WORLD BANK MIDDLE EAST AND NORTH AFRICA REGION MENA ECONOMIC UPDATE APRIL 2020 How Transparency Can Help the Middle East and North Africa WORLD BANK MIDDLE EAST AND NORTH AFRICA REGION MENA ECONOMIC UPDATE APRIL 2020 How Transparency Can Help the Middle East and North Africa 2020 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 23 22 21 20 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved.. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/ by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Arezki, Rabah; Lederman, Daniel; Abou Harb, Amani; El-Mallakh, Nelly; Fan, Rachel Yuting; Islam, Asif; Nguyen, Ha; Zouaidi, Marwane. 2020. “How Transparency Can Help the Middle East and North Africa.” Middle East and North Africa Economic Update (April), Washington, DC: World Bank. Doi: 10.1596/978-1-4648-1561-4. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. ISBN (electronic): 978-1-4648-1561-4 DOI: 10.1596/978-1-4648-1561-4 Cover: Billion Photos / Shutterstock HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Table of Contents Acknowledgements  iv Abbreviationsv Prefacevi Chapter I: The Dual Shocks of the Novel Coronavirus and the Oil Price Collapse 4 I.1 The spread of Covid-19 4 I.2 The Collapse of Oil Prices 6 I.3 Toward a Sequencing of Policy Responses to the Dual Shocks 8 I.4 Quantifying the Effects of the Dual Shocks 9 I.5 MENA’s Chronic Low-Growth Syndrome 14 I.6 Enhancing MENA’s Transparency Can Accelerate Growth 15 Chapter II: External Imbalances, Fiscal Sustainability, and Data Transparency in MENA 20 II.1 Current Account Sustainability 20 II.2 Fiscal Sustainability: Lack of Transparency Obfuscates Existing Methods of Analysis 26 Chapter III: Data Gaps, Definitions, and the Measurement of Labor Market Outcomes 27 III.1 The Measurement of Unemployment in MENA 27 III.2 Female Labor-Force Participation: A Generational Issue 33 III.3 The Missing Piece: Measuring Informality in MENA 38 III.4 Conflict and Female Labor Force Participation  39 Chapter IV. Summary of Findings 41 References42 Appendix  44 Appendix A: Estimating the Relationship between Statistical Capacity and Economic Growth  44 Appendix B: MNACE’s Current Account Model  46 Appendix C: Fiscal Sustainability  50 TABLE OF CONTENTS i MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 List of Figures Overview: How Transparency Can Help the Middle East and North Africa Chapter I: The Dual Shocks of the Novel Coronavirus and the Oil Price Collapse Figure I.1 Negative Supply and Demand Shocks in MENA  6 Figure I.2 Global Oil Demand Forecasts for 2020 6 Figure I.3 Brent Oil Price and Futures 7 Figure I.4 Rough Calculations of the Income Effect of the Oil-Price Collapse across MENA Economies 7 Figure I.5 Correlates of the Costs of the Crisis: Growth Downgrades, Oil Export Exposure and Health Security 12 Figure I.6 Fluid Estimates of the Costs of the Crisis — Changes in World Bank Growth Forecasts  13 Figure I.7 MENA’s Chronic Low-Growth Syndrome 14 Figure I.8 Regional Development and Statistical Capacity 17 Figure I.9 Statistical Capacity Index across MENA 18 Chapter II: External Imbalances, Fiscal Sustainability, and Data Transparency in MENA Figure II.1 Unexplained Current Account Balances for MENA countries 22 Figure II.2 The Relationship between Primary Fiscal Balances and Past Debt – MENA and the Rest of the World since 1990  26 Chapter III: Data Gaps, Definitions, and the Measurement of Labor Market Outcomes Figure III.1 Unemployment Rates by Urban and Rural Locations 32 Figure III.2 Unemployment Rates by Education 32 Figure III.3 Unemployment Rates by Age Groups 33 Figure III.4 Female labor force participation rates  33 Figure III.5 Predicted Female Labor Force Participation in Egypt by Age Groups 34 Figure III.6 Predicted Female Labor Force Participation Rates by Age Groups (GMD)  34 Figure III.7 Women’s Educational Attainment by Age Cohorts  36 Figure III.8 Labor Force Participation Rates in the United States since 1890 39 Figure III.9 Labor Force Participation in Yemen, 1990-2019 39 Chapter IV. Summary of Findings ii LIST OF FIGURES HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA List of Tables Chapter I: The Dual Shocks of the Novel Coronavirus and the Oil Price Collapse Table I.1. Standard Deviation of Private-Sector Forecasts for 2020 GDP Growth across MENA Economies 9 Table I.2. Uncertain Forecasts: World Bank’s Growth, Current Account and Fiscal Balance Forecasts 10 Table I.3. Changing Estimates of the Costs of the Crisis: World Bank Growth Forecasts Relative to October 2019 11 Chapter II: External Imbalances, Fiscal Sustainability, and Data Transparency in MENA Table II.1. Primary and Structural Fiscal Balances versus Debt-Stabilizing Primary Fiscal Balances in MENA, 2018 and 2019  24 Table II.2 Debt Reporting in MENA Countries 26 Chapter III: Data Gaps, Definitions, and the Measurement of Labor Market Outcomes Table III.1. Definitions of Employment and Unemployment from the U.S. BLS, the French INSEE and the ILO 27 Table III.2. Consistency of Employment and Unemployment Definitions across MENA 28 Table III.3. Definitions of Employment and Unemployment 29 Table III.4. Unemployment Rates in Egypt in 2018  30 Table III.5. Unemployment Rates in Jordan in 2016 30 Table III.6. Unemployment Rates in Tunisia in 2014 31 Table III.7. Decomposing the Gap in FLFP Rates between Younger and Older Cohorts  37 Table III.8. Informal Employment in Egypt, Jordan and Tunisia  38 Appendix A: Estimating the Relationship between Statistical Capacity and Economic Growth  Table A1. Macro-economic Loss in GDP due to Statistical Capacity Index Decline in MENA (2005-2018) 45 Table A2. Definitions of the Statistical Capacity Measure 45 Appendix B: MNACE’s Current Account Model  Table B1. Summary Statistics  47 Table B2. MNACE Model Estimates of the Fundamental Drivers of Current Account Balances  49 Appendix C: Fiscal Sustainability  Table C1. The Relationship between Primary Balance and Debt  52 List of Boxes Chapter I: The Dual Shocks of the Novel Coronavirus and the Oil Price Collapse Box I.1. Transparency and the Statistical Capacity Index 16 Chapter III: Data Gaps, Definitions, and the Measurement of Labor Market Outcomes Box III.1. The World Bank’s Global Micro Database and household data for seven MENA countries 35 LIST OF TABLES AND BOXES MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Acknowledgements The Middle East and North Africa (MENA) Economic Update is a product of the Office of the Chief Economist for the Middle East and North Africa Region (MNACE) of the World Bank Group. The report was written by Rabah Arezki (Regional Chief Economist), Daniel Lederman (Deputy Chief Economist, Team Leader), Nelly El-Mallakh, Asif Mohammed Islam, Amani Abou Harb, Rachel Yuting Fan, Ha Minh Nguyen, and Marwane Zouaidi. The team received invaluable comments on preliminary results that appear in Chapters II and III from our World Bank MENA macroeconomists—including Kevin Carey and Eric Le Borgne (Practice Managers), Khaled Alhmoud, Sara B. Alnashar, Sona Varma, Bledi Celiku, Damir Cosic, Wissam Harake, Majid Kazemi, Dalia Al Kadi, Naoko C. Kojo, Wael Mansour, Ashwaq Natiq Maseeh, Khalid El Massnaoui, Mamadou Ndione, Harun Onder, Saadia Refaqat, Abdoulaye Sy, Hoda Youssef, and other participants at a technical workshop held on January 21, 2020 in Washington D.C. Subsequently the team received invaluable feedback from Amatalalim Al-Soswa (Sr Consultant, MNACE, World Bank), Robert Bou Jaoude (on behalf of World Bank staff from our Cairo Office), and particularly Johannes Hoogeveen (Practice Manager, EMNPV) who shared academic literature on the link between data transparency and development outcomes. The team is indebted to Najy Benhassine (MENA Regional Director) for his critique of the econometric models presented in this report; Anna Bjerde (MENA Director for Strategy and Operations) for encouraging us to clarify the link between transparency on the one hand and poor fiscal and labor-market outcomes on the other hand; and Ferid Belhaj (MENA, Vice President) for pushing the team to quantify the potential economic costs of lack of transparency. Our Country Directors, Jesko Hentschel, Kanthan Shankar, and Marina Wes demanded deeper and broader analyses of the role of transparency in development and World Bank operations. The authors also gratefully acknowledge the comments and constructive criticism received from other participants in a meeting of the World Bank’s MENA Regional Leadership Team held on March 11, 2020. We thank Swati Raychaudhuri for providing administrative support and James L. Rowe Jr for editing the manuscript. Help from Translation and Printing & Multimedia Unit from The World Bank’s Global Corporate Solutions is acknowledged. Last but not least, Nate Rawlings, Ashraf Al-Saeed, and Radhia Achouri provided editorial assistance on the Overview. Without their constant prodding the report’s messages would have ended up even more murky than they are. All remaining errors and omissions are the authors’ responsibility. iv ACKNOWLEDGEMENTS HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Abbreviations BLS United States Bureau of Labor Statistics CAPMAS Egypt’s Central Agency for Public Mobilization and Statistics CPI Consumer Price Index EAP East Asia and the Pacific ENCDM National Survey on Household Consumption and Expenditure FLFP Female Labor Force Participation GCC Gulf Cooperation Council GDP Gross Domestic Product GHS Global Health Security Index GMD Global Micro Database HIES Households Income and Expenditure Survey IEA International Energy Agency ILO International Labor Organization ILOSTAT International Labor Organization Statistics INSEE France-Institut Nationale de la statistique et des etudes economiques IMF International Monetary Fund MENA Middle East and North Africa MNACE Middle East and North Africa Chief Economist Office MPO Macro and Poverty Outlook OPEC Organization of the Petroleum Exporting Countries PPP Purchasing Power Parity SDG Sustainable Development Goals SOE State-owned enterprises SSA Sub-Saharan Africa UAE United Arab Emirates ABBREVIAITONS v MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Preface As the Coronavirus pandemic sweeps across the Middle East and North Africa (MENA), uncertainty and fear are gripping the streets. While citizens have turned to their governments to act, decades of lack of transparency has bred distrust and undermined our State credibility. People cannot be certain if daily reporting and updates are true. As someone aptly described the leadership response to the Coronavirus: “When you lose people's trust, even when you tell the truth, people won't believe you.” As if the spread of a global pandemic during a time of social unrest were not enough, more than any other region of the world, MENA is confronting two distinct but related shocks. Alongside the spread of the virus, oil prices collapsed, putting pressure on incomes and fiscal accounts of oil exporters, and indirectly but heavily affecting the developing economies of the region that rely on worker remittances, foreign direct investment, and transfers from their high-income neighbors of the Gulf Cooperation Council. The World Bank Group is committed to helping governments weather the dual shocks with the intention of leaving no one behind. We have put together a Covid-19 emergency financing facility of 14 billion dollars, and we are working tirelessly to ramp up our operations in a time of rising financing needs. On March 25, the World Bank Group committed 160 billion dollars when David Malpass presented our plans to our Board of Directors. This will finance support operations over the next 15 months tailored to the needs of each country, but with a strong poverty focus and an emphasis on policy-based financing and protecting the poorest households and the environment. In addition, on March 24 the World Bank and the International Monetary Fund asked for debt relief for the poorest countries, a plea that was repeated in President Malpass’s statement delivered to the Group of 20 on March 26. As we fight the spread of the novel virus across the world, policies designed to contain the spread and mitigate its impact on public health systems, such as the closing of large swaths of the economy, are clearly having at least short term recessionary consequences with potentially grave social costs beyond the deterioration of public health. We are committed to help by offering financing and technical expertise. Yet soon, together with our partners in MENA, we will come out of emergency mode. The question is whether we will come out stronger than ever, with a hopeful vision for a brighter future for MENA. To bring a new hope to our citizens, we must learn and change. After all, when the virus arrived in the region, and I mean all of it – its leaders, its entrepreneurs, its educated youth, the broader civil society – all of the region, was already engaged in difficult debates about the past and future development of their countries. All aspects of society seemed to be at stake, from the nature of political systems to technical aspects of social and macroeconomic policies. With this report, a product of our regional Chief Economist Office, we aim to contribute to an emerging constructive yet candid public discourse about what we can do better together after the immediate recessionary impacts of the shocks wither away. I cannot think of anything more important in terms of its scope and reach, than to begin immediately to discuss the transparency with which public sectors operate. After all, if there is a single lesson to draw from the pandemic, it is that transparency in the provision of public information can save lives and improve economic outcomes, partly by enhancing societal trust in the state. It is unfortunate that the region has under-performed for years if not decades in the transparency department. PREFACE HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA In fact, as shown in this report, since the beginning of the 21st century, growth of output per capita across MENA has been lower than what is typical for economies with the same levels of development. The authors argue that if the region had grown at the typical rate observed in the rest of the world, the region would be at least 20 percent richer than it is today. The lack of data and transparency in the region could be at least partly responsible for the region’s chronic low-growth syndrome. Indeed, as this report demonstrates, MENA stands out as the only region of the world to experience an absolute decline in their index of data transparency (the “statistical capacity index”) between 2005 and 2018. Many MENA countries have either lagged in their capacity to generate data or have prevented access to data. But reliable data and transparency not only help improve public policies over time but also enhance people’s trust in the state. Lack of transparency hurts even more when systems are under stress by potent threats such as the ongoing pandemic. The authors argue that the region’s declining data transparency has resulted in losses of income per person ranging between 7% and 14%. It is thus plausible that the lion’s share of income losses accumulated during the 21st century relative to the typical growth rates of the rest of the world were due to lack of transparency. Although there is no ironclad econometric model, and as a trained lawyer I cannot opine about such technical matters, the evidence in this report deserves serious attention. Lack of data and transparency hinders credible analyses of many important issues, including the performance of state- owned enterprises, public procurement, the allocation of precious assets such as land, the attraction of private foreign investment, and even obfuscate the maladies affecting the macroeconomies and labor markets of our countries. Since economic policies will only be as good as the information they are based on, logic dictates that lack of transparency in MENA deters effective policy making. I, for one, firmly believe that evidence-based policy debates can accelerate the pace of long-term economic development. This report makes a valiant effort to both raise issues of lack of transparency and to show how key pieces of information are missing in the regional policy dialogue. It is difficult to think of more important long-term challenges for the region than raising the pace of economic growth, solving fiscal vulnerabilities, and improving the performance of labor markets across MENA. Yet, the authors argue, in these areas we see either missing information or ambiguity in the published indicators. In short, this report sheds light in dark corners of crucial ongoing economic policy debates. The grievances that sparked protests across the region can only be addressed by rebuilding trust. The Coronavirus pandemic has put in stark relief what is at stake: Nothing less than human lives and prosperity. Now more than ever, a new social contract is needed, and the process of healing starts with transparency and accountability. The report makes a compelling case that transparency can, in turn, lead to growth and prosperity across MENA in the years and decades to come. I invite you to study this report and decide for yourselves if more sunlight is needed to bring a prosperous future to MENA with enhanced societal trust in the state. Ferid Belhaj Vice President Middle East and North Africa Region The World Bank Group PREFACE MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Overview: How Transparency Can Help the Middle East and North Africa The Middle East and North Africa (MENA) region is, once again, consumed by social unrest. The inability of many governments to deliver quality, affordable public services – from healthcare and education to water and electricity – erodes the possibility of improvement. Corruption and mismanagement are twin culprits, and the public perception that the state cannot improve things has created an urge to impose a new system of accountability. People are taking to the streets to express their frustration and demand change. In the words of one protester criticizing the ruling elites: “It’s been the same people for 30 years. The main point of this revolution is to do something for the poor—jobs, services, education.”1 Perhaps the most important word is “revolution.” The surge of societal frustration is not surprising. After all, MENA has struggled with low growth, macroeconomic fragility, and stagnant labor markets for decades. This report brings attention to one of the root causes of that frustration—the lack of transparency in the region, defined as the paucity of published data that meets the minimum accepted international standards regarding definitions of key economic and social indicators. The report provides technical analyses of the region’s notable lack of transparency and how it relates to the challenges of low growth, macroeconomic fragility, and stagnant labor markets. Lack of transparency hurts even more when systems are under stress by potent threats such as the ongoing Covid-19 pandemic. The ramifications of the lack of trust, forged by limited transparency, come into stark relief when citizens are confused about what to believe. And as we have seen several times over many decades, regaining credibility is not easy. As one citizen in the region aptly described the leadership response to Covid-19: “When you lose people's trust, even when you tell the truth, people won't believe you.”2 Losing credibility during a crisis such as a pandemic can be deadly. The lack of transparency across MENA has taken several forms—from a dearth of overall data, to a lack of accessibility and questionable accuracy of data that does exist. The situation has been allowed to fester and become deeply embedded in various institutions within the region. Now, it may have severe consequences. One area where we will likely see the effects of Covid-19 is economic growth, at least in the near future. Due to the dual shocks of the spread of the virus and lower oil prices, World Bank economists expect output of MENA to decline in 2020. This is in sharp contrast to the growth forecast of 2.6 percent published in October 2019. The growth downgrade of 3.7 percentage points is arguably a measure for the costs associated with the dual shocks of Covid-19 and the oil price collapse. Moreover, such estimates of the costs are highly uncertain and likely to change over the course of the year as new information comes to light. The report provides an analysis of recent growth forecasts by both private-sector and World Bank economists. Perhaps more importantly, since the beginning of the 21st century, growth of output per capita across MENA has been lower than typical growth for economies with the same levels of development. Had MENA’s growth of output per capita 1 Reuters, 2019. “Lebanon a 'beautiful idea' in need of a reboot, say protesters” (November 7, 2019). 