Tunisia’s Jobs Landscape Tunisia’s Jobs Landscape This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank 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. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Photography: Cover photo by Dennis Sylvester Hurd. Design and Layout: Circle Graphics, Inc. Additional material relating to this report can be found on the World Bank Mauritius website (www. worldbank.org/mauritius). The material includes a fact sheet. © 2022 International Bank for Reconstruction and Development / International Development Association or The World Bank Group 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 www.worldbankgroup.org  iii Contents EXECUTIVE SUMMARY........................................................................................................ 1 INTRODUCTION................................................................................................................... 5 CHAPTER 1  Economic Growth, Structural Transformation, and Employment................. 7 Highlights....................................................................................................................7 Growth, Poverty Reduction, and Job Creation.............................................................8 Economic Transformation and Sources of Growth.....................................................15 References Chapter 1.................................................................................................29 Annex Chapter 1........................................................................................................30 CHAPTER 2  Access to the Labor Market: A Spotlight on Women and Youth............... 31 Highlights..................................................................................................................31 Demographics and Projections...................................................................................32 Trends in Access to the Labor Market........................................................................37 Gender Gaps..............................................................................................................46 Constraints on Women’s Participation in the Labor Force..........................................69   Contextual Factors.................................................................................................70  Endowments...........................................................................................................81   Preferences and Choices..........................................................................................84 Youth.........................................................................................................................87 References Chapter 2...............................................................................................102 Annex Chapter 2......................................................................................................105 CHAPTER 3  Employment and Wage Outcomes........................................................... 117 Highlights................................................................................................................117 Public Sector, Formal, and Informal Employment.....................................................118 Wage Trends, Wage Gaps, and Returns to Education................................................129   Trends in Wages....................................................................................................135   Gender Wage Gaps...............................................................................................137   Wage Gaps Among Sectors...................................................................................142   Conditional Wage Gaps Between Public and Private Sector Workers....................144   Young University Graduates: Unemployment and the Public Sector    Wage Premium..................................................................................................144   Conditional Wage Gaps Between Formal and Informal Workers in    the Private Sector..............................................................................................149   Returns to Education and Other Correlates of Wages...........................................150 References Chapter 3...............................................................................................155 CHAPTER 4  Job Creation: Sectoral, Spatial, and Enterprise Transformation.............. 157 Highlights................................................................................................................157 Structural and Spatial Transformation......................................................................158 Enterprise Transformation and Productivity.............................................................170 References Chapter 4...............................................................................................180 Annex Chapter 4......................................................................................................181 ivContents List of Figures Trends in GDP per capita, Tunisia and Middle East and North Africa, Figure 1.1.  1990–2020........................................................................................................ 9 Impact of COVID-19 on Annual GDP Growth, Overall and by Broad Figure 1.2.  Sector, 2019–20................................................................................................ 9 Annualized Change in Employment, Labor Force, and Working-Age Figure 1.3. Population, by Education, 2006–17................................................................. 9 Employment Deficit, by Education, 2006–17................................................... 9 Figure 1.4.  Figure 1.5. Employment to Growth Elasticity, by Sector and Subperiod, 2006–17........ 10 Employment to Growth Elasticity, Tunisia and Comparator Countries, Figure 1.6.  and Average Among Middle-Income Countries, 2011–17............................ 10 Figure 1.7. Annual Employment Creation, by Sector and Subperiod, 2006–17.............. 11 Trends in the Poverty Headcount Ratio, Tunisia and Middle East Figure 1.8.  and North Africa ($1.90 Poverty Line), 2000–15............................................ 12 Trends in Poverty Headcount Rate Overall and by Area (National Figure 1.9.  Poverty Line), 2000–15................................................................................... 13 Figure 1.10. Trends in Inequality Overall and by Area (Gini Index), 2000-15.................. 13 Annualized Growth of Per Capita Consumption Expenditures Figure 1.11.  by Percentile, 2000–15................................................................................. 13 Employment Type: Distribution of Employed Population, by Quintile Figure 1.12.  of per Capita Household Expenditure, 2010 and 2015............................... 14 Figure 1.13. Trends in External Factors as a Share of GDP, 2000–19.............................. 16  rends in the Export Volume Index, Tunisia and Comparator Figure 1.14. T Countries, 2000–19....................................................................................... 16 Economic Complexity Index Ranking, Tunisia and Comparator Figure 1.15.  Countries, 1995–2018................................................................................... 16 Figure 1.16. Trends in Gross Fixed Capital Formation, by Sector, 2000–19.................... 17 Gross Fixed Capital Formation in Tunisia and Comparator Countries, Figure 1.17.  2019.............................................................................................................. 17 Trends in the Sectoral Distribution of Gross Fixed Capital Formation, Figure 1.18.  2000–14........................................................................................................ 18  rends in the Sectoral Distribution of Foreign Direct Investment, Figure 1.19. T 2005–19........................................................................................................ 18 Figure 1.20. Trends in the Current Account Balance as a Share of GDP, 2000–19.......... 18 Trends in General Government Debt and Budget Balance as a Figure 1.21.  Share of GDP, 2000–19................................................................................. 18 Figure 1.22. Factor Decomposition of GDP Growth, by Subperiod, 1990–2018............ 19 Factor Decomposition of GDP Growth with Human Capital, Tunisia Figure 1.23.  and Comparators, 201018............................................................................ 19 Contents v Trends in GDP, Total Factor Productivity, Labor Productivity, Figure 1.24.  and Capital per Worker, 2000–19................................................................. 20 Trends in the Incremental Capital Output Ratio, Tunisia and Figure 1.25.  Comparator Countries, by Subperiod, 2000–19.......................................... 20 Labor Productivity Gaps Overall and by Sector, Tunisia and Figure 1.26.  Comparator Countries, 2017........................................................................ 21 Figure 1.27. Trends in Output per Worker, 2000–19........................................................ 22  atio of Output per Worker in Tunisia vs. Middle East and North Africa, Figure 1.28. R 2000–19........................................................................................................ 22 Figure 1.29. Labor Productivity, by Sector, 2006 and 2017............................................. 22 Figure 1.30. Trends in the Sectoral Distribution of Employment, 2006–17..................... 22 Figure 1.31. Trends in the Sectoral Distribution of Value Added, 2006–17..................... 23 Sectoral Distribution of Employment and Value Added, Tunisia Figure 1.32.  and Middle-Income Countries, 2017............................................................ 24 Figure 1.33. Labor Productivity and Employment Intensity, by Sector, 2017.................. 25 Decomposition of Changes in per Capita Value Added, Figure 1.34.  by Subperiod, 2006–17................................................................................ 27 Annualized Change in Labor Productivity and Employment, Figure 1.35.  by Sector, 2006–17....................................................................................... 27  ecomposition of Changes in per Capita Value Added in Tunisia Figure 1.36. D and Comparator Countries, 2011–17........................................................... 28 Figure 1.37. Sectoral Contributions to Growth in Output per Worker, 2006–17............ 29 Figure 2.1. Total, Child and Old-age Dependency Ratios, 1971–2075............................ 34 Figure 2.2. Population Pyramid, by Five-year Age-group, 2020...................................... 34 Population Pyramid, by Five-year Age-group, 2040 (Medium Variant Figure 2.3.  Projection)....................................................................................................... 34 Figure 2.4. Literacy Rates, by Birth Cohort, 2015............................................................ 35 Figure 2.5. Educational Level, Distribution by Cohort, 2017........................................... 35 Educational Level, Distribution Among the Working-age Population, Figure 2.6.  2006 and 2017................................................................................................ 35 Figure 2.7. Mathematics Scores, Tunisia and Comparator Countries, Circa 2015........... 36 Figure 2.8. Science scores, Tunisia and Comparator Countries, Circa 2015.................... 36 Figure 2.9. PISA and TIMSS Test Scores, by Sex, 2011 and 2015.................................... 36 Figure 2.10. Labor Market Structure, Tunisia, 2017......................................................... 37 Distribution of Working-age Individuals, by Labor Status, Age-group, Figure 2.11.  Sex, Educational Level, and Residence, 2017.............................................. 41 Figure 2.12. Distribution of the Employed Population, by Broad Sector, 2006–17......... 42 Distribution of the Employed Population, by Type of Employment, Figure 2.13.  2006–17........................................................................................................ 43 Figure 2.14. Correlates of Employment, by Sex, 2017.................................................... 44 viContents Figure 2.15. Correlates of Employment, by Region (Coastal vs. Inland), 2017............... 45 Trends in Selected Labor Market Indicators, by Sex, Q1, 2019–Q1, Figure 2.16.  2021.............................................................................................................. 46 Women’s Labor Force Participation Rates (Ages 15+), Tunisia Figure 2.17.  and the Rest of the World, 1990–2019........................................................ 48 Figure 2.18. Labor Force Participation Rates, by Sex, 2006–17...................................... 49  emale Labor Force Participation Rates, by Region and Urban Figure 2.19. F and Rural Areas, 2017................................................................................... 51 The Role of Observable Characteristics of Women in Gaps in Women’s Figure 2.20.  Labor Force Participation Across Governorates, 2017................................ 51 Figure 2.21. Labor Force Participation Rates, by Sex and Age-group, 2006–17............. 52 Female Labor Force Participation Rates, by Cohort Over the Life Figure 2.22.  Cycle, 2006–17............................................................................................. 52 Female Labor Force Participation Rates, by Educational Level Over Figure 2.23.  Time and Over the Life Cycle, 2006–17....................................................... 53  emale Labor Force Participation Rates, by Marital Status Over Time Figure 2.24. F and Over the Life Cycle, 2006–17................................................................ 54 Figure 2.25. Reasons for Not Working Over the Life Cycle, by Sex, 2015...................... 55 Share of Inactive Women Reporting Household Duties as a Main Reason Figure 2.26.  for Not Working Over the Life Cycle, by Quintile of Household Expenditure, 2015........................................................................................ 55 Figure 2.27. Correlates of Labor Force Participation, by Sex, 2017................................ 58 Oaxaca-Blinder Decomposition of the Gender Gap in Labor Force Figure 2.28.  Participation and Counterfactual Labor Force Participation Rates, 2006–17........................................................................................................ 58 Figure 2.29. Employment-to-Population Ratios, by Sex, 2006–17................................... 59 Figure 2.30. Unemployment Rates, by Sex, 2006–17...................................................... 60 Figure 2.31. Employment-to-Population Ratios, by Sex and Age-group, 2006–17......... 63 Figure 2.32. Unemployment Rates, by Sex and Age-group, 2006–17............................. 63 Employment-to-Population Ratios, by Educational Level and Sex, Figure 2.33.  2006–17........................................................................................................ 64 Figure 2.34. Unemployment Rates, by Educational Level and Sex, 2006–17.................. 64 Figure 2.35. Employment Category Distribution, by Sex, 2006–17................................ 65 Figure 2.36. Sectoral Distribution of Wage Workers, by Sex, 2006–17........................... 66  ccupational Distribution of Wage Workers, by Sector and Sex, Figure 2.37. O 2006–17........................................................................................................ 67 Educational Level Distribution of Wage Workers, by Sector and Sex, Figure 2.38.  2006–17........................................................................................................ 70 Distribution of Wage and Nonwage Workers, by Sex and the Number Figure 2.39.  of Hours Worked per Week, 2015................................................................ 71 Contents vii Distribution of Women Wage Workers in the Public and Private Sectors, Figure 2.40.  by the Number of Hours Worked per Week, 2015...................................... 73 Framework for the Constraints on Women’s Labor Market Figure 2.41.  Participation.................................................................................................. 73 Women, Business, and the Law Ranking, Tunisia and other Middle East Figure 2.42.  and North Africa Countries.......................................................................... 74 Figure 2.43. Women, Business and the Law, by Domain................................................. 74  ultural Traditions and custom Assign Men and Women Figure 2.44. C Traditional Roles........................................................................................... 77 Figure 2.45. Correlates of More Gender Egalitarian Views............................................. 79 Self-reported Frequency of Sexual Harassment in the Neighborhood, Figure 2.46.  % of Adult Men and Women........................................................................ 81 Primary and Secondary Gross Enrollment Rates and the Gender Figure 2.47.  Parity Index................................................................................................... 81 Figure 2.48. Gender Gaps in the Ownership of Productive Assets, 2017....................... 82 Figure 2.49. Gender Gaps in Access to Finance.............................................................. 83 Figure 2.50. Gender Gaps in Unpaid Work, 2017............................................................ 84 Key Labor Market Indicators, by Age-group, Youth (Ages 15–24 Figure 2.51.  and 25–29) and Adults (Ages 30–64), 2006–17............................................ 88  orrelates of the Probability of Unemployment Among Youth, by Sex, Figure 2.52. C 2015.............................................................................................................. 92 Figure 2.53. Activity Status of Youth, by Age and Sex, 2015.......................................... 93 Youth NEET Rates by Age-group, Sex, Educational Level, Region of Figure 2.54.  Residence, and Quintile of per Capita Household Expenditure, 2015........ 94 Correlates of the Probability of Inclusion in NEET Among Youth, Figure 2.55.  by Sex and Educational Level, 2015............................................................. 95 The Main Reason for Being Out of the Labor Force and not Looking Figure 2.56.  for Jobs Among Youth, by Sex and Educational level, 2015....................... 96  nnualized Change in the Population, Labor Force, and Employment, Figure 2.57. A by Age-group, 2006–17................................................................................ 97 Change in the Number of University Graduates, the Employed, Figure 2.58.  and the Employed in High-end Jobs, Circa 2011–17................................... 98 Distribution of Graduates, by Field of Study and Academic Year, Figure 2.59.  2012/13–2017/18.......................................................................................... 98 Change in the Number of Wage Workers Employed in High-end Jobs, Figure 2.60.  by Occupation, 2012–17.............................................................................. 99  uration of School-to-work Transitions and Probability of Never Figure 2.61. D Transiting from School to Work, 2013.......................................................... 61 Figure 3.1. The Composition of Employment, 2019...................................................... 118 Informality Rates and the Contribution to Total Employment, Figure 3.2.  by Type of Employment, 2019..................................................................... 127 viiiContents The Distribution of Wage Employment, by the Formality Status Figure 3.3.  of Workers and Firms, 2019......................................................................... 127 Marginal Effect of Selected Covariates on the Probability of a Specific Figure 3.4.  Type of Employment, 2019.......................................................................... 131  rends in Real Monthly and Hourly Wages, Average and Median Values, Figure 3.5. T 2012–19........................................................................................................ 135 Trends in Real Average Monthly Wages, by Broad Industrial Sector, Figure 3.6.  2012–19........................................................................................................ 136 Figure 3.7. Trends in Real Average Monthly Wages, by Sector, 2012–19...................... 136 Trends in the Real Average Monthly Wage, by Educational Level, Figure 3.8.  2012–19........................................................................................................ 137 Unconditional Gender Differentials in Hourly Wages, by Quantile Figure 3.9.  and Sector, 2012 and 2019........................................................................... 138  axaca-Blinder Decomposition: Mean Gender Hourly Wage Figure 3.10. O Differential, by Sector and Characteristics, 2012–19................................. 140  axaca-Blinder Decomposition: Gender Hourly Wage Differential Figure 3.11. O at Selected Percentiles, Private Sector, 2012–19....................................... 141  axaca-Blinder decomposition: Gender Hourly Wage Differential Figure 3.12. O at Selected Percentiles, Public Sector, 201219.......................................... 143  robability Density and Cumulative Distribution Functions of Real Figure 3.13. P Monthly Wages, by Sector, 2019................................................................ 144  axaca-Blinder Decomposition: Mean Hourly Wage Differential, Figure 3.14. O Wage Workers in the Public and Private Sectors, 2012–19....................... 145  nconditional Mean Monthly Wage Gap, by Sector, 25–34 Age-group Figure 3.15. U with Tertiary Education, 2019..................................................................... 146  axaca-Blinder Decomposition: Mean Hourly Wage Differential Figure 3.16. O Between Wage Workers Ages 25–34 with Tertiary Education and Employed in the Public and Private Sector, 2019...................................... 146  outh Ages 25–34 with Tertiary Education, Employed in Public Figure 3.17. Y Administration and in the Private Sector, by Type of Degree, 2015......... 147  rofiles of Youth Ages 25–34 with Tertiary Education, Employed Figure 3.18. P Outside Public Administration, by Type of Employment and Quintile of Household per Capita Expenditure, 2015............................................. 147  hare of Unemployed and Inactive Youth Ages 25–34 with Tertiary Figure 3.19. S Education, by Quintile of Household per capita Expenditure, 2015......... 148  istribution of Unemployed and Inactive Youth Ages 25–34, with Figure 3.20. D Tertiary Education, by Quintile of Household per Capita Expenditure and Relation to the Household Head, 2015............................................... 148  axaca-Blinder Decomposition: Mean Hourly Wage Differential Figure 3.21. O Between Formal and Informal Wage Workers, Private Sector, 2019........ 150 Returns to Education, Wage Workers Ages 15–64, 2012, 2015, Figure 3.22.  and 2019..................................................................................................... 151 Contents ix Returns to Education, by Sector and Sex, Wage Workers Ages 15–64, Figure 3.23.  2012–19...................................................................................................... 152 Returns to Education Among Formal and Informal Wage Workers Figure 3.24.  in the Private Sector, Ages 15–64, by Sex, 2019....................................... 153 Figure 3.25. Correlates of Hourly Wages, Wage Workers Ages 15–64, 2019............... 154 Figure 4.1. Changes in Employment and Employment Shares, by Sector, 2006–17..... 159 Employment Levels and Employment Growth, by Secondary Figure 4.2.  and Tertiary Subsectors, 2006–17................................................................ 159 Figure 4.3. Sectoral Composition of Exports, Tunisia, 2006 and 2018.......................... 161 Figure 4.4. Distribution of Region-Level Employment, by Sector, 2006 and 2017........ 161 Figure 4.5. Share of Sectoral Employment, by Region, 2017........................................ 162 Trends in Employment, Employment Shares, and Growth Rates, Figure 4.6.  by Region, 2006–17...................................................................................... 163 Effect of Geographical Location on the Probability of Working Figure 4.7.  in Different Types of Job, Marginal Effects.................................................. 164 Figure 4.8. Distribution and Growth Rate of Registered Firms, by Region, 2003–19... 166 Figure 4.9. Trends in the Distribution of registered Private Sector Firms, 2003–19..... 172 Figure 4.10. Distribution of Wage and Overall Employment, by Firm Size, 2019......... 172  istribution of Registered Firms with 100 Formal Wage Workers Figure 4.11. D or More, by Sector, 2019............................................................................ 173  istribution of registered Firms and Formal Employment, by Regime Figure 4.12. D (Onshore/Offshore) and Size of Firms, 2019.............................................. 173 Change in the Contribution to Formal Wage Employment Creation, Figure 4-13.  by Size Among Registered Firms, 2011–19............................................... 174 Share of Registered Firms Entering and Exiting, by Size and Year, Figure 4.14.  2003–19...................................................................................................... 175 Figure 4.15. Correlation Between Measures of Productivity and Firm Size, 2020........ 176 Figure 4.16. Correlation Between Measures of Productivity and Firm Age, 2020........ 177 Cumulative Distribution Functions of Sales per Worker and Value Figure 4.17.  Added per Worker Over Time, 2013 and 2020......................................... 178 Share of Firms Reporting Various Business Environment Constraints Figure 4.18.  as Major or Severe, 2013 and 2020............................................................ 179 Share of Firms Investing in Human and Physical Capital and Figure 4.19.  Innovating, 2013 and 2020......................................................................... 180 Figure 1.A.1. Trends in Employment, by Sector, 2006–17............................................... 30 Figure 1.A.2. Sectoral Contributions to Employment Growth, 2006–17......................... 30  etailed Oaxaca-Blinder Decomposition of the Gender Gap Figure A 2.1. D in Labor Force Participation, by Year, 2006–17........................................ 112 Figure A 2.2. Sectoral Distribution of Unpaid Family Workers, by Sex, 2006–17......... 113  ducational Level Distribution of Employers and Own-Account Figure A 2.3. E Workers, by Sex, 2006–17........................................................................ 114 xContents Educational Level Distribution of Unpaid Family Workers, by Sex, Figure A 2.4.  2006–17..................................................................................................... 115 Unemployment Rates Among Youth, by Year, Age-group, Figure A 2.5.  Educational Level, and Region, 2006–17.................................................. 116 Figure B 1.1.1. Decomposition of per Capita GDP Growth............................................. 26 Female labor Force Participation Rates, by Marital Status Figure B 2.3.1.  and Quintile of Household Consumption Expenditure, 2015................. 56 Sectoral Distribution of Employers and Own-Account Workers, Figure B 2.2.  by Sex, 2006–17.......................................................................................... 72 Figure B 3.1.1. Trends in the Number of Civil Servants, by Category, 2011–17........... 119 Figure B 3.1.2. Distribution of Civil Servants, by Sex and Category, 2017.................... 121 List of Maps Map 2.1. Female Labor Force Participation Rates, by Governorate, 2017..................... 50 Map 2.2. Unemployment Rates of Women, by Governorate, 2017................................. 61  ensity of Registered Firms (Number of Firms per 1,000 People), Map 4.1. D by Governorate, 2019..................................................................................... 167 Map 4.2. Distribution of Registered Firms, by Size and Delegation, 2019.................... 168 Map 4.3. Poverty Headcount Ratios, by Delegation, 2015............................................ 171 List of Tables Table 1.1. Average Annual GDP per Capita Growth Rates by Period, 1981–2019........... 8 Table 1.2. Key Labor Market Indicators, 2006–17............................................................ 11  nnual Growth and Contribution to GDP Growth, by Expenditure Table 1.3. A Category and Subperiod, 2000–19.................................................................. 15  istribution of the Population, by Region, Urban or Rural Area, Table 2.1. D and Share of Urban Population, 2006 and 2017.............................................. 34  ey Labor Market Indicators, by Sex, Age-Group, Educational Attainment, Table 2.2. K and Urban or Rural Location, 2006–17............................................................. 38 Table 2.3. Key Labor Market Indicators, 2006–17................................................................ 39  rends in Employment, by Industry, 2006, 2011, and 2017............................. 42 Table 2.4. T Table 2.5. Trends in Employment, by Occupation, 2006, 2011, and 2017....................... 44  abor Force Participation Rate and Unemployment Rate of Youth by Table 2.6. L Age-group, Sex, Educational Level, Decile of Household per Capita Expenditures, Georgraphical Area, and Profiles of Youth, by Age-group, 2017......................................................................................... 90 Contents xi Distribution of Public Sector, Formal and Informal Workers by Individual Table 3.1.  and Household Characteristics, 2019............................................................ 122 Distribution of Public Sector, Formal, and Informal Workers, Table 3.2.  by Job Characteristics, 2019.......................................................................... 125 Informal Employment, by Type and Contribution and by Individual Table 3.3.  and Household Characteristics, 2019............................................................ 128  hare of Informal Employment, by Type and Contribution and by Job Table 3.4. S Characteristics, 2019...................................................................................... 130 Table 4.1. Trends in Migration Balances, by Region, 1989–2014................................... 166 Table 4.2. Distribution of Regional-Level registered Firms, by Industry, 2019.............. 170 Table 4.3. Transition Matrices of Formal Firms Across Employment Size...................... 175 Annualized Growth Rate in Average Productivity, by Type of Firm Table 4.4.  and Productivity Measure, 2013–20............................................................... 178 Table A 2.1. Tunisia Snapshot, Women, Business and the Law 2021............................. 106 Childcare Centers and Preprimary Schools (Public and Private), Table A 2.2.  by Governorate........................................................................................... 107  onthly and Registration Fees and Opening Days/Hours of Surveyed Table A 2.3. M Private Day-Care Centers, by Governorate, April 2021............................. 108 Table A 2.4. Main Active Labor Market Policies for Youth............................................. 109 Estimates of Firm-Level Characteristics and Measures of Productivity, Table A 4.1.  2013 and 2020............................................................................................ 181 Correlation Between Educational Level of Heads and Spouses, Table B 2.3.1.  by Quintile of Household Consumption Expenditure, 2015..................... 57 Married Women Employed, by Type of Wage Employment and Table B 2.3.2.  Quintile of Household Consumption Expenditure, 2015.......................... 57 List of Boxes Box 2.1. Definitions of Key Labor Market Concepts........................................................ 33 Box 2.2. Digital Labor Platforms....................................................................................... 47 Women’s Labor Market Participation Along the Household Welfare Box 2.3.  Distribution......................................................................................................... 56 Box 2.4. Internal Migration and Two Secondary Cities in Tunisia.................................... 62 Box 2.5. Gender Gaps in Self-Employment...................................................................... 71 Box 2.6. Child Day-Care Centers and Preprimary Schools in Tunisia............................... 86 Box 3.1. Civil Service: Hiring and Compensation Mechanisms...................................... 119 Box 4.1. State-Owned Enterprises................................................................................. 171  xiii Acknowledgments This report was prepared by Marco Ranzani (Economist, EMNPV), with the support of Isis Gaddis (Senior Economist, HGNDR) in the analysis of constraints to women’ labor market participation. Giuseppe Grasso (STC, EMNPV) and Dan Pavelesku (ETC, EMNPV) provided excellent research and analytical support. The team would like to thank Nancy Lozano (Senior Economist, SMNDR) and Mahdi Barouni (Senior Economist, HMNSP) for sharing their work and views on internal migration, Michael Drabble (Senior Economist, HEAED) for the work concerning the role of technical and vocational education and training, and Gladys Lopez-Acevedo (Lead Economist, EMNPV) for insightful comments. The team is extremely grateful for the support and collaboration of the Tunisia National Institute of Statistics, in particular Adnen Lassoued (General Director, INS), Nadia Touihri (Head of the Demographic and Social Statistics Department, INS), Sofiene Derbali (Director ENPE, INS), Yamen Hlel (Former Director ENPE, INS), Fadia Bougacha (Deputy Director ENPE, INS), and Mohamed Salah Traidi (Deputy Director ENPE, INS). The report was prepared under the guidance of Johannes Hoogeveen (Practice Manager, EMNPV), Jesko Hentschel (Country Director, MNC01), Tony Verheijen (Country Manager at CN review stage, MNCTN), Alexandre Arrobbio (Country Manager, MNCTN). The team would like to express its grati- tude to Lantoniaina Ramanankasina (Program Assistant, EMNPV) for the assistance provided during the preparation of the report. The report was peer reviewed by Daniel Lederman (Deputy Chief Economist, MNACE) and Federica Saliola (Lead Economist, HSPJB).  1 EXECUTIVE SUMMARY I n the aftermath of the revolution, Tunisia embarked on performance of the labor market to be able to single out a complex political transition that has been marked the key policy levers that need to be pulled. by setbacks and is yet to be completed, but has also allowed the country to be celebrated as the only demo- This report argues that the main driver of the sluggish cratic success story of the 2011 Arab Spring. These events employment performance is low-grade economic growth, brought about a change in economic policy as well. To which has been a constant feature of the decade fol- accommodate social demands, which, together with the lowing the 2011 revolution. The high employment-to- desire for political freedom, had sparked the uprising, growth elasticity observed in the postrevolution period, economic policies became more inclusive and consensus well above the average in middle-income countries, indi- driven. Public sector recruitment was expanded, and public cates that a slightly higher economic growth rate would wages were raised, while public transfers, including the have generated an equally higher rate of employment Programme National d’Aide aux Familles Nécessiteuses creation. It is worth noting though that about 20 percent (National Program of Assistance to Needy Families) and of the net employment added over the period 2011–17 is access to health insurance at reduced prices, were rapidly ascribable to the expansion of employment in the public scaled up. Yet, the engine of economic growth started to sector as well as in health care and education services, lose steam and has, ever since, been slow compared with and therefore it might not be a sustainable path in the income peers, not least because of increased uncertainty, medium term. partly a consequence of security incidents that took a toll on tourist arrivals. The export-oriented model based The study pinpoints several important stylized facts, on low-technology manufacturing and tourism-related which are briefly summarized below and developed in the activities that had been the main driver of the economy rest of the overview. A full analysis may be found in the before the revolution faced headwinds. The lack of prog- main report. First, fewer than 1 working-age individual ress along the path of structural reforms contributed to in 2 actively participates in the labor market, that is, is deterioration in the business environment, which became either employed or looking for a job. Tunisia’s human less conducive to investments. The geographical inequali- capital is thus largely underutilized, and the public invest- ties between rural and urban areas as well as between ments in education that have led to considerable improve- inland and coastal regions have persisted. Labor market ments in education in past decades are not carrying over outcomes have sometimes been sluggish. Thus, labor into employment opportunities. Two groups in particular force participation rates are strikingly low, particularly stand out because of their low participation and employ- among women. Employment creation is meager; univer- ment rates: women and youth. sity graduates continue to face high unemployment rates; and a large share of workers are employed informally. In In the case of women, despite some improvements spear- parallel, increasing public expenditures, driven primarily headed by youngsters with tertiary education, partici­ by a rising wage bill, pushed up the fiscal deficit, which, pation remains, on average, extremely low. Weak labor combined with an expanding current account deficit, has demand, assigned gender roles, and the limited availability highlighted the unsustainability of the economic develop- of affordable childcare services are plausible drivers ment model. Then came the COVID-19 pandemic, which of the persistently low labor force participation among has worsened the economic outlook and exacerbated women. In addition, a sizable gender wage gap in the pri- existing imbalances. vate sector that effectively translates among women into the equivalent of almost three months of free labor per year Today, Tunisia is faced by limited economic growth, fiscal contributes to the low participation rates among women. and current account deficits, and labor market outcomes Indeed, a large wage difference per hour worked between that are unsatisfactory for the majority of the popula- men and women might provide an economic incentive, in tion and that are nurturing a sense of frustration. It is the context of household bargaining between spouses, for therefore important to identify the culprits of the subpar wives to bear most of the household burden in housework 2 Tunisia’s Jobs Landscape and family care while their husbands work, thus reinforc- education a premium of 26.1 percent relative to secondary ing assigned gender roles. In the case of youth, over the education. In addition, returns to tertiary education are past decade, unemployment has been a steady and seri- considerably higher in the public sector and have increased ous issue among university graduates. The sluggish creation over time, while they have declined in the private sector. of high-end jobs is one of the main reasons for the high This raises a question about the sustainability of wage unemployment rate among youth with tertiary educa- growth in the public sector. tional attainment, together with a skills mismatch as the curricula selected by many youth are not in line with pri- Except for the low participation rates and gender gaps, the vate sector demand, but are rather more suitable to the evidence identifies limited distortions in the labor market profile of civil servants. More importantly, the large wage and high employment-to-growth elasticity. The key issue gap between university graduates employed in the public to address in seeking to foster job creation is therefore why sector and those employed in the private sector is almost economic growth has been so low over the past decade. entirely attributable to youth’s characteristics. A young The answer is not trivial, and multiple factors may be Tunisian who holds a university degree and is employed in in play. Most of the recent economic growth has arisen the public sector does not earn, on average, a higher salary because of increases in employment; little has been asso- relative to a youth with the same characteristics working ciated with labor productivity growth. The modest gains in the private sector. Yet, public sector jobs are associated in labor productivity have been largely attributable to the with additional benefits, such as job security, guaranteed movement of labor from lower than average to higher salary increases, allowances, a wide range of annual leave than average productivity sectors, as opposed to growth options, long maternity leave, and flexible working hours, in labor productivity within sectors. The study advances that can make them more attractive, particularly among two complementary hypotheses linked to fiscal and regu- women. In addition, many unemployed university gradu- latory policies. First, the high and rising fiscal and current ates can afford to wait while living with their parents. account deficit generated by the expansionary fiscal policy Moreover, active labor market policies consist of wage sub- in the aftermath of the revolution and by the decline in sidies that provide temporary employment opportunities to exports and continued increase in imports, respectively, beneficiaries at the cost of significant deadweight loss and has increased the cost of capital and contributed to a reduc- substitution effects, but do not lead to more job opportu- tion in investments, together with a deterioration in the nities in the long term. business environment. Second, despite high entry and exit rates, particularly among small firms, firms are not growing Second, a sizable share of workers are employed infor- in size after entry. The lack of private sector dynamism can mally, that is, they do not have access to social insurance be blamed on several factors. A key element is the limited or they operate unincorporated businesses that are not market contestability. Politically connected private firms registered with the tax authorities or other formal public and state-owned enterprises (SOEs) do not respond to any accounting procedures. Among wage workers, informality logic of efficiency because they are shielded from compe- is more widespread among men, youth, and workers with tition thanks to direct support and financing guaranteed little education in rural areas and inland regions. However, by the state, the imposition of tariffs, limits on foreign while workers with such profiles face difficulties in access- direct investment, and price controls. Such effects are ing public sector jobs or formal jobs in the private sector not restricted to the markets in which advantaged firms and are not protected against the risks covered by social operate; they extend to upstream and downstream mar- insurance (such as health events, old age, unemployment, kets, further dampening productivity growth and employ- and disability), they do not suffer wage penalties. Most of ment creation. the wage differential between formal and informal wage workers in the private sector derives from differences in The analysis presented in the report takes advantage of workers’ and jobs’ characteristics. several data sources produced by the Tunisia National Institute of Statistics (INS) that include public use data Third, returns to education are sizable in Tunisia relative to files, restricted use data files, and reports published by the middle- and high-income countries. In 2019, workers with INS based on microenterprise surveys and the national primary education enjoyed a premium of about 12.6 per- business register. The analysis would not have been pos- cent per hour worked relative to workers with no schooling. sible without the data collection effort and the excellent Secondary education yielded an additional premium of support and collaboration of INS. The analysis of wages about 9.1 percent relative to primary education, and tertiary stands out as an example of collaboration and of how Executive Summary 3 important data production, analysis, and dissemination of specific factors that make transitions from school to are to the understanding of trends and patterns of labor work difficult, including the skills mismatch, the quality of market outcomes and, ultimately, of changes in living stan- education and training, labor regulations, and active labor dards. The study is a testament to the tireless work of the market policies, can support policy makers in prioritizing INS in collecting high-frequency survey data and repre- and tailoring actions aimed at reducing the number of indi- sents a plea to continue on the virtuous path of strengthening viduals not in education, employment, or training (NEET) the production and dissemination of data and statistics. and facilitating labor market entry and retention among More and more high-quality data, together with wide data university graduates. Labor force survey data, adminis- access, are key to informing evidence-based public debate trative data from technical and vocational education and and policy making. training and academic institutions, and from institutions and line ministries in charge of active labor market policies The report identifies some areas that merit further research. will be required to conduct such in-depth analyses. Modest An in-depth analysis of the link between the degree of improvements in the labor market participation of women product market contestability and the lack of firm-level and the persistent large gaps in educational attainment dynamism that appears to be a key driver of the meager call for attention to factors that might help raise women’s economic growth in the country can shed light on the engagement in the labor market, such as childcare services, policy levers required to promote the growth of firms assigned gender roles and cultural barriers, and preferences and job creation. This may also foster greater participa- for certain types of work. Labor force survey data, admin- tion and employment among women and youth. This will istrative data on childcare facilities, and the collection of ad require access to microdata from the national business hoc microdata on roles, preferences and cultural barriers register and firm-level surveys. Assessing the importance would inform this research agenda. 5 INTRODUCTION U nderstanding the link between economic transfor- the tireless work done by the INS with the collection of mation, growth and jobs is the key to economic high-frequency survey data and a plea to continue on the development, particularly in middle-income coun- virtuous path of strengthening production and dissemi- tries. First, economic growth is necessary to job creation as nation of data and statistics. More and high-quality data well as to raise labor income and ultimately living standards. together with wide data access are key to inform evidence- In middle-income countries at more advanced state of eco- based public debate and policy-making. nomic transformation, shifts of labor across sectors are less relevant to economic growth, whereas within-sector The report is organized in four chapters. Chapter  1 productivity gains are of paramount importance. More describes trends in growth, productivity, demography, individuals are typically working for a wage as opposed employment, and living standards to inform the analysis to be employed on their own-account or in the family busi- of labor supply and labor demand carried out in the chap- ness as contributing family workers. And workers are on ters that follow. The chapter starts by depicting aggregate average better educated and might have high reservation trends in economic growth and living standards of the wages, so low labor force participation and high unem- Tunisian population, the drivers of growth (e.g. remit- ployment rates are typically more important challenges tances and migration, FDI, exchange rate, productivity, in such context. While a thriving private sector is key to etc.), and broad structural changes in terms of job creation and labor productivity growth. Chapter 2 provides an generating more and better jobs, economic growth alone overview of the composition of the labor market and how is not sufficient unless the demand for the type of labor a it has changed over time, including demographics and labor country’s workforce can supply does not increase. More- force participation, employment and employment compo- over, for economic growth to translate into more inclusive sition in terms of type of job, industrial sector, occupation jobs, barriers to economic participation as well as distor- both at the aggregate level and for different population tions in the labor, credit, and product market that favor a groups based on gender, age, educational level, and geor- group of insiders, often politically connected, over others graphical location. It turns the spotlight on two groups shall be removed. that face particular difficulties in accessing the labor mar- ket, namely women and youth, and advances hypotheses The objective of this report is to provide a comprehensive regarding key barriers to their engagement in the labor jobs diagnostic that can inform policies to generate more, market. Chapter 3 shifts the focus to one of the most rele- better, and inclusive jobs. The report illustrates characteris- vant dimensions that characterize the Tunisian labor market, tics, constraints, and dynamics of the Tunisian labor market namely the distinction between public sector, formal and over the past 15 years and offers a description of demand informal employment. The chapter investigates how indi- and supply side dynamics that determine labor market vidual characteristics are correlated with the probability outcomes with attention to changes observed before and of working in different types of employment; it provides after the 2011 Jasmine revolution, geographical dispari- an overview of recent trends in wages and of conditional ties between coastal and inland regions, and labor market wage gaps along a number of dimensions (men/women, developments following the outbreak of the COVID-19 public/private, formal/informal employment); and it illus- pandemic. trates how wage workers with different characteristics, in particular different educational endowments, benefit The analysis would not have been possible without the from the labor market. Finally, building on the findings of data collection effort as well as the excellent support and Chapter 1, Chapter 4 examines recent trends in the pat- collaboration of INS. The analysis presented in the report terns of structural and spatial transformation along the takes advantage of several data sources produced by the employment and firm dimension. It provides an overview INS that include public use data files, restricted use data of the firm landscape in terms of size, industrial sector, files, as well as reports published by Tunisia National Insti- geographical area as well as recent trends in firms’ perfor- tute of Statistics (INS) based on microenterprise surveys mance, dynamics, labor decisions and capital investments, and the national business register. The study is a praise to as well as constraints and opportunities firms face. 7 CHAPTER 1 Economic Growth, Structural Transformation, and Employment HIGHLIGHTS ◾ Recent economic growth has been subpar as Tunisia’s economic development has recently shifted to a less sustainable model that has substituted investments and exports with domestic demand ◾ Poverty reduction has continued largely thanks to an expansion of public transfers ◾ Economic growth was driven primarily by gains in labor productivity before the revolution and by employment creation thereafter ◾ In the years leading up to the revolution, gains in labor productivity occurred in monopolistic or non-contestable markets dominated by SOEs and in public administration thanks to rapidly rising public expenditures, i.e. wages ◾ In the aftermath of the revolution, although employment creation became the main driver of eco- nomic growth, it was insufficient to keep up with a growing labor force, particularly of university graduates ◾ Although structural transformation contributed to labor productivity growth between 2011 and 2017, it will not be able to support economic growth going forward unless capital and efficiency increase in the sectors with higher than average labor productivity that have attracted more workers 8 Tunisia’s Jobs Landscape Growth, Poverty Reduction, (−21.3 percent). The economy continued to suffer from the effects of the global downturn in the third and fourth quar- and Job Creation ters (−5.7 percent and −6.1 percent, respectively, compared with the same quarters in 2019). In 2020, GDP per capita T unisia’s growth has been historically on par with is estimated at $9,728 (measured in 2017 PPP). Estimates regional and income group peers. During 1981–2000, for the second quarter of 2021 indicate an increase of 16.2 economic growth averaged 4.2 percent a year in relative to the same quarter of 2020, driven by accommo- Tunisia, 3.8 percent in the Middle East and North Africa dation and food service activities, textiles, oil refining, and region (excluding high-income countries), and 4.1 percent construction. In Q3 2021, the economy posted a small in both lower- and upper-middle-income countries. After increase (0.3 percent) relative to the same quarter in 2020, accounting for differences in population growth, Tunisia’s as growth in most sectors faded. historical performance, estimated at 2.4 percent per year on average between 1981 and 2000, is superior to regional Modest economic growth has been insufficient to keep up (1.4 percent) and income group comparators (2.1 percent with the increase in the size of the labor force, particu- and 2.8 percent among lower- and upper-middle-income larly university graduates. In 2006–17, the economy cre- countries, respectively) (Table 1.1). ated employment at an annualized rate of 1.4 percent on average (Figure 1.3). Over the same period, the labor force Since the Jasmine revolution of January 2011, economic increased at a rate of 1.7 percent per year, and the number growth has weakened. Economic growth, measured by of individuals of working age rose by 1.2 percent per year. average annual growth in per capita gross domestic product Thus, Tunisia had an average net employment deficit of (GDP), has lost steam in comparison with the historical about 18,000 jobs a year. Aggregate numbers hide impor- trend of the country and with income group and regional tant differences by educational level. Employment creation comparators (see Table 1.1). The subpar growth that fol- for Tunisians with no schooling or a primary school certifi- lowed Tunisia’s graduation to the upper-middle-income cate was faster than their entry into the labor force, thereby group in 2010 pushed the country back to the lower-middle- contributing to a decrease in unemployment among indi- income group five years later. In 2019, GDP per capita viduals with low educational attainment (Figure 1.4). By is estimated at $10,756 (measured in 2017 purchasing contrast, employment among Tunisians with secondary power parity [PPP]) compared with a regional (excluding and, particularly, tertiary education was not sufficient to high-income countries) average of $10,172 (Figure 1.1). keep up with their growing number in the labor force; this affected university graduates disproportionately. The COVID-19 pandemic and economic downturn have further worsened the economic outlook. Tunisia’s The employment to growth elasticity has picked up since GDP declined by 8.8  percent in 2020, with the largest the revolution, largely thanks to manufacturing, tourism, reduction observed in the services sector, particularly in and other services. Estimated at 0.9 percent before the rev- accommodation and food service activities and in trans- olution, the annualized growth rate of employment acceler- port (Figure  1.2). Quarterly data of the National Insti- ated to 1.6 percent in 2011–17. The employment to growth tute of Statistics (Institut National de la Statistique; INS) elasticity, which measures the increase in employment for indicate that GDP took a hit in the second quarter of every 1 percentage point increase in GDP growth, rose from 2020, following the enforcement of a national lockdown 0.28  percentage points in 2006–11 to 0.89 in 2011–17 TABLE 1.1. Average Annual GDP per Capita Growth Rates by Period, 1981–2019 1981–90 1991–2000 2001–05 2006–10 1981–2010 2011–19 Tunisia 1.0 3.1 3.1 3.5 2.4 0.7 Lower-middle-income 1.0 1.0 4.1 4.2 2.1 3.7 Upper-middle-income 1.4 2.0 4.6 5.6 2.8 3.8 Middle East and North Africa 0.4 1.3 2.4 2.8 1.4 0.7 (excluding high income) Source: Based on data from the World Development Indicators, World Bank. Economic Growth, Structural Transformation, and Employment 9 FIGURE 1.1. Trends in GDP per Capita, Tunisia and FIGURE 1.2. Impact of COVID-19 on Annual GDP Middle East and North Africa, 1990–2020 Growth, Overall and by Broad Sector, 2019–20 Constant International $ (2017 PPP) 12,000 –8.8 Total 10,000 –6.3 Nontradable activities 8,000 –13.3 Tradable services 6,000 Industry (exlcuding –8.8 manufacturing) 4,000 –9.3 Manufacturing 2,000 4.4 Agriculture and shing – –15.0 –10.0 –5.0 0.0 5.0 10.0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 Percent Tunisia Middle East & North Africa (excluding high income) Source: Based on data of the World Development Indicators, World Bank and INS. FIGURE 1.3. Annualized Change in Employment, FIGURE 1.4. Employment Deficit, by Education, Labor Force, and Working-Age Population, by 2006–17 Education, 2006–17 Total (17,933) 1.4 Total 1.7 1.2 Tertiary (16,324) 4.2 Tertiary 5.7 5.4 Secondary (7,639) 1.6 Secondary 1.9 0.9 1.0 Primary 4,807 Primary 0.5 0.8 –2.2 None 1,215 None –2.4 –0.7 ) ) 0) ) – 0 00 0 0 00 00 00 00 00 ,0 ,0 5, –4.0 –2.0 0.0 2.0 4.0 6.0 8.0 0, 5, 0, 10 (5 (2 (1 (1 Percent Percent Employment Labor Force Working age population Source: Based on data from the Labor Force Survey (ENPE), INS. 10 Tunisia’s Jobs Landscape FIGURE 1.5. Employment to Growth Elasticity, FIGURE 1.6. Employment to Growth Elasticity, by Sector and Subperiod, 2006–17 Tunisia and Comparator Countries, and Average Among Middle-Income Countries, 2011–17 2.5 2.0 2.0 1.5 1.8 Elasticity 1.0 1.6 0.5 1.4 0.0 1.2 Elasticity –0.5 1.0 –1.0 0.8 Total Agriculture Manufacturing Mining and Energy Construction Trade Telecommunication and food services Financial services Pa/Education/Health Other services Accommodation 0.6 Transport and 0.4 Services 0.2 0.0 MICs Tunisia Jordan Albania Malaysia Romania Morocco Indonesia 2006–17 2006–11 2011/17 Source: Based on data from the Labor Force Survey (ENPE), INS; World Development Indicators, World Bank. (Figure 1.5). Thus, 1 percentage point of growth was asso- Growth of the working-age population has weakened, and ciated with an increase in employment by 0.9 percentage employment and labor force participation rates remain low. points. This elasticity is higher compared with Tunisia’s his- Between 2006 and 2017, the working-age population (ages torical figures (0.61 and 0.57 percentage points in 1980–89 15 and above) increased on average by 1.3 percent a year, and 1990–99, respectively), with the subregional (North from about 7.5 million to 8.7 million, and the labor force Africa) income group (0.35 percentage points in 2011–17), grew by 1.6 percent per year (Table 1.2). Both increased at and with global estimates (0.51 and 0.3 percentage points a lower rate beginning in 2011, and the labor force partici- in 1999–2003, respectively), as well as with estimates for pation rate declined slightly, from 47.2 percent in 2011 to comparator countries (except Jordan) (Figure 1.6).1 Manu- 47 percent in 2017. The employment ratio hovered around facturing was the sector with the highest employment elas- 40  percent over the entire period; only 4 individuals of ticity since 2011, followed by other services (including real working age in 10 had a job. In 2017, the employed popu- estate, business support services, and social and cultural lation was estimated at almost 3.5 million. Unemployment activities), accommodation and food services, and financial decreased both in level and rate beginning in 2011 among services (Figure 1.5). the overall population and among youth. At 35  percent in 2017, the youth unemployment rate was considerable. Construction before the revolution and trade, public admin- Most of the employed population was working for a wage. istration, education, and health services thereafter posted The transition toward wage employment continued, with the largest increases in employment. The annualized rate an annual growth rate of 2.2 percent between 2006 and of employment creation was considerable in construction 2017 and a share as high as 75 percent in 2017. (4.5 percent) in 2006–11, whereas banking and insurance services (5.2  percent) and accommodation and food ser- Job-related international migration increased. According to United Nations data, international migration from Tunisia vices (3.2  percent) posted the highest growth rates in increased by 75  percent over the past 30  years. As of 2011–17. However, thanks to the initially large size, public 2019, over 810,000 Tunisians were living abroad, largely administration, health, and education services (+11,700 per in Western Europe (France, Germany, and Italy).2 Of the year), trade (+11,550 per year), and manufacturing (almost 640,000 Tunisians living in countries of the Organisation +10,000 per year) contributed over 60 percent of employ- ment creation since the revolution (Figure 1.7). 2 Trends in International Migrant Stock (dashboard), Population Division,   Department of Economic and Social Affairs, United Nations, New York, 1 Historical estimates for Tunisia are from Mouelhi and Ghazali (2014),   http://www.un.org/en/development/desa/population/publications/migration/ whereas estimates for North Africa are from Kapsos (2005). migrant-stock-2013.shtml. Economic Growth, Structural Transformation, and Employment 11 FIGURE 1.7. Annual Employment Creation, by Sector and Subperiod, 2006–17 Total Other services PA/Education/Health services Financial services Accommodation and food services Transports & Telecommunication Trade Construction Mining and Energy Other manufacturing 2006–11 2011–17 Textile Agriculture (20,000) – 20,000 40,000 60,000 Change in employment Source: Based on data from the Labor Force Survey (ENPE), INS. TABLE 1.2. Key Labor Market Indicators, 2006–17 2006–11 2011–17 annualized annualized change change 2006 2008 2009 2011 2013 2015 2016 2017 (%) (%) Working-age population 7,525,883 7,807,036 7,931,938 8,146,651 8,315,665 8,480,590 8,580,953 8,694,333 1.6 1.1 (‘000s) Labor force (‘000s) 3,434,562 3,603,788 3,689,246 3,844,646 3,943,658 3,991,403 4,047,211 4,084,204 2.3 1.0 Labor force participation 45.6 46.2 46.5 47.2 47.4 47.1 47.2 47.0 0.7 −0.1 rate (%) Employment (‘000s) 3,004,893 3,155,349 3,198,925 3,139,771 3,315,283 3,386,337 3,417,581 3,458,104 0.9 1.6 Employment-to- 39.9 40.4 40.3 38.5 39.9 39.9 39.8 39.8 −0.7 0.6 population ratio (%) Unemployment (‘000s) 429,668 448,439 490,321 704,876 628,375 605,066 629,630 626,100 10.4 −2.0 Unemployment rate (%) 12.5 12.4 13.3 18.3 15.9 15.2 15.6 15.3 7.9 −2.9 Youth unemployment 177,709 181,578 186,082 266,475 214,907 179,626 195,917 194,076 8.4 −5.1 (‘000s) Youth unemployment 27.7 28.4 30.9 42.3 34.7 35.0 34.9 34.9 8.8 −3.2 rate (%) Wage employment (‘000s) 2,048,643 2,186,590 — 2,234,490 2,386,063 2,465,767 2,466,635 2,598,651 1.8 2.5 Wage employment (%) 68.2 69.3 — 71.2 72.0 72.8 72.2 75.2 0.9 0.9 Source: Based on data from the Labor Force Survey (ENPE), INS. 12 Tunisia’s Jobs Landscape for Economic Co-operation and Development (OECD), FIGURE 1.8. Trends in the Poverty Headcount Ratio, 24 percent had postsecondary or higher educational attain- Tunisia and Middle East and North Africa ($1.90 Poverty ment (DIOC database 2015/16).3 Over 70 percent of Tuni- Line), 2000–15 sian international migrants in 2009–14 left their country 7 to look for better job opportunities, and most of them 6 6 were youth ages 25–34. 5 3.7 3.8 Percent 4 3.2 3.4 Population projections indicate the need for more rapid 3 2.1 2 employment creation. Although accurate predictions of the 2 future size of the labor force are challenging, it is none- 1 0.2 theless a useful exercise, particularly to illustrate the need 0 for creating more jobs, especially good jobs that are 2000 2005 2010 2015 attractive to the increasingly well-educated labor force. MENA - $1.90 Tunisia - $1.90 Using population projections of the United Nations and Source: Based on data from PovcalNet, World Bank. assuming the rate of employment creation is equal to the rate observed between 2011 and 2017, employment-to- population ratios will remain at 40 percent or less until dynamics of consumption in 2005–15.4 Average consump- 2030 and increase up to 44 percent by 2040. Given the tion increased by 2.1 percent per year between 2000 and employment-to-growth elasticity estimated over the same 2015 and by 3.6 percent per year among households in period, the economic growth rate required to achieve such the bottom 40  percent of the consumption distribution employment creation would be 1.8  percent per year or (the bottom 40). Between 2010 and 2015, growth was higher. By contrast, to increase the employment ratio to even more pro-poor as the gap in the consumption growth 60 percent by 2040, GDP per capita would need to grow rate at the mean (0.9 percent) and among the bottom 40 at an annualized rate of 3  percent, Tunisia’s economic (4.4  percent) increased to 3.5  percentage points (Fig- performance before the revolution. This is purely a math- ure 1.11, panels a and b). After a modest increase between ematical exercise. Any changes in labor productivity and 2000 and 2005, inequality, as measured by the Gini index, in employment-to-growth elasticity can alter the magni- declined from 40.8 in 2005 to 36.5 in 2015, slightly above tude of the economic growth necessary to achieve specific the average of 35.4 in the Middle East and North Africa employment levels and ratios. region (Figure 1.10).5 Considerable progress in living standards was achieved National level estimates hide sizable geographical dis- within 15  years. Measured against the $1.90-a-day per parities in living standards. Although poverty reduction capita line, the poverty headcount ratio declined from occurred in both urban and rural areas, the pace of reduc- 6 percent in 2000 to 0.2 percent in 2015, while poverty in tion in rural areas picked up only between 2010 and the Middle East and North Africa region fell from 3.7 per- 2015. In 2015, about 26 percent of the population in cent to 2.1 percent during the first decade and then bounced rural areas was poor, compared with 10 percent in urban back to 3.8 percent in 2015, largely because of conflicts areas (see Figure  1.9). Regional gaps are sizable, too. in Syria and Yemen (World Bank 2020) (Figure 1.8). Mea- A recently completed poverty map indicates that poverty sured against the national poverty line, the poverty head- is high in the Center-West (30.8 percent) and North-West count ratio was estimated at 15.2 percent in 2015, down (28.4 percent) regions of Tunisia (World Bank and INS from 25.4  percent 10  years earlier (Figure  1.9). Health 2020). Although the incidence in the coastal regions— and education outcomes improved: the human develop- ment index (HDI) increased by 14 percent between 2000 and 2019, and the country ranked 94th (UNDP 2020). Economic growth was pro-poor in 2000–15 thanks to the 4  Between 2000 and 2005, no difference is detected between the growth rate of mean consumption and of consumption among the bottom 40 percent of the consumption distribution (the bottom 40). 