Crisis and Recovery Learning from COVID-19’s Economic Impacts and Policy Responses in East Asia Table of Contents Reader’s Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii • Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 by Alvaro Gonzalez, Lydia Kim, and Maria Ana Lugo Pre-Pandemic Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 The Economic Impact of the Pandemic and Government Efforts to Mitigate It . . . . . . . . . . 12 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Annex I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 • CHAPTER 1  Cambodia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 by Kyung Min Lee, Isabelle Salcher, Wendy Karamba, Kimsun Tong, and Trang Thu Tran Timeline of the COVID-19 Pandemic and Government Measures . . . . . . . . . . . . . . . . . . . . . 40 Impacts on the Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Employment Impacts: Shocks and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Fiscal Support to Firms and Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Annex 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 • CHAPTER 2  Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 by Rabia Ali, Aufa Doarest, Ade Febriady, and Bayu Agnimaruto Timeline of the COVID-19 Pandemic and Government Measures . . . . . . . . . . . . . . . . . . . . . 67 Employment Impacts: Shocks and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Fiscal Support to Firms and Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Annex 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 i ii CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA • CHAPTER 3  Malaysia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 by Smita Kuriakose, Ririn Salwa Purnamasari, Trang Thu Tran, Sarah Waltraut Hebous, and Zainab Ali Ahmad Timeline of the COVID-19 Pandemic and Government Measures . . . . . . . . . . . . . . . . . . . . . 87 Employment Impacts: Shocks and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Fiscal Support to Firms and Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Annex 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 • CHAPTER 4  Mongolia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 by Nona Karalashvili and Ikuko Uochi Timeline of the COVID-19 Pandemic and Government Measures . . . . . . . . . . . . . . . . . . . 111 Impacts on the Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Employment Impacts: Shocks and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Fiscal Support to Firms and Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Annex 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 • CHAPTER 5  The Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 by Edgar Avalos, Irene Jo Estigoy Arzadon, Jaime Frias, Karl Robert Lasmarias Jandoc, Jesica Torres, and Trang Thu Tran Timeline of the COVID-19 Pandemic and Government Measures . . . . . . . . . . . . . . . . . . . 135 Employment Impacts: Shocks and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Fiscal Support to Firms and Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Annex 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 • CHAPTER 6  Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 by Sarah Hebous, Shawn W. Tan, Trang Thu Tran, Matthew Wai-Poi, and Judy Yang Timeline of the COVID-19 Pandemic and Government Measures . . . . . . . . . . . . . . . . . . . 161 Impacts on the Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Employment Impacts: Shocks and Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Fiscal Support to Firms and Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 Annex 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 TABLE OF CONTENTS iii • CHAPTER 7  Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 by Pablo Fleiss Weinberger, Alvaro Gonzalez, Lydia Kim, and Maria Ana Lugo How Did Governments Respond to the COVID-19 Economic Slowdown? . . . . . . . . . . . . . 186 Research Agenda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 LIST OF FIGURES Figure I.1 Report framework: Impact of the COVID-19 shock on households and firms . . . . . . . . 2 Figure I.2 WEF Global Competitiveness Index 2019 – Select indicators of business environment . . . 7 Figure I.3 Financial development index in six countries in East Asia, 2009–19 . . . . . . . . . . . . . . . . . 8 Figure I.4 Poverty in East Asia and Pacific 2018 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure I.5 Inequality in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure I.6 Typology of social protection systems in East Asia and the Pacific . . . . . . . . . . . . . . . . . 12 Figure I.7 Number of reported COVID-19 cases, by region, January 2020–January 2022 . . . . . . 13 Figure I.8 Stringency and mobility indexes, by region, January 2010–January 2022 . . . . . . . . . . 14 Figure I.9 2020 GDP Index and projected 2020 GDP Index (2019 GDP = 100) . . . . . . . . . . . . . . . 16 Figure I.10 Percent of fully or partially-open businesses, by sector and firm size, 2020–21 . . . . 18 Figure I.11 Percent of fully or partially open businesses during the pandemic, by period (weighted) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure I.12 Average reported change in sales in 30 days before interview, relative to same period in 2019, by country and firm size (weighted) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure I.13 Percent of firms that began using or increased use of digital platforms, by period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure I.14 Percent of establishments in six East Asian countries that started or increased their use of digital platforms, by sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure I.15 Employment-to-population ratio in 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure I.16 Work stoppages in 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Figure I.17 Characteristics of workers and work stoppages in selected countries in East Asia and Pacific, May 2020–December 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure I.18 Non-farm business income vs. wage income losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Figure BI.2.1. Global comparisons of work stoppages and household income losses . . . . . . . . . . 26 Figure I.19 Household coping mechanisms adopted during the pandemic in six countries in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure I.20 Composition of COVID-19 support packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure IA.1 Timing of surveys conducted by the World Bank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure 1.1 Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Cambodia, 2020–22 . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 1.2 Firms experiencing changes in sales revenue during the 30 days before the survey, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 iv CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Figure 1.3 Reasons why workers in Cambodia stopped working, 2020–21 . . . . . . . . . . . . . . . . . . 44 Figure 1.4 Work stoppages in Cambodia, by level of education . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Figure 1.5 Loss of labor income in Cambodia, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Figure 1.6 Change in employment relative to before the COVID-19 pandemic in Cambodia, by sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Figure 1.7 Share of firms in Cambodia that granted leaves of absence, cut hours, or cut wages in the 30 days before the survey, May 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 1.8 Firm adoption or increase in use of digital technology in Cambodia, May 2021 . . . . 48 Figure 1.9 Correlation between changes in employment and labor productivity in Cambodia . . . 49 Figure 1.10 Share of working adults in Cambodia who switched jobs or sectors, May 2020–March 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 1.11 Household access to COVID-19 cash transfer program in Cambodia, 2020–21 . . . . . 52 Figure 1.12 Formal firms’ access to COVID-19-related public support in Cambodia, by firm size, 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Figure 1.13 Household access to COVID-19 cash transfer in Cambodia with and without non-farm business, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Figure 1.14 Correlation between access to COVID-19-related support and employment change in Cambodia, 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Figure 1.15 Types of COVID-19-related public support to firms in Cambodia, 2021 . . . . . . . . . . 55 Figure 1.16 Perceptions of beneficiary households in Cambodia of the importance of COVID-19 cash transfers to their economic well-being, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . 56 Figure 1.17 Coping strategies adopted by households in response to the COVID-19 crisis in Cambodia, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Figure 2.1 Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Indonesia, 2020–22 . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Figure 2.2 Operating status of firms and changes in sales in Indonesia, June 2020– August 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Figure 2.3 Changes in employment and income in Indonesia, May 2020–October 2021 . . . . . . 70 Figure 2.4 Characteristics of firms in Indonesia that experienced severe sales shocks . . . . . . . . 71 Figure 2.5 Digital adoption by firms in Indonesia between June 2020 and August 2021 . . . . . . . 73 Figure 2.6 Size and composition of fiscal packages in select countries in 2021 . . . . . . . . . . . . . . . 74 Figure 2.7 Correlation between size of fiscal package and number of COVID-19 cases in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Figure 2.8 Rollout of COVID-19 support programs for households and firms in Indonesia, January 2020–January 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Figure 2.9 Budget allocation and implementation of support programs in Indonesia, 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Figure 2.10 Beneficiaries of social assistance programs and other relief measures in Indonesia, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Figure 2.11 Targeting accuracy of household support in Indonesia, by program and income level, March 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 TABLE OF CONTENTS v Figure 2.12 Probability of a firm receiving assistance in indonesia based on decline in sales, June 2020 and September 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Figure 2.13 Beneficiaries of social assistance programs and other relief measures in Indonesia between March 2020 and October 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Figure 2.14 Correlates of firing and hiring workers in Indonesia in 2020–21 . . . . . . . . . . . . . . . . . 81 Figure 2.15 Percent of firms in Indonesia that were aware of government support programs for firms, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Figure 2.16 Reasons why Indonesian firms did not participate in government support programs, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Figure 3.1 Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Malaysia, 2020–22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Figure 3.2 Share of firms laying off workers and share of firms reducing hours and/or wages in Malaysia, October 2020–March 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Figure 3.3 Employment Dynamics of Malaysians who were working before the pandemic, February 2020–October/November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Figure 3.4 Correlation between change in sales and share of Malaysians working in 2021, by sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Figure 3.5 Change in employment in Malaysia, October 2020–March 2022, by firm size . . . . . 91 Figure 3.6 Share of workers in Malaysia that stopped working or suffered income shock, May/June–October/November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Figure 3.7 Share of Malaysians that stopped working during the pandemic . . . . . . . . . . . . . . . . . 92 Figure 3.8 Correlation between change in employment between October 2020 and March 2022 and pre-pandemic labor productivity in Malaysia . . . . . . . . . . . . . . . . . . . . . . . . . 93 Figure 3.9 Correlation between working from home during July 2021 and household access to fixed internet connection during May-June 2021 in Malaysia, by region . . . . . . . . 94 Figure 3.10 Access to government support in Malaysia, October 2020–March 2022, by firm size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Figure 3.11 Correlation between receipt of government support in Malaysia and change in sales, October 2020–March 2022, by sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Figure 3.12 Share of Malaysian households that received cash transfers, May/June–October/ November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Figure 3.13 Coping strategies by Malaysian households that did and did not experience economic shocks, May/June–October/November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Figure 3.14 Correlation between various factors and the likelihood of laying off workers in Malaysia, January-February 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Figure 3.15 Share of households in Malaysia partially or fully unable to cover current monthly basic needs, October/November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Figure 4.1 Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Mongolia, 2020–22 . . . . . . . . . . . . . . . . . . . . . . . . . 112 Figure 4.2 Characteristics of firms in Mongolia that experienced the largest declines in sales and employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 vi CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Figure 4.3 Differential impact on foreign-owned and exporting firms in Mongolia . . . . . . . . . 115 Figure 4.4 Change in employment in Mongolia between May 2020 and June 2021 among respondents who worked before the pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Figure 4.5 Employment status of men and women in Mongolia between December 2020 and June 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Figure 4.6 Work stoppages among respondents working pre-pandemic, May 2020– June 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Figure 4.7 Correlation between change in employment and baseline labor productivity in Mongolia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Figure 4.8 Work stoppages in Mongolia, May 2020–June 2021, by welfare quintile . . . . . . . . . 119 Figure 4.9 Components of Mongolia’s COVID-19 responses, 2000 through Mid-2021 . . . . . . 121 Figure 4.10 Coverage and types of government support in Mongolia among surviving firms . . . . 122 Figure 4.11 Size of Mongolia’s child money program and one-time cash payment, by wealth quintile (MNT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Figure 4.12 Percent of households in Mongolia that received government assistance, by welfare quintile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Figure 4.13 Correlation between baseline labor productivity in Mongolia and probability of receiving government support in August 2020 and February 2021 . . . . . . . . . . . . . . . . . . 124 Figure 4.14 Use of Mongolia’s child money program and the one-time cash payment by recipients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Figure 4.15 Effectiveness of government support to firms in Mongolia . . . . . . . . . . . . . . . . . . . . 125 Figure 4.16 Panel regression results: Receipt of one-time government cash transfer in Mongolia in April 2021 and likelihood of being food insecure . . . . . . . . . . . . . . . . . . . . . . . 126 Figure 4.17 Perceived effectiveness of the child money program and the one-time cash transfer in mitigating the effects of the COVID-19 shock in Mongolia . . . . . . . . . . . . . . . . . . 127 Figure 4.18 Types of government support firms in Mongolia cited as most needed, February 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Figure 4.19 Household opinions on government responses to the COVID-19 pandemic in Mongolia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Figure 4.20 Labor force participation and the employment-to-population ratio in Mongolia, 2019–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Figure 4.21 Activity status of registered business entities in Mongolia, 2019–21 . . . . . . . . . . . . 130 Figure 5.1 Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in the Philippines, 2020–22 . . . . . . . . . . . . . . . . . . . . . 136 Figure 5.2 Net job layoffs and net job creation by firms in the Philippines in November/ December and May 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Figure 5.3 Labor force participation and unemployment rates in the Philippines, 2019–22 . . 139 Figure 5.4 Index of number of workers in the Philippines in select sectors, 2019–21 . . . . . . . . 140 Figure 5.5 Index of employment in the Philippines, by type of employment and gender, 2019–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Figure 5.6 Unemployment rate in the Philippines, by years of formal schooling, 2019–21 . . . 141 TABLE OF CONTENTS vii Figure 5.7 Share of firms in the Philippines that report increasing their use of select technologies in the second quarter of 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Figure 5.8 Correlation between firm size and likelihood of receiving financial support or wage subsidy in the Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 Figure 5.9 Government support of households in the top 40 and bottom 60 percent in the Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Figure 5.10 Percent of micro, small, and medium-size enterprises in the Philippines that borrowed under the CARES program, by region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Figure 5B.1 Social assistance coverage in the Philippines in 2020, by wealth quintile, area, and gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Figure 6.1 Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Vietnam, 2020–22 . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Figure 6.2 Changes in sales in Vietnam between 2020 and 2022 relative to same period in 2019 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 Figure 6.3 Share of firms in Vietnam making employment adjustments, June 2020– January 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Figure 6.4 Underemployment and unemployment rates in Vietnam, by gender and age group, 2010–20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 Figure 6.5 Self-reported changes in household income in Vietnam, 2020 and 2021 . . . . . . . . 166 Figure 6.6 Percentage change in average earnings of women and men in Vietnam in third and fourth quarters of 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 Figure 6.7 Change in sales in Vietnam, by sector, June 2020–January 2022 . . . . . . . . . . . . . . . . 169 Figure 6.8 Average nominal monthly income of Vietnamese households in agriculture, industry, and services, fourth quarter 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 Figure 6.9 Quarterly informal employment rate in Vietnam (excluding agriculture), 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 Figure 6.10 Index of household income in Vietnam by people with and without formal sources of income, June/July 2020–March 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Figure 6.11 Average change in sales in Vietnam, by region, June 2020–January 2022 . . . . . . 172 Figure 6.12 Adjustment mechanisms adopted by firms in Vietnam, June 2020– January 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Figure 6.13 Correlation between changes in employment and labor productivity in Vietnam’s manufacturing sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 Figure 6.14 Adoption of digital platforms in Vietnam, by firm size, June 2020–January 2021 . . . . 174 Figure 6.15 Government spending and forgone fiscal revenue in response to COVID-19 in select countries in Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Figure 6.16 Access to support policies in Vietnam, June 2020–January 2022 . . . . . . . . . . . . . . . 