‫املستقبل‬ (reads “Al Mustaqbal”: The Future) YEMEN COUNTRY ECONOMIC M E M O R A N D U M 2022 Yemen Country Economic Memorandum – 2022 ‫المستقبل‬ (reads “Al Mustaqbal”: The Future) Glimmers of Hope in Dark Times This work was originally published by the World Bank in English as “Yemen Economic Memorandum 2022” in April 2023. In case of discrepancies, the original language will govern. This report was also co-financed by the Yemen Resilience, Recovery, and Reconstruction Trust Fund (also known as Yemen Fund), which was established by the World Bank in 2022, with the United Kingdom’s Foreign, Commonwealth and Development Office (FCDO) as a seed contributing partner, to support Yemen’s transition over a 10-year horizon. The entire World Bank team would like to express their sincere gratitude for this important and timely contribution. Table of Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii 1. Introduction: Change and Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Pre-War Yemen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Patronage Systems Dominated Pre-War Yemen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 The War Reshaped Yemen’s Political Economy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Yemen Remains Heavily Dependent on Oil Production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Remittances Remain Critical to Macroeconomic Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 12. Costing the Conflict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Collapse of Real Economic Activity and Productivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Bifurcated Monetary Policy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 The Weaponization of Trade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Agricultural Shifting Production Patterns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Transportation and Mobility Constraints. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Underemployment and Vulnerability Are on The Rise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 The Public Sector Payroll Remains a Complex Challenge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 The Private Sector Has Proven Remarkably Resilient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Aid Dependency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 The Hydrocarbon Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Renewables Potential. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 3. Yemen’s Transformed Political Economy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Fragmented Institutions, Heterogeneous Outcomes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 iii iv ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Old Problems in New Forms: Rent-Seeking, Patronage, and Corruption. . . . . . . . . . . . . . . . . . . . 61 The Economic War and Its Power Dynamics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 The Emergence of New Political Economy Actors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 The Fight for Control of Trade Corridors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Sectoral Governance during the Conflict: The Case of Liquefied Petroleum Gas . . . . . . . . . 67 Decentralization and Entrepreneurship: Ingredients for Optimism? A Comparative View. . . . . . 70 The Segmentary State and Tribal Capitalism in Abu Dhabi . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Traditional Institutions and Peace in Sierra Leone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Tribes, Accountability, and Growth in Somaliland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Uneven Progress and Solid Growth in Colombia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4. SHAPING YEMEN’S FUTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Scenarios for Yemen Through 2030 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Political Economy and Macroeconomic Scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Divergent Path to Recovery Sub-Scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Comparative Context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Key Policy Takeaways. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Scenario 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Political Economy – Status Quo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Macroeconomic – Base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Scenario 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Political Economy – No War, No Peace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Macroeconomic – Limbo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Comparative Context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Scenario 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Political Economy – A Recentralized State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Macroeconomic – Limited Recovery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Comparative Contexts – Abu Dhabi, Rwanda, Somaliland. . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Scenario 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Political Economy – An Imperfect Peace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Macroeconomic – “Enhanced Recovery”. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Comparative Lessons – Colombia, Sierra Leone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 A. Restoring Macroeconomic Stability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Price Stabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Hydrocarbon Production and Exports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 B. Boosting Productivity Growth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Transportation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Social Protection and Household Welfare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 mproving Transparency, as well as Data Quality and Availability. . . . . . . . . . . . . . . . . . . . . . . . 89 C. I Table of Contents v List of Figures Figure 1 Reported Fatalities, 2015–2022. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Figure 2 Conflict Damage by Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Figure 3 Oil Production (thousands of Barrels Per Day). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Figure 4 Estimated Imports, 2014-2022* (% of GDP). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure 5 Remittances and Fuel Exports, 2014–2023 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure 6 GDP and GDP Per Capita, 2011–2022. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 7 GDP Per Capita, Yemen and Comparators, 1990-2021. . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 8 Real GDP Index (2010 = 100). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 9 Estimated Economic Activity Based on Nighttime Illumination, 2012-2021. . . . . . . . . . . 16 Figure 10 Monthly Gas Flaring and Non-Flaring Illumination, 2012-2022. . . . . . . . . . . . . . . . . . . . . 17 Figure 11 Contribution to Potential GDP Growth (Solow Decomposition). . . . . . . . . . . . . . . . . . . . 17 Figure 12 Human Capital Index Scores, 2020. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Figure 13 Palmer Drought Severity Index by District. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 14 Average Rial-to-Dollar Exchange Rate, January 2019 – October 2022. . . . . . . . . . . . . . . 22 Figure 15 Prices for the Basic Food Basket by District, 2016–2021 (overlapping). . . . . . . . . . . . . . . 22 Figure 16 Decomposition of Inflation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 17 District-Level Food-Basket Prices in IRG and Houthi-Controlled Areas. . . . . . . . . . . . . . 23 Figure 18 Quarterly Real Gasoline Prices in Urban Areas, 2016–2021. . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 19 Changes in Fuel Prices Following the Announced Air and Sea Blockade . . . . . . . . . . . . 24 Figure 20 The Impact of the Announced Air and Sea Blockade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Figure 21 District-Level Food-Basket Prices (blue) and Reported Fatalities (orange). . . . . . . . . . . . 27 Figure 22 Changes in the Area Under Qat Cultivation in Nine Selected Governorates. . . . . . . . . . 28 Figure 23 Qat Cultivation in Otmah District, Dhamar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 24 Intra-Urban Road Damage by City. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Figure 25 Most Important Source of Household Income . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 26 Public Employees by Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure 27 Most Recent Payment for Public Employees. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure 28 Number of Months That Public-Sector Employees Received Payment Over the Last 12 Months. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure 29 Share of Respondents Working During the Seven Days Before the Survey. . . . . . . . . . . 37 Figure 30 Employment Status of Respondent or Main Income Earner . . . . . . . . . . . . . . . . . . . . . . . 37 Figure 31 Sector of Employment of Respondent or Main Income Earner. . . . . . . . . . . . . . . . . . . . . 38 Figure 32 Occupation of Respondent or Main Income Earner. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Figure 33 Employment Sector of Respondent or Main Income Earner. . . . . . . . . . . . . . . . . . . . . . . 38 Figure 34 Occupation of Self-Employed and Own-Account Workers. . . . . . . . . . . . . . . . . . . . . . . . 39 Figure 35 Employment Type of Working Respondent or Main Income Earner. . . . . . . . . . . . . . . . . 39 Figure 36 Frequency of Payment Delays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Figure 37 Duration of Payment Delays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Figure 38 Job Satisfaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 39 Reasons for Job Dissatisfaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 40 Employment Income Sufficient to Cover Expenses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 vi ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 41 Food-Security Status. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 42 Employment Status and Food Security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Figure 43 Employment Sector and Food Security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Figure 44 Worked in the Last Seven Days and Food Security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Figure 45 Job Type and Food Security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Figure 46 Employment by Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Figure 47 Fiscal Revenue and the Public-Sector Wage Bill, 2006–2020. . . . . . . . . . . . . . . . . . . . . . . 43 Figure 48 The US Dollar Value of Teachers’ Salaries in IRG-Controlled Areas, 2014 and 2021 (US$). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Figure 49 Number of Monthly Payments Received by Public-Sector Workers Over the Last 12 Months by Area of Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Figure 50 Public Employment by Area of Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Figure 51 Last Payment Received by Public-Sector Workers Area of Control. . . . . . . . . . . . . . . . . . 45 Figure 52 Net FDI Inflows, 2007–2020 (% of GDP). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 53 The Conflict-Induced Economic Contraction and the Expanding Role of the Private Sector, 2012–19. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 54 Population-Adjusted Nighttime Illumination in IRG and Houthi Areas Each January, 2015 to 2018. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Figure 55 Nighttime-Illumination-Based GDP Estimates for Hajja, Saada, Al Raymah and Mahwit versus the National Baseline, 2012-2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Figure 56 Nighttime Illumination and GDP Estimates, Marib Versus National Baseline . . . . . . . . . 61 Figure 57 Nighttime Illumination and GDP Estimates, Al-Marah Versus National Baseline . . . . . . 61 Figure 58 Nighttime Illumination and GDP Estimates, Taiz and Comparators. . . . . . . . . . . . . . . . . 62 Figure 59 Nighttime Illumination and GDP Estimates, Aden Versus National Baseline. . . . . . . . . . 63 Figure 60 Nighttime Illumination and GDP Estimates, Various Governorates Versus the National Baseline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Figure 61 Correlation between Nighttime Illumination and Oil Output. . . . . . . . . . . . . . . . . . . . . . 66 Figure 62 Projected Growth and Other Macroeconomic Indicators by Scenario. . . . . . . . . . . . . . . 77 Figure 63 Projected Sectoral GDP Growth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Figure 64 Grant Aid for Selected Scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 List of Tables Table 1 Economic and Humanitarian Indicators 2014–2022 (averages) . . . . . . . . . . . . . . . . . . . . . 2 Table 2 OLS Regression Analysis of Nighttime Illumination, GDP, and Oil Production . . . . . . . . 15 Table 3 US Dollar Price Changes for Basic Food Items in Yemen and Comparator Countries, 2020–21. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Table 4 Paved Roads, Population, and Land Area by Governorate, 2015 . . . . . . . . . . . . . . . . . . . 29 Table 5  Challenges and Opportunities for the Private Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Table 6 Potential Renewable Energy by Source. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Table 7 Variations in Production Explained by Nighttime Illumination. . . . . . . . . . . . . . . . . . . . . . 65 Table of Contents vii Table 8 Monthly Average Price for One LPG Cylinder in Sanaa and Aden, May 2021–April 2022. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Table 9 Yemen Scenarios Through 2030. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Table 10 Growth Assumptions Across Macroeconomic Path to Recovery Scenarios. . . . . . . . . . . 78 List of Boxes Box 1 An Overview of The World Bank’s 2015 CEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Box 2 The Role of Women in Yemen’s Economy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Box 3 Alternative Economic Opportunities in Wartime Yemen. . . . . . . . . . . . . . . . . . . . . . . . . . 51 Box 4 Main Model Assumptions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 List of Maps Map 1 Geographic Distribution of Economic Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Map 2 Nighttime Illumination, February 2015 (Left) and May 2015 (Right). . . . . . . . . . . . . . . . . . 16 Map 3 Precipitation Data at the District-Level. Millimeters (average of 5km Input Pixels, Summed for the Year) in a Regular 10km Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Map 4 Transportation Routes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Map 5 Hawban to Taiz (left) and Aden to Taiz (right) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Map 6 Major Transport Routes Before the Conflict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Map 7 Current Status of Major Transport Routes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Map 8 Fuel Supply and Distribution Routes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Map 9 Zones of Control, January 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Map 10 Population-Adjusted Nighttime Illumination in Western Yemen Each January, 2015 to 2018. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Map 11 LPG Supply Routes in Sanaa, Taiz, and Aden . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Preface Working in situations of fragility, conflict, and vio- are unsurprising. A far-reaching reform agenda will lence is a complex challenge, and Yemen is no be critical to bolster the resilience of Yemen’s mac- exception. A growing body of literature on Yemen’s roeconomic framework and strengthen the country’s wartime economy highlights the binding limitations institutional governance. Large-scale investments imposed by a lack of uniformly collected and verifi- in fixed assets and human capital development will able data, and recent analyses often present a grim be vital to restore economic opportunity for a gen- view of the country’s prospects. By contrast, this eration of young people who have grown up amid report uses innovative, technologically advanced a devastating war. Other conclusions are perhaps data-gathering methodologies to overcome data less obvious. Advanced data analytics, for example, constraints. The analysis focuses on opportunities indicate that Yemen’s de facto decentralization can for recovery and renewed growth. This approach help ground its future growth, corroborating a per- has allowed Bank staff to overcome (albeit with spective repeatedly voiced in informant interviews. important caveats) many of the limitations of ear- The World Bank stands ready to continue working lier work. The availability of data has shaped the in cooperation with the government and its other content of the CEM. development partners to support the Yemeni people This report is designed to inform the efforts of during a delicate transition towards stability, peace, policymakers and other stakeholders as they build and security; restore the economy to a sound foot- the foundation for long-term growth and shared ing; and advance the ambition of the Yemeni people prosperity in Yemen. Some of the report’s findings to achieve renewed prosperity and a brighter future. ix Acknowledgments The Yemen Country Economic Memorandum 2022 Goldblatt (Chief Scientist, Geospatial Data), (YCEM22) was prepared by a large team of world- Safa Ali Qassim Almoayad (Social Scientist), Saki class experts, led by Gianluca Mele (Lead Economist Kumagai (Governance Specialist), Selamawit T. and Program Leader; Task Team Leader) and Sharad Ghebremicael (Geospatial Research Associate). A Tandon (Senior Economist; co-Task Team Leader). very special thanks goes to Katia Stefanova (Program Peter Salisbury’s (Political Economy Lead) role as Assistant) for her unwavering administrative and co-editor has been pivotal: his highly esteemed logistical support during the entire preparation of knowledge of Yemen, clear-sighted contributions the report, and to Sean Lothrop, for his highly val- and hard work permeate the whole report. The ued contribution in professionally editing the final team of experts includes (in alphabetical order): Ali draft. The team worked under the strategic leader- Ibrahim Almelhem (Economist and Data Scientist), ship of the Country Management Unit (CMU), led Alia Aghajanian (Economist), Anthony Biswell by Marina Wes (Country Director) and Tania Mayer (Senior Economist), Benjamin Herzberg (Senior (Country Manager, Yemen), and with the tireless and Private Sector Specialist), Eli Kravitz (Data Scientist, much-valued contributions of Mark Ahern (Lead Geospatial Data), Fiona Davis (Economist, Public Economist and Program Leader) and Omar Al-Aqel Policy and Institutional Development), Garrett (Operations Officer). The team benefited from Tate (Senior Data Scientist, Geospatial Data), constant technical guidance from Eric Le Borgne Geoff Handley (Senior Public Sector Management (Practice Manager, MTI) and Hans Hoogeveen Specialist), James Robinson (Professor and Director (Practice Manager, POV). The team is especially of The Pearson Institute, University of Chicago), grateful for the valuable feedback from many col- Hans Lofgren (Senior Economist, Economic leagues and experts working in parallel develop- Modeling), Henry Thompson (Senior Economist), ment – multilateral and bilateral – institutions. An Martin Cicowiez (Professor, Universidad Nacional de initial workshop on the conceptualization of the La Plata, Argentina), Mohamad Alabsi (Economist), YCEM took place in May 2022, and was attended Mohammad Al Akkaoui (Economist), Naif Abu- by high-level officials from the Yemen “Economic Lohom (Senior Water Resources Specialist), Ran Working Group”, which includes officials from the xi xii ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES United Nations, the United States Department of cartography unit, and especially Marlee Jennean State, the United States Treasury, the United States Beers (Cartographer), for their excellent coopera- Agency for International Development (USAID), the tion, as well as the Resource Management team International Monetary Fund, the External Action of the World Bank, in the person of Florencia Toro Service (EEAS) of the European Commission, the (Resource Management Officer), for seamlessly Foreign, Commonwealth & Development Office taking care of all budget-related aspects. A very of the United Kingdom. Constructive cooperation special and heartfelt thanks goes to the United was also received from the Yemeni authorities, to Kingdom FCDO, in the persons of Jocelyn Waller whom the team is deeply grateful, and special (Head, Economy and Recovery Team, FCDO Yemen thanks go to Dr. Afrah Al-Zouba (Managing Director office) and Victoria James (Economic Adviser, of the Executive Bureau for the Acceleration of FCDO Yemen office), for steadily and generously Aid Absorption and Support for Policy Reforms), supporting this report financially, through various Dr. Mansour Rageh (Deputy Governor of the phases of its development, as well as to Tatiana Central Bank of Yemen, CBY, and Chief Economist Weber (Senior Operations Officer), for her tireless at the Sana’a Center for Strategic Studies), and and constructive work on partnerships and financial Dr. Shuaib Ali Abdullah Al-Zaghir (Director of support. The team is also grateful to the col- Technical Office at Yemen’s Ministry of Planning leagues of the International Monetary Fund, and and International Cooperation). The YCEM22 ben- particularly Sascha Buetzer (Economist), under the efited from broad-ranging and extensive consulta- leadership of Brett Rayner (IMF Mission Chief tions with hundreds of local experts, private sector for Yemen) for truly setting the bar of an representatives, informal sector representatives, outstanding WB-IMF cooperation and for the academics, and other key informants, including constant exchange of views and data. last but experts from SanaaCenter, ACAPS, CrisisGroup, not least, the team would like to thank the peer DeepRoot, and Pragma, among others. The reviewers who took the time to share their team would also like to thank the World Bank helpful suggestions and insights on this report: Zeljko Bogetic (lead Economist), Kevin Carey (Practice Manager, Global Macroeconomics and Trade Unit), and Marek Hanusch (Senior Economist). Executive Summary Change and Continuity national level, and the limited available data are often contested for political reasons. While these Yemen’s economy has been transformed by data constraints are not insurmountable, coping eight years of violent conflict. War has shattered with them demands a more innovative and holistic the country’s already fragile economic equilibrium, approach to growth analysis. This Country Economic touching upon virtually every aspect of life. The Memorandum (CEM) uses novel data-collection compounded shocks of the COVID-19 pandemic methods and analytical techniques, triangulating its and rising global prices have only deepened the findings with traditional approaches and direct data economic and humanitarian disaster precipitated collection to close the economic knowledge gap. by the war. Since the start of the conflict, economic Information sources include extensive key-informant analyses have tended to focus on the deteriora- interviews, household phone surveys, and remotely tion of macroeconomic indicators, the sharp rise sensed geospatial data based on satellite imagery, in poverty and food insecurity, and the destruction including nighttime illumination data. Econometric of infrastructure and the capital stock, but relatively simulations of structural transformation produced little attention has been paid to the current structure by a general-equilibrium model calibrated for the of the economy or what prospects can be envisaged Yemeni economy complete the diverse array of ana- for the country. Also, it is important to situate this lytical approaches used in this report. analysis within the political economy dynamics of This CEM also combines an in-depth political the country (including key political and economic economy analysis with economic development players, institutions and centers of powers, both for- investigation. Given Yemen’s history of recidivism mal and informal) which majorly affect the economic to political instability and clashes, this CEM matches development challenges of the country. a rigorous and extensive political science analy- Data constraints and the unique character- sis—regarded as one of the pivotal elements to istics of Yemen’s recent experience limit the support Yemen’s path towards permanent peace, effectiveness of traditional growth-analysis and ultimately economic development—with more methodologies. Consistent data are rare at the traditional macro and microeconomic analysis. xiii xiv ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Yemen’s Fragile Context military intervention. The Houthis would ultimately consolidate control over Sanaa and the main popu- Before the outbreak of war in late 2014, Yemen lation centers in northwest Yemen. However, former was already a poor country with weak gover- President Saleh was killed by the Houthis following nance that had recently sustained a series of a schism in December 2017. economic and political shocks. A centralized patrimonial state in Sanaa relied on the dwindling Yemen at War returns from the country’s hydrocarbon sector to generate revenues in a broadly underperforming The conflict has accelerated Yemen’s political economy. Oil and gas exports also financed a rising and economic fragmentation. Since early 2015, import bill that accounted for about 40 percent of Yemen has been divided into multiple zones of the consumer price index, while the state heavily political and military control. The conflict extends subsidized electricity and used government jobs beyond the often simplistic, binary view of a Houthi- as a form of welfare. A culture of rent-seeking and led government in Sanaa, referred to throughout patronage severely hindered meaningful economic this report as the de facto authority (DFA), and reforms, and as oil output fell from the early 2000s Yemen’s internationally recognized government onward, Yemen’s elites shifted from a posture of (IRG) based in Aden. Disagreements among groups cooperative rent-seeking to competition over a nominally united under either the Houthis or IRG shrinking pool of resources. In 2011, an increas- banner are often as sharp as the conflict between ingly high-stakes, zero-sum struggle among elites the two overarching factions. contributed to a political crisis that resulted in the The war has not just divided the country ouster of longtime President Ali Abdullah Saleh. territorially; it has splintered the public sector The transition process was troubled from the and solidified subnational identities. Over the start, and the incoming administration failed to course of the conflict, the institutions and processes implement vital macroeconomic and governance of formal civil administration have fragmented and reforms. By 2014, when the World Bank’s last CEM morphed into new hybrid governance arrange- was being prepared,1 Yemen was at an inflection ments, often fused with informal practices and tradi- point. Worsening fiscal and external balances, fal- tional norms. The line between state- and nonstate tering public service delivery, mounting insecurity in institutions (both civil and military) has become cities, and open conflict in Yemen’s hinterlands had increasingly blurred, particularly since the formation brought the country to the brink of the humanitarian of the Political Leadership Council (PLC) in April disaster that would shortly materialize. 2022. The PLC is an eight-member council compris- In mid-2014, a coup initiated under the cover ing the leaders of military forces that have fought of popular grievances preceded a full-scale war. against the Houthis but which, in many cases, did Amid regular power outages and fuel shortages, a not fully recognize the authority of the administra- sudden cut to subsidies in July sparked widespread tion of President Hadi, who stepped down in April protests, which the Houthi movement, popularly 2022 to make way for the PLC. known as the Houthis, used as a motivation for The costs of the war have not been evenly its takeover of Sanaa in September. Working with distributed across these different zones of con- forces loyal to former President, Ali Abdullah Saleh, trol, reflecting asymmetries in institutional and the Houthis pursued then-President Abed Rabbo Mansour Hadi to the southern port city of Aden, 1 The CEM was published in 2015 due to delays caused by the launching air strikes that precipitated a Saudi-led outbreak of war. Executive Summary xv quasi-institutional capacity, as well as disparities high, while revenue efficiency has worsened. After in access to financial resources. New state-like plunging in 2015–17, public spending as a percent- capacity has emerged in some parts of Yemen and age of GDP has rebounded, though it remains vola- across the economic system, fostering surprisingly tile and far below pre-war levels. Utility subsidies robust governance and development in previously continue to weigh heavily on IRG finances, generat- deprived and insecure areas. In certain areas, local ing arrears and widening budget gaps. Yemen has and traditional institutions have filled the gaps become more import-dependent than ever, increas- left by weak national governance. In other areas, ing its vulnerability to external shocks, while govern- however, already poor living standards have plum- ment debt has more than doubled since 2015. meted amid social atomization and administra- tive neglect. Some private firms have managed to An Economic Conflict flourish despite the conflict, while many previously dominant players have withered. Both major factions are attempting to assert The civil war has had complex ramifications control over the national economy. Yemen has for women, some clearly and deeply negative, been broadly divided into two competing economic but others positive or at least potentially so. zones, each with its own currency and governance Women were active participants in the popular arrangements—however limited, arbitrary, or ad uprising that removed Saleh from power, but men hoc—which generated a regulatory vacuum in the have dominated the wartime leadership and peace financial sector and precipitated a massive depre- negotiations, and the recently formed PLC has no ciation in the value of the Yemeni rial (YRI) in areas female members. Nevertheless, women were active under IRG control. In the rest of the country, under participants in the popular uprising that removed Houthi control, economic information is extremely Saleh from power and made up one-quarter of limited; these regions also experienced high levels all participants in the 2014 National Dialogue of inflation, although far less than the IRG ones, Conference (NDC). The civil war has also created likely due to the absence of monetary financing of new opportunities for women by enabling them to the fiscal deficit. assume social and economic responsibilities for- International and domestic constraints on merly dominated by men. the movement of people and goods have nega- Pre-war patronage networks have been tively affected growth. Repeated standoffs over uprooted and reordered. The factors driving this fuel shipments into Houthi-controlled areas have process include the Houthi takeover of Sanaa and disrupted supplies, created space for price goug- its seizure of assets from rival groups, the freezing of ing, and driven up the cost of fuel. Over the past oil exports and their later resumption under an IRG- seven years, establishing control over revenue-gen- backed revenue-sharing scheme, and the formation erating ports, informal trade routes, and oil infra- of multiple local power centers with no meaningful structure has become a key economic, military and ties to the Saleh administration. However, these political objective. Road closures have increased the changes have not fundamentally transformed a complexity of supply chains, compounding other political culture that continues to be marked by domestic and external inflationary pressures and clientelism and corruption. contributing to a massive cost-of-living crisis with The conflict has not altered the core struc- deeply negative implications for poor and vulner- ture of the public finances. Despite disruptions in able households. production and exports, oil remains vital to Yemen’s A complex and hostile operating environ- economy. Fiscal dependence on oil exports remains ment has discouraged trade and investment, xvi ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES sharply limiting opportunities for young work- Nevertheless, the economy has clearly contracted ers. The deteriorating business climate has greatly while the population has grown. Oil production and reduced economic activity outside of commodity exports have declined, undermining the govern- trading and construction. Most forms of employ- ment’s ability to support a significant share of the ment offer low wages, poor working conditions, population through public employment. Looking and little or no job security. A lack of economic at fixed capital, the conflict is estimated to have opportunity continues to push young people to damaged or destroyed more than one-third of the join armed groups, which frequently offer higher country’s homes, schools, hospitals, and water and and more stable prospects than the private sector. sanitation facilities. Productivity indicators were International aid also plays an increasingly already weak before the conflict, and productivity important role in Yemen’s political economy. plunged as the violence intensified. Since 2015, Yemen has been designated as one Despite production disruptions, oil remains of the world’s worst humanitarian crises, placing crucial to the economy. Oil production collapsed it alongside fragile and conflict-affected countries in 2015, and exports did not resume until August such as Afghanistan, the Democratic Republic 2016. Production and exports rates remain below of Congo, South Sudan, and Syria. The country pre-conflict levels.4 In 2014, Yemen produced an has received billions of dollars2 in aid since 2015. estimated 167,000 barrels per day of oil. However, International organizations have faced allegations government estimates suggest that oil production that aid is being diverted by local actors to reward averaged just 50,000–60,000 barrels per day between their preferred constituencies and deprive less 2020 and 2022.5 favored groups. Just as pertinently, on account of Remittances are increasingly vital to sustain money’s fundamental fungibility, key actors in posi- consumption, and annually they are estimated tions of power have allegedly been able to redirect to one-quarter—or more—of Yemen’s GDP. In available funds from service delivery to political 2021, official sources estimated remittance inflows patronage. The recent decline in aid inflows reflects 2 Before the civil war, ODA had already risen from an average of a rapidly shifting geopolitical context, growing com- US$628 million (1.84 percent of GDP) in 2011–12 to US$1.23 bil- petition among global humanitarian and develop- lion (2.93 percent of GDP) in 2013–14 (Figure 1). As the war inten- ment priorities, and increasing hesitancy among sified, ODA inflows increased significantly, averaging US$2.55 donors to engage with an opaque and unreliable billion (8.19 percent of GDP) over 2015–17. In 2018, Yemen’s public administration. ODA inflow recorded a peak of US$5.45 billion or 25.24 percent of GDP; since then, it has been on the decline. Between 2019 and 2021, the trend shows a clear decrease in the requirement- Costing the Conflict: Shattered Growth, to-incoming funding ratio, with 2021 having a coverage of only Productivity, and Wellbeing roughly 30% of funds required by the UN system. It should be stressed that data on total aid commitments is inconsistent, as The conflict has catalyzed negative trends and commitments are made by each country or entity through a reversed positive ones, shrinking economic out- variety of mediums that are hard to track, including speeches, put and making future growth harder to achieve conferences, social media, and closed-door meetings. even if peace is secured. World Bank estimates3 3 Reflecting official exchanges with country authorities and based on figures in use also by the International Monetary Fund. suggest that real GDP contracted by approximately 4 https://www.reuters.com/article/yemen-energy-masila- 50 percent between 2011 and 2021. Remotely idUKL8N1AS4E8; World Bank, “Yemen Economic Monitor: sensed data broadly corroborate these estimates, Clearing Skies Over Yemen?,” Spring 2022. while suggesting a greater level of improvements 5 CBY (2022) Quarterly Bulletin: Economic and Monetary De- in GDP than the available statistics indicate. velopments, Issue no. 6, March. Executive Summary xvii at about US$4 billion, though the actual figure could Yemen is acutely vulnerable to exogenous be more than twice as high.6 Even at just US$4 bil- shocks, especially climate change and health- lion, remittances would have clearly supplanted related crises, which pose additional threats to hydrocarbon exports as the main source of foreign productivity and growth. Yemen is the 22nd most exchange. Remittances finance imports of essen- vulnerable country and the 14th least ready country tial commodities, enabling households to mitigate to adapt in the Notre Dame Global Adaptation the negative impact of the socioeconomic shocks Initiative (ND-GAIN) Index8 . Recurrent flooding has that the country has suffered over the course of the extensively damaged roads and bridges, render- conflict. However, remittance income is not evenly ing the movement of people and goods virtually distributed across the population. impossible in several governorates, with especially Despite declining agricultural output due in severe floods occurring between April and August part to climate-change-related extreme weather of 2020. The pandemic compounded the shock of events, agriculture has grown as a share of the conflict, and almost two years after the global overall output and Yemeni workers are increas- rollout of COVID-19 vaccines, only about 1 percent ingly dependent on agriculture for employment, of the population has received a single dose. despite the precariousness of the sector. Men The agricultural sector has shifted from and women are both most likely to say they work growing food crops to Qat, a mild stimulant, in agriculture. Labor market conditions and food posing a series of threats to food security and security conditions are substantially different in IRG nutrition. Machine-learning models and remotely and Houthi areas of control, likely reflecting the sensed data enable an estimate of changes in qat more regular payment of government salaries in IRG cultivation areas in selected Yemeni regions from areas. Yet in aggregate, most workers are employed 2016 to 2021, revealing that the area devoted to in the private sector, primarily in agriculture. qat increased by over 40 percent during the period. The prevalence of agricultural work repre- Qat cultivation has expanded faster than overall sents both a step back from efforts to shift the cultivation, suggesting that Qat is expanding at economy towards value-added services, and the expense of food crop lands. Higher food prices deepening insecurity in the labor market due may therefore reflect a shift in cultivation from food to climate change. A geospatial data analysis crops