5 The Middle East and North Africa average is based on the latest avail-   able data for the following economies: Algeria (2011), Djibouti (2013), the Arab Republic of Egypt (2015), the Islamic Republic of Iran (2015),   3 DIOC (Database on Immigrants in OECD and non-OECD Countries: Iraq (2012), Israel (2014), Jordan (2010), Morocco (2013), Syria (2004), DIOC) database 2015/16, OECD, https://www.oecd.org/els/mig/dioc.htm. Tunisia (2015), West Bank and Gaza (2011), and Yemen (2014). Economic Growth, Structural Transformation, and Employment 13 FIGURE 1.9. Trends in Poverty Headcount Rate FIGURE 1.10. Trends in Inequality Overall and by Overall and by Area (National Poverty Line), 2000–15 Area (Gini Index), 2000–15 45 40.4 40.8 38.5 40.4 45 40 38.8 38.2 38.5 35.9 36.5 36 40 35.7 36.4 34.5 35 35 35 31.9 30 26 25.4 30 Percent 23.1 Gini Index 25 20.5 25 20 16.6 14.8 15.2 20 15 12.6 10.1 15 10 10 5 5 0 0 2000 2005 2010 2015 2000 2005 2010 2015 National Urban Rural National Urban Rural Source: Based on data from the EBCNV 2000, EBCNV 2005, EBCNV 2010, EBCNV 2015, INS. FIGURE 1.11. Annualized Growth of per Capita Consumption Expenditures by Percentile, 2000–15 a. 2000–05 b. 2005–10 4 4 3 3.5 Percent Percent 2 3 1 2.5 0 0 20 40 60 80 100 0 20 40 60 80 100 Expenditure percentile Expenditure percentile Annualized expenditure growth rate Annualized expenditure growth rate Growth rate in mean expenditure Growth rate in mean expenditure c. 2010–15 8 6 4 Percent 2 0 –2 0 20 40 60 80 100 Expenditure percentile Annualized expenditure growth rate Growth rate in mean expenditure Source: Based on data from the EBCNV 2000, EBCNV 2005, EBCNV 2010, EBCNV 2015, INS. 14 Tunisia’s Jobs Landscape FIGURE 1.12. Employment Type: Distribution of Employed Population, by Quintile of per Capita Household Expenditure, 2010 and 2015 100.0 80.0 60.0 Percent 40.0 20.0 0.0 2010 2015 2010 2015 2010 2015 2010 2015 2010 2015 1 2 3 4 5 employer ag employer nonag own-account ag own-account nonag wage worker ag wage wortker nonag unpaid work ag upaid worker nonag the Greater Tunis Metropolitan Area (5.3 percent), North- in nonagricultural and wage employment was observed East (11.6 percent), and Center-East (11.5 percent)—was among individuals in the bottom 20 percent of the house- low compared with the rest of the country, there were hold consumption distribution (Figure 1.12). For example, some delegations (districts) with relatively high incidence. in the lowest quintile, there was an increase from 47.6 per- Similarly, geographical gaps persist in terms of inequality. cent to 57.0 percent in the share of household members A large part of the inequality is driven by disparities within employed in salaried jobs outside agriculture, a decline of urban and rural areas and between regions. As of 2015, over 50 percent in the share of unpaid family workers, and in urban areas, the Gini index is estimated at 35.0, while, an increase in the share of nonagricultural employer and in rural areas, it stood at 31.9, and the gap has widened own-account workers.6 Although no data are available over time. on labor income among nonwage workers, the shift out of agriculture and toward wage employment was likely The acceleration in the pace of poverty reduction between associated with higher income from labor and, therefore, 2010 and 2015 seems to be largely associated with an an improvement in standards of living. expansion in public transfers. As low-income households largely rely on public transfers (69 percent of household Gaps in living standards are largely explained by differ- income in 2014), namely, pensions and social assistance ences in household endowments, but differences in returns (Krafft and Davis 2021), they have a considerable impact on matter between urban and rural areas as well as across the welfare of the poor relative to other sources of income. urban areas in different regions. Differences in house- In the aftermath of the revolution, Tunisia scaled up the hold consumption can be decomposed into differences in cash transfer program, the Programme National d’Aide aux household characteristics and differences in the returns Familles Nécessiteuses. The number of beneficiary house- to these characteristics.7 Results from this decomposition holds increased dramatically, from 176,000 in 2011 to 234,000 in 2015, and the amount of the transfer was 6  Sectoral labor productivity differentials are discussed below. 7  The Oaxaca-Blinder decomposition can be used to estimate differences in raised from TD 72 in 2010 to TD 150 in 2015 (real terms) welfare across regions or urban and rural areas and understand the main (CRES, AfDB, and ADF 2017). In addition, health insurance components (Blinder 1973; Oaxaca 1973). The first step consists of estimating at reduced prices (AMG2) was provided to an increasing log-consumption equations as a function of a set of household character- istics, including head’s age, marital status, educational level, labor force number of vulnerable households, and generous consumption status and sector of employment, household size, household demographic subsidies continued to shield purchasing power. A second structure (share of household members in different age-groups), household educational structure (share of household members with different educa- channel might have contributed to lift some households tional level), the main source of heating of the dwelling, the main source of out of poverty. The employment composition changed con- drinking water, and the distance to the nearest health center and the nearest commercial center. The second step implements the Oaxaca-Blinder decom- siderably with the shift out of agriculture and an increase position to estimate gaps between areas of interest and obtain the explained in the share of wage employment. The largest increase (endowments) and unexplained (returns) components. Economic Growth, Structural Transformation, and Employment 15 provide insights about the best approach to reducing wel- in 2011–19 from 85 percent in 2000–10, while the con- fare gaps. Urban-rural differences within regions are largely tribution of investments declined from 74.3  percent to explained by gaps in endowments. Gaps between Greater −6.6 percent. Tunis and other urban areas are also ascribable to different endowments, except for the North East, Center East, and The external sector, particularly exports, has been weak South West regions. where returns play a nonnegligible role. since the 2008 financial crisis and the 2011 political Gaps among rural areas in leading (North-East, Center- unrest. Although all the components of aggregate demand East, and Greater Tunis) and lagging regions largely derive have recently declined, the fall in exports has been par- from returns. Efforts to improve the productive character- ticularly large. Exports as a share of GDP increased on istics of lagging regions as well as of rural populations are average by about 2.3 percent per year between 2000 and therefore key to continuing to raise living standards. This 2010; they declined between 2011 and 2016, when they includes expanding access to basic services and to quality reached a low of 40 percent (Figure 1.13). Over the past health and education services. Gaps across urban/rural two decades, export growth (in volume) has been subpar; areas of different regions can be narrowed through a com- only Indonesia ranks lower than Tunisia (Figure 1.14). By bination of policies aimed at improving the characteristics contrast, remittances have hovered around 4.5  percent of local populations and efforts that improve the connec- of GDP over the last two decades. Receipts from tourism tivity of these areas. Policies aimed at encouraging economic have decreased with the political uncertainty triggered by activities and job creation in lagging locations through fiscal the 2011 revolution as well as with the impact of security and financial incentives have not proven successful in other incidents. They were estimated at 4.5 percent of GDP in countries. The Tunisian experience also points to lack of 2019, compared with 8.4 percent in 2000. Recent trends success in the use of incentives to reduce regional disparities paint a gloomy picture for Tunisia, which, as a small open (World Bank 2014). economy, needs a strong external sector to thrive. Firms need to sell to foreign markets to grow, benefit from econ- omies of scale, and boost job creation. Exporting firms can Economic Transformation also drive productivity growth because they are exposed to international competition. and Sources of Growth Exports have become increasingly diverse and complex, Economic growth has been increasingly driven by con- but have limited spillover on the rest of the economy. The sumption, while the contribution of trade and investments product space of exports has become more diversified and has faded. Over the past two decades, public and private more complex over time (Figure 1.15). Tunisia ranks 46th consumption has gained importance in GDP growth in the economic complexity index (ECI), which represents (Table 1.3), implying a shift of the economy to a less sus- an improvement of 8 positions over the past decade and tainable path of economic development as opposed to a leaves it behind only its aspirational peers. And yet the econ- growth led by sustainable factors, such as investment and omy is not taking full advantage of trade openness. Compa- trade. The contribution of both investments and net trade nies in the offshore sector benefit from tax exemptions and to GDP growth began declining in 2011. The contribu- reductions and simplified administrative procedures, but tion of exports to GDP growth dropped to 44.5 percent have limited connection with the rest of the economy, with TABLE 1.3. Annual Growth and Contribution to GDP Growth, by Expenditure Category and Subperiod, 2000–19 Average annual growth Contribution to GDP growth 2000−10 2011−19 2000−19 2000−10 2011−19 GDP growth 4.2 2.2 3.0 Exports 3.6 1.0 12.0 84.6 44.5 Imports 4.0 0.0 2.5 94.3 40.2 Consumption 4.5 2.5 3.6 106.1 111.7 Government 0.9 0.6 0.8 22.3 27.3 Private 3.5 1.9 2.8 83.8 84.4 Investment 3.1 −0.1 0.6 74.3 −6.6 16 Tunisia’s Jobs Landscape FIGURE 1.13. Trends in External Factors as a Share of GDP, 2000–19 70.0 60.0 Percent (share of GDP) 50.0 40.0 30.0 20.0 10.0 0.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Exports Imports Remittances FDI, net in ows International tourism, Net ODA received receipts Source: Based on data from the World Development Indicators, World Bank. FIGURE 1.14. Trends in the Export Volume Index, FIGURE 1.15. Economic Complexity Index Ranking, Tunisia and Comparator Countries, 2000–19 Tunisia and Comparator Countries, 1995–2018 100 Export Volume Index (2000=100) 90 80 70 60 Ranking 50 40 30 20 10 0 1995 2000 2005 2010 2018 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Tunisia Morocco Albania Indonesia Malaysia Romania Tunisia Morocco Albania Jordan Malaysia Romania Jordan Indonesia Source: Based on data of World Development Indicators, World Bank; the Atlas of Economic Complexity, Harvard University. Economic Growth, Structural Transformation, and Employment 17 implications in terms of the size of technological transfers FDI inflows were invested in manufacturing. The services and employment creation. sector continues to attract less than 10 percent of all FDI inflows, despite its importance as a key to improving job Structurally low investments have declined further over the opportunities among university graduates. last decade. Underpinning the recent meager economic performance, Tunisia has low levels of investment. In the The current account deficit is now above 10 percent of 1990s and in the first decade of the 2000s, investments GDP, pushed by the subpar performance of exports. The hovered around 24 percent of GDP (Figure 1.16). Since continued deterioration in net exports, coupled with a the revolution, both public and private investments have decline in tourism receipts and a roughly constant flow started a steady decline and reached 18 percent of GDP in of remittances, increased Tunisia’s current account deficit 2019. This is below the average among income group and above 10 percent of GDP (Figure 1.13 and Figure 1.20). regional comparators, which was estimated at 28.6 per- The COVID-19 pandemic and the global economic down- cent and 21.9 percent, respectively, in 2019 (Figure 1.17). turn contributed to narrowing the deficit temporarily to Investments in Tunisia also fall short of the levels observed 6.8  percent of GDP because of lower import demand in structural and aspirational peers, with the exception of and resilient remittances, despite a strong hit on exports Jordan (Figure 1.17). Over two-thirds of investments were and collapsing tourism receipts (IMF 2021). The current concentrated in the services sector, which was protected account deficit will remain a concern as current account from international competition (Figure 1.18). balances reflect the net saving rate of the economy, and deficits are not sustainable in the long run. Foreign direct investment (FDI) inflows have lost momentum, but they have gradually shifted toward manufacturing. In addition, Tunisia’s fiscal stance has deteriorated sig- Relative to both regional and income group comparators, nificantly. The fiscal deficit increased from an average of at around 3.8 percent, FDI inflows were a sizable share −2.4 percent of GDP over 2000–10 to −5 percent of GDP of GDP before the 2008 crisis. FDI inflows picked up and on average in 2011–19 because of high recurrent expen- reached 2.1 percent of GDP in 2019, well below the levels ditures, including the public wage bill, energy and food of the first decade of the 2000s (Figure 1.13). While, in the subsidies, and transfers to state-owned enterprises (SOEs) past, FDI inflows were mainly targeted at the energy sector (see Figure 1.21). Recurrent expenditures increased from (and, in some years, to telecommunication), they have par- 17.9  percent of GDP in 2010 to 21.4  percent of GDP tially shifted to manufacturing, which is key for growth, in 2011 and continued to grow, reaching 25.4 percent jobs, and exports (Figure  1.19). In 2019, 50  percent of of GDP in 2013, before declining slightly, to 24.7 percent of FIGURE 1.16. Trends in Gross Fixed Capital FIGURE 1.17. Gross Fixed Capital Formation in Formation, by Sector, 2000–19 Tunisia and Comparator Countries, 2019 50% Middle East & 21.6 45% North Africa 40% Middle income 28.6 35% Romania 23.6 30% Percent 25% Malaysia 23.0 20% Morocco 27.7 15% 10% Jordan 17.1 5% Indonesia 32.3 0% Albania 22.5 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 Tunisia 17.7 Private and Public Corps. Investment (LHS) G.Gov. Investment (LHS) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Percent Source: Based on data from the World Development Indicators, World Bank. 18 Tunisia’s Jobs Landscape FIGURE 1.18. Trends in the Sectoral Distribution of FIGURE 1.19. Trends in the Sectoral Distribution of Gross Fixed Capital Formation, 2000–14 Foreign Direct Investment, 2005–19 100% 100% 80% 80% 60% 60% Percent Percent 40% 40% 20% 20% 0% 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2005 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Agriculture Manufacturing Agriculture Manufacturing Other secondary Trade Energy Tourism and real estate Transports/ Hotels and food services Telecommunications Financial services Telecommunications Other services PA/Education/Health and domestic services Source: Based on data from the National Accounts, INS; Central Bank of Tunisia. GDP in 2017 (World Bank and AfDB 2020). The economic 15 percent of GDP. Expansion of public spending has helped downturn triggered by COVID-19 and the fiscal response address the challenge of insecurity and social demands that pushed up both the fiscal deficit and public debt in 2020. followed the 2011 revolution. With the 2012 law promoting According to the International Monetary Fund (IMF), the access to public administration by people wounded in the fiscal deficit (excluding grants) is estimated at 11.5 percent revolution and those covered by the amnesty of 2011, of GDP, and central government debt is estimated to have public sector hiring and the salaries of civil servants rose increased to 87.6 percent of GDP in 2020. considerably (Brockmeyer, Khatrouch, and Raballand 2015; INS 2017; OECD 2018). The wage bill rose from The civil servant wage bill is one of the highest in the world, 10.5 percent of GDP in 2000 to 14.6 percent of GDP in absorbing almost 50 percent of public expenditures and 2019, and it is estimated to have reached 17.6 percent of FIGURE 1.20. Trends in the Current Account Balance FIGURE 1.21. Trends in General Government Debt as a Share of GDP, 2000–19 and Budget Balance as a Share of GDP, 2000–19 0.0 0.0 90.0 –1.0 80.0 –2.0 –2.0 70.0 –4.0 60.0 –3.0 Percent Percent Percent 50.0 –6.0 –4.0 40.0 –5.0 –8.0 30.0 Budget balance –6.0 20.0 Gross debt –10.0 –7.0 10.0 –12.0 –8.0 0.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Source: Based on data from the World Economic Outlook, International Monetary Fund. Economic Growth, Structural Transformation, and Employment 19 GDP in 2020, following additional public sector hiring, of FIGURE 1.22. Factor Decomposition of GDP Growth, which about 4 in 10 in the health sector, and increases in by Subperiod, 1990–2018 salaries (IMF 2021). In 2020, the level was about twice the 100% median of 8.7 percent of GDP in non–oil-producing devel- oping markets and ranks Tunisia as the highest among non– 80% 40 50 45 oil-producing developing markets (IMF 2021). This bloated wage bill crowds out other public expenditures. In 2020, it 60% consumed about 75 percent of tax revenues, and it was also 95 Percent 40% 34 25 almost three times the size of public investment and almost 29 six times the spending on social programs (IMF 2021). 20% 26 30 21 25 The twin deficit is not sustainable in the long term. Fiscal 0% balances reflect the net savings rate of the public sector, and –20 current account balances reflect the net savings rate of the –20% 1990–2018 1990–2000 2000–2010 2010–2018 whole economy (Arezki et al. 2019). Tunisia’s public debt Capital Stock (α * gK) Labor ((1-α) * gL) risks becoming unsustainable in the medium term unless a Total Factor Productivity (gA) Total Period number of reforms aimed at reducing the fiscal deficit are adopted, including lowering the public wage bill, reduc- Source: Based on background paper for the 2021 Tunisia SCD Update. ing energy subsidies, strengthening the targeting of social protection spending, and making the tax system more effi- could be attributed to improvements in TFP. Since 2010, cient and fair. The positive correlation that exists in Tunisia the role of capital accumulation has increased to 95 pre- between the current account and the fiscal deficit raises the cent; labor has continued to account for about 25 per- question of the sustainability of the first, particularly in cent, while TFP has had a negative effect on growth situations in which the latter is motivated by expenditures (−20 percent). Although the levels are low, the expansion that do not have large multiplier effects. in investments and employment increasingly account for most of the growth, indicating the existence of flaws in Recent economic growth has been hampered by modest the economy. Controlling for human capital, one finds gains in total factor productivity (TFP). Before the revo- that the contribution of capital, labor, and human capital lution, the accumulation of capital and labor contributed is estimated at 56, 15, and 9 percent, respectively. Tunisia on average 45 percent and 25 percent to growth, respec- has the lowest total labor contribution to economic growth tively (Figure 1.22). The remaining 30 percent of growth among comparator countries (Figure 1.23). The country’s FIGURE 1.23. Factor Decomposition of GDP Growth with Human Capital, Tunisia and Comparators, 201018 120% 100% 18% 25% 25% 80% 9% 9% 30% 6% 13% 60% 15% 25% 17% Percent 18% 48% 40% 56% 20% 38% 43% 45% 25% 0% –20% –14% –20% –40% TUN MAR JOR MYS IDN Capital Stock Labor Human Capital TFP Source: Based on background paper for the 2021 Tunisia SCD Update. 20 Tunisia’s Jobs Landscape FIGURE 1.24. Trends in GDP, Total Factor Productivity, FIGURE 1.25. Trends in the Incremental Capital Labor Productivity, and Capital per Worker, 2000–19 Output Ratio, Tunisia and Comparator Countries, by Subperiod, 2000–19 200.0 14.0 150.0 12.0 Index (2000=100) 10.0 ICOR Ratio 100.0 8.0 6.0 50.0 4.0 2.0 0.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 – 2000–05 2006–10 2011–19 GDP Labor productivity Albania Indonesia Jordan Capital per worker TFP Malaysia Morocco Romania Tunisia Source: Based on data from the Penn World Table version 10.0. Source: Based on data from the Economist Intelligence Unit and Tunisian Institute of Competitiveness and Quantitative Studies. human capital index, a measure of the productivity of the next generation, is 0.51 compared with 0.56 for Jordan, 0.5 for Morocco, and as high as 0.62 for a benchmark aspi- However, the fact that private investments have continued rational country, such as Malaysia. Similarly, the role of TFP is a signal that returns to investments have remained high. is the lowest (and negative) among comparator countries There are also factor distortions linked to the barriers to (Figure 1.23). entry and exit of firms and regulatory failures. Capital is not flowing to the most efficient firms. Although factor The weak performance and recent decline of TFP might accumulation is appropriate for a country such as Tunisia, suggest a misallocation of resources. TFP performance which has a large stock of untapped human capital, produc- over the past decade has been negative and subpar relative tivity growth is necessary to generate more wealth per capita to rapidly growing economies, while labor productivity and ultimately more rapid jobs creation. Had TFP growth has continued to increase, although at a lower rate (Fig- been higher, labor productivity gains might have translated ure 1.24).8 One possible explanation for this trend is ineffi­ into higher output growth. Employment to growth elastic- ciency in the use of capital that could derive from allocative ity would not matter as much because employment growth inefficiency. Thus, capital may not flow to the most produc- would be greater. tive sectors of the economy. Or it may arise because of tech- nical inefficiency, that is, weak capacity in converting inputs Aggregate labor productivity in Tunisia is above that of into output.9 The rate of capital accumulation has slowed structural peers, but it below the regional average because recently, and capital per worker has increased at a lower of lagging secondary sector. Another commonly used indi- rate since 2011. The trend in the incremental capital output cator of productivity is labor productivity, which is strictly ratio indicates that capital investments are decreasing correlated with changes in living standards through wages. in efficiency and produce marginal returns (Figure 1.25). Labor productivity measures gross value added per unit of labor input and indicates how efficiently labor is used 8  Many developed countries had TFP growth of over 50 percent between in production.10 In Tunisia, labor productivity reached 1950 and 1970, with growth rates of TFP above 2 percent per year (Caves, Christensen. and Swanson 1980). 9  However, this might not be the only explanation of a TFP decline because TFP 10  First, in this report, labor productivity is measured as gross value added per may decrease because of a global decline in TFP. In 2000–19, many industrial- worker because information on hours worked is not readily available. This is ized countries posted modest TFP growth or a decline, whereas large devel- problematic if large gaps in working hours exist across sectors and if sizable oping economies such as China, Indonesia, Nigeria, the Philippines, and the changes occurs over time. Second, changes in labor productivity result from Russian Federation posted considerable TFP growth. In addition, the method- the combined effects of various factors, including technological change, capital ology adopted to construct the TFP productivity measure used in the analysis accumulation, the capacity of workers, and the intensity of their efforts; it can assumes that factor shares are constant across countries and over time. therefore be difficult to isolate the contribution of each component. Economic Growth, Structural Transformation, and Employment 21 FIGURE 1.26. Labor Productivity Gaps Overall and by Sector, Tunisia and Comparator Countries, 2017 3.5 3.0 2.5 2.0 Ratio 1.5 1.0 0.5 – Albania Jordan Indonesia Morocco Malaysia Romania Agriculture Industry Services Total Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS; World Development Indicators, World Bank. $36,650 (constant 2017 PPP) in 2019, above the average meager increase in labor productivity in the services sector. among middle-income countries ($27,850 constant 2017 The secondary sector was on average as productive as the PPP), but still below the regional average ($41,650 con- services sector in 2006, but its subpar performance over stant 2017 PPP) (Figure 1.27 and Figure 1.28). The pro- time translated into a decline of about 15 percent relative ductivity of Tunisia’s agricultural sector is relatively high to services. Within industry, mining, other manufacturing, among income group comparators and its structural peers and construction posted the largest drop in output per (Figure 1.26) as is the productivity of the services sector. worker, whereas, in services, all sectors except accommo- By contrast, Tunisia’s secondary sector has lost ground dation and food activities and other services were more and is considerably less productive than in Tunisia in other productive in 2017 relative to 10 years earlier. middle-income countries.11 Although structural change has been slow over the past Labor productivity gaps have narrowed over time, largely decade, Tunisia is ahead of the average middle-income thanks to productivity gains in agriculture. Labor produc- country in terms of nonagricultural employment. Between tivity gaps between agriculture and other sectors are the 2006 and 2017, structural transformation proceeded at driving force of the process of structural transformation a pace slightly below the average in other middle-income that pushes labor from low- to high-productivity sectors. countries.12 In 2017, agriculture accounted for 15 percent In Tunisia, labor productivity gaps have narrowed over of total employment, down from 19.2  percent in 2006 time (Figure  1.29). Such reduction is the by-product of (−4.4 percentage points) (Figure 1.30). This share is below three factors: (1) a rapid increase in agricultural labor pro- the average in middle-income countries, at 28.1 percent ductivity both before and after the revolution, (2) a decline in 2017 (Figure 1.32, panel a). The employment share of in productivity in secondary sectors since 2011, and (3) a the secondary sector increased slightly from 32 to 33 per- cent (+1.3 points), well above the middle-income countries average of 20.1 percent, and the services sector contrib- 11  The productivity gaps described here reflect differences in average labor productivity. What matters is productivity at the margin that, with well- uted 52 percent of total employment relative to 48 percent functioning markets and no constraints, should be equalized. Under a in 2006 (+ 3.1 points). The latter is in line with the share of Cobb-Douglas production function, the marginal productivity of labor equals the average productivity, multiplied by the employment share. So, 51.9 percent among middle-income countries. Within the if labor shares differ greatly across sectors, comparing average labor pro- secondary sector, the share of manufacturing declined to ductivities can be misleading. For example, average productivity in the mining sector is high. This is likely to be a reflection of the fact that the employment share of value added in this capital-intensive sector is small. 12 The average change in agricultural employment in middle-income   Nonetheless, other sectors, such as agriculture, manufacturing, construc- countries over the period 2006–17 was −6  percentage points, whereas tion, public administration, and health and education services, have a com- the average change in the secondary and tertiary sectors was +0.6 and parable employment share, and gaps in average productivity can therefore +5.4 percentage points, respectively (based on data of World Development approximate gaps in marginal productivity reasonably well. Indicators and ILO employment modeled estimates). FIGURE 1.27. Trends in Output per Worker, 2000–19 FIGURE 1.28. Ratio of Output per Worker in Tunisia vs. Middle East and North Africa, 2000–19 60,000 GDP per worker (constant 50,000 2.5 40,000 2017 PPP$) 2.0 30,000 1.5 Ratio 20,000 1.0 10,000 – 0.5 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 – 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Tunisia Middle East & North Africa Middle income Tunisia/MENA Tunisia/Middle income Source: Based on data from the World Development Indicators, World Bank. FIGURE 1.29. Labor Productivity, by Sector, 2006 and 2017 180,000 160,000 140,000 TD (2017 prices) 120,000 100,000 80,000 60,000 40,000 20,000 – Agriculture Agro-food Construction Mechanic and industries Textile manufacturing Mining and Construction Trade communication Accomodation Financial PA/Health/ Other services and restaurant Chemical services industries Education Transport and materials, Utilities electric Other Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS. FIGURE 1.30. Trends in the Sectoral Distribution of Employment, 2006–17 2017 14.8 18.4 1.1 13.8 13.3 5.4 3.7 1.0 19.1 9.3 2016 14.8 18.5 1.1 13.7 13.4 5.5 3.5 0.9 19.2 9.5 2015 14.8 18.6 1.1 13.4 14.0 5.7 3.5 0.9 19.3 8.8 2013 15.4 18.9 1.3 13.3 12.4 6.0 3.6 0.9 19.5 8.9 2011 16.4 18.5 1.0 14.2 12.5 5.6 3.4 0.8 18.8 8.8 2009 18.3 17.8 1.1 13.0 12.0 6.0 4.1 0.9 18.4 8.5 2008 17.8 19.3 1.2 12.7 11.6 5.8 4.0 0.8 18.4 8.3 2006 19.2 19.0 1.1 11.9 11.6 5.6 3.9 0.9 18.8 8.0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percent Agriculture and Fisheries Manufacturing industry Mining and Utilities Construction Trade Transport and communication activities Accomodation and restaurant activities Financial services PA/Health/Education Other services Source: Based on data from the Labor Force Survey (ENPE), INS. Economic Growth, Structural Transformation, and Employment 23 18.4 percent, although some industries posted an increase job in a different sector might imply moving to a different such as production of agrifood and mechanical and elec- location or might require a completely different set of skills. trical goods. The construction sector’s share increased In 2017, 60 percent of Tunisian workers were employed in by about 2 percentage points. Within the services sector, sectors with below average productivity (Figure 1.33). The transport and communication, as well as accommodation low productivity sectors include construction (13.8 percent and food services recorded a small decline in share, while of employment), agriculture (14.8 percent), trade (13.3 per- trade and other services rose by 1.7 and 1.3 percentage cent), and manufacturing (18.4 percent). In addition, about points, respectively. 19 percent of workers are employed in public administra- tion, education, and health services, which are sectors with On the production side, the sectoral structure of the Tuni- a productivity level only slightly above the average. Except sian economy is in line with the average middle-income for mining, which is typically a capital-intensive sector that country. Agriculture accounted for about 11  percent of employs a small share of workers, high productivity sec- value added in 2018, while industry contributed some tors, including transport and telecommunication, financial 25 percent, and the service sector took the lion’s share with services, accommodation and food services activities, and over 63 percent (Figure 1.31). This puts Tunisia broadly in other services, employ less than 20  percent of Tunisian line with the average middle-income country in terms of workers (Figure 1.33). structural change measured on the production side (Fig- ure 1.32, panel b). Economic growth was underpinned by labor productivity gains before the revolution and by employment creation Yet, 6 workers in 10 are still employed in sectors with below thereafter. To ascertain the contribution of structural average productivity. Workers have an incentive to move transformation to economic growth, the analysis carried from lower to higher labor productivity sectors as long as out a decomposition of GDP per capita growth. GDP labor productivity gaps persist across sectors and such gaps growth is decomposed into the contribution of changes in are reflected in the wages paid to workers. However, markets demographics, in employment, and in labor productivity are not always competitive. Labor productivity can differ (Box 1.1). Between 2006 and 2011, output per worker, from wages within sectors, and workers might face barriers a measure of labor productivity, increased by almost to mobility across sectors, for example because getting a 3  percent per year. It was the main driver of economic FIGURE 1.31. Trends in the Sectoral Distribution of Value Added, 2006–17 100% 90% 80% 70% 60% Percent 50% 40% 30% 20% 10% 0% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Agriculture and Fisheries Manufacturing industry Mining and Utilities Construction Trade Accomodation and restaurant activities Transport and communication Financial services PA/Health/Education activities Source: Based on data from the Statistical Yearbook, INS. 24 Tunisia’s Jobs Landscape FIGURE 1.32. Sectoral Distribution of Employment and Value Added, Tunisia and Middle-Income Countries, 2017 a. Employment b. Value added ZWE ZWE TZA TZA NPL NPL LAO BEN PNG UZB VUT GUY BTN FSM KEN COM MRT KIR COM KHM MMR MMR AGO NGA ZMB KEN TLS TZA CMR GHA IND CIV GEO PAK GNQ ALB CIV TON BGD VNM VNM PNG BEN TLS PAK MRT SLB NIC ALB KGZ FJI SEN AZE IND NGA CMR MAR MDA KHM MNG COG DMA DJI MAR GAB EGY MDA BGD GHA HND HND IDN GTM SUR SEN PRY THA BTN WSM XKX ARM BOL IDN MHL BOL UKR NIC GTM MNG BLR ECU ECU PER STP CHN TUN LKA FJI PHL PHL UZB WSM TON DZA EGY BLZ KGZ SWZ BWA ZWE NAM THA TKM MNE PRY MYS STP PSE TUR MKD LBY ZMB IRQ CPV SLV ARG CUB AGO GUY TUR LCA VCT IRN IRN BLZ LKA SRB CHN KAZ GEO BIH SRB JAM COL COL AZE MKD JAM UKR BIH TUN PER LBN GRD CPV CRI MEX SLV SWZ NAM CRI DOM VCT BRA MYS GAB BLR COG DZA LSO DOM LBN BRA MDV LSO KAZ MDV BGR MNE RUS VEN CUB SUR JOR BGR MEX PSE LCA RUS ZAF ZAF BWA JOR GNQ ARG IRQ 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Agriculture Industry Services Agriculture Industry Services Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS; World Development Indicators, World Bank. Economic Growth, Structural Transformation, and Employment 25 FIGURE 1.33. Labor Productivity and Employment Intensity, by Sector, 2017 140,000 25.0 120,000 19.1 18.4 20.0 100,000 TD (2017 prices) 14.8 13.8 13.3 15.0 80,000 Percent 60,000 Average labor 9.3 productivity, 10.0 TD 27,432 5.4 40,000 3.7 5.0 20,000 1.1 1.0 – 0.0 restaurant activities Financial services PA/Health/ Other services Agriculture and Manufacturing Mining and Utilities Construction Trade communication Accomodation and Education Transport and industry activities Fisheries Labor productivity Employment intensity Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS. growth, with a contribution of 104.6  percent. Demo- beginning of the period, while others, including mechani- graphics, captured by the share of population of working cal goods manufacturing, public administration, health age, contributed about 20  percent. Employment con- and education, and agriculture, were sectors with initial tributed negatively (−24.5 percent) to economic growth levels of productivity slightly above or below average. because employment creation fell short of the increase Except for agriculture and textiles, where employment in the working-age population (Figure 1.34). Following declined, productivity gains were not the by-product of the revolution (2011–17), labor productivity gains faded a reduction in the number of workers, but the results of (+0.2 percent per year on average) as value added growth the ability of firms to combine inputs more efficiently and was outpaced by employment creation. Employment rose increase value added. By contrast, in the second period at a rate of 1.7 percent per year on average and contrib- (2011–17), labor productivity increased only in a handful uted 80.4 percent to economic growth, becoming the main of sectors, whereas employment grew across the board, driver of growth. Demographics had a modest negative with the exception of textiles and chemical industries (Fig- effect on growth as the number of elderly increased more ure 1.35-, panel b). In 2 sectors out of 3 where output per rapidly than the population of working age (Figure 1.34). worker decreased, the decline is ascribable to a more rapid Over this period, Tunisia was the country with the small- growth of employment relative to value added as opposed est contribution of labor productivity to economic growth to a decline in value added. The sectors with the largest among comparators (Figure 1.36). It stands out though as productivity gains were all sectors with productivity levels the country with the largest positive effect of employment below the average in 2011, namely, agriculture, trade, and creation on economic performance. public administration, and health and education services. Efficiency gains in the use of labor achieved before the Growth in labor productivity was driven by within-sector revolution were lost to increases in employment levels. gains in productivity before the revolution. Changes in labor Before 2011, labor productivity increased in many sectors, productivity can be unpacked into changes in output per except for accommodation and food services, construc- worker within sectors and changes in output per worker tion, the chemical industry, and other manufacturing (Fig- ascribable to shifts in labor across sectors (see Box 1.1). The ure 1.35, panel a). Some of the sectors that posted large latter is one of the ways to measure the process of structural gains in output per worker, such as mining, financial ser- transformation, that is, the reallocation of economic activity vices, and transport, were high-productivity sectors at the across sectors that accompanies modern economic growth BOX 1.1. Shapley Decomposition of Changes in Value Added per Capita The methodology decomposes value added per capita growth using several consecutive steps. In a first step, growth in value added per capita is decomposed into changes in employment ratio, changes in output per worker (or labor productivity), and demographic changes, as follows (Figure B 1.1): Y Y E A = × × . (B1.1.1) N E A N Y = total value added; N = total population; E = total employment; A = total population of working-age; Y/E = v − → output per worker; E/A = e − → share of working-age population that is employed; A/N = a − → share of the total working-age population. In the second step, employment changes, De, are further decomposed into changes in employment by sectors: ∑ S De = Dei . (B1.1.2) i =1 The third step decomposes changes in output per worker into changes linked to changes in output per worker within sectors and changes linked to structural transformation or the reallocation of workers across sectors by noting as follows: Y Yi Ei ∑ ∑ S S = × or v = v i si , (B1.1.3) E s =1 Ei E s =1  Y   E  where v i = i  is output per worker in sector i; and si = i  is employment share in sector i. Taking differences of equa-  Ei   E  tion B1.1.3 between the final year (t) and the initial year (t − s), one obtains the following: ∑ ∑ S S Dv t = si Dv i,t + v i Dsi,t , (B1.1.4) i =1 i =1 Within Structural component transformation where Dvi,t and Dsi,t are the changes between period t and (t − s) in output per worker and employment share in sector i, respectively. Thus, changes in output per worker are the weighted sum of changes in output per worker in all sectors, where the weights are the employment shares of each sector. The weights of each sector are calculated as averages over the two periods of the shares in employment and the shares in output per worker in each sector. A fourth step goes further in understanding the role played by each sector on the aggregate effect of employment shifts across sectors. Increases in the share of employment in sectors with above-average productivity will increase overall productivity and contribute positively to the structural transformation term. By contrast, movements of labor out of sectors with above-average productivity will have the opposite effect. Similarly, increases in the share of employment in sectors with below average pro- ductivity will reduce growth, while reductions in their share will contribute positively to growth. If a sector has productivity below average and its employment share shrinks, then its contribution will be positive; Thus, outflows of workers from this low- productivity sector will have contributed positively to the increase in output per worker. If the same sector sees an increase in its employment share, such inflows of workers into this low-productivity sector will contribute negatively to output per worker and thus have a negative effect on the structural transformation term. The magnitude of the effect will be proportional to (a) the difference in the sector’s productivity with respect to the average and (b) the size of the employment shift. The last step combines all the elements together to calculate how much each factor contributes to GDP per capita growth. FIGURE B 1.1.1. Decomposition of per Capita GDP Growth ∆(Y/N) Change in GDP per capita ∆(Y/E) ∆(E/A) ∆(A/N) Change in Change in Change in working output per worker emp-to-working age pop ratio age to total population Changes between sectors (or structural change) Sectoral pattern of employment creation Changes within sectors Economic Growth, Structural Transformation, and Employment 27 FIGURE 1.34. Decomposition of Changes in per negative effect because of the reallocation of labor toward Capita Value Added, by Subperiod, 2006–17 sectors with below average productivity and away from sectors with above-average productivity. 140.0 120.0 19.9 Gains in labor productivity achieved before the revolution 100.0 were lost to increases in employment. Between 2011 and 80.0 80.4 2017, the within-sector component contributed negatively Percent 60.0 104.6 because sectors posted a decline in productivity, with a 40.0 few exceptions (public administration, health and educa- 20.0 28.4 tion services, agriculture, trade, and financial services). The 0.0 –8.9 –24.5 between-sector component that captures the effect because –20.0 of the reallocation of labor from sectors with lower than –40.0 average labor productivity to sectors with higher than aver- 2006–11 2011–17 age labor productivity explained the largest share of the Output per worker Y/E Employment Ratio E/A modest labor productivity gains (Figure  1.37, panel b). Demographics A/N However, the structural change observed during this period Source: Based on data from the Labor Force Survey (ENPE) and will not be able to drive economic growth going forward. Statistical Yearbook, INS; World Development Indicators, World Bank. Labor shifted to sectors with above-average productivity, but with negative productivity growth. (Herrendorf, Rogerson, and Valentinyi 2014). This step of the decomposition indicates that, between 2006 and The Tunisian economy is stuck in a low productivity equi- 2011, the within-sector component explained virtually all librium and operates below potential. Before the revolution, the productivity gains. The increase in output per worker sector level productivity increased the most in mining, was particularly high in mining, utilities, financial services, utilities, public administration, agriculture, transport and public administration, health and education services, agri- communication, and manufacturing. With the exception culture, manufacturing, and transport and communication of manufacturing, the productivity gains obtained before (Figure 1.37, panel a). Structural change exerted a small the revolution occurred in monopolistic or noncontestable FIGURE 1.35. Annualized Change in Labor Productivity and Employment, by Sector, 2006–17 a. 2006–11 12.0 Mining and Utilities Annualized change in output per worker (percent) 10.0 Financial services Mechanic and Construction electric industries 8.0 materials, ceramic PA/Health/Education and glass industries 6.0 Transport and Agriculture communication and Fisheries activities 4.0 Agro-food industries 2.0 Other services Textile Trade 0.0 –2.0 Accomodation and Construction –4.0 restaurant activities Chemical industries Other manufacturing –6.0 –4.0 –3.0 –2.0 –1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Annualized change in employment level (percent) (continued) 28 Tunisia’s Jobs Landscape FIGURE 1.35. Annualized Change in Labor Productivity and Employment, by Sector, 2006–17 (continued) b. 2011–17 10.0 Agriculture and Fisheries Construction materials, ceramic PA/Health/ and glass industries Annualized change in output per worker (percent) Education 5.0 Trade Financial Chemical services industries 0.0 Textile Other Agro-food Construction services –5.0 industries Other manufacturing Mechanic and Transport and electric industries –10.0 communication activities Accomodation and restaurant activities Mining and Utilities –15.0 –2.0 –1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Annualized change in employment level (percent) Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS; World Development Indicators, World Bank. Note: Circle size is proportional to sectoral labor productivity at the start of the period. FIGURE 1.36. Decomposition of Changes in per Capita Value Added in Tunisia and Comparator Countries, 2011–17 250.0 200.0 14.8 150.0 37.9 30.0 10.5 18.1 1.3 100.0 196.9 11.0 5.2 Percent 80.4 27.4 50.0 93.5 106.0 84.3 87.8 54.5 28.4 0.0 –8.9 –23.5 –43.9 –50.0 –111.8 –100.0 –150.0 Tunisia Albania Jordan Indonesia Morocco Malaysia Romania Output per worker Y/E Employment Ratio E/A Demographics A/N Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS; World Development Indicators, World Bank. Economic Growth, Structural Transformation, and Employment 29 FIGURE 1.37. Sectoral Contributions to Growth in Output per Worker, 2006–17 a. 2006–11 b. 2011–17 30.0 26.3 400.0 24.5 269.0 25.0 300.0 215.9 19.3 200.0 20.0 100.0 75.7 15.0 12.5 12.4 3.8 Percent Percent 0.0 10.0 –15.8 5.6 4.7 –100.0 –47.6 5.0 –20.6 1.5 –200.0 –93.6 –11.5 127.7 0.0 –300.0 –1.7 –1.0 –400.0 –5.0 –4.2 –403.0 –10.0 –500.0 Agriculture and Fisheries Manufacturing industry Mining and Utilities Construction Transport and communication activities Accomodation and restaurant activities Financial services PA/Health/Education Other services Strcutural transformation Agriculture and Fisheries Manufacturing industry Mining and Utilities Construction Trade Transport and communication activities Accomodation and restaurant activities Financial services PA/Health/Education Other services Strcutural transformation Trade Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS; World Development Indicators, World Bank. markets dominated by SOEs, where productivity reflects 2011 revolution, productivity gains have declined consid- rents more than increases in efficiency. State-controlled erably across the board, and only public administration, enterprises still operate in banking, mining, and utilities agriculture, and trade posted a sizable increase in output (OECD 2018). Private sector participation is still restricted per worker. The reduction in TFP points to misallocation of in the case of some agricultural products (Arezki et al. 2020). resources that are largely not captured by the most produc- Furthermore, the productivity growth in public administra- tive firms and sectors. The consequences are reflected in the tion is the effect of a more rapid increase in public expen- subpar economic performance of the country and in the lack ditures, that is, wages, relative to employment. Since the of employment creation in high productivity sectors. REFERENCES CHAPTER 1 Arezki, Rabah, D., Lederman, A.  A., Harb, R.  Y., Fan, and H., Nguyen. Article-IV-Consultation-Press-Release-Staff-Report-and-Statement-by- 2019. “Reforms and External Imbalances: The Labor-Productivity the-50128. Connection in the Middle East and North Africa, Middle East and INS (Institut National de la Statistique, National Institute of Statistics). North Africa.” Economic Update (April), World Bank, Washington, DC. 2017. Caractéristiques des Agents de la Fonction Publique et Leurs Arezki, Rabah, M., A., Ait Ali Slimane, A., Barone, K., Decker, D., Detter, Salaires 2001–2015. Tunis. R.Y., Fan, H., Nguyen, G., Miralles Murciego, L., Senbet. 2020. “Reach- Kapsos, S. 2005. “The Employment Intensity of Growth: Trends and ing New Heights: Promoting Fair Competition in the Middle East and Macroeconomic Determinants.” ILO Employment Strategy Paper 12. North Africa.” Middle East and North Africa Economic Update Krafft, C. and E.E., Davis 2021. “The Arab Inequality Puzzle: The Role (October), Washington, DC: World Bank. of Income Sources in Egypt and Tunisia.” Middle East Development Blinder, A. S. 1973. “Wage Discrimination: Reduced Form and Structural Journal 13 (1): 1–26. Estimates.” Journal of Human Resources 8 (4): 436–55. Mouelhi, R. and M., Ghazali 2014. “The Employment Intensity of Output Brockmeyer, A., M., Khatrouch, and G., Raballand. 2015. “Public Sector Size Growth in Tunisia and Its Determinants.” ERF Working Paper 857, and Performance Management: A Case-Study of Post-Revolution Tunisia.” Economic Research Forum, Giza, Egypt. Policy Research Working Paper 71589, World Bank, Washington, DC. Oaxaca, R.  L. 1973. “Male-Female Wage Differentials in Urban Labor Caves, Douglas W., Laurits R. Christensen. and Joseph A. Swanson. 1980. Markets.” International Economic Review 14 (3): 693–709. “Productivity in U.S. Railroads, 1951–1974. Bell Journal of Economics OECD (Organisation for Economic Co-operation and Development). 11 (1): 166–81. 2018. “OECD Economic Surveys: Tunisia.” March, Paris. CRES (Centre de Recherches et d’Etudes Sociales), AfDB (African Develop- UNDP (United National Development Program). 2020. Human Develop- ment Bank), and ADF (African Development Fund). 2017. Évaluation ment Report 2020: The Next Frontier, Human Development and the de la performance des programmes d’assistance sociale en Tunisie, May. Anthropocene. New York. Herrendorf, B., R., Rogerson, and A. Valentinyi. 2014. “Growth and Struc- World Bank. 2014. The Unfinished Revolution. Bringing Opportunities, tural Transformation.” In Handbook of Economic Growth, chapter 6, Good Jobs, and Greater Wealth to All Tunisians. Development Policy edited by P. Aghion and S. N. Durlauf. Review. Washington, DC: World Bank. IMF (International Monetary Fund). 2021. “Article IV Consultation: Press World Bank. 2020. Poverty and Shared Prosperity 2020: Reversals of For- Release; Staff Report; and Statement by the Executive Director for tune. Washington, DC: World Bank. Tunisia.” Country Report 2021/044 (February), Washington, DC. https:// World Bank and AfDB (African Development Bank). 2020. “Tunisia Public www.imf.org/en/Publications/CR/Issues/2021/02/26/Tunisia-2020- Expenditure Review.” World Bank Other Operational Studies 33854. 30 Tunisia’s Jobs Landscape ANNEX CHAPTER 1 FIGURE 1.A.1. Trends in Employment, by Sector, 2006–17 4,000 3,500 3,000 Employment (thousands) 2,500 2,000 1,500 1,000 500 – 2006 2008 2009 2011 2013 2015 2016 2017 Agriculture Manufacturing Construction Other secondary Trade Transports/Telecommunications Hotels and food services PA/Education/Health Services Other Services Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 1.A.2. Sectoral Contributions to Employment Growth, 2006–17 a. 2006–11 b. 2011–17 120.0 105.4 50.0 40.0 37.3 100.0 80.0 30.0 26.6 25.8 60.0 20.0 13.7 12.7 31.1 Percent 40.0 6.5 Percent 17.9 16.1 10.0 3.3 20.0 3.3 4.4 3.1 0.1 0.0 0.0 –10.0 –20.0 3.6 –13.0 –13.6 –20.0 –40.0 –60.0 –30.0 –54.8 –40.0 –29.7 –80.0 Agriculture and Fisheries Manufacturing industry Mining and Utilities Construction Trade Transport and communication activities Accomodation and restaurant activities Financial services PA/Health/Education Other services Agriculture and Fisheries Manufacturing industry Mining and Utilities Construction Trade Transport and communication activities Accomodation and restaurant activities Financial services PA/Health/Education Other services Source: Based on data from the Labor Force Survey (ENPE) and Statistical Yearbook, INS; World Development Indicators, World Bank. 31 CHAPTER 2 Access to the Labor Market: A Spotlight on Women and Youth HIGHLIGHTS ◾ Tunisia has realized considerable progress in educational outcomes, particularly enrollment, accom- panied by a reversal in the gender gap ◾ Nonetheless, 1 Tunisian of working age in 5 has no schooling, and the country lags comparators in the quality of learning ◾ With only 1 working-age individual in 2 participating in the labor market, Tunisia underutilizes its human capital, particularly youth and women ◾ Despite some improvements over the past decade spearheaded by youngsters with tertiary education, women’s labor force participation is extremely low ◾ Weak labor demand, assigned gender roles, limited availability of affordable childcare, and gender gaps in the ownership of productive assets and in private sector wages are among the main barriers to greater women’s participation in the labor market ◾ With about 4 youth ages 15–29 in 10 not in education, employment, or training (NEET) and high unemployment rates among university graduates, 1 youth in 3 is unemployed; the youth challenge is of paramount importance ◾ Inactivity seems to be a matter of exclusion among young men with little education, of lack of jobs among young men with tertiary education, and a combination of modest job creation and assigned gender roles among young women with university degrees ◾ Sluggish job creation, together with skill mismatch and a sizable public sector wage premium, seems to be the factor driving high unemployment among university graduates ◾ Wage subsidies providing temporary employment opportunities at the cost of significant dead- weight loss and substitution effects have little impact on long-term job creation ◾ The government and stakeholders need to boost the participation and employment of women and youth to take advantage of a small, but open demographic window 32 Tunisia’s Jobs Landscape C hapter  1 sets the stage by providing the macro­ population. The total dependency ratio, which captures economic context and trends in terms of growth, the ratio of nonworking-age to working-age population, living standards, and aggregate labor market out- declined from 84 (per 100 people of working age) in 1980 comes over past decades. Chapter 1 documents a gradual to 44.4 in 2011. It gradually increased (48.9 in 2019) and shift of the Tunisian economy toward a less sustainable is projected to continue to rise slowly over the next two economic development model that is based on domestic decades (Figure  2.1). A rising old-age dependency ratio demand and that includes the context of a decline in aggre- that will outweigh a falling child dependency ratio will gate productivity that points to technical and allocative slowly push the total dependency ratio up from 49.6 in inefficiencies. It shows that structural transformation has 2020 to 51.6 in 2040 (Figure 2.1). continued slowly, and the majority of workers are still employed in sectors with below average productivity. It also The demographic window is narrow, but still open. The illustrates that the gains in labor productivity achieved youth population, which is the main contributor to new before the revolution have been lost to employment labor market entries, will hover around 21 percent as a growth since 2011, a growth that has nonetheless not kept share of the total population over the next two decades up with the expansion in the working-age population. and gradually decline thereafter (Figure 2.2; Figure 2.3). (See Box 2.1 for definitions of selected terms.) The population of working age (15–64) will shrink mod- estly, from 66.8 percent to 66.0 percent, by 2040. The This chapter starts with an overview of the evolution of latter is the by-product of a modest increase in the share demographics in Tunisia in recent decades, including trends of youth (15–29), a decline by 5.4 percentage points in the in the age structure and educational attainment of the share of individuals ages 30–44, and an increase by 4 per- population, and how demographics will likely change in centage points in the share of individuals ages 45–64. By future. The chapter illustrates labor market trends, such as contrast, the share of the elderly, ages 65 and above, will recent changes brought about by the COVID-19 pandemic. nearly double, from 8.9 to 16.0 percent, largely thanks to The focus is on access to the labor market at the aggregate rising women’s life expectancy. Creating more jobs for a still level and according to individual characteristics, as well as sizable and increasingly well-educated working-age popula- among groups that are at a disadvantage. The chapter turns tion will be key to taking advantage of the demographic divi- a spotlight on the labor market participation of women and dend. Raising participation and employment rates among youth. It highlights how specific groups of women fare groups that are currently lagging will be an important chal- better than others in terms of access to the labor market, lenge that the government of Tunisia and other stakeholders and it provides a summary of the constraints on women’s will need to tackle given the aging of the population. participation in the labor market based on a desktop review of the available academic and grey literature. The chapter Over 60 (70) percent of the (urban) population is located also focuses on youth unemployment and inactivity, par- in coastal regions. The share of the population located ticularly among university graduates, and investigates the in Tunisia’s coastal regions, namely, Greater Tunis, the main barriers to a smooth transition from school to work. North-East, and the Center-East, rose from 60  percent to 62 percent between 2006 and 2017 (Table 2.1). These regions have also become more urbanized, and, in 2017, Demographics and Projections they were home to almost 72 percent of the urban popu- lation of the country. Inland regions host 28 percent of Tunisia is at the later stages of a demographic transition. the country’s urban population. The share of the urban Changes in demographics have important effects on labor population increased from about 65 percent in 2006 to market developments, economic growth, and living stan- 68 percent in 2017. All regions posted an expansion in dards in all countries. Thus, in Tunisia, the total fertility the rate of urbanization except the South-West, where the rate, which measures the number of births per woman, share has remained constant, at 68 percent. has fallen by more than half over the past 40 years, from 5.1 in 1980 to 2.2 in 2018 (WDI), and life expectancy In recent decades, Tunisians have achieved substantial increased from 62 years in 1980 to over 76 years today. progress in educational outcomes. In 2014, about 8 Tuni- These changes impacted the population growth rate, which sians in 10 were literate, compared with fewer than 1 in declined from 2.6 percent in 1980 to about 1.1 percent 2 three decades earlier. The literacy rate is comparable over the last five years, as well as the age structure of the with the average rate in the region (79.3 percent in 2019) Access to the Labor Market: A Spotlight on Women and Youth 33 BOX 2.1. Definitions of Key Labor Market Concepts Labor market status Population of working age All individuals ages 15 and above Labor force All individuals of working age who were either employed or unemployed during the reference week Employed The employed population consists of individuals of working age who have worked for pay, profit, or household gain for at least one hour during the reference week. It includes individuals who are temporarily absent from work for reasons such as working time arrangements, the nature of their work, public holidays, annual leave, sick leave, or maternity/paternity leave. Unemployed The unemployed population comprises all individuals of working age who were not employed during the reference week, looked for work during the past month, and were available for work during the reference week. Out of the labor force The population out of the labor force includes individuals who were neither employed nor unemployed during the reference week. NEET Youth, ages 15–24, who are not in employment, education or training. Type of employment Wage worker or employee A wage worker or employee is a person who works for pay for someone else, even in a temporary employment. Apprentice An apprentice is a person being trained for a job or trade. The individual may be paid or may receive some pocket money; a paid apprentice is considered in employment. Unpaid apprentices are considered as out of labor force. Employer An employer is a person who operates his/her own business or trade and hires one or more employees. Own-account worker An own-account worker is a person who operates his/her own business or trade and does not hire employees. He/She may be working alone or with the help of contributing family workers. Unpaid or contributing family A contributing family worker is a person who works without pay in a market-oriented worker enterprise operated by a household member. Public/private employment Public sector employment Employment in the public sector comprises all employees working in a public establish- ment or in a public company. Private sector employment Employment in the private sector includes all employees not working in a public establishment or in a public company, as well as all employers, own-account workers, and unpaid family workers. Formal/informal employment Informal employment Informal employment includes (a) employees and apprentices who work for an employer who does not contribute to social security on their behalf or, in the case of missing answers, if they do not benefit from paid annual leave and paid sick leave; (b) own-account workers and employers who run informal sector economic units (as defined below); (c) all contributing family workers. Informal sector The informal sector includes own-account workers and employers who run non­ incorporated private enterprises without a tax identification number, or with a tax identification number, but without a formal bookkeeping system. 34 Tunisia’s Jobs Landscape FIGURE 2.1. Total, Child and Old-Age Dependency Ratios, 1971–2075 100 90 Total dependency ratio 80 Old-age dependency ratio Dependency ratio 70 Child dependency ratio 60 50 40 30 20 10 Projections 0 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013 2016 2019 2022 2025 2028 2031 2034 2037 2040 2042 2045 2048 2051 2054 2057 2060 2063 2066 2069 2072 2075 Source: Based on data from the World Development Indicators, World Bank; World Population Prospects 2019 (database), Population Division, Department of Economic and Social Affairs, United Nations, New York, https://population.un.org/wpp/. FIGURE 2.2. Population Pyramid, by Five-Year FIGURE 2.3. Population Pyramid, by Five-Year Age-Group, 2020 Age-Group, 2040 (Medium Variant Projection) 100+ 100+ 90–94 90–94 80–84 80–84 70–74 70–74 Age-group Age-group 60–64 60–64 50–54 50–54 40–44 40–44 30–34 30–34 20–24 20–24 10–14 10–14 0–4 0–4 600 400 200 0 200 400 600 600 400 200 0 200 400 600 Thousands Thousands Female Male Female Male Source: Based on data of World Population Prospects 2019 (database), Population Division, Department of Economic and Social Affairs, United Nations, New York, https://population.un.org/wpp/. TABLE 2.1. Distribution of the Population, by Region, Urban or Rural Area, and Share of Urban Population, 2006 and 2017 National Urban Rural Share urban 2006 2017 2006 2017 2006 2017 2006 2017 Greater Tunis 22.9 24.3 32.2 32.8 5.1 6.3 92.3 91.7 North-East 14.0 14.0 13.0 13.3 15.7 15.6 61.1 64.5 North-West 12.0 10.4 6.9 6.4 21.7 18.8 37.7 42.0 Center-East 22.8 23.8 25.1 25.5 18.4 20.3 72.1 72.8 Center-West 13.5 13.0 6.8 6.8 26.4 26.1 32.7 35.6 South-East 9.2 9.1 10.1 9.9 7.6 7.5 71.5 73.6 South-West 5.7 5.5 5.9 5.5 5.2 5.5 68.1 68.1 Total 100.0 100.0 100.0 100.0 100.0 100.0 65.4 68.0 Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 35 FIGURE 2.4. Literacy Rates, by Birth Cohort, 2015 FIGURE 2.5. Educational Level, Distribution by Cohort, 2017 90 70 70 Percent 50 50 Percent 30 30 10 10 50 9 9 9 9 00 –5 –6 –7 –8 19 0 50 60 70 80 –2 50 9 9 9 9 00 –5 –6 –7 –8 e 19 19 19 19 90 19 0 r 50 60 70 80 fo –2 19 e be 19 19 19 19 90 r fo 19 be Men Women None Primary Secondary Tertiary Source: Based on data from the EBCNV 2015, INS; and from the Labor Force Survey (ENPE), INS. and below the average among middle-income countries 52 percent as productive when they grow up as they might (86.4 percent in 2019). The gross secondary-school enroll- have been if they had enjoyed complete education and full ment ratio rose from 25.1 percent in 1980 to 92.7 percent health. The human capital index in Tunisia is below the in 2016, and the gross tertiary enrollment ratio increased regional average, but higher than the average among lower- from 5.0 percent in 1980 to 31.8 percent in 2019. This middle-income countries. The index is higher in Tunisia than compares with 81.7 (41.0) percent and 77.2 (36.9) per- in Morocco, but lower than in Albania (63), Indonesia cent for the secondary (tertiary) enrollment ratio among (54), Jordan (55), Malaysia (61), and Romania (58). In regional and income group comparators, respectively. 