, this effect has that such households were poor. doubled. Moreover, an additional elderly person fur- ther increases the odds of being poor by 4 percent in Household size and dependency are significant pre- 2012, implying that pension receipts may not have dictors of poverty in Iraq, and in general, increase the been sufficient to overcome the effect of increased probability of being poor. In Kurdistan, it is the pres- dependency within the household. ence of children below the age of 18 that significantly increases the odds of being poor along with house- Living in a rural area generally increases the odds of hold size (Figure 122). In 2012, each additional child poverty compared to otherwise similar households ex- increased the odds of being poor by 3 percent on av- cept in Baghdad and in the Central division in 2012, erage. In Baghdad, while the overall effect of house- where there is no difference (Figure 123). While the hold demographics has remained stable, in 2012, the rural disadvantage has remained stable in the North presence of dependent children has become impor- and in the South, where it increases the marginal tant compared to 2007. In the North and in the Cen- likelihood of being poor by 6 and 12 percent respec- tre, an increase in the number of household members tively, it has fallen sharply in Kurdistan (from 13 to and in the number of children significantly increases 5 percent between 2007 and 2012) and has disap- the likelihood that a household is poor. In 2007, the peared entirely in the Centre. In 2012, households presence of an additional elderly household member whose heads had been migrants (lived elsewhere for reduced the likelihood that a household was below 6 months or more) were significantly less likely to the poverty line by 5 percent in the Centre. In the be poor in the North and the South. Male headed South, household size and composition were im- households faced higher odds of being poor in 2012 portant predictors of poverty in 2007, but their role in Baghdad, the Centre and the South, which may be seems to have become even more important in 2012. reflecting an expansion in social protection transfers This primarily stems from dependency: in 2007, an towards widow-headed households. A higher number FIGURE 122: Marginal Effects: Household Size FIGURE 123: Marginal Effects: Household and Composition Characteristics 10 25 5 20 0 15 –5 –10 10 –15 5 –20 –25 0 –30 –5 –35 –40 –10 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Kurdistan Baghdad North Central South Kurdistan Baghdad North Central South Number of employed working age males Number of elderly Head of household lived elsewhere Number of children 7–17 for at least 6 months Number of children 0–6 Male headed household Number of household members Living in an urban area Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. 106 THE UNFULFILLED PROMISE OF OIL AND GROWTH of employed working age males is associated with a has completed secondary education, by 19 percent if small decrease in the odds of poverty everywhere in the head has completed higher secondary education 2012 except in Kurdistan and the South. and by 24 percent if the head has tertiary education. By far the most consistent and striking correlate of The relationship between employment sector and poverty is the education of the head of the house- poverty and varies across divisions (Figure 125). hold. Even in Kurdistan, where the magnitudes of For instance, in 2007, households in Kurdistan with the partial effects are the lowest, primary and in- heads employed in utilities, transport and storage, termediate schooling in themselves each lower the and commerce and retail were 4 to 5 percent less likelihood that a household is poor by 5 and 10 likely to be poor compared with otherwise similar percent respectively; and secondary education and households. In 2012, in contrast, no sector of em- higher reduce these odds further by more than 7 ployment significantly lowered the odds of poverty percent each (Figure 124). In Baghdad in 2007, it relative to non-employment; and construction jobs was tertiary education that really distinguished the actually increased the likelihood of being poor. In non-poor from the poor, but the picture has become Baghdad, in 2012, commerce and retail, transport, more in line with the national pattern in 2012, with finance and public administration jobs lower the like- the likelihood of being poor falling with each addi- lihood of poverty in 2012. In the North, no employ- tional level of education. In the South, where edu- ment sector affected the odds of a household being cation levels are the lowest, education reduces the poor in 2012. In the Central division, in contrast, no odds the poverty the most. In 2012, household with employment sector increases the likelihood that the heads with even complete primary education were household is poor—public administration and com- 15 percent less likely to be poor (compared to 9 merce and retail lower poverty in 2007 and 2012. percent less likely in 2007) relative to similar house- In the South, where labor market outcomes are the holds with illiterate heads. Higher education starkly poorest, and where the male employment to work- reduces the likelihood that a household is below the ing age ratio has fallen between 2007 and 2012, the poverty line in the South: by 16 percent if the head FIGURE 125: Marginal Effects: Sector of Employment of the Head of FIGURE 124: Marginal Effects: Education of Household the Head of Household 40 0 20 –10 –20 0 –30 –20 –40 –50 –40 –60 –60 –70 –80 –80 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 –90 Kurdistan Baghdad North Central South –100 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Other Public administration, Kurdistan Baghdad North Central South Finance, insurance health and education and professional services Transport, storage Tertiary Intermediate Commerce and retail and communication Higher secondary Complete primary Utilities Construction Secondary Incomplete primary Mining and Quarrying Manufacturing Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 107 relationship between employment and poverty has poverty associated with living in that governorate become stronger. Mining, utilities, public adminis- relative to the excluded or reference governorate tration and finance in 2012—which are predomi- within the division (among households with similar nantly public sector jobs—lower the likelihood of size and composition, education and employment living below the poverty line by 13, 11, 9 and7 per- of the head of household, etc). In 2007, households cent respectively, as do commerce and manufactur- in both Sulaimaniya and Erbil were 8 and 5 per- ing. In contrast, households with heads employed in cent less likely to be poor, but by 2012, only house- construction are 10 percent more likely to be poor holds in Sulaimaniya enjoyed a small advantage over compared to similar households. households living in other governorates (Figure 126). This is consistent with the fact that Duhouk It is also interesting to note how the relative posi- was the only governorate within Kurdistan to signif- tions of different governorates within each division icantly reduce poverty headcount rates. The Central have altered between 2007 and 2012. In these pro- division, like the Kurdistan region, has witnessed bit regressions, we also include dummy variables a remarkable convergence across governorates. In for governorates within that division, and the co- 2007, living in any governorate outside Najaf in- efficients on these dummies measure the effect on creased the likelihood of a household being poor. governorate-specific factors. In other words, these By 2012, this was true only of Diyala and Wasit. coefficients estimate the higher or lower odds of FIGURE 126: Marginal Effects: Governorate Effects Kurdistan (relative to Duhouk) Centre (relative to Najaf) 0 40 35 –2 30 –4 25 20 –6 15 10 –8 5 –10 0 Sulaimaniya Erbil Diyala Anbar Babylon Karbala Wasit North (relative to Salahadin) South (relative to Basra) 20 40 15 35 10 30 5 25 20 0 15 –5 10 –10 5 –15 0 –20 –5 –25 –10 Nineveh Kirkuk Qadisiya Muthanna Thi Qar Missan 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. 108 THE UNFULFILLED PROMISE OF OIL AND GROWTH In the North, Kirkuk began with a relative advan- have changed over time, and try to understand why tage compared to Salahadin in 2007, but the gap these disparities exist. Figure 127, panel A presents appears to have bridged (Salahadin was one of the the mean differences in the welfare ratios between governorates to reduce poverty significantly). In urban and rural areas within each region in 2007 contrast, in Nineveh, where households were less and 2012 (Baghdad is excluded because it is an likely to be poor relative to Salahadin by more than overwhelmingly urban governorate).36 In Iraq, as 20 percent in 2007, five years later, they were 16 in most other countries, households in urban ar- percent more likely to be poor (consistent with the eas have higher levels of welfare than those living increase in poverty in Nineveh). In the South, in in rural areas, even after taking into account cost of 2007, after accounting for the effect of household living differences. The largest differential between characteristics, only households in Missan had a 5 urban and rural welfare is in Kurdistan followed by percentage point lower likelihood of being poor the South. In three of the four divisions, the urban- relative to Basra. By 2012, each governorate in the rural welfare gap declined between 2007 and 2012 South was associated with significantly higher odds while it actually widened in the South, where rural of poverty relative to Basra (from 20 percent in Mis- poverty increased at a faster rate than urban poverty san to 33 percent in Qadisiya). in three governorates. Table 27 summarizes the marginal or partial ef- These differences in welfare between rural and ur- fects of the same set of characteristics—household ban areas can be due to differences in portable (or size and composition, the education and sector of non-geographic, mobile) household characteristics employment of the head of household and other such as education or household composition or household characteristics—on the probability that due to differences in returns to these characteris- a household belongs to the bottom 40 percent of tics, i.e., their marginal effects. The estimated de- the consumption distribution. In terms of four sets compositions of welfare differences between urban of factors—living in an urban area, household size and rural areas within divisions reveal that charac- and dependency, education of the head of house- teristics explain a larger share of the welfare differ- hold—there is a remarkable overlap in terms of pov- ences within regions (see panel b in Figure 127). erty and belonging to the bottom 40 percent, with Differences in household characteristics explain significant increases in the magnitudes of the coeffi- around 60% of the welfare differences between the cients. This suggests that the same set of factors that urban and rural areas in 2007, and their contribu- are highly correlated with poverty are also associ- tion has increased to more than 70% in 2012. Thus, ated with belonging to the bottom 40 percent, or in rural-urban welfare differences within each division other words, that these are very similar households. in Iraq are driven by the (increasing) concentra- The relationship with sectors of employment is not tion in urban areas of individuals with a higher such a linear and straightforward one. For example, level of endowments. For instance, individuals who in 2012, construction jobs for the household head live in urban areas have significantly higher levels are associated with higher odds of poverty in Kurd- of education which appears to explain a large part istan and the South; but they also increase the likeli- of the differential in welfare levels. In Kurdistan, hood that a household belongs in the bottom 40 this factor contributes more than 40% of the dif- only in Kurdistan. ference explained by characteristics in both years. 36 The welfare ratio is the ratio of the household’s expendi- Explaining Rural – Urban Welfare ture to the contemporaneous poverty line in the region of Disparities Within Divisions residence of the household. The welfare ratio as defined is a We now explore the scale urban-rural welfare dis- number that measures the standard of living as a multiple parities within divisions in Iraq, examine how these of the poverty line. See Annex for details on methodology. Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 109 TABLE 27: Probability of Being in the Bottom 40 Percent* Marginal probability effects: Partial effects of each Kurdistan Baghdad North Centre South explanatory variable (evaluated at mean values) on the probability that a household belongs to the bottom 40 percent of the consumption distribution 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Living in an urban area –21.5 –11.5 –11.2 –6.1 –19.5 –12.1 –12.0 Household size and Number of household members 7.4 7.2 21.5 17.1 10.2 5.8 8.1 7.7 10.9 9.4 dependency Number of children 0–6 10.5 7.2 6.2 6.8 6.4 4.7 6.2 3.7 5.6 Number of children 7–17 7.2 7.8 3.7 2.2 5.0 3.7 6.5 3.5 6.9 Number of elderly –9.3 –6.7 –6.1 Male headed household 16.0 6.9 12.9 Head of household lived elsewhere for at least 6 months 17.8 Number of employed working age males –6.6 –5.5 –8.7 –3.9 Sector of Agriculture and fishing –26.1 –16.0 –15.3 11.7 employment of the Mining and Quarrying –29.2 –21.7 head of household (relative to non- Manufacturing –8.5 –11.4 employed head of Utilities –22.7 household) Construction 14.0 21.0 11.8 16.0 Commerce and retail –15.7 –14.9 –12.6 –11.4 Transport, storage and communication –23.5 –8.6 Finance, insurance and professional services –14.5 –11.3 –17.2 –12.8 –13.0 Public administration, health and education 11.1 –21.0 –10.1 –8.6 –12.1 –12.4 Other –18.6 Education level Incomplete primary –12.3 –8.4 –24.2 –6.2 –11.0 –11.8 of the head of Complete primary –9.9 –16.1 –7.3 –9.0 –12.0 –8.7 –14.5 –19.5 household (relative to illiterate head of Intermediate –19.8 –16.7 –30.4 –12.2 –17.3 –21.0 –16.0 –26.5 –24.6 household) Secondary –21.4 –27.0 –34.4 –18.9 –19.3 –23.6 –23.8 –23.4 –26.3 Higher secondary –29.0 –22.9 –26.0 –30.0 –28.5 –23.7 –23.6 –21.4 –24.9 –28.5 Tertiary –33.6 –28.5 –35.2 –35.7 –14.8 –30.9 –27.4 –30.7 –30.4 –35.7 Source: Authors’ calculations, IHSES 2007 and 2012. NOTE: * This table summarizes results of probit regressions (marginal effects) reported in Tables A 4.11 to A 4.15 in the Annex. Another pattern that is important to note is that in There are two potential explanations as to why dif- the division where headcount rates fell, the Centre, ferences in characteristics between urban and rural the contribution of characteristics has more than areas of the same region may be large. First, it could doubled: differences in individual endowments be- be that the nature of productive activities in urban tween rural and urban areas explained less than 40 and rural areas may require inherently different percent of welfare differences in 2007, compared characteristics. Farming activities in rural areas, for to 87 percent in 2012. In general, this pattern instance, require little formal education and might holds in every division where rural-urban welfare be carried out more efficiently by households with differences have fallen. In contrast, in the South, more family members. However, a lower education where poverty increased, there has been almost no level and larger number of family dependents would change in the relative contribution of characteris- be less likely to lead to better economic outcomes tics and returns. in an urban setting. A second, complementary 110 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 127: Differences Within Regions – Urban Versus Rural Areas A. Welfare di erences B. Decomposition 0.5 2012 South 0.4 2007 2012 Center 0.3 2007 0.2 2012 Kurdistan North 2007 0.1 2012 2007 0 Kurdistan North Center South 0 20 40 60 80 100 2007 2012 Characteristics Returns Source: Authors’ calculations, IHSES 2007 and 2012. possibility is that people sort themselves across space and rural areas in each division against the rural based on their characteristics. That is, individuals South (the poorest rural sub-division). migrate between urban and rural areas within re- gions to the location where they can earn the high- Figure 128 presents the mean differences in the est returns for their set of characteristics. welfare ratios between divisions in 2007 and 2012. Comparing the welfare ratios between divisions, Why did these rural-urban differences in charac- not surprisingly, Kurdistan has the highest welfare teristics become more important over time? While of any division (panel A in Figure 128) or any of it is unlikely that the nature of economic activities the urban areas of Iraq’s divisions. The South is the changed enough to explain this trend in the five year poorest division relative to Kurdistan in both 2007 time period considered here, it is more probable that and 2012 though the gap seems to have declined mobility between rural and urban areas increased, somewhat in 2012 in urban areas (panel B in Fig- potentially induced by existing income differentials, ure 128) while it has increased in rural areas (panel so as to reduce the welfare premium associated with C in Figure 128). Moreover, the rural areas of the living in urban areas, holding all else equal. South appear to have lower welfare ratios than the rural areas of other divisions and this has worsened over time. In line with the rapid welfare improve- Explaining Welfare Differences between ments in the Centre, the welfare difference be- Divisions tween Kurdistan and the Centre have come down To carry out the “between-region” comparisons, sharply in 2012, in both rural and urban areas. we first compare the differences in welfare in each division (urban and rural areas pooled) against The decomposition of welfare differences “be- Kurdistan, which has the lowest poverty levels. Ac- tween” divisions in Iraq reveals that differences in knowledging that the pooling of urban and rural returns to characteristics of households play a larg- areas into one regional aggregate may be “mixing er role in explaining welfare differentials than they apples with oranges”, we also construct additional did between urban and rural areas within the same comparisons of urban areas in each division against divisions. This result is robust to different types of urban Kurdistan (the least poor urban sub-division) comparisons. Thus, welfare differences between the Conflict, Revival and Neglect: Understanding Spatial Disparities in Welfare 111 FIGURE 128: Welfare Differences between FIGURE 129: Explaining Welfare Differences Divisions between Regions A. Divisions vs. Kurdistan A. Regions vs. Kurdistan 0 2012 South 2007 –0.1 2012 Center 2007 2012 Baghdad North –0.2 2007 2012 –0.3 2007 Baghdad North Center South –150 –100 –50 0 50 100 150 200 B. Urban areas in each division vs. urban Kurdistan B. Urban areas in region vs. urban Kurdistan 0 2012 –0.1 South 2007 2012 Center 2007 –0.2 2012 Baghdad North 2007 –0.3 2012 Baghdad North Center South 2007 –100 –50 0 50 100 150 200 C. Rural areas in each division vs. rural South (excluding Baghdad) 0.4 C. Rural areas in region vs. rural South 0.3 2012 Center 2007 0.2 2012 North 0.1 2007 2012 Kurdistan 0 Kurdistan North Center 2007 2007 2012 0 20 40 60 80 100 Source: Authors’ calculations, IHSES 2007 and 2012. Characteristics Returns Source: Authors’ calculations, IHSES 2007 and 2012. urban areas of other divisions and urban Kurdistan are mostly due to differences in returns to charac- 2012 (see panel B of Figure 129). Differences in teristics rather than due to differences in charac- welfare between the urban areas in the North and teristics. That is, people living in urban Kurdistan the urban areas in Kurdistan appear to be explained have roughly comparable characteristics to urban for the most part (close to 60%) by differences in residents of the other divisions, but the latter re- the characteristics of the household endowments ceive much lower returns for these characteristics. in these areas. An exception is the case of the North region in 112 THE UNFULFILLED PROMISE OF OIL AND GROWTH The results of the analysis also reveal some inter- In this chapter, we argue that spatial disparities in esting changes in the primary factors explaining welfare across Iraq are explained in part by eco- inter-division welfare disparities over time. In 2012, nomic revival in relatively stable and peaceful gov- welfare differences across urban areas in different ernorates in the Centre, in the Kurdistan region divisions and urban Kurdistan seem to be less due and in Basra, in part by the immediate effect of the to differences in the returns compared to 2007. For post-2003 violence (which continues to this day) example, in 2007 about 16 percent of the welfare in Baghdad and the Northern governorates, and to differential between urban areas in the South and the continued neglect of the southern governor- Kurdistan could be attributed to differences in char- ates. Violence and insecurity has been a pervasive acteristics with the remaining 84 percent attributed feature of many parts of the country for a long to differences in returns. By 2012, differences in time, and are evident in long term trends of dis- characteristics seem to play a bigger role (just over placement and a stalling of progress in health and 40 percent), while differences in returns seem to be- education. Since 1990 however, beginning from come less important (just below 60 percent). a common legacy of persecution and neglect dur- ing the Saddam era, Kurdistan and the South have When comparing differences in welfare across rural been on divergent trajectories. Poverty trends mask areas in Kurdistan, the North and the Centre with the improvements in education and labor market respect to the rural South, differential returns to outcomes that have been experienced in the three characteristics account for around four-fifths of the Kurdish governorates, which have also experienced welfare differences in both years. Thus, rural indi- a significant increase in population. In the southern viduals in the South with a certain set of characteris- governorates, and with the exception of Basra, the tics earn much lower returns than similar individuals last five years appear to have compounded the ne- in rural areas of other divisions. glect of the past, with declining male employment and labor force participation, declining female em- Taken together, these suggest that urban-rural mo- ployment in agriculture, and with young people bility within divisions may be bridging the welfare falling further behind in human capital. Relative gap; and that to some extent a similar trend is true peace and stability has not been sufficient for eco- (although much smaller in scale) between urban nomic revival. But the absence of peace and secu- areas of different divisions. However, differential rity has implied little change in welfare in Baghdad returns to characteristics continue to be very impor- and the North, where the post-2003 violence was tant in explaining welfare differential across rural ar- concentrated. It is only in the Centre where peace eas in different divisions. In the South in particular, and stability have to some extent combined with an the rural-urban welfare gap has increased, accompa- improvement in economic activity, and where job nied by an increasing disparity between rural areas growth has outpaced the growth in the male work- in the rest of the country and the rural South. ing age population. Understanding the Drivers of Poverty Reduction 5 T o understand the drivers of poverty reduc- The results of the decomposition exercise mask a lot of tion, we decompose the distributional changes heterogeneity across different parts of Iraq. In some of in consumption and income over the 2007 to the areas where poverty fell, changes in labor income 2012 period, and examine the size and influence of and employment explain more than the half of the re- various factors in driving these changes. The reduc- duction in poverty, and in the others, they represent the tion in poverty observed between 2007 and 2012 was second most important factor, explaining more than a mainly driven by changes in labor income, i.e., growth quarter of the reduction in poverty. In those areas that in earnings rather than growth in employment. Had experienced an increased in poverty, changes in labor everything else stayed the same, the change in labor income and employment moved in the opposite direc- income alone would have generated a further reduc- tion to the other forces and contributed negatively to tion in poverty headcount rates to 18 percent in 2012 poverty increase, or rather, mitigated the increase in instead of the observed 19.8 percent. Demographic fac- poverty. tors, in particular, the declining share of adults per household among the poor, as well as the consumption- Higher employment contributed to poverty reduction income ratio worked against poverty reduction. only in 2 out of 5 sub-divisions where headcount rates fell, contrary to findings for the whole country. In the Non labor income components and public and private other 3 sub-divisions where poverty fell, lower levels transfers including pensions and domenstic remittanc- of employment among the poor actually increased es contributed to poverty reduction. Domestic and in- poverty. The same pattern was found among those ternational remittances among households contributed divisions where headcount rates increased: changes to 13 percent of the total reduction in poverty, although (decreases) in employment contributed to increases most of the change in consumption that is explained in poverty. by private transfers came from domestic remittances. Capital income and the flow of services from owner-oc- Although changes in labor income are the main con- cupied dwellings also contributed so as to reduce poverty tributor to poverty reduction in most divisions, non- over the period. Considering public transfers, on the labor incomes also mattered. In those sub-divisions one hand, pensions and other public transfers like social with significant decreases in poverty, changes in other safety net compensations explained almost 27 percent of private and public transfers compensated for the de- the reduction in poverty. On the other hand, the decline cline in implicit PDS transfer incomes. in implicit incomes from PDS transfers acted counter to poverty reduction. In fact, the reduction of ration In order to unpack the dynamics of poverty reduc- transfers would have led to a 43 percent lower reduc- tion in Iraq between 2007 and 2012, this chapter de- tion in poverty, if everything else has been held constant. composes the distributional changes in consumption 114 THE UNFULFILLED PROMISE OF OIL AND GROWTH and income over this period, and examines the size 2. Growth in labor income as consequence of and influence of various components in driving these changes in employed members, movements in distributional changes. In particular, this chapter an- their earnings or a combination of both; swers the following types of questions: 3. Growth in non-labor income, mainly in the form of public or private transfers; and • Was poverty reduction a result of demographic 4. Changes in consumption or saving patterns. changes that led to a lower dependency rate? • Was the observed poverty reduction the result Before undertaking the decomposition analysis, we of higher employment or higher labor income first examine the trends for each of these underly- due to improved labor market conditions? ing components of consumption over the 2007 to • Did welfare improve due to improved and more 2012 period. effective social protection policies, or perhaps due to an increase in private transfers or capital First, changes in the demographic composition of income? households can play a role in determining house- hold welfare by altering the dependency ratio or the Although these decompositions do not allow for the number of earners relative to the number of con- identification of casual effects, they help to focus the sumers in a household. Indeed, the rate of popula- attention to the quantitative elements that are most tion growth has increased in Iraq over the period, important in describing changes in poverty. The in- particularly for individuals younger than 15 years sights provided by a deeper understanding of income old. This cohort grew faster than the rest of the and poverty dynamics can contribute to the evidence population. These demographic changes, however, base for Iraq’s policy-making going forward. It also translate into an almost constant average household identifies areas for in-depth analysis in the second size as a result of a decrease on average in the share half of the report that immediately follows. of adults per household (Figure 131 Panel A and B). A lower number of adults per household generally We begin by discussing potential sources of the ob- implies a higher dependency rate and consequently, served distributional changes in labor and non-labor lower consumption per-capita, assuming that adult income, employment and population components employment rates remain unchanged. occurring over the period for the country as a whole and for the different divisions. We then provide an However, these overall trends for the nation as a introduction to the model of consumption under- whole do not necessarily hold in each of the divi- lying the poverty decompositions and explore the sions. Among the five divisions, the average house- main decomposition results. Finally, we examine hold size as well as the number of adults over 15 some of the implications of these findings. years old grew in the North. Despite these increases, the age-dependency ratio has expanded as a con- sequence of an even higher rate of growth among Potential Sources of Consumption younger cohorts (Figure 132 panel A). The oppo- Growth and Poverty Reduction in Iraq site trend was experienced in Kurdistan where both the average household size and number of adults There are at least four factors that could have influ- decreased, while the share of adults per household enced consumption growth and therefore, poverty (inverse of the dependency ratio) increased (Figure reduction over the period: 132 panel A). These trends still mask considerable heterogeneity across households within each region. 1. The demographic composition of the house- Most importantly, note that in most divisions ex- hold; as measured by the share of adults per cluding Kurdistan, the share of adults per household household (the inverse of the dependency ratio); among the poor decreased more than on average, Understanding the Drivers of Poverty Reduction 115 FIGURE 130: Demographic Changes Population Pyramids +70 +70 61–65 61–65 51–55 51–55 41–45 41–45 31–35 31–35 21–25 21–25 10–15 10–15 0–5 0–5 –4000000 –2000000 0 2000000 4000000 –4000000 –2000000 0 2000000 4000000 2007 2012 Males Female Source: Authors’ calculations, IHSES 2007 and 2012. implying that demographic changes among poor that despite the observed population growth, both households were potentially acting against poverty labor force participation and the employment-to- reduction across much of Iraq (Figure 132 panel B). population ratio decreased over the period, particu- larly for women. At the household level, the share Second, growth in labor income could be the main of working adults (ages 15–64) slightly decreased driver of the observed changes in poverty. This (Figure 131 panel B), pointing to a potential de- could be due to increases in the number and share of cline in consumption attributable to lower work- employed members or increases in labor income or force participation as a result of discouragement on a combination of both. As described in greater de- the one hand and to lower employment rates for tail in next chapter, simple summary statistics reveal women on the other. FIGURE 131: Demographic Characteristics A. Household size remains constant… B. ….while the share of adults has decreased 10 70 8.6 8.4 60.3 59.0 60 8 50 6 39.9 39.3 5.1 4.8 40 4 30 20 2 10 0 0 2007 2012 2007 2012 Average household size Share of adults (age>=15) per household Number of adults (age>=15) per household Occupied adults (as a share of number of adults) Source: Authors’ calculations, IHSES 2007 and 2012. 116 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 132: Share of Adults Per Household A. Average B. Poor 66 56 54.8 63.9 64.5 64 54 62.0 61.9 62 52 59.8 51.1 50.9 60 58.0 58.0 50 49.0 49.4 58 57.2 48.8 56.7 56.1 47.8 48 46.9 46.7 56 46.6 54 46 52 44 50 42 Kurdistan Baghdad North Center South Kurdistan Baghdad North Center South 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. This negative trend in the share of working adults Having said this, earnings moved to counteract holds true not only among divisions but also across these lower employment rates. There is evidence households within each geographical division (Fig- that labor incomes per adult increased at the bot- ure 133). The exception is Kurdistan which indeed tom of the distribution in most divisions except experienced the opposite trend, with an increase in the South (Figure 134). Unfortunately, we can- its share of working adults over the period for poor not determine wether this is due to higher earn- and non poor households. In each of the other di- ings per hour or due a to greater number of hours visions, even among poor households, movements worked or a combination of the two. In any case, followed the same trend as the average, although in most parts of the country, the incomes that the the magnitude of the changes are not significant ex- poor derive from their work have increased over cept for poor households in the South and North, the period. who experienced a significant decline. FIGURE 133: Share of Working Adults Per Household A. Average B. Poor 44 43.3 45 42.1 40 39.4 40.3 39.0 37.9 42 41.5 36.8 36.3 36.5 36.5 40.5 35 33.0 33.2 40.3 40 39.3 39.0 30 25 38 37.5 37.1 36.6 20 36 15 10 34 5 32 0 Kurdistan Baghdad North Center South Kurdistan Baghdad North Center South 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Understanding the Drivers of Poverty Reduction 117 FIGURE 134: Labor Income by Region – Growth Incidence Curves A. Kurdistan: Pro-poor B. Baghdad: Pro-rich 20 5 18 4 16 3 14 2 % Change % Change 12 1 10 0 8 –1 6 –2 4 –3 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 Percentile of per capita income Percentile of per capita income C. North: Neutral D. Central: Pro-poor 25 14 12 20 10 % Change % Change 15 8 10 6 4 5 2 0 0 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 Percentile of per capita income Percentile of per capita income E. South: Pro-rich 3 2 1 % Change 0 –1 –2 –3 0 10 20 30 40 50 60 70 80 90 Percentile of per capita income Source: Authors’ calculations, IHSES 2007 and 2012. Poverty reduction could have been related to 2007 to less than 4.75 ID trillions in 2012 in real growth in different non-labor income compo- terms).37 Government spending for subsidies have nents. Figure 135 panel A shows the public trans- decreased as a share of the GDP while pensions fers have decreased by almost 6 percentage points as a share of the GDP over the last 5 years. This 37 Government social expenditures were deflated using GDP represents a reduction of a 2.75 percent per annum deflator for General Government in order to express it in in absolute terms (from almost 5.5 ID trillions in 2007 prices. 