to qat, a mild stimulant widely consumed conducted for this report confirms that Yemen’s across Yemen. territory has become increasingly prone to climate- Qat use and production on such a massive related shocks. Alternating periods of drought and scale likely decreases Yemen’s economic pro- flooding likely weakened agricultural productivity, ductivity. While it is reasonable to assume that exacerbating the overall decline in productivity increased Qat production has a positive effect on across the economy. Farmers in rural areas experi- overall economic output—since the value added of enced a considerable loss of income, buying inputs at considerably higher prices and selling produce 6 Remittances are originated mostly from GCC countries and at significantly lower prices because of transport Egypt, and in North America. Source: Cash Consortium of Yemen, disruptions. Consequently, many were forced to Remittances Tracker (https://reliefweb.int/report/yemen/cash- consortium-yemen-ccy-remittances-tracker​-june-2022). abandon their farms or limit production for sub- 7 Yemen’s agriculture in distress, Conflict and Environment Ob- sistence purposes.7 Households without working -agriculture- servatory, October 2020, https://ceobs.org/yemens​ members or with working members employed as in-distress/. seasonal agricultural labor are the most likely to be 8 ND-Gain https://gain.nd.edu/our-work/country-index/ra​ food insecure. nkings/. xviii ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES the qat is likely higher than the agriculture planta- over time conflict-induced losses in physical, finan- tions that would have theoretically otherwise been cial, and human capital will inflict permanent scarring cultivated—Qat poses risks for large productivity on Yemen’s economic potential. losses, hinging on significant health risks and lack Some of the most critical economic chal- of basic nutrition. lenges faced by Yemeni households stem from Productivity—a key driver of wealth and job currency depreciation, barriers to market creation—was historically sluggish and plum- access, and a shift from food to cash crops. meted as the conflict progressed. The destruction The bifurcation of monetary policy and foreign- of fixed capital exacerbated negative growth dynam- exchange flows, coupled with the monetization of ics, and has slowly been eroding human capital. fiscal deficits, have driven the steep depreciation Labor productivity continued an already worrying of the rial since the beginning of the war, and the pre-war trend of gradual thinning. According to the rising real cost of living has played a major role in Human Capital Index, a child born in the Republic of pushing Yemen’s population into a humanitarian Yemen in 2020 will be only 37 percent as productive crisis. A factor decomposition analysis of infla- as an adult than he/she would have been if they had tion shows that the depreciation of the riyal has enjoyed complete education and full health. This is played a central role in the rising cost of living for lower than the average for Middle East & North Africa the Yemenis. International and internal barriers to region and also slightly lower than the average for movement have further increased the cost of living. Low Income Countries. Violence, including airstrikes, shelling, and Emergency food-security assessments high- the widespread use of landmines has damaged light the household-level impact of declining and destroyed roads and bridges, creating economic activity since 2015. Most of the popu- internal frontiers between warring factions. At lation has experienced a steep decline in living the same time, key international land borders have standards since the start of the conflict, and the been closed. De facto barriers around the city of Taiz deepening humanitarian crisis has severely eroded have increased the cost of transporting basic goods productivity and earnings. Working conditions are by up to 100 percent, while the cost of transport via poor, employment income is not sufficient to cover the Aden-Ar Rahidah road appears to have roughly basic needs, and underemployment is increasingly doubled since the beginning of the war. pervasive. A comparison of households before and The public sector was once a major employer after the Houthi takeover of Sanaa in 2014 shows a of Yemenis and is still an important source of substantial increase in the share of household enter- income. In a recent phone survey, 17 percent of prises but a sharp decline in the average earnings of respondents had a public-sector job before the war, those enterprises—indicating that entrepreneurship while this had gone down to 8 percent of current was being used as a coping strategy.9 Such evidence employment. However, public-sector wages are paid is consistent with an increase in underemployment irregularly if at all, especially in Houthi areas. rather than an increase in livelihood opportunities. In the absence of crucial governmental Labor-market data suggest that both in Houthi and functions, Yemen’s private sector has assumed IRG areas, workers remain engaged in increasingly tenuous and low-quality employment rather than 9 The 2014 Household Budget Survey continued uninterrupted relying exclusively on external support, and the pri- despite the Houthi takeover of Sanaa. See: Tandon, S. 2019. vate sector continues to provide livelihood oppor- “When Rebels Attack: Quantifying the Impacts of Capturing Ter- tunities despite massive disruptions. While Yemeni ritory from the Government of Yemen.” World Bank Economic firms and workers have proven remarkably resilient, Review 33 (2), 328–352. Executive Summary xix an ever-greater role in managing the country’s In other words, Yemen’s immediate future interrelated security, economic, and demo- remains uncertain, and the path to robust and graphic challenges. The private sector contributes inclusive growth is extremely narrow, but with to the provision of social services and has proven ground for some qualified optimism. Despite the surprisingly resilient, but Yemen’s firms face enor- complexity of the conflict, the ongoing struggle mous challenges. Recent estimates suggest that the for control of the economy, and the sheer scale of private sector incurred US$27 billion in direct and Yemen’s humanitarian crisis and economic dete- indirect losses during the first three years of war.10 rioration, there are reasons for qualified optimism. Moreover, 25 percent of all businesses closed after One of Yemen’s most promising latent assets is its 2014, and over 80 percent of firms experienced a class of entrepreneurs, who have survived decades drastic drop in sales, signaling further waves of busi- of political and economic turbulence. Yemen is ness closures. Based on the 2020 Yemen Business surrounded by wealthy markets that share cultural Climate Survey Report, profits fell among 45 percent ties and import most of their consumption goods. of businesses between 2017 and 2019. Given peace, stability, and adequate investment, The conflict has also dampened international the combination of a resilient private sector and trade and deterred foreign investment. An inhos- a range of viable export markets could enable a pitable environment for investment and trade dis- broad-based recovery led by agro-processing and courages foreign investors and commercial traders. light manufacturing. Yemen’s foreign direct investment (FDI) dynamics It is critical to stress that Yemen had a history are very weak, even by the standards of conflict- of conflict well before the current one, which affected countries and low-income economies. stresses the importance of linking the politi- Net FDI inflows have turned negative since 2011, cal science analysis to economic development. indicating that foreign entities are disinvesting from Large economic literature shows that such a history Yemen. of repeated conflicts is also a predictor of a higher likelihood of falling back into new conflicts.11 Once An Uncertain Outlook again, this is why the political economy analysis of Yemen is a subject of key focus within this CEM, Yemen’s political economy is highly volatile, and so as to look into the elements that could help the while a peace agreement would be a critical country, once peace is reached, to make it lasting step towards recovery, it would not guarantee and sustainable, and support broad-based, inclu- a return to macroeconomic stability. Beyond an sive growth. end to the war, an increase in hydrocarbon produc- While a political settlement will not be suf- tion, the reopening of the country’s main transport ficient to ensure renewed growth, peace is a routes, and better integration in regional and prerequisite for Yemen’s recovery. Despite the global trade markets would all be vital to renewed fundamentally unpredictable nature of the conflict, growth. Moreover, Yemen’s social and political frag- the international experience with crisis, conflict, mentation is likely to persist into the post-conflict period, with divided territorial control and multiple 10 Central Statistical Organization (CSO), 2020. “Promoting overlapping formal and informal administrative Partnership with the Private Sector,” Socio-Economic Update, frameworks. In this context, the international com- Issue number 53, October. munity’s cooperation, assistance, and continued 11 Among others, see “Conflict Relapse and the Sustainability presence in Yemen will be critical to secure a last- of Post-Conflict Peace”, World Development Report 2011 Back- ing peace. ground Paper, World Bank, by B. F. Walter. xx ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Yemen Scenarios Through 2030 Political Economy Scenarios Macroeconomic Scenarios Business as Usual (Scenario 1) Status Quo (Scenario 1) Base scenario The war continues along a similar trajectory to the past eight years, with Economic growth is slow and living conditions do not improve, with GDP violence ebbing and flowing. growth of 2 percent from 2023 onwards. Poverty remains high. (Scenario 2) “No War, No Peace” (Scenario 2) Limbo Conflict intensity dwindles, perhaps because of a ceasefire, but there is no Growth is marginally better than in the base scenario, around 3.1 percent political settlement. on average. Exports and imports increase marginally relative to 2022. Poverty remains high. Path to Recovery (Scenario 3) Restored State (Scenario 3) Limited Recovery De facto and/or de jure authority are consolidated around a single central- GDP grows at twice the rate of the base scenario, and public investment ized set of state institutions. This authority is periodically contested by increases. Poverty decreases marginally. subnational actors. (Scenario 4) Imperfect Peace (Scenario 4) Enhanced Recovery A political settlement either enshrines de jure decentralization or leads to Growth is three times higher than the base scenario. Government and de facto decentralization that is accepted by national elites. private sector investment increases and poverty falls to levels approaching those seen pre-war. and political turbulence yields important lessons rule and informal institutions can be assets that for Yemen. The first is the transition from conflict to support post-conflict recovery and develop- peace in Sierra Leone since the end of the civil war ment, not barriers to progress. In Yemen as else- in 2001. Sierra Leone is not a development miracle, where, classic state-centric approaches to economic but it is a peace and stability miracle. The reason is and political development are unlikely to represent the strength and legitimacy of its traditional institu- a comprehensive solution. A best-case outcome tions. The traditional institutions of Yemen are differ- for Yemen is that its post-conflict order matches ent, but they are equally as legitimate and strong. the “Red Queen Effect” described as the path to The second example is Somaliland, Yemen’s near inclusive institutions by Acemoglu and Robinson neighbor, which harnessed the legitimacy of local (2019). Acemoglu and Robinson argue that the institutions to build a culturally relevant but modern best way to build a prosperous and inclusive soci- accountable state and provide the basis for eco- ety is by balancing the power of state and society. nomic growth. The third is Abu Dhabi, where tribal If the state become too powerful and dominates ties underpinned rather than impeding state build- society then this leads to a “despotic Leviathan” ing as many Eurocentric paradigms would assume which has some potential for growth but where the would happen in a tribal society. And the fourth uncontrolled power of the state inevitably leads to is Colombia, which has experienced intermittent overreach that dulls innovation. If society becomes political turmoil and internal conflict since a ten- too powerful relative to the state, then this leads to year civil war that began in 1948 but nonetheless an “Absent Leviathan” with very limited economic has experienced solid economic growth, providing potential. It is in the middle, where society and state a less optimistic but also likely very realistic addi- balance each other, that a dynamic emerges which tional comparison for Yemen. can lead to a “Shackled Leviathan”. In this dynamic The examples of Sierra Leone, Somaliland, process the state and society complete with each and Abu Dhabi all highlight how local customary other, and in the process, both become stronger Executive Summary xxi in ways which lead to public good provision and not representing development forecasts. Based on inclusive economic growth, and puts the country key informant interviews and a review of the litera- into what Acemoglu and Robinson describe as the ture on Yemen, the report outlines the following “narrow corridor” of sustainable economic develop- scenarios (see table above). ment. The negotiation in Somaliland is a real-world While the scenario analysis does not rep- example of how a society can bootstrap their way resent concrete predictions or forecasts, the into such narrow corridor of coexistence and coop- exercise does provide some useful policy take- eration between the state and society. Despite the aways. First, Yemenis witness at least some mod- immense damage incurred by the conflict, there is est improvements in their economic situation in some evidence that the collapse of the patrimonial all non-Status Quo / Base scenarios, including a state after 2014 created opportunities for good pessimistic No War, No Peace scenario, in which governance and entrepreneurship by replacing an the conflict is frozen but there is no progress oppressive and sclerotic central administration with towards a political settlement. In other words, new local power centers. Formalizing and integrat- if the fighting is significantly reduced or com- ing these decentralized systems could establish the pletely stops, life gets better for Yemenis even if foundation for an effective post-conflict governance the political outcome is unsatisfactory. Second: model. peace, decentralization13 and increased public investment efficiency can yield dramatic gains in Imagining Yemen’s Future employment and poverty reduction. And third, the private sector makes an important contribution to Any attempt to propose policy directions for growth across scenarios. Yemen is complicated by the fundamental uncer- In the most optimistic reading of this tainty surrounding the country’s political future. analysis, if the conflict subsides, the donor Yemen possesses a broad roadmap for the future in community is supportive, and the economy the form of the outcomes of the National Dialogue responds positively to increased investment in Conference, a 10-month series of broadly inclusive infrastructure and human capital, Yemen could talks held in Sanaa in 2013 and 2014.12 However, achieve considerable gains in renewed growth given the country’s extreme territorial and institu- and poverty reduction by 2030. Given Yemen’s tional fragmentation, deep political polarization, and low levels of infrastructure and human develop- ongoing competition among rival power centers, it is ment, investments in these areas should yield unclear what importance decision-makers and politi- high returns. However, effective public spending cal elites attach to these outcomes. Policymaking is will require stronger governance, including more further complicated by lack of reliable official data efficient public expenditure management and tax and by suboptimal coordination on data collection collection, with adequate support for poor house- among local and international partners. holds and positive incentives for the private sector. Data constraints notwithstanding, some International aid will be vital to finance investments broad political and economic scenarios can be in Yemen’s development. sketched out based on Yemen’s macroeconomic context and the political economy analysis that 12 National Dialogue Conference Outcomes Document, Feb is found throughout this report. In Chapter 4, 2014, Collated by the UN Special Envoy for Yemen’s Office. alternative political economy and macroeconomic 13 if, as a large platform of key informants and the macroeco- scenarios are presented for 2022–2030, reflecting dif- nomic scenarios assume, this leads to improved security and ferent institutional and economic trajectories, while stability. CHAPTER 1 Introduction: Change and Continuity Yemen’s economy has been transformed by Yemen’s reconstruction needs are estimated eight years of violent conflict . The war has at US$20–25 billion.14 More than one-third of all plunged Yemen into a dire humanitarian crisis. homes, schools, hospitals, and water and sanitation Already deeply fragile before the conflict, Yemen facilities have been damaged or destroyed during has effectively been tipped into state failure. the war. Data from a 2020 World Bank Dynamic Economic output has fallen by half, while the Needs Assessment (DNA)15 found that the housing country’s infrastructure has been devastated (see sector had been the most severely affected by the Table 1). The poverty rate has shot to over 66 per- conflict, with about 40 percent of the housing stock cent. The supply of basic services has plummeted, suffering physical damage (see Figure 2).16 About government effectiveness is weak or nonexistent in much of the country, and levels of food insecu- 14 World Bank’s Yemen Dynamic Needs Assessment, 2020 rity, severe malnutrition, and physical violence are Update; https://documents1.worldbank.org/curated ​ / en/ among the highest in the world. 490981607970828629/pdf/Yemen-Dynamic​-Needs-Assessment- The conflict has exacted a devastating Phase-3–2020-Update.pdf. human toll. The war’s intensity has fluctuated over 15 The World Bank conducted a multi-phased Dynamic Needs Assessment (DNA) in Yemen, first in September 2015, to assess the course of eight years of fighting, but on aver- the impacts of the ongoing crisis in the country and to provide age more than a thousand people have been killed a preliminary estimate of recovery and reconstruction needs. every month, with Armed Conflict Events Database A first phase, completed in early 2016, covered the education, (ACLED) estimates of reported fatalities rarely dip- energy, health, housing, transport, and WASH sectors in Sanaa, ping below that number (see Figure 1). According to Aden, Taiz, and Zinjibar. The second phase, finalized in June 2017, ACLED, by mid-2022 150,000 people had died as a updated the first-phase results and expanded the assessment to 10 cities and seven sectors. The third phase, conducted in 2018, direct result of violence, including more than 15,000 further expanded the number of cities and sectors assessed. Its civilians. In 2019, the UN Development Program findings were updated in early 2020 to reflect conflict dynamics. (UNDP) estimated that more than 133,000 people 16 Damage estimates are based on: (i) the replacement cost per had died from the war’s indirect impacts, including asset; (ii) number of assets damaged in each category assessed a lack of access to food and health services. within the sector; and (iii) physical status of the damaged assets. 1 2 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Table 1 Economic and Humanitarian Indicators 2014–2022 (averages) 2002–2009 2010–2014 2015–2022 GDP growth (annual %) 3.9 0.4 –6 GDP per capita (constant 2015 US$) 2,519 2,393 911 FDI, net inflows (% of GDP) 2.0 -0.4 –1.2* Human Development Index (HDI) 0.49 0.51 0.46** Gender Development Index (GDI) 0.66 0.64 0.50** Percentage of undernourished people 26.8 32.4 44.2*** Source: World Development Indicators, WB; UNDP; FAO. *data available up until 2019 ; ** data available up until 2021; ***data available up until 2020. Figure 1 Reported Fatalities, 2015–2022 Figure 2 Conflict Damage by Sector Beginning 100 4% 3% 1% 13% of the truce 90 3K 80 60% 70 62% 61% 59% 71% Reported fatalities 75% 60 Percent 2K 50 77% 40 30 28% 1K 39% 20 32% 36% 19% 29% 10 9% 3% 2% 1% 6% 1% 10% 0 0K Education Health Housing ICT Power Trasnport WASH Apr-15 Oct-15 Apr-16 Oct-16 Apr-17 Oct-17 Apr-18 Oct-18 Apr-19 Oct-19 Apr-20 Oct-20 Apr-21 Oct-21 Apr-22 Destroyerd Partially damaged No damage Unknown Source: ACLED Source: World Bank DNA. 38 percent of the country’s water supply, sanitation, some form of humanitarian assistance. Of these, and hygiene (WASH) infrastructure also suffered 12.9 million are currently in acute need of  assis- physical damage, with about 10 percent of all physi- tance. Moreover, the conflict has displaced roughly cal WASH assets having been destroyed outright. 4.3  million people. The UN has estimated that In addition, 38 percent of the country’s health facili- US$4.3 billion will be required to meet the coun- ties were either damaged or destroyed, along with try’s humanitarian needs in 2022, but Yemen has 34 percent of education facilities and 29 percent of received just US$2.3 billion in donor support. fixed transportation assets. The conflict has touched upon virtually every Yemen’s vast humanitarian needs are largely aspect of social and economic life. Basic public unmet. About 24.3 million people, representing health and education services have been severely more than two-thirds of the population, require disrupted, infrastructure has been damaged or Introduction: Change and Continuity 3 destroyed, the private sector is a fraction of its for- Pre-War Yemen mer size, and the country’s civil service has gone unpaid for years at a time. The COVID-19 pandemic Prior to the war, Yemen was already a very and subsequent international price shocks have poor country facing severe governance chal- compounded Yemen’s economic challenges, and lenges. In 2014, Yemen’s GDP per capita was about the authorities have been ill-equipped to cope with US$1,600.19 The country was dominated by a cen- these challenges. Yemen is also increasingly vul- tralized, patrimonial state in Sanaa, characterized nerable to the effects of climate change, including by weak accountability and poor governance that increasingly frequent and severe droughts, flood- distorted incentives and stifled growth.20 Yemen ing, erratic rainfall, agricultural diseases and pests, was a case of what Migdal (1988) called a “strong and rising sea levels. The war’s impact has been society and a weak state.” Outside of the country’s so severe and sustained that even if the war were urban centers, informal local governance arrange- to suddenly end, a post-conflict Yemen could face ments and customary law were as common as state- stagnant growth and productivity. provided services, if not more so. Data constraints compound Yemen’s policy Yemen’s pre-war economy hinged on the challenges. For example, while many humanitar- country’s dwindling hydrocarbon resources. The ian and development actors have warned of high increasing importance of the hydrocarbon sector levels of hunger and even a potential famine in was the dominant trend in Yemen’s pre-war econ- Yemen since 2017, others have argued that these omy. Oil and gas directly contributed 12.5 percent of claims rest on unreliable data.17 Similarly, reports value added between 1994 and 2014. By the 2010s, of a banking system in disarray and a market trun- hydrocarbons accounted for roughly 90 percent of cated by blockades are belied by evidence of ongo- exports, about 75 percent of government revenue, ing commercial transactions and imports of food, and roughly 20 percent of GDP. Meanwhile, growth fuel, and other vital commodities.18 Addressing and productivity stagnated outside of the hydrocar- these and other apparent contradictions under- bon sector and the domestic activities were financed scores the importance of this Country Economic by oil and gas exports. A growth accounting analysis Memorandum (CEM) and other innovative analyti- conducted by the World Bank in 201421 found that cal work in Yemen. productivity growth was driven by capital accumula- The World Bank’s last Yemen CEM was tion and by a labor force that was increasing in size completed in 2015, when the country was pro- while its marginal productivity stagnated. The econ- foundly different than it is today. An updated omy struggled to create jobs or allocate resources CEM, providing an in-depth analysis of Yemen’s to high-value-additive activities, and most workers socioeconomic context, political economy, and were engaged in low-productivity agriculture.22 conflict dynamics is crucial to support effective Along with remittances, oil and gas exports country engagement. However, several analytical underwrote a rising import bill that included instruments that the World Bank typically lever- ages to produce a CEM are not available due to 17 See: Tandon and Vishwanath (2021) and Vuylsteke (2021). data constraints and the limited policy dialogue 18 For example, see International Crisis Group (2021). 19 Source: World Bank Databank, https://data.worldbank​ with domestic authorities. In response, this CEM .org/indicator/NY.GDP.PCAP.CD?locations=YE. draws on a combination of traditional and innova- 20 World Bank, Yemen Country Economic Memorandum, 2015. tive data sources, as well as qualitative assessments 21 Unlocking the Potential for Economic Growth, Yemen CEM, and perspectives, to fill knowledge gaps and assess October 2015, World Bank. economic conditions. 22 Ibid. 4 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES about 90 percent of Yemen’s staple foods, government elected not to renew its PSA with a 40  percent of basic consumer goods, and a Canadian firm for the Masila oil field in Hadramawt, growing share of government revenue. Prior then the country’s most productive field. Combined to the conflict, oil revenues were estimated at with mounting insecurity, including attacks on oil US$7.9 billion, while remittances were estimated at and gas infrastructure, these actions had a deeply US$3.4 billion, though the latter were likely much negative impact on investor sentiment, and several higher than reported.23 Elevated global oil prices post-2007 bidding rounds failed to draw significant and rising remittance inflows were not enough credible interest. to offset mounting import demand, and Yemen Fiscal rigidities limited government oppor- ran a negative current-account balance from tunities to maximize the hydrocarbon sector’s 2007 onward.24 The government’s share of export potential. A series of fiscal crises and the prioritiza- revenue from the Yemen LNG Project25 totaled tion of short-term cashflow management contrib- approximately US$665 million in 2014, representing uted to chronic underinvestment in the oilfields 22 percent of total government hydrocarbon export the government had taken over, despite their vital revenue. Income from LNG exports was forecast to importance as a source of revenue. Before 2009, oil- more than triple by 2018, when it would account for field development policies and PSAs did not include more than 45 percent of government hydrocarbon provisions for natural gas extracted as a byproduct export revenue.26 of oil production. Numerous analyses proposed that Yemen’s dependence on the hydrocarbon such gas could be used to produce cheap, clean sector proved unsustainable. Oil production energy for domestic consumption, but gas-based peaked at 457,000 barrels a day in 2002, but a electricity generation projects struggled to progress decade later output had fallen by more than 65 due in part to confusion over who owned Yemen’s percent. Nevertheless, government oil revenue gas.28 Because oil firms and the government could increased due to rising export prices, an increas- not agree on how to share the profits, gas was ing share of export royalties, and the addition of either reinjected into oilfields or flared. By this point, some LNG revenue. Yemen’s gas reserves were esti- Yemen’s oil-export pipelines and terminals, which mated at about 430 billion cubic feet in 2014. The Yemen LNG project encompassed about two-thirds 23 IMF, WB data. of proven reserves, while the remaining reserves 24 The CA balance went from an average of +3.2 percent of stayed untapped or were flared—in part because GDP (surplus) in 2000–2006 to a deficit of –6.6 percent of GDP, the country’s legislative framework for oil produc- in the remainder years of the decade (2007–2010), and onwards. tion did not account for gas deposits. By 2014, 25 The Yemen LNG project was launched in 2005 as a partner- ship between the government and several major international gas exports were generating around 13 percent of oil companies. It started producing and exporting LNG from total hydrocarbon revenues, but initiatives to use a non-associated gas field in Marib in 2009. Cost-recovery the remaining gas reserves to produce electricity provisions in the development agreement combined with stalled.27 low gas export prices limited government revenue during the To boost its oil and gas income, the gov- initial phase. ernment started increasing public ownership 26 Increases were calculated on the basis of an end to the cost- recovery process on the Yemen LNG project, the expiration and of some of the largest oilfields in 2005. As a renegotiation of long-term LNG export contracts, and expected Production Sharing Agreement (PSA) with a U.S. shifts in market pricing. firm came to an end, the government elected 27 Estimates sourced from IMF (2014) and World Bank (2015). to take full control of one of the country’s big- 28 Yemen – A natural gas incentive framework, World Bank, gest oilfields, the Safer field in Marib. In 2011, the 1 June 2006. Introduction: Change and Continuity 5 Map 1 Geographic Distribution of Economic Activity REP6#LIC OF YEMEN REFINERIES LAND USE INDUSTRIES INDUSTRIES SAUDI ARABIA CRUDE TERMINAL PASTURES PETROLEUM REFINING FISHERY PORTS LNG LIQUEFACTION TERMINAL ARABLE LAND METAL PROCESSING I IRON OUT OF SERVCE GAS PIPELINES IRRIGATED FARMING SHIP CONTAINER TERMINAL CROPS OUT OF SERVICE CRUDE PIPELINES NOMADIC HERDING C CEMENT COFFEE CRUDE PIPELINES WASTELAND QAT LIGHT INDUSTRY GAS/OIL FIELDS OASIS TEXTILES SELECTED CITIES/TOWNS FOOD PROCESSING OMAN INTERNATIONAL BOUNDARIES Saada Al Masila Sana’a R ed Midi Al Hazm Al Ghaydah Se C Hajjah Amran Nishtun a Al Mahwit As-Salif Ma’rib Bajil Al-Hodeidah Al Mahrah Fishing Areas Al Jabin Dhamar Ataq Al Mukalla Al-Hodeidah Fishing Areas Zabid Al Bayda Catch Per Region Ibb ERITREA Al Dali Hadramawt Fishing Areas 20% Taiz Shuqrah 10% 40% Mocha Zinjibar Lahj 30% Aden C I ETHIOPIA Hadibu 'Abd al Kuri Samha Socotra Gulf of Aden Darsa The Brothers DJIBOUTI IBRD 46564 | APRIL 2022 This map was produced by the Cartography Unit of the World Bank Group. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of the SOMALIA World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. Source: World Bank (reflecting status in the 2020s). had been built before unification in 1990, were in also an issue of extreme political sensitivity, and dire need of rehabilitation. attempts to cut subsidies in 2005 and in 2014 led Sizeable energy subsidies impose an unsus- to mass protests.31 tainable burden on the public finances. Even as it exported crude oil and LNG, Yemen increas- ingly relied on imported refined fuel products from 29 Republic of Yemen : Unlocking the Potential for Economic abroad, which the government resold at subsidized Growth, World Bank, October 2015. 30 See Engelke et. al., Petroleum Subsidies in Yemen: Leverag- prices to meet local energy needs.29 Energy subsi- ing Reform for Development, World Bank, June 2013; Yemen dies cost the equivalent of 7.2 percent of GDP in Corruption Assessment, USAID, September 2016. 2013 and 6.4 percent in 2014. This subsidy system 31 See: Yemen subsidy riots leave 16 dead, Associated Press, remains in place. It is regressive and reportedly 22 July 2006; Yemen protests erupt after fuel price doubled, Al subject to pervasive corruption.30 However, it is Jazeera English, 30 July 2014. 6 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Patronage Systems Dominated Pre-War and extractive companies, and between Yemen tribes and the government. It has suc- cessfully prevented and resolved conflicts Yemeni society is highly heterogeneous: the over resources, development services, country encompasses multiple religious con- and land, and has sometimes managed fessions, tribal structures, political and societal to contain complex revenge-killing cases. elites, and business networks. These multiple Nationally, tribal mediators have played overlapping identities have profoundly influenced an important role in promoting political state-building efforts. Paul Dresch, among the most dialogue and building consensus among eminent scholars of modern Yemen, writes that political groups. (….) Most Western observers and urban Yemenis believe that [T]he idea of Yemen as a natural unit has tribes and the tribal system have been an been embedded in literature and local impediment to state building and devel- practice. Unified power has not. Political opment in Yemen. And, indeed, there structures through the nineteenth century have been tribal leaders who traded the were defined by reference to religion or needs of their people for political influ- dynasty, not territory, and a list of rulers ence and who were part of the corrupt would be indefinitely long for their claims patronage network that undermined the overlap in both time and space…. Much state. However, tribes have played an of Yemen’s history through the twentieth important role in holding Yemen together century connects with efforts to form [a] in the face of increasing political conflicts state, which was finally established in 1990. and harsh economic conditions. Moreover, Before that there were two states, North as the tribal system is increasingly under and South, with their capitals at Sanaa and stress, tribal leaders and citizens in… at Aden, each with its view of the country’s tribal areas…are eager to see legitimate past and future. [In] the years around 1900 and functioning state institutions in their there were myriad little centers of power— areas and have the desire and willingness hence myriad different histories…and a to contribute to achieving that.33 few great claimants, two of which were foreign empires.32 Pre-war Yemen was marked by a complex and deeply embedded patronage system (see After the unification of north and south Box 1). Throughout the 1990s and 2000s, corrup- Yemen in 1990, the central government in tion and rent-seeking dominated the public and Sanaa struggled to establish sound governance private sectors. To remain stable, this system was arrangements in areas outside of the coun- forced to accommodate a growing number of cli- try’s urban centers. Nadwa al-Dawsari, a leading ents as the population grew and previously margin- Yemeni scholar of tribal society, argues that alized groups gained power.34 Perceptions of state 32 Dresch, Paul, A History of Modern Yemen, Cambridge Uni- Yemenis have relied on indigenous tribal versity Press, 2000. traditions to regulate conflict and estab- 33 Al-Dawsari, Nadwa, Tribal Governance and Stability in Yemen, lish justice for centuries, if not millennia. Carnegie Endowment for International Peace, 24 April 2012. Tribal law has effectively handled conflicts 34 Republic of Yemen : Unlocking the Potential for Economic between various tribes, between tribes Growth, World Bank, October 2015. Introduction: Change and Continuity 7 BOX 1: AN OVERVIEW OF THE WORLD BANK’S 2015 CEM Prior to the start of the war, Yemen was already the poorest country in the Middle East and North Africa (MENA) region. Home to over 30 million people, Yemen has one of the highest birth rates in MENA at about 4.4 births per woman in 2014. Life expectancy is only 66 years, and infant mortality is approximately 46 per 1,000 live births, 77 percent higher than the average for Arab states. The poverty rate in 2014 was estimated at about 48 percent of the population. Yemen is a net importer of food, fuel, livestock, and machinery. It exports some crude oil, qat, coffee, and seafood. In 2014, the World Bank prepared a CEM for Yemen just as the political crisis that had started in 2011 was metastasizing into a civil war. The CEM highlighted the country’s lackluster economic performance and heavy dependence on the hydrocarbon sector, noting that oil and gas production remained the key source of rents and the engine of the economy, and that the government had proven unable to diversify economic output beyond the extractive industries. The CEM also identified the country’s deeply entrenched patronage system as a major obstacle to reform. While patronage was particularly acute in the hydrocarbon sector, it was also prevalent in agriculture, WASH, telecommunications, and financial services. The CEM argued that the struggle for control of rents had distorted economic policies and undermined the influence of the private sector and civil society. The CEM found that few major enterprises were held by nonelite actors and that elites routinely used the power of the state to undermine the competitiveness of firms they did not control. Arbitrary discrimination in favor of elite interests represented a substantial barrier to private-sector development, worsening an already weak business environment marked by the poor quality of infrastructure and services. In line with the outcomes of the National Dialogue Conference (NDC), the CEM proposed the adoption of a federalism framework to strengthen accountability, oversight, and citizen engagement while reducing corruption and promoting economic competition. corruption in Yemen were among the highest in the The transitional period failed to foster world and helped set the stage for the conflict.35 improvements in economic governance. Political Falling oil output intensified intra-elite competition, elites in the new unity government accelerated as rival factions sought to capture larger shares of a their competition for institutional and financial shrinking pool of resources.36 As oil rents dwindled, resources, as historical rivals and emerging fac- elites fought over what remained rather than imple- tions sought to improve their positions and weaken menting painful but necessary reforms to promote their opponents.39 Tens of thousands of employees growth outside the oil sector. were added to state payrolls, particularly in the In 2011, a political crisis further weakened security and military sectors, while elites reputedly the macroeconomic environment. Mass pro- sabotaged oil, gas, and electricity infrastructure in tests against the government of then-President an effort to undermine one another.40 An inclusive Ali Abdullah Saleh precipitated an open conflict National Dialogue Conference (NDC) organized between rival military units and private militias in by the United Nations provided the basis for a new Sanaa and other major cities.37 The resulting secu- 35 In 2010, Yemen was ranked 146 out of 183 countries for rity vacuum was exploited by rebels in the north, perceived corruption by Transparency International. In 2011, secessionists in the south, and other extremist meanwhile, Gallup found that seven out of 10 Yemenis polled elements, who seized control of provincial capi- on drivers of that year’s uprising said that corruption was wide- tals. Civil war was averted when neighboring Gulf spread in government and was a major cause of the unrest. See Cooperation Council (GCC) states brokered a Gallup, ‘Widespread Corruption in Government’, June 2011. transitional initiative that led to the formation of 36 Ibid. 37 Yemen:  Enduring  Conflicts, Threatened Transition, Interna- a unity government and the ouster of President tional Crisis Group, 3 July 2014. Saleh.38 Interim President Abed Rabbo Mansour 38 Ibid. Hadi was appointed to oversee a two-year transi- 39 Yemen: Corruption, Capital Flight and Global Drivers of tion to democracy while implementing economic Conflict, Chatham House, 1 September 2013. and security-sector reforms. 