2010–20, the index in Tunisia declined slightly, from 0.53 to 0.52, largely because of a drop in harmonized test scores. Yet, 1 Tunisian of working age in 5 has not obtained any These scores measure performance in international testing school certificate. The progress achieved in recent decades programs. Students in Tunisia score 384 (405 in 2010) on is reflected in the educational level among different cohorts of the working-age population. The literacy rate increased from about 17 percent and 51 percent among women and FIGURE 2.6. Educational Level, Distribution Among men born before 1950, respectively, to over 95 percent among the Working-Age Population, 2006 and 2017 younger generations (born in the 1990s) (Figure 2.4). Sim- ilarly, educational attainment has improved considerably over time. About 1 Tunisian in 3 born in the 1990s is a uni- versity graduate, compared with 1 in 4 among Tunisians born 2006 23.5 31.0 35.4 10.0 between 1980 and 1989 and less than 2 percent in the cohort born before 1950 (Figure 2.5). The number of additional uni- versity graduates peaked during the 2010/11 academic year and started to decline thereafter, to reach about 52,000 in 2017 20.0 30.1 34.4 15.6 2018/19. Meanwhile, as of 2017, about 20 percent of the working-age population has no education (23.5 percent in 2006); about 30 percent has primary education; 34 percent has secondary education; and 15.6 percent has tertiary edu- 0 10 20 30 40 50 60 70 80 90 100 cation (10 percent in 2006) (Figure 2.6). Percent None Primary Secondary Tertiary The quality of education lags in Tunisia relative to com- parator countries. Children born in Tunisia today will be Source: Based on data from the Labor Force Survey (ENPE), INS. 36 Tunisia’s Jobs Landscape FIGURE 2.7. Mathematics Scores, Tunisia and FIGURE 2.8. Science Scores, Tunisia and Comparator Comparator Countries, Circa 2015 Countries, Circa 2015 Tunisia Tunisia Romania Romania Morocco Morocco Malaysia Malaysia Jordan Jordan Indonesia Indonesia Albania Albania 0 200 400 600 0 200 400 600 8th grade 4th grade 8th grade 4th grade Source: Based on data of TIMSS (Trends in International Mathematics and Science Study) (data repository), International Association for the Evaluation of Educational Achievement, Amsterdam, https://www.iea.nl/data-tools/repository/timss. Note: Administration year changes across countries: Indonesia (2015, 4th; 2011, 8th), Jordan (2015), Malaysia (2015), Morocco (2015), Romania (2011), Tunisia (2011). a scale on which 625 represents advanced attainment and The rapid improvement in educational outcomes was 300 minimum attainment. According to the 2015 Program accompanied by a closure and, in some cases, a reversal for International Student Assessment (PISA), Tunisia of gender gaps. The literacy rate among women rose from performed well below the OECD average in reading 35.8 percent to 72.2 percent between 1984 and 2014, and, (361 vs. 490), science (386 vs. 491), and mathematics among younger cohorts, women are today on par with men (367 vs. 487). There was also a decline in scores in 2012–15. (see Figure 2.4). About 17 percent of women of working The Trends in International Mathematics and Science age have tertiary education, compared with 13.9  per- Study (TIMSS) provides data on mathematics and science cent among men. The gap is considerably larger among achievement among students at grades 4–8 every four years youth ages 25–29 (38.2 percent vs. 23.7 percent among since 1995. In 2011, the average mathematics scores of young women and young men, respectively). Over the 4th and 8th graders in Tunisia were 359 and 425, respec- past decade, about 70 percent of university graduates are tively, and the average science scores were 346 and 439 women. The human capital index is higher among women (Figure 2.7; Figure 2.8). This is below comparator coun- (54) than men (50). In 4th grade, girls outperform boys tries for 4th grade students and below aspirational peers in science assessments (2011 TIMSS), but, in 8th grade, (Malaysia and Romania) for 8th grade students. boys do slightly better than girls both in mathematics and FIGURE 2.9. PISA and TIMSS Test Scores, by Sex, 2011 and 2015 500 400 300 Score 200 100 0 Reading Mathematics Science Mathematics Science Mathematics Science PISA (2015) TIMSS 4th grade (2011) TIMSS 8th grade (2011) Women Men Source: Based on data of PISA (Programme for International Student Assessment) (dashboard), Organisation for Economic Co-operation and Development, Paris, http://www.oecd.org/pisa/pisaproducts/; TIMSS (Trends in International Mathematics and Science Study) (data repository), International Association for the Evaluation of Educational Achievement, Amsterdam, https://www.iea.nl/data-tools/repository/timss. Note: Scores range from 0 to 1000. Some apparent differences between estimates may not be statistically significant. Access to the Labor Market: A Spotlight on Women and Youth 37 science. Among 15-year-old students, girls outperform but considerably low compared with the average middle- boys in reading, while boys and girls perform equally in income country (64.9 percent in 2017). mathematics and science (2015 PISA) (Figure 2.10). Labor force participation among women is unusually low, particularly among women with little education. Fewer Trends in Access to the than 3 women in 10 participates in the labor market. At 26.5  percent, compared with 68.3  percent among men Labor Market (2017), women’s labor force participation rose modestly during the decade (24.4  percent in 2006) (Table  2.2). In Tunisia, human capital is underutilized: more than Youth also showed lower than average participation, 1 Tunisian of working age in 2 are not employed and not largely thanks to increases in secondary and tertiary enroll- looking for work. The working-age population ages 15 ments. The activity rate of people with no education was and above, comprises 8.7 million people (76 percent of the exceptionally low. In 2017, the participation rate among total population) who can contribute productively to the these people was estimated at 18.1 percent, down from economy (2017) (see Figure 2.10). About 47 percent of 24.7  percent in 2006. This was mainly ascribable to the working-age population is active in the labor market, women ages 30–44 and 45–64 with no education. and 53  percent is neither employed nor looking for work, corresponding to 4.6 million people. Among the About 15  percent of the labor force is unemployed, inactive, more than 8 in 10 (or 3.7 million people) are also not a rate higher than a decade ago. About 0.6 million people in education. Tunisia’s labor force participation rate is above looking for work were unable to find job in 2017 (see Fig­ the average in the Middle East and North Africa region ure  2.10). This corresponds to an unemployment rate (43.2 percent in 2017, excluding high-income countries), of 15.3  percent, almost three times as high as the rate FIGURE 2.10. Labor Market Structure, Tunisia, 2017 Wage workers 0.21 million (43%) Own-account Agriculture workers/employers 0.5 million (15%) 0.23 million (45%) Contributing family workers Employed 0.06 million (12%) 3.5 million (85%) Labor force 4.1 million (47%) Wage workers Working age 2.38 million (81%) population (15+) Unemployed 0.6 million (15%) 8.7 million (76%) Total population Not in the Non agriculture Own-account 11.4 million Iabor force 3 million (85%) Not in working workers/employers 4.6 million (53%) age population 0.52 million (18%) Out of school 2.7 million (24%) 3.7 million (81%) Contributing family workers 0.03 million (1%) Source: Based on data from the Labor Force Survey (ENPE), INS. Note: The percentages in brackets are calculated as a share of the level displayed in the higher-level cell. Estimates of public sector employment differ from administrative data possibly due to measurement error in information about place of work reported by respondents in the labor force survey. 38 Tunisia’s Jobs Landscape TABLE 2.2. Key Labor Market Indicators, by Sex, Age-Group, Educational Attainment, and Urban or Rural Location, 2006–17 2006 2008 2009 2011 2013 2015 2016 2017 Labor force participation rate By sex Men 67.3 68.0 68.7 70.1 70.0 68.8 68.5 68.3 Women 24.4 24.7 24.8 24.9 25.6 26.0 26.6 26.5 By age-group Youth 15–29 40.1 41.7 41.0 44.5 45.3 41.6 43.2 42.2 Adults 30–54 60.3 60.9 62.0 60.5 61.1 62.9 62.5 62.7 Older workers 55+ 22.4 20.5 20.5 21.0 20.3 21.5 21.1 21.2 By educational level No education 24.7 22.4 23.1 22.0 18.0 19.4 18.4 18.1 Primary 54.3 55.6 55.5 53.8 52.6 52.0 51.9 51.2 Secondary 46.6 47.0 46.4 47.8 52.8 50.6 51.7 51.3 Tertiary 64.6 65.5 67.3 69.8 63.5 67.2 66.1 66.6 By area Urban areas 46.3 47.5 46.6 47.8 49.0 49.2 49.4 48.8 Rural areas 44.3 43.4 46.2 46.0 44.1 42.5 42.2 43.0 Employment-to-population ratio By sex Men 59.5 60.4 61.0 59.6 60.7 60.3 60.0 59.8 Women 20.7 20.8 20.1 18.1 19.7 20.3 20.4 20.5 By age-group Youth 15–29 30.2 31.0 29.4 27.5 30.1 27.8 28.7 28.2 Adults 30–54 56.4 57.4 58.1 55.2 56.2 57.2 56.9 57.1 Older workers 55+ 21.7 20.0 20.2 20.5 19.9 21.1 20.8 20.9 By educational level No education 23.1 21.5 21.7 20.2 17.2 18.4 17.4 17.4 Primary 47.2 49.7 49.7 47.2 47.4 47.4 47.4 46.9 Secondary 40.8 40.7 39.9 38.0 44.2 42.4 43.2 43.2 Tertiary 53.7 52.4 52.5 49.4 44.3 49.2 47.2 47.2 By area Urban areas 40.2 41.4 40.6 39.2 41.1 41.5 41.6 41.0 Rural areas 39.3 38.3 39.7 37.1 37.3 36.6 36.0 37.0 Unemployment rate By sex Men 11.5 11.1 11.3 15.0 13.3 12.4 12.4 12.4 Women 15.1 15.9 18.8 27.4 23.0 22.2 23.5 22.6 By age-group Youth 15–29 24.6 25.6 28.2 38.2 33.5 33.1 33.5 33.2 Adults 30–54 6.4 5.6 6.3 8.7 8.0 9.0 9.0 9.0 Older workers 55+ 2.8 2.2 1.5 2.1 1.8 2.0 1.5 1.8 (continued) Access to the Labor Market: A Spotlight on Women and Youth 39 TABLE 2.2. Key Labor Market Indicators, by Sex, Age-Group, Educational Attainment, and Urban or Rural Location, 2006–17 (continued) 2006 2008 2009 2011 2013 2015 2016 2017 By educational level No education 6.3 4.2 6.1 8.0 4.7 5.5 5.6 4.3 Primary 13.0 10.6 10.4 12.4 9.9 8.8 8.6 8.3 Secondary 12.5 13.4 14.0 20.6 16.2 16.3 16.5 15.6 Tertiary 16.9 20.0 21.9 29.2 30.2 26.8 28.5 29.1 By area Urban areas 13.0 12.8 12.9 17.9 16.1 15.7 15.9 15.9 Rural areas 11.4 11.7 14.0 19.3 15.4 13.9 14.8 13.9 Source: Based on data from the Labor Force Survey (ENPE), INS. observed in middle-income countries and about 2.5 per- Similarly, young university graduates face high unemploy- centage points higher than the regional average. During ment rates (29.1  percent in 2017, roughly stable since the years preceding the 2011 revolution, the unemploy- 2011) relative to individuals with secondary (15.6  per- ment rate rose from 12.5 percent in 2006 to 18.3 percent cent) or lower education (8.3 percent, primary education; in 2011. Since then, it has gradually declined, and, yet, it 4.3 percent, no schooling). remains above the rate a decade ago and is significantly higher than the average among regional (10.9 percent) and Rural areas and inland regions lag in all labor market income group comparators (5.6 percent) (Table 2.3). outcomes. Individuals in rural areas are engaged in the labor market less than their urban counterparts: in 2017, Unemployment is higher among women and university about 4 in 10 Tunisians in rural areas participated in the graduates. The average unemployment rate of 15.3 per- labor market relative to almost 5 in 10 in urban areas cent masks considerable heterogeneity. Although women (see Table 2.3). The gap in the activity rate expanded from and men alike posted a decline in unemployment begin- 2 percentage points in 2006 to almost 6 percentage points ning in 2011, the unemployment rate is higher among in 2017 because of both a constant increase in urban women than among men. In 2017, it was estimated at areas and a decline in rural areas since 2011. At 15.9 per- 22.6 percent among women and 12.4 percent among men. cent in 2017, the unemployment rate in urban areas was Large gaps exists across age-groups and educational levels. above the rate in rural areas (13.9 percent), while both The unemployment rate among youth ages 15–29 was declined beginning in 2011 and are still above the level 33.2 percent in 2017, on the decline relative to the level observed in 2006. Wide gaps exists across regions (see reached in 2011 (38.2 percent). This compares with less Table  2.3). The more deprived inland regions showed than 1 in 10 among individuals of prime age (30–54) and both lower participation rates (42.3 percent vs. 49.8 per- 1.8 percent among older individuals (ages 55 or more). cent in inland and coastal regions, respectively, in 2017) TABLE 2.3. Key Labor Market Indicators, 2006–17 2006 2008 2009 2011 2013 2015 2016 2017 MICs MENA Labor force participation rate 45.6 46.2 46.5 47.2 47.4 47.1 47.2 47.0 64.9 43.2 Labor force participation rate, women 24.4 24.7 24.8 24.9 25.6 26.0 26.6 26.5 45.2 18.0 Employment-to-population ratio 39.9 40.4 40.3 38.5 39.9 39.9 39.8 39.8 61.3 38.1 Unemployment rate 12.5 12.4 13.3 18.3 15.9 15.2 15.6 15.3 5.6 12.9 Share of wage employment, % of total employment 68.2 69.3 — 71.2 72.0 72.8 72.2 75.1 47.6 62.6 Share of nonagricultural employment, % of total 80.9 82.3 81.9 83.8 84.7 85.2 85.3 85.3 30.7 20.4 employment Source: Based on data from the Labor Force Survey (ENPE), INS; and World Development Indicators, World Bank. Note: The data on middle-income countries (MICs) and countries in the Middle East and North Africa (MENA) refer to 2017 and are based on national estimates with the exception of the share of employment in agriculture in both MICs and MENA countries and the overall and female labor force participation rate in MICs, which are based on modeled estimates o the International Labour Organization. The data on the MENA region exclude high-income countries. 40 Tunisia’s Jobs Landscape and higher unemployment (20.1 percent vs. 12.9 percent Tunisians are moving out of agriculture and are increas- in inland and coastal regions, respectively, in 2017). The ingly employed in the services sector.14 Of about 3.5 million unemployment rate was 25.6  percent and 24.3  percent employed in 2017, 85 percent were working in nonagricul- in the South-West and South-East regions, respectively, tural sectors, mostly in services (52 percent) (Figure 2.12). followed by the Center-West and North-West regions, at The share of workers in agriculture declined by 23 per- 17.4 percent and 16.7 percent. The North-East and Center- cent (more than 60,000 individuals) over the decade, from East regions exhibited unemployment rates of around 19.2 percent in 2006 to 14.8 percent in 2017, and the pace 10.0 percent, whereas the Greater Tunis region reached of the transition accelerated beginning in 2011. The share 17.0 percent. Unemployment modestly declined beginning of the secondary sector rose from 32.0 percent to 33.3 per- in 2011 across all regions, but only the North-East in 2017 cent in 2006–17, but this was below the peak reached had an unemployment rate lower than a decade previous in 2011 (33.7 percent). Within the secondary sector, food (10.4 percent in 2017 vs 14.2 percent in 2006) thanks to manufacturing posted a growth of over 44.0 percent (or steady growth in the number of the employed. more than 28,000 individuals), followed by construction, with an increase of about 34.0 percent, and other manu- The unemployed are largely youth, men, individuals with facturing (26.7 percent) (Table 2.4). By contrast, textiles up to primary or secondary education, and urban residents. shed jobs, and employment declined by almost 10 percent Of 0.6 million unemployed in 2017, about 2 in 3 were (a loss of 24,500 individuals). The services sector continued youth ages 15–29 (31 percent in the 15–24 age-group and a slow yet steady expansion, from 48.8 percent in 2006 to 32  percent among the 24–29 age-group) (Figure  2.11). 51.9 percent in 2017, adding over 330,000 workers, which This is particularly concerning because, over the next two is an increase of about 23 percent, compared with a growth of decades, the share of youth in the total population will 20 percent in the secondary sector. Public administration, remain roughly constant, at about 23  percent. Around together with the education and health sector, contributed 58 percent of the unemployed were men, and almost 60 per- about 20 percent to total employment, with approximately cent had at best obtained a certificate of secondary educa- 658,000 employed. Trade was the second largest sector. tion: 38 percent had secondary education; 18 percent had Trade was also the main contributor to employment cre- primary education; and about 2 percent had no schooling. ation in the services sector (+111,300), followed by public Individuals with tertiary education contributed 42 percent administration, health and education services, real estate to total unemployment. Most of the unemployed live in and professional services, and transport. urban areas and are predominantly located in the Greater Tunis region (30.5 percent) and in the Center-East region Most Tunisians work for a wage. About 75 percent of the (15 percent).13 employed work for a wage (2017), an increase by more than 10 percent over the decade, with steady and stable The inactive population is prevalently composed of youth, growth both before and after the 2011 revolution (Fig- women, individuals with up to primary education, and ure 2.13). The rate is above the average in middle-income urban residents. About 40  percent of the inactive are countries (47.6 percent) and in the region (62.6 percent, youth ages 15–29, of whom 31 percent are ages 15–24 excluding high-income countries). The share of unpaid and largely in school (see Figure 2.11). About 20 percent family workers decreased considerably, from nearly 7.0 per- of the inactive population is ages 30–44, and 42 percent cent in 2006 to less than 3.0 percent in 2017, as did the are ages 45 or more. The large majority of inactive indi- share of own-account workers, from about 19.0  per- viduals are women (71 percent) and have little education. cent to 15.5 percent. By contrast, employers had gained Almost 1 in 3 had no schooling, 26 percent has primary importance; in 2017, they contributed about 6 percent to education, and 34 percent have secondary education. Less total employment. In agriculture, only 43 percent of the than 9 percent of the inactives have a university degree. employed are wage workers, compared with 81 percent in Over 2 inactives in 3 live in urban areas, and over 1 in nonagricultural sectors (see Figure 2.10). Self-employment 2 lives in coastal regions. represents the majority of agriculture workers: 45 percent 14  Movements of labor indicate individual transitions across sectors, which   Combined, the two regions contributed about 48 percent of the total 13 are typically rare, particularly movements out of agriculture and a net addi- working-age population. tion of more workers in services relative to agriculture. Access to the Labor Market: A Spotlight on Women and Youth 41 FIGURE 2.11. Distribution of Working-Age Individuals, by Labor Status, Age-Group, Sex, Educational Level, and Residence, 2017 a. By age-group b. By sex Employed 10.5 12.5 43.1 32.2 1.8 Employed 73.7 26.3 Unemployed 31.0 32.0 32.4 4.50 Unemployed 57.6 42.4 Inactive 27.7 7.5 19.2 27.4 18.2 Inactive 29.3 70.7 0 20 40 60 80 100 0 20 40 60 80 100 Percent Percent 15–24 25–29 30–44 Men Women 45–64 65 and above c. By educational attainment d. By urban or rural residence Employed 8.7 35.4 37.3 18.5 0. Employed 70.9 29.1 Unemployed 17.7 38.2 41.9 0. Unemployed 74.0 26.0 2.2 0. Inactive 30.8 27.7 31.6 Inactive 66.3 33.7 9.8 0 20 40 60 80 100 0 20 40 60 80 100 Percent Percent None Primary Secondary Urban Rural Tertiary Not stated e. By region 4.6 Employed 27.0 16.2 9.7 25.1 10.2 7.3 Unemployed 30.5 10.4 10.8 14.9 11.8 12.9 8.6 Inactive 22.9 13.0 11.2 23.5 13.9 9.8 5.8 0 20 40 60 80 100 Percent Greater Tunis North East North West Center East Center West South East South West Source: Based on data from the Labor Force Survey (ENPE), INS. 42 Tunisia’s Jobs Landscape FIGURE 2.12. Distribution of the Employed Population, by Broad Sector, 2006–17 100 80 48.8 49.0 49.8 49.9 51.3 52.1 51.9 51.9 60 Percent 40 32.0 33.2 32.0 33.7 33.4 33.1 33.4 33.3 20 19.2 17.8 18.3 16.4 15.4 14.8 14.8 14.8 0 2006 2008 2009 2011 2013 2015 2016 2017 Agriculture Industry Services Source: Based on data from the Labor Force Survey (ENPE), INS. TABLE 2.4. Trends in Employment, by Industry, 2006, 2011, and 2017 Change Percentage 2006 2011 2017 in level change 2006 2011 2017 Level 2006–17 Share (%) Agriculture 572,689 510,022 509,924 −62,765 −11.0 19.1 16.2 14.7 Food manufacturing 63,719 72,339 92,016 28,297 44.4 2.1 2.3 2.7 Textile manufacturing 256,935 236,036 232,477 −24,458 −9.5 8.6 7.5 6.7 Other manufacturing 243,334 269,732 308,348 65,014 26.7 8.1 8.6 8.9 Construction 355,266 441,686 475,592 120,326 33.9 11.8 14.1 13.8 Other secondary 32,331 31,597 37,590 5,259 16.3 1.1 1.0 1.1 Trade 346,074 388,130 457,393 111,319 32.2 11.5 12.4 13.2 Transports 136,364 175,284 187,140 50,776 37.2 4.5 5.6 5.4 Hotels and restaurants 115,262 106,116 127,977 12,715 11.0 3.8 3.4 3.7 Financial services 26,805 25,737 34,978 8,173 30.5 0.9 0.8 1.0 Real estate and professional services 105,517 133,607 174,050 68,534 65.0 3.5 4.3 5.0 Public administration and health/ 558,063 587,332 657,610 99,547 17.8 18.6 18.7 19.0 education services Other services 163,678 139,648 147,475 −16,203 −9.9 5.4 4.4 4.3 Not defined 28,857 22,505 15,534 −13,323 −46.2 1.0 0.7 0.4 Total 3,004,893 3,139,770 3,458,104 453,211 15.1 100.0 100.0 100.0 Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 43 FIGURE 2.13. Distribution of the Employed Population, by Type of Employment, 2006–17 0.0 0.2 0.0 0.0 0.0 0.6 0.0 100 80 68.2 69.3 71.2 72.0 72.8 72.2 75.1 60 Percent 40 6.9 4.4 4.3 3.8 2.6 2.2 2.8 20 19.3 19.4 18.6 17.0 16.9 17.9 15.5 5.5 5.8 5.8 7.1 6.7 6.1 6.2 0 2006 2008 2011 2013 2015 2016 2017 Employer Own-account worker Contributing family worker Apprentice/Wage worker Other Not stated Source: Based on data from the Labor Force Survey (ENPE), INS. were employed as own-account workers or employers, have primary and secondary education, respectively. About and 12 percent as contributing family workers. 18.5 percent of workers hold a university degree. Around 66 percent are urban residents. Almost 60 percent live in A growing majority of workers (about 80 percent in 2017) coastal areas. The largest share (23.5  percent) is in the are employed in midlevel and low-end occupations. Between Center-East. A multivariate regression of the probability 2006 and 2017, the number and share of high-end occupa- of employment conditional on being in the labor force, tions, including managers, professionals, technicians, and separately by sex, confirms that youth, university gradu- associate professionals, declined by 3.8 percent (27,000) ates, and individuals in inland regions have a lower prob- because of a reduction in the number of managers and ability of being employed (Figure 2.14).15 Conditional on technicians, while the number of professionals rose sizably participating in the labor market, single men and women (Table 2.5). The number of employed in midlevel occupa- have a lower probability of being employed. A simi- tions increased by almost 222 percent (370,000) thanks to lar exercise conducted separately on coastal and inland growth in the number of services and sales workers as regions indicates that men are considerably more likely well as craft and trade workers. The number of workers to be employed relative to women in inland regions, and in elementary occupations expanded by about 17 percent university graduates are significantly penalized in access (99,000). Overall, the number of high-end occupations to jobs in inland regions. Among coastal regions, residing declined from 24 percent to 20 percent, and the share of outside the Greater Tunis area increases the likelihood of midlevel occupations rose by about 3 percentage points to being employed as does living in the North-West among reach around 60 percent in 2017. the interior regions (Figure 2.15). The employed population is largely composed of individuals ages 30–64, men, individuals with primary and secondary 15 A probit regression was estimated on the sample of working-age indi-   education, and residents of urban areas and coastal regions. viduals in the labor force. The dependent variable was equal to 1 for indi- viduals who reported that they were employed in the reference week. The Over 75 percent of the employed population is ages 30 or set of covariates included in the model are as follows: a set of dummies for more; 43 percent is ages 30–44; and they are predominantly different age-groups, a dummy for women, a set of dummies for marital status, a set of dummies for educational level, region of residence, and men (see Figure  2.11). Only about 9  percent of work- cohort of birth, and a dummy for urban residency. In addition, the number ers have no education, whereas 35 percent and 36 percent of children ages 0–5 and 6–15 are included in the specification. 44 Tunisia’s Jobs Landscape TABLE 2.5. Trends in Employment, by Occupation, 2006, 2011, and 2017 Chane in Percentage 2006 2011 2017 level change 2006 2011 2017 Level 2006–17 Share (%) Managers 292,776 211,056 158,985 −133,791 −45.7 9.7 6.7 4.6 Professionals 188,021 188,980 349,803 161,782 86.0 6.3 6.0 10.1 Technicians and associate professionals 244,295 233,765 189,064 −55,231 −22.6 8.1 7.4 5.5 Clerical support workers 167,420 182,317 147,568 −19,853 −11.9 5.6 5.8 4.3 Services and sales workers 304,662 477,734 632,911 328,249 107.7 10.1 15.2 18.3 Skilled agricultural workers, forestry and 490,696 380,211 363,054 −127,641 −26.0 16.3 12.1 10.5 fishing Craft and related trade workers 380,573 414,273 489,465 108,892 28.6 12.7 13.2 14.2 Plant and machine operators and 356,232 383,308 436,565 80,333 22.6 11.9 12.2 12.6 assemblers Elementary occupations 577,235 663,413 676,245 99,010 17.2 19.2 21.1 19.6 Not stated 2,983 4,713 14,443 11,460 384.1 0.1 0.2 0.4 Total 3,004,893 3,139,770 3,458,104 453,211 15.1 100.00 100.00 100.00 Source: Based on data from the Labor Force Survey (ENPE), INS. COVID-19, lockdowns, and the economic crisis have had still above the levels estimated before the pandemic, par- deleterious effects on the labor market. Compared with ticularly among men (Figure  2.16, panel d). Similarly, the first quarter (Q1) of 2020, employment dropped by the youth unemployment rate rose and was estimated at 4.5 percent in Q2, and, after a partial rebound in Q3, it 42.4 percent in Q3 2021, relative to 34.2 percent before continued to decline in the last quarter of 2020 as well as the pandemic. in the first three quarters of 2021 (Figure 2.16, panel b). Total employment was estimated at 3.38 million in the The effects on the labor market translated into a deterio- third quarter of 2021, which is about 3.8 percent (or ration in living standards among Tunisian households.16 almost 133,000 workers) below the level observed one Public sector workers were the least affected by the pan- year earlier. The reduction in employment was signifi- demic, mainly because of a reduction in working hours cantly larger in relative terms among women between Q1 or delays in wage payments (based on data relative to and Q2 of 2020. Women’s employment bounced back November 2020). Formal wage workers were less likely more rapidly, and, in Q3 2021, it was above the prepan- to be temporarily or permanently laid off, whereas infor- demic level (Q1 2020). By contrast, men continued to mal workers suffered a higher probability of being perma- experience job losses, and their employment level was nently laid off (35 percent). Employers and own-account more than 8 percent below the level observed in Q1 2020. workers faced a number of difficulties, mainly because of As of Q3 2020, the partial rebound in employment was a loss in demand (76 percent), difficult access to customers largely ascribable to the dynamic of informal employ- arising from mobility restrictions (74 percent), difficult ment, which rose by 2.6  percent relative to the same access to suppliers (71 percent), and limited availability quarter of 2019. Informal employment increased more of inputs and price increases (75  percent).17 Although rapidly among women than among men (5.5 percent vs. farmers accounted for a small share of total employment, 2.0 percent, respectively; Figure 2.16, panel c). A sectoral they experienced severe challenges, including reductions breakdown indicates that, between Q1 and Q4 2020, the largest reduction in employment (in relative terms) was in 16  The data presented here are based on household phone surveys collected by agriculture and fishing (−9.8 percent), followed by manu- the INS, in collaboration with the World Bank (5 rounds conducted between facturing (−8.1 percent; about 54,000 jobs lost), and the April and October 2020), and on household phone surveys collected by the INS, in collaboration with the Economic Research Forum (one round services sector (−0.9 percent). By contrast, other secondary collected in October/November 2020). The data refer to the population of sectors, mainly construction, posted an increase of about mobile phone owners. Alfani et al. (2021) and Krafft, Assaad, and Marouani (2021) have analyzed the data collected by the INS and the World Bank and 1.8  percent. With the reduction in employment, unem- by the INS and the Economic Research Forum, respectively. ployment rates increased, and, as of Q3 2021, they were 17  Responses refer to the 60 days prior to the survey week. Access to the Labor Market: A Spotlight on Women and Youth 45 FIGURE 2.14. Correlates of Employment, by Sex, 2017 25–29 30–44 45–64 Married Widowed Divorced Primary Secondary Tertiary Not stated Number of children ages 0–5 Number of children ages 6–15 North-East North-West Center-East Center-West South-East South-West Rural –.3 –.2 –.1 0 .1 .2 Marginal effect Men Women FIGURE 2.15. Correlates of Employment, by Region (Coastal vs. Inland), 2017 Men=1 25–29 30–44 45–64 Married Widowed Divorced Primary Secondary Tertiary Not stated Number of children ages 0–5 Number of children ages 6–15 North-East Center-East Rural Center-West South-East South-West –.2 –.1 0 .1 .2 Marginal effect Coastal Inland Source: Based on data from the Labor Force Survey (ENPE), INS. Note: The reference categories are as follows: age 15–24, single, no schooling, residence in Greater Tunis, in urban areas. 46 Tunisia’s Jobs Landscape FIGURE 2.16. Trends in Selected Labor Market Indicators, by Sex, Q1, 2019–Q1, 2021 a. Population in the labor force, by sex b. Employed population, by sex Thousands (total employment) 4,500.0 3,500.0 (Men and women employment) 4,000.0 3,000.0 (Men and women active) Thousands (total active) 4,400.0 3,900.0 4,300.0 3,000.0 3,800.0 2,500.0 4,200.0 2,500.0 3,700.0 Thousands 4,100.0 3,600.0 2,000.0 Thousands 4,000.0 2,000.0 3,500.0 3,900.0 3,400.0 1,500.0 3,800.0 1,500.0 3,300.0 3,700.0 3,200.0 1,000.0 1,000.0 3,600.0 3,100.0 3,500.0 500.0 3,000.0 500.0 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q1 Q2Q3Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 2019 2020 2021 2019 2020 2021 All Men Women All Men Women c. Employed population, by formality status d. Unemployment rates, by sex 2,400.0 26.0 2,200.0 24.0 2,000.0 22.0 Thousands 20.0 Percent 1,800.0 18.0 1,600.0 16.0 1,400.0 14.0 1,200.0 12.0 1,000.0 10.0 Q2 Q3 Q4 Q3 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 2019 2020 2019 2020 2021 Formal Informal All Men Women Source: Based on data from the Labor Force Survey (ENPE), INS. in inputs and a drought that led to a drop in revenues, and 77  percent of farmers experienced or expected smaller Gender Gaps harvests). Together with employers, own-account workers, LABOR FORCE PARTICIPATION and informal wage workers, farmers posted large declines in household income, and only 18 percent of the relevant There are important economic and social equity arguments respondents reported in February 2020 that they had received for improving labor market conditions among women. some form of government assistance. The main coping mech- From a merely economic perspective, the low participa- anisms consisted of savings (50 percent), social networks tion of women constitutes an underutilization of human (national, 45 percent; international, 10 percent), financial resources, particularly in light of the humongous progress assistance from banks or other lenders (11 percent), or sale achieved in education whereby girls are outstripping boys of assets (15 percent). A major consequence of the nega- in educational outcomes. Lower participation of women tive effects on the labor market, combined with difficulties contributes to lower incomes and living standards and can in access to services and increases in prices, was a wors- translate into higher poverty rates. Population aging exac- ening in living standards. About 1 Tunisian household in erbates the issue. A rising share of elderly and a shrinking 2 reported a decline in welfare in October relative to before workforce will need to be counterbalanced with rising the pandemic; the proportions were more than 6 in 10 participation among groups with low activity rates. Raising among households in the bottom 40. About 1 household labor force participation among less well educated women, in 5 declared that there had been a worsening throughout many of whom are among the poor and the bottom 40, the pandemic, that is, between May and October. Mean- can also make growth more inclusive. In addition, gender while, the pandemic accelerated a digital transformation equality and women’s empowerment are fundamental (Box 2.2). rights and important development objectives on their own, Access to the Labor Market: A Spotlight on Women and Youth 47 BOX 2.2. Digital Labor Platforms The outbreak of COVID-19 and the lockdowns introduced in many countries to contain the spread of the virus accelerated a digital transformation that has been under way for decades. Millions of citizens worldwide moved online. Children with internet access at home attended virtual classes. Many employees, particularly those in midlevel and high-end occupations, started to work from home. Many firms adopted digital business models to continue their operations and minimize revenue losses. At the same time, digitalization helped contain the pandemic, for example through the use of mobile applications developed to track and trace infected individuals and their contacts. Yet, the pandemic also exacerbated inequalities associated with gaps in access to digital technologies across countries and, within countries, across less and more affluent households. Telephone penetration is greater in Tunisia than in most of the developing countries in the Maghreb, excluding Algeria. At about 86 percent, 4G coverage is the second highest in Tunisia after Morocco (AUC and OECD 2021). The share of enter- prises with a website is estimated at 66 percent on average, with a peak at 81 percent among large businesses. Together with Morocco, at 69 percent, these are the highest rates in the Maghreb (AUC and OECD 2021). The share declines with firm size. Thus, it ranges between 81 percent among large businesses and 59 percent among small businesses (AUC and OECD 2021). This excludes small, informal production units, among which the share with a website is estimated at less than 2 percent. About 9.4 percent report that they use the internet, and a similar share report that they use computers (10.5 percent).a In addition to good internet coverage and high mobile phone penetration, the provision of electronic payment methods, well-developed fintech, and adequate transport infrastructures is key to fostering the development of digital platforms for the exchange of goods, services, and labor. This could help sustain a service-led growth model, whereby global innovator services, such as information and communication technology (ICT), finance, and professional services, coexist with low-skill domestic and tradable services to create more higher-productivity job opportunities for all. Global innovator services are intensive in skilled labor, but generate positive spillover effects in other sectors, including manufacturing, thanks to their links and to the greater demand induced by higher incomes (Nayyar, Hallward-Driemeier, and Davies 2021). Low-skill services may find more opportunities on larger markets because of digital platforms and the incentive to scale up based on intangible capital (Nayyar, Hallward-Driemeier, and Davies 2021). Digital labor platforms provide a new way to boost labor demand and labor supply and also expand labor demand by increasing the size of the market. On online web-based platforms, tasks may be performed remotely, for instance, in legal and financial services, software development, translation services, programming, and data analysis, and there is no geographical limit to the size of the market. In the case of location-based platforms, work is carried out in person in physical locations identified by workers; this may include, for example, taxi services, delivery and home services, domestic work, and care services. The first type of plat- forms are likely to expand job opportunities among well-educated workers, whereas location-based platforms can provide an additional and more efficient way to match the demand or and supply of low-skill labor. Both have the potential to foster labor force participation and employment among women. According to surveys conducted by the International Labour Organization (ILO), the majority of workers on digital labor plat- forms are young (ages below 35) and well educated (ILO 2021). Women, too, are on such platforms, though they contribute 4 workers in 10 on online web-based platforms and only 1 worker in 10 on location-based platforms because of the sectoral composition of jobs on location-based platforms (ILO 2021). Evidence also shows that greater flexibility, better pay, and lack of alternative job opportunities are the main factors that push workers to use location-based platforms. Virtually all workers find their main sources of labor income on these platforms, and earnings can be higher than in traditional sectors, as in the case of app-based taxi and delivery services. Digital platforms are no panacea, however, and the opportunities they provide are often accompanied by important challenges. Regularity of work and incomes, working conditions, access to social protection, the right to collective bargaining, and discrimi- nation and harassment are examples of issues commonly reported by workers engaged on digital labor platforms (ILO 2021). In addition, unfair competition is often cited as the most important issue traditional businesses raise because some platforms are not subject to the conventional tax and regulatory framework. To take full advantage of digital platforms, the regulatory framework should guarantee low barriers to market entry to check the market power of digital incumbents and to allow new entrants to keep their incentives to compete (World Bank 2019). Moreover, in the case of location-based labor platforms that can help mediate labor in traditional sectors, such as low-skill labor in construction or food services, the degree of competition in the sectors in which employers are seeking to hire labor through digital platforms is crucial. Intermediating labor through digital platform will likely not generate more or higher- quality jobs if there is only one or a handful of firms operating in the sector that can therefore command the prices of the products, services, and labor they use. Some governments have successfully introduced regulatory responses that can help improve the working conditions on digital platforms,. Several have extended social security to platform workers, including coverage of accident insurance costs paid by platforms, the extension of social security, and the provision of work injury and death benefits and sick and unemployment benefits (ILO 2021). (continued) 48 Tunisia’s Jobs Landscape BOX 2.2. Digital Labor Platforms (continued) In Tunisia, the social contract that has for decades hinged on a large public sector and state-owned enterprises (SOEs) to deliver on the promise of job creation has failed. High unemployment rates among university graduates and the large number of workers employed informally with low incomes and little protection represent an urgent call to action. Job creation cannot be a responsibility of the public sector and a few well-connected incumbents. A new way forward might take advantage of e-commerce, fintech, digital labor platforms, and of the digital economy more generally to establish a new equilibrium, whereby new markets are reached and more job opportunities with better working conditions are created. Yet, competition on and off digital platforms in shielded sectors and the regulatory framework of digital platforms are critical to reducing the risk of generating new cohorts of unprotected workers with low earnings. a. Figures based on the 2016 Survey of the Economic Activities of Micro-Enterprises (ENAE). The sample is extracted from the business registry and covers nonagricultural microproduction units—that is, fewer than six employees and revenues below TD 1 million a year—that have tax IDs, operate from fixed premises, and do not undertake precise accounting. as established by the United Nations Sustainable Develop- regional average (18 percent in 2017) (Figure 2.17). Over ment Goals. Globally, gender gaps in labor market out- time, the participation of women in Tunisia has increased. comes have been remarkably resistant to change, despite In 2017, 26.5 percent of working-age women participated progress in other dimensions of gender equality (Klasen in the labor market, compared with 24.4 percent in 2006 2019a; World Bank 2011, 2014a). (Figure 2.18). Estimated at 41.8 percentage points in 2017, the gender gap in labor force participation rates is strik- Despite some progress, women’s labor force participation is ing, although it narrowed by about 1 point relative to 2006 low in Tunisia relative to international standards and rela- thanks to a less rapid increase in participation rates among tive to men. On average, labor force participation among men relative to women (1.5 percent vs. 8.6 percent, respec- women has been around 25  percent over the decade, tively, between 2006 and 2017). which is about half the average rate among OECD coun- tries (51.5 percent), about 20 percentage points below the Women’s labor force participation is particularly low average among middle-income countries (45.2  percent in the center and south as well as in rural areas. The average in 2017), and about 8 percentage points higher than the women’s labor force participation hides significant disparities FIGURE 2.17. Women’s Labor Force Participation Rates (Ages 15+), Tunisia and the Rest of the World, 1990–2019 100 80 LFP rate - Women 60 40 20 0 6 8 10 12 Log - GDP per capita (2017 PPP) Rest of the World Tunisia MENA Lower middle income countries C.I. Quadratic t Singapore Source: Based on data from the World Development Indicators, World Bank. Access to the Labor Market: A Spotlight on Women and Youth 49 FIGURE 2.18. Labor Force Participation Rates, by (Figure 2.18).18 Composition-free and in-sample predicted Sex, 2006–17 probabilities line up well around the 45-degree line. In par- ticular, governorates with the lowest women’s labor force 100 participation rates have a larger number of young women, 90 80 married women, and households with children ages under 15. 70 In addition, considerable differences in the share of women 60 with tertiary education are observed between the north and Percent 50 south of the country. However, a few governorates, namely, 40 Beja, Bouzide, Jendouba, Kairouan, Kasserina, Sidi Mahdia, 30 and Siliana, have low female participation rates that cannot 20 be explained by women’s observable characteristics. These 10 governorates are located in the north and center of the 0 country, and governorate-specific characteristics, in addi- 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 tion to unobservable characteristics of women, seem to play All Female Male an important role in this case. Source: Based on data from the Labor Force Survey (ENPE), INS. The gender gap in labor force participation widens with age and peaks among individuals ages 30–44. Labor force across governorates and regions and between urban and participation among both men and women is low among rural areas. Governorates in the center and south show sig- the youngest age-group, ages 15–24, thanks to school nificantly lower women’s participation rates. For example, attendance. In the 15–24 age-group, about 4 boys in 10 par- in Kasserine, the rate is estimated at 15.7 percent relative ticipated in the labor market in 2017 compared with to 35.4 percent in Tunis (Map 2.1). Substantial differences 2 girls in 10 (Figure 2.21). The gap is ascribable to a higher exist across regions as well as between urban and rural school attendance rate among girls, but also to a larger areas overall and between urban and rural areas within share of girls who are not in education and not in the labor each region (Figure 2.19). At the regional level, gender par- force. The gap widens in the next two age-groups, 25–29 ticipation gaps range from 35.3 percent in Greater Tunis and 30–44, where it reached 35 and 53 percentage points to 49.4 percent in the South-East. However, the largest in 2017 because more women fail to enter or exit the labor geographical differences are detected between urban and market as they grow older. However, the gap narrowed rural areas. The average participation rate of women in over the decade thanks to sizable increases in participa- rural areas is 18.3 percent, compared with 30.3 percent in tion rates among women in the corresponding age-groups urban areas. The gender gap is much wider in rural areas, at (about 6 and 7 percentage points), whereas participa- over 51 percentage points, than in urban areas, where it is tion among men of the same age declined by almost 1 per- estimated at 37.5 percent. (This is also ascribable to slightly centage point. The gender gap in participation is even larger lower rates among men in urban areas relative to rural among the population ages 45–64 (59 percentage points areas.) In rural areas of southern regions and the North- in 2017) and has remained constant over time because of a West, the gender gap is estimated at 54–55 percentage points similar reduction in labor market engagement among both (Figure 2.19). men and women in the same age-group. Differences in women’s observable characteristics are a key The younger cohorts of women exhibit higher participa- factor in explaining geographic differences in women’s labor tion rates than older cohorts of women. Average figures force participation, but other factors play a role in some hide important variations across cohorts and throughout governorates. A simple exercise to understand whether most the life cycle (Figure 2.22). Among women, both cohort of the differences in participation rates among women across governorates are driven by women’s characteristics rather than other factors, such as infrastructure, institu- 18  The exercise consists in estimating separate equations for each governor- ate of the probability of women participating in the labor market based on tions, discrimination, or social norms, indicates that individual, household, and geographical characteristics. In a second step, a large part of the gaps can be ascribed to differences probabilities of participating are predicted for each governorate based on the governorate-specific estimate coefficients in the sample (the characteris- in demographics, including age, marital status, the pres- tics of women living in each governorate) and the characteristics of women ence and number of children, and educational attainment out of sample, that is, in the entire country. 50 Tunisia’s Jobs Landscape MAP 2.1. Female Labor Force Participation Rates, by Governorate, 2017 (34,40] (28,34] (22,28] (16,22] Bizerte [10,16] 26.2 Ariana 32.3 Manouba Tunis Beja 28.6 35.4 Jendouba Nabeul 20.9 Ben Arous 16.2 36.1 36.4 Zaghouan 30.4 Le Kef Siliana 29.5 21.9 Sousse 26.8 Kairouan Monastir 19.6 32.9 Mahdia Kasserine 20.8 15.7 Sidi Bouzide 16.9 Sfax 26.5 Gafsa 21.6 Tozeur 20.1 Gabes 20.9 Kebili 21.3 Mednine 16.9 Tataouine 19.8 Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 2.19. Female Labor Force Participation Rates, by Region and Urban and Rural Areas, 2017 Urban Rural 100 100 80 80 60 60 Percent Percent 40 40 20 20 0 0 Greater Tunis North-East North-West Center-East Center-West South-East South-West Greater Tunis North-East North-West Center-East Center-West South-East South-West Female Male Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 2.20. The Role of Observable Characteristics of Women in Gaps in Women’s Labor Force Participation Across Governorates, 2017 .45 NABEUL .4 ZAGHOUAN MONASTIR BEN AROUS LE KEF TUNIS SILIANA Composition-free probability .35 BEJA MANOUBA BIZERTE ARIANA MAHDIA SFAX KAIROUAN .3 SIDI BOUZIDE SOUSSE JENDOUBA KASSERINE TATAOUINE .25 TOZEUR GAFSA GABES KEBILI MEDNINE .2 .15 .15 .2 .25 .3 .35 .4 .45 In-sample probability Source: Based on data from the Labor Force Survey (ENPE), INS. 52 Tunisia’s Jobs Landscape FIGURE 2.21. Labor Force Participation Rates, by Sex and Age-Group, 2006–17 15−24 25−29 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 100 100 80 80 60 60 Percent Percent 40 40 20 20 0 0 30−44 45−64 100 100 80 80 60 60 Percent 40 Percent 40 20 20 0 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 All Female Male Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 2.22. Female Labor Force Participation Rates, by Cohort Over the Life Cycle, 2006–17 a. Men b. Women 100 100 90 90 80 80 70 70 60 60 Percent Percent 50 50 40 40 30 30 20 20 10 10 0 0 15 20 25 30 35 40 45 50 55 60 15 20 25 30 35 40 45 50 55 60 Age Age 1960–69 1970–79 1960–69 1970–79 1980–89 1990–99 1980–89 1990–99 Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 53 and life-cycle effects are important, whereas cohort effects higher educational attainment typically participate in larger do not play much of a role among men. Men tend to reach numbers in the labor market and exhibit a greater degree high participation rates, above 90 percent, and younger of attachment to labor. Tunisian women are no exception (older) cohorts of men increase (decrease) their participa- (Figure 2.23, panel a). In 2017, fewer than 1 woman in 10 tion as they enter (exit) the labor market, whereas par- with no schooling participated in the labor market. This ticipation rates among middle cohorts remain constant at compares with over 2 women in 10 with primary educa- high levels (Figure 2.22, panel a). By contrast, a look at tion, almost 3 women in 10 with secondary education, and female labor force participation at various ages indicates more than 6 women in 10 with tertiary education. The last that, at any age, women born more recently participate in is close to the average rate among men (68.3 percent). Over the labor market in greater numbers than women in pre- the decade, participation rates declined among women vious cohorts. For example, about 45 percent of women with no schooling or with primary education, while they born in the 1990s were participating in the labor force increased among well-educated women, particularly among at age 25, while fewer than 40 percent of women born women with tertiary education (7 percentage points). The a decade earlier were participating in the labor force at additional key element has been the change in the composi- age 25 (Figure 2.22, panel a). Similarly, at age 35, 41 per- tion of the working-age population by educational level. cent of women born in the 1980s are active in the labor The share of women with tertiary education in the female market, while the corresponding share at age 35 among working-age population rose by over 8 percentage points, women who were born in the 1970s is 35 percent. About from 8.9 percent in 2006 to 17.1 percent in 2017. At the 31  percent of women born in the 1970s were active in same time, the share of women with no education declined the labor market at age 45, compared with about 26 per- from 32 percent to 27 percent. cent of women who were born in the 1960s. This is not an age effect given that the various cohorts are compared High educational attainment leads to a higher degree of at the same age. This seems to point to a set of factors labor market attachment throughout the life cycle. In that are positively correlated with women’s participation addition to participating in the labor market on average that improved over time. Educational attainment, norms almost as much as men, women with tertiary education and attitudes, and conditions of employment are potential maintain an attachment to the labor market throughout candidates. the life cycle that is similar to the attachment of men (see Figure  2.23, panel b). The association between tertiary Higher educational levels among women contribute to education and labor market participation remains strong greater participation in the labor market. Individuals with at various ages. Women with a university degree enter the FIGURE 2.23. Female Labor Force Participation Rates, by Educational Level Over Time and Over the Life Cycle, 2006–17 a. Over time b. Over the life cycle 100 100 90 90 80 80 70 70 60 Percent 60 Percent 50 50 40 40 30 20 30 10 20 0 10 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 0 2017 15 20 25 30 35 40 45 50 55 60 65 None Primary None Primary Secondary Tertiary Secondary Tertiary Source: Based on data from the Labor Force Survey (ENPE), INS. 54 Tunisia’s Jobs Landscape FIGURE 2.24. Female Labor Force Participation Rates, by Marital Status Over Time and Over the Life Cycle, 2006–17 a. Over time b. Over the life cycle 100 100 90 90 80 80 70 70 60 Percent 60 Percent 50 50 40 40 30 30 20 20 10 0 10 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 15 20 25 30 35 40 45 50 55 60 65 Single Women Married Women Single Women Married Women Men Men Source: Based on data from the Labor Force Survey (ENPE), INS. labor market at older ages relative to their less well edu- that marriage, pregnancy, and childcare deter women from cated counterparts, but their participation rates rise rap- (re)entering the labor market because of prevailing social idly to over 80  percent by age 30. It increases again at norms and gender roles that attribute to women the pri- ages 30–44 and hovers around 86 percent on average (the mary role of caregiver. This seems to be corroborated by rate among male university graduates is 97 percent at ages the main reason for not being engaged in the labor mar- 30–44). The association, however, is not as strong at lower ket that most women report. Most inactive women ages levels of education, where women are outperformed by 25–29 and over 90  percent of women ages 30 or more men by at least 30 percentage points, suggesting that high mention household responsibilities, whereas men rarely education is one of the key factors that can help close the mention housework as a reason for not looking for work participation gap by sex. (Figure  2.25). The share of women reporting household responsibilities as the main reason for not working differ Marriage, childbirth, and childcare responsibilities are key considerably at young ages across quintiles of household explanatory factors of female labor force participation. consumption expenditures (Figure 2.26; Box 2.3). This is Women’s marital status can explain large differences largely ascribable to the fact that young women in richer in women’s participation rates. The average age at first households can afford to stay in school longer. The dif- marriage is 22.5 (based on the 2014 population census), ference in the share of women reporting household duties and about 80 percent of women are married by age 36.19 narrows considerably as women grow older. Participation rates among single and married women increased virtually at the same rate over the decade. The A multivariate analysis confirms the correlation between share of married women who participate in the labor observable characteristics and women’s labor force par- force thus remains about 6 percentage points below that ticipation. All the bivariate correlations illustrated so far of single women (estimated at 27.2 percent and 33.5 per- can be combined in a multivariate analysis of labor force cent, respectively, in 2017) (Figure 2.24, panel a). The gap participation (Figure  2.27).20 Women’s participation rises between married and single women widens over the life at young ages and then progressively slows at older ages. cycle, with a more rapid rise in participation among single women (Figure 2.24, panel b). This gap expands during 20  A probit model is fit by regressing a dummy for participation—taking the the early stages of the life cycle, reaching a peak of about value of 1 if a woman participates in the labor market and 0 otherwise—onto 31 percentage points in the late 20s and fluctuating between a set of individual and household characteristics, including a second-degree polynomial in age, dummies for year of birth cohort, dummies for residing in 1 and 12 points after age 40. This pattern might indicate each of the Tunisian regions, a dummy for urban areas, dummies for marital status, dummies for educational level, household age composition (presence of children ages under, presence of children ages 2–5, presence of children ages 19  This includes women who, by age 35, are divorced or widowed. 6–15, and presence of household members ages 65 or more). Access to the Labor Market: A Spotlight on Women and Youth 55 FIGURE 2.25. Reasons for Not Working Over the Life Cycle, by Sex, 2015 a. Women b. Men 100 100 90 90 80 80 70 70 60 60 Percent Percent 50 50 40 40 30 30 20 20 10 10 0 0 16–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 16–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 Student Housework Student Housework Source: Based on data from the EBCNV 2015, INS. Note: Respondents could also mention other reasons. Responses therefore do not sum to 100 percent. Married women are about 12 percent less likely to enter the that plays the most important role. It is estimated that, rela- labor market relative to single women (the negative effect tive to having no education, having tertiary education was declines in more recent years), while divorced or separated associated with a 40 percent greater chance of entering the women are relatively more likely to do so (about 6 percent labor market in 2017, which is higher than the effect esti- in 2017; again, the effect is attenuated in more recent years). mated 10 years earlier. The effect of secondary education The important role of the level of education, particularly ter- is sizable, too (16  percent in 2017), and it has increased tiary education, in shaping women’s participation decisions over time. Household composition matters. The presence is confirmed. While every woman with some level of educa- of young children (ages less than 2) reduces the probability tion is more likely to participate in the labor force compared of participating by about 3 percent. Similarly, the presence of with a woman with no schooling, it is tertiary education children ages 2–5 reduces participation by about 2 percent. The presence of children ages 6–15 has a similar negative FIGURE 2.26. Share of Inactive Women Reporting effect. By contrast, the presence of elderly people (ages 65+) Household Duties as a Main Reason for Not Working in the household does not seem to be a major obstacle to Over the Life Cycle, by Quintile of Household women’s engagement in the labor market in most years, Expenditure, 2015 but its effect is significant and estimated at −1.3 percent in 2017. Women in rural areas as well as women in all 100 regions except the North-East are less likely to participate 90 in the labor market relative to women in urban areas and 80 in Greater Tunis. 70 60 Percent 50 Observable characteristics leave a large part of the gender 40 gap in labor force participation unexplained. This is because 30 individual and household characteristics controlled for 20 in the analysis are not able to account for most of the 10 variation observed in women’s labor force participation. 0 This is corroborated by a Blinder-Oaxaca decomposition of the participation gap between men and women (Fig- 16–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 ure 2.28, panel a).21 Nearly all the difference in participation Poorest Q2 Q3 Q4 Richest  Figure A 2–1 in annex illustrates the effect of covariates broken down by 21 groups, including demographics, marital status, educational level, house- Source: Based on data from the EBCNV 2015, INS. hold composition, and location. 56 Tunisia’s Jobs Landscape BOX 2.3. Women’s Labor Market Participation Along the Household Welfare Distribution Close to 70  percent of the Tunisian population ages 25–54 is married. The share of married couples declined slightly, from 78 percent in 2006 to 77 percent in 2017, and the share of two-earner households among married couples hovered around 19 percent over the decade. Hence, for many workers, decisions concerning labor force participation and employment take place with a partner within a household. Household economic status plays an important role (and is endogenous) to the labor market decision of household members. For example, in poor households, the employment of an additional member can raise household income considerably and potentially push the household above the poverty line. By contrast, women in affluent households might be able to afford to be inactive because of the prevalence of an income effect. In Tunisia, the 2015 household budget survey allows an investigatio into the patterns of labor market participation among women along the distribution of household consumption expenditure (Figure B 2.3-1).a First, labor force participation among single women monotonically decreases along the consumption distribution. Thus, a larger share of single women in the poorest households participate in the labor market (62.7 percent) compared with women in richer households (52.5 percent in the top quintile). These figures include all women ages 15 and above who are single, excluding widows and divorced women. This group largely consists of young women who still live with their parents: 86 percent are single women ages 15–34, and over 45 percent are still in education. The declining pattern along the welfare distribution likely captures an income effect, whereby young single women among more affluent households can afford to continue pursuing educational goals. This pattern is also consistent with qualita- tive interviews, which show that young women’s engagement in the labor market, especially in work that is considered below their educational qualifications, is often driven by the economic needs of their households (World Bank 2014a). FIGURE B 2.3.1. Female Labor Force Participation Rates, by Marital Status and Quintile of Household Consumption Expenditure, 2015 100 80 60 Percent 40 62.7 60.0 59.9 58.1 52.5 47.5 20 37.3 40.0 40.1 41.9 0 1 2 3 4 5 Quantile of Household Consumption Single Women Married Women Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. Second, married women display the opposite pattern. Their labor force participation rate increases along the consumption dis- tribution, from 37.3 percent in the first quintile to 47.5 percent in the richest quintile. While it is difficult to unpack the reasons behind such a trend, some hypotheses may be advanced. Married women in richer households are typically more well educated relative to women living in the poorest households, and, because of assortative mating, they also tend to be married to well- educated husbands (Table B 2.3-1).b This might lead to the prevalence of the substitution effect over the income effect, whereby the price of leisure is higher among well-educated women. (continued) Access to the Labor Market: A Spotlight on Women and Youth 57 BOX 2.3. Women’s Labor Market Participation Along the Household Welfare Distribution (continued) TABLE B 2.3.1. Correlation Between Educational Level of Heads and Spouses, by Quintile of Household Consumption Expenditure, 2015 Spouse’s educational level Bottom quintile None Primary Secondary Tertiary Total Head’s None 31.0 3.6 0.8 0.1 35.4 educational Primary 22.0 22.2 5.0 0.4 49.6 level Secondary 3.7 5.6 4.0 0.3 13.6 Tertiary 0.3 0.5 0.4 0.3 1.4 Total 57.0 31.9 10.1 1.0 100.0 Top quintile None Primary Secondary Tertiary Total None 5.7 2.0 0.4 0.0 8.1 Primary 5.7 12.1 5.2 1.3 24.3 Secondary 2.6 10.0 19.4 5.7 37.6 Tertiary 0.3 2.2 8.9 18.6 30.0 Total 14.2 26.2 34.0 25.7 100.0 Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. Well-educated women also have access to formal wage jobs that offer greater security and protection and provide access to maternity leave that can make reentry into the labor force after a pregnancy easier. The share of married women in wage employment increases along the welfare distribution from 58 percent in the lowest quintile to 84 percent in the top quintile at the expense of unpaid family work and own-account work (Table B 2.3.2). Among wage workers, the share of married women holding a public sector job ranges between 14 percent in the lowest quintile and 63 percent in the top quintile (Table B 2.3.2). Similarly, the share of married women employed in formal wage jobs in the private sector rises monotonically along the welfare distribution from 27 percent in the lowest quintile to 76 percent in the top quintile, respectively, monotonically increasing along the welfare distribution. Childcare is likewise more affordable among more affluent households because both parents have higher incomes from labor relative to parents in the poorest quintile. Perceptions of social norms might also be softer among well-educated couples that feel less constrained by traditional gender roles that assign wives the primary role as caregivers. TABLE B 2.3.2. Married Women Employed, by Type of Wage Employment and Quintile of Household Consumption Expenditure, 2015 Wage work Public sector Private sector formal Quintile % of all employment % of wage employment 1 58.2 13.7 26.7 2 64.4 21.8 40.9 3 71.8 30.9 55.1 4 75.2 45.1 64.7 5 84.4 63.4 76.3 Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. a. Estimates of participation rates among men and women derived from the 2015 household budget survey and the 2015 labor force survey (second quarter) differ to some extent. Men’s participation rate is estimated at 68 percent in the household budget survey and at 68.8 percent in the labor force survey, whereas women’s participation rate is estimated at 30.2 percent in the former and at 26 percent in the latter. b. Following Fortin and Schirle (2006), assortative mating is defined as the likelihood of a person in labor income decile i to be married to a spouse in the same labor income decile, according to their respective labor income distribution. Lack of data on labor income in the household budget survey and relying on the strong correlation between educational level and income from labor at the individual level, Table B 2.3 1 shows the percentage of married couples sorted by the husband and wife’s educational levels in 2015. The degree of assortative mating is captured by the percentage of couples along the main diagonal. 58 Tunisia’s Jobs Landscape FIGURE 2.27. Correlates of Labor Force Participation, by Sex, 2017 age age squared Married Widowed Divorced Primary Secondary Tertiary Not stated Presence of children ages 0–1 Presence of children ages 2–5 Presence of children ages 6–15 Presence of children ages 65 and over North-East North-West Center-East Center-West South-East South-West Rural –.2 0 .2 .4 Marginal effect Women Men Source: Based on data from the Labor Force Survey (ENPE), INS. Note: The reference categories are as follows: single, no schooling, Greater Tunis region, urban areas. FIGURE 2.28. Oaxaca-Blinder Decomposition of the Gender Gap in Labor Force Participation and Counterfactual Labor Force Participation Rates, 2006–17 a. Participation gap, men and women b. Men and women assumed to have the same characteristics 0 80 –5 70 –10 Percentage points 60 –15 Percent 50 –20 –25 40 –30 30 –35 20 –40 2006 2008 2009 2011 2013 2015 2016 2017 2006 2008 2009 2011 2013 2015 2016 2017 Men Difference Explained Unexplained Women with Men’s Characteristics Women with Women’s Characteristics Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 59 rates—about 36 percentage points in 2017—is driven by FIGURE 2.29. Employment-to-Population Ratios, the unexplained component; fewer than 2 percentage points by Sex, 2006–17 are accounted for by observable characteristics. The unex- 100 plained part captures differences in the estimated parameters 90 of the labor force participation equation of men and women 80 as well as differences in unobservables. A large part of 70 the differences in coefficients is ascribable to demographics 60 Percent and marital status (annex Figure A 2.1), factors that strongly 50 40 correlated with participation in the bivariate correlations 30 illustrated above. In addition, the regression does not con- 20 trol for some important factors, such as the supply and cost 10 of childcare and eldercare services within regions, which 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 could account for some of the unexplained component. In addition, the choice of curricula women decide to pursue All Female Male in school might partly account for the difficulty in findings jobs and might, in the medium term, discourage women Source: Based on data from the Labor Force Survey (ENPE), INS. from entering the labor market. Nonetheless, the magnitude of the unexplained component might also point to factors employment increased by over 20 percent, which is about other than socioeconomic and demographic characteris- three times the rate of increase in the female population of tics. Cultural values and social norms assign to women a working age. Among men, employment growth was slightly traditional role as the main providers of child and elder- above the rate of increase in the population of working age. care, household chores, and other nonmarket activities and Therefore, beginning in 2011, the employment-to-population might dominate over the empowering effect of education ratio increased modestly among men and by about 2.5 per- among women with less than tertiary education. A simple centage points among women (from 18.1 percent in 2011 econometric exercise shows that, even if women had exactly to 20.5 percent in 2017), and the gender gap narrowed by the same characteristics observed among men, including almost 2 percentage point, but was still slightly above the level age, educational level, marital status, and so on, the partici- observed in 2006. pation rate predicted by the multivariate regression would increase a little, but would still be far lower than that of men Women continue to lag men in terms of jobs access, and (Figure 2.28, panel b).22 unemployment rates are considerably larger among women than among men. In 2006, women’s unemployment rate was estimated at 15.1 percent, compared with 11.5 percent EMPLOYMENT AND UNEMPLOYMENT among men, with a gender gap of about 3.6  percentage Despite a still large gender gap, women’s access to jobs has points (Figure  2.30). The latter expanded over time and improved since 2011. In 2006–17, the rate of employment peaked at over 12  percentage points in 2011 and then creation was not sufficient to keep up with the increase in gradually declined to about 10 points in 2017. This trend the working-age population, particularly among women. is largely ascribable to a drastic increase in unemployment Yet, an important distinction needs to be made between the among women between 2006 and 2011, when the number pre- and post-2011 periods (Figure 2.29). Between 2006 of unemployed women increased by 100 percent to reach and 2011, the rate of employment creation was almost over 280,000 unemployed women in 2011, and then began half the rate of growth of the working-age population, a slow decline thereafter. The number of unemployed men and the total number of employed women declined from increased by 46  percent during the first time span, and 787,000 to 745,700. Between 2011 and 2017, women’s it declined more rapidly than among women thereafter (−15 percent between 2011 and 2017), thus contributing to a reduction in the gender gap in unemployment rates relative to the level reached in 2011. 