176 Figure 6.17 Composition of support to firms in Vietnam, January 2021–January 2022 . . . . . . 177 Figure 6.18 Average change in sales by Vietnamese firms that did and did not receive public support, June 2020–January 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 Figure 6.19 Percent of firms receiving COVID–related government support in Vietnam, by firm size, June 2020–January 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 viii CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Figure 6.20 Correlation between probability of receiving support and labor productivity in Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 Figure 6.21 Public perceptions of government response to COVID-19 in Vietnam . . . . . . . . . . 180 Figure 7.1 Correlation between per capita GDP and size of announced fiscal spending in 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Figure 7.2 Fiscal spending on COVID-19 relief to households in six countries in East Asia in 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 Figure 7.3 Spending on income support to households and income losses in 2020 and 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Figure 7.4 Share of households that received government assistance for COVID-19 relief in six countries in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Figure 7.5 Percent of households that received social assistance in the past 30 days in six countries in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 Figure 7.6 Coverage and adequacy of cash transfers by region . . . . . . . . . . . . . . . . . . . . . . . . . . 196 Figure 7.7 COVID-19-related government support for firms in 2020 and 2021 by type of assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Figure 7.8 COVID-19-related government support for firms in six countries in East Asia in 2020 and 2021, by category of support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 LIST OF TABLES Table I.1 Select pre-pandemic country-level characteristics, circa 2019 . . . . . . . . . . . . . . . . . . . . . 5 Table I.2 Support to households and workers provided by six countries in East Asia . . . . . . . . . . 29 Table 1.3 Support to firms provided by six countries in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Table I.4 Financial sector policy measures to provide relief to borrowers . . . . . . . . . . . . . . . . . . . . 32 Table IA.1 Dates and sampling methods of surveys used in this report . . . . . . . . . . . . . . . . . . . . . . 35 Table 1.1 Select policy instruments Cambodia used to mitigate effects of the COVID-19 pandemic, 2020–21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Table 1A.1 Dates and sample sizes of high-frequency phone surveys of households in Cambodia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Table 1A.2 Dates and sample sizes of business pulse survey in Cambodia . . . . . . . . . . . . . . . . . . . 62 Table 1B.1 Timeline of actions January 2020–November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Table 2.1 Major social and business support programs in Indonesia . . . . . . . . . . . . . . . . . . . . . . . . 76 Table 2A.1 Dates and sample sizes of high-frequency phone surveys of households in Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Table 2A.2 Dates and sample sizes of business pulse surveys in Indonesia . . . . . . . . . . . . . . . . . . 85 Table 3.1 Instruments Malaysia introduced to support firms and households during the pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Table 3A.1 Dates and sample sizes of high-frequency phone surveys of households in Malaysia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Table 3A.2 Dates and sample sizes of business pulse surveys in Malaysia . . . . . . . . . . . . . . . . . 103 Table 3B.1 Type of assistance to firms and households based on various assistance packages throughout the pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 TABLE OF CONTENTS ix Table 3C.1 Correlations between firm characteristics and the likelihood of receiving government support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Table 4A.1 Dates and sample sizes of high-frequency phone surveys of households in Mongolia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Table 4A.2 Dates and sample sizes of the the COVID-follow up enterprise surveys in Mongolia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Table 5.1 Labor market responses to the COVID-19 pandemic by the government of the Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Table 5.2 Allotments, disbursements, and use of COVID-19 budget by select government agencies in the Philippines, 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Table 5A.1 Dates and sample sizes of high-frequency phone surveys of households in the Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Table 5A.2 Dates and sample sizes of business pulse surveys in the Philippines . . . . . . . . . . . . 154 Table 5B.1 Summary of programs by the government of the Philippines targeting workers and firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Table 6.1 National policy instruments Vietnam used to support firms and households during the pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 Table 6.2 Social assistance Vietnam provided in first COVID-19 relief package (percent of households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Table 6A.1 Dates and sample sizes of high-frequency phone surveys of households in Vietnam . . . . 183 Table 6A.2 Dates and sample sizes of business pulse surveys in Vietnam . . . . . . . . . . . . . . . . . . . . . 183 Table 7.1 Fiscal space variables, 2019 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 Table 7.2 Structural indicators in the labor market and dependency ratios, 2019 . . . . . . . . . . 189 Table 7.3 Percent of surveyed firms reached by government support, by survey wave . . . . . . 201 Table 7.4 Percent of surveyed firms reached by government support in six countries in East Asia, by firm size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 Table 7.5 Announcements of support to sectors by governments in six countries in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 LIST OF BOXES Box I.1 Household and firm survey data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Box I.2 Global comparisons of household work stoppages and income losses using high-frequency phone survey data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Box 5.1 Emergency social transfers in the Philippines during the COVID-19 pandemic . . . . . 145 Abbreviations 4Ps Pantawid Pamilyang Pilipino Program (Philippines) AICS Assistance to Individuals/Families in Crisis AKAP Abot Kamay ang Pagtulong APIS Annual Poverty Indicators Survey ARDB Agricultural and Rural Development Bank (Cambodia) ASEAN Association of Southeast Asian Nations BKC Bantuan Khas COVID–19 (Malaysia) BKK Bantuan Kanak–Kanak (Malaysia) BKM Bantuan Keluarga Malaysia BKM Bantuan Keluarga Malaysia BLT DD–Bantuan Langsung Tunai Dana Desa (Indonesia) BOKU Bantuan Orang Kurang Upaya (Malaysia) BOT Bantuan Orang Tua (Malaysia) BPJS Badan Penyelenggara Jaminan Sosial (Indonesia) BPN Bantuan Prihatin Nasional (Malaysia) BPR Bantuan Prihatin Rakyat (Malaysia) BPS Business pulse survey BPUM Bantuan Produktif Usaha Mikro (Indonesia) BSH Bantuan Sara Hidup (Household Living Aid; Malaysia) BSP Bangko Sentral ng Pilipinas (Philippines) BST Bantuan Sosial Tunai (Indonesia) BSU Bantuan Subsidi Upah (Indonesia) CAMP COVID–19 Adjustment Measures Program (Philippines) CARES COVID–19 Assistance to Restart Enterprises (Philippines) CCT Conditional Cash Transfer CGCC Credit Guarantee Corporation of Cambodia CMP Child Money Program (Mongolia) COA Commission on Audit (Philippines) CRM Customer Relationship Management CSSE Center for Systems Science and Engineering CTPPT Comprehensive and Progressive Agreement for Trans–Pacific Partnership DHS Davis, Haltiwanger, and Schuh DOLE The Department of Labor and Employment DOLE AKAP–The Department of Labor and Employment’s Abot Kamay ang Pagtulong (Philippines) DSWD Department of Social Welfare and Development (Philippines) xi xii CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA DTI Department of Trade and Industry (Philippines) DTKS Data Terpadu Kesejahteraan Sosial (Indonesia) EAP East Asia and Pacific EC European Commission EIS Employment Insurance System EMIs Electronic Money Issuers EPF Employees’ Provident Fund (Philippines) ERP Enterprise Resource Planning ESA Emergency Shelter Assistance (Philippines) ESCS Economic, Social, and Cultural Status ESP Emergency Subsidy Program FNI Food and Non–Food Items FSRF Financial Subsidy to Rice Farmers (Philippines) GCI Global Competitiveness Index GDP Gross Domestic Product GIS Geographic Information Systems GNI Gross National Income GSO General Statistics Office (Vietnam) HFPS High–Frequency Phone Surveys HISBA Household Income and Basic Amenities Survey (Malaysia) HRDF Human Resources Development Fund ICT Information and Communications Technology IDPoor Identification of Poor Households (Cambodia) IMF International Monetary Fund IT Information Technology JKP Jaminan Kehilangan Pekerjaan (Indonesia) KHR Cambodian Riel KUR Kredit Usaha Rakyat (Indonesia) LGUs Local Government Units (Philippines) LFS Labor Force Surveys LSMS+ Living Standards Measurement Study Plus MNT Mongolian Tugrik MSMEs Micro, Small, and Medium-Size Enterprises NEDA National Economic Development Authority NIS National Institute of Statistics NPLs Non–Performing Loans NSO National Statistical Office OECD Organization for Economic Cooperation and Development OEVS Occupational, Employment and Vacancy Survey OFW Overseas Filipino Workers OWWA Overseas Workers Welfare Administration ABBREVIATIONS xiii OxCGRT Oxford COVID–19 Government Response Tracker PISA Program for International Student Assessment PKH Program Keluarga Harapan (Family Hope Program–Indonesia) PKTD Padat Karya Tunai Desa (Indonesia) PSA Philippine Statistics Authority RCEP Regional Comprehensive Economic Partnership RFAP Rice Farmers Financial Assistance Program (Philippines) SAP Social Amelioration Program (Philippines) SBWS Small Business Wage Subsidy (Philippines) SMEs Small and Medium–Sized Enterprises SMS Short Message Service SOCSO Social Security Organization (Malaysia) SRM Supplier Relationship Management SSCs Social Security Contributions SSS Social Security System (Philippines) STUFAP Student Financial Assistance Programs (Philippines) TESDA The Technical Education and Skills Development Authority (Philippines) TUPAD Tulong Panghanapbuhay sa Ating Disadvantaged/Displaced Workers (Philippines) TVET Technical and Vocational Education and Training UCT Unconditional Cash Transfer program (Philippines) UNCTAD United Nations Conference on Trade and Development UN ESCAP United Nations Economic and Social Commission for Asia and the Pacific VAT Value–Added Tax (Mongolia) Acknowledgements This report is the product of a collaboration between the World Bank’s Finance, Competitiveness and Innovation and Poverty and Equity units for East Asia and the Pacific region. It was led by Asya Akhlaque (until December 2021), Alvaro Gonzalez (since December 2021), and Maria Ana Lugo under the general direction of Hassan Zaman, Rinku Murgai, Zafer Mustafaoglu, and Cecile Thioro Niang. The report team included Bayu Agnimaruto, Zainab Ali Ahmad, Rabia Ali, Edgar Avalos, Anh Thi Bao Tran, Aufa Doarest, England Rhys Can, Irene Jo Estigoy Arzadon, Ade Febriady, Pablo Enrique Fleiss Weinberger, Jaime Frias, Nadia Belhaj Hassine Belghith, Sarah Hebous, Nona Karalashvili, Wendy Karamba, Lydia Kim, Smita Kuriakose, Karl Robert Lasmarias Jandoc, Kyung Min Lee, Ririn Salwa Purnamasari, Isabelle Salcher, Francesco Strobbe, Shawn W. Tan, Kimsun Tong, Jesica Torres Coronado, Trang Thu Tran, Ikuko Uochi, Matthew Wai-Poi, Judy Yang, Tatiana Didier Brandao, and Rekha Reddy. Barbara Karni provided editing services. Mildred Gonsalvez, Susana Rey, and Marinella Yadao provided excellent assistance throughout the process. The team is grateful for the guidance and advice on the overall report from Mary Hallward-Driemeier, Ruth Hill, Andrew Mason, Denis Medvedev, Phillipe De Meneval, Philip O’Keefe, Sharon Faye Alariao Piza, Tobias Pfutze, Habib Rab, and Jonathan Timmis. The report was edited and typeset by Circle Graphics, Inc., Reisterstown, MD. xv Reader’s Guide This edited volume examines the economic consequences and government responses to the COVID-19 pandemic in six East Asian economies: Cambodia, Indonesia, Malaysia, Mongolia, the Philippines, and Vietnam. These countries, while geographically close to each other, experienced the crisis differently and implemented a variety of measures to contain the spread of the virus and mitigate economic impacts with varying success. The COVID-19 pandemic began as a health crisis that plunged these six countries into a sudden and severe economic downturn. The virus, which was first reported in Wuhan, China, in December of 2019 and declared a global pandemic by the World Health Organization in March of 2020, has significantly affected global health, economies, and daily life. To combat the spread of the virus, countries implemented measures such as lockdowns, social distancing, mask mandates, and travel restrictions. However, these measures also led to a sudden and widespread disruption of economic activity. Economic output declined sharply due to falling demand and supply chain disruptions, leading to job losses and business closures, particularly in the travel, tourism, and hospitality industries. The pandemic disproportionately impacted low-income workers and those in the informal sector, exacerbating existing economic inequalities and exposing weaknesses in economic systems. The report analyzes how the six East Asian economies coped with the COVID-19 crisis and how their governments responded. Using rapid surveys of enterprises and households conducted by the World Bank, the chapters in the report focus on the fate of businesses and households in each country. The Introduction provides background information on all six countries, while each country chapter (chapters 1 through 7) tells its own story and can be read independently for a more in-depth understanding of each particular case. Chapters follow a similar structure and use similar indicators to facilitate comparisons across the countries. Chapter 8 offers a comparative analysis of each government’s responses to the economic crisis. It is important to note that the long-term economic impact of the pandemic is still uncertain. These country studies offer an initial glimpse into how the global shock has, or has not, changed how businesses operate and people work, and may serve as a starting point for further understanding the pandemic’s far- reaching consequences. xvii Introduction by Alvaro Gonzalez, Lydia Kim, and Maria Ana Lugo More than three years after the first COVID-19 case was discovered in the East Asia and Pacific (EAP) region, it is time to take stock of the lasting effects—and opportunities—of the pandemic and identify which policies may have helped stem the economic losses suffered by households and firms. To do so, this regional report examines the economic impact of the COVID-19 pandemic on households and firms in six countries: Cambodia, Indonesia, Malaysia, Mongolia, the Philippines, and Vietnam. This edited volume examines: (a) the links between impacts on firms and households, in particular through the employment channel, and (b) governments’ fiscal responses to the COVID crisis, through transfers, subsidies, and taxes.1 It identifies and explains changes in household well-being by examining the economic effects of the pandemic on labor markets. As the source of employment and wage income, businesses have a direct role in determining jobs and earnings, and, indirectly, welfare, poverty, and inequality. When faced with a shock, firms responded by adjusting employment, reducing wages, increasing prices, and reducing services provided. All of these channels directly affected households’ wellbeing. For this reason, the report focuses on firms in addition to households. Governments responded through various instruments, providing transfers and subsidies and lowering the tax burden to both households and firms. A schematic representation of the links between households, firms, and governments analyzed in the report is presented in Figure I.1. The volume presents findings on the impacts of the crisis on six economies, as well as government responses to it, highlighting the specific characteristics of each of the economies studied and the challenges that emerged as countries embarked on recovery. The individual chapters reveal what governments did and offer insights into where policy responses may need to go next, as countries look toward recovery or a “new normal” of living with the pandemic. This volume will first introduce the regional trends and the six country-specific characteristics and impacts. The objective is to both set the stage for the country chapters that follow and to highlight some of their findings related to businesses and households. 1  The distinction between firms and households is not always apparent, particularly in low-income settings. Self-employment and small family firms are prevalent in developing countries, not only in rural agricultural activities but also in petty trade and other activities. The household is often a family firm, making production and consumption decisions simultaneously. The framework adopted in this report is therefore merely a schematic representation of channels of impacts. The analysis of some countries, particularly Cambodia and Indonesia, emphasizes this continuum, combining information from business pulse surveys and High-Frequency Phone Surveys (HFPS). Although the performance indicators are not entirely comparable, they yield a more comprehensive economic impact of the pandemic across different types of establishments. 