118 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 135: Public and Private Transfers A. Public transfers B. Rental prices increased faster than food prices 12 170 10 2.3 160 3.3 150 8 Percent of GDP 2007 = 100 3.7 4.3 140 6 5.7 5.8 3.3 130 4 3.6 3.5 2.5 120 2.1 2 2.9 2.0 110 2.2 1.6 1.6 2.3 0 0.6 100 2007 2008 2009 2010 2011 2012 2007 2008 2009 2010 2011 2012 Subsidies PDS Pensions CPI Food Rent Source: Authors’ calculations, PER (2013) and CSO (2014). have partially compensated this trend by increas- Change in the Consumption-to- FIGURE 136:  ing 75 percent in real terms between 2007 and Income Ratio 2012 (from 1.3 ID trillions to more than 2.3 ID trillions in real terms for 2007 and 2012 respec- 2% tively). Other social benefits in the form of trans- 1% fers associated with the Public Distribution System (PDS) were cut in more than half over the 5 years 0% period. In addition to public sources of transfers, –1% flows of services from dwellings (i.e. the implicit rental income earned by living in owner-occu- –2% pied households) have increased over the period. –3% Rental values have in fact grown much faster than Decile 1 2 3 4 5 6 7 8 9 10 prices of food and non-food items over the 2007 Deciles of per capita consumption to 2012 period (Figure 135 panel B). Source: Authors’ calculations, IHSES 2007 and 2012. Finally, in the absence of measurement error, changes in consumption-based poverty could also be related to changes in consumption and saving distribution, while it fell for those at the top over patterns. Faced with growing incomes, households the period. could either increase consumption porportionate- ly or they could increase their savings. However, However, this trend was not homogeneous across given measurement errors in income and expendi- divisions. Indeed, the ratio fell across the whole ture aggregates in households surveys, and the low distribution for Kurdistan and for almost all deciles rates of formal savings and credit in Iraq, it is diffi- of per capita consumption in Baghdad and North cult to differentiate between changes in household except the lowest decile. These behaviors in con- consumption on account of real behavioral shifts sumption patterns could be related to lower pov- versus changes due to measurement. Figure 136 erty rates and relatively higher welfare levels. The shows that in Iraq the consumption-to-income ra- rest of the country matches the behavior of the tio increased for households at the bottom of the country except for the Central division where the Understanding the Drivers of Poverty Reduction 119 Changes in the Consumption-to- FIGURE 137:  Where n is the total number of members in house- Yh Income Ratio by Division = θh is the C hh, hold consumption-to-income ratio, and Yh represents nthe total income of household h. 6% C h = θh n 4% We decompose the contribution of changes in real 2% per capita consumption expenditures to poverty re- 0% duction following the Paes de Barros et.al. (2006) –2% methodology. In particular, poverty reduction is di- –4% vided into 4 main components: a household’s pro- pensity to consume, adult population in the house- –6% hold, labor income per adult, non-labor income –8% Decile 1 2 3 4 5 6 7 8 9 10 per adult (Figure 138). This decomposition helps Deciles of per capita consumption to estimate the relative contributions of each of the different components to changes in real per capita Kurdistan Baghdad North Central South consumption and, consequently, to the observed poverty reduction over the last 5 years. Source: Authors’ calculations, IHSES 2007 and 2012. Figure 139 summarizes the contributions of the dif- consumption-to-income ratio increased across the ferent factors to poverty reduction and the direction distribution (Figure 137). of their influence, so that the total contributions add up to (explain) a 100 percent of the total observed All in all, each of the sources of change described change in poverty. Changes in labor income played above could have contributed in a positive or nega- the most significant role in reducing poverty in Iraq tive manner to the observed reduction in poverty over the 2007–2012 period. Moreover, it was the over the period for the Iraq and each division in growth in labor income that contributed the most particular. The question we turn to next is how im- (145 percent) and more than compensated for the portant the contribution of each of these forces was. effect of the reduction in the number of workers or jobs (–2 percent) measured as the shared of oc- cupied adults. In other words, had everything else Decomposing Poverty Reduction: stayed the same, the change in labor income alone 2007–2012 would have generated a further reduction in pov- erty headcount rates to 18 percent instead of the In contrast to methods that focus on aggregate sum- observed 19.8 percent. However, we are not able mary statistics such as the growth and redistribu- to disentangle whether the increase in earnings was tion analysis in Chapter 2, the micro-decomposition due to improvements in quality of jobs, changes in methods applied in this chapter generate a series of productivity, or simply due to longer hours. simulations of entire counterfactual distributions to account for the contributions of different factors The increase in the flow of the dwelling’s services such as demographics, labor income, and non-labor has also contributed to poverty reduction. Improve- incomes to poverty reductions. Underlying the de- ments in the economic situation and the significant composition is a simple model of household con- increase in rental values would positively impact the sumption. In particular, consumption per capita in actual value of the dwelling. In a country such as household h is defined by: Iraq where more than 70 percent of the households own their dwellings, imputed rental values have Yh played a significant role in poverty reduction by ex- C h = θh n plaining more than a quarter of its change. 120 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 138: Different Components of Consumption Per Capita Propensity to consume & measurement Share of error Θ occupied adults Share of adults n_A/n n_A/n Consumption Household Labor income per capita Income per adult Income per adult Labor income per Non labor occupied adult Number of income per members (n) adult Source: Inchauste, G and others (2014). There were also important contributions from dif- Considering public transfers, on the one hand, pen- ferent non-labor income components which acted sions and other public transfers like social safety net in different directions. From the private viewpoint, compensations explained almost 27 percent of the domestic and international transfers or remittances reduction in poverty. On the other hand, the reduc- among households contributed to 13 percent of the tion in the number of ration items transferred to the total reduction in poverty. Notice that even though poor did not help to reduce poverty. The reduction international transfers increased significantly over of ration transfers would have led to a 43 percent the period, most of the change in consumption that reduction in poverty if they had remained at their is explained by private transfers came from domestic 2007 levels. remittances. Capital income slightly contributed so as to reduce poverty over the period. Demographic factors as well as the consumption- income ratio worked to counter poverty reduction. In the case of the first factor, the share of adults per households among the poor decreased, there- FIGURE 139: Contributions to Poverty fore increasing poverty, pointing to an unequalizing Reduction – Total Iraq force in the country. These negative contributions suggest that if these factors had remained constant Imputed rent 26.9 over the period, poverty reduction could have been Other Private transfers 13.2 larger than actually observed.38 Capital 0.7 Other public transfers 7.7 Ration –42.8 The result that labor income growth has been Labor income 18.9 the main contributor to poverty reduction dur- Occupation share 145.6 Adult population –2.1 ing 2007–2012 period is in line with the results Pension –37.4 obtained for similar work undertaken for other Consumption-income ratio –30.6 countries over the last decade. Figure 140 shows –120 –60 0 60 120 180 these simple accounting decompositions for a set of % of poverty reduction countries in which there was a substantial decline Source: Authors’ calculations, IHSES 2007 and 2012. Notes: “Capital” includes all incomes from property such as rent from land, non- residential buildings, equipments, shares and profits, interests, among others; “Other 38 In the case of the consumption to income ratio there is mea- public transfers” refers to social protection network compensation and other public transfers in cash and in kind; and “Other private transfers” refers to domestic and surement error in both of its components, interpretations international remittances, zakat and other private transfers in cash and in kind. about changes in this ratio must be treated with caution. Understanding the Drivers of Poverty Reduction 121 FIGURE 140: Growth in Labor Income is the Main Contributor Factor to Changes in Moderate Poverty 140 120 100 80 60 40 20 0 –20 –40 Bangladesh, 2000–10 Ghana, 1998–2005 Nepal, 1996–2003 Cambodia, 2007–2010 Mongolia, 2007–2011 Philippines, 2006–2009 Sri Lanka, 2002–09 Vietnam, 2004–2010 Argentina, 2000–10 Brazil, 2001–09 Chile, 2000–09 Colombia, 2002–10 Costa Rica, 2000–08 Ecuador, 2003–10 Honduras, 1999–2009 Panama, 2001–09 Paraguay, 1999–2010 Peru, 2004–10 Thailand, 2000–09 Moldova, 2001–10 Romania, 2001–09 US$1.25 a day US$4 a day US$5 a day Share of working-age family members Employment+earnings Nonlabor income Consumption-to-income ratio Source: Inchauste, G.and others (2014), World Bank, Washington DC. Notes: “Labor income” refers to the change in employment and earnings per adult; “nonlabor income” refers to transfers, pensions, capital, and other income not from labor. Consumption- based measures of poverty are used in the case of Bangladesh, Ghana, Nepal, Peru, Thailand, Moldova, and Romania. Income-based measures of poverty are used elsewhere. Data from SEDLAC, RIGA and National Household Surveys. in poverty. In particular, in 12 of the 21 countries Central 1: Diyala and Anbar), changes in labor in- changes in labor incomes and employment explain come and employment explain more than the half of more than half of the change in poverty, and in an- the reduction in poverty, and in the other 2 sub-divi- other 6 countries, they account for more than 40 sions (i.e. Salahuddin and Central 2: Kerbala, Wasit, percent of the reduction in poverty. The number of Najaf and Babylon), they represent the second most occupied adults increased, contributing to poverty important factor, explaining more than a quarter of reduction due to increased employment. But, it was the reduction in poverty. Additionally, in those sub- increases in earnings per occupied adult that made divisions which experienced an increased in poverty the largest contribution to poverty reduction similar (i.e. Ninevah and Rest South: Qadisiya, Muthanna, to what has been experienced in Iraq. Thi-Qar and Missan), changes in labor income and employment moved in the opposite direction to the The results of the decomposition exercise mask a lot other forces and contributed negatively to poverty of heterogeneity across different parts of Iraq. We increase, or rather, mitigated the increase in poverty. perform the same exercise for those governorates and sub-divisions which experienced a significant Higher employment contributed to poverty reduc- change in poverty, either positive or negative, over tion only in 2 out of 5 sub-divisions where head- the period, and find that labor income still repre- count rates fell (i.e. Duhouk and Central 1), con- sents one of the most important factors which con- trary to findings for the whole country. In the other tribute to poverty reduction (Figure 141). However, 3 sub-divisions where poverty fell (i.e. Basra, Cen- the size of this relationship varies. In 3 out of 5 sub- tral 2 and Salahuddin), lower levels of employment divisions where poverty fell (i.e. Duhouk, Basra and among the poor actually increased poverty (Table 122 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 141: Decomposition of Significant Poverty Changes by Divisions 200 150 100 22.5 41.3 142.3 18.2 25.9 50 74.3 47.6 65.4 52.6 48.1 48.8 2.2 28.7 0 –20.0 –15.4 –75.1 –50 –100 Duhouk Ninevah Salahuddin Central 1 Central 2 Rest South Basra North Central South Consumption-to-income ratio Adult population Employment+earnings Nonlabor income Imputed rent Source: Authors’ calculations, IHSES 2007 and 2012. Notes: Include only those divisions with significant change in poverty reduction. Central 1 = Diyala and Anbar; Central 2 = Kerbala, Wasit, Najaf and Babylon and Rest South: Qadisiya, Muthanna, Thi-Qar and Missan.“Capital” includes all incomes from property such as rent from land, non-residential buildings, equipments, shares and profits, interests, among others; “Other public transfers” refers to social protection network compensation and other public transfers in cash and in kind; and “Other private transfers” refers to domestic and international remittances, zakat and other private transfers in cash and in kind. A5.2). The same pattern was found among those increase in poverty in Rest of South (41.3 percent) divisions where headcount rates increased: changes and Ninevah (22.5 percent). (decreases) in employment contributed to increases in poverty. In general, the increase in earnings of Finally, changes in the consumption-to-income ra- workers was relatively more important in reducing tio generally helped to reduce poverty in 3 out of poverty than the change in the number of workers 5 divisions: Salahuddin, Central 2 and Basra; where or jobs. As mentioned, we cannot differentiate be- the ratio increased at the bottom of the distribu- tween improvements in quality of jobs, productivity tion. However, the reduction in the consumption- or longer hours of work. to-income ratio during the period in Rest of South and Ninevah explained more than 50 percent of the Although changes in labor income are the main increase in poverty. In other words, poverty did not contributor to poverty reduction in most divisions, fall as it would have had consumption remained a non-labor income also mattered. In those sub-divi- constant share of income. It is important to notice sions with significant decreases in poverty, changes that these results must be interpreted with caution in non-labor income components have contributed to the extent that there is measurement error in positively, i.e., in favor of poverty reduction (i.e. in both components of the ratio. Duhouk, Basra, Salahuddin and Central 2). More- over, the reduction in ration items was overcompen- To conclude, the reduction in poverty observed sated with increases in pensions, other public trans- between 2007 and 2012 was mainly driven by fers and domestic remittances (Figure 142 panel changes in labor income, i.e., growth in earn- A). On the other hand, divisions with increases in ings rather than growth in employment. Non la- poverty also experienced the reduction in transfers bor income components and public and private from rations but not the counterbalancing effect of transfers including pensions and domenstic remit- sufficient increases in other non-labor income com- tances contributed to poverty reduction. At the ponents (Figure 142 panel B). As a result, changes same time, demographic changes have worked in in non-labor income explained a significant part of the opposite direction, limiting poverty reduction. Understanding the Drivers of Poverty Reduction 123 FIGURE 142: Public and Private Transfers A. Reduction in poverty B. Increase in poverty Other private transfers Other private transfers Capital Capital Other public transfers Other public transfers Ration Ration Pension Pension –50 0 50 0 20 40 Central 1 Central 2 Salahuddin Basra Duhouk Ninevah Rest South Source: Authors’ calculations, IHSES 2007 and 2012. Notes: Include only those divisions with significant change in poverty reduction. Central 1 = Diyala and Anbar; Central 2 = Kerbala, Wasit, Najaf, and Babylon and Rest South: Qadisiya, Muthanna, Thi-Qar and Missan.“Capital” includes all incomes from property such as rent from land, non-residential buildings, equipments, shares and profits, interests, among others; “Other public transfers” refers to social protection network compensation and other public transfers in cash and in kind; and “Other private transfers” refers to domestic and international remittances, zakat and other private transfers in cash and in kind. Thus, the labor market in Iraq, as in other coun- economic growth suggests that the relationship be- tries appears to be the most important channel tween growth, employment and earnings is weak. linking growth and welfare. However, the record We next turn to an indepth exploration of these on modest poverty reduction in the face of strong issues. The Growth-Employment Nexus 6 Iraq experienced steady and strong GDP growth, av- benefits and working hours. Earnings have dispropor- eraging a rate of 7 percent per year over the 2007 to tionately increased among public sector employees. A 2012 period but only modest poverty reduction; im- private sector worker still earned almost 30 percent less plying a negative but weak relationship between eco- than public workers in 2012, before including other nomic growth and poverty reduction. For economic benefits associated with public sector employment. growth to have a positive impact on poverty, it needs to generate employment and income for those who need it The expansion of the public sector has adversely affect- the most. Recent economic growth was driven mainly ed the labor market in Iraq by creating economy wide by growth in mining or oil, which represents half of disincentives to labor force participation, especially for the total GDP, the bulk of government revenues, and the youth. An increase of one percentage point in the almost all export revenues. However, it employs a tiny public employment rate of the governorate where the share of the labor force (1%) and has a low output- youth was born contributes to a significant reduction elasticity of employment (–0.2). Across other sectors, (around 50 and 40 percent for 2007 and 2012 respec- in general, employment generation has remained low tively) in their participation rate, keeping everything despite output growth. In contrast, earnings are much else constant. This disincentive to labor force partici- more responsive to output growth in general and for pation is also significant among adults. In general, the mining sector in particular; although the latter reduces the participation rate of both adult women is not likely to directly benefit the poor. Earnings have and men by about 30 percent, holding everything grown at an annual rate of 8.3 percent in the min- else constant. For women, in addition, a significant ing sector, which employs 1 percent of the labor force, gender wage gap further disincentivizes labor force compared to –2.5 percent in agriculture or 0.8 percent participation. in construction, which employ more than a fifth of the population. The nature of public sector expansion has also altered the incentives for education. With the expansion in Although more than 750,000 new jobs were created public sector jobs for less educated men, the returns to over the five year period, these were not enough to ab- education in the public sector level off after primary sorb all new participants in the labor market, and school, and do not pick up again until after tertiary four-fifths of these new jobs were generated by the public education. In 2012, an average male in the working- sector. The expansion of public sector employment has age population received 20 percent more in the public increased the incentives for ‘wait unemployment’ and sector by having complete primary education relative low labor force participation. Having a public sector to being illiterate. However, the premium for getting job translates into having a “better” job relative to the an additional level is almost negligible thereafter un- private sector on several dimensions including wages, til he completes tertiary education. 126 THE UNFULFILLED PROMISE OF OIL AND GROWTH The type of employment growth generated by Iraq over for employment creation for any given growth rate the last five years has only weakly impacted poverty, because its employs a tiny share of the labor force and the expansion of public sector employment has oc- (1%) and has a low output-elasticity of employment curred in sectors where a minority of the poor work. (–0.2).39 This implies that a 1 percent increase in For instance, one-quarter of the poor work in Con- output will generate a 0.2 percent reduction in em- struction, where only 7 percent of jobs are public sec- ployment in the mining sector. Given the growth rate tor jobs. This does not imply more intervention of the experienced in mining since 2007, this means that public sector in these sectors but the contrary. In fact, employment in the mining sector has reduced by 1.3 the expansion of the public sector has adversely affected percent or 0.07 percent of the labor force. the labor market in Iraq by creating economy wide disincentives to labor force participation; inhibiting This pattern is evident not only in mining but the development of the private sector; and distorting across almost all economic sectors. Even when a incentives to invest in education. sector’s GDP has grown over the period, employ- ment generation has remained low. This implies that employment is relatively inelastic to output The Links between Growth and growth overall (Table 28). Among sectors which Employment in Iraq have grown between 2007 and 2012, manufactur- ing is one of the most responsive sectors in terms As mentioned in Chapter 2, Iraq experienced of employment. It employs 10 percent of the labor steady and strong GDP growth, averaging a rate force and it has one of the highest output elastici- of 7 percent per year over the 2007 to 2012 period ties of employment of 0.6. Having experienced an but only modest poverty reduction (3.8 percentage annual growth rate of 9.5 percent over the period, points over the entire five year period). This points this indicates that employment in manufacturing to a negative but weak relationship between eco- sector has increased by 6 percent, or 2.4 percent of nomic growth and poverty reduction, which may the labor force. be due to weak links between economic growth and employment or between employment and earnings, In contrast, earnings are much more responsive to or both. On the other hand, the decomposition output growth in general and for the mining sec- analysis from Chapter 5 shows that the growth of tor in particular; although the latter is not likely to labor income is the most important contributory directly benefit the poor. Earnings have grown at factor to poverty reduction, but that in Iraq, it an annual rate of 8.3 percent in the mining sector, was not driven by employment growth, but rather which employs 1 percent of the labor force, com- by growth in earnings. Thus, an economy where pared to –2.5 percent in agriculture or 0.8 percent growth goes hand in hand with job creation and in- in construction, which employ more than a fifth of come generation, i.e., one where the links between the population (Table 28). Moreover, the output growth and welfare are strong, has the potential elasticity of earnings is also relatively high in pub- to deliver significant poverty reduction. And so, lic administration, the financial services sector, and we turn our attention to the relationship between other services, all of which are primarily public sec- growth, employment and earnings in Iraq. tor jobs. Within the private sector, only transport, storage and communications has a relatively high For economic growth to have a positive impact on earnings elasticity; and accounts for 11 percent of poverty, it needs to generate employment and in- employment. come for those who need it the most. Recent eco- nomic growth was driven mainly by growth in mining 39 We are referring to the direct effect of mining on employ- which represents almost half of the total GDP. How- ment; we are not considering the spillover effect on other ever, of all the sectors, this has the lowest potential sectors. The Growth-Employment Nexus 127 TABLE 28: Growth, Employment and Labor Income Relations Employment Earnings Annual Annual Output Annual Output growth rate Labor Share growth rate elasticity of growth rate elasticity of Economic Activities (%) 2007–12 (%) 2012 2007–12 labor 2007–12 earnings Agriculture & fishing 2.69 8.94 –4.78 –1.78 –2.54 –0.95 Mining & quarrying 7.80 1.02 –1.29 –0.17 8.26 1.06 Manufacturing Industry 9.45 9.67 5.96 0.63 4.74 0.50 Utilities 16.08 2.21 1.87 0.12 5.94 0.37 Construction 15.30 13.63 2.96 0.19 0.82 0.05 Transport, storage & communication 7.96 11.43 3.65 0.46 6.65 0.84 Commerce and retail 12.39 15.61 0.17 0.01 6.93 0.56 Financial, insurance 3.61 14.37 15.51 4.30 3.23 0.89 Public administration 5.66 17.23 –2.81 –0.50 6.95 1.23 Other services 4.00 5.89 1.90 0.47 4.64 1.16 Total 7.09 100.00 1.95 0.27 4.77 0.67 Source: Authors’ calculations, IHSES 2007 and 2012. New Jobs in the Economy: Public challenge for the labor market particularly in an Versus Private Sector economy with relatively weak links between growth and employment. Moreover, the relatively small im- Notwithstanding the weak relationship between provement in human capital of younger generations economic growth and employment, more than 750 poses another challenge for the country. thousand new jobs were created over the five year period. However, these were not enough to absorb The private sector contributed around 60 percent of all new participants in the labor market. In the face total jobs in Iraq in 2007 and 2012. However, it did of significant labor market discouragement, this has not lead new job creation: in the last five years 80 per- translated into stagnant participation and employ- cent of the new jobs were generated by the public sec- ment rates for the working age population during tor. While in 2007, public jobs accounted for 35 per- the period.40 cent of total jobs in Iraq, five years later, their share jumped to almost 40 percent (Figure 143 panel A). Demographic trends in Iraq create opportunities as These types of jobs grew not only across space—in well as challenges for the future growth and poverty reduction. The working-age population has been expanding at an average growth rate of 2.2 percent 40 In chapter 4 “For young women, the main reason cited for per year indicating an 11 percent increase in the not looking for work is social reasons and being a house- wife. In contrast, for young men, the bulk those who fall working-age population between 2007 and 2012. in this category say they are not actively looking for a job The “bulge” among five to fourteen year-olds in because they cannot find a job, and this is another indicator the 2012 population indicates that the working-age of labor market discouragement among young men”. population growth trend will increase over the next 41 Individuals between 5 and 15 year-olds grew annually al- decade.41 While a growing labor force can be an as- most 1 percentage point faster (3 percent per year) than the set for income generation and growth, absorbing a annual growth rate of the working-age population between large wave of new entrants every year poses a major 2007 and 2012 based on IHSES 2007 and 2012. 128 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 143: Public Sector Employment A. Share of Public Jobs B. Economic Sector (% of total jobs) Kurdistan Other services 50 40 Public administration 30 Financial, insurance Iraq Baghdad 20 Transport, storage & 39.1 34.7 10 Commerce and retail – Construction Utilities Manufacturing South North Mining & quarrying Agriculture & shing Central – 20 40 60 80 100 2007 2012 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Baghdad, the North and the South—but also among Additionally, being in a public sector job implies economic sectors such as Financial, Insurance and working fewer hours (about 2 hours less) than in the professional services (with more than half of them in private sector (Figure 146 panel A). The hours gap 2012), Manufacturing and Transport (Figure 143). between both sectors has been increasing between 2007 and 2012. However, there are differences In 2007, the public sector already employed more across space: while this gap increases significantly in than 70 percent of highly educated workers, those Kurdistan and Baghdad, it remains almost constant with higher secondary and tertiary education; and in the South and Central division and reduces in the around 40 percent of total public jobs were con- North (Figure 146 panel B). For instance, in Kurdis- centrated among them (Figure 144 panel A). While tan the difference between the median hours of work the dominance of the public sector as a source of in private and public sector was 2 hours in 2007. This employment for the highly educated has continued, represents 22 percent less hours of work relative to new public sector jobs have also absorbed less edu- the median hours of work in the private sector for cated workers. Most of the new public jobs went that year. In 2012, this gap increases up to 27 percent mainly those with lower than intermediate level of of the median hours of work in the private sector (i.e. education. In particular, 60 percent of the new pub- 9.5 hours). This translates into almost 2.6 less hours lic sector jobs were distributed among workers with of work in a public job relative to a private one. less than primary complete (Figure 144 panel B). Thus, what little job creation has taken place in Iraq Having a public sector job translates into having a over the last five years has been driven by the public “better” job relative to the private sector on several different dimensions such as retirement plan. Figure 42 According to last consultation with country counterparts 145 shows that around 90 percent of public work- held in Erbil–May 2014, all public jobs include health in- ers have a retirement plan irrespective of the year surance. However, the 2012 IHSES presents inconsisten- compared to their private counterparts who have no cies in the variable that measures this benefit; thus results benefits at all.42 are not presenting in this report. The Growth-Employment Nexus 129 FIGURE 144: Public Jobs – Education Levels A. Share of Public Jobs among levels B. Distribution of new public jobs 80 70 60 10.7 23.4 50 11.7 40 30 6.4 20 –0.7 10 11.9 36.5 – Incomplete primary Complete primary Intermediate Secondary Higher Secondary Tertiary Illiterate Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary 2007 2012 Tertiary Source: Authors’ calculations, IHSES 2007 and 2012. FIGURE 145: Private and Public Jobs: Retirement population, and limit active job search as people wait and queue for openings in the public sector. More- 100 over, the private sector’s ability to attract talent is 90 80 also severely constrained by an expanding and in- 70 creasingly well-paying public sector. In the medium 60 term, the anticipation of public sector employment 50 40 can also distort education investments, by catering 30 to certain fields of study that may be irrelevant to 20 the private sector. Below we quantify some of these 10 adverse consequences of the expansion of public – Retirement private Retirement public sector employment on youth, male and female labor force participation; on education investments; and 2007 2012 on the potential for private sector development. Source: Authors’ calculations, IHSES 2007 and 2012. Youth Participation in the Workforce sector. A large and expanding public sector can cre- Youth labor force participation in Iraq is low and ate significant distortions in the labor market, espe- stagnant. Only 40 percent of Iraqis between the ag- cially in a country where there is already low labor es of 15 and 24 who are not in school, are employed force participation, even among men, and labor or looking for work. This means that 6 out of every market discouragement is widespread. The increas- 10 young Iraqis who are not in school are neither ing possibility of a public sector job, with greater working nor actively seeking work. However, labor job security, benefits and fewer hours of work, can force participation rates vary significantly by the raise the reservation wage among the working age level of education of the youth. It ranges from 25 130 THE UNFULFILLED PROMISE OF OIL AND GROWTH Private and Public Jobs: Gaps on Hours of Work Per Day FIGURE 146:  A. Median hours of work per day B. Gaps of median hours of work by region 10 Kurdistan 9 30 25 8 20 7 Iraq 15 Baghdad 6 10 5 5 – 4 3 2 South North 1 – 2007 2012 Gap Central Private Public 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. percent for illiterate individuals to 66 percent for a job in 2007 compared to 35 percent in 2012 (Fig- those with tertiary education (Figure 147 panel A). ure 147 panel B). Conversely to the participation rate, the employment rate does not vary much by Given these low participation rates, it is not surpris- education. This implies much higher unemploy- ing to find low levels of employment among this ment rates among the highest educated youth. In young cohort. 