40 Ibid. 8 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES national constitution, but its contents—and par- at the head of another faction battling the Houthis ticularly its provisions related to federalism—were on the western Red Sea coast.43 disputed and at times contradictory. Yemen is divided among multiple competing By 2014, Yemen was at an inflection point. power centers. The Houthis’ fatal split from Saleh Worsening fiscal and external balances, faltering ser- in December 2017 left them as the dominant force vice delivery, mounting insecurity in cities, and open in northwest Yemen, where the bulk of the popula- conflict outside the capital had pushed the country tion lives and GDP is produced. The international to the brink of a humanitarian disaster. In July 2014, community has largely upheld Hadi’s Internationally a sudden reduction in subsidies amid regular power Recognized Government (IRG) as Yemen’s rightful outages and fuel shortages sparked widespread sovereign authority; on the other hand, the Houthi protests, which the Houthi movement, popularly forces are the De Facto Authority (DFA) in large por- known as the Houthi rebellion, used as a motivation tions of the country. Hadi sustained alliances with for seizing Sanaa in September of the same year.41 political, tribal, and military forces in Taiz, Marib, The Houthis and their allies placed Hadi under house eastern Abyan, northern Shebwa, and northern arrest after he attempted to ratify the new constitu- Hadramawt, but his forces were pushed out of Aden tion, then moved to consolidate their control over by the pro-independence Southern Transitional the entire country. The Houthi movement joined Council (STC) in 2019. Other independent factions forces with Saleh’s loyalists, and when Hadi fled to include Salafist militants and the forces of Tareq the southern port city of Aden, the Houthi-Saleh Saleh on the Red Sea coast and elsewhere. alliance pursued, launching airstrikes that precipi- The conflict has divided the country physi- tated a Saudi-led military intervention. The Houthis cally and administratively, as formal state institu- would ultimately consolidate control over Sanaa and tions have fragmented and morphed into new the main population centers in northwest Yemen. hybrid governance arrangements. Yemen’s formal However, former President Saleh was killed by the and informal public administration has been trans- Houthis following a schism in December 2017. formed over the course of the conflict, and the lines between state and nonstate military and civilian The War Reshaped Yemen’s Political institutions is increasingly blurred. New state-like Economy capacity has emerged in some parts of the country, fostering surprisingly robust governance and signifi- The conflict has exacerbated preexisting social, cant development in previously marginalized and political, economic, and geographic disparities.42 insecure areas. Elsewhere, local institutional capac- The Houthi-Saleh campaign faltered outside of ity has withered, while public service delivery has the country’s northern highlands, as the alliance partially or completely ceased. Meanwhile, some struggled to make progress in areas where reli- areas that were already deprived have seen further gious, tribal, and political affiliations supported declines in living standards. organized resistance. Tribal groups in Marib (to A Political Leadership Council (PLC) formed Sanaa’s east) and al-Bayda (southeast of Sanaa), as to replace Hadi in April 2022 has struggled to well as Salafist and pro-independence fighters in overcome national and subnational divides, the formerly independent south, along with mili- tary and tribal forces in the industrial and cultural 41 Yemen: Stemming the Rise of a Chaos State, Salisbury, 2016. hub of Taiz formed various resistance movements. 42 Regionalism and Rebellion in Yemen, Stephen Day, 2013. In 2018, shortly after the Houthis killed former 43 The Yemen Matrix: Untangling the Relationships that Drive President Saleh, his nephew Tareq Saleh emerged the War, DeLozier, 2020. Introduction: Change and Continuity 9 while governance quality has not significantly Figure 3 Oil Production (thousands improved. The council includes senior leaders of Barrels Per Day) from the STC; Tareq Saleh the governor of Marib; a powerful Salafist-led faction; the pre-2011 ruling 500 party, the General People’s Congress; and, less 450 directly, Islah, the country’s main pre-war Sunni 400 Islamist political bloc. Just over three months 350 after the PLC’s formation, Salafist and local forces 300 pushed Islah-aligned military and security units out 250 of northern Shebwa, while the STC seized control of 200 eastern Abyan and began to threaten a complete 150 takeover of Hadramawt. Conflicts between PLC fac- 100 tions continue, and of US$3.3 billion in economic 50 support pledged by Saudi Arabia and the UAE 0 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 2020 to the PLC only $1bn has materialized, and only 15 percent of this amount has been dispersed to Source: BP annual statistical review 2021a. the government.44 a See: https://www.bp.com/en/global/corporate/energy-economics/ statistical-review-of-world-energy.html. Yemen Remains Heavily Dependent on Oil Production reports indicate that such revenues may exceed IRG income.46 Citing internal Houthi documents, a 2021 Despite disruptions in production and exports, report by a UN expert panel estimated formal Houthi oil remains vital to Yemen’s economy. Oil pro- revenues at US$1.8 billion in 2019.47 Local authorities duction collapsed in 2015, and exports ceased until in Marib and Hadramawt also collect and redistrib- August 2016. Production and exports remain below ute their own revenues, which are included in the pre-conflict levels, and both were threatened by national accounts. Fiscal oversight of Yemen’s ports Houthi attacks on IRG oil export infrastructure in and oilfields is uneven, and actual revenues could be October and November 2022.45 Oil output is only far higher than these estimates suggest. around one-third of its pre-war average: In 2014, Public spending has fallen far below pre-war Yemen produced an estimated 167,000 barrels per levels. IRG revenues and expenditures dropped day of oil. However, government estimates suggest by 35 percent and 26 percent (nominal YR terms), that oil production averaged just 50,000–60,000 respectively, between 2014 and 2021. Meanwhile, barrels per day between 2020 and 2022, while gas the rial depreciated by about 75 percent, compound- exports dropped to zero (see Figure 3). ing the fiscal contraction. No reliable estimates are Fiscal dependence on oil exports remains high, while revenue efficiency has worsened. 44 This report has been completed as of end-January 2023 and Official IRG estimates suggest that the government reviewed before publication in March. earned about US$479 million from oil exports in 2021, 45 https://www.reuters.com/article/yemen-energy-masila- idUKL8N1AS4E8; World Bank, “Yemen Economic Monitor: representing about 27 percent of total government Clearing Skies Over Yemen?,” Spring 2022. revenue, standing at $1.7bn, the same year, per offi- 46 How Huthi-Saudi Negotiations Will Make or Break Yemen, cial figures. These figures do not include revenues Crisis Group, 2022. earned by the Houthis from customs and tax collec- 47 Final Report of the UN Panel of Experts for Yemen, United tion, as no reliable estimates exist, though informal Nations, 29 January 2021. 10 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES available for Houthi spending, and some regions Figure 4 Estimated Imports, under nominal IRG control collect and redistribute 2014–2022* (% of GDP) their own revenues, complicating efforts to form a complete picture of revenues and expenditures 60% nationwide.48 50% Utility subsidies weigh heavily on IRG 40% finances, generating arrears and widening bud- get gaps. Although the government has not sub- 30% sidized fuel sold on the open market since 2017, it 20% does subsidize fuel used to produce electricity. In 2021, the IRG spent an estimated US$400 million 10% on subsidies, representing about one-third of total 0% 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 expenditures and a lion’s share of oil export revenues. Of which: Fuel While figures are not yet available for 202249, high Of which: Food international fuel prices may have tripled subsidy Imports of goods and services spending. Despite enormous public spending on Source: IMF, World Bank, IRG authorities. (*2022-23 figures are projected). the sector, the electricity supply remains extremely unreliable in towns and cities under IRG control, prompting allegations of widespread fuel theft and Remittances Remain Critical to corruption in the electricity sector (see Chapter 3). Macroeconomic Stability As consumers are rarely billed for electricity, most public spending on electricity production is never Remittances, along with official development recouped.50 assistance (ODA), play a vital role in financing Yemen is more import-dependent than ever imports and preventing further currency depre- before (see Figure 4). After falling in 2015, the offi- ciation. Remittances are also an object of competi- cial import bill has risen steadily over the course of tion between the Houthis and IRG. Key informant the conflict. According to official estimates, the coun- interviews indicated that the dollars used to pay for try spent US$9.2 billion on imports in 2021, with food imports of essential commodities, including food and fuel accounting for 65 percent. The government and fuel, are sourced from remittances from orga- expects imports to rise to US$12.5 billion in 2022, a nizations, firms, migrant workers, and expatriates.52 35 percent increase, which would make this the first Remittance payments are not distributed evenly year since the conflict began that the cost of imports across the population. A large percentage of remitted exceeded the pre-war peak of US$12.26  billion in funds flows to households in more populous Houthi- 2014. However, total exports in 2014 were almost controlled northwestern governorates. The IRG has three times higher than the government’s projection repeatedly attempted to assert its authority over for 2022.51 Anecdotal evidence from World Bank staff money-transfer services and banks, and to monitor and key informant interviews strongly suggests that official data significantly undercount imports due 48 See: Brokering a Ceasefire in Yemen’s Economic Conflict, International Crisis Group, January 2022. to high levels of informal trade. As imports strongly 49 This report was concluded prior to end-2022. correlate with overall economic activity, this is cir- 50 World Bank staff based on consultant Key informant inter- cumstantial evidence—also suggested by Night views, January-October 2022. Time Light activity—that economic activity may be 51 Source: National authorities and IMF data. stronger than official figures indicate. 52 Key informant interview, September 2022. Introduction: Change and Continuity 11 and direct flows through the IRG’s Central Bank of Figure 5 Remittances and Fuel Yemen in Aden (CBY-Aden), albeit with mixed results. Exports, 2014–2023 Annual remittances may amount to one- quarter of GDP or more. In 2021, remittances were 5,000 25% about US$4 billion according to official estimates, 4,000 20% but the real figure could be as much as twice this amount. A  simple back-of-the-envelope calcula- 3,000 15% US$ mln tion gives a sense of the volume of transactions and the challenge of precisely estimating remittance 2,000 10% flows. Yemen’s import bill for 2021 totalled roughly 1,000 5% US$9.2 billion.53 While official IRG figures show US$1.2 billion in receipts from oil exports for 2021, much of 0 0% 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 this amount was used to cover external operating costs and finance fuel imports for electricity produc- Remittances Fuel exports tion. CBY-Aden provided approximately US$110 Remittances as % GDP Fuel exports as % GDP million in foreign exchange to the open market over Source: IMF and World Bank Staff Calculations. the course of nine auctions during 2021. ODA to Yemen averaged approximately US$2.9  billion per year between 2015 and 2021, around 40 percent of remittances.55 This may indicate a further consolida- which (about US$1.2 billion) was to cover the govern- tion of a shift in remittance channels among Yemeni ment’s operating expenses.54 With no other formal source of foreign exchange available to the market, 53 National authorities and IMF data. the total value of remitted funds could exceed US$7 54 Source: World Bank staff calculations, based on: UN, OECD, billion, especially since import volumes are likely OCHA, USAID, US State Department, and UK FCDO data undercounted (see Figure 5). sources. 55 Key informant interviews, September 2022. As outlined in the Even at the low-end estimate of US$4 bil- ACAPS report, The impact of remittances on Yemen’s economy, lion, remittances have clearly replaced hydrocar- published in October 2021: “In the case of Yemeni migrants and bon exports as Yemen’s main source of foreign expatriates based in Saudi Arabia, official/formal remittance exchange. Over the two decades prior to the con- options are: (1) SWIFT transfers through the formal banking flict, foreign exchange from oil exports was at least system, (2) international money transfers through global money double the officially recorded value of remittances, transfer and remittance companies (e.g. Western Union, Mon- but the war has reversed this relationship. In addition eyGram, and Xpress Money), (3) electronic remittance services, and (4) licensed money exchange companies, such as Bin Yaala to financing imports and stabilizing the exchange Exchange.” The ACAPS report also outlines the different form rate, remittances support essential household con- of electronic or digital remittance services provided by licensed sumption and have helped mitigate the negative operators in Saudi Arabia: “An electronic remittance is carried impact of the socioeconomic shocks that the country out via a Saudi bank that the remitter has an account with and has suffered over the course of the conflict. that offers its own, official/formal electronic remittance services, The formalization of remittance flows may be such as Quick Pay (provided by Al-Ahli Bank), Enjaz (provided by Albilad Bank), Ersal (provided by Saudi Post and Alinma affecting the balance of economic power inside Bank), and Fawri (provided by Al Jazeera Bank). Yemeni banks, Yemen. According to key informants in the financial specifically Tadhamon Bank, Al-Kuraimi Islamic and Microfinance services sector and migrant workers in Saudi Arabia, Bank, and Yemen Kuwait Bank (YKB), complete the transfer and official digital remittance services are replacing infor- pay out electronic remittances to recipients in Yemen. All three mal service providers as the preferred channel for banks are licensed by the Saudi Central Bank to operate with 12 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES migrant workers and expatriates in Saudi Arabia country as a percentage of their income. Remitters observed between March 2020 and June 2021. The often opt to send additional remittances via unof- growing prevalence of official digital remittances ficial channels to avoid detection.57 services reflects the lower vendor fees charged by Saudi and Yemeni banks versus those charged by unofficial or informal service providers; the imple- Saudi banks and offer their respective electronic remittance mentation of Saudi Arabia’s digitalization strategy services via mobile phone application technology (Tadhamon of financial services; partnerships between licensed Pay by Tadhamon Bank, Kuraimi Express by Al-Kuraimi Islamic money-exchange services and Saudi and Yemeni and Microfinance Bank, and Yeah Money by Yemen Kuwait banks; and digital remittances provided directly Bank). These three Yemeni banks are also authorised to provide by licensed money-exchange services.56 Bolstered other formal remittance services in the form of either SWIFT by the formalization of remittance services and the transfers or as an official agent of Western Union, MoneyGram, continued economic recovery in the aftermath of and Express Money. For SWIFT transfers, the remitter must have an account with a Saudi bank that then processes the SWIFT the pandemic, official data suggest that 2022 are transfer with the correspondent Yemeni bank. For Western estimated to be a record year for remittance inflows. Union and MoneyGram, the Yemeni banks operate as official Remittances sent through formal networks agents of these global money transfer and remittance compa- likely benefit financial service providers in IRG- nies;” https://www.acaps.org/sites/acaps/files/slides/files/20​ controlled areas. Although Yemeni migrant work- 211015_acaps_yemen_analysis_hub_impact_of_remitta​ ers in Saudi Arabia are increasingly using official nces_on_yemens_economy.pdf. 56 Source: Saudi Arabia official government website; digital remittance services, estimates suggest that https://www.my.gov.sa/wps/portal/snp/aboutksa/digit​ a still sizeable volume of remittances continues to altransformation. be sent through informal channels. One of the main 57 This analysis was produced by World Bank staff and con- reasons is that Saudi Arabia caps the total value of sultants based on key informant interviews conducted during remittances that an individual can transfer out of the July-September 2022. CHAPTER 2 Costing the Conflict Since the start of the war, most economic ana- to triangulate estimates and projections wherever lysis in Yemen has focused on broad macroeco- possible using remotely sensed data, humanitar- nomic indicators, the humanitarian crisis, and ian data, and extensive key informant interviews, the destruction of infrastructure and capital. especially at the sub-national level. While these Relatively little attention has been paid to the cur- approaches have limitations—for example, house- rent structure of the economy or its evolution since hold surveys often have limited coverage in con- 2014. Data constraints limit the effectiveness of tra- flict-afflicted areas—taken together they provide a ditional growth-analysis methodologies, as official strong foundation for updated economic analysis. data are sparse and often inconsistent. The Central While innovative methodologies can help Statistical Office58 (CSO) has been unable to col- overcome data limitations, their findings are lect data necessary to update key indicators since subject to important caveats. Some of the data before the conflict escalated in 2015.59 Survey data sources used in this CEM conflict with one another: collected by international institutions and nongov- for example, World Bank projections suggest that ernmental organizations is sometimes alleged to be GDP has contracted sharply since the start of the manipulated by local vested interests. war, yet remotely sensed nighttime lights suggest These data constraints are not insurmount- that economic activity has rebounded to pre-war able, but addressing them demands a more inno- levels. Similarly, official oil exports are much lower vative and holistic approach to growth analysis. than an analysis of remotely observed gas flaring at More data are being collected in Yemen, and at a oil production sites would indicate. Underestimated higher frequency, than in other comparable con- GDP and oil output figures coupled with persistently texts. However, statistics on economic aggregates high poverty rates imply that inequality may be far after 2015 are based on traditionally sourced data more severe than official estimates show. While that may not fully capture the economic transfor- mation caused by the war. Given the limitations of 58 The CSO was based in Sanaa until 2022, when a separate standard analytical approaches in data-poor envi- Aden-based CSO was established. ronments like Yemen’s, the World Bank has sought 59 See, e.g., World Bank (2017). 13 14 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 6 GDP and GDP Per Capita, Figure 7 GDP Per Capita, Yemen and 2011–2022 Comparators, 1990-2021 (% of U.S. GDP Per Capita) 50,000 1,800 45,000 1,600 1.0 AUS 40,000 1,400 USA 35,000 0.8 1,200 GBR 30,000 1,000 DEU 2021 Indexed 25,000 0.6 800 CTP 20,000 KOR 600 0.4 ESP 15,000 400 URY PRT 10,000 TUR CHNKAZ 0.2 CHL 5,000 200 BWA ARG MEX 0 0 COLGABBRA 0.0 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 0.0 0.2 0.4 0.6 0.8 1.0 1990 Indexed GDP (US$m, real) GDP per capita (US$) 0.10 Source: World Bank staff estimates based on official figures and IMF data. ALB VNM the analysis presented below is based on the best IDN DZA 0.05 information available, its conclusions cannot be MRT GHA NGA cited with absolute certainty. BEN COM BGE YEMEN HTI KEN Collapse of Real Economic Activity and 0.0 Productivity 0.0 0.05 0.10 Source: IMF, World Bank, IRG authorities. The war in Yemen has imposed an immense human and economic cost. World Bank estimates60 its regional and structural peers. While the sharp suggest that real GDP contracted by approximately drop in hydrocarbon production has had the greatest 50 percent between 2011 and 2021. While analysis impact on GDP, several other factors have contrib- later in this report suggests these figures may uted to the decline in economic activity, including overstate the contraction, the country’s ability to lower agricultural output, weakening institutional produce and export oil and gas has clearly been quality, and systematic interference in the private diminished along with its ability to support a large sector by rival factions.61 Between 1990 and 2021, share of households through public employment declining economic output caused Yemen’s growth (see Figures 6). The conflict has catalyzed negative trends and trajectory to diverge from other economies. In 1990, reversed positive ones, reducing economic out- Yemen’s per capita GDP was equal to about 7 percent put and making future growth harder to achieve. of U.S. per capita GDP, but it has since fallen to just Economic literature identifies institutional quality and 2 percent. Some commodity-dependent economies, productivity growth in non-resource sectors as vital 60 Reflecting official exchanges with country authorities and to the sustained development of commodity export- based on figures in use also by the International Monetary Fund. ers. Even before the war, Yemen performed poorly in 61 FAO (2019); Favari et al. (2021); and ICG (2021). See also both areas, and since 2010 it has fallen further behind Figure 4 in Annex 3. Costing the Conflict 15 including Nigeria, Algeria, Botswana, Mauritania, Figure 8 Real GDP Index (2010 = 100) and Mexico, also experienced decline relative to the United States. But others, such as Kazakhstan, 140 Ghana, and Indonesia, remained in lockstep with more advanced economies (see Figures 7, 8). 105 Remotely sensed data corroborate traditional growth estimates during the early years of the war but suggest a faster recovery in economic 70 output. Nighttime illumination is demonstrably cor- related with overall economic activity and can be used as a proxy for GDP growth in situations where 35 economic output is difficult to measure.62 As the con- flict escalated in March 2015, nighttime illumination 0 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 2020 dropped dramatically, a pattern that is broadly con- sistent with official statistics and pre-war World Bank Yemen Middle East & North Africa Low income estimates.63 Light emissions fell by over two-thirds FCSs World between February and May 2015 before gradually recovering in subsequent years and ultimately return- ing to pre-conflict levels in 2021. A regression analysis of nighttime illumination, GDP, and oil production shows a positive, statistically significant relationship 62 Yemen in Transition, Henderson et al, 2011. across all specifications, suggesting that every 1 63 This is corroborated by both Fetzer (2017) and World Bank (2017). percent change in illumination reflected a 0.35 per- 64 This finding is supported by established studies by Henderson cent change in GDP.64 This finding is supported by (2012) which use panel data for 72 countries, and find this elas- established studies by Henderson (2012) which use ticity to be in a similar range, at 0.265. Henderson used panel panel data for 72 countries, and find this elasticity to data for 72 countries and found a growth elasticity of nighttime be in a similar range, at 0.265 (see Table 2). illumination of 0.265. Table 2 OLS Regression Analysis of Nighttime Illumination, GDP, and Oil Production Dependent Variable: (1) log(RGDP) (2) RGDP (3) Oil Prod. (1) Index (2) Index (3) Mb/d Independent. Variable (1) (2) (3) log(NTLs) 0.349*** (0.002) NTLs 0.0357** (0.004) Flaring NTLs 0.648*** (0.025) Observations 23 23 113 R 2 0.799 0.788 0.856 Note: *p<0.1; **p<0.05; ***p<0.01. 16 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Remotely sensed data indicate that the col- Figure 9 Estimated Economic lapse in output between 2015 and 2016 was Activity Based on Nighttime driven by falling oil production, the departure Illumination, 2012-2021 of extractive-industry firms, and the overall impact of the civil war.The decrease in nighttime illumination was concentrated in Sanaa and in the 120 oil-producing western regions. But nighttime illumi- nation data show signs of a faster-than-understood 100 economic recovery from 2017 to 2021, which may 80 reflect increased oil production and a resurgent pri- vate sector (see Figure 9). 60 Satellite-based data suggest that oil produc- tion is higher than most official estimates. The 40 brightness of gas-flaring sites was relatively stable 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 before the conflict, but flaring levels plunged in 2015 as foreign oil companies withdrew. However, Total NTLs Total GDP production appears to have recovered more quickly Source: WB staff calculations. than official data suggests. Oil production esti- mates based on nighttime illumination consistently exceed those provided by the United States Energy dollars over several years, which cannot reasonably Information Administration (EIA). These findings are be explained by statistical errors or methodological corroborated by large differences in fuel imports differences (see Figure 10). and exports between Yemen and its trading part- Emergency food-security assessments ners. Although modest disparities are common reveal the household-level impact of slow- when using mirror data, in this case the magnitude ing economic activity. Most of the population appears to be in the hundreds of millions of US has experienced a substantial decline in living Map 2 Nighttime Illumination, February 2015 (Left) and May 2015 (Right) Source: WB staff, based on VIIRS satellite data. Costing the Conflict 17 Figure 10 Monthly Gas Flaring and Figure 11 Contribution to Potential Non-Flaring Illumination, GDP Growth (Solow 2012-2022 Decomposition) 700 10 Growth rate (percentage points) 600 0 500 Brightness (Millions) –10 400 300 –20 200 –30 100 –40 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Capital stock (α*gK) Labor ((1–α)*gL) Total Flares No flares Total factor productivity (gA)) Real GDP (gY) Source: WB staff calculations. Source: WB staff calculations. standards since the start of the war. By June 2016, more than half of the population was facing emer- gency or crisis levels of food insecurity.65 The World Figure 12 Human Capital Index Food Program’s November 2016 Emergency Food Scores, 2020 Security and Nutrition Assessment (EFSNA), the Human capital index 2020 country’s first major food-security assessment 1.0 based on in-person interviews, estimated that 65 Productivity relative to benchmark percent of the population had poor or borderline 0.8 food consumption, a 57 percentage-point increase from the World Bank’s 2014 Household Budget 0.6 Survey.66 While the rate of poor or borderline food security fell to around 36 percent in February 2019,67 0.4 a phone survey conducted by the World Bank in September/October 2022 found that this share had 0.2 6 8 10 12 risen to 50 percent. Log real GDP per capita at PPP Productivity—a key driver of wealth and job creation—was already weak before the Source: World Bank Human Capital Index. war, and plummeted as the conflict intensified. Limited innovation and the inefficient allocation of 65 UN (2016). These estimates are based on the UN’s Integrated resources contributed to low levels of total factor Phase Classification (IPC) system, which is designed to estimate the severity of food emergencies. productivity (TFP), which deteriorated further after 66 EFSNA (2017); World Bank (2017a). The 2014 estimates are 2014. The destruction of fixed capital exacerbated derived from a longer food consumption module in the 2014 negative growth dynamics, and the conflict badly Household Budget Survey. eroded labor productivity. According to the World 67 According to the latest publicly available report from the Bank’s Human Capital Index (HCI), a child born in WFP’s mobile Vulnerability Analysis and Mapping survey. 18 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Yemen in 2020 is expected to reach just 37 percent livelihoods depend on favorable climatic condi- of what her lifetime productive potential would have tions.72 Aden is ranked as the world’s 6th most vul- been had she enjoyed a complete education and nerable city to rising sea levels and storm surges.73 full health. Yemen’s HCI score is below the average Compounding its exposure to natural hazards, for MENA and slightly below the average for low- weak governance impairs Yemen’s ability to adapt income countries worldwide (see Figures 11, 12). to climate change. The worsening humanitarian crisis has fur- A geospatial data analysis conducted for ther dampened Yemen’s long-term outlook. The this report confirms that Yemen’s territory has deterioration of physical and human capital dimin- become increasingly prone to climate-related ishes future output and earnings. The war’s adverse shocks. Cross-referencing the Palmer Drought effects on nutrition, health, and education, as well Severity Index (PDSI) with precipitation data at the as the psychological and physical trauma associ- district level reveals extreme fluctuations in rainfall ated with the conflict, will be felt for generations.68 (see Figure 13). From 2014 to 2015, the climate was Since the conflict escalated in March 2015, very dry, and several districts ended this phase at fuel shortages have been added to a growing < –5 on the PDSI. From 2016 to 2017, the climate list of growth and productivity bottlenecks. Legal was dry-to-normal, with more precipitation in competition over control of the economy, and of the IRG-controlled territories. From 2018 to 2020, pre- fuel sector specifically, was also pursued through cipitation levels rose rapidly, particularly in areas regulatory efforts. Decree 75 increased oversight of controlled by the IRG. These alternating periods of fuel imports, while Decree 49 required the payment drought and flooding likely weakened agricultural of duties and taxes before shipments entered Yemeni productivity, exacerbating the overall decline in ports. 69 Disagreements among IRG and Houthi productivity across the economy (see Map 3). authorities have repeatedly prevented fuel vessels Yemen also faces acute local environmental from docking at the Port of Al Hodeida, resulting risks. Since the beginning of the conflict, the UN in reported fuel shortages in northern Yemen.70 has sought to address the status of the FSO Safer, Supplied by imports from the south, a parallel fuel an aging oil tanker that has not been inspected or market has emerged, with pump prices up to three or undergone maintenance since 2015. The Safer may four times higher than the national average. Elevated contain up to 1 million barrels of oil. According to the global oil prices have exacerbated fuel shortages, International Maritime Organization, a spill from the further suppressing economic activity while eroding Safer “would be a major humanitarian and environ- household welfare. mental disaster. A significant spill is likely to heavily Climate change poses additional threats to productivity and growth. According to 68 Assessing the impact of the war, Moyer et al., UNDP, 2019. the Notre Dame Global Adaptation Initiative 69 Cfr. Brokering a Ceasefire in Yemen’s Economic Conflict, (ND-GAIN) Index71 for 2019, Yemen is the 22nd most International Crisis Group (ICG), January 2022. vulnerable country and the 14th least ready coun- 70 As a part of the recent truce, it was agreed that 18 fuel ships try to adapt to climate change. Recurrent flood- would enter Hodeidah, starting April 2022. ing has extensively damaged roads and bridges, 71 ND-Gain https://gain.nd.edu/our-work/country-index/ra​ nkings/. severely restricting the movement of people and 72 UNDP; see https://www.adaptation-undp.org/explore/ar​ goods in several governorates. These supply-chain ab-states/yemen. disruptions sharply increased prices for essential 73 Climate Change and the Future Impacts of Storm-Surge commodities. Over 75 percent of the population Disasters in Developing Countries, Center for Global Develop- is engaged in farming or pastoralism, and their ment, 2009. Costing the Conflict 19 Map 3 Precipitation Data at the District-Level. Millimeters (average of 5km Input Pixels, Summed for the Year) in a Regular 10km Grid (continued on next page) 20 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Map 3 Precipitation Data at the District-Level. Millimeters (average of 5km Input Pixels, Summed for the Year) in a Regular 10km Grid (continued) Source: WB staff analysis and graphic representation based on satellite data. impact the north-western coastline of Yemen, includ- 12,000 COVID-19 cases and 2,150 deaths have been ing the Yemeni Islands in the Red Sea, and Kamaran reported in Yemen since 2020, but these figures are Island in particular. There is also a potential for oil clearly underestimates. Years of conflict have badly to drift and impact neighboring countries, including damaged Yemen’s healthcare facilities, and a large Djibouti, Eritrea and Saudi Arabia.”74 share of medical professionals have left the country The pandemic further damaged Yemen’s or stopped practicing. COVID-19 testing is expen- economy. A year and a half after the global rollout sive by local standards, and limited transportation of COVID-19 vaccines, only about 1 percent of the infrastructure reduces access to testing. population had received a dose—one of the world’s 74 FSO SAFER oil spill risk – addressing the threat, International lowest vaccination coverage rates.75 In the areas Maritime Organization, 23 September 2022. under Houthi control, where roughly 80 percent of 75 Source: WHO Eastern Mediterranean Regional Office; https://​ the population resides, COVID-19 vaccines are not tw​i​t​t​er​.com​/WH​O​E​M​RO​/st​a​t​us​/15​5​9​1​3​1​5​6​3​4​0​9​0​0​2496. allowed. According to official statistics,76 only about 76 Source: WHO. Costing the Conflict 21 fiscal deficit. The collapse of oil production and Figure 13 Palmer Drought Severity exports in 2015 placed Yemen’s fiscal and monetary Index by District position under significant pressure. The country’s 15 overall fiscal balance worsened, as the deficit wid- ened from a pre-war average of –5.5 percent of 10 GDP79 to an average of –6.8 percent.80 The authori- ties increasingly turned to monetary financing, District mean PDSI 5 the CBY-Aden’s claims on the IRG increased by an average of 36 percent per year between 2015 and 0 2021, while currency in circulation rose by 24 per- cent per year.81 Hard-currency reserves were quickly –5 depleted, prompting the central bank to abandon the de facto dollar reference, and macroeconomic 2010 2012 2014 2016 2018 2020 stability deteriorated amid rapid inflation and cur- Overall average DFA average IRG average rency depreciation. As the conflict has progressed, Yemen’s economy increasingly developed into a de Source: authors computations. Values < 0 indicating drier climates with values < -4 indicating extreme drought, and with values > 0 indicating facto dual economy, with IRG- and Houthi-controlled wetter climates and values > +4 indicating extremely wet climates. areas effectively acting as separate national econo- mies. This split is most clearly demonstrated in the bifurcation of the exchange rate in IRG and Houthi Bifurcated Monetary Policy areas. Figure 14 shows that following the split of the currencies, the rial continued to depreciate in IRG- Of all the economic challenges posed by the controlled regions while remaining relatively stable conflict, the bifurcation of monetary policy and in Houthi-controlled regions. foreign-exchange flows and the monetization of Currency depreciation has played a major fiscal deficits are among the most consequential. role in Yemen’s burgeoning humanitarian cri- In 2016, the IRG moved the central bank’s headquar- sis.82, 83 The average cost of the basic foods84 Yemeni ters to Aden. CBY-Aden was, in effect, a newly formed institution that retained access to markets and mon- etary instruments but lacked adequate foreign- 77 World Bank and IMF (2021). exchange liquidity. Meanwhile, Houthi-controlled 78 ACAPS (2020). Sanaa remained Yemen’s banking and trading center, 79 Calculated over the period 2011–2014. and due to the larger population living in Houthi- 80 Calculated over the period 2015–2020. 81 WB staff calculations based on national authorities and IMF data. areas it benefitted from greater remittance inflows 82 IPC (2018); IPC (2020); Favari et al. (2021). and aid-agency payments executed via the formal 83 In the 2018 currency crisis, the entire country was affected, banking system.77 The rial briefly stabilized in early but in the current currency crisis that began in 2020 and was 2019, following a large deposit of hard currency at continuing as of early 2022, only the approximately one-third the CBY-Aden by Saudi Arabia.78 As this deposit was of the population in IRG-controlled areas are affected by the gradually depleted, the Houthis announced that it currency crisis. However, there are other factors that are push- ing the entire country towards a humanitarian crisis, including would no longer accept the new banknotes printed increases in global food prices and declines in humanitarian by the IRG, effectively severing monetary policy assistance to Houthi-controlled regions (e.g., Favari et al. 2021). between the two regions (see Figure 14). 