22 A probit regression is estimated by regressing the participation dummy   upon the same set of controls described above. Estimated coefficients are then used to generate a prediction of the probability of women participat- Lack of job opportunities and limited geographical mobility ing in the labor market as if they had men’s characteristics. Thus, counter- factual women’s participation rates are predicted by applying coefficients translates into large inland and coastal gaps in unemploy- estimated for women onto the distribution of men’s characteristics. ment rates among women. Large gaps exist in labor market 60 Tunisia’s Jobs Landscape FIGURE 2.30. Unemployment Rates, by Sex, 2006–17 Thus, women ages 15–24 and 25–29 exhibited unemploy- ment rates of 37.9 percent and 41.9 percent in 2017 (Fig- 30 ure 2.32). In addition, gender gaps in the 25–29 and 30–44 25 age-groups expanded considerably. Virtually no difference 20 was detected in unemployment rates within the oldest age- group, ages 45–64; the rates remained roughly constant Percent 15 over the decade. In 2006–11, the number of unemployed 10 rose rapidly across all age-groups and among both men and women. Beginning in 2011, the number of young 5 unemployed women started to decline, from over 89,000 0 to 65,500 (ages 15–24) and from 111,500 to 99,600 (ages 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 25–29). By contrast, the number continued to rise among the older age-groups. A similar pattern is observed among All Female Male unemployed young and middle-age men. Source: Based on data from the Labor Force Survey (ENPE), INS. In addition to being more likely to participate in the labor outcomes among both men and women across regions and market, women with tertiary education are also more likely governorates. This is driven by the different employment to land a job. Large gaps in employment-to-population opportunities and limited economic development of inland ratios exist across women (and men) at different educational regions. However, geographical unemployment gaps among levels (Figure  2.33). For example, in 2017, the employ- women reach peaks of 50 percentage points (21 percentage ment ratio of women with no schooling was estimated points in the case of men in 2017) across governorates at 8.4 percent, compared with 19 percent among women (Map 2.2). In the governorate of Monastir, located along with primary education and 37.7 percent among women with the coast (Center-East region), the unemployment rate is university degrees. Employment ratios declined across all estimated at 8.4 percent, compared with 58.6 percent in the educational levels, except secondary education, which governorate of Kebili, in the South-West region. In the case bounced back beginning in 2011. The largest gender gap is of women, the rapid improvement in educational outcomes observed among workers with primary and secondary edu- is reflected in more severe issues in terms of labor market cation, where women employment ratios stood at 19 per- insertion, particularly in rural areas and inland regions that cent and 23  percent, respectively, relative to 70  percent have fewer job opportunities. Young women face more dif- and 59 percent among men. The large rise in the supply of ficulties in moving to areas where economic opportunities new cohorts of women with university degrees outpaced flourish and end up being more constrained than men by the the capacity of the economy to absorb it. Unemployment lack of opportunities local labor markets offer (Box 2.4). rates among women with tertiary education skyrocketed from 25.5 percent to 41.8 percent in 2006–11 and hovered Young women lag older women and men in access to jobs. around 40 percent thereafter (Figure 2.34). The gender gap Employment-to-population ratios are the lowest in Tunisia was constant at around 20 percentage points. among adolescent girls (Figure 2.31). Ratios of 12.1 percent and 28.0 percent in the 15–24 and 25–29 age-groups, respec- The majority of working women are employed in wage tively, compared with estimates of 32.5 percent in the 30–44 work. Most of the employed population works for a wage age-group and 17.8 percent in the oldest age-group (45–64) in Tunisia, and wage employment has gained importance in 2017. The gender gap declined significantly between men over time, from 67.8 percent in 2006 to 75 percent in 2017 and women ages 25–29 and 30–44, whereas it stayed virtu- (Figure 2.35). Women are employed as wage workers in ally unchanged with respect to the youngest and oldest an even greater proportion than men, 85 percent in 2017, age-groups. Between 2011 and 2017, employment among compared with 71.4  percent among men. The share in female youngsters started to increase rapidly, more rap- wage employment rose at a much more rapid rate among idly than the growth in the total population within the women than among men over the decade; the shift occurred corresponding age-group. This occurred in parallel with at the expense of lower contributions of women to house- an ongoing decline in employment among young men. hold duties (3.7 percent in 2017) and own-account work Similarly, women ages 25–29 posted the highest unem- (8.0 percent in 2017), while the share of women employers ployment rate across age-groups in comparison with men. increased marginally (2.5 percent in 2017). Access to the Labor Market: A Spotlight on Women and Youth 61 MAP 2.2. Unemployment Rates of Women, by Governorate, 2017 (49,60] (38,49] (27,38] (16,27] Bizerte 15.9 Ariana [5,16] 14.9 Manouba Tunis Beja 27.1 22.6 Nabeul Jendouba Ben Arous 34.2 12.0 37.4 23.5 Zaghouan 11.9 Le Kef Siliana 19.0 29.1 Sousse 14.3 Kairouan Monastir 18.0 8.4 Mahdia Kasserine 21.3 35.4 Sidi Bouzide 28.7 Sfax 15.0 Gafsa 55.9 Tozeur 35.9 Gabes 50.5 Kebili 58.6 Mednine 43.7 Tataouine 50.6 Source: Based on data from the Labor Force Survey (ENPE), INS. 62 Tunisia’s Jobs Landscape BOX 2.4. Internal Migration and Two Secondary Cities in Tunisia According to the INS, between 2009 and 2014, almost 690,000 individuals moved across delegations (Tunisian districts), which is the definition of internal migration. Over 62 percent of the internal migrants moved across governorates, accomplishing long distance moves. Coastal governorates remain the main recipients of population inflows thanks to a concentration of public and private investments, services, and economic activities (Figure B 2.4.1). Between 2009 and 2014, the migration balance was posi- tive in Greater Tunis, the North-East, and the Centre-East and negative in the other regions (World Bank 2021b). Urban-to-urban migration is dominant. Over 80 percent of intergovernorate migration takes place among urban areas. The departure from rural areas to cities contributes less than 7 percent to all intergovernorate migration (World Bank 2021b). MAP B 2.4.1. Net Interdelegation Flows (> 200), 1994 and 2004 2004 200 km Source: World Bank 2021b. A recent study conducted by the World Bank (2021b) on internal migration in Jendouba and Kairouan, two secondary cities in the two poorest internal regions of Tunisia, point to some interesting findings. Both cities have a weak industrial structure, with a predominance of agriculture, and therefore face difficulties in offering economic opportunities. Both cities are at the top of outmigration flows in favor of coastal cities. The flows are not unidirectional because the cities are also receiving large inflows of migrants from rural areas and from distant delegations within the same governorates. In 2009–14, the city of Jendouba attracted over 4,000 migrants, of which more than 1 in 2 was from urban areas, whereas the city of Kairouan received about 10,000 migrants, with over 7 in 10 migrating from other urban areas. Based on focus group discussions, the World Bank (2021b) finds that many migrants moved to Jendouba or Kairouan before securing a job, and better job opportunities, together with superior access to services and security, are the main drivers of migra- tion. Migrants though encounter challenges in finding jobs and often rely on informal channels in seeking jobs. In Jendouba, men typically find jobs as ay laborers in construction, whereas women are more likely to work in irrigated agricultural areas outside the city. In Kairouan, men migrants find jobs as waiters, and women migrants work as nannies, craftswomen, or in garment factories or agrifood processing if they have specialized skills given the presence of an industrial sector in the city. Most migrants face precarious working conditions with no labor law protection or social security coverage. In addition, women migrants have to bear the double burden of work and family care without the possibility of relying on social or extended family networks. FIGURE 2.31. Employment-to-Population Ratios, by Sex and Age-Group, 2006–17 15−24 25−29 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 100 100 80 80 60 60 Percent Percent 40 40 20 20 0 0 30−44 45−64 100 100 80 80 60 60 Percent Percent 40 40 20 20 0 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 All Female Male Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 2.32. Unemployment Rates, by Sex and Age-Group, 2006–17 15−24 25−29 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 50 50 45 45 40 40 35 35 30 30 Percent Percent 25 25 20 20 15 15 10 10 5 5 0 0 30−44 45−64 50 50 45 45 40 40 35 35 30 30 Percent Percent 25 25 20 20 15 15 10 10 5 5 0 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 All Female Male Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 2.33. Employment-to-Population Ratios, by Educational Level and Sex, 2006–17 None Primary 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 100 100 80 80 60 60 Percent Percent 40 40 20 20 0 0 Seconday Tertiary 100 100 80 80 60 60 Percent Percent 40 40 20 20 0 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 All Female Male Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE 2.34. Unemployment Rates, by Educational Level and Sex, 2006–17 None Primary 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 50 50 45 45 40 40 35 35 30 30 Percent Percent 25 25 20 20 15 15 10 10 5 5 0 0 Secondary Tertiary 50 50 45 45 40 40 35 35 30 30 Percent Percent 25 25 20 20 15 15 10 10 5 5 0 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 All Female Male Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 65 FIGURE 2.35. Employment Category Distribution, by Sex, 2006–17 100 0.1 0.2 0.0 0.0 0.0 0.1 0.0 0.0 2.8 0.0 0.2 2.0 3.2 0.0 0.0 0.2 0.0 0.0 0.2 1.6 0.2 0.1 0.0 0.0 0.1 80 66.8 66.8 68.4 68.6 70.6 67.8 69.5 68.6 71.5 69.0 71.0 71.7 72.7 72.0 75.0 60 75.3 79.3 80.9 81.8 81.6 85.0 Percent 40 0.3 0.3 0.2 4.4 3.4 3.1 0.3 0.1 0.2 0.4 0.3 2.9 2.1 2.0 0.1 2.5 0.5 6.9 4.4 0.2 0.3 0.1 4.3 3.8 0.1 0.1 2.6 14.1 0.4 0.1 2.2 2.8 20 21.6 22.0 21.3 19.6 19.5 20.8 18.1 7.2 8.0 0.2 6.3 0.2 0.1 0.1 19.3 19.4 18.6 17.0 16.9 17.9 15.5 4.0 2.7 3.7 13.1 12.2 10.1 9.2 9.2 9.8 8.0 6.8 7.1 6.9 8.5 8.0 7.4 7.5 5.5 5.8 5.8 7.1 6.7 6.1 6.2 1.8 2.0 2.5 3.1 2.9 2.6 2.5 0 06 08 11 13 15 16 17 06 08 11 13 15 16 17 06 08 11 13 15 16 17 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Male Female All Employer Own-account worker Contributing family worker Apprentice Wage worker Other Not stated Source: Based on data from the Labor Force Survey (ENPE), INS. The services sector employs almost 60 percent of women public administration and in health and education services, wage workers and less than 50 percent of men wage workers. followed by other services (8.6 percent), trade (7.3 per- As of 2017, about 6 percent of women working for wages cent), and real estate and professional services (4.8 percent). were employed in agriculture, compared with over 9 per- Within services, men showed larger shares of wage workers cent among men; both shares increased over time by over in trade, transport, and hotels and food services. 20 percent (Figure 2.36). The textile sector, which attracted many more women than men, shed jobs and employed Although still largely employed in low- and mid-skill jobs about 20 percent of women wage workers in 2017, while in the private sector, the share of women wage workers the rest of the secondary sector—other manufacturing, in high-skill jobs was rising.23 Although the trend was construction, and utilities—accounted for about 15 percent toward a decline in the share of women employed for a of women’s wage employment (Figure 2.36, panel b). The wage in low- and medium-skill jobs, more than 8 women share of women wage workers employed in the secondary in 10 were still performing elementary or medium-skill sector declined by over 15 percent, while the share of men jobs in 2017, respectively 59.0  percent and 22.7  per- increased by 5 percent thanks to the role played by the cent (Figure  2.37, panel b). Among women performing construction sector. The opposite trend was detected in medium-skill jobs, the share of service and sales workers services in 2006–17. The share of women wage workers and of skilled agricultural workers increased, whereas in the services sector rose by 10 percent, while the share of men declined by about 7 percent. Trade, real estate and 23  Jobs are classified as high, medium, or low skill based on the occupation, professional services, public administration, and health that is, the type of work performed as reported by workers in the labor and education services were the drivers of the growth in force survey. High-skill jobs include managers, professionals, technicians; medium-skill jobs cover clerks, service and sales workers, skilled agricul- service sector wage employment among women. In 2017, tural workers, craft and related trade workers, and plant and machine 32  percent of women wage workers were employed in operators and assemblers; low skill jobs include elementary occupations. FIGURE 2.36. Sectoral Distribution of Wage Workers, by Sex, 2006–17 a. Men 3.1 2006 7.6 3.0 3.6 10.5 21.1 2.0 7.3 5.1 5.9 25.1 4.70 1.2 3.0 2008 7.5 3.1 3.9 10.6 22.7 2.1 6.4 5.1 5.9 23.4 4.60 1.1 3.1 2011 7.3 2.8 3.6 10.0 25.2 1.8 6.7 5.9 5.1 23.2 3.40 1.1 3.5 2013 8.9 2.9 3.2 10.3 23.4 2.1 6.7 6.2 5.0 23.7 2.7 1.0 3.7 2015 9.8 2.9 3.0 10.5 23.5 1.8 7.9 5.7 4.8 22.9 2.6 1.1 4.3 2016 8.5 3.1 3.1 10.5 24.1 1.8 7.9 5.3 4.7 22.4 2.7 1.0 4.3 2017 9.3 3.1 2.8 10.3 23.9 1.8 8.1 5.7 5.1 21.6 2.6 1.0 0 10 20 30 40 50 60 70 80 90 100 Percent Agriculture Agro-food industry Textile Manufacturing Other Manufacturing Construction Other Secondary Trade Transports Hotels and Restaurants Financial Services Real Estate/Professional Activities PA/Health/Education Other Services Not De ned b. Women 0.6 1.6 2006 4.9 2.0 30.5 8.0 5.6 1.0 31.1 9.8 0. 0.5 2.2 2.0 0.8 2.3 2.6 2008 5.4 1.9 27.9 9.1 5.6 31.2 9.6 0. 0.6 1.0 1.3 0.7 1.6 2011 5.5 2.6 25.0 9.8 13.9 0.4 5.9 2.2 6.1 32.5 8.0 0. 0.7 1.9 2013 5.1 2.2 24.8 9.5 6.1 4.1 32.8 7.5 0. 0.7 2.3 1.7 0.6 1.9 7.4 2015 5.1 2.7 22.6 10.5 7.3 4.6 32.7 0. 0.8 2.4 1.4 1.0 1.7 2016 5.1 2.7 21.2 10.1 7.3 2.3 4.4 33.0 8.6 0. 0.6 1.5 0.8 2.0 2017 6.2 2.9 20.5 10.4 7.3 2.3 4.8 31.9 8.6 0. 0.6 1.9 0 10 20 30 40 50 60 70 80 90 100 Percent Agriculture Agro-food industry Textile Manufacturing Other Manufacturing Construction Other Secondary Trade Transports Hotels and Restaurants Financial Services Real Estate/Professional Activities PA/Health/Education Other Services Not De ned Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 67 FIGURE 2.37. Occupational Distribution of Wage Workers, by Sector and Sex, 2006–17 a. Private sector, men 2006 2.62.1 4.4 4.2 11.5 5.8 23.7 13.2 32.4 0. 2.0 2011 2.13.6 4.6 10.5 4.9 23.8 12.9 35.4 0. 2013 2.6 2.5 3.5 4.2 10.1 5.1 22.9 13.1 36.0 0. 2015 2.4 3.5 5.2 2.9 12.1 4.7 24.2 12.4 32.5 0. 2016 2.4 3.5 4.2 3.1 12.7 4.6 24.9 12.7 31.7 0. 2017 2.1 3.6 4.5 2.8 14.0 6.2 23.8 12.9 30.1 0. 0 10 20 30 40 50 60 70 80 90 100 Percent Managers Professionals Technicians Clerks Service and sales workers Skilled agricultural Craft workers Machine operators Elementary occupations Not stated b. Private sector, women 2006 5.7 6.4 10.8 7.6 2.8 6.0 37.4 25.8 0. 2011 2.9 7.4 11.7 8.4 4.0 3.4 36.9 24.1 0. 2013 2.1 3.9 7.3 11.2 8.9 3.8 3.8 35.4 23.5 0. 2015 1.8 6.5 7.6 8.6 9.1 3.2 6.3 33.6 23.2 0. 2016 2.4 7.9 7.5 8.5 10.7 3.2 6.8 31.3 21.5 0. 2017 2.3 8.3 7.8 7.0 10.5 4.6 5.9 31.0 22.7 0. 0 10 20 30 40 50 60 70 80 90 100 Percent Managers Professionals Technicians Clerks Service and sales workers Skilled agricultural Craft workers Machine operators Elementary occupations Not stated (continued) 68 Tunisia’s Jobs Landscape FIGURE 2.37. Occupational Distribution of Wage Workers, by Sector and Sex, 2006–17 (continued) c. Public sector, men 2006 7.2 18.6 18.2 10.5 18.8 0.6 4.5 7.0 14.2 0. 2011 8.9 16.0 15.3 9.1 20.7 0.6 4.6 7.7 17.2 0. 2013 9.5 17.0 14.4 9.2 19.4 0.7 4.6 6.8 18.3 0. 2015 8.5 24.7 8.4 9.4 25.2 0.7 3.4 6.0 13.5 0. 2016 8.5 22.0 7.8 8.5 28.5 0.7 3.2 7.4 13.3 0. 2017 8.2 22.1 7.7 8.4 30.0 0.68.3 6.2 13.3 0.0. 0 10 20 30 40 50 60 70 80 90 100 Percent Managers Professionals Technicians Clerks Service and sales workers Skilled agricultural Craft workers Machine operators Elementary occupations Not stated d. Public sector, women 2006 3.2 27.8 40.0 16.9 0.5 3.6 7.9 0. 2011 4.2 29.7 37.2 16.7 4.40.5 0.5 7.2 2013 4.7 26.7 35.9 15.7 4.80.5 11.3 2015 4.9 47.1 14.2 15.2 0.4 5.5 0.2 12.0 2016 5.5 47.7 15.2 13.0 0.2 5.5 0.4 11.6 11.2 2017 5.4 49.6 15.7 12.1 0.3 5.0 0.2 0 10 20 30 40 50 60 70 80 90 100 Percent Managers Professionals Technicians Clerks Service and sales workers Skilled agricultural Craft workers Machine operators Elementary occupations Not stated Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 69 clerks, craftworkers, and machine operators declined con- Women work, on average, fewer hours in wage employ- siderably. By contrast, the share of men wage workers in ment compared with men. Women working for wages were medium-skill jobs rose modestly, from 58.5  percent to employed an average of about 41 hours a week relative to 59.7 percent, mainly because of an increase in the numbers 44 hours a week worked by men wage workers (Figure 2.39, of service and sales workers and skilled agricultural workers panel a). Similarly, self-employed women work an average (Figure 2.37, panel a). The share of women in high-skill of 38 hours a week compared with 46 hours worked by men jobs, including senior officials, professional staff, and tech- (Figure 2.39, panel b; Box 2.5). Among the self-employed, nicians, increased from 10.0 percent in 2006 to 18.4 per- contributing family workers work short hours on average; cent in 2017 and is considerably higher than the share women work 36 hours on average, and men 42. among men working for a wage (at 10.2 percent in 2017, Figure 2.37, panel a). Wage jobs in the public sector and nonwage jobs provide more flexibility in working hours. Women (and men) working About 7 in 10 women wage workers employed in the public in the private sector work on average longer hours com- pared to their counterparts in the public sector. Precisely, sector perform high-skill jobs. The share of women working women wage workers in the public sector work on average in the public sector and performing high-skill jobs remained 33 hours, where those employed in the private sector are constant at about 71  percent in 2006–17, a share that at work for almost 44 hours on average. Further, the public is more than 30 percentage points higher than the share sector provides the possibility of working less than full-time among men (Figure  2.37, panels c and d). Key was the as a large number of women work about 20 hours per week rise in the share of professionals, while the share of tech- (Figure 2.40). About 32 percent of women work between nicians declined. Over 2006–17, the share of women in 40 and 47 hours a week and 22 percent 48 hours a week medium-skill jobs fell from 21 percent to 18 percent, and or more. In the private sector the share of women working the share of women in low-skill jobs rose from 7.9 percent 48 hours a week or more is considerably higher, estimated to 11.2 percent. Similar to the pattern in the private sector, at 52 percent. Nonwage jobs too offer the possibility of the share of men performing high-skill jobs declined from working a lower number of hours per week, and almost 1 44 percent to 38 percent, while medium-skill jobs gained in 2 women self-employed work less than 40 hours (Fig- importance thanks to the rise in the share of service and ure 2.39-panel b). sales workers from 8.8 percent to 30 percent. Women in both the private sector and the public sector Constraints on Women’s are, on average, more well educated than men. Over time, improvements in the educational level of the population Participation in the Labor Force have become reflected in the employed population. Girls This section summarizes the constraints on women’s par- are now outstripping boys in educational outcomes, and, ticipation in the labor market in Tunisia based on a desk- on average, working women are more well educated than top review of the available academic and grey literature working men. Restricting the sample to wage workers, the and the analysis of data, including in databases maintained large majority of the employed population, over 20.0 per- by the OECD and the World Bank. The starting point is cent of women in the private sector had tertiary education the conceptual framework described by Chakravarty, Das, in 2017, up from 7.8 percent in 2006; the share was half and Vaillant (2017), which distinguishes between three as much among men (10.3 percent in 2017) (Figure 2.38, broad categories of constraints: (1) contextual factors, panels a and b). The increase in the share of tertiary educated (2) endowments, and (3) preferences (Figure 2.41). The women corresponded to a decline in the share of working latter two (endowments and preferences) are constraints women with primary and secondary education, whereas the to women’s labor supply, while the former (contextual share of women with no schooling declined more slowly, factors) includes both demand side and supply side con- from 12.0 percent to 10.7 percent. Relative to the private straints. In the spirit of Pimkina and de la Flor (2020), sector, the share of tertiary educated women in the public the discussion of contextual factors is enriched with the sector was much higher, estimated at 63.3 percent in 2017, inclusion of macroeconomic forces, such as broader eco- compared with 53.6 percent in 2007 (Figure 2.38, panels c nomic trends and structural change, which may—in some and d). Among men, the share stayed constant over time contexts—pull women into the labor force. Embedded in and was estimated at 32,1 percent in 2017, over 30 percent- this conceptual framework is the notion of dynamic feed- age points lower than among women. back loops. For example, changes in cultural traditions 70 Tunisia’s Jobs Landscape FIGURE 2.38. Educational Level Distribution of Wage Workers, by Sector and Sex, 2006–17 a. Private sector, men b. Private sector, women 2006 9.9 46.6 36.5 6.90. 2006 12.1 38.4 41.7 7.8 0. 2008 9.7 46.5 36.3 7.5 0. 2008 11.1 36.1 41.9 10.9 0. 2011 9.2 45.0 36.3 9.4 0. 2011 10.1 31.7 44.0 14.2 0. 2013 7.4 46.2 37.6 8.7 0. 2013 9.1 33.6 42.4 14.7 0. 2015 9.0 43.7 36.5 10.6 0. 2015 9.5 31.4 41.6 17.0 0. 2016 7.9 44.6 37.2 10.2 0. 2016 8.8 30.2 40.4 20.3 0. 2017 7.4 43.4 38.8 10.3 0. 2017 10.7 30.5 37.9 20.9 0. 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percent Percent None Primary Secondary None Primary Secondary Tertiary Not stated Tertiary Not stated c. Public sector, men d. Public sector, women 2006 3.5 17.9 45.5 32.8 0. 2006 3.1 6.6 36.4 53.6 0. 2008 3.7 17.4 44.0 34.6 0. 2008 2.9 5.2 35.4 56.2 0. 2011 3.0 18.2 43.4 35.2 0. 2011 2.7 5.6 30.7 60.5 0. 2013 2.6 18.6 46.1 32.5 0. 2013 3.7 7.4 35.3 53.5 0. 2015 3.6 19.0 42.5 34.7 0. 2015 4.3 8.3 28.9 58.4 0. 2016 4.2 21.4 43.2 30.9 0. 2016 4.9 9.1 26.2 59.4 0. 2017 3.1 19.4 45.3 32.1 0. 2017 3.9 8.7 24.1 63.3 0. 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percent Percent None Primary Secondary None Primary Secondary Tertiary Not stated Tertiary Not stated Source: Based on data from the Labor Force Survey (ENPE), INS. regarding the role of men and women in society can affect and developments, discrimination, cultural traditions, and women’s preferences for time use and family formation. public safety considerations., is presented. These factors Likewise, structural change (that is, changes in sectoral have in common that, even though they are beyond the con- labor demand) may create incentives to invest in skills or trol of the individual, they may, by affecting endowments alter cultural traditions about the types of jobs that are and preferences, have a profound direct or indirect impact appropriate for women. on women’s decisions to participate in the labor market. CONTEXTUAL FACTORS Legal Framework First, a review and discussion of contextual factors, which Tunisia’s legislative framework is considered progressive by here includes the legal framework, macroeconomic trends regional standards. Important principles of gender equality Access to the Labor Market: A Spotlight on Women and Youth 71 FIGURE 2.39. Distribution of Wage and Nonwage Workers, by Sex and the Number of Hours Worked per Week, 2015 a. Wage workers b. Nonwage workers .15 .04 .03 .1 Density Density .02 .05 .01 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Weekly hours worked Weekly hours worked Women Men Women Men Source: Based on data from the EBCNV 2015, INS. BOX 2.5. Gender Gaps in Self-Employment In Tunisia, the share of the self-employed, which includes employers, own-account workers, and unpaid family workers, is declining. Between 2006 and 2017, the share of self-employment in total employment fell by over 20 percent, from 32.0 percent to 24.4 percent. The shift from self-employment to wage-employment was more rapid among women than men. The share of self-employed women fell by half and was esti- mated at 14 percent of total women’s employment in 2017, while the share is twice as large among men. The reduction in self-employment occurred thanks to a sizable reduction in the share of unpaid family workers and own-account workers, whereas the share of employers rose among both men and women. In 2017, about 2.5 (8.0) percent of working women were employers (own-account workers) relative to 7.5 (18.0) percent of men. Despite a substantial drop, the prevalence of unpaid family workers is still higher among women relative to men: 3.7 percent vs 2.5 percent in 2017. Women working as employers and on their own account have shifted out of agriculture into the services sector (Figure B 2.5.1). In 2017, 60 percent of such women were employed in services, especially in trade, 20 percent in agriculture, and 20 percent in the secondary sector, mainly in textiles and agrifood production. The share of men employers and own-account workers in agriculture also declined; the share was estimated at 32 percent in 2017. Although a similar reallocation is observed among unpaid family workers (annex Figure A 2.2), unpaid men and women family workers are still largely employed in agriculture: 74.0 percent of women and 59.4 percent of men in 2017. The shift toward the services sector translated mainly into an increase in the share of unpaid workers in trade. Employers and own-account workers are distributed by educational level similarly to the fully employed population by sex, though there are important differences among unpaid family workers (annex Figure A 2.3; Figure A 2.4). Among men, there is a large concentration of workers with primary and secondary education. The share of unpaid men family workers with secondary education rose from about 42 percent to over 57 percent in 2006–17; this compares with 35 percent among the overall employed population. Among women, while the distribution of unpaid family workers by educational level does not differ considerably relative to the general employed population, 19 percent of women in this type of employment hold university degrees. This shows that unpaid family workers are a highly heterogeneous group. (continued) BOX 2.5. Gender Gaps in Self-Employment (continued) FIGURE B 2.2. Sectoral Distribution of Employers and Own-Account Workers, by Sex, 2006–17 a. Men 2006 36.8 0.3 5.9 1.1 5.7 0.0 26.5 8.4 2.0 1.1 6.7 1.43.90. 2008 34.6 1.1 1.6 6.1 5.3 0.0 27.6 8.9 2.0 3.1 7.0 1.3 3.70. 2011 34.0 1.0 2.0 6.3 0.0 3.3 29.4 9.3 10.1 7 7.7 0. 1.53.1 2013 30.4 1.1 1.7 6.6 4.50.1 29.4 10.4 2.6 5.2 8.0 0. 1.73.2 2015 30.0 1.1 1.3 6.4 5.4 0.0 29.6 10.7 2.5 6.1 8.1 0. 1.83.1 2016 29.5 1.1 2.5 6.3 5.4 0.1 29.4 10.4 2.0 6.1 8.1 0. 1.93.0 2017 32.2 1.1 1.2 5.4 4.70.0 30.9 9.5 2.6 0.1 7.5 0. 1.43.1 0 10 20 30 40 50 60 70 80 90 100 Percent Agriculture Agro-food Industry Textile Manufacturing Other Manufacturing Construction Other Secondary Trade Transports Hotels and Restaurants Financial Services Real Estate/Professional Activities PA/Health/Education Other Services Not De ned b. Women 2006 46.6 0.9 16.6 0.0 1.2 18.0 2.52.6 0.3 1.2 10.9 0 2008 42.8 1.4 13.4 0.0 1.2 21.0 0.0 1.22.3 2.2 12.4 0 2011 28.8 2.5 14.2 0.0 1.2 26.8 0.0 1.3 5.2 4.3 15.0 1.7 2013 20.1 3.4 15.7 0.4 2.0 29.1 0 5.5 1.17.3 4.1 16.0 0 2015 19.7 3.2 14.6 1.1 0.0 0.4 29.6 6 6.3 1.10.3 6.0 15.8 0 2016 19.1 3.8 16.0 1.1 0.0 0.9 1.1 29.5 0.1 1.0 4 5.8 6.5 13.7 0 2017 19.8 5.3 13.1 0.0 0.4 1.1 32.1 22 1.0 .1 5.7 6.4 13.4 0 0 10 20 30 40 50 60 70 80 90 100 Percent Agriculture Agro-food Industry Textile Manufacturing Other Manufacturing Construction Other Secondary Trade Transports Hotels and Restaurants Financial Services Real Estate/Professional Activities PA/Health/Education Other Services Not De ned Source: Based on data from the Labor Force Survey (ENPE), INS. Access to the Labor Market: A Spotlight on Women and Youth 73 FIGURE 2.40. Distribution of Women Wage Workers Tunisia’s score improved significantly—from 58.75 to the in the Public and Private Sectors, by the Number of current 67.5 in 100 in 2018—after the Tunisian parliament Hours Worked per Week, 2015 had passed landmark legislation (Act No. 2017–58) in 2017 aimed at eliminating violence against women and .15 girls (UN Women 2017).25 The law, which considers dif- ferent types of violence (that is, physical, economic, sexual, .1 political, and psychological violence) and provides institu- tional mechanisms for the protection of victims, has been Density widely recognized as a milestone, though implementation .05 and enforcement may be less stringent than the law itself (Boukhayatia 2018). 0 Despite these improvements in the legal framework, there are 0 20 40 60 80 100 still several areas of legislation that disadvantage women’s Weekly hours worked economic opportunities relative to those of men. The Public Private country performs well on the subscores for mobility, workplace, and pension (100/100), but significantly less Source: Based on data from the EBCNV 2015, INS. well on the indicators pay (25/100), marriage (60/100), parenthood (40/100), entrepreneurship (75/100), and are enshrined in the country’s constitution of 2014 (Cham- assets (40/100) (Figure 2.43). Below are examples of laws bers and Cummings 2014; Sinha 2011). According to that restrict women’s economic opportunities (see annex Women, Business, and the Law 2021 data, Tunisia has Table A 2-1): more gender equitable laws than most other countries in the region (with the exception of Djibouti, Malta, Morocco, Restrictions on the type of employment women can per- Saudi Arabia, and the United Arab Emirates) (Figure 2.42).24 form: Current laws restrict women’s work at night (Code FIGURE 2.41. Framework for the Constraints on Women’s Labor Market Participation Laws Macroeconomy Discrimination Endowments: Women's Preferences • Skills labor • Time & family • Productive assets market • Occupation • Networks participation • Mobility Cultural traditions & customs Safety Source: Modified based on Chakravarty, Das, and Vaillant 2017; Pimkina and de la Flor 2020. 24 The data include an index on 190 economies that is structured around   25  The reforms led to significant improvements in the indicators developed the life cycle of a working woman. In total, 35 questions are scored across from the following questions: “Is there legislation on sexual harassment eight indicators. Overall scores are then calculated by taking the average in employment?” “Are the criminal penalties or civil remedies for sexual of each indicator, with 0 representing the lowest, and 100 the highest pos- harassment in employment?” “Is there legislation specifically addressing sible score. Data refer to the laws and regulations that are applicable to the domestic violence” (see annex Table A1). main city of business (in the case of Tunisia, Tunis) ( World Bank 2021c). See WBL (Women, Business, and the Law 2021) (dashboard), World Bank, Washington, DC, https://wbl.worldbank.org/en/wbl. 74 Tunisia’s Jobs Landscape FIGURE 2.42. Women, Business, and the Law Ranking, Tunisia and other Middle East and North Africa Countries WBL index, MENA countries 100 88.8 82.5 80.0 75.6 68.1 67.5 80 57.5 55.6 52.5 WBL INDEX 50.0 46.9 60 45.0 45.0 36.9 35.6 31.3 29.4 28.8 26.9 40 26.3 20 0 Malta United Arab Emirates Saudi Arabia Morocco Djibouti Tunisia Algeria Bahrain Lebanon Libya Jordan Iraq Egypt, Arab Rep. Syrian Arab Republic Oman Iran, Islamic Rep. Qatar Kuwait Yemen, Rep. West Bank and Gaza Source: Based on data of World Bank 2021c; WBL (Women, Business, and the Law 2021) (dashboard), World Bank, Washington, DC, https://wbl.worldbank.org/en/wbl. du Travail, articles 66 and 68-2) and in the primary sector of maternity leave, which falls significantly short of the (agriculture, mining) (Code du Travail, articles 77 and standard of 14  weeks of maternity leave recommended 375).26 While ostensibly geared to protect women, singling by the International Labour Organization (ILO). Across out women for special protections is increasingly being countries, the length of maternity leave strongly correlates viewed as out-of-date and inconsistent with principles with women’s employment in the private sector (see Amin of nondiscrimination and equal treatment (OECD 2017; and Islam 2019). Furthermore, there is no legislation in Politakis 2001). Tunisia prohibiting the dismissal of pregnant workers. Maternity leave and protection of pregnant workers: Parental leave: While Tunisian law provides 30 days of Tunisia provides mothers with approximately 30  days maternity leave and one day of paternity leave, there is no FIGURE 2.43. Women, Business and the Law, by Domain Tunisia - Scores for Women, Business and the Law 2021 Mobility Workplace Pay Marriage Parenthood Entrepreneur- Assets Pension WBL 2021 Index ship Score 100 100 25 60 40 75 40 100 67.5 Source: Based on data of World Bank 2021c; WBL (Women, Business, and the Law 2021) (dashboard), World Bank, Washington, DC, https://wbl.worldbank. org/en/wbl. “Tunisie: Code du travail, 1996,” NATLEX (Database of National   26 Labour, Social Security, and Related Human Rights Legislation), Inter- national Labour Organization, Geneva, https://www.ilo.org/dyn/travail/ docs/778/Labour%20Code%20Tunisia.pdf. Access to the Labor Market: A Spotlight on Women and Youth 75 system of paid parental leave, that is, leave that is either change, and female labor force participation (for example, shared between mother and father or an individual entitle- Gaddis and Klasen 2014; Goldin 1990, 1995; Klasen et al. ment that each parent can take regardless of the other to 2021; Mammen and Paxson 2000). It goes beyond the care for small children. Evidence from other, mostly high- purpose of this chapter to review this literature in detail, income countries suggests that a well-designed parental but four key insights that have potential relevance for the leave system, especially if it includes certain elements to Tunisian country context are now summarized. incentivize take-up by fathers (for example, bonus months or daddy quotas), can lead to a more equitable sharing of First, there are examples of countries, notably in Asia, paid and unpaid work between parents (Patnaik 2019). where new opportunities in growing sectors of the econ- omy have been associated with increases in female labor Property ownership: As in most countries in the region, force participation (Klasen 2019b). For example, it has Tunisia’s inheritance laws are based on Islamic Sharia law been argued that the expansion of light manufacturing and do not provide for equal inheritance rights among (for instance, textiles, clothing, footwear) can be a driving male and female surviving spouses or sons and daughters. force of rising female labor force participation in parts of In November 2018, the Tunisian cabinet adopted a draft bill East and South Asia (Heath and Mobarak 2015; Seguino to amend the Personal Status Code to provide for gender 2000). There is likewise some evidence that growth equality in inheritance as a default. However, the bill, which in service sectors and occupations has created employ- was presented to Parliament in February 2019, failed to ment opportunities for highly educated women in Latin garner the necessary support (HRW 2018; Tanner 2020). America and India (Gasparini and Marchionni 2015; In addition, the default marital regime is separation of Klasen and Pieters 2015). If the conditions are favor- property, and there are no laws explicitly recognizing non- able, rising female employment because of new economic monetary contributions to marital property (for instance, opportunities may set off a virtuous cycle of incentives the contribution of a stay-at-home spouse taking care of to invest in skills and delay age at marriage (for instance, children or the household). Separation of property regimes, see Heath and Mobarak 2015 on Bangladesh). Moreover, where all property is individually owned, are generally the increase in women’s employment may lead to higher less favorable to women than community property regimes levels of women’s decision-making within the household, whereby most property acquired during the marriage is and this could challenge traditional gender roles (Majlesi owned jointly. Community of property, which recognizes 2016). women’s role in the accumulation of marital property through child-rearing and other unpaid work, is especially Second, despite these possibilities for positive feedback important in legal systems that do not provide for equal loops and the positive experiences of a few (mostly Asian) inheritance rights between males and females, because countries, the changes in female labor force participation widows cannot automatically claim ownership of their that can be traced to economic growth and structural deceased husband’s estate (Deere and Doss 2006). Overall, change are typically rather small. Gaddis and Klasen the international evidence suggests that more gender equi- (2014) use data on 200 countries to investigate the empiri- table laws on property ownership strongly correlate with cal relationship between sectoral growth and female labor the likelihood of women owning land and housing property force participation between 1980 and 2005 and simulate (Gaddis, Lahoti, and Swaminathan 2020). the portion of the change in female labor force participa- tion over this period that can be linked to sectoral growth. Antidiscrimination laws: While Tunisia’s legal code pro- The results suggest that slightly less than 10 percent (that hibits discrimination in employment based on gender, is, 1 percentage point) of the 11 percentage point increase there are no specific provisions mandating equal work in female labor force participation among the countries for equal value to protect against wage discrimination. in the sample is linked to structural change. Similarly, Likewise, there are no laws that prohibit discrimination in changes in overall GDP per capita (even if accounting for access to credit based on gender. a nonlinear relationship) explain little of the variation in female labor force participation at the country level, compared with country fixed effects (Gaddis and Klasen Macroeconomic Factors 2014). This suggests there are important historical deter- A rich academic literature investigates the relationship minants of female labor force participation that are highly between economic growth and development, structural persistent over time and dwarf changes associated with 76 Tunisia’s Jobs Landscape growth and structural change (Klasen 2019b).27 Similarly, sector could offer potential opportunities for employ- the World Bank (2020b) shows that sustained increases ment growth, especially in tourism and the care economy, in female labor force participation from a low base are given the county’s aging population. Moreover, these sec- relatively rare. tors could provide employment opportunities for women with low levels of education. However, women’s employ- Third, some studies document the countercyclicality of ment growth in tourism has so far been constrained by female labor force participation (Bhalotra and Umaña- the sector’s negative reputation, restrictions on women’s Aponte 2010; Serrano et al. 2019). Similarly, female employ- geographic mobility, and limited family support services. ment and labor force participation have often been found Together, this suggests that growth and structural change to increase during times of economic downturn and reces- could contribute to raising labor force participation sion (Sabarwal, Sinha, and Buvinic´ 2011; Lim 2000).28 One among women in the near future, particularly if accompa- explanation of this phenomenon is the added worker effect, nied by an alleviation of some of the other constraints to which refers to a temporary increase in married women’s women’s participation. In the medium and long run, the labor supply because of a job or income loss by their income effect ascribable to rising living standards associ- husbands (Lundberg 1985). In the context of developing ated with the higher earnings of men might mitigate such countries, strong added worker effects have been docu- positive impacts on women’s labor force participation, mented in Latin America (for example, Cardona-Sosa, particularly among women with little education and weak Flórez, and Zurita 2016 on Colombia; Fernandes and labor market attachment. de Felicio 2005 on Brazil; Skoufias and Parker 2006 on Mexico). However, there appear to be no studies on the Middle East and North Africa.29 Discrimination If employers discriminate against women, this could mute Fourth and related to the previous point, labor force par- the potential for a boost in women’s employment because ticipation of women, especially poor women with low of rising labor demand, especially if the discrimination levels of education, often declines as the incomes of other occurs in the growing sectors. While it is difficult to find household members rise (Klasen et al. 2021). This rela- direct evidence of discrimination, there are some indications tionship, which is consistent with standard labor supply that discriminatory practices exist and may disadvantage theory, seems to be one of the main forces behind the women. Kärkkäinen (2011) states that 60 percent of hotels decline in female labor force participation in India (Klasen and 40 percent of information and communication tech- and Pieters 2015). Overall, these last two points suggest nology (ICT) companies interviewed specified the desired that poor women in many developing countries have a sex of the applicants during the process of recruiting staff weak attachment to the labor market and often withdraw even though this practice is outlawed. Moreover, many if it becomes affordable to do so. companies apparently expressed a preference for single women rather than married women because married What does this imply for Tunisia? Economic growth in women were considered more costly and less produc- Tunisia has been modest recently, and structural change tive, especially during maternity. While the study is some- has proceeded slowly. Moreover, employment in the tex- what dated, it seems plausible that at least some of these tile and garment sectors, which have been associated with practices have continued. Similarly, World Bank (2014b) rising female labor force participation in parts of Asia, has shows that more than 60 percent of young women in rural declined in recent years because of stiff competition from Tunisia expressed the concern that women were discrimi- Asian manufacturers with lower wage costs. The service nated against in seeking work in the private sector; the share is lower in work in the public sector (44 percent), but still high.30 27  For example, Alesina, Giuliano, and Nunn (2013) show that traditional agricultural practices are correlated with current differences in traditions and customs related to gender roles, which may explain some of the persis- Discrimination can be reinforced by lack of legal reme- tent cross-country differences in female labor force participation. dies among victims. Tunisia’s legal code does not contain 28  The COVID-19 pandemic-induced recession is different, however. In many countries, female employment declined disproportionately during the pandemic. A possible explaining factor is that school closures during periods of lockdown raised the demand for caregiving and reinforced tradi- tional gender roles at home (Alon et al. 2020; Kugler et al. 2021). 30 Among young rural men, 44 (32) percent perceived that there was dis-   29  Ilkkaracan (2012) documents an added worker effect in Turkey. crimination against women in the public (private) sector. Access to the Labor Market: A Spotlight on Women and Youth 77 provisions to mandate equal work for equal value or pro- 69 percent of women and 74 percent of men agree with hibit discrimination in access to credit based on gender the statement, “a preschool child suffers with a working (World Bank 2021c). mother” (Figure 2.44, panel b). Support is somewhat less for the statement, “men make better executives than women,” with which 37 percent of women and 52 percent of men Cultural Traditions and Customs agree, and the statement, “university is more important for Besides gender inequalities in the legal code, cultural tradi- a boy than a girl,” with which 19 percent of women and tions and customs assign men the role of the household 31 percent of men agree (Figure 2.44, panels c and d). The breadwinner, while women are expected to take care of shares also show that, even though conservative views are children and provide other unpaid family work. According prevalent among both men and women, women espouse to data of the 2019 World Values Survey in Tunisia, 58 per- relatively more gender egalitarian views. cent of women and 73 percent of men agree with the state- ment, “when jobs are scarce, men should have more right Disaggregating the results of the World Values Survey to a job than women” (Figure  2.44, panel a). Similarly, further shows that young women, in particular, are less FIGURE 2.44. Cultural Traditions and Custom Assign Men and Women Traditional Roles a. Agreement with the statement, “when jobs are scarce, men should have more right to a job than women,” % Female 36.7 21 22.2 15.1 Male 46.1 26.7 15.9 8.6 0 10 20 30 40 50 60 70 80 90 100 Agree strongly Agree Neither agree nor disagree Disagree Disagree strongly b. Agreement with the statement, “a preschool child suffers with a working mother,” % Female 38.1 30.9 25 4.9 Male 36.4 36.9 23.5 2.9 0 10 20 30 40 50 60 70 80 90 100 Agree strongly Agree Disagree Strongly disagree No answer or missing c. Agreement with the statement, “men make better executives than women,” % Female 20.2 16.4 43.4 18.7 Male 26.3 25.3 38.2 9.7 0 10 20 30 40 50 60 70 80 90 100 Agree strongly Agree Disagree Strongly disagree No answer or missing d. Agreement with the statement, ”university is more important for a boy than a girl,” % Female 10 9.1 47.7 32.3 Male 14.5 16.8 46.6 21.7 0 10 20 30 40 50 60 70 80 90 100 Agree strongly Agree Disagree Strongly disagree No answer or missing Source: Based on data of WVS (World Values Survey), WVS Wave 7 (2017–2020): Tunisia 2019 (dashboard), King’s College, Old Aberdeen, United Kingdom, https://www.worldvaluessurvey.org/WVSDocumentationWV7.jsp. 78 Tunisia’s Jobs Landscape likely to agree with conservative attitudes about gender not driven by a social norm, but rather by customs and roles. Figure 2.45 plots the coefficients of a linear regres- traditions.31 sion analysis, whereby an increase in the dependent vari- able indicates more gender egalitarian views. Women ages 15–29 are significantly more likely than older women Safety (ages 50+) to disagree with statements that men have more Women’s economic activity can be constrained by the risk rights to jobs then women, that preschool children suffer of harassment and violence or in the public space, which with a working mother, and that men make better execu- is often exacerbated by a lack of safe public transport tives than women. There is no significant relationship (Pimkina and de la Flor 2020). Sexual harassment and between a woman’s age and the likelihood of her disagree- safety concerns are sometimes reported as constraints on ing with the statement that university is more important women’s labor force participation (Assaad and Barsoum for a boy than for a girl, but support for this statement is 2019; World Bank 2014b). However, there are few quan- generally low, especially among the female population (see titative data on the prevalence of this association. In terms Figure 2.44, panel d). Among men, there are no significant of perceptions, the 2019 World Values Survey in Tunisia age differences in the views, indicating that traditional shows that almost 27 percent of women and 30 percent gender role beliefs remain entrenched among younger of men report that sexual harassment frequently occurs in men. However, there is evidence from the regressions that their communities (Figure 2.46). the upper levels of educational attainment (at the tertiary level) correlate with less conservative attitudes, an associa- Qualitative data support the notion that public harassment tion that is observed among both men and women, though on streets, which may include unwanted yelling or touching, the coefficients are often only slightly below the 5 percent is a significant concern among young women, especially margin of statistical significance. There is no strong rela- on public transport, and that safety concerns limit young tionship between conservative attitudes and the number of women’s mobility after dark (Jesse 2017). Kärkkäinen children in the household, except for the statement “univer- (2011) documents that concerns over sexual harassment sity is more important for a boy than a girl,” with which at the workplace constrain the types of jobs for which women with children are more likely than women without women apply in the tourism industry; many women, for children to disagree. example, avoid jobs in bars or kitchens. These concerns can be amplified by cultural traditions that consider women as While the World Values Survey data show that there is upholders of social propriety, morality, and family honor. widespread support for traditional gender role models, For instance, jobs that involve direct contact with male they do not necessarily indicate that these beliefs are clients, colleagues, or superiors may be regarded, prima driven by social norms. Recent research by the World facie, as threats to women’s reputation (ILO 2018b; Jesse Bank behavioral science team explores whether time allo- 2017). cations of men and women across paid and unpaid work are driven by social norms or cultural traditions and cus- As in many countries, violence against women in Tunisia, toms (World Bank 2021a). While, colloquially, the term specifically, abuse of women and domestic violence against social norm is often used in a way that encompasses indi- women, seems to have increased during the COVID-19 vidual beliefs and cultural traditions, the strict sense of pandemic. According to an online survey conducted in the term requires evidence of high empirical and norma- June–July 2020, 37  percent of women and 28  percent tive expectations (Bicchieri 2017). Individual behaviors of men reported that violence in their communities had need to be conditioned by the beliefs and behaviors of increased during COVID-19 (World Bank 2021a). While others to qualify as a social norm. To understand whether this trend is deeply concerning, the longer-term conse- this is the case in Tunisia, the team used vignettes to assess quences, including on women’s labor market behavior, are the extent to which individuals might alter their behaviors uncertain. based on the views and behaviors of other community members. Overall, the results provide little evidence for such conditionality. In particular, both men and women 31 Caution is necessary because the composition of the sample is skewed   toward relatively young, well-educated, single women in urban areas. Such viewed men’s participation in housework favorably, which women typically exhibit high labor force participation rates and are poten- suggests that gender differences in time allocation are tially less affected by social norms and custom in their behavior. Access to the Labor Market: A Spotlight on Women and Youth 79 FIGURE 2.45. Correlates of More Gender Egalitarian Views a. Disagrees: men have more rights to jobs than women Disagrees Men have more rights to jobs than women Age group (Ref.: 50+ years) 15–29 years 30–49 years Number of children (Ref.: no children) 1 child 2 children 3 children 4+ Children Educational attainment (Ref.: Lower) Middle Upper –.5 0 .5 1 male female b. Disagrees: preschool child suffers with working mother Disagrees: Pre-school kid suffers with working mother Age group (Ref.: 50+ years) 15–29 years 30–49 years Number of children (Ref.: no children) 1 child 2 children 3 children 4+ Children Educational attainment (Ref.: Lower) Middle Upper –.4 –.2 0 .2 .4 .6 male female (continued) 80 Tunisia’s Jobs Landscape FIGURE 2.45. Correlates of More Gender Egalitarian Views (continued) c. Disagrees: men make better executives than women Disagrees: Men make better executives than women Age group (Ref.: 50+ years) 15–29 years 30–49 years Number of children (Ref.: no children) 1 child 2 children 3 children 4+ Children Educational attainment (Ref.: Lower) Middle Upper –.5 0 .5 1 male female d. Disagrees: university is more important for a boy than a girl Disagrees: University more important for a boy than a girl Age group (Ref.: 50+ years) 15–29 years 30–49 years Number of children (Ref.: no children) 1 child 2 children 3 children 4+ Children Educational attainment (Ref.: Lower) Middle Upper –.4 –.2 0 .2 .4 .6 male female Source: Based on data of WVS (World Values Survey), WVS Wave 7 (2017–2020): Tunisia 2019 (dashboard), King’s College, Old Aberdeen, United Kingdom, https://www.worldvaluessurvey.org/WVSDocumentationWV7.jsp. Note: Coefficients after OLS estimation. Dependent variable is coded 1-5 (strongly agree, agree, neither agree nor disagree, disagree, strongly disagree) for the question on job scarcity and 1-4 (strongly agree, agree, disagree, strongly disagree) for all other questions. 95 percent confidence interval. Access to the Labor Market: A Spotlight on Women and Youth 81 FIGURE 2.46. Self-Reported Frequency of Sexual Harassment in the Neighborhood, % of Adult Men and Women Female 6.6 19.9 28.5 43.7 Male 7.3 22.6 31.2 37.8 0 10 20 30 40 50 60 70 80 90 100 Very Frequently Quite frequently Not frequently Not at all frequently No answer or missing Source: Based on data of WVS (World Values Survey), WVS Wave 7 (2017–2020): Tunisia 2019 (dashboard), King’s College, Old Aberdeen, United Kingdom, https://www.worldvaluessurvey.org/WVSDocumentationWV7.jsp. ENDOWMENTS because of an impressive expansion in access to basic edu- cation in recent decades, gender gaps in enrollments have Another potential constraint on women’s participation closed at the primary level and even reversed at the sec- in the labor market is gender differences in human and ondary level (Figure 2.47).32 This implies that more girls physical endowments. The term endowments is considered than boys are graduating with secondary degrees. Like- rather broadly as human capital endowments (education wise, women significantly outnumber men at the univer- and skills), property ownership, control and access to pro- sity level; 149 women are enrolled at the tertiary level for ductive resources (assets, credit), and access to networks every 100 men (World Development Indicators). and information. Data from international assessments, though somewhat dated, show that learning outcomes are also better among Human Capital and Skills adolescent girls than among boys. In the 2012 PISA assess- Historically, Tunisia has exhibited large gender gaps in ment, a significantly smaller share of female than male school enrollments and educational attainment. However, 15-year-old students achieved scores that classify them as FIGURE 2.47. Primary and Secondary Gross Enrollment Rates and the Gender Parity Index a. Primary education b. Secondary education Primary Secondary 140 1.2 140 1.2 Primary school enrollment Primary school enrollment 1.1 120 1.1 Gender parity index 120 Gender parity index 100 1 100 1 rate (gross) rate (gross) .9 .9 80 80 .8 .8 60 60 .7 .7 40 40 .6 .6 20 .5 20 .5 0 .4 0 .4 70 75 80 85 90 95 00 05 10 15 20 70 75 80 85 90 95 00 05 10 15 20 19 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 19 20 20 20 20 20 year year GER female GER male GER female GER male Gender parity index Gender parity index Source: Based on data from the World Development Indicators, World Bank.   32 For individuals born after 1985, basic schooling constitutes primary and preparatory school, for a total of nine years of mandatory education (Assaad, Ghazouani, and Krafft 2017). 82 Tunisia’s Jobs Landscape FIGURE 2.48. Gender Gaps in the Ownership of Productive Assets, 2017 30 27 25 20 20 15 Percent 15 13 13 9 10 10 6 6 7 4 4 4 5 5 5 3 3 2 3 2 2 2 2 2 0 Agricultural land Land for Real estate Non-agricultural Car Motorcycle construction business Rural Men Rural Women Urban Men Urban Women Source: Based on data from ILO 2018b. Note: The figure shows the share (%) of adult men and women who report that they own the asset indicated solely or jointly. low performers (below PISA proficiency level 2) in all three men in on-the-job learning. For example, young women subjects, reading, mathematics, and science (OECD 2015). interviewed during focus group discussions in 2012 reported In the 2012 and 2015 assessments, girls performed signifi- that family concerns about women’s safety and societal cantly better than boys in reading, while boys showed a restrictions on female mobility prevented them from taking slightly better performance in mathematics, a pattern that up the casual, short-term filler jobs that often help young is found in many countries. There was, however, no signifi- men gain relevant skills and entry-level labor market expe- cant gender gap in the performance in science (Figure 2.9). rience (World Bank 2014b). In the 2015 assessment in Tunisia, more female than male students (40  percent vs 29  percent) reported they had science-related career aspirations (OECD 2018). This is Ownership of Property and Other Productive consistent with the observation that, in many Middle East Assets and Access to Capital and North African countries, the shares of women in sci- ence, technology, engineering, and mathematics is com- There are significant gender gaps in the control over pro- paratively high, often exceeding OECD averages (OECD ductive resources. Research conducted by the ILO (2018b) 2020; World Bank 2009). in 2017 using the women’s empowerment in agriculture index methodology shows that women are much less likely Despite the impressive increase in female school enroll- than men to report ownership of productive assets, espe- ments, low levels of educational attainment may be a cially agricultural land and motorized transport.33 In rural significant constraint on the labor force participation of areas, 27 percent of men reported that they owned agri- older women, who did not benefit from the expansion in cultural land, compared with only 6 percent of women. access to basic education over the past two decades. Illit- In transportation, 13 percent of rural men and 20 percent eracy levels are significantly lower among working-age of urban men reported that they owned a car, compared women than among working-age men, especially among with only 2 percent of rural women and 5 percent of urban older cohorts (see Figure 2.4). Projections suggest it will women (Figure 2.48). These gaps, which are similar for the take approximately two more decades to eliminate gender ownership of motorcycles, may be a factor explaining why gaps in educational attainment among the adult popula- women are typically less geographically mobile than men. tion (Evans, Akmal, and Jakiela 2020). Because educa- The gender gaps in the ownership of land for construction, tional attainment is positively linked to women’s labor real estate, and nonagricultural household businesses are force participation (see Figure 2.23, panel a), the changes smaller, but also favor men. Gender gaps in the ownership in the educational composition of the female working-age of property are partly related to the gender inequalities population in the next decades can be expected to raise female labor force participation. 33 Even though the survey is nationally representative at the household   level, it oversampled household heads and spouses for individual-level In addition to gender gaps in formal education, there is interviews. The survey is therefore not necessarily representative at the some evidence that women are disadvantaged relative to population level. Access to the Labor Market: A Spotlight on Women and Youth 83 FIGURE 2.49. Gender Gaps in Access to Finance 46 50 40 34 28 Percent 30 21 23 20 12 9 14 11 6 3 5 2 10 1 0 2 0 2014 2017 2014 2017 2014 2017 2014 2017 … a nancial institution … saved at a nancial … borrowed to start, … a mobile money account institution operate, or expand a account farm or business Men Women Source: Based on data of Global Findex (Global Financial Inclusion Database), World Bank, Washington, DC, https://globalfindex. worldbank.org/#data_sec_focus. Note: The figure shows the share (%) of men and women ages 15+ who report that they participated in the activity indicated. in the legal system that put women at a disadvantage in women’s rights, for example, a lack of legal provisions that accumulating property through marriage and inheritance. prohibit gender discrimination in access to credit. Women are much less likely than men to use financial products, such as savings accounts, credit, and loans. Networks Thus, 46 percent of men reported that they had accounts Women are disadvantaged in access to networks and infor- at financial institutions, compared with only 28 percent of mation. The share of rural women who report that they are women (Figure 2.49). Likewise, many fewer women than free to join various groups, such as religious groups, civic men saved at financial institutions (14 percent vs 23 per- groups, cultural associations, sports clubs, and political par- cent) or borrowed to start, operate, or expand a busi- ness (11 percent vs. 5 percent). Mobile money accounts ties, ranges between 32 percent and 38 percent, compared are not commonly used in Tunisia and were reported by with 53 percent to 56 percent among men (ILO 2018b).34 only about 2 percent of men and women. There is also no These gender gaps are smaller in urban areas, but still indication that gender gaps in access to finance are declin- observable. There are almost no gender gaps in the ability ing as overall access increases. While financial inclusion is to join groups among youth, except in sports clubs, where greater overall in 2017 than in 2014, gender gaps are as women continue to be disadvantaged. This indicates that large or slightly larger. these constraints disproportionately affect older genera- tions of women, especially in rural areas. Gender gaps in access to finance partly reflect social and cultural attitudes toward women and men’s roles in society, Even among younger Tunisians, however, access to networks which prioritize men’s labor market engagement and pro- may disadvantage women job-seekers. Young women ductive investments over those of women. For example, graduates report that they have few opportunities to 18  percent of rural men and 21  percent of urban men socialize and network outside the household and market- report that they are free to borrow from a financial insti- place, which may put them at a disadvantage in access- tution, compared with only 8 percent of rural women and ing labor market opportunities (World Bank 2014b). 6 percent of urban women. A recent impact evaluation of Similarly, interviews with young women in the ICT and a cash grant project targeted at poor women in rural Tuni- tourism sectors revealed that many felt disadvantaged by sia found that the intervention did not significantly affect standard recruitment practices, whereby information about women’s income, but may have positively affected the vacancies was shared informally or by personal contacts income-generating activities of other household members. (Kärkkäinen 2011). This suggests that the funds, though ostensibly targeted at women, were primarily used to promote the income- generating activities of the husbands and other household 34 Even among men, a significant share reported that they are not free   members (Ferrah et  al. 2021). These cultural traditions to join groups. More research is needed to clarify the constraints on the are reinforced by the lack of a robust legal protection for choices of men in this context. 