2 https://www.pbs.org/newshour/show/in-iran-government-distrust-rises-amid-deadly-outbreak-of-novel-coronavirus 1 OVERVIEW HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA been the same as that of a typical (median) peer economy over the past two decades, the region’s real output per capita would be at least 20 percent higher than what it is today. And this benchmark is, by definition, mediocre. A large part of MENA’s low growth is arguably due to a lack of data and transparency. Many MENA countries have either lagged in their capacity to generate data or have prevented access to data, both of which lead to poor policies. Just as lack of data about the spread of a virus hampers public policy and societal responses, lack of data and imprecise indicators of public debt and unemployment hamper policy actions to deal with these long-standing development challenges. This report argues that reliable data and transparency can help improve public policies over time and enhance people’s trust in the state. In fact, since 2005, the report finds a strong empirical association between statistical capacity – including the regular publication of microeconomic and macroeconomic information – and economic growth. That effect appears to be at least as large (if not larger) than the empirical association between education and growth. Moreover, the evidence from five models suggests that an observed decline in MENA’s transparency between 2005 and 2018 is associated with an expected loss of the region’s income per capita ranging from 7 to 14%. MENA’s macroeconomic fragility has also come to the forefront during recent years. The report presents a battery of tests on current account and fiscal account vulnerabilities in the region and the rest of the world; however, the credibility of these analyses critically depends on data transparency. MENA countries generally do not report net public debt, a crucial metric to assess debt sustainability. Even for gross public debt, MENA countries vary greatly in their reporting standards. World Bank economists and other external analysts do not have access to vital information about many types of public debt. Thus, any assessment of the sustainability of the region’s public debt needs to be interpreted with a grain of salt. With that in mind, considering the lack of transparency in debt indicators, the analyses presented in this report suggest that MENA countries continue to face notable macroeconomic challenges. First, three developing MENA economies appear to have unsustainable current account deficits, which in turn are due to low GDP per working age population. In other words, low growth has brought external macroeconomic fragility to some MENA economies. Second, in 2019, 11 MENA countries appeared to be on unsustainable fiscal paths—that is, the primary fiscal balances were insufficient to stabilize their reported gross debt-to-GDP ratios. On the other hand, we did find some encouraging evidence (also imperfect, due to data constraints) that suggests that MENA’s developing countries, as a group, have been on a more sustainable fiscal path than the rest of the world, on average, in recent years. Either way, for good or for bad, the lack of transparency of debt indicators in MENA limits our ability to reach firm conclusions about the region’s fiscal vulnerabilities. The labor market is another area where MENA faces both notable challenges and constraints imposed by lack of transparency in the form of imprecise indicators of labor-market outcomes. Among the challenges are persistently high reported unemployment rates and low rates of female labor force participation. These issues are related to countries’ unsustainable current account imbalances and fiscal paths because the low employment of the working age population is associated with both external imbalances and inadequate revenue bases. One of the chief constraints regarding labor markets is the lack of uniformity in the definition of "employment". A review of MENA’s standards and definitions of labor-market outcomes shows that countries rely on varying and often uncertain definitions of employment, which in turn affect indicators of unemployment and informality – with little harmonization – either across the region, or with respect to international standards. In fact, for countries with independent sources of OVERVIEW 2 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 nationally representative labor-force data for recent years (Egypt. 2018; Jordan, 2016; and Tunisia, 2014), the authors were not able to replicate the official reported unemployment rates. This report argues that the differences are not innocuous because the discrepancies disproportionately distort the role of women and rural areas in national labor markets. It suffices to say that it would be desirable to have access to all countries’ official labor force surveys in order to have an informed dialogue about definitions and methods. In turn, the report assesses the key challenge of raising MENA’s female labor force participation, which is critical to increasing GDP per working age people. The evidence (relying on internationally accepted definitions of employment and unemployment) suggests that female labor force participation might be a generational issue, as young women tend to have high rates of labor market participation. A large part of the difference is due to education. Because younger women are increasingly more educated than previous generations, they are more likely to join the labor force. In our analyses, education explains between 5 and 12 percentage points of the difference in labor force participation across generations, while we found that family structure (such as marriage and children) is less important. These findings are good news for MENA’s future. It is noteworthy that these findings are not due to imprecise indicators, since the analyses rely on internationally accepted and harmonized definitions of female employment and unemployment. The question is whether the region can afford to wait for the national rate of female labor force participation to rise with the advent of new generations of educated young women. Likewise, historical evidence from an advanced economy and recent evidence from Yemen indicate that female labor force participation tends to rise during periods of armed conflict, when fewer men are in the labor force. It is plausible that the increase in female labor force participation might be a thin silver lining in conflict economies; however, it remains to be seen whether female labor force participation will remain high after peace arrives, as it did in the United States after World War II. The grievances that sparked protests across the region can only be addressed by rebuilding trust. A new social contract starts with transparency and accountability, which could, in turn, lead to growth, more robust analyses of fiscal sustainability, and improved policies for tackling the macroeconomic and employment challenges that have been allowed to fester for years—if not decades. Through transparency, prosperity with societal trust can be achieved across MENA in the years and decades to come. 3 OVERVIEW HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA CHAPTER I: THE DUAL SHOCKS OF THE NOVEL CORONAVIRUS AND THE OIL PRICE COLLAPSE Chapter I takeaways: • Economies in the Middle East and North Africa face unprecedented dual shocks from the spread of the novel coronavirus (Covid-19) and the collapse of oil prices. • The spread of Covid-19 causes severe negative demand and supply shocks. • The fall in oil prices depresses income in the MENA region. It is directly felt by oil exporters and indirectly by oil importers through reduced remittances, foreign investment, and sovereign lending. • A preliminary estimate of the crisis’ economic costs in 2020 is 3.7% of regional output. • The estimate remains tentative because the costs will depend on both policy and societal responses to the dual shocks. • But MENA has suffered from a low-growth syndrome for decades when compared to the rest of the world in terms of GDP per capita growth rates. • The region’s low long-term growth rates are partly attributable to the region’s declining data transparency. Economies in the Middle East and North Africa (MENA) face unprecedented dual shocks from the spread of Covid-19 and the collapse in oil prices. As the world struggles with an emerging recession, MENA was already facing social discontent manifested in street protests. The dual shocks bring negative supply, demand and income shocks, further aggravating pre-existing economic and social challenges. I.1 The spread of Covid-19 Chinese authorities first alerted the World Health Organization (WHO) of a new coronavirus infection on December 31, 2019. This new virus, Covid-19, can produce flu-like symptoms that sometimes are more severe with a higher chance of death than the flu. Besides the human toll, the virus affects MENA economies via four channels: the deterioration of public health, falling global demand for the region’s goods and services, declines in MENA’s domestic supply and demand, and importantly, falling oil prices. Ì The Deterioration of Health The virus has spread to more than 177 countries and territories—with over 360,000 cases and more than 15,000 deaths as of March 23, 20203 . The pandemic has caused severe disruptions of economic activity across the globe. The virus has already spread to MENA countries, of which Iran, has been hardest hit with more than 23,000 reported cases and 1,400 deaths as of March 23. Other MENA countries have also reported infections. Governments are closing schools and imposing restrictions on businesses and public gatherings. These efforts, while believed necessary to slow the spread 3 https://www.worldometers.info/coronavirus/ CHAPTER I 4 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 of the virus—especially because of limited real-time information on its magnitude and spread—will undoubtedly hurt economic activity. The ability to contain the virus depends on the strength of the public health systems of the MENA countries and their health policy responses. Public health policy responses, in turn, depend on the number of infections and where they are spreading. The outbreak can be in clusters, with some places becoming hot spots. Transparency of data and sharing of information are vital as governments and citizens must work together to change social behaviours to flatten the infection curve. Otherwise public health systems could be overwhelmed. Consequently, societies subject to opaque information- sharing by the state are also among the most vulnerable to a pandemic. This is evident in the relative success of South Korea in containing the virus. As South Korea’s Foreign Minister Kang Kyung-wha put it in an interview: “The basic principle is openness, transparency, and fully keeping the public informed” (BBC, 2020). The MENA region is particularly vulnerable. MENA scores second-lowest among all regions in the overall Global Health Security Index4, while ranking last in both “epidemiology workforce” and “emergency preparedness and response planning”. The situation is more dire in conflict-affected environments. Wars in Syria and Yemen will almost certainly impede the proper functioning of their health systems. Ì Falling Global Demand Global economic difficulties and the disruption of global value chains will reduce demand for the region’s goods and services, most notably oil and tourism. The implications of collapsing oil prices are discussed in section I.2. The tourism sector in MENA is affected in two ways. First, many MENA countries and many other nations are imposing travel restrictions. Second, the global economic slowdown and social distancing efforts imply fewer tourists travelling to other countries, including within the MENA region. Moreover, the sharp drop of global travel will further depress already tanking oil prices. Ì Declines in MENA’s Domestic Supply and Demand The spread of Covid-19 also generates a negative demand shock from the abrupt reduction in regional business activities and travel due to concerns about the spread of the virus. In addition, uncertainty associated with the spread of the virus and aggregate demand could further dampen the region’s investment and consumption. Collapsing oil prices further depress demand in MENA; the oil and gas sector is the most important one in many of the region’s economies. Finally, potential financial market volatility could have real effects and further disrupt aggregate demand in the region. Similar effects were observed in the 1918 Spanish Flu pandemic in the United States. Correia and others (2020) find that the Spanish Flu pandemic also created both negative supply and demand shocks. Localities in the United States that had higher mortality rates also had larger declines in manufacturing activity, bank assets and spending on durable goods. Perhaps more importantly from a policy perspective, the evidence from 1918 indicates that localities that implemented 4 https://www.ghsindex.org/ 5 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Figure I.1 Negative Supply and Demand Shocks in MENA tougher containment and mitigation policies such as social P distancing and limitations on social interactions were S1 also the localities that experienced the fastest economic S0 recoveries afterwards. The related issue of the sequencing of policy responses is explored further in section I.3. The negative demand and supply shocks are illustrated in D1 D0 Figure I.1. Original demand and supply curves are D0 and S0 and the original output is Q0. Because of the spread of Q1 Q0 Q the virus to the region, both curves shift left, generating a new equilibrium output Q1, which implies a reduction of output and income even though the net impact on price inflation is ambiguous (Figure I.1 is drawn to reflect a scenario in which prices fall). The extent of the declines in supply and demand is uncertain. The drop-offs depend on the disclosure of information about the spread of the virus and the health policy responses by the countries. Consequently, any forecasts or estimates of future economic activity must be treated with great caution and are likely to change as new information comes to light. I.2 The Collapse of Oil Prices The oil price collapse exacerbates the impact of Covid-19 and brings severe negative income shocks to many MENA economies. After tallies of death rates in China Figure I.2 Global Oil Demand Forecasts for 2020 were reported, oil prices declined sharply. Because 1400 KBD of China’s increasingly important role in global 1200 commodity markets, any setbacks to its economy are 1000 800 expected to significantly reduce global demand. In 600 addition, the global fear and uncertainty regarding 400 the spread of virus is likely to hurt investment 200 decisions in China and in other countries, further 0 weighing on demand prospects and lowering oil -200 Aug-19 Sep-19 Oct-19 Nov-19 Dec-19 Jan-20 Feb-20 Mar-20 prices. The March 2020 report from the International Source: International Energy Agency Energy Agency (IEA) projected that in 2020 global Note: Horizontal axis indicates different vintages of IEA’s 2020 global demand forecasts. oil demand growth would fall for the first time since 2009 (see Figure I.2). The IEA in March forecasts a 0.09 million barrel per day (mbd) decline daily demand in 2020, 1.1 mbd lower than its forecast a month earlier. In response to falling demand for oil, the Organization of the Petroleum Exporting Countries (OPEC) on March 5 proposed a 1.5 mb/d production cut for the second quarter of 2020—of which 1 mb/d would come from OPEC countries and 0.5 mb/d from non-OPEC but aligned producers, most prominently Russia. The following day, Russia rejected the proposal, prompting Saudi Arabia—the world’s largest oil exporter—to boost production to 12.3 mb/d, its full capacity. Saudi Arabia also announced unprecedented discounts of almost 20 percent in key markets (Arezki and Fan, 2020). The boost in supply, coupled with falling demand, resulted in a collapse in oil prices, to as low as $31.1 a barrel on the following CHAPTER I 6 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Figure I.3 Brent Oil Price and Futures (U.S. dollars a barrel; expiration dates on x-axis) 90 Spot Brent oil price 9/25/19 80 1/21/20 3/9/20 70 3/18/20 60 50 40 30 Dec 2022 20 Jan-2018 Jul-2018 Jan-2019 Jul-2019 Jan-2020 Jul-2020 Jan-2021 Jul-2021 Jan-2022 Jul-2022 Source: Bloomberg, L.P. Note: The black line indicates the spot price of Brent crude oil. The colored lines illustrate the futures prices of Brent crude oil on, respectively, September 25, 2019; January 21, 2020; March 9, 2020, after the disintegration of the OPEC+ alliance; and March 18, 2020. Monday, March 9, and to about $25 a barrel as of March 18 (see Figure I.3). The upward pointing futures curve suggests the market still expects oil prices to recover, but slowly—reaching about $43 per barrel by the end of 2022. But, of course, such forecasts are uncertain. The oil price collapse produced a massive Figure I.4 Rough Calculations of the Income Effect of the Oil-Price Collapse direct negative income effect for MENA across MENA Economies oil exporters. It also hurt oil importers, even though as a general rule, lower 10% prices are good for oil-importing economies and bad for oil exporters. 5% A simple way to get a sense of the size 0% of the real income effect is to multiply the difference between production and -5% consumption (net oil exports) as a share -10% of GDP by the percentage point decrease in the oil price (see Figure I.4). Based -15% on a hypothetical scenario in which oil -20% prices stay 48 percent below their 2019 -25% level, Kuwait, whose net oil exports Iran, Islamic Rep. Lebanon Kuwait Algeria Bahrain Djibouti Tunisia Saudi Arabia United Arab Emirates Egypt, Arab Rep. Qatar Iraq Oman Yemen Morocco Jordan account for 43 percent of GDP, would experience a decline in real income of about 20 percent of GDP, while oil- importer Morocco would experience an increase in real income equivalent to 3 percent of GDP. Source: Authors’ calculations based on data from the World Economic Outlook database. Note: Oil prices are from March 13, 2020 with a forecast of $33.4 per barrel for 2020, 48% lower than the price of 2019 ($64 per barrel). However, in MENA it is likely that lower oil prices will also indirectly hurt oil 7 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA importers through reduced foreign direct investment, remittances, and grants from the region high-income exporters. The economic fortunes of MENA’s oil importers and exporters are connected. The two shocks of Covid-19 and oil price collapse are intertwined, yet distinct. On one hand, the demand component of the oil shock is linked to the sharp reduction in oil consumption stemming from the negative supply and demand shocks associated with the spread of Covid-19. The speed of that recovery will depend on how swiftly and decisively governments take measures to mitigate the economic and financial dislocations from the health crisis. On the other hand, the supply component of the oil shock is likely to persist and keep downward pressure on oil prices for some time. I.3 Toward a Sequencing of Policy Responses to the Dual Shocks To deal with the dual shocks, a two-pronged approach can be pursued. One is immediate and related to the health emergency. The other concerns forward looking policy reforms. Authorities could sequence and tailor their responses to the severity of the shocks. That is, it might be desirable to focus first on responding to the health emergency and the associated economic contraction. Fiscal consolidation and structural reforms associated with the persistent drop in oil prices and pre-existing challenges are also very important, but with proper external support, can wait until the health emergency subsides. After the crisis, budget-neutral reforms such as debt transparency and reforms of state-owned enterprises will also come to the fore. In responding to Covid-19, fiscal authorities could prioritize health spending—including producing or acquiring test kits, mobilizing and paying health workers, adding health-care infrastructure, and preparing for vaccination campaigns. The authorities could use target cash transfers on vulnerable households and support the private sector, including informal firms (see Arezki and Nguyen, 2020). During this time, it is paramount to reach the large number of workers in the informal sector, which provides no safety net. Chapter III provides estimates of the informal labor force in MENA, which can be as high as 70 percent of the labor force. Many of them work hand-to-mouth. Given the large number of informal workers and borrowing constraints in many developing countries in MENA, targeted assistance is vital and should be larger relative to the economy than similar efforts in advanced economies. Successful models of quickly deploying technology to fight Covid-19 and target assistance can be analyzed and replicated5. Freeing information flows, increasing transparency, and data disclosure to reduce leakages, are crucial elements in targeted cash transfers—which are essential to ensuring a flattening of the spread of the virus, hastening the economic recovery, and limiting the rise in poverty. To reduce the risk of financial instability, the relevant policy authorities in the MENA region can reduce interest rates and inject liquidity into the banking system. Where inflation is low, liquidity injection and targeted cash transfers could even be financed by “helicopter money,” that is, essentially, money printed by central banks (Gali, 2020). The battle against the spread of the novel coronavirus and its economic and social consequences will be made more difficult by empty government coffers. Many MENA countries are facing large balance of payments and fiscal deficits. Many also carry high sovereign-risk premiums. For those countries, additional foreign borrowing on private markets will be difficult. Moreover, countries with fixed exchange rates will find it difficult to use helicopter money because of the tension between money-printing and maintenance of a peg. The region will need much international support to help it navigate an extremely rough patch. 5 See Foreign Affairs (2020) for the experience of Taiwan. CHAPTER I 8 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 I.4 Quantifying the Effects of the Dual Shocks One key indicator of the expected costs of the dual shocks is arguably the change in growth expectations relative to previous growth forecasts prior to the the dual shocks. However, while it is relatively straightforward to predict that MENA’s growth will decline, it is much more difficult to forecast future growth precisely. For one thing, economists generally have difficulties forecasting severe economic downturns (An and others 2018). For another, rare and large negative shocks, so-called “black swan” events, are even harder to predict than run-of-the-mill recessions (Vegh and others 2018). During crisis periods, the variance across different forecasts tends to grow larger, reflecting economists’ diverging thinking about future economic developments in situations where anything can seem plausible. Table I.1 shows the standard deviation of private sector forecasts for a set of MENA countries’ output growth in 2020—made in March 2020, after the pandemic and the collapse of oil prices, and in December 2019, before the pandemic. For most countries, projected 2020 GDP growth rates were revised downward while the standard deviation (an indicator of forecast uncertainty) increased. Table I.1. Standard Deviation of Private-Sector Forecasts for 2020 GDP Growth across MENA Economies Consensus Forecasts Standard Deviation March December March December Difference Difference forecasts forecasts forecasts forecasts Algeria 1.5 1.7 -0.23 0.7 0.6 0.07 Egypt 5.5 5.6 -0.05 0.3 0.4 -0.02 Iraq 3.4 3.8 -0.39 2.1 1.7 0.39 Jordan 2.3 2.3 -0.05 0.2 0.3 -0.01 Lebanon -1.2 1.1 -2.27 3.1 1.0 2.09 Morocco 3.0 3.0 -0.05 0.5 0.5 -0.01 Tunisia 2.3 2.1 0.17 0.7 0.9 -0.13 Bahrain 1.7 2.0 -0.26 0.5 0.7 -0.17 Kuwait 1.9 2.2 -0.31 0.8 0.7 0.13 Oman 2.0 2.4 -0.43 1.1 1.5 -0.42 Qatar 2.0 2.5 -0.45 0.6 0.4 0.14 Saudi Arabia 1.6 2.0 -0.43 0.5 0.4 0.13 UAE 2.0 2.4 -0.40 0.6 0.4 0.23 Source: Authors’ calculations based on data from Focus Economics (2020). Given the dual shocks to the region, World Bank economists downgraded the growth forecast for MENA in 2020 by 3.7 percentage points relative to our forecast published in October 2019 (see Tables I.2 and I.3). The growth downgrade of 3.7 percentage points is arguably a rough measure of the expected costs of the dual shocks, because they are the dominant developments since October 2019. Nevertheless, it cannot be overstated that the forecasts remain fluid. To illustrate the uncertainty surrounding the forecasts, Table I.3 shows the differences between two recent rounds of World Bank’s forecasts separated by less than two weeks (in March 19, 2020 and April 1, 2020 respectively) with the October 2019 forecasts. For most countries, the new forecasts in April 1, 2020 are substantially lower than those done in March 19, 2020, reflecting adjustments due to newly available information. 