84 As defined by the WFP at www.wfp.org/wfp-food-basket; Faced with mounting financial pressures, including grain staple such as wheat flour or rice, a bean or the IRG resorted to monetary financing of the legume, vegetable oil, sugar, and salt. 22 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 14 Average Rial-to-Dollar Figure 15 Prices for the Basic Exchange Rate, January Food Basket by District, 2019 – October 2022 2016–2021 (overlapping) Food basket price by district 1,400 1,200 4,000 1,000 800 3,000 600 YER 400 200 2,000 Juamuary-19 March-19 May-19 Jly-19 September-19 November-19 Juamuary-20 March-20 May-20 Jly-20 September-20 November-20 Juamuary-21 March-21 May-21 Jly-21 September-21 November-21 Juamuary-22 March-22 May-22 Jly-22 September-22 1,000 2017 2018 2019 2020 2021 Sana’a Aden IRG DFA Sources: Country authorities, IMF and WB data. nected to the bifurcation of the banking sector (see Figure 15). Figure 16 Decomposition of Inflation A factor-decomposition analysis of inflation confirms the central role that the rial’s depreciation 100 has played in raising the cost of living. Between the 80 second half of 2016 and 2022, a 63 percent correla- 60 tion was observed between changes in the rial-to-dol- Percent 40 lar exchange rate and consumer price inflation. This strong pass-through effect is due to Yemen’s reliance 20 on imports, the prevalence of exchange-rate-based 0 pricing, and the progressive loss of confidence in the –20 local currency. Decomposing changes in inflation85 into Sep 2018 Nov 2018 Jun 2019 Mar 2019 May 2019 Jul 2019 Sep 2019 Nov 2019 Jun 2020 Mar 2020 May 2020 Jul 2020 Sep 2020 Nov 2020 Jun 2021 Mar 2021 May 2021 Jul 2021 Sep 2021 Nov 2021 regional, structural-comparator, and country-specific effects demonstrates that whereas in previous years High inflation countries (structural factors) inflation was primarily caused by external factors, since MNA countries (regional factors) Idiosyncratic componet 2019 inflation has increasingly been driven by internal Yemen (recorded) inflation factors, including but not limited to monetary financ- ing of the fiscal deficit (see Figures 16, 17). The bifurcation of monetary policy has inten- households rely upon has increased substantially in sified price divergence and economic differences every district since 2015. In 2016, the average price between IRG- and Houthi-controlled territories. for the basic food basket was roughly YRI 1,400, but Exchange-rate differentials have weakened purchasing over the next few years it doubled or more in most districts, reaching as high as YRI 4,000 in Socotra, 85 Based on methodology developed by Krusper (2012) based for example. The rising cost of living is closely con- on Stock and Watson (2002). Costing the Conflict 23 Figure 17 District-Level Food-Basket Prices in IRG and Houthi-Controlled Areas IRG DFA 4,000 4,000 3,000 3,000 YER YER 2,000 2,000 1,000 1,000 2017 2018 2019 2020 2021 2017 2018 2019 2020 2021 Note: Modeled prices are shown in blue. Observed prices for each district are shown in grey. Table 3 US Dollar Price Changes for Basic Food Items in Yemen and Comparator Countries, 2020–21 Column1 Djibouti Iraq Lebanon Oman Syria Yemen Sana’a Aden Wheat –1% 18% –28% 26% –31% 38% 43% 43% Wheat Flour 5% 5% 32% 7% 8% 50% 67% 53% Rice –3% 0% –12% 1% –11% 4% 20% 4% Lentils 22% 17% 28% 28% 22% 74% 91% 71% Onions 46% –6% –20% 10% 10% –20% –12% –21% Beans 42% 7% –1% 23% 4% 68% 83% 67% Source: WFP and WBG staff calculations. power and distorted relative prices, contributing to prices reflect trade isolation, friction in maritime and inefficient factor allocation and intensifying inequality. inland trade, elevated insurance and transportation Currency devaluation, however, does not rep- costs, and a potential increase in market concentra- resent the whole story of Yemen’s cost-of-living tion due to Yemen’s challenging operating environ- and humanitarian crises: structural rigidities in ment (see Figures 18, 19). price determinants also pay a significant role. The Although currency crises can affect a popula- cost of goods in dollar terms has been rising faster tion through several channels (beyond prices), it in Yemen than in other countries. When comparing price changes for six essential food items during 86 The group is composed of three MNA FCV countries (Iraq, 2020–21 across a representative comparator group Lebanon, and Syria) and two neighboring countries (Djibouti of countries,86 Yemen experienced the greatest price and Oman). increase for five out of six items.87 Higher structural 87 Excludes one outlier. 24 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 18 Quarterly Real Gasoline is unlikely that many Yemenis living in IRG-held Prices in Urban Areas, regions have sustainably gained from the currency 2016–2021 depreciation. Remittances from abroad and internal remittances from Houthi-controlled areas increased 160 in value as the rial depreciated in IRG-held regions. Fuel blockade 140 Workers and businesses producing export goods 100 could have benefited from relative price changes in Q3 2016 = 100 foreign markets.88 However, most workers in IRG areas 80 are either underemployed or dependent on fixed 60 government salaries.89 A recent phone survey con- 40 ducted by the World Bank found that the main source of income for 64 percent of households is wages, not 20 remittances or export-oriented cash crops. 0 The monetary situation in IRG areas showed Q3–2016 Q4–2016 Q1–2017 Q2–2017 Q3–2017 Q4–2017 Q1–2018 Q2–2018 Q3–2018 Q4–2018 Q1–2019 Q2–2019 Q3–2019 Q4–2019 Q1–2020 Q2–2020 Q3–2020 Q4–2020 Q1–2021 Q2–2021 Q3–2021 Q4–2021 signs of improvement in late 2021. A successful new foreign-exchange auction mechanism, a reshuf- Hudaydah Marib Sa’ada Sana’a Taiz Aden fling of the central bank’s management, and a Presidential Decree requesting an audit of central Source: WFP data and WB Staff calculations. bank operations since 2016 have bolstered confi- dence in the government’s capabilities and oversight of monetary policy, albeit starting from a low base. Renewed attempts at macroeconomic stabilization helped relieve pressure on the exchange rate and Figure 19 Changes in Fuel Prices caused the rial in Aden to appreciate to about YRI Following the Announced 1,100:US$1 by the beginning of 2022. During the Air and Sea Blockade course of 2022, the announcement of a US$3.3 bil- 0.5 lion dollar package by Saudi Arabia and the United Arab Emirates had an immediate positive effect on 0.4 the exchange rate, though these gains were partially reversed when the promised funds failed to materi- 0.3 alize. The subsequent stabilization of the rial in IRG areas may also be due in part to the growing formal- 0.2 ization of remittances from Saudi Arabia to Yemen 0.1 and the IRG’s increased pressure on currency traders and banks to comply with local regulations. However, 0.0 some of these benefits may be zero sum for Yemen as a whole: for example, greater remittances into IRG –0.1 areas may deplete hard currency in Houthi regions. September October November December January February 95% CI – Upper 95% CI – Lower Estimate 88 See Chen (2006) for an example where a subset of workers Notes: This figure presents the change in fuel prices leading up to benefitted from Indonesia’s currency crisis in the late 1990’s, and following the complete blockade in Nov 2017. The estimates rep- resent the percentage difference in prices relative to average prices and a subset were especially harmed. in Aug 2017. 89 Yemen: 2019 Humanitarian Needs Overview, OCHA, 2019. Costing the Conflict 25 The Weaponization of Trade changes occurring even when supply and demand dynamics are not meaningfully affected. Conflict escalation has at times been accom- The announced blockade had the greatest panied by the conflict parties’ imposition of impact on better-off households, which tended barriers to movement of key commodities, to spend more on fuel. Better-off households including a partial blockade of the ports that experienced the greatest deterioration in food supply Houthi-controlled regions. The strength access after the blockade was announced. At the and targets of these barriers have changed over time, approximately one-quarter of Yemeni house- time. At times, maritime inspections to verify that holds used generators as their primary electricity commercial shipments do not contain weapons source, and this share was highest among house- have restricted imports, and armed forces have peri- holds with adequate access to food. By contrast, odically blocked fuel distribution within Yemen.90 households with inadequate access to food were Although it is difficult to precisely gauge the unlikely to spend money on fuel. Combined, these intensity of the disruption to supply chains, reports patterns suggest that large changes in fuel prices suggest that immediately following the escalation are more likely to affect better-off households. of the conflict in 2015 imports in Hodeidah—a Some types of fuel spending, such as for genera- port that supplies nearly the entirety of Houthi- tors or to reach jobs and basic services, may be controlled Yemen—were immediately reduced to 40 price inelastic, at least in the short term, and house- percent of pre-conflict levels.91 This partial blockade holds using fuel would experience large short-term was particularly harmful, as the country imported increases in the share of resources devoted to fuel. approximately 90 percent of food staples like wheat, In such an event, these better-off households might and 100% of its rice, prior to the conflict.92 have to adjust their food consumption to cope until While data constraints make it difficult to pre- they can pivot from fuel consumption over time. cisely measure how the trade barriers imposed Individual events can impact price levels over the course of the conflict affected prices, even if they do not directly disrupt supply. For behavioral changes provide important evidence. example, during IRG-backed military advances on After Houthi forces took credit for a missile strike on the Houthi-controlled port of Hodeida in 2018, Riyadh’s international airport, the Saudi-led coali- food prices started to increase rapidly nationwide. tion backing the IRG announced a nationwide air These increases peaked in October, clearly coincid- and sea blockade in November 2017.93 Though the ing95 with a spike in reported conflict fatalities. The announced blockade lasted for just eight days in offensive in Hodeida started in June 201896 and was IRG areas and 12 days in Houthi areas, and trade followed by an increase in commodity prices in July. data show that shipments to key ports continued, A UN-sponsored truce was signed in mid-Decem- households and markets adopted coping strategies ber,97 and prices returned to the baseline in January in anticipation of shortages and high prices. Fuel 90 Consequences of Forced Displacement in Active Conflict, prices rose in all parts of the country in response Shandon, Favari et al., 2021. to the announcement and did not return to earlier 91 In blocking arms to Yemen, Saudi Arabia squeezes a starving levels once the blockade ended. However, there was population, Gebrekidan and Saul, 2017 92 WFP (2012). little immediate impact on food prices, which expe- 93 World Report 2017: Yemen, Human Rights Watch (2017. rienced small month-to-month increases similar to 94 Tandon and Vishwanath (2020). previous changes during the conflict.94 This pattern 95 Within a 1-month lag from each other. indicates that fuel prices may be especially sensitive 96 The Wall Street Journal (2018, June 13). to market-moving information, with significant price 97 The New York Times (2018, December 13). 26 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 20 The Impact of the Announced Air and Sea Blockade a. Share with poor food access b. Share with deprivations in non-food dimensions 65 65 60 60 55 55 50 50 45 45 40 40 35 35 30 30 25 25 20 20 15 Water Primitive No municipal Deprivation – 15 insufficient electricity waste 3 or more Poor food access – Poor food access – source collection Food Consumption Score (FCS) Reduced Coping Strategies Index (rCSI) Pre-blockade Post-blockade Pre-blockade Post-blockade c. Average price of fuel d. Average price of common food items 345 345 325 325 305 305 285 285 265 265 245 245 225 225 10–2017 11–2017 12–2017 1–2018 2–2018 10–2017 11–2017 12–2017 1–2018 2–2018 e. Average food access measures by pre-existing welfare status f. Persistence of food security impact on better off households 60 62 55 60 50 58 45 56 40 54 35 52 30 50 25 48 20 46 15 44 FCS rCSI FCS rCSI Pre-blockade nov Post-blockade nov 12–2017 1–2018 2–2018 3–2018 4–2018 Bottom of welfare distribution Top of welfare distribution Pre-blockade Post-blockade Notes: This figure reproduces key findings from Tandon and Vishwanath (2020) and Favari et al. (2021), which documents how wellbeing and markets were impacted by the complete air and sea blockade implemented in November 2017. All figures are the authors’ calculations using the monthly WFP mobile phone surveys. Costing the Conflict 27 Figure 21 District-Level Food-Basket Prices (blue) and Reported Fatalities (orange) 4,000 3,000 YER 2,000 1,000 0 2017 2018 2019 2020 2021 Note: Increased fatalities between Jul 2018 and Dec 2018 (dashed lines) coincide with the port of Hodeidah conflict. 2019. Traders across Yemen, not just in Houthi areas, sector. These shifts in environmental factors and appear to have anticipated future supply shortages production patterns are less obvious but have con- and increased prices accordingly (see Figures 20, 21). tributed to the rising cost of living. Higher food prices Other factors that may have contributed to may reflect a shift in cultivation from food crops to price increases, including alleged price manipula- qat, a mild stimulant widely consumed across Yemen. tion by authorities in Houthi-controlled regions,98 Machine-learning models and remotely sensed data are more difficult to assess. But some general enable an estimate of changes in qat cultivation areas effects of price increases can be inferred. The in selected Yemeni regions from 2016 to 2021, reveal- diversion of imports to more distant ports across the ing that the area devoted to qat increased by over front lines of the conflict further increases transpor- 40 percent during the period. Figure 22 (below) shows tation costs, albeit in many cases not to the extend the range of qat area changes, as a percent change that explains price increases in Houthi areas.99 In a from the baseline (indexed to 100) in 2016. While country without a functioning electric grid, high fuel estimates vary at the district level, the overall trend prices limit the ability of businesses and households is clear, as the median increase is 80 percent. to rely on diesel generators, which provide more Qat cultivation has expanded faster than reliable energy than solar panels.In addition, busi- overall cultivation, suggesting that Qat is nesses are forced to pay higher costs for intermedi- expanding at the expense of food crop lands. ate inputs, further diminishing the competitiveness While qat cultivation grew by 40 percent in the of domestic goods and services.100 selected governorates, total cropland increased Agricultural Shifting Production Patterns 98 How Yemen’s War Economy Undermines Peace Efforts, Inter- national Crisis Group, 2021. Shifting production patterns towards cash com- 99 ACAPS (2020). modities are also present in the agricultural 100 Favari et al. (2021). 28 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 22 Changes in the Area Under governorate, which was independently identified Qat Cultivation in Nine by several local Yemeni experts as a major area of Selected Governorates qat production. In this case, the increase has been massive, as our model predicts a doubling of qat 200 cultivated areas in the district (see Figures 22, 23). Qat use and production on such a massive Percent change in predicted qat area 150 scale likely decreases Yemen’s economic pro- 100 ductivity. While it is reasonable to assume that increased Qat production has a positive effect 50 on overall economic output—since the value added of the qat is likely higher than the agri- 0 culture plantations that would have theoretically –50 otherwise been cultivated—Qat poses risks for large productivity losses, hinging on significant 2016 2017 2018 2019 2020 2021 health risks and lack of basic nutrition. While data Median 20th to 80th percentile limitations do not allow for precise estimates of Note: Changes are shown as a percent deviation from the baseline (indexed qat consumption, it is possible to assume that an to 100) in 2016. 30m pixel predictions for each district were aggregated increase in qat production equates to an increase into an average for the nine selected governorates. in qat use. Considerable amounts of anecdotal evidence suggest that agricultural land is being converted from food crops to qat plantations at a large scale, reducing the domestic food supply Figure 23 Qat Cultivation in Otmah while consuming significant water resources in an District, Dhamar already severely water-scarce environment. 125 Transportation and Mobility Constraints Percent change in predicted qat area 100 Prior to the conflict, Yemen’s road network 75 was basically adequate, and its scope and qual- ity was improving. Internal transport corridors 50 expanded significantly after the unification of 25 Yemen in 1990.102 By 2015, the paved road network had more than tripled to about 17,000 kilometers 0 –25 101 Cropland estimates come from the Moderate Resolution Imaging Spectroradiometer (MODIS), a group of satellites 2016 2017 2018 2019 2020 2021 orbiting around the Earth, viewing the entire planet’s sur- face once every 1 to 2 days, and acquiring data on land, oceans, and the lower atmosphere. NASA Website, https:// modis.gsfc.nasa.gov. by 22 percent.101 In relative terms, qat plantations 102 According to data from Yemen’s Public Corporation for expanded from 16 percent of total crop area in 2016 Roads and Bridges, after reunification, the country counted to roughly 20  percent in 2020. The figure below 5,000 kilometers of asphalt roads, 2,100 gravel roads, and shows the trend for Otmah district, in the Dhamar 44,000 dirt roads. Costing the Conflict 29 Table 4 Paved Roads, Population, and Land Area by Governorate, 2015 Governorate Paved Roads in KM Population (2015 Est.) Approximate Area in KM2 Abyan 535.9 590,977 21,939 Aden 211.0 818,004 1,114 Ad Dali’ 238.4 637,285 4,786 Al Bayda 654.5 778,013 11,193 Al Hudaydah 1176.5 2,934,568 17,509 Al Jawf 415.4 595,542 30,620 Al Mahrah 960.0 120,987 122,500 Al Mahwit 403.1 668,746 2,858 Amran 1012.0 1,188,300 9,587 Dhamar 910.5 1,794,719 10,495 Hadramawt 2866.5 1,405,552 191,737 Hajjah 750.7 1,995,789 10,141 Ibb 1123.9 2,867,076 6,484 Lahij 603.6 979,590 15,210 Ma’rib 443.7 322,224 20,023 Raymah 235.9 531,421 3,442 Sa’dah 548.3 938,734 15,022 Sanaa (incl. Amanat al Asimah) 1921.4 3,653,000 15,178 Socotra (Socotra was part of Hadramawt in 2015) 2,359 Shabwah 1086.7 636,117 47,728 Taiz 1202.3 3,230,218 12,605 Total 17300.2 26,686,862 572,530 Source: World Bank staff based on CSO data. and extended to most large urban centers.103 The produce the bulk of agricultural output and where World Bank-funded Rural Access Program was most economic activity occurs. The country’s largest launched in 2001 to provide equitable all-weather markets, manufacturing centers, and business hubs road access to the rural population. As a result of are concentrated in these same areas, but most of these and other efforts, Yemen’s pre-war road net- the local population is spread across 133,000 vil- work was better than those of several comparator lages and small rural settlements, where tenuous countries, with 60 percent of the paved network road access constrains the flow of goods and the rated as “good” or “very good,” leading to the movement of people, driving up the cost of liv- rise of a substantial vehicle fleet and high traffic ing.105 Because Yemen depends heavily on imported volumes on the arterial network (see Table 4).104 goods, and because there are no alternatives to Connectivity is vital to trade, productiv- ity growth, and poverty reduction in Yemen. 103 CSO Statistical Yearbook (2015); World Bank (2014). A large share of the population lives in rugged 104 World Bank (2019a). areas in the western and central highlands, which 105 World Bank (2019b). 30 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES overland trucking for freight transport, prices and Figure 24 Intra-Urban Road Damage competitiveness are highly sensitive to the quality by City and accessibility of the road network.106 Airstrikes, shelling, and landmines have dam- 100 aged and destroyed roads and bridges across the 30% country.107 Internal frontiers between different areas 80 49% 49% 60% 60% 64% 63% 68% 69% of political-military control have severed road access, 72% 73% 71% 75% 79% 80% and overland borders that previously accounted for 60 85% 90% Percent a large share of trade and tourism have been closed. 70% 40 The conflict has sharply curtailed international and 51% 51% domestic market access, increased supply-chain 40% 40% 37% 36% 20 31% 32% length and complexity, and left hundreds of thou- 29% 28% 27% 25% 21% Ad-Dhale 20% Dhamar 15% Khoka 10% sands of Yemenis trapped in a state of partial siege 0 in towns and cities. Destroyed road infrastructure Aden Al Hazm Amram Bayhan Hodeidah Lahj Lodar Ma’rib City Mocha Rada’a Sa’da Sana’a Taiz Total and physical barriers have intensified the truncation of the economy while exacerbating Yemen’s physical No damage Damaged and political fragmentation. The war has reversed hard-won improvements in transportation access for Source: World Bank estimates. poor rural communities, and most of the population still lives in areas that are barely serviced by road corridors (see Map 4).108 via the branch roads to the main north-south and Roads and bridges along key transport corri- east-west highways. Only one entrance links the dors have been extensively damaged, especially southwestern part of the city to the district of At in the governorates Sadah, Amran, Sanaa, Taiz, Turbah, which provides road access to other parts Marib and Aden.109 The main highways connect- of the country under IRG control. The Houthis con- ing Yemen’s major northern and southern cities— trol Hawban, Taiz’s economic hub, which includes including the N1 route linking Taiz, Ibb, Dhamar, and some of Yemen’s major food industries and where Sanaa with Aden, as well as the smaller Aden-Ad most Taiz residents are employed. Citing security Dali highway, which is the shortest route between concerns, the Houthis have blocked the 6km road Aden and Sanaa—have suffered significant dam- route linking Taiz with Hawban, forcing travelers age and are severed at key points. Houthi forces to take an alternate 100km route. As a result, what have destroyed bridges and other infrastructure was once a 10-minute commute can now take five and have blocked access to roads and highways to eight hours. Moreover, transporting goods from for defensive purposes or to channel trade through Aden to Taiz may take several days due to poor road taxable routes. About two-thirds of Yemen’s popula- conditions and a complicated network of alternative tion resides in those governorates, and the highway system is vital to their livelihoods (see Figure 24).110 106 Yemen’s Economy: Oil, Imports and Elites, Chatham House, 2011. Restricted movement in and around the city 107 See, e.g., al-Tairi, Nabil, The Road Transport Sector in Yemen: Critical Issues and Policy Priorities, Development Champions, of Taiz and between Taiz and neighboring gov- March 2022. ernorates has affected about 4 million people. 108 Ibid. Between 2015 and 2017, continual conflict over 109 Yemen Transport Sector, Input to the Yemen Policy Note no. the densely populated city of Taiz ended with the 4 on Inclusive Services Delivery, World Bank, 2019. Houthis controlling the city’s main access points 110 World Bank, 2014. Costing the Conflict 31 Map 4 Transportation Routes Roads Main Grain Storage/ Airports Main roads Silo Location Open Secondary roads Source: WFP-Logistics Cluster Main border crossings Closed SAUDI ARABIA Open Closed National Capital Ports Roads Governorate Capitals Open Closed Main towns Closed Difficult to access International Boundaries Source: WFP-Logistics Cluster Open Governorate Boundaries OMAN Al Mazyunah [open] Al Abr/Al Wadiah SAADA [open] Harad/At Tuwal Saada Hawf [open] [closed] AL AL JAWF HADRAMAUT MAHRAH Washa’ha Qarah Al Ghaydah R Kushar AMRAN Al Hazm Tareem ed HAJJAH Al Jamimah As Saoum Khulna Ash Sharaf Al Qatn Seiyun Se AMANAT a Hajjah Amran AL-ASEMAH Shibam MA’RIB Al Mahwit Ma’rib As Salif AL MAHWIT SANA’A Ras Isa (SBM)Al Haymah Ad Dakhili Bani Ras Isa Manakha SANA'A Matar SHABWA (Shore Tanks) Sa'fan Al Haymah Al Khariji Al Jabin DHAMAR Markhah Al-Hodeidah Baihan Assufla Ash Shihr RAYMAH Dhamar Ataq REP. OF Markhah Nisab Al Mukalla YEMEN AL BAYDA Al Olya AL-HODEIDAH IBB Hateeb AL Al Bayda Ibb DALI Balhaf ERITREA Taiz Al Dali ABYAN TAIZ Mokha LAHJ Zinjibar Lahj ETHIOPIA Aden Gulf of Aden Hadibu ADEN S O QAT R A Socotra 'Abd al Kuri Samha Darsa DJIBOUTI The Brothers 0 50 100 150 Kilometers IBRD 42861 | APRIL 2018 This map was produced by the Cartography Unit of the World Bank Group. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of the 0 50 100 Miles SOMALIA World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. Source: World Bank Group. routes, and the costs of transporting goods from study111 found that total transportation expenses Aden to Taiz using the 128km Aden–Ar-Rahidah increased 175–195 percent when trailers and heavy road are estimated to have roughly doubled since trucks used secondary roads instead of one and the beginning of the war. Goods traveling from two-lane national roads and by 298–330 percent Aden to other Houthi-controlled areas via Taiz when they used difficult feeder roads. Moreover, must now move along poorly maintained mountain this analysis focused on the increase in mileage only, routes to the south of Taiz before navigating the and it was conducted before the war, when truck- Hawban detour route (see Map 5). ing was not disrupted by unexpected blockades or Longer and more complex road routes have outbreaks of fighting. Due to the risks associated changed the economics of overland freight. Longer travel times increase transportation costs, 111 Phase II, Transport-Cost Report, Development of a National both for the delivery trip and the backhaul. A 2009 Port Strategy for Yemen, the Cornell Group, March 2009. 32 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Map 5 Hawban to Taiz (left) and Aden to Taiz (right) Source: International Crisis Group. Map 6 Major Transport Routes Before the Conflict Costing the Conflict 33 with the job and a worsening shortage of skilled times make it infeasible to transport perishable workers, experienced truck drivers command high goods. Intra-city transport rates have increased by and rising wages. an average of 12–14 percent over the course of the Most goods entering the Aden port and war, and these increased costs are passed through bound for Houthi areas now pass through the to consumers (see Maps 6, 7, and 8).112 northern city of Marib, as do goods imported Before the war, around 50 percent of via other IRG-controlled ports and land borders Yemen’s non-extractive exports and 60 percent with Saudi Arabia and Oman. The 160km road of its imports transited through the At Tuwal connecting Sanaa with Marib remains closed due overland border crossing, which connected to fighting, forcing travelers and truckers to take Yemen with Saudi Arabia. Most of these exports alternative routes through Al Bayda, adding nearly were agri-food products and fishery products.113 400km and 7–10 hours to the journey, or across The total value of the cross-border trade passing the desert via northern Marib and Al Jawf. These routes include dirt roads in poor condition, adding 112 For cost calculations, see Annex B. to safety concerns, while elevated costs and travel 113 UN Comtrade Data, https://comtrade.un.org/data. Map 7 Current Status of Major Transport Routes 34 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Map 8 Fuel Supply and Distribution Routes through At Tuwal was estimated at US$1.5bn. The 1.5 million Yemeni travelers used these three land closure of this border since 2015 has imposed borders annually for inbound and outbound trips.115 major constraints on transportation, trade, and The closure of overland borders has compounded the economic activity.114 At present, the only overland economic losses caused by the conflict and pushed border crossing between Yemen and Saudi Arabia a large share of Yemen’s population into poverty. that remains open is Al Wadiah, in the eastern governorate of Hadramawt, about 600km from Underemployment and Vulnerability Are At Tuwal. on The Rise Based on data from the Ministry of Tourism, more than 53 percent of the 1.3 million tourists To investigate how households are responding who entered Yemen on average each year from to the challenging context, this section reports 2012–14 passed through the overland border 114 Source: Yemen CSO, Foreign Trade, Table 13, Exports and crossings of At Tuwal, Al Buqa, and Alab, all of Re-Exports by Customs Ports, and Table 12, Imports by Customs which are now closed. Most tourists were Saudi Ports: 2013 to 2017. nationals, who would be unlikely to travel to Yemen at 115 Source: Yemen CSO, Tourism, Transport, and Travel, for the present even if the borders were open. An additional years 2012 to 2014. Costing the Conflict 35 on the results of a random-digit phone survey not waiting for the conflict to end while relying conducted by the World Bank during August on external support. Instead, despite multiple and September 2022. The survey contacted about economic shocks, the widespread destruction 1,300 respondents across Yemen,116 with questions of infrastructure, and the persistence of everyday focusing on livelihoods and labor-market condi- violence, Yemenis continue to work. The private tions. While representativeness is always a chal- sector continues to provide limited livelihood lenge for phone surveys, sampling weights have opportunities, but wages are low and underem- been calibrated to population projections, alleviat- ployment is pervasive. ing some (though not all) potential biases. Given the The survey highlights the large share of scarcity of traditional statistics, this dataset presents households that depend on wage income pro- a unique opportunity to understand household vided by a single household member. According living conditions and labor-market dynamics (see to the survey, 84 percent of households report hav- Figure 25). ing at least one employed household member. In The results indicate that while households addition, 64 percent of respondents report that rely heavily on labor income, employment wages are their most important source of income. alone is insufficient to cover their basic needs. Only 23 percent of households report having a sec- Labor-market conditions and food-security indi- ond source of income, and this secondary source cators differ substantially between the IRG and Houthi areas, due in part to the more reliable 116 A full description of the survey methodology and results can payment of government salaries by the IRG. Job be found in a forthcoming report – Yemen Mobile Phone Survey quality is poor in both areas, but workers are Monitoring Round I, World Bank 2022. Figure 25 Shares of Important Sources of Household Income Wages 64 20 Profit from non-household based enterprise 8 5 Profit from household based enterprise 13 28 Sale of food assistance 1 3 Cash transfers from humanitarian agencies 0 4 International remittances 1 3 National remittances 01 Renting of land, buildings or machinery 11 Support from family or friends 3 13 Pensions 2 4 Other social benefits 1 3 Begging 1 0 Regular government salary 2 2 Irregular government salary 2 11 Other 1 2 0 10 20 30 40 50 60 70 Main income source Secondary income source Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. 36 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 26 Public Employees by Status Figure 27 Most Recent Payment for Public Employees 20 19 Percentage of respondents for each group (weighted) 80 Percentage of respondents for each group (weighted) 18 17 70 16 15 70 14 60 12 48 10 50 46 10 8 40 37 37 8 7 6 30 21 4 20 2 11 13 10 8 0 Working in public sector Currently doing work for 0 before the conflict this public sector job Earlier this Between 2017 Before 2017 All DFA IRG year and 2021 All DFA IRG Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. is also predominantly wages. Household and non- months and 54 percent report having worked in household-based enterprises, such as the sale of the last seven days. By contrast, 90 percent of male crops or animal products, also appear to be impor- tant income sources. The public sector was once a major employer and remains an important source of income. Figure 28 Number of Months That Seventeen percent of respondents report being Public-Sector Employees employed in the public sector before the war, and Received Payment Over 8 percent continue to work in the public sector. Of the Last 12 Months these, almost half of respondents report that the 9 household member employed in the public sec- 8 tor was paid earlier this year, though rates range 8 from 70  percent in IRG areas to just 37 percent 7 in Houthi areas. Public-sector employees in IRG 6 6 areas received their salary an average of eight of 5 the last 12 months, while public-sector employees 4 in Houthi areas received their salary an average 4 of four out of the last 12 months (see Figures 26, 3 27, and 28). 2 While 84 percent of respondents report hav- 1 ing a job, only 70 percent report having worked in the week before the interview. Gender gaps 0 All DFA IRG are substantial. Eighty percent of female respon- dents report having had a job during the last 12 Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Costing the Conflict 37 Figure 29 Share of Respondents Figure 30 Employment Status of Working During the Seven Respondent or Main Days Before the Survey Income Earner 100 100 90 90 85 85 90 80 30 28 33 80 80 70 70 70 60 54 60 50 50 40 40 70 72 67 30 30 20 20 10 10 0 0 All Male Female All Male Female Have a job Worked in week before survey Employed Self-employed Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. respondents report having had a job in the last 12 women report being employed in the public sector months, and 85 percent report having worked in (see Figures 30, 31).118 the last seven days. Given the limitations of tele- Job quality is poor, and low and unreliable phone surveys, these results should be interpreted wage payments are combined with a lack of job with caution and cannot be construed as national security. Respondents describe 33 percent of jobs employment rates. For example, a phone survey is as permanent in nature, with rates ranging from more likely to sample better-off households who 40 percent in IRG areas to 29 percent in Houthi areas. own phones, and within each respondent house- Conversely, respondents in Houthi areas are more hold the person who answers the phone is likely likely to report being engaged in seasonal and occa- to be the working member. Nevertheless, the data sional work (28 and 18 percent, respectively) than shed light on important issues around employment respondents in IRG areas (21 and 7 percent, respec- and livelihoods (see Figure 29) . tively). Less than half of working respondents or main Among working household members, most income earners report being paid on time. More are employed as elementary occupation work- than one-quarter report up to six late payments over ers in the private sector.117 Men and women are the last 12 months; 6 percent report more than 6 late both most likely to cite agriculture as the sector payments; and 15 percent report payments being of employment. Construction is also a significant employer, albeit primarily among male workers. 117 In 81 percent of cases the respondent had a job, even if they Across the private sector, most respondents report had not worked in the last week. If the respondent did not have working in enterprises with five employees or fewer, a job, we asked them to respond on behalf of the main income and women are especially likely to be self-employed. earner, this applied to 103 cases. In 11 percent of cases there Public employment remains substantial, particularly were no working members in the household. among women, and about 15 percent of working 118 Al-Ammar et al 2019. 38 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 31 Sector of Employment Figure 32 Occupation of of Respondent or Main Respondent or Main Income Earner Income Earner 7 Manager 8 7 Female 84 15 1 9 Professional 8 11 Associate professional 5 and clerical worker 6 4 Service and sale 15 Male 90 7 3 16 workers 15 6 Skilled workers 5 6 Elementary occupation 46 All workers 49 87 11 2 43 10 Other 7 14 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 Private sector Public sector/government Other All Male Female Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. late all the time. The average delayed payment is one month late (see Figures 32–36. Most workers reported being satisfied with Figure 33 Employment Sector of their jobs,119 but among the 18 percent who Respondent or Main report being dissatisfied, the main reasons are Income Earner low pay and underemployment. Only 9 percent of households with a working member report that Female 23 5 7 23 12 15 15 employment income is sufficient to cover their expenses, while 35 percent report not being able to cover expenses at all, and 30 percent report Male 29 5 15 21 10 9 11 being able to cover less than half of all household expenses. Reported rates of job satisfaction and abil- All 26 5 12 22 11 12 13 ity to cover expenses are significantly higher among 0% 20% 40% 60% 80% 100% households living in IRG areas (see Figures 37–45. Many households appear to be turning to Agriculture Manufacturing, mining, and other industrial activities self-employment in household enterprises as Construction Wholesale and retail trade, transportation, and accommodation 119 It should be noted that feedback from the survey team dem- Information, financial, real estate, professional, onstrated that the question on job satisfaction was not cultur- scientific, and technical activities Public administration ally appropriate and results might overestimate how satisfied Other services workers are. Nonetheless, results are included here because the reasons for dissatisfaction are important. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Costing the Conflict 39 Figure 34 Occupation of Self- Figure 35 Employment Type of Employed and Own- Working Respondent or Account Workers Main Income Earner Other 45 10% 40 40 35 33 33 Manager 29 Professional 30 27 28 16% 25 6% 24 25 21 18 20 14 Elementary 15 occupation workers Service 10 7 38% and sale 5 workers 18% Associate 0 Permanent Temporary Seasonal Occasional professional and clerical All DFA IRG workers Skilled workers 4% Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. 8% Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Figure 36 Frequency of Payment Figure 37 Duration of Payment Delays Delays 29 49 A couple of days 27 Always on time 48 32 51 37 A month 40 29 30 Sometimes it is late 28 27 (Up to 6 times) 31 1–6 months 25 31 More often that 6 2 6–12 months 1 not it is late 7 2 (More than 6 times) 3 I have not yet 6 been paid 7 15 4 It is always late 15 14 0 10 20 30 40 50 0 10 20 30 40 50 60 All DFA IRG All DFA IRG Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. 