84 Tunisia’s Jobs Landscape Given the cultural restrictions on women’s mobility out- FIGURE 2.50. Gender Gaps in Unpaid Work, 2017 side the household, women could potentially benefit from 5 ICT access. This would require that access to ICT be rela- 4 3.8 tively gender equitable. The evidence is mixed. A much 4 higher share of young men than women reported that 3 2.6 they use the internet regularly (78 percent vs 32 percent, 2.1 respectively). However, there are no marked gender gaps 2 1.4 in participation in online discussions (46 percent among 0.9 1.1 0.8 1 men vs 43 percent among women) or Facebook accounts (99 percent among men vs. 98 percent among women), 0 indicating that most young women do have some access to Hours spent on Hours of caring for domestic work on household members ICT, though perhaps not as regularly as men. a regular day on a regular day Rural Men Rural Women Urban Men Urban Women PREFERENCES AND CHOICES Source: Based on data from ILO 2018b. This section discusses gender differences in the types of jobs Note: The figure shows the share of hours per day of unpaid work men and women seem to prefer. These gender differences in among adult men and women. preferences and choices are clearly not rooted in biology, but arise from gendered social and cultural traditions. married women reported that the time they had spent on domestic chores had risen since the onset of the pandemic, compared with only 14 percent of married men (World Time Use and Family Formation Bank 2021a). In childcare, the self-reported increase was even more; 12 percent of married men and women Single women are more likely to be in the labor force than reported that they spent more time on the activity now married women, and the gap widens over the life cycle than before the pandemic. (see Figure 2.24). This suggests that marriage and mother- hood are associated with women’s decisions to leave or These gender differences in time allocation reflect cultural not enter the labor force. Moreover, traditions and cus- toms assign women in Tunisia with broad responsibilities traditions that assign women broad responsibilities for for providing childcare and other unpaid domestic work. unpaid work inside the household. Among women, 2 in Thus, women spend significantly more time than men on 3 believe that a preschool child suffers if a mother works unpaid work. Women report that they spend approxi- (see Figure 2.44). Similarly, in the 2018 Arab Barometer mately 4  hours a day on domestic work and an addi- data, 58 percent of women agree with the statement that it tional 2.1 hours (urban areas) to 2.6 hours (rural areas) is better for a household if a woman has the main respon- on care work (Figure 2.52). Over the course of a week, sibility for taking care of the home and the children rather this amounts to more hours than would be required by a than a man.35 This suggests that these norms are often full-time job: 42.7 hours in urban areas and 46.2 hours in internalized by women themselves. Women’s decision to rural areas. Conversely, men spend less than one hour a pursue employment opportunities outside the household day on domestic work and between 1.1 hours (urban) and may be considered subversive, with potentially negative 1.4 hours (rural) a day on care work (Figure 2.50). This consequences for the households and the women, including leaves the men with significantly more time to pursue paid because of diminished marriage prospects (World Bank work (see Figure 2.39). The gender gap in time spent on 2014b). In such an environment, gendered cultural tradi- unpaid care work observed in Tunisia is large by interna- tions and customs are difficult to separate from individual tional standards (ILO 2018a; Samman, Presler-Marshall, preferences. and Jones 2016). Women in rural areas are also dispro- portionately engaged in agricultural production for own Another factor that may influence the amount of time consumption (Hanmer, Tebaldi, and Verner 2017). women spend on unpaid activities is the availability and There is strong evidence that the COVID-19 pandemic 35  Arab Barometer Public Opinion Survey Series (database), Inter-university Consortium for Political and Social Research, Institute for Social Research, exacerbated existing gender differences in time use. In an University of Michigan, Ann Arbor, MI, http://www.icpsr.umich.edu/ online survey conducted in June–July 2020, 35 percent of icpsrweb/ICPSR/series/508. Access to the Labor Market: A Spotlight on Women and Youth 85 affordability of childcare services. International evidence Even among single women, there is a strong preference strongly suggests that improved access to childcare services for public sector jobs because these are seen as a positive (through crèches and preschools, longer elementary school signal in the marriage market. Krafft and Assaad (2020) days, and so on) significantly increases women’s employ- show that public sector employment significantly acceler- ment and labor force participation (Buvinic ´ and O’Donnell ates marriage among women in Tunisia. Qualitative evi- 2016; De Henau 2019; Halim, Johnson, and Perova 2019; dence also shows that households expect young women to Mateo Díaz and Rodríguez-Chamussy 2013; Padilla-Romo take up only employment that is considered appropriate, and Cabrera-Hernández 2018). In Tunisia, day-care atten- that is, work that is commensurate with a woman’s quali- dance is rare among children ages 3–36 months (Box 2.6). fications and preferably in the public sector to enhance Among children ages 3–6 years, about one in two attends a marriage prospects (World Bank 2014b). day-care center, with significantly higher rates of attendance among more affluent households compared with the poorer However, with rising female education levels, this strategy quintile (71 percent vs. 17 percent). While this income gra- is coming under pressure because the range of employ- dient might reflect a variety of factors, including more con- ment opportunities in the public sector is increasingly out servative gender role attitudes among poorer (often rural) of sync with the number of women graduates looking for households, it may also signal that cost and affordability such jobs. To the extent that formal or informal employ- constrain the access of poor households to childcare. There ment in the private sector and self-employment are not is substantial variation in day-care fees across regions and considered acceptable for educated women, unemploy- service providers, but the average fee of TD 140 amounts ment and inactivity may remain the only options. to over 30 percent of the median wage of working women with primary educational attainment. Moreover, in the public sector, workplaces with more than 50 workers are Geographic Mobility required to have on-site childcare facilities, and this may There is evidence that women are less geographically mobile partly explain women’s preference for public sector jobs than men, which further limits their opportunities. For (Moghadam 2017). example, the ICT companies interviewed by Kärkkäinen (2011) reported that women’s lack of mobility poses a barrier to the recruitment and promotion of women within com- Occupation and Sector panies. Similarly, interviews with women themselves reveal Women’s preferences for certain types of jobs and occupa- regional differences in willingness to travel, particularly tions are shaped by societal views of what types of jobs are over long distances, which may be a factor contributing to considered acceptable for women and would allow women differences across regions in women’s labor force participa- to combine employment with household responsibilities. tion rates (Hanmer, Tebaldi, and Verner 2017). As in many countries in the region, women in Tunisia have a strong preference for public sector employment over the Several factors may explain differences in women and men’s private sector (Mouelhi and Goaied 2018; Stampini and willingness to take up jobs that would require either a move Verdier-Chouchane 2011). Women are significantly overrep- to a different city, or longer commuting times. Many women resented in the public sector workforce. Employment in the avoid using public transport, especially at night. In addi- public sector generally offers more favorable working condi- tion, significantly fewer women than men own motorized tion, including shorter hours, greater job security, social secu- transport (see Figure 2.48). Qualitative interviews of the rity coverage, paid annual and sick leave, and better access World Bank (2014a) show that conservative gender roles to childcare services (Moghadam 2018) (see Figure 2.40). may limit young women’s ability to take up employment These features make it easier for women to balance paid and that would require moving out of the household. unpaid work and are therefore relatively more important for women than for men (Assaad and Barsoum 2019).36 Mobility constraints may be particularly severe for women with lower levels of education or from poorer back- grounds. Research in Jordan shows that there is a nega- 36  Feld, Nagy, and Osman (2020) use an experimental design to elicit from job-seekers their valuations of various job attributes in Egypt. They show tive correlation between female labor force participation that women are more sensitive to long commutes and value flexible work and commuting times at the district level, but only among schedules more than men do. While these results do not relate specifically to Tunisia, they provide support for the notion that women value the work- women with less than high school education (Kasoolu life balance more than men in a similar cultural context. et al. 2019). The research shows that women with lower 86 Tunisia’s Jobs Landscape BOX 2.6. Child Day-Care Centers and Preprimary Schools in Tunisia Overview of Child Day-Care Centers, Kindergartens, and Preprimary Schools Children ages 3 months–3 years attend day-care centers (or crèches). Children attend preprimary school from age 3 to age 6 (Table B 2.6.1 and annex Table A 2.2). Preprimary education is provided by the following: • Preprimary schools and kindergartens: These are socioeducational institutions privately or publicly owned or run by special- ized associations. Those publicly owned are municipal institutions under the Ministry of Local Affairs and Environment and the Ministry of Women, Family, Children, and Elderly. Private early childhood education institutions need to register and be approved by the Ministry of Women, Family, Children, and Elderly. Controls are carried out by inspectors to verify compliance with health and safety standards. • Kouttabs: These are religious institutions providing care services for children ages 3–5 years. Kouttabs introduce children to the Koran and teach them how to read, write, and count. They are under the aegis of the Ministry of Religious Affairs. • Primary schools: They offer preparatory courses with a reception class for children ages 5–6. This is an integral part of basic education, but not compulsory. The courses are delivered under the purview of the Ministry of Education, and they are offered in both public and private primary schools. • To date, virtually all day-care centers and kindergartens are privately owned and managed. Only about 1 percent of all children ages 3–36 months attend day-care centers.a The net attendance rate among children ages 3–6 is 51 percent. The rate is higher in urban areas (63 percent) than in rural areas (28 percent). The rate is also higher among children from affluent households (71 percent), compared with children living in the poorest households (17 percent). Private sector supply accounts for 94.0 percent of the facilities available among the kindergartens, compared with 5.9 percent provided by the public sector. The net school attendance rate among children ages 5–6 in preparatory primary school (reception class) is 90 percent: 94 per- cent in urban areas and 83 percent in rural areas.b Access to Public Kindergartens The cost of public kindergartens is between TD 25 and TD 45 per month. Since 2010, a program run by the Ministry of Women, Family, Children, and Elderly has allowed needy and low-income households to enroll their children in kindergartens for free for two years. The number of beneficiaries is about 6,000 children per year. Access to public kindergartens is based on information about the social status of the household and the pay slips of parents. For nonneedy households, registration is based on avail- able places and the location of residence. Access to public preparatory classes is provided free of charge. Cost of Private Day-Care Centers The cost of private day-care centers varies by service provider. To estimate an average cost of private centers, a survey of private day-care centers was conducted across the 24 governorates of Tunisia in April 2021. One center was selected ran- domly in each governorate, for a total of 21 private crèches.c Monthly fees are in the range of TD 40–TD 350 (Table B 2.6.2 and annex Table A 2.3). The average is about TD 140, but considerable variation exists across regions. The northern regions have the highest monthly fees. This was around 27 percent of the median monthly wage in 2019. It was about 35 percent of the median monthly wage of working women with primary education. TABLE B 2.6.1. Childcare Centers and Preprimary Schools (Public and Private), by Region Crèche Kindergarten Preprimary school Kouttab Region (ages 3–36 months) (ages 3–5 years) (ages 3–5 years) (ages 3–6 years) Greater Tunis 139 1337 579 444 North-East 33 771 0 206 North-West 30 378 22 124 Center-East 105 1322 697 384 Center-West 8 479 37 151 South-East 27 527 306 210 South-West 18 441 0 93 Total 360 5255 1641 1612 Source: Based on data of the Ministry of Women, Family, Children, and Elderly and the Ministry of Religious Affairs. (continued) Access to the Labor Market: A Spotlight on Women and Youth 87 BOX 2.6. Child Day-Care Centers and Preprimary Schools in Tunisia (continued) TABLE B.2 6.2. Monthly Fees at Private Day-Care Centers, by Region, April 2021 Region Lowest (TD) Highest (TD) Average (TD) Greater Tunis 210 350 263 North-East 130 320 210 North-West 80 140 100 Center-East 90 200 165 Center-West 40 120 80 South-East 70 110 90 South-West 40 90 65 Source: Based on data collected through phone interviews with one random day-care center in each governorate. a. Data provided by the Ministry of Women, Family, Children, and Elderly. b. Based on data of the 2018 Multiple Indicator Cluster Surveys, INS. See MICS (Multiple Indicator Cluster Surveys) (dashboard), United Nations Children’s Fund, New York, http://mics.unicef.org/. c. A few centers were not reached by the survey. levels of education rely disproportionately on public trans- challenge, particularly in light of population aging. By port, while more well educated women are more likely to 2040, about 16 percent of the Tunisian population will use private transport, which allows them to circumvent the be ages 65 or more. This section provides an overview of lack of safe public transport. Moreover, a comparison of the labor market situation of Tunisian youth. It focuses commuting times of single men and women, a segment of on key labor market indicators and on the difficulty of the population with comparatively few domestic respon- the school-to-work transition among university graduates, sibilities, shows that the commuting times of women are thereby providing evidence on key constraints to the significantly shorter than those of men. A plausible expla- employment of youth. nation is that concerns over harassment and safety on public transport reduce women’s willingness to take up Youth fare poorly in the labor market relative to prime- jobs that would involve longer commutes. age individuals. First, at 29.2 percent, the employment- to-population ratio among youth is over 15  percentage points below the level observed among prime-age workers Youth (ages 30–54). The large gap is driven primarily by the low employment ratios among youth ages 15–24, estimated Youth represent a large share of the Tunisian population at 19.7 percent in 2017 (Figure 2.51, panel a). This is of working age. The youth population is expected to hover ultimately ascribable to increasing school attendance around 21 percent of the total population and 33 percent rates. Among youth ages 15–24 and 25–29, an estimated of the working-age population (ages 15–64) in the next 46.5 percent vs. 5.5 percent, respectively, were attending two decades and decline gradually thereafter.37 Youth’s school in 2019. At 44  percent in 2017, the gap in the share in the population in Tunisia is modestly above the employment ratio between youth and adults is also sizable average among OECD countries (19.1 percent in 2020) among youth ages 25–29. The employment ratio has been and below regional (24.1  percent) and income group declining among youth, while it has been increasing among (23.7  percent) comparators. Over the coming 20  years, prime-age and older workers. Second, youth fare poorly demographic trends will ease some of the pressure on job relative to the rest of the working-age population also creation deriving from a rising share of youth. Yet, such in unemployment. About 1 in 3 youth ages 15–29 was trends make youth employment an extremely relevant unemployed in 2017. The share peaked at 38.2 percent in 2011 and has hovered around 33.0  percent ever since. Unemployment rates are higher among younger age-groups. 37 World Population Prospects 2019 (database), Population Division,   Department of Economic and Social Affairs, United Nations, New York, It is estimated at 34.9 percent among youth ages 15–24 https://population.un.org/wpp/. and 31.7 percent among youth ages 25–29 (Figure 2.51, 88 Tunisia’s Jobs Landscape FIGURE 2.51. Key Labor Market Indicators, by Age-Group, Youth (Ages 15–24 and 25–29) and Adults (Ages 30–64), 2006–17 a. Employment-to-population ratio 100 90 80 70 60 Percent 50 40 30 56.4 57.4 58.1 58.2 57.2 58.9 67.1 55.2 49.0 49.9 48.0 45.1 45.3 44.9 44.6 44.0 20 31.8 30.8 30.8 31.2 33.6 33.1 33.4 30.7 10 22.1 22.0 20.5 21.7 18.8 20.2 10.7 18.5 0 2006 2008 2009 2011 2013 2015 2016 2017 Youth 15–24 Youth 25–29 Prima age 30–54 Older workers 55–64 b. Unemployment rate 50 40 30 Percent 42.3 20 34.5 34.7 35.0 34.9 34.9 30.0 32.4 31.6 32.4 31.7 27.7 28.4 25.7 21.1 22.9 10 8.7 8.0 9.0 9.0 9.0 8.4 5.0 6.3 3.6 2.4 1.8 2.3 2.0 2.2 1.8 2.0 0 2006 2008 2009 2011 2013 2015 2016 2017 Youth 15–24 Youth 25–29 Prima age 30–54 Older workers 55–64 (continued) Access to the Labor Market: A Spotlight on Women and Youth 89 FIGURE 2.51. Key Labor Market Indicators, by Age-Group, Youth (Ages 15–24 and 25–29) and Adults (Ages 30–64), 2006–17 (continued) c. Inactivity rate 100 90 80 70 60 Percent 50 40 69.4 67.0 69.2 68.5 70.4 68.6 68.0 60.6 66.0 68.2 71.1 65.6 69.0 68.3 69.7 68.0 30 39.7 39.1 38.0 39.6 38.9 37.1 37.6 37.3 20 37.9 35.3 35.4 33.0 34.4 34.0 36.5 31.1 10 0 2006 2008 2009 2011 2013 2015 2016 2017 Youth 15–24 Youth 25–29 Prima age 30–54 Older workers 55–64 Source: Based on data from the Labor Force Survey (ENPE), INS. panel b). This compares with 9.0 percent among prime- among youth in the same age-group with tertiary education. age workers and 2.0 percent among older workers. Third, Young women ages 25–29 with university degrees face at 70 percent, inactivity is high among youth ages 15–24 an even greater risk of unemployment than young men because of high attendance rates in school. 38 Among (57.5 percent vs. 40.3 percent). Similarly, the unemploy- youth ages 25–29, the rate of inactivity is estimated at ment rate among youth ages 15–24 was 15.3  percent if 35.5 percent (Figure 2.51, panel c), which compares with they did not have school certificates, and the rate rose to 37.3 percent and 66.0 percent among prime-age and older 32.4 percent and 63.9 percent, respectively, if they have workers, respectively. secondary or tertiary education. Third, youth living in the more deprived areas of the country, namely, the North- Youth with tertiary education, young women, and youth West, the Center-West, and the southern regions, face a living in inland regions and urban areas face more difficul- higher probability of unemployment. For example, among ties in accessing jobs. First, the unemployment rate among youth ages 25–29, the unemployment rate was estimated young women is higher in both age-groups, 15–24 and at 51.3 percent in the South-West region, 45.2 percent in 25–29, relative to young men. The gender gap expands the South-East region, 39.9  percent in the North-West, from 3.4 to over 15.0 percentage points as women grow and 35.7 percent in the Center-West. This compares with older (ages 25–29) (Table 2.6). Second, youth unemploy- 20.9 and 22.5 percent in the Center-East and North-East ment rates rise with educational level from 16.2 percent regions, respectively. A sizable gap is also detected between among youth ages 25–29 with no education to 51.0 percent youth in urban and rural areas. In the 25–29 age-group, urban youth have an unemployment rate of 33.4 percent, compared with 27.3 percent among rural youth. Although  The large gap disappears after accounting for the large number of youth 38 who are attending school. In 2019, inactivity rates calculated among youth indicators have improved since the 2011 revolution, youth ages 15–24 not attending school were estimated at 23 percent. with secondary and tertiary education and particularly 90 Tunisia’s Jobs Landscape TABLE 2.6. Labor Force Participation Rate and Unemployment Rate of Youth by Age-Group, Sex, Educational Level, Decile of Household per Capita Expenditures, Georgraphical Area, and Profiles of Youth, by Age-Group, 2017 Youth 15–24 Youth 25–29 Profile Profile Unemployment rate All Unemployed Unemployment rate All Unemployed Sex Women 37.2 49.9 33.8 41.2 52.1 49.7 Men 33.8 50.1 66.2 25.9 47.9 50.3 Educational level No education 15.3 3.5 0.8 16.2 5 0.9 Primary 26.9 16 18.7 16.1 19.1 8.7 Secondary 32.4 61.8 56.3 23.3 44.5 33.5 Tertiary 63.9 18.5 24.2 51 31.2 56.9 Not stated 12.4 0.2 0 12.1 0.1 0 Decile of household consumption per capita – 2015 Lowest 54.6 10.1 13.5 42.3 7.7 7.1 2 45.6 10.2 12.1 41.8 9.1 9.5 3 40.9 10.1 10.5 42.4 9.4 10.5 4 43.4 10.6 12.2 39 10.4 10.7 5 41.7 10.1 11.1 36.8 10.2 10.6 6 44.3 10.9 11.9 37.9 10.4 11.3 7 41.5 10.4 9.6 34.2 9.8 9.9 8 36.7 10 8 35 11.3 12 9 35.6 9.8 6.3 31.4 10.9 10.4 Highest 38 7.8 4.9 24.2 10.9 7.9 Region Greater Tunis 46.7 23.1 29.8 33.4 26.7 30.6 North-East 23.9 13.5 11.2 22.5 13.5 10.5 North-West 45.8 9.4 12.5 39.9 8.3 10.4 Center-East 20 25.7 15.2 20.9 24.4 15.4 Center-West 36.7 13.6 12.2 35.7 11.8 10.6 South-East 46.9 9.3 12.7 45.2 9.4 12.8 South-West 50.2 5.4 6.6 51.3 5.9 9.6 Location Rural 30.1 32.8 31.5 27.3 29.9 23 Urban 37.7 67.2 68.5 33.4 70.1 77 Source: Based on data from the 2017 Labor Force Survey (ENPE) and the 2015 Household Budget Survey (EBCNV), INS. youth living in inland regions faced higher unemployment also reflects the profile of the youth population at large rates in 2017 than in 2006 (see annex Figure A 2-5). (see Table 2.6). The incidence of unemployment is greater among youth with tertiary education—this has doubled However, the majority of unemployed youth are women, over the past decade—and among youth in inland regions. have up to secondary education, and live along the coast Other groups account for larger shares of unemployed or in urban areas. A look at the distribution of unemployed youth. Youth with tertiary education contribute about youth by characteristics helps clarify the profile of the largest 41  percent of total youth unemployment: 24.2  percent groups. In addition to the incidence of unemployment, this among youth ages 15–24 and 56.9 percent among youth Access to the Labor Market: A Spotlight on Women and Youth 91 ages 25–29. Youth with up to secondary education make +31.9  percent among young women). Once youth are up the remaining 59 percent. Youth with secondary edu- active in the labor market, that is, looking for jobs, they cation contribute 44.7 percent (more than 176,000 indi- are less likely to be unemployed if they are married. This viduals) of total youth unemployment: 56.3  percent in link is more evident among young men than among young theyounger age-group and 33.5 percent in the older age- women (−16.9 vs −6.5 percent, respectively). Similarly, the group. Similarly, youth in inland regions contribute about number of children (ages 0–4 and 5–14) in the household is 43.7 percent of total youth unemployment, whereas the associated with the likelihood that youth household mem- largest share is located in Greater Tunis: 30.2  percent bers will be looking for jobs or be employed, especially the or about 120,000 individuals. The second largest group young men. The location of residence, too, has a sizable is found in the Center-East (15.3 percent). The profile is impact. Young men and young women living near Greater similar in the two age-groups. The prevalence of urban Tunis are less likely to be unemployed relative to youth in unemployed youth is largely driven by the georgraphical other regions. Youth in more affluent households are more distribution of the youth population, although, among likely to search for job and, once they search, are more likely youth ages 25–29, the urban concentration of unemployed to find jobs relative to youth in the poorest households. Sta- youth is higher (77 percent of urban unemployed youth, tistically, the marginal effect of living in a household in the compared with 70 percent of all urban youth). second or fifth quintile, respectively, relative to living in a household in the bottom quintile of the expenditure distri- Almost 6 youth in 10 have been unemployed for a year or bution rises from −10.2 percent to −26.8 percent among longer. The long-term unemployment rate among youth is young men and ranges from −3.9 percent to −20.9 percent too high. In 2015, 58 percent of youth ages 15–29 were among young women. estimated to be among the long-term unemployed, that is, searching for jobs for at least 12 months. The rate is About 4 youth in 10 are NEET: the NEET rate is the higher among young women (62.0 percent) and among highest among young women, youth with little education, well-educated youth: 64.0  percent among youth with youth in inland regions, and youth in poorer households. tertiary education relative to 45.0 percent among youth The share of NEET youth hovered around 40 percent of with primary education. Youth in Tunis (63.0 percent), the the population ages 15–29 over the decade, a rate that is North-West (71.0 percent), and in the southern regions above the OECD average (12.9 percent in 2019) and the (74.0 percent in the South-West and 61.5 percent in the average among regional and income group comparators. South-East) face higher long-term unemployment rates. NEET rates differ considerably across groups of youth. A small difference is detected across quintiles of household The NEET rate among boys and young men peaked at age consumption expenditure. In 2015, the long-term unem- 24 (44 percent) and then declined to 35 percent by age 29 ployment rate was estimated at 55 percent and 56 percent in 2015. Among young women, the rate increased with in the bottom two quintiles and at about 60 percent in the age and peaked at 67 percent at age 28 (Figure 2.53). On top three quintiles. average, the NEET rate among young women is estimated at 40  percent, almost 10  percentage points higher than Educational attainment, marital status, geographical loca- among young men (30.7  percent). Among young men, tion, and household welfare are strong correlates of the unemployment is the largest component of the NEET probability of unemployment. The patterns depicted above group, whereas, among young women, the largest share are corroborated by a multivariate analysis estimated sepa- of NEET is accounted for by young women who are out rately by sex among youth who participate in the labor of school and not engaged in the labor market. Youth ages market, that is, either as employed or as unemployed. Con- 25–29 and youth with no schooling (66.0 percent) or pri- trolling for a set of individual and household characteris- mary education (54.7 percent) are more likely to be not in tics, namely, age, educational attainment, marital status, education, not in employment, and not looking for jobs household size, number of children ages 0–4 and 5–14 in (Figure 2.54, panels a, b, c). NEET rates are higher among the household, region of residence, urban and rural loca- youth who reside in the Center-West (44.7 percent), South- tion, and quintile of household expenditure distribution, West (42.4 percent), and South-East (38.9 percent) (Fig- a subset of characteristics emerge that are significantly ure 2.54, panel d). NEET rates are higher at the bottom correlated with the probability of unemployment (Fig- of the household expenditure distribution. NEET rates ure 2.52; see Figure 2.55). The probability of unemploy- decline from 41.2 (63.4) percent in the lowest quintile to ment is greater among university graduates relative to youth 15.6 (36) percent in the highest quintile among youth ages with no schooling (+21.5 percent among young men and 15–24 (25–29) (Figure 2.54, panel e). 92 Tunisia’s Jobs Landscape FIGURE 2.52. Correlates of the Probability of Unemployment Among Youth, by Sex, 2015 ages=16 ages=17 ages=18 ages=19 ages=20 ages=21 ages=22 ages=23 ages=24 ages=25 ages=26 ages=27 ages=28 ages=29 Primary Secondary Tertiary married Household size Number of children ages 0–4 Number of children ages 5–14 North East North West Center East Center West South East South West urban-1 quintile 2 quintile 3 quintile 4 quintile 5 –0.6 –0.4 –0.2 0 0.2 0.4 Marginal effect Men Women Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. Youth living in coastal regions, youth with secondary edu- household, governorate of residence, urban or rural loca- cation, and young women contribute the largest shares to tion, and quintile of household expenditure distribution. the NEET population. The prevalence of NEET among It found that age, marital status, number of children, and specific groups of youth does not coincide with the shares position in the welfare distribution were the most impor- in the NEET population. The largest contributor to the tant correlates of the probability of inclusion in NEET, NEET population is women (56 percent), youth with sec- with some distinctions by sex and educational level (Fig- ondary education (52 percent), and youth in coastal regions ure 2.55). Among young men with secondary or tertiary (Greater Tunis, 20 percent; North-East, 12.1 percent; and education, age is an important correlate of NEET status. Center-East, 22 percent). The probability of inclusion in NEET increases monotoni- cally with age up to around age 24 (+36 percent relative to Age, marital status, urban location, and household wel- a 15-year-old) and then starts to decline a bit. A status as fare are strong correlates of the probability of inclusion married and living in a household with children ages 5–14 in NEET. The patterns depicted above are corroborated reduces the probability of inclusion in NEET. The same is by a multivariate analysis estimated separately by sex and true of residence in a more affluent household relative to educational level (up to primary education vs. secondary the poorest households (first quintile). By contrast, among or tertiary education). The analysis controlled for a set young men with no schooling or no primary education, of individual and household characteristics, namely, age, age does not seem to play any role, whereas the negative marital status, number of children ages 0–4 and 5–14 in the effects of marital status, number of children, and quintile Access to the Labor Market: A Spotlight on Women and Youth 93 FIGURE 2.53. Activity Status of Youth, by Age and Sex, 2015 a. Men 100 90 80 70 60 Percent 50 40 30 20 10 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Age Work only Education only Work and Education Not in the labor force Unemployed b. Women 100 90 80 70 60 Percent 50 40 30 20 10 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Age Work only Education only Work and Education Not in the labor force Unemployed Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. 94 Tunisia’s Jobs Landscape FIGURE 2.54. Youth NEET Rates by Age-Group, Sex, Educational Level, Region of Residence, and Quintile of per Capita Household Expenditure, 2015 a. By age-group b. By age-group and sex 100 100 90 90 80 80 70 70 60 60 Percent Percent 50 50 40 40 30 30 62.4 50.2 20 20 37.4 28.1 28.5 27.8 10 10 0 0 Youth 15–24 Youth 25–29 Youth 15–24 Youth 25–29 Women Men c. By age-group and educational level d. By age-group and region of residence 100 100 90 90 80 80 70 70 60 60 Percent Percent 50 50 40 40 73.3 30 60.7 55.6 53.4 49.1 46.9 30 59.6 57.9 61.5 20 20 47.5 44.0 55.6 34.3 38.8 31.8 44.8 10 24.3 18.4 10 27.7 24.7 30.9 21.6 0 0 Youth 15–24 Youth 25–29 Youth 15–24 Youth 25–29 None Primary Secondary Tertiary Greater Tunis North-East North-West Center- Center-West South-East South-West East e. By quintile of per capita household expenditure 100 90 80 70 60 Percent 50 40 30 63.4 57.3 51.5 46.3 20 41.2 32.1 36.0 27.4 22.8 10 15.6 0 Youth 15–24 Youth 25–29 Poorest Quintile 2 Quintile 3 Quintile 4 Richest Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. Access to the Labor Market: A Spotlight on Women and Youth 95 FIGURE 2.55. Correlates of the Probability of Inclusion in NEET Among Youth, by Sex and Educational Level, 2015 a. Young men b. Young women ages=16 ages=16 ages=17 ages=17 ages=18 ages=18 ages=19 ages=19 ages=20 ages=20 ages=21 ages=21 ages=22 ages=22 ages=23 ages=23 ages=24 ages=24 ages=25 ages=25 ages=26 ages=26 ages=27 ages=27 ages=28 ages=28 ages=29 ages=29 married married Household size Household size Number of children ages 0–4 Number of children ages 0–4 Number of children ages 5–14 Number of children ages 5–14 ARIANA ARIANA BEN ANOUS BEN ANOUS MANOUSA MANUOSA NABEUL NABEUL ZAGHOUAN ZAGHOUAN BIZENIE BIZENIE BEJA BEJA JENDOUBA JENDOUBA LE KEF LE KEF SILIANA SILIANA SOUSSE SOUSSE MONASIN MONASIN MAHJIA MAHJIA SFAX SFAX KAIROUAN KAIROUAN KASSENNE KASSENNE SIDI BOUZIO SIDI BOUZIO GABES GABES MEDENNE MEDENNE TATAOUINE TATAOUINE GAFSA GAFSA TOZEUN TOZEUN KEBILI KEBILI urban-1 urban-1 quintile 2 quintile 2 quintile 3 quintile 3 quintile 4 quintile 4 quintile 5 quintile 5 –0.4 –0.2 0 0.2 0.4 –0.5 0 0.5 Marginal effect Marginal effect Men-up to primary Men-secondary or higher Women-up to primary Women-secondary or higher Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. of household welfare persist. In the case of young women in a governorate besides the governorate of Tunis. The with secondary and tertiary education, the effect of age on position in the household welfare distribution retains its the probability of inclusion in NEET is positive and rises negative effect on the chance of inclusion in NEET. How- up to around age 27. The marginal effects are also larger ever, the magnitude of the effect is considerably smaller with respect to young men. Unlike young men, a status except among the highest quintile. Among young women as married and living in a household with children ages with no schooling or only primary education, the effect of 0–4 increases the chances of inclusion in NEET, possibly age vanishes, whereas married status preserves its positive because of the care responsibilities that women typically effect, together with the dummy for urban residence and take on when they marry and have children. The number its negative effect. The effect of the position in the house- of children ages 5–14 has a negative effect on the likeli- hold expenditure distribution loses its significance except hood of inclusion in NEET. Urban residence decreases the among young women in the top quintile. Among young chances of inclusion in NEET among young women with men, the combined effect of marital status and quintile of higher education. The availability of job opportunities and household expenditure indicates a negative effect of mar- high-quality childcare services may act as a pull factor riage on the probability of inclusion in NEET at any quin- toward the labor market. Similar is the effect of residing tile relative to the status as not married and in the lowest 96 Tunisia’s Jobs Landscape FIGURE 2.56. The Main Reason for Being Out of the Labor Force and Not Looking for Jobs Among Youth, by Sex and Educational Level, 2015 100 2.0 1.6 0.3 1.2 3.9 1.5 6.0 7.0 8.5 8.7 4.0 1.6 90 9.0 7.0 27.2 15.4 80 3.5 5.0 2.4 70 45.8 22.3 23.1 60 75.7 79.9 79.1 Percent 50 59.1 10.6 40 78.1 30 56.6 50.7 20 1.5 10.9 35.0 2.8 6.0 10 4.4 13.9 10.7 10.7 7.7 0 Men Women Men Women Men Women Men Women None Primary Secondary Tertiary No jobs available Not willing to work Household duties Unable to work Other Source: Based on data from the 2015 Household Budget Survey (EBCNV), INS. quintile.39 The magnitude of the effects increases at higher Lack of jobs seems to be a constraint only among young quintiles. A similar negative effect, smaller in magnitude, is women with tertiary education. A stated preference for estimated for single young men. By contrast, in the case of not engaging in the labor market is rare among young young women, the estimated effect is negative and increas- women, at less than 9  percent; the shares are larger, at ing in magnitude along the distribution for single women about 20 percent, among young women with primary and and positive and roughly stable along the distribution for secondary education.40 married women. Inactivity seems to be associated with exclusion in the The self-reported reasons for inactivity differ by sex and minds of young men with little education. Among young educational level. About 8 in 10 young men with no men with tertiary education, a lack of jobs is the main schooling report that the main reason they are not looking reason for inactivity. Among young women with univer- for jobs is their inability to work. Young men with higher sity degrees, the reason is a combination of lack of jobs and educational attainment mention primarily lack of jobs, gender roles. Three facts can be derived from the evidence and this share rises from 50.7 percent among young men obtained through the multivariate analysis and self-reported with primary education to more than 78.0 percent among reasons for inactivity and not looking for jobs. First, age is young men with tertiary education (Figure 2.56). House- key among youth with tertiary education. The probability hold duties are a key factor among more than 70 percent of of inclusion in NEET among this group increases as time young women on average, with peaks at 80 percent among young women with primary and secondary education. 40  Gender identity norms might be so fully internalized that they become part of one’s self-conception, thereby shaping preference. Behaviors and  Estimates of the interaction between marital status and quintile of house- 39 choices may be affected by concerns about social image and the reputational hold expenditure are available on request. consequences of deviating from the prescribed behavior (Bertrand 2020). Access to the Labor Market: A Spotlight on Women and Youth 97 FIGURE 2.57. Annualized Change in the Population, Labor Force, and Employment, by Age-Group, 2006–17 6.0 5.3 5.5 5.2 Population Labor Force Employment 4.5 4.8 4.8 4.0 3.4 3.5 3.6 2.4 2.1 1.7 2.2 2.3 2.0 1.7 1.7 1.9 2.0 1.5 1.5 1.3 0.7 1.0 Percent 0.0 –0.5 –0.6 –0.3 –2.0 –1.6 –1.8 –4.0 –3.3 –3.7 –6.0 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 Source: Based on data from the Labor Force Survey (ENPE), INS. passes likely because of the more lengthy school-to-work 1.4 percent, while the labor force grew at a rate of 1.7 per- transition arising from the lack of jobs, particularly among cent a year (Figure 2.57). Youth ages 25–29 were the most university graduates. Second, household responsibilities act affected by the slow employment creation. On average, in opposite directions among young men and young women their total employment declined by −0.3  percent a year in line with assigned gender roles. In the case of young men, (−13,100 overall), while the number of such youth in the household responsibilities translate into a motive to obtain labor force increased at an annualized rate of about 1 per- a job to support household members, whereas, in the case cent (68,500 overall). The number of youth ages 15–24 in of young women, household duties are the main reason for the labor force declined, as did the number of such youth withdrawing from the labor market and taking on the role who were employed, thanks to higher enrollment rates of caregiver. Third, inclusion in NEET is not a luxury that and longer attendance in school. By contrast, among older only middle-class and affluent youth can afford, but rather individuals, the economy generated employment at a more an issue of exclusion among young men with little education. rapid average rate than the rate of expansion in the labor If they are married, young women experience a greater prob- force. Although the rate of employment creation accelerated ability of inclusion in NEET regardless of their position along after the revolution, it was not sufficient to absorb the large the welfare distribution. Young men, by contrast, engage in number of university graduates. As employment increased the labor market to support their newly formed families. by about 53,000 jobs a year on average, almost 65,000 addi- tional youth graduated, on average, between academic year Drivers of youth unemployment and idleness can be classified 2012/13 and academic year 2017/18 (Figure 2.58). If one into three groups: labor demand, labor supply, and institu- takes into account the occupational composition of employ- tional factors. Evidence on the first two groups of factors are ment, the deficit is striking. Between 2011 and 2017, the presented below. On the institutional component, this study number of high-end jobs, including managers, professionals is limited to an overview of active labor market policies, and technicians, and associate professionals, that university whereas other factors such as social insurance and assistance graduates can aspire to obtain, rose by less than 11,000 a systems, employment protection laws, wage-setting mecha- year (about 64,000 overall). nisms, and minimum wages are not examined. On the supply side, the quality of learning may constrain Sluggish employment creation, especially in high-end jobs, the ability of university graduates to land a job. Consid- is one of the main drivers of unemployment and inactivity erable progress has been achieved in Tunisia in enroll- among youth, particularly university graduates.41 In 2006–17, ment and completion at secondary and tertiary level. total employment rose at an average annualized rate of Girls have outstripped boys in these areas. The quality of learning and the relevance of education are among the main reasons for the lack of capacity of the coun- 41  In addition to economic growth, the presence of a well-paying public sec- tor is often mentioned as one of the possible causes of high unemployment try to produce employable graduates, that is, individuals rates among university graduates. Chapter 3 provides evidence on wage with the skills and qualifications needed to find a job gaps between private and public sector jobs among university graduates and on the labor market status and household characteristics of university regardless of the educational attainment. The quality of graduates not employed in public administration. learning in Tunisia is below comparators countries, for 98 Tunisia’s Jobs Landscape FIGURE 2.58. Change in the Number of University Graduates, the Employed, and the Employed in High-End Jobs, Circa 2011–17 University graduates 64,540 (2012/13–2017/18) Total employment 53,056 (2011–2017) Total employment in high-end 10,675 occupation (2011–2017) – 10,000 20,000 30,000 40,000 50,000 60,000 70,000 Net change Source: Based on data from the Ministry of Higher Education and Scientific Research; Labor Force Survey (ENPE), INS. FIGURE 2.59. Distribution of Graduates, by Field of Study and Academic Year, 2012/13–2017/18 100% 4.1 3.2 3.6 3.3 3.6 3.7 9.0 9.1 8.7 10.4 9.7 9.1 80% 16.9 18.5 19.3 20.1 19.4 19.4 60% Percent 26.5 25.3 25.2 23.8 23.9 23.7 40% 1.5 1.6 1.4 1.6 1.5 1.6 23.8 25.6 26.3 25.5 27.3 28.6 20% 0.7 0.2 0.3 0.4 0.6 0.5 17.5 16.6 15.2 14.9 14.1 13.5 0% 2012/2013 2013/2014 2014/2015 2015/2016 2016/2017 2017/2018 Letters and arts Teaching Social sciences, commerce and law Agriculture Science Engineering and construction Health and social protection Services Source: Based on data from the Ministry of Higher Education and Scientific Research. example, in mathematics and science test scores (see Fig- social sciences, including business and law, in the 2017/18 ure 2.7; Figure 2.8). academic year (Figure 2.59). While the number of grad- uates in the humanities declined from over 11,500 in University graduates tend to select curricula that are not in 2012/13 to about 7,800 in 2017/18, the number graduating line with private sector demand. There is an important gap in the social sciences rose by about 1,000 during the period between the competencies required by the labor market and (about 16,600 in 2017/18). The share of graduates in the sci- the student demand for higher education. About 35.0 per- ences declined to 23.7 percent (about 13,800 in 2017/18), cent of employers in Tunisia identify an inadequately and the share in engineering and construction rose from educated workforce as a major constraint to business 17 percent to over 19 percent (about 11,300 in 2017/18).43 operation and firm growth. The share had increased from In 2012–17, the number of wage workers employed as 29.1  percent in 2013 and is above the average in the managers, professionals, and technicians increased by Middle East and North Africa (20.4 percent).42 The cur- about 110,000 or 21 percent. The number of teachers, ricula selected by university graduates are not aligned with the needs of the labor market. About 4 university 43 Young women graduated predominantly in the humanities. A smaller   graduates in 10 obtained a degree in the humanities or share enrolled and graduated in science, technology, engineering, and mathematics, including statistics, construction, and ICT. See TLMPS (Tuni- sia Labor Market Panel Survey 2014) (dashboard), Economic Research 42  2020 data of Enterprise Surveys (dashboard), World Bank, Washington, DC, Forum, Gza, Egypt, http://www.erfdataportal.com/index.php/catalog/105/ https://www.enterprisesurveys.org/. data-dictionary. Access to the Labor Market: A Spotlight on Women and Youth 99 particularly primary- and secondary-school teachers, rose than 1 percent according to 2015 household budget survey by more than 100,000 (more than 90 percent). The number data. This may contribute to a less smooth transition from of ICT, legal, social, and cultural professionals increased school to work and be positively correlated with a larger by 34,000, and science and engineering occupations added share of NEET youth and a longer duration in the transi- 5,000 workers (+20 percent) (Figure 2.60). The number of tion (Manacorda et al. 2017). Third, Tunisia stands out health professionals and associate professionals expanded for markedly long transitions from school to first employ- by about 5.0  percent and 14.5  percent, respectively. By ment, an average of 35 months and a median of 29 months, contrast, the number of business, administration, and legal below only to Jordan and to West Bank and Gaza) (Fig- associate professionals fell by over 50,000 (40 percent). ure 2.61, panel a). The transitions are particularly long This means the choices of youth are not aligned well with among women (a median of 41.6 months vs. 21.2 months the needs of the private sector. Graduates in the social sci- among men). Fourth, a large share of youth are expected ences and law face more challenges in obtaining jobs given never to transit to first employment: 23 percent on average the decline in the number of employed associate profes- in Tunisia compared with 17 percent in the Middle East sionals in these fields. The rise in the number of wage jobs and North Africa; the share is disproportionately larger as science, engineering, and health professionals was lim- among women (35  percent vs. 12  percent among men) ited. Graduates in the humanities can reasonably expect to (Figure 2.61, panel b). This evidence is particularly con- find jobs given the continuous expansion of public sector cerning because the probability of finding a job among hiring in this field. However, the rise in public sector hiring youth falls as the duration of the transition rises in both is not sustainable, and hiring has recently diminished. developing and advanced countries. This phenomenon is known as negative duration dependence. This contributes to lengthy school-to-work transitions, which many youth, particularly young women, do not Assigned gender roles constraints the labor market partici- complete. First, by age 21, 50 percent of youth ages 15–29 pation of young women after marriage. Recent research by leave school in Tunisia. This is similar to the outcome the World Bank’s behavioral science team shows that both observed in advanced economies and higher than the men and women view men’s participation in housework average age in middle-income countries (OECD 2015). favorably, which suggests that gender differences in time Second, few Tunisian youth combine work and study, less allocation are driven by customs and traditions. FIGURE 2.60. Change in the Number of Wage Workers Employed in High-End Jobs, by Occupation, 2012–17 120,000 Change in wage employment 100,830 100,000 80,000 60,000 40,000 34,141 20,000 5,392 7,021 674 816 – (20,000) (40,000) (60,000) (52,773) (80,000) administrative, legal, IC (business, administrative, Science and engineering Health associate Teaching professionals Science and engineering Health professionals professionals (business, associate professionals professionals Other professionals Other associate professionals technicians) legal, ICT) Source: Based on data from the Labor Force Survey (ENPE), INS. Note: High-end jobs include managers, professionals, technicians, and associate professionals. See ISCO-08 classification, ISCO (International Standard Classification of Occupations), International Labour Organization, Geneva, https://www.ilo. org/public/english/bureau/stat/isco/. 100 Tunisia’s Jobs Landscape FIGURE 2.61. Duration of School-to-Work Transitions and Probability of Never Transiting from School to Work, 2013 a. Estimated average and median duration b. Predicted probability of never transiting 60.0 60.0 35.3 Mean Median 50.0 50.0 28.7 Number of months 40.0 40.0 Percent 30.0 30.0 23.0 20.0 20.0 10.0 10.0 0.0 0.0 Cambodia Madagascar Peru Samoa Ukraine Moldova, Rep. Nepal Benin Uganda Russian Fed. Brazil Armenia El Salvador Vietnam Kyrgyz Rep. Tanzania Togo Jamaica Egypt Tunisia Palestine Jordan Vietnam Cambodia Madagascar Uganda Togo Tanzania Russian Fed. Peru Jamaica Moldova, Rep. Kyrgyz Rep. Egypt Ukraine Tunisia Nepal Jordan Armenia El Salvador Palestine Benin Samoa Source: Based on data from Manacorda et al. 2017. Most active labor market policies in Tunisia target youth • The Programme d’accompagnment des promoteures who have secondary and tertiary education, consist of des petites enterprises (Programme for Mentoring Pro- wage subsidies, and lack monitoring and evaluation. moters of Small Enterprises) promotes entrepreneur- Reviews indicate that the large majority of the govern- ship among youth. Participants may benefit from free ment’s active labor market policies target university grad- training, orientation and coaching services, and finan­ uates (Boughzala 2019). In 2018, four policies represented cing. The Agence Nationale pour l’Emploi et le Travail the largest component of active labor market policies (see Independant (National Agency for Employment and annex Table A 2-4). Independent Work, ANETI) and the Tunisian Bank of Solidarity participate in the program. The latter is the • The Stage d’initiation à la vie professionnelle (Initia- main provider of financing. The program is relative tion into Work Program) was introduced in 1987. It small compared with the others. It covers fewer than has been renamed Contrat d’initiation à la vie profes- 4,000 youth a year. sionnelle (Contract for Integration into Working Life). It is aimed at facilitating job access by helping youth Wage subsidies provide temporary employment oppor- acquire professional experience. It offers university tunities to beneficiaries and labor to firms at lower cost, graduates social security coverage, along with a mini- often at the expense of significant deadweight loss and mum stipend of TD 300 a month, of which 50 percent substitution effects. The goal of wage subsidies is to is paid by the government and 50 percent by participating stimulate the demand for labor by subsidizing the associ- enterprises that commit to hiring at least 50 percent of ated cost among firms. This can support young workers, the youth they have supported under the program. whose productivity may be low initially. Because the cost • The Contrat d’adaptation et d’insertion professionnelle of hiring is reduced, employers may become more keen to (Adapting to the Workplace Insertion Contract) is the employ the target groups. In Tunisia, given the abundant analogue of the Contract for Integration into Working supply of university graduates, the relative price of their Life for youth who have not graduated. Target youth labor should be adjusted downward. However, the exis- receive a lower stipend and cannot participate more tence of collective wage agreements limits the potential for than once in the program, unlike the case of the Con- such adjustments, thus incentivizing hiring youth infor- tract for Integration into Working Life. mally. Wage subsidies can play a role by making formal • Service civile volontaire (Voluntary Civil Service) was employment of youth more attractive to employers. Wage introduced in 2010 to meet the special needs of inland subsidies take several forms depending on how they are set regions, which are less urbanized and have relatively (for example, a reduction in social security contributions fewer formal firms. or payments of a fraction of the wage), who receives them Access to the Labor Market: A Spotlight on Women and Youth 101 (workers or employers), who is eligible (all workers, new evaluations through the use of administrative data. Well- hires, first-time job-seekers, and so on), and the type of designed, monitored, and evaluated wage subsidies for conditionalities on employers (Angel-Ardinola, Nucifora, private sector employment have the potential to support and Robalino 2015). Recent evaluations indicate that, in temporary employment among youth and allow youth to general, wage subsidies are an effective tool for raising acquire work experience, thus playing the role of stepping employment rates among eligible individuals, but mainly stone to more permanent employment. This can also help as a way to provide work experience in the short term as contain the negative effects associated with deadweight opposed to permanent employment, particularly among losses and substitution effects and mitigate the fiscal cost the long-term unemployed (Card, Kluve, and Weber 2018; of public sector recruitment. McKenzie 2017). In addition, wage subsidies seem to suc- ceed in providing support to firms and preventing them In general, active labor market policies have modest posi- from shedding workers in the event of temporary shocks. tive effects that are smaller than those typically expected by There are two key issues with wage subsidies: deadweight beneficiaries and policy makers. Active labor market poli- losses, that is, the risk of subsidizing jobs that would have cies are an instrument governments have long adopted to been created anyway, and substitution effects, that is, the intervene in the labor market with the goal of generating possibility that employers substitute nonsubsidized workers more and better employment opportunities for workers, with subsidized workers. Evaluations in Europe indicate often workers with little other opportunities. In addition to wage subsidies and public works that share the same objec- that deadweight and substitution effects can affect around tive, the policies typically operate on the labor supply side 90 percent of the jobs (Martin 2000). by increasing the employability of workers (training pro- grams, including business training to foster self-employ- The degree of targeting, the extent to which the subsidies ment), and on matching between labor demand and labor affect new hires, whether they are standalone or part of supply though job search and matching programs (job a comprehensive package, and the existence of effective search assistance and matching). A review of recent evalu- monitoring and evaluation are crucial to the success of ations of active labor market policies in developing coun- wage subsidies. In Tunisia, the various wage subsidies tries finds that skill training, wage subsidies, and job search introduced in recent decades have helped youth land a first assistance programs have modest impacts in most cases job and gain work experience. This might have helped mit- (McKenzie 2017). By contrast, expectations of the impact igate the increase in the youth unemployment rate, which of such programs among participants and policy makers might have been higher without the wage subsidies for are typically overoptimistic. While the small effects of most youth. Difficulties in defining and enforcing the target- of the programs might be ascribable to the fact that labor ing, the assignment of trainees to the relevant tasks within markets ultimately function relatively well, particularly in firms, and enforcing the conditionalities to retain trainees urban areas, it is also possible that other constraints limit and workers after the end of the subsidy cast doubt on the job creation (McKenzie 2017). If this is so, then nontradi- efficacy of the wage subsidies. An evaluation of the Con- tional active labor market policies that address, for example, tract for Integration into Working Life Program, based on sectoral and spatial mismatches whereby workers are stuck a graduate tracer study with a sample of 4,700 youth who in occupations or locations that do not meet demand may had graduated in 2004 and were interviewed in 2005 and be more effective at tackling unemployment. 2007, finds that the program reduced the joblessness rate of university graduates by 8  percentage points, assum- Encouraging technical and vocational education and ing no deadweight or substitution effects, with an esti- training (TVET) may be a promising avenue for boosting mated unitary cost of about TD 18,000 (Broecke 2013). youth employment in the long term. Because the creation However, self-selection into the program, with subsidies of high-end jobs has been weak and the number of univer- allocated on a first-come first-served basis, and a lack of sity graduates is well above the capacity of the economy, combination with other services, such as training, coun- TVET could be a viable avenue to boost employment selling, and job search assistance, raise questions about growth in Tunisia. The number of TVET graduates is con- the efficacy of the program in the medium term (Broecke siderably below the number of university graduates. In 2013). Going forward, it is critical that the program target 2017, about 27,500 individuals graduated from TVET, graduates with the highest risk of unemployment, the eligi- and over twice as many (56,279) graduated from univer- bility criteria are accurately met, and adequate monitoring sity, even though the enrollment numbers are quite similar of conditionalities is implemented, together with ex post (52,075 enrolled in public TVET institutes and 57,503 102 Tunisia’s Jobs Landscape in universities).44 According to the World Bank (2020a), same time, a study conducted by the National Observa- TVET is perceived as an unattractive option given the tory for Employment and Skills (ONEQ 2017) finds that, low status of technical schools (collèges techniques) at the four years after their graduation, about 65  percent of lower-secondary level and the lack of high-quality alterna- TVET graduates were employed, though with a gender gap tives at the upper-secondary level (OECD 2015). At the that favors men because only 20 percent of the employed men graduates were looking for a job compared with 41 percent of the women graduates. In addition, the study finds that TVET graduates in mechanic arts, construc- 44  For TVET graduates, see ONEQ (2019); the information on university graduates is taken from data of the Ministry of Higher Education and tion, transport, electricity, and fish and aquaculture had a Scientific Research. higher likelihood of getting a job. REFERENCES CHAPTER 2 Alesina, Alberto Francesco, Paola Giuliano, and Nathan Nunn. 2013. “On Card, David E., Jochen Kluve, and Andrea Weber. 2018. “What Works? the Origins of Gender Roles: Women and the Plough” Quarterly Journal A Meta-analysis of Recent Active Labor Market Program Evaluations.” of Economics 128 (2): 469–530. Journal of the European Economic Association 16 (3): 894–931. Alfani, F., Dhrif, D., V., Molin, D., Pavelesku, and M., Ranzani. 2021. “Living Cardona-Sosa, Lina, Luz Adriana Flórez, and Leonardo Fabio Morales Standards of Tunisian Households in the Midst of the COVID-19 Pan- Zurita. 2016. “Intra-Household Labor Supply after an Unemploy- demic.” Policy Research Working Paper 9581, World Bank,, Washing- ment Event: The Added Worker Effect.” Borradores de Economía 944, ton, DC. Economic Studies Department, Central Bank of Colombia, Bogotá, Alon, T., Coskun, S., Doepke, M., Koll, D., and Tertilt, M. 2020. “From Colombia. Mancession to Shecession: Women’s Employment in Regular and Pan- Chakravarty, S., S. Das and J. Vaillant. 2017. “Gender and Youth in Sub- demic Recessions.” NBER Working Paper 28632, National Bureau of Saharan Africa. A Review of Constraints and Effective Interventions.” Economic Research, Cambridge, MA. Policy Research Working Paper 8245, World Bank, Washington, DC. Amin, M., and A. Islam. 2019. “Paid Maternity Leave and Female Chambers, V. and C. Cummings. 2014. “Building Momentum: Women’s Employment: Evidence Using Firm-Level Survey Data for Develop- Empowerment in Tunisia.” Case Study Report, Women’s Empower- ing Countries.” Policy Research Working Paper 8715, World Bank, ment, Overseas Development Institute, London. Washington, DC. De Henau, Jérôme. 2019. “Investigating in Free Universal Childcare in South Angel-Ardinola, Diego F., Antonio Maria Nucifora, and David Alejandro Africa, Turkey and Uruguay: A Comparative Analysis of Cost, Short- Robalino, eds. 2015. Labor Policy to Promote Good Jobs in Tunisia: Term Employment Effects, and Fiscal Revenue.” With Debbie Budlender, Revisiting Labor Regulation, Social Security, and Active Labor Market Fernando Filgueira, Ipek Ilkkaracan, Kijong Kim, and Rafael Mantero. Programs. Directions in Development: Human Development Series. Discussion Paper 28 (February), UN Women, New York. Deere, C. D., and C. R. Doss. 2006. “The Gender Asset Gap: What Do We Washington, DC: World Bank. Know and Why Does it Matter?” Feminist Economics 12 (1–2): 1–50. Assaad, R. and G. Barsoum. 2019. “Public Employment in the Middle East Evans, D., M. Akmal, and P. Jakiela. 2020. “Gender Gaps in Education: The and North Africa.” IZA World of Labor 2019: 463. Long View.” CGD Working Paper 523, Center for Global Develop- Assaad, R., S. Ghazouani and Caroline Krafft. 2017. “The Composition ment, Washington, DC. of Labor Supply and Unemployment in Tunisia.” ERF Working Paper Feld, Brian, AbdelRahman Nagy, and Adam Osman. 2020. “What Do 1150, Economic Research Forum, Giza, Egypt. Jobseekers Want? Estimating Reservation Wages and the Value of Job AUC (African Union Commission) and OECD (Organisation for Eco- Attributes.” ERF Working Paper 1414, Economic Research Forum, nomic Co-operation and Development). 2021. “Africa’s Development Giza, Egypt. Dynamics 2021: Digital Transformation for Quality Jobs.” AUC, Addis Fernandes, Reynaldo, and Fabiana de Felicio. 2005. “The Entry of the Wife Ababa; OECD, Paris. into the Labor Force in Response to the Husband’s Unemployment: Bertrand, M. 2020. “Gender in the Twenty-First Century.” AEA Papers and A Study of the Added Worker Effect in Brazilian Metropolitan Areas.” Proceedings 110 (May): 1–24. Economic Development and Cultural Change 53 (4): 887–911. Bhalotra, S. and M. Umaña-Aponte. 2010. “The Dynamics of Women’s Ferrah, S., J. Gazeaud, N. Khan, E. Mvukiyehe, V. Shrotri, O. Sterck, and Labour Supply in Developing Countries.” IZA Discussion Paper 4879, S. Zineb. 2021. “Enhancing Female Entrepreneurship through Cash Institute for the Study of Labor, Bonn, Germany. Grants: Preliminary Evidence from a Randomized Controlled Trial in Bicchieri, C. 2017. Norms in the Wild: How to Diagnose, Measure, and Rural Tunisia.” Endline Report, March. Change Social Norms. Oxford, UK: Oxford University Press. Fortin, N. M., and T. Schirle. 2006. “Gender Dimensions of Changes in Boughzala, M. 2019. “Marché du travail, dynamique des compétences et Earnings Inequality in Canada.” In Dimensions of Inequality in Canada, politiques d’emploi en Tunisie.” European Training Foundation, Turin, edited by David A. Green and Jonathan R. Kesselman, 307–46. Italy. Vancouver: UBC Press. Boukhayatia, Rihab. 2018. “Tunisia’s Violence against Women Law Seemed Gaddis, I., and S. Klasen. 2014. “Economic Development, Structural Perfect on Paper: In Practice, It Hasn’t Been.” Huffpost: World News, Change, and Women’s Labor Force Participation.” Journal of Popula- December 13, 2018. https://www.huffpost.com/entry/tunisia-violence- tion Economics 27: 639–81. against-women-law-58_n_5c126f7be4b002a46c14e61a. Gaddis, I., R. Lahoti, and H. Swaminathan. 2020. “Women’s Legal Rights Broecke, Stijn. 2013. “Tackling Graduate Unemployment in North Africa and Gender Gaps in Property Ownership in Developing Countries.” through Employment Subsidies: A Look at the SIVP Programme in Policy Research Working Paper 9444, World Bank, Washington, DC. Tunisia.” IZA Journal of Labor Policy 2 (1): 1–19. Gasparini, L., and M. Marchionni. 2015. “Bridging Gender Gaps? The Buvinic,´ Mayra, and Megan O’Donnell. 2016. “Revisiting What Works: Rise and Deceleration of Female Labor Force Participation in Latin Women, Economic Empowerment, and Smart Design.” November, America: An Overview.” CEDLAS Working Paper 185, Center for Dis- Center for Global Development, Washington, DC. http://www. tributive, Labor, and Social Studies, La Plata, Argentina. womeneconroadmap.org/sites/default/files/CGD-Roadmap-Update_ Goldin, C. 1990. Understanding the Gender Gap: An Economic History of v4.pdf. American Women. New York: Oxford University Press. Access to the Labor Market: A Spotlight on Women and Youth 103 Goldin C. 1995. “The U-shaped Female Labor Force Function in Economic Majlesi, K. 2016. “Labor Market Opportunities and Women’s Decision Development and Economic History.” In Investment in Women’s Making Power within Households.” Journal of Development Economics Human Capital, edited by T. P. Schultz. Chicago: University of Chi- 119: 34–47. cago Press. Mammen, K., and C. Paxson. 2000. “Women’s Work and Economic Devel- Halim, Daniel Zefanya, Hillary C. Johnson, and Elizaveta Perova. 2019. opment.” Journal of Economic Perspectives 14 (4): 141–64. “Preschool Availability and Female Labor Force Participation: Evi- Manacorda, Marco, Furio Camillo Rosati, Marco Ranzani, and Giuseppe dence from Indonesia.” Policy Research Working Paper 8915, World Pio Dachille. 2017. “Pathways from School to Work in the Developing Bank, Washington, DC. World.” IZA Journal of Labor and Development 6 (1). Hanmer, L., E. Tebaldi, and D. Verner. 2018. “Women in the Tunisian Labor Martin, J. P. 2000. “What Works among Active Labour Market Policies: Market.” In The Tunisian Labor Market in an Era of Transition, edited Evidence from OECD Countries’ Experiences.” OECD Economic by Ragui Assaad and Mongi Boughzala. Oxford, UK: Oxford Univer- Studies 30: 79–113. sity Press. Mateo Díaz, María Mercedes, and Lourdes Rodríguez-Chamussy. 2013. Heath, R., and A. M. Mobarak. 2015. “Manufacturing Growth and the Lives “Childcare and Women’s Labor Participation: Evidence for Latin of Bangladeshi Women.” Journal of Development Economics 115: 1–15. America and the Caribbean.” IDB Technical Note IDB-TN-586, Inter- HRW (Human Rights Watch). 2018. “Tunisia: Parliament Should Back American Development Bank, Washington, DC. Gender Equality in Inheritance.” News, December  4, 2018. https:// McKenzie, David J. 2017. “How Effective Are Active Labor Market Poli- www.hrw.org/news/2018/12/04/tunisia-parliament-should-back- cies in Developing Countries? A Critical Review of Recent Evidence.” gender-equality-inheritance#. World Bank Research Observer 32: 127–54. Ilkkaracan, I. 2012. “Why So Few Women in the Labor Market in Turkey?” Moghadam, V. M. 2017. “Dynamics of Gender Inequality and Women’s Feminist Economics 18: 1–37. Labour-Force Participation in Tunisia: Some Key Issues.” Female ILO (International Labour Organization). 2018a. “Care Work and Care Employment and Dynamics of Inequality Research Network, ESRC, Jobs for the Future of Decent Work.” ILO, Geneva. SOAS and GCRF. ILO (International Labour Organization). 2018b. “Women and Youth Moghadam, V. M. 2018. “Gender Inequality and Economic Inclusion in Empowerment in Rural Tunisia. An Assessment Using the Women’s Tunisia: Key Policy Issues.” Issue Brief, September 3, Baker Institute Empowerment in Agriculture Index (WEIA).” Taqeem Impact Report for Public Policy, Rice University. Series 11, ILO, Geneva. Mouelhi, Rim Ben Ayed, and Mohamed Goaied. 2018. “Women in the ILO (International Labour Organization). 2021. “World Employment and Tunisian Labor Market.” In The Tunisian Labor Market in an Era of Social Outlook 2021: The Role of Digital Labour Platforms in Trans- Transition, edited by Ragui Assaad and Mongi Boughzala. Oxford, forming the World of Work.” ILO, Geneva. UK: Oxford University Press. Jesse, A. 2017. “Staring Down Street Harassment: Women’s Perspectives Nayyar, Gaurav, Mary Hallward-Driemeier, and Elwyn Davies. 2021. At of Street Harassment in Tunisia.” Independent Study Project Collec- Your Service? The Promise of Services-Led Development. Washington, tion, 2668. DC: World Bank. Kärkkäinen, Outi, ed. 2011. “Women and Work in Tunisia: Tourism and OECD (Organisation for Economic Co-operation and Development). 2015. ICT Sectors; A Case Study.” European Training Foundation, Turin, “The ABC of Gender Equality in Education: Aptitude, Behaviour, Con- Italy. fidence.” PISA, OECD, Paris. Kasoolu, Semiray, Ricardo Hausmann, Tim O’Brien, and Miguel Angel OECD (Organisation for Economic Co-operation and Development). Santos. 2019. “Female Labor in Jordan: A Systematic Approach to 2017. Women’s Economic Empowerment in Selected MENA Coun- the Exclusion Puzzle.” CID Faculty Working Paper 365 (October), tries: The Impact of Legal Frameworks in Algeria, Egypt, Jordan, Center for International Development, Harvard University, Cam- Libya, Morocco and Tunisia. Competitiveness and Private Sector bridge, MA. Development Series. Paris: OECD. Klasen, Stephan. 