9 CHAPTER I Table I.2. Uncertain Forecasts: World Bank’s Growth, Current Account and Fiscal Balance Forecasts Real GDP Growth Real GDP per capita Growth Current Account Balance Fiscal Balance CHAPTER I (percent) (percent) (percent of GDP) (percent of GDP) 2018 2019e 2020f 2021f 2022f 2018 2019e 2020f 2021f 2022f 2018 2019e 2020f 2021f 2022f 2018 2019e 2020f 2021f 2022f MENA 1.1 0.3 -1.1 2.1 2.8 -0.6 -1.1 -2.6 0.4 1.5 4.1 1.2 -7.2 -4.6 -3.4 -3.0 -4.7 -9.7 -8.0 -7.0 Developing MENA 0.2 -0.2 -1.8 2.3 3.1 -1.3 -1.5 -3.4 0.7 1.6 -0.8 -3.1 -6.8 -5.2 -4.9 -2.8 -5.4 -10.0 -8.6 -8.0 Oil Exporters 0.4 -0.5 -1.6 1.7 2.3 -1.4 -2.0 -3.2 0.0 1.0 6.3 2.6 -7.8 -4.8 -3.3 -2.1 -4.2 -10.3 -8.4 -7.3 GCC 2.0 0.9 -0.4 1.8 2.5 0.1 -0.8 -1.9 0.1 1.4 8.5 5.1 -7.6 -4.0 -1.7 -3.2 -4.0 -9.3 -7.4 -5.9 Bahrain 1.8 1.8 -2.5 3.0 2.3 1.3 1.8 -2.5 -3.4 2.3 -5.9 -3.0 -9.2 -7.3 -5.3 -11.9 -10.6 -16.6 -12.5 -10.2 Kuwait 1.2 0.7 0.0 1.6 2.2 -0.2 -0.5 -1.2 0.2 0.8 15.1 8.1 -6.4 -5.2 -3.0 -3.0 -13.6 -25.6 -19.8 -14.7 Oman 1.8 0.5 -3.5 2.7 2.5 -2.3 -3.0 -6.3 0.4 2.5 -5.5 -5.2 -15.2 -11.4 -7.7 -7.9 -6.9 -17.9 -14.5 -12.0 Qatar 1.5 1.4 0.4 1.5 2.4 -0.6 -0.4 -1.3 -0.2 0.7 8.7 2.6 0.0 1.9 2.8 2.2 1.3 -3.0 -1.0 0.0 Saudi Arabia 2.4 0.3 0.2 2.1 2.6 0.6 -1.4 -1.4 0.5 1.1 9.0 5.4 -10.3 -5.4 -5.3 -5.9 -4.2 -7.5 -6.4 -6.4 United Arab 1.7 1.7 -1.1 1.2 2.3 0.2 0.3 -2.4 0.0 2.3 9.1 7.5 -5.7 -2.5 3.9 1.2 -1.6 -7.0 -5.5 -3.0 Emirates Developing Oil -2.4 -3.1 -3.9 1.4 1.9 -4.0 -4.2 -5.5 -0.3 0.2 2.4 -1.9 -8.3 -6.1 -5.9 -0.2 -4.5 -11.9 -10.2 -9.8 Exporters Algeria 1.4 0.9 -3.0 1.1 1.8 -0.3 -0.7 -4.5 -0.3 0.3 -9.8 -10.2 -18.8 -17.0 -17.0 -9.6 -11.5 -16.3 -16.5 -14.8 HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Iran -4.7 -8.2 -3.7 1.3 1.5 -5.7 -9.1 -4.6 0.4 0.7 5.3 -0.4 -2.5 -2.1 -1.9 -1.4 -5.1 -6.5 -6.8 -7.1 Iraq -0.6 4.4 -5.0 1.9 2.7 -3.3 3.5 -8.3 -1.6 -0.8 6.9 2.5 -12.1 -5.9 -5.7 11.2 3.0 -19.4 -12.3 -11.4 Developing Oil 3.8 3.5 0.6 3.3 4.5 2.2 1.9 -0.8 1.9 3.2 -6.6 -5.0 -4.8 -4.0 -3.7 -7.4 -6.7 -7.6 -6.4 -5.9 Importers Djibouti 8.4 7.5 1.3 9.2 8.2 6.8 5.9 -0.1 7.7 6.7 13.4 18.5 16.5 18.4 18.4 -2.5 -0.5 -2.9 -2.1 -2.0 Egypt 5.3 5.6 3.7 3.8 5.8 3.4 3.7 1.9 2.1 4.1 -2.4 -3.6 -3.7 -3.4 -3.3 -9.7 -8.1 -8.2 -7.3 -6.5 Jordan 1.9 2.0 -3.5 2.0 2.2 0.1 0.5 -4.5 1.3 1.9 -7.0 -2.9 -3.9 -3.7 -3.3 -3.4 -4.7 -4.4 -4.1 -3.4 Lebanon -1.9 -5.6 -10.9 -6.3 -3.4 -2.5 -6.1 -11.4 -6.8 -2.1 -24.3 -12.5 -7.0 -6.5 -7.5 -11.0 -10.6 -12.1 -11.4 -12.0 Morocco 3.0 2.3 -1.7 5.5 4.2 1.9 1.2 -2.7 4.4 3.1 -5.5 -4.6 -7.5 -4.2 -3.2 -3.7 -3.6 -6.0 -3.3 -3.1 Tunisia 2.7 1.0 -4.0 4.2 2.2 1.6 -0.4 -4.8 3.2 1.3 -11.2 -8.8 -7.2 -7.0 -6.8 -4.8 -4.1 -5.0 -3.8 -2.9 West Bank & Gaza 1.2 0.9 -2.5 2.1 2.4 -1.4 -1.7 -5.0 -0.5 -0.3 -10.2 -9.9 -7.3 -6.3 -6.0 -2.5 -4.4 -6.1 -3.9 -3.8 Memorandum Libya 15.1 2.5 -19.4 NP 1.4 13.4 1.0 -20.5 20.7 0.3 21.4 11.6 -29.3 -9.0 -8.2 -7.0 1.7 -36.7 -2.6 -2.5 Sources: Authors’ calculations based on data from World Bank Macro and Poverty Outlook. Data are as of April 1, 2020. Note: e=estimate, f=forecast and NP=not presented. Data are rounded up to a single digit. Data for Egypt correspond to its fiscal year (July-June). Libya, Syria and Yemen are not included in the regional and sub-regional averages due to lack of reliable data. 10 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table I.3. Changing Estimates of the Costs of the Crisis: World Bank Growth Forecasts Relative to October 2019 Percentage points Difference (March 19, Difference (April 1, October Forecast Real GDP Growth, percent 2020 - October 2019) 2020 - October 2019) 2019e 2020f 2021f 2019e 2020f 2021f 2019e 2020f 2021f MENA 0.6 2.6 2.9 -0.3 -2.1 -0.3 -0.2 -3.7 -0.8 Developing MENA 0.0 3.0 3.1 -0.2 -2.8 -0.5 -0.2 -4.8 -0.7 Oil Exporters -0.4 2.1 2.3 -0.2 -2.4 -0.3 -0.1 -3.7 -0.7 GCC 1.1 2.2 2.7 -0.3 -1.4 -0.2 -0.2 -2.6 -0.9 Bahrain 1.8 2.1 2.3 -0.1 -1.3 -0.7 0.0 -4.6 0.7 Kuwait 1.5 2.5 2.8 -0.8 -2.5 -1.2 -0.8 -2.5 -1.2 Oman 0.3 3.5 4.0 0.2 -3.3 -3.0 0.2 -7.0 -1.3 Qatar 2.0 3.0 3.2 -0.6 -1.6 -0.9 -0.6 -2.6 -1.7 Saudi Arabia 0.5 1.6 2.2 -0.2 -0.8 1.3 -0.2 -1.4 -0.1 United Arab Emirates 1.8 2.6 3.0 -0.3 -1.8 -1.7 -0.1 -3.7 -1.8 Developing Oil Exporters -3.3 1.8 1.7 0.1 -4.1 -0.4 0.1 -5.7 -0.3 Algeria 1.3 1.9 2.2 -0.4 -3.2 -0.9 -0.4 -4.9 -1.1 Iran -8.7 0.1 1.0 0.5 -1.9 0.1 0.5 -3.8 0.3 Iraq 4.8 5.1 2.7 -0.4 -9.0 -0.8 -0.4 -10.1 -0.8 Developing Oil Importers 4.1 4.4 4.6 -0.6 -1.2 -0.6 -0.6 -3.8 -1.3 Djibouti 7.2 7.5 8.0 0.3 -0.5 0.2 0.3 -6.2 1.1 Egypt 5.6 5.8 6.0 -0.1 -0.3 -0.4 -0.1 -2.1 -2.2 Jordan 2.2 2.3 2.5 -0.2 -0.4 -0.3 -0.2 -5.8 -0.5 Lebanon -0.2 0.3 0.4 -5.4 -8.1 -6.1 -5.4 -11.2 -6.8 Morocco 2.7 3.5 3.6 -0.4 -1.8 0.0 -0.4 -5.2 1.9 Tunisia 1.6 2.2 2.6 -0.6 -1.4 -0.8 -0.6 -6.2 1.6 West Bank & Gaza 1.3 -1.1 -0.4 -0.4 0.0 2.5 -0.4 -1.4 2.5 Source: Authors’ calculations based on data from World Bank Macro and Poverty Outlook and Arezki et al. (2020). Note: Libya, Syria and Yemen are not included in the regional and sub-regional averages due to lack of reliable data. As with any economic forecasts during periods of unexpected and large negative shocks, these forecasts have large margins of error. This speaks to the possibility of different future states of the world – “multiple equilibria” in technical jargon. Depending on the future spread of the virus, the health policy responses, societal responses, and future developments in global oil markets, several plausible scenarios could emerge for any or all MENA economies. The actual growth rates could reflect different equilibria than the ones presented here. The uncertainty about the future is further exacerbated by incomplete current information about the spread of the novel coronavirus—which could be due to lack of testing, lack of disclosure, or both. With this caveat, we examine whether the World Bank’s growth downgrades correlate with a country’s exposure to oil exports—measured as net crude oil exports as a fraction of GDP in 2019—and a country’s Global Health Security (GHS) index—which captures its capability to prevent and mitigate epidemics and pandemics, including compliance with international norms and the reporting of health information6. 6 The index was jointly developed by the Nuclear Threat Initiative, the Johns Hopkins Center for Health Security, and the Economist Intelligence Unit. Data were released in 2019. The index consists of six categories: prevention; detection and reporting; rapid response; health system; compliance with international norms; and risk environment (see https://www. ghsindex.org/). 11 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Fifteen countries are included. 7 Overall, growth downgrades between April 1, 2020 and October 2019 are positively correlated with the GHS Index and not significantly correlated with oil export exposure. In other words, for countries with a stronger capability to prevent and mitigate pandemics, economic growth is expected to fall relatively less than for countries not so well situated (see the partial correlation scatterplots in Figure I.5). The fact that the downgrades are not significantly correlated with oil export exposure could reflect the fact that even oil importers in MENA can suffer from the decline in oil prices, as mentioned above. However, a similar exercise conducted with the March 19 forecasts showed a negative relationship between the growth downgrades and oil export exposure. This relationship disappeared with the April 1 forecasts. Figure I.5. Correlates of the Costs of the Crisis: Growth Downgrades, Oil Export Exposure and Health Security Panel A: Growth Downgrades and Health Security Index Panel B: Growth Downgrades and Oil Export Exposure 4 4 Saudi Arabia Egypt Kuwait 2 Qatar Algeria 2 Egypt Qatar United Arab Emirates Kuwait Saudi Arabia Iran 0 Algeria Bahrain Morocco Lebanon Jordan Djibouti Tunisia 0 Iran Tunisia Bahrain Djibouti United Arab Emirates -2 Lebanon Morocco Oman -2 Jordan -4 Iraq Iraq Oman -6 -4 -20 -15 -10 -5 0 5 10 15 -.3 -.2 -.1 0 .1 .2 .3 .4 Health residual Oil residual Growth residual Fitted values Growth residual Fitted values Source: Authors’ calculations based on data from World Bank Macro and Poverty Outlook. The evidence suggests that the changes in the World Bank’s forecasts are systematically related to initial pre-existing conditions, particularly each economy’s health system. It cannot be overstated, however, that all economic forecasts from any source are highly uncertain. This said, the challenge of low growth was not brought to MENA by the dual shocks. The region has been suffering from low-growth syndrome for decades. An important and related question is whether the growth slowdown will be transitory or permanent. In general, a transitory economic shock that does not permanently affect other economic variables will imply a transient growth downgrade, leading to a fast recovery afterwards. This would be a “V-shaped” pattern. If instead, the shocks are permanent or if transitory shocks interact with domestic conditions to make the impact permanent, then the over-time path of growth could follow an “L” shape. In the current context, as mentioned above, the spread of the virus could have potentially catastrophic, but transitory, effects on public health . However, it could have more durable effects on the economy if it interacts with existing economic vulnerabilities. On the other hand, the oil price collapse might be more long-lasting if it reflects permanent structural changes in the global oil market due to technological progress (e.g., shale and renewables) and consequent changes in the market power of OPEC+. 7 The Global Health Security Index is not available for West Bank and Gaza. Oil export data are not reliable for Libya. CHAPTER I 12 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 A close inspection of the changes in the World Bank’s growth forecasts can shed light on the World Bank economists’ belief about the nature of the dual shocks’ impacts on our growth expectations for MENA in 2020-2021. Panel A of Figure I.6 shows the changes in the growth forecasts between those conducted in April 1, 2020 and those published in October 2019, for MENA as a whole, the GCC, developing oil exporters, and developing oil importers as per the classification in Table I.2 above. The sharpest drop in 2020 growth forecasts corresponds to the developing oil exporters, followed by the developing oil importers and then by the GCC countries. Despite the oil price collapse, significant growth downgrades of oil importers indicate the economic fortunes of the region’s oil exporters and oil importers are connected. The new forecasts assume that the price of oil will be around $30 dollars per barrel in 2020 and rise to about $40 in 2021. Both are notably below the price forecasts in October 2019. In addition, a part of the recovery of growth forecasts for 2021 relative to those in October is due to the expected recovery in oil price between 2020 and 2021. Nevertheless, the World Bank expects MENA growth rates in 2021 to be below the forecasts in October 2019. That is, our 2021 growth forecasts have by and large been revised downward. And this is true for both oil exporters and oil importers (Panel A of Figure I.6). A decomposition analysis suggests that once we remove the direct positive effect of the expected oil price recovery, the 2021 growth for oil exporters would be even lower. This implies that our economists expect somewhat durable effects of the dual shocks, above and beyond the direct income effect of oil price changes. Note that the forecasts are fluid and subject to change when new data become available. Panel B of Figure I.6 illustrates this. Between March 19 and April 1, 2020, World Bank economists sharply increased their estimates of the cost of the dual shocks. Growth downgrades for 2020 and 2021 for MENA, the approximate costs of the dual shocks, went from -2.1 to -3.7 percentage points for 2020, and from -0.3 to -0.8 percentage points for 2021. To the extent that the full repercussions of the dual shocks have not been fully captured in economic forecasts, it is safe to conclude that our estimates of the costs of the crisis are conservative; they can be interpreted as lower-bound estimates of the costs.. Figure I.6 Fluid Estimates of the Costs of the Crisis — Changes in World Bank Growth Forecasts Panel A: Changes in Forecasts (April 1, 2020 - October Panel B: Changes in Forecasts (April 1, 2020 - October 2019) across MENA Country Groups, 2019-2021 2019) and (March 19, 2020 – October 2019) MENA Developing Oil Exporters Developing Oil Importers GCC MENA (April 1, 2020 - October 2019) MENA (March 19, 2020 - October 2019) 1.0 0.0 2019E 2020F 2021F -0.5 0.0 2019E 2020F 2021F -1.0 -1.0 -1.5 -2.0 Percentage Percentage -2.0 -3.0 -2.5 -4.0 -3.0 -5.0 -3.5 -6.0 -4.0 Source: Authors’ calculations based on data from World Bank Macro and Poverty Outlook and Arezki et al. (2020). Note: Libya, Syria and Yemen are not included in the regional and sub-regional averages due to lack of reliable data. 13 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA I.5 MENA’s Chronic Low-Growth Syndrome The analysis of current challenges and Figure I.7 MENA’s Chronic Low-Growth Syndrome risks underscores a sobering fact: The 2000-2022 region has suffered from chronic low- growth syndrome for decades. In fact, 4 Lower Middle MENA’s per-capita growth has been low 3 Upper Middle even relative to the mediocre benchmark 2 of the typical (median) growth rate High Income in the rest of the world. When each 1 country’s growth performance since 0 the beginning of the 21st century -1 into our forecast horizon is compared -2 to the median (or typical) economy in -3 their corresponding income groups, West Bank and Gaza Bahrain Kuwait Lebanon Algeria Iran, Islamic Rep. Djibouti Tunisia Saudi Arabia United Arab Emirates Egypt, Arab Rep. Yemen, Rep. Oman Qatar Iraq Libya Jordan Morocco the evidence indicates that most MENA economies underperformed relative to that mediocre benchmark. Figure I.7 shows average growth rates in per capita GDP for each MENA economy (represented by blue diamonds) as well as the median growth rates of Sources: World Bank, Macro and Poverty Outlook and World Development Indicators; International Monetary Fund, World Economic Outlook; and World Bank staff calculations. Data are as of October 2019. their corresponding income groups (represented by red horizontal lines) over the period 2000–2022. All Gulf Cooperation Council (GCC) countries and upper-middle-income MENA countries have grown more slowly than the typical high-income and upper-middle income country during this period. Among lower-middle-income countries, only in Djibouti and Morocco does long-term growth outperform that of a typical peer. The long-standing underperformance of economic growth in the region is not only notable but large. A rough calculation suggests that if all MENA economies had grown at the median rate of their respective income groups, the region would be, on average, at least 20 percent richer than it is today. These findings raise the question of what factors underlie this low-growth syndrome, which has been accompanied by outbursts of social unrest. This report argues that lack of transparency is a strong candidate. The following section discusses the relationship between lack of transparency and long-term growth. CHAPTER I 14 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 I.6 Enhancing MENA’s Transparency Can Accelerate Growth There is a widening data gap between MENA and advanced economies: while advanced economies characteristically have modern and well-coordinated data collection systems that are accessible to the research community, many economies in the MENA region have either lagged in their capacity to generate data or have prevented the research community, the independent media, and civil society from accessing its data. While concerns over privacy are real, little attention has been paid to the costs of opaque data systems that hamper external knowledge generation. But some of these costs are becoming apparent. Lack of transparency hurts even more when systems are under stress by potent threats, such as the ongoing Covid-19 pandemic. An optimal societal response requires open and direct communication across several actors in society—the government, health care systems, civil society, and various institutions. Information needs to be collected in real time to enable governments and public health officials to take timely, decisive actions. Citizens need to report cases and respond to behavioral changes requested by the government. The flow of data is the oil of the engine of this system of interactions and responses. When data is not made public or is misused, the engine can fail. The ramifications of the lack of trust, forged by limited transparency, comes into stark relief when citizens are confused about what to believe. And as we have seen several times over many decades, regaining credibility is not easy. As one citizen in the region aptly described the leadership response to Covid-19: “When you lose people's trust, even when you tell the truth, people won't believe you.” The responsibilities of countries in the data agenda has been well established. At their core, the U.N.’s Sustainable Development Goals (SDGs) emphasize the need for countries to generate socioeconomic indicators within the limits of each country’s capacity. The costs of such efforts are not trivial, and the tradeoffs between investing in data capacity and systems and other pressing needs are difficult for developing country governments to ascertain. The long-run benefits of transparency, however, are considerable. It is uncontroversial that economies in the MENA region need to make substantial investments, while adopting best practices to catch up in terms of data transparency. 15 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Box I.1. Transparency and the Statistical Capacity Index The evolution and widespread use of the word “transparency” in terms of governance can be largely attributed to supranational and non-governmental organizations. In the early 1990s, Peter Eigen, a former World Bank manager established Transparency International as an alternative way to address corruption. The organization’s mission is to study the effects of corruption on citizens and advocate policy reforms in global institutions to address corrupt practices. “Integrity International” and “Honesty International” were considered as names for the organization, but “transparency” won out as it was understood to convey the term “openness” (Ball, 2009). Consequently, the term “transparency” spread across the World Bank, the OECD, and congressional directives to the International Monetary Fund. Academicians adopted the term “transparency” and formalized its meaning, especially in the field of International Studies. Finel and Lord (1999) defined transparency as comprising “the legal, political, and institutional structures that make information about the internal characteristics of a government and society available to actors both inside and outside the domestic political system. Transparency is increased by any mechanism that leads to the public disclosure of information, whether a free press, open government, hearings, or the existence of nongovernmental organizations with an incentive to release objective information about the government.” Mitchell (1998) used the definition: “Transparency constitutes the demand for information, the ability of citizens to obtain information, and the supply and actual release of information by government and NGOs.” The World Bank’s statistical capacity index goes beyond its name by capturing many of the elements of transparency consistent with the definitions above (see table A2 for detailed definition of the statistical capacity index and its components). The availability and regular publication of micro and macro data as well as whether production of such data adheres to international standards goes to the heart of “openness,” the ability of citizens to obtain information, and actual release of information by the government. The measure goes beyond statistical capacity— highly competent statistical offices can be penalized if they do not publish statistics. The statistical capacity index captures transparency by using data-centric, objective, and verifiable measures—and is unique in that it is not dependent on perceptions of transparency by survey respondents as is typically the case in many transparency indicators. The statistical capacity index can be re-interpreted as a statistical or data transparency index. Figure I.8 shows a fitted line between overall statistical capacity and the level of development for 149 mostly developing economies in 2005 (see left panel) and in 2018 (see right panel). Note that the positive relationship between statistical capacity and GDP per capita would likely be stronger were advanced economies included in the sample. The circles in figure I.8 shows the regional averages. While the MENA region (excluding the GCC countries) was already underperforming in 2005 relative to its level of development, it became the region with the lowest statistical capacity in 2018. The weakening relationship between statistical capacity and development between 2005 and 2018 could be because poorer economies succeeded in developing statistical capacity or richer economies restricted data access, perhaps due to privacy concerns. In theory, there are at least three channels through which data transparency affects development. First, credible and timely data serve as the basis for policy formulation and reforms. Policies can only be as good as the empirical evidence on which they are based. At a fundamental level, data are about records. Take the example of a business. A manager has the primary goal of raising profits. To achieve this, performance must be benchmarked historically and compared with that of competitors. Collateral must be evaluated and leveraged to obtain financing to pursue new ventures. Risk and reward must be balanced. Investors need to be enticed. Without record-keeping, many CHAPTER I 16 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 of these goals could not be achieved. A similar set of challenges face governments in the MENA region. Countries need to grow, and, to expand options, data must be reliably transparent to provide guidance. Countries with high quality and broadly accessible information can make better decisions. Through data and evaluation, existing policies may be reformed and refined, while new policies may be experimentally evaluated. Figure I.8. Regional Development and Statistical Capacity Note: The overall statistical capacity measure captures availability of data (micro and administrative), adherence to international standards in terms of methodology, and periodicity and timeliness of statistical capacity (see Appendix Table A2 and Box I.1). The fitted line is based on 149 economies, although only regional averages are displayed. The MENA sample excludes GCC economies. The West Bank and Gaza is included only in the 2018 graph due to lack of data for 2005. Since the West Bank and Gaza has better statistical capacity than most economies in the region, its omission in 2005 may indicate that MENA was performing worse in terms of statistical capacity than the 2005 graph indicates. Second, data that are accessible to the broader civil society can generate better policies and reforms. Substantial expansions in the frontier of knowledge occur when data are available to a large base of analysts. Researchers test hypotheses, debate and dispute findings, establish robust facts and relationships to facilitate the emergence of the best ideas for addressing challenges. It is not surprising that richer economies are researched more than poorer economies. Publications in top-ranking journals in the economics profession are skewed towards wealthier economies—a fact that could be partly explained by lack of data accessibility in lower income economies (see the evidence presented by Das and others 2013)8. Chapter II of the report shows that the lack of transparency concerning public debt stocks in MENA could hamper credible analyses on debt sustainability, an important topic as stimulus measures are being adopted to respond to the Covid-19 and oil price dual shocks. After the crisis, it would be good to have a clear understanding of the debt situation in the region. Third, when data are of questionable quality or unavailable, the gap between perceptions and reality may grow. Important reforms may lead to real welfare improvements yet have little impact on public perceptions because there is limited data tracking such improvements. These perceptions may foster a narrative that results in frustrations that manifest themselves in social protests and unrest. Similarly, if data are of dubious quality, the public may lose confidence in such information and may not alter their perceptions despite positive findings from the data. More important, once a government walks down the path of unreliable or limited data accessibility, it may be difficult to regain credibility. The public may be less willing to trust information from the government, which makes it difficult for a changed government to change public perceptions. The result is economies that are more prone to social upheavals. In fact, for MENA, there has been a long-standing mystery of why social unrest is so prevalent in societies with relatively little income inequality. Recent research suggests that this might be due to other factors, perhaps lack of transparency, which create a wedge 8 It is not surprising that richer economies are researched more than poorer economies. Publications in top-ranking journals in the economics profession are skewed towards wealthier economies—a fact that could be partly explained by lack of data accessibility in lower income economies (see the evidence presented by Das and others 2013). 