1 2 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.1  Report framework: Impact of the COVID-19 shock on households and firms External REST OF WORLD Remittances Demand Demand Employment Supply FIRMS Wages Formal HOUSEHOLDS Informal Prices Family Finance Services Uncertainty Transfer/ Transfer/ Subsidies Subsidies GOVERNMENT Taxes Taxes Source: World Bank staff elaboration. The rest of the volume is organized as follows: Chapters 1–6 provide detailed analyses of the effect of the pandemic, the individual governments’ responses, and lessons learned in each of the case study countries. Chapter 7 provides a comparative perspective of the scale and composition of governments’ responses and summarizes the lessons learned from their experience, in order to help policy makers—within the region and elsewhere—improve their policies going forward. The report is based on a rich set of high-frequency household and firm monitoring data collected by the World Bank Group by telephone in multiple rounds, beginning in May 2020. The surveys administered a standardized questionnaire to collect timely information on the effects of the pandemic on businesses and households. See Box I.1 for a description of the surveys, and Annex IA for survey dates and sampling methods. Pre-Pandemic Context Economic crises expose and exacerbate weaknesses in economies. Established strengths help economies recover. The nature and type of crises associated with the COVID-19 pandemic were quite unique. In early 2020, the COVID-19 outbreak quickly became a global pandemic, and countries around the world imposed border closures and national lockdown measures to contain its spread. The containment measures had a profound economic effect by restricting travel, disrupting trade, and limiting the economic activity and provision of services that required face-to-face interactions. INTRODUCTION 3 BOX I.1  Household and firm survey data The business pulse survey was implemented in all of the countries included in this study except Mongolia. In most countries, interviews were conducted over the phone or administered online. The samples covered formal micro, small, medium-size, and large businesses across all main sectors: manufacturing, retail, and other services. Data for Mongolia was collected as a follow-up of the World Bank enterprise survey, using a questionnaire that excluded some questions from the standard version of the business pulse survey; these interviews were conducted face-to-face. The business pulse survey and the enterprise survey were designed to provide up-to-date information on how firms were coping with the pandemic-induced recession. As the surveys sampled formal firms, they describe a particular part of the economy that may not be representative of the whole, particularly in countries in which informality is widespread. The business surveys focused on five channels of impact: lockdown effects, changes in supply, changes in demand, financial disruptions, and uncertainty. They also collected information on firms’ adjustment strategies and access to public support programs (Apedo-Amah and colleagues, 2020). Similar to the efforts related to firms, in collaboration with national statistics offices and development partners, country teams designed and implemented high-frequency phone surveys (HFPS) of households. These surveys collected information on employment, income, food insecurity, and access to public services. They also provided near real-time information on individual behavior changes, or coping mechanisms, and the reach of government relief measures in response to the pandemic. Integrating the analysis of households and firms provides insights that can inform policy design and implementation. HFPS were implemented across all regions in the world. In EAP, teams across 11 countries implemented a total of 37 surveys. This report covers six countries, for which there were multiple rounds of both business and household surveys. Household surveys differed slightly across countries. In some cases, the sample was drawn from previous household surveys implemented by the countries’ national statistics offices. In other countries, samples were taken from a registry of telephones. Surveys also differed in terms of the representativeness of the type of workers. In Indonesia, for instance, the sample included only the main household breadwinner. (See Annex IA for more details). Several factors likely determined the severity of the recession and the speed with which countries recovered. These include a country’s economic structure; the quality of its regulations and institutions; the depth, breadth, and health of its financial markets; and its business environment. For this pandemic-induced recession, ongoing economic recovery is also a function of the number, severity, and duration of waves of the virus; the quality of health care systems; the strictness, duration, and severity of lockdown measures; and the fiscal and monetary response to mitigate the impact of the virus. This section examines pre-COVID economic trends and conditions throughout the EAP region.2 It highlights conditions in the six case study countries. 2  The analysis excluded high-income countries in the region. The countries covered include Cambodia, China, Fiji, Indonesia, Kiribati, the Lao People’s Democratic Republic (Lao PDR), Malaysia, the Marshall Islands, Micronesia, Mongolia, Myanmar, Nauru, Palau, Papua New Guinea, the Philippines, Samoa, the Solomon Islands, Thailand, Timor-Leste, Tonga, Tuvalu, Vanuatu, and Vietnam. 4 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Economic performance Employment and growth Before the pandemic, EAP was a relatively high-growth region. Between 2000 and 2019, regional per capita gross domestic product (GDP) grew at an average annual rate of 3.7 percent. In the two years before the pandemic, the region grew at an average annual rate of 4.4 percent. The selected country cases had similar, but above average, rates of economic growth for the region. Cambodia entered the pandemic with strong economic and employment growth. Annual economic growth exceeded 7 percent in both 2018 and 2019, thanks to continued export growth and strong construction activity. Between 2010 and 2019, the economy added more than 1 million jobs, and employment grew by almost 15 percent. However, deterioration of external conditions even before the pandemic hit was expected to lead to a slowdown in growth to below 7 percent in 2020, indicating an economic outlook with significant risks. Despite external headwinds, including capital flow reversals, Indonesia grew more than 5 percent in 2018, with robust domestic demand offsetting the decline in net exports. GDP growth was expected to rise to the level of potential growth, 5.3 percent, over the medium term, underpinned by strong domestic demand. Indonesia added over 23 million jobs between 2010 and 2019, an increase of over 21 percent. Before the pandemic, Malaysia’s economy was stable and growing. Annual growth averaged almost 5 percent in the previous five years, raising per capita income. As a result of this strong performance, employment grew nearly 30 percent between 2010 and 2019. Mongolia’s economic outcomes were tied closely to its exports to China, particularly coal and copper ore. Annual economic growth rose from approximately 1.5 percent in 2016 to 7.7 percent in 2018, driven primarily by the mining sector. Although pre-pandemic growth remained robust, there were significant risks with respect to global commodity demand, given Mongolia’s dependence on trade. Between 2010 and 2019, Mongolia’s rapid economic growth resulted in a nearly 16 percent increase in employment. The Philippines was on a steady trajectory toward improving the management of its economy. Prior to 2020, the Philippines’ economy performed well and was in a favorable position to address its socioeconomic challenges. Reforms during successive administrations delivered sustained growth, low inflation, financial stability, and strong external and fiscal positions; however, levels of poverty and inequality remained high. Additionally, although the economy may not have created enough jobs for its young workforce, it did create nearly 7 million new employment opportunities between 2010 and 2019, an increase of almost 19 percent. Economic growth in Vietnam reached a 10-year high of 7.1 percent in 2018. The economy remained resilient in the face of rising trade tensions and growth, and the expansion was broad-based. The income and consumption habits of the growing and urbanizing middle class were strong, and the agricultural and manufacturing sectors were surging. This strong economic momentum continued into 2019, aided by competitive labor costs and solid fundamentals, including a diversified trade structure and new free trade agreements, which spurred reforms. Between 2010 and 2019, the number of new jobs rose by over 4 million people, an 8.2 percent growth rate, the lowest rate of job creation among the six economies studied in this report. Before the pandemic, despite relatively low revenue mobilization, compared to other developing countries, the six economies were in reasonably strong positions by international standards. The expectation was, therefore, that these economies would fare better than other countries and emerge more quickly from INTRODUCTION 5 the pandemic-induced recession. Generally, they did so, although there have been significant variances across the six economies. These differences may reflect institutional endowments, the predominance of certain sectors, and the policies and programs that the public sector put in place to address the damage wrought by the pandemic and ensuing recession. Structure of economies The effect of the pandemic-induced recession differed across countries. Economies dominated by essential industries that required low levels of interpersonal interactions, such as many forms of manufacturing, suffered less than economies dominated by sectors that required a high level of interpersonal interaction, such as tourism, restaurants, and hospitality. As lockdowns were imposed across the globe, economies more exposed to trade and tourism suffered more than countries that were less exposed. Seven country-level indicators were used to describe the six economies and the regional average for all developing EAP countries, excluding high-income EAP countries (Table I.1). They were selected to jointly reflect the level of development, characteristics of employment, economic structure, and Internet usage. Some of these indicators were chosen in hindsight. For example, the Internet and digital technologies played a prominent part in how households and businesses coped with the lockdowns and the recession. This indicator was included because the importance of digital technology became apparent when looking at the survey data collected during the pandemic. Tourism and hospitality were highlighted because they were the sectors most affected by the lockdowns. Countries more reliant on global trade were more exposed to the immediate effects of the global contraction, although more diversified exports made them less vulnerable. Additionally, employment indicators reflect individual reliance on activities that were affected by the COVID crisis. TABLE I.1  Select pre-pandemic country-level characteristics, circa 2019 East Asia Cambodia Vietnam Philippines Indonesia Mongolia Malaysia and Pacific GDP per capita (constant 2015 US$) 1,297 2,331 3,335 3,598 4,073 10,682 7,280 Percent of employment that is informal 89.4 69.7 — 80.5 44.1 16.8–30.5* — Percent of employment in agriculture 37.8 40.4 25.7 30.8 27.9 11.2 29.0 Percent of employment in trade, transportation, hospitality, 27.0 25.4 40.1 33.7 28.2 44.1 — and business services Trade as percent of GDP 125.6 196.7 66.0 39.8 111.5 129.0 45.6 Export diversification index (lower values 5 more diversified) 0.77 0.55 0.54 0.56 0.85 0.44 — Percent of population using the Internet 27.8 58.9 41.0 33.5 33.3 79.1 51.6 Gross government debt as percent of GDP 28.5 42.9 38.6 30.4 66.1 57.2 — Regulatory capital to risk-weighted assets 21.8 11.8 15.2 21.6 — 18.6 — Non-performing loans to gross loans 1.6 1.8 2.0 2.3 — 1.5 — Return on assets 1.7 1.2 1.5 2.5 — 1.5 — Sources: ILO, 2019; UNCTAD 2019; IMF Fiscal Monitor Database and IMF Financial Soundness Indicators 2019. Other statistics are from the World Bank’s World Development Indicators, 2022. Note: The regional averages are unweighted and exclude high-income economies. Diversification index indicates to what extent the structure of exports by product of a given economy differs from the world pattern. — indicates that data are not available. For Malaysia, the National Statistical Office (NSO) reported an informal employment rate of 16.8%, but alternative estimates based on social insurance coverage suggested a rate of about 30.5%. 6 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Malaysia was the wealthiest country of the six in this study. It relied heavily on trade, with a highly diversified exports portfolio, and a large share of its employment was engaged in tourism and trade activities, sectors that were hardest hit by the pandemic. At the same time, Malaysia’s high Internet penetration helped businesses and households to navigate health and economic disruptions. At the other extreme, the two poorest countries of the six, Cambodia and Vietnam, were heavily exposed to economic events in China, because of their reliance on trading partnerships. In 2019, about 15 percent of the value of Vietnam’s output was dependent on Chinese capital and intermediate goods, and 17 percent of its economy relied on demand from China. On the positive side, more than a third of workers in the two countries worked in agriculture, a sector that was less affected by the pandemic than other sectors. One important difference between the two countries was the level of diversification of their exports. Vietnam’s exports were relatively well diversified and recorded a strong expansion of product diversification in the past decades (WTO-ITC-UNCTAD 2017). On the other hand, Cambodia was in its initial phases of diversification. In between these two groups of countries lie the Philippines, Indonesia, and Mongolia. The Philippines was highly reliant on tourism and related activities, in terms of both employment—40 percent of total employment—and contribution to GDP, at 8.7 percent of GDP, according to the World Trade and Travel Council 2022, second only to Cambodia. Indonesia was less dependent on the rest of the world for trade and tourism than many other countries, but a third of its labor force worked in the service economy—which includes trades, hospitality, and transportation—where mobility restrictions had severe impacts. Combined with high levels of informality, this feature of the economy made workers more vulnerable as a result of the COVID recession. Mongolia’s economy was closely tied with that of China and potentially vulnerable to declines in demand in key commodities, particularly in the mining sector, the backbone of the Mongolian economy. At the same time, the relatively high level of formalization of its labor force meant that the system was better prepared to deal with labor shocks through standard labor market instruments. Business environment The quality of an economy’s business environment affects how well firms can absorb and adjust to recessions and respond to the opportunities of an impending recovery. A look at the quality of the business environment across the six economies seemed to indicate that Malaysia may have had the best business environment among the six, while Cambodia and Mongolia had the least favorable environments. In sum, a more favorable business environment better prepares firms to adjust to shocks from a recession and opportunities during recovery. For example, higher levels of competition condition firms to respond to competitive threats and move quickly to exploit opportunities before competitors do. Hence, as a result of the better business environment, firms in Malaysia were better prepared to deal with the pandemic-induced disruption and take advantage of the available opportunities during the recovery. To assess the quality of the pre-pandemic business environment, indicators from the World Economic Forum’s (WEF) Global Competitiveness Index (GCI) were used. Competition, domestic and foreign, is a key measure of the quality of an economy’s business environment. For that reason, competition across the six East Asian economies was the first indicator examined. Figure I.2 displays a breakdown of some pre- pandemic measures of competition—one component was domestic market competition and the other was INTRODUCTION 7 FIGURE I.2  WEF Global Competitiveness Index 2019 – Select indicators of business environment 100 90 80 70 60 50 40 30 20 10 0 Domestic competition Trade openess Labor flexibility Ease of entry/exit Institutions Macroeconomic stability East Asia and Pacific Cambodia Indonesia Malaysia Mongolia Philippines Vietnam Source: WEF Global Competitiveness Index 2019 Note: Scores range from 0 (worst) to 100 (best) foreign competition. On the domestic side, the indicator summarizes whether product market regulation and other rules tilted the playing field in favor of some domestic competitors, thwarting and distorting competition. In addition, to the extent that market concentration distorted competition, these measures were also used to suggest where markets were less competitive. The indicator of trade openness, in turn, measures the level of import tariffs and non-tariff barriers, the complexity of the tariff code, and the efficiency of customs clearance. A more open, transparent, and less complex trade regime would also lead to more nimble markets and agile firms. Figure I.2 also shows WEF indicators that measure how well economic resources can move across markets. For example, the indicator of labor market flexibility summarizes the ease of hiring and firing and the non-wage costs associated with it, wage flexibility, quality of employer-employee relations, workers’ rights, the degree of worker spatial mobility, and availability of active labor maker policies.3 Lastly, ease of entry and exit by firms captures the administrative requirements related to the costs and time to start a business, as well as to the insolvency framework and recovery rates of insolvent firms. Private sector dynamism also relies on the quality of institutions. As the set of rules and norms in a society, institutions help markets work, as they determine the constraints and incentives faced by agents.4 The WEF indicator on institutional quality is related to public sector performance and transparency, strength of property rights and checks and balances, future orientation of government, security, and social capital. Finally, a solid macroeconomic framework is crucial to provide the predictability and stability required for businesses to operate, as captured by the second indicator, in terms of low inflation and debt dynamics. The conclusion that the Malaysian economy had the most favorable business environment was based on several indicators. As seen in Figure I.2, for all indicators, Malaysia performed better than the regional East  For a discussion on labor market flexibility and its relation to firm performance, see Davis and Haltiwanger (2014). 