2019a. “From ‘MeToo’ to Boko Haram: A Survey of Levels OECD (Organisation for Economic Co-operation and Development). and Trends of Gender Inequality in the World.” Discussion Paper 263 2018. PISA 2015: PISA Results in Focus. Paris: OECD. https://www. (July), Courant Research Center, Georg-August-Universität, Göttingen, oecd.org/pisa/pisa-2015-results-in-focus.pdf. Germany. OECD (Organisation for Economic Co-operation and Development). Klasen, Stephan. 2019b. “What Explains Uneven Female Labor Force Par- 2020. Changing Laws and Breaking Barriers for Women’s Eco- ticipation Levels and Trends in Developing Countries?” World Bank nomic Empowerment in Egypt, Jordan, Morocco and Tunisia. Paris : Research Observer 34 (2): 161–97. OECD. Klasen, Stephan, T. T. Ngoc Le, J. Pieters, and M. Santos Silva. 2021. “What ONEQ (Observatoire National de l’Emploi et des Qualifications, National Drives Female Labour Force Participation? Comparable Micro-Level Observatory for Employment and Skills). 2017. “Etude sur l’insertion Evidence from Eight Developing and Emerging Economies.” Journal of professionnelle des diplômés du dispositif national de la formation Development Studies 57 (3): 417–42. professionnel.” November, ONEQ, Tunis. Klasen, Stephan, and Janneke Pieters. 2015. “What Explains the Stagnation ONEQ (Observatoire National de l’Emploi et des Qualifications, National of Female Labor Force Participation in Urban India?” World Bank Observatory for Employment and Skills). 2019. “La formation profes- Economic Review 29 (3): 449–78. sionnelle en chiffres 2017.” April, ONEQ, Tunis. Krafft, Caroline, and Ragui Assaad. 2020. “Employment’s Role in Enabling Padilla-Romo, M., and F. Cabrera-Hernández. 2018. “The Effect of and Constraining Marriage in the Middle East and North Africa.” Children’s Time in School on Mothers’ Labor Supply: Evidence Demography 57: 2297–2325. from Mexico’s Full-Time Schools Program.” Working Paper 2018-04, Krafft, Caroline, Ragui Assaad, and Mohamed Ali Marouani. 2021. “The Haslam College of Business, Knoxville, TN. Impact of COVID-19 on Middle Eastern and North African Labor Patnaik, A. 2019. “Reserving Time for Daddy: The Consequences of Markets: Vulnerable Workers, Small Entrepreneurs, and Farmers Bear Fathers’ Quotas.” Journal of Labor Economics 37 (4): 1009–59. the Brunt of the Pandemic in Morocco and Tunisia.” ERF Policy Brief Pimkina, S., and L. de la Flor. 2020. “Promoting Female Labor Force Par- 55 (February), Economic Research Forum, Giza, Egypt. ticipation.” Review Note, World Bank, Washington, DC. Kugler, M., M. Viollaz, D. Duque, I. Gaddis, D. Newhouse, A. Palacios- Politakis, G. P. 2001. “Night Work of Women in Industry: Standards and Lopez, and M. Weber. 2021. “How Did the COVID-19 Crisis Affect Sensibility.” International Labour Review 140 (4): 403–28. Different Types of Workers in the Developing World?” World Bank, Sabarwal, S., N. Sinha, and Mayra Buvinic. ´ 2011. “How Do Women Washington, DC. Weather Economic Shocks? What We Know.” Economic Premise 46, Lim, J.Y. 2000. “The Effects of the East Asian Crisis on the Employment World Bank, Washington, DC. of Women and Men: The Philippine Case.” World Development 28 (7): Samman, E., E. Presler-Marshall, and N. Jones. 2016. “Women’s Work: 1285–1306. Mothers, Children, and the Global Childcare Crisis.” With T. Bhatkal, Lundberg, S. 1985. “The Added Worker Effect.” Journal of Labor Economics C. Melamed, M. Stavropoulou, and J. Wallace. Overseas Development 3 (1, Part 1): 11–37. Institute, London. 104 Tunisia’s Jobs Landscape Seguino, S. 2000. “Accounting for Gender in Asian Economic Growth.” World Bank. 2014a. “Gender at Work: A Companion to the World Devel- Feminist Economics 6 (3): 27–58. opment Report on Jobs.” Working paper, Report 89273 (February 20), Serrano, J., L. Gasparini, M. Marchionni, and P. Glüzmann. 2019. “Eco- World Bank, Washington, DC. nomic Cycle and Deceleration of Female Labor Force Participation in World Bank. 2014b. “Tunisia: Breaking the Barriers to Youth Inclusion.” Latin America.” Journal for Labour Market Research 53: 13. World Bank, Washington, DC. Sinha, S. 2011. “Women’s Rights: Tunisian Women in the Workplace.” Jour- World Bank. 2019. “A New Economy for the Middle East and North nal of International Women’s Studies 12 (3): 185–200. Africa.” MENA Economic Monitor, October, World Bank, Wash- Skoufias, E., and S. Parker. 2006. “Job Loss and Family Adjustments in ington, DC. Work and Schooling during the Mexican Peso Crisis.” Journal of Popu- World Bank. 2020a. “The Role of Technical and Vocational Education and lation Economics 19 (1): 163–81. Training.” Tunisia Skills Development for Employment, May, World Stampini, M., and A. Verdier-Chouchane. 2011. “Labor Market Dynamics Bank, Washington, DC. in Tunisia: The Issue of Youth Unemployment.” IZA Discussion Paper World Bank. 2020b. “Women’s Economic Participation in Iraq, Jordan, DP 5611, Institute of Labor Economics, Bonn, Germany. and Lebanon.” State of the Mashreq Women Flagship 1, Report Tanner, Victor. 2020. “Strengthening Women’s Control over Land: Inheri- AUS0001349, World Bank, Washington, DC. tance Reform in Tunisia.” DAI: Developing Alternatives, February 20, 2020. https://dai-global-developments.com/articles/strengthening- World Bank. 2021a. “Household Production and Gender Roles in the Time womens-control-over-land-inheritance-reform-in-tunisia. of COVID-19: Insight from a Rapid Online Survey in Tunisia.” Internal UN Women. 2017. “Tunisia Passes Historic Law to End Violence against Draft, World Bank, Washington, DC. Women and Girls.” News and Events, August  10, 2017. https:// World Bank. 2021b. “Secondary Cities and Migrants: The Tunisian Case.” www.unwomen.org/en/news/stories/2017/8/news-tunisia-law- Background report for the study “Migrants, Markets and Mayors – on-ending-violence-against-women. Rising Above the Employment Challenge in Africa’s Secondary Cities,” World Bank. 2009. “The Status and Progress of Women in the Middle East prepared for EFO 1505 for Cities Alliance Program, “Secondary Towns and North Africa.” World Bank, Washington, DC. and Migrants,” World Bank, Washington, DC. World Bank. 2011. World Development Report 2012: Gender Equality World Bank. 2021c. Women, Business, and the Law 2021. Washington, and Development. Washington, DC: World Bank. DC: World Bank. https://wbl.worldbank.org/en/wbl. Access to the Labor Market: A Spotlight on Women and Youth 105 ANNEX CHAPTER 2 TABLE A 2.1. Tunisia Snapshot, Women, Business and the Law 2021 Question Answer Legal Basis Mobility Can a woman choose where to live in the same way Yes No restrictions could be located as a man? Can a woman travel outside her home in the same Yes No restrictions could be located way as a man? Can a woman apply for passport in the same way Yes Loi No. 1975-40, Arts. 8 et 13; Passport applica- as a man? tion procedures Can a woman travel outside the country in the Yes No restrictions could be located same way as a man? Workplace Can a woman get a job in the same way as a man? Yes No restrictions could be located Does the law prohibit discrimination in employ- Yes Code du Travail, Art. 5 bis ment based on gender? Is there legislation on sexual harassment in Yes Loi organique No. 2017-58 du 11 août 2017, employment? relative à I’élimination de la violence à l’égard des femmes, Art. 15(Art. 226 ter) Are there criminal penalties or civil remedies for Yes Criminal: Loi organique No. 2017-58 du 11 août sexual harassment in employment? 2017, relative à I’élimination de la violence à l’égard des femmes, Art. 15 (Art. 226 ter) Civil: No applicable provisions cold be located Pay Does the law mandate equal remuneration for No No applicable provisions could be located work at equal value? Can a woman work at night in the same way as a man? No Code du Travail, Arts. 66 et 68-2 Can a woman work in a job deemed dangerous in Yes No restrictions could be located the same way as a man? Can a woman in an industrial job in the same way No Code du Travail, Art. 77, 375 as a man? Marriage Is there no legal provision that requires a married Yes No applicable provisions could be located woman to obey her husband? Can a woman be head of household n the same No Code du Statut Personnel, Art. 23 way as a man? Is there legislation specifically addressing domestic Yes Loi organque No. 2017-58 du 11 août 2017, violence? relative à I’élimination de la violence à I’égard des femmes Can a woman obtain a judgment of divorce in the Yes No restrictions could be located same way as a man? Does a woman have the same rights to remarry as No Code du Statut Personnel, Arts. 34 and 35 a man? Parenthood Is paid leave of at least 14 weeks available to mothers? No Code du Travail, Art. 64 Does the government pay 100% of maternity leave Yes Loi No. 196-3 du 14 décembre 196, Arts. 78, 82 benefits? et 88 Is paid leave available to fathers? Yes Code du Travail, Art. 122 Is there paid parental leave? No No applicable provisions could be located Is dismissal of pregnant workers prohibited? No No applicable provisions could be located (continued) 106 Tunisia’s Jobs Landscape TABLE A 2.1. Tunisia Snapshot, Women, Business and the Law 2021 (continued) Question Answer Legal Basis Entrepren. Does the law prohibit discrimination in access to No No applicable provisions could be located credit based on gender? Can a woman sign a contract in the same way as a Yes No restrictions could be located man? Can a woman register a business in the same way Yes No restrictions could be located as a man? Can a woman open a bank account in the same Yes No restrictions could be located way as man? Assets Do men and woman have equal ownership rights Yes Code du Statut Personnel, Arts. 23 et 24 to immovable property? Do sons and daughters have equal rights to inherit No Code du Statut Personnel, Arts. 92, 96, 98, 103 assets from their parents? et 104 Do female and male surviving spouses have equal No Code du Statut Personnel, Arts. 101 et 102 rights to inherit assets? Does the law grant spouses equal administrative Yes Code du Statut Personnel, Arts. 23 et 24 authority over assets during marriage? Does the law provide for the valuation of non­ No No applicable provisions could be located monetary contributions? Pension Is the age at which men and women can retire Yes Women: Décret No. 74-499 du 27 Avril 1974, Art. 15 with full pension benefits the same? Men: Décret No. 74-499 du 27 Avril 1974, Art. 15 Is the age at which men and women can retire with Yes Women: No applicable provisions could be located partial pension benefits the same? Man: No applicable provisions could be located Is the mandatory retirement age for men and Yes Women: No applicable provisions could be located women the same? Man: No applicable provisions could be located Are periods of absence due to child care Yes Décret No. 74-499 du 27 Avril 1974, Art. 2(c) accounted for in pension benefits? Source: World Bank (2021b). Access to the Labor Market: A Spotlight on Women and Youth 107 TABLE A 2.2. Childcare Centers and Preprimary Schools (Public and Private), by Governorate Kindergarten Preprimary school Nursery Kouttab Region Governorate (3–5 years old) (3–5 years old) (starting 3 months) (3–6 years old) Greater Tunis Tunis 485 422 69 186 Ariana 299 NA 29 58 Ben Arous 372 NA 35 113 Manouba 181 157 6 87 Total 1337 579 139 444 North-East Nabeul 409 NA 18 118 Zaghouan 101 NA NA 40 Bizerte 261 NA 15 48 Total 771 0 33 206 North-West Béja 101 NA 24 30 Jendouba 99 NA NA 20 El Kef 97 6 6 34 Seliana 81 16 NA 40 Total 378 22 30 124 Center-East Sousse 352 229 45 96 Monastir 381 163 25 106 Mahdia 147 26 3 86 Sfax 442 279 32 96 Total 1322 697 105 384 Center-West Kairouan 177 37 NA 68 Kasserine 139 NA 3 47 Sidi Bouzid 163 NA 5 36 Total 479 37 8 151 South-East Gabes 212 124 14 45 Mednine 293 169 13 119 Tataouine 22 13 NA 46 Total 527 306 27 210 South-West Gafsa 187 NA 8 44 Tozeur 110 NA 5 24 Kebili 144 NA 5 25 Total 441 0 18 93 Grand Total 5255 1641 360 1612 Source: Based on data from the Ministry of Ministry of Women, Family, Children, and Elderly and the Ministry of religious Affairs. 108 Tunisia’s Jobs Landscape TABLE A 2.3. Monthly and Registration Fees and Opening Days/Hours of Surveyed Private Day-Care Centers, by Governorate, April 2021 Region Location Monthly fee + registration fee Opening days/hours Greater Tunis TUNIS TD 230/month + TD 180 Monday to Friday 7:00AM–6:00PM & Saturday 7AM–1PM TUNIS TD 350/month + TD 350 Monday to Friday 7:30AM–6:30PM & Saturday 7AM–1PM TUNIS TD 210/month + TD 270 Monday to Friday 7:00AM–6:00PM & Saturday 7AM–1PM North-East BIZERTE TD 130/month +TD 100 Monday to Friday 6:30AM–4:30PM & Saturday 8AM–1PM NABEUL TD 180/month+TD 300 Monday to Friday 7AM–6PM ZAGHOUAN TD 320/month + TD 200 Monday to Friday 7AM–7PM North-West JENDOUBA TD 80/month + TD 150 Monday to Friday 7AM–5PM SILIANA TD 80/month + TD 80 Monday to Friday 8AM–5PM KEF TD 140/month + TD 80 Monday to Friday 7:30 AM–6:30PM BEJA TD 100/month + TD 50 Monday to Friday 7AM–6PM Center-East SOUSSE TD 170/month + TD 100 Monday to Friday 7AM–6PM MONASTIR TD 200/month + TD 120 Monday to Friday 7AM–6PM & Saturday 7AM–1PM MAHDIA TD 90/month + TD 40 Monday to Friday 8AM–5PM & Saturday 8AM–1PM SFAX TD 200/month + TD 450 Monday to Friday 7AM–6PM & Saturday 7AM–2PM Center-West KAIROUAN TD 40/month Monday to Friday 7AM–5PM & Saturday 7AM–1PM KASSERINE TD 80/month + TD 60 Monday to Friday 8AM–6PM SIDI BOUZID TD 120/month + TD 100 Monday to Friday 7AM–5PM & Saturday 7AM–1PM South-East MEDININE TD 70/month + TD 100 Monday to Friday 8AM–5PM GABES TD 110/month + TD 210 Monday to Friday 7AM–6PM & Saturday 7AM–1PM TATAOUINE NA NA South-West KEBILI NA NA GAFSA TD 40/month + TD 30 Monday to Friday 7AM–5PM TOZEUR TD 90/month Monday to Friday 7AM–5:30PM Source: Based on data collected through phone interviews to one random day-care centers in each governorate. Access to the Labor Market: A Spotlight on Women and Youth 109 TABLE A 2.4. Main Active Labor Market Policies for Youth Program name Responsible agency Program description Target population Contrat d’initiation à la vie Agence Nationale pour The contract for integration into working life aims to First-time job-seekers with professionnelle (contract l’Emploi et le Travail meet the needs of private sector companies and to a university degree or for integration into Indépendant (ANETI) - help job-seekers acquire professional skills in order BTS (Brevet technicien working life. (Named until National Agency for to facilitate their integration into working life. supérieur) 2019 Stage d’initiation à Employment and The duration of the contract is 12 months. However, OR la vie professionnelle) Self-Employment ANETI may extend the duration of the contract for an additional maximum period of 12 months. Job-seekers with disabilities with a university degree ANETI pays: or BTS (Brevet technicien • a monthly allowance of TD 200 for holders of a higher supérieur) education diploma or a BTS, OR • a monthly allowance of TD 150 for other diplomas, • an additional monthly grant of TD 50 in case of Young people with a disabled persons, minimum of 7 years of • social coverage of trainees, basic education (second • employer’s contribution to social security. year of secondary school) In addition, the host company is exempted from the payment of social security contributions and must grant the beneficiary a monthly allowance of: • TD 200 for the holder of a higher education diploma or a BTS • TD 150 for the other levels. The company can only take on new trainees under this contract if it has previously recruited at least 50% of all the trainees who have completed their work experience during the last three years preceding the year in which the new application is submitted. If the company does not achieve the above-mentioned rate, it can only take on new trainees after at least one year has elapsed since the end of the last contract. Contrat d’insertion The purpose of the Contrat d’Insertion des Diplômés Job-seekers with a des diplômés de de l’Enseignement Supérieur (Contract for the university degree who l’enseignement supérieur Integration of Higher Graduates) is to allow the have been unemployed (CIDES) - Insertion con- beneficiary to acquire professional qualifications by for over three years from tract for higher education alternating between a private company and a public the date of graduation. graduates or private training structure, in accordance with the requirements of a job position. ANETI covers the cost of: • the training of trainees up to a of 400 hours • a monthly allowance of TD 150 served to the trainee + an additional allowance of TD 50 to the trainee who resides outside the governorate of the company • social security coverage of the trainee • a recruitment bonus fixed at TD 1,000 to the company after one year of work • employer’s contribution to social security in case of recruitment of the trainee for the first 7 years (at a declining rate) The company commits to • pay the trainee an additional monthly allowance of TD 150 • hire the trainee who has completed the internship contract (continued) 110 Tunisia’s Jobs Landscape TABLE A 2.4. Main Active Labor Market Policies for Youth (continued) Program name Responsible agency Program description Target population Contrat d’adaptation et ANETI The objective is to allow the beneficiary to acquire pro- Job-seekers without a d’insertion profession- fessional qualifications in line with the requirements university degree nelle (CAIP) of a job offer presented by a private company and which has not been fulfilled due to the unavailability of manpower. ANETI covers: • a monthly allowance of TD 100 to the trainee, • social coverage of the trainee, • the cost of the training within a limit of 400 hours. The company: • pays the trainee a monthly allowance of TD 50, • recruit the beneficiary who has completed the training contract. Service civil volontaire ANETI The Civil Service Contract aims to enable graduates First-time job-seekers (SCV) of higher education, who are first-time job-seekers, who hold a University to carry out an activity within associations and pro- degree or an equivalent fessional organizations to develop their skills and diploma and who have competences and to acquire practical abilities. been unemployed over The duration of the contract is 12 months. The asso- 12 months from the date ciation can exceptionally extend the duration of of the diploma. the contract for an additional maximum period of 12 months. ANETI covers: • a monthly allowance of TD 200, • an additional monthly grant of TD 50 in case of disabled persons, • social coverage of the trainees, • cost of the training within the limit of 400 hours, • employer’s contribution to social security. Programme ANETI and Banque This program allows the promoter to identify a project Any entrepreneur d’accompagnement des tunisienne de solidarité idea, to develop the project study and related busi- promoteurs des petites ness plan and to ensure the necessary support to entreprises (PAPPE) the entrepreneur to succeed in the project. ANETI covers: • a scholarship up to TD 200 for internship in a company (for a period of 3 months, renewable once), • the cost of adaptation sessions of up to 200 hours, • the cost of management sessions up to a maximum of 120 hours, • the costs of technical adaptation sessions up to a maximum of 400 hours, • the costs of technical assistance up to a maximum of 12 days. (continued) Access to the Labor Market: A Spotlight on Women and Youth 111 TABLE A 2.4. Main Active Labor Market Policies for Youth (continued) Program name Responsible agency Program description Target population Le programme ANETI The KARAMA program aims to encourage private First-time job-seekers « contrat-dignité », sector companies to recruit first-time job-seekers with a university degree KARAMA with higher education degrees and to improve the or technician diploma supervision. (BTS – Brevet technicien The job-seeker benefit from a minimum monthly supérieur) salary of TD 600 paid by the company for maximum 24 months. OR The company benefits from: Job-seekers with disabilities with a university degree • financial support by the national fund of employment, or BTS (Brevet technicien for two years from the date of recruitment, in the supérieur) amount of 50% of the net salary and within the limit of TD 400 per month. • financial support by the national fund of employment, for two years from the date of recruitment, of the employer’s share of contributions to social security. Programme d’action ANETI This program consists in organizing training sessions Job-seekers with a minimum d’adaptation pour to the benefit of job-seekers in order to improve of 7 years of basic educa- l’amélioration de their employability and facilitate their integration in tion (second year of l’employabilité - companies where work requires additional training secondary school) Employability Adjustment or adaptation. Action Program • To the beneficiary: ANETI pays during the training period a monthly allowance of: • TD 200 for graduates of higher education or equivalent diplomas or holders of a technician diploma (BTS), • TD 150 for all other levels, • TD 50 in case of disabled persons, • social security contributions. • To the company: ANETI pays for: • the cost of training or additional adaptation for a maxi- mum of 6 months and up to 600 hours per beneficiary, • the cost of soft-skills training up to 60 hours per beneficiary, • the cost of language certification for each beneficiary up to TD 400, • the cost of certification in the field of ICT or other technical specializations capped at TD 1,000. Note: A first-time job-seeker is defined as (a) an individual who has not been in employment for a continuous period of over 24 months after obtaining the last diploma; (b) an individual who has not been in employment for over 36 months in a discontinuous way after obtaining the last diploma. 112 Tunisia’s Jobs Landscape FIGURE A 2.1. Detailed Oaxaca-Blinder Decomposition of the Gender Gap in Labor Force Participation, by Year, 2006–17 a. Explained component 20.0 0.0 Percentage points –20.0 –40.0 –60.0 –80.0 2006 2008 2009 2011 2013 2015 2016 2017 Demographics Marital status Education Household composition Location b. Unexplained component 20.0 0.0 Percentage points –20.0 –40.0 –60.0 –80.0 2006 2008 2009 2011 2013 2015 2016 2017 Demographics Marital status Education Household composition Location Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE A 2.2. Sectoral Distribution of Unpaid Family Workers, by Sex, 2006–17 a. Men 2006 66.2 0.7 0.9 2.92.2 0.0 13.3 0.9 1.92.6 0.9 0.3 8.6 2008 71.8 0.2 0.72.91.9 0.0 14.9 1.6 2.0 0.0 2.21.1 0.4 0 2011 67.0 0.6 1.3 3.61.2 0.0 18.9 1.8 1.9 2.7 0.0 0.2 2013 62.9 1.2 0.1 0.6 5.1 2.2 20.2 2.91.42.6 0.0 0.1 2015 66.2 0.83.8 2.1 1.2 0.0 19.0 1.4 2.20.0 2.9 2016 55.6 0.84.5 3.7 2.0 0.1 20.8 2.4 3.6 0.14.30.9 1.2 2017 59.4 1.03.31.8 0.9 0.0 24.4 7.9 0. 0 10 20 30 40 50 60 70 80 90 100 Percent Agriculture Agro-food industry Textile Manufacturing Other Manufacturing Construction Other Secondary Trade Transports Hotels and Restaurants Financial Real Estate/Professional PA/Health/Education Other Services Not De ned Services Activities b. Women 2006 73.7 2.4 0.3 0.0 0.4 0.6 0.4 0.0 0.25.2 0.1 16.3 2008 88.5 0.4 1.4 0.30.0 6.1 0.0 0.90.0 0.4 2011 85.3 2.1 0.2 10.0 0.0 0.4 .4 7.4 0.60.9 2013 81.9 1.0 0.8 1.70.1 9.6 0.0 0.2 0.7 1.60.0 2015 79.3 0.0 0.6 1.1 0.9 11.4 0.2 0.0 3.5 0.6 2016 73.2 2.40.0 1.3 0.7 0.4 14.7 2.6 0.9 0.0 0. 1.21.7 2017 73.6 1.3 2.52.3 0.0 0.1 14.3 7.9 0. 0 10 20 30 40 50 60 70 80 90 100 Percent Agriculture Agro-food industry Textile Manufacturing Other Manufacturing Construction Other Secondary Trade Transports Hotels and Restaurants Financial Real Estate/Professional PA/Health/Education Other Services Not De ned Services Activities Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE A 2.3. Educational Level Distribution of Employers and Own-Account Workers, by Sex, 2006–17 a. Men 2006 19.6 43.3 31.3 5.7 0. 2008 16.8 45.0 32.0 6.1 0. 2011 15.9 44.1 31.8 8.2 0. 2013 11.4 45.9 35.0 7.6 0. 2015 12.8 44.7 34.1 8.2 0. 2016 11.8 46.7 33.8 7.5 0. 2017 12.8 46.4 32.8 7.9 0. 0 10 20 30 40 50 60 70 80 90 100 Percent None Primary Secondary Tertiary Not stated b. Women 2006 41.9 29.2 23.0 5.7 0. 2008 36.7 28.6 26.5 8.1 0. 2011 27.5 30.2 31.4 10.7 0. 2013 20.9 29.5 36.8 12.6 0. 2015 19.7 29.7 33.7 16.5 0. 2016 20.8 31.6 29.6 17.6 0. 2017 20.1 30.5 30.4 19.0 0. 0 10 20 30 40 50 60 70 80 90 100 Percent None Primary Secondary Tertiary Not stated Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE A 2.4. Educational Level Distribution of Unpaid Family Workers, by Sex, 2006–17 a. Men 2006 6.8 49.5 41.7 1.7 2008 5.0 46.4 45.5 2.9 2011 4.0 39.5 52.8 3.8 2013 2.2 41.7 53.0 2.9 2015 3.7 41.6 51.1 3.5 2016 3.1 35.9 55.9 4.9 2017 3.8 35.6 57.2 3.4 0 10 20 30 40 50 60 70 80 90 100 Percent None Primary Secondary Tertiary Not stated b. Women 2006 58.3 30.9 9.8 0.9 2008 52.1 34.3 12.3 1.2 2011 56.2 27.2 13.2 3.4 2013 48.1 32.8 16.2 2.8 2015 44.0 35.0 17.9 3.0 2016 35.8 38.9 20.7 4.4 2017 40.1 35.5 19.1 5.4 0 10 20 30 40 50 60 70 80 90 100 Percent None Primary Secondary Tertiary Not stated Source: Based on data from the Labor Force Survey (ENPE), INS. FIGURE A 2.5. Unemployment Rates Among Youth, by Year, Age-Group, Educational Level, and Region, 2006–17 a. By age-group and educational level 100 90 80 71.8 70 67.5 63.9 60.8 60 58.9 52.9 53.8 51.7 51.7 52.6 Percent 50.7 51.0 50 49.0 41.8 42.2 39.8 38.4 40 36.6 34.4 35.7 31.4 32.7 34.3 33.6 32.4 29.3 29.5 30 27.7 27.6 26.7 27.1 26.1 27.8 26.9 25.3 24.6 26.1 25.4 25.3 24.3 23.3 20.5 19.4 20.1 20.5 21.5 19.2 19.9 19.0 20 17.8 17.1 18.4 18.2 17.3 15.3 16.4 15.0 16.9 16.1 16.2 12.6 13.4 13.8 10.6 10 0 2006 2008 2009 2011 2013 2015 2016 2017 2006 2008 2009 2011 2013 2015 2016 2017 Youth 15–24 Youth 25–29 None Primary Secondary Tertiary b. By age-group and region 100 90 80 70 60 Percent 50 40 30 20 10 0 2006 2008 2009 2011 2013 2015 2016 2017 2006 2008 2009 2011 2013 2015 2016 2017 Youth 15–24 Youth 25–29 Greater Tunis North-East North-West Center-East Center-West South-East South-West Source: Based on data from the Labor Force Survey (ENPE), INS. 117 CHAPTER 3 Employment and Wage Outcomes HIGHLIGHTS ◾ An important divide exists between public sector workers and formal and informal workers in the private sector. ◾ Over 1.5 million workers are employed informally, and the informality rate is estimated at about 44 percent. ◾ The profiles of public, formal, and informal workers point to significant selection into the three seg- ments; vulnerable workers are more exposed to informality. ◾ In addition to a number of benefits, the public sector pays, on average, twice the hourly wage paid by the private sector, and a sizable part of the gap derives from differences in observable characteristics. ◾ Informal workers are paid on average about 16 percent less per hour worked than their formal counterparts, and the gap is largely ascribable to differences in observable characteristics. ◾ More than 50 percent of university graduates not employed in public administration are either unemployed or inactive. Most of the inactive are young married women in affluent households, whereas the majority of the unemployed are young men living with their parents. ◾ The nonmonetary benefits and job security provided by public sector jobs may contribute to the high rates of nonemployment (unemployment and inactivity) observed among university graduates. ◾ Assigned gender roles are strengthened by a sizable gender wage gap in the private sector. Women are paid, on average, $0.82 for every $1.00 paid to men per hour worked, and most of the wage gap arises because of a different wage structure or to unobserved characteristics that would, on average, make men more productive than women. ◾ In the public sector, by contrast, women make, on average, about one-third more than men per hour worked, and a large part of the wage premium is ascribable to more productive characteristics of women. ◾ Returns to education are sizable: tertiary education yields a premium of about 26 percent per hour worked relative to secondary education among wage workers. ◾ Returns to tertiary education in the private sector have started to decline because the demand for well-educated workers has been sluggish. By contrast, returns to tertiary education are considerably higher and on the rise in the public sector. 118 Tunisia’s Jobs Landscape C hapter  2 offers an overview of trends in demo- of different types of employment, (2) provides an over- graphics and shows that Tunisia’s demographic view of recent trends in wages and of conditional wage window is narrow, but still open, and that the gaps along a number of dimensions (men/women, public/ quality of learning still lags comparator countries despite private, formal/informal employment), and (3) illustrates significant progress in school enrollments. The chapter how wage workers with different characteristics, in par- describes recent trends in labor market indicators and ticular different educational endowments, benefit from the highlights the underutilization of human capital; only labor market. 50 percent of Tunisians of working age are participating in the labor market. It focuses on two groups that face par- ticular difficulties in accessing the labor market, namely, Public Sector, Formal, and women and youth. It documents the modest improvements Informal Employment over the past decade, which are mainly ascribable to young cohorts of women and progress in educational attainment About 21  percent of the employed population work in among the working-age population. Weak labor demand, public administration or in a public company. In 2019, traditional behaviors, and the limited availability of child- about 750,000 workers were employed in public admin- care are among the barriers to the greater engagement of istration or in state-owned enterprises (SOEs) according women in the labor market. About 4 youth ages 15–29 in to labor force survey data (Figure  3.1). In 2019, over 10 are NEET, and 1 youth in 3 is unemployed. While inac- 650,000 people were employed in public administra- tivity rates are higher among young men and women with tion, and about 95,000 in SOEs. This represented about little education and seems to be ascribable to exclusion, 21 percent of total employment. unemployment rates are high among university graduates and seem to be largely ascribable to sluggish job creation. Public sector hiring increased in the aftermath of the 2011 revolution and, together with wage increases, absorbs a large This chapter shifts the focus to one of the most relevant share of tax revenues. To address the challenge of insecu- dimensions that characterizes the Tunisian labor market, rity and social demands that followed the 2011 revolution, namely, the distinction among public sector, formal, and public sector hiring rose considerably with the 2012 law informal employment. The chapter (1) investigates how promoting access to public administration among people individual characteristics are correlated with the probability injured during the revolution and covered by the amnesty FIGURE 3.1. The Composition of Employment, 2019 Public sector Wage workers employment 1.12 million (including public companies) Employers 0.75 million (13%) 0.068 million Own-account workers 0.043 million Total Formal employment employment Wage workers 1.23 million 3.53 million 0.76 million Employers 0.174 million Own-account Informal employment workers 1.55 million 0.52 million Contributing family workers 0.093 million Source: Based on data from the Labor Force Survey (ENPE), INS. Employment and Wage Outcomes 119 of 2011 (Brockmeyer, Khatrouch, and Raballand 2015; INS the entire period and an average of 1.9 percent per year) 2017; OECD 2018) (Box 3.1). In 2012, the number of civil (INS 2017, 2019). The expansion in the number of civil servants increased by over 88,000 (almost 20 percent) com- servants, together with wage increases, led to growth in the pared with 2011, and, between 2011 and 2017, the number wage bill from about 11.9 percent of gross domestic product of civil servants rose by almost 200,000 (45 percent over (GDP) in 2011 to 14.6 percent in 2019 and an estimated BOX 3.1. Civil Service: Hiring and Compensation Mechanisms The public sector in Tunisia comprises central and regional administrations, local authorities, and state-owned enterprises (SOEs). Public sector employees are governed by one of seven employment regimes: the general regime for public employees in central and local state institutions of administrative nature (le Statut Général de la Fonction Publique) and six distinct regimes for the judiciary, members of the administrative court, members of the court of auditors, internal security forces, the military, and customs agents (Brockmeyer, Khatrouch, and Raballand 2015). In addition, a separate regime governs employment in SOEs. The general civil service includes public employees in central and local state institutions of an administrative nature. It covers most civil service employees. Civil servants are divided into officials (fonctionnaires), workers (ouvriers), and temporary staff (personnel temporaire). Hiring Procedures The seven regimes are divided into approximately 130 professional groups or corps (for example, teachers, financial inspectors, engineers, and so on). Within each group, employees are classified according to level of education and are recruited under one of four categories, A, B, C, or D. Category A is the highest and includes three subcategories (A1, A2, A3). Category A requires at least some university (category A1, a master’s degree; category A2, a bachelor’s degree; and category A3, two years of uni- versity). Category B requires a high school diploma. Category C requires some high school (four years after elementary school). Category D requires elementary school. Workers (ouvriers) are classified into units ranging from 1 to 3 according to their level of education. The system is career based. An employee is recruited at a specific grade that corresponds to the worker’s education. The worker receives seniority-based pay increases, and is tenured within a short time (Boutar 2018). Officials are recruited through competitive recruitment procedures (concours) that are based on tests or on background (application- based selection).a The details of each procedure are determined by specific statute and eligibility criteria and typically include the exact degree required rather than the minimum education degree. About half the positions in categories A, B, and C are filled through external recruitment, and the other half is filled through internal promotions. The number of positions is nego- tiated between each ministry and the Ministry of Finance based on the annually allocated budget.b Workers (ouvriers) are recruited using a simplified version of the recruitment system for officials. Workers are hired to permanent positions through tests or professional examinations. Temporary workers are recruited by direct appointment on a revocable basis for a deter- mined period either to fill a vacancy caused by a lack of permanent staff or to replace a staff member. Contract workers are recruited by contract for specific projects for a limited period. The 2011 general amnesty and several exceptional provisions in 2012 were approved to allow for direct recruitment, in addi- tion to the regularization of contract and temporary workers in 2012 and 2013 (Brockmeyer, Khatrouch, and Raballand 2015). The provisions led to a sizable increase in recruitment, which more than doubled between 2010 and 2011 and remained at a high level until 2013 (Figure B 3.1.1). FIGURE B 3.1.1. Trends in the Number of Civil Servants, by Category, 2011–17 700.0 23.4 19.2 20.6 22.7 21.3 Number of civil servants 600.0 4.3 123.2 105.9 107.1 113.9 500.0 101.6 21.0 92.0 (Thousands) 400.0 71.7 300.0 457.8 464.6 474.2 485.9 496.3 200.0 436.8 352.2 100.0 – 2011 2012 2013 2014 2015 2016 2017 Of cials Workers Others Source: Based on data from INS 2017, 2019. (continued) 120 Tunisia’s Jobs Landscape BOX 3.1. Civil Service: Hiring and Compensation Mechanisms (continued) Promotions There are three ways civil servants may obtain a promotion (Boutar 2018). The first consists of the successful completion of a continuous training cycle. The second is an internal competitive recruitment procedure (concours interne); only employees with five years of seniority in their current grade are eligible for this type of promotion. The third relies a point-based system and is used for 10 percent of civil servants with a minimum of 10 years of seniority in their current grade; it is a last resort option for promotion and is offered once in an employee’s career. Evaluation Civil servants are evaluated using a double rating system that comprises an annual professional rating and a quarterly perfor- mance rating linked to a bonus paid, in addition to the salary (Brockmeyer, Khatrouch, and Raballand 2015). The professional rating is the sum of 5 grades out of 20 for the following criteria: work quality, work quantity, interpersonal relationships and conduct, attendance, and perseverance. The employee’s immediate supervisor is responsible for making the assessment. The professional rating is typically not transparent or objective because no goals or objectives exist on which the rating might be based, and no benchmarks for evaluating the quantity and quality of work are set. In addition, the majority of civil servants receive ratings between 95 and 100, thus making the distinction between high and low performers impossible. Similarly, under the performance bonus rating, most employees receive the highest possible grade unless they have been absent, arrive late to work, or take sick leave. The performance bonus is considered an additional fixed compensation because supervisors do not want to risk queries, grievances, and internal conflicts by differentiating grades. The compensation policy has therefore been revised to include a two-thirds share of the performance bonus as a fixed part of the monthly salaries of a number of professional groups. Supervisors also have the possibility to sanction nonperforming employees. First-degree sanctions include blaming and rep- rimanding. Second-degree sanctions, which must be handed down by a disciplinary committee, include a one-year delay in advancement, temporary suspension, transfer with change of residence, or dismissal. Brockmeyer, Khatrouch, and Raballand (2015) show that, besides warning and rebuke, second-degree sanctions are rare. Compensation and Benefits The compensation of civil servants includes a base salary and a number of allowances, comprising common allowances, corps- specific allowances, and special allowances, in addition to a quarterly performance bonus. Employees with managerial posi- tions receive managerial allowances. The performance bonus is a small fraction of the total compensation and is affected only by attendance and not by the quality of work. Each corps has a salary grid, consisting of two types of base salary increases: a seniority-based increase and a promotion-based increase. The seniority-based increase is automatic, and the increase rate and frequency are determined by statute. In general, there is an increase every year for the first four years and every two years thereafter. Specific allowances by professional group are large and make staff reassignment across professional groups or ministries complicated because this might imply a sizable monetary loss for the employee. A promotion comes with a promotion-based increase, but resets the seniority-based level to the first level. Civil servants are entitled to a wide range of annual leave options, including administrative leave, leave for health reasons, training leave, unpaid leave, and leave to create a business. In addition, women benefit from two-month maternity leave at full pay and can be granted a postnatal leave at half-pay for up to four months. Women with a dependent children ages under 16 can request a special part-time work regime at two-thirds of full-time wage. The duration of the regime is set at three years as long as the conditions are met and may be renewable twice during the administrative career of the agent and under the same conditions. Civil servants benefiting from the special part-time work regime retain full rights for advancement, promotion, leave, and social security. Civil servants also benefit from flexible working hours, up to a half-hour before or after the scheduled entry time, and employees with one or more dependent children ages under 16 benefit from the flexibility of up to an hour and a half, subject to compensation on the same day. Data from the 2019 labor force survey confirm that public sector workers enjoy nonmonetary benefits that are not available to most workers in the private sector. Almost 88 percent of public sector workers had open-ended contracts, compared with 61 percent of formal wage workers and 22 percent of informal workers. Almost 1 in 2 (45.6 percent) of the latter did not have any contracts. Similarly, virtually all workers in the public sector had regular permanent jobs in 2019 compared with about 87.0 percent of formal wage workers and 55.0 percent of informal wage workers. Among the latter, about 4 workers in 10 had temporary or casual jobs, and about 5 percent had seasonal jobs. (continued) Employment and Wage Outcomes 121 BOX 3.1. Civil Service: Hiring and Compensation Mechanisms (continued) Compared with the private sector, the civil service is a more favorable employer for women. In 2017, the share of women in the private sector was about 24 percent compared with 36 percent in the civil service. Although the overall share of women in the civil service declined by about 3.5 percentage points between 2011 and 2017, their share in the top three categories (A1, A2, and A3) remained constant or increased (at about 50 percent) in 2017. Within the group of officials, women were largely employed in the top two categories (A1 and A2) in 2017; almost 75 percent of women were in category A1 and A2, compared with about 53 percent of men (Figure B 3.1.2). FIGURE B 3.1.2. Distribution of Civil Servants, by Sex and Category, 2017 Men 25.6 27.7 9.6 15.0 13.8 8.4 Women 33.8 40.5 13.6 6.6 4.9 0.6 – 20.0 40.0 60.0 80.0 100.0 Percent Category A1 Category A2 Category A3 Category B Category C Category D Source: Based on data of INS 2019. Source: The box draws on Boutar 2018; Brockmeyer, Khatrouch, and Raballand 2015; UN Women 2017. a. The selection procedure applies to positions in categories A, B, and C, whereas recruitment for category D positions takes place only externally. b. Direct recruitment is allowed only among students from approved schools, such as l’Ecole National d’Administration. 17.6 percent in 2020. According to the International Mon- (11 percent), and unpaid family workers (6 percent) (see etary Fund (IMF 2021), most of the enlargement in the Figure 3.1). wage bill is ascribable to salary increases, including wage boosts in 2016–18, an additional wage increase agreed in Vulnerable groups are more highly exposed to informality 2019 and delivered in three tranches in 2019–20 for a total and less likely to be employed in the public sector or formal of 1.5 percent of GDP, and an additional jump equivalent private sector jobs. To clarify whether formal and informal to 0.3 percent of GDP agreed in 2020. This bloated wage workers, both wage workers and nonwage workers, dif- bill crowds out other public expenditures. In 2020, it con- fer in observable characteristics, a profile of the employed sumed about 75 percent of tax revenues, and it was almost population along demographic, household-level, geo- three times the size of public investment and almost six graphical, and job-related characteristics is provided as of times the amount of public spending on social programs 2019 (Table 3.1). Separate profiles are presented for the (IMF 2021). public sector, formal and informal wage workers, and formal and informal nonwage workers, including employers, own- About 44 percent of the employed are in informal jobs. account workers, and unpaid family workers. In 2019, almost 2.8 million workers were employed in the private sector, and, among these, about 1.6 million • Sex. Relative to the overall distribution of the employed were informal workers (see Figure  3.1). In 2019, the population by sex (26.4 percent women), a larger share informality rate was estimated at 43.9 percent overall and of women are employed as public sector (32.4 percent) 55.7 percent among private sector workers. At 87.7 per- and private sector formal (33.6 percent) workers. This cent, the rate is considerably higher among nonwage compares with 23.0  percent among informal wage workers relative to wage workers (29.0 percent). In 2019, workers and even lower shares among formal and infor- 1 informal worker in 2 was a wage worker (49 percent), mal nonwage workers (17.7 percent and 15.7 percent, followed by own-account workers (33 percent), employers respectively). Among nonwage workers, the gender gap 122 Tunisia’s Jobs Landscape TABLE 3.1. Distribution of Public Sector, Formal and Informal Workers by Individual and Household Characteristics, 2019 Public wage Formal wage Informal wage Formal nonwage Informal nonwage Indicator workers workers workers workers workers Sex Women 32.4 33.6 23.0 17.7 15.7 Men 67.6 66.4 77.0 82.3 84.3 100.0 100.0 100.0 100.0 100.0 Age-group 15–24 3.1 10.0 21.1 2.4 8.3 25–34 21.0 28.8 29.2 17.1 17.7 35–44 34.1 31.3 24.6 33.5 25.9 45–54 29.0 20.1 15.2 24.3 23.6 55–64 12.1 8.7 8.5 16.8 17.7 65+ 0.8 1.1 1.5 5.9 6.8 100.0 100.0 100.0 100.0 100.0 Relation to head Head 56.3 51.7 43.8 72.0 65.7 Spouse 24.0 17.8 8.1 11.9 9.4 Children 18.1 26.9 44.2 14.8 23.3 Grandchildren 0.1 0.3 0.5 0.1 0.2 Daughter-/son-in-law 0.3 0.5 0.3 0.2 0.3 Parents/parents-in-law 0.1 0.0 0.0 0.0 0.1 Other relatives 0.8 1.6 2.0 0.7 1.0 Other nonrelatives 0.2 1.2 1.1 0.2 0.2 100.0 100.0 100.0 100.0 100.0 Marital status Single 20.6 31.3 48.7 18.1 25.3 Married 76.5 66.5 48.8 79.2 71.5 Widowed 1.2 0.9 1.2 1.4 1.8 Divorced 1.8 1.4 1.3 1.3 1.4 100.0 100.0 100.0 100.0 100.0 Educational level No education 4.6 5.6 11.5 3.4 14.3 Primary 15.6 34.3 43.0 29.1 45.5 Secondary 34.4 38.5 38.5 36.8 33.7 Tertiary 45.4 21.6 7.0 30.7 6.5 Not stated 0.1 0.1 0.0 0.1 0.1 100.0 100.0 100.0 100.0 100.0 (continued) Employment and Wage Outcomes 123 TABLE 3.1. Distribution of Public Sector, Formal and Informal Workers by Individual and Household Characteristics, 2019 (continued) Public wage Formal wage Informal wage Formal nonwage Informal nonwage Indicator workers workers workers workers workers Decile of household consumption per capita, 2015 Lowest decile 3.1 4.3 12.8 2.9 10.1 2 4.7 7.3 12.0 5.8 11.1 3 5.6 8.4 13.1 7.3 11.1 4 7.1 9.1 11.4 7.7 11.9 5 8.4 9.8 11.6 9.5 10.0 6 9.5 10.7 10.4 10.3 10.1 7 11.3 11.7 8.5 11.2 11.0 8 12.8 12.1 8.7 12.9 9.5 9 16.6 12.9 6.9 14.5 8.7 Highest decile 21.0 13.7 4.8 17.8 6.6 100.0 100.0 100.0 100.0 100.0 Source: Based on data from the Labor Force Survey (ENPE) 2019 and Household Budget Survey (HBS) 2015, INS. Note: Statistics are based on data from the second quarter of the Labor Force Survey (ENPE) 2019. Statistics by decile of per capita household consumption are based on the Household Budget Survey (HBS) 2015 and refer to the employed population ages 18 and above because of an age-based skip pattern in the question about affiliation with social security. is considerably smaller in formality status and more which include employers, own-account workers, and driven by the smaller share of women in charge of their unpaid family workers. Both youth and older workers own business rather than working for a wage. are overrepresented in employment types at higher risk • Age. Youth ages 15–24 are more likely to work infor- of informality; thus, greater shares of unpaid family mally for a wage relative to prime-age workers, but less workers are young, while many employers and own- likely to be formal or informal nonwage workers. The account workers are ages 65 or more. share of youth among wage workers in the public sector • Education. Educational level is key to accessing public is smaller than other age-groups, except for workers sector jobs. Almost 1 worker in 2 (45.4 percent) employed ages 65 or more, who, by that age, are retired from in the public sector has tertiary education, compared the civil service. Older workers ages 65 or more con- with 21.6 percent of formal wage workers and 7.0 per- tribute to nonwage employment, particularly informal cent of informal wage workers. Workers with tertiary employment, more than to wage employment. About education also contribute a large share of formal non- 6.0  percent and 6.8  percent of formal and informal wage employment (30.7 percent), but only 6.5 percent nonwage workers, respectively, are ages 65 or more of informal nonwage employment. Over 1 worker in relative to about 1 percent of wage workers (0.8 per- 2 in informal employment has no schooling or only cent of public sector wage workers and 1.1  percent primary education: 54.6 percent in the case of wage and 1.5  percent among private sector formal and employment and 59.8 percent in the case of nonwage informal wage workers, respectively). Workers ages employment. The share of workers with secondary edu- 25–64 account for the overwhelming majority of public cation does not differ substantially across groups. sector employment (96.1 percent), formal wage employ- • Marital status. The differences in the composition of ment (88.9 percent), and informal wage employment employment across groups by marital status is largely (about 77.4  percent). The shares reach 91.6  percent attributable to the age distribution of employment. and 84.9 percent among formal and informal nonwage The share of single workers is larger in informal wage employment, respectively. This life-cycle pattern is employment (48.7 percent) than in formal (31.3 per- partly ascribable to the composition of employment by cent) and public sector (20.6  percent) wage employ- type over the life cycle and to variation in the incidence ment or in formal nonwage employment (18,1 percent), of informality across employment types. Informality reflecting the younger ages of workers in these employ- rates in Tunisia are higher among nonwage workers, ment types. By contrast, the share of single workers in 124 Tunisia’s Jobs Landscape formal nonwage employment is the lowest, at 18.1 per- poverty rate of 15.1 percent at the national level, the cent, mirroring the small share of youth ages 15–24 in poverty rate among fully formal households was consid- that group. erably lower, at 9.5 percent, and the rate among mixed • Relation to the head of household. The composition of households was estimated at 12.4 percent. Completely employment by relation to the household head across informal households were significantly more likely to groups largely reflects the sex and age structure of each be poor (26.7 percent). group. In particular, the share of spouses, virtually all women, is smaller among informal wage (8.1 percent) and nonwage (9.4  percent) workers, compared with PUBLIC SECTOR, FORMAL, AND INFORMAL formal wage (17.8 percent), formal nonwage (11.9 per- JOBS HAVE DISTINCT CHARACTERISTICS cent), and public sector (24 percent) workers. Youth • Geographical location. About 72 percent of the employed ages 15–24 contribute 44.2 percent to informal wage population is located in urban areas (Table 3.2). This employment relative to 26.9 percent, 18.1 percent, and compares with 82.3 percent of public sector workers, 14.8 percent of formal wage, formal nonwage, and public about 80 percent of formal wage workers and 85.8 per- sector employment, respectively. cent of formal nonwage workers. By contrast, informal • Household welfare. Over 1 public sector worker in 2 and employment is more equally distributed between urban 1 formal wage or nonwage worker in 2 is in a household and rural areas; rural shares reach 62.5  percent and in the top four deciles (61.7 percent, 50.4 percent, and 58.5 percent of wage and nonwage informal employ- 56.5 percent, respectively).45 Only about 20.5 percent ment, respectively. The regional distribution of employ- of wage workers in the public sector, 29.0 percent of ment indicates that about 1 public sector wage workers formal wage workers, and 23.7 percent of formal non- in 3, 1 formal private sector wage worker in 3, and about wage workers live in households in the bottom 40. 45 percent of formal nonwage workers are in Greater Informal workers are not all less well off, however, Tunis. The corresponding share is around 20  percent particularly in the case of informal nonwage workers. in the case of informal wage and nonwage workers. The distribution of informal nonwage workers across Formal wage workers are located largely in the most deciles is roughly even up to the 8th decile; then the share highly developed areas of the country (over 84 percent), declines modestly to reach a minimum of 6.6 percent in namely, the North-East and the Center-East, including the highest decile. However, to capture the welfare impli- Greater Tunis. By contrast, the share of informal work- cations of work in formal and informal jobs, the house- ers located in these areas is around 60 percent (62.5 per- hold dimension is revealing. Besides households with no cent among informal wage workers and 58.5 percent working members, which, in Tunisia, represented about among informal nonwage workers), and their presence 22 percent of all households in 2015, households may is sizable in the western and southern regions of Tunisia. have members employed formally or informally, and • Industry. Most workers in the public sector are civil households with more than one working member may servants employed in public administration or in health exhibit a degree of formality or informality. Excluding care, education, and social services (77.2 percent). Some households with no employed members, households are employed in SOEs in transport, utilities, manufac- have therefore been classified as completely formal, turing, agriculture, and mining. There are differences completely informal, or mixed. About 56  percent of in the contribution to formal and informal employ- households were completely formal; 26.8 percent com- ment in the various industrial sectors. Agriculture, con- pletely informal; and 16.8 percent mixed. Relative to a struction, and trade are the largest contributors to informal employment. Manufacturing contributes over 40 percent of formal wage employment, while trade contributes over one-third to formal nonwage 45  The distribution of workers along the distribution of per capita consump- tion is based on data from the 2015 household budget survey, as opposed employment. Construction and agriculture contribute to the rest of the analysis, which relies on the 2019 labor force survey. the largest share to informal employment. Construc- Moreover, because of a lack of information in the household budget survey, the definition of informal employment only accounts for workers who are tion dominates informal wage employment with a share affiliated with social security rather than the official definition introduced of 39.4 percent. Agriculture, at 36.2 percent, and trade, at by INS in the 2019 labor force survey that accounts for access to paid and sick leave in cases in which access to social scurity is not reported by 29.5 percent, are the largest contributors to informal non- respondents. wage employment. The contribution of sectors, such as Employment and Wage Outcomes 125 TABLE 3.2. Distribution of Public Sector, Formal, and Informal Workers, by Job Characteristics, 2019 Public wage Formal wage Informal wage Formal nonwage Informal nonwage Indicator workers workers workers workers workers Region Greater Tunis 31.9 33.2 19.8 45.1 19.5 North-East 11.5 20.0 16.1 9.1 14.4 North-West 11.4 4.8 9.9 7.3 14.3 Center-East 18.9 31.2 26.6 25.7 24.4 Center-West 10.3 4.0 13.0 4.6 15.3 South-East 8.4 5.1 10.1 6.0 7.1 South-West 7.7 1.8 4.6 2.2 5.1 100.0 100.0 100.0 100.0 100.0 Location Rural 17.7 20.6 37.5 14.2 41.5 Urban 82.3 79.4 62.5 85.8 58.5 100.0 100.0 100.0 100.0 100.0 Industry Agriculture 1.5 5.1 16.5 4.7 36.2 Mining 0.9 0.8 0.1 0.2 0.0 Manufacturing 2.4 40.5 12.5 14.0 9.2 Utilities 3.4 0.4 0.2 0.4 0.6 Construction 0.6 13.9 39.4 5.8 6.8 Trade/repair 1.3 11.3 11.6 33.8 29.5 Transportation and storage 6.0 3.8 2.4 8.0 7.3 Accommodation and food service 0.5 6.9 6.6 5.0 2.4 activities Information and communication 1.4 2.7 0.5 1.4 0.6 Financial, insurance and real estate 1.7 2.1 0.3 1.2 0.3 activities Professional, scientific and technical 0.5 1.9 0.6 8.7 1.0 activities Administrative and support service 0.9 2.8 1.2 2.1 0.4 activities Public administration and defense, 36.9 1.0 0.3 0.4 0.2 social security Education, health, and social work 40.3 4.3 2.6 8.9 1.2 activities Activities of households as employers 0.0 0.9 2.7 0.2 0.2 Other services 1.7 1.7 2.6 5.2 4.1 100.0 100.0 100.0 100.0 100.0 (continued) 126 Tunisia’s Jobs Landscape TABLE 3.2. Distribution of Public Sector, Formal, and Informal Workers, by Job Characteristics, 2019 (continued) Public wage Formal wage Informal wage Formal nonwage Informal nonwage Indicator workers workers workers workers workers Occupation Managers 7.3 3.7 0.3 22.7 4.4 Professionals 32.2 7.7 1.9 14.2 1.7 Technicians and associate professionals 9.6 8.0 1.9 4.2 1.2 Clerical support workers 10.5 6.3 1.8 0.5 0.2 Service and sales workers 22.0 11.8 15.0 28.9 26.0 Skilled agricultural workers 0.8 2.7 9.0 4.1 32.7 Craft and related trades workers 2.1 15.6 21.2 15.3 17.8 Plant and machine operators, and 4.6 25.9 7.6 8.6 8.4 assemblers Elementary occupations 11.0 18.2 41.3 1.5 7.5 100.0 100.0 100.0 100.0 100.0 Type of contract Fixed-term contract 7.7 24.8 26.1 Open-ended contract 85.7 57.5 20.1 No contract 5.8 17.2 52.7 Not stated 0.8 0.4 1.2 100.0 100.0 100.0 Firm size 1–5 5.7 24.9 64.8 75.7 89.0 6–9 3.2 6.5 10.4 5.7 3.5 10–49 37.5 21.1 12.0 9.2 2.2 50+ 44.6 38.6 7.1 3.6 0.5 Not stated 9.0 8.9 5.8 5.8 4.8 100.0 100.0 100.0 100.0 100.0 Source: Based on data from the Labor Force Survey (ENPE) 2019, INS. Note: Statistics are based on data from the second quarter of the Labor Force Survey (ENPE) 2019. ICT services, finance, insurance and real estate activi- About 7.0 (4.1) percent of informal nonwage (wage) ties, professional, scientific, and technical activities, and workers are managers, professionals, or technicians, administrative and support services is about twice as which compares with 49.0  percent of public sector large in formal employment compared with informal workers, 41.0 percent of formal nonwage workers, and employment. 19.4  percent of formal wage workers. The relatively • Occupation. In line with their educational level, the large share of plant and machine operators and assem- large majority of informal workers are active in medium blers among formal workers (25.9 percent) is largely and low-end occupations, which also account for ascribable to the sizable number of formal workers about 51.0 percent of public sector workers, 58.9 per- employed in manufacturing. cent of formal nonwage workers, and 80.0 percent of • Contract type. The share of workers with open-ended informal wage workers. More than 4 informal wage contracts is significantly larger among public sector workers in 10 are employed in elementary occupations; workers (over 85.0 percent) compared with formal 21.2 percent are employed as craft and trade workers; wage workers (57.5 percent) and informal wage workers and 15.0 percent as services and sales workers. In the (20.1 percent). Fixed-term contracts are more common case of informal nonwage workers, the largest share is among formal wage workers (24.8 percent) and infor- made up of skilled agricultural workers, followed by mal workers (26.1 percent) relative to workers in the service and sales workers and craft and trades workers. public sector (7.7 percent). Informal workers are more Employment and Wage Outcomes 127 FIGURE 3.2. Informality Rates and the Contribution to Total Employment, by Type of Employment, 2019 a. Informality rates b. Contribution to total employment 100 25.0 90 80 20.0 70 60 15.0 Percent Percent 50 100.0 92.4 40 10.0 21.6 72.0 30 14.7 20 43.9 5.0 29.0 10 4.9 2.6 0 0.0 All Wage Employers Own- Unpaid Wage Employers Own- Unpaid workers account family workers account family workers workers workers workers Source: Based on data from the Labor Force Survey (ENPE), INS. exposed to the threat of job loss. More than 1 in 2 is the largest, at 21.6 percent, followed by 14.7 percent does not have a contract. The corresponding share is of own-account workers, 4.9 percent of employers, and estimated at 17.2 percent among formal workers and 2.6 percent of unpaid family workers (Figure 3.2, panel b). 5.8 percent among public sector workers. • Enterprise size. Informal wage and nonwage workers Most formal employees work in formal production units. are typically employed in microenterprises. Almost Combining the concept of formality at the worker level 65.0 percent of informal wage workers and 89.0 per- among wage workers with that of formality among the cent of informal nonwage workers are employed in economic units that employ them reveals that the overlap enterprises with fewer than six employees, compared between the formality status of wage workers and firms with 24.9 percent of formal wage workers and fewer is high. Overall, 59.0 percent of formal wage workers are than 6  percent of public sector workers. Most for- active in formal enterprises; only 11.0 percent of formal mal nonwage workers are active in microbusinesses, wage workers are employed in informal firms (Figure 3.3). which is partly a consequence of the large share of Informal wage workers in informal firms contribute about own-account workers in this group. The share of wage workers in small firms (six–nine employees) is FIGURE 3.3. The Distribution of Wage Employment, slightly larger in the case of informal workers relative by the Formality Status of Workers and Firms, 2019 to formal workers (10.4 percent vs 6.5 percent). About 19.0 percent of informal wage workers are employed 70.0 in firms each with more than nine employees, whereas 60.0 59.0 the shares are 59.7 percent and 82.1 percent among formal and public sector wage workers, respectively. 50.0 40.0 Percent Own-account workers exhibit the highest informality rate; yet, informal wage employment is the largest contributor 30.0 24.4 to total employment. At 29.0 percent, informality rates are 20.0 considerably lower than the average (43.9 percent) among 11.0 wage workers, and they are much higher among employers 10.0 5.6 and own-account workers (Figure 3.2, panel a).46 However, 0.0 because of the large number of wage workers in Tunisia, the in informal in formal in informal in formal contribution of informal employees to total employment rms rms rms rms informal wage workers formal wage workers The informality rate among unpaid family workers if, by definition,   46 100 percent. Source: Based on data from the Labor Force Survey (ENPE), INS. 128 Tunisia’s Jobs Landscape 24.0 percent to total wage employment, while informal wage are typically higher in nonwage employment than among workers in formal firms contribute only about 5.6 percent. wage workers. There are some important differences by worker characteristics. First, women and men employed Informality rates are heterogenous across worker character- as nonwage workers do not show considerably different istics and types of employment. Among wage workers, infor- informality rates (84 percent vs. 82 percent). Second, young mality rates are higher among men than women (32.9 percent nonwage workers have higher informality rates than other vs. 22.8 percent), among youth ages 15–25 (55.6 percent), age-groups, though the gap is smaller relative to wage among workers with no schooling (48.8 percent) or with employment. Third, virtually all nonwage workers with no primary education (40.7  percent), and among workers schooling are informal (95.5 percent), and the informality who live in households in the bottom three deciles of the rates are also high among workers with primary (88.9 per- consumption distribution (68.7 percent, 53.8 percent, and cent) and secondary education (82.4  percent). The dif- 51.8 percent, respectively) (Table 3.3). Informality rates ference is great in the informality rate between wage TABLE 3.3. Informal Employment, by Type and Contribution and by Individual and Household Characteristics, 2019 Wage workers Nonwage workers Contribution to Contribution to Indicator Informality rate total employment Informality rate total employment Sex Women 22.8 30.1 82.0 16.0 Men 32.9 69.9 84.0 84.0 Age-group 15–24 55.6 11.3 94.6 7.3 25–34 32.7 26.7 84.2 17.6 35–44 24.4 30.1 79.9 27.1 45–54 21.5 21.2 83.3 23.7 55–64 26.3 9.6 84.4 17.5 65+ 38.8 1.1 85.5 6.7 Educational level No education 48.8 7.1 95.5 12.5 Primary 40.7 31.5 88.9 42.8 Secondary 30.8 37.3 82.4 34.2 Tertiary 8.6 24.0 52.1 10.4 Not stated 18.5 0.1 88.4 0.1 Decile of household consumption per capita, 2015 Lowest decile 68.7 3.2 77.8 1.4 2 53.8 3.0 67.2 1.5 3 51.8 3.3 61.3 1.5 4 44.5 2.8 61.9 1.6 5 41.6 2.9 52.6 1.4 6 36.5 2.6 50.5 1.4 7 29.0 2.1 50.2 1.5 8 28.2 2.2 42.9 1.3 9 21.4 1.7 38.8 1.2 Highest decile 13.7 1.2 27.5 0.9 Source: Based on data from the Labor Force Survey (ENPE) 2019 and Household Budget Survey (HBS) 2015, INS. Note: Statistics are based on data from the second quarter of the Labor Force Survey (ENPE) 2019. Statistics by decile of per capita household consumption based on the Household Budget Survey (HBS) 2015 refer to the employed population ages 18 and above due to an age-based skip pattern concerning the question about affiliation to social security. Employment and Wage Outcomes 129 and nonwage workers with tertiary education (8.6  per- in the public sector (3.2 percent) and the private sector as cent vs. 52.1 percent). Informality rates are the highest at formal wage workers (9.2 percent) or unpaid family workers the bottom of the welfare distribution, but, unlike in the (1.2  percent) compared with men, whereas women have case of wage employment, nonwage workers at the top of a smaller chance of working as a nonwage workers or as the distribution show high informality rates, estimated at informal wage workers (−1.7 percent). Higher educational 42.9 percent, 38.8 percent, and 27.5 percent in the 8th, attainment is positively correlated with the probability of 9th, and 10th deciles. Such a pattern might indicate that employment as public or private formal wage workers and informality among nonwage workers, particularly employers negatively with the probability of employment as informal and own-account workers, may be more a choice than a wage workers. Workers with tertiary education show a last resort option. 