17 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA between people’s perceptions of their relative socioeconomic standing in society and their actual position. Clementi and others (2019) studied Morocco before and after the Arab Spring and found evidence of such skewed perceptions. However, to inform policy, availability and accessibility of good quality (credible) data are necessary, but not enough. Think tanks, the media and policymakers play an important role in facilitating debate fostered by data to ensure that policy is an outcome of the process and the public takes ownership of the debate. The process typically goes as follows: As produced, data are not easily digestible by the public, so academics generate knowledge from the data that fosters debate among themselves. Think tanks and policymakers join in. Media institutions communicate the information to the public who then participate and take ownership of the debate. The outcome of this process hopefully is an optimal set of policies that promotes overall welfare. Each of the institutions in the process is critical to fostering and galvanizing public debate. Failure of the media to inform the public or the absence of think tanks would limit the value of information in guiding policy and diminish the returns of investing in data. All parts must move together for the data ecosystem to be effective. That is, transparency requires not only the production and publication of reliable data on a regular basis, but it also requires that key stakeholders are empowered to express their voice openly. Ì Data Transparency in MENA In 2005, the MENA region exceeded the East Asia and the Pacific (EAP) and the sub-Saharan Africa (SSA) regions in terms of overall “statistical capacity,” a term that includes aspects of data quality as well as accessibility—two important ingredients of transparency. Since then, both the EAP and SSA countries have overtaken the MENA region. By 2018, the MENA score was the lowest of all regions (see Figure I.8). MENA was also the sole region to experience a decline in statistical capacity between 2005 and 2018. Conflicts might have played a role in the decline. There are considerable differences in statistical capacity among the countries in the MENA region (see Figure I.9). Egypt is the best performer, followed by Iran and Jordan. Statistical Figure I.9 Statistical Capacity Index across MENA capacity has been steadily Statistical Capacity score (overall average), 2018 increasing since 2005 in these 100 90 economies. At the other end are 90 Libya, Syria, and Yemen. All three 80 73 74 78 economies are overwhelmed by 70 63 64 67 60 conflict and their data systems 60 51 57 have drastically deteriorated 50 38 40 since 2005. Nonetheless, even in 29 33 30 the seemingly high performers, 20 deterioration in freedom of 10 expression has probably become 0 an impediment to harnessing Libya Syrian Yemen, Iraq Algeria Djibouti Tunisia Lebanon West Morocco Jordan Iran, Egypt, Arab Rep. Bank Islamic Arab the upside from the production Republic and Gaza Rep. Rep. of reliable data. Furthermore, as discussed in following chapters, Source: The World Bank, http://datatopics.worldbank.org/statisticalcapacity/. CHAPTER I 18 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 even in Egypt there are issues concerning the lack of availability of key information on public debt, as well as issues concerning the precision of the definitions used to compute labor-market indicators. Low statistical capacity in the MENA region is also reflective of a lack of micro data, especially regarding enterprises and prices. For instance, the lack of micro business statistics in the region implies that the structure of the economy in terms of the types of firms in each sector is unknown. That inhibits the advancement of discussion and policies framed around firm dynamics, particularly market concentration and competition policy. Perhaps more important, business statistics are also crucial inputs for private sector firms and potential investors. Ì The Empirical Link between Data Transparency and Economic Growth There is a positive correlation between data statistical capacity and subsequent economic growth during 2005–18, a sample of 146 economies shows. The relationship between growth and statistical capacity holds across various econometric models after accounting for several confounding factors such as level of development, sectoral composition, human capital, and political institutions (see Appendix A for details on the econometric models). The magnitude of the association between statistical capacity and growth is at least as large (if not larger) than the association between education and growth9,10. The statistical capacity index decline experienced in the MENA region between 2005 and 2018 may have resulted in a loss of GDP per capita between 7 and 14 percent, depending on the econometric model employed (see Appendix Table A1). The results suggest that the availability and frequency of compilation of the administrative and micro data is a key predictor of economic growth, although the findings are somewhat susceptible to the empirical methodology. These findings complement other studies that have found positive correlation between the statistical capacity indicator and a wide variety of governance and service provision outcomes (Hollyer and others, 2011; Hoogeveen, 2018; Islam 2006; Williams 2009). This is a first step in the analysis of the relationship between data transparency and economic growth. Several caveats apply to the findings. Issues of endogeneity such as simultaneity between data capacity and growth are not completely obviated and there is always the challenge of omitted variable bias driven by conflict and resource dependency of economies. In future research, these issues might be addressed with more sophisticated analytical tools. Ì What Next? Bridging the MENA data-transparency gap requires a multipronged approach to developing sustainable data ecosystems. The ongoing pandemic has put this issue at the forefront. Where governments lack the capacity to generate data, investments are needed to build that capacity. Where governments are unwilling to share data, agreements need to be developed in concert with a clear agenda that highlights the benchmarks of good data ecosystems and the crucial role of data in generating good policies and social harmony. And when important topics are under-researched in the region and require specific data, investments should be made in data collection activities to set up a baseline of knowledge. The immediate goal is to bring the data transparency challenge to the table. Future initiatives will dig deeper into these issues, acknowledging country-specific contexts. The following chapters turn to the role played by specific data and transparency issues that afflict MENA in two other areas where they hurt—macroeconomic fragility and labor markets. 9 Magnitudes are derived by looking the effects of increasing the variables by 1 standard deviation. The endogeneity concerns regarding statistical capacity also apply for education. 10 Other studies examined the mechanisms why greater data transparency could help growth. For example, greater data transparency could lower the costs of external borrowing (Cady, 2015 and Kubota and Zeufack, 2020). 19 CHAPTER I HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA CHAPTER II: EXTERNAL IMBALANCES, FISCAL SUSTAINABILITY, AND DATA TRANSPARENCY IN MENA Chapter II takeaways: • Current account deficits of several MENA economies are not explained by fundamentals; transparency issues do not affect these estimates. • In 2019, 11 MENA countries seemed to be on unsustainable fiscal paths: their reported primary fiscal balances were insufficient to stabilize their gross-debt-to-GDP ratios. • Fiscal sustainability assessments are hampered by lack of transparency concerning public debt stocks. This chapter assesses the sustainability of current account and fiscal deficits across MENA countries. It relies on the best available data that are comparable across countries. We begin with a discussion of the current account and then move on to a battery of tests assessing fiscal sustainability—hampered by a lack of transparency concerning stocks of public debt. II.1 Current Account Sustainability In April 2019, the MENA Chief Economist’s office presented a model to determine whether current account imbalances in some MENA countries are sustainable (see Appendix B). The model determines how a country’s current account balance is related to fundamental determinants drawn from academic literature—demography (dependency ratios, which measure pressure on the working age population, and aging speed), expected changes in economic growth, GDP per working-age population, and exposure to commodity price fluctuations. The determinants can be connected to current account imbalances in a variety of ways:11 • Demography and savings. As dependency ratios fall, national savings rise and improve the current account. Because the savings associated with fewer children or fewer older dependents are likely to be of different magnitudes, the model includes two dependency ratios. Young-age dependency captures the ratio of those younger than 15 to the working-age population (15-64 years of age). Old-age dependency captures the ratio of those older than 64 to the working-age population. The third proxy for demography is aging speed, which is the annual change in the old-age dependency ratio. When this ratio changes rapidly, family savings can rise in anticipation of future expenditures associated with the elderly. Hence it is plausible that current accounts can improve in economies with a rapidly aging population relative to the rest of the world. Demographic statistics, including the projections, account for refugees, who make up a large share of population in such countries as Lebanon and Jordan. The United Nations Population Division's total 11 Other research in this area, such as IMF (2013), has a longer list of fundamentals. This report considers fundamentals that are likely unaffected by an economy’s short-term economic performance. CHAPTER II 20 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 population estimates and forecasts incorporate migration data. These estimates include refugee inflows and outflows (United Nations 2017). • Forecast growth and domestic savings. If an economy’s growth is expected to accelerate, it would likely run a current account deficit because it could use future resources derived from faster expected growth to pay for today’s investment or consumption. The model includes a proxy of expected growth acceleration which relies on historical data from International Monetary Fund (IMF) forecasts. • GDP per working-age population and net savings. Aggregate labor productivity is simply the ratio of GDP to the working-age population. Economies with high labor productivity relative to other countries are likely to have higher domestic savings. Thus, unless improvements in output per worker are accompanied by disproportionate increases in domestic consumption, improvements in aggregate labor productivity are associated with improvements in the current account. The model utilizes the lagged ratio of an economy’s output (measured in terms of purchasing power parity, or PPP) to the size of its working- age population relative to the United States (the economy assumed to be at the “frontier” of highest productivity). However, if capital flows into less productive economies, it is possible that that such inflows can be associated with declines in the current account because inflows raise domestic investment and consumption. Consequently, the effect of productivity on the current account might itself depend on the openness of the capital account. The model thus includes the interaction between the openness of the capital account and relative labor productivity. • Commodity prices and the trade balance. The trade balance of an economy can be determined by fluctuations in commodity prices. When prices rise, trade balances improve for net exporters of commodities and deteriorate for net importers. In turn, when the trade balances change, so do the current account balances. As a result, the model takes commodity prices and commodity-trade balances into account.12 This variable is particularly relevant for MENA countries, because many are major oil exporters and significant food importers. By the same token, if food prices increase, the current account positions of food-importing countries can be expected to deteriorate. For example, the widening of the current account deficit in Tunisia in 2007 and 2008 caused by food imports should be captured by the index. • Exchange rate regimes. Fixed exchange rate regimes could be subject to real exchange rate misalignments, which affect the current account. For example, the real exchange rate could become undervalued in good times and overvalued in bad times because of the inability of the nominal exchange rate to adjust when domestic prices do not respond quickly to changes in demand. The MNACE model controls for three types of exchange-rate regimes: fixed exchange rate regimes, crawling pegs or managed floats, and free floats. In addition, the exchange rate regime variables interact with relative labor productivity to reflect the extent to which the regimes affect the response of current accounts to changes in labor productivity. The results from the MNACE model are broadly consistent with the predictions. In addition, results from an auxiliary model on national savings rates suggest that the selected explanatory variables affect the current account through their influence on national savings. The key issue, however, is whether MENA’s observed current account balances are fully explained by the fundamentals. 12 See Appendix B for details concerning the construction of the index. 21 CHAPTER II HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA There are six MENA economies that have current account balances significantly lower than the model’s predictions. These unexplained current accounts are the residuals (that is, the difference between the predicted value and the observed value) of the MNACE model. The residuals are computed by subtracting the predicted current account balance from the actual current account balance. Figure II.1 represents the 95 percent confidence interval of the residuals and groups the economies based on whether or not their current account balance is significantly lower than the model’s predicted current account balance. Panel A consists of countries whose reported current account balances are statistically significantly lower than what was predicted by the model—Algeria, Bahrain, Lebanon, Oman, Qatar, and Tunisia. Panel B consists of those whose current account balances are not statistically significantly lower than the model’s predictions— Djibouti, Egypt, Iran, Iraq, Jordan, Kuwait, Morocco, Saudi Arabia, the UAE, and Yemen. Figure II.1. Unexplained Current Account Balances for MENA countries Panel A: Countries with CA balances lower than what predicted Panel B: Countries with CA balances in line or larger than by fundamentals what predicted by fundamentals Algeria Bahrain Lebanon Djibouti Egypt Iran Iraq 35 20 25 15 10 5 -5 0 -15 -10 Current Account Balance (% of GDP) Current Account Balance (% of GDP) Jordan Kuwait Morocco Saudi Arabia -20 35 25 15 -30 5 -5 Oman Qatar Tunisia -15 20 2010 2012 2014 2016 2018 2020 2010 2012 2014 2016 2018 2020 10 UAE Yemen 0 35 25 -10 15 5 -20 -5 -15 -30 2010 2012 2014 2016 2018 2020 2010 2012 2014 2016 2018 2020 2010 2012 2014 2016 2018 2020 2010 2012 2014 2016 2018 2020 2010 2012 2014 2016 2018 2020 Residuals Upper Bound Residuals Upper Bound Residuals Residuals Lower Bound Residuals Forecast Lower Bound Residuals Forecast Source: Authors’ calculations based on the MNACE’s current account model (see Appendix B). Libya, Syria and West Bank & Gaza are not included because of the lack of data. Forecasts of current account are as of October 2019 Countries can reduce their large unexplained external imbalances by increasing their GDP per working-age population (Arezki and others, 2019). This is the only path that could help reduce external imbalances and improve fiscal sustainability while reducing the need for socially painful fiscal austerity. Improvements in aggregate labor productivity are associated with increased domestic savings. Barring a concurrent increase in labor productivity and domestic consumption, improvements in labor productivity should reduce current account deficits. (Arezki and others 2019 discuss these issues in more detail). In addition, because output per working-age population is likely to enlarge the public-sector’s revenue base, there will be some alleviation of fiscal challenges. Raising GDP per working-age population can only be achieved by putting working-age adults to work, or by raising private investment, because governments are facing severe fiscal constraints. Below we discuss methods and data that can help assess the fiscal situation of MENA economies. CHAPTER II 22 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 II.2 Fiscal Sustainability: Lack of Transparency Obfuscates Existing Methods of Analysis In this section, we adopt three approaches to examining fiscal sustainability in MENA countries: • Method 1: We calculate the required primary fiscal balance that stabilizes the public debt-to-GDP ratio in a given year and compare it to the observed balance. • Method 2: We construct and evaluate the structural fiscal balance by removing the components of revenues and expenditures that are automatically connected the business cycle, which is then compared to the observed and required balances mentioned above. • Method 3: We estimate a relationship between the primary fiscal balance and public debt in the previous year across a global sample of countries. Ì Required primary balance for debt stabilization The first approach is commonly used by academics and multilateral institutions such as the IMF and the World Bank to conduct debt sustainability analyses (see, for example, Debrun and others, 2019). The higher a country’s debt, or the higher the interest rate on the debt, the larger then is the primary balance required to stabilize the debt. Conversely, if a country has higher growth, it can afford a smaller required primary balance. In mathematical terms (details of which can be found in Appendix C1), the required primary balance to stabilize debt relative to output is where pbt is the required primary balance (as a share of output) for year t; dt-1 is debt-to-output ratio of the previous year. gt is nominal output growth and rt is nominal interest rate in local currency. Ì Estimating structural fiscal primary balances In the second approach, we determine each country’s structural primary fiscal balance—what remains after removing the components of revenues and expenditures that are connected to the inevitable ups and downs of the economy, such as additional tax revenue that comes from increased output (see Appendix C2). It is arguably more precise to assess solely the structural fiscal balance because it captures the fundamental (structural) fiscal condition of a country. IMF (2011) provide a step by step summary of the methodology used for calculating structural balances. 