3  See, for instance, North (1990) and Acemoglu, Johnson, and Robinson (2005) 4 8 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Asia average. Indonesia also performed higher than the East Asia average—which included all countries of East Asia, in addition to the 6 countries that were the focus of this study—in entry and exit regulation and macroeconomic stability. The Philippines did well with respect to macroeconomic stability but fell below the regional average on all other indicators. The other six economies had an overall business environment no better than the regional average. Cambodia and Mongolia had the least favorable business environments among the six economies. Development of the financial market Credit markets—which are central to the process of efficiently allocating resources—were growing in the region before the pandemic. Because financial markets buffer shocks, countries that lack financial markets and institutions that are widely accessible, deep, and efficient may experience longer and deeper recessions than countries with better-developed financial markets. The International Monetary Fund (IMF) developed an index that assesses financial institutions and financial markets in terms of depth, as in size and liquidity; access, or the ability of individuals and enterprises to access financial services; and efficiency— the ability of financial institutions to provide financial services at low cost, with sustainable revenues, and the level of activity in capital markets. The index covers 180 countries, including the six covered in this study. The data for 2008–19 suggests that Malaysia was the most developed financial market among the six, and Mongolia the least developed (Figure I.3). According to this index, the level of financial development in Vietnam, Cambodia, Indonesia, and the Philippines was converging on a point in between the two extremes. Even in well-developed financial markets, many small businesses will emerge from the pandemic less financially stable than they were in 2019, and some will have to close. Better-developed financial systems will be better able to help revive firms that are viable but in need of a fresh injection of capital, expedite the restructuring of businesses that were deeply disrupted yet still viable, and close firms that are unviable. FIGURE I.3  Financial development index in six countries in East Asia, 2009–19 5.5 4.5 3.5 Index 2.5 1.5 0.5 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Cambodia Indonesia Malaysia Mongolia Philippines Vietnam Source: IMF Financial Development Index database, accessed June 24, 2022. INTRODUCTION 9 Relative to other countries in the region, the financial sector in Malaysia was more likely to have helped absorb more of the shock of the recession and may help in the recovery. The financial sectors in the other countries were shallower and thus less able to help firms emerge from the crisis. Fiscal space Countries with more fiscal space were thought to be better able to react to the economic crisis, through more aggressive fiscal stimulus, without jeopardizing long-term debt sustainability or market access (Kose and colleagues, 2017). While there are many measures of fiscal space, government gross debt as a proportion of GDP is one of the most cited ones, and thus presented here. Based on this measure of fiscal space, Mongolia entered the COVID-19 pandemic in the most precarious situation. In 2017, gross government debt as a percent of GDP had peaked at just over 100 percent and fell to 66.1 percent by 2019 (Table I.1). No other country out of the six breached the 60 percent mark. Malaysia consistently kept gross government debt at around 55 percent of GDP from 2015 through 2019. Cambodia and Indonesia were consistently below 30 percent, and Vietnam and the Philippines were 37 and 45 percent of GDP, respectively. Poverty and inequality Record-level poverty reduction Over the past three decades, the EAP region has made extraordinary progress in reducing poverty, nearly eliminating extreme poverty. The share of the population living on less than the international poverty line of US$1.90 per day, in 2011 purchasing power parity, plummeted from 61 percent in 1990 to about 1 percent in 2019, according to the World Bank’s Poverty and Inequality Platform (2022). At the US$3.20 per day poverty line, the rate fell sharply from 85.0 percent in 1990 to 6.3 percent in 2019. At both poverty lines, the region now has lower poverty rates than Latin America and the Caribbean and the Middle East and North Africa and is making its way toward high-income standards of poverty. Among the countries included in the report, Indonesia and the Philippines had the highest headcount rates, with about 20 percent of the population living on less than US$3.20 per day (Figure I.4, panel a). Given their relatively large populations, the two countries accounted for half of the people living on less than US$3.20 per day in the six countries (Figure I.4, panel b). Mongolia, whose per capita gross national income (GNI) is similar to that of Indonesia and the Philippines, had significantly lower poverty rates, reflecting the more equal distribution of incomes there. In Vietnam, just 7 percent of the population lived on $3.20/day or less, despite much lower per capita GNI than Indonesia and the Philippines. In Malaysia, the only upper- middle income country in the group, there were virtually no people living below this threshold. Economic growth has been the main driving force behind poverty reduction in most EAP economies. Analysis of long-term relationships between growth and poverty shows EAP countries, except for most Pacific Island countries, experienced near- or above-average reductions in poverty for their level of growth. Labor market shifts away from agriculture into more productive nonagricultural sectors have likely been a driving factor in the translation of growth into poverty reduction. In the Philippines and Vietnam, for example, household labor incomes increased as workers shifted away from agriculture and into nonagricultural wage activities, and labor productivity rose (World Bank 2022b; forthcoming a). Combined with the decline in dependency ratios, this shift caused poverty rates to fall. 10 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.4  Poverty in East Asia and Pacific 2018 a. Poverty headcount rate b. Percent of extremely poor and poor population 60 100 50 80 Share of poor (%) 40 60 Poverty rate (%) 30 40 20 20 10 0 $1.90 $3.20 1 Indonesia Philippines Mongolia Vietnam Other EAP Malaysia Indonesia Other EAP Philippines $1.90 $3.20 $5.50 Vietnam Mongolia Malaysia Source: World Bank 2021; World Bank Poverty and Inequality Platform, 2022. Note: Poverty lines are in 2011 PPP. Poverty rates may vary according to the year of the household survey. Data for Cambodia is not available. In panel a, “Other EAP” is the population-weighted average for China, the Lao People’s Democratic Republic (PDR), Myanmar, Thailand, and Timor-Leste. In panel b, poverty rates are lined up to 2018 based on growth projections. Despite this shift, agriculture remained an important source of employment in the region. Almost 40 percent of workers in Cambodia and Vietnam and a quarter of workers in Indonesia, the Philippines, and Mongolia were attached to the sector (see Table I.1). Agricultural workers were more likely than the average worker to be poor. Agriculture may have served as a refuge activity for the poor during COVID-19 restrictions, which primarily affected urban, nonagricultural activities, as the country chapters show. Inequality as a growing concern Economic transformation and high rates of growth were accompanied by rising levels of inequality in many EAP countries in the early 2000s (World Bank 2021b, 2022a). In several countries in the region, however, inequality declined between 2010 and 2019, although levels remained relatively high (Figure I.5, panel a). Inequality in EAP is significantly lower than in Latin America and the Caribbean and in Sub-Saharan Africa, although it is higher than in Eastern Europe and Central Asia and in South Asia. Achieving shared prosperity emerged as an important goal in several EAP countries in recent years. In the Philippines, despite declining inequality in recent years, concerns over inequality have risen, and support for redistribution is increasing (World Bank forthcoming a). In Malaysia, inequality had declined consistently since the 1980s, unlike many other countries in the region, but the Gini index remained high for an aspiring high-income country (Record and colleagues, 2021). In Vietnam, where the Gini index was below the global average, the share of consumption attributed to the richest 10 percent of the population increased substantially since 2014 (World Bank 2022b). There is room for more active inequality-reducing policies, as redistribution through fiscal policy remained limited in the six countries, except for Mongolia (Figure I.5, panel b). The reduction in the Gini coefficient associated with fiscal policy in the Philippines and Indonesia was smaller than in similar countries at similar levels of development, partly because social assistance was less generous than in other INTRODUCTION 11 FIGURE I.5  Inequality in East Asia and Pacific a. Gini index, 2000–18 b. Reduction in Gini coe cient through fiscal policy 55 20 18 50 16 45 14 Redistributive e ect 12 Gini index 40 10 8 35 6 Mongolia 30 4 2 Indonesia Philippines 25 China 0 Myanmar Philippines (i) Malaysia (i) Indonesia (c) Vietnam (c) Mongolia (c) EAP 3 3.5 4 4.5 5 5.5 Circa 2000 Circa 2010 Circa 2018 Log(per capita GDP) in 2017 Source: World Bank Poverty and Inequality Platform. Source: World Development Indicators; Fuchs, Sosa, and Wai-Poi, 2021. Note: The figure for EAP is an unweighted average for 12 developing countries. The welfare aggregate used in Note: The redistributive effect is the gap between the Gini coefficient of disposable income each country to measure poverty is indicated in parentheses. “i” means that the welfare aggregate used is income, and the Gini coefficient of market income, where disposable income equals market income and “c” refers to consumption. plus direct transfers and minus direct taxes. middle-income regions, such as Latin America and Eastern Europe. In addition, relatively low reliance on direct taxes, for example, on property, assets, and income, limited the role that fiscal policy could play in reducing the Gini coefficient. Social protection systems Spending on social assistance in East Asia and Pacific tended to be lower than in other regions. This was in part because of a traditional preference of countries in the regions for development-oriented programs with a “productivist” emphasis; meaning, social protection was intended to serve broad economic goals (O’Keefe and colleagues, forthcoming) rather than the generous welfare states found in many countries in the Organization for Economic Cooperation and Development (OECD). This preference was reflected in low levels of spending as a share of GDP. Based on the latest available information, developing countries in the EAP spent, on average, about 1.3 percent of GDP on social assistance—less than the 1.9 percent allocated in Europe and Central Asia, and the 1.8 percent allocated in Latin American and the Caribbean. As a result, except for a few countries, including Mongolia, benefits of social assistance were more modest, safety net coverage was weaker, and the impact on poverty reduction was more limited than in other regions. Yet, social protection systems have evolved in the past decades in many countries, from rudimentary delivery systems to sophisticated digitalized platforms. Considering the scale and diversity of instruments in the case of social assistance and the coverage of social insurance, social protection systems in the region can be divided into three groups, based on their maturity (Figure I.6). Cambodia has nascent or incipient social assistance and insurance systems, with relatively low spending and narrow coverage. Mongolia has well-established social protection systems, with a long legacy of assistance programs and large coverage in both assistance and insurance systems. Indonesia, Malaysia, the Philippines, and Vietnam fall in between, 12 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.6  Typology of social protection systems in East Asia and the Pacific Nascent Emerging Established Cambodia, Myanmar, Indonesia, Malaysia, Lao PDR, Tonga, Social Philippines, Vietnam, Solomon Islands, Papua Mongolia, China Assistance New Guinea, Samoa, Thailand, Timor-Leste, Fiji Vanuatu, Kiribati Cambodia, Indonesia, Philippines, Thailand, Social Myanmar, Lao PDR, Mongolia, Vietnam, Solomon Islands, Insurance Tonga, Papua New Samoa, Vanuatu Malaysia, China, Fiji Guinea Source: O’Keefe and colleagues (forthcoming). with social programs created and expanded as a response to the Asia financial crisis in 1997/98 and the 2008 global crisis. Yet, some variation in maturity is evident across countries within this group. Malaysia made strides in recent years that have put it closer to the established category; Vietnam is at the lower end of the spectrum, although not considered nascent. The capacity of safety nets to respond to the pandemic depended partly on the status of the social assistance system before the pandemic hit. In Mongolia, where coverage of the system was broad before the pandemic, emergency packages were among the largest in the world (see chapter 4). In Cambodia, where the system was small before the pandemic, efforts remained limited, even if well targeted to the structurally poor (see chapter 1). Labor programs are relatively underdeveloped in the region. Spending tends to be very low and often available only to people with formal sector jobs, who represent a small share of all workers in many of the regions’ countries. According to some estimates, only about 3 percent of the labor force across East Asia and South Asia has access to active labor market policies, well below the share in Latin America and in Eastern Europe and Central Asia (O’Keefe and colleagues, forthcoming). Unemployment insurance is extremely limited in four of the six countries; it does not exist in Cambodia or Indonesia. The Economic Impact of the Pandemic and Government Efforts to Mitigate It The pandemic has taken a comparatively modest toll on human health in EAP, with the region recording proportionally fewer cases and deaths than other regions (Figure I.7, panel a). Many countries in the region were effective in preventing large-scale COVID-19 outbreaks (Figure 1.7, panel b). By the end of 2021, INTRODUCTION 13 FIGURE I.7  Number of reported COVID-19 cases, by region, January 2020–January 2022 a. By region b. By country in East Asia 300 300 COVID-19 cases (per 100,000 people) COVID-19 cases (per 100,000 people) 200 200 100 100 0 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 East Asia & Pacific (middle income) East Asia & Pacific (high income) Cambodia Indonesia Malaysia Europe & Central Asia Latin America & Caribbean Mongolia Philippines Vietnam Middle East & North Africa United States & Canada South Asia Sub-Saharan Africa Source: Our World in Data, accessed May 2022. Note: Figures are seven-day rolling averages. developing EAP countries had recorded an average of about 3,100 cases and 36 deaths per 100,000 people, far fewer than the global average of 6,500 cases and 98 deaths, according to data from Our World in Data.5 The pandemic unfolded at different paces and levels of severity across countries within the region, but three phases can be distinguished: The initial shock, in which to contain COVID surges, most countries imposed lockdowns, border closures and stay-at-home measures in the second quarter of 2020, which led to a sharp reduction in mobility, even when cases were low. A second phase of recovery ensued, with a gradual progression to normalcy, as restrictions were lifted, and economic activity resumed. Finally, in a third phase of resurgence, new COVID outbreaks emerged, starting in late 2020. Restrictions were reinstated, but mobility did not decline as much as in the initial phase, representing both pandemic fatigue and the rollout of vaccines. All countries experienced these three phases of the pandemic, but the timing and duration of each phase varied significantly across countries. In Indonesia and the Philippines, COVID cases increased steadily early on, and both countries sustained prolonged health crises earlier than other countries in the region. Malaysia and Mongolia experienced increases in infections late in 2020, followed by some of the largest spikes in the region. Cambodia and Vietnam managed to avoid any outbreaks for an extended period, until undergoing a surge in cases in early to mid-2021. 5  Underreporting of COVID–related deaths may be significant in some countries. For instance, COVID-related deaths in Malaysia and Mongolia followed a similar evolution and levels than excess deaths (measured as the difference between the death rate observed in 2020 and 2021 and the death rate trends in pre-pandemic years). Conversely, COVID-related deaths and excess deaths rates vary substantially in Philippines (see The Economist’s excess death tracker at: https://www.economist.com/graphic-detail/coronavirus-excess-deaths-tracker). Excess death data is not available for the other three countries in the report. 14 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA The relative success of the EAP region in preventing more severe COVID outbreaks may stem from its early and effective enforcement of containment measures and sustained compliance. The stringency index developed by the Blavatnik School of Government of Oxford University quantifies the strictness of lockdown style policies by country (2022). On average, both middle- and high-income EAP countries responded faster than most other countries in the world (Figure I.8, panel a). FIGURE I.8  Stringency and mobility indexes, by region, January 2010–January 2022 a. EAP region and rest of the world Stringency index Mobility index 80 90 60 Mobility index (14-day average) 60 Stringency index (14-day average) 30 40 0 20 –30 –60 0 Jan-20 Jul-20 Jan-21 Jul-21 Jan-22 Jan-20 Jul-20 Jan-21 Jul-21 Jan-22 East Asia & Pacific (middle income) Rest of world b. By country in East Asia Stringency index Mobility index 100 100 Stringency index (14-day average) 80 Mobility index (14-day average) 50 60 0 40 –50 20 0 –100 Philippines Indonesia Vietnam Malaysia Mongolia Cambodia Philippines Malaysia Cambodia Indonesia Vietnam Mongolia Sources: Oxford University, Google Mobility Reports, accessed May 2022. Note: The mobility index is the average percentage change in mobility to retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. In panel a, figures are 14-day rolling averages. In panel b, the period over which both indices are observed is from March 2020 to April 2022. Boxplots indicate the interquartile range, median, minimum, and maximum values and outliers (dots). INTRODUCTION 15 While the region was quick to enact mobility restrictions, the degree of severity of these measures varied across countries. Among the six studied in this report, mobility restrictions in the Philippines and Indonesia were generally the most stringent and protracted (Figure I.8, panel b), although compliance waned toward the end of 2021 and into 2022 in Indonesia. In both countries, the severity of lockdown measures was largely consistent throughout the pandemic, a reflection of relatively early and prolonged outbreaks. On the other hand, mobility restrictions were more varied during the first two years of the pandemic in Vietnam, Malaysia, Mongolia, and Cambodia. Vietnam was quick to enforce severe lockdown measures at the start of the pandemic and in the latter half of 2021, when cases rose rapidly, yet successful containment for much of 2020–2021 meant that after the initial months, mobility levels were overall close to pre-pandemic levels. Malaysia had long periods of high stringency—as did Indonesia and the Philippines—but removed restrictions earlier. The duration of strict lockdowns was comparatively short and more intermittent in Cambodia, and particularly Mongolia, occurring at crucial points in time when new outbreaks were imminent. How Severe was the economic impact of COVID-19? Like other regions, EAP experienced a large contraction in 2020 and began to recover in 2021 (World Bank 2022c). The region is now projected to grow 5.2 percent in 2022–23, the second-fastest regional growth rate in the world, after South Asia (World Bank 2022a). However, there are sizeable differences in performance within the region and among the six countries examined in this report. In the region, the Pacific Island countries were severely impacted, given their dependence on tourism, while China only contracted in the initial months of 2020 and ultimately exceeded its pre-pandemic output level in 2020. Among the six countries covered in the report, the Philippines suffered the steepest economic contraction in 2020, suffering output losses significantly larger than the global median (figure I.9). Despite recovering to pre-pandemic levels by 2022, the decline in GDP was still greater than in most countries globally. Malaysia and Mongolia also faced sharp contractions in 2020, and while Malaysia recovered to pre-pandemic levels by 2022, the relatively slow pace of recovery in Mongolia meant that it was the only country among the six studied that still had not rebounded to pre-pandemic levels in 2022. At the other extreme, Vietnam grew throughout the pandemic, albeit at a slower rate than in 2019. Indonesia and Cambodia also had strong recoveries after smaller drops in output in 2020. The EAP Economic Updates provide a useful framework for explaining the potential factors behind the heterogenous impacts and recovery across the region (World Bank 2021a, 2021b, 2022c). Economic performance across countries has been shaped by three broad factors: ⦁ The scale of the COVID-19 shock and the efficiency of containment measures: Countries that were relatively more successful in avoiding large-scale outbreaks relied on prompt, effective lockdowns and border closures, as well as early and extensive testing in the first years, and timely vaccine rollout since 2021. ⦁ The ability to take advantage of buoyant external conditions: The pandemic had different effects on different sectors. Tourism-dependent economies experienced the sharpest declines in growth, and specialization in manufactured goods eased economic pain. ⦁ The capacity of the government to provide support: As will be discussed in more detail in chapter 9, government support to households and firms greatly aided the economic rebound across the world and in several of the EAP economies. 