39.0 percent greater probability of working in the public sector relative to workers with no schooling. The correla- Informality rates differ across job characteristics, particu- tion with the probability of working in nonwage employ- larly in wage employment. Informality rates among wage ment is economically insignificant. Geographical location in workers are higher in the North-West, Center-West, and terms of region or urban or rural area matters. Residing in South-East (from 35.9 percent to 45.8 percent) and in rural an urban area raises the probability of working as a public areas (45.2 percent), agriculture (66.0 percent), construc- (4.6 percent) or private formal (3.8 percent) wage worker tion (66.5 percent), and household services (68.6 percent) and is associated with a lower probability of employment as (Table 3.4). This is reflected in the informality rates by occu- an informal own-account worker (−4 percent) or an unpaid pation. Skilled agricultural workers, workers employed in family worker (−2.5 percent). The likelihood of employment elementary occupations, and craft and trade workers display as a formal wage worker in the private sector is negatively the highest rates of informality among wage workers. Wage associated with all regions, except for the North-East com- workers with no contracts (64.1 percent), and workers in pared with Greater Tunis. The opposite holds in the case of microenterprises (61.8  percent) and small (46.3  percent) informal wage workers in Greater Tunis relative to all other firms exhibit informality rates that are considerably higher regions. The regression also controls for household demo- than the average (43.9 percent) among wage workers. Non- graphics, namely, marital status and number of children ages wage workers show higher informality rates than wage 0–4 and 5–14 in the household, with the aim of capturing workers, and the rates are particularly high in rural areas, the effect of more household-friendly and flexible working in agriculture, construction, household services, trade, and arrangements in the case of formal wage jobs in the pub- transport, and among workers employed in medium and lic and private sectors. Marital status seems to be positively low-end occupations and in microenterprises and small associated with formal wage employment in the public and businesses. private sectors. The microdeterminants of the probability of employment in specific categories suggest uneven access to quality jobs in Wage Trends, Wage Gaps, and the public sector and to formal jobs by age, sex, educational Returns to Education level, and geographical location. Figure 3.4, panels a–f, illus- trates the marginal effects of a number of covariates derived This section takes advantage of wage data collected by the from estimating a multinomial logit regression on the deter- National Institute of Statistics (INS) with the labor force minants of working in one of eight categories, namely, public survey and illustrates how workers benefit from the labor sector wage work, formal wage work, informal wage work, market and which individual characteristics are correlated formal or informal employer, formal or informal own- with wages. Although no information on labor income is account worker, and unpaid family worker. Age is a strong available for workers employed as own-account workers correlate of working as a wage worker in the public sector and employers, wages are the main source of income from and in the formal or informal private sector. For example, labor for the majority of the population because more workers ages 25–29 exhibit a 4.2 percent greater likelihood than 3 workers in 4 in Tunisia work for wages. This sec- of working in the public sector relative to youth ages 15–24 tion (1) presents recent trends in wages at the aggregate and a 1.3 percent greater probability of working as formal level and by worker and job characteristics; (2) investi- wage workers in the private sector. The chances turn nega- gates the existence and correlates of gender wage gaps in tive, however, among older workers in the case of formal the private and public sectors, wage gaps among wage employees. Women have a higher probability of working workers in the public and private sectors, and wage gaps 130 Tunisia’s Jobs Landscape TABLE 3.4. Share of Informal Employment, by Type and Contribution and by Job Characteristics, 2019 Wage workers Nonwage workers Contribution Contribution to to nonwage Indicator Informality rate wage employment Informality rate employment Region Greater Tunis 20.5 28.8 68.9 23.7 North-East 29.3 16.4 89.0 13.5 North-West 35.9 8.2 91.0 13.1 Center-East 30.2 26.3 83.0 24.6 Center-West 45.8 8.5 94.4 13.6 South-East 40.0 7.6 86.0 6.9 South-West 31.8 4.3 92.1 4.6 Location Rural 45.2 24.8 93.7 37.1 Urban 24.8 75.2 77.8 62.9 Industry Agriculture 66.0 7.4 97.6 31.0 Mining 5.8 0.7 57.4 0.1 Manufacturing 17.7 21.2 77.2 10.0 Utilities 4.9 1.2 88.2 0.6 Construction 66.5 17.7 85.7 6.7 Trade/repair 40.5 8.5 81.7 30.2 Transportation and storage 17.7 4.0 82.3 7.4 Accommodation and food service activities 39.7 4.9 71.2 2.8 Information and communication 9.0 1.6 68.3 0.7 Financial, insurance and real estate activities 5.8 1.4 55.7 0.5 Professional, scientific and technical activities 16.0 1.1 37.2 2.2 Administrative and support service activities 20.6 1.8 50.0 0.7 Public administration and defense, social security 0.9 11.1 65.6 0.2 Education, health, and social work activities 5.5 14.2 41.0 2.5 Activities of households as employers 68.6 1.2 85.7 0.2 Other services 39.0 2.0 80.3 4.3 Occupation Managers 2.2 3.7 49.9 7.4 Professionals 4.4 13.0 38.1 3.7 Technicians and associate professionals 8.4 6.6 59.8 1.7 Clerical support workers 8.6 6.1 65.2 0.2 Service and sales workers 28.5 15.7 82.2 26.5 Skilled agricultural workers 67.1 4.0 97.6 28.1 Craft and related trades workers 47.3 13.4 85.7 17.4 Plant and machine operators, and assemblers 15.8 14.3 83.4 8.4 Elementary occupations 53.6 23.0 96.2 6.5 (continued) Employment and Wage Outcomes 131 TABLE 3.4. Share of Informal Employment, by Type and Contribution and by Job Characteristics, 2019 (continued) Wage workers Nonwage workers Contribution Contribution to to nonwage Indicator Informality rate wage employment Informality rate employment Type of contract Fixed-term contract 38.4 20.3 Open-ended contract 11.0 54.4 No contract 64.1 24.5 Not stated 46.3 0.8 Firm size 1–5 61.8 31.3 85.8 86.9 6–9 46.3 6.7 75.7 3.9 10–49 15.5 23.1 55.4 3.4 50+ 6.8 30.9 39.8 1.0 Not stated 21.6 8.0 80.9 5.0 Source: Based on data from the Labor Force Survey (ENPE) 2019, INS. Note: Statistics are based on data from the second quarter of the Labor Force Survey (ENPE) 2019. FIGURE 3.4. Marginal Effect of Selected Covariates on the Probability of a Specific Type of Employment, 2019 a. Age .042 25–29 .1 Public employee 30–44 .19 45–64 .013 25–29 –.018 Formal employee 30–44 –.084 45–64 –.074 25–29 –.16 Informal employee 30–44 –.22 45–64 .004 25–29 .014 Formal employer 30–44 .02 45–64 .013 25–29 .027 Informal employer 30–44 .044 45–64 .0059 25–29 .011 Formal own-account 30–44 .014 45–64 .018 25–29 .058 Informal own-account 30–44 .089 45–64 –.021 25–29 –.038 Unpaid worker 30–44 –.045 45–64 –.2 –.1 0 .1 .2 Marginal effect (continued) 132 Tunisia’s Jobs Landscape FIGURE 3.4. Marginal Effect of Selected Covariates on the Probability of a Specific Type of Employment, 2019 (continued) b. Sex .032 Public employee .092 Formal employee –.017 Informal employee –.014 Formal employer –.045 Informal employer –.0077 Formal own-account .053 Informal own-account .012 Unpaid worker –.05 0 .05 .1 Marginal effect c. Educational attainment Primary Public employee Secondary Tertiary Not Stated Primary Formal employee Secondary Tertiary Not Stated Primary Informal employee Secondary Tertiary Not Stated Primary Formal employer Secondary Tertiary Not Stated Primary Informal employer Secondary Tertiary Not Stated Primary Formal own-account Secondary Tertiary Not Stated Primary Informal own-account Secondary Tertiary Not Stated Primary Unpaid worker Secondary Tertiary Not Stated –.4 –.2 0 .2 .4 Marginal effect (continued) Employment and Wage Outcomes 133 FIGURE 3.4. Marginal Effect of Selected Covariates on the Probability of a Specific Type of Employment, 2019 (continued) d. Region of residence North-East North-West Public employee Center-East Center-West South-East South-West North-East North-West Formal employee Center-East Center-West South-East South-West North-East North-West Informal employee Center-East Center-West South-East South-West North-East North-West Formal employer Center-East Center-West South-East South-West North-East North-West Informal employer Center-East Center-West South-East South-West North-East North-West Formal own-account Center-East Center-West South-East South-West North-East North-West Informal own-account Center-East Center-West South-East South-West North-East North-West Unpaid worker Center-East Center-West South-East South-West –.3 –.2 –.1 0 .1 .2 Marginal effect (continued) FIGURE 3.4. Marginal Effect of Selected Covariates on the Probability of a Specific Type of Employment, 2019 (continued) e. Urban residence .018 Public employee .008 Formal employee –.038 Informal employee .01 Formal employer .0024 Informal employer .0035 Formal own-account .01 Informal own-account .025 Unpaid worker –.05 0 .05 Marginal effect f. Marital status and number of children 0–4 and 5–14 married Public employee Number of children ages 0–4 Number of children ages 5–14 married Formal employee Number of children ages 0–4 Number of children ages 5–14 married Informal employee Number of children ages 0–4 Number of children ages 5–14 married Formal employer Number of children ages 0–4 Number of children ages 5–14 married Informal employer Number of children ages 0–4 Number of children ages 5–14 married Formal own-account Number of children ages 0–4 Number of children ages 5–14 married Informal own-account Number of children ages 0–4 Number of children ages 5–14 married Unpaid worker Number of children ages 0–4 Number of children ages 5–14 –.1 –.05 0 .05 Marginal effect Source: Based on data from the Labor Force Survey (ENPE), INS. Note: The reference categories are the following: 15–25 years of age; men; no schooling; Greater Tunis; rural areas; not married. Employment and Wage Outcomes 135 FIGURE 3.5. Trends in Real Monthly and Hourly Wages, Average and Median Values, 2012–19 a. Monthly wage, 2012–19 b. Hourly wage, 2012–19 700.0 5 600.0 Monthly wage (2019 prices) 4 Hourly wage (2019 prices) 500.0 3 400.0 300.0 2 200.0 1 100.0 0.0 0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 Median Mean Median Mean Source: Based on data from the Labor Force Survey (ENPE), INS. among formal and informal wage workers in the private monthly wage was estimated at about TD 583. It increased sector; and (3) analyzes correlates of wages, in particular to about TD 644 in 2017. Since 2017, the rising trend in returns to education.47 wages has been reversed, and average monthly wages have declined by over 3 percent a year. The average in 2019 stood at around TD 598 a month (Figure 3.5, panel a). A similar TRENDS IN WAGES trend is observed among average hourly wages. In 2012, an employee made, on average, TD 3.5 per hour worked. This Between 2012 and 2019, average monthly wages did not had increased to about TD 4 by 2017, and subsequently increase considerably, and, in 2019, the average monthly declined to TD 3.7 in 2019 (Figure  3.5, panel b). Over wage was estimated at about TD 600. In 2012, the average 2012–19, the annualized growth rate of mean hourly wages was higher compared with monthly wages (about 0.8 per- 47  The wage statistics presented in this chapter refer to monthly and hourly wages expressed in 2019 prices for employees ages 15–64. Wages exclude cent compared with 0.4 percent) because of a decline in the bonuses and benefits in kind. As workers are allowed to report their last average number of working hours from about 45.3 hours pay according to a reference period of choice, wages are first converted per week in 2012 to 43.1 hours per week in 2019. into monthly values, which is the most commonly used reference period in the survey. Wages reported by the week or by the day are converted into monthly values by multiplying weekly and daily values by the number of The secondary and services sectors pay, on average, more weeks and days worked during the previous month, respectively. If information about weeks and days worked is missing, wage values are than agriculture, but the sectoral wage gap has been reduced multiplied by 4.33 (the number of weeks in a month) and 22 (number over time. In 2012, the average monthly wage was about of working days in a month), respectively. This means one assumes that workers work full time during the previous month. In addition, fewer than 54  percent higher in the secondary sector than in agri- 0.05 percent of wage workers, with the exception of 2014 (4.4 percent) and culture, and the services sector paid more than twice 2019 (0.5 percent), report their last wage according to an unknown time unit. The reported wage amount has been used without applying any con- as much as agriculture (123 percent) (Figure 3.6, panel a). version factor in this case. Hourly wages have been calculated by dividing Between 2012 and 2015, the sectoral wage gap declined monthly wage values by the number of hours worked the week preceding considerably, and, since then, it has expanded modestly. In the interview, multiplied by 4.33. The question regarding the number of hours worked per week refers to all jobs in all survey rounds until 2018 2019, the wage gap was estimated at about 48 percent in and to the main job only starting with the 2019 round of the survey. For the secondary sector and about 93 percent in the services the sake of comparability, wage workers with two jobs have been excluded from the analysis. This leads to the exclusion of fewer than 0.5 percent of sector. The dynamic observed in the wage gap is the wage workers in each year. All hourly wage values that are more (or less) by-product of two factors. Average monthly wages in agri- than 3 standard deviations away from the mean in each survey year are identified as outliers and are excluded from the analysis. Wage statistics culture increased more rapidly until 2017 and subsequently are based on wages among the population of wage workers who report declined less than in the other two sectors. Over the entire their wage during the interview. The share of wage workers who did not respond to the question concerning the wage amount increased from about period (2012–19), agricultural wage workers had a cumu- 25 percent in 2012 to about 53 percent in 2019. lative gain in their average wages of about 16  percent, 136 Tunisia’s Jobs Landscape FIGURE 3.6. Trends in Real Average Monthly Wages, by Broad Industrial Sector, 2012–19 a. Average monthly wage, secondary and services b. Cumulative growth rate, average monthly wage, sectors, ratio of average monthly wage in by broad industrial sector, 2012–19 agriculture, 2012–19 2.50 25.0 Wage Ratio (/agriculture) 2.00 15.0 1.50 5.0 Percent –5.0 1.00 –15.0 0.50 –25.0 0.00 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 Agriculture Secondary Services Secondary Services Source: Based on data from the Labor Force Survey (ENPE), INS. whereas secondary sector workers posted a cumulative higher in public administration and public companies than in rise of about 11 percent. By contrast, the average monthly the private sector (Figure 3.7, panel a). After a slowdown in wage in the services sector remained virtually at the level the aftermath of the 2011 revolution, average monthly wages of 2012 (Figure 3.6, panel b). in public administration and public companies started to rise rapidly until 2017. Over 2018–19, average wages declined Workers employed in public administration and in public in both the private and public sectors (Figure 3.7, panel b). companies are paid, on average, more than private sector The overall increase in average monthly wages was about workers and posted larger increases in wages over time. 9.2 percent in the private sector and about 6.7 percent in Average monthly wages of workers have been about 1.8 times public administration and public companies. FIGURE 3.7. Trends in Real Average Monthly Wages, by Sector, 2012–19 a. Ratio of average monthly wage in the public b. Cumulative growth rate in average monthly sector to the private sector, 2012–19 wages, by sector, 2012–19 2.00 25.0 1.80 15.0 Wage Ratio (/private) 1.60 1.40 1.20 5.0 Percent 1.00 –5.0 0.80 0.60 –15.0 0.40 0.20 –25.0 0.00 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 Private sector Public sector Source: Based on data from the Labor Force Survey (ENPE), INS. Employment and Wage Outcomes 137 FIGURE 3.8. Trends in the Real Average Monthly Wage, by Educational Level, 2012–19 a. Ratio of average monthly wages, workers with b. Cumulative growth rate in average monthly education to workers with no education, 2012–19 wage, by educational attainment, 2012–19 3.00 25.0 2.50 Wage Ratio (/no education) 15.0 2.00 5.0 Percent 1.50 –5.0 1.00 –15.0 0.50 –25.0 0.00 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 None Primary Primary Secondary Tertiary Secondary Tertiary Source: Based on data from the Labor Force Survey (ENPE), INS. The wage premium of workers with tertiary education GENDER WAGE GAPS declined considerably. As expected, the average monthly Gender gaps in labor force participation are partly the wage increases with the level of education attained by by-product of a household-level bargaining process, whereby workers. For example, in 2012, a worker with tertiary edu- gender gaps in earnings and other factors determine the cation made, on average, almost 2.8 times the amount made bargaining power of men and women within the house- by a worker with no education, while a worker with sec- hold. This section investigates the existence of gender ondary education made about 1.6 times the amount made by a worker with no education (Figure 3.8, panel a). The wage gaps, thus restricting the analysis to the popula- wage dynamic differs considerably according to worker tion of wage workers, which, in the case of Tunisia, rep- educational attainment. While workers with no schooling resents about 75  percent of the employed population or only primary education experienced large gains and then, and over 85 percent of employed women. In Tunisia, more recently, modest reductions in the monthly wage, women are more likely than men to be employed for workers with secondary education and, especially, workers a wage in the public sector (see Table  3.1). There are with tertiary education posted modest gains and then, more important differences in educational attainment and the recently, sizable reductions in monthly wages (Figure 3.8, distribution across occupations between men and women panel b). The average monthly wages of workers with ter- employed in the private and public sectors (see chap- tiary education were 8 percent lower in 2019 relative to ter 2). For example, women employed as wage workers 2012. Workers with secondary education benefited from an are, on average, more well educated than men, and the increase in average wages until 2017, but this was sub­ share of women with tertiary education is much higher sequently undone. Workers with primary education enjoyed in the public sector than in the private sector. The share a cumulative increase of 14 percent, and workers with no of women with tertiary education in the public sector has schooling benefited from a cumulative growth of above also increased over time. Among men, the correspond- 19 percent. These trends are consistent with a labor market ing share remained constant over time and is estimated characterized by an abundance of individuals with secondary to be about 30 percentage points lower than the share and tertiary education who cannot be absorbed given a lack among women. In the private sector, women are largely of demand for these types of workers, and many of these employed in low- and mid-skill jobs, whereas most women workers end up unemployed.48 in the public sector perform high-skill jobs. To account for these important differences, the analysis of gender 48 The rest of the chapter highlights that returns to tertiary education, con-   wage gaps is conducted separately for private and public trolling for other observable worker characteristics, have declined over time. sector wage workers. 138 Tunisia’s Jobs Landscape In the private sector, women make, on average, $0.83 per minimum estimated at −13 percent around the median (Fig- hour to the $1.00 made by men, and the gap expands ure 3.9, panel a). along the wage distribution. The unconditional gender wage gap, which captures the gender differences in wages In the public sector, women make, on average, 46 percent without accounting for differences in the characteristics of more than men per hour of work, and the gap was more the pool of employed men and women, indicates that, on than twice as large among high earners. Women employed average in Tunisia in 2019, a woman employed in the pri- in the public sector earned about 9  percent more per vate sector earned about 16 percent less per hour worked month than men in 2019. The gap in favor of women rose relative to a man. The unconditional gender gap in hourly from 5 percent in 2012. The unconditional difference is wages increased along the distribution of hourly wages considerably larger in the hourly wage. In the public sec- from about −16 percent among the bottom 20 to about tor in 2019, women made about 46  percent more than −24 percent among the top 20 in 2019 (Figure 3.9, panel b). men, and the gap had expanded from 26 percent in 2012. In 2012, the gender gap had an inverted U shape, with a The larger gap in hourly wages relative to monthly wages FIGURE 3.9. Unconditional Gender Differentials in Hourly Wages, by Quantile and Sector, 2012 and 2019 a. Private sector, 2012 b. Private sector, 2019 30 30 20 20 10 10 Percent Percent 0 0 –10 –10 –20 –20 –30 –30 0 20 40 60 80 100 0 20 40 60 80 100 Percentile ∈[p5, p95] Percentile ∈[p5, p95] c. Public sector, 2012 d. Public sector, 2019 130 130 110 110 90 90 70 70 Percent Percent 50 50 30 30 10 10 –10 –10 –30 –30 0 20 40 60 80 100 0 20 40 60 80 100 Percentile ∈[p5, p95] Percentile ∈[p5, p95] Source: Based on data from the Labor Force Survey (ENPE), INS. Employment and Wage Outcomes 139 is ascribable to differences in working hours. Women and to unobserved characteristics or different treatments employed in the public sector work, on average, fewer of men and women, that is, the unexplained component hours relative to men. The gap expands along the distri- (Figure 3.10, panels a and b).50 In the public sector, the bution, and the premium in favor of women has expanded explained and unexplained components work in opposite over time (see Figure 3.9, panels c and d). In 2019, the directions. Differences in observable characteristics exert unconditional gender gap in hourly wages ranged from a positive effect on the gender hourly wage gap that shores about 3 percent at the 5th percentile to about 25 percent it up in favor of women. Among observable characteris- at the median. It peaked at over 110 percent at around the tics, the fact that women in the public sector are more 75th percentile. It declined to around 80 percent among well educated than men and are employed in high-end workers at the 85th percentile. occupations, such as managers and professionals, pushes the gender gap in favor of women (Figure 3.10, panel d). Controlling for observable characteristics, the analysis finds This is in line with the stylized facts presented in chapter 2, that, in 2019, women employed in the private sector earned, whereby women in the public sector are relatively more on average, about 18.5  percent less than men per hour concentrated in high-end occupations and have, on aver- worked; in the public sector, women made 33 percent more age, a higher educational level. The increase in the gen- than men per hour worked. Given the sizable differences der gap over time is largely ascribable to improvements in some of the characteristics of men and women employed in these characteristics among women employed in the in the public and private sectors, namely, in educational public sector relative to men. By contrast, the unexplained attainment and occupation, the unconditional gender wage component has a negative effect on the hourly wage, but gap is not a good indicator of the extent of discrimination the effect is modest. This component is associated with a in the labor market. Women and men working in the public different wage structure or with unobserved characteris- sector or the private sector are endowed with a set of char- tics that would, on average, make men more productive acteristics that make them more or less productive. The con- than women. ditional gender hourly wage differentials, that is, the hourly wage gaps obtained after controlling for a set of worker In the private sector, most of the average gender wage gap characteristics and estimated through wage equations, are is unexplained. In the private sector, too, the two compo- reported in Figure 3.10.49 The results indicate that women nents (explained and unexplained) run in opposite direc- in the private sector are paid hourly wages significantly tions in most years (Figure 3.10, panel b). The largest share lower compared with men. The gap was estimated at about of the hourly wage difference is ascribable to the unex- 18.5 percent in 2019, on average, and has hovered around plained component that exercises a large negative effect on this level over time. In the public sector, women received an the gender gap. Although the explained component pushes hourly wage premium of about 33 percent in 2019, up from the gender gap in positive territory in most years, its con- about 18 percent in 2012. tribution is relatively small, and the overall gender gap is negative. Among observable characteristics, differences A large part of the average gender gap in the public sector in educational level and occupation contribute positively, is explained by differences in the characteristics of men whereas differences in demographics, particularly a larger and women. Estimates from a twofold Blinder-Oaxaca share of working men in rural areas and at ages above 45, decomposition indicate the extent to which the differ- exert a large negative effect on the wage differential. ences observed in hourly wages between men and women are ascribable to differences in the observable character- istics of the two groups, that is, the explained component, 50  The Blinder-Oaxaca decomposition is used to gauge the extent to which differentials in hourly wages between men and women are ascribable to dif- ferences in the observed and unobserved characteristics of the two groups. The effect associated with the first difference constitutes the explained 49  Regressions control for a second-degree polynomial in age and individual component of the differential, also known as characteristics, composi- dummies for year-of-birth cohorts, educational level, region and urban or tion, or endowment effect, in that it reflects the portion of the differential rural location of residence, occupation, industry and sectoral category, that associated with group differences in individual observable attributes (for is, the domain of employment in the public sector (public administration, example, educational level, sector of activity, industry, occupation). The SOEs) and in the private sector (Tunisian or foreign or mixed privately effect related to the second difference is referred to as the unexplained owned company, private household business, and so on), type of contract component. This embodies the portion of the wage gap stemming from (fixed-term, open-ended, or no contract), and affiliation to social security the differential valuation of women and men’s characteristics in the labor (National Social Security Fund, National Social Security Fund, other, or market that arise because of differences in unobservable characteristics or no affiliation). unequal pay structures between the two groups. 140 Tunisia’s Jobs Landscape FIGURE 3.10. Oaxaca-Blinder Decomposition: Mean Gender Hourly Wage Differential, by Sector and Characteristics, 2012–19 a. Private sector b. Public sector 50.0 50.0 40.0 40.0 30.0 30.0 20.0 20.0 10.0 10.0 Percent Percent 0.0 0.0 –10.0 –10.0 –20.0 –20.0 –30.0 –30.0 –40.0 –40.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 difference explained unexplained difference explained unexplained c. Private sector: observable characteristics d. Public sector: observable characteristics 50.0 50.0 Percent Percent 0.0 0.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 demographics education demographics education job characteristics occupation job characteristics occupation industry industry Source: Based on data from the Labor Force Survey (ENPE), INS. In the private sector, the conditional gender hourly wage gender gap into negative territory at the median, particu- differential expands from about -16 percent at the bottom larly in recent years, and at the top (Figure 3.11, panels d, e, to over -25 percent at the top of the distribution.51 In the and f). The unexplained component contributes the largest private sector, the large bulk of the gender hourly wage share to the gender gap along the distribution and par- difference is concentrated in the upper half of the distribu- ticularly at the bottom. Job characteristics, followed by tion (Figure 3.11, panels a, b, and c). It was estimated at occupation and educational level, are the main drivers of –15.6 percent at the 10th percentile, −14.7 percent at the the effect in favor of women of the explained component median, and -24.9 percent at the 90th percentile in 2019. at the bottom. In the middle and at the top of the distri- In 2012–19, it declined at the bottom and increased at bution, differences between men and women in job char- the median, while it stayed roughly constant at the top. acteristics, industrial sector, and demographic differences The decomposition results on this sector show that the push the gender gap in favor of men, and the positive effect explained and unexplained components operate in oppo- played by educational level and occupation is too modest site directions at the bottom, whereas they both push the to overturn the effect of other characteristics. In the public sector, the gender gap differentials are lowest at 51  To unpack the gender gap along the distribution, an unconditional quan- tile regression is estimated at selected percentiles using the rifreg command the tails of the distribution and rise in the middle, displaying in Stata. an inverted-U shape pattern. In 2019, at the 10th percentile, Employment and Wage Outcomes 141 FIGURE 3.11. Oaxaca-Blinder Decomposition: Gender Hourly Wage Differential at Selected Percentiles, Private Sector, 2012–19 a. Percentile 10 b. Percentile 50 50.0 50.0 40.0 40.0 30.0 30.0 20.0 20.0 10.0 10.0 Percent Percent 0.0 0.0 –10.0 –10.0 –20.0 –20.0 –30.0 –30.0 –40.0 –40.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 difference explained unexplained difference explained unexplained c. Percentile 90 d. Percentile 10: observable characteristics 50.0 50.0 40.0 30.0 20.0 10.0 Percent Percent 0.0 0.0 –10.0 –20.0 –30.0 –40.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 demographics education difference explained unexplained job characteristics occupation industry e. Percentile 50: observable characteristics f. Percentile 90: observable characteristics 50.0 50.0 Percent Percent 0.0 0.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 demographics education demographics education job charactiristics occupation job charactiristics occupation industry industry Source: Based on data from the Labor Force Survey (ENPE), INS. 142 Tunisia’s Jobs Landscape the conditional hourly wage gap was estimated at about roles can contribute to the gender wage gap. Data of the 6  percent, in positive territory compared with 2012 most recent available round of the labor force survey show (−9.6 percent), at about 36.0 percent at the median (up that, in 2017, women and men tended to work in different from 17.8 percent in 2012), and at about 20.4 percent at sectors in the economy. Manufacturing (textiles, mechani- the 90th percentile (similar to the gap observed in 2012, cal goods, and electrical equipment), trade, and social and 19.2 percent) (see Figure 3.12, panels a, b and c). At the cultural services are the three most important sectors of bottom and at the median, the two components exert a women’s employment in the private sector, and construc- positive effect—that is, in favor of women—on the gender tion, agriculture, and trade dominate among men in the wage gap. Meanwhile, at the top, the unexplained com- private sector. Gender differences in the industrial sector ponent plays a negative effect, but the size of the effect of employment, occupation, and enterprise size contribute is not sufficiently large to undo the positive effect of the to the observed gender wage gap. observable characteristics of women. In addition, at the top and at the median, the explained component plays the largest role, and, at the bottom of the distribution, the WAGE GAPS AMONG SECTORS unexplained component contributes 60 percent to the gen- Workers employed in the public and the private sectors der wage gap. Similar to the case at the mean, differences have different observable characteristics as do formal and in educational level and occupation are the main drivers of the positive effect of the explained component at each informal workers employed in the private sector. This of the selected percentiles, and the differences in industrial section investigates the existence of wage gaps between composition exert a positive effect only at the top (Fig- wage workers in the public and private sector and the ure 3.12, panels d, e, and f). existence of wage gaps between formal and informal workers conditional on a set of observable individual Overall, these results suggest that, in the private sector, the and job characteristics. wage gap in favor of men is, to some extent, ascribable to systematic differences in the jobs to which women have Formal (informal) private sector workers made, on average, access, and women’s higher educational levels are not suf- $0.65 ($0.50) to the $1.00 made by public sector workers ficient to compensate for such obstacles. The main factor in 2019. There were sizable unconditional average wage behind the gender wage gap in the private sector remains, gaps between public and private sector wage workers and, however, related to differences in unobservable character- in the private sector, between formal and informal wage istics or in the pay structure, that is, an unequal pay struc- workers. The density of the monthly wages of wage workers ture to the disadvantage of women. By contrast, women in employed in the public sector was shifted to the right of the public sector receive a wage premium relative to men the density of the monthly wages of private sector workers, thanks to the considerably more productive endowments and there is only a partial overlap (Figure 3.13-panel a). of women and even though the unexplained component There was thus a larger share of public sector workers with operates in favor of men at the top of the distribution. high wages compared with workers in the private sector. By contrast, the density function of formal and informal Low wages among women in the private sector may reduce workers in the private sector shows a significant overlap. women’s incentives to join the labor force. Women are The density of the wages of informal workers slightly likely to continue to bear most of the household burden shifted to the left of the density of formal workers. A larger in housework and family care because of assigned gen- share of informal wage workers earn low wages relative to der roles (World Bank 2021a). Such activities compete formal wage workers. The median monthly wage among for women’s time spent with work on the labor market. public sector workers is about TD 1,000, which compares The wage gap might keep some women out of the labor with about TD 540 and TD 435 in the case of formal and market or push some women to look for less competitive informal wage workers, respectively. The large differences and less remunerative career paths and greater flexibility are evident by looking at the cumulative distribution func- at work. Reducing and eliminating the gender wage gap tions, which denote the proportion of wage workers whose in the private sector has the potential of helping increase wages fall below a given level, illustrated in Figure 3.13, women’s participation and contributing to promoting panel b. For example, less than 25 percent of public sector inclusive growth and achieving the full potential of the wage workers have a wage of TD 600 or lower, whereas the economy, particularly among women with little education. percentage is about 67 percent and 80 percent in the case In addition, labor market segregation and assigned gender of formal and informal wage workers. Employment and Wage Outcomes 143 FIGURE 3.12. Oaxaca-Blinder Decomposition: Gender Hourly Wage Differential at Selected Percentiles, Public Sector, 201219 a. Percentile 10 b. Percentile 50 50.0 50.0 40.0 40.0 30.0 30.0 20.0 20.0 10.0 10.0 Percent Percent 0.0 0.0 –10.0 –10.0 –20.0 –20.0 –30.0 –30.0 –40.0 –40.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 difference explained unexplained difference explained unexplained c. Percentile 90 d. Percentile 10: observable characteristics 110.0 50.0 90.0 70.0 50.0 Percent Percent 30.0 0.0 10.0 –10.0 –30.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 demographics education difference explained unexplained job characteristics occupation industry e. Percentile 50: observable characteristics f. Percentile 90: observable characteristics 50.0 50.0 Percent Percent 0.0 0.0 –50.0 –50.0 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 demographics education demographics education job characteristics occupation job characteristics occupation industry industry Source: Based on data from the Labor Force Survey (ENPE), INS. 144 Tunisia’s Jobs Landscape FIGURE 3.13. Probability Density and Cumulative Distribution Functions of Real Monthly Wages, by Sector, 2019 a. Probability density function of real monthly b. Cumulative distribution function of real monthly wages, by sector, 2019 wages, by sector, 2019 .003 1 .8 Wage Workers (%) .002 .6 Density .001 .4 .2 0 0 1000 2000 3000 4000 5000 0 Monthly wage - main job 0 0 0 0 0 0 0 0 0 0 20 40 60 80 00 20 40 60 80 1, 1, 1, 1, 1, Public Private Formal Private Informal Monthly wage Public Private Informal Private Formal Source: Based on data from the Labor Force Survey (ENPE), INS. Note: CDFs are truncated at the 99th percentile. CONDITIONAL WAGE GAPS BETWEEN A large part of the average public-private hourly wage PUBLIC AND PRIVATE SECTOR WORKERS gap is explained by differences in the characteristics of the wage workers employed in the two sectors. Estimates from In 2019, conditional on observable characteristics, wage a twofold Blinder-Oaxaca decomposition indicate that the workers employed in the public sector earned, on average, explained component contributes in all years, except 2012, over twice as much as wage workers in the private sector 50  percent or more to the estimated gap (Figure  3.14, per hour worked. Considerable differences exist between panel b). Differences in observable characteristics exert a wage workers employed in the public and private sectors, positive effect on the hourly wage gap that shores it up in including sex, age, educational level, geographical loca- favor of public sector workers. Among observable charac- tion, industry, and occupation. Therefore, it is crucial to teristics, the type of occupation and other job characteris- investigate the existence of wage gaps while controlling tics seem to play the largest role, followed by educational for observable characteristics. The conditional hourly level (Figure 3.14, panel c). There are also unobservable wage differentials are estimated through wage equations characteristics or a wage premium in favor of workers in and are reported in Figure 3.14, panel a.52 The results indi- the public sector. cate that, in 2019, wage workers in the public sector were paid, on average, over twice the amount paid per hour worked YOUNG UNIVERSITY GRADUATES: to workers in the private sector. The estimated public- UNEMPLOYMENT AND THE PUBLIC SECTOR private hourly wage gap declined from about 112  per- WAGE PREMIUM cent in 2012 to 89 percent in 2015 and then increased to around 106 percent in 2016–19. In 2019, the average monthly salary of youth ages 25–34, with tertiary education and employed in public adminis- tration was estimated at about TD 1,030, which compares with about TD 1,244 in the case of SOEs and with TD 734 52  Regressions control for a second-degree polynomial in age and individual dummies for year-of-birth cohorts, educational level, region and urban or and TD 466 in the case of a formal and informal employee rural location of residence, occupation, industry and sectoral category, that in the private sector, respectively (Figure 3.15). A univer- is, the domain of employment in the public sector (public administration, SOEs) and in the private sector (Tunisian or foreign or mixed privately sity graduate could thus expect to make, on average, about owned company, private household business, and so on), type of contract 40 percent more as a civil servant compared with a formal (fixed-term, open-ended, or no contract), and affiliation to social security (National Social Security Fund, National Social Security Fund, other, or employee in the private sector and 1.2 times the amount no affiliation). gained by informal employees. The gap in hourly wages was Employment and Wage Outcomes 145 FIGURE 3.14. Oaxaca-Blinder Decomposition: Mean Hourly Wage Differential, Wage Workers in the Public and Private Sectors, 2012–19 a. Conditional mean hourly wage gap between b. Oaxaca-Blinder decomposition: explained and public and private sector wage workers unexplained components 120.0 100% 27.9 23.7 33.7 100.0 80% 36.6 43.8 50.9 80.0 60% 89.3 Percent Percent 106.4 60.0 40% 72.1 76.1 66.4 63.4 56.2 49.1 40.0 20% 10.7 20.0 0% –6.2 0.0 –20% 2012 2013 2014 2015 2016 2017 2018 2019 2012 2013 2014 2015 2016 2017 2018 2019 Explained Unexplained c. Oaxaca-Blinder decomposition: breakdown of the explained component 100% 50% Percent 0% –50% 2012 2013 2014 2015 2016 2017 2018 2019 Demographics Education Industry Job characteristics Occupation Source: Based on data from the Labor Force Survey (ENPE), INS. even higher because employees in the public sector tend to in observable characteristics, conditional wage gaps are a work shorter hours relative to employees in the private sector. better indicator of whether working in the public sector The average hourly salary of a university graduate ages pays higher hourly wages than working in the private sector, 25–34 was estimated at about TD 8.0 in public adminis- all else being equal. In 2019, wage workers ages 25–34 tration and TD 3.8 and TD 2.7 as a formal and informal with tertiary education employed in the public sector, which employee in the private sector, respectively. This is in addi- comprises public administration and SOEs, were paid, on tion to the allowances, job security, and more favorable average, about 120 percent more per hour worked relative leave policies that the public sector provides (see Box 3.1). to workers in the private sector. This is slightly higher than the average gap estimated among all wage workers (esti- Among youth with tertiary education, the public sector mated at 106 percent in 2019). pays, on average, about 120 percent more per hour worked relative to the private sector. Because workers employed in The gap is largely ascribable to differences in observable the public sector and the private sector differ considerably characteristics. The twofold Oaxaca-Blinder decomposition 146 Tunisia’s Jobs Landscape FIGURE 3.15. Unconditional Mean Monthly Wage FIGURE 3.16. Oaxaca-Blinder Decomposition: Mean Gap, by Sector, 25–34 Age-Group with Tertiary Hourly Wage Differential Between Wage Workers Education, 2019 Ages 25–34 with Tertiary Education and Employed in the Public and Private Sector, 2019 1,400 Average monthly wage (TD) 1,200 100.0 10.2 1,000 80.0 800 600 60.0 Percent 400 89.8 200 40.0 – administration Public company Private formal Private informal 20.0 Public 0.0 2019 Explained Unexplained Source: Based on data from the Labor Force Survey (ENPE), INS. Source: Based on data from the Labor Force Survey (ENPE), INS. of the conditional hourly wage gap indicates that about compared with 12.5 percent in the private sector, followed 90 percent of the difference is ascribable to differences in by 16.2 percent with a degree in economics, management, observed characteristics. The characteristics that matter the or law, significantly lower compared with the same group most in explaining the wage gap are occupation, type of (25.5 percent) in the private sector (Figure 3.17). The share contract, place of work, and access to social security. Thus, of engineers, at 5.8  percent, was three times smaller in youth with tertiary education are more likely to gain access public administration, compared with the private sector to jobs that have specific characteristics positively corre- (19.7 percent), whereas the share of youth with a degree lated with wages in the public sector relative to the private in medicine or pharmacy was about twice as large in public sector. Although there are also unobservable characteristics administration relative to the private sector (4.8  percent and a wage premium in favor of wage workers in the public vs. 1.9 percent). The share of youth with a master’s degree sector, this component contributes about 10  percent to was estimated at 10.3  percent in public administration the wage gap (Figure 3.16). This raises the question about and 15.8 percent in the private sector, whereas the share whether expectations of a higher salary as a civil servant, of employees with a doctorate degree was about four times largely ascribable to the characteristics of the relevant jobs, larger in public administration (4.3 percent vs. 0.8 percent). raise the reservation wage of young university graduates and incentivize them to queue for a public sector job while Fewer than 1 university graduate in 2 not working as a civil idle.53 So, the question is what youth with the same level of servant is employed, and about 1 in 4 holds a formal wage education as civil servants might do. job in an SOE or in the private sector. About 46 percent of youth ages 25–34 with tertiary education and not working About 1 youth ages 25–34 in 2 employed in public admin- as civil servants were employed in 2015. Of this pool, istration in 2015 obtained a certificate of tertiary education. almost 55 percent were employed in SOEs (17 percent) Among these, 17.2 percent had a degree in humanities, or formally in the private sector (38 percent) (Figure 3.18, panel a). About 29.2  percent were informal employees, 53 It is difficult to provide direct evidence regarding queuing for jobs in   and the rest were roughly evenly split between employers public administration and the size of the queue because data on the number and own-account workers, plus a residual 1.8 percent of of job applications and openings would be required to construct measures of the size of the queue as the number of applicants per job posting. An unpaid family workers. It is difficult to ascertain whether example is the study of Krueger (1988), who finds that the application rate some of the youth employed outside public administration for federal jobs in the United States increases as the ratio of government to private sector earnings increases, although the rate is not related to the saw their jobs as a temporary buffer while waiting for their relative level of fringe benefits. preferred job in the public sector or as a good employment Employment and Wage Outcomes 147 FIGURE 3.17. Youth Ages 25–34 with Tertiary Education, Employed in Public Administration and in the Private Sector, by Type of Degree, 2015 private sector 12.5 25.5 9.1 19.7 1.9 14.7 15.8 0.8 public administration 17.2 16.2 13.5 5.8 4.8 28.0 10.3 4.3 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 Percent Humanities Economics, management, law Sciences Engineering Medicine, pharmacy Other diplomas Master Doctorate Source: Based on data from the 2015 Household Budget Survey (HBS), INS. outcome. The distribution by quintile of household per The rest are either unemployed or inactive, and their inci- capita expenditure indicates that the share of informal dence is larger at the bottom of the distribution. Among workers, particularly informal employees and unpaid family university graduates who were not employed in public workers, is considerably larger at the bottom of the distri- administration in 2015, about 46  percent had jobs else- bution (Figure 3.18, panel b). It cannot be ruled out that where, and the remaining 54 percent were either unemployed some of the youth who belong to less affluent households (39.6 percent) or inactive (13.9 percent) (Figure 3.19). Youth work in the private sector, while waiting to gain access to a with tertiary education predominantly belong to the middle job in public administration. and upper class, and, as a consequence, the distribution FIGURE 3.18. Profiles of Youth Ages 25–34 with Tertiary Education, Employed Outside Public Administration, by Type of Employment and Quintile of Household per Capita Expenditure, 2015 a. By type of employment, 2015 b. By quintile of household per capita expenditure, 2015 100.0 Unpaid family worker 1.8 80.0 Informal own-account worker 3.9 60.0 Percent Formal own-account worker 3.1 40.0 Informal employer 2.1 20.0 Formal employer 4.5 0.0 Informal employee 29.2 1 2 3 4 5 Formal employee 38.1 Quintile SOE employee Formal employee SOE employee 17.2 Informal employee Formal employer 0.0 10.0 20.0 30.0 40.0 50.0 Informal employer Formal own-account worker Percent Informal own-account worker Unpaid family worker Source: Based on data from the 2015 Household Budget Survey (HBS), INS. 148 Tunisia’s Jobs Landscape FIGURE 3.19. Share of Unemployed and Inactive Youth Ages 25–34 with Tertiary Education, by Quintile of Household per Capita Expenditure, 2015 80.0 70.0 14.4 12.8 60.0 16.3 50.0 13.9 16.8 Percent 40.0 11.5 30.0 57.7 58.9 47.9 20.0 39.6 39.1 28.1 10.0 0.0 Total 1 2 3 4 5 Quintile Unemployed Inactive Source: Based on data from the 2015 Household Budget Survey (HBS), INS. of unemployed and inactive youth is also skewed toward The pool of inactive youth is more evenly split between the middle and upper tail of the distribution. The rate of those who live with their parents (47.6 percent) and those unemployed or inactive youth, by quintile of the distribution who live with spouses (42.7 percent) (Figure 3.20, panel b). of household per capita expenditure, declined from about This is largely ascribable to the sex and age composition 72 percent in the lowest quintile to about 40 percent in the of the pool of inactive youth with tertiary education who highest quintile. comprise a larger share of women, particularly of women ages in the early 30s relative to the pool of unemployed University graduates who are unemployed or inactive live youth. The share of spouses peaks at the top of the distri- with their parents or, in the case of women, are married to a bution: 55.0 percent relative to 25.8 percent in the lowest working man. The majority of unemployed youth live with quintile, while the share of children shifts in the opposite their households of origin (about 76 percent), and the share direction from 70.7 percent to 37.5 percent (Figure 3.20, declines progressively from the bottom (88.7 percent) to the panel b). Furthermore, most unemployed youth complain top (71.0 percent) in favor of spouses (Figure 3.20, panel a). about a lack of jobss, whereas inactive youth report that FIGURE 3.20. Distribution of Unemployed and Inactive Youth Ages 25–34, with Tertiary Education, by Quintile of Household per Capita Expenditure and Relation to the Household Head, 2015 a. Unemployed youth, 2015 b. Inactive youth, 2015 100.0 5.7 100.0 3.5 5.0 8.8 9.4 9.6 11.2 9.0 9.7 11.2 7.5 90.0 90.0 16.2 80.0 80.0 70.0 70.0 37.5 47.6 40.2 60.0 60.0 70.7 50.2 Percent Percent 74.3 71.0 80.5 50.0 75.9 81.0 74.1 50.0 88.7 40.0 40.0 30.0 30.0 55.0 42.7 48.7 20.0 20.0 33.6 10.0 20.0 20.0 10.0 25.8 15.3 1.9 9.4 14.7 14.5 0.0 0.0 Total 1 2 3 4 5 Total 1 2 3 4 5 Quintle Quintle Spouse Son/Daughter Other Spouse Son/Daughter Other Source: Based on data from the 2015 Household Budget Survey (EMNVB), INS. Employment and Wage Outcomes 149 household duties are the main reason they are not engaged amnesty of 2011 and the exceptional provisions approved in the labor market. This may be linked to the fact that in 2011/12, which allowed for direct recruitment, as well women tend to take up most of the burden of house- as the regularization of contract and temporary workers hold responsibilities as they marry. The large majority of between 2011 and 2014 may have contributed to favor- both unemployed and inactive youth have never worked ing well-connected candidates over qualified candidates. before (more than 78  percent and 71  percent, respec- Data from the QoG Expert Survey indicate that there tively), and those who did work in the past only did so were no improvements in the meritocratic recruitment for short periods, typically up to six months during the indicator in Tunisia between 2013 and 2019.54 year preceding the survey. The limited creation of suitable jobs in the private sector, CONDITIONAL WAGE GAPS BETWEEN assigned gender roles, and a sizable public sector wage pre- FORMAL AND INFORMAL WORKERS mium are plausible causes of the high rate of nonemploy- IN THE PRIVATE SECTOR ment observed among university graduates. Available data In 2019, formal wage workers made, on average, 16 per- do not allow a definite answer to the question on why more cent more than informal wage workers per hour worked, than 1 university graduate ages 25–34 in 2 is looking for a and over 80 percent of the difference was explained by job or inactive. Nonetheless, the evidence illustrated above observable characteristics. The conditional formal-informal helps corroborate some hypotheses. First, the inactive youth hourly wage gap was estimated at 16.2 percent, on average, with a university degree are predominantly women in in 2019 (Figure 3.21, panel a). About 84 percent of the dif- their late 20s and early 30s, married to working men, in ference is ascribable to differences in observable character- the middle class, and living in affluent households. The istics (Figure 3.21, panel b). Formal wage workers are more inactivity status is likely attributable to the role women well endowed than informal wage workers with charac- have in the household as they get marry, together with the teristics positively correlated with wages. In particular, job limited creation of private sector jobs suitable for univer- characteristics (including place of work, type of contract, sity graduates. Second, unemployed graduates are largely and enterprise size), type of occupation, and educational young men living with their parents, which shows a con- level, together, explain over 90  percent of the explained siderable incidence at the bottom of the distribution. This component (Figure 3.21, panel c). The unexplained compo- may indicate that the main reason for the high unemploy- nent acts in the same direction as the observable character- ment rates are scarcity of suitable jobs in the private sector, istics, that is, in favor of formal wage workers. In addition, together with the expectation of gaining access to high- considering that informal workers do not pay income taxes paying jobs in the public sector. The greater incidence of on their wages, the unexplained wage gap may be even unemployment at the bottom of the distribution may be smaller in the upper half of the wage distribution. The per- explained by the employment of household members in sonal income tax system entails a zero tax area up to TD public administration. Estimates of a multinomial regres- 5,000 per year, which corresponds to about TD 417 per sion indicate that, after controlling for other character- month (slightly below the median wage of informal wage istics, the employment of household members ages 35 workers).55 or more in the public sector is positively correlated with the probability of university graduates working in public administration. The possible explanations of these effects may include both push and pull factors. In the first case, 54 The QoG Expert Survey is a survey of 1,294 public sector experts in   a civil servant in the household may provide more incen- 159 countries. The survey asks experts about the structure and behavior tives to participate in concours. In the second case, the pres- of public administration, such as hiring practices, politicization, profes- sionalization, and impartiality. An indicator gauges whether civil servants ence of a household member employed in the public sector are appointed and evaluated according to professional criteria. Tunisia’s may increase the chances of being hired in the public sector score has been constant at 50 over time. A 50 score is earned if any of the following conditions apply: (a) not all civil servants are appointed because thanks to a network effect. Based on qualitative interviews, of their merits, (b) not all appointees are free of conflicts of interest, and Brockmeyer, Khatrouch, and Raballand (2015) argue that, (c) performance evaluations are not always based on standard benchmarks. See QoG Expert Survey (Quality of Government Expert Survey), Quality before the 2011 revolution, the civil service recruitment of Government Institute, University of Gothenburg, Gothenburg, Sweden, system was functioning well, despite occasional interfer- https://www.gu.se/en/quality-government/qog-data/data-downloads/qog- expert-survey. ence in favor of candidates with support from the dominant 55 The following income bracket, between TD 5,000 and TD 20,000 per   political party; the system deteriorated thereafter. The general year, is taxed at a rate of 26 percent. 150 Tunisia’s Jobs Landscape FIGURE 3.21. Oaxaca-Blinder Decomposition: Mean Hourly Wage Differential Between Formal and Informal Wage Workers, Private Sector, 2019 a. Conditional mean hourly wage gap between b. Oaxaca-Blinder decomposition: explained and informal and formal wage workers unexplained component 0.0 90.0 83.6 80.0 –5.0 70.0 –10.0 60.0 Percent 50.0 Percent –15.0 40.0 –16.2 30.0 –20.0 16.2 20.0 –25.0 10.0 0.0 –30.0 explained unexplained c. Oaxaca-Blinder decomposition: breakdown of explained component 50.0 44.3 40.0 32.4 30.0 Percent 20.0 15.9 8.9 10.0 –1.5 0.0 –10.0 Demographics Education Industry Job characteristics Occupation Source: Based on data from the Labor Force Survey (ENPE), INS. RETURNS TO EDUCATION AND OTHER Returns to education, particularly tertiary education, are CORRELATES OF WAGES sizable and do not increase monotonically with the level of education. In 2019, workers with primary education enjoyed This subsection describes correlates of wages and docu- a premium of about 12.6 percent per hour worked relative ments the returns to education overall and separately in to workers with no schooling (Figure 3.22). Secondary edu- the public and private sectors and by sex.56 cation yielded an additional premium of about 9.1 percent (21.7 percent – 12.6 percent) relative to primary education, and tertiary education a further premium of 26.1  percent 56 The returns estimated in this subsection are a coarse measure of private   (47.8 percent – 21.7 percent) relative to secondary education. returns to schooling. Some of the categories lump together more than one certificate or diploma. In particular, tertiary education comprises multiple edu- This is consistent with existing evidence from developing cational attainments such as bachelor’s, master’s, and doctorate degrees. For countries (Psacharopoulos and Patrinos 2018), although the these reasons, the returns to one additional year of schooling at different levels of education are not provided. Instead, a coarse measure of the returns to dif- increment from secondary to tertiary education is quite large ferent levels of education is presented in this analysis. in the case of Tunisia and contrasts with the limited evidence Employment and Wage Outcomes 151 FIGURE 3.22. Returns to Education, Wage Workers Ages 15–64, 2012, 2015, and 2019 100.0 80.0 60.0 Percent 40.0 20.0 0.0 2012 2013 2014 2015 2016 2017 2018 2019 Primary Secondary Tertiary Source: Based on data from the Labor Force Survey (ENPE), INS. available from earlier studies on Tunisia in that returns to the explanation. Technological change allows workers per- education do not increase monotonically with the level of forming repetitive tasks to be replaced and can also create education (Limam and Ben Hafaiedh 2017; Zouari-Bouatour new functions that require a combination of technologies 1987; Zouari-Bouatour, Boudhraa, and Zouari 2014). and high-skill workers able to perform abstract tasks. In Tunisia, the average employee performs fewer nonroutine Between 2012 and 2019, the returns to primary educa- interpersonal and analytical tasks relative to the average tion increased, whereas the returns to tertiary education wage worker in Germany (World Bank 2021b). In 2017, the declined. An important observed trend is that returns to average routine task intensity was below the level observed tertiary education have declined over time. In 2012, the in 2000, in contrast with trends observed in developed returns to tertiary education relative to secondary educa- economies (Marouani and Minh 2021). tion were estimated at about 33  percent and gradually declined to 26 percent by 2019. The premium associated Relative to the private sector, the returns to tertiary edu- with secondary education on top of the premium of pri- cation are about three times as high in the public sector, mary education hovered around 9 percent over the period. where they have also expanded over time. Large differences By contrast, the returns to primary education increased over are estimated in the returns to education per hour worked time from 9.6 percent to 12.6 percent. This seems to be between the public and the private sectors (Figure 3.23). consistent with a growing supply of individuals of working The returns to primary education are not considerably dif- age with university degrees that is not matched by an equally ferent between the two sectors. In 2019, workers with pri- large growth in the demand for this type of educational mary education made, on average, about 11 percent more attainment, particularly in the private sector. Another than workers with no schooling, whereas, in the public explanation of the declining returns to tertiary education sector, the primary education premium was estimated at concerns changes in the composition of jobs across edu- about 13 percent. The returns to secondary education are cation groups. If the supply of university graduates rises considerably higher in the public sector. In 2019, a wage and demand does not follow, some graduates might look worker with secondary education in the public sector made for jobs that require a skill level below their qualifications about 27 percent more than a worker with primary educa- and thus contribute to a decline in the returns to tertiary tion per hour worked, which compares with a premium of education. Marouani and Minh (2021) find that the share about 5.4 percent in the private sector. The premium also of medium- and low-skill jobs performed by workers with rose over time in the public sector from 20.8 percent rela- tertiary education increased at the expense of high-skill tive to those with primary education in 2012. Similarly, the jobs. The hypothesis of skill-biased technological change returns to tertiary education in the public sector were about and changes in the task-content of occupations driven by twice as high relative to the private sector in 2012, and the the information technology revolution might contribute to premium increased over time and was estimated at about 152 Tunisia’s Jobs Landscape FIGURE 3.23. Returns to Education, by Sector and Sex, Wage Workers Ages 15–64, 2012–19 a. Public sector 120 100 80 Percent 60 40 20 0 Women Men Women Men Women Men 2012 2015 2019 Public sector Primary Secondary Tertiary b. Private sector 200 180 160 140 120 Percent 100 80 60 40 20 0 Women Men Women Men Women Men 2012 2015 2019 Public sector Primary Secondary Tertiary Source: Based on data from the Labor Force Survey (ENPE), INS. three times as high in 2019. Public sector wage workers The returns to education increase monotonically with the enjoyed a premium of about 46 percent relative to wage level of education in the public sector. The pattern identi- workers with secondary education per hour worked, which fied at the aggregate level whereby returns to education fol- compares with about 14.4  percent in the private sector. low a U-shaped curve does not hold in both the public and Such large differences in the returns to tertiary education private sectors. In the private sector, the pattern detected at likely contribute, together with job stability, social security the aggregate level is respected, and the benefit of secondary coverage, and other benefits, to the attractiveness of public education over primary is lower than the benefit of primary sector jobs among university graduates. over no schooling. By contrast, in the public sector, the Employment and Wage Outcomes 153 returns to education increase with the level of education, Second, a different pattern is detected between formal and which is consistent with the findings of previous studies on informal workers by sex. In the case of men, the returns Tunisia. In other words, the benefits of additional educa- to education are higher among formal workers relative to tion are higher at higher levels of education. informal ones. For example, in 2019, the returns to primary education relative to no education were about 14.0 per- The returns to education are higher among men relative cent per hour worked in the case of a man formal wage to women in the public sector. In the public sector, returns worker and 7.7  percent in the case of a man informal to education are higher for men relative to women and the worker. The difference is considerably smaller in the case gap increases monotonically with the educational level (Fig- of secondary education (0.9 percentage points in favor of ure 3.24). For example, in 2019, women with tertiary edu- formal workers) and widens at the level of tertiary educa- cation enjoyed a premium of 32.6 percent per hour worked tion (12.1 percentage points in favor of formal workers). relative to women with secondary education. In the case The opposite pattern, with higher returns to education of men, the corresponding premium is estimated at about among informal workers, is detected among women, with 58.0 percent, which means a gap of almost 26 percentage the exception of primary education. Also, the formal- points in favor of men. Similarly, women with secondary informal difference in returns is smaller in the case of women, education had a premium of about 18 percent relative to at about −0.5 percentage points at secondary education and women with primary education, whereas in the case of men at −4.8 percentage points at tertiary education. Third, the the premium was over 10  percentage points higher. The returns to education are higher among men than among gender difference in the wage premium associated with a women at any level of education, except among informal primary education relative to no education was estimated at workers with tertiary education. Among the latter, a gap of around 7 percentage points in 2019. Over time, the gender about 11.7 percentage points is estimated in favor of women. gap in the returns to education seems to have expanded. In the case of the private sector, the gender differences in the Other important correlates of hourly wages are industrial returns to education are in favor of women and have nar- sector, occupational level, type of contract, and geographical rowed over time (Figure 3.24). location. Figure 3.25 illustrates the estimated coefficients of a wage regression of the logarithm of hourly wage on a set of Women informal workers have higher returns to education individual and job characteristics. The industry of employ- relative to women formal workers. First, the U-shaped pat- ment has sizable effects on hourly wages after controlling for tern observed in the private sector holds both among for- human capital and a range of job characteristics. In 2019, mal and informal wage workers as of 2019 (Figure 3.24). for example, the interindustry wage differentials were as FIGURE 3.24. Returns to Education Among Formal and Informal Wage Workers in the private Sector, Ages 15–64, by Sex, 2019 50.0 45.0 40.0 35.0 30.0 Percent 25.0 20.0 15.0 10.0 5.0 0.0 Formal Informal Formal Informal Women Men Primary Secondary Tertiary Source: Based on data from the Labor Force Survey (ENPE), INS. 154 Tunisia’s Jobs Landscape FIGURE 3.25. Correlates of Hourly Wages, Wage Workers Ages 15–64, 2019 .18 .029 male –.036 age .12 age square/100 .2 Primary .39 Secondary –.093 Tertiary –.28 Public company –.24 Private company - Tunisian –.48 Private company - Foreign/Mixed –.42 Private premises –.29 Dwelling –.25 Hawker –.17 Farm land –.45 Building site .037 Other .25 International org/embassy –.18 Professionals –.33 Technicians –.41 Clerks –.41 Service and sales workers –.34 Skilled agricultural –.37 Craft workers –.43 Machine operators .23 Elementary occupations .21 Mining .1 Manufacturing .25 Utilities .21 Construction .23 Trade/repair .27 Transport/storage .18 Accommodation/food services .33 Info/comm .14 Fin/Ins/RealEst Prof/Scie/Tech act .19 .052 Admin/support serv act .2 PA/Def/SocSec .49 Education/Health/SocWrk .16 Household act –.0085 Other services .13 Fixed-term contract –.16 Open-ended contract –.23 CNSS –.24 Not af liated –.087 Other –.19 North-East –.013 North-West –.13 Center-East .046 Center-West .1 South-East .031 South-West Urban –0.5 0 0.5 Marginal effect Source: Based on data from the Labor Force Survey (ENPE), INS. Employment and Wage Outcomes 155 high as 39 percent in financial, insurance, and real estate have a large effect on hourly wages. For instance, wage services, 31.5 percent in accommodation and food services, workers employed in a private company made between 28.8 percent in construction, 25.7 percent in mining and 21 percent and 24 percent less per hour worked, depending quarrying, 25.7 percent in transportation, 23.6 percent in on whether the company was owned by Tunisians or was a trade, and 23.2 percent in manufacturing. Similarly, worker mixed foreign and local ownership company, than the same occupation contributed to large gaps in hourly wages. Rela- worker employed in public administration. An open-ended tive to managers, professionals and associate professionals contract increases hourly wages by 13.3 percent, compared made almost 29 percent more per hour worked in 2019, with no contract, and lack of access to social security also whereas workers employed in all other occupations gained contributes to the penalty. The marginal effects of geograph- between 16 percent and 35 percent less than managers per ical location on hourly wages are also considerable: living in working hour. For example, workers performing elementary urban areas increases hourly wages by about 3 percent, and occupations made about 35  percent less than managers, living in northern or central regions has a negative effect of machine operators about 31 percent less, and craftworkers between 1 percent and 17 percent on hourly wages relative about 29 percent less than managers. The place of work, to workers in Greater Tunis, whereas workers in southern including, for example, public administration, public com- regions benefit from a wage premium of between 4.7 percent panies, private companies, worker dwelling, and so on also (South-East) and 11 percent (South-West). REFERENCES CHAPTER 3 Boutar, L. 2018. “Performance in the Civil Service Incentive Structure: A Case Marouani, A.  