40 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 38 Job Satisfaction Figure 39 Reasons for Job Dissatisfaction 100 I am not using 14 my skills 3% 90 18 20 I am not qualified enough/ 80 work is too difficult 4% Work is 70 27 29 25 dangerous 6% 60 50 40 Too little 30 work Pay is too 54 54 56 16% low 20 54% 10 Too much work 0 12% All DFA IRG Dissatisfied Neither satisfied nor dissatisfied Satisfied Lack of flexibility 5% Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Figure 40 Employment Income Figure 41 Food-Security Status Sufficient to Cover Expenses 100 90 100 9 9 9 80 4 4 4 50 48 80 55 22 22 21 70 60 60 30 31 28 40 50 40 25 20 25 35 34 37 30 26 0 All DFA IRG 20 25 28 Not at all Less than half of 10 18 household expenses 0 Around half of More than half of All DFA IRG household expenses household expenses Poor Borderline Acceptable It is enough to cover expenses Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Costing the Conflict 41 Figure 42 Employment Status and Figure 43 Employment Sector and Food Security Food Security 100 15 12 11 Other services 90 25 80 24 28 Public administration 70 26 23 Information, financial, 60 real estate, professional, scientific, and technical 50 activities Wholesale and retail 40 trade, transportation, 64 61 and accommodation 30 59 52 20 Construction 10 Manufacturing, mining, 0 and other industrial All Poor Borderline Acceptable activities Employed Self-employed Not working Agriculture Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. 0 5 10 15 20 25 30 Acceptable Borderline Poor All Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Figure 44 Worked in the Last Seven Figure 45 Job Type and Food Security Days and Food Security 45 41 80 74 40 70 70 33 70 35 32 32 62 28 30 26 27 24 25 25 60 25 22 22 20 20 18 50 14 15 40 10 10 30 5 0 20 Permanent Temporary Seasonal Occasional All Poor Borderline Acceptable 10 Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. 0 All Poor Borderline Acceptable Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. 42 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES a coping strategy. The 2014 Household Budget Figure 46 Employment by Sector Survey, which spanned the Houthi takeover of Sanaa, revealed a substantial increase in the share of households that owned a household enterprise but a sharp decrease in the average earnings from 11.4 19.6 such enterprises.120 These findings suggest that household enterprises are often used as a means to diversify income or to compensate for lost labor income. 31 30.6 The phone survey includes a food-consump- tion score, which reflects the frequency and 6.9 diversity of food consumption over the previ- ous seven days. The score is used to categorize Private sector Public sector households into poor, borderline, and acceptable Own-account worker Employer of others levels of food security. One-quarter of households Contributing family worker have poor food security, and another quarter have Source: ILO Labor Force Survey, 2013–14. borderline food security, consistent with the results of other analyses.121 Food security is an important welfare indicator and can be considered a proxy population.122 Roughly one-third of these workers, for poverty. Differences across zones of control are some 477,653, were civilian employees of admin- substantial, with 28 percent of households in Houthi istrative departments, while a similar number, areas having poor food security versus just 18 per- 459,991, were employed by the defense ministry. cent in IRG areas. Across the country, households Another 182,361 were employed by the interior without working members or with working members ministry, 10,400 in political security, and 122,678 in employed as seasonal agricultural labor are the other roles (see Figure 46). most likely to be food insecure. While these results Historically, Yemen’s public sector has been only point to associations between food security the country’s largest employer. Before the cur- and employment outcomes, they do indicate that rent conflict, the public sector employed more households with less stable livelihood options than 1.2  million workers and supported millions outside of agriculture are more likely to be food of dependent family members. In 2014, the pub- insecure. lic sector’s total wage bill was YRI 1.14 trillion, or about US$5 billion.123 Some 48 percent of the wage The Public Sector Payroll Remains a bill went to administrative civilian staff, 26 percent Complex Challenge to staff of the defense ministry, and 11 percent to staff of the interior ministry. Outside the military, The public sector was the largest employer before the war. According to International Labor 120 Tandon, S. 2019. “When Rebels Attack: Quantifying the Im- Organization statistics, more than 30 percent of pacts of Capturing Territory from the Government of Yemen.” World Bank Economic Review 33 (2), 328–352. Yemen’s pre-war workforce was employed in the 121 World Bank (2022). public sector. Official state figures recorded some 122 Yemen publishes first labour force survey in over 15 years, 1.25 million people on the public payroll. These International Labour Organization, 27 October 2015. salaries may have supported as many as 6 mil- 123 This figure is calculated at the prevailing exchange rate for lion Yemenis, or more than one-fifth of the prewar 2014. Costing the Conflict 43 the education and health sectors accounted for Figure 47 Fiscal Revenuea and the the largest shares of public employment. An esti- Public-Sector Wage Bill, mated 350,000 civil servants were employed in the 2006–2020 education and health sectors, accounting for about 70 percent of the civilian payroll.124 2,500 In 2015, Yemen’s large public payroll already 2,000 posed risks to fiscal sustainability, and the bud- getary stress of the conflict has exacerbated the 1,500 YER bn unaffordability of the wage bill. Between 2006 and 2014, the public-sector wage bill grew by an 1,000 average of 12 percent per year, while annual revenue growth averaged just 8 percent. In 2014, the wage 500 bill totaled YRI 977 billion (about US$4.5 billion),125 0 representing 42 percent of revenues or 10 percent 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 of GDP. After hostilities escalated in 2015, the pub- Hydrocarbon Other revenues Salaries lic payroll swiftly came to exceed the government’s total annual revenue (see Figure 47).126 Source: WB staff calculations based on IMF data in 2010, 2014 and 2021. The public wage bill was widely believed to a The historical data for revenues from 2006 to 2014 cover the whole of Yemen. The World Bank relies on the available IMF revenue data from be inflated by “ghost workers,” employees who 2015 onward. The IMF data include IRG revenue information only. existed only on paper and whose salaries pro- vided additional income for officials and other elites. During the 2012–204 transitional period, controlled by the Houthis, causing the value of the UNDP led an effort to reduce budget leakages by rial to diverge significantly between the two territo- introducing new biometric identification technology ries. Between 2017 and 2020, the Houthis reportedly for public-sector workers. However, the onset of the made irregular salary payments to civil servants in the war halted this initiative. areas under its control. A total of 20 half-month sal- The large public-sector wage bill, the revenue ary payments were made in Houthi-controlled areas shock caused by the conflict, the relocation of over the period, with total outlays equivalent to 10 the government, and the enormous fiscal cost months of full salary payments (see Figures 48, 49).127 of the conflict all contributed to the suspension While the total number of civil servants in of civil-service salary payments in mid-2016. By Yemen cannot be reliably estimated, by 2018 early 2017, the IRG had resumed salary payments in roughly 250,000 civil servants in IRG-controlled areas under its control. Although the monetization areas were receiving salaries, while about of the fiscal deficit enabled the authorities to con- 225,000 civil servants in Houthi-controlled areas tinue paying salaries, the resulting inflation eroded household purchasing power. IRG salary payments to 124 These estimates are based on 2012 biometric registration teachers and healthcare workers in Houthi-controlled data for the education sector, the 2014 wage bill, and the 2016 areas largely ceased in September 2016, when the UNDP rapid assessment. 125 This figure reflects an exchange rate of YRI 215 per US dollar. IRG relocated the central bank to Aden. For a short 126 World Bank (2020). period in 2018 and 2019, the IRG paid the salaries 127 World Bank. 2020. Joint World Bank/UNDP/UNOPS/UNICEF of 50,000 civil servants in Houthi-controlled areas, Discussion Note: Proposed Approach for Facilitating the Pay- but this initiative was suspended once the use of rial ment of Civil Servants in Yemen. Washington, DC: World Bank; banknotes printed after 2016 was banned in areas Stakeholder interviews in November to December 2021. 44 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 48 The US Dollar Value of Figure 49 Number of Monthly Teachers’ Salariesa in IRG- Payments Received by Controlled Areas, 2014 and Public-Sector Workers Over 2021 (US$) the Last 12 Months by Area of Control 400 349 25 350 21 300 20 19 18 250 15 14 200 11 150 10 8 7 7 103 100 5 50 0 0 2014 2021 IRG Working in public sector Currently doing work for before the conflict this public sector job Source: Data shared with World Bank staff by UNICEF. All DFA IRG Contested a Figure used for teachers is YER 74,798 which is the average of the 2014 education payscale. World Bank data suggests that many teachers may Source: WB staff, based on household phone survey 2022. have been paid a lower amount of YER47,972. were not being paid regularly.128 This situation is indirectly. As noted above, 8 percent of phone- ongoing. For example, the CSO has indicated that survey respondents report that a household mem- its staff receives, on average, only one month’s sal- ber is currently employed in the public sector and ary every six months. The continued nonpayment that almost half these public-sector workers were of civil-service salaries in the 11 governorates con- paid earlier this year. IRG areas lead Houthi areas trolled by the Houthi has led to absenteeism and in public employment and recent payment, and to informal payments being demanded at schools over the last 12 months public-sector employ- and health centers. UNICEF reports that 171,000 ees in IRG areas received their salary twice as teachers, or about 64 percent of all teachers in often as their counterparts in Houthi areas (see Yemen, have not received regular salary payments Figures 50, 51). since 2016.129 In the health sector, UNDP and media reports indicate that 26,000 to 30,000 health workers 128 IGC. 2020. Public sector salaries in Yemen: Towards building are not receiving regular salary payments.130 In areas confidence in the peace process. London, UK. where payments were made, the 2014 pay scale 129 UNICEF. ND. “Support for Teachers in Yemen: Education remained in effect until 2019, when the IRG added a Teachers Incentive and Rural Female Teachers.” Presentation. 30 percent allowance to compensate for the impact 130 UNDP. 2016. “Yemen Rapid Assessment Report for December 2016: Measuring the impact of the public sector wage suspension of inflation. Nevertheless, the purchasing power of on basic service delivery in the healthcare and education sec- civil-servant salaries has fallen drastically since 2014. tors.” See also: “UN cuts extra pay for health workers in Yemen Despite these challenges, the public sec- just as COVID-19 hits” in New Humanitarian May 7, 2020. https:// tor remains an important source of income for www.th​enewhumanitarian.org/news/2020/05/07/coronavirus​ many Yemeni households, both directly and -health-yemen-unpaid-world-health-organisation-cuts. Costing the Conflict 45 Figure 50 Public Employment by Area Figure 51 Last Payment Received by of Control Public-Sector Workers Area of Control 80 71 10 70 9 9 60 47 48 47 8 50 38 7 6 40 33 33 6 6 30 23 19 5 20 14 10 4 3 10 5 3 0 Earlier this year Between 2017 Before 2017 2 and 2021 1 All DFA IRG Contested 0 All DFA IRG Contested Source: WB staff, based on household phone survey 2022. Source: WB staff, based on household phone survey 2022. The UN has sought to negotiate an agree- In 2020, the UN revived the question of sal- ment on salary payments between the Houthis ary payments during talks over an initiative to and IRG since at least 2016. The then-Hadi prevent a battle for the oil-rich governorate of government’s subsequent decision to move the Marib, but the draft Four-Point Plan produced by Yemeni central bank headquarters to Aden that the talks was never ratified. During these rounds September was motivated by the government’s of negotiations, both sides had agreed that joint belief the Houthis were using the bank to under- payments should be for civilian public employees write the cost of their war effort. When the gov- only, but another setback occurred Throughout ernment resumed payment of salaries In 2017, it these different rounds of negotiation, the central largely focused on doing so in areas under its con- focus was restoring payments to civil servants listed trol. Citing a lack of liquidity of Yemeni Riyal notes in the government’s 2014 payroll. While the govern- and the loss of oil revenues, the Houthis paid civil ment and the Houthis debated who counted as a service salaries only sporadically and usually in half- civilian—the Houthis claimed that state security month increments. UN-led economic mediation employees including paramilitary forces under the efforts received a boost in December 2018 when, authority of the interior ministry were civilians while as part of the Stockholm Agreement that pre- the government argued that only administrative vented a battle for the Red Sea port of Hodeida, staff working outside of the interior and defense the parties agreed to pay salaries nationwide using ministries should be paid—they appeared to agree a mix of revenues from Hodeida and customs, tax, that defense ministry employees should not be paid and oil-export income. However, subsequent talks under the joint mechanism. But this changed when, over the mechanisms for collecting and distribut- in late September 2022,when the Houthis began ing funds stalled, and no salaries were paid under insisting that oil-export revenues should be used the agreement. to pay defense personnel. 46 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES The Private Sector Has Proven sector provided more than 60 percent of healthcare Remarkably Resilient services. The private sector is also responsible for 100 percent of agricultural output, which provides In the absence of crucial governmental functions, about 20–25 percent of domestic food consumption. the private sector has assumed an ever-greater In 2016, the private sector oversaw 899 basic and sec- role in managing the country’s interrelated secu- ondary schools and 101 higher-education institutions rity, economic, and demographic challenges. with 83,177 students. The collapse of public service Yemen’s economy is highly labor-intensive, with delivery has been accelerated by the war, the public- labor contributing 60.7 percent of GDP. The private sector salary crisis, and the suspension of operational sector contributed about 63 percent to GDP from financing for public facilities in most governorates. 2006 to 2014, since 2015 this share has risen to about However, the private sector has continued to provide 70 percent as the public sector’s contribution fell education and health services. Moreover, humanitar- amid declining oil and gas production.131 The pri- ian organizations rely on Yemeni businesses to provide vate sector’s average contribution to total gross out- logistics and supply intermediate goods and services, put rose from 64 percent in 2010–14 to 82 percent including vital medical supplies during the cholera in 2015–17.132 According to the CSO, 97 percent of 131 Promoting Partnership with the Private Sector, Central the private sector is made up of micro, small, and Statistical Organization (CSO), Socio-Economic Update, Issue medium-sized enterprises (MSMEs), which employ number 53, October 2020. a combined 73 percent of the Yemeni workforce.133 132 CSO, 2017. The private sector plays a major role in social 133 Yemen Country Report, Economist Intelligence Unit, January service provision. Prior to the conflict, the private 2021. MSMEs are defined as firms employing less than 25 people. BOX 2: THE ROLE OF WOMEN IN YEMEN’S ECONOMY Female labor-force participation in Yemen was low even at the start of the civil Figure B2  Female Labor-Force Participation war. The last labor-force survey, conducted in 2013–14, showed 6 percent of Rates, Yemen and Comparators women over the age of 15 engaging in the labor market. This is the lowest female labor force participation rate amongst countries in the Middle East and North 46 50 Africa Region, which already sees some of the lowest participation rates in the 45 world. Yemen has also ranked last out of the 144 countries included in the World 40 34 Economic Forum’s Global Gender Gap Index since 2016.a 35 31 29 29 However, these statistics mask some modest progress in terms of women’s 30 25 21 political empowerment. Women long played an important role in civil society in 17 19 20 15 16 Yemen, leading key groups and civic movements since the 1990s. The 2000s 13 15 11 saw the first female minister appointed. Most notably, women were active 10 6 participants in the popular uprising that removed Saleh from power. During the 5 National Dialogue Conference of 2013–2014, women made up 28 percent of 0 Egypt, Arab Rep. Iran Jordan Lebanon ILibya Oman West Bank and Gaza Saudi Arabi Sudan Syrian Arab Republic Yemen, Rep. Middle East & North Africa World delegates. The dialogue’s outcomes included a number of positive gains for women, including a minimum threshold of 30 percent participation for women in electoral politics, which was signed into law in 2014, and an increase in the minimum age of marriage to 18. However, setbacks have limited the progress made over last two decades in increasing women’s participation in political life. The formation of a recent government to include the STC in late 2020 had sidelined women’s (continued on next page) Costing the Conflict 47 BOX 2: THE ROLE OF WOMEN IN YEMEN’S ECONOMY (continued) participation, infuriating women activists. This was seen as a weak commitment to women on the political front and disrespectful of the sacrifices of women who had fought hand in hand with men during the demonstrations against Saleh.b The absence of women in the peace negotiations, which has been heavily male dominated, was yet another negative signal, undermining the role of women when deciding the future of Yemen.c Beyond the peace negotiations, the civil war has had very complex impacts on women and girls with lasting and often harmful effects on their economic and social well-being. Women have been disproportionately affected by the war’s direct and indirect effects.d Some 36 percent of girls are out of school,e and women make up around three-quarters of the estimated 4.3 million people displaced over the course of the war—a number that has risen since 2020 due to fighting in Marib and Shabwa.f Because many men have been injured, died in combat, or lost income from public sector jobs, women have had to bear the full economic burden of sustaining large families. Unfortunately, there seems to have been significant increases in child marriage and gender-based violence.g At the same time, the civil war has opened up new opportunities for women, who have assumed responsibilities previously reserved for men. By analyzing the Yemen Household Budget Survey of 2014 and the Yemen Human Development Survey of 2021, we find improved intrahousehold female empowerment in districts of Southern Yemen that experience medium levels of conflict (Ishak et al 2022). This is explained by explained by changes in household composition, whereby men are more likely to leave the household in conflict-affected districts, leaving women in charge of household decisions. There is also emerging evidence of changes to women’s involvement in economic life. For example, a 2018 survey found that women were more likely to be working for NGOs than men and “During the conflict, women have entered professions previously closed to them by cultural restrictions, like waitressing and retail” Al-Ammar et al. (2019, p. 6). Key informant interviews also point to an increase in women setting up microenterprises, such as bakeries, within their homes. This is also confirmed by the Yemen Mobile Phone Survey, reported in Chapter 2, which show that women are either likely to work in the public sector or are self-employed. The international experience suggests that such shifts in norms can lead to large permanent gains for women. Gay (2021) shows how the First World War in France led to large increases in female labor force participation to compensate for the “missing men” in the army and those who were killed in the fighting, and this led to permanent gains for women in wages and employment status. While there has been limited evidence of permanent gains for women in conflict-affected countries outside of Europe, Tripp (2015) finds that in Africa the active role of women in civil war is associated with significant post conflict gains in political representation and officeholding. In Yemen, women have been directly involved in the fighting on all sides.h The end of the war will, therefore, likely represent a critical juncture for women where well targeted interventions may have a powerful effect on consolidating some of the benefits they have secured during the conflict, and past political gains can be restored. Understanding in more detail what these are, and how they can be preserved, is an important analytical task that should be taken forward in the near future. While there are some positive signs for a changing economic role for women, the evidence thus far does not paint the full picture. Moreover, the context of Yemen, with low levels and unequal human capital, is very different to a post-war Europe which saw lasting social change. a Global Gender Gap Report 2022, World Economic Forum, 13 July 2022. b Shujaldeen, 2020. c https://www.civicus.org/index.php/media-resources/news/interviews/5658-yemen-women-are-completely-absent-from-decision-making-bodies-political​ly-we-don-t-exist. d See Rohwerder 2017 for a review of available evidence. e International Rescue Committee (https://www.rescue.org/article/4-ways-war-yemen-has-impacted-women-and-girls). f Source: United Nations (https://news.un.org/story/2020/03/1058591). g Al-Ammar et al. 2019. h Ibid. outbreak in April–September 2017.134 Women play the private sector’s resilience to shocks and crises. an important role in the Yemen economy (see Box 2). Meanwhile, inadequate infrastructure and support In addition to macroeconomic imbalances, services, including limited electricity coverage, inad- weak institutions, and an unfavorable business equate water supplies, and poor transportation and environment, Yemen’s private sector has histori- 134 Private Sector: Vital Role in Times of War, Yemen So- cally suffered from structural vulnerabilities. The cio-Economic Update Issue 35, Central Statistical predominance of the family-based business model Organization (CSO), July 2018, https://reliefweb.int/sit​ and the large share of MSMEs have weakened es/reliefweb.int/files/resources/YSEU35_English_Final​.pdf. 48 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES telecommunications systems, have undermined the at ports of entry, and increased insurance costs for private sector’s competitiveness. Insufficient work- ships coming into Yemen.138 force skills and a wide skills gap are serious chal- Some private actors have struggled, others lenges to productivity, along with the crowding-out have rebounded, and others yet have thrived. effect of government spending on private consump- The fragmentation of Yemen, the country’s economy, tion, the absence of a robust financial market, low and state-institutions have opened several market levels of innovation, and the minimal use of informa- opportunities for private sector actors to provide tion and communications technology (ICT). services that used to be provided by the state— Recent estimates suggest that the private most notably, as discussed above, fuel imports and sector incurred US$27 billion in direct and indi- distribution, the distribution of LPG’, and other rect losses during the first three years of the transportation services. The following table outlines war.135 Moreover, 25 percent of all businesses closed patterns of consistency and changes for five key sec- after 2014, and over 80 percent of firms experienced tors in Yemen—the food, fuel, financial, transport, a drastic drop in sales. Key informant interviews telecommunications, and LPG sectors. identified security issues in conflict zones, economic The conflict has dampened international devastation and financial recession, and the physi- trade and deterred foreign investment. Foreign cal destruction of firms as the main reasons for firm investors and commercial traders find the invest- closures, while surviving firms faced interruptions of ment and trade environments hard to navigate. FDI key services and a dwindling customer base. Firms trends are poor even by the standards of low-income downsized in response, with 52 percent shifting and conflict-affected countries, and net FDI inflows from either large to medium, medium to small, have been negative since 2011. As Yemen has a sub- or small to micro during the conflict, and another stantial binding overhang in its WTO commitments, 20 percent of firms relocated their operations, either it retains the ability to raise tariffs. The highest tariffs domestically or abroad.136 are applied to fish, fruits and vegetables, beverages The 2020 Yemen Business Climate Survey and tobacco, coffee and tea, and animal products. Report found that profits fell at 45 percent However, due to porous borders and widespread of businesses between 2017 and 2019. Survey informal trade, tariff revenue is likely well below what respondents cited exchange-rate instability and a the statutory schedule would indicate. In addition, diminished number of customers as the main rea- non-tariff measures have increased trade barriers sons for declining profits, and the non-payment of since 2015 and became especially acute during the public salaries has weakened consumer purchasing COVID-19 crisis, when restrictions were imposed power and negatively impacted business activities.137 on imports of goods deemed essential for the fight The disruption of public service delivery, costly against the pandemic, including certain chemicals domestic and imported inputs, loan restrictions, and (see Table 5; Figures 52, 53). substantial spending on taxes and other regulatory fees have placed a heavy burden on the private sec- 135 Promoting Partnership with the Private Sector, Yemen’s tor. Firms also face political instability, insecurity, the Central Statistical Organization (CSO), Socio-Economic Update, periodic suspension of exports, import restrictions, Issue number 53, October 2020. 136 The Impact of Conflict on Private Enterprises Report, WBG, widening fiscal and monetary imbalances, rising 2019. levels of crime and informality, which forced almost 137 Yemen Business Climate Survey Report 2020, Impact of 20 percent of firms to relocate their operations, Conflict and COVID-19 on Private Sector Activity, SMEPS, DFID. both domestically and abroad. Businesses also face 138 The Impact of Conflict on Private Enterprises Report, WBG, challenges pertaining to fuel scarcity, lengthy delays 2019. Costing the Conflict 49 Table 5 Challenges and Opportunities for the Private Sector  Sector Key Points Food importa • Major food importers that dominated the market before the conflict maintained and even increased their market share during the conflict. • The decline of local wheat production increased dependency on wheat imports and created additional market opportunities for major wheat importers. • The IRG and CBY-Aden were criticized for their management of the US$2 billion deposit Saudi Arabia provided in March 2018 and for the letters- of-credit system that used funds from the deposit to underwrite critical food imports. Criticism included accusations of favoritism shown to certain food importers. Similar accusations of favoritism were levelled against the Houthis and the Payments and Foreign Currency Committee that manages access to foreign currency on behalf of the Houthis and in coordination with CBY-Sanaa. Fuel importb • The privatization of fuel imports and distribution in Yemen saw private sector actors fill the gaps left by the state-run Aden Refinery Company and Yemen Petroleum Company. • Previously marginalized or unknown fuel importers gained an increased market share in competition with fuel traders established before the conflict began. • New players that own and operate prominent exchange companies entered the market. • Political dynamics and competition between the Houthis and IRG heavily impacted fuel import and distribution dynamics. • Diversion of fuel from non-Houthi to Houthi areas continued parallel-market trading. Financialc • De-risking measures applied against Yemen banks weakened the banking sector, reduced rial and foreign-currency liquidity, and undermined confidence among Yemeni businesses and citizens. • Banks that are less impacted by the challenges facing the formal banking sector continue to grow, including banks that offer Islamic banking services and did not invest in government Treasury bonds, as well as banks that record higher foreign-currency inflows either in the form of remittances or humanitarian funding. • Money-exchange services have grown rapidly, with a rising number of licensed and unlicensed exchange companies. • The increased importance of money-exchange companies for trade financing and salary payments has led to the emergence of working relationships between certain exchange companies and groups of traders, businesspeople, and foreign patrons. • Money exchangers offer credit and saving services that are technically beyond the scope of what their licenses permit. • CBY-Aden awarded several prominent market exchangers licenses to operate as microfinance banks in 2021. • Financial service providers, particularly money exchangers, are frequently accused of currency speculation and arbitrage. Transportd • Local conflict dynamics and economic competition between the Houthis and IRG negatively impact the movement of goods. Major supply routes have been severed, and the same goods may be taxed both by the IRG and the Houthis. • Private logistics, transport, and trucking companies are growing, as illustrated by the expanding distribution of fuel and cooking gas. • Fuel and liquified petroleum gas (LPG) truckers are gaining increased leverage due to the scarcity of skilled workers. Telecome • The IRG and the Houthis have both attempted to extract rents from prominent Yemeni telecom companies Sabafon and MTN Yemen. • Sabafon relocated its headquarters to Aden after consistent pressure from the Houthis, which then appointed its own executives to operate Sabafon’s systems. • MTN sold shares to Emerald International Investment (2021) after deciding to pull out of Yemen due to the adverse business environment. LPGf • The state-run Yemen Gas Company (YGC) has been marginalized in Houthi areas. • The number of actors along the LPG supply chain has increased, with a growing number of private LPG transport companies and the imposition of Houthi supervisors and local leaders in the distribution and sale process in Houthi areas. • LPG intended for non-Houthi areas has been diverted to Houthi areas. • LPG is now imported but was not imported before the conflict. Source: World Bank Group research based on key informant interviews. a https://cdn.odi.org/media/documents/Food_security_in_Yemen_1_-_the_private_sector_and_imported_food_uTu5YXS.pdf; https://cdn.odi.org/media/ documents/Food_security_in_Yemen_2_-_the_role_of_the_private_sector_in_promoting_domesti_oHhlb93.pdf; https://sanaacenter.org/publications/ analysis/13577; https://www.acaps.org/sites/acaps/files/products/files/20201216_acaps_yemen_analysis_hub_food_supply_chain_0.pdf; https://documents1. worldbank.org/curated/en/376891524812213584/pdf/125815-WP-PUBLIC-YemenReportFinalAprilcompressed.pdf. b https://sanaacenter.org/files/Rethinking_Yemens_Economy_No4_En.pdf; https://sanaacenter.org/publications/analysis/10323; https://d2071andvip0wj.cloud- front.net/231-yemen-economic-conflict.pdf; https://www.acaps.org/sites/acaps/files/products/files/20210817_acaps_yemen_analysis_hub_fuel_prices_0_0.pdf. c ACAPS, “Yemen: Yemeni Financial Sector in December 2021,” January 2022; https://sanaacenter.org/publications/analysis/9948; https://cdn.odi.org/media/ documents/Impact_of_conflict_on_the_financial_sector_in_Yemen_Jg6sNMp.pdf. d https://devchampions.org/uploads/publications/files/Rethinking_Yemens_Economy_No11_En-1.pdf; https://sanaacenter.org/files/The_Gas_Crisis_in_Taiz_ War_Mismanagement_and_Exploitation_en.pdf. e https://sanaacenter.org/files/Rethinking_Yemens_Economy-policy_brief_21.pdf; https://developingtelecoms.com/telecom-business/operator-news/12332- mtn-group-exits-yemen-via-stake-sale-to-emerald-international.html. f https://sanaacenter.org/files/The_Gas_Crisis_in_Taiz_War_Mismanagement_and_Exploitation_en.pdf. 50 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 52 Net FDI Inflows, 2007–2020 Figure 53 The Conflict-Induced (% of GDP) Economic Contraction and the Expanding Role of the 7 Private Sector, 2012–19 6 5 20 4 10 3 0 2 –10 1 –20 0 –30 –1 –2 –40 –3 –50 2012 2013 2014 2015 2016 2017 2018 2019 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Real GDP growth rate (%) Yemen, Rep. Low income Private sector growth rate (%) Fragile and conflict affected situations Public sector growth rate (%) Source: World Bank, World Development Indicators. Source: CSO, National Accounts Bulletin 2019. While the conflict has had an extensive and international organizations such as UNICEF and deeply negative impact on the business environ- the World Food Program. Private firms worked ment, about 75 percent of firms are still operat- closely with international agencies to address the ing, and the private sector dominates economic COVID-19 pandemic through the International activity. The public sector’s negative contribution to Initiative on COVID-19 in Yemen. GDP growth has been nearly twice that of the private sector, and the private sector’s share in real GDP Aid Dependency increased from 62.3 percent in 2014 to 70 percent in 2016. Data from the CSO show the private sector’s Yemen continues to experience one of the world’s growing involvement in numerous activities, including most severe humanitarian crises, which has been manufacturing, construction, and services, despite the marked by extreme levels of food insecurity. enormous challenges faced by firms. The resilience Senior aid officials have repeatedly warned of the of the private sector limited the economic downturn potential for famine if vital trade routes are blocked caused by the conflict, and the continued provision or if donors do not fund multibillion dollar aid of goods, services, and job opportunities has helped appeals conducted by the UN and nongovernmental shore up food security and mitigate the damage to organizations.139 Fundraising for Yemen peaked in physical assets and human capital (see Box 3). 2018, when the UN response plan and other fund- The private sector has also bolstered com- ing sources mobilized some US$5.2 billion in aid. By munity resilience and supplied vital logistical, 2022, with COVID-19 driving a pullback in overseas storage, and distribution services for humani- development assistance and the war between Russia tarian relief actors. Private firms execute cash transfers from donors to beneficiaries, including 139 Nichols, Michelle, ‘Countdown to catastrophe’ in Yemen members of vulnerable groups, in partnership with as U.N. warns of famine – again, Reuters, 11 November 2020. Costing the Conflict 51 BOX 3: ALTERNATIVE ECONOMIC OPPORTUNITIES IN WARTIME YEMEN The private sector has not replaced enough of the jobs lost during Yemen’s war to prevent a major unemployment crisis. Yemen’s economic decline has placed enormous pressure on Yemeni households, and especially on young people, to pursue whatever income opportunities are available. These include joining an armed group, engaging in various illicit or informal activities, and starting a new business. • Joining an armed group: For many fighters in Yemen’s war, the decision to take up arms was a matter of financial necessity, particularly among those with no previous military experience. In many cases, armed groups appear to offer more reliable pay than civilian state jobs or private firms. For example, while the Houthis have paid civil servants irregularly or not at all, it has paid its fighters’ salaries consistently and offered additional incentives, such as providing cooking- gas cylinders at the lower “official price” to their households. In Yemen’s central, southern, and eastern governorates, anti-Houthi forces have received varying degrees of foreign patronage during the conflict, providing an added incentive to take up arms. From 2016 to 2019, UAE-backed groups offered some of the highest and most regular salaries. However, this changed after 2019, when the UAE scaled back its involvement in the conflict. • Engaging in illicit or informal activities: The marginalization of the YPC and YGC has given rise to a flourishing parallel market for fuel in Houthi areas and, to a lesser extent, in IRG areas as well. Firms and individuals not previously involved in either the transport, distribution, or sale of fuel have entered different stages of the supply chain and seek to generate profit from selling fuel in local markets where the pump price exceeds the official price. For example, an individual who owns a suitable truck can purchase fuel in Aden or Mukalla and ship it to Houthi areas where fuel prices are higher. Within Houthi areas, roadside vendors sell petrol or diesel in plastic bottles. Smuggling weapons, people, and drugs predates the conflict but appears to have grown in magnitude amid rising poverty rates and weak rule of law. The smuggling of illicit material out of Yemen has expanded to include historical artifacts and exotic plants and animals. As the volume of illicit trade has increased, so too have opportunities in the sector, one of the few that offer a substantial and reasonably stable income. • Starting a new business: With public-sector employment unreliable and various sectors struggling to maintain pre-conflict output levels, many workers have turned to entrepreneurship to diversify their income or replace lost wages. While starting a business is often a coping strategy, several sectors have flourished during the conflict. For example, solar energy expanded by over 50 percent between 2015 and 2017 as the national electricity grid deteriorated, though the subsector’s growth has since levelled off. Solar energy grew fastest in the northern and northwestern governorates under Houthi control. and Ukraine forcing Western countries to reprioritize either in-kind assistance or cash transfers. In 2020, aid, support to Yemen had plunged to US$2.4 bil- the World Bank estimated that cash transfers dis- lion, leading aid agencies to issue renewed warnings tributed by the World Bank and UNICEF benefitted about the potential for a humanitarian disaster.140 1.4 million households. In-kind food assistance pro- The billions of dollars of aid delivered to vided by the World Food Program (WFP) reached Yemen since 2015 have greatly boosted the coun- another 1.4 million households, while other aid try’s economy. In six of the eight years of the war, agencies provided assistance to an additional aid has exceeded 10 percent of Yemen’s estimated 182,000 households.143 World Bank staff found that GDP, peaking at 24 percent in 2018. While no reliable “the combined reach of humanitarian and devel- disaggregation of aid expenditures is available, relief opment programs is enough to cover the entire agencies estimate that about 20–30 percent of exter- Yemeni population” but noted that overlapping nal funding enters Yemen in the form of hard currency assistance to some households meant that other to cover payroll, operating costs, and local procure- households did not receive assistance.144 ment, as well as cash transfers.141 These inflows likely represented credits to Yemen’s current-account bal- 140 Based on OCHA estimates of funding for aid routed through ance of between US$460 million and US$1.5 billion the UN and other aid agencies, and World Bank GDP estimates. 