23 CHAPTER II HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Table II.1. Primary and Structural Fiscal Balances versus Debt-Stabilizing Primary Fiscal Balances in MENA, 2018 and 2019 Panel A: 2018 Primary Fiscal Balance in 2018 (% of GDP) Country Observed Required Structural Algeria -4.92 -1.65 -4.67 Iran, Islamic Rep. -5.13 -9.71 -5.58 Iraq 9.44 -6.00 10.50 Yemen, Rep. -6.27 -27.27 -6.11 Bahrain -7.36 -2.29 -7.39 Kuwait -2.79 -2.93 -2.30 Oman -5.87 -2.96 -5.75 Qatar 4.35 -4.13 4.63 Saudi Arabia -5.65 -1.90 -5.66 United Arab Emirates 2.19 -0.75 2.36 Djibouti -1.61 -1.60 -1.54 Egypt, Arab Rep. 0.11 -13.71 0.14 Jordan 0.01 -0.05 0.05 Lebanon -1.13 1.02 -1.11 Morocco -1.19 -0.01 -1.21 Tunisia -2.13 -3.53 -2.21 Panel B: 2019 Primary Fiscal Balance in 2019 (% of GDP) Country Observed Required Structural Algeria -5.32 0.11 -5.24 Iran, Islamic Rep. -4.98 -4.14 -4.49 Iraq -3.32 -1.00 -2.63 Yemen, Rep. -5.84 -11.67 -5.71 Bahrain -3.69 3.06 -3.64 Kuwait -6.14 0.13 -5.48 Oman -5.06 3.19 -4.26 Qatar 3.56 5.84 3.86 Saudi Arabia -5.86 1.16 -5.49 United Arab Emirates -0.97 0.47 -0.69 Djibouti 0.61 -3.50 0.65 Egypt, Arab Rep. 1.94 -5.81 1.89 Jordan 0.85 -0.72 0.91 Lebanon 0.26 0.71 0.40 Morocco -1.20 0.22 -1.06 Tunisia -1.23 -2.41 -1.26 Source: World Bank, Macro and Poverty Outlook; real GDP data, World Economic Outlook. Note: The observed primary balance and our calculated required and structural primary balances in 2018 and 2019 are all presented as a percentage of GDP. The required primary balance is one that stabilizes the debt- to-output ratio. The structural balance is the fiscal balance after the components connected to economic fluctuations are removed. For most countries, structural primary balances were close to observed primary balances because their output was close to potential. Panel A of Table II.1 shows that in 2018, the observed primary balances of Algeria, Bahrain, Lebanon, Morocco, Oman and Saudi Arabia were lower than the balances required for debt stabilization. These findings remain true even after accounting for business cycle drivers of the primary balance. For these six countries, the structural primary balance is also lower than the required primary balance. Panel B of Table 2.2.1 presents the structural and required primary balances for 2019. Fiscal sustainability worsened for MENA relative to 2018—11 of 16 MENA countries in our sample had a required primary balance that was larger than their observed primary balance. CHAPTER II 24 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table II.1 requires some qualification related to debt data transparency, however. First, dt-1 in equation (1) should be net public debt. But because data is lacking on net public debt, we use gross public debt to estimate the primary balances. This substitution might have inflated the required primary balance for countries with substantial public assets (such as sovereign wealth funds), because their net debt could be substantially smaller than their gross debt. Moreover, gross public debt is not reported consistently across MENA countries. As Table II.2 shows, MENA countries do not report many components of public debt. Although the reporting of all sources of public debt might result in higher debt-to-GDP ratios, it is not obvious that debt sustainability analyses would also deteriorate, because what is at stake is the stabilization of the debt-to-GDP ratio over time, even if it is higher than in the reported data. It should be clear by now that macroeconomic data issues in MENA hinder efforts to understand the region’s macroeconomic fragility. Second, Table II.1 shows the required primary balance for 2018 and 2019 after the fact, when data on debt, interest rates and growth are all realized. Estimating the expected future required primary balances, while more meaningful for policy discussions, is more challenging and subject to greater uncertainty. To illustrate the point, we take an example in Appendix C1 where there are two types of public debt: foreign debt (denominated in U.S. dollars) and domestic debt. The expected required primary balance depends on the weighted average of the expected nominal interest rates (in domestic currency) of both the foreign debt and the domestic debt. Without data on the composition of public debt, it would be impossible to calculate the weights of each debt component and hence the expected weighted average nominal interest rate. Furthermore, the expected nominal interest rate of the foreign debt equals the interest rate in dollars multiplied by the expected depreciation of the exchange rate. A larger unexpected depreciation implies larger interest payments in domestic currency terms and hence a larger required primary balance. For example, Alnashar (2019) shows that Egypt’s public debt dynamics are not only driven by fiscal policies, but also exchange rate fluctuations. In sum, lack of debt data transparency impedes meaningful analyses of future fiscal sustainability. Third, even for countries whose required fiscal balance is smaller than the observed balance in 2018 (such as Egypt), the finding should be treated with cautious optimism. As discussed above, current sustainability does not guarantee future sustainability. Interest rates, growth rates, and exchange rates could change and complicate fiscal sustainability. Ì The relationship between debt and the primary fiscal balance In the third approach, fiscal balance sustainability is assessed by estimating the relationship between the primary fiscal balance and debt-to-GDP ratios in previous years (that is, lagged debt). Following Mendoza and Ostry (2008), when the partial correlation between the primary fiscal balance and the previous year’s debt-over-GDP ratio is positive, the fiscal path is interpreted as being “sustainable.”13 Appendix C3 presents the analytical framework in detail and contrasts MENA—including the GCC and developing countries in MENA—with the rest of the world. Figure II.2 shows that for the rest of the world, the primary balance has a negative and statistically significant relationship with lagged debt on average; this suggests that when lagged debt increases, the primary balance deteriorates (see Appendix Table C1).14 This is clearly not a sign of fiscal sustainability, but the trend in recent years points toward greater sustainability compared to the situation at the beginning of the 21st century.15 The primary balance for the GCC 13 This is by no means a guarantee of fiscal sustainability. Rather, this is one of several exercises to examine fiscal sustainability. 14 Figure II.2 corresponds to Appendix Table C1 15 This finding is concerning given the buildup of debt in emerging economies, but it paints a more favorable picture of debt sustainability around the world and particularly for the developing economies of MENA than recent analyses such as Kose and others (2020). Still, it remains to be seen whether the picture would change if we had access to data required to compute the total gross and net debt of the public sector across countries. 25 CHAPTER II HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA countries has an even more negative relationship with lagged debt than the rest of the world, although the relationship is becoming less negative for the GCC (see Figure II.2 The Relationship between Primary Fiscal Balances and Past Debt – MENA and the Rest of the World since 1990 the red line in Figure II.2). The good news is that the primary balance for developing MENA has a positive relationship with the lagged .5 debt, suggesting that their situation is fiscally sustainable. .25 0 Finally, it cannot be overstated that the -.25 findings regarding MENA’s debt sustainability should be interpreted with caution given -.5 the incomplete reporting of debt data. For -.75 good or bad, they might change when more complete data become available. As with lack -1 2011 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2012 2013 2014 2015 2016 2017 2018 of data during a pandemic, obfuscation of year debt information hampers open policy debates GCC Developing MENA in search of solutions. The following chapter Rest of the World turns our attention to labor markets, where inconsistencies in definitions substantially Note: The vertical axis of the figure shows recursive estimates of econometric regressions between primary and lagged debt with other controls (please see Appendix C3 for details). The estimating windows are gradually expanded. For example, the point estimate distort labor market indicators and thus might for 1999 is the result of the estimation with the sample from 1990 until 1999. Similarly, the point estimate for 2018 is the result of the estimation with the sample from 1990 until 2018. But the changes over time in the estimated coefficients are the result of the inclusion of data from the last year in the sample. The bands indicate 10 percent confidence intervals. lead to ineffective policies. Table II.2 Debt Reporting in MENA Countries Source: MENA country World Bank. Note: Table II.2 follows the public debt reporting template of World Bank-IMF’s Debt Sustainability Framework (see IMF, 2017).a indicates the country reports the type of debt (for both domestic and external debts); × indicates the country has the type of debt but does not report it; n/a = not applicable and indicates that the country might not have this type of debt; blank cells indicate that World Bank economists do not have information regarding whether the country has the type of debt but does not report it, or that the country does not have the type of debt, or that the debt might be included in total government debt. Debt reporting is as of 2019. CHAPTER II 26 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 CHAPTER III: DATA GAPS, DEFINITIONS, AND THE MEASUREMENT OF LABOR MARKET OUTCOMES Chapter III takeaways: • MENA countries rely on imprecise definitions of employment, which blur the lines between unemployment and informality. • For countries with independent sources of nationally representative historical labor-force data, the authors were unable to replicate the official reported unemployment rates. • The discrepancies distort the role of women and rural areas in national labor markets. • Using precise definitions of employment and unemployment, statistical evidence suggests that female labor force participation might be a generational issue, because it is high among educated young women. • Historical evidence from an advanced economy and recent evidence from Yemen indicate that female labor force participation tends to rise during periods of armed conflict This chapter studies the role of transparency in the measurement of aggregate labor market outcomes in MENA. It analyzes unemployment rates, female labor force participation rates, and, to a lesser extent, informality. III.1 The Measurement of Unemployment in MENA The labor market is another area in which data gaps and measurement inconsistencies could pose problems to policymaking. Countries around the world usually follow the ILO’s definitions of employment and unemployment, which are considered the gold standard, and are consistent with definitions adopted by other developed countries, such as the United States (see Table III.1). In MENA, official employment and unemployment rates are inconsistently reported across countries. Many MENA countries do not follow the ILO’s definition of employment or unemployment—or do not clearly specify whether they do (see Table III.2). Table III.1. Definitions of Employment and Unemployment from the U.S. BLS, the French INSEE and the ILO Definitions Employment Unemployment Source Someone, aged 16 or over, who has either (1) worked at least 1 hour as a paid employee or (2) in their own Someone, aged 16 or over, https://www. United States -Bureau of business, profession, trade, or farm, or (3) was not who (1) does not have a job, bls.gov/cps/ temporarily absent from their job, business, or farm, (2) has actively looked for one definitions.htm Labor Statistics (BLS) whether or not they were paid for the time off, or (4) in the past 4 weeks, and (3) is worked without pay for a minimum of 15 hours in a available for work. business or farm owned by a member of their family. France- Institut Individuals who worked for any amount of time, even https://www. All people aged 15 and older insee.fr/en/ Nationale de la if only for an hour in the course of the reference week. Only individuals in the working-age population (between who do not have a job and are metadonnees/ statistique et des etudes definitions looking for one. economiques (INSEE) 15 and 64 years of age) are considered. All those of working age (15 years and over) who, during All those of working age (15 a short reference period, were engaged in any activity years and over) who were not to produce goods or provide services for pay or profit. in employment, carried out https://www.ilo. International Labor org/ilostat-files/ Organization (ILO) Other They comprise employed persons "at work", that is, who activities to seek employment Documents/ worked in a job for at least one hour; and employed during a specified recent Statistical%20 Stakeholders persons "not at work" due to temporary absence from period, and were available to Glossary.pdf a job, or to working-time arrangements (such as shift take up employment given a work, flextime and compensatory leave for overtime). job opportunity. 27 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Table III.2. Consistency of Employment and Unemployment Definitions across MENA Follow ILO unemployment Follow ILO employment Age of working Country definition definition population Morocco YES unspecified 15 and above Algeria YES YES unspecified Tunisia YES unspecified unspecified Libya YES unspecified 15 and above Egypt YES YES 15 and above Lebanon YES unspecified YES West Bank and Gaza YES unspecified 15 and above Jordan YES unspecified 15 and above Saudi Arabia NO unspecified 15 and above Oman Unspecified unspecified unspecified UAE YES unspecified YES Qatar YES unspecified 15 and above Bahrain YES YES unspecified Kuwait YES unspecified YES Iran YES YES NO Djibouti NO NO unspecified Iraq unspecified unspecified unspecified Syria and Yemen unavailable unavailable unavailable Source: Authors’ summary based on information from national statistics websites. This section exploits the availability of Labor Market Panel Surveys (which gather data for the same subject over a period of time) for Egypt, Jordan, and Tunisia. It uses them to recompute unemployment rates following the ILO’s definitions and compares the recomputed figures with the national estimates of unemployment reported by the three countries. The results are reported in Tables III.4 to III.6. The analysis relies on the most recent Labor Market Panel Survey in each country. For Egypt, the analysis is for 2018; for Jordan, 2016; and for Tunisia, 2014. The difference between the various computed unemployment rates in these tables relies on the definitions of “employment” and “unemployment.” Employment can be defined according to the market definition, which considers as employed only individuals engaged in market economic activities or to the extended definition, which considers as employed individuals who engage in market and subsistence economic activity. On the other hand, the standard definition of unemployment requires individuals to be actively searching for a job to be considered as unemployed, while the broad definition of unemployment does not require active job search for the individual to be counted as unemployed. According to the broad definition, an individual is unemployed if he or she did not work in the reference period, wanted to work and was not attached to a job (whether or not the individual was actively searching for employment). Table III.3 summarizes the definitions. CHAPTER III 28 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table III.3. Definitions of Employment and Unemployment Definitions Market definition of employment only considers individuals who are employed in market economic activities as employed considers individuals who engage in market and subsistence economic activities as Extended definition of employment employed Standard definition of unemployment requires an individual to be actively searching for employment Broad definition of unemployment does not require an individual to be actively searching for employment Panel A of Table III.4 indicates that in Egypt there is a gap of 5.2 percentage points between the total unemployment rate relying on the market definition and the extended definition. To determine the extent to which this gap is due to a change in female or male unemployment rates, the gender disaggregated unemployment rates are computed relative to the total labor force rather than each gender’s unemployment rate (see Panel B and Panel C).16 The results suggest that 3.8 percentage points, or 73 percent of the gap, is due to definition-based changes in female labor force participation rates, while 1.4 percentage points, or 27 percent of the gap, is due to changes in male unemployment rates. Interestingly, Panel C indicates a great variability in female unemployment rates that stems from the large proportion of Egyptian women who engage in subsistence work. It is no surprise then that a market definition of employment (which excludes subsistence work) suggests that that female unemployment rates in Egypt are very large (ranging between 20 and 30 percent). Furthermore, relative to the various estimated unemployment rates relying on the Egypt Labor Market Panel Survey (ELMPS) data, the official 6.8 percent male unemployment rate published by Egypt’s statistical office, the Central Agency for Public Mobilization and Statistics (CAPMAS), is inflated. In Jordan, Panel A of Table III.5 shows that the gap between the largest computed total unemployment rate according to the market definition (search not required) and the smallest computed total unemployment rate according to the extended definition (search required) is 2.5 percentage points. Unlike in Egypt, where most of this gap was due to changes in female unemployment rates, in Jordan changes in male and female unemployment rates equally contribute to this gap. Under these two definitions, female and male unemployment rates change by 1.2 percentage points and 1.3 percentage points, respectively. Jordan’s male and female unemployment rates are much higher than Egypt’s: while female unemployment rates in Egypt were subject to very large variability depending on the definition of employment (market versus extended), female unemployment rates change only slightly in Jordan, which suggests that Jordanian women do not engage in subsistence work or do so only marginally. It is also important to note that the official unemployment rates derived from the ILO’s ILOSTAT database are consistently lower than those estimated using the Jordan Labor Market Panel Survey (JLMPS). 16 The computed male and female unemployment rates relative to the total labor force are available upon request 29 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Table III.4. Unemployment Rates in Egypt in 2018 Panel A: Total Unemployment Rates Definition of employment Market definition Extended definition No 11.4 8.2 Actively searching requirement Yes 8.2 6.2 National estimate 9.9 Panel B: Male Unemployment Rates Definition of employment Market definition Extended definition No 5.8 5.5 Actively searching requirement Yes 4.9 4.7 National estimate 6.8 Panel C: Female Unemployment Rates Definition of employment Market definition Extended definition No 30.0 13.4 Actively searching requirement Yes 20.6 9.2 National estimate 21.4 Note: This table presents unemployment rates in Egypt according to four definitions using the 2018 Egypt Labor Market Panel Survey (2018), which is a nationally representative survey. The computed unemployment rates are weighted. National estimates refer to official unemployment rates and come from Egypt’s statistical office, the Central Agency for Public Mobilization and Statistics. Table III.5. Unemployment Rates in Jordan in 2016 Panel A: Total Unemployment Rates Definition of employment Market definition Extended definition No 21.3 20.9 Actively searching requirement Yes 19.2 18.8 National estimate 15.3 Panel B: Male Unemployment Rates Definition of employment Market definition Extended definition No 15.7 15.5 Actively searching requirement Yes 14.2 14.1 National estimate 13.3 Panel C: Female Unemployment Rates Definition of employment Market definition Extended definition No 41.2 39.2 Actively searching requirement Yes 37.6 35.7 National estimate 24.1 Note: This table presents unemployment rates in Jordan according to four definitions using the 2016 Jordan Labor Market Panel Survey (2016), which is a nationally representative survey. The computed unemployment rates are weighted. National estimates refer to official unemployment rates and come from International Labor Organization’s ILOSTAT database. CHAPTER III 30 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table III.6. Unemployment Rates in Tunisia in 2014 Panel A: Total Unemployment Rates Definition of employment Market definition Extended definition No 15.8 13.5 Actively searching requirement Yes 12.7 10.9 National estimate 15.0 Panel B: Male Unemployment Rates Definition of employment Market definition Extended definition No 12.6 11.8 Actively searching requirement Yes 10.3 9.7 National estimate 12.3 Panel C: Female Unemployment Rates Definition of employment Market definition Extended definition No 24.5 17.0 Actively searching requirement Yes 19.5 13.6 National estimate 21.6 Note: This table presents unemployment rates in Tunisia according to four definitions using the 2014 Tunisia Labor Market Panel Survey, which is a nationally representative survey. The computed unemployment rates are weighted. National estimates refer to official unemployment rates in 2014 and come from Tunisia’s statistical office, the National Institute of Statistics. For Tunisia, Panel A of Table III.6 indicates that the gap in total unemployment rates between the unemployment rates relying on the market and the extended definition is 4.9 percentage points. An estimation of the gender disaggregated unemployment rates relative to the total labor force reveals that 53 percent of the gap, or 2.6 percentage points, is due to changes in the male unemployment rate and 47 percent, or 2.3 percentage points, is due to changes in the female unemployment rates. While in Egypt the gap between these two estimated unemployment rates was mostly driven by changes in female unemployment rates, in Jordan and Tunisia, differences in male and female unemployment rates contribute equally to this gap. These results are surprising because female labor force participation rates (LFPR) are much lower than that of men, according to both official and our own estimates. That is, since the unemployment rate is computed relative to the sum of employed and unemployed individuals, women are under-represented in the denominator of unemployment rates. As such, an equal contribution to differences in the total unemployment rate belies are much larger impact of the definitions of employment. Indeed, female unemployment rates vary