33 percent of this population found 2012, unemployment rises to almost 25 percent FIGURE 147: Youth Participation and Employment Rates, by Education Level A. Participation rate B. Employment rate 100 100 90 90 80 80 70 70 60 60 40.6 40.1 50 50 34.6 33.3 40 40 30 30 20 20 10 10 – – Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. The Growth-Employment Nexus 131 FIGURE 148: Type and Sector of Work, by Education Level A. Salaried B. Public sector 100 100 90 90 80.4 80 80 70.4 70 70 60 60 50 50 40 40 30 30 19.4 19.2 20 20 10 10 – – Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. among youth with tertiary education in comparison Financial, Insurance and professional services (Fig- to about 5 percent for illiterate. ure 149 panel A). As a result of this sectoral shift among the youth, the education levels within each The majority of employed youth were engaged in sector have remained largely unchanged, with the salaried jobs irrespective of their level of education. most educated youth are still being employed in In 2007, salaried employment accounted for 70 services and the less educated are absorbed main- percent of all employed youth with a variation of 20 ly by Construction (Figure 149 panels B and C). percentage points between the lowest and highest However, average education levels within the pub- education level. In 2012, the share of salaried em- lic administration and mining sectors has increased, ployment increased at all levels of education (or in with more young workers having relatively high other words, self-employment increased) except ter- education. tiary (Figure 148 panel A). Salaried workers consist of two distinct groups, public and private workers, with public sector jobs making up a fifth of salaried What contributes to labor force employment. The former has remained at almost 20 participation among youth? percent over the five year period. Between 2007 and To investigate the hypothesis that youth labor force 2012, there was a 17 percentage point decrease in participation rates are low because youth are queu- the share of tertiary educated workers employed in ing for a job in the public sector, we calculate the the public sector; but since their share in the youth marginal effects of the probability of being active population is so small, the overall rates of public sec- conditioned on several individual and household tor employment remain largely unchanged (Figure characteristics of the young population for 2007 148 panel B). and 2012. We find that an increase of one percent- age point in the public employment rate of the gov- Between 2007 and 2012, the youth have moved out ernorate where the individual was born contributes mainly from Agriculture, Public Administration and to a significant reduction (around 50 and 40 percent Commerce to Construction, Manufacturing and for 2007 and 2012 respectively) in the participation 132 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 149: Sectors of Employment, Youth I. Share of employment by economic sector Other services Public administration Financial, insurance Transport, storage & Commerce and retail Construction Utilities Manufacturing Mining & quarrying Agriculture & shing 0 5 10 15 20 25 30 2007 2012 II. 2007: Education structure by economic sector III. 2012: Education structure by economic sector 100% 3.1 30.7 7.8 21.9 5.0 9.9 9.1 10.7 31.5 7.3 100% 2.2 9.1 5.2 13.0 3.8 9.0 9.7 13.5 39.8 7.4 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 58.3 21.9 42.7 22.8 47.3 34.5 32.4 38.3 22.7 41.9 10% 67.4 28.9 52.3 39.8 57.0 43.9 46.1 38.0 18.8 48.1 0% 0% Agriculture & shing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & Financial, insurance Public administration Other services Agriculture & shing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & Financial, insurance Public administration Other services Illiterate + Incomplete primary Prim. Complete + Intermediate Secondary and above Source: Authors’ calculations, IHSES 2007 and 2012. rate among the youth, keeping everything else con- public employment rates: minimum or low, mean stant (Table A 6.1).43 and maximum (Figure 150). Being born in a governorate where the public sec- These graphs provide clear evidence in favor of the tor contributes a larger share of the jobs in the local hypothesis that youth queue for a job in the pub- economy relative to other governorates creates dis- lic sector independent of their education level. Both incentives to participate in the labor market among the youth. To illustrate by how much the public 43 To avoid problems associated with the endogeneity of mi- employment rate affects the labor participation of gration, we use the jurisdiction of birth rather than that of this cohort, we calculate the probability to partici- current residence to calculate the relevant public employ- pate by education level considering three levels of ment rate as Asaad (2013) The Growth-Employment Nexus 133 FIGURE 150: Youth Participation Rate by Public Employment Rate in the Governorate of Birth A. 2007 B. 2012 0.8 0.8 0.7 0.7 Probability to participate 0.6 0.6 in the labor market 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 0 Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Low Mean Max Source: Authors’ calculations, IHSES 2007 and 2012. graphs show on the one hand, that the probability member working in the public sector as well as living to participate in the labor market increases with the in urban areas, in households where the net income level of education of the young independent of the is relatively high and where the household head has share of public employment. For instance, in 2012, more than incomplete primary education; reduces around 40 percent of the youth with less than inter- the probability to participate in the labor market.44 mediate level education were willing to participate in the labor market if they were born in a governorate On the other hand, there are many factors which with low public employment rate. These participa- encourage youth labor force participation in 2007 tion rate levels increase to 70 percent for youth with and 2012. Among these, we find that being male higher secondary or more in similar governorates. increases participation significantly by about 70 per- cent for both years. Being relatively older and living On the other hand, the higher the share of employ- in a rich governorate increases the probability by ment in the public sector, the lower the participa- about 3 and 8 percent respectively. The individual’s tion rates among the youth irrespective of their level level of education also has a positive effect on the of education. For example, the participation rate probability to participate. The higher the education among those with tertiary education decreases 30 level, the higher the impact on participation, rang- percentage points (from 70 to 40 percentage) if the ing from 5 up to 20 percent higher with respect to youth were born in a governorate with low pub- being illiterate (Figure 151). lic employment rates relative to a governorate with high rates of public employment. In other words, the bigger the public sector, the higher the disincen- Female Participation in the Labor Force tive to participate among the youth. Only 15 percent of adult Iraqi women of working age participate in the labor force; well below the already We also note that there are other factors which con- tribute to lower labor force participation among 44 Net income includes labor income from other household youth. For instance, having another household members and from non-labor sources. 134 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 151: Marginal Effects of Being Active 2007 and 2012 (Figure 152 panel B). Conse- – Youth Population by Education quently, the most educated women also have the Level highest rates of unemployment of around 5 to 6 percent. 0.30 Marginal e ect of being active 0.25 Overall, 70 percent of employed women work in 0.20 salaried employment (Figure 153 panel A). Among 0.15 the few less educated women who work, the major- 0.10 ity are self-employed: about a third of women with primary education or less work in salaried jobs. In 0.05 contrast, with higher education, the likelihood of 0 working in a salaried job more than doubles: for Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary instance, almost all employed women who have secondary or higher education are employed in a salaried job. This pattern is driven by differences in access to public jobs by education level. Almost 2007 2012 two-thirds of employed women work in the public sector, varying from around a fifth of women with Source: Authors’ calculations, IHSES 2007 and 2012. primary education or less to almost all employed women with secondary or higher education (Figure 153 panel B). low rates of female labor force participation in the Middle East and North Africa region of around 25 Between 2007 and 2012, adult women have moved percent. The share of women aged 25 to 64 who primarily from the public administration, health are either employed or actively seeking work has and education services sector, which still employs remained stagnant between 2007 and 2012. Par- 40 percent of women, to mostly public sector jobs ticipation varies sharply by education. Among adult in financial, insurance and professional services, and women with intermediate or lower education, labor private sector jobs in commerce and retail. In ad- force participation rates are below 10 percent. With dition, there was a small decrease in employment secondary education, these rates more than double in agriculture (Figure 154 panel A). Almost all to 24 percent, and increase six-fold with secondary women who work in agriculture have low levels of and tertiary education. Well educated Iraqi women education; in contrast, almost 90 percent of women are therefore at least six times as likely to work or to employed in the public administration sector have be looking for work compared to the vast majority secondary and higher education. As a result of shift of Iraqi women (Figure 152 panel A). in women’s employment towards financial services and commerce, the share of highly educated women This pattern is almost entirely reflected in employ- in the former has come down, whereas the share of ment rates for adult women. Almost all of labor highly educated women in commerce has increased force participation comprises of employment, with (Figure 154 panels B and C). negligible rates of unemployment at each level of education. Less than 10 percent of women with intermediate education or less are employed, but What contributes to labor force almost 60 percent of women with higher second- participation among adult women? ary or tertiary education work. However, there has As in the case of the youth, we analyze how the pub- been an 8 percentage point decline in employment lic sector affects adult female participation in the la- among the most highly educated women between bor market. To do so, we consider women between The Growth-Employment Nexus 135 FIGURE 152: Participation and Employment Rates, Adult Women A. Participation rate B. Employment rate 100 100 90 90 80 80 66.4 64.7 62.2 61.9 70 70 60 60 50 50 40 40 30 30 16.0 15.3 15.1 14.2 20 20 10 10 – – Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. FIGURE 153: Type and Sector of Work, Adult Women A. Salaried B. Public sector 100 100 90 90 80 80 70.5 70.5 66.1 63.6 70 70 60 60 50 50 40 40 30 30 20 20 10 10 – – Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. 25 and 64 years old and we divide them into two percentage point in the public employment rate of groups accordingly to their marital status: currently the governorate where the woman was born would married and non-married (including single, widowed contribute to a significant reduction, in general and separated); and skill level: less educated—less about 30 percent, in her participation holding every- than primary complete, and highly educated—more thing else constant (see Table A 6.2 and A 6.3). The than primary complete. This results in four mutu- exception to this result is married women with high ally exclusive groups. We find that an increase of one education levels in 2012. One possible explanation 136 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 154: Economic Sector of Employment, Adult Women A. Share of employment by economic sector 60 50 40 30 20 10 0 Agriculture Manufacture Commerce Financial Public 2007 2012 B. 2007: Education structure by economic sector C. 2012: Education structure by economic sector 100% 6.7 39.0 34.9 60.7 91.2 100% 3.6 42.8 42.7 56.8 89.5 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 65.0 31.0 32.1 13.3 3.0 0% 68.2 25.1 32.2 23.8 2.2 Agriculture Manufacture Commerce Financial Public Agriculture Manufacture Commerce Financial Public Illiterate + Incomplete primary Prim. Complete + Intermediate Secondary and above Source: Authors’ calculations, IHSES 2007 and 2012. of this is that the overall improvement in economic women who are more qualified for public sector conditions over the last five years may have increased jobs, this effect acts as an incentive to participate the pool of highly educated women who were also possibly because of better access to networks and competing for the relatively scare number of public connections to get these types of jobs. sector jobs. This is measured through the negative coefficient on the local labor supply measure. At the Among the factors that contribute to higher labor same time, the opportunity cost of working appears force participation, we find evidence that social and to be increasing for these women. This is captured gender norms within the household, which are mea- by the stronger relationship of having children on sured by the presence of a working woman in the the decision to participate. household, matter. This factor increases the labor force participation by at least 10 percent. The de- In general, the higher the net income, the lower the cision to participate is unaffected by the education participation rate among women. We also note that of the spouse or the household head. Finally, own the presence of another member employed in the education independently promotes the labor force public sector acts as a disincentive to participate for participation even after controlling for a range of less educated women. However, for highly educated household and individual characteristics. The Growth-Employment Nexus 137 Male Participation in the Labor Force panel A). A majority of men work in salaried em- More than four-fifth of adult Iraqi men of work- ployment at each level of education, and the share ing age are either employed or actively looking for of self-employment has fallen at each level of educa- work. While these rates of labor force participation tion between 2007 and 2012. At the highest lev- are much higher than those among adult women els of education, more than four-fifths of employed and have remained almost stagnant between 2007 men work in salaried jobs. Unlike for women, this and 2012. Unlike for women, male labor force par- is not entirely driven by public sector employment, ticipation rates are on average higher among the less except for the highly educated. While the share educated than among the better education. For in- of employment in the public sector has increased stance, in 2012, while 87 percent of men aged 25 to across education levels, less than 50 percent of men 64 participated in the labor force, this drops 10 per- with secondary education in the public sector. In centage points among illiterate educated men and contrast, 70 percent of tertiary educated men work increases to more than 90 percent among secondary in the public sector. However, on average, 58 per- educated men (Figure 155 panel A). cent of adult men are employed in the private sec- tor; driven by a greater dependence on the private As with women, the variation in labor force partici- sector as a source of employment among the major- pation by education is reflected in employment rates ity of adult men who have relatively low levels of at different levels of education. Overall, 85 percent education (Figure 156 panel B). of adult men are employed, a rate that varies from 70 percent among the less educated men to 90 per- In 2012, three sectors—commerce and retail, public cent among men with tertiary education (Figure administration, health and education; and financial, 155 panel B). insurance and professional services—together ac- counted for almost half of all employment for men More than 70 percent of employed men work in (Figure 157 panel A). Between 2007 and 2012, adult salaried employment, which represents an increase men have moved away from agriculture, commerce of 6 percentage points relative to 2007 (Figure 156 and retail and public administration, predominantly FIGURE 155: Participation and Employment Rates, Adult Men A. Participation rate B. Employment rate 100 100 87.4 86.1 84.9 82.7 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 – – Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. 138 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 156: Type and Sector of Work, Adult Men A. Salaried B. Public sector 100 100 90 90 73.2 80 80 67.5 70 70 60 60 42.0 50 50 34.8 40 40 30 30 20 20 10 10 – – Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Total 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. into the financial services sector, which has almost conditioned on several individual and household doubled its share in male employment to 15 percent. characteristics. The results show that being born in a governorate with high public employment rates is a The majority of male workers in agriculture and con- powerful disincentive to participate in the labor mar- struction are less educated: more than 80 percent of ket, lowering it by almost 30 percent (Table A 6.4). male workers have intermediate or lower education. In contrast, more than half of male employees in Another factor that operates in the same direction mining and public administration have secondary is higher levels of household net income which or higher education levels. With the shift towards reduces by almost 4 percent the likelihood of par- male employment in the public sector-dominated ticipating in the labor market. In other words, the financial services sector, the average education level higher the household net income, the lower the ne- in this sector has come down, suggesting that these cessity to work or look for work for adult men. As new jobs attracted relatively less educated men (Fig- expected, older men, living in urban areas, in richer ure 157 panel B and C). A similar trend is noticeable governorates, head of household, married and with in the utilities sector, which is also predominantly high levels of education have higher labor force par- public sector. ticipation rates. At a time when labor market outcomes have ex- What contributes to labor force hibited very little dynamism, the primary mover in participation among adult men? the labor market has been the public sector. But The expansion of jobs was led by the public sector, irrespective of the group of analysis, the presence which accounts for more than 40 percent of adult of a large public sector in the local labor market male employment. We now explore whether queu- has acted as a significant deterrent to labor force ing for public sector jobs inhibits adult male partici- participation, which has potentially further weak- pation. In order to do so, we estimate the likelihood ened the relationship between economic growth to participate in the labor market for this cohort and employment. The Growth-Employment Nexus 139 FIGURE 157: Sector of Economic Activity, Adult Men I. Share of employment by economic sector Other services Public administration Financial, insurance Transport, storage & Commerce and retail Construction Utilities Manufacturing Mining & quarrying Agriculture & shing 0 5 10 15 20 25 2007 2012 II. 2007: Education structure by economic sector III. 2012: Education structure by economic sector 100% 100% 90% 16.6 55.0 35.0 55.2 15.1 30.0 22.4 45.4 52.7 24.1 90% 11.9 52.9 34.4 32.9 11.6 28.4 21.0 28.5 55.8 38.2 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 41.6 8.3 20.4 11.9 27.5 21.1 21.1 14.0 13.7 26.3 0% 49.4 11.0 23.5 29.1 40.7 26.9 29.8 25.3 11.6 26.4 Agriculture & shing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & Financial, insurance Public administration Other services Agriculture & shing Mining & quarrying Manufacturing Utilities Construction Commerce and retail Transport, storage & Financial, insurance Public administration Other services Illiterate + Incomplete primary Prim. Complete + Intermediate Secondary and above Source: Authors’ calculations, IHSES 2007 and 2012. What Explains Differences in 1000 firms across the country.45 The BES is a na- Earnings? tionally representative, firm-level survey conducted periodically by the World Bank Group in the devel- At the same time, another dimension in which the oping world. It collects data from non-agricultural, public sector affects the labor market in the medium private-owned, registered firms with more than term is through distorting the incentives to invest in 5 employees. The survey covers a broad range of education by catering to certain fields of study that business environment topics including investment may be irrelevant to the private sector. To provide as well as their balance sheets and performance some evidence from the demand side on this partic- ular problem, we use the 2011 Iraq Enterprise Sur- 45 The BES was collected in Iraq between March 2011 and vey (BES), which collects information over almost October 2011. 140 THE UNFULFILLED PROMISE OF OIL AND GROWTH measures. Given that this survey is representative at there are significant differences in how these returns the national level, it is ideal for obtaining a broader have varied across levels of education for male and picture of the investment environment in Iraq. female workers over time. The BES finds that one of the major business con- We calculate the returns to education for the work- straints is inadequately educated labor force. 34 per- ing-age population (between 15 to 65 years old) by cent of the firms identified this as a major constraint, gender in 2007 and 2012. We see that returns have which is almost 13 percentage points higher than the increased significantly for almost all levels of edu- average in the region and 7 percentage points higher cation irrespective of being male or female (Figure than among all countries (Figure 158). This could be 159). However, for males, the gap has been expand- a reflection of an important supply side challenge for ing over the last five years in a faster rate than that the growth of the private sector. Iraq’s private sector observed for females (Figure 159 panel A). Note also faces a range of other challenges to growth and also that the magnitude of differences in returns be- job creation (See Box 4 for the main findings from tween low and high educated women is bigger than the 2012 Investment Climate Assessments). for men. Moreover, these differentials in returns seem to not have significantly changed for women Earning functions for the Iraqi labor market are es- with less than secondary education but have wid- timated in order to understand the factors that ac- ened for women with higher secondary or tertiary count for differences in earnings levels and growth education. This may be related to labor market seg- rates. As expected, the main factors include sector of mentation between low and high educated women occupation, education, experience and geographic (Figure 159 panel B). location. Even though the educational structure of the Iraqi working age population has not signifi- So far it appears that the labor market is generating cantly changed over the last five years; the returns to the correct incentives to invest in education. How- education have increased for all workers. However, ever, these dynamics seem to differ in the public FIGURE 158: Business Constraints Electricity Practices of the informal sector Political instability Access to land Major business constraints Inadequately educated workforce Business licensing and permits Labor regulations Access to nance Tax rates Customs and trade regulations Corruption Crime, theft and disorder Tax administration Transportation 0 5 10 15 20 25 Iraq Middle East & North Africa All countries Source: Authors’ calculations, World Bank Enterprise Survey, Iraq, 2011. The Growth-Employment Nexus 141 BOX 4: Doing Business in Iraq – Main Messages from the Investment Climate Assessment 2012 The private sector has limited presence and incentives for its expansion are nearly absent in Iraq. Accordingly to the last Investment Climate Assessment (ICA) for Iraq, this is a consequence of decades of socialist economic policy which have tightly bound Iraq’s economy to the state. Although the potential for private sector engagement in post-conflict Iraq is significant, the country’s private sector is both under-developed and unsupported. Most private businesses in Iraq are very small and informal, mainly operating in retail and trade, construction and transportation services, as well as in light industry in the textile, food, engineering, and chemicals fields. The Enterprise Survey suggests significantly higher costs of doing business in Iraq than elsewhere in the region. Respondents to this survey listed electricity, instability, and corruption as the three most severe obstacles to doing business in Iraq. The average firm suffered nearly 600 power interruptions per year, with some firms reporting three outages per day. This unstable power supply results in significant sales losses. As a reflection of the political instability in Iraq, security and crime are also major concerns, with many surveyed firms spending considerably on protection. Note however, that the importance of these problems often varies significantly between regions and sectors. Corruption further acts as an economic burden on firms surveyed, with firms reporting that bribes are commonplace, often to smooth frequent inspections from authorities, including government agencies and municipal police forces. Among other problems identified by this report are the following: poor access to finance and land; and competition from State-Owned Enterprises (SOEs) and the informal sector. Credit systems and access to finance are severely limited because of the state’s long-term dominance. Nearly half of the firms surveyed reported having great difficulty in obtaining financing; fewer than 7% reported having a loan. Some small and micro-enterprises have no access to banks with markedly difficulties for the smaller ones. Firms also reported poor access to land for investment in general caused by widespread government ownership of prime land and inefficient methods of allocation to the private sector. With regard to competition from the informal sector, firms report difficulties confronting competition from unregistered companies, many of which infringe upon the rights of registered businesses, and face lower costs due to their noncompliance with formal rules. These are few of multiple constraints that private firms faced on doing business in Iraq such as proper transport or access to adequate labor force. It is not surprising that the ICA report emphasizes the fact that The World Bank’s Doing Business 2012 Report ends ranking Iraq 164 out of 183 economies. Iraq ranks second to last amongst the economies in the Middle East and North Africa region (MENA), significantly behind regional best performers on a global level such as Saudi Arabia (12), United Arab Emirates (33), and Qatar (36). Source: Investment Climate Assessment—Iraq, World Bank (2012). FIGURE 159: Evolution of Returns to Education by Gender A. Males B. Females 0.40 1.4 0.35 1.2 0.30 1.0 0.25 0.8 0.20 0.6 0.15 0.10 0.4 0.05 0.2 0 0 Incomplete primary Complete primary Intermediate Secondary Higher Secondary Tertiary Incomplete primary Complete primary Intermediate Secondary Higher Secondary Tertiary 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. 142 THE UNFULFILLED PROMISE OF OIL AND GROWTH and private sector labor markets. From the public only get 6 percent higher returns for having com- sector’s viewpoint, the incorporation of less edu- pleted primary school. However, the private sector cated workers has produced an expected adjustment catches up for the highest educated males by offer- on their returns relative to highly educated public ing similar premiums as the public sector. Note that workers irrespective of gender. Figure 160 shows the private sector pays much less than the public sec- that there was no significant increase in the returns tor irrespective of gender. of the highly educated workers while the less edu- cated ones experienced the opposite relative to il- A similar story is true for Iraqi women (Table 29). literate workers. On the other hand, what occurred For the few women who are employed, the pre- in the private sector was that the gap was widened mium for getting primary school is higher than for significantly only for highest educated males. males, increasing from 11 percent up to 25 percent relative to illiterate women in the private and public The returns to education in the labor market can sector respectively. In contrast to their male coun- profoundly affect the decision to invest in education terparts, women received an increase again of about and the level at which education is stopped. When 30 percent more by obtaining secondary school ir- we examine the earnings gains for each additional respective of the sector (Table 29). This would be education level, we see that in 2012 an average male mainly driven by the types of activities (i.e. secretar- in the working-age population receives 20 percent ies, assistants, etc.) that most of these workers likely more in the public sector by having complete pri- perform in comparison to males. The next big jump mary education relative to being illiterate. How- in women’s education premium happens when they ever, the premium for getting an additional level is get up to tertiary. almost negligible thereafter except when he com- pletes tertiary education, when he obtains roughly Hand in hand with the growth in employment in an additional 11 percent increase (Table 29). The the public sector, earnings have disproportionately private sector is not as generous as the public sector increased among public sector employees, from al- at lower levels of education and a similar male will ready high levels compared to the private sector. At FIGURE 160: Evolution of Returns to Education by Gender – Public Sector A. Males B. Females 0.40 0.9 0.35 0.8 0.30 0.7 0.25 0.6 0.5 0.20 0.4 0.15 0.3 0.10 0.2 0.05 0.1 0 0 Incomplete primary Complete primary Intermediate Secondary Higher Secondary Tertiary Incomplete primary Complete primary Intermediate Secondary Higher Secondary Tertiary 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. The Growth-Employment Nexus 143 TABLE 29: Returns to Education by Gender characteristics. Figure 161 panel A shows that, even and Sector – 2012 after controlling for these differences, about 13 per- cent of the gap remains unexplained in 2012. How- Private Public ever, almost 50 percent of the gap is explained by Male Female Male Female differences in characteristics among similar workers Incomplete primary 0.0136 0.0610 0.108*** 0.199 in both sectors. This may reflect the fact that the type of activities performed by private workers de- Complete primary 0.0635*** 0.116 0.189*** 0.253** mand lower levels of qualification than their public Intermediate 0.0691** 0.170 0.196*** 0.175 counterparts. This pattern has not changed over the Secondary 0.129*** 0.405 0.225*** 0.509*** period of analysis. Higher Secondary 0.118** 0.275 0.247*** 0.728*** Tertiary 0.354*** 0.902*** 0.356*** 0.773*** To disentangle these differences between private Source: Authors’ calculations, IHSES 2007 and 2012. and public workers, we performed the same analysis for men and women workers separately. We see that the same time, between 2007 and 2012, even pri- the unexplained factor is the smallest contributory vate sector workers experienced an increase in their factor for men and the differences in human capital earnings although they continue to be at a disadvan- characteristics (education, experience, sector of em- tage in the labor market relative to their public sec- ployment, etc) increases their contribution to almost tor counterparts. A private sector worker still earned 60 percent in 2012. In other words, there seems to almost 30 percent less than public workers in 2012, be no significant discrimination factors that explain before including other benefits associated with pub- differences in earnings between private and public lic sector employment. A portion of this difference workers for men (Figure 161 panel B), but rather, is likely due to differences in the endowments of public sector employees have better characteristics public and private sector workers like education lev- relative to those employed in the private sector. els, experience, among others. In order to under- stand the factors underlying public/private income On the other hand, women seems to be more ex- differentials, we utilize the Oaxaca-Blinder decom- posed to labor segmentation between public and position method. This approach is typically used in private activities given the lack of common support order to disentangle the share of the wage gap at- among these types of workers. Figure 161 panel C tributable to characteristics from the share attribut- shows that more than 60 percent of the gap among able to an unexplained component (which may be women is explained by this factor. As we have al- due, in part, to discrimination). In the Iraqi context, ready seen women with low levels of education the most problematic assumption of this approach is are mainly concentrated in private sector activities the failure to recognize the potential for deep seg- which are mainly informal (i.e. with almost no ac- mentation between the kinds of jobs that public and cess to health insurance or retirement plan) and in private workers perform. Construction of counter- economic sectors such as agriculture and commerce factuals, i.e. private workers’ earnings if rewarded at that pay much less than their public counterparts. public prices, may be meaningless if private workers On the other hand, we find that it is the highest cannot be matched to public workers with similar educated women who have access to “better” jobs, characteristics. As described in Nopo (2008), failure in terms of access to benefits as well as high levels of to recognize this problem implies in overestimation earnings; mainly concentrated in public sector. of the unexplained component of the wage gap. Another gap relevant to our analysis is the difference In order to take this problem into account, the between male and female earnings. If there is a sig- earning decomposition is undertaken after “match- nificant gender wage gap, so that women with simi- ing” public and private workers that have “similar” lar characteristics systematically earn less than men 144 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 161: Decomposing Differences in to women irrespective of the sector they are work- Earnings ing in. In general, in 2012, males earned 20 per- cent more than women if they were working in the A. Public versus private earnings gap public sector. Moreover, in contrast to 2007, when 100% up to 26 percent of the gap could be explained by 90% 80% 38.1 human capital characteristics, by 2012, the contri- 49.6 70% bution of this factor is less than 13 percent, despite 60% 50% the slight improvement in women levels of educa- 40% tion (Figure 162 panel A). The primary reason why 30% 20% women earned less than men is largely unexplained 10% and not attributable to the idea that women lack 0% qualifications for being public servants; perhaps is 2007 2012 attributable to discrimination against them. This B. Public versus private earnings gap-Men gender wage gap dramatically increases up to almost 100% 8 times of that observed in the public sector when 80% considering workers in the private sector, which may 53.5 60% 57.1 explain why women who have high enough educa- 40% tion work primarily in the public sector (Figure 162 20% panel B). Again, the main contributor to this differ- ence is the unexplained factor. 