141 Key informant interviews, Jan-June 2022. during 2016–2021. In addition, aid agencies finance 142 Al-Ammar, Ali; Ghorpade, Yashodhan, Social Protection at about 12 percent of Yemen’s cereal imports, easing the Humanitarian-Development Nexus: Insights from Yemen , the hard-currency burden on importers.142 World Bank, April 2021. A large share of foreign aid finances emer- 143 Ibid. gency relief. Millions of Yemeni households receive 144 Ibid. 52 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Yemen badly needs aid. But a growing body Liquefied natural gas (LNG) exports boosted of evidence suggests that large-scale humani- export revenues in the 2010s. The Yemen LNG tarian aid delivered over an extended period project launched as a partnership between the can weaken growth and foster dependency.145 Yemeni government and several major interna- While continued aid is vital to sustain basic con- tional oil companies in 2005. The project repre- sumption and prevent a humanitarian disaster the sented the single largest FDI inflow in Yemen’s history country risks falling into an aid dependency trap, at about US$4–5 billion, and it started producing with significant ramifications for the post-conflict and exporting LNG from a non-associated gas field macroeconomy and humanitarian situation; and for in Marib in 2009. By 2014, gas exports were gener- local markets and development. In addition, in-kind ating about 13 percent of hydrocarbon revenues,148 assistance can distort and weaken local markets, but international oil and gas investors had become although cash transfers are often less effective wary of Yemen. As noted in the 2014 CEM, “invest- when local markets are failing. Cash transfers are ment in production is limited by a deficient legis- also cheaper to deliver than in-kind aid. To prevent lative framework, empty government coffers, and aid dependency from becoming more widespread security threats to exploration firms… The business and entrenched, aid agencies can (i) pivot from environment is characterized by weak governance in-kind to cash-based assistance, (ii) work with the and corruption, as well as by the influence of well- private sector to address local needs and explore entrenched networks.” The impact on foreign inves- opportunities for collaborative development; and tors was noticeable—successive post-2007 bidding (iii) expand local development and cash-for-work rounds failed to draw substantial and credible inter- programming.146 est. The Yemen LNG project encompassed about two-thirds of Yemen’s 430 billion cubic feet of proven The Hydrocarbon Sector gas reserves, and was prioritized over a proposed gas-to-power system for the remaining third could Prior to the war, the oil and gas sector was the have produced inexpensive electricity that allevi- main source of government revenue, exports, ated Yemen’s dependence on imported heavy fuel and foreign exchange. The hydrocarbon sector oil. Political gridlock stalled progress on the gas- accounted for 88 percent of pre-war FDI, which in to-power project in the years before war broke out. turn accounted for 1.5 percent of GDP,147 but the An underdeveloped downstream refining government missed numerous opportunities to and distribution industry led to rising import maximise the sector’s potential. In 2005 and 2011, the dependence. The country’s only major refinery, in government refused to renew Production-Sharing Aden, was not equipped to process the heavier Agreements (PSAs) with international oil firms and grades of crude oil produced at nearby oilfields instead took direct control of several oilfields, appar- and met less than half of local liquid fuel demand. ently in pursuit of increased short-term revenues. Liquefied petroleum gas (LPG) produced at a simple Fiscal crises and the prioritization of short-term cash- flow contributed to chronic underinvestment in pro- 145 See for e.g., Clemens et. al., Counting Chickens when they ductive infrastructure and capital assets. Meanwhile, Hatch: Timing and the Effects of Aid on Growth, The Economic Journal, 122(561), 2012. oil firms and the government failed to reach an 146 https://www.wilsoncenter.org/article/democratizing- agreement on how to share profits from associated development-yemen-beyond-food-aid. gas, which was either reinjected into oilfields or flared 147 Yemen: Unlocking the Potential for Economic Growth, World rather than being used to produce electricity, easing Bank, October 2015. the fiscal burden and improving sector efficiency. 148 IMF (2014). Costing the Conflict 53 refinery in Marib met most local gas needs. As However, internal political disagreements and Yemen became increasingly dependent on imported security threats to LNG facilities have prevented fuel, exorbitant subsidies imposed a mounting fiscal the resumption of LNG exports.149 burden. In 2013, income from domestic oil and gas As international firms withdraw, Yemeni sales covered only 70 percent of the cost of subsi- firms have stepped up investments in the oil and dies. Some fuel subsidies were lifted in 2014, but gas sector. Since 2019, one of Yemen’s largest busi- too quickly and without adequate compensatory ness conglomerates, has expanded its domestic oil measures in place, sparking violent public demon- and gas holdings.150 The company has increased its strations which eventually led to the conflict. stake in a functioning oilfield and is investing heavily Oil output remains far below pre-war levels in an effort to double its output. The company has but could increase significantly given adequate also bought the licenses for two additional fields investment. In 2014, Yemen produced an estimated and intends to invest to bring them to commercial 167,000 barrels per day of oil, but by 2021 oil pro- viability. This and other companies are discussing duction had fallen to less than 60,000 barrels per further acquisitions in the sector. day and LNG exports remained frozen. Government Foreign firms are unlikely to invest in Yemen and industry officials estimate that developing four until peace treaties are signed and visibly effec- oil blocks in Shabwa and Hadramawt governorates tive, and a functioning government has been in could boost oil output by around 20,000 barrels per place for a protracted period. Even under these day, more than one-third of current production, but circumstances, international oil companies may be will require a significant capital investment. The reluctant to operate in Yemen. For the Yemeni pri- state-owned oil companies Safer Petroleum and vate sector, however, the environment is different, as PetroMasila claim to lack the funds to do more than national firms can draw on their local knowledge of sustain basic production and export operations. the political, legal, and security context to mitigate Further increases in output are likely to be limited risk. As one Yemeni businessman stated in an inter- and contingent on the availability of investment, view conducted for this CEM, “We know who to talk technical capacity, and equipment. Yemen currently to and how to talk to them, that is our advantage.” lacks all three. At least seven international oil and Moreover, Yemeni firms are not obliged to comply gas firms have either relinquished their stakes in with international standards, which lowers their oilfields to the government or sold their interests. operating costs and exposure to reputational risk. Some of these firms claim that they did so after the IRG repeatedly blocked their efforts to resume Renewables Potential exports. International oilfield-service companies vacated Yemen at the outset of the war. Some A growing body of anecdotal evidence sug- removed drilling and other equipment from the gests that fuel and electricity shortages and country, while others sold drilling rigs to local firms. rising energy costs have encouraged the adop- LNG represents an unrealized potential tion of renewable energy across the country, revenue stream. With a nameplate capacity of 6.7 million tons per year, the Yemen LNG project’s 149 Bank, consultant, key informant interviews, January 2022-October 2022. revenues could have exceeded US$4 billion in 2022 150 The Hayel Saeed Anam Group of companies is Yemen’s based on average international benchmark prices. largest multinational company by far. Still a family owned and While the IRG would have received only a fraction managed business, within Yemen it manages assets in import of this sum, its fiscal contribution would have been / export, banking, insurance and food manufacturing among considerable given the IRG’s budget constraints. others. See: www.hsagroup.com. 54 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Table 6 Potential Renewable Energy by Source Technical Potential Renewable Energy Source Output (MW) Gross (MW) Practicable (MW) Biomass – Landfill Gas 10 8 6 Geothermal 304,000 29,000 2,900 Hydropower – Major Wadis 12–31 11–30 — Solar Electric (CSP) 2,466,000 1,426,000 18,600 Solar Thermal 3,014 278 278 Wind 308,722 123,429 34,860 Source: Ministry of Electricity and Energy “Renewable Energy Strategy and Action Plan Study Report” (2008). a a Oliver Hiel, Werner Klaus, Richard Lawless, Tobias Leschinsky, Romeo Pacudan, Kuno Schallenberg, Axel Slogsnat and Bernhard Wunder, “Ministry of Electricity The Republic of Yemen: Renewable Energy Strategy and Action Plan. Task 1: Renewable Energy Resource Assessment,”draft report, Lahmeyer International, August 2006, https://moee-ye.com/site-ar​ /wp-content/uploads/2019/06/. particularly in Houthi areas. Even before the of necessity, as the public electricity grid effectively conflict, weak public electricity utilities, inconsis- ground to a halt. The high cost of fuel and recur- tent fuel supplies and pricing, and falling prices ring supply issues made private generator usage for photovoltaic solar panels had encouraged less feasible. For households, transitioning to solar the growth of renewables. This shift has acceler- power or upgrading existing solar energy systems ated over the course of the conflict, as the public would ease pressure on household budgets. The electricity grid largely ceased to function, while Ministry of Electricity and Energy’s 2008 Renewable the high cost of fuel made private generators less Energy Strategy and Action Plan still offers a useful feasible.150 In the agricultural sector, many farmers guide for future efforts to increase renewable energy have turned to solar-powered pumping systems. usage in Yemen. The country has significant wind Nongovernmental organizations also used solar and geothermal energy potential, though develop- energy to power health and education facilities ing geothermal power would require a considerable (see Table 6).151 upfront investment and a long lead time. Renewable Investments in renewable energy could help energy sources other than solar are unlikely to be reduce the cost of electricity. Renewable energy developed while the conflict continues. offers an opportunity for the IRG and the Houthis to reduce the fiscal burden of electricity production, 151 Ibid. which is subsidized both directly and indirectly via 152 https://reliefweb.int/report/yemen/solar-energy-helps​ low rates of collection efficiency. Increased solar -yemeni-hospitals-save-lives; https://yemen.un.org/en energy usage, particularly in the northern and /20029-solar-power-can-safeguard-future-thousands-ye​meni- north-western governorates, was largely borne out children. CHAPTER 3 Yemen’s Transformed Political Economy Fragmented Institutions, Heterogeneous Institutions have become increasingly het- Outcomes erogeneous over the course of the conflict. The country’s different areas of control increasingly Yemen is now divided into at least six distinct resemble autonomous regions, each with its own areas of political-military control.153 These include: unique military, security, and economic manage- (i) the northwest, where the Houthis and their allies ment arrangements, as well as varying levels of rule Yemen’s major population centers; (ii) Taiz, which public service capabilities. Key informants describe is controlled by tribal, political and military factions substantial differences in security standards, service formerly allied with the Hadi government and with provision, economic transactions, law enforcement, close ties to Islah; (iii) Marib governate and northern and the overall governance culture in different parts Hadramawt governorate, which are held by a similar of the country. coalition of forces; (iv) Lahij, Aden, Abyan gover- National institutions are divided and have lim- norates and Socotra, where the pro-independence ited authority. Prior to the war, state institutions were Southern Transitional Council (STC) and broadly largely centralized and managed by line ministries in aligned groups are the dominant force; (v) southern Sanaa. After the Houthi takeover, the Houthis largely Hadramawt governorate, Shebwa and al-Mahra gov- kept the existing institutions in place. While many tech- ernorate, which are contested by the STC but main- nical staff predate the conflict, more senior positions tain their own complex political arrangements;154 and have become increasingly “Houthified” over time, (vi) the Red Sea port city of Mokha and surround- creating an intricate network of formal and informal ing areas, which are held by Tareq Saleh’s National Resistance Forces. The Presidential Leadership 153 Based on maps from the Sanaa Center for Strategic Studies, Council formed in April 2022 ostensibly unites the Yemen Polling Center, International Crisis Group, European Council on Foreign Relations, Washington Institute for Near different factions in non-Houthi areas into a single East Policy. governing body. But recent literature on the subject, 154 For example, southern Hadramawt is controlled by political- infighting over Shebwa in July 2022, and key infor- military factions aligned with the Hadramawt Elite Forces, which mant interviews suggest that the divisions described have sought to remain relatively neutral in disputes between above remain in effect (see Map 9). the IRG and the Southern Transitional Council. 55 56 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Map 9 Zones of Control, January 2022 Source: International Crisis Group, May 2022). power relationships.155 These institutions’ author- have filled some of these gaps.157 But aid flows have ity is not recognized in non-Houthi areas, where a fallen, in part due to fears over Houthi aid manipu- patchwork of national and local authorities hold sway. lation and diversion, and challenging relations with The Houthis are perceived to have imposed some international organizations. relatively uniform rules in their areas, establish- The IRG has not fully rebuilt central state ing through intimidation and force what appears institutions in a uniform manner.158 Ministries to be a hybrid police state that incorporates both pre-war public institutions and a system of 155 Bank staff, consultant key informant interviews, Jan. 2022-Oc- tober 2022. See also Joshua Rogers, Changing Local Gover- Houthi-centric supervision.156 While official data nance in Yemen: District and governorate institutions in the are not available, anecdotal evidence describes the areas under Ansar Allah’s control, Berghof Foundation Working Houthis as a highly efficient collector of taxes and Paper, December 2020. custom duties. However, the Houthis lacks access 156 Truce Test: The Houthis and Yemen’s War of Narratives, to oil and gas, the principal source of Yemen’s pre- International Crisis Group, 29 April 2022. war fiscal revenue. Civil-service salaries are paid 157 Bank staff, consultant key informant interviews, Jan. 2022- Oct. 2022. See also: Entrenched Power: The Houthi System of sporadically at best, and public investment is neg- Governance, Sanaa Center for Strategic Studies, June 2022. ligible. The Houthis have largely neglected public 158 Bank staff, consultant key informant interviews, Jan. 2022-Oct. service delivery and many of the basic functions of 2022. See also: Letter dated 22 January 2021 from the Panel of the state, except security. UN agencies and a small Experts on Yemen addressed to the President of the Security number of foreign nongovernmental organizations Council, 22 January 2021. Yemen’s Transformed Political Economy 57 struggled to link branches throughout the country, Figure 54 Population-Adjusted and key informants reported that the government Nighttime Illumination in was unable to establish full authority over local IRG and Houthi Areas branches and personnel.159 When governance and Each January, 2012 to 2023 security are relatively strong in IRG areas, it is usu- ally because local authorities, in coordination with 4,000 160 local power centers including tribes, have taken 3,500 140 control. Governance is best where local resources 120 Brightness (Millions) 3,000 are available to fund expenditures, or external 100 actors have chosen to underwrite the cost of local 2,500 80 security and governance. Even when ministries have 2,000 60 managed to retain authority over local branches, 40 they often lack the resources to finance projects 1,500 20 or even to cover the cost of essential procurement 1,000 and personnel.160 When the government has been 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 able to access funds, questions have been raised about the efficacy and transparency of their distribu- Houthi (right) Non-Houthi tion.161 For example, the mechanism for collecting Source: World Bank. and distributing oil-export revenues is opaque, with government officials claiming that these revenues primarily finance operating costs, fuel imports for far below pre-conflict levels. All areas of Yemen electricity generation, and a recent series of foreign- experienced a sharp reduction in nighttime illumi- exchange auctions. nation at the onset of the conflict; however, while Key informants dispute the popular claim both curves have slowly recovered, non-Houthi areas that Houthi areas are better governed than have recovered faster than their Houthi counterparts. other parts of the country, citing the authoritari- Satellite images show that the area around Aden anism of the Houthis’ rule, as well as its unwill- was among the fastest to recover between 2015 and ingness or inability to provide public goods. 2018. Information on food security and employment Satellite data on the economic recovery, cited outcomes also indicate better living conditions in IRG below, further undermine such claims; and support areas than in Houthi areas (see Figure 54). the idea that while the Houthis’ governance is harsh Nighttime illumination data at the subnational and therefore highly visible, it is unable or unwilling level indicate that the Houthis have neglected the to deliver public goods. less politically or strategically important areas Likely due to a combination of differences under its jurisdiction, likely worsening poverty in governance capabilities and uneven access to hydrocarbon resources, nighttime illumina- 159 Bank staff, consultant key informant interviews, Jan. 2022-Oct. tion data indicate that the economic recovery in 2022. See also: Letter dated 22 January 2021 from the Panel of non-Houthi areas has outpaced improvements in Experts on Yemen addressed to the President of the Security Council, 22 January 2021. Houthi territory. Subnational GDP estimates based 160 IMF Staff Concludes Visit on Yemen, IMF, 3 June 2021; World on nighttime illumination indicate that economic Bank key informant interviews, Jan. 2022-Oct. 2022. activity has recovered faster in non-Houthi areas. 161 Beyond the Business as Usual Approach: Combatting While light emissions in some regions rebounded Corruption in Yemen, Sanaa Center for Strategic Studies, relatively quickly, illumination in other areas remains November 2018. 58 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES and food security among marginalized commu- with the Houthis and frontline fighting has been nities. Nighttime illumination has declined in four most intense. In Taiz, local conflict and the severing key Houthi-controlled areas: Al Raymah, Al Mawhit, of road access have constrained livelihoods oppor- Hajja, and Saada governorates The first two are the tunities, especially among residents of Taiz city. least populous governorates under full Houthi ter- With travel into Taiz and between Houthi and non- ritory. The latter two are an important political base Houthi areas of Taiz governorate restricted amid for the Houthis, but they are also situated along the the continued threat of violence, access to essential border with Saudi Arabia, which has been an active services has been sharply curtailed, and the price of front throughout the conflict. goods in Taiz city has increased significantly. While IRG areas include several notable growth the Houthis extract considerable tax revenue from centers. Governance and development outcomes Taiz due to their control of the industrial center of in IRG areas have been extremely heterogeneous. In Hawban, the conflict has severely undermined the Marib, for example, local authorities have overseen local economy. Nevertheless, nighttime illumina- the rapid development of infrastructure, healthcare, tion data show that Taiz city has staged a gradual and education, while managing to provide robust economic recovery, though it continues to lag other local security, by drawing on revenues from nearby governorates in Houthi and non-Houthi areas (see oil and gas fields and refineries. In late 2015, Marib Figures 56–58). decided to withhold locally generated revenues Elsewhere, the picture is more mixed. In from the CBY, which was headquartered in Sanaa southern Yemen,166 the STC has focused largely on at the time. These revenues were primarily from securing its areas and has struggled to establish crude oil and liquified petroleum gas produced at robust governance or deliver essential services. As Marib’s Block 18, which was managed by the state- a demonstration of trust and good faith toward the run Safer Exploration & Production Operations STC and the population under its control, the IRG Company (SEPOC), commonly known as SAFER. and the PLC have worked to reestablish the elec- Marib authorities did not start sharing locally gen- tricity supply in Aden, which is the IRG’s temporary erated revenues with CBY-Aden until mid-2019 (see capital but has been under de facto STC control Map 10).162 The eastern governate of Al-Mahra, which 162 https://devchampions.org/uploads/publications/files​ borders Oman and Saudi Arabia, has become a /Rethinking_Yemens_Economy_No2_En.pdf. In mid-2017 Marib key transit point for maritime and cross-border followed in neighbouring Hadramawt’s footsteps and secured a trade. The volume of maritime traffic and trade negotiated, individual agreement with IRG defining that 20 per- flows across Al-Mahra’s land border with Oman has cent of the export revenues generated from crude oil produced at Marib’s Block 18 (and then exported via the Rudum/Nushayma increased significantly during the war.163 Rising trade export terminal in Shabwa) would go to Marib’s local authorities. has boosted locally collected import taxes and cus- 163 Source: Sanaa Center for Strategic Studies, Ye- toms revenue,164 and Al-Mahra’s authorities halted men Arabian Sea Ports Corporation, ACAPs. https:// revenue transfers to CBY-Aden for an extended sanaac ​ e nter.org/publications/analysis/7606; https:// period between 2018 and 2020.165 Nighttime illumi- ​ c.co/en/statistics/port-of-nashtoon/; https://yemen.ye​ www.yasp nation data suggest that Al-Mahrah is a rare exam- ti.acaps.org/imports-exports/; shorturl.at/fsLN0. 164 See al-Sewari, Yahya, Yemen’s Al-Mahra: From Isolation to ple of a governorate in which economic growth has the Eye of a Geopolitical Storm. significantly surpassed pre-war levels since 2015 165 Ibid. Bank staff, consultant key informant interviews, Jan. (see Figure 55). 2022-Oct. 2022. Economic losses in IRG territory have been 166 Bank staff, consultant key informant interviews, Jan. 2022- concentrated in areas where control overlaps Oct. 2022. Yemen’s Transformed Political Economy 59 Map 10 Population-Adjusted Nighttime Illumination in Western Yemen Each January, 2015 to 2018 Note: Units are nanoWatts/cm2/sr per capita. Source: World Bank, TBC. since 2019. Along with other groups such as the yet to recover. Shabwa governorate, which was also National Resistance Forces on the Red Sea coast contested by STC and IRG forces in 2019, subse- and the Salafist-oriented Giants Brigades, the STC quently experienced significant growth, likely due to appears to rely on external support both for its mili- an IRG revenue-sharing scheme with local authori- tary operations and civilian governance functions.167 ties. Economic output in al-Dhale, the site of one Nighttime illumination data show significant of the conflict’s most active fronts, appears to remain variations in the growth patterns of other south- below its 2014 level. Hadramawt governorate, which has ern governorates. Abyan governorate, which was not been directly affected by the national conflict but contested by STC- and IRG-affiliated forces from 2019 through 2022, suffered apparent economic 167 Bank staff, consultant key informant interviews, Jan. 2022- contractions in 2015 and 2019, from which it has Oct. 2022. 60 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 55 Nighttime-Illumination-Based GDP Estimates for Hajja, Saada, Al Raymah and Mahwit Versus the National Baseline, 2012-2022 Al Mahawit Al Maharah 600 500 500 Index (april 2012 = 100) Index (april 2012 = 100) 400 400 300 300 200 200 100 100 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Amaran Ad Dali’ 600 400 500 Index (april 2012 = 100) Index (april 2012 = 100) 300 400 300 200 200 100 100 0 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index National Baseline 120 Index (april 2012 = 100) 100 80 60 40 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Date Total NTLs Total GDP Source: World Bank Yemen Mobile Phone Survey Monitoring Round I. Yemen’s Transformed Political Economy 61 Figure 56 Nighttime Illumination and GDP Estimates, Marib Versus National Baseline Ma’rib National Baseline 600 140 Index (april 2012 = 100) Index (april 2012 = 100) 500 120 400 100 300 80 200 60 100 40 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Total NTLs index Total GDP index Source: World Bank, TBC. Figure 57 Nighttime Illumination and GDP Estimates, Al-Marah Versus National Baseline Al-Marah National Baseline 140 500 Index (april 2012 = 100) Index (april 2012 = 100) 120 400 100 300 80 200 60 100 40 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Total NTLs index Total GDP index Source: World Bank, TBC. suffered a series of local battles between 2015 and 2016, Old Problems in New Forms: Rent-Seeking, appears to have experienced a modest contraction and Patronage, and Corruption then quickly recovered to pre-war output levels, likely due to its combination of a busy port and oil and gas In pre-war Yemen, political stability was predicated on resources. The island of Socotra, meanwhile, has seen rent sharing among elites, which limited the incentives its economic output nearly double over the course of for open conflict.168 Rents included access to economic the conflict, likely due to significant investment from the Saudi-led coalition and a growing military presence on 168 See Yemen, Unlocking the Potential for Economic Growth, the island (see Figures 59–61). World Bank, October 2014. 62 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 58 Nighttime Illumination and GDP Estimates, Taiz and Comparators Taiz Ibb 175 150 250 Index (april 2012 = 100) Index (april 2012 = 100) 125 200 100 150 75 100 50 25 50 0 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index NTL index GDP index Lahj National Baseline 140 200 Index (april 2012 = 100) Index (april 2012 = 100) 120 150 100 100 80 50 60 40 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Total NTLs index Total GDP index Source: World Bank, TBC. opportunities via the granting of import licenses and as the country’s dwindling hydrocarbon resources. other permits, the allocation of political positions to Lower oil revenues coupled with the fragmentation family members, and the diversion of resources to key of territory and institutions have driven a shift in the leaders through the corruption of the military pay- ways rival parties generate income, seek rents, and roll.169 Most patronage rested, directly or indirectly, distribute patronage. Oil rents can be extracted on oil rents. The system privileged military and tribal relatively easily, though exports can only be sold leaders who had the capacity to directly threaten the legally by the IRG. By contrast, tax rents primarily authority of the state, and it created strong incen- reflect the size of the economy under each group’s tives for fuel theft, smuggling, bribery, extortion, and control. Most locally generated income likely goes abuse of the public investment budget. to sustaining high-level political and tribal alli- While the conflict has reshaped Yemen’s ances and to prosecuting the war. Key informants patronage systems around different actors and report that the clearest evidence of the existence power centers, corruption remains prevalent. Since 2016, rival factions have increasingly com- 169 Ibid. See also: Yemen Corruption Assessment, United States peted to control financial and trade flows, as well Agency for International Development, September 205 2006. Yemen’s Transformed Political Economy 63 Figure 59 Nighttime Illumination and GDP Estimates, Aden Versus National Baseline Aden National Baseline 140 140 Index (april 2012 = 100) Index (april 2012 = 100) 120 120 100 100 80 80 60 60 40 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Total NTLs index Total GDP index Source: World Bank, TBC. Figure 60 Nighttime Illumination and GDP Estimates, Various Governorates Versus the National Baseline Abyan Ad Dali 600 500 500 Index (april 2012 = 100) Index (april 2012 = 100) 400 400 300 300 200 200 100 100 0 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Hadramawt Shabwah 120 500 Index (april 2012 = 100) Index (april 2012 = 100) 100 400 300 80 200 60 100 40 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Socotra Source : World Bank, TBC. National Baseline (continued on next page) 500 140 2 = 100) 2 = 100) 120 400 120 500 Index (april 2012 = 100) Index (april 2012 = 100) 100 400 300 80 64 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES 200 60 100 40 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Figure 60 Nighttime Illumination and GDP Estimates, Various Governorates Versus the National Baseline (continued) NTL index GDP index Socotra National Baseline 500 140 Index (april 2012 = 100) Index (april 2012 = 100) 120 400 100 300 80 200 60 100 40 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 NTL index GDP index Total NTLs index Total GDP index Source: World Bank, TBC. of patronage networks is the stability of Houthi oil production. Gas flaring strongly predicts oil rule in tribal areas that historically resisted both production, and nighttime illumination captures the IRG and the pre-war government.170 85  percent of the variation in output. Oil pro- GDP data alone are insufficient to explain duction estimates based on nighttime illumina- the relative economic strengths and weaknesses tion consistently exceed those provided by the of different groups. Based on nighttime illumi- United States Energy Information Administration nation and phone surveys, the IRG areas appears (EIA). These findings are corroborated by large to enjoy higher economic growth rates. However, differences in fuel imports and exports between the economy in Houthi-controlled territory is more Yemen and its trading partners. Although modest diversified and opportunities broader than in IRG disparities are common when using mirror data, areas. In Houthi areas, taxation of goods, markets in this case the magnitude appears to be in the and businesses has become the principal source hundreds of millions of US  dollars over several of rents, which are reportedly redistributed to local years, which cannot reasonably be explained by authorities and leaders, including in the military statistical errors or methodological differences. sector.171 The strong share of economic output in While the observed discrepancies may be due IRG calculated via nighttime illumination is at least in part to poor local statistical capacity, they may in part explained by its control over key oil and gas also indicate corruption in resource management fields, from which it does not extract the maximum and rent sharing. value, as it is forced to share at least one-fifth of its income with local groups and because it does 170 Bank staff, consultant key informant interviews, Jan. 2022- not commercialize associated gas.172 Allegations Oct. 2022. of oil and fuel theft are pervasive, meanwhile, as 171 Bank staff, consultant key informant interviews, Jan. 2022- Oct. 2022. See also: Brokering a Ceasefire in Yemen’s Economic are claims of fuel theft from subsidized electricity War, the International Crisis Group, January 2022; Yemen: Al production in IRG areas (see Table 7). Hodeidah fuel import and fuel price modelling, 10 April 2022. An analysis of gas flaring reveals signifi- 172 Bank staff, consultant key informant interviews, Jan. 2022- cant discrepancies in the official reporting on Oct. 2022. Yemen’s Transformed Political Economy 65 Table 7 Variations in Production Explained by Nighttime Illumination Dependent Variable: (1) log(RGDP) (2) RGDP (3) Oil Prod. (1) Index (2) Index (3) Mb/d Independent. Variable (1) (2) (3) log(NTLs) 0.349*** (0.002) NTLs 0.0357** (0.004) Flaring NTLs 0.648*** (0.025) Observations 23 23 113 R 2 0.799 0.788 0.856 Note: *p<0.1; **p<0.05; ***p<0.01. According to official data, the IRG continues The Economic War and Its Power to spend about one-quarter of its budget on Dynamics electricity subsidies, the same share as before the conflict.173 Interviews with government officials The Emergence of New Political Economy indicate that payments for fuel imports to produce Actors electricity are a major drain on foreign-exchange Yemen’s war extends beyond the battlefield to reserves, but key informants claim electricity supply the economic sphere, further immiserating a is no better than it was before the subsidies were population already traumatized by conflict. All put in place. While a thorough investigation of factions use a mix of coercive economic tools to subsidy spending is beyond the scope of this CEM, improve their economic positions and weaken their the international experience shows that electricity rivals. Yemen has been divided into two competing subsidies tend to benefit better-off households, zones of monetary control with separate regulatory distort prices and encourage inefficiency, and institutions. Banking services, foreign exchange, enable corruption. Indeed, previous subsidy-reform seaports, oil fields, and road access across domestic efforts in Yemen were presented as anti-corruption and internal borders are all objects of contention initiatives.174 between rival factions. Imports can also present opportunities for An unstable regulatory environment, rising patronage and rent-seeking. The Houthis have con- remittances, and skyrocketing demand for for- sistently denied receiving fuel from foreign countries eign exchange have given rise to a new cadre as a form of financial support. However, a 2019 UN of unregulated but systemically important expert panel reported that a small number of compa- nies operate as fronts for fuel imports designed to ben- 173 Official government budgets for 2021 and 2022 presented to efit the Houthis.175 In addition, the Houthis earn much the World Bank and IMF show spending on subsidies represent- of their income taxing imports into the areas it con- ing 22 per cent and 32 per cent of all expenditures. 174 Petroleum Subsidies in Yemen: Leveraging Reform for De- trols, which include the country’s largest markets and velopment, Clemens Breisinger, Wilfried Engelke, Olivier Ecker, population centers. The IRG appears to have sought World Bank Group, February 2011. to reroute trade and financial flows bound for Houthi 175 Yemen’s Economic Outlook- Fall 2016, World Bank, October areas to IRG-controlled ports and financial institutions. 2016. 66 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 61 Correlation between Nighttime Illumination and Oil Output 1e6 300 260 6 275 240 220 5 250 GDP index 225 200 4 200 180 175 160 3 150 140 125 120 2 100 100 2 3 4 5 6 1990 1995 2000 2005 2010 2015 2020 No flares GDP index No flares (right) 1e6 1e6 9 300 260 275 240 8 250 220 7 GDP index 225 200 6 200 180 5 175 160 4 150 140 125 120 3 100 100 2 2 3 4 5 6 7 8 1990 1995 2000 2005 2010 2015 2020 Total GDP index Total (right) 1e6 Total petroleum and other liquids (Mb/d) 5.5 25,000 5.0 200 4.5 175 20,000 4.0 150 125 15,000 3.5 100 3.0 10,000 75 2,5 50 5,000 2.0 25 1.5 10.0 10.5 11.0 11.5 12.0 2012 2013 2014 2015 2016 2017 2018 2019 2020 Flares Total petroleum and other liquids (Mb/d) Flares (right) Source: WB computations based on remote-sensing, gas-flaring analysis. money-transfer agents.176 The huge increase in exchange rates across regions, as well as the need the number and prominence of money exchangers for intermediaries to execute financial transactions since 2015, coupled with the bifurcation of monetary between IRG and Houthi areas.178 The large public policy, has further compromised the capacity of the deposits of local and foreign currency that these central banks in Aden and Sanaa to oversee or exert regulatory control over these new entities.177 Some 176 Ibid. money exchangers have almost certainly benefit- 177 Ibid. ted from the volatility of the rial and the differential 178 Ibid. Yemen’s Transformed Political Economy 67 money exchangers can hold enables them to capital- Road blockades and other restrictions on ize on local exchange-rate dynamics,179 and though the overland movement of goods are part of the their operations are often technically illegal, the CBY- same zero-sum economic competition. Internal Aden has limited capacity to identify and respond to barriers to movement limiting the number of road instances of currency arbitrage and speculation.180 routes available to transportation firms and per- vasive deployment of formal and informal check- The Fight for Control of Trade Corridors points by tribes, armed groups, security and military forces among others, have created opportunities Since 2015, most major battles have been for armed groups and criminal gangs to demand fought over the control of ports, oil and gas rents from truckers.184 Territorial fragmentation has production facilities, and key trade routes. The presented arbitrage opportunities for traders at the IRG and the Saudi-led coalition initially attempted expense of consumers, enabling powerful actors a naval blockade of the Houthi-controlled port of to exploit price differentials created by economic Hodeida, though they were ultimately persuaded truncations and the bifurcation of monetary policy. to allow shipments to enter the port upon inspec- All parties to the conflict have contributed tion and approval by the IRG and coalition.181 Since to the deterioration of governance and the then, different restrictions and embargoes have fragmentation of the national economy, but the been implemented on Hodeida and neighboring Houthis’ policies have been especially distor- Houthi-controlled ports. In 2018, coalition forces tive. While the Houthis have blamed price spikes launched a campaign to seize control of Hodeida on embargo-driven shortages, it has also shuttered outright, which was only halted when the UN bro- many of the publicly owned fuel stations under its kered a deal to demilitarize the port and its sur- control while directing fuel supplies to privately roundings. During 2018 and 2019, the STC took owned gas stations to be sold at higher rates. The control of port facilities in Aden in south Yemen. Houthis have also imposed quotas on the amount of The fuel embargo limited the Houthis’ access fuel consumers could purchase at the official price to revenue while increasing IRG income at ports within a one-week period. The Houthi-run YPC coor- under its control. Fuel imports to Hodeida fell, while dinated the supply of fuel to traders willing to pay inflows to IRG ports increased. Fuel imported to premiums to secure larger quantities of fuel to cover IRG areas was trucked overland into Houthi areas,182 which disrupted supply chains, increased pump 179 Yemen: Yemeni Financial Sector in December 2021, ACAPS, prices in Houthi areas, and led to double taxation.183 January 2022. 