depending on the employment definition we use. This finding therefore suggests that, as in Egypt, women in Tunisia largely engage in non-market economic activities. Official male unemployment rates in Egypt are inflated relative to those estimated using the ELMPS. In Jordan, unemployment rates are consistently lower than official figures. In Tunisia, official unemployment rates are the closest to those computed using the Tunisia Labor Market Panel Survey. It is notable that male unemployment rates are much larger in both Jordan and Tunisia relative 31 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA to Egypt, where male unemployment rates range between 5 and 7 percent. This difference might be the result of unemployment benefits being available in Jordan and Tunisia, while there are none in Egypt.17 Disaggregating unemployment rates across urban and rural areas in Jordan, Tunisia, and Egypt reveals diverse patterns (Figure III.1). In Tunisia and Egypt, male unemployment rates are higher in urban areas; in Jordan, male unemployment rates are higher in rural areas. In Jordan, rural areas consistently have higher unemployment rates. Tunisia and in Egypt have high female unemployment rates in rural areas, which are even larger when the market definition of employment, which excludes subsistence work, is used. The large variability in female unemployment rates in Tunisia and Egypt and the large increase in unemployment rates in rural areas when relying on the market definition of employment are due to the wider prevalence of subsistence work in rural areas. Figure III.1. Unemployment Rates by Urban and Rural Locations Female Unemployment Male Unemployment 50.0 25.0 45.0 40.0 20.0 35.0 Broad unemployment, market Broad unemployment, market 30.0 definition 15.0 definition 25.0 Broad unemployment, extended Broad unemployment, extended definition definition 20.0 10.0 Standard unemployment, market Standard unemployment, market 15.0 definition definition 10.0 Standard unemployment, extended 5.0 Standard unemployment, definition extended definition 5.0 0.0 0.0 Urban Rural Urban Rural Urban Rural Urban Rural Urban Rural Urban Rural Jordan 2016 Tunisia 2014 Egypt 2018 Jordan 2016 Tunisia 2014 Egypt 2018 Source: Egypt Labor Market Panel Survey 2018, Jordan Labor Market Panel 2016, Tunisia Labor Market Panel 2014. Note: The analysis is restricted to working age individuals (15 to 64 years of age). The computed unemployment rates are weighted. Figure III.2. Unemployment Rates by Education Female Unemployment Male Unemployment 50.0 20.0 45.0 18.0 40.0 16.0 35.0 14.0 Broad unemployment, market Broad unemployment, market 30.0 12.0 definition definition 25.0 10.0 Broad unemployment, extended Broad unemployment, extended definition definition 8.0 20.0 Standard unemployment, market Standard unemployment, market 6.0 15.0 definition definition 4.0 Standard unemployment, extended 10.0 Standard unemployment, extended definition definition 2.0 5.0 0.0 0.0 Low High Low High Low High Low High Low High Low High educated educated educated educated educated educated educated educated educated educated educated educated Jordan 2016 Tunisia 2014 Egypt 2018 Jordan 2016 Tunisia 2014 Egypt 2018 Source: Egypt Labor Market Panel Survey 2018, Jordan Labor Market Panel 2016, Tunisia Labor Market Panel 2014. Note: The analysis is restricted to working age men (15 to 64 years of age). The low-educated are those with less than secondary education, while the high-educated are those with secondary education or more. The computed unemployment rates are weighted. Figure III.2 presents disaggregated male and female unemployment rates by educational attainment: low-educated individuals (who have less than a secondary education) and high-educated individuals (who have a secondary education or more). In Tunisia and Egypt, unemployment rates rise as educational attainment rises; in Jordan, unemployment rates are higher for individuals with low educational attainment. Combining this finding with the results from the urban- rural disaggregated data reveals that male unemployment rates in Jordan are higher in rural areas and among the less 17 Across the three countries, using the search-required definition, we find that sons or daughters of the household head accounted for the greatest proportion of unemployed individuals (42 percent in Egypt in 2018, 59 percent in Jordan in 2016, and 74 percent in Tunisia in 2014). On the other hand, 24 percent of unemployed individuals in Egypt, 19 percent in Jordan and 14 percent in Tunisia are household heads, while 31 percent of the unemployed in Egypt are listed as spouses, 19 percent in Jordan, and 9 percent in Tunisia. CHAPTER III 32 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 educated, whereas in Egypt and Tunisia male unemployment rates are higher in urban areas and among the highly educated individuals. As for women, the data suggest that female unemployment rates in the three countries are higher among highly educated women.18 More importantly, there is high variability of female unemployment rates across the definitions of unemployment. This large variance demonstrates how vague definitions of employment can distort the picture of the labor market, particularly for women. Figure III.3. Unemployment Rates by Age Groups Female Unemployment Male Unemployment 70.0 35.0 65.2 31.6 30.6 60.0 30.0 50.0 25.0 42.5 15-24 38.6 38.3 15-24 40.0 25-34 20.0 25-34 30.5 30.7 35-44 15.3 35-44 30.0 45-54 15.0 45-54 11.9 11.7 20.0 55-64 55-64 15.0 15.0 15.4 10.0 7.8 8.1 9.2 6.3 6.6 5.4 10.0 5.0 2.8 1.9 5.2 5.0 0.3 2.1 3.1 3.3 2.1 2.0 0.0 Jordan 2016 Tunisia 2014 Egypt 2018 0.0 Jordan 2016 Tunisia 2014 Egypt 2018 Source: Egypt Labor Market Panel Survey 2018, Jordan Labor Market Panel 2016, Tunisia Labor Market Panel 2014. Note: The analysis is restricted to working age men (between 15 and 64 years of age). Unemployment rates are reported using the market definition of employment (search required). The market definition only considers as employed individuals who engage in market economic activities and excludes subsistence workers. The standard definition of unemployment requires active job search. The computed unemployment rates are weighted. Disaggregating male and female unemployment rates by age groups in Figure III.3 shows that unemployment rates are particularly high among younger cohorts (between 15 and 35 years of age). This is true for both men Figure III.4 Female labor force participation rates and women across the three countries. It is important Female Labor Force Participation Rates to highlight that in all three countries few of the 45.0 unemployed individuals between 15 and 24 years of 39.9 40.0 38.1 age were students at the time of the survey. In Egypt, 35.0 33.7 only 4.2 percent of unemployed men and 1.5 percent 32.5 30.0 of unemployed women in the 15-24 age bracket were 25.5 Broad unemployment, market definition 24.1 studying. In Jordan, it was 0.8 percent of unemployed 25.0 23.4 20.6 Broad unemployment, extended definition men and 1.3 percent of unemployed women. In Tunisia 20.0 17.7 18.5 16.7 17.6 Standard unemployment, market definition Standard unemployment, extended it was 1.9 percent of unemployed men and none of the 15.0 definition unemployed women. 10.0 5.0 III.2 Female Labor-Force Participation: A 0.0 Generational Issue Jordan 2016 Tunisia 2014 Egypt 2018 Female labor force participation (FLFP) rates are Sournce: Egypt Labor Market Panel Survey 2018, Jordan Labor Market Panel 2016, Tunisia Labor Market Panel 2014. generally very low across the three countries.19 Figure Note: The analysis is restricted to working age women (between 15 and 64 years of age). The computed unemployment rates are weighted. III.4 indicates that FLFP rates are lowest in Jordan, where they range between 17 percent and 19 percent, 18 There is no systematic correlation between being highly educated and the wealth score in Egypt, Jordan, or Tunisia. The correlation coefficient between being highly educated (with secondary education and above) and the household wealth score is of 0.31 (P-value=0.00). 19 Generally, due to many economic and cultural reasons, female labor force participation in MENA is low (see World Bank, 2013). 33 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA depending on the definition of employment. Because the incidence of subsistence work among women is particularly high in Tunisia and Egypt, a recalculation of FLFP rates relying on the extended definition of employment shows that FLFP rates are much higher than the official estimates (about 34 percent in Tunisia and 40 percent in Egypt when using the broad unemployment definition). Figure III.5. Predicted Female Labor Force Participation in Egypt by Age Groups Note: The analysis is restricted to working age women (between 15 and 64 years of age). Labor force participation is defined according to the market definition of employment and the standard definition of unemployment. The following variables are used to predict female labor force participation rates: year fixed effects; age-group dummies; three dummies for an individual’s highest level of educational attainment (less than secondary, secondary, and higher than secondary education); a dummy for being married; rural dummy, region fixed effects, three dummies for a father’s highest level of educational attainment, three dummies for a mother’s highest level of educational attainment (less than intermediate, intermediate and above intermediate) and wealth quintile dummies. Figure III.5 shows that Egypt’s and Jordan’s FLFP rates follow an inverted U-shape curve with respect to age, peaking at around 40–45 years in Egypt and around 30–35 in Jordan. These results are in line with Blagrave and Santoro (2017), who use data from Chile and find that labor force participation is low for youth, increases during prime age, and decreases again as retirement approaches. They find that both male and female labor force participation rates follow these patterns in Chile; however, they find a gender gap that persists Figure III.6 Predicted Female Labor Force Participation Rates along the entire life cycle. Using an alternative definition by Age Groups (GMD) of FLFP (following the extended definition of employment and broad definition of unemployment) does not change the inverted U-shape relationship between FLFP and age. It is interesting to highlight the evolution of these predictions in both countries over time. In Egypt, FLFP rates were the highest for each age group in 2006, when Egypt witnessed very high GDP per capita growth rates (approximately 5 percent). Similar patterns are also observed in Jordan where predicted FLFP rates are found to be higher in 2010 and lower in 2016. The average annual growth rate of GDP per capita in Jordan in the 5 years preceding the survey in 2010 was 3 percent, while it was –2 percent in the period between 2011 and 2015 Note: The analysis is restricted to working-age women (between 15 and 64 years of age). An individual is (preceding the 2016 survey). The procyclicality of labor considered a member of the labor force whether or not she is employed. The following variables are used to predict female labor force participation rates: country fixed effects, age groups, level of educational attainment, force participation with respect to the business cycle is in marital status, a rural dummy, the number of children in a household, and electricity and water access. line with Blagrave and Santoro’s results on Chile. CHAPTER III 34 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Box III.1. The World Bank’s Global Micro Database and household data for seven MENA countries TThe World Bank’s Global Micro Database (GMD) provides access to harmonized microdata from around the world. Its main objective is to improve access to socioeconomic statistics that are comparable over time and across countries. The harmonization of this data facilitates statistical research across various years and different countries, permitting us to draw comparable conclusions for the MENA region. This internal World Bank database covers more than 150 countries—representing most of the developing world. Furthermore, GMD’s focus on household survey data on poverty, health, and education is relevant and applicable to labor market research being conducted within the Office of the Chief Economist of MENA. MENA, as a region, faces a significant challenge because of a lack of statistical capacity and data transparency. As a result, some countries’ microdata are publicly unavailable, resulting in a smaller sample of countries from the region. Seven MENA countries are included in the GMD: Djibouti, Egypt, Jordan, Morocco, Tunisia, West Bank & Gaza, and Yemen. For the seven countries there are multiple years’ worth of data, which allows for a more thorough picture through the evolution of households. To paint the most accurate current picture, this analysis is restricted to the most recent data available for each country in the database (Djibouti, 2017; Egypt, 2015; Jordan, 2010; Morocco, 2013; Tunisia, 2015, West Bank & Gaza, 2016; and Yemen, 2014). In Djibouti, a household expenditure survey—Enquête Djiboutienne Auprès des Ménages (EDAM 2017)—was designed and implemented, with fieldwork, concluding in December 2017. In Egypt, the Household Income, Expenditure, and Consumption Survey (HIECS) is a multi-topic survey done every two years, which covers households, demographics, education, employment, consumption of food, ownership of assets, and disability. In Jordan, the Households Income and Expenditure Survey (HIES) covers average income and expenditure for urban and rural households at both the provincial and country level. Similarly, in Morocco, the National Survey on Household Consumption and Expenditure (ENCDM in French) provides a detailed description of living standards and consumption expenditures across different socioeconomic groups. The same can be said for the remaining surveys from Palestine, Tunisia, and Yemen, each of whose objective is to collect accurate socioeconomic data on a consistent basis. The GMD’s role in harmonizing these surveys into a single master dataset facilitates the amount of research done on labor markets and labor conditions in the developing world. Number of female Country Years Survey Name observations in the Percentage latest year of survey Djibouti 2012, 2013, 2017 EDAM 15886 7.61 Egypt 2004, 2008, 2010, 2012, 2015 HIECS 25982 12.44 Jordan 2006, 2008, 2010 HEIS 30521 14.61 Morocco 2000, 2006, 2013 ENCDM, ENNVM 38917 18.63 West Bank & Gaza 2011, 2016 PECS 9998 4.79 Tunisia 2005, 2010, 2015 NSHBCSL 53530 25.63 Yemen 2004, 2014 HBS 34028 16.29 Total 208862 100.00 35 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA The finding is robust to a broader set of countries. Figure III.6 shows the results of a similar analysis using the World Bank’s Global Micro Database (GMD) that compiles household survey data from official sources. The database includes seven MENA countries—Djibouti, Egypt, Jordan, Morocco, Tunisia, West Bank & Gaza, and Yemen – for which World Bank teams have been granted access (see Box III.1). In these surveys, female labor force participation peaks at the 40–44 year-old cohort. The GMD database does not contain sufficient information, however, to allow for the calculation of different definitions of FLFP, rendering impossible the robustness checks of the relationship between FLFP and age for alternative definitions of FLFP. Once again it is abundantly clear that data obfuscation hampers proper analyses of one of the most important policy challenges facing MENA, namely low reported rates of FLFP Women in MENA are becoming more educated, as illustrated in Figure III.7, which shows the highest level of educational attainment for each age group in Egypt, Jordan, and Tunisia. In all three countries, the dark blue bar, which represents the share of women with no educational degree, is the smallest among those between ages 25 and 34 of age and is larger among the older age groups. This suggests that lack of education is particularly high among older women. For instance, in Tunisia, 88 percent of women between 55 and 64 years of age have no educational degree, while 32 percent of women between 25 and 34 years of age report having no education. On the other hand, the share of women with secondary education or less and the share of women with post-secondary education are found to be the highest among younger women. We also observe similar phenomenon using the GMD. Ì The Roles of Education and Family: Results from Blinder-Oaxaca Decompositions. To systematically examine the differences in Figure III.7 Women’s Educational Attainment by Age Cohorts female labor force participation rates, the gap in female labor force participation rates 100 7 2 1 between younger cohorts (25–35 years of age) 11 12 8 11 8 15 14 26 24 21 19 25 19 20 21 and older cohorts (36– 64 years of age) is 80 42 23 42 33 42 decomposed using a technique based on ones 46 60 41 39 47 44 51 52 proposed separately by Blinder (1973) and 51 40 37 79 88 Oaxaca (1973). It decomposes differences in 72 62 52 55 50 labor force participation between two groups, 20 43 21 33 37 32 24 38 28 34 based on linear regression models, into those 0 “explained” by observable characteristics 25-34 35-44 45-54 55-64 25-34 35-44 45-54 55-64 25-34 35-44 45-54 55-64 25-34 35-44 45-54 55-64 Jordan 2016 Tunisia 2014 Egypt 2018 GMD-7 and those “unexplained” by them. This section assesses whether the low female labor force participation rates in the region No education Secondary or less Above secondary are a generational issue, because women’s Source: Jordan Labor Market Panel Survey 2016; Tunisia Labor Market Panel Survey 2014; Egypt Labor Market Panel Survey 2018; education in the MENA region has been on World Bank Global Micro Database. Note: The GMD–7 are Djibouti, Egypt, Jordan, Morocco, Tunisia, West Bank & Gaza, and Yemen. the rise. That is, as women become more educated, will female labor force participation rates increase? The explained differences between younger and older women fall into three areas: education, family (which includes a woman’s marital status as well as the number of children below 19 years old in a household), and “other factors,” such as wealth, region of residence, urban or rural area of residence, and parental education. CHAPTER III 36 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Columns (1) and (2) of Table III.7 report the results for Egypt in 2018. FLFP rates are comparable between the two groups—approximately 25 percent. By contrast, in Jordan, the difference in FLFP rates between younger and older cohorts is 16 percentage points—29.1 percent for younger women, 13.3 percent for older women. The decomposition results show that 9 percentage points of this gap is due to observed differences between the two groups, while 7 percentage points are unexplained. Interestingly, roughly half, of the gap is the result of educational differences between the two groups. The results for Tunisia are more striking. The FLFP rate for the younger cohort is 34.8 percent compared with 21.4 percent for the older cohort. The gap in female labor force participation rates between the younger and older age cohort is approximately 13 percentage points. The Blinder-Oaxaca decomposition results show that the entire gap is due to observed differences in the two groups’ characteristics. More important, the gap is almost solely explained by education differences between younger and older women (12 percentage points). Finally, data from the GMD for the seven MENA countries present a similar story. The FLFP rate for the younger cohort is 35 percent compared to 27 percent for the older cohort. A large part of this gap is explained by education (5.3 percentage points) and family factors (4.7 percentage points). Table III.7. Decomposing the Gap in FLFP Rates between Younger and Older Cohorts Egypt 2018 Jordan 2016 Tunisia 2014 GMD–7 Variables Overall Explained Overall Explained Overall Explained Overall Explained Old cohort 0.251*** 0.133*** 0.214*** 0.274*** [0.005] [0.006] [0.008] [0.002] Young cohort 0.244*** 0.291*** 0.348*** 0.349*** [0.006] [0.009] [0.015] [0.003] Difference 0.007 -0.158*** -0.134*** -0.075*** [0.008] [0.010] [0.017] [0.003] Explained -0.041*** -0.085*** -0.133*** -0.096*** [0.005] [0.007] [0.011] [0.002] Unexplained 0.048*** -0.073*** -0.001 0.020*** [0.009] [0.010] [0.018] [0.003] Education -0.056*** -0.070*** -0.123*** -0.053*** [0.004] [0.005] [0.010] [0.001] Family 0.023*** -0.001 -0.016*** -0.047*** [0.003] [0.003] [0.005] [0.001] Other factors -0.008*** -0.014*** 0.007* 0.005*** [0.002] [0.004] [0.004] [0.001] Number of Observations 12,864 12,864 6,359 6,359 3,466 3,466 85,350 85,350 Source: Jordan Labor Market Panel Survey 2016; Tunisia Labor Market Panel Survey 2014; Egypt Labor Market Panel Survey 2018; World Bank Global Micro Database. Note. This table relies on the Blinder-Oaxaca technique to decompose the gap in female labor force participation rates between younger women (between 25 and 35 years of age) and older women (between 36 and 64 years of age) The GMD–7 are Djibouti, Egypt, Jordan, Morocco, Tunisia, West Bank & Gaza, and Yemen. The regressions include the following variables: a dummy variable for being married, the number of children aged less than 19 years old, three dummies for an individual’s educational attainment (no educational degree, secondary education or less and above secondary education), a rural dummy, region fixed effects, three dummies for a mother’s highest level of educational attainment, three dummies for a father’s highest level of educational attainment (less than intermediate, intermediate and above intermediate) and wealth quintile dummies. The education vector therefore includes three dummies for the individual’s highest level of educational attainment and a family vector includes a married dummy and the number of children at the household level. 