0% –20% 2007 2012 The type of growth generated by Iraq over the last five years has only weakly impacted poverty. The C. Public versus private earnings gap-Women analysis of the previous chapter shows that it was 100% 3.9 mainly through the increase in labor income rather 80% than with the creation of new jobs that some of the 60% poor jumped the threshold. Throughout this chap- 40% ter, we also learned that even though the new jobs, 20% which were mainly created by the public sector, are 0% much “better” in terms of hours and benefits, these –20% –114.6 did not benefit many of the poor. Figure 163 shows –40% 2007 2012 that the poor are concentrated in those economic sectors where the public sector generates a smaller Unexplained share of employment. For instance, one-quarter of Di erences due to inability to match workers the poor work in the construction sector, where 93 Di erences in characteristics of matched workers percent of jobs are in the private sector. This does not imply more intervention of the public sector in Source: Authors’ calculations, IHSES 2007 and 2012. these sectors but the contrary. In fact, the expan- sion of the public sector has adversely affected the labor market in Iraq by creating economy wide dis- with similar characteristics, this may be a further dis- incentives to labor force participation; inhibiting the incentive for women to work. We perform two de- development of the private sector; and distorting in- compositions of the gender wage gap based on the centives to invest in education. In the next chapter, sector of employment: public and private. Men earn we turn to a closer examination on the labor market a substantial premium in the labor market relative facing the poor. The Labor Market for the Poor: The Rural-Urban Divide 145 FIGURE 162: Decomposing Gender Differences in Earnings A. Male versus Female earnings gap-Public B. Male versus Female earnings gap-Private 100% 100% 26.2 12.7 80% 80% 38.0 60% 60% 40% 40% 20% 20% 0% 0% –6.9 –20% –20% 2007 2012 2007 2012 Unexplained Di erences due to inability to match workers Di erences in characteristics of matched workers Source: Authors’ calculations, IHSES 2007 and 2012. FIGURE 163: Evolution of Employment by Economic Sector A. Share of public jobs by economic sector B. Distribution of poor by economic sector Other services 43.6 Other services 6.1 42.8 5.4 Public administration 96.4 Public administration 8.9 95.3 13.3 Financial, insurance 65.7 Financial, insurance 10.5 35.0 4.9 Transport, storage & 24.3 Transport, storage & 12.1 17.8 8.8 Commerce and retail 4.8 Commerce and retail 10.0 3.6 12.3 Construction 7.7 Construction 25.2 9.3 19.4 Utilities 76.0 Utilities 2.8 84.2 1.1 Manufacturing 33.9 Manufacturing 7.8 23.5 9.2 Mining & quarrying 86.8 Mining & quarrying 0.4 91.7 0.8 Agriculture & shing -2.2 Agriculture & shing 16.2 24.8 – 20 40 60 80 100 0 5 10 15 20 25 30 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. The Labor Market for the Poor: The Rural-Urban Divide 7 T he labor market for the poor looks significantly between 2007 and 2012, rural Iraq does not show signs different from that facing the non-poor in of catching up and continues to lag behind significantly. Iraq, and it varies considerably across rural and urban areas. Poverty is not only correlated with The urban labor market and poverty: lower rates of employment and labor force participa- tion, but also with important differences in the types Between 2007 and 2012, the employed urban poor be- of economic activities. Comparing urban and rural came increasingly concentrated in three sectors, in ad- households, not only are the characteristics of rural dition to commerce and retail—construction; trans- households starkly different—larger household sizes port, storage and communication; and financial, and lower educational attainment, for instance—, insurance and professional services—which together but even for the same characteristics, poverty rates are accounted for approximately 58 percent of the urban much higher for rural households. employed poor. Most of the urban poor work in the pri- vate sector. Poverty rates among households with heads In terms of the labor market, the most important employed in the public sector are significantly lower headline indicators that distinguish non-poor house- than among other urban households. Households with holds from poor households are higher participation heads employed in the public sector earn the highest per and employment rates, and in particular, public sec- capita income, driven by the highest per capita labor tor employment rates. In 2007, 76 percent of the ur- income, compared to other types of urban households. ban poor and 81 percent of the rural poor worked in The results of the Oaxaca-Blinder decomposition the private sector, compared with 60 percent and 72 method show that households with heads employed in percent of the urban non-poor and the rural non-poor. the public sector had on average better characteristics, Over time however, the role of the public sector as a which were associated with lower poverty. source of employment has increased, especially for the non-poor, but also for the poor. In 2012, 27 percent of The rural labor market and poverty: the employed urban poor and 22 percent of the em- ployed rural poor worked in the public sector. In terms of the type of employment, the most important change between 2007 and 2012 among rural house- The differences in the sectors of employment are only one holds has been in the large increase in the share of the of the many differences in characteristics between urban rural poor working in salaried non-farm employment and rural households. Rural households have on aver- rather than in self-employed farm work. There has age much lower levels of education than urban house- been a shift away from agriculture among the rural holds, and while educational attainment among the poor: in 2007, 47 percent of the rural poor worked in urban working age population has remained stagnant this sector, compared to 30 percent in 2012. 148 THE UNFULFILLED PROMISE OF OIL AND GROWTH In 2007, agriculture was the main sector of employ- demographics or whether welfare improved only for ment for 27 percent of employed Iraqi women and 10 particular types of households, perhaps because the percent of employed Iraqi men. By 2012, agriculture’s returns to employment in certain sectors improved. share in female and male jobs had fallen to 23 percent for women and 7 percent for men. The bulk of this declining dependence on agriculture occurred in the Where do the Poor Work? North, Centre and the South; which all recorded sharp declines in agricultural households accompanied by Compared with non-poor households, labor force no change in or decreases in the share of diversified participation rates and employment rates are lower households. among men and women in poor households. Fe- male labor force participation rates among the rural Overall, within the 2007 to 2012 period, rural poverty poor were almost three and a half times higher than reduction was driven by households who were diversi- among the urban poor in 2007; and although ru- fied—with at least one member employed in agricul- ral female workforce participation has declined by 6 ture, and at least one working off the farm. Poverty percentage points since then, it is still higher than in rates fell sharply (almost halved) among diversified urban areas (able 30). While male participation and households, while there was little change in the welfare employment rates are fairly similar among the rural of other types of households. The primary reason why ag- poor and the urban poor, one important difference riculture households are poorer than diversified house- is the decrease in male participation and employ- holds is largely explained by differences in the coefficients ment rates among the urban poor between 2007 and not attributable to differences in characteristics. and 2012; and the decrease in female participation Welfare improvements experienced by households who and employment rates among the rural poor during were diversified were limited to Kurdistan, the North the same period. and the Centre. In Baghdad and the South, poverty increased for almost all types of rural households. In terms of the type of employment, the most im- portant change between 2007 and 2012 has been Overall, poverty reduction has been largely ex- in the large increase in the share of the rural poor plained by increases in labor income, and over the working in salaried non-farm employment rather five year period between 2007 and 2012, new job than in self-employed farm work. This is turn is re- creation and significant increases in labor earnings flected in the shift away from agriculture among the have been concentrated in the public sector. Nei- rural poor: in 2007, 47 percent of the rural poor ther employment nor earnings have expanded as worked in this sector, compared to 30 percent in fast in the private sector, and in particular in sec- 2012. Instead, the rural poor have shifted into con- tors where the poor work. At the same time, while struction (23 percent) and the financial, insurance Iraq is a relatively urbanized country, poverty reduc- and professional services sector (9 percent). The tion in Iraq between 2007 and 2012 was faster in urban poor have become increasingly dependent rural areas, and trends in rural poverty also drove on construction, with 28 percent employed in this trends in headcount rates within the country. In this sector in 2012; but have also moved to transport, chapter, we take a closer look at the labor market storage and communication (15 percent) and the for the poor, and examine whether the drivers of financial, insurance and professional services sector. poverty reduction were different across rural and urban Iraq, given the differences in the main sectors The differences in the type and sector of economic of economic activity across the country. We also try activity notwithstanding, the employed poor are and understand whether welfare improvements have much more dependent on the private sector as a been accompanied by an improvement in education source of labor earnings (Figure 164). In 2007, 76 and labor market outcomes and favorable changes in percent of the urban poor and 81 percent of the The Labor Market for the Poor: The Rural-Urban Divide 149 rural poor worked in the private sector, compared on predominantly public sector jobs in the financial with 60 percent and 72 percent of the urban non- sector in both rural and urban areas. In rural areas, poor and the rural non-poor. Over time however, agricultural employment has declined sharply among the role of the public sector as a source of employ- the poor, whereas in urban areas, employment in the ment has increased, especially for the non-poor, but commerce and retail sector and the manufacturing also for the poor. In 2012, 27 percent of the em- sector has decreased among the poor. ployed urban poor and 22 percent of the employed rural poor worked in the public sector. Differences in Economic Activity are Only Thus, in terms of the labor market, the most impor- One of Many Dimensions of Differences tant headline indicators that distinguish non-poor between Urban and Rural Households households from poor households are higher par- These differences in the sectors of employment are ticipation and employment rates, and in particular, only one of the many differences in characteristics public sector employment rates. Within the poor, between urban and rural households. Rural house- urban and rural households are employed in differ- holds have on average much lower levels of educa- ent types of economic activity, with an increasing de- tion than urban households, and while educational pendence on construction, and to a smaller extent, attainment among the urban working age popula- tion has remained stagnant between 2007 and 2012, rural Iraq does not show signs of catching up and TABLE 30: Labor Market Outcomes for the continues to lag behind significantly (Figure 165). Urban and Rural Poor, 2007 to 2012 While there are minor increases in the share of work- ing age individuals in rural Iraq with incomplete and Urban poor Rural poor primary education, less than a fifth of the rural work- 2007 2012 2007 2012 force has intermediate or higher education (in con- Labor force Male 74.6 72.7 71.9 69.9 trast to 38 percent of the urban workforce). participation Female 4.4 3.6 14.8 8.6 Employment Male 68.2 64.9 65.1 65.6 to working Female 3.7 3.2 14.3 8.5 age ratio FIGURE 164: Share of Employment in the Labor relation Salaried farm 1.9 0.7 3.7 3.0 Private Sector, Urban and Rural, Salaried nonfarm 76.7 81.2 44.3 62.0 2007 to 2012 Self-employed farm 2.1 1.8 43.2 26.3 Self-employed 19.4 16.2 8.8 8.7 90 nonfarm 80 Sector of Agriculture & fishing 3.9 2.9 46.9 30.4 70 employment 60 Mining & quarrying 1.3 0.3 0.2 0.4 50 Manufacturing 14.7 9.6 3.5 6.0 40 Utilities 1.4 3.7 0.9 1.8 30 Construction 21.3 27.7 17.5 22.6 20 10 Commerce and retail 20.0 13.9 4.1 5.9 0 Transport, storage & 9.4 14.8 8.3 9.2 2007 2012 2007 2012 Financial, insurance 6.4 12.1 3.3 8.9 Urban Rural Public 14.1 8.4 12.4 9.3 Poor 76.2 73.1 81.4 77.8 administration Non poor 59.3 55.7 72.4 66.2 Other services 7.7 6.5 2.9 5.6 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. 150 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 165: Education Levels, Working Age Population, Urban and Rural Households, 2007 and 2012 40 35 30 25 20 15 10 5 0 Urban Rural Urban Rural 2007 2012 Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Source: Authors’ calculations, IHSES 2007 and 2012. Rural households continue to be significantly larger in rural and urban areas, with declines in the share than urban households. Moreover, while average of household members of working age. Finally, this household size in urban areas has come down from has been accompanied by a decline in the share of 8.2 to 7.8 persons, accompanied by a similar decline household members who are occupied or employed in the average number of adults aged 15+, average adults (Figure 167). household sizes and the number of adults in rural areas has remained unchanged (Figure 166). At the Thus, it appears that demographics have also not same time, there has been an increase in dependency been in favor of poverty reduction across Iraq, but especially in rural areas, where dependency has been FIGURE 166: Household Size and Composition Across Urban and FIGURE 167: Dependency and Employment Rural Areas, 2007 and 2012 Across Urban and Rural Areas, 2007 and 2012 12 70 9.65 9.78 59.03 57.86 10 60 52.11 50.94 8.21 7.79 50 8 40 6 5.07 5.19 5.15 4.68 30 23.92 21.52 23.39 4 20.12 20 2 10 – – Urban Rural Urban Rural Urban Rural Urban Rural 2007 2012 2007 2012 Average household size Share of working age population per household Number of adults (age>=15) per household Share of occupied adults Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. The Labor Market for the Poor: The Rural-Urban Divide 151 increasing without a commensurate increase in earn- evident in urban areas. Urban non-poor households ing adults. In any case, the rural working age popu- are significantly smaller in size, about 7 members, lation of Iraq is younger than the urban working age less than half of whom are children; and with much population: in 2012, 41 percent of the rural working more educated heads: only 31 percent of heads had age population is between 15 and 25 compared to 38 less than primary education. percent in urban areas; while 16 percent is between 46 and 65 in rural areas, compared to 19 percent in urban areas. This suggests that investments in educa- For Any Given Characteristic, Rural tion and improving labor market outcomes are even Households Face Higher Poverty Rates more important in rural Iraq, if young entrants into Than Urban Households the labor market are to be able to find productive Despite these similarities between urban and rural employment and contribute to household welfare. poor households in terms of their average character- istics, for any given characteristic, headcount pov- While the differences between the urban and rural erty rates in rural areas are much higher. poor in terms of the type and sector of work are large, in terms of household size and composition, Rural households and poor households are typically as well as the education of working age adults, a much larger in size than non-poor urban households. typical poor household in rural Iraq in 2012 appears Headcount poverty rates increase steeply among to be more similar to a poor urban household than larger households, but more so in rural areas than to a rural non-poor household (Table A 7.1). in urban areas (Figure 168). Poverty rates among households of 4 or less are around 10 percent in rural The average poor rural household has 11 members, areas, less than 3 percent in urban areas. Rural house- 7 of whom are dependents. Almost 85 percent of holds with 10 or 11 members face headcount rates of heads of households have primary education or around 40 percent, while similarly large households less, and more than a third are illiterate. More than in urban areas experience poverty rates around 10 60 percent of the heads of poor rural households percentage points lower. Between 2007 and 2012, are either not employed (32 percent) or work in headcount rates of poverty among large rural house- agriculture (21 percent) or construction (11 per- holds have significantly reduced, while they have re- cent). A typical non-poor household in rural Iraq mained stable among large urban households. is smaller, with 9 members and around 5 children, and 70 percent of heads have primary education or Poverty also declines starkly with education, especial- less (with a quarter being illiterate). In terms of the ly in rural areas. Between 2007 and 2012, headcount employment status of heads of households, this is rates have declined at almost every level of educa- somewhat similar to those of poor households in tion in both urban and rural areas, but at each level rural areas, with 44 percent being non-employed or of education, poverty is almost double in rural ar- working in agriculture; but with 12 percent holding eas (Figure 169). For instance, in 2012, headcount jobs in the public administration sector. rates among urban households with heads with in- complete primary education were around 15 per- Poor urban households on average had 10 mem- cent; a little lower than poverty rates among rural bers, 6 of who were children, and with 80 percent households with secondary education. 16 percent of of household heads having primary education or urban households with illiterate heads were poor in less, and 32 percent of heads being non-employed. 2012, as compared to 32 percent of similar house- A similar pattern of larger household sizes, higher holds in rural areas. Note that while education levels dependency, the relative importance of construction of household heads are fairly similar across poor and and transport as sectors of employment for the poor, rural households, the incidence of poverty in rural ar- and of lower levels of education for the poor is also eas is much higher, irrespective of level of education. 152 THE UNFULFILLED PROMISE OF OIL AND GROWTH  overty Headcount Rates by Household Size, Urban and Rural Households, 2007 FIGURE 168: P and 2012 0.8 0.6 0.4 0.2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 or higher Rural households, 2007 Rural households, 2012 Urban households, 2007 Urban households, 2012 Source: Authors’ calculations, IHSES 2007 and 2012. This pattern is mirrored in the relationship between 24 percent, which is significantly lower than pov- poverty rates and the household head’s employment erty among rural households with employed heads, status. While poverty has declined among house- which is almost 30 percent. holds with employed heads in both rural and urban areas, there has been an increase in poverty among Thus, not only do the rural poor have different rural households with unemployed heads. However, types of opportunities for employment, these are irrespective of the employment status of the head, also accompanied by large human capital gaps and headcount rates are higher in rural areas (Figure significant differences in household size and compo- 170). In fact, households with unemployed heads sition compared to the urban poor, each of which is face the highest rates of poverty in urban areas, also correlated with lower welfare. FIGURE 169: Poverty Headcount Rates by Education, Urban and Rural Households, 2007 and 2012 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0 2007 2012 2007 2012 Urban Rural Illiterate Incomplete primary Complete primary Intermediate Secondary Higher secondary Tertiary Source: Authors’ calculations, IHSES 2007 and 2012. The Labor Market for the Poor: The Rural-Urban Divide 153 FIGURE 170: Poverty Headcount Rates by rates increased, however, the negative association Employment Status, Urban and between consumption and rural areas has become Rural Households, 2007 and 2012 even stronger between 2007 and 2012. 0.5 Table A 7.3 shows the results of multivariate analysis 0.4 that estimates estimate the marginal effects of these characteristics in predicting poverty and similar pat- 0.3 terns are evident there as well. Across both survey 0.2 years and for urban and rural households, larger household sizes and more children increase the like- 0.1 lihood of poverty. But the marginal effect of having an additional household member on whether the 0 2007 2012 2007 2012 household is poor or not has substantially declined Urban Rural in rural areas, and is now equal in magnitude to that Full time employed Unemployed in urban areas. Education of the head of household Part time employed Out of the labor force significantly lowers the odds of poverty. For instance, in 2012, tertiary education reduced the likelihood Source: Authors’ calculations, IHSES 2007 and 2012. that an otherwise similar household was poor by 19 percentage points in urban areas and by 31 percent- age points in rural areas. It is worth noting that the Next, we examine how these household charac- overall relationship between poverty and education teristics correlate with consumption across urban is stronger in rural areas, where education levels are and rural households (Table A 7.2). Among urban the lowest and has been increasing over time. households, the relationship between household size, composition and education with consumption Employment for the head of household in the pub- are very similar to the national average, and similar lic administration sector significantly reduces the to those for rural households (although education odds of being poor, by 3 to 4 percentage points has stronger effects in urban areas). In both years, in urban areas. In rural areas, public administra- employment in manufacturing, commerce and re- tion jobs did not significantly lower poverty risks tail, finance, insurance and professional services, in 2007, but in 2012, they lower the likelihood and in public administration, health and education that a household is poor by almost 12 percent. are associated with higher per capita consumption In 2012, employment in mining and quarrying relative to non-employment, whereas construction and commerce and retail also lowered the risk of becomes negatively associated with consumption in poverty for urban households while construction 2012. In rural areas, employment in construction increased the likelihood. In rural areas, in 2012, is negatively correlated with per capita consump- all sectors of employment except manufacturing, tion expenditures in both years relative to house- utilities, commerce, finance and public administra- holds whose heads are unemployed or out of the tion had no effect on the likelihood of being poor. labor force. The presence of elderly household Overall, public sector employment—public admin- members in general increases consumption, except istration and mining and quarrying in urban areas; among rural households in 2012, perhaps suggest- and public administration, finance and utilities in ing that pension receipts no longer compensate for rural areas—lower the odds of poverty; however, the increased dependency rates. Comparing 2007 agriculture and construction, which offer primarily and 2012, living in a rural area is by and large no private sector jobs, have a weak relationship with longer associated with lower consumption in the poverty or actually increase the likelihood that the governorates where poverty fell. Where poverty household is poor. 154 THE UNFULFILLED PROMISE OF OIL AND GROWTH In urban areas, the limited improvement in welfare FIGURE 171: Share of the Urban Poor, by may be related to the decline in employment rates Sector of Employment for men, which may have been counteracted by the prevalence of public sector jobs in mining, utilities Other services and public administration. Moreover, as we show Public administration below, public sector employment and the expansion of public transfers, especially pensions, is associ- Financial, insurance ated with lower poverty in urban areas. What lim- Transport, storage & ited welfare improvements that have occurred have Commerce and retail happened without any perceptible improvement in Construction private sector employment and earnings. Among Utilities the rural population, despite higher poverty rates at each education level relative to urban areas, poverty Manufacturing reduction has taken place despite any perceptible Mining & quarrying change in education or in labor market outcomes. Agriculture & shing Indeed, rural households with employed heads face higher poverty rates than urban households with un- 0 5 10 15 20 25 30 employed heads, in part because two major sources 2007 2012 of male employment—agriculture and construction are not associated with lower odds of poverty. Source: Authors’ calculations, IHSES 2007 and 2012. Public Sector Employment, Public households with heads employed in the public sec- Transfers and Urban Poverty tor are significantly lower than among other urban households, and have reduced from 13.6 percent to Between 2007 and 2012, the employed urban poor 11 percent between 2007 and 2012. While house- became increasingly concentrated in three sectors, holds with heads employed in the private sector and in addition to commerce and retail—construction; non-employed heads both had relatively high rates transport, storage and communication; and finan- of poverty in 2007, the decline in poverty has been cial, insurance and professional services—which to- sharper among households with non-employed gether accounted for approximately 58 percent of heads—from 19 percent in 2007 to 16 percent in the urban employed poor. With the exception of 2012. The bulk of the urban poor, almost four- financial, insurance, and professional services, which fifths belong to households whose heads are not have become dominated by the public sector, most employed in the public sector. Overall, though, be- of the urban poor work in the private sector. tween 2007 and 2012, the distribution of the urban poor across these categories has changed little. To further understand urban poverty, we there- fore consider three types of households: those with Households with heads employed in the public sec- heads employed in the public sector, those with tor earn the highest per capita income, driven by heads employed in the private sector; and those with the highest per capita labor income, compared to non-employed heads (including heads of household other types of urban households. Between 2007 who are not of working age). Figure 172 shows the and 2012, these households have also experienced poverty headcount rates for each of these types of the largest average increases in per capita labor in- households, the trends between 2007 and 2012, come. Non-labor incomes on the other hand, have and Figure 173 graphs their contribution to the declined over time, primarily because the increase total urban poor population. Poverty rates among in pension incomes, domestic remittances and other The Labor Market for the Poor: The Rural-Urban Divide 155 transfers has not compensated for the decline in im- because of relatively low per capita labor earnings plicit incomes from ration receipts. In line with the (earned by other household members), and both findings in the previous chapter, welfare improve- have grown slowly (Table 32). However, these types ments among these types of households are prob- of households receive the highest levels of non- ably related to the large and increasing earnings and labor income, more than twice the levels received benefits associated with public sector employment. by households with public sector employed heads. Moreover, these have increased by 16 percent over On the other hand, households with non-employed the 2007 to 2012 period; compared with a decline heads have the lowest levels of per capita income in non-labor incomes among households with heads working in the public sector. While ration incomes have declined over time, these have been more than FIGURE 172: Headcount Rates, by Types of compensated by a 45 percent increase in pension Urban Households incomes, which now makes up the single largest source of non-labor incomes and a doubling of 20 incomes received as domestic remittances. These 18 types of households therefore, receive much larger 16 public and private transfers compared to other ur- 14 ban households, and have likely allowed household 12 10 heads to remain non-employed; and the increases 8 in these transfers over time has probably led to the 6 observed welfare improvements. 4 2 0 Public Private Non employed TABLE 31: Labor and Non-Labor Income, 2007 2012 Households with Heads Employed in the Public Sector Source: Authors’ calculations, IHSES 2007 and 2012. Per capita Income by Percentage components 2007 2012 change Total 158.46 196.85 24.23 FIGURE 173: Share of the Poor, by Types of Labor 112.10 145.18 29.50 Urban Households Non Labor 21.16 19.58 –7.46 Imputed Rent 25.19 32.10 27.40 Non Labor Capital 1.73 1.36 –21.35 Income Pensions 2.09 2.75 31.58 2012 components Remittances Intl’ 0.73 0.23 –68.86 Domestic 2.58 3.15 22.12 Social protection 0.05 0.24 365.29 Other Public 2.03 3.39 66.69 2007 Transfers Private 0.35 0.08 –76.56 Rations 11.58 8.35 –27.88 0% 20% 40% 60% 80% 100% Zakat 0.01 0.03 341.84 Public Private Non employed Public 15.76 14.73 –6.50 Private 5.40 4.85 –10.24 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. 156 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 32: Labor and Non-Labor Income, On average, urban households with heads working Households with Non-Employed in the private sector earn lower per capita incomes Heads and per capita labor incomes relative to those with public sector employed heads and higher per cap- Per capita Income by Percentage ita incomes and labor incomes compared to those components 2007 2012 change with non-employed heads. On average, both have Total 141.73 168.62 18.97 increased by slightly more than 20 percent be- Labor 77.23 84.36 9.23 tween 2007 and 2012. However, there has been Non Labor 38.00 44.02 15.86 only a negligible increase in non-labor incomes, as Imputed Rent 26.50 40.23 51.84 the increase in pensions and domestic remittances Non Labor Capital 3.76 3.35 –10.98 has barely compensated for the decline in ration Income components Pensions 13.50 19.61 45.28 incomes. Remittances Intl’ 1.57 0.54 –65.82 Domestic 4.00 8.01 100.21 Breaking down this aggregate picture by different Social protection 0.52 1.49 189.47 employment sectors within the urban private sec- Other Public 2.19 1.61 –26.67 tor, Figure 174 shows that headcount rates have Transfers significantly increased in the sectors where more of Private 0.49 0.36 –26.68 Rations 11.90 8.96 –24.70 the urban poor are now concentrated—in construc- Zakat 0.06 0.09 55.26 tion and financial, insurance and professional ser- Public 28.11 31.68 12.68 vices. Note that poverty rates have almost doubled among urban households with heads employed in Private 9.88 12.34 24.90 private sector jobs in financial, insurance and profes- Source: Authors’ calculations, IHSES 2007 and 2012. sional services. In contrast, poverty has fallen among households with heads working in commerce and TABLE 33: Labor and Non-Labor Income, Households with Heads Employed in the Private Sector FIGURE 174: Headcount Rates by Employment of the Head (Urban Per capita Income by Percentage components 2007 2012 change Private Sector), 2007–2012 Total 150.02 181.70 21.12 35 Labor 104.66 128.46 22.74 30 Non Labor 22.53 23.63 4.85 25 Imputed Rent 22.83 29.62 29.74 20 Non Labor Capital 1.88 2.29 22.04 Income 15 Pensions 3.75 5.55 47.93 components Remittances Intl’ 0.61 0.29 –51.61 10 Domestic 2.38 4.31 80.76 5 Social protection 0.26 0.83 221.97 0 Agriculture & nishing Manufacturing Construction Commerce and retail Transport, storage & Financial, insurance Others services Other Public 1.77 1.62 –8.74 Transfers Private 0.25 0.10 –60.42 Rations 11.62 8.58 –26.15 Zakat 0.01 0.05 366.30 Public 17.41 16.58 –4.73 2007 2012 Private 5.13 7.04 37.37 Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. The Labor Market for the Poor: The Rural-Urban Divide 157 FIGURE 175: Decomposing Differences in Headcount Rates between Private and Public Sector Employed Heads of Household as reference Private 2007 76 25 2012 53 30 as reference Public 2007 75 24 2012 70 47 –30 20 70 120 –30 20 70 120 Characteristics Returns Interaction Source: Authors’ calculations, IHSES 2007 and 2012. retail and manufacturing, both of which now account and the part explained by differences in coefficients for a smaller share of the poor than they did in 2007. (which in this case measure the strength and nature of the relationship between the characteristic and poverty). In order to do that, we estimate the prob- Explaining the Differences in Welfare ability of being poor or not on a set of characteristics between Public and Private Sector of the household and household head among others The vast majority of poor households in Iraq, 70 for both types of households in urban Iraq.46 percent, have employed heads of household. With- in these types of urban households, poverty is low- The results of this exercise show that in both years, er among those with heads employed in the public differences in characteristics between these two types sector; and these rates have come down faster, by of urban households explain the bulk of differences 2.5 percentage points between 2007 and 2012 (as in headcount rates (Figure 175). In other words, opposed to 1.5 percentage points among house- households with heads employed in the public sec- holds with heads in the private sector). Were these tor had on average better characteristics, which were changes driven by differences in endowments or associated with lower poverty. In 2012, in addition, due to other factors? In order to understand the the returns associated with having similar character- factors underlying poverty headcount differen- istics appear to have become more important, and tials, we utilize the Oaxaca-Blinder decomposition may explain the faster welfare improvements among method. This approach is typically used in the la- households with heads in the public sector. This is bor market literature to disentangle the share of in line with the increasing returns on the labor mar- the wage gap attributable to characteristics from ket from public sector employment observed earlier. that attributable to coefficients (which may be due Next, we turn to the labor market facing the poor in to discrimination). rural Iraq and begin with a brief description of the nature of the agricultural sector, on which so many In this case, we are applying this method to explain of the rural poor still depend. the poverty headcount difference between house- holds with heads employed in the private and public sector in 2012. The idea is to quantify the part of that 46 We use the nldecompose command to perform the Oaxaca- difference explained by differences in characteristics Blinder decomposition for non-linear model. 