A., and P. L. Minh. 2021. “Inequality and Occupational Study of Tunisia.” Wharton Research Scholars 163, https://repository. Change in Times of Revolution: The Tunisian Perspective.” Document upenn.edu/wharton_research_scholars/163. de travail 2021–06. Brockmeyer, A., M., Khatrouch, and G., Raballand. 2015. “Public Sector Size OECD (Organisation for Economic Co-operation and Development). and Performance Management: A Case-Study of Post-Revolution Tunisia.” 2018. “OECD Economic Surveys: Tunisia.” OECD, Paris. Policy Research Working Paper. 7159, World Bank, Washington, DC. Psacharopoulos, G., and H. A. Patrinos. 2018. “Returns to Investment IMF (International Monetary Fund). 2021. “Article IV Consultation: in Education: A Decennial Review of the Global Literature.” Policy Press Release; Staff Report; and Statement by the Executive Director Research Working Paper 8402, World Bank, Washington, DC. for Tunisia.” Country Report 2021/044 (February), Washington, DC. UN Women. 2017. Présence des femmes dans la fonction publique et accès https://www.imf.org/en/Publications/CR/Issues/2021/02/26/Tunisia- aux postes de décision en Tunisie.” December. 2020-Article-IV-Consultation-Press-Release-Staff-Report-and-Statement- World Bank. 2021a. “Household Production and Gender Roles in the Time by-the-50128. of COVID-19: Insight from a Rapid Online Survey in Tunisia.” Internal INS (Institute National de Statistiques). 2017. “Caractéristiques des agents draft, World Bank, Washington, DC. de la fonction publique et leurs salaires 2011–2015.” Edition 2017. World Bank. 2021b. Transforming Markets for More and Better Jobs in INS (Institute National de Statistiques). 2019. “Caractéristiques des agents MENA. Washington, DC: World Bank. de la fonction publique et leurs salaires 2013–2017.” Edition 2019. Zouari-Bouatour, S. 1987. “Capital Humain et Salaires: le Cas de la Tunisie.” Krueger, Alan B. 1988. “The Determinants of Queues for Federal Jobs.” Imprimerie Officielle de la République Tunisienne, Tunis, Tunisia. ILR Review 41 (4): 567–81. Zouari-Bouatour, S., L., Boudhraa, and S. Zouari. 2014. “Evolution of Limam, I., and A. B. Hafaiedh. 2017. “Education, Earnings, and Returns to Rates of Return to Schooling in Tunisia: 1980–1999.” Eurasian Jour- Schooling in Tunisia.” Economic Research Forum Working Paper 1162. nal of Social Sciences 2 (3): 28–47. 157 CHAPTER 4 Job Creation: Sectoral, Spatial, and Enterprise Transformation HIGHLIGHTS ◾ The process of structural transformation has continued slowly over the past decade. ◾ Construction, agrifood and mechanical and electrical goods manufacturing in the secondary sector, and real estate, trade, and transportation activities in services have spearheaded employment creation. ◾ Structural transformation has not been accompanied by considerable spatial transformation: economic activities and employment opportunities remain clustered in the coastal regions of Tunisia. ◾ The economic landscape is dominated by microfirms: firms with fewer than six employees contribute over 98 percent of firms, of which the majority are single-person firms, and almost 50 percent of total employment. ◾ Small firms create the most jobs thanks to considerable firm entry, but they are also more likely to exit the market because firm mobility is limited. ◾ The relationship between firm size and performance is weak, and average productivity does not increase with firm age and, in fact, decreases among older firms. ◾ The busines environment has deteriorated and has become less conducive to investment in human and physical capital and innovation, which are key to job creation. 158 Tunisia’s Jobs Landscape E nhancing productivity growth and job creation together with the best available technology to produce are challenges of paramount importance for many high–value added output. countries. Productivity is a key driver of growth and is cruxial to improving living standards through higher This chapter assesses the channels for job creation by earnings, particularly among the bottom 40, among whom focusing on two main transformations: (1) structural and earnings are the main source of income. Labor productivity spatial transformation, with an analysis of changes in sec- growth can be achieved in two main ways: (1) within eco- toral and spatial patterns of employment, and (2) the trans- nomic sectors, through capital accumulation, technological formation of the landscape of private firms, with attention change, or the improved allocation of resources across to changes in firm structure, productivity, and the business plants, and (2) through labor movement from sectors with environment. lower productivity to sectors with higher productivity. As economies develop, labor reallocation across sectors, that is, structural transformation, becomes less and less impor- Structural and Spatial tant in raising labor productivity. Fostering productivity growth is the principal channel and key engine.57 Transformation Over the decade from 2006 to 2017, the process of struc- The process of structural transformation entails a shift of tural transformation continued slowly. Structural trans- labor out of agricultural toward the secondary sector and formation proceeded at a pace slightly below the average services, and it is also accompanied by a process of spatial in other middle-income countries.58 In 2006, agricul- transformation because jobs in manufacturing and services ture accounted for 19.0  percent of total employment; are typically concentrated in and around urban areas. In the share had declined to 14.8 percent (−4.3 points) by some countries, this occurs with rapid urbanization and 2017 (Figure 4.1, panel b). The movement of labor away agglomeration, and, in others, the process involves a spatial from agriculture was concentrated in the period 2006–11, shift toward secondary cities. In Tunisia, the natural advan- when about 63,000 employed were lost in the sector. Labor tage provided by coastal access on the Mediterranean and moved toward secondary sectors as well as toward the vicinity to Europe, Tunisia’s largest export market, has led services sector, with an acceleration between 2011 and to a concentration of economic activities and population 2017. The services sector’s employment share increased by in coastal areas, especially the North-East and Center-East over 3 percentage points over 2006–17 and reached 51.7 per- regions, including Greater Tunis. According to the World cent in 2017. Secondary sectors posted a rise in their share Bank (2014), over 92 percent of industrial firms are within between 2006 and 2011 and a small decline thereafter. an hour’s drive of the cities of Tunis, Sfax, and Sousse. Overall, about 195,000 employed were added in secondary sectors, and over 334.000 in the services sector, of which The increase in productivity within sectors requires a trans- almost 70 percent occurred between 2011 and 2017 (Fig- formation at the enterprise level, entailing a shift toward ure 4.1, panel a). more sophisticated and higher–value added production of goods and services as well as a shift from informal to Construction, together with agrifood and mechanical and formal firms. Dynamic and high-growth enterprises are electrical goods manufacturing, was the driver of employ- important to the job creation and productivity growth ment growth in the secondary sectors. While textiles that can translate into widely shared improvements in continues to contribute the largest share to employment living standards through jobs of better quality. For this in manufacturing, its relative importance has declined. to happen, a conducive business environment is neces- and between 2006 and 2017, the sector shed about sary, whereby labor and capital are allocated to the most 25,000 workers (Figure 4.2, panel a). This is the long- productive firms and used in the most efficient manner, term consequence of the end of the Multi-Fiber Agreement in 1994, China’s joining the World Trade Organization, 57  Productivity gaps discussed in this study refer to average productivity. However, marginal productivity gaps matter and are expected to decrease as economies develop. Marginal labor productivity equals average pro- 58  The average change in agricultural employment in middle-income coun- ductivity, multiplied by the share of labor input under the assumption tries over 2006–17 was −6 percentage points, whereas the average change of a Cobb-Douglas production function. Comparisons of average labor in the secondary and tertiary sectors was +0.6 and +5.4 percentage points, productivity are therefore meaningful only in the absence of large differ- respectively (based on data of World Development Indicators, ILO employ- ences in labor shares across sectors. ment modelled estimates). Job Creation: Sectoral, Spatial, and Enterprise Transformation 159 FIGURE 4.1. Changes in Employment and Employment Shares, by Sector, 2006–17 a. Changes in employment b. Changes in employment shares 400 4.0 3.3 335 Change in employment shares 1.8 2.1 300 2.0 1.2 1.5 Change in employent (percentage points) 231 194 (thousands) 200 0.0 104 100 95 –0.4 100 –2.0 –1.5 –2.8 – –4.0 (0) –4.3 (100) –63 (63) –6.0 2006–11 2011–17 2006–17 2006–11 2011–17 2006–17 Agriculture Industry Services Agriculture Industry Services Source: Based on data from the Labor Force Survey (ENPE), INS. China’s consequent rise as a manufacturer of exports, and intensity (Ghali and Nabli 2020; Joumard, Dhaoui, and the pressures on low-technology exports of other coun- Morgavi 2018; Figure 4.3). Construction has created jobs tries. This raises questions about the jobs among low- at a rapid pace, and the employment share rose from 12 per- skilled women, who are disproportionately employed in cent to 14  percent between 2006 and 2017, with the the textile sector. The decline in textile employment has addition of about 20,000 workers per year. The pace of slowed since 2011, however (Figure 4.2, panel c). In par- employment creation in the construction sector is thought allel, employment in agrifood and mechanical and elec- to have lost steam since the revolution. trical goods manufacturing has increased, contributing to an additional 87,000 jobs, representing more than three In the services sector, employment creation was more rapid times the number of workers shed in the textile sector. in real estate activities, trade, and transport. Within the ser- Employment growth in new manufacturing sectors was vices sector, trade and public administration, together with driven by the diversification and sophistication of Tuni- health and education services, employ the largest share of sian exports, particularly the mechanical and electrical workers (see Figure 4.2, panel b). In 2017, the employ- products associated with a medium level of technological ment share of trade was estimated at about 13.2 percent, FIGURE 4.2. Employment Levels and Employment Growth, by Secondary and Tertiary Subsectors, 2006–17 a. Secondary subsectors: employment levels 600 476 442 400 355 Thousands 257 236 232 200 154 119 95 88 72 88 92 88 64 37 24 32 35 28 32 39 28 38 – 2006 2011 2017 Agri-food manufacturing Construction material, ceramics and glass manufacturing Mechanical and electrical manufacturing Chemical manufacturing Textile Other manufacturing Construction Other Secondary Sectors (continued) 160 Tunisia’s Jobs Landscape FIGURE 4.2.  Employment Levels and Employment Growth, by Secondary and Tertiary Subsectors, 2006–17 (continued) b. Tertiary subsectors: employment levels 800 658 558 587 600 457 Thousands 388 400 346 164 175 187 174 200 136 134 140 128 147 115 106 106 27 26 35 – 2006 2011 2017 Trade Transports Hotels and Restaurant Financial Services Real Estate/Professional Activities to Enterprises Public Administration/Health & Education Services Other Services c. Secondary subsectors: employment growth 80.0 62.3 60.0 44.4 40.0 33.9 Percent 25.8 27.2 29.0 24.3 13.5 17.5 19.0 16.2 16.3 20.0 11.1 7.7 5.5 0.6 0.2 – (0.4) (2.3) (1.1) (1.5) (5.0) (8.1) (9.5) (20.0) 2006–11 2011–17 2006–17 Agri-food manufacturing Construction material, ceramics and glass manufacturing Mechanical and electrical manufacturing Chemical manufacturing Textile Other manufacturing Construction Other Secondary Sectors d. Tertiary subsectors: employment growth 100.0 65.0 35.9 37.2 50.0 32.2 28.5 26.6 30.3 30.5 Percent 17.8 20.6 17.8 12.2 5.2 6.8 12.0 11.0 5.6 – (7.9) (4.0) (9.9) (14.7) (50.0) 2006–11 2011–17 2006–17 Trade Transports Hotels and Restaurant Financial Services Real Estate/Professional Activities to Enterprises Public Administration/Health & Education Services Other Services Source: Based on data from the Labor Force Survey (ENPE), INS. Job Creation: Sectoral, Spatial, and Enterprise Transformation 161 FIGURE 4.3. Sectoral Composition of Exports, Tunisia, 2006 and 2018 a. 2006 b. 2018 Source: Atlas of Economic Complexity (dashboard), Growth Lab, Center for International Development, Harvard University, Cambridge, MA, https://atlas.cid.harvard.edu/. and the share of employment in public administration and slow pace over the decade, some regions achieved greater health and education services was as high as 19 percent. progress than others. In particular, the North-West and However, employment rose at a more rapid rate in real Center-West regions, which started with a larger share of estate services, trade, and transport (Figure 4.2, panel d): agricultural employment in 2006, were able to progress a cumulative growth of 65 percent in real estate services more quickly than others in labor reallocation away from and above 34 percent in trade and transport between 2006 agriculture (Figure 4.4, panels a and b). In both regions, and 2017. Trade alone added over 18,000 workers a year, the increase in the share of nonagricultural employment second only to construction. was larger in the services sector compared with secondary sectors. Employment in services rose from 37.7 percent The structural transformation of the economy has pro- to 43.1 percent in the North-West and from 36.9 percent ceeded unevenly across regions. Although the process of to 45.5  percent in the Center-West region. As of 2017, structural transformation has continued, on average, at a Greater Tunis, the South-East, and the South-West had the FIGURE 4.4. Distribution of Region-Level Employment, by Sector, 2006 and 2017 a. 2006 b. 2017 Greater Tunis 3.1 30.3 66.5 Greater Tunis 3.1 28.6 68.3 North-East 20.9 39.1 40.0 North-East 18.7 40.8 40.5 North-West 43.9 18.4 37.7 North-West 34.6 22.3 43.1 Center-East 13.2 42.5 44.2 Center-East 10.8 42.8 46.3 Center-West 43.1 20.0 36.9 Center-West 29.7 24.8 45.5 South-East 11.7 30.8 57.5 South-East 9.5 33.6 56.9 South-West 19.5 25.4 55.1 South-West 24.7 23.8 51.6 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percent Percent Agriculture Industry Services Agriculture Industry Services Source: Based on data from the Labor Force Survey (ENPE), INS. 162 Tunisia’s Jobs Landscape largest share of employment in the services sector, whereas largely because of transportation costs and access to ports the North-West and Center-West had the largest share of for exports: 23.2 percent in Greater Tunis, 19.9 percent in agricultural employment despite the progress achieved the North-East, and 32.3  percent in the Center-East. By over the decade. A finer sectoral breakdown indicates that contrast, agricultural employment is concentrated in inland the composition of the services sector is somewhat different regions, particularly in the North-West and Center-West among the three areas with the largest share of services in (22.8 percent and 20.4 percent, respectively), although a 2017. In Greater Tunis, banking and insurance services, significant share is located in the North-East. real estate activities, and social and cultural services play a larger role relative to the southern regions. In the South- Over the decade, employment creation occurred dispro- West, the private sector contributes considerably less to portionately in the coastal regions. Between 2006 and employment in services, and the share of employment in 2017, Greater Tunis alone added over 220,000 employed public administration, education, and health services is individuals, almost 50  percent of all employment gen- significantly larger (28.9 percent) compared with the other erated in the country (Figure 4.6, panel a). The North- two areas (Greater Tunis and the South-East). East and Center-East follow with about 110,000 and 97,000 employed, respectively, contributing 24.4  per- Employment in the services sector is concentrated in Greater cent and 21.5 percent of the total employment created Tunis and in the coastal regions. Greater Tunis, including the between 2006 and 2017. The North-West stood out for governorates of Ariana, Ben Arous, Manouba, and Tunis, a negative contribution to employment growth, with a contributes more than 35  percent to total service sector loss in employment of over 36,000, whereas the rest of employment, followed by the Center-East at 22.4 percent the regions contributed between 2 percent and 6 percent and the North-East at 12.7 percent (Figure 4.5). Similarly, of the total employment added between 2006 and 2017. employment in industry is considerable in the coastal regions This is evident in the regional annualized growth rates in FIGURE 4.5. Share of Sectoral Employment, by Region, 2017 40 35.4 32.3 30 27.0 25.1 23.1 22.8 22.4 20.5 20.4 19.9 Percent 20 18.3 16.2 12.7 9.7 10.2 10 8.9 8.1 7.6 7.3 8.0 7.3 7.6 6.5 5.7 4.6 4.5 4.6 3.3 0 Greater Tunis North-East North-West Center-East Center-West South-East South-West Agriculture Industry Services Total Source: Based on data from the Labor Force Survey (ENPE), INS. Job Creation: Sectoral, Spatial, and Enterprise Transformation 163 FIGURE 4.6. Trends in Employment, Employment Shares, and Growth Rates, by Region, 2006–17 a. Trends in employment 1,000 933 868 816 800 776 770 710 Employment (Thousands) 600 561 498 450 400 373 361 336 342 330 352 251 222 225 200 158 137 134 0 Greater Tunis North-East North-West Center-East Center-West South-East South-West 2006 2011 2017 b. Employment share (2006) and annual employment growth, by region, 2006–17 25 3 2.5 2.0 20 2 Employment growth (%) Employment share (%) 1.2 15 1.1 1.1 1 10 0.3 0 5 –0.9 0 –1 s t t t t st t as es as es es ni a Tu E -E -E -W -W W h- er th h- er th er t or nt u ut at or nt So Ce So N re Ce N G Employment share (2006) Annual employment growth (2006–17) Source: Based on data from the Labor Force Survey (ENPE), INS. 164 Tunisia’s Jobs Landscape employment (Figure 4.6, panel b). Greater Tunis and the other regions, though workers in the North-East, Center- North-East posted employment growth rates of 2.5 per- East, and South-East have a greater chance of employment cent and 2 percent per year, respectively. The North-West as wage workers relative to their counterparts in other shed employment at an annual rate of 0.9 percent, and the regions. Working in agriculture, either as a wage worker Center-East created employment at a rate similar to that or self-employed, is more common in the western regions of the southern regions. as is the probability of being a contributing family worker. Private sector formal wage jobs are more likely to appear Wage employment, particularly public sector and formal in Greater Tunis; workers are more likely to be employed wage employment, is also clustered in coastal areas. High- in the public sector if they reside in the North-West, quality employment, including public sector and formal Center-West, or southern regions relative to Greater Tunis. wage employment, is concentrated in Greater Tunis and Working as a nonwage worker or an informal employee the coastal regions. The probability of employment in is typically more likely outside Greater Tunis (except for different types of jobs is thus strongly correlated with the North-East, North-West, and South-West in the case region of residence (Figure 4.7, panels a and b). The esti- of informal employees). mates control for worker characteristics and indicate that the likelihood of employment as a wage worker in non­ The concentration of employment in the coastal regions has agricultural sectors is higher in Greater Tunis relative to all helped shape the patterns of internal migration. According FIGURE 4.7. Effect of Geographical Location on the Probability of Working in Different Types of Job, Marginal Effects a. Types of employment: agricultural wage worker, nonagricultural wage worker, agricultural self-employed, nonagricultural self-employed, unpaid family worker, 2017 .082 North-East .027 North-West .0085 Center-East .024 Agricultural wage worker Center-West .021 South-East .081 South-West –.083 North-East –.15 North-West –.059 Non-agricultural wage worker Center-East –.14 Center-West –.058 South-East –.14 South-West .028 North-East .079 North-West .031 Agricultural self-employed Center-East .055 Center-West .017 South-East .085 South-West –.15 –.1 –.05 0 .05 .1 Marginal effect (continued) Job Creation: Sectoral, Spatial, and Enterprise Transformation 165 FIGURE 4.7. Effect of Geographical Location on the Probability of Working in Different Types of Job, Marginal Effects (continued) b. Types of employment: public employee, formal employee, informal employee, employer, own-account worker, unpaid family worker, 2015 North-East North-West Public employee Center-East Center-West South-East South-West North-East North-West Formal employee Center-East Center-West South-East South-West North-East North-West Informal employee Center-East Center-West South-East South-West North-East North-West Formal employer Center-East Center-West South-East South-West North-East North-West Informal employer Center-East Center-West South-East South-West North-East North-West Formal own-account Center-East Center-West South-East South-West North-East North-West Informal own-account Center-East Center-West South-East South-West North-East North-West Unpaid worker Center-East Center-West South-East South-West –.3 –.2 –.1 0 .1 .2 Marginal effect Source: Based on data from the Labor Force Survey (ENPE) 2017 and Household Budget Survey 2015, INS. to the World Bank (2021a), historical internal migration the North-East and Center-East (World Bank 2021a). The trends persist over time, and the coastal areas of Tunisia largest negative balance is detected in the Center-West, at absorb the largest share of migrants. The governorates about −44,400 between 2009 and 2014. located along the eastern maritime border, including Ariana (40,100), Ben Arous (26,600), Manouba (8,500), Medenine Similar geographical patterns paint the landscape of (2,700), Monastir (11,900), Nabeul (12,800), Sfax (9,600), firms; registered private sector firms are increasingly clus- and Sousse (19,200), were net receivers of immigrants over tered in eastern Tunisia.59 In 2003, Greater Tunis and 2009–14. Although the governorate of Tunis experienced a negative migration balance, more than 1 migrant in 2 living 59  The INS (National Statistics Institute) maintains a national business reg- in Tunis moved to neighboring governorates that are part istry (Répertoire National des Entreprises) that provides a list of all private sector businesses registered with the tax authority, together with the num- of Greater Tunis. As illustrated in Table 4.1, migrants move ber of wage workers employed in the businesses and registered with the from the western and southern regions to Greater Tunis and National Social Security Institute. 166 Tunisia’s Jobs Landscape TABLE 4.1. Trends in Migration Balances, by Region, year, on average. As a result, more than 15 years later, the 1989–2014 landscape was even more geographically clustered. Over 60  percent of registered firms were located in Greater 1989–1994 1999–2004 2009–2014 Tunis and the Center-East. The North-East was stable Greater Tunis 44,380 57,396 47,788 with a share of about 13 percent. All other regions posted North East −265 3,407 5,708 a decline in share. North West −33,332 −42,384 −38,,112 Center East 17,314 47,757 37496 Firm density confirms the geographical concentration of Center West −22,221 −53,965 −44,382 registered firms. Given differences in population size across South East −2,537 −2,126 −1,965 geographical areas, an indicator that captures the concen- South West −3,338 −10,085 −6,532 tration of firms relative to the resident population can pro- vide a better picture of the landscape of firms. The density Source: World Bank 2021a. of registered firms per 1,000 people confirms that northern and eastern Tunisia have the highest concentration of reg- istered firms, with the governorates of Tunis (179), Ariana the Center-East accounted for more than 55 percent of (120), Sousse (116), Ben Arous (111), and Sfax (110) leading all registered firms in Tunisia (Figure  4.8). The North- the ranking in 2019 (Map 4.1). By contrast, the North- East followed with a share of 13.2 percent, and all other West and Center West lagged, with an average of about regions contributed less than 10.0 percent each, with the 60 registered firms per 1,000 residents. smallest share in the South-West (4.7 percent). Between 2003 and 2019, the average annualized growth rate of Both microenterprises and larger registered firms are clus- the number of registered firms was about twice as high tered along Tunisia’s northeastern coastline; microenter- in Greater Tunis and the Center-East, compared with the prises are more prevalent in the inland regions. First, the southern regions (Figure  4.8). For example, in Greater panorama of firms is dominated by micro and small firms Tunis, the number of registered firms rose at a rate of (Map 4.2, panels a and b). The share of firms with fewer about 4.1 percent per year, on average, relative to about than six employees is above 96 percent in every gover- 2 percent in the North-West and South-West. The South- norate, with peaks of 99 percent. Second, the geographical East posted a considerable increase of 3.8  percent per distribution of registered firms does not show large dif- ferences between micro (fewer than six employees) and larger firms (six employees or more). Third, in the North- FIGURE 4.8. Distribution and Growth Rate of East and Center-East, the ratio between nonmicro firms Registered Firms, by Region, 2003–19 and microenterprises is higher than in the rest of the 40 3 country. In 2019, there were about 3.5 nonmicro firms 4.1 4.2 or every 100 microenterprises, whereas, in the rest of the Registered lm growth (%) 3.8 Registered lm share (%) 3.7 country, the ratio ranged between 1.1 and 1.7 per 100. 30 2 2.8 Patterns in the location of firms differ by sector. The 20 2.0 1 1.9 largest share of registered firms operate in the services sector across all regions (on average, a share of around 10 0 80  percent or higher), whereas manufacturing contrib- utes between 8 percent and 15 percent (Table 4.2). Within 0 –1 the services sector, there are differences across regions. In South-West Center-West South-East Greater Tunis North-West Center-East North-East particular, the North-East is home to more high–value added services, including information and communica- tion, finance and insurance activities, real estate, and professional, scientific, and technical service activities. By Registered lm share (2003) contrast, the Center-East has a comparative advantage in Annual registered lm growth (2003–19) manufacturing, and the remaining regions host traditional Source: Based on data from the National Business Registry (RNE), INS. sectors, including trade and transport activities. Job Creation: Sectoral, Spatial, and Enterprise Transformation 167 MAP 4.1. Density of Registered Firms (Number of Firms per 1,000 People), by Governorate, 2019 (99,179] (5) (83.5,99] (7) (68,83.5] (6) [55,68] (6) Bizerte 87 Ariana 120 Manouba Tunis Beja 91 179 Jendouba Nabeul 81 Ben Arous 98 70 111 Zaghouan 82 Le Kef Siliana 66 62 Sousse 116 Kairouan Monastir 66 99 Mahdia Kasserine 86 55 Sidi Bouzide 59 Sfax 110 Gafsa 56 Tozeur 79 Gabes 75 Kebili 79 Mednine 99 Tataouine 19.8 Source: Based on data from the National Business Registry (RNE), INS. 168 Tunisia’s Jobs Landscape MAP 4.2. Distribution of Registered Firms, by Size and Delegation, 2019 a. Firms with six or more employees, by governorate (1583,5511] (6) (308.5,1583] (6) (176.5,308.5] (6) [87,176.5] (6) Bizerte 728 Ariana 1778 Manouba Tunis Beja 436 5511 Nabeul Jendouba 229 Ben Arous 201 1743 1895 Zaghouan 302 Le Kef Siliana 131 104 Sousse 1992 Kairouan Monastir 287 1423 Mahdia Kasserine 403 198 Sidi Bouzide 164 Sfax 2905 Gafsa 189 Tozeur 160 Gabes 315 Kebili 141 Mednine 520 Tataouine 87 (continued) Job Creation: Sectoral, Spatial, and Enterprise Transformation 169 MAP 4.2. Distribution of Registered Firms, by Size and Delegation, 2019 (continued) b. Firms with fewer than six employees, by governorate (45037,135268] (6) (21743.5,45037](6) (12090,21743.5] (6) [5946,12090] (6) Bizerte 34303 Ariana 54233 Manouba Tunis 26444 135268 Nabeul Beja Jendouba 16779 Ben Arous 55736 18392 51261 Zaghouan 9912 Le Kef Siliana 11169 9297 Sousse 56119 Kairouan Monastir 25201 38813 Mahdia Kasserine 23947 15935 Sidi Bouzide 16443 Sfax 73089 Gafsa 13011 Tozeur 5946 Gabes 19540 Kebili 8995 Mednine 32437 Tataouine 8003 Source: Based on data from the National Business Registry (RNE), INS. 170 Tunisia’s Jobs Landscape TABLE 4.2. Distribution of Regional-Level Registered Firms, by Industry, 2019 North-East North-West Center-East Center-West South-East South-West Agriculture 0.6 1.0 0.6 1.1 0.5 1.1 Manufacturing 10.8 7.9 14.9 8.8 10.4 10.2 Food industry 1.8 2.1 2.1 2.3 2.3 2.6 Textile 2.0 1.4 3.4 1.4 1.8 2.4 Chemical and pharmaceutical industry 0.4 0.2 0.5 0.2 0.2 0.2 Computer, electronic, optical, and electrical 0.3 0.1 0.3 0.1 0.1 0.0 products manufacturing Other manufacturing 6.3 4.2 8.4 4.8 6.1 5.0 Construction 5.0 4.9 6.1 5.6 6.5 6.8 Trade 39.0 46.8 41.1 45.1 42.9 46.6 Transport and storage 11.5 18.4 13.6 21.4 16.4 13.2 Accommodation and food service 6.1 6.9 4.5 5.1 6.0 5.1 activities Information and communication 2.8 1.0 1.5 1.0 1.2 1.1 Financial and insurance activities 0.4 0.1 0.2 0.1 0.2 0.1 Real estate activities 1.0 0.3 0.4 0.1 0.6 0.1 Professional, scientific, and technical service 9.5 3.0 5.3 3.0 3.4 3.0 activities Administrative and support service 3.6 1.5 2.3 1.0 2.2 2.6 activities Education, health, and social services 4.5 3.0 4.6 3.5 4.4 4.8 Repair of computers and other personal and 0.9 0.8 0.9 0.7 0.9 0.7 household goods Other personal services 3.0 3.4 2.7 2.6 3.1 3.0 Other services 1.4 1.1 1.2 0.9 1.3 1.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source: Based on data from the National Business Registry (RNE), INS. The coastal-interior regional divide is reflected in sizable disparities in living standards that persist across regions, Enterprise Transformation despite considerable progress in poverty reduction. A snap- and Productivity shot of poverty outcomes at delegation level is provided in Map  4.3: higher poverty headcount ratios are esti- Microenterprises dominate the panorama of firms. Provid- mated in rural Tunisia and in inland areas of the country. ing a comprehensive overview of firms operating in Tunisia For example, in 2015, a poverty rate of 53.5 percent was is challenging.60 In 2019, over 780,000 private enterprises estimated in Hassi Ferid in the Center-West, followed by were registered with the tax authority, of which about Djedeliane (53.1 percent) and El Ayoun (50.1 percent) in the same region. Coastal areas had, on average, lower poverty headcount ratios: 11.9  percent in the North- 60  The national business registry is an excellent source of information about East and 11.7 percent in the Center-East, although a few the number of formal firms operating in Tunisia and the profile of the for- mal workforce. Yet, no information is available on the number of infor- pockets of poverty were present in rural areas of these mal production units, including small unincorporated firms and household regions. Similarly, the same regions had lower unemploy- businesses not registered with the tax authority. Every five years, the INS conducts a survey of microenterprises using the business registry as a sam- ment rates thanks to a higher concentration of economic pling frame. The survey covers production units registered with the tax activities. authority that employ fewer than six wage workers and that show an annual Job Creation: Sectoral, Spatial, and Enterprise Transformation 171 MAP 4.3. Poverty Headcount Ratios, by Delegation, 87 percent were single-person firms, that is, own-account 2015 workers, or production units with no formal employees. Overall, about 97 percent of the registered businesses had fewer than 6 formal employees; 2.3  percent were small firms (between 6 and 49 employees); 0.3  percent were medium-size firms (between 50 and 199 employees), and the remaining 0.1 percent were firms with 200 or more formal employees. Box 4.1 offers a profile of state-owned enterprises (SOEs). Over 2003–19, the distribution of registered firms did not change significantly, although an increase in the share of microenterprises from 96.5 per- cent in 2003 to 97.2 percent in 2019 was recorded. This is attributable to the rapid growth in the number of regis- tered self-employed from about 373,500 (85.2 percent) in 2003 to over 679,700 (86.9 percent) in 2019, which may partly arise because of the low cost of registration and penalties for noncompliance, in addition to the special tax BOX 4.1. State-Owned Enterprises State-owned enterprises (SOEs) are a defining feature of Middle East and North Africa economies, and Tunisia is no exception. SOEs are rooted in the set of companies inherited from colonial regimes and the policies that fol- lowed in the aftermath of independence (IMF 2021). According to the World Bank (2021c), 195 SOEs are recorded in official statistics, for total revenues in 2018 equivalent to about 20 percent of GDP and employing about 190,000 workers. SOEs operate in 40 of the 44 offi- cial sectors and subsectors, well above the average in other countries (between 22 in developed economies and 26 in developing countries). SOEs operate both in infrastructure and noninfrastructure sectors, and often 0.2 – 5.0 the government control is indirect. Most of the large 5.0 – 8.5 SOEs in Tunisia are highly indebted and deliver losses. 8.5 – 12.0 In 2018, 21 of the 31 largest SOEs recorded losses of 12.0 – 16.7 over TD 6 billion or 6  percent of GDP. The remaining 16.7 – 20.7 10 produced profits and contributed to 88 and 75 per- 20.7 – 25.9 25.9 – 33.4 cent of all SOE revenues and employment, respectively 33.4 – 52.7 (World Bank 2021c). Source: World Bank and INS 2020. In Tunisia, most SOEs operate in commercial sectors (17 of the 31 largest SOEs), although there is no eco- nomic rationale for state ownership in commercial sectors, such as, for example, manufacturing and construction because markets are contestable, and private businesses can provide goods or services more efficiently. In Tunisia, turnover below a specified threshold (TD 1 million in 2016). The survey commercial SOEs benefit from state support in the form provides information about the informal sector, which is defined as the set of of subsidies and therefore compete unfairly with private production units that do not keep formal accounts, but are registered with operators and are also protected from competition by the tax authority. In addition, the quarterly labor force survey provides infor- regulation as the sectors in which they operate have limits mation about the total number of workers in the country, including formal and informal employees and formal and informal household businesses and on foreign direct investment or price controls (World Bank own-account workers. Combining these sources of information allows a 2021c). more accurate snapshot of the distribution of firms by size. 172 Tunisia’s Jobs Landscape FIGURE 4-9. Trends in the Distribution of Registered FIGURE 4-10. Distribution of Wage and Overall Private Sector Firms, 2003–19 Employment, by Firm Size, 2019 100% 100% 7.6 7.0 99% 90% 98% 80% 27.0 24.9 Percent 97% 70% 96% 60% 22.3 Percent 95% 23.5 50% 94% 40% 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 30% 41.8 45.8 Large (200 and above) Medium (50–199) 20% Small (6–49) Micro (0–5) 10% Source: Based on data from the National Business Registry (RNE), INS. 0% Wage Employment Total Employment regime (régime forfaitaire) for microenterprises (Rijkers Micro (0–5) Small (6–49) Medium & Large (50 and above) Not stated et al. 2014).61 The share of small, medium, and large firms modestly declined because of the their less rapid growth in Source: Based on data from the labor force survey (ENPE), INS. numbers relative to microenterprises (Figure 4.9). In addi- tion, almost 520,000 informal own-account workers were estimated to be active in 2019 according to labor force More than 1 registered large firm in 2 operates in the survey data. This means that the distribution of firms is manufacturing sector. In 2019, among registered firms further skewed to the left, with over 98 percent of firms employing 100 or more employees, 56 percent were manu- falling in the micro category. No information is available facturing enterprises (Figure 4.11). At 23.7 percent, textile about firms operating without registering with the tax manufacturing was the sector with the greatest number authority. of large firms, followed by manufacturers of computer, electronic, and optical products (6.5 percent), and food Microenterprises contribute almost 50  percent of total manufacturers (6.3 percent). Outside the manufacturing employment In terms of employment, the contribution sector, enterprises operating in accommodation and food of firms of different sizes is considerably different from services, enterprises providing administrative and support the snapshot provided so far (Figure 4.10). First, although services, and enterprises in the trade sector contributed medium and large firms represent a small share of pro- 8.2  percent, 7.5  percent, and 6.9  percent of all regis- duction units, they accounted for about 25 percent of tered large enterprises, respectively. However, within employment (and 27  percent of wage employment) in manufacturing, microenterprises (firms with fewer than 2019. Second, small firms contributed about 22.3 percent 10 employees) made up the largest share (93.6 percent) to total employment (and 23.5 to wage employment), and of enterprises operating in the sector. Only in two sub- microenterprises contributed about 45.8 percent to total sectors, namely, automobile manufacturing and computer, employment (and 41.8  percent to wage employment).62 electronics, and optical products manufacturing, did regis- Overall about 1 worker in 2 is employed in firms with tered large firms contribute a considerable share (12.9 per- fewer than 10 workers. Such a pattern, whereby most cent and 10.4 percent, respectively) to the total number of firms are small and medium and large firms contribute firms operating in the subsector. a sizable share of employment, is not unique to Tunisia. About 1 private sector formal wage worker in 3 is employed in offshore firms, and almost 90 percent of offshore sector 61  Based on the 2016 microenterprises survey, about 65 percent of registered employees are employed in medium and large firms. Over nonagricultural microenterprises did not keep formal accounts. 62 The distribution of employment by firm size is based on self-reported   the decade, the share of offshore firms, that is, enter- information by respondents in the 2019 labor force survey. prises producing for the export market, increased from Job Creation: Sectoral, Spatial, and Enterprise Transformation 173 FIGURE 4-11. Distribution of Registered Firms with 100 Formal Wage Workers or More, by Sector, 2019 Computer, electronic and optical Other manufacturing, manufacturing, Construction, 3.6 Textile 6.5 manufacturing, 20.5 23.0 Trade, 6.9 Food manufacturing, 6.3 Agriculture, Forestry, and Transports and Fishing, 2.4 storage, 2.5 Other services, 9.1 Accomodation Administrative Education, health, and support service and food service social work activities, 7.5 activities, 8.2 activities, 3.5 Source: Based on data from the National Business Registry (RNE), INS. 2.9 percent in 2009 to 4 percent in 2019 (31,060 units). about 3.4  percent and 1.3  percent of the production The employment share of offshore firms also expanded, units in the offshore sector. Relative to the distribution of from 32.8 percent to 34.7 percent over the period, con- onshore firms, there is a larger share of offshore firms in tributing a total of about 397,200 formal employees in the medium and large size category. More than 86 percent 2019. Similar to onshore firms, offshore firms are predom- of formal employees in the offshore sector are employed inantly microenterprises. About 87  percent of offshore in large and medium firms, which compares with about firms are microenterprises, and 8.2 percent are small firms 54.7 percent in the onshore sector (Figure 4.12, panel b). (Figure 4.12, panel a). Medium and large firms represent With the objective of facilitating the integration of firms FIGURE 4.12. Distribution of Registered Firms and Formal Employment, by Regime (Onshore/Offshore) and Size of Firms, 2019 a. Distribution of rms b. Distribution of employment 100% 0.1 0.2 100% 1.3 2.1 98% 98% 3.4 35.3 96% 96% 94% 94% 60.1 92% 92% Percent Percent 8.2 19.4 90% 90% 97.6 88% 88% 86% 86% 27.7 26.3 84% 87.1 84% 82% 82% 17.6 12.1 80% 80% 1.5 Onshore Offshore Onshore Offshore Micro Small Medium Large Micro Small Medium Large Source: Based on data from the National Business Registry (RNE), INS. 174 Tunisia’s Jobs Landscape FIGURE 4-13. Change in the Contribution to Formal Wage Employment Creation, by Size Among Registered Firms, 2011–19 Large (200 and above) 2.4 Medium (50–200) –4.5 Small (6–49) 0.4 Micro (less than 6) 1.7 –5.0 –4.0 –3.0 –2.0 –1.0 0.0 1.0 2.0 3.0 Percentage points Source: Based on data from the National Business Registry (RNE), INS. operating for the domestic and the export market, the from recent data of the national business registry point in preferential tax regime for the offshore sector was elimi- the same direction: the share of entering firms is signifi- nated for newly established firms in January 2019 (World cantly larger among microenterprises (fewer than 6 formal Bank 2021c). In January 2021, the reform was extended employees) relative to small, medium, and large firms. For to all offshore firms that now pay between 10 and 25 per- example, in 2019, about 9 percent of firms with fewer than cent of corporate tax, with potentially negative transitory 6 employees entered the registry as a share of all firms of effects on the competitiveness and profit margins of exist- that size, which compares with 1 percent and 0.4 percent ing offshore firms (World Bank 2021c). among firms with 6–49 and firms with 50 or more formal employees, respectively (Figure 4.14, panel a). Recent trends indicate a deepening of structural patterns in firms and employment composition. Since 2011, the Firm mobility is limited, and small firms are more likely to relative gains in employment creation of registered enter- die. The majority of registered firms do not grow, even in prises have occurred at the tails of the distribution of firm the long run (a 14-year period, 1997–2010). For example, size. The employment share increased by 1.7 percentage fewer than 4 percent and only 2 percent of all firms with points among microenterprises and by 2.4  percentage between 10 and 49 employees in 1996 employed between points among large enterprises (Figure 4.13). By contrast, 50 and 99 or more than 100 workers by 2010, respec- enterprises in the middle of the size distribution posted tively (Rijkers et al. 2014). Transition matrices built using either a negative contribution to formal employment cre- data from the 2020 round of Enterprise Surveys confirm ation (−4.5 percentage points in the case of medium enter- this pattern.63 Between 2016 and 2019/20 virtually all prises) or a modest positive contribution in the case of medium and large firms did not grow, whereas 1.2 percent small enterprises (0.4 percentage points). of small firms managed to turn into large firms. Consider- ing the time period since the start of each firm’s operations, Small firms create the most jobs; yet, this is driven by firm only about 1 percent of small firms grew into large ones, entry, and most entrants are small. Previous work based though about 22  percent of initially small firms passed on business registry data covering the period 1997–2010 indicates that aggregate job creation was largely driven by firm entry (Rijkers et al. 2014). Virtually all net new 63 The matrices are constructed using recall data on size at the time of the   jobs were created by entering firms, and particularly by enterprise survey (2020), three fiscal years before (2016), and at the time the business was established. Firms that entered or exited within this time the entry of one-person firms, that is, own-account work- period cannot be accounted for. See Enterprise Surveys (dashboard), World ers. After entry, these firms post modest growth. Evidence Bank, Washington, DC, https://www.enterprisesurveys.org/. Job Creation: Sectoral, Spatial, and Enterprise Transformation 175 FIGURE 4.14. Share of Registered Firms Entering and Exiting, by Size and Year, 2003–19 a. Firm entry b. Firm exit 12.0 14.0 10.0 12.0 10.0 8.0 Percent 8.0 Percent 6.0 6.0 4.0 4.0 2.0 2.0 0.0 0.0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 less than 6 6–49 50 and above less than 6 6–49 50 and above Source: Based on data from the National Business Registry (RNE), INS. TABLE 4.3. Transition matrices of formal firms across employment size Size in 2019 Micro and small Medium Large (1–19) (20–99) (100 and more) Total Size at start Micro and small (1–19) 76.8 22.2 1.0 100.0 Medium (20–99) 7.5 59.9 32.6 100.0 Large (100 or more) 1.0 13.0 86.0 100.0 Size in 2016 Micro and small (1–19) 91.4 7.5 1.2 100.0 Medium (20–99) 2.0 96.2 1.9 100.0 Large (100 or more) 0.0 2.2 97.8 100.0 Source: Based on data from the Enterprise Survey, World Bank. the threshold to become medium-sized firms (Table 4.3).64 medium and large firms (Rijkers et  al. 2014). Similarly, Smaller firms are more likely to die. Between 7  percent over a long time period, almost 1 self-employed individ- and 8 percent of firms with fewer than six workers exit ual in 6 who registered in 1996 exited the market after the market after one year (on average over 1996–2010) 14 years compared with fewer than 20 percent of firm with compared with between 1.6  percent and 3.8  percent of 1,000 employees or more (Rijkers et al. 2014). Recent data confirm the findings: the share of firms exiting the market is considerably larger among firms with fewer than six 64  Firm size and age are correlated. Therefore separate transition matrices by workers relative to larger firms (see Figure 4.14, panel b). firm age in 2016 would provide superior information. However, the small sample size does not allow statistically meaningful transitions to be estimated by age-group. Small, young firms (between 1 and 9 years old in 2016) are The process of creative destruction is weak, suggesting more likely to become medium-sized firms (11.2 percent made the transi- tion between small and medium size in 2016–20) compared with middle-age the presence of distortions. Previous research using the (1 percent accomplish the same transition) and older firms (4.5 percent grow Tunisia national business registry indicates that allocative into medium-sized firms). Medium-sized firms in 2016 tended to maintain their size, and only a relatively few managed to grow larger (2 percent and efficiency, understood as the relationship between size and 3 percent of medium-size middle-age and older firms, respectively). performance, is low. In addition, average productivity does 176 Tunisia’s Jobs Landscape FIGURE 4.15. Correlation between Measures of Productivity and Firm Size, 2020 a. Sales per worker (log) vs. employment (log) b. Value added per worker (log) vs. employment (log) Lowess smoother Lowess smoother log(Sales per worker (in USD 2009)) 12 11 worker (in USD 2009)) log(Value added per 11 10 10 9 9 8 8 7 0 2 4 6 8 2 4 6 8 log(Number of workers) log(Number of workers) bandwidth = .8 bandwidth = .8 c. TFP (log) vs. employment (log) Lowess smoother 1.5 (based on VAKL model)) log(Estimate of TFPR 1.0 0.5 0 –0.5 2 4 6 8 log(Number of workers) bandwidth = .8 Source: Based on data of Enterprise Surveys (dashboard), World Bank, Washington, DC, https://www.enterprisesurveys.org/. Note: Observations with productivity below the 5th and above the 95th percentile are dropped from the sample. not increase rapidly with firm age and, in fact, declines among firms that have been established for more than four price variation can be decomposed into differences in input prices, differ- ences in market power, and differences in quality and other factors affecting years (Rijkers et al. 2014). According to the World Bank the demand for the product. What follows is grounded on revenue-based (2021a), there is little correlation between various proxies productivity and can potentially confuse increasing market concentra- tion with efficiency gains. TFP captures the portion of output that is not for firm performance and whether or not a firm exited explained by the amount of inputs utilized (see Francis and Karalashvili in six Middle East and North Africa countries, including 2021). Labor productivity is defined either as sales per worker or value Tunisia.65 Using recent data from Enterprise Surveys, the added per worker and indicates how efficiently labor is used in production. However, changes in labor productivity result from the combined effects correlation between size and different measures of produc- of multiple causes, including technological change and capital accumula- tivity is not linear (Figure 4.15, panel a-c).66 Productivity, tion, as well as the capacity of workers and the intensity of their efforts. It is therefore challenging to isolate the contribution of each variable. Large gaps in labor productivity across sectors might suggest that it would be 65  Using the panel component of the 2013 and 2020 rounds of Enterprise possible to achieve efficiency by reallocating workers to what appears to Surveys, exit rates are defined in two ways: whether the discontinuation of be sectors with higher productivity growth. To the extent that such differ- the firm was confirmed (conservative definition) or whether the firm was ences are attributable to firm rents, then the analysis would argue in favor unreachable (extended definition). of reallocating labor toward the more highly concentrated and distorted 66  The Enterprise Surveys collect data that allow labor, capital, and total fac- sectors of the economy as opposed to the most productive. Value added per tor productivity (TFP) to be calculated in a comparable manner across a worker and TFP are calculated only for manufacturing firms, whereas sales large number of countries. TFP estimates are revenue based, whereby sales per worker is a measure available for all firms in the survey. The relation- are measured in local currency as opposed to physical units, and therefore ship illustrated in Figure 4.15, panel a, does not change by restricting the production efficiency cannot be separated from the effects of prices. The sample to manufacturing firms. Job Creation: Sectoral, Spatial, and Enterprise Transformation 177 measured as sales per worker and value added per worker, correlations between different measures of producivity seems to be higher among large firms (400 workers and and firm age. Mature firms are more productive in terms above) relative to microenterprises and small and medium of sales per worker or value added per worker relative to firms. In the case of TFP, the nonlinear relationship dis- young firms. However, TFP measures seem to be higher plays an inverted U-shaped pattern, with TFP rising among young firms and modestly decline among older firms from microenterprises to small firms and then declining (Figure 4.16, panel c). A regression analysis confirms the modestly among firms with more than 50 workers. A relationship between productivity and age; geographical regression analysis (annex Table A 4.1) indicates the exis- location and export status do not appear to be significantly tence of a positive linear relationship between sales per correlated with productivity measures, whereas foreign- worker and size and a nonlinear relationship between owned firms are less productive relative to those owned by value added per worker or TFP and size, with measures of Tunisians (annex Table A 4.1). productivity rising with size at a decreasing rate. No productivity gap is detected between formal firms Productivity increases with firm age and is higher among managed by men and those managed by women. In line food manufacturers compared with other sectors (with with existing evidence on the Middle East and North the exception of TFP). Figure  4.16 illustrates bivariate Africa region, no statistically and economically significant FIGURE 4.16. Correlation between Measures of Productivity and Firm Age, 2020 a. Sales per worker (log) vs. age (log) b. Value added per worker (log) vs. age (log) Lowess smoother Lowess smoother log(Sales per worker (in USD 2009)) 12 11 worker (in USD 2009)) log(Value added per 11 10 10 9 9 8 8 7 1 2 3 4 5 1 2 3 4 5 log(Years of establishment) log(Years of establishment) bandwidth = .8 bandwidth = .8 c. TFP (log) vs. age (log) Lowess smoother 1.5 (based on VAKL model)) log(Estimate of TFPR 1.0 0.5 0 –0.5 1 2 3 4 log(Years of establishment) bandwidth = .8 Source: Based on data of Enterprise Surveys (dashboard), World Bank, Washington, DC, https://www.enterprisesurveys.org/. Note: Observations with productivity below the 5th and above the 95th percentile are dropped from the sample. 178 Tunisia’s Jobs Landscape FIGURE 4.17. Cumulative Distribution Functions of Sales per Worker and Value Added per Worker Over Time, 2013 and 2020 a. Sales per worker b. Value added per worker 1 1 CDF CDF .5 .5 0 0 0 50000 100000 150000 200000 250000 0 50000 100000 150000 Sales per worker (in USD 2009) Value added per worker (in USD 2009) 2013 2020 2013 2020 Source: Based on data of Enterprise Surveys (dashboard), World Bank, Washington, DC, https://www.enterprisesurveys.org/. Note: Cumulative distribution functions are truncated at the 5th and 95th percentiles. productivity gap is detected between formal firms man- TABLE 4.4. Annualized Growth Rate in Average aged by men and those managed by women in Tunisia, Productivity, by Type of Firm and Productivity using the 2020 round enterprise survey data (EBRD, EIB, Measure, 2013–20 and World Bank 2016; Islam et  al. 2020; World Bank Value 2021b). There was though a slight uptick in the man- added agement (from 8.5 percent to 10.4 percent) and owner- Sales per per ship (from 2.7 percent to 7.7 percent) of formal firms by Type of firm worker worker TFP women between 2013 and 2020. Overall −8.7 −8.4 −4.3 Industry Labor productivity decreased over time, particularly in the Food manufacturing −13.2 −20.5 −5.6 manufacturing sector.67 Between 2013 and 2020, average Textile manufacturing 1.7 12.4 −6.2 labor productivity, measured by value added per worker, Other manufacturing −14.7 −13.9 −1.5 dropped from $40,767 to $22,013 in 2020 (2009 prices), Construction −9.6 implying an annualized reduction of about 8.4 percent.68 Trade −7.7 The estimated reduction in average values is detected Hotels and restaurants −5.9 along the entire distribution of productivity. It widens around the middle of the distribution, and it peaks in the Other services −4.1 case of value added per worker (Figure 4.17). The decline Size is ascribable to a larger reduction in annual sales relative Micro and small (1–19) −8.2 −17.7 −4.6 to the costs of raw materials and intermediate goods and Medium (20–99) −8.2 −0.9 −3.5 employment levels. Except for textile manufacturing, firms Large (100 and more) −12.2 −17.9 −4.8 operating in manufacturing posted the largest reduction Age in productivity level as did firms employing 100 or more < 5 years −26.8 −33.1 −3.3 employees and firms established fewer than five years 6–14 years −8.8 2.6 0.1 before the survey (Table 4.4). 15+ years −7.3 −14.7 −5.5 Exporting status 67  Data derived from the 2013 and 2020 Enterprise Surveys may be affected by political uncertainty following the Jasmine revolution and the effect of Exporter −10.0 −6.6 −5.1 the COVID-19 outbreak and lockdown imposed in the country, respec- Nonexporter −7.9 −16.3 −2.8 tively. Therefore, statistics should be interpreted with caution. 68  Sales per worker declined on average from $77,127 in 2013 to $40,897 Source: Based on data of Enterprise Surveys (dashboard), World Bank, in 2019. Washington, DC, https://www.enterprisesurveys.org/. Job Creation: Sectoral, Spatial, and Enterprise Transformation 179 FIGURE 4.18. Share of Firms Reporting Various Business Environment Constraints as Major or Severe, 2013 and 2020 Competitors in informal sector Corruption Political instability Access to nance Transport Tax rates Inadequately educated workforce Customs and trade regulations Electricity Business licensing and permits Access to land Tax administrations Crime, theft and disorder Labor regulations Courts 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 2020 2013 Source: Based on data of Enterprise Surveys (dashboard), World Bank, Washington, DC, https://www. enterprisesurveys.org/. The COVID-19 pandemic has exacerbated the downward regulations (+23.6  percentage points), access to finance trend in productivity. According to the Business Pulse (+23.4 percentage points), and electricity (+23.3 percentage survey, over 80 percent of firms posted a reduction in annual points). The top three obstacles in 2020 were competition turnover in July 2020 relative to April of the same year.69 from the informal sector, corruption, and political instability. Older firms seem to have been more affected by the downturn More than 1 firm in 2 mentioned each of these issues as relative to younger firms, whereas chemical and pharmaceu- severe or major obstacles. tical firms posted an increase in sales. Reductions in cash- flows and difficulties in gaining access to credit are the main Fewer firms are investing in human and physical capital reasons reported by firms for the decline in business. The and in innovation. According to Kim and Loyaza (2019), largest cashflow reductions reported by firms were in accom- together with education, market efficiency, infrastructure, modation and food service activities, transport and storage and institutions, innovation is among the most important services, and mechanical and electrical goods manufacturing. determinants of TFP growth. The share of firms investing in physical capital over the previous fiscal year before the In addition, the business environment has deteriorated; an survey declined from 44 percent to 31 percent between 2013 increasing share of firms report major or severe obstacles and 2020; similarly the share of firms offering formal training in daily operations. Firms captured in the 2019 round of to workers contracted by 10 points and was estimated at the Enterprise Survey report, on average, a deterioration about 19 percent in 2020 (Figure 4.19). Ghali and Nabli of the business environment along all dimensions except (2020) document emerging pockets of innovation that have political instability, which is not surprising considering led to more sophisticated exports over time, particularly in that the previous survey round (2013) was conducted the mechanical, electrical, and pharmaceutical sectors. Yet, the share of firms investing in research and development in the aftermath of the 2011 revolution (Figure  4.18). declined from 18 percent to 7 percent between 2013 and Particularly striking is the increase in the share of firms 2020, as did the share of firms investing in new products or reporting the following as major or severe constraints: services (from 28 percent to 14 percent) or in new processes transport (+37 percentage points), competition from the (from 35 percent to 4 percent) (Figure 4.19). informal sector (+30 percentage points), business licensing and permits (+23.5 percentage points), customs and trade Cronyism and political connections remain a distinctive feature of the Tunisian private sector landscape. Politi- 69 COVID-19 Business Pulse Survey Dashboard, World Bank, Wash-   ington, DC, https://www.worldbank.org/en/data/interactive/2021/01/19/ cal connections undermine market contestability and fair covid-19-business-pulse-survey-dashboard. competition in a number of ways. For example, politically 180 Tunisia’s Jobs Landscape FIGURE 4.19. Share of Firms Investing in Human and Physical Capital and Innovating, 2013 and 2020 50.0 44.2 2013 2020 45.0 40.0 35.2 35.0 30.5 30.0 27.6 28.9 Percent 25.0 18.0 19.1 20.0 15.0 14.0 10.0 6.7 4.4 5.0 0.0 Investment in Introduction Investment in Investment in Offer of formal xed assets of new a new process R&D training product/service Source: Based on data of Enterprise Surveys (dashboard), World Bank, Washington, DC, https://www. enterprisesurveys.org/. connected firms are able to benefit from easier access to Africa region as well as in middle-income countries.70 The credit and to access sectors with barriers to entry or where World Bank (2021b) estimates that politically connected the existence of privileges can deter unconnected firms firms in the Middle East and North Africa region are more from entry. In Tunisia, politically connected firms are found likely than firms in other regions to be part of a business to have abused entry regulations for their own gain and to organization and to have access to external finance. be more likely to avoid import tariffs (Rijkers, Baghdadi, and Raballand 2017; Rijkers, Freund, and Nucifora 2017). 70 The 2020 round of the Tunisia Enterprise Survey included a question   About 28 percent of formal firms declare that they have that asks respondents the following: “Has the owner, CEO, top manager, or any of the board members of this firm ever been elected or appointed to a political connection in Tunisia, a figure considerably a political position in this country?” This is the strict definition of political higher than the average in the Middle East and North connections used in the analysis. REFERENCES CHAPTER 4 EBRD (European Bank for Reconstruction and Development), EIB (Euro- in Developing Countries? Evidence from Tunisia.” Labour Economics pean Investment Bank), and World Bank. 2016. What’s Holding Back 31 (December): 84–102. the Private Sector in MENA? Lessons from the Enterprise Survey. Rijkers, Bob Martinus Johannes, Leila Baghdadi, and Gaël Raballand. 2017. Washington, DC: World Bank. “Political Connections and Tariff Evasion: Evidence from Tunisia.” Francis, D. C., and N., Karalashvili. 2021. “Firm Level Productivity Esti- World Bank Economic Review 31 (2): 459–82. mates: Methodological Note.” April 26. Rijkers, Bob Martinus Johannes, Caroline L. Freund, and Antonio Maria Ghali, S., and M. K. Nabli. 2020. “Export Diversification and Sophisti- Nucifora. 2017. “All in the Family: State Capture in Tunisia.” Journal cation and Industrial Policy in Tunisia.” ERF Working Paper 1415, of Development Economics 124 (January): 41–59. Economic Research Forum, Giza, Egypt. World Bank. 2014. “Tunisia Urbanization Review: Reclaiming the Glory of IMF (International Monetary Fund). 2021. “State-Owned Enterprises in Carthage.” June, World Bank, Washington, DC. Middle-East, North Africa, and Central Asia: Size, Role, Performance, World Bank. 2021a. “Secondary Cities and Migrants: The Tunisian Case.” and Challenges.” DP/2021/19, IMF, Washington, DC. Background report for the study “Migrants, Markets and Mayors: Rising Islam, Asif, Isis Gaddis, Amparo Palacios-López, and Mohammad Amin. above the Employment Challenge in Africa’s Secondary Cities,” World 2020. “The Labor Productivity Gap between Formal Businesses Run Bank, Washington, DC. by Women and Men.” Feminist Economics 26 (4): 228–58. World Bank. 2021b. Transforming Markets for More and Better Jobs in Joumard, I., S. Dhaoui, and H. Morgavi. 2018. “Tunisia’s Inclusion in MENA. Washington, DC: World Bank. Global Value Chains and the Role of Offshore Companies.” OECD World Bank. 2021c. “Tunisia Country Private Sector Diagnostic. Moving Economics Department Working Paper 1478, Organisation for Eco- toward a More Competitive, Inclusive, and Resilient Private Sector.” nomic Co-operation and Development, Paris. World Banl, Washington, DC. Kim, Y. E., and N. V. Loayza. 2019. “Productivity Growth: Patterns and World Bank and INS (Institut National de la Statistique, National Institute Determinants across the World.” Policy Research Working Paper 8852, of Statistics). 2020. Carte de la pauvreté en Tunisie. February. Tunis: World Bank, Washington, DC. INS and World Bank. Rijkers, Bob Martinus Johannes, Hassen Arouri, Caroline L. Freund, and Antonio Maria Nucifora. 2014. “Which Firms Create the Most Jobs Job Creation: Sectoral, Spatial, and Enterprise Transformation 181 ANNEX CHAPTER 4 TABLE A 4.1. Estimates of Firm-Level Characteristics and Measures of Productivity, 2013 and 2020 2013 2020 2013 2020 2013 2020 Value added Sales per worker per worker TFP Firm sector Textile manufacturing −1.848*** −1.206*** −1.404*** −1.119*** 0.137 0.119 (0.210) (0.201) (0.226) (0.195) (0.136) (0.211) Other manufacturing −0.716*** −0.371** −0.580*** −0.319* 0.011 0.368* (0.173) (0.184) (0.183) (0.177) (0.115) (0.185) Construction −0.726*** −0.013 (0.256) (0.248) Trade −0.261 0.524*** (0.183) (0.174) Hotels and Restaurants −1.546*** −0.420* (0.220) (0.228) Other services −0.884*** −0.063 (0.190) (0.204) Firm size Employment (log of) 0.127 0.344* −0.256 0.621** −0.079 0.886*** (0.163) (0.190) (0.278) (0.288) (0.160) (0.329) Employment squared (log of) −0.018 −0.037 0.030 −0.081** 0.010 −0.097** (0.022) (0.025) (0.034) (0.037) (0.020) (0.040) Firm age 6–14 years −0.144 0.799*** −0.374* 1.260*** 0.024 −0.103 (0.152) (0.241) (0.224) (0.355) (0.136) (0.345) 15 years and over −0.188 0.988*** −0.415** 1.241*** 0.213* −0.019 (0.142) (0.232) (0.206) (0.345) (0.119) (0.327) Firm location Sfax 0.365*** −0.063 0.176 −0.389** −0.326** −0.024 (0.137) (0.105) (0.232) (0.174) (0.135) (0.177) North-East −0.259** −0.265* −0.386* −0.829*** −0.502*** 0.133 (0.118) (0.158) (0.228) (0.242) (0.134) (0.214) South Coast/South-West −0.086 −0.247 −0.451** −0.463* −0.306** −0.269 (0.126) (0.164) (0.223) (0.276) (0.132) (0.416) Interior −0.077 0.225 −0.170 (0.195) (0.332) (0.194) Foreign-owned firm −0.171 0.003 −0.271 −0.520*** 0.172 −0.468*** (0.135) (0.159) (0.206) (0.168) (0.118) (0.163) Exporting firm −0.013 −0.029 −0.232 −0.150 −0.076 0.289 (0.096) (0.124) (0.147) (0.148) (0.085) (0.176) Constant 11.771*** 8.651*** 12.474*** 8.810*** 1.147*** −0.783 (0.439) (0.558) (0.700) (0.802) (0.406) (0.898) Observations 544 483 239 197 228 84 R-squared 0.248 0.277 0.303 0.272 0.129 0.275 Note: Reference category: firms operating in food manufacturing; firms established less than 6 years before the survey time; firms located in Tunis; firms owned by Tunisians; not exporting firms. Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1