37 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA III.3 The Missing Piece: Measuring Informality in MENA Table III.8. Informal Employment in Egypt, Jordan and Tunisia EGYPT 2018 JORDAN 2016 TUNISIA 2014 Definition Male informal employment No work contract or no social 79.4 64.8 71.5 (% of total employment) security Female informal employment No work contract or no social 62.8 43.7 68.4 (% of total employment) security Male informal employment No work contract and no social 62.8 29.0 44.3 (% of total employment) security Female informal employment No work contract and no social 30.0 13.5 33.9 (% of total employment) security Male informal employment 67.3 46.5 57.4 No work contract (% of total employment) Female informal employment 32.2 32.1 39.0 No work contract (% of total employment) Male informal employment 69.3 41.6 50.3 No social security (% of total employment) Female informal employment 41.1 22.8 48.2 No social security (% of total employment) Sources: The Egypt Labor Market Panel Survey in 2018, the Jordan Labor Market Panel Survey in 2016, and the Tunisia Labor Market Panel Survey in 2014. Informal employment in MENA is a challenging issue. Table III.8 shows informal employment as percent of total employment for both women and men in Egypt, Jordan, and Tunisia. Four definitions of informality are explored. An individual is considered informal if he or she has: • no work contract or no social security • no work contract and no social security • no work contract • no social security. Informality rates are the highest when relying on the first definition—no work contract or no social security. The second definition, no work contract and no social security, yields the smallest estimated informality rates. Under the second definition, male informality rates are the highest in Egypt (63 percent), followed by Tunisia (44 percent) and Jordan (30 percent). The highest female informality rates are observed in Tunisia (34 percent), followed by Egypt (30 percent), while the lowest female informality rates (13.5 percent) are observed in Jordan. Using the third or the fourth definitions which define informality based on having a work contract or based on having social security, respectively, yields informality rates that are somewhat in the middle range between those estimated using the most restrictive and least restrictive definitions. CHAPTER III 38 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 III.4 Conflict and Female Labor Force Participation Figure III.8. Labor Force Participation Rates in the United States since 1890 One of modern feminism’s most recognizable images is that of Rosie the 100 Riveter, Howard Miller’s 1942 poster 90 of a confident female worker sporting 80 a polka-dotted red bandana, sleeve 70 rolled up to showcase a flexed bicep 60 while encouraging women to work, 50 saying “We can Do It!” Given the low 40 levels of female labor force participation 30 in the early 20th century, this image 20 is a reminder of the ways in which 10 war may distort the composition of a 0 country’s labor force. Of the 16 million 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 American men who joined the Armed ▬ Female ▬ Male Forces during World War II, 73 percent Sources: Acemoglu and Auto (2004) until 1990 and World Development Indicators for 2000 and 2010. Note: Data include individuals over 14 years of age prior to 1950; over 16 from 1950–1990 and over 15 since 2000 were deployed overseas (Acemoglu and Autor, 2004). Given the concurrence of a shortage of male labor driven by Figure III.9. Labor Force Participation in Yemen, 1990-2019 military conscription and heightened Year on Year Changes in Labor Force Participation in Yemen demand for military equipment driven 0.15 by war, employers turned to women to fill industrial vacancies (Milkman 1982). As a result, female labor force 0.1 participation increased by about 50 percent—with an estimated 6.7 million 0.05 women entering the labor market during the war (Rose, 2018 and Figure 0 III.8). For those reasons, World War II is largely considered to have created a sea change in female labor force -0.05 participation in the United States. On a paradigmatic and societal level, the -0.1 decrease in male labor supply led to 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 the overturn of “marriage bars.” Prior ▬ Male ▬ Female to the 1940’s, and particularly during Source: Author’s calculations based on data from the World Development Indicators. Depression, employers did not hire married women and fired single women who married (Goldin, 1991). While many of the women who entered the labor force due to the war eventually dropped out of the labor force by the 1950s, the largest proportional increase in female labor force participation is attributable to the 1940s, according to Acemoglu and Autor. The positive shift in female labor supply occurred for a variety of reasons, some circumstantial. For married women, the income effect of a conscripted husband, as well as the reduction of housekeeping duties, increased their 39 CHAPTER III HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA probability of entering the labor market. Patriotism could have also played a role in women’s decision to work and support the war effort. Acemoglu and Auto study women’s labor supply by exploiting differences in WWII mobilization rates across different states. They find that in 1950 female labor force participation rates were higher in states with higher rates of working-age male conscription in 1940. The impact of conflict on female labor force participation in MENA is probably diverse and country specific. In Yemen, some evidence suggests that conflict has led to an increase in female labor force participation. Historically, female labor force participation in Yemen has been low relative both to male labor force participation and to female labor force participation rates in countries with similar levels of income per capita. The income effect of men's imprisonment or participation in active military combat during the wars of 1994 and 2015 in Yemen led to an increase in the number of female-headed households, as is illustrated by the increases in the annual growth rates of female labor force participation during these years (see Figure III.9). A deeper analysis of the impact of the war on labor composition J. Howard Miller/War Production Co-Coordinating Committee, “"We Can Do It!" HERB: Resources for Teachers, accessed February 4, 2020, https://herb.ashp.cuny.edu/items/show/1192. in the country is required; however, the lack of comprehensive official labor statistics, including sectoral data, makes this difficult. The most accessible sources of labor statistics in Yemen are the ILO’s ILOSTAT database and the World Bank’s World Development Indicators, both of which are based on modeled estimates and projections. Prior to the 2011 Syrian conflict, female labor force participation in Syria was about 13 percent. A study by the RAND Corporation suggests that female Syrian refugees significantly increased their labor force participation rates after leaving Syria, with rates as high as 25 percent in Turkey and Lebanon and 50 percent in Jordan (Constant and others, 2019). While this increase happened outside the conflict zone, it was arguably triggered by changing socioeconomic conditions that arose from the conflict. Women and men surveyed in Constant and others suggested that female work opportunities were empowering and more socially accepted. Iraq is an interesting example of the ways in which conflict can affect female labor force participation in diverse ways. Before the 1991 Gulf War, female labor force participation in Iraq was increasing— especially during the Iran-Iraq war (1980-88), when a shortage of working-age men led to an increase in female participation in the labor market and in civil service (Human Rights Watch, 2003). Following the Gulf War, a combination of economic, legal, and political factors reversed the advancement of women in Iraq. For instance, UN sanctions on Iraq disproportionately affected women and children, especially young girls, because when faced with economic difficulties, families sent their boys to school and kept their girls at home, according to Human Rights Watch. This suggests that while conflict may induce women to join the labor force, the ultimate effect of conflict on female labor force participation in MENA can be hard to predict. CHAPTER III 40 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 CHAPTER IV. SUMMARY OF FINDINGS World Bank economists expect output of MENA to decline in 2020. This is in sharp contrast to the forecast in October 2019 when the regional economies were expected to grow at 2.6 percent this year. The 3.7 percentage point growth downgrade reflects the costs associated with the Covid-19 and oil price collapse dual shocks. The growth forecasts do not change the picture of the region’s struggle with the triplet challenges of lackluster long- term growth of GDP per capita, macroeconomic fragility and poor labor market outcomes. The report argues that the region’s lack of data and transparency contributed to these long-term outcomes. Many MENA countries are lagging in their data generation capacity or have prevented access to data. Reliable, accessible data combined with clear definitions of key indicators, the essential ingredients of transparency, not only helps improve policymaking over time but also reduce distrust and social unrest. Indeed, the report finds a strong and positive empirical association between a country’s statistical transparency and its subsequent long-run economic growth. The report highlights two areas– macroeconomic fragility and labor market measures - where the lack of data and transparency in MENA countries weakens credibility and hampers good policymaking. First, the report presents a battery of tests on current account and fiscal account vulnerabilities in MENA and the rest of the world. Whereas the analyses of the sustainability of external imbalances is plausibly immune to lack of transparency, the credibility of any debt-sustainability analyses critically depends on data transparency. MENA countries vary greatly in their public debt reporting. For many types of public debt, World Bank economists and other external analysts do not have access to relevant information. Consequently, analytical results about MENA’s macroeconomic fragility have to be taken with a grain of salt given that disclosure of relevant data can considerably change the findings, both good and bad. Second, current analyses of labor market outcomes in MENA are subject to data and definition inconsistencies. MENA countries rely on various definitions of employment, which affect unemployment and informality indicators, with little harmonization within the MENA as well as with international standards. It was not possible to replicate officially reported unemployment rates using independent sources of nationally representative historical labor-force data in three countries where they are available. In the meantime, gaining access to official labor force surveys, upon which official reported unemployment rates rest, remains a daunting task for the World Bank, needless to say to MENA’s broader civil society. Better data and transparency is needed in MENA countries to enable credible economic analyses and informed public discourse, which arguably lead to better policymaking, enhanced societal trust and faster economic development in the long run. Data for development is a priority agenda for the World Bank as evidenced by the upcoming World Development Report 2021 that will be dedicated to the role of data in development. The World Bank can serve as a catalyst for better data in MENA. It can provide arguments and evidence about how data transparency can aidpolicymaking. Benchmarking countries’ statistical capacity could provide a powerful incentive for countries to improve their data ecosystem. The World Bank has and can continue to provide technical assistance to improve countries’ statistical capacity, such as helping with conducting firm and household level surveys, or digitizing data for public access while also anonymizing micro data to protect citizen and enterprise privacy. The World Bank also benefits directly from better statistical capacity, as better data enables more accurate economic analyses and consequently improves the quality of policy recommendations. The evidence presented in this report and future collaboration can help us to make transform MENA into a more transparent, prosperous and peaceful society in the hopefully near future. 41 SUMARY OF FINDINGS HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA References Acemoglu, Daron and David H. Autor. 2004. “Women, War, and Wages: The Effect of Female Labor Supply on the Wage Structure at Midcentury.” Journal of Political Economy 11(3): 497-551. Alnashar, Sara. 2019. “What are the drivers of Egypt’s government debt?” ERF Working Paper No. 1376 An, Zidong, João Tovar Jalles and Prakash Loungani. 2018. "How Well Do Economists Forecast Recessions?" IMF Working Papers 18/39 International Monetary Fund. Arezki, Rabah; Lederman, Daniel; Abou Harb, Amani; Fan, Rachel Yuting; Nguyen, Ha. 2019. “Reforms and External Imbalances: The Labor-Productivity Connection in the Middle East and North Africa” Middle East and North Africa Economic Update (April), World Bank, Washington, DC. Arezki, Rabah; Ait Ali Slimane, Meriem; Barone, Andrea; Decker, Klaus; Detter, Dag; Fan, Rachel Yuting; Nguyen, Ha; Miralles Murciego, Graciela; Senbet, Lemma. 2020. “Reaching New Heights: Promoting Fair Competition in the Middle East and North Africa.” Middle East and North Africa Economic Update (October), Washington, DC: World Bank Arezki, Rabah and Rachel Yuting Fan. 2020. “Oil price wars in a time of Covid-19”. VoxEU blog. https://voxeu.org/article/ oil-price-wars-time-covid-19 Arezki, Rabah and Ha Nguyen. 2020. “Coping with a Dual Shock: COVID-19 and Oil Prices”. In Baldwin, Richard and Beatrice Weder di Mauro (Ed.) Economics in the Time of COVID-19. CEPR Press VoxEU.org eBook Ball, Carolyn. 2009. “What is Transparency?” Public Integrity, 11 (4): 293-308 BBC. 2020. Andrew Marr Show: Interview with South Korean Foreign Minister. http://news.bbc.co.uk/2/shared/bsp/hi/ pdfs/15032002.pdf Blinder, Alan S. 1973. "Wage Discrimination: Reduced Form and Structural Estimates". Journal of Human Resources. 8 (4): 436–455 Blagrave, Patrick and Marika Santoro. 2017. “Labor Force Participation in Chile: Recent Trends, Drivers, and Prospects”. IMF Working Paper WP/17/54 Cady, John. 2005. "Does SDDS Subscription Reduce Borrowing Costs for Emerging Market Economies?" IMF Staff Papers, 52(3): 1-6. Clementi, Fabio, Haider Khan, Vasco Molini, Francesco Schettino and Khalid Soudi. 2019. "Polarization and Its Discontents: Morocco before and after the Arab Spring." World Bank Policy Research Working Paper Series 9049. Constant, Louay, Shanthi Nataraj, and Fadia Afashe. 2019. “As Refugees Syrian Women Find Liberation in Working.” RAND Corporation. Accessed 30 January 2020. https://www.rand.org/blog/2019/02/as-refugees-syrian-women-find- liberation-in-working.html Correia, Sergio, Stephan Luck and Emil Verner. 2020. “Pandemics Depress the Economy, Public Health Interventions Do Not: Evidence from the 1918 Flu”, mimeo Das, Jishnu, Quy-Toan Do, Karen Shaines, and Sowmya Srikant. 2013. "U.S. and them: The Geography of Academic Research" Journal of Development Economics 105: 112–130 Debrun, Xavier, Jonathan D. Ostry, Tim Willems, and Charles Wyplosz. 2019. “Public Debt Sustainability.” CEPR Discussion Paper DP14010. Finel, Bernard I., and Kristin M. Lord. 1999. “The Surprising Logic of Transparency.” International Studies Quarterly 43, no. 2:315–339. Focus Economics. 2020. Focus Economics Consensus Forecast: Middle East & North Africa – March 2020 Foreign Affairs. 2020. “How Civic Technology Can Help Stop a Pandemic” (March 20, 2020). https://www.foreignaffairs. com/articles/asia/2020-03-20/how-civic-technology-can-help-stop-pandemic Gali, Jordi. 2020. “Helicopter money: the time is now”. VoxEU blog. https://voxeu.org/article/helicopter-money-time-now REFERENCES 42 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Goldin, Claudia D. 1991. “The Role of World War II in the Rise of Women's Employment.” The American Economic Review 81(4):741-756. Hoogeveen, Johannes. 2018. “A social contract indicator for sub-Sahara Africa”. Unpublished. Hollyer, James, B. Peter Rosendorff, and James Raymond Vreeland. 2011. “Democracy and Transparency.” Journal of Politics 73 (4): 1191–205. Human Rights Watch. 2003. “Background on Women’s Status in Iraq Prior to the Fall of the Saddam Hussein Government.” Accessed 30 January 2020. https://www.hrw.org/report/2003/11/21/background-womens-status-iraq-prior-fall-saddam- hussein-government Islam, Roumeen. 2006. “Does More Transparency Go along with Better Governance?” Economics and Politics 18 (2): 121–67. Ilzetzki, Ethan, Carmen Reinhart and Kenneth Rogoff. 2019. “Exchange Arrangements Entering the 21st Century: Which Anchor Will Hold?” Quarterly Journal of Economics, 134:2, 599–646 International Monetary Fund (IMF). 2011. “When and How to Adjust Beyond the Business Cycle? A Guide to Structural Fiscal Balances”. Technical Notes and Manuals. IMF: Fiscal Affairs Department. ———2013. “External Balance Assessment (EBA) Methodology” ———2017. “Guidance note on the Bank-Fund Debt Sustainability Framework for Low Income Countries” Lopez-Cordova, Ernesto. 2020. “A Slowdown of China’s Economy and its Impact on the Demand for Tourism Services”, Brief Kose, M. Ayhan; Peter Nagle, Franziska Ohnsorge, and Naotaka Sugawara. 2020. Global Waves of Debt: Causes and Consequences. Washington, DC: World Bank. Kubota, Megumi and Albert Zeufack. 2020. “Assessing the Returns on Investment in Data Openness and Transparency”. Policy Research Working Paper WPS 9136. Washington, D.C.: World Bank Group. Mendoza, Enrique G. and Jonathan D. Ostry. 2008. "International evidence on fiscal solvency: Is fiscal policy "responsible"?” Journal of Monetary Economics 55(6): 1081-1093. Milkman, Ruth. 1982. “Redefining "Women's Work": The Sexual Division of Labor in the Auto Industry during World War II.” Feminist Studies 8(2):336-372. Mitchell, Ronald B. 1998. “Sources of Transparency: Information Systems in International Regimes.” International Studies Quarterly 42(1):109–130. Oaxaca, Ronald. 1973. "Male-Female Wage Differentials in Urban Labor Markets". International Economic Review. 14 (3): 693–709 Reuters, 2019. “Lebanon a 'beautiful idea' in need of a reboot, say protesters” (November 7, 2019). https://www.reuters. com/article/us-lebanon-protests-nation/lebanon-a-beautiful-idea-in-need-of-a-reboot-say-protesters-idUSKBN1XH1KC Rose, Evan K. 2018. “The Rise and Fall of Female Labor Force Participation during World War II in the United States.” The Journal of Economic History 78(3):1-39. United Nations. 2017. “World Population Prospects: The 2017 Revision” Department of Economic and Social Affairs, Population Division. Vegh, C A, G Vuletin, D Riera-Crichton, J P Medina, D Friedheim, L Morano, and L Venturi. 2018. From Known Unknowns to Black Swans: How to Manage Risk in Latin America and the Caribbean. LAC Semiannual Report, World Bank. Williams, Andrew. 2009. “On the Release of Information by Governments: Causes and Consequences.” Journal of Development Economics 89: 124–38. World Bank. 2013. Opening Doors: Gender Equality and Development in the Middle East and North Africa. Washington, DC: World Bank. Reuters. 2020. “Iranian oil minister confirms OPEC agreed a 1.5 million bpd cut”. https://www.reuters.com/article/us- opec-meeting-iran/iranian-oil-minister-confirms-opec-agreed-a-15-million-bpd-cut-idUSKBN20S1JA 43 REFERENCES HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Appendix Appendix AppendixA: Appendix E s tim a tin g t A: Estimating he R the elationship bbetween Relationship ta tistic a l C a etween SStatistical d Eco pacity anand Capacity no m ic Economic G ro w th Growth The relationship between the statistical capacity index is explored using several estimation models. The sample consists of 146 economies between 2005 and 2018. Data for West Bank and Gaza are available only from 2009 to 2018. A cross-sectional model is estimated using the log difference of GDP per capita between 2005 and 2018 as the outcome variable regressed on the 2005 levels of the statistical capacity index and other covariates (Model: OLS - Cross-section) as presented in in equation (A1). !"#$ℎ!,#$%&' = ( + *& +,$,!,#$ + *( !+-./!,#$ + *& 01!!,#$ + *& 2!!,#$ + *& 3/ℎ!,#$ + *& 4",56!,#$ + *) -789:$!,#$ + ;! (21) where Grwth is GDP per capita growth (log difference) for economy i between 2005 and 2018. Data represents the overall statistical capacity of economy i in 2005. All explanatory variables are for the year 2005. These include the level of GDP per capita (GDPpc), the sectoral composition of the economy—share of manufacturing value added over GDP (MFG) and share of agricultural value added over GDP (AG), primary school enrollment (Sch), and trade as a share of GDP (Trade). Polnst is a vector of political institutions obtained from the World Governance Indicators that include political stability, voice and accountability, rule of law, and control of corruption. Finally, ε is the error term. An alternate estimation can be obtained by employing panel estimations using random country effects and year fixed effects (equation A2). Grwth*+ = α + β& Data*,+ + β( GDPpc*,+%& + β, MFG*,+ + β- AG*,+ + β$ Sch*,+ + β. Trade*,+ + β/ Polnst *,+ + ν* + τ+ + ε*+ (A2) Grwth is the log difference in GDP per capita for economy i between time t and t − 1. Data is the overall statistical capacity of economy i at time t. GDPpc*,+%& is the log GDP per capita in time t − 1. All the other covariates are the same as in equation A1 at time t. Finally, ν* is the random country effect, τ+ is the year fixed effect, and ε*+ is the error term (Model: Country RE & Year FE). Alternatively, we estimate equation A2 with lagged statistical capacity and other covariates as presented in equation A3 (Model: Country RE & Year FE– (all covariates lagged)). Grwth*+ = α + β& Data*,+%& + β( GDPpc*,+%& + β, MFG*,+%& + β- AG*,+%& + β$ Sch*,+%& + β. Trade*,+%& + β/ Polnst *,+%& + ν* + τ+ + ε*+ (A3) Furthermore, given that the country-level effects are likely to be correlated with the lagged GDP per capita resulting in inconsistent estimates, we estimate systems GMM dynamic panel estimators with specifications where the statistical capacity index and other covariates are either included concurrently (Model: Systems GMM) or lagged (Model: Systems GMM - (all covariates lagged)). Across all models and specifications, the results show a positive relationship between statistical capacity and economic growth. Regression results are available upon request. The macro-economic loss due to the decline in the statistical capacity index in MENA is reported for all the models in Table A1. APENDIX A 60 44 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table A1. Macro-economic Loss in GDP due to Statistical Capacity Index Decline in MENA (2005-2018) Model Loss in GDP per Capita (%) Country RE & Year FE 7.4 Country RE & Year FE – (all covariates lagged) 9.2 Systems GMM 10.6 Systems GMM– (all covariates lagged) 9.9 OLS - Cross-section 13.8 *Regression results and calculations are available upon request from authors. Model details explained in Appendix A. Table A2. Definitions of the Statistical Capacity Measure Statistical Capacity Measure (2005-2018) Definitions Average of three sub-indicators: Source data, Methodology , and Periodicity and Overall Statistical capacity (0 – 100) timeliness of socioeconomic indicators Called "Source data assessment of statistical capacity" in the statistical capacity dataset. Source data reflects whether a country conducts data collection activity in line with internationally recommended periodicity, and whether data from Source data (micro data availability and administrative systems are available and reliable for statistical estimation purposes. periodicity) (scale 0 - 100) Specifically, the criteria used are the periodicity of population and agricultural censuses, the periodicity of poverty and health related surveys, and completeness of vital registration system coverage. Called “Methodology assessment of statistical capacity" in the statistical capacity dataset. Statistical methodology measures a country’s ability to adhere to internationally recommended standards and methods. This aspect is captured by assessing guidelines and procedures used to compile macroeconomic statistics and Methodology (international standards) social data reporting and estimation practices. Countries are evaluated against a (scale 0 - 100) set of criteria such as use of an updated national accounts base year, use of the latest balance of payments manual, external debt reporting status, subscription to International Monetary Fund’s Special Data Dissemination Standard, and enrollment data reporting to the United Nations Educational, Scientific, and Cultural Organization. Called "Periodicity and timeliness assessment of statistical capacity" in the statistical capacity dataset. Periodicity and timeliness looks at the availability and periodicity of key socioeconomic indicators, of which nine are MDG indicators. This Periodicity and timeliness of key dimension attempts to measure the extent to which data are made accessible to socioeconomic indicators (scale 0 - 100) users through transformation of source data into timely statistical outputs. Criteria used include indicators on income poverty, child and maternal health, HIV/AIDS, primary completion, gender equality, access to water and GDP growth. 