158 THE UNFULFILLED PROMISE OF OIL AND GROWTH Rural Employment and Agriculture MAP 2: Percentage of Land Used for in Iraq, 2007 to 2012 Agriculture Per Square Kilometer Iraq’s historical endowment of agriculture has been steadily and it appears, irrevocably eroding. In 1971, agriculture accounted for 16 percent of value added, around a fifth of GDP and 55 percent of total em- ployment. At the time, it was recognized as a lagging sector, with little investment towards increasing pro- ductivity and building the necessary infrastructure. A 1974 World Bank report notes that “agricultural yields, reflected centuries of abuse and neglect of the land, much of which had been allowed to deteriorate to the point of being uncultivable.”47 $JULFXOWXUDOVKDUHRIDUHD During Saddam Hussein’s early years, the state at-   tempted to promote private sector investment in ʹͲͳͶǦͳͲ agriculture through the distribution of high yield- Source: Fritz et al. 2011. ing variety seeds, higher output prices, expanded subsidies to agriculture and heavy investments in irrigation. While area and production expanded and physical infrastructure, especially the irrigation throughout the 1980s, cereal yields continued to network, were degraded as a result of inadequate stagnate.48 The Iran-Iraq war diverted labor and maintenance and funding. In the South, the adverse investment away from the sector, and caused sig- impact on livelihoods was compounded by the mas- nificant damage to infrastructure in the southern sive drainage of the Mesopotamian marshlands in governorates on the border with Iran. Following the 90s.49 the invasion of Kuwait by Saddam Hussein, wide- ranging sanctions were imposed, and the inability The spatial variation in agricultural activity across to export oil severely limited access to imports of Iraq is quantified and visualized below. In order to food and agricultural inputs. While the introduc- measure agricultural land use in Iraq, we use the tion of the Public Distribution System thereafter Global Hybrid dataset (0611–2012 V2) produced guaranteed some degree of food security to the by Fritz et al. (2011) which estimates the percent- Iraqi population, it introduced significant disin- age share of land used for agriculture within a one centives for cereal production, depressing producer square kilometer pixel (Map 2). By multiplying this prices and private investment. percentage by the total pixel area we derive a mea- sure of total agricultural land use area. By aggregat- The establishment of the autonomous Kurdistan re- ing the agricultural area of each pixel we calculate gion in the early 90s comprising a large part of the agricultural land use at a district level. This allows northern rain fed agricultural zone was followed by us to determine the total share of agricultural land decades of relative peace and stability in the three Kurdish governorates. In the rest of Iraq, agricul- tural activity runs along and between the Tigris and 47 World Bank (1974). Current Economic Position and Pros- Euphrates rivers, and is dependent on irrigation. pects of Iraq. Report No. 419a-lRQ. Salinization has historically been a major challenge, 48 http://digital.library.unt.edu/ark:/67531/metacrs7073/ given the low and saline water table in this zone, m1/1/high_res_d/RS21516_2003May13.pdf. and it became widespread as agricultural services 49 Joint World Bank FAO Agriculture Sector Note, 2011. The Labor Market for the Poor: The Rural-Urban Divide 159 MAP 3: Share of Agricultural Land Use of at the district level: the bi-monthly grid mean, bi- Total District Land Area monthly grid standard deviation, and bi-monthly grid maximum. Each of these bi-monthly statistics is then aggregated in time to produce analysis at the annual level. Map 4 displays the district level variation in 2003, 2005, 2007 and 2009 that shows high greenness in the far North as well as the dis- tricts near Baghdad and along the Tigris and Eu- phrates rivers. Because of climate conditions, the Northern districts of Iraq have on average higher levels of greenness than the Southern districts. Vari- ation in annual precipitation levels over the period of our analysis also affects the measure of green- $JULFXOWXUDOVKDUHRIGLVWULFW  ness. For example, the 2007 and 2008 drought in   the Northern districts of Iraq corresponds to lower measures of greenness.  ! ʹͲͳͶǦͳͲ Source: Staff calculations based on Fritz et al. 2011. The post-2003 conflict in Iraq led to a further diver- sion of resources and widespread destruction of in- frastructure. Violence was predominantly focused in use within the total district land area, shown in Map Baghdad and the North, and in the governorates of 3, which clearly shows the concentration of agricul- Anbar and Diyala in the Center; while the rest of the tural activity in Kurdistan and around and between Centre, Kurdistan and the South remained relatively the Tigris and Euphrates rivers in Baghdad, Diyala, peaceful. The improvements in the security situa- Babylon, Wasit, Kerbala, Qadisiya and Thi Qar. tion in the countryside have been accompanied by a revival of the rural economy in some parts of the But agricultural activity can also vary over time, es- country, as we will show below; but in the south- pecially as a response to weather shocks and con- ern governorates, poverty among households de- flict. Below we use a measure of greenness or the pendent on agriculture has risen sharply, and while intensity or density of vegetation within a district people are leaving agriculture, they have nowhere to over time, which identifies areas vegetation, includ- go as the local economy continues to stagnate. ing agricultural land and forest cover. The most common measure of greenness is the Normalized Differentiated Vegetation Index (NDVI), which is Agricultural Jobs: Evolution Across Time derived from remote sensing data. The NDVI cal- and Space culates greenness values between of –1, (indicating In 2007, agriculture was the main sector of em- complete absence of vegetation,) and 1, (indicat- ployment for 27 percent of employed Iraqi women ing the greatest intensity of vegetation).50 In this and 10 percent of employed Iraqi men. By 2012, analysis, we use NDVI data constructed by the U.S. National Aeronautics and Space Administration (NASA) Global Inventory Modeling and Mapping 50 Using the NDVI from multiple Landsat satellite images to Studies (GIMMS) at a bi-monthly frequency be- estimate the total cultivated area for a portion of Iraq, Gib- son et al. (2012) present the decline of cultivated area from tween 2003 and 2009 that measure greenness over the Late Sanctions period derived from NDVI calculated 8 square kilometers pixels available for the entire ar- from images between 2000 to 2003 compared to the Op- ea of Iraq (Zhu et al. 2013). From each bi-monthly eration Iraqi Freedom (OIF) period derived from NDVI pixel data, we derive three statistical measurements calculated from images between 2008 to 2011. displays the district level variation in 2003, 2005, 2007 and 2009 that shows high greenness in the far North as well as the districts near Baghdad and along the Tigris and Euphrates rivers. Because of 160 climate conditions, the Northern districts of Iraq have on average higher levels of greenness than the THE UNFULFILLED PROMISE OF OIL AND GROWTH Southern districts. Variation in annual precipitation levels over the period of our analysis also affects the measure of greenness. For example, the 2007 and 2008 drought in the Northern districts of Iraq corresponds to lower measures of greenness. MAP 4: NDVI Changes in Greenness Over Time MAP 4: NDVI CHANGES IN GREENNESS OVER TIME : Authors’Authors’ Source: Source calculations. calculations agriculture’s share in female and male jobs had 212agriculture combined with growing non-agricultur- fallen to 23 percent for women and 7 percent for al employment. It is due to an absolute decline in men. In line with these, there has been a decline 51 agricultural employment for both men and women in the share of the population belonging to house- holds with heads employed in agriculture from 10 51 In 2007, agriculture employed 8 percent of the male work- percent in 2007 to 7 percent in 2012. The declin- ing age population and 3 percent of the female working age ing importance of agriculture as a sector of em- population. By 2012, these estimates had fallen to 6 percent ployment is not the result of static employment in and 2 percent respectively. The Labor Market for the Poor: The Rural-Urban Divide 161 FIGURE 176: Jobs in Agriculture for Men and Women, 2007 and 2012 700,000 587,900 600,000 500,000 437,240 400,000 300,000 264,621 230,045 200,000 100,000 0 –1,000,000 –34,576 –150,660 –2,000,000 2007 2012 2007 2012 Female Male Female Male Change (2012 relative to 2007) Source: Authors’ calculations, IHSES 2007 and 2012. FIGURE 177: Changes in Number of Jobs in Agriculture and Total Jobs, Men and Women, by Division 300,000 250,000 200,000 150,000 100,000 50,000 0 –50,000 –100,000 Kurdistan Baghdad North Centre South Male, agricultural jobs Male, total jobs Female, agricultural Jobs Female, total Jobs Source: Authors’ calculations, IHSES 2007 and 2012. between 2007 and 2012, with 34,500 fewer jobs These changes are reflected in the spatial distribu- for women and 150,600 fewer jobs for men of tion of agricultural jobs across Iraq (Figure 178). working age (Figure 176). In 2007, Kurdistan, Baghdad and the North to- gether accounted for 35 percent of all agricultural This has happened for the most part, due to a de- jobs for women. By 2012, each witnessed an in- cline in the number of people working in agriculture: crease in their share and now account for almost men in the North, Centre and South, and women half the jobs for women in agriculture. In contrast, in the Centre and South. While in Kurdistan, the the importance of the Centre and the South in ag- North and the Centre, it has been accompanied by ricultural jobs for women has declined significantly, a significant increase in the number of jobs for men where it was a much larger employer in 2007 (ac- between 2007 and 2012, in the South, there was counting for more than 40 percent and 23 percent little additional job creation to compensate for the respectively), especially in the South, where its decline in agricultural employment (Figure 177). share has halved. 162 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 178: Share of Agricultural Jobs in Each FIGURE 179: Share of Agriculture as a Source Division, 2007 and 2012 of Employment for Men and Women Within Each Division, 0.12 0.21 0.17 2007 and 2012 0.23 0.37 0.40 0.33 0.41 0.38 0.35 0.18 0.30 0.23 0.13 0.23 0.25 0.13 0.07 0.10 0.20 0.07 0.15 0.20 0.16 0.11 0.15 2007 2012 2007 2012 0.10 Women Men 0.05 Kurdistan Baghdad North Centre South 0 Kurdistan Baghdad North Central South Source: Authors’ calculations, IHSES 2007 and 2012. Women, 2007 Men, 2007 Women, 2012 Men, 2012 Source: Authors’ calculations, IHSES 2007 and 2012. A similar trend is true somewhat for men, with an increase in Kurdistan and Baghdad’s share in agri- cultural jobs for men, although from relatively low levels; no change in the North (around 23 percent), for non-farm diversification; we define four types of and declines in the Centre (by 5 percentage points) (mutually exclusive) households: and in the South (by 4 percentage points). But still, the Centre and the South accounted for half of all 1. Non-agricultural household: A household where agricultural jobs for men and women in 2012. no employed member works in agriculture 2. Agricultural household: A household where all Across the country, with the exception of Baghdad, employed members work in agriculture agriculture is an important source of jobs for the few 3. Diversified household: A household where at women who work: in 2007, it accounted for a quar- least one employed member works in agricul- ter of employed women in Kurdistan, more than 35 ture and at least one works outside agriculture percent in the North and the Centre, and more than 4. Non-employed household: A household where 30 percent in the South. While its share has some- no member is employed what declined, with a sharp decrease in the South, where overall female employment fell, it is still one Non-agricultural households account for a large of the most important sectors of work for women. majority of the population of Iraq: 83 percent of For men, on the other hand, only 15 percent of em- the population in 2012, and 92 percent of the ur- ployed men in the North and the Centre work in ag- ban population in 2012 belong to non-agricultural riculture, and less than 10 percent in other divisions. households (Table 34). In rural areas, there was a The decline in agriculture’s role in male employment large increase in the share of these households, from by 2012 is evident across all divisions (Figure 179). 47 percent in 2007 to 63 percent in 2012; and a substantial decline in the share of households at- tached to agriculture. For instance, the share of the Rural Poverty and Non-Farm Diversification population in agricultural households fell by 10.5 To better understand the rural economy, the chang- percentage points; and those in diversified house- ing role of agriculture and examine the opportunities holds fell by 5 percentage points. In 2012, less than The Labor Market for the Poor: The Rural-Urban Divide 163 TABLE 34: Share of Different Types of Households in Urban and Rural Areas, 2007 and 2012 Total Rural Urban Change (percentage point) Share of population belonging to 2007 2012 2007 2012 2007 2012 Total Rural Urban Non-agricultural household 79.03 82.90 47.08 63.38 91.96 91.91 3.87 16.30 –0.05 Agricultural household 7.92 5.27 24.60 14.10 1.16 1.20 –2.65 –10.50 0.04 Diversified household 7.34 6.02 20.66 15.52 1.95 1.63 –1.32 –5.14 –0.32 Non-employed household 5.71 5.81 7.65 7.01 4.92 5.26 0.10 –0.64 0.34 Source: Authors’ calculations, IHSES 2007 and 2012. 30 percent of the rural population belonged to a the shares of both these types of households fell, household where at least on member was employed although the larger decline occurred among agri- in agriculture, compared to 45 percent in 2007. In cultural households in the Centre and among di- addition, almost 6 percent of the population and 7 versified households in the South. In contrast, in percent of the rural population in 2012 belonged to Baghdad and Kurdistan, while the share of non- households where no member was employed. agricultural households increased somewhat, the share of agricultural households fell, but the share The bulk of this declining dependence on agri- of diversified households increased. culture occurred in the North, Centre and the South; which all recorded sharp declines in agri- In rural areas, non-agricultural households had the cultural households accompanied by no change in lowest poverty rates in 2007 and in 2012, and their or decreases in the share of diversified households headcount rates fell by 4 percentage points in the (Figure 5). In the North, the share of the popu- intervening period (Table 35). However, because of lation in households where all employed members a large increase in their share, they now make up were working in agriculture fell by 11 percentage 60 percent of the rural poor. Neither agricultural points while there was almost no change in diver- households nor non-employed households expe- sified households. In the South and the Centre, rienced substantial welfare improvements over the FIGURE 180: Share of Different Types of Households in Each Division, 2007 and 2012 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007 2012 2007 2012 2007 2012 2007 2012 2007 2012 Kurdistan Baghdad North Centre South Non-agricultural household Diversi ed household Agricultural household Non-employed household Source: Authors’ calculations, IHSES 2007 and 2012. 164 THE UNFULFILLED PROMISE OF OIL AND GROWTH  overty Rates of Different Types of TABLE 35: P was driven by welfare improvements experienced by Households in Rural Areas, 2007 diversified households. and 2012 These overall trends mask significant improvement Headcount Share and worsening in different rural parts of the country poverty rates of the poor Rural (Figure 181): households 2007 2012 2007 2012 Non-agricultural 33% 29% 40.26 59.75 • decline in poverty among all rural households household in Kurdistan, and in particular among non-em- Agricultural 41% 40% 25.71 18.22 ployed households; household • very large and significant declines in poverty Diversified 50% 27% 26.33 13.45 household rates (between 47 and 61 percent) among all rural households in the Central division; Non-employed 39% 37% 7.69 8.57 household • sharp decreases in headcount rates for diversi- Source: Authors’ calculations, IHSES 2007 and 2012. fied households and substantial increases in poverty for non-employed households in the rural North; five year period, and continue to have high rates of • increases in poverty in rural Baghdad; and poverty in 2012, 40 percent and 37 percent respec- • large and significant increases in poverty head- tively. The share of agricultural households among count rates among rural agricultural and non- the rural poor has come down from 26 percent in employed households in the South. 2007 to 18 percent in 2012 because their share in the population has shrunk. The largest welfare im- Thus, welfare improvements experienced by house- provements occurred among diversified rural house- holds who were diversified were limited to Kurdis- holds, whose headcount rates almost halved, from tan, the North and the Centre. In Baghdad and the 50 percent to 27 percent; as did their share in the South, poverty increased for almost all types of rural rural poor, from 26 percent to 13 percent. Thus, it households, but especially among rural households appears that the observed reduction in rural poverty in the South who were completely dependent on FIGURE 181: Trends in Headcount Rates of Different Types of Households, by Division, 2007 and 2012 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Kurdistan Baghdad North Centre South Kurdistan Baghdad North Centre South 2007 2012 Non-agricultural household Diversi ed household Agricultural household Non-employed household Source: Authors’ calculations, IHSES 2007 and 2012. The Labor Market for the Poor: The Rural-Urban Divide 165 agriculture for employment: from an already high 47 all rural households; while in the South the opposite percent in 2007 to a whopping 75 percent in 2012. was true, and especially for agricultural households. Next, we try to understand the factors behind some These divisional trends in rural poverty are even of these changes to the extent possible. more apparent when disaggregated to the gover- norate level (Table A.4). In the four southern gov- ernorates of Qadisiya, Missan, Thi Qar and Muthan- Human Capital: Education na, more than 70 percent of agricultural households Perhaps the patterns and trends in poverty are sim- were poor in 2012. These represent a significant ply representing differences in the human capital increase in poverty relative to 2007 among these endowments of these households. In other words, households, more than 15 percentage points in perhaps households dependent on agriculture are Muthanna and Qadisiya, and more than 30 percent- poorer because they are less educated; and the age points in Missan and Thi Qar. Even among non- improvement in welfare among diversified house- agricultural and diversified households, headcount holds represents a shift in the composition of these rates were above 50 percent in 2012 in each of these households towards higher education. Among rural governorates, as a result of increases in poverty in households, it does appear to be the case that agri- Qadisiya, Thi Qar and Missan, and despite decreas- cultural households have on average lower educa- es in headcount rates among these households in tion levels, with almost 90 percent having primary Muthanna. Within the Central divisions, large de- education or less (Figure 182). On the other hand, creases in poverty rates among agricultural and di- diversified households have relatively higher levels versified households were experienced in all gover- of education than agriculture households but lower norates, with the exception of Najaf, where poverty than non-agriculture households. However, there fell modestly from relatively low levels. does not appear to be a significant improvement in education levels among diversified households be- It is also not always the case that poverty is higher tween 2007 and 2012. among agricultural households compared to those who diversify out of agriculture. In Najaf and Er- Just as the overall picture on poverty and the role bil, for instance, headcount rates among diversified of agriculture in the rural economy hides significant households are almost twice those of households spatial variation, so does education, with relatively solely in agriculture. In contrast, in Basra, half of the better and improving education in Kurdistan, and households completely dependent on agriculture with the lowest education levels in the South. In are poor, relative to a fifth of diversified households. Kurdistan, there has been a significant improvement It is also striking that almost without exception, in the education levels of all types of rural house- overall rural poverty declined in governorates where holds between 2007 and 2012, with shifts from the welfare of diversified households and non-agri- primary education or less to intermediate education cultural households improved; while rural poverty and higher. Among agricultural households in ru- increased when that was not the case (Table A 7.5). ral Kurdistan, the share of individuals with primary education or less was 82 percent in 2012, a decline Overall, within the 2007 to 2012 period, there ap- of 8 percentage points since 2007, while in the other pears to have been a shift in rural households away divisions, there has been little change. Rural agricul- from agriculture. At the same time, poverty rates fell tural households in the South have the lowest levels sharply (almost halved) among diversified house- of education, with 50 percent of individuals illiter- holds, while there was little change in the welfare ate, and another 30 percent with incomplete primary of other types of households. Looking across Iraq, schooling in 2012. Thus, in the South, the increases these patterns and trends mask significant variation: in poverty among all types of rural households ap- in the Centre for example, poverty declined among pear to be unrelated to changes in education. 166 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 182: Educational Attainment of Different Types of Rural Households, 2007 and 2012 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007 2012 2007 2012 2007 2012 2007 2012 Non-agricultural household Agricultural household Diversi ed household Non-employed household Tertiary Higher secondary Secondary Intermediate Complete primary Incomplete primary Illiterate Source: Authors’ calculations, IHSES 2007 and 2012. Sector of employment poor and non-poor diversified households appears Next, we turn to the sectors of employment among primarily to be in a greater dependence on manu- the rural individuals, and in particular, among diver- facturing and construction among the poor, com- sified households, to understand whether non-farm pared to public administration, financial, insurance diversification was concentrated in certain sectors. and other services, and commerce and retail among Among all rural employed individuals, the share the non-poor. This is in line with the strengthened working in agriculture fell from 44 percent to around association between public sector employment and a quarter between 2007 and 2012 (Table 36). This lower poverty in rural areas in 2012 in the probit was compensated by an increase in employment in regressions discussed earlier in the chapter. manufacturing, construction, financial, insurance and professional services and other services. In gen- In order to identify the role of different potentially eral, individuals from poor and non-poor households correlated characteristics in predicting diversifica- followed the same pattern. However, construction tion or dependence on agriculture, we model first, absorbed relatively more workers among the poor the decision of individuals of choosing among sec- while financial, insurance and professional services tors of employment; and second, the household’s absorbed more among non-poor households. occupation type. Both models are conditioned on a range of individual and household characteristics The share of agriculture in employment among and are estimated for 2007 and 2012. The sample is individuals belonging to rural diversified house- restricted to rural areas in all governorates excluding holds remained relatively steady at a little above Baghdad (given its small rural sample). 50 percent. Individuals belonging to poor diversi- fied households were similar to other rural poor in The results in the Annex (Table 7.6a and 7.6b and that the relative importance of manufacturing and 7.7a and 7.7b) are reported as relative risk ratios, financial services in employment went up. In con- i.e., for a unit change in the characteristic or predic- trast, they were less likely to work in construction tor variable (such as age), by how much the relative which is significantly correlated with higher pov- risk of being in a certain category (for example, be- erty. In fact, the share of construction in employ- ing a diversified household), relative to the reference ment among these diversified households declined. group (agricultural household) is expected to change The difference in sectors of employment between given all other characteristics are held constant. The Labor Market for the Poor: The Rural-Urban Divide 167 TABLE 36: Sectors of Employment Individual in Rural Areas, 2007 and 2012 Share of poor rural employed Share of non-poor rural Share of poor rural individuals (diversified employed individuals (diversified employed individuals households) households) 2007 2012 Difference 2007 2012 Difference 2007 2012 Difference Agriculture & fishing 50.14 32.46 –17.68 55.79 53.02 –2.77 55.22 52.23 –2.99 Mining & quarrying 0.20 0.42 0.22 0.26 0.02 –0.24 0.26 0.42 0.16 Manufacturing 3.35 6.06 2.71 3.51 7.98 4.47 1.94 4.69 2.75 Utilities 0.85 1.74 0.89 0.41 0.93 0.52 0.87 1.09 0.22 Construction 16.51 21.89 5.38 15.88 14.09 –1.79 11.91 9.83 –2.08 Commerce and retail 4.02 5.95 1.93 2.84 2.46 –0.38 4.87 4.99 0.12 Transport, storage & 7.64 8.74 1.10 6.78 5.45 –1.33 5.31 6.25 0.94 Financial, insurance 3.12 8.45 5.33 1.89 7.39 5.50 3.40 9.38 5.98 Public administration 11.47 8.86 –2.61 10.85 6.04 –4.81 12.48 8.54 –3.94 Other services 1.53 5.43 3.90 1.79 2.62 0.83 3.72 2.57 –1.15 Source: Authors’ calculations, IHSES 2007 and 2012. Thus, a relative risk ratio of 1 implies that a unit in- engaged in agriculture relative to other sectors. Be- crease in the characteristic increases the likelihood of ing male vastly increases the odds of employment in being in the category by the same amount as of be- each sector relative to agriculture, especially in con- ing in the reference category. Similarly, a relative risk struction, transport, storage and communication, ratio greater than 1 implies that for example, a unit and other services.53 Education significantly raises increase in education increases the probability of be- the odds of employment in every sector relative to ing a diversified household relative to an agricultural agriculture; but especially in manufacturing, finance household, and vice versa for a ratio less than 1. and public administration. Having higher per capita cultivated area in Kurdistan especially increases the The first model (Tables A 7.6a and A 7.6b) predicts odds of being employed in agriculture relative to all the decision of the employed individual among dif- other sectors. ferent economic sectors conditioned on individual characteristics including age, education, gender The second model (Tables A 7.7a and A 7.7b) pre- and on household size and demographics, as well dicts whether a household is non-agricultural, diver- as household characteristics such as per capita land sified, agricultural or non-employed based on sev- owned, cultivated and per capita public and private eral household head and household characteristics. transfers.52 The reference or base category is agricul- The reference or base category is households where tural employment. all members are employed in agriculture. The results suggest that agriculture is more likely to There appears to have been a shift in the effect of be the occupation for young females in rural areas, the demographic composition of different types with low educational attainment, and who belong to the households with large dependency ratios. Indi- 52 Mining and quarrying and utilities are excluded as they are viduals belonging to a household with a larger num- very small ber of children have higher odds of being engaged 53 The large increase in the male coefficient on finance, insur- in agriculture relative to any other sector in 2007 ance and professional services in 2012 is in line with the and in 2012; while age reduces the risk of being significant increase in male employment in this sector. 168 THE UNFULFILLED PROMISE OF OIL AND GROWTH between 2007 and 2012. In 2007, an increase in FIGURE 183: Decomposing Differences in household size raised the odds of a rural household Headcount Rates between being non-agricultural or diversified while an in- Agriculture and Diversified crease in the number of dependent members low- Households – 2012 ered the relative odds; by 2012, this relationship had been weakened. An increase in the educational as reference Agriculture attainment of the household head significantly in- 2012 31 57 creases the relative likelihood that a household would be non-agricultural rather than agricultural. While in 2007, having higher secondary and tertiary as reference Diversi ed education for the head of household were the only 2012 43 69 education levels that distinctly increased the odds of diversification, in 2012, having primary and in- complete primary education also have the same ef- –25 –5 15 35 55 75 95 115 fect. As with the individual level regressions, larger Characteristics Returns Interaction cultivated land per capita in Kurdistan appears to increase the relative likelihood of being agricultural Source: Authors’ calculations, IHSES 2012. households. One reason why agricultural and diversified house- What Explains the Lower Poverty Rates of holds may have different coefficients on similar levels Diversified Households? of characteristics is that diversified households are en- Starting from a situation in 2007 where diversified gaged in different sectors of employment and these rural households experienced almost 10 percentage may be associated with different returns on the labor points higher poverty rates compared to households market. While we cannot directly introduce employ- completely dependent on agriculture, by 2012, ment sectors in the regression models and the de- headcount rates among diversified households had composition above, the coefficients associated with halved while agricultural households remained at households or individual characteristics may change the same poverty levels. This differential pattern due to the indirect effects of changes in diversifica- over the 2007 to 2012 period could be explained tion sectors and the associated earnings. There is by differences in the endowments or characteristics some evidence from descriptive data that this is the of these types of households, or may be attributable case. Table 37 focuses on the main employment sec- to other factors. We again use the Oaxaca-Blinder tors for individuals belonging to non-poor diversi- decomposition method to explain the poverty head- fied households in rural areas. In terms of the non- count difference between agriculture and diversified agricultural employment sectors, there is a marked households in 2012. shift towards manufacturing and financial, insurance and professional services (an increase of 5.5 and 12 The results show that the primary reason why ag- percentage points respectively); and a lower depen- riculture households are poorer than diversified dence on construction and public administration (a households is largely explained by differences in the decline of 6 and 10 percentage points respectively) as coefficients and not attributable to differences in sources of employment. The sectors into which the characteristics (Figure 183). We find that only one- non-poor have moved, manufacturing and financial, third of the difference in poverty among these types insurance and professional services, have both been of households is explained by their characteristics in associated with a large increase in per capita labor 2012. These results do not vary significantly if we earnings, 78 percent in the former and 68 percent in change the reference category. the latter. At the same time, earnings have increased The Labor Market for the Poor: The Rural-Urban Divide 169 TABLE 37: Changes in Non-Agricultural Employment and Labor Earnings for Non-Poor Rural Diversified Households, 2007 and 2012 Non-agricultural Median per capita Employment employment share labor earnings sector, change Earnings, (Percentage change 2007 2012 2007 2012 point) (Percent) Manufacturing 4.34 9.82 63.9 113.7 5.48 77.84 Construction 26.59 20.57 125.1 121.7 –6.02 –2.66 Commerce and retail 10.87 10.44 105.5 125.1 –0.43 18.54 Transport, storage & communication 12.00 13.09 106.2 107.0 1.22 0.78 Financial, insurance & professional 7.60 19.64 79.5 133.6 12.04 67.99 Public administration, health & education 27.88 17.88 122.9 129.5 –10.00 5.43 Other services 8.31 5.39 85.5 109.3 –2.92 27.89 in almost every other non-agricultural sector. These especially pensions, in urban areas have also per- patterns suggest that among diversified households, haps led to some limited welfare improvements.55 non-poor households were increasingly likely to be But the largest section of the urban poor belong employed in certain sectors that were associated with to the private sector, and here, the sectors where higher earnings.54 This in turn may well be evident the poor work have seen an increase in head count in the magnitude and signs of coefficients on char- rates; counteracted by the move of some poor ur- acteristics that predict poverty in the decomposition ban household heads from private sector employ- exercise. ment to public sector work, which has been associ- ated with higher earnings. To conclude, the labor market for the poor looks significantly different from that facing the non- While urban households as a whole have experi- poor in Iraq, and it varies considerably across ru- enced limited welfare gains, rural poverty reduction ral and urban areas. Poverty is not only correlated has been more marked, and has been driven by a with lower rates of employment and labor force significant welfare improvement among households participation, but also with important differences where individuals are employed in agriculture as in the types of economic activities. These differ- well as in other types of economic activity. In con- ences are compounded by lower levels of human trast, households that are wholly dependent on agri- capital and by urban-rural differences. Compar- culture have seen little welfare improvements. Here ing urban and rural households, not only are the spatial differences are again salient: in some parts characteristics of rural households starkly differ- ent—larger household sizes and lower educational attainment, for instance—, but even for the same 54 A similar comparison for poor rural diversified households characteristics, poverty rates are much higher for also shows a shift towards manufacturing and financial, in- rural households. This ‘characteristic deficit’ is surance and professional services which is associated with higher earnings. The difference between the poor and the accompanied by differences in the types of eco- non-poor within diversified households appears to be a nomic activity that rural and urban households are greater dependence on construction and lower earnings engaged in and how these are related to welfare. within each employment sector, which are likely associated In urban areas, public sector employment is as- with differences in characteristics between these households. sociated with lower poverty and correlated with 55 We take up the role of public and private transfers in great- higher labor earnings. Increases in public transfers, er detail in the next chapter. 