180 Bank staff, consultant key informant interviews, Jan. 2022-Oct. Military struggles over oilfields reflect the 2022. See also: Economic Warfare: Government and Houthi crucial economic importance of hydrocarbon Struggle for Control Deepens, 3 January 2020; Brokering a exports. Houthi forces attempted to seize oil-rich Ceasefire in Yemen’s Economic Conflict, the International Crisis Marib and Shebwa in 2015 but were rebuffed. In Group, January 2022. 2020, the Houthis again launched a major offensive 181 In Hindsight: The Story of the UN Verification and Inspection in Marib, aimed at seizing its oil and gas produc- Mechanism in Yemen, Security Council Report, September 2016. tion facilities. STC forces attempted to gain control 182 Al Hodeida fuel import and price modelling, ACAPS, 4 April 2022. of oil-rich Shebwa in 2019, and in 2022 other forces 183 Bank staff, consultant key informant interviews, Jan. 2022-Oct. perceived to be STC-aligned took Shebwa. The 2022. See also: Al Hodeida fuel import and price modelling, STC then publicized its intention to take control of ACAPS, 4 April 2022. neighboring Hadramawt, another major hydrocarbon 184 Bank staff, consultant key informant interviews, Jan. 2022- production hub. Oct. 2022. 68 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES their operational needs. Estimates calculated by the susceptible to diversion, disruption, delays, and International Crisis Group suggest that the Houthis profiteering from multiple actors. Key informants earned more from fuel sales in 2020 and 2021 than report that LPG transport and distribution compa- in previous years despite apparent shortages. nies, truck drivers, armed groups, and local retailers who operate independently of the YGC have all Sectoral Governance during the Conflict: benefitted from the breakdown of governance in The Case of Liquefied Petroleum Gas the LPG subsector. The rising cost of imported fuel, intermittent Dynamics in Yemen’s domestic LPG market import-supply disruptions, and falling prices for highlight the enormous burden the economic domestically produced LPG have boosted LPG conflict imposes on ordinary households. While demand. The use of LPG for personal, commercial, Yemen imports most of its energy supply, the bulk and industrial purposes has increased significantly of the LPG sold on the local market is produced during the conflict, and the resulting competition domestically. The state-run, Marib-based Safer Oil has opened new avenues for rent-seeking. LPG Exploration and Production Company (SAFER) pro- supplies earmarked for households have been sys- duces an estimated 23,000 barrels per day of LPG tematically diverted and sold to energy-intensive at a small onsite refinery at an oilfield in Marib.185 businesses,189 which are willing to pay premiums Before the war, LPG distribution was over- to secure the quantities they need. Cross-regional seen by the state-run Yemen Gas Company (YGC), arbitrage offers attractive profit margins, and LPG which set prices and sales quotas. LPG was allo- has been regularly diverted to the domestic markets cated to governorate and district-level distribution that offer the highest prices, most of which are in centers based on the size of the local population, Houthi areas.190 industry, and demand. Consumers bought LPG The influence of armed groups over the sup- directly from YGC sale points or from privately ply and distribution of LPG has not been con- owned sale points certified by the YGC.186 As the fined to Houthi areas. Three of the five bottling YGC fragmented along with other state institutions, the regulation of the sector deteriorated, and even- 185 Bank staff, consultant key informant interviews, Jan. 2022-Oct. tually the YGC was no longer able to set prices or 2022. See also: Al Hodeida fuel import and price modelling, guarantee LPG availability across the country. The ACAPS, 4 April 2022. See also: Conflict Escalation in Marib number of private actors involved in the transporta- and Potential Humanitarian and Economic Impacts: Scenario, tion, distribution, and sale of LPG grew significantly, ACAPS, July 2021. and new market entrants competed with established 186 Based on World Bank background paper on Yemen’s gas LPG transport and distribution companies. In Houthi sector. See also: Yemen: Unlocking the Potential for Economic Growth. areas, the YGC has been further marginalized by the 187 Bank staff, consultant key informant interviews, Jan. 2022-Oct. deployment of Houthi supervisors (mushrafeen) and 2022. See also: Yemen: Impacts of Fuel Dynamics and Fuel Price neighborhood bosses (aqel al-harah) to oversee Structures, ACAPS. August 2021. local LPG distribution (see Table 8).187 188 Bank staff, consultant key informant interviews, Jan. 2022-Oct. While pre-conflict allocation quotas remain 2022. See also: Yemen: Impacts of Fuel Dynamics and Fuel Price in place, the distribution of benefits within the Structures, ACAPS. August 2021. LPG sector has changed drastically. An estimated 189 Bank staff, consultant key informant interviews, Jan. 2022-Oct. 2022. See also: Yemen: Impacts of Fuel Dynamics and Fuel Price 60–70 percent of daily LPG production is allocated Structures, ACAPS. August 2021. to Houthi areas in accordance with pre-conflict 190 https://sanaacenter.org/files/The_Gas_Crisis_in_Taiz​ quotas.188 However, LPG supply networks are now _War_Mismanagement_and_Exploitation_en.pdf. Yemen’s Transformed Political Economy 69 Table 8 Monthly Average Price for One LPG Cylinder in Sanaa and Aden, May 2021–April 2022 Ratio (domestic Date Sana’a (YER)a USD equivalent Exchange rate den (YER) USD equivalent Exchange rate currency based) 21-May 10,419 17.5 593.7 4500 5 898.6 2.3 21-Jun 11,700 19.6 595.6 4500 4.8 932.7 2.6 21-Jul 11,724 19.6 597.1 4500 4.6 980.7 2.6 21-Aug 12,355 20.6 598.4 4500 4.5 1001.5 2.7 21-Sep 11,333 18.9 599.9 4500 4.1 1096.3 2.5 21-Oct 11,774 19.6 600 4500 3.5 1270.1 2.6 21-Nov 11,917 19.8 600.2 11500 7.6 1517.7 1.0 21-Dec 12,371 20.6 600 14438 12.3 1175.5 0.9 22-Jan 14,597 24.3 600.5 7666.7 6.8 1114.5 1.9 22-Feb 15,179 25.3 600.4 14773.5 13.2 1117.2 1.0 22-Mar 18,016 29.9 601.6 15500 12.4 1247.5 1.2 22-Apr 14,167 25.4 557.4 16000 16.7 958.7 0.9 Sources: Authors research based on ACAPS and WFP figures, coupled with results from key informant interviews. Notes: ª The data for Sana’a reflects the average monthly parallel market price applied. The data for Aden reflects a combined official and parallel market price provided by WFP. Map 11 LPG Supply Routes in Sanaa, Taiz, and Aden 70 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES facilities that serve Taiz governorate and are out- crisis and economic deterioration, there are rea- side of Houthi control have repeatedly suspended sons for qualified optimism. One of Yemen’s most activity due to alleged security risks.191 Frequent promising assets is its class of entrepreneurs, who disruptions have been reported at the Al Farshah have survived decades of political and economic bottling facility in Lahj, which is a vital link in the turbulence. Yemen is surrounded by wealthy mar- LPG supply chain that runs from Marib to Taiz via kets that share cultural ties and import most of their Shabwa, Abyan, Aden, and Lahj governorates.192 consumption goods, content for which could be The supply of LPG to areas outside Houthi control produced in Yemen.194 Given peace, stability, and has been disrupted due to: (i) the diversion of LPG adequate investment, the combination of a resilient to Houthi-controlled parts of Taiz and possibly to private sector and a range of viable export markets other Houthi areas; (ii) the diversion of LPG ear- could enable a broad-based recovery led by agro- marked for households to businesses; (iii) security processing and light manufacturing. threats at bottling facilities serving non-Houthi On balance, the immense damage wrought areas; (iv) the intervention of local military, security, by the conflict, the collapse of the patrimonial and political officials at various points along the state after 2014 opened up new spaces for good supply chain; and (v) the challenging access routes governance and entrepreneurship. Formal and that LPG truck drivers must take via southeastern informal local institutions have demonstrated surpris- Taiz to reach non-Houthi areas (see Map 11). ing capabilities in some parts of the country, often by IRG efforts to consolidate control over local marrying tribal and other customary institutions with non-LPG fuel markets as part of their post-Hode- local government. The private sector has been able ida strategy have also reportedly disrupted fuel to operate in a highly challenging context, ensuring supply in both nominal IRG and Houthi areas. that key basic goods are available in local markets After requiring importers to sell a large share of and often replacing the state in delivering services. their fuel to the IRG in 2021, the YPC attempted Consolidating and expanding these pockets of to coordinate the onward distribution and sale of strength, resilience, and innovation could help build LPG. These efforts were met with strong resistance the foundation for a successful post-conflict state. by some fuel traders, including the country’s single Post-conflict Yemen can draw on lessons from other largest fuel importer.193 Multiple importers refused fragile situations, as well as historical growth models to dock their vessels at Aden Oil Harbour, disrupting based on community organization coupled with fund- the flow of LPG to Aden and from Aden to Houthi ing from the diaspora.195 areas. However, beginning in mid-January 2022 Yemen has effectively undergone an unman- these constraints eased, as fuel shipments entered aged decentralization process that is unlikely to and unloaded at Aden Oil Harbour. be reversed even after the conflict has ended. Decentralization and Entrepreneurship: 191 Ibid. Ingredients for Optimism? 192 https://www.alayyam.info/news/8WS1GBY5-OA64F0-AAA7. A Comparative View 193 Bank consultant key informant interviews in January 2022. 194 Chatham House (2011). 195 An interesting example is that of the Local Development Yemen’s immediate future remains uncertain, Associations (see Lackner, 2019, and Cohen, Hébert, Lewis and and the path to robust and inclusive growth is Swanson, 1981). These were the ways that local communities extremely narrow. Despite the complexity of the raised resources and provided public goods. Though they at- conflict, the ongoing struggle for control of the econ- rophied in the 1990s, in the years of the patrimonial state, they omy, and the sheer scale of Yemen’s humanitarian seem to have grown dramatically since 2014. Yemen’s Transformed Political Economy 71 Prior to 2014, a highly centralized state dominated recovers and moves toward convergence with peripheral regions.196 The rule of law was applied neighboring states, or remains locked in a cycle unevenly and often in the service of national elites, of low growth and chronic instability. The gov- but in many parts of the country customary law and ernment and other stakeholders can draw on the local institutions provided a more robust framework experience of other places that have successfully of legal and quasi-legal protections for citizens.197 emerged from a period of profound crisis. Four But in recent years the periphery has taken advan- such countries demonstrate that Yemen’s power- tage of the collapse of central control. Power has ful and heterogeneous social structures should coalesced around the 21 governorates, especially not be viewed as an overwhelming obstacle to the outside of Houthi-controlled areas. creation of an effective non-patrimonial state. Such The post-war equilibrium appears likely to structures have been made compatible with states shift power to political authorities that offer of varying degrees of effectiveness. greater accountability and more abundant public The first, best-case outcome is to be found in goods. In Marib, for example, key informant inter- Abu Dhabi, where tribal ties underpinned rather views indicate that the especially active role of local than impeded the process of state-building, as leaders in the public administration has fostered Eurocentric paradigms would tend to assume. more effective governance. Key informants argue The second, still optimistic, is the consolidation that the state can be reconstructed at the subna- of peace in Sierra Leone since the end of the civil tional level because regional and local authorities war in 2001. While Sierra Leone remains deeply are more “connected to the community” and gov- underdeveloped, it has successfully avoided a return ernors “are from the community, so people have to conflict for over two decades by leveraging the ways to influence them.”198 These views are con- strength and legitimacy of its traditional institutions. sistent with the outcomes of the National Dialogue The third case study is the example is Somaliland, Conference of 2013–2014, which offers important Yemen’s near neighbor, which embedded tradi- insights into Yemeni views on governance and tional governance systems into the foundation of accountability. a modern, accountable state that has created the Numerous obstacles stand in the way of a conditions for economic growth—despite lacking more prosperous post-conflict Yemen. The first international recognition. challenge will be to strike an appropriate balance The fourth country is Colombia which repre- of power between the national and subnational sents the least optimistic but perhaps the most governments, including the division of revenues relevant case for Yemen. In Colombia, growth and from natural resources, customs, and tax collec- increasing economic sophistication in some parts of tion. Another challenge will be to counter the the country including its urban centers, have been pervasive rent-seeking behavior and extensive achieved in tandem with high levels of violence patronage networks that would inevitably hinder and illicit trade in the often ungoverned interior. reform in a post-conflict setting. Ultimately, the Colombia offers a plausible if unsatisfying glimpse fundamental prerequisite for stability and renewed at Yemen’s potential future, where cities and some growth is peace. An unbalanced division of power regions thrive even as other areas remain chroni- or unchecked rent-seeking could undermine even cally disordered. the most well-intentioned settlement. The end of the conflict will be a critical 196 Consultant key informant interview, January 2022. juncture for Yemen. The country’s political and 197 See for example, al-Dawsari (2018). security situation will determine whether it fully 198 Consultant key informant interview, January 2022. 72 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Yemen differs in important ways from each opportunities among the different Bani Yas tribes, of these countries and territories, and the case effectively mapping the traditional social structure studies provide broad strategic guidance on onto a modern state. While this approach would consolidating peace and stability in fragile states appear to formalize patronage, in practice it can rather than a copy-and-paste model for political help diffuse political tensions by ensuring that tribal organization and growth in Yemen. Abu Dhabi, and political elites outside the ruling coalition retain Somaliland, and Sierra Leone are territorially smaller access to a meaningful degree of independent politi- and less populous than Yemen. Their societies are cal and economic power.200 relatively heterogeneous compared to Yemen’s, and Many states in the developing world, Abu each country’s formal political systems and traditional Dhabi included, are more like what Southall institutions vary enormously, as to their economic (1956) first called “Segmentary”—a fusion of kin- bases. While these caveats are important to bear in ship and fictive kinship groups with more modern mind, several key lessons emerge from the cross- state institutions and functions. To cite another country experience. The most important of these is relevant example outside of the Middle East, the that classic state-centric approaches to economic effectiveness of the Rwandan state today rests not on and political development are not the only way to the existence of a “Weberian bureaucracy”, but on achieve growth and development, and that local the fact that the government can leverage traditional customary governance and informal institutions can mechanisms of reciprocity deeply embedded in the be an asset in a post-conflict environment, not a bar- construction of the pre-colonial state (Heldring and rier to progress. Robinson, 2022). The Segmentary State and Tribal Traditional Institutions and Peace in Capitalism in Abu Dhabi Sierra Leone Abu Dhabi offers an example of centralized Sierra Leone shows how decentralization can governance that maps on to traditional social be foundational to stability. Sierra Leone is not a orders. It demonstrates that powerful social struc- “development miracle” but it could be described tures are not necessarily an obstacle to significant as a “stability miracle”. Since the end of its civil war development and growth. The political evolution of in 2002, Sierra Leone has not only completely re- Abu Dhabi shows that traditional social structures imagined national institutions from the bottom up can be compatible with modern governance. Abu but the strength of the traditional institutions at the Dhabi emerged as a separate polity in the late 18th local level has addressed fragility, with impressive century, when the Al-Nahyan family became leaders of the Bani Yas, a confederation of about 30 tribes,199 199 Davidson, 2018. but the country’s remarkable political transition did 200 “[P]ositions are … distributed to representatives of those not begin until after 1966, when Zayed bin Sultan other Bani Yas sections and tribes that have been historically Al-Nahyan replaced his brother as leader. supportive of the Al-Nahyan…. Many powerful government As in the other case studies, balance between and military portfolios remain in the hands of the non-ruling tribal elites. The advent of the new multi-sector economy and local social constituencies and the emerging state all of its constituent new parastatals and joint ventures, means was the key to Abu Dhabi’s development. While that there is a superfluity of fresh opportunities for this elite’s Abu Dhabi was highly centralized, and relied heav- inclusion in the establishment. Consequently, centuries-old ily on rising income from oil exports, Al-Nahyan alliances are being incorporated into a present-day system of divided control of ministries and allocated economic tribal capitalism” (Davidson, 2018, p. 111). Yemen’s Transformed Political Economy 73 results. The current homicide rate is lower than that international assistance is absent. In Somaliland, of the U.S. The legitimate political order at the local clans and the state arrived at a broad equilib- level allowed demobilized combatants to be reinte- rium of power over the course of years of formal grated into society. and informal negotiations. Social contracts have Post-colonial, pre-civil war Sierra Leone underpinned Somaliland’s stability and develop- suffered from many of the same pathologies ment over the past three decades. Somali society that affected Yemen prior to 2015, including is accustomed to negotiation and agreement, even an autocratic, corrupt, unaccountable, and during active conflict. Amid recurrent conflict in the highly patrimonial state with little administra- 1980s and 1990s, clan elites tapped into this tradi- tive capacity.201 In the 1890s, the British created tion and organized a series of conferences, which a system of paramount chiefs based on pre- often lasted for months, to try to build consensus colonial tribal institutions, in which social norms on how to govern. Over time, they developed an and customary practices provided a degree of innovative new clan-based political institution that accountability. The British appear to have coopted formed the basis of what now is the upper house traditional political systems because they retained of the Somaliland legislature. popular support and legitimacy,202 and these sys- While Somaliland has never had a strong cen- tems remained largely intact after independence. tral government, the cultural tradition of voluntary Indeed, the legitimacy of traditional systems was social agreements (heer204) and the fundamental evident at the end of the civil war in 2002. Despite egalitarianism of Somali society have provided intense pressure from outsiders to radically change the basis for state formation. After independence, local political institutions after the conflict, the new political elites were able to counterbalance the government re-constituted the system in much the power of clans and gradually shift authority towards same form as before the conflict. the state. In 1993, Somaliland’s first president used Past conflict is a strong predictor of future loans from foreign businesses to finance the demo- conflict, but Sierra Leone has remained at peace bilization of armed militias and leveraged customs for over two decades. The legitimacy of the politi- revenue to strengthen his political position. The cal order at the local level allowed demobilized government slowly but systematically consolidated combatants to be reintegrated into society. The its position until it had reached an equilibrium with international community has also played an impor- clan leaders and other subnational power centers. In tant role, and the World Bank-sponsored post-civil 2000, the government asserted its unique authority war decentralization plan enabled power-sharing to levy taxes by banning local revenue-generating between the two major political parties.203 However, the strength of traditional institutions at the local 201 Reno, 1995, and Acemoglu and Robinson, 2012. level has likely been the decisive factor in prevent- 202 Little, 1965 and 1966. ing a renewed conflict. 203 Zhou, 2009. 204 “[Heer’s] closest equivalents in English are compact, contract, Tribes, Accountability, and Growth agreement, or treaty in a bilateral sense. Thus several men or in Somaliland parties are said to be of the same heer when their relations are regulated by an agreement ... As need arises the terms of contracts are abrogated, existing treaties modified or rescinded, Somaliland shows how a mix of state-building, and new agreements made. The majority of heer agreements decentralization, customary rule and gradual binding groups relate principally to collective defense and institutional assimilation can offer a long- security and to political cohesion in general.” Lewis (1961) pp. term path to growth and stability, even when 161–162. 74 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES checkpoints, and centrally appointed governors opened a negotiating table. Nevertheless, the sub- disrupted de facto clan control. While the ruling sequent demobilization of the FARC has not led to a party favored centralization, the political opposition new Colombia. There have been many demobiliza- championed the interests of clans and other local tions and re-mobilizations in Colombian history. New institutions, and the ongoing discourse between armed groups have emerged and coca production, at them helped normalize the peaceful contestation of an all-time high, flourished in ungoverned rural areas, power through the political process.205 while politics remains clientelistic. Somaliland has achieved stability and eco- This model of a state, what Acemoglu and nomic development without significant interna- Robinson (2019) call a “Paper Leviathan” was tional aid. As Phillips (2020) notes, “Indeed, not compatible with economic growth, at least in only did Somaliland receive no ODA, it received some areas. Roldán (2002) shows in detail how (almost) no international charitable contributions, the core-periphery relationship that characterizes no support for state-building or peace-building, the entire country, emerged within the depart- no peacekeeping missions, no political support for ment of Antioquia. In the urban Medellín, in central belligerents in its civil war, no military assistance, Antioquia, business invested and grew. In the rural no access to the licit international weapons trade, periphery, anarchy and disorder ruled but in an no significant involvement in extractive indus- equilibrium with the core which did not undermine tries, minimal foreign investment, and no access its incentives. The economy grew at 5 per cent a to international loans. And it was unencumbered year on average even during the civil war. Since with international debt repayments. Other than then, despite repeated waves of conflict, growth the remittances Somalilanders were sending home has been steady and positive. from abroad, the country was remarkably free from Despite waves of violence and conflict, external involvement of any kind.” Colombia has managed to achieve significant economic sophistication and entrepreneurship, Uneven Progress and Solid Growth in although not equally distributed. As Acemoglu Colombia and Robinson (2019) discuss, sophisticated eco- nomic elites are able to exploit the weakness of The experience of Colombia can help demonstrate the state in the periphery by illegally accumulating how weakened central governance and a history land and assets (Oxfam, 2013). They are also able to of conflict can coexist with economic growth and use it to establish monopolies. This has meant that uneven stability. In Colombia, after a civil war in the most of the direct benefits of growth have passed mid-20th century, the re-constituted state reached on to a relatively small elite. Colombia is one of an equilibrium with guerilla groups in the country’s the most unequal countries in the world, by most interior. By the 1970s paramilitary groups and drug conventional measures, second only to Honduras cartels flourished. The balance was upset in the in Latin America. 1980s and 1990s when one major cartel launched a campaign of terror against the state. This led to a 205 “By using a discourse related to state and constitution, the massive military crackdown on cartels and a rewrit- government tried to keep the opposition in check. By using a ing of the constitution. In the 2000s, the government discourse related to clan institutions, the opposition tried to effectively defeated its main paramilitary rival and displace the government’s legitimacy.” Renders (2012) p. 222. CHAPTER 4 Shaping Yemen’s Future Scenarios for Yemen Through 2030 Sanaa or elsewhere is likely to be either heavily contested by groups who feel left out of the new Yemen is in a highly volatile and unpredict- national compact; or achieved primarily through able situation. While a political settlement coercive force, potentially in a way that precludes would be a critical step towards recovery, it international aid and development assistance. would not guarantee a return to stability or Key informants also use the colloquial term broad-based growth. Yemen’s social, political “no war, no peace” to describe the current situ- and institutional fragmentation is likely to persist ation in Yemen, the continuation of which is seen into any post-conflict period, with territorial con- as the most likely alternative to a return to war. trol divided into de facto zones of authority and Since early 2022 fighting has ebbed to a low level of multiple overlapping formal and informal admin- intensity, initially because of the UN-brokered truce istrative frameworks. and recently because of a tacit agreement among Nor is even limited peace guaranteed. Many the Yemeni and regional parties to the conflict to well-informed Yemenis and outside observers see a informally maintain the truce. But an internationally wide range of potential trajectories for Yemen that mediated, sustainable political settlement to end do not correspond with conventional understand- the conflict remains a hope. ings of “peace”. These include a continuation of stop-start national conflict both between the IRG Political Economy and Macroeconomic and the Houthis, in line with patterns observed Scenarios over the past eight years; more infighting among the forces currently nominally aligned with the IRG; Yemen’s future may be impossible to predict or both. but this does not mean that a range of scenarios Key informants widely share the view that built around the above context cannot be used only a de facto or de jure decentralized gover- to sharpen policy thinking. To support planning nance structure can prevent a slide back into efforts, the World Bank commissioned two comple- conflict. Any effort to rebuild a central state in mentary scenario analyses to accompany this report, 75 76 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES one reflecting possible political economy outcomes in future growth. Contested state legitimacy for Yemen’s future, the other a complementary set and continued violence with or without a political of potential macroeconomic outcomes. settlement would starve Yemen of the stability that The political economy scenarios build on a would foster much-needed improvements in liv- rigorous political economy analysis (presented ing standards via public and private consumption in Chapter 3), largely substantiated by exten- and investment, and broad economic growth. (See sive key informant interviews conducted for this Figure 62). report. It is critical to stress that the scenario analy- The macroeconomic CGE scenarios, on the ses should be read as broad indications of poten- other hand, build on the best available data tial outcomes rather than forecasts of political on Yemen for through 2020, and are based on economy or economic developments. Nor should the new Social Accounting Matrix (SAM), out- they be read as an endorsement of any one politi- lined in Chapter 2. A new SAM, updated to 2020 cal outcome. (more details about the SAM are available in the Annex), is the main data source for the CGE model, Divergent Path to Recovery the Sustainable Development Goals Simulation Sub-Scenarios Integrated Model (SDGSIM), which was used for the scenario analysis. Like other CGE models, it In the political economy analysis, the two Path to imposes rigor based on first economic principles, Recovery scenarios are broken into sub-scenar- considering the links between the domestic actors ios, reflecting the different ways the outcomes (households, producers, and the government), the they describe (a Restored State or Imperfect constraints under which they operate, and their Peace) might be achieved. Scenario 3: Restored interactions with the outside world through trade, State could be achieved either via complete territo- private and public transfers, and financial flows rial victory by one of the warring parties, or through (see Box 4). a UN-brokered political settlement that leads to The analysis provides a broad sense of the the formation of a new central government. It is political, international, and macroeconomic assumed that the new national authorities in either context across two categories of scenarios. of these sub-scenarios would face a considerable Under the first two “Business-as-Usual” scenar- number of challenges to their authority and that the ios—Status Quo and No War, No Peace—violent situation on the ground would remain shaky. In the conflict either continues with occasional periods case of an outright or near outright Houthi victory, of military inertia, or lapses into truce or ceasefire donors would be less inclined to providing aid and but without a political settlement that sustainably grants (see Figure 62). ends the war. The second set of scenarios unfold Key Informant interviews indicate that in a context where political / military develop- Scenario 4: “Imperfect Peace”—the most stable ments allow for either de jure or de facto central- end-state—is most likely to occur in the context ized (Restored State) or decentralized (Imperfect of decentralization. Again, this can happen in dif- Peace) state authority, creating space for a “Path ferent ways: via local authorities insisting on their to Recovery”. de facto autonomy regardless of the nature of a Each scenario has important ramifications settlement: or as part of a mutually agreed de jure for internal political and macroeconomic stabil- political framework that explicitly decrees decen- ity, investment and consumption, and interna- tralization. In the first sub-scenario, donors may feel tional aid and grant support, key determinants that they have no choice but to work with central Shaping Yemen’s Future 77 Figure 62 Projected Growth and Other Macroeconomic Indicators by Scenario Growth in other macro indicators by scenario (% pt. deviation from base) 14 12.5 12 11.4 11.4 10 8.0 8 6.5 6.5 5.9 6.0 6 5.1 4.0 4.0 4 3.2 3.3 3.1 2.2 2.6 2.0 2.0 2.2 2.1 2.2 2 1.3 1.1 1.2 1.1 1.2 1.1 0.7 0.9 0.2 0 Limbo Limited rec Limited rec (+aid) Enhanced rev Enhanced rev (+aid) Best case scenario Hhd cons’on Private inv’ment Exports Imports GDP at f.c. Table 9 Yemen Scenarios Through 2030 Political Economy Scenarios Macroeconomic Scenarios Business as Usual Status Quo Scenario 1 Base Scenario The war continues along a similar trajectory to the past eight years, with Economic growth is slow and living conditions do not improve, with GDP violence ebbing and flowing. growth of 2 percent from 2023 onwards. Poverty remains high. “No War, No Peace” Scenario 2 Limbo Conflict intensity dwindles, perhaps because of a ceasefire, but there is no Growth is marginally better than in the base scenario, around 3.1 percent political settlement. on average. Exports and imports increase marginally relative to 2022. Poverty remains high. Path to Recovery Restored State Scenario 3 Limited Recovery De facto and/or de jure authority are consolidated around a single central- GDP grows at twice the rate of the base scenario, and public investment ized set of state institutions. This authority is periodically contested by increases. Poverty decreases marginally. subnational actors. Imperfect Peace Scenario 4 Enhanced Recovery A political settlement either enshrines de jure decentralization or leads to Growth is three times higher than the base scenario. Government and de facto decentralization that is accepted by national elites. private sector investment increases and poverty falls to levels approaching those seen pre-war. authorities, and may limit the funds they deliver, Mirroring this split, the Path to Recovery while in the latter they may feel more confident macroeconomic scenarios are also subdivided: about directly supporting local authorities, particu- Into “Limited Recovery” and “Limited Recovery larly those they share an affinity with. Plus”; and “Enhanced Recovery” and “Enhanced 78 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES BOX 4: MAIN MODEL ASSUMPTIONS The Path to Recovery scenarios assume that the government achieves a higher level of efficiency. This is implemented in the model through real govern- ment consumption, which grows at a fixed annual rate of 2 percent in the baseline, while it is assumed to stay unvaried in both the Limited Recovery and Enhanced Recovery scenarios, in the sectors outside of health, education and infrastructure. In other words, both the Limited Recovery and Enhanced Recovery scenarios assume some level of efficiency gains, as the government is expected to be able to channel more resources to high-priority spending (such as health, education and infrastructure) without increasing taxes or tapping into foreign aid. More specifically, for the Enhanced Recovery scenarios, the increases in government health and education spending (split into consumption [current] and investment [capital]) during the period 2023–2026 are defined to permit the country to deliver services typical of the 90th LIC percentile starting from 2026. This is assumed to generate a more gradual increase in Yemen’s human-capital index (HCI) to the 90th LCI percentile during the period 2023–2030 as a growing part of the labor force benefits from improved health and education. In its turn, labor productivity grows relative to the 2022 level at a rate that, according to cross-country evidence, is expected given the HCI increase. For infrastructure, the Enhanced Recovery scenarios impose gradual increases in operations and maintenance and investment spending that, on average may be sufficient for low- and middle-income countries to achieve the infrastructure-related sustainable development goals (SDGs; Rozenberg and Fay 2019; Foster et al. 2022). Drawing on cross-country estimates (Gupta et al. 2014, p. 171), it is assumed that the infrastructure capital stock has a marginal productivity (MP) of 0.30 for Enhanced Recovery and Enhanced Recovery Plus. The gains are realized in the form of increases in total factor productivity (TFP); more specifically, once installed and other things being equal, value-added increases by 0.30 (0.60) YER for every additional YER of infrastructure capital with the gains spread across all sectors in proportion to their initial shares in value added. The central assumption of the Path to Recovery “Plus” scenarios is that Yemen receives additional foreign grant aid: in each year 2023–2026 foreign grants increase by 1 percent of 2022 GDP (i.e., reaching 4 percent in 2026); the increases are reversed during the period 2027–2030 so that, in 2030, foreign grants are back at the level of the base scenario. Similarly, the assumed increase in social transfers from the government to households would permit Yemen to reach the 90th percentile among LICs in terms of spending on social safety nets, assuming that Yemen’s initial level was in the bottom half among LICs (more specifically in the 25th percentile). For remittances, the baseline assumes that they grow at the same rate as the population, 2 percent, i.e., that they are unchanged in per-capita terms. For the enhanced recovery scenarios, it is assumed that, starting from 2023, the annual growth rate increases by 1 percentage point. A practical condition for this situation would materialize neighboring countries in the Gulf Cooperation Council (GCC) provide preferential treatment to worker migrants from Yemen and facilitate their remittance back home. In our model, the Yemeni households primarily use the increases in social transfers and remittances to boost their consumption. Table 10 Growth Assumptions Across Macroeconomic Path to Recovery Scenarios Scenario Shock area Changes compared to base a Limited recoveryb Enhanced recoveryc Health & education Government health & education spending (% of base GDP) (evenly split +2.7 +5.4 between consumption and investment) Labor productivity in 2030 (%) — gradual increase 2027–2030 +7.9 +15.8 Infrastructure Government infrastructure spending (% of base GDP) +3.3 +6.7 (32.5% consumption, 67.5% investment) Social transfers Transfers from government to households (% of GDP) +0.6 +1.2 Remittances Transfers to households from abroad (US$) (% annual growth) (2023–2030) +0.5 +1.0 FDI Foreign direct investment (% of base GDP; in US$) +3.45 +6.9 Oil-gas output Change in oil & gas resource extraction levels (% annual growth) +1.3 +2.4 a Unless otherwise noted, changes are introduced gradually 2023–2026 and after that kept in place up to 2030. b The difference between the two subcenarios within ‘limited recovery’ is that the latter (aid) enjoys foreign grant increases that in 2023–2026 gradually increase from 0.5 to 2 percent of base scenario 2022 GDP and in 2027–2030 gradually decline, in 2030 being back at the 2030 base level. c The difference the subscenarios within ‘enhanced recovery’ is that the latter two cases enjoy foreign grant increases that are twice as large as for first subscenario, peaking at 4 percent of 2022 GDP in 2026. Specifically the best-case scenario differs from all other recovery scenarios in the MP of government capital is 0.6 instead of 0.3. Shaping Yemen’s Future 79 Recovery Plus”. Under the two “Plus” macroeco- Peace, decentralization 206 and increased nomic scenarios foreign grant aid increases faster public investment efficiency can yield dramatic between 2022 and 2026, and aid remains at a gains in employment and poverty reduction. In higher level up to 2029. The increase in foreign aid all scenarios, GDP growth reduces unemployment and grants is associated with the most acceptable rates relative to the base by 2030. GDP growth and political outcome for the international community employment rates are closely linked. In the Imperfect (e.g. the restoration of the state or establishment of Peace/ Enhanced Recovery scenario, female unem- a new decentralized order via UN-brokered settle- ployment falls slightly below its 2022 level. The pov- ment rather than the imposition of de facto author- erty rate drops below the base even in Scenario 2: ity at the national/ local level) (see Tables 9, 10). No War, No Peace/Limboscenario, and even more in all other scenarios. Improved health and education Comparative Context services, better infrastructure, and greater physical security would yield additional gains in household Building on the comparative context section of welfare that are not captured by the model. Chapter 3, each “Path to Recovery” scenario Across all non-Base scenarios, growth accel- includes a short concluding note on lessons drawn erates for all sectors, especially for manufactur- from other cases outlined in Chapter 3. As in ing. Particularly in the Scenario 4 (Imperfect Peace/ Chapter 3, this comparative work focuses on four Enhanced Recovery), the expansion of manufac- countries: Colombia, Sierra Leone, Somaliland, and turing is strongest due to the combined impact of Abu Dhabi / the United Arab Emirates. higher income elasticities of demand from house- holds (compared to agriculture) and the absence Key Policy Takeaways of sectoral natural resource constraints. Apart from in the No War, no Peace / Limbo scenario, growth These scenarios contain some important lessons is driven by other industries, with an especially for policymakers. Perhaps most importantly: even strong expansion in construction spurred by ris- in Scenario 2: No War, no Peace / Limbo—where the ing public and private investment. For extractive conflict remains on pause despite a lack of political industries, output rises fastest in Scenario 4. The progress—Yemenis witness some modest improve- growth of government services is also fastest in ments in their economic situation compared to the imperfect Peace scenario, with a policy-driven Scenario 1: Status Quo/Base, under which conflict increase in education and health spending and continues. In other words, if the fighting stops, life unchanged spending on public administration. gets better for Yemenis even if the political outcome Sectoral changes in export growth are highly is unsatisfactory. correlated with changes in sectoral GDP growth, Other than in Scenario 1: Status Quo/ Base, though less so in agriculture, where exports are gains in productivity, driven by improved secu- also affected by domestic demand (see Figure 63). rity and policy changes, have an almost imme- Aid and grants in particular will continue diate effect in boosting GDP (fig 68). These to play a vital role in Yemen, even in the best- productivity gains also boost income levels and case scenarios for peace and recovery over the they generate higher household consumption, sav- medium term. During the coming decade, foreign ings, private investment, exports, and imports, with the investment increases feeding back to higher 206 A central assumption is that this, as key informants and growth. In the Path to Recovery scenarios, private the macroeconomic scenarios suggest, can lead to improved investment growth is boosted by increases in FDI. security and stability. 