45 APPENDIX A HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA B: MNACE’s AppendixB: Appendix urrent AAccount MNACE’s CCurrent o del ccount MModel The framework of this part is based on the IMF’s External Balance Assessment (2013). Using data from various sources, we assembled a panel dataset of major economic indicators for the world’s economies.19 Specifically, we set out to identify current account imbalances that cannot be explained by a country’s fundamental indicators. To do so, we ran the following regression: CA*,+ = β# + YoungDep*,+ + OldDep*,+ + AgingSpeed*,+ + EGrowth*,+ + RelativeGDP*,+%& + RelativeGDP*,+%& ∗ ChinnIto*,+ + ChinnIto*,+ + ∆NetComPI*,+%& + ∆NetComPI*,+ + fe* + fe+ + ε*,+ (B1) The dependent variable, CA*,+ , is the current account balance as a percentage of GDP. Data are from the World Economic Outlook (WEO). YoungDep*,+ captures the percentage of young dependents (less than 15 years of age) to the working population (15–64 years of age). The regression also includes old-age dependency, OldDep*,+ which captures the percentage of dependents aged 64 years or older to the working population. It also includes an aging speed variable which measures the annual change in the old-age dependency. Data are from the United Nations (2017) EGrowth*,+ captures a country’s expected growth acceleration, by taking the difference between the growth forecast for the following year and the growth forecast for the current year. Data for the growth forecast are from the historical forecasts of the WEO. RelativeGDP*,+%& is a country’s real GDP per worker (in purchasing power parity, or PPP) relative to that of the United States at time t-1. It captures relative productivity. RelativeGDP*,+%& ∗ ChinnIto*,+ captures the idea is that capital flows to poor countries also depend on a country’s financial openness. Data of real PPP GDP are from the WEO. The Chinn-Ito Index is a measure of capital account openness (Chinn and Ito 2006). ∆NetComPI*,+ and ∆NetComPI*,+%& are the log change in the commodity price index and its first lag. The variable is constructed as follows. First, following Bruckner and Arezki (2012), a commodity price index is 1! calculated as ComPI+ = ∏2 Price0,+ where θ0 is the long-run exposure of the country to commodity c. θ0 is calculated as the average share of the country’s net exports of commodity c over country GDP. Price0,+ is the world price of commodity c at time t. Next, again following Bruckner and Arezki, we generate the change in the commodity price index as ∆log (ComPI)+ =log(ComPI+ ) − log (ComPI+%& ). Commodity prices are from the IMF; trade data are from UN Comtrade; and nominal GDP data are from World Development Indicators. fe* are country fixed effects; fe+ are time fixed effects. ε*,+ is the residual of the regression, which is the “unexplained” component of the current account. Exchange rate regime data are from Ilzetzki and others (2019). We recode their classification of 1 as “fixed exchange rate regimes,” 2 and 3 as “managed floats” and 4 and 5 as “free floats.” See Table B1 for the summary statistics. 19 We retrieved the data from the following sources: the World Economic Forum, the World Development Indicators, the Penn World Table, Chinn and Ito (2006) and the U.S. Federal Reserve. 62 APPENDIXB 46 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table B1. Summary Statistics World Number of N Mean Mediam Min Max countries Young Age Dependency % 189 11083 62.062 65.791 14.898 113.702 Old Age Dependency % 188 11083 10.186 7.513 .874 37.5 Table B1 Summary Statistics Aging WorldSpeed % 188 10897 .095 .046 -1.154 1.563 Predicted Changes in Growth 188 Number5203 of N.457 Mean .213 Median -230.834Min 151.992Max GDP/worker relative to USA (t-1) 181 countries 6619 .354 .2 .011 6.452 ChinnYoung Age Dependency % Ito Index 175 189 7319 11083.452 62.062 .416 65.791 0 14.898 113.702 1 Old Age Dependency % 188 11083 10.186 7.513 .874 37.5 Log Change in net commodity Aging Speed % price (t) 188 6470 188 10897 0 .095 0 .046 -.357 -1.154 .28 1.563 Predicted Changes in Managed Float Exchange Rate Growth 183 188 10430 5203.356 .457 0 .213 0 -230.834 151.992 1 GDP/worker relative to USA (t-1) 181 6619 .354 .2 .011 6.452 Exchange Float Chinn Rate Ito Index 183 10430 175 7319.075 .452 0 .416 0 0 1 1 MENA Log Change in net commodity price (t) 188 6470 0 0 -.357 .28 Managed Float Exchange Rate 183 10430 .356 0 0 1 Number of Float Exchange Rate countries N 183 10430 Mean .075 Mediam Min0 0 Max 1 Young Age Dependency % 19 1199 65.64 69.678 15.237 113.702 MENA Old Age Dependency % 19 1199 6.013 6.043 .874 15.748 Aging Speed % 19 Number of 1180 N.024 Mean.015 Median -.653 Min Max .602 countries Predicted Changes in Growth 18 561 .145 .198 -230.834 151.992 Young Age Dependency % 19 1199 65.64 69.678 15.237 113.702 GDP/worker relative to USA Old Age Dependency % (t-1) 18 19689 1199.719 6.013.313 6.043 .04 .874 6.452 15.748 Aging Chinn Ito Speed Index % 19 19934 1180.561 .024.656 .015 0 -.653 1 .602 Predicted Changes in Growth 18 561 .145 .198 -230.834 151.992 Log Change in net commodity price (t) 18 782 .001 0 -.315 .236 GDP/worker relative to USA (t-1) 18 689 .719 .313 .04 6.452 Managed Float Chinn Ito Exchange Rate Index 19 1172 19 934.397 .561 0 .656 0 0 1 1 Float LogExchange Change Rate 19 in net commodity price (t) 1172 18 782.015 .001 0 0 0 -.315 1 .236 Managed Float Exchange Rate 19 1172 .397 0 0 1 Float Exchange Rate 19 1172 .015 0 0 1 The MNACE current-account determinants model has three specifications. The within specification has both time and country fixed effects. The time fixed effects capture the effects of common world factors in a given year on all countries’ current account positions. The country fixed effects capture the effects of unobservable country- specific time-variant factors (such as consumption preferences) on each country's current account position. This specification considers the effects of the fundamentals within a country. The pooled specification has only time fixed effects and no country fixed effects. This allows us to examine the effects of fundamentals on current account positions across countries as well as over time. The between specification takes the average of the current account position and the fundamentals across years within a country, and then examines the effect of the average fundamentals on the average current account across countries. The residuals of the regressions reflect the portion of the the current accounts that are unexplained by the fundamentals. In the pooled specification (the first column of Table B2), the fundamentals have the expected signs. The coefficients of young age and old age dependencies are negative. The coefficient of –0.148 implies that a 1 percent increase in old age dependency is associated with a 0.15 percentage point decrease in the current account balance, measured as a percent of GDP . To put this in perspective, MENA’s (simple average) old-age dependency ratio went from 5.82 percent in 2007 to 6.29 percent in 2017. Note that old-age dependency has a larger 47 APPENDIX B negative association with the current account balance than does young-age dependency. A 1 percentage increase in aging speed is associated with a 4.4 percent increase in the current account balance. A 1 percentage point specification considers the effects of the fundamentals within a country. The pooled specification has only time fixed effects and no country fixed effects. This allows us to examine the effects of fundamentals on current account positions HOW across countries TRANSPARENCY CAN HELPas well THE as over MIDDLE EAST AND The time. between AFRICAspecification takes the average of the current account NORTH position and the fundamentals across years within a country, and then examines the effect of the average fundamentals on the average current account across countries. The residuals of the regressions reflect the portion of the the current accounts that are unexplained by the fundamentals. In the pooled specification (the first column of Table B2), the fundamentals have the expected signs. The account of coefficients youngNote balance. age that andin theage old pastdependencies 10 years, MENA’s negative. are simple coefficient The labor average of –0.148 productivity relative implies that a 1 to the United percent increase States has been in old age dependency declining, from about associated is56 percent in with 2007 a 0.15 percentage to about 46 percent point decrease in 2017. in the Given thecurrent same levelaccount of relative balance, productivity, measured a completely as a percent of GDP open . Tocapital account put this (that is, Chinn-Ito in perspective, MENA’stakes (simplethe average) value of 1) is associated old-age with dependency lowerfrom ratio awent current balance 5.82 of 7in percent percentage 2007 to points compared 6.29 percent into a completely 2017. closed Note that capital account, old-age dependency becausehas capital a larger inflows are expected to be higher. A 1 percent increase in the net commodity index negative association with the current account balance than does young-age dependency. A 1 percentage increase in the current year is associated with a in aging 59 percentage speed pointwith is associated increase a 4.4in the current percent accountin increase balance, and a 1 the current percentbalance. account increase in A the net commodity 1 percentage point index in the previous year is associated with a 38 percentage point increase in the growth acceleration is associated with a 0.3 percent decrease in of the current account balance. When the capital current account balance. Interestingly, none of the exchange rate regime variables are statistically significant, implying no systematic account is completely closed (the Chinn-Ito Index takes the value of 0), a 1 percentage point decrease in relative differential impact of exchange-rate regimes on the current account. productivity compared to the United States is associated with a 0.16 percentage point decline in the current account In thebalance. the past 10 Note that in (columns other specifications years, 2 and 3 ofMENA’s Table B2), simple the average laborhave fundamentals productivity relative largely similar to the United impacts, with States onehas been declining, exception. from about In the between 56 percent specification, in 2007 forecast 63 to future about growth46 percent has a in positive2017. Given assocation the with same the level current of account. relative This means productivity, that when aopen a completely country has, on capital average, account a larger (that growth acceleration, is, Chinn-Ito takes the value it tends tois of 1) sustain a larger associated with current a lower currentaccount balance. balance of 7 percentage points compared to a completely closed capital account, because capital inflows are expected to To ascertain whether the be higher. A 1 percent model’s increase results reflect thein the net commodity influence index in of fundamentals onthe current national year isrates, savings associated we a 59 percentage with estimated point an auxiliary increase model in the current on national savingsaccount rates. The balance, effects of theafundamentals and 1 percent increase on savingin the ratesnet commodity are broadly index in the similar to previous associated year is account those on current with 20 balances. 38 percentage a Old-age point increase and young-age dependenciesin thearecurrent account significantly balance. correlated with lowernone Interestingly, savingof exchange the Higher rates. rate aggregate relative regime variables are statistically labor productivity significant, is associated implying with a higher no rate, systematic saving and differential given the same of impact exchange-rate relative productivity regimes level, anon the current open account. capital account is associated with lower saving rate (thanks to capital inflows). Similarly, an increase in the commodity index is associated with a large increase in saving rates. In the other specifications (columns 2 and 3 of Table B2), the fundamentals have largely similar impacts, with Thus the evidence suggests that the fundamental drivers of the current account probably work through the exception. one national In the savings between specification, forecast future growth has a positive assocation with the current rate. account. This means that when a country has, on average, a larger growth acceleration, it tends to sustain a larger current account balance. To ascertain whether the model’s results reflect the influence of fundamentals on national savings rates, we estimated an auxiliary model on national savings rates. The effects of the fundamentals on saving rates are broadly similar to those on current account balances.20 Old-age and young-age dependencies are significantly correlated with lower saving rates. Higher relative aggregate labor productivity is associated with a higher saving rate, and given the same relative productivity level, an open capital account is associated with lower saving rate (thanks to capital inflows). Similarly, an increase in the commodity index is associated with a large increase in saving rates. Thus the evidence suggests that the fundamental drivers of the current account probably work through the national savings rate. Table B2 MNACE Model Estimates of the Fundamental Drivers of Current Account Balances (1) (2) (3) VARIABLES Pooled Within Between Young Age Dependency % -0.0223* -0.0509 -0.0771* (0.0117) (0.0351) (0.0435) Old Age Dependency % -0.148*** 0.166 -0.420*** (0.0510) (0.114) (0.138) Aging Speed % 4.425*** 1.912*** 7.943 20 The results of the model on national savings rates are available upon request. Table B2 MNACE Model Estimates of the Fundamental Drivers of Current Account Balances 64 (1) (2) (3) APPENDIX B VARIABLES Pooled Within Between 48 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 Table B2. MNACE Model Estimates of the Fundamental Drivers of Current Account Balances (1) (2) (3) VARIABLES Pooled Within Between Young Age Dependency % -0.0223* -0.0509 -0.0771* (0.0117) (0.0351) (0.0435) Old Age Dependency % -0.148*** 0.166 -0.420*** (0.0510) (0.114) (0.138) Aging Speed % 4.425*** 1.912*** 7.943 (0.767) (0.700) (5.107) Predicted Changes in Growth -0.293 -0.361 1.572** (0.289) (0.295) (0.748) Relative Productivity (t-1) 15.75*** 39.70*** (2.437) (11.50) Chinn Ito Index 0.129 5.992*** -1.000 (0.602) (1.603) (2.060) Chinn Ito x Relative Productivity (t-1) -7.005*** -28.74*** 3.423 (2.443) (7.267) (3.944) Net Commodity Price Index, (t) 59.67*** 56.97*** 1,964*** (8.015) (7.407) (601.0) Net Commodity Price Index, (t-1) 38.38*** 36.13*** -2,080*** (6.958) (5.991) (709.0) Managed Float Exchange Rate 0.326 -1.129 1.662 (0.520) (0.796) (1.530) Float Exchange Rate 0.279 -0.828 -2.023 (1.108) (1.263) (3.534) Managed Float Exchange Rate x Relative Productivity 2.366 -4.170 3.895 (t-1) (1.515) (2.784) (2.466) -2.141 -3.571 -0.116 Float Exchange Rate x Relative Productivity (t-1) (1.796) (3.094) (3.673) Relative Productivity (t-1) 4.481 (3.957) -5.200*** -1.884 1.532 Constant (1.640) (3.498) (4.560) Observations 4,254 4,254 168 R-squared 0.236 0.469 0.505 Time Fixed Effects Yes Yes No Country Fixed Effects No Yes Yes Number of Countries 162 162 163 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: Data are for 163 countries. To ensure change in the commodity index is exogenous, we drop large countries—the United States, China, India, Japan, and Russia—which are important commodity consumers and whose economic activity could sway world commodity prices. The period of consideration is from 1990–2017. Eighteen MENA countries are included in our analysis. West Bank and Gaza is not included because it lacks sufficient GDP, predicted changes in growth, and Chinn-Ito data. The Chin-Ito index for Iraq is assumed to gradually converge to MENA’s average in 2018. 49 APPENDIX B HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA A p p e n d ix Appendix C: F isc a l S C:Fiscal u sta in a b ility Sustainability C1. Debt Arithmetic: Primary Balance Required to Stabilize Debt • One type of debt: First, we consider the baseline scenario with one type of debt. We start with a government’s budget constraint: (1 + r+ )D+%& − PB+ = D+ (C1) which means that new debt (D+ ) equals old debt plus interest, (1 + r+ )D+%& , minus new primary balance, PB+ . All are denominated in domestic currency. Dividing the equation by the country’s nominal output is equivalent to: D+ D+%& Y+%& PB+ = (1 + r+ ) − (C2) Y+ Y+%& Y+ Y+ &34" or, d+ = d+%& − pb+ (C3) &35" where d+ is debt-to-GDP ratio, r+ is nominal interest rate, g + is nominal GDP growth. For debt-to-GDP ratio to stabilize, namely d+ = d+%& , (C3) implies 1 + r+ pb+ = o − 1p d+%& (C4) 1 + g+ Similarly, for future debt-to-GDP ratio to stabilize, namely, d+3& = d+ , (C4) implies that the future primary balance depends on the expected future interest rate and expected future output growth. E(1 + r+3& ) pb+3& = o − 1p d+ (C5) 1 + Eg +3& • Multiple types of debt: Multiple types of debt can be collapsed to total debt and weighted average interest rate. We start with an accounting identity & )D& ( ( (1 + r+ +%& + (1 + r+ )D+%& − PB+ = D+ (C6) where D& ( +%& and D+%& are two different types of debts (such as domestic debt and foreign debt). (C6) is equivalent to: & & 6$ ( 6% &34 d+ = &35 t(1 + r+ ) 6$ "#$ % + (1 + r+ ) 6$ "#$ % u d+%& − pb+ , or d+ = &35" d+%& − pb+ , where r+ is the " "#$ 36"#$ "#$ 36"#$ " $ % & 6 ( 6 weighted average interest rate, that is, 1 + r+ = t(1 + r+ ) 6$ "#$ 36% + (1 + r+ ) 6$ "#$ 36% u. "#$ "#$ "#$ "#$ Similarly, if future debt-to-GDP ratio is to stabilize, the accounting identity (C6) implies a required future primary balance: & (1 + Er+3& )ω& ( ( + + (1 + Er+3& )ω+ pb+3& = v − 1x d+ (C7) 1 + Eg +3& 6& where ω& ( * + and ω+ are weights of the two types of debts, that is, ω+ = " . $ 6" 36%" 66 APPENDIX B 50 MIDDLE EAST AND NORTH AFRICA REGION ECONOMIC UPDATE APRIL 2020 (C7) implies that the required future primary balance will rise with expected interest rates and decline with a higher growth rate. This means if the interest payments are expected to rise, the country must run a bigger fiscal surplus to finance the interest payment. Take an example where one type of debt, for example, D&+ , is foreign debt. The expected nominal interest rate equals the interest rate in foreign currency times the expected depreciation of the exchange rate, that is, & ∗ (1 + Er+3& ) = (1 + r+3& ) × E(∆E+3& ). (C7) becomes ∗ (1 + r+3& ) × E(∆E+3& )ω& ( ( + + (1 + Er+3& )ω+ pb+3& =| − 1} d+ (C8) 1 + Eg +3& C2- Calculating Structural Primary Fiscal Balance Calculating the structural primary fiscal balance is based on the following two equations: 8 9 : 9 log  ∗ Ä = θ8 log  ∗Ä and log  ∗ Ä = θ: log  ∗ Ä, 8 9 : 9 where Y and Y ∗ are real output and the Hodrick-Prescott -filtered trend component of output; R and G are real primary revenue and expenditure (excluding interest revenue and interest payments), R∗ and G∗ are real structural revenue and expenditure (IMF, 2011). Following the literature and IMF (2011), we set θ8 = 1 and θ: = 0 to calculate R∗ and G∗ . The argument for θ8 = 1 is that to the extent that tax rates remain unchanged, revenue is a constant share of output. On the contrary, it is difficult to think of a component of government expenditure that is automatically connected to the business cycle, so θ: = 0. Structural fiscal balance PB∗ equals R∗ − G∗ . C3. Relationship between Debt and Primary Balance Following Mendoza and Ostry (2008), we examine how the primary fiscal balance reacts to debt in the previous period. The specification is as follows: Ç pb*+ = β# + fe* + β& g Ç ;+ + β( y;+ + ρd*,+%& + fe+ + fe* + ϵ*+ (C9) Ç pb*+ is primary fiscal balance (as a share of output); g Ç ;+ and y;+ are temporary fluctuations in government !," : !," 9 expenditure and GDP. Specifically, g Ç ;+ = log á:∗ à and y;+ = log á9∗ à are the HP-filtered cylical components Ç !," !," of real government expenditure and real output. d*,+%& is debt as a share of output in the previous period. fe+ and fe0 are time and country fixed effects. A positive ρ shows signs of fiscal sustainability, because primary balance improves when debt increases. To compare MENA and the rest of the world, we apply the following regression: pb*+ = β# + fe* + β& g Ç Ç ;+ + β( y;+ + ρd*,+%& + ρ( MENA × d*,+%& + fe+ + fe* + ϵ*+ (C10) 67 51 APPENDIX C HOW TRANSPARENCY CAN HELP THE MIDDLE EAST AND NORTH AFRICA Table C1. The Relationship between Primary Balance and Debt VARIABLES Primary Balance/GDP Output gap 54.93*** 56.23*** (8.346) (8.324) Expenditure gap -54.41*** -54.51*** (2.654) (2.646) Debt/GDP (t-1) -0.0471*** -0.0471*** (0.00735) (0.00733) MENA x Debt/GDP (t-1) 0.0301 (0.0230) GCC x Debt/GDP (t-1) -0.135*** (0.0409) Developing MENA x Debt/GDP (t-1) 0.0971*** (0.0267) Constant -4.768 -5.213 (3.662) (3.651) Observations 3854 3854 r2 0.258 0.263 Country FE Y Y Year FE Y Y Note: Standard errors in parentheses, * p<0.10 ** p<0.05 *** p<0.01 APPENDIX C 52 WORLD BANK MIDDLE EAST AND NORTH AFRICA REGION MENA ECONOMIC UPDATE APRIL 2020