170 THE UNFULFILLED PROMISE OF OIL AND GROWTH of the country, people are leaving agriculture and employment in agriculture is declining along with being absorbed into other sectors, and diversifica- male labor force participation. It appears as though tion is associated with better returns. In other parts the local labor market is barely creating adequate of the nation, notably the South, poverty has in- opportunities for diversification in the South, as creased especially among those who have not diver- even diversified and non-agricultural households sified outside of agriculture, while at the same time, continue to face high rates of poverty. Transfers, Safety Nets, and Poverty 8 T he poor in Iraq are disproportionately de- With the exception of the PDS, public transfers also pendent on non-labor incomes, and lacking cover a small proportion of the poor. Pension incomes assets, in particular, on transfers includ- reach less than 20 percent of the poor; social protec- ing through the Public Distribution System (PDS). tion network transfers cover only a tenth of the poor. Despite an increase in the share of income from the Per capita ration receipts were higher among house- labor market to 49 percent in 2012 from 42 percent holds with non-employed heads, and receipts decline in 2007, public and private transfers still account steadily with the increases in the education of the head for 36 percent of total income for the bottom decile, of household. Receipts are also higher for rural house- of which more than 80 percent is comprised of public holds, and in every division relative to Kurdistan, es- transfers. pecially in the South. The bigger the household size, the less it receives from the PDS. However, rations do not The dependence on transfers as a source of income also decline evenly as household size increases. The loss in varies widely across space, especially for the bottom de- transfers received from rations gets up to 30 percent cile. For instance, while labor incomes account for 58 when households are bigger than 12 members. On the percent of the total income of the bottom 10 percent other hand, the richer the household, the more it re- in Kurdistan, its share falls to 42 percent for the bot- ceives irrespective of it size. tom 10 percent living in the South. PDS transfer re- ceipts as a source of non-labor income are also smaller The Public Distribution System (PDS) remains the in Kurdistan, accounting for less than 20 percent of overwhelming source of calories for the poor and bot- non-labor incomes on average while it rise up to 48 tom 40 percent, accounting for 74 and 64 percent of percent of total non-labor incomes in the South. The their total caloric consumption respectively in 2012. relatively low share of ration transfers in Kurdistan At the same time, it accounts for 30 percent of food is compensated by relatively high shares of pension and expenditures for the poorest 10 percent of Iraqi house- capital income. holds, and 16 percent of total expenditures. In terms of the self-reported impact of the decline in rations, more Private transfers are relatively small in size and cover than 80 percent of households reported experiencing a a minority of the poor. International remittances cover decline in incomes as a result, and while 80 percent less than 1 percent of the poor and more than 90 percent reduced food stocks, and 70 percent cut back on food of the recipients are non-poor. Zakat transfers cover on- purchases, 20 percent had to increase food purchases to ly 2.4 percent of the poor, but a third of zakat recipients compensate for the lack of PDS items. are below the poverty line. Domestic remittances com- prise almost a third of poor and non-poor households; Overall, ration and free market items are essential in although only 20 percent of the recipients are poor. the consumption basket of Iraqis with the exception of 172 THE UNFULFILLED PROMISE OF OIL AND GROWTH free market oils. In other words, Iraqi households are in 2012, with a small decrease in the share of non- almost non-responsive in terms of altering demand to labor incomes, and within non-labor income, a small changes in food prices of ration items and their free decline in the importance of public transfers. market equivalents. Richer households are more re- sponsive to variations in prices of ration items than Among the poor, and especially among the bottom poorer households while the opposite is for free market 10 percent of the consumption distribution, the de- goods: less-well off households are more responsive to pendence on non-labor incomes, especially on pub- changes in prices of free market goods than those lo- lic transfers increases sharply. Despite an increase in cated in the upper part of the distribution. In general, the share of income from the labor market to 49 most rations items are marginally “inferior” goods percent in 2012 from 42 percent in 2007, public in the Kurdistan region irrespective of the level of per and private transfers still account for 36 percent of capita consumption. As the economy evolves and the total income for the bottom decile, of which more levels of income increases across the distribution, and than 80 percent is comprised of public transfers. as the rest of the country approaches the higher wel- While the share of public transfers in non-labor fare levels of Kurdistan, these types of ration goods will incomes remains above four-fifths for each of the be less demanded in the short run. Eliminating the bottom 4 deciles, there is a distinct shift towards a ration system would be approximately equivalent to higher share of labor income as households move increasing the price of ration items up to the market out from the bottom 10 percent. price levels given the low own price elasticities of ra- tion goods and the positive income elasticities for these The dependence on transfers as a source of income goods for much of the population. This will affect di- also varies widely across space, especially for the bot- rectly consumer’s welfare levels by reducing them by tom decile. For instance, while labor incomes account one-fifth to one-third for the upper quintiles in urban for 58 percent of the total income of the bottom 10 areas and up to 60% for the lowest quintile in urban percent in Kurdistan, its share falls to 42 percent for areas. However, where income levels are higher, local markets are more evolved, and where rations are not universally consumed, as in Kurdistan and Egypt, the TABLE 38: Main Sources of Income (Share greater flexibility in consumer response suggests that of Total), National Average and welfare impacts may be smaller when a similar envi- Bottom 4 Deciles, 2007 and 2012 ronment is created across the rest of the country. Non labor income Non Imputed Iraq Labor Labor Rent Public Private Transfers, Safety Nets and Poverty 2007 66.18 19.49 14.34 78.41 21.59 Deciles Faced with limited opportunities for employment 1 42.29 40.73 16.98 91.22 8.78 and earnings in the labor market, the poor in Iraq are 2 55.20 29.46 15.34 87.79 12.21 disproportionately dependent on non-labor incomes, 3 60.20 24.74 15.06 87.70 12.30 and lacking assets, in particular, on transfers includ- ing through the Public Distribution System (PDS). 4 60.77 23.92 15.31 87.01 12.99 In 2007, labor earnings accounted for 66 percent of 2012 68.00 16.40 15.60 75.62 24.38 total income for the average Iraqi, with another 20 Deciles percent from non-labor income (private and public 1 49.24 35.59 15.17 83.52 16.48 transfers), and 14 percent from implicit rental in- 2 61.99 24.18 13.83 85.01 14.99 come from owner occupied dwellings. Almost 80 3 64.15 21.15 14.70 82.46 17.54 percent of non-labor income was made up of public 4 65.93 18.92 15.16 81.35 18.65 transfers. This pattern remained roughly unchanged Source: Authors’ calculations, IHSES 2007 and 2012. Transfers, Safety Nets, and Poverty 173 the bottom 10 percent living in the South. At the of ration incomes in non-labor income from almost same time, ration or PDS transfer receipts as a source 60 percent to less than 40 percent between 2007 of non-labor income are also smaller in Kurdistan, and 2012. At the same time, there was an increase accounting for less than 20 percent of non-labor in absolute and relative terms in the contributions incomes on average, and 42 percent of non-labor due to pension incomes and domestic remittances, incomes among the bottom decile. In contrast, 48 which together accounted for more than 40 percent percent of total non-labor incomes in the South of non-labor incomes on average in 2012. accrue from implicit incomes associated with PDS receipts, and this share increases to 62 percent for the lowest 10 percent. The relatively low share of Who Receives Transfers? ration transfers in Kurdistan is compensated by rela- In what follows, we consider in greater detail six tively high shares of pension and capital income in different types of transfers and non-labor income: Kurdistan. Overall, pensions make up between 22 to 30 percent of average non-labor incomes, but are roughly half as important for the bottom ten percent PDS and Pensions as a Share of FIGURE 184:  of the consumption distribution. GDP, 2007 to 2012 12 These spatial differences are also important because poverty increased in those parts of the country where 10 2.3 3.3 the decline in implicit ration transfers between 2007 8 Percent of GDP and 2012 was not compensated by increases in oth- 3.7 4.3 6 5.7 er public and private transfers (as Chapter 5 estab- 5.8 3.3 4 3.6 lishes). Between 2007 and 2012, while the size of 3.5 2.5 2.1 pension transfers has steadily increased, following a 2 2.0 policy change in the PDS that reduced the num- 0 2.2 2.9 1.6 1.6 2.3 0.6 ber of items to be distributed in 2008–09, expendi- 2007 2008 2009 2010 2011 2012 tures on the PDS as a share of GDP have fallen from Subsidies PDS Pensions around 6 percent to 2 percent. In household level data, this change is reflected in a fall in the share Source: Authors’ calculations, IHSES 2007 and 2012. TABLE 39: Sources of Non-Labor Income Across Iraq, Overall and Bottom Decile, 2012 Iraq Kurdistan Baghdad North Centre South Share in total Labor income Overall 68.00 69.16 63.02 71.81 68.02 68.71 income, 2012 Lowest decile 49.24 58.18 52.80 49.82 49.21 41.70 Share of Rations Overall 38.92 19.72 41.45 42.20 38.69 48.21 non-labor income, Lowest decile 59.96 42.04 53.72 64.09 59.53 62.10 2012 Pensions Overall 26.25 30.40 33.33 23.87 24.39 21.78 Lowest decile 13.27 19.40 18.10 9.72 16.86 11.43 Domestic Overall 14.45 12.09 16.18 13.30 17.10 12.24 remittances Lowest decile 11.43 13.39 18.36 9.11 8.61 12.72 Capital income Overall 8.27 14.41 3.84 8.13 11.14 3.89 Lowest decile 3.39 6.31 1.81 5.66 4.30 1.62 Source: Authors’ calculations, IHSES 2007 and 2012. 174 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 185: Sources of Non-Labor Income, In general, private transfers are relatively small in 2007 and 2012 size and cover a minority of the poor. International remittances cover less than 1 percent of the poor and less than 2 percent of the non-poor; with more than 90 percent of the recipients being non-poor. Zakat 2012 26.25 14.45 38.92 transfers are also small, covering only 2.4 percent of the poor, but a third of zakat recipients are below the poverty line. Domestic remittances, which have 2007 19.97 8.76 49.67 become increasingly important over time, cover al- most a third of poor and non-poor households; al- though only 20 percent of the recipients are poor 0% 50% 100% (Figure 186). Capital Pensions International remittances Domestic remittances Turning to public transfers, pension incomes, which Social protection Other public transfers are not explicitly designed to as an anti-poverty Other private transfers Rations transfer, reach less than 20 percent of the poor, and Zakat more than a quarter of the non-poor. About 85 percent of pension recipients belong to non-poor Source: Authors’ calculations, IHSES 2007 and 2012. households. Social protection network transfers, on the other hand, do involve some categorical target- private transfers—domestic and international remit- ing of households, and while a quarter of the poor tances, and zakat receipts; as well as public transfers receive some form of such transfers, the program still in the form of pensions, transfers from the social covers only a tenth of the poor. Transfers through protection network, and implicit incomes associ- the PDS are, in contrast, almost universal, and cover ated with receipts of subsidized food items from the more than 99 percent of the poor. While less than Public Distribution System (PDS). a fifth of PDS beneficiaries are poor households; it FIGURE 186: Share of Poor and Non-Poor Individuals Receiving Public and Private Transfers, 2012 100 90 80 70 60 50 40 30 20 10 0 Rations Pensions Social Domestic Zakat International protection remittances remittances Public transfers Private transfers % of poor receiving 99.39 19.50 10.60 28.62 2.42 0.81 % of non-poor receiving 98.58 26.73 7.55 29.90 1.33 1.80 % recipients poor 19.95 15.29 25.77 19.13 31.12 9.95 Source: Authors’ calculations, IHSES 2007 and 2012. Transfers, Safety Nets, and Poverty 175 remains the only safety net that comprehensively household sizes and dependency ratios, belonging covers Iraq’s poor population. to a household with a less educated head, a female head, and a non-employed head, all increase the In order to further understand the role of these pri- likelihood of receiving zakat. Poorer households are vate and public transfers in household welfare, we un- also more likely to receive these transfers; although dertake multivariate analysis to identify the household households everywhere except in the South are less characteristics associated with whether a household likely to be beneficiaries compared to those living receives a particular transfer; and with the level of per in Kurdistan. While per capita zakat receipts de- capita receipts (results in the Annex Tables 8.1–8.6). cline in general with household size in 2012, they are larger for very large households with more than 20 members, for female headed households and Private transfers for households with heads employed in the private In general, female headed households and house- sector. Rural households and households living in holds with non-employed heads appear to be more Baghdad receive smaller zakat transfers on average; likely to receive international remittances; and as do households with non-employed heads. are also more likely to receive higher per capita amounts. While households living in Kurdistan are more likely to receive international remittances Public transfers compared to those living in other divisions, in terms Pensions are not designed to be anti-poverty trans- of the amount received per person, households in fers, and as expected, the presence of a widow or a the South receive more, while those in the Centre person of pensionable age in the household signifi- receive less. While rural households are less likely cantly increases the likelihood of the household re- to be recipients, among the ones who do receive ceiving pensions by 12 and 23 percent respectively. these remittances, per capita receipts are on average, Urban households are also, as one may expect, more higher than among urban recipients. likely to receive pensions; as are households living in the Kurdistan region; while households with less Domestic remittances are more likely to go to house- educated heads receive smaller amounts per person. holds with higher dependency ratios, although larg- Once these factors are taken into account, larger er households are less likely to be recipients and also households are more likely to be receiving pensions receive less per capita, relative to households with 1 (as these households tend to have more eligible to 4 members. Households headed by females and members) but receive less per capita. Moreover, by those not employed in the public sector are 5 and households with heads who are not employed in the 10 percent more likely to receive these transfers; and public sector are more likely to have someone in the receive larger per capita amounts. In 2012, house- household who is receiving pensions; and also have holds in rural areas were less likely to benefit from higher receipts per person. domestic remittances; while those with less educated heads were more likely to. While households in the Transfers from the social protection network, while Centre were 7 percent more likely to be recipients of small, do appear to be relatively pro-poor. On av- domestic remittances relative to those in Kurdistan, erage, the likelihood of receiving these transfers in- and those in the South were 9 percent less likely; per creases with household size, is higher for households capita receipts were the lowest in Kurdistan, and the whose heads are not employed in the public sector, highest in the South. However, overall, remittance whose heads are less educated, for households with amounts per person increase with wealth. a widow or eligible pensioner, and for households with lower consumption expenditures. However, per Zakat transfers, although very small, appear to capita receipts decline with household size and in- be quite progressive and well-targeted. Larger crease with household consumption expenditure. On 176 THE UNFULFILLED PROMISE OF OIL AND GROWTH the other hand, they are higher for households with social protection transfers do appear to be relatively less educated heads, for households with a widow, progressive. The coverage of the poor under these and for households with heads who are not employed transfers however, is also in part an outcome of hav- in the public sector. While households in the Cen- ing received the transfer. For instance, domestic re- tre and Kurdistan appear to be more likely to receive mittances may only cover a small proportion of the such transfers; among recipients, per capita transfers poor, because on average, the size of these transfers are higher in Baghdad, the Centre and the South. may be large enough so that the receipt of these transfers enables a household to consume above the Finally, we examine the correlates of incomes re- poverty line. While causal inference of the poverty ceived in the form of subsidized food through the mitigating impact of these transfers are beyond the Public Distribution System (PDS). Per capita ration scope of the analysis here, we can try to get a sense receipts were higher among households with non- of the correlation between the receipt of these trans- employed heads, and receipts decline steadily with fers and the likelihood of a household being poor or the increases in the education of the head of house- belonging to the bottom two quintiles of the con- hold. Receipts are also higher for rural households, sumption distribution. and in every division relative to Kurdistan, espe- cially in the South. However, larger households re- Annex Table 8.8 estimates the marginal effects, or ceive lower receipts per capita, as do female headed the change in the likelihood of these two outcomes, households, and households belonging to the bot- associated with the receipt of transfers, in addition tom 20 percent of the consumption distribution. to a range of household characteristics. The rela- tionship between household demographics and the In principle, PDS rations are supposed to be allo- education and labor market characteristics of the cated on a per person basis, so that per capita re- head with the two outcomes; as well as the relation- ceipts should be invariant to household size, and in ship between place of residence and welfare are as particular, to the number of people recorded on the expected; and in line with the diagnostic analysis in household’s ration cards. Therefore, we restrict at- Chapter 2. Higher dependency, larger household tention to the majority of households for whom the sizes, the household head’s employment in the pri- number of members reported on the ration cards is vate sector or the household head being non-em- identical to the number of household members. For ployed, living in a rural area, living in any division this sample of households, per capita receipts de- other than Kurdistan and the Centre, are all associ- cline with household size but increase with wealth ated with higher poverty odds. (Annex Table 8.7). To explore this apparent regres- sivity, we interact household size with consumption The receipt of pensions, international remittances in the final specification. The finding shows that the and higher levels of per capita PDS receipts are bigger the household size, the less it receives from all associated with lower likelihood of poverty and the PDS. However, rations do not decline evenly belonging to the bottom 40 percent. While this as household size increases. The loss in transfers should not be causally interpreted, it may suggest received from rations gets up to 30 percent when that among otherwise similar households, the size households are bigger than 12 members. On the of these transfers are large enough so that those who other hand, the richer the household, the more it receive them are likely to have significantly higher receives irrespective of it size. welfare. In contrast, the relatively progressive za- kat transfers and social protection payments do not Overall, most public and private transfers received lower the likelihood of being poor. In fact, they are by households tend to be small and cover a minor- associated with a higher likelihood of poverty and ity of the poor, with the important exception of belonging to the bottom 40 percent. Again, this the PDS. While very small in magnitude, zakat and does not imply that the receipt of these transfers Transfers, Safety Nets, and Poverty 177 increases poverty, but perhaps instead, that while among better off households, and especially among these are associated with poorer households, they the top 2 quintiles. In 2012, PDS consumption ac- do not bridge the gap sufficiently. counted for only a third of calories consumed by the richest 20 percent of the consumption distribution. That being said, the PDS remains the overwhelm- The Public Distribution System ing source of calories for the poor and bottom 40 percent, accounting for 74 and 64 percent of their Iraq’s Public Distribution System, the largest pub- total caloric consumption respectively in 2012. licly subsidized food distribution system in the world, remains the only safety net covering the PDS expenditures account for 30 percent of food poor and vulnerable in the country. The reform in expenditures for the poorest 10 percent of Iraqi 2009–2010 that cut the number of items distrib- households, and 16 percent of total expenditures uted through the PDS by and large left the caloric (Figure 188). The share of PDS expenditures de- content of the PDS basket unchanged, dropping clines to 12 percent for the 2nd decile, to 7 percent items such as detergents, soap, milk (for adults), among the 5th decile, and to less than 2 percent for tea, and tomato paste. Thus, much of any observed the top decile. In the consumption aggregate, con- changes in caloric consumption attributable to the sumption of PDS items are valued at the national PDS reflect a reduction in consumption of ration median of the prices reported by ration agents in items rather that a change due to the reduction in response to the question: “If you could buy this the number of items. Figure 187 plots the share of [ITEM] in the market, how much would you have total food calories from the PDS for households, for to pay for it?” Even though these prices are signifi- each of the quintiles of the consumption distribu- cantly higher than the official (subsidized) prices for tion. In 2007, three-quarters of the calories of the ration items, they still represent a significant under- bottom 20 percent came from consumption of PDS estimate of the shadow cost of the ration bundle items; while for the top 20 percent, this share was (See Box 5 for a brief description of some of the 45 percent. Between 2007 and 2012, while depen- challenges in valuing PDS rations). dence on the PDS as a source of calories has changed little for the bottom 40 percent, it has come down FIGURE 188: PDS Expenditures, Food and FIGURE 187: Share of Calories from the PDS, Non-Food Expenditures by by Consumption Quintile, 2007 Consumption Quintile, 2012* and 2012 100% 0.8 0.750.74 80% 0.7 0.68 0.64 0.6 60% % of total calories 0.5 0.45 40% 0.4 0.33 35.40 36.82 37.15 37.25 36.35 36.71 35.73 36.40 35.53 34.29 0.3 20% 0.2 0% 16.52 11.63 9.59 8.03 6.87 5.96 5.05 4.12 3.27 1.93 0.1 1 2 3 4 5 6 7 8 9 10 0 1 (poorest) 2 3 4 5 (richest) Ration Non-ration Non-food 2007 2012 Source: Authors’ calculations, IHSES 2007 and 2012. Note: Rations are valued at national median ration agent prices, which is a large Source: Authors’ calculations, IHSES 2007 and 2012. underestimate of the shadow cost of the ration bundle 178 THE UNFULFILLED PROMISE OF OIL AND GROWTH BOX 5: PDS Ration Items and the Valuation Problem The IHSES surveys collect information about the quantity of ration items received, consumed and purchased. The 2007 methodology used a notion of ‘net quantity received’ and purchases of ration items recorded in the diary on a monthly basis (very few transactions) to measure the quantity of rations consumed. The former is the quantity of ration items received, net of amounts bartered, sold, or given away. However, this measure has no clearly defined recall period, such as the last week or the last month. Moreover, since receipts are not consumption, they may not reflect utility. Two households who receive the same amount of rations, but consume very different amounts, derive different utility from rations. The IHSES surveys also include a direct question on consumption of ration items within the last 30 days. This is a more accurate measure of consumption, with a clear recall period, and equal consumption implies equal utility derived for households. This is the primary measure of consumption of ration items in the revised methodology. Purchases of ration items in the diary (over the last week) are converted into monthly equivalents, and also included, as households who purchase additional rations on the market must be assigned higher consumption and thereby utility. How is this important component of food consumption to be valued? In principle, goods and services ought to be valued equal to their infra-marginal benefit; i.e.; the market price faced for the marginal unit consumed. In the case of Iraq, ration items are rarely traded and in this sense, a market-equivalent price does not exist. A few transactions are recorded in the diary but these are insufficient to calculate unit values, and moreover, are associated with a select few households who are quantity constrained. So these unit values cannot be used to value all ration consumption. Another possibility is to use official prices for ration items, which are very low, nominal prices paid by consumers. Using these heavily subsidized prices would artificially suppress the value of food expenditures stemming from rations. Moreover, rations should be valued at a price close to one at which we expect these items to be traded; and official prices are not the prices at which households can procure unlimited quantities. Is there a close substitute to ration items that are traded in the market? In the case of Iraq, unit values for these substitutes are significantly higher for some items, especially rice, suggesting important quality differences. This implies that market prices for commercially available items cannot be used because they are not perfect substitutes. The only remaining candidate to value rations is a question that asks households their opinion on how much they would pay for ration-equivalent items in the market. In practice, few households expressed an opinion, and enumerators approached the local ration agent in the cluster, in a manner akin to a price survey. However, there were variations in these prices that may reflect uncertainty, noise and local variations in supply, demand and quality. In order to ensure that all those who consume exactly the same amount of a ration item are assigned the same expenditure; and that this expenditure increases with higher consumption; it was decided to use the national median values of prices reported by ration agents to value ration items. Source: Poverty in Iraq: 2007–2012—Methodological Note. In this context, how did households cope with the to jobs, death/injury, security, and other—jobs and decline in transfers in the form of implicit ration security related shocks affected more than 9 per- incomes? We begin by first broadly characterizing cent and more than 7 percent of the population as the shocks experienced by households in 2007 and whole; and prevalence did not vary by poverty sta- 2012, and relate these to the main coping response. tus. In 2012, the incidence of these types of shocks In particular, in 2012, the survey asks about wheth- had increased to 24 percent among the poor and er a household experienced a shock due to the loss 28 percent among the non-poor.56 The higher in- of rations, the loss of other government assistance, cidence of shocks among the non-poor is related to or due to a decline in remittances. Only 3 percent of households experience covariate shocks related to households reported having experienced this type of shock, primarily driven by lower rations. 56 It should be noted that the questionnaires are not strictly In 2007, 17 percent of all poor and non-poor Iraqi comparable between 2007 and 2012. While the 2007 mod- households reported having suffered from a shock ules lists a possible set of 11 shocks, including “Another in the past 12 months—these included the loss of huge problem”, the 2012 module is more detailed, asking about 23 possible shocks, including in particular, agricul- a job or of wages, or of the closure of a business; ture-related shocks and shocks affecting the local economy sickness, injury or the death of a household mem- and community. It is likely that the longer list of possible ber; theft, violence, kidnapping and other types of options in 2012 elicited a better response, and as a result problems. Of each of these types of shocks—related 2012 prevalence rates are significantly higher. Transfers, Safety Nets, and Poverty 179 agriculture—including drought, the loss of assets or FIGURE 190: Incidence of Shocks by Poverty livestock, reduced agricultural water quality, pests Status, 2012 and diseases and reduced availability of grazing ar- eas. These are also reflected in 15 percent of rural 28% household reporting having experienced an agricul- 24% ture related shock, especially in Kurdistan and the 17% North, with prevalence rates of 8 and 10 percent re- 16% 14% 13% spectively. These may be reflecting the drought ex- 11% 9% perienced in northern Iraq and Syria between 2007 8% 8% 7% 5% 5% and 2009. In 2012, as in 2007, the prevalence of 4% 3% 3% 1% 1% different types of shocks does not vary by poverty status, except as noted already, covariate shocks re- Any shock Idiosyncratic Covariate Agriculture Security Job Death/Injury Transfers Local lated to agriculture. While the prevalence of shocks was higher in urban areas in 2007, explained by higher job-related shocks, in 2012, rural areas expe- Non poor Poor rienced higher shocks, driven by agricultural shocks. Source: Authors’ calculations, IHSES 2007 and 2012. In 2012, households in Baghdad were most likely to report having experienced a shock in the last 12 months. About a third of households had experi- shocks were relatively high in Kurdistan, the North enced at least one shock, and a fifth reported a ‘lo- and the Centre. cal’ shock—an aggregation that includes reduced drinking water quality and availability; an unusu- The 2012 IHSES data also includes information on ally high level of human disease; or unusually high the effect of the shocks on households, i.e., house- prices of food and other essential commodities. holds report whether they increased, decreased or These local shocks were also quite high across the did not alter food purchases, food stocks, food pro- other divisions, and in addition, agriculture-related duction, assets and income as a result of each type FIGURE 189: Incidence of Shocks by Poverty FIGURE 191: Prevalence of Shocks in Urban Status, 2007 and Rural Areas, 2007 18% 17% 17% 15% 10% 10% 9% 9% 8% 7% 7% 6% 6% 6% 5% 5% 3% 2% 2% 2% Any shock Job Security Death/Injury Other Any shock Job Security Death/Injury Other Non poor Poor Rural Urban Source: Authors’ calculations, IHSES 2007 and 2012. Source: Authors’ calculations, IHSES 2007 and 2012. 