16 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.9  2020 GDP Index and projected 2020 GDP Index (2019 GDP 5 100) 120 Vietnam 110 Indonesia Cambodia 2022 GDP Index Malaysia 100 Philippines Mongolia 90 80 80 90 100 110 120 2020 GDP Index Source: World Bank 2022a. Note: Dotted lines indicate global medians. Cambodia and Vietnam’s relatively effective containment of the outbreak early on through short-lived stringent measures, comprehensive testing, and contact tracing allowed domestic activity to resume while suffering relatively smaller costs. Still, Cambodia’s main engines of growth, manufacturing of garments and footwear for export, and tourism, contracted severely in 2020 as global demand collapsed, while Vietnam’s highly diversified export-orientation (including electronics) in high demand from advanced economies helped the country continue to grow in 2020, as well as to rebound quickly in 2021. Despite Indonesia’s prolonged health crisis and limited containment, the country’s relatively low dependence on trade isolated the economy from a severe shock in 2020. Yet, the continuous outbreak and lockdowns hampered the domestic economy in the first year of the pandemic, as well as the recovery in 2021. At the same time, the government’s relatively strong response to provide relief and to later stimulate employment and growth, while containing spending and public debt, may have helped the economy recover. Malaysia experienced a severe dip in 2020, driven by high levels of stringency and a fall in exports, despite their relative success in containing the virus. In 2021, Malaysia’s economy recovered as external demand on manufactured goods and domestic demand resumed, following strong vaccine rollout and the lowering of mobility restrictions. Export diversification likely aided the recovery. Mongolia, while vastly different from Malaysia, also experienced a severe dip in 2020, driven by high levels of stringency and a fall in exports. However, in the case of Mongolia, the economy’s high dependence on Chinese demand meant that as China instituted strict lockdown measures that slowed their economy, this in turn crippled Mongolia’s own economic recovery. Export concentration linked Mongolia’s economic recovery to the fate of the Chinese economy. INTRODUCTION 17 Finally, prolonged and severe mobility restrictions in the Philippines, without much containment, took a toll on the domestic economy. In addition, the country’s strong reliance on tourism and related service exports greatly impacted the economy early on. Given that global demand for tourism remained subdued in 2021, recovery was also limited. Impact on firms A growing body of literature has emerged on the impact of the pandemic on the economy, but only a few studies look at developing countries.6 Beck and colleagues (2020) examined a sample of 500 firms across 10 developing countries. They found that most firms limited layoffs and payroll reductions by reducing investment. Apedo-Amah and colleagues (2020) analyzed a sample of firms in 51 developing countries. The data showed that the COVID-19 shock was severe for small and medium-size enterprises (SMEs), with persistent negative impacts on sales. The impact of the COVID-19 pandemic varied in intensity across countries in EAP and changed over the course of the crisis. Many EAP countries are integrated into the global economy through trade and tourism. In some ways, this integration exacerbated their vulnerability to the pandemic. Falling commodity prices and the tightening of financial conditions added to the disruption of exporting countries, such as Malaysia and Mongolia. However, negative shocks through trade channels did not affect Vietnam as much. Trade may have hurt some countries, but diversified trade may have helped in the medium-term. How trade affected countries in the downturn, yet helped others in the recovery, remains an open question. During the recovery phase, the share of operating firms gradually increased across all firm sizes and countries, though it declined again in the resurgence phase in 2021 (Figure I.10). Throughout all survey waves, at more than 90 percent, the country with the largest share of firms open was Vietnam; the Philippines, which had more severe outbreaks and lockdown restrictions than other countries, had the largest share of closed firms (Figure I.11). Of the six countries studied, the Philippines experienced the steepest decline in sales, with 6 out of 10 firms reporting sales declines of more than 90 percent—twice the share of Philippine firms in the sample. Nearly 6 out of 10 firms that declared increases in sales were in Malaysia, about twice the share of its firms in the sample. The largest drop in sales took place between October and December 2020. The contrast between Malaysia and Philippines is notable, despite having similar GDP declines and mobility restrictions. The pandemic had an uneven impact across firms within countries, with smaller firms, especially microenterprises, experiencing greater proportional losses than larger firms (Figure I.12). These differences persist after controlling for initial labor productivity; the age and location of firms, such as in the capital region or elsewhere in the country; and firm linkages to international markets via imports or exports. Approximately 20 percent of firms in the business pulse survey and enterprise survey samples reported no variation in sales in the 30 days before the survey relative to the same period in 2019. Firms in the top 10th percentile of sales declines experienced reductions of 90 percent or more. The median decline in sales was 35 percent. Across countries, sales declined in 70 percent of businesses, remained constant in 20 percent, and increased in 10 percent. 6  Most research has been focused on China, the United States, and other advanced economies. Dai and colleagues (2014) document revenue loss, business closures, and layoffs in China. Acharya, and Steffen (2020) and Fairlie (2020) document the evidence on the United States. Demmou and colleagues (2021a, 2021b), Banerjee and Kharroubi (2020), and IMF (2021) study advanced economies. 18 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.10  Percent of fully or partially-open businesses, by sector and firm size, 2020–21 a. By sector b. By firm size 100% 100% 90% 90% Percent Percent 80% 80% 70% 70% 60% 60% 0 0 20 21 1 1 02 02 0 0 20 21 21 1 02 02 20 02 20 02 02 20 n2 t2 20 n2 p2 20 t2 ar n2 p2 ec Oc ar u Jun u ec Se Oc –M r–J t–D r–J u Se – –M t–D – r–J r– Jul – Ap Jan Jul Ap – Jul Oc Jan Jul Ap Ap Oc Manufacturing Retail and wholesale Micro (1–4) Small (5–19) Hospitality Other services Medium (20–99) Large (100+) Source: World Bank estimates using business pulse survey data, 2020–21. Note: T-bars around markers in panel b represent estimates at a 95 percent confidence interval. FIGURE I.11  Percent of fully or partially open businesses during the pandemic, by period (weighted) 100% 80% 60% Percent 40% 20% 0% Apr-Jun 2020 Apr-Jun 2021 Jan-Mar 2021 Jul-Sep 2020 Apr-Jun 2020 Jan-Mar 2021 Oct-Dec 2020 Jan-Ma r2021 Jul-Oct 2021 Jul-Sep 2020 Oct-Dec 2020 Jan-Mar 2021 Jul-Sep 2020 Apr-Jun 2021 Jul-Sep 2020 Oct-Dec 2020 Apr-Jun2020 Jan-Mar 2021 Jul-Oct 2021 Jul-Sep 2020 Oct-Dec 2020 Cambodia Indonesia Malaysia Mongolia Philippines Vietnam Source: Data from business pulse surveys and enterprise surveys, 2020–21. Note: Data was not available for all countries in all periods. INTRODUCTION 19 FIGURE I.12  Average reported change in sales in 30 days before interview, relative to same period in 2019, by country and firm size (weighted) a. By country b. By firm size (employees) Philippines Indonesia Mongolia Vietnam Malaysia 20 21 0 20 1 21 02 02 20 20 20 20 2 2 ar c un un ep Oct–Dec 2020 Oct–Dec 2020 Oct–Dec 2020 Oct–Dec 2020 De ct Jan–Mar 2021 Jan–Mar 2021 Jan–Mar 2021 –M Apr–Jun 2020 Apr–Jun 2020 Jul–Sep 2020 Jul–Sep 2020 Jul–Sep 2020 Jul–Sep 2020 Apr–Jun 2021 –O r–J r–J –S Jul–Oct 2021 Jul–Oct 2021 t– Jan Ap Oc Ap Jul Jul 0 0 –10 –10 –20 Percent –20 –30 Percent –30 –40 –40 –50 –50 –60 Micro (1–4) Small (5–19) –60 Medium (20–99) Large (100+) Source: Data from business pulse surveys and enterprise surveys, 2019-2020. Note: The data represents an average across all sampled firms. Data combining multiple countries can only be aggregated by rounds, not periods, which may not always coincide. For panel a, data on Cambodia was not available, as questions on change in sales were not included. In panel b, whiskers around markers in panel b represent estimates at a 95 percent confidence interval. The slowdown caused by the COVID-19 outbreak and lockdown measures made it difficult for companies to meet their financial obligations. It is likely that many otherwise sound companies faced liquidity constraints that eventually led to insolvency. Across survey rounds, the proportion of firms in arrears or expecting to fall into arrears within six months held relatively steady in all countries. The threat of insolvency was particularly acute in Mongolia and the Philippines. Across all countries, the largest share of cash-constrained firms were small firms and microenterprises. However, trends improved over time, suggesting that the number of firm closures is likely to decline. To help mitigate the impact of the pandemic, businesses adjusted their use of labor, reducing the number of workers and/or hours and days worked. Employment adjustment by formal sector firms included reducing wages/hours and granting leave, the intensive margin, as well as laying off workers, the extensive margin. According to the business pulse survey data, Malaysia, the Philippines, and Mongolia experienced much more labor displacement than did Indonesia and Vietnam. There was also significant heterogeneity in patterns of employment over time. In some countries, the drop in employment worsened over time. In Malaysia, for example, the share of firms that sought to reduce their labor costs—through either reducing hours or shedding workers—was roughly the same in the fourth quarter of 2020 and February 2021. In contrast, in Vietnam, the share of firms reducing labor costs had declined by March 2021. Businesses in Mongolia, Malaysia, and the Philippines were the most likely to reduce working days and hours to reduce their wage bills. Firms also responded to the pandemic, lockdowns, and economic downturn by adopting digital technologies. The uptake of digital technology by firms in the six countries studied was impressive. Globally, the share of firms that increased the use of digital technologies rose from 31 percent in the early months of 20 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA the pandemic to 44 percent some 7 to 12 months into the crisis. In the six East Asian countries, 38 percent of firms used digital technology in the early months, but four to six months later, the figure had risen to 55 percent. As a result, there was an uptick in the use and adoption of digital technologies recorded by each successive wave of the survey, coinciding with the progression of the pandemic and economic contraction (Figure I.13). Only Malaysia—where the use of these technologies was the highest among the six economies and higher than the regional average before the pandemic started—did not experience this change. Although there was a wide variation among countries and waves, the tendency to increase the use of these platforms as the pandemic progressed was robust. There are sector-specific patterns to digital adoption across the six East Asian economies. Figure I.14 displays industry-level data on the use of digital platforms. On average, establishments in ICT and education increased their use of the Internet by 30 percent more than average in the economy. At the opposite end of the spectrum, use of digital platforms by firms in the food services sector was 12 percent below average, and the manufacturing sector lagged by 7 percent. The pandemic affected all firms, but it affected some more than others. Several characteristics are correlated with performance in terms of sales losses during the pandemic: ⦁ Age: Younger firms were somewhat more likely to suffer large declines in sales than older firms. Young firms represented 26 percent of the firms surveyed and 30 percent of firms that performed worst in terms of sales losses. Older firms represented 42 percent of the firms surveyed and 46 percent of the best performers. ⦁ Size: Smaller firms, meaning those with less than 19 employees, were overrepresented among the worst performers. These firms represented 31 percent of all firms, 54 percent of the worst-performing firms, and 16 percent of the best-performing firms. Large firms, defined as firms with more than FIGURE I.13  Percent of firms that began using or increased use of digital platforms, by period 100 80 60 Percent 40 20 0 Vietnam Indonesia Cambodia Malaysia Philippines Mongolia Apr–Jun 20 Jul–Sep 20 Oct–Dec 20 Jan–Mar 21 Apr–Jun 21 Jul–Oct21 Source: Data from business pulse and enterprise surveys 2020-21. INTRODUCTION 21 FIGURE I.14  Percent of establishments in six East Asian countries that started or increased their use of digital platforms, by sector Information and communication Education Other services Financial activities and real estate Health Accommodation Retail and wholesale Construction and utilities Agriculture, fishing, and mining Transportation and storage Manufacturing Food services 0.8 0.9 1 1.1 1.2 1.3 Source: Data from the business pulse survey. Note: Data is scaled by economy and survey wave and normalized so that the average rate of adoption for the economy is equal to 1.0. 100 employees, represented 21 percent of the firms surveyed, 6 percent of the worst-performing firms, and 37 percent of the best-performing firms. The average worst-performing firm was significantly smaller than other firms, at 63 versus 185 employees, whereas the average best-performing firm was significantly larger than other firms, with 268 versus 162 employees. ⦁ Engagement in exporting: For the sample, exporting firms made up 30 percent of the surveyed firms, 12 percent of the worst-performing firms, and 45 percent of the best-performing firms. An exception was Mongolia, where exporters and foreign firms were more, rather than less, likely to close (see chapter 3). ⦁ Gender of owner: There were no notable statistically significant differences in the performance of firms owned by women versus those owned by men. ⦁ Sector: Accommodation and food preparation services were overrepresented among the worst- performing firms and underrepresented among the best-performing ones. The opposite was observed for manufacturing; financial services; and, to a lesser extent, agriculture and health. Retail and wholesale businesses were underrepresented among the worst-performing firms but not overrepresented among the best-performing firms. ⦁ Adoption of digital technologies: Firms that did not invest in digital technologies were prevalent among the worst performers. The share of firms that invested in technology was 46 percent among all firms, 23 percent among the worst performers, and 63 percent among the best performers. ⦁ Bankruptcy: The worst-performing firms were predominantly among firms that closed during the pandemic, with 17 percent of all firms, 62 percent of the worst-performing firms, and 7 percent of the best-performing firms closing. Three-quarters of the worst-performing firms expected to enter arrears in the following six months. Impact on households Firms’ adjustment strategies at the extensive margin were reflected in declining employment across EAP countries. In 2020, employment declined substantially in many EAP countries (Figure I.15), and unemployment 22 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.15  Employment-to-population ratio in 2020 and 2021 1 0.2 0.2 Percentage point di erence relative to 2019 0 –1 –1.3 –1.1 –1.0 –2 –1.8 –1.9 –2.1 –1.9 –2.4 –2.5 –2.2 –3 –2.6 –3.2 –3.0 –4 –5 –4.8 –6 Philippines Vietnam Cambodia Indonesia Malaysia Mongolia World Lower-middle income countries 2020 2021 Source: Labor force surveys via the International Labor Organization’s ILOSTAT, accessed May 2022. Note: Modelled estimates by the International Labor Organization were used for Cambodia, the world and lower-middle income countries. reached new highs. The Philippines, which suffered the largest reductions in firm sales in the region, also experienced the largest drop in employment in 2020, with the employment-to-population ratio declining by 4.8 percentage points relative to 2019, and the official unemployment rate peaking at 18 percent in April 2020, according to the Philippines Statistics Authority. Other countries also faced drops in employment, which tended to be larger in 2021—and generally above world averages—when many countries faced large-scale outbreaks and continued lockdowns. In some countries, employment impacts were softened by the ability of workers to retreat into mostly low-skilled, informal sectors. In the Philippines, employment recovered to pre-pandemic levels in 2021, yet the increase was largely due to a reallocation of employment into lower-skilled sectors, such as agriculture and retail. Employment in key service sectors, such as tourism, remained substantially lower than pre- pandemic levels. In Malaysia, which, like the Philippines, also faced a significant drop in output in 2020, labor markets were quicker to adjust, and employment declines in 2020 were not as severe as those in other countries in the region (Figure I.15). The comparatively muted initial effect on employment in Malaysia, however, can largely be attributed to both workers and firms taking advantage of new opportunities created by the informal gig economy during the lockdown, such as food delivery and courier services. In Indonesia, labor force survey data showed that in 2020, despite significant unemployment, previously inactive women in their 30s or older entered the workforce, mostly in unskilled, informal, and agricultural jobs, to support household income (Halim and colleagues, 2021). As economies continue their paths to recovery, it remains to be seen whether the reallocation of employment into lower-skilled, informal jobs will lead to a structural shift in the labor market—namely, whether economic growth in the post-COVID era results in less formal jobs compared to pre-COVID. Labor market policies can play a crucial part in promoting formal job creation as countries recover, as they did in Malaysia. At the height of the pandemic, the Malaysian government created incentives for firms to hire new workers and extended social insurance to informal workers to bring them into the formal economy. These policies played an important role in cushioning impacts in the labor market while promoting formal job creation. INTRODUCTION 23 The HFPS provide a more nuanced picture of employment impacts throughout 2020 and 2021. For example, official estimates indicate that overall, employment in Mongolia was not affected in 2020. In contrast, the HFPS data shows that work stoppages increased significantly in November to December of 2020 (Figure I.16), when the government imposed a nationwide lockdown. The surveys also reveal that although the Philippines experienced the largest overall drop in employment in 2020 among the six countries, work stoppages had already abated somewhat by August 2020, and employment continued to improve into 2021. For several countries, including Indonesia, the HFPS data on employment between 2020 and 2021 differs from estimates based on labor force surveys. The difference may reflect the fact that work stoppages might capture normal fluctuations in employment from seasonality and trends depending on the sample of respondents used to calculate work stoppages.7 Work stoppages affected some groups more than others, highlighting the potential for increased inequalities (Figure I.17). Although results differed by country and period, women, poorer and less educated workers, and workers employed in sectors that required face-to-face interactions were more likely to stop working during the pandemic than other workers: ⦁ Gender: On average, women were 6 percentage points, or 40 percent, more likely than men to stop working. This gender gap was evident in almost all EAP countries. In some countries, such as Mongolia and the Philippines, women were significantly more likely to stop working, due to caregiving responsibilities at home. In Indonesia, as mentioned, previously inactive women entered the labor force to support household income, albeit in generally more informal, agricultural activities (Halim and colleagues, 2021). FIGURE I.16  Work stoppages in 2020 and 2021 50 Percent of respondents working pre-pandemic 40 30 20 10 0 Mongolia Philippines Cambodia Malaysia Indonesia Vietnam May–June 2020 January–March 2021 July–September 2020 April–July 2021 October–December 2020 October–December 2021 Source: Kim and colleagues, 2021, based on high-frequency phone surveys. Note: Work stoppages are defined as the share of respondents working before the pandemic that were no longer working in the week preceding the survey. 