80 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Figure 63 Projected Sectoral GDP Growth Sector GDP growth by scenario (% pt. deviation from base) 9 8 7 6 5 4 3 2 1 0 –1 Limbo Limited rec Limited rec (+aid) Enhanced rec Enhanced rec (+aid) Best case scenario Agriculture Mining Manufacturing Other industry Private services Gov’t services aid will be critical to enable the government to set and achieve realistic taxation goals while sup- porting household consumption, investment, and Figure 64 Grant Aid for Selected savings. Without adequate foreign aid, the gov- Scenarios ernment will be forced to increase taxation at the Grant aid for selected simulations (% of GDP) expense of household welfare, possibly sparking 15 renewed unrest, or to undersupply public goods 14 and services. Faced with the country’s ongoing 13 humanitarian crisis, a temporary increase in aid may 12 be necessary to protect the poorest and most vul- 11 10 nerable households while bolstering prospects for a 9 sustainable resolution to the conflict (see Figure 64). 8 Evidence gathered through interviews and 7 literature suggests there are some reasons to be 6 optimistic about the possibility of a process of the 2022 2023 2024 2025 2026 2027 2028 2029 2030 re-formulation of a unified state integrating local Rec-t Rec-ta Rec+t Rec+ta and customary institutions. However, this report Source: CSO, National Accounts Bulletin 2019. only evaluates the likelihood, rather than the desirability of one outcome versus others. The high strength and legitimacy of tribal and local institutions Scenario 1 is often associated with the process of disarmament Political Economy – Status Quo and re-integration. There is strong evidence that tribal structures are still very important in solving routine Under the political economy “Status Quo” problems at various degrees across regions in Yemen. scenario, the patterns of political and military Shaping Yemen’s Future 81 conflict of the past eight years remain largely annual rates close to 2 percent with mining being unchanged through 2030, and Yemen’s politi- the only exception, reflecting the central assump- cal and military fragmentation either expands tion that extraction rates remain relatively stagnant. or remains largely unchanged. Intermittent peri- Unemployment rates increase both for men and ods of intense fighting and some territorial gains women. Aggregate export growth is 1 percent, or losses are interspersed lengthy pauses for reca- kept in check by a negative growth rate for mining, libration and political outreach. Governance is still which in 2022 represented around 62 percent of total heterogeneous as are local economic outcomes. exports. For the non-mining sectors, the growth rates Oil and gas production remains overall subdued, are around 3 percent. As shares of GDP, the changes and largely similar to current output levels. Aid in size of the government and its different receipt remains largely focused on the humanitarian crisis. and spending items are minor. Foreign aid, which amounts to around 14 percent of GDP, is assumed Macroeconomic – Base to be almost entirely in grant form. Government for- eign financing (borrowing net of interest) payments In the Base macroeconomic scenario, current pat- is close to zero, generating a virtually unchanged terns continue. The scenario assumes that: (a) all ratio between foreign government debt and GDP. international (export and import prices) are constant in real terms; and (b) for most institutional payments, Scenario 2 including receipt and spending items in the govern- Political Economy – No War, No Peace ment budget, the GDP shares are constant, kept at the 2020 shares. In the political economy No War, No Peace sce- The picture for the Base: scenario is one nario, a cessation of hostilities is brokered but of stable GDP growth, averaging 2 percent, talks over a political settlement do not progress starting from 2023. To facilitate comparisons to significantly. This is a marginal but nonethe- non-base scenarios, GDP growth is simulated at a less meaningful improvement over the Status stable annual rate of 2.0 percent between 2022 and Quo scenario, and creates space for economic 2030.207 Absorption (i.e., the sum of domestic final growth. Current de facto administrative divisions demands) is boosted by Yemen’s trade deficit, per- remain in place, however, and the peace is occasion- mitting it to exceed—and by a wide margin—GDP ally punctured by low-to-medium intensity conflict. at purchasers’ prices and, even more so, GDP at Oil and gas production remains largely similar to factor cost. Within domestic demand, household current output levels. However, unlike the Base consumption dominates, followed by government scenario, private sector and initiative develop more, thanks to security improvements. consumption (which includes consumption financed by donors and managed by NGOs and others), pri- vate investment, and government investment. Macroeconomic – Limbo Poverty remains high, at over 70 percent, Macroeconomic gains in the macroeconomic through 2030, and unemployment is bound to Limbo scenario are limited but mark an improve- gradually increase. A continuation of war and frag- ment over the Base scenario. GDP growth through mentation does not allow the growth rate for house- 2030 is 3.1 per cent on average (a 1.1 per centage hold consumption to exceed the population growth point improvement on the base). Both exports and rate, leaving per capita consumption unchanged with a headcount poverty rate of 74.4 percent both 207 Unless otherwise noted, the growth rates are geometric in 2022 and 2030. The aggregate sectors grow at averages for the period 2022 and 2030. 82 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES imports increase, albeit only marginally compared to current Houthi dominance.208 Given the extent the 2022 (growing by 2.3 per cent and 2 per cent respec- political, military and institutional fragmentation, tively). Private sector and government fixed invest- however, it is possible that one or more local actors ment levels remain low (growing 3.1 per cent and resist or actively rejects a central state’s author- 2.8 percent respectively), and the headcount poverty ity, leading to political turmoil or outright conflict. rate remains above 70 per cent (at 72 per cent). Macroeconomic – Limited Recovery Comparative Context The macroeconomic Limited Recovery and Limited Recovery Plus scenarios project strong GDP Of the comparative contexts discussed in and investment growth increases compared to Chapter 3, this scenario has elements that resem- the Base scenario. GDP grows by 2 percentage ble Somalia’s recent past, suggesting potential points more than in the base scenario (doubling for individual development success stories at the to 4  percent), and fixed investment grows by subnational level like Somaliland’s. 6.5 percentage points, with particularly strong growth in public investment. Headcount poverty Scenario 3 experiences a marginal decrease compared to the Base scenario under Limited Recovery scenario and Political Economy – A Recentralized State by slightly more in the Limited Recovery Plus scenario, with increased aid flows accounting for the gap. Under In the two Recentralized State scenarios, a both scenarios, Yemen closes half of the gap between nationwide military victory leaves the Houthis its current government performance and that of the or the IRG in control of the entire country; or a 90th percentile of Low Income Countries (LICs). political settlement leads to a recentralization of state functions in Sanaa. Patronage networks Comparative Contexts – Abu Dhabi, are expanded to include (fully or partially) former Rwanda, Somaliland rivals. The oil sector expands. Particular attention is paid to increasing rents in the form of increased oil A unified state may be possible if it integrates exports and restored LNG exports. The new elites local and customary institutions rather than seek to sanction positively or negatively private sector attempting to assert its authority from the get-go. agents, depending on affiliation networks and inter- The high strength and legitimacy of tribal institutions ests, leading to a selection of the wartime oligarchy. is often associated with the process of disarmament Aid and grand flows increase, particularly if the new 208 Among the most common interpretations of the Houthi re- authorities build trust and constructive relationships gime, is that it represents an attempt to build a theocratic state with regional and international donors. in north of Yemen, potentially a hybrid of the current Republican In this and other scenarios, the sustainabil- system and the religious and class structure of the Imamate with ity—or not—of the Houthi institutional setting overtones of other politico-religious institutions like those found in the North of the country will be one of the in Iran.Truce Test: The Huthis and Yemen’s War of Narratives, In- crucial factors that determines the institutional ternational Crisis Group, 29 April 2022. Most available literature concludes that the Houthis have an ideological project which shape of the future Yemen. Whether the ultimate they intend to institutionalize through a state in the areas that situation involves all the 21 governorates as constitu- they currently control—and through future territorial expansion ent units of the future Yemen, or a portion of these, both within Yemen—and, potentially, into neighboring states. If either de jure or de facto, rests on several factors, but this is the case, it is likely to have significant consequences for in large measure by the nature and sustainability of international aid and development support for Yemen in future. Shaping Yemen’s Future 83 and re-integration. There is strong evidence that Scenario 4 tribal structures are still very important in solving routine problems at various degrees across regions. Political Economy – An Imperfect Peace How Yemen’s tribal structures can over time be institutionalized within the state is a critical In the political economy Imperfect Peace sce- question. A comparative look can help learn from nario, a political settlement is agreed that either other contexts. There are interesting lessons to be enshrines or de facto recognizes current levels learned from Somaliland, for example. There, the of decentralization, affording considerable pow- government slowly but systematically consoli- ers to governorates and/or local authorities. This dated its position until it had reached an equi- in effect cements the status of the rival parties on librium with clan leaders and other subnational the ground and leads to a new round of coalition- power centers. building among former rivals as each group seeks Examples of state construction in Abu Dhabi to outbalance key rivals via cooperation with for- and also Rwanda suggest that it is possible to mer enemies. Resource sharing likely becomes an build a segmentary state on the basis of legiti- increasingly contentious issue, potentially leading to fresh rounds of violence. Oil exports increase. mate traditional institutions that can be highly The shadow of autonomy and secession developmental. While the complexity of doing this hangs over this scenario, which is again bro- in Yemen should not be underplayed, there is no ken into two sub-cases. In the first sub-scenario, reason to believe it would not be possible to build either decentralization happens because local the type of highly effective states seen in the UAE, in powers force it into de facto existence despite the sociological context of Yemen. The devil is in the efforts to reform a central state in Sanaa; and/or details of how the tribes are to be integrated into the one or more parties attempts to gain indepen- state and how tasks and resources are to be divided. dence leading to a standoff over aid and develop- While Yemen’s tribal structures can be institu- ment assistance to one or more areas of Yemen. In tionalized within the state, the credibility of the the second, some form of federated or decentral- process will be pivotal. The experiences with state- ized state is formed as part of an internationally building in the UAE, Sierra Leone, and Somaliland sanctioned political process and all regions of show that it is possible to build a modern government the country remain within the new state structure, that incorporates tribal allegiances and traditional leading to higher aid and grant inflows. institutions. The complexity of such a process in Yemen Macroeconomic – “Enhanced Recovery” should not be underestimated. But the country could eventually build a highly effective state with In the two sub-cases for the macroeconomic embedded tribal elements along the lines of the UAE. Enhanced Recovery scenarios, growth is higher However, the specific conditions under which the tribes than in any other context, rising to 6 per cent are to be integrated into the state and the division of on average between 2022 and 2030. This sce- power and resources among them will be enormously nario also sees the greatest increase in investment consequential, as will the participants’ ability to estab- (government and private sector fixed investment lish mutual trust. As the issue of credibility will remain rises to 30.9 per cent and 13.6 per cent respectively at the core of any future constitutional process, lever- under both sub-scenarios) and productivity (3 per aging traditional institutions designed to foster trust cent under both scenarios) and fall in the headcount could increase the likelihood of a successful outcome. poverty rate. The differentiating factor between 84 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES “Enhanced Recovery” and “Enhanced Recovery inimical to economic growth. Despite repeated Plus” scenarios is the lever of international aid and waves of conflict, growth in Colombia has been grants to Yemen (which as is argued elsewhere is steady and positive. But it has not been equally likely to be contingent on the nature and durability distributed. One of the paradoxes of Colombia is of a political settlement). In the Enhanced Recovery how a country can have such economic sophistica- Plus scenario, Yemen realizes improvements that tion and entrepreneurship in parallel with horrific bring the country up to outcomes typical of the violence. The answer is that these things have 90th percentile among LICs. adjusted to each other over the past 70 years. Sophisticated economic elites are able to exploit Comparative Lessons – Colombia, Sierra Leone the weakness of the state in the periphery by ille- gally accumulating land and assets (Acemoglu and Evidence gathered through interviews and a Robinson, 2019; Oxfam, 2013). They are also able review of the international literature yields rea- to use it to establish monopolies. This has meant sons to be optimistic about the possibility of that most of the direct benefits of growth have establishing a decentralized state that integrates passed on to a relatively small elite. Colombia is tribal allegiances and traditional institutions. one of the most unequal countries in the world, The international experience shows that strong and by most conventional measures, second only to legitimate traditional institutions can support a post- Honduras in Latin America. This may be the path conflict process of disarmament and re-integration, that Yemen follows in the event of a political set- and power-sharing agreements between different tlement, with its private sector and post-conflict tribal polities have proven effective in certain cases. elites reaping the benefits of peace while ordi- However, any post-war political and economic order nary Yemenis remain stubbornly stuck in poverty. will reallocate power and resources based on the gov- Despite a history of violent conflict, mean- ernance capacity, strategic positioning, and military while, Sierra Leone has remained at peace for capabilities of local authorities. over two decades. Local political legitimacy and Even if a stable equilibrium is reached, some the integration of local customary and formal parts of the country will inevitably fare better authorities into the country’s system of governance than others. Some local interests might be better has allowed a gradual reconstitution of the soci- served under an alternative scenario—such as the ety. The international community has also played “restored state”—even if it is suboptimal for the an important role, and the World Bank-sponsored country as a whole. Other factors, such as the con- post-civil war decentralization plan contributed to tinued bifurcation of monetary policy, could widen enable power-sharing between the major political regional disparities and complicate the peace. parties. However, the strength of traditional institu- Colombia demonstrates that weak central tions at the local level has likely been the decisive rule and waves of conflict are not necessarily factor in preventing a renewed conflict. CHAPTER 5 Conclusions The comparative and scenario analysis above level; and (iv) fostering greater transparency and indicates that Yemen’s diffuse traditional institu- data availability. tions could be leveraged for stability, and that a resolution to the conflict based on the decentral- A. Restoring Macroeconomic Stability ization of political power could create the condi- tions for a robust post-conflict recovery. A highly Reestablishing a credible macroeconomic frame- centralized state would be acutely vulnerable to the work will help guide the transition away from historical challenges of elite rent-seeking and patron- the wartime economy. While returning Yemen to age, which would weigh on growth while undermin- a stable growth path will require an end to the war ing macroeconomic stability and national cohesion. followed by a comprehensive reform effort, the In any scenario, rebuilding Yemen’s physi- authorities can take immediate action to improve the cal and institutional infrastructure will require macroeconomic environment. Key measures include extensive efforts by the authorities and the strengthening fiscal and monetary policy to achieve international community to address financial price stability; bolstering hard-currency inflows by needs, capacity gaps, macroeconomic imbal- increasing oil and even gas exports; and reunifying ances, and the erosion of human capital. state institutions so that monetary policy and finan- Unanticipated shocks could complicate policy for- cial regulation are the purview of a single institution mulation and execution. Any additional challenges rather than two or more competing institutions. will make it harder to reconcile growth policies with social protection in a context of limited fiscal space. Price Stabilization In the short-to-medium term, the authorities should focus on four priority high-level directions: (i) restor- As noted in Chapters 2 and 3, one of the key ing macroeconomic stability, and establishing stable drivers of Yemen’s cost-of-living crisis has been monetary-exchange regime and financial sector; the depreciation of the Yemeni rial and the (ii) enabling private-sector-led growth and produc- resulting high rates of inflation. The rial’s depre- tivity gains; (iii) building capacity at the subnational ciation, particularly in IRG areas, has been caused 85 86 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES by multiple factors, including the general effects in IRG areas (page 81). These subsidies continue to of the war, the loss of hard-currency inflows from represent the largest burden on the state budget, oil exports, the central bank split in 2016, the mon- and the political cost of lifting them may be too high etization of IRG fiscal deficits, and the worsening for the IRG to contemplate in the coming months. If regulatory and security environment. Economic ending the policy is not feasible, strengthening col- competition between the IRG and the Houthis has lection efficiency in the electricity sector could at least also had a deleterious effect on the country’s fiscal somewhat alleviate the burden on the public finances. and monetary stability. Reunifying monetary policy across IRG and In 2022, high international oil prices boosted Houthi areas, along with hard-currency man- revenue and forex accumulation, but one-off wind- agement and financial regulation, could further falls are not a substitute for structural reforms. shore up price stability. The bifurcation of mon- Dysfunctional macroeconomic policy inaction remains etary policy has intensified price differentials and one of the most significant risks to Yemen’s outlook economic divergence between IRG- and Houthi- (see page 22–23). Recent reforms have strengthened controlled areas (page 37). While the UN envoy’s the IRG’s credibility and effectiveness in addressing office has mooted negotiations on economic policy macro-fiscal vulnerabilities. Since November 2021, reintegration, the World Bank and IMF could play IRG-overseen hard-currency auctions have helped sta- an important role in fostering coordination and bilize the exchange rate, with weekly sales increasing encouraging the uptake of best practices by the from US$20 million to US$30 million. The authorities two central banks. Yemen’s development partners have also begun using the market exchange rate for should continue to offer support while remaining oil revenues, which immediately increased govern- mindful of the risks of entrenching the fragmenta- ment revenues. Historically high international oil tion of public institutions. prices have also contributed to fiscal sustainability and helped avert further monetary financing of the fiscal Hydrocarbon Production and Exports deficit. While favorable oil prices and the prudence of the CBY-Aden have established a basis for quali- Price stability will also require predictable fied economic optimism, structural reforms remain a and sustainable hard-currency inflows. Yemen’s precondition for sustainable growth. current-account deficit has been widening and is Refraining from monetary financing of the expected to double as a share of GDP and triple fiscal deficit (see page 34–35) and continuing to in nominal US dollar terms by the end of 2022. enhance the transparency and independence of Declining official development assistance and remit- the central bank would help restore monetary tance receipts, as neighboring countries apply more stability. Monetary conditions have tightened protectionist labor policies, have offset the predicted during 2022, as the CBY-Aden drastically reduced increase in oil exports. Yemen’s difficulties in financ- fiscal-deficit monetization to levels consistent with ing its external deficit remain significant, and while its stated money-supply growth targets. The CBY- the country has thus far relied on successfully liqui- Aden’s policy shifts effectively bolstered exchange- dating previously frozen foreign-currency deposits, rate stability, and inflation fell to single digits in the delivery of promised external financial assistance 2022, even as prices for food and other commodi- will be crucial to sustain foreign-currency reserves. ties increased by up to 30 percent. Boosting hydrocarbon production and Energy subsidies are an expensive and inef- export could greatly improve Yemen’s access to fective policy. The conflict has effectively ended hard currency in the near term. There are three all subsidies except for indirect electricity subsidies options to increase income from the hydrocarbons Conclusions 87 sector (pages 67–68). The first is to ensure that all infrastructure and human capital, and enhanc- the oil and gas produced in the country is sold on ing subnational governance capacity. As budget domestic and international markets via formal chan- constraints limit the scope for expansionary fiscal nels. The second is to bring shuttered production policy, the authorities should invest in productiv- capacity online, which could increase output by over ity-enhancing technologies while strengthening 30 percent. The third is to restore LNG production financial access for small and microenterprises. A and exports by resuming work on the Yemen LNG formalization program for small firms could also project. Utilizing flared gas to generate electricity improve tax collection while fostering a more com- could generate significant fiscal savings by sup- petitive business environment. planting fuel-oil imports, but attracting investment Rebuilding human capital will enable renewed in fixed infrastructure will require greater security. productivity growth. The deterioration of the edu- Increasing hydrocarbon revenue through cation system due to the destruction of schools, the any of these three strategies could prove politi- lack of funding for teachers’ salaries, and the physi- cally challenging and would entail significant cal danger of the conflict has eroded the productive risks. Increasing oil production may require capital capacity of Yemen’s future workforce (page 30). To expenditure that the private and state sectors are address these losses, the authorities should focus unwilling to provide absent stringent guarantees. on enhancing instructional quality, lowering drop- The Yemen LNG project is exposed to security out rates, and expanding access to early childhood threats from numerous armed actors, while a corrupt development services. institutional environment poses serious reputational Special attention should be paid to creating risks. Indeed, the re-emergence of rent-seeking and opportunities for young workers, who may other- patronage is a central risk associated with increasing wise join armed groups as a matter of economic hydrocarbon production and export. The authorities necessity. These efforts should include job train- should strive to ensure that the extractive industries ing and internship programs. Tighter connections benefit the whole economy by de-risking the sector, between employers and the school system can help increasing transparency, and creating an enabling ensure that relevant skills are included in the curricu- environment for FDI. A comprehensive effort to lum and students have access to information about secure transparency and accountability over all labor demand. Training in business skills and micro- sectoral revenue would be an important first step. grant programs could help foster entrepreneurship. In parallel, Yemen’s recent gains in renew- Institutional quality is fundamental to growth able energy should be accelerated. Building on and productivity, especially at the subnational existing studies of Yemen’s renewables potential level. Since the start of the war, governorate and (page 69), significant investment should be made district-level institutions have played a crucial role in into promoting inexpensive, equitably distributed maintaining a minimal level of public administration energy. Solar energy should be prioritized given and service delivery. The donor community can do the relative maturity of the solar electricity sector in more to support subnational institutions and allow Yemen, local familiarity with solar, and well-estab- them to assume greater responsibility for rebuild- lished retail supply chains. ing infrastructure and restoring the supply of public goods. The quality of subnational institutions may B. Boosting Productivity Growth ultimately determine whether post-conflict recon- struction and state reintegration efforts are success- Improving growth and productivity will require ful, and significant resources should be devoted to facilitating private investment, rebuilding strengthening them. 88 ‫ – املستقبل‬GLIMMERS OF HOPE IN DARK TIMES Capacity-building and localization within costs while lowering farm-gate prices, devastating subnational institutions could help alleviate rural livelihoods. Delays and detours also waste Yemen’s deep and persistent regional disparities. hundreds of millions of hours, and restoring road Building the capacity of local authorities to imple- connectivity would greatly reduce travel times. ment development programs, provide services, and design investment plans that reflect local needs and Social Protection and Household Welfare participatory processes could improve household welfare both during and after the conflict. Customary Affordability is fundamental to food security, law and non-state local leaders can help improve especially in the near term. The authorities should security, resolve conflicts, and normalize economic shift focus from wheat-centric food security to holis- activity. Over time, these customary practices and tic nutrition security. Diversifying food imports, both traditional institutions can be integrated into the in terms of goods and source countries, and diversi- formal public sector. fying domestic agricultural and livestock production could strengthen resilience to shocks. Trade-finance Transportation options, such as establishing dollar-denominated supply arrangements with importers or offering Reopening roads would significantly reduce partial guarantees in cooperation with external freight costs, enhancing the competitiveness of partners, could help boost the import supply. local businesses. Longer and more complex road Increasing domestic production is a long- routes have changed the economics of overland term goal that will require raising agricultural freight (page 46). Restoring Yemen’s road transport productivity and moving away from input- network and ending the closure of major highway intensive to knowledge-intensive production corridors will be crucial to support the economic practices, including the use of improved seeds recovery, reverse cumulative losses incurred dur- and agricultural technologies. More data on the ing the war, and boost domestic production, trade, spread of qat plantations and their effects on food private consumption, and investment. Ending road production would help inform agricultural policy, closures would also restore severed value chains and while additional information on domestic qat usage increase economic activity along large sections of could underpin a strategy to counter its adverse closed roads, where isolation has led to a massive health effects. These goals will be difficult to achieve loss of income-generating opportunities. Greater during wartime. connectivity will offer trade-driven growth opportu- Facilitating and formalizing remittance flows nities to entrepreneurs and existing businesses alike. would also contribute to economic security. High Improved connectivity would have especially levels of informality and weak regulatory oversight positive implications for agriculture and food of the financial sector have allowed remittance security. Delays due to road closures are among networks and local authorities to charge exorbi- the main reasons behind the massive drop in pro- tant fees. Improving the regulatory framework, duction, sales, and exports of perishable foods and strengthening market infrastructure, and embracing other agricultural goods. Before the war, agricul- innovative technologies could increase the share tural and fishery products contributed more than of remittances that accrues to Yemeni households. 20 percent of GDP and were Yemen’s second largest Restoring the provision of basic services, export. The decline in agricultural production is also especially for the poorest households, will a major cause of food insecurity, as the disruption require sustained improvements in local gov- of transportation networks raised agricultural input ernance capacity. The lack of essential services Conclusions 89 is taking an immense human toll and may cause Houthis and IRG authorities has led to mistrust of long-term economic scarring. The near-total lack their governance practices and contributed to the of medical facilities and services, combined with widespread perception of elites as rapacious, self- limited access to clean water and sanitation, have serving, and unconcerned with the public interest. led to a sharp increase in morbidity and mortality. Sound public policy requires reliable and According to the UN, every 10 minutes one Yemeni comprehensive data. A lack of trustworthy data child dies from preventable diseases.209 obscures priorities and complicates development An inclusive dialogue will be vital to forge a planning. Data constraints are pervasive in Yemen, new social contract based on equal citizenship and reconciliating numerous, often inconsistent, rights, civic trust, and public accountability. data sets could be a starting point to improve Ensuring the participation of women and vulner- data quality. The international community can offer able groups will be a key challenge, but building a greater support in this area. national vision that acknowledges and accommo- Improved coordination around data collection dates a diverse range of identities is crucial to foster across agencies could reduce duplication and free social cohesion. In Houthi areas, formal policies and up resources to collect data on underexplored informal norms have reduced women’s participation issues. While this CEM demonstrates that it is pos- in the labor market and restricted their movement sible to conduct thorough and rigorous analysis by without guardians.210 Empowering women to fully using innovative and non-traditional approaches, the participate in all aspects of social and economic life intrinsic value of strengthening the production and will help maximize the value of the country’s limited availability of traditional data cannot be overstated. human capital. Yemen is in a unique situation, as every data source is likely flawed and multiple shocks are occurring Improving Transparency, as Well as C.  simultaneously. As the country emerges from conflict, Data Quality and Availability establishing a new baseline will require extensive efforts to expand and enhance data collection, as Transparency is crucial to preventing the sound analytical underpinnings will be vital to effec- reconstitution of the patronage networks that tive policymaking in a new era of peace and recovery. thwarted reform efforts, slowed productivity gains, and hindered inclusive growth in pre- 209 https://www.unocha.org/fr/story/yemen-what-people-need- conflict Yemen. Corruption still appears to be most. pervasive, and each faction distributes rents to 210 Key informant interviews by World Bank staff, February- important clients. A lack of transparency by both the September 2022. 1818 H Street, NW Washington, DC 20433