180 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 192: Prevalence of Shocks in Urban For most other types of shocks, in general, the pri- and Rural Areas, 2012 mary impact seems to be in terms of a reduction in incomes and in food purchases, with food stocks, production and assets being less affected. While dif- 30% ferences in food production effects are understand- 23% able (as those experiencing agricultural shocks are 19% also likely to be those engaged in food production), 15% 15% 14% on average, between 20 and 30 percent of house- 11% 10% 10% holds experience declining assets in the face of other 8% 6% types of shocks. In terms of transfer shocks, includ- 4% 4% 3% 2% 2% 1% 1% ing the decline in rations, more than 80 percent of households reported experiencing a decline in Any shock Idiosyncratic Covariate Agriculture Security Job Death/Injury Transfers Local incomes as a result, and while 80 percent reduced food stocks, and 70 percent cut back on food pur- Rural Urban chases, 20 percent had to increase food purchases to compensate for the lack of PDS items. Source: Authors’ calculations, IHSES 2007 and 2012. The main coping strategies of households appear to be fairly similar across idiosyncratic and covari- of shock. Irrespective of whether the shock was id- ate shocks, with about a third of households rely- iosyncratic, i.e., household or individual specific, or ing on their own savings, and less than a fifth doing covariate, i.e., community or locality specific, around nothing. One important difference, however, is that four-fifths of households reported a reduction in in- households were more likely to reduce the quan- come, around two-thirds reported a reduction in tity, quality and variety of food or purchase food on food purchases, 60 percent reported reducing food credit in response to covariate shocks; whereas in stocks. Reduction in assets and food production was response to individual or household specific shocks, less likely, while 12 percent of households increased loans, credit and assistance from friends and relatives food purchases. becomes more important. These patterns are re- flected in responses to agricultural and local shocks, For households faced with agriculture-related which are primarily covariate shocks. In contrast, the shocks, around 60 percent reduced food purchases, dependence on social and family networks is more stocks, and production; a similar proportion experi- important in responding to individual or household enced a reduction in assets, and almost 90 percent level shocks including job-related shocks, or those experienced falling incomes. No other shock results related to death or injury of a family member. In in such a decrease across food, assets and income. the case of shocks related to transfers, rations and TABLE 40: Prevalence of Shocks, 2012 Any shock Agriculture Security Job Death/Injury Transfers Local Kurdistan 24% 8% 2% 3% 10% 0% 8% Baghdad 31% 2% 1% 5% 7% 3% 19% North 27% 10% 2% 2% 6% 6% 8% Centre 27% 7% 1% 6% 9% 3% 8% South 15% 4% 0% 2% 6% 1% 6% Source: Authors’ calculations, IHSES 2007 and 2012. Transfers, Safety Nets, and Poverty 181 FIGURE 193: Household Perceptions of the Effects of Different Shocks, 2012 Idiosyncratic shocks Covariate shocks Food purchases Food purchases Food stock Food stock Food production Food production Assets Assets Income Income 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Increase Decrease Not change Source: Authors’ calculations, IHSES 2007 and 2012. FIGURE 194: Household Perceptions of the Effects of Each Type of Shock, 2012 Agriculture shock Local shock Food purchases Food purchases Food stock Food stock Food production Food production Assets Assets Income Income 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Job shock Security shock Food purchases Food purchases Food stock Food stock Food production Food production Assets Assets Income Income 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Transfer shock Death/Injury shock Food purchases Food purchases Food stock Food stock Food production Food production Assets Assets Income Income 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Increase Decrease Not change Source: Authors’ calculations, IHSES 2007 and 2012. remittances (which primarily comprises rations), the with households, with 50 percent of households re- food response seems to be particularly important; ducing the quality and quantity of food in response; 182 THE UNFULFILLED PROMISE OF OIL AND GROWTH FIGURE 195: Main Response to Each Shock, 2012 Local 23% 16% 16% 23% 15% 7% Transfer 14% 18% 17% 24% 27% Death injury 38% 16% 13% 6% 6% 12% 10% Job shock 36% 11% 13% 10% 7% 7% 10% 6% Security 32% 33% 15% 9% 6% 5% Agriculture 38% 20% 24% 7% 6% 5% Covariate 28% 19% 19% 18% 12% 5% Idiosyncratic 34% 17% 14% 10% 9% 9% 7% Relied on own savings Did not do anything Other Reduced quantity of food Reduced quality and variety of food Received help from relatives/friends Loans from relatives/friends Purchased food on credit Sold livestock Source: Authors’ calculations, IHSES 2007 and 2012. and only 14 percent relying on savings. Thus, faced going beyond food subsidies, the government of with declining ration items, many households fur- Iraq is considering further reforms to the PDS.57 ther cut back on food consumption in terms of quantity, quality and variety. In this section, we analyze the impacts of changes in the PDS on household welfare under different re- form scenarios. To do that, we use a partial equilib- Simulation of the Welfare Impact of the PDS rium setting which will allow us to estimate demand Even though the total expenditure on the Public responses that are essential in predicting outcomes Distribution System has reduced as a share of GDP of various policy reforms and in undertaking projec- over the last 5 year period (Figure 184), and the tions of food demand. This framework will answer total amount expended by the government has de- questions such as: how will consumers throughout creased in real terms from 2.3 to 1.1 ID trillions, it the entire consumption distribution adjust their still represents a fiscal burden for the government demand for rice and other food items if the effec- budget. While the PDS provides a level of broad tive price of rice is increased as a consequence of food security to the poor and vulnerable in Iraq, it the reduction in PDS rice distribution?; or, what will also covers more than 95 percent of the non-poor, be the effect on market demand of vegetable oil, and is therefore, a very expensive safety net program. brown flour or sugar? Moreover, in its current form, it suffers from sig- nificant inefficiencies in procurement, distribution, The model we use for this analysis, the Mixed De- and management, and implies significant macro- mand model, attempts to capture the consumption economic distortions because of its heavy reliance on food imports and its universal nature. For this 57 The Government of Iraq is currently considering moving to a and many other reasons, including the need for the ‘smart card’ system of delivery of PDS entitlements, and po- introduction of a comprehensive safety net system tentially a means-targeted eligibility criterion for the PDS. Transfers, Safety Nets, and Poverty 183 structure of Iraqis households, given the particular Budget Shares of Ration Items by TABLE 41:  characteristics of ration items and the distortions Decile of Per-Capita Consumption, these goods impose on their own markets and on 2012 those of other freely traded market goods.58 The two major empirical difficulties tackled by this approach, Total rations (% of total Brown relative to more traditional demand systems such as expenditure) Rice wheat Sugar Vegetable oil the Almost Ideal Demand System (AIDS) are the 1 16.52 6.43 2.67 3.95 3.28 following: the lack of enough variability in the price 2 11.63 4.41 1.91 2.78 2.31 of ration goods makes it impossible to identify price 3 9.59 3.64 1.58 2.31 1.95 effects; and the identification of demand from ob- 4 8.03 2.99 1.26 1.99 1.67 served purchases given that the quantity supplied to 5 6.87 2.58 1.08 1.69 1.42 each household is fixed.59 6 5.96 2.15 0.93 1.50 1.26 However, this methodological approach makes as- 7 5.05 1.79 0.78 1.32 1.10 sumptions which are common to these types of 8 4.12 1.41 0.63 1.09 0.93 demand systems in order to identify demand pat- 9 3.27 1.11 0.47 0.88 0.77 terns. Among these, it assumes that all goods are 10 1.93 0.58 0.26 0.56 0.49 purchased by households. This assumption has Source: Authors’ calculations, IHSES 2007 and 2012. empirical implications: given that not all house- holds within IHSES 2012 consumed or purchased all items, it generates an unbalanced sample across expenditure are larger than one percentage point in commodity groups. To solve this problem, we es- almost all deciles. timate the model at the stratum level for different quintiles in different geographic areas (i.e. urban Secondly, we consider four goods which are sub- and rural or Kurdistan and other regions), instead stitutes or complements for the ration items which of household level. The choice of aggregating over are traded in the free market. These are rice, cere- the stratum level generates a loss in information in als, sweets and oils, which are composite goods, i.e., the data compared to an analysis at the household goods for which prices within each group of com- level, but it avoids the need to adopt more sophis- modities move in parallel, so that the corresponding ticated procedures for dealing with multiple corner group can be treated as a single good.62 In sum, our solutions in demand systems.60 system is estimated for a total of eight items: four rations and four non-rations. Another choice which is essential for this kind of analysis is the type and number of goods included. 58 The Almost Ideal Demand System (AIDS) is not the best In the Iraqi case, the type of goods is easily defined methodological approach to use. This is essentially because by their nature: ration and non-ration items. It is the existence of food subsidies in the consumer demand the number of items which represents a problem. system is associated with individual consumption quotas, Larger demand systems are harder to deal with than which introduce nonlinearities in demand functions (see smaller ones; the more goods, the greater the com- Annex for further details on the model). putational burden, and the harder it is to report the 59 In particular cases such as Iraq where rationing quotas are results.61 To avoid this problem, we firstly included conditioned upon observed characteristics of the household, some degree of variability in purchased quantity is possible. four out of eight ration items with significant bud- However, we opted for the Mixed Demand Model. get shares: rice, brown wheat, sugar and vegetable 60 See Ramadan and Thomas (2011). oil. Table 41shows that these items contribute more than 98 percent of total rations expenditure across 61 Deaton (1997). the entire distribution; and their share in total 62 Deaton and Muellbauer (1980). 184 THE UNFULFILLED PROMISE OF OIL AND GROWTH Own-Price Elasticities of Ration Items by Quintile of Per Capita Consumption and FIGURE 196:  Area, 2012 Rural Urban 0.00 0.00 –0.01 –0.01 –0.02 –0.02 –0.03 –0.03 –0.04 –0.04 –0.05 –0.05 –0.06 –0.06 –0.07 –0.07 1 2 3 4 5 1 2 3 4 5 Brown our Rice Sugar Vegetable oil Source: Authors’ calculations, IHSES 2007 and 2012. Note this methodological approach is quite useful a one percentage change in price holding everything to simulate variations in prices and quantities which else constant. When this relationship is positive, then are not significant enough to make individuals shift these goods are considered as luxury goods. from their initial utility levels. Thus, significant changes in prices will need a general equilibrium Figure 196 and Figure 197 present the own-price framework rather than this type of analysis, because elasticities of the four ration and free market items it would include all links and implications between in urban and rural Iraq respectively. Overall, the es- macro and micro levels sectors. However, it’s major timates suggest that these particular items are es- constraint is the intense information demand which sential in the consumption basket of Iraqis with the makes them difficult to apply in most developing exception of free market oils. In other words, Iraqi countries. Another caveat of the mixed demand households are almost non-responsive in terms of model is that requires a closed form expression of altering demand to changes in food prices of ration utility functions.63 In other words, results depend items and their free market equivalents. If, for ex- on the assumed function used to represent house- ample, the price of ration brown flour increases 10 holds’ satisfaction levels (i.e. utility). percent, consumers living in rural areas would de- crease their demand for ration by 0.3 percent if they are in the bottom 40 percent of the consumption How does household demand respond to distribution. Similarly, if prices of sugar increase by changes in own prices?64 In general, it is expected that quantities demanded 63 We are using the Gorman Polar form suggested by Mos- of a particular good will decrease in response to an chini and Rizzi (2007). increase in its own price (law of demand) and these 64 Estimates of the cross-price elasticities are not presented types of goods are classified as normal. Price elastic- in this report. See Ramadan, Krishnan and Olivieri (2014) ity of demand (own-price elasticity) gives the per- for estimates of cross-price elasticities for 2012 IHSES and centage change in quantity demanded in response to their interpretation. Transfers, Safety Nets, and Poverty 185 Own-Price Elasticities of Free Market Goods by Quintile of Per Capita Consumption FIGURE 197:  and Area, 2012 Rural Urban 0.00 0.00 0.01 0.00 –0.01 –0.10 0.00 –0.10 –0.01 –0.20 –0.01 –0.20 Cereals, Sweets, Rice Cereals, Sweets, Rice –0.01 –0.30 –0.02 –0.30 –0.02 –0.40 –0.02 –0.40 Oil Oil –0.02 –0.50 –0.03 –0.50 –0.03 –0.60 –0.03 –0.60 –0.03 –0.70 –0.04 –0.70 –0.04 –0.80 –0.04 –0.80 –0.04 –0.90 1 2 3 4 5 1 2 3 4 5 FM cereals FM rice FM sweets FM oils Source: Authors’ calculations, IHSES 2007 and 2012. 10 percent, irrespective of consumption quintile or How does income affect consumption of living in rural or urban areas, consumers reduce patterns? their demand by only 1 percent or less. However, if After price and quantity adjustments (like substitu- the price of free market oil increases by 10 percent tion) have taken place, households may end up with in rural areas, demand for oil by Iraqi consumers net positive or negative income. The income or ex- would decrease by 8 percent in the lowest quintile. penditure elasticity of demand measures the respon- siveness of the demand for a good to a change in the Having said this, there is some variation of elastici- income/expenditure of the people demanding that ties levels across quintiles especially for free market particular good, holding everything else constant. goods (Figure 197). The elasticities of three out of A negative expenditure elasticity of demand is as- four free market products (i.e. rice, sugar and oil) de- sociated with “inferior” goods while a positive value crease as consumption per capita increases. In other with “normal” goods. words, less well-off households are more responsive to changes in prices of free market goods than those Table 42 presents the expenditure or income elas- located in the upper part of the distribution. Howev- ticities of ration and free markets goods by area and er, the opposite happens for ration items particularly quintiles. Overall, most products show a positive ex- for rice and brown flour. Richer households are more penditure elasticity in both areas and across quintiles. responsive to variations in prices of ration items than This implies on the one hand that these are normal poorer households. Well-off households may have goods i.e. their consumption increases when expen- other options like selling their quotas in the market diture increases, and on the other hand these are nec- or substituting them for better quality goods than essary goods, which are reflected by values less than less well-off households. Finally, elasticity for ration one. Another takeaway is that more expensive food oil and sugar and for free market cereals varies little items such as free market goods have relatively high by consumption quintile, and the demand response expenditure elasticities for all quintiles relative to ra- to changes in price are uniformly close to zero. tion items. At the same time, less well-off households 186 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 42: Expenditure Elasticities by Quintile of Per Capita Consumption and Area, 2012 Ration Products Equivalent Free Market Products Quintiles Brown Flour Rice Sugar Oil Brown Flour Rice Sugar Oil Rural 1 0.015 –0.024 0.005 0.021 0.086 0.353 0.359 –0.017 2 0.084 –0.005 0.006 0.023 0.054 0.227 0.259 –0.021 3 0.059 –0.009 0.005 0.027 0.045 0.155 0.213 –0.022 4 0.027 0.003 0.005 0.030 0.037 0.118 0.180 –0.013 5 0.263 0.082 0.017 0.044 0.033 0.072 0.166 –0.046 Urban 1 –0.057 –0.008 0.000 0.019 0.060 0.359 0.292 0.087 2 –0.073 –0.014 0.000 0.022 0.044 0.256 0.224 0.085 3 –0.093 –0.004 –0.001 0.025 0.035 0.175 0.187 0.061 4 –0.088 0.012 –0.001 0.028 0.029 0.129 0.162 0.042 5 –0.146 0.035 –0.002 0.041 0.023 0.082 0.152 0.017 Source: Authors’ calculations, IHSES 2007 and 2012. are more responsive than their well-off counterparts constant. Thus, we consider how households would for free market goods, which is a common pattern. adjust their consumption patterns over time, as wel- fare levels improved, by comparing current demand However, there are exceptions like ration brown responses in Kurdistan and the rest of Iraq. We take flour and free market oils. The expenditure elastici- Kurdistan as the reference region because their cur- ties for ration brown flour are negative in urban ar- rent consumption levels of ration items are the low- eas and positive in rural areas for all quintiles. This est in the country and because their per capita ex- indicates that ration brown flour is a marginally “in- penditure levels are the highest on average. ferior” good in urban Iraq: if household incomes increased by 10 percent, demand for ration brown Similar consumption responses to changes in own- flour would decrease from 0.5 up to 1.5 percent in prices of ration and free market goods are seen for urban areas. The opposite happens for free market Kurdistan and for the rest of Iraq relative to previ- oil which is marginally a “normal” good in urban ous findings for urban and rural area (Figure 198). areas and the opposite in rural areas. Overall, most goods are ordinary goods meaning that demand for these type of goods decrease when there is an increase in their own prices. Not surpris- How would consumers adjust their ingly, ration items are much less elastic than free mar- consumption responses over time? ket goods. However, all levels responses are higher Given the lack of information of future consumer in Kurdistan than in the rest of Iraq and also higher responses to changes in prices and expenditure, one than the estimates for urban areas shown above. At way to understand behavior over time is by exploit- the same time, well-off households in Kurdistan re- ing the rich and vast spatial disparities that Iraq has. gion are much more responsive to variations in prices In other words, consumer behavior in better off re- of ration goods and the opposite for their free mar- gions may be a rough approximation of how worse ket equivalents than in the rest of Iraq and in urban off regions today will behave in the future as their Iraq. In other words, in line with higher welfare lev- welfare levels improved, holding everything else els in Kurdistan relative to urban Iraq, and in urban Transfers, Safety Nets, and Poverty 187 Own-Price Elasticities of Ration Items by Quintile of Per Capita Consumption and FIGURE 198:  Area, 2012 Kurdistan: Ration goods Rest of Iraq: Ration goods 0.00 0.00 –0.01 –0.01 –0.02 –0.02 –0.03 –0.04 –0.03 –0.05 –0.04 –0.06 –0.05 –0.07 –0.08 –0.06 –0.09 –0.07 1 2 3 4 5 1 2 3 4 5 Brown our Rice Sugar Vegetable oil Kurdistan: Free market goods Rest of Iraq: Free market goods 0.02 0.02 0.02 0.02 0.01 0.01 –0.18 –0.08 0.00 Cereals, Sweets, Rice 0.01 –0.38 Cereals, Sweets, Rice –0.18 –0.01 0.00 –0.58 –0.02 –0.28 Oil Oil –0.03 –0.01 –0.78 –0.38 –0.04 –0.01 –0.98 –0.48 –0.05 –0.02 –1.18 –0.06 –0.58 1 2 3 4 5 1 2 3 4 5 FM cereals FM rice FM sweets FM oils Source: Authors’ calculations, IHSES 2007 and 2012. Iraq relative to rural Iraq, the flexibility of consumer “inferior” goods in the Kurdistan region irrespec- demand to changes in prices increases. Thus, as the tive of the level of per capita consumption. In other economy grows, consumers face greater options and words, as household expenditures increase by 10 ability to substitute away from ration items and in- percent, demand will fall by between 0.4 and 3.4 crease their consumption of free market goods. percent for brown flour and by around 0.7 percent for rice. Opposite responses are obtained in the Rest This pattern in consumer behavior is quite clear of Iraq: ration items are considered “normal” goods. when inspecting demand responses for goods to In sum, as the economy evolves and the levels of in- variations in total household expenditure and in- come increases across the distribution, and as the come. In general, most rations items are marginally rest of the country approaches the higher welfare 188 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 43: Expenditure Elasticities by Quintile of Per-Capita Consumption and Area, 2012 Ration Products Free Market Equivalents Quintiles Brown Flour Rice Sugar Oil Brown Flour Rice Sugar Oil Kurdistan 1 –0.04 –0.07 0.00 0.03 0.04 0.29 0.21 0.16 2 –0.10 –0.11 0.00 0.03 0.04 0.18 0.20 0.17 3 –0.02 –0.07 0.00 0.04 0.03 0.12 0.18 0.06 4 –0.13 –0.07 0.00 0.04 0.03 0.09 0.15 0.08 5 –0.34 –0.07 –0.01 0.07 0.02 0.06 0.14 0.05 Rest of Iraq 1 0.03 0.01 0.00 0.02 0.09 0.46 0.37 –0.13 2 0.05 0.02 0.00 0.03 0.06 0.35 0.25 –0.09 3 0.00 0.02 0.00 0.03 0.04 0.25 0.21 –0.07 4 –0.02 0.03 0.00 0.03 0.03 0.20 0.17 –0.06 5 0.06 0.09 0.00 0.05 0.03 0.12 0.16 –0.06 Source: Authors’ calculations, IHSES 2007 and 2012. levels of Kurdistan, these types of ration goods will TABLE 44: Own Price Elasticities by Subsidy be less demanded in the short run. Products, Egypt 1997 and Iraq 2012 Thus we see that in part, demand elasticities for Egypt Iraq ration items are likely to become larger as welfare Subsidy plus Cooking oil –0.030 –0.002 Quantity ration levels improve allowing for a decline in consump- Sugar –0.120 –0.008 tion of ration items when faced with price increases Brown Flour –0.042 and a greater consumption of free market goods as Rice –0.037 incomes rise. Another piece of the puzzle is how Subsidy only Wheat flour –0.060 big these price elasticities would be in an economy Bread –0.120 where developed markets of ration goods exist. Giv- Source: Authors’ calculations, IHSES 2007 and 2012. en that ration goods are universally distributed in Iraq, the spatial framework used until now is rela- tively uninformative. addition, wheat flour and bread are subsidized uni- We obtain suggestive evidence by comparing Iraq versally, and there is no quantity rationing so that with Egypt, where a public distribution system ex- households can acquire as much as they would like ists for food but it is not a universal system. Table to consume at the subsidized price. Perhaps because 44 presents own-price elasticities for ration items of these differences, Egypt has a far more developed in Egypt. There are several differences between the market for ration items and own-price elasticities PDS system implemented in Egypt and in Iraq. For are far (almost 15 times) higher than in Iraq, where instance, while in Iraq, quantities of ration items are there is negligible trade in PDS items. a function of the number of members included in the ration card, Egypt distributes fixed quotas and 65 Further details on the Egypt system: see Ramadan and subsidies for cooking oil and sugar and consumers Thomas (2011) and Ahmed, Bouis, Gutner and Lofgren can complement them from the free market.65 In (2001). Transfers, Safety Nets, and Poverty 189 PDS reform scenarios and their simulated Table 45 shows the average change in the welfare ag- welfare impacts gregate for both scenarios and each simulation step The analysis so far suggests that faced with changes in by quintile of the consumption aggregate in urban the effective prices of ration items, households will, areas. According to the previous discussion, elimi- on average, have very limited change in their demand nating the ration system would be approximately for these items. However, there are also some indica- equivalent to increasing the price of ration items up tions that ration goods slowly become less preferred to the market price levels given the low response of at least for those households who have larger bud- consumers to variations in prices and expenditures. gets. For instance, brown flour, which accounts for This will affect directly consumer’s welfare levels by the largest share of expenditures within ration items, reducing them by one-fifth to one-third for the up- is an inferior good in urban areas as well as in Kurd- per quintiles in Scenario A and up to 60% for the istan. Moreover, rice from rations also is an inferior lowest quintile in urban areas in Scenario B. good in Kurdistan. On the other hand, as welfare lev- els improve in rural areas, people will not significantly As the country grows, households would become lower their demand for ration brown flour, sugar and better-off and the demand for ration items would oil in the short run, which is also true for urban areas, reduce as consequence of being “inferior” goods. Kurdistan and the Rest of Iraq. However, as a result To account for this possibility, we focus on the of higher prevailing welfare levels in urban areas and results based on the Kurdistan-Rest of Iraq com- particularly in Kurdistan region, as well as the pres- parison. Table 46 shows the average change in ence of relatively well-functioning markets, consum- the welfare aggregate for both scenarios and each ers do exhibit more flexible consumption patterns. simulation step by quintile of the consumption ag- gregate in Kurdistan. Given the higher elasticities Taking these into account, we propose two reform to changes in own prices and in incomes in Kurdis- scenarios of the Public Distribution System which tan, the welfare impact of this type of reform of the may minimize the social effort by removing ration items from those households who may need them the least or who could easily adjust their consump- TABLE 45: Average Change in Total Expenditure tion or a combination of both. The first scenario by Quintile in Urban Areas (Scenario A) consists in targeting rations only to the bottom 60 percent of the urban (or Kurdistan) con- Rice + sumption distribution and entire rural (or Rest of Rice + Sugar + Scenario A Rice Rice + Oil Sugar + Oil Oil+ Wheat Iraq) population. The second scenario (Scenario B) 1 (Poorest) 0% 0% 0% 0% involves in targeting rations only to rural (or Rest 2 0% 0% 0% 0% of Iraq) areas. The simulation process is the same for both scenarios: we start removing ration items, 3 0% 0% 0% 0% from the smallest to the largest ration item, one by 4 –8% –14% –23% –34% one according to the importance of the item mea- 5 (Richest) –7% –10% –17% –20% sured by its share in the total expenditure. Rice + Rice + Sugar + Oil Note, in both scenarios, simulations do not consider Scenario B Rice Rice + Oil Sugar + Oil + Wheat spillover effects and/or effects among and between 1 (Poorest) –12% –24% –39% –60% households who live in the same or different geo- 2 –10% –21% –33% –51% graphic areas. These are strong assumptions, given 3 –9% –17% –28% –43% that when such policies are implemented may result 4 –8% –14% –23% –34% in black markets and leakages without simultaneous 5 (Richest) –7% –10% –17% –20% implementation of effective targeting policies. Source: Authors’ calculations, IHSES 2007 and 2012. 190 THE UNFULFILLED PROMISE OF OIL AND GROWTH TABLE 46: Average Change in Total Expenditure impacted by PDS reforms because they already by Quintile in Kurdistan Region consume lower quantities in Kurdistan and can adjust their demand for ration items more easily Rice + when both initial levels of welfare are higher and Rice + Sugar + markets are relatively well developed. Scenario A Rice Rice + Oil Sugar + Oil Oil+ Wheat 1 (Poorest) 0% 0% 0% 0% To conclude, the poorer segments of the consump- 2 0% 0% 0% 0% tion distribution in Iraq are disproportionately de- 3 0% 0% 0% 0% pendent on public and private transfers to supple- 4 –8% –11% –18% –27% ment their relatively low earnings on the labor 5 (Richest) –7% –6% –10% –12% market. Most of these transfers are however, small, Rice + and cover a fraction of the poor. The single exception Rice + Sugar + Oil is the Public Distribution System, which guarantees Scenario B Rice Rice + Oil Sugar + Oil + Wheat a minimum amount of caloric consumption for not 1 (Poorest) –11% –20% –32% –46% just the poor, but the whole population. Given its 2 –10% –18% –28% –42% universal nature, large fiscal costs, and the significant 3 –8% –14% –23% –34% distortions the PDS introduces in the economy as 4 –8% –11% –18% –27% a whole, we try to estimate the welfare impact of 5 (Richest) –7% –6% –10% –12% reforming the PDS by targeting it to a section of Source: Authors’ calculations, IHSES 2007 and 2012. the population. Given the universality of consump- tion, the lack of a market for ration items, and the low levels of income for much of the population, PDS system is relatively smaller than what might household consumption of PDS items is relatively be expected for urban areas as a whole. The larg- inelastic to changes in price, and for much of the est impact is experienced by the poorest quintile population, these goods are not inferior, but rather in Scenario B, a reduction in welfare levels by half, normal goods. However, there are some signs that compared to a 60 percent decrease in expenditures with improvements in welfare levels, and faced with for the poorest 20 percent in urban areas. In Sce- well-functioning markets, some segments of the nario A, where the top 40 percent are excluded population are substituting away from the PDS and from receiving PDS items, the welfare impact on increasing their consumption of market substitutes. the richest quintile in Kurdistan is a decline in Overall, this suggests that while any one-shot reform average expenditures by 12 percent compared to will have adverse and sizeable welfare impacts, over 20 percent for the richest quintiles in urban Iraq. time, and with increases in incomes, some house- This may reflect the fact that consumers are less holds may not be as significantly affected. Policy Implications: Learning from the Past to Build a Better Future 9 M ore than in most countries, poverty and which welfare gains will be difficult to realize and deprivation in Iraq are symptomatic of sustain; its daunting legacy of violence and fragil- • Sequencing will be very important to secure gains ity, of a skewed economy heavily dependent on oil and from reform, and supply and demand side con- on public intervention, and of decades of lost develop- straints will need to be relaxed simultaneously: ment. Addressing the twin goals in this context and for instance, investing in education access and with continuing violence will require a concerted and quality will not bear fruit absent adequate pro- multi-sectoral approach, combining short-term and ductive employment opportunities that generate medium-term efforts. appropriate returns on the labor market; • Taking advantage of spillovers and complemen- Establishing and maintaining peace and security tarities, such as spatially targeted public works across the nation is a pre-requisite for sustained and programs to generate employment and increase healthy economic growth. Moreover, Iraq will need incomes which also have the potential to redress to strengthen the natural links between growth and the infrastructure deficit and build skills for fu- welfare through better management of oil resources, ture jobs; non-oil diversification of the economy, bridging spa- • Finally, while there is clear scope for urgent ac- tial inequities, private sector development and imple- tion, other efforts are also needed that will require menting a well-designed and comprehensive safety net a medium to long term horizon. system. An economic growth process that creates em- ployment and increases earnings, while maintaining A comprehensive policy framework for enhancing the incentives for labor participation and investments in welfare of Iraq’s citizens will require a three-pronged education, is the least costly path to poverty reduction approach: and inclusion. At the same time, Iraq faces a large deficit in infrastructure, services and human capital, • Establishing and maintaining peace and security which will need to be redressed. Moreover, targeted is a fundamental pre-requisite. interventions are called for, to address long-standing • Maintaining economic growth, managing oil inequities across space and groups. revenues, promoting the diversification of the economy in favor of non-oil and private sector led In identifying areas for policy action, a few key prin- activities. While these will bear fruit in the me- ciples are critical to keep in mind: dium and long terms, the seeds must be sown now. • Implementing an effective and comprehensive sys- • Peace, security and economic growth are foun- tem of safety nets to address the multiple depriva- dational pre-requisites for development, without tions and vulnerabilities of the population, while 192 THE UNFULFILLED PROMISE OF OIL AND GROWTH redressing the human capital deficit and spatial for instance, investing in education access and inequities. quality will not bear fruit absent adequate pro- ductive employment opportunities that generate The analysis presented in this report thus far points appropriate returns on the labor market; to the myriad and deep-rooted challenges to pov- • Taking advantage of spillovers and complemen- erty reduction and broad-based prosperity in Iraq. tarities, such as spatially targeted public works More than in most countries, poverty and depriva- programs to generate employment and increase tion in Iraq are symptomatic of its daunting legacy of incomes which also have the potential to redress violence and fragility, of a skewed economy heavily the infrastructure deficit and build skills for fu- dependent on oil and on public intervention, and ture jobs; of decades of lost development. Addressing the twin • Finally, while there is clear scope for urgent ac- goals in this context and with continuing violence tion, other efforts are also needed that will re- will require a concerted and multi-sectoral approach, quire a medium to long term horizon. combining short-term and medium-term efforts. The current strategy towards poverty reduction de- pends overwhelmingly on public intervention, either Policy Imperatives through the creation of jobs in the public sector, or through the distribution of transfers, primarily uni- Establishing and maintaining peace and security versal food subsidies. This former approach is nei- across the nation is a pre-requisite for sustained and ther creating adequate jobs, nor promoting private healthy economic growth. Moreover, Iraq will need sector development, and is in fact posing significant to strengthen the natural links between growth and disincentives to labor force participation. The Public welfare through better management of oil resourc- Distribution System remains the only safety net that es, diversified development of the economy, en- adequately covers the poor, and while it does pro- abling the private sector and implementing a well- vide a base level of food security, childhood malnu- designed and comprehensive safety net system. An trition remains a concern, and the PDS is inadequate economic growth process that creates employment to address the multiple deprivations in human capi- and increases earnings, while maintaining incentives tal and labor market opportunities faced by the less for labor participation and investments in educa- well-off segments of the Iraqi population. If the sta- tion, is the least costly path to poverty reduction tus quo—the current relationship between, growth, and inclusion. At the same time, Iraq faces a large redistribution and poverty reduction—remains un- deficit in infrastructure, services and human capital, changed, it will take a sustained and significant effort which will need to be redressed. Moreover, targeted in terms of growth and/or redistribution to achieve interventions are called for, to address long-stand- meaningful poverty reduction in the medium term. ing inequities across space and groups. In identifying areas for policy action, a few key prin- Growth, Redistribution and Poverty ciples are critical to keep in mind: Reduction In order to understand the relationship between • Peace, security and economic growth are foun- consumption growth, redistribution and poverty re- dational pre-requisites for development, with- duction, we undertake a micro-simulation exercise out which welfare gains will be difficult to real- that illustrates the combinations of redistribution- ize and sustain; neutral consumption growth and redistribution • Sequencing will be very important to secure gains needed to attain certain hypothetical poverty-reduc- from reform, and supply and demand side con- tion targets. The roads leading to sustainable pov- straints will need to be relaxed simultaneously: erty reduction can be varied. For the purposes of this Policy Implications: Learning from the Past to Build a Better Future 193 FIGURE 199: Isopoverty Curves A. Standard B. Perfect targeting 30 7 6 25 5 20 Tax rate (alfa) Tax rate (alfa) 4 15 3 10 2 5 1 0 0 0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Annual growth rate (g) Annual growth rate (g) Reduction 75% Reduction 50% Reduction 25% Source: Authors’ calculations, IHSES 2012. exercise, we simplify the issue by thinking of poverty (α), while the horizontal axis measures the annual reduction as the result of either neutral per capita consumption growth rate between 2012 and 2030 income (or consumption) growth, or redistributive (g). Each point in the isopoverty curve corresponds policies, or a combination of both. Of course, reality to a combination of redistribution policy with a tax is much more complex: there might be no policy in- rate α and neutral growth at rate g needed to reduce strument that increases productivity proportionally poverty (25, 50 or 75 percent) from 2012 to 2030.67 for all the population, while redistributive policies may take a significant toll on efficiency, and hence on The curves in Figure 199 panel B are relatively incomes. However, it is still illustrative to know what ‘flatter’ than those of panel A, implying that the the effort is needed in terms of neutral consumption growth and simple non-distortionary redistributive policies to attain a certain poverty target. This in- 66 See Annex for details on the methodology. formation is useful at least to have an idea of the 67 The position of an isopoverty curve shows how easy or dif- ‘distance’ of the country from the poverty target in ficult is for Iraq to meet the poverty-reduction target: the terms of consumption growth and redistribution.66 closer to the origin an isopoverty curve lies, the less con- sumption growth and income transfers are required to reach the target. The isopoverty curves are negatively sloped, In Figure 199, we present three isopoverty curves, indicating that it is possible to substitute or trade off con- combinations of distribution-neutral growth rates of sumption growth for income redistribution, and convex, consumption and simple redistributive policies, cor- indicating that the marginal rate of substitution between responding to the goals of reducing poverty 25, 50 consumption growth and consumption redistribution is and 75 percent from the current levels over the next decreasing. The horizontal intercept indicates how much consumption growth the country needs in order to meet eighteen years. In all cases, we take the official pov- the poverty target with no additional consumption redistri- erty line in Iraq, and we use household per capita bution. The vertical intercept informs how much consump- consumption as the individual well-being measure. tion redistribution the country needs in order to achieve the The vertical axis measures the consumption tax rate poverty target with no growth in consumption. 