7  For most countries, the head of household is predominantly presented in the sample. To the extent that the head of household differs from the average worker, results may not be representative of the entire labor force. 24 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE I.17  Characteristics of workers and work stoppages in selected countries in East Asia and Pacific, May 2020–December 2021 0.06 Female 0.01 Urban Employment sector (Reference: agriculture) 0.06 Industry 0.04 Services: hospitality, transport –0.02 Services: other Educational attainment (Reference: less than primary) –0.02 High school –0.08 Tertiary Welfare group (Reference: bottom 40) –0.01 Middle 40 –0.03 Top 20 –0.10 0.00 0.10 Coefficient of work stoppage on worker characteristics Source: Data from 33 rounds of high-frequency phone surveys across 11 countries in East Asia and Pacific. Note: Figures show coefficients of pooled regressions of work stoppages on worker characteristics: one regression per characteristic, not controlling for other factors. All regressions include country fixed effects. ⦁ Location: For the region, urban and rural workers were generally as likely to be affected by work stoppages. In several countries, however, including Mongolia, Vietnam, and Indonesia (early on), urban workers were more likely to be affected, partly because of higher participation in sectors that were heavily affected by the pandemic. ⦁ Sector: Employment across the region, particularly in mining and construction, as well as tourism, hospitality, and transportation, was heavily affected. Work stoppages were less severe in other service sectors, such as public administration, finance, and professional activities, as well as in agriculture, although losses in agricultural income were substantial in many countries throughout the region. ⦁ Educational attainment: On average, workers who had completed secondary school were 2 percentage points, or 12 percent, less likely than workers who had completed less than secondary school to stop working, and workers who completed tertiary education were 8 percentage points, 41 percent, less likely to do so. These findings partially reflect the fact that better-educated workers are more likely to work in sectors that are more conducive to working from home, but differences persist even within sectors, indicating that other factors, such as job security, formality, and type of occupation, may also play a role. ⦁ Wealth: When mobility restrictions were strict, workers experienced work stoppages across a wide range of sectors, and employment impacts were distributed relatively evenly across the welfare distribution. As mobility restrictions were relaxed and businesses reopened, work stoppages became more selective. INTRODUCTION 25 Poorer workers—often employed in less secure, informal employment—were less likely to return to work than workers in the top 20 percent (Kim and colleagues, 2021). On average, workers in the top 20 percent of the wealth distribution were 3 percentage points, or 16 percent, more likely to stop working than workers in the bottom 40 percent. Losses in labor income were more widespread than losses in employment, corroborating findings that firms were more likely to reduce wages, work hours, and returns to capital than to lay off workers. Across six countries, in which at least one HFPS was conducted in both 2020 and 2021 (Cambodia, Indonesia, Lao PDR, Mongolia, the Philippines, and Vietnam), nearly 70 percent of households experienced reductions in labor income relative to before the pandemic. This share declined to 55 percent in 2021, signaling a gradual, albeit not yet full, recovery. Throughout the pandemic, self-employment income losses were more prevalent than losses in wages (Figure I.18), indicating relatively large reductions in sales among businesses. However, in general, EAP countries fared better in terms of work stoppages and income losses than many other countries (Box I.2). Shocks to labor income were widespread in the first year of the pandemic and were more likely to persist among poorer households. Across EAP countries, differences in the likelihood of experiencing a reduction in labor income were marginal in 2020. In contrast, in 2021, the average household in the bottom 40 percent was about 14 percent more likely than the average household in the top 60 percent to earn less than before the pandemic. Poorer households may also have more difficulty recovering from prolonged lockdowns and economic downturns, despite significant social assistance in many countries. FIGURE I.18  Non-farm business income vs. wage income losses 80 Percent of households that experienced wage/business 60 income reduction 40 20 0 Mongolia Philippines Cambodia Malaysia Indonesia Vietnam Non-farm business Wages Source: Data from high-frequency phone surveys. Note: Estimates are among households with businesses or wage income, respectively. Income losses are those following the previous survey round; for countries with more than one round, figures show the average across rounds. 26 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA BOX I.2  Global comparisons of household work stoppages and income losses using high-frequency phone survey data Global comparisons from the World Bank 2022 Poverty and Shared Prosperity Report show that many EAP countries suffered lower work stoppages and household income losses than most other countries with high-frequency data available. With the exception of Mongolia, all EAP countries examined in this report faced levels of work stoppages below the global median (Figure BI.2.1) In addition, in all six countries, with the exception of Cambodia, for which only the first round of data are available, households were less likely to suffer income losses than in the global median. FIGURE BI.2.1. Global comparisons of work stoppages and household income losses Work stoppages VNM KHM PHL MNG MYS IDN 0 20 40 60 80 Household income losses or work stoppages IDN VNM MNG PHL KHM MYS 0 20 40 60 80 Source: Data from high-frequency phone surveys. Note: Average values of work stoppages and income losses used in the World Bank 2022 Poverty and Shared Prosperity Report may differ from those presented in this report due to differing numbers of periods included in the sample and differing harmonized definitions. Reductions in remittances were also widespread in 2020 and showed limited improvement in 2021. The share of recipient households that experienced a reduction in remittances in mid-2020, relative to before the pandemic, was as high as 75 percent in Cambodia and Indonesia, where about 10–20 percent of households received remittances. In Cambodia, the share of households receiving remittances declined across rounds of the HFPS. In Indonesia, remittances showed little sign of improvement in 2021. For countries such as Mongolia, in which most remittances come from domestic sources and are received in cash, mobility restrictions prevented transfers between households. However, evidence from 2020 shows that use of bank transfers and other digital methods of sending remittances increased significantly in Mongolia, particularly among the bottom 40 (World Bank forthcoming b). To cope with the impacts of the pandemic, households engaged in a variety of coping mechanisms, some of which could have long-term implications for their welfare. In most countries, the most frequently INTRODUCTION 27 FIGURE I.19  Household coping mechanisms adopted during the pandemic in six countries in East Asia Cambodia Indonesia Malaysia Mongolia Philippines Vietnam 0 20 40 60 80 100 Percent of households Low-cost Potentially scarring Less food consumption Less non-food consumption Source: Data from high-frequency phone surveys Note: Low-cost coping mechanisms include engaging in additional income-generating activities, borrowing or receiving aid from friends/family, relying on savings, relying on government or NGO assistance, and using insurance. Potentially scarring coping mechanisms include selling farm/non-farm assets, taking children out of school, taking out loans from informal/formal institutions, buying on credit, and delaying payment obligations. used coping strategies were relatively low-cost strategies, such as drawing on savings, relying on government assistance, and engaging in more income-generating activities (Figure I.19). However, a significant share of households also deployed potentially scarring coping strategies, including selling productive assets, increasing indebtedness, reducing food expenditure, and taking children out of school. As many as two- thirds of households in the Philippines relied on some of these mechanisms, for example, and these negative coping strategies were more prevalent than low-cost ones in Cambodia. How did governments respond? Countries in EAP mounted some of the largest and most timely expansions in support to households and firms in response to the pandemic. According to recent estimates by the IMF based on the IMF Fiscal Monitor Database, the COVID-19 policy response in EAP was greater than in any other developing region. From the onset of the crisis, East Asian governments implemented a variety of measures to preserve jobs, smooth household consumption, and prevent widespread bankruptcy among firms. Beyond mitigating the immediate impacts of the pandemic to households and firms, they also put in place measures to help workers and firms reintegrate into the economy as countries recovered, improve productivity, and prevent detrimental long-term consequences to human capital and economic prospects. Government efforts have helped countries cushion the observed poverty impacts. This introduction will present a brief description of the measures put in place in the six countries of study, while the analysis and reflection on these measures is relegated to chapter 7, after the reader has had the opportunity to review the country-specific analysis presented in chapters 1 to 6. 28 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Spending on support measures to households and firms reached unprecedented levels in many countries in the region, with many EAP countries doubling or even tripling their spending on social protection, albeit from low levels in some countries, particularly early in the pandemic (World Bank 2021b). This is even though domestic revenue mobilization—as measured by the total revenue from taxes and social contributions as a share of GDP—in many of these countries, such as Indonesia, Malaysia and the Philippines, has been traditionally low relative to their level of development (UNU-WIDER Government Revenue Dataset 2022 and WDI 2022. Mongolia recorded the highest spending on pandemic-related support as a share of GDP, about a quarter, in 2020 and 2021 (Figure I.20), with support to households and liquidity measures as the main components. Vietnam provided the smallest package, 5 percent to households and firms, and relied largely upon liquidity and revenue measures, such as tax and contributions deferrals and rates reductions, mostly in 2021. Vietnam’s response is consistent with the fact that the economic disruption there appears to have been the mildest of the six countries studied. Cambodia spent 5.4 percent of GDP supporting households and firms, but relative to Vietnam, three times as much support went to households, potentially as a reaction to the large income losses and work stoppages experienced in 2020. Malaysia and the Philippines relied heavily on support to firms in the form of liquidity, credit, and lending—“below-the-line” measures. Indonesia provided significant support to households, 3.2 percent of GDP, largely in the form of direct support through transfers and subsidies. Chapter 7 provides a more detailed analysis of the differences in size, composition, and effectiveness of COVID-19 support packages. FIGURE I.20  Composition of COVID-19 support packages 35 30 25 Percent of GDP 20 15 10 5 0 Mongolia Malaysia Indonesia Philippines Cambodia Vietnam Contingent liabilities and equity injections Other above-the-line Public investment Health Support to firms Support to households Source: World Bank staff estimates based on the latest data from governments as of February 2022. Note: Estimates are based on the latest data from national governments as of February 2022. Where disbursement data was not available, data on planned budgets was used. “Support to households” includes income support and revenue measures to households, such as direct transfers, employment programs, utility subsidies, and tax exemptions. “Support to firms” includes income support and revenue measures to firms, such as direct transfers; wage subsidies; and tax and contribution deferrals, reductions, and exemptions. “Other above-the-line” items include uncategorized categories in which a disaggregation could not be made. INTRODUCTION 29 Support to households Countries in the region adopted various instruments to support households during the pandemic (Table I.2). Social insurance systems by design are well-suited for disruptions such as those brought about by the pandemic, providing automatic compensation to displaced workers and covering health costs for those that fell ill. But in a context where only a fraction of workers contributes to the formal system, social assistance can play an important role in alleviating the losses of workers and households that have suddenly lost the ability to work, either because of stay-at-home orders, the inability to offer face-to-face services, or a slump in demand for their products. All six countries in this study used cash-based transfers, complemented in most cases by in-kind support and utilities or financial support, as did most countries of the world (World Bank 2022d). These social assistance measures reached a wide swath of the population, including vulnerable, but not initially poor, households and informal sector workers. Cambodia, Indonesia, and the Philippines also provided support to informal workers, such as temporary jobs and cash-for-work opportunities, a relatively little-utilized instrument in the region, as only a quarter of 21 EAP countries did so. For formal sector workers, governments extended the reduction in social security contributions and, to a lesser degree, paid leave and unemployment benefits. They also adopted revenue measures, such as tax deductions/exemptions and other payment deferrals, to provide support to firms. Additionally, the six countries used labor market interventions aimed at protecting jobs and strengthening workers’ skills, such as job training programs and wage subsidies. As the pandemic progressed, the delivery of support improved, particularly to vulnerable households and informal workers. In many countries, however, social assistance was short-lived, with support to households rolled back in all countries in 2021, despite evidence of ongoing income losses. More than half of households surveyed in Indonesia, Lao PDR, and the Philippines reported lower levels of labor income in 2021 than before the pandemic. Scaling back government assistance despite continuing widespread income losses could increase poverty or impede efforts to reduce it. TABLE I.2  Support to households and workers provided by six countries in East Asia Percent of all 21 EAP countries Type of support that used measure Cambodia Indonesia Malaysia Mongolia Philippines Vietnam Social assistance Cash-based transfers 86 ✓ ✓ ✓ ✓ ✓ ✓ Public works 24 ✓ ✓ ✓ In-kind support 62 ✓ ✓ ✓ ✓ ✓ Utilities and financial support 86 ✓ ✓ ✓ ✓ ✓ Social insurance Paid leave/unemployment 48 ✓ ✓ ✓ ✓ Health insurance support 29 ✓ ✓ ✓ Pension and disability benefits 29 ✓ ✓ Social security contributions 76 ✓ ✓ ✓ ✓ ✓ Source: Gentilini and colleagues, 2022. 30 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Support to firms Decisions on which policies and instruments to apply reflected several country-specific aspects, including the public sector’s administrative capacity to provide timely support, the level of targeting, and the problems the policy or instrument was designed to address, for example, temporary cash shortages, insolvency, bankruptcy, or the need to preserve a highly specialized labor force. Targeting can be based on firm characteristics, such as size, sector, and whether the firm is an essential input supplier to other industries or earns foreign exchange by exporting. Preservation of linkages between firms and workers The transitory nature and severity of the shock called for measures beyond traditional policy responses to help preserve employment linkages. In response to mild economic disruptions, it may be sufficient to provide some liquidity support to firms while strengthening the social protection system. The highly disruptive nature of the pandemic crisis, especially on the supply side, made measures to preserve employment linkages critical. These instruments are often thought of as temporary and time-bound, with flexibility for extension if needed. Measures such as wage subsidies and restrictions on lay-offs help prevent the loss of firm-specific human capital, which can be costly over the medium term, and ease liquidity pressures on stressed firms. They help buffer economic activity in the short term, even if some workers receiving subsidies work at reduced capacity or productivity. All six of the countries studied provided support for subsidized wages to formal sector firms (Table I.3). Mongolia and the Philippines provided subsidies to allow firms to maintain workers at a reduced work schedule. Indonesia and the Philippines made changes to labor regulations. Indonesia gave more flexibility to employers to cut salaries and offer unpaid leave and made it easier to terminate employees. The Philippines instituted legislative changes on COVID-19-related sick leave, social security contributions, and unemployment insurance. Liquidity support and borrower relief Firms often delay investments during a recession. However, in this downturn, investments in digital technologies were critical, and many firms made them. Investments in digital technology allowed many firms to reach customers and suppliers despite mobility restrictions. In addition, investments in information technology (IT) allowed more employees to work-from-home. All six countries provided liquidity support for or directly provided short-term lending. However, in general, such lending support went almost entirely to larger firms, as will be shown in chapter 7. Legally sanctioned forbearance of payment for financial obligations also predominantly benefited formal firms. Tax deferrals, tax relief, and deferrals of interest and/or principal payments on loans benefit firms that pay taxes and/or have access to credit; forbearance of utility bills benefits only businesses registered to receive electricity. All of these measures almost always exclude informal firms. All countries except Cambodia made long-term loans to firms. It is unclear how effective these measures were because it is not yet clear how firms used these loans and what benefits they received from them. If the proceeds were used to support the adoption of digital technologies to reach customers and suppliers and allow workers to work-from-home, such loans may have been beneficial. If, however, their investments were inefficient, this type of liquidity support may not have been helpful. INTRODUCTION 31 TABLE I.3  Support to firms provided by six countries in East Asia Measure Cambodia Indonesia Malaysia Mongolia Philippines Vietnam Labor market support Wage subsidy ✓ ✓ ✓ ✓ ✓ ✓ Training ✓ ✓ ✓ ✓ ✓ Labor regulation adjustment ✓ ✓ Reduced work time subsidy ✓ ✓ Liquidity support Short-term lending ✓ ✓ ✓ ✓ ✓ Support policies for short-term lending ✓ ✓ ✓ ✓ Foreign exchange operations ✓ Credit creation Financial sector lending/funding ✓ Loan guarantees ✓ ✓ ✓ ✓ ✓ ✓ Direct long-term lending Long-term lending ✓ ✓ ✓ ✓ ✓ Forbearance ✓ ✓ ✓ Equity support ✓ ✓ Income support Payment deferrals (taxes, utilities, rent and loan payments) ✓ ✓ ✓ ✓ ✓ ✓ Source: Asian Development Bank, COVID-19 Policy Database, accessed April 19, 2022. All six countries provided loan guarantees, with differing eligibility criteria, which incentivized banks to make loans they otherwise may have viewed as too risky. Credit guarantees are often more efficient than direct government support because the transaction cost of distributing subsidies or loans to multiple beneficiaries is likely to be higher. Experience with credit guarantees indicates that the least risky firms are usually afforded better terms of lending when the loans are guaranteed. However, it is unclear that banks actually made loans that they would not have made without the guarantee. Mongolia was the only one of the six countries studied that supported credit creation through direct government support. Malaysia and Vietnam provided equity to keep firms solvent. This type of support may be best for firms that provide essential input on which a network of other firms depends. It can also be effective to ensure that exporters continue to operate, providing much needed foreign exchange to the economy. A government equity injection immediately improves the balance sheet of the firm and reduces funding costs and risks, allowing the firm to continue operating and investing. The government is a shareholder with a voice to ensure value for taxpayers’ money, including by setting conditions for the equity injection—which could also be legally imposed. Equity is junior to debt, however, and may be excessively risky if the crisis is prolonged. In addition to the previously addressed liquidity support, governments implemented a series of pandemic crisis support measures to provide borrower relief and stem insolvencies. Authorities in the six East 32 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE I.4  Financial sector policy measures to provide relief to borrowers Area Policy measure Countries Funding Fresh funds Cambodia, Indonesia, Malaysia, Philippines, and Vietnam Capital injection in banks Cambodia Regulatory relief Capital buffers Cambodia, Malaysia Capital: risk weights Philippines Liquidity buffers Cambodia and Malaysia Liquidity: reserve requirements Indonesia, Malaysia, Mongolia, and Philippines Moratoria, classification and provisioning Cambodia, Indonesia, Malaysia, Philippines, and Vietnam Credit reporting Indonesia Financial infrastructure Asset management company creation Changes to the insolvency regime Source: Adapted from table 2, page 47 of Non-Performing Loans in East Asia and the Pacific: Practices and Lessons in Times of COVID-19, World Bank. 2021. Asian economies instituted moratoria-like regulatory forbearance and repayment directly aimed at keeping credit flowing during the crisis. Specifically, the measures implemented can be grouped into two categories. Firstly, authorities have given some degree of regulatory relief to financial institutions so that they are able to keep the flow of credit channeled to the economy. Secondly, none of the authorities have promoted changes to elements of the financial infrastructure, such as reforming insolvency procedures, so that it better accommodates challenges arising from the COVID-19 crisis.8 One group of financial sector measures included the provision of fresh funds to borrowers, predominantly firms (Table I.4). For instance, in Cambodia, the government injected capital into the state-owned Rural Development Bank to grant loans to SMEs in the agricultural sector. In Indonesia, SMEs with a good credit record and the capacity to repay are eligible for loans provided through public funds, with the government paying a portion of the loan interest for certain sectors. As for the second group of measures, which involved regulatory relief, some of the six EAP countries allowed financial institutions to make use of the flexibility that already existed in the regulatory framework (Table I.4). The first area where flexibility has been extended is through measures to allow banks to use existing capital and liquidity buffers. By dipping into the buffers, banks can make use of the funds to provide additional credit or to absorb losses without the pressure to immediately replenish those cushions. Cambodia and Malaysia issued explicit guidance in this respect. In addition, Cambodia, Indonesia, Mongolia, and the Philippines reduced liquidity reserve requirements, a tool that is traditionally used to relieve short-term liquidity pressures and to expand loan funds. In some cases, authorities granted regulatory relief by lowering minimum capital requirements for some borrowers, thus deviating from international good practice standards. Some jurisdictions have lowered the credit risk weight for loans to SMEs to a level below the ratio required by the Basel Framework.9 8  This section cites, paraphrases, and provides a summary of chapter 2 of the World Bank (2021) Non-Performing Loans in East Asia and the Pacific: Practices and Lessons in Times of COVID-19. World Bank, Kuala Lumpur. © World Bank. https://openknowledge.worldbank.org/ handle/10986/36522. For a more complete exposition on this topic, the reader is referred to this document. 9  The Basel Framework refers to the work of the Basel Committee on Banking Supervision (BCBS) that, developed international regulatory capital standards through a number of capital accords and related publications, which have collectively been in effect since 1988. The most recent set of reforms to the Basel Framework, Basel III, is a comprehensive set of reform measures, developed by the BCBS, to strengthen the regulation, supervision, and risk management of the banking sector. The measures include both liquidity and capital reforms. INTRODUCTION 33 For instance, the Philippines implemented such reductions in the risk weight until the end of 2020, with this measure intended to support lending to these firms (Table I.4). Moratoria were implemented by five of the six countries as extraordinary relief measures, with these jurisdictions adopting a range of different designs. In most cases, the authorities tried to limit moral hazard by making the measures time bound and by defining the sectors and loans included in the initiatives. Nevertheless, in some cases, a blanket moratorium has been applied automatically to all borrowers within a certain group. For instance, in Malaysia, an automatic six-month moratorium on loan repayments, comprising principal and interest for individuals and SMEs, was implemented on April 1, 2020, to be replaced by targeted borrower assistance after a six-month period. The Philippines also introduced some degree of forbearance in applying the prudential classification rules and provisions. In the Philippines, a temporary relaxation of provisioning requirements was allowed, subject to regulatory approval. All six countries have yet to reform some elements of the financial infrastructure to address insolvencies or non-performing loans (NPLs) that will result from the COVID-19 economic crisis. While Singapore, for example, revised elements of the insolvency frameworks to accommodate and update some areas of the legal framework for insolvencies, none of the six countries in this study followed that example. To avoid massive numbers of corporate bankruptcies, Singapore suspended the obligation on directors to file for bankruptcy. In anticipation of an increase in the number of NPLs that may have to be resolved, the creation of an asset management company could also be considered to resolve NPLs. These reforms to the financial infrastructure could be important. While banks have processes in place to manage distressed assets, the scale and complexity of the possible increase in corporate vulnerabilities could strain that capacity. Furthermore, delayed resolution of distressed loans may generate unnecessary losses and channel resources to keep “zombie” firms alive—that is, by allocating funding to weak businesses that have little or no prospect of returning to health and fully paying off their debts. This means diverting resources from viable firms, with higher productivity and growth potential, thus ultimately hindering economic activity. In addition, the lack of transparency regarding the extent of the build-up of corporate risks in banks’ balance sheets can weaken the trust in the financial sector, potentially hindering financial intermediation in the medium term. While stability risks to the financial sector seem to be relatively contained, close monitoring may be warranted. In addition to widespread repayment moratoria, countries that implemented regulatory forbearance during the pandemic crisis, including relaxing the rules defining an NPL, could be more at risk of financial instability. Financial institutions have incentives to underplay the true extent of their exposure to under-performing loans to avoid a supervisory or market response. The phasing-out of these financial sector support measures could unmask significantly larger credit risks in the banking sector associated with corporate vulnerabilities. Policymakers thus face the challenge of interpreting opaque balance sheets (World Bank 2021e and WDR 2022). Conclusions The COVID-19 crisis inflicted unequal losses on both firms and households. Small firms suffered proportionately greater earnings losses and were more likely to close than larger firms (Apedo-Amah and colleagues, 2020; Cirera and colleagues, 2021; Karalashvili and Viganola 2021), and poor households suffered greater employment and income losses than wealthier ones (Kim and colleagues, 2021; Agrawal and colleagues, 2021). 34 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Will the recovery be inclusive? Governments can use the lessons from the analysis of what governments did or did not do to soften the blow of the pandemic-induced recession on firms and households to turn the pandemic into a transformative moment for reducing inequality. These are questions that the report aims to explore, by providing an analysis of the impacts and government actions in six EAP countries. Country chapters are written so that they are self-contained and can be read on their own, for those interested in a particular case. At the same time, chapters follow a similar structure, including key concepts and definitions, to allow for an easy comparison. To conclude the analysis, chapter 7 brings these experiences together, focusing specifically on the government responses and the lessons that can be drawn from the cross-country comparisons at this stage. INTRODUCTION 35 Annex I Annex IA.  Survey Dates and Sampling Methods TABLE IA.1  Dates and sampling methods of surveys used in this report Dates of survey rounds Sampling method Country Household survey Firm survey Household survey Firm survey Cambodia R1: May 2020 R1: June–July 2020 Subsample of the 2017 Sample from listings from R2: August–September 2020 R2: September 2020 Cambodia Socio-Economic the National Institute of R3: October–November 2020 R3: January–February 2021 Survey Statistics (NIS) and the R4: December 2020– R4: May 2021 Cambodia Chamber of January 2021 Commerce R5: Mar 2021 Indonesia R1: May 2020 R1: June 2020 Subsample of the 2018 Sample from R2: May–June 2020 R2: October–November 2020 Urban Perception, 2018 Manufacturing Directory R3: July–August 2020 R3: March 2021 Rural Poverty, and the 2020 2017 and Economic Census R4: November 2020 Digital Economy Household Directory 2016 R5: Mar 2021 surveys R6: October 2021 Malaysia R1: May–June 2021 R1: October 2020 Random digit dialing Sample from R2: October 2021 R2: January–February 2021 from national numbering comprehensive listing plans list by Malaysian of businesses in all Communications and sectors and sizes, across Multimedia Commission Peninsular and East (provided by the survey firm) Malaysia Mongolia R1: May 2020 R1: August 2020 Subsample of 2018 Sample of the population R2: September 2020 R2: February 2021 Household Socio–Economic of formal businesses with R3: December 2020 Survey five or more employees R4: April 2021 obtained from enterprise R5 June 2021 survey Philippines R1: August 2020 R1: July 2020 Phone survey based on Online survey reweighted R2: December 2020– R2: November– survey firm sample frame; with data from 2018 Listing January 2021 December 2020 self–administered web of Establishments R3: May 2021 R3: May 2021 surveys Vietnam R1: June–July 2020 R1: June–July 2020 Subsample from Sample from 2018 R2: July–August 2020 R2: September–October 2020 representative 2018 survey Establishment Census R3: September 2020 R3: January–February 2021 (GSO) Statistical Office R4: January 2021 R5: March 2021 Source: Based on World Bank’s high-frequency phone surveys, business pulse surveys and enterprise survey for the six respective countries. Note: R = Round 36 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE IA.1  Timing of surveys conducted by the World Bank May-20 Jun-20 Jul-20 Aug-20 Sep-20 Oct-20 Nov-20 Dec-20 Jan-21 Feb-21 Mar-21 Apr-21 May-21 Jun-21 Jul-21 Aug-21 Sep-21 Oct-21 Household R1 R2 R3 R4 R5 Cambodia Firm R1 R2 R3 R4 Household R1 R2 R3 R4 R5 R6 Indonesia Firm R1 R2 R3 Household R1 R2 Malaysia Firm R1 R2 Household R1 R2 R3 R4 R5 Mongolia Firm R1 R2 Household R1 R2 R3 Philippines Firm R1 R2 R3 Household R1 R2 R3 R4 R5 Vietnam Firm R1 R2 R3 Source: Based on World Bank’s high-frequency phone surveys, business pulse surveys and enterprise surveys for the six respective countries. There were approximately 58,000 interviews carried out to compile the firm-level data. For firm-level analysis in the introduction, for reasons of data compatibility and completeness, data for rounds 1 through 4 of the business pulse surveys were used. However, an additional round of business pulse survey data for Vietnam, round 5, was used in that country chapter. Finally, another round of business pulse survey data was to be collected in Indonesia by the end of 2022 but was unavailable for this report. INTRODUCTION 37 References Acharya, Viral, and Sascha Steffen. 2020. “‘Stress Tests’ for Banks as Liquidity Insurers in a Time of COVID.” VoxEU, March 22, https://cepr.org/voxeu/columns/stress-tests-banks-liquidity-insurers-time-covid. Agrawal, Sarthak, Alexandra Cojocaru, Veronica Montalva, Ambar Narayan, Tom Bundervoet, and Andrey Ten. 2021. COVID-19 and Inequality: How Unequal Was the Recovery from the Initial Shock? Washington, DC: World Bank. https://openknowledge.worldbank.org/bitstream/handle/10986/35867/COVID-19-and-Inequality-How-Unequal- Was-the-Recovery-from-the-Initial-Shock.pdf. Apedo-Amah, Marie Christine, Besart Avdiu, Xavier Cirera, Marcio J. Cruz, Elwyn Davies, Arti Grover, Leonardo Iacovone, Umut Kilinc, Denis Medvedev, Franklin Okechukwu Maduko, Stavros Poupakis, Jesica Torres Coronado, and Trang Thu Tran. 2020. “Unmasking the Impact of COVID-19 on Businesses: Firm Level Evidence from across the World.” Policy Research Working Paper 9434, World Bank, Washington, DC. http://documents.worldbank.org/ curated/en/399751602248069405/Unmasking-the-Impact-of-COVID-19-on-Businesses-Firm-Level-Evidence- from-Across-the-World. Banerjee, Ryan, and Enisse. Kharroubi. 2020. “The Financial Vulnerabilities Driving Firms to the Exit,” BIS Quarterly Review, December 2020. https://www.bis.org/publ/qtrpdf/r_qt2012e.pdf. Flynn, Burton, Mikaal. Homanen, and Thorsten Beck (eds). 2020. “COVID-19 in Emerging Markets: Firm-Survey Evidence,” Covid Economics: Vetted and Real-Time Papers 38, July. https://cepr.org/node/390565. Cerutti, Eugenio, Stijn Claessens, and Luc Laeven. 2017. “The Use and Effectiveness of Macroprudential Policies: New Evidence.” Journal of Financial Stability 28: 203–24. https://doi.org/10.1016/j.jfs.2015.10.004. Cirera, Xavier, Marcio Cruz, Elwyn Davies, Arti Grover, Leonardo Iacovone, Jose Ernesto López Córdova, Denis Medvedev, Franklin. Okechukwu Maduko, Gaurav Nayyar, Santiago Reyes Ortega, Jesica Torres. 2021. “Policies to Support Businesses through the COVID-19 Shock: A Firm-Level Perspective.” Policy Research Working Paper, No. 9506. World Bank, Washington, DC. https://documents1.worldbank.org/curated/en/188461610547807479/pdf/ Policies-to-Support-Businesses-through-the-COVID-19-Shock-A-Firm-Level-Perspective.pdf. Dai, Bibing., Di Fu, Guangteng Meng, Bingsheng Liu, Qi Li, and Xun Liu . 2014. “The Effects of Governmental and Individual Predictors on COVID-19 Protective Behaviors in China: A Path Analysis Model.” Public Administration Review 80(5): 797–804. https://doi.org/10.1111/puar.13236. Demmou, Lilas, Guido Franco, Sara Calligaris, and Dennis Dlugosch. 2021a. “Liquidity Shortfalls During the COVID-19 Outbreak: Assessment of Risks and Policy Responses.” OECD Economics Department Working Paper 1647, Organization for Economic Cooperation and Development, Paris. https://t4.oecd.org/officialdocuments/ publicdisplaydocumentpdf/?cote=ECO/WKP(2021)1&docLanguage=En. Demmou, Lilas, Sara Calligaris, Guido Franco, Dennis Dlugosch, Müge Adalet McGowan, and Sahra Sakha. 2021b. “Insolvency and Debt Overhang Following the COVID-19 Outbreak: Assessment of Risks and Policy Responses.” OECD Economics Department Working Paper 1651, Organization for Economic Cooperation and Development, Paris. https://read.oecd-ilibrary.org/view/?ref=916_916885-2mmnsxy9du&title=Insolvency-and-debt-overhang- following-the-COVID-19-outbreak-assessment-of-risks-and-policy-responses. Fairlie, Robert. 2020. “The Impact of COVID-19 on Small Business Owners: Evidence from the First 3 Months after Widespread Social-Distancing Restrictions.” Journal of Economics & Management Strategy. Winter;29(4):727–740. https://doi.org/10.1111/jems.12400. Fuchs, Alan, Mariano Sosa, and Mathew Wai-Poi. 2021. “Progressive Resources Mobilization for a COVID-19 Recovery.” Policy Note, Washington, DC: World Bank. Gentilini, Ugo, Mohamed Almenfi, Ian Orton, and Pamela Dale. 2022. Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures. World Bank, Washington, DC. https://openknowledge.worldbank.org/ handle/10986/33635. Halim, Daniel Zefanya, Hillary C. Johnson, Elizaveta. Perova. 2017. “Could Childcare Services Improve Women’s Labor Market Outcomes in Indonesia?” East Asia and Pacific Gender Policy Brief, No. 1. World Bank, Washington, DC. http://documents.worldbank.org/curated/en/855851490958133680/Could-childcare-services-improve-women- s-labor-market-outcomes-in-Indonesia. IMF (International Monetary Fund). 2021. World Economic Outlook: Recovery During a Pandemic: Health Concerns, Supply Disruptions, Price Pressures. Report, October 2021. Washington, DC: IMF. https://www.imf.org/en/Publications/ WEO/Issues/2021/10/12/world-economic-outlook-october-2021. 38 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA International Labour Organization. 2022. ILO modelled estimates database, ILOSTAT Labour Force Statistics (LFS). Available from https://ilostat.ilo.org/data/. Kannan, Prakash. 2009. “The Lingering Effects of Financial Crises.” VoxEU.org, November 19, 2022. https://cepr.org/ voxeu/columns/lingering-effects-financial-crises. Karalashvili, Nona, and Domenico Viganola. 