194 THE UNFULFILLED PROMISE OF OIL AND GROWTH poverty-reduction impact of even a small transfer assuming no growth in consumption (g=0). For in- program is equivalent to that of many percentage stance, in order to halve poverty, Iraq would need points in accumulated consumption growth. For in- to transfer almost 2 per cent of non-poor individu- stance, an annual growth rate of almost 4 per cent als’ total consumption to the poor people under the between 2012 and 2030 is equal, in terms of pov- targeted scheme, if consumption were not to grow erty reduction, to an income transfer to poor people between 2012 and 2030. The same poverty-reduc- of more than two percentage points of the non- ing effect could be achieved with no income redis- poor individuals’ consumption (under the perfect tribution along with an average annual consumption scheme). In general, the curves are flatter for tar- growth rate of 3 per cent between 2012 and 2030. geted transfer policies than for a simple redistribu- tive policy, as the latter imply a greater fiscal effort Although the effort required through consump- to achieve the poverty-reduction goal. tion growth seems relatively small, over the last five years, consumption growth among the bottom 40 Columns (i) and (ii) of Table 47 show the intercepts percent of the distribution has hovered around 1 of the isopoverty curves with the horizontal and ver- percent, and has been below 2 percent for the popu- tical axis for every kind of transfer program. Halving lation as a whole. This has occurred despite signifi- poverty through a simple redistributive linear policy cant rates of GDP growth, on average, 7 percent per demands an incremental tax rate of 20 percent of annum between 2008 and 2012. In other words, everyone’s consumption and then distributing the if nothing else were to change, and Iraq wanted to revenues equally across the population, or alterna- halve poverty by 2030, an average annual rate of tively an annual rate of consumption growth of 1.12 GDP growth of roughly 7 percent would have to be percent. On the other hand, if Iraq were able to im- maintained over the next 15 years, which is a chal- plement a perfectly targeted system of transfers, the lenge for any country. If a more ambitious target fiscal effort to halve poverty would be a much small- were conceived, GDP would have to grow even er by taxing 2 percent of non-poor’s consumption. faster. The fundamental challenge is that the rela- tionship between GDP growth and consumption Column (iii) shows the amount of consumption trans- growth is weak; and strengthening the relationship ferred from non-poor individuals to poor individuals between the two, which will require a change in the as a percentage of the country’s total consumption, composition and drivers of growth, will translate into larger gains in terms of poverty reduction for the same rate of GDP growth. TABLE 47: Trading Off Consumption Growth and Redistribution for Poverty Reduction Strengthening the Relationship between Growth and Welfare Cost (g = 0) X axis Y axis (% of national (α = 0) (g = 0) consumption) The foundational pre-requisite to economic growth Standard and how and whether it translates into shared pros- 25 0.52 11.09 perity and poverty reduction is the establishment 50 1.12 20.38 and maintenance of peace and security. That being said, an inclusive and broad-based development pro- 75 2.12 30.63 cess can mitigate the risks of recurring violence, and Targeted Transfer vice versa, the deterioration of socio-economic con- 25 2.74 1.79 1.63 ditions can severely test a fragile peace. The gover- 50 3.15 2.05 1.86 norate of Nineveh, which had witnessed significant 75 3.99 2.34 2.13 sectarian violence since 2003, and where poverty Policy Implications: Learning from the Past to Build a Better Future 195 headcount rates have increased by 11 percentage to publish all revenues from the oil sector under the points between 2007 and 2012, is now reeling un- Extractive Industries Transparency Initiative (EITI). der an insurgency, and half a million people have Further, the 2012 Iraq Country Economic Memo- fled the conflict there. randum and Public Expenditure Review recommend the creation of a sovereign “parking fund” and a fis- A comprehensive policy framework for enhancing cal stabilization fund to minimize the impact of oil the welfare of Iraq’s citizens will require a three- revenue volatility on expenditure policy; improving pronged approach: the strategic orientation of public expenditure and enhancing the links between macro-fiscal policies 1. Establishing and maintaining peace and security and resource allocation; reorienting spending to- is a fundamental pre-requisite wards capital investment, and increasing the efficien- 2. Maintaining economic growth, managing oil cy of current capital expenditure. revenues, promoting the diversification of the economy in favor of non-oil and private sec- Moreover, while Iraq’s major challenges are me- tor led activities. While these will bear fruit in dium term, government response has been primar- the medium and long terms, the seeds must be ily short-term, pushing for a rapid expansion of oil sown now. production, without an accompanying vision for 3. Implementing an effective and comprehensive diversification of the economy. To address the mul- system of safety nets to address the multiple de- tiple development challenges facing the country, an privations and vulnerabilities of the population, expansion in job creation led by a diversified pri- while redressing the human capital deficit. vate sector will be essential, which in turn requires a favorable investment climate, the availability of fi- nance on competitive terms, and flexible labor mar- Economic Growth, Oil Management and ket mechanisms and institutions. Diversification Over and above the restoration of peace and security, While Iraq does have some of the key elements for a the achievement of significant and sustained welfare diversified economy—varied geography and climate, gains will be predicated on economic growth and natural agglomeration centers, and a fairly large do- the diversification of the economy away from oil. mestic market—these prospects have been eroded In Iraq, as in many other developing countries, the over time. That being said, there are opportunities heavy reliance on oil for growth, exports and gov- for an expanded role of the private sector in construc- ernment revenues also occurs within the context of a tion, banking, industry and tourism, if the appropriate narrow and non-inclusive power structure that privi- enabling environment is made available. Agriculture, leges ‘short-term private enrichment over longer- which has long been neglected, will require signifi- term collective welfare enhancement’.68 As a World cant investments in infrastructure, technology and Bank (2012) report highlights, the translation of extension services, but has the potential to protect natural resource rents into development gains will rural livelihoods and stimulate the local economy. In require credible intertemporal commitment on the time, these investments can lay the foundation for past of government to both extractive companies future competitiveness. Finally, the linkages between and its citizens, as well as a more inclusive process the oil sector and manufacturing can be strength- of decision making and public accountability so that ened, by encouraging, for instance, private producers resource rents are allocated towards investments in of light manufacturing or less sophisticated industrial service delivery and public good provision. machinery used in the oil industry. A first key step is greater oil revenue transparency, and 68 Rents to Riches? The Political Economy of Natural-Re- in 2008, the Government of Iraq indeed committed source Led Development. World Bank 2012. 196 THE UNFULFILLED PROMISE OF OIL AND GROWTH It will be imperative for the public sector not to with concerns about corruption and the lack of a crowd out the private sector in being able to com- level playing field to hinder investment and growth. pete for talent and business. Rather, the key chal- lenge for government is to put its significant rev- Barriers to firm entry and competition derive, as in enues to efficient use, making critical investments in many other countries in the region, from regulatory health, education, infrastructure and regional devel- barriers and discriminatory implementation and en- opment, and in creating a well-targeted and com- forcement of rules and regulations. These also cre- prehensive system of safety nets for the poor and ate opportunities for corruption and rent-seeking. vulnerable (World Bank, 2012, Iraq PER). The 2012 Investment Climate assessment recom- mends the clarification and simplification of existing In the latter context, direct distribution of resource rules, the strengthening of transparency in public rents to citizens has sometimes been advocated, procurement, and enhancing the accountability of which require the presence of mechanisms to iden- public institutions that deal with the private sector tify beneficiaries and guarantee payment proce- to promote rules over discretion in the implementa- dures. For instance, Mongolia’s Motherland Gift tion of policies and regulation. Fund pays dividends from its mining revenues to its citizens, while Alaska’s Permanent Fund distrib- Iraq’s large informal sector also poses an important utes USD 1 billion annually to its 600,000 citizens. challenge. Streamlining registration and licensing These types of Direct Dividend Payments (DDP) procedures will reduce the regulatory burden of en- are being increasingly advocated because improve- try into the formal sector; as will a reform of the tax ments in technology have made large scale transfers and regulatory regime to assure ease of entry and increasingly feasible, because these are potentially compliance for small firms. Moreover, expanding more effective at enhancing welfare and less distor- the access to finance and business support services tionary than other types of transfers, and because will enable informal firms to compete in the formal they could strengthen the citizen-state relationship economy. In the end, firms choose to formalize when coupled with taxation.69 However, given the based on a rational comparison of costs and ben- current social contract where the state is seen as the efits; and the formal economy can grow when these sole provider of jobs and subsidies, these types of benefits outweigh the costs. transfers may further weaken the relationship be- tween the Iraqi state and its citizens. Moreover, the Private sector growth has also been impeded by the success of these types of dividend distributions will lack of power and transport infrastructure; limited depend on the capacity of local markets and the local access to land and financing, and the absence of a economy to translate the cash influx into improved skilled labor force. There is a strong potential role welfare as opposed to a purely inflationary effect. for the private sector to play in partnership with the public sector to enhance infrastructure quality and access, but this will require reforms of the regula- Private Sector Led Job Creation tory environment to attract adequate investment. The creation of a large and diversified set of jobs Similarly, financial sector reforms will be needed to remains one of the fundamental development chal- expand access to competitive financing for private lenges for Iraq. Across the world, poverty reduction businesses. Easing the regulations governing land is most often driven by an expansion in employ- and facilitating registration and the use of property ment and increases in labor earnings; and the pri- as collateral will also be important. vate sector is usually the primary engine of job cre- ation. The fundamental constraint to private sector 69 Devarajan et al, The Case for Direct Transfers of Resource growth in many parts of the country is the environ- Revenues in Africa, Centre for Global Development, ment of insecurity and instability; which combines Working Paper 333, July 2012. Policy Implications: Learning from the Past to Build a Better Future 197 Agglomeration economies in Iraq are on the one natural resource development. Iraq has strong po- hand being stymied by the insecure local environ- tential for this approach as there are large resource ment in many parts of the country, and on the endowments adjacent to high poverty governorates. other, by the uneven implementation of regulations In addition, the country is historically aligned along and the severe infrastructure deficit. In part a conse- its rivers, while oil and gas fields are more dispersed quence of insecurity and violence, individuals in Iraq vis-à-vis the major population centers (Baghdad and seem to find it difficult to move across the country Mosul). A resource corridor approach can focus on to earn the appropriate returns on their character- mapping the oil and gas value chain into the demo- istics. Rather, the inability to migrate in search of graphic and geographic structure of Iraq in a way remunerative work, implies that otherwise equally that allows the domestic private sector to leverage capable individuals earn significantly less because of large scale investments by international companies where they live. and the government into downstream entrepre- neurship and job opportunities. The resource corri- Strengthening the education and vocational train- dor approach will also help Iraq capitalize on activi- ing system in partnership with the private sector ty spillovers from key trade routes, though of course and curricula reform to better reflect the technical the full benefits of land trade can only be realized skills sought by the private sector will be important when historically important routes to Syria and the in bridging the skills gap. As it stands now, the pay, Gulf are reopened. Where political risk is impeding hours and benefits premium earned by public sec- investment, guarantees may be an effective way to tor employees relative to those who work in the pri- induce potentially transformative investments from vate sector inhibits the ability of the latter to attract the private sector. and retain talent. In the short term, a combination of tax incentives, wage credits and on the job skills training and internship programs can encourage the Enhancing Inclusion and Social private sector to expand hiring; but in the medium Protection term, there is no substitute for easing firm entry and exit, promoting competition, guaranteeing a level Intergovernmental Fiscal Relations playing field and putting in place the enabling infra- The spatial variations in poverty outcomes points structure for growth. to the impact of the relationship between layers of government in Iraq. As noted in Chapter 1, the The overarching context for private sector develop- current model is asymmetric devolution: a unitary ment is fragility and conflict. Manifestations of this state with one semi-autonomous region. Gover- in terms of spatial fragmentation of the country and norates not in a region are mostly managed by and infrastructure deficiencies have already been high- accountable to, the central government. Iraq has lighted. It is important therefore to tap knowledge been moving in the direction of greater decentral- gained from private sector development projects ization, but the powers for governorates envisaged in other FCS. Given Iraq’s vast oil and gas endow- in a 2013 law have not been actualized. The spatial ment, the “resource corridor” approach holds par- differentials are a warning sign that simply giving ticular promise, especially as lessons emerge from a more powers to governorates will not necessarily resource corridor approach in Afghanistan. The key be equalizing. These differentials (especially those idea is to use a set of complementary investments related to service delivery) seem associated with by international and domestic private investors and variation in capacity at the governorate level, e.g. the government to manage natural resource devel- for the historically weak and disempowered gov- opment in a way that generates spillover activities. ernorates in southern Iraq. On the other hand, Of course, parallel interventions are necessary to the central government also lacks the instruments maintain security and promote an inclusive mode of to play an equalizing role. While the PDS is an 198 THE UNFULFILLED PROMISE OF OIL AND GROWTH effective national subsistence program, there is no Redressing the Human Capital Deficit analogous instrument capable of targeting com- As a consequence of decades of violence and insta- mon outcomes in terms of service delivery or trans- bility, Iraq has suffered from a significant depletion fers. Iraq does not have a tax and transfer system of its human capital stock, and with it, the loss of of the depth needed to protect the poor, nor has cadres of skilled teachers, academics and medical it been able to address capacity deficiencies at the professionals. Moreover, the deterioration and de- front-line service level. struction of education infrastructure, the financial constraints faced by many households, the lack of Although prospects for implementation of the interest in education expressed by many, and the 2013 provincial powers law are uncertain, there limited returns on the labor market have meant are a number of tracks along which the layers of many Iraqi children do not go beyond primary government could be strengthened and coordina- schooling. This represents a significant erosion of tion between them improved. First, “bottom-up” Iraq’s once promising human capital endowment. accountability can be improved by providing citi- The health sector has not fared much better: insti- zens with more relevant information about budget tutional arrangements were directly affected by the flows from allocation to facility level, so that they sanctions regime, and the system, overwhelmed by can have a better understanding of the relationship the ensuing violence, remains in crisis. Despite uni- between funding and needs. Second, “top-down” versal food distribution, a third of children belong- accountability can be strengthened by a joint com- ing to poor families are born stunted. Other aspects mitment from the central government, governor- of service delivery remain hampered, with less than ates, and governorate councils to address the most half the population covered by garbage collection glaring gaps on social and public services; this and sewage services, and very few rural households. should include a consensus on targets and budgets needed to achieve them. Third, the central govern- Significant investments are needed to invest in ment should revamp its current capacity-building health, education and service infrastructure to guar- efforts to make them more operational, for example antee a basic level of access for all Iraqi citizens, and by linking capacity building directly to spending these need to be staffed by qualified personnel. The activities or the development of targets; too often, expressed lack of interest of many children in con- capacity building has taken the form of training or tinuing their education is worrying and suggests workshops delinked from day-to-day tasks. Finally, that education quality may have deteriorated, that existing initiatives to provide more funding to the curriculum may need to be revised to be relevant to governorates should be reassessed for consistency the needs of today, and that investments are needed with spending capacity. There is a clear risk in the in teacher training and improved teaching methods. current “petrodollar” distribution to the governor- ates (US $1 per barrel produced in a governorate, A more nuanced approach will need to be put in with a proposed increase to US$5) that funding will place to ensure girls enroll in school and stay in outstrip capacity to spend. It would be preferable to school. Over and above understandable concerns agree a set of spending assignments for all gover- about safety, the oft-cited ‘social reasons’ to not go norates based on capacity and need, and then have to school or to drop out suggest that norms about funding follow these assignments. At a later stage, the value of educating girls and concerns about their revenue assignments could be reconsidered when honor and reputation play a role in limiting girls’ the fiscal system is better developed. As expenditure education. Simply building a school will not ensure and revenue assignments should be aligned with ac- that girls go to school: there may be a need in ad- countability mechanisms, the assignments can be dition for gender-sensitive design of the schooling designed in conjunction with the implementation system including developing a cadre of qualified fe- of the provincial powers law. male teachers, separate toilets for girls and boys, and Policy Implications: Learning from the Past to Build a Better Future 199 where necessary, the provision of safe and reliable In terms of the impact of CCTs on improving child transportation. Absent an adequate cadre of trained health outcomes the evidence is mixed: they reduced female teachers, the skills of local educated woman the incidence of low child height for age or stunting could be developed to serve as supplementary teach- only in some countries and only among some popu- ers. For instance, the Community Support Program lations. Nicaragua’s Red de Proteccion Social, which (CSP) in rural Balochistan in Pakistan relaxed the increased by about 0.17 points the height-for-age Z educational requirement for teachers in govern- score for children younger than 5 years of age; and ment schools to ensure an adequate supply of fe- Colombia’s Familias en Acción, which improved the male teachers. These teachers were given additional Z scores of treated children younger than 2 years of training to help bridge the gap, and as a result, the age by 0.16 points, implying a 7–percentage point program raised girls’ enrolment rates (Kim, Alder- reduction in the probability of stunting are two ex- man and Orazem, 1998 and World Bank 2005). amples of success (Maluccio and Flores, 2005 and Attanasio et al., 2005). Other evaluations of pro- The fact that many children, especially boys, do not grams in Honduras, Ecuador and Nicaragua find no attend school or drop out because of inadequate fi- effects (Macours, Schady, and Vakis, 2008, Hoddi- nancial resources or to find work to help their fami- nott, 2008 and Paxson and Schady, 2008). lies, suggests that there may be a role for financial incentives conditional on school attendance, and on Finally, some CCT programs that have successfully completing different levels of education. Similarly, incentivized girls to stay in school have also either targeted interventions may also be needed to keep explicitly conditioned benefits on girls remaining girls in school and overcome norms that place little unmarried or have led to lower rates of early mar- value on girls’ education. A range of conditional riage. In 1994, Bangladesh introduced the Female cash transfer (CCT) programs have been found to Stipend Program (FSP), a conditional cash transfer have had significant and, in some cases, large effects program that gives a monthly stipend to female sec- on school enrollment and attendance; and to some ondary school students contingent on maintaining extent, increases in the use of preventive health a minimum attendance rate and test score, and re- services. maining unmarried. Over the next ten years, girls’ enrollment in secondary schools almost quadrupled, Mexico’s Opportunidades program, for instance, enabling Bangladesh to achieve gender parity in was effective in ensuring more children transitioned education. The Zomba Cash Transfer program in from primary to secondary school in rural areas, and Malawi, which offers cash transfers conditional on also appears to have positive spillovers, with school girls attending school is another good example. The enrollments increasing even among ineligible chil- evaluation found that adolescent girls who were not dren.70 Moreover, these types of CCT programs in school at the beginning of the program, and were may disproportionately benefit poorer households, who begin with lower educational attainment, face significant financial constraints, and potentially 70 Bobonis and Finan (2008) argue that the increase was a higher returns to schooling at the margin. Several result of peer effects—barely ineligible children in Opor- studies find this to be the case. In Cambodia, a pro- tunidades communities were more likely to enroll because gram that gave scholarships to girls between the last their eligible peers were in school. year of primary school and the first year of second- 71 In Mexico, Behrman, Sengupta, and Todd (2005) argue that ary school (the Japan Fund for Poverty Reduction) Oportunidades program effects are largest for children with the lowest propensities to enroll in school at baseline. Finally, increased enrollment by approximately 50 percent- Oosterbeek, Ponce, and Schady (2008) show that the BDH age points for girls from less well-off families, com- program in Ecuador had a significant effect on enrollment pared with 15 percentage points for girls in the rich- for children around the 20th percentile of the proxy means, est families (Filmer and Schady, 2008).71 but no effect among children around the 40th percentile. 200 THE UNFULFILLED PROMISE OF OIL AND GROWTH offered conditional cash transfers, were 40 percent interventions to bridge the gap in human capital less likely to marry after one year than girls in the and to ensure equitable access to basic services. control group which did not receive the transfers. Moreover, some parts of the country have suffered This indicates that in this setting, schooling did have from decades of neglect, and spatially targeted pov- a protective effect for girls who are at higher risk of erty reduction programs will be essential to address early marriage. These links between extending girls’ welfare and vulnerability. All of these must be in education and lower rates of early marriage may also place before a reform of the PDS is considered. imply that girls, when they get married, are better able to care for themselves and their children. Given A nationwide public works program, with spatially the positive correlation between early motherhood differentiated components has the potential to pro- and adverse child malnutrition outcomes in Iraq, this vide employment and incomes in the immediate may have spillover effects on the next generation. term, to rebuild local infrastructure and service de- livery and skills for future employment. Certain key However, investing in human development will on- design and implementation elements are critical for ly reap dividends in the medium term when it is ac- success. First, the wage must be set appropriately, companied by the generation of a large and diversi- low enough that the wage that does not crowd out fied set of productive job opportunities that rewards the private sector, and effectively self-targets those these investments in the labor market. The current who need supplementary income the most. While policy of absorbing less educated workers into the in rural areas, the focus of the program could be public sector and the expansion of jobs in the secu- to rebuild and restore local infrastructure, in urban rity sector have stalled the returns to education at areas, physical labor works could be supplemented primary level, and therefore, the incentives to invest by basic service delivery. Public works may include in further schooling. the creation, maintenance, or reconstruction of existing infrastructures, like roads, schools, health posts, sanitation improvements; environmental and Public Works and Spatially Targeted Programs agricultural projects such as, irrigation, soil conser- The Public Distribution System is currently the only vation and watershed development; cleaning roads safety net covering the poor, and therefore, has been and other public facilities; and social services includ- politically very difficult to reform. This is in large ing day care, food preparation and so on. part due to the absence of any other viable system of social protection, which has made large parts of In general, the high rate of joblessness in Iraq, es- the population very dependent on food rations, and pecially among the youth, and the lack of visible resistant to reform. Meanwhile, the continuing in- improvements in day to day life despite massive rev- stability and insecurity in the country have made it enues from oil, has the potential to further marginal- difficult to put in place alternative means of secur- ize segments of Iraqi society. In other post-conflict ing livelihoods, either through public intervention and fragile contexts, programs have deliberately or through the revival of local economic activity. As targeted young at-risk groups to limit the risks of a result, without putting other programs in place, their reverting to violence. The Government of any one-shot reform to the PDS including target- Liberia, in collaboration with an NGO, Action on ing, will have adverse consequences on welfare, and Armed Violence and the United Nations success- will be hard to implement. fully helped ex-combatants move into full-time farm work by offering them training, start-up capital and Therefore, there is need for a phased implementa- counselling. Men who participated in the program tion of a comprehensive safety net system which were more likely to spend more time farming, and includes a variety of programs to address differ- less time in illegal activities; however, the combina- ent dimensions of deprivation. These must include tion of capital and training was critical to generate Policy Implications: Learning from the Past to Build a Better Future 201 impact.72 An evaluation of Uganda’s Youth Oppor- of malnutrition and the greater dependence on the tunities Program, which offered cash transfers to PDS as a source of calories in these parts of the groups of youth to increase employment and reduce country, nutritional supplements and enhancing the conflict, generated a shift from agricultural work to- nutritional content of food rations, may be consid- wards skilled trades and strong increases in income.73 ered as well. In Iraq as well, similar programs could be designed The Government of Iraq is considering a transi- to specifically target young men who have the high- tion to `smart cards’ linked to biometric identifica- est rates of joblessness, and should prepare them tion to deliver PDS entitlements. If well-conceived for future employment in formal labor markets. As and implemented, the series of programs proposed part of the program, young people may be encour- above which involve a combination of investments aged or required to complete training in vocational in infrastructure and service delivery and individu- or professional skills or apprentice and intern with ally or spatially targeted cash transfers conditional businesses. Women must also be integrated into im- on health, education or work, could be integrated proving local service delivery, and offered training into a unified delivery system based on a universal as supplementary teachers and basic health service registry. This will be fundamental to track benefi- providers. The program will have to be simple and ciaries, ensure the systems of delivery are working, transparent in design to enable easy roll-out and and prevent corruption and leakages. However, this monitoring in a low-capacity context. It should not will not guarantee that benefits accrue to those who provide permanent employment but rather the op- need it most; this remains a critical design challenge. tion of temporary work when needed. An effective monitoring system, perhaps linked to the proposed transition to ‘smart cards’ for PDS entitlements, will 72 Can Employment Reduce Lawlessness and Rebellion? A be essential in this regard. Field Experiment with High-Risk Youth in a Fragile State,” May 2014, Christopher Blattman (Columbia University) and Jeannie Annan (International Rescue Committee). In areas where poverty rates are very high, espe- 73 Blattman, Christopher and Fiala, Nathan and Martinez, cially in rural parts of the South, universal programs Sebastian, Generating Skilled Self-Employment in De- would have enormous benefits. The poverty map veloping Countries: Experimental Evidence from Uganda exercise that is underway will help identify pockets (November 14, 2013). Quarterly Journal of Economics, of severe poverty where spatially targeted programs Forthcoming. Available at SSRN: http://ssrn.com/ab- can be implemented. Given the higher prevalence stract=2268552 or http://dx.doi.org/10.2139/ssrn.2268552. 1818 H Street, NW Washington, DC 20433