2021. “The Evolving Effect of COVID-19 on the Private Sector.” Global Indicators Briefs, No. 1. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35869. Kim, Lydia. Y., Maria Ana Lugo, Andrew D. Mason, and Ikuko Uochi. 2021. “Inequality under COVID-19: Taking Stock of High-Frequency Data for East Asia and the Pacific.” Policy Research Working Paper WPS 9859, World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/36635. Kose, M. Ayhan, S. Kurlat, F. Ohnsorge, and N. Sugawara. 2017. “A Cross-Country Database of Fiscal Space.” Policy Research Working Paper 8157, World Bank, Washington, DC. https://openknowledge.worldbank.org/ handle/10986/27964. O’Keefe, Phillip and colleagues. Forthcoming. Diverse Paths: The Dynamic Evolution of Social Protection in Asia and the Pacific. World Bank, Washington, DC. Record, Richard, J. L., Achim D. Schmillen, Kenneth Simler, Bradely R. Larson, Mahama A.S.S Bandaogo, Norman V. Loayza, Rajni Bajpai, Shakira B. Teh Sharifuddin, Smita. Kuriakose, and Yew K. Chong. 2021. Aiming High: Navigating the Next Stage of Malaysia’s Development. Country Economic Memorandum. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35095. UNU-WIDER Government Revenue Dataset’. Version 2022. https://doi.org/10.35188/UNU-WIDER/GRD-2022. World Bank. 2021a. World Bank East Asia and Pacific Economic Update, April 2021: Uneven Recovery. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/35272. World Bank. 2021b. World Bank East Asia and Pacific Economic Update, October 2021: Long COVID. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/36295. World Bank. 2021c. Macro Poverty Outlook. October. Washington, DC: World Bank. World Bank. 2021d. Cambodia Economic Update: Living with COVID–Special Focus: The Impact of the COVID-19 Pandemic on Learning and Earning in Cambodia. Washington, DC: World Bank. http://documents.worldbank.org/curated/ en/099350012062137172/P1773400f35a770af0b4fa0781dffcd517e. World Bank. 2021e. Non-Performing Loans in East Asia and the Pacific: Practices and Lessons in Times of COVID-19. World Bank, Kuala Lumpur. http://hdl.handle.net/10986/36522. World Bank. 2022a. Global Economic Prospects, January 2022. Washington, DC: World Bank. World Bank. 2022b. From the Last Mile to the Next Mile–2022 Vietnam Poverty and Equity Assessment. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/37952. World Bank. 2022c. Braving the Storms: East Asia and Pacific Economic Update, April 2022. World Bank. Washington, DC. https://www.worldbank.org/en/region/eap/publication/braving-the-storms-global-launch-of-the-east-asia-and- pacific-economic-update-april-2022. World Bank. 2022d. Poverty and Shared Prosperity Report 2022: Correcting Course.. Washington, DC: World Bank. World Bank. 2022e. World Development Report 2022: Finance for an Equitable Recovery. Washington, DC: World Bank. https://www.worldbank.org/en/publication/wdr2022. World Bank. Forthcoming a. Inequality in the Philippines: Past, Present & Perspectives for the Future. Washington, DC, World Bank. World Bank. Forthcoming b. Mongolia 2020 Poverty Report: A Decade of Progress and Stagnation in Poverty Reduction. Washington, DC: World Bank. WDI 2022 World Development Indicators. Washington, D.C.: The World Bank. WTO (World Trade Organization), ITC (International Trade Center), and UNCTAD (United Nations Conference on Trade and Development). 2017. World Tariff Profiles 2017. Geneva: WTO. https://www.wto.org/english/res_e/ booksp_e/tariff_profiles17_e.pdf. CHAPTER 1 Cambodia by Kyung Min Lee, Isabelle Salcher, Wendy Karamba, Kimsun Tong, and Trang Thu Tran At the onset of the pandemic, Cambodia recorded fewer cases of COVID-19 and related deaths than other countries in the region. To contain COVID-19, Cambodia’s authorities implemented stringent measures, including, but not limited to, banning travel from high-incidence countries, closing schools, and restricting religious activities and gatherings of more than 50 people (World Bank 2020). Citizens largely complied with health directives, including minimizing social interactions and staying at home. Despite relatively few cases of COVID-19 cases at the onset, the pandemic had large impacts on the Cambodian economy, which prompted unprecedented policy interventions. In addition to stringent containment measures, the shock to domestic and global demand triggered by the pandemic hit the manufacturing, tourism, construction, and real estate sectors—Cambodia’s growth engines—hard, leading to the country’s first economic contraction in 25 years. To support firms and households during this crisis, the government spent 3.0 percent of GDP in 2020 and an estimated 4.9 percent in 2021. In the formal sector, most government support took the form of fiscal exemptions, rent deferrals, and wage subsidies. These interventions largely excluded microenterprises and small firms, firms outside the garment or tourism sector, and informal firms. Given this limited coverage and the pervasiveness of informality in Cambodia, scaled-up social assistance played an important role in reaching broad parts of the population experiencing pandemic-induced employment and income shocks. A cash transfer program targeting poor and vulnerable households was the largest social assistance instrument implemented during the pandemic. This chapter examines the effects of the COVID-19 pandemic on employment in Cambodia and the policy response to support households and firms during the crisis. It draws on five rounds of a household- level survey and four rounds of a firm-level phone survey. The pandemic created an urgent need for timely data to help monitor effects on firms and households and inform policy response (see Annex 1). In Cambodia, the World Bank conducted a business pulse survey (BPS) and high-frequency phone surveys (HFPSs) of households between May 2020 and May 2021, with asynchronous timing of household- and firm-level data collection (Figure 1.1 and Annex 1A). The firm survey provides information on firms’ operational status, adjustment strategies, and receipt of public support during the COVID-19 crisis. The household survey covers topics such as employment and income, coping strategies, and public safety nets. The two surveys complement each other by shedding light on labor demand and supply. 39 40 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 1.1  Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Cambodia, 2020–22 HFPS R1 HFPS R2 HFPS R3 HFPS R4 HFPS R5 80 125 Inverted mobility index (14-day average) Stringency index (14-day average) 60 100 Stringency index 40 75 20 50 Inverted Mobility index 0 25 –20 0 BPS R1 BPS R2 BPS R3 BPS R4 6 COVID-19 cases/deaths (per 100,000 people) COVID-19 cases 4 2 COVID-19 deaths 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Sources: Mobility data comes from Google Mobility Reports, April 2022. Stringency data comes from the Oxford COVID-19 Government Response Tracker (OxCGRT); COVID-19 cases and deaths were derived from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. Note: The mobility index is the average percentage change in time spent at retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. For each day of the week, the baseline is defined as the median value from the five-week period January 3–February 6, 2020. The figure shows 14-day rolling averages of the mobility index, which is inverted so that higher values indicate lower mobility. The stringency index records the strictness of lockdown-style policies; 14-day rolling averages are shown. COVID-19 cases and deaths are smoothed by a seven-day average. Phases of the pandemic are determined by identifying structural breaks in COVID-19 cases and the mobility index. See the introduction for further details. Timeline of the COVID-19 Pandemic and Government Measures After detecting its first case of COVID-19, on January 27, 2020, Cambodia experienced three phases of the pandemic, characterized by the number of cases and the stringency of the government’s measures (Figure 1.1): ⦁ Phase 1 (February–May 2020): Introduction of restrictive measures after the first case was detected ⦁ Phase 2 (June 2020–January 2021): Easing of mobility restrictions as the number of cases fell and reintroduction of some restrictions as cases rose ⦁ Phase 3 (February–December 2021): Reinstatement of restrictive measures as cases resurged In phase 1 of the pandemic, the government adopted stringent containment measures early on to curb the spread of COVID-19 (Table 1B.1). Beginning in March 2020, the government closed educational institutions nationwide, restricted international arrivals and domestic travel, closed garment factories and CHAPTER 1: CAMBODIA 41 entertainment venues, imposed restrictions on public gatherings, and cancelled major public holidays (World Bank 2020). It also conducted massive COVID-19 testing, complemented with contact tracing. These containment measures sharply reduced mobility, as citizens largely complied with health directives, including minimizing social interactions and staying at home. Cambodia also received significant financing support from international health partners to help strengthen its healthcare system (Nit and colleagues, 2021). In phase 2, the government lifted most restrictions, given the limited spread of infection. Travel bans were lifted, and most schools had reopened by September 2020, through a phased approach. Mobility slowly picked up, although it remained below the January 2020 baseline level. After the first known community transmission cluster was detected in Phnom Penh in November 2020, schools in Phnom Penh and Kandal province closed again. The government also reinstated some restrictions on social gatherings and activities. Cambodia contained the COVID-19 outbreak through February 2021, recording only about 500 infections. It reported no COVID-19-related deaths until March 2021. In phase 3, COVID-19 cases surged, and the government reinstated strict mobility restrictions while racing to vaccinate the populace. Following an outbreak in Phnom Penh in February 2021, Cambodia registered almost 20,000 cases in less than three months.10 In response, high-risk cities introduced an 8 pm–5 am curfew, and Cambodia’s first lockdown was later implemented in these areas. All nonessential travel and business activity ceased. Over the next months, new virus variants contributed to growing case numbers, and restrictive measures—including suspended border travel, curfews, and lockdowns—remained in place during much of 2021, even as new infections started to decline. Launched in early February 2021, Cambodia’s vaccination program could not prevent the large outbreak later that month but permitted the country to work toward reopening later that year (World Bank 2021). As of mid-February 2022, almost 82 percent of Cambodians had been fully vaccinated against the virus, one of the highest coverage rates in the region. Mobility gradually increased, as global and domestic stringency measures were relaxed, yet by mid-February 2022, Cambodia had confirmed over 120,000 cases and over 3,000 deaths in total. Impacts on the Economy As a result of the pandemic, Cambodia’s economy contracted for the first time in 25 years. Real gross domestic product (GDP), which had grown by 7.1 percent in 2019, contracted by 3.1 percent in 2020. The shock to and global demand triggered by the pandemic, in conjunction with stringent measures to contain its spread, hit the manufacturing, tourism, construction, and real estate sector hard. Serving as the main engines of growth, these sectors accounted for more than 70 percent of GDP growth and about 40 percent of total paid employment in 2019 (World Bank 2020). The travel and tourism sector provided an estimated 2 million jobs and constituted a quarter of GDP before the pandemic. Because of travel restrictions, this sector virtually collapsed, with international arrivals falling 80  percent in 2020. Exports of garments, footwear, and travel goods—Cambodia’s key merchandise exports—contracted 8.1 percent in 2020, as garment factories closed and global demand weakened. At the same time, the European Union partially withdrew  Data on COVID-19 cases and deaths come from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) 10 at Johns Hopkins University. 42 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA tariff preferences.11 Steel imports fell 35 percent, and cement imports 49 percent in 2020, as a result of contractions in the construction and real estate sector, which had previously contributed over a third of real growth in 2019 (World Bank 2021). Coupled with Cambodia’s heavy reliance on a few sectors, the high prevalence of informality increased Cambodia’s vulnerability to the COVID-19 shock and slowed its recovery. The pandemic affected employment in both formal and informal sectors. Declines in mobility profoundly affected the services sector, and COVID-19-related business closures impacted livelihoods. Disruptions to employment and incomes weakened consumer demand domestically and abroad. The informal sector—which provided about 88 percent of all employment in Cambodia in 2019 (National Institute of Statistics and Ministry of Planning 2020)—experienced a larger decline in employment and a slower recovery. Before the pandemic, the agricultural sector accounted for 37 percent of informal employment in Cambodia, the wholesale and retail trade sector for 18 percent, manufacturing for 13 percent, and construction for 11 percent. Because of their unofficial status, informal firms also faced challenges receiving the public support intended to help them cope with, and recover from, the pandemic-induced shock. For example, tax exemptions—a central, firm-targeted policy instrument during the pandemic—did not reach informal businesses that were not registered in the business registry and did not pay taxes. Cambodia’s recovery remains fragile. Cambodia’s real GDP is estimated to have grown 2.2 percent in 2021, despite a COVID-19 outbreak earlier in the year that slowed the recovery that had begun in late 2020. The reopening of the economy beginning in November 2021—made possible by Cambodia’s high vaccine coverage—improved growth prospects. Growth is projected to reach 4.5 percent in 2022, with traditional growth drivers and the agricultural sector underpinning the recovery (World Bank 2021). However, travel and tourism remain subdued. A slowdown in global demand could also hurt export-oriented sector, which are Cambodia’s growth driver, and lead to a much slower recovery of the tourism sector (World Bank 2021). Employment Impacts: Shocks and Recovery Even before the pandemic, Cambodia already had a significantly less-favorable business environment than its regional peers (WEF 2019).12 Relatively weak competitive forces in the product market, both domestically and via international trade, add to the low business dynamism and limit creative disruption in the present market. Time and costs to start a new business in the country are among the highest in the world. While hiring and firing practices and access to foreign professionals provide adequate labor flexibility, broader institutional development, in the form of public and corporate governance, intellectual property protection, and systems of checks and balances, is still limited. 11  The withdrawal of preferential access to the EU market—which went into effect on August 12, 2020— affects about 20 percent of Cambodia’s exports to the European Union, worth over US$1 billion (EC 2020). Cambodia’s garment and footwear exports are estimated to have declined by US$510 million, or 5.4 percent in 2018 as a result. In 2018, Cambodia exported US$9.5 billion of garment and footwear products, a third of which were imported by the European Union. The estimated decline in Cambodia’s milled rice exports to the EU market was US$65.1–144.1 million, or 36.6– 81.2 percent, according to the World Bank (2019). 12  World Economic Forum (WEF) 2019. The Global Competitiveness Report, 2019. Available online at: https://www3.weforum.org/docs/ WEF_TheGlobalCompetitivenessReport2019.pdf. CHAPTER 1: CAMBODIA 43 In phase 1, the onset of the COVID-19 pandemic disrupted formal firms’ operations and business revenues. In addition to stringency measures, weakened global and domestic demand hit the tourism and export-oriented manufacturing sector hard. These shocks led to widespread business closures and revenue declines. Data from the BPS conducted in June 2020 shows that 19 percent of formally registered firms reported that they had either permanently or temporarily suspended their business operations. Restrictive social distancing measures also affected business revenue. In June 2020, about 9 in 10 formally registered firms reported a decline in sales during the previous 30 days (Figure 1.2). Data from the Cambodia HFPS of households shows that the pandemic had a similar impact on largely informal household enterprises. In May 2020, about 8 in 10 household businesses reported earning no revenue or less revenue than the preceding month, with weak consumer demand cited as the main driver of revenue losses. The pandemic also adversely affected informal firms, most of which operate in the wholesale and retail sector. In May 2020, 87 percent of household businesses in these sectors reported no revenue or reduced revenue relative to the previous month. Mobility declined by more than 40 percent from the pre-pandemic norm before the first restrictions were imposed in March 2020. In addition to weakened consumer demand, these restrictions likely profoundly affected the wholesale and retail sector, which accounts for two-thirds of all non-farm household businesses in Cambodia. Household businesses operate in personal services, construction, transportation, and the manufacturing sector to a much lesser extent. None of the surveyed household businesses were active in the hospitality sector. Shocks to firm operations and revenues disrupted employment in phase 1 of the pandemic. In May 2020, it was estimated that by causing Cambodia’s main growth drivers to collapse, the pandemic put at least FIGURE 1.2  Firms experiencing changes in sales revenue during the 30 days before the survey, 2020–21 87.3 87.5 80 71.4 change in revenue in previous 30 days 60 Percent of firms experiencing 49.4 41.0 40 18.8 20 10.1 9.8 9.6 9.7 2.6 2.8 0 June 2020 September 2020 January/February 2021 May 2021 Increased Remained the same Decreased Source: World Bank data sourced from business pulse survey rounds 1–4. Note: Sample includes firms that were formally registered with a government authority at the time of the survey. 44 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA 1.76 million Cambodian jobs directly at risk (World Bank 2020). Of the 81 percent of formal firms that were still operating in June 2020, 37 percent reported that they had reduced the number of employees in the previous 30 days. Large employment reductions were not limited to the formal sector. According to the HFPS of households, which captured both formal and informal sector employment, 14 percent of adults stopped working in May 2020, when strict social distancing measures were in place (Figure 1.3). Among individuals who stopped working, 60 percent cited reasons related to COVID-19 restrictions, such as business closures, temporary layoffs, and movement restrictions. The pandemic also affected less-educated workers more than better-educated workers: 22 percent of adults with incomplete primary education had stopped working, but only 12 percent of adults with complete primary education and 11 percent of adults with secondary education or higher had done so (Figure 1.4). Work stoppages did not vary with gender or urban/rural status in May 2020. Household labor income declined as a consequence of employment disruptions. Nearly all livelihoods in Cambodia depend on labor income, especially from wage employment, non-farm businesses, or farming. Of these households, 79 percent experienced a decline in labor income between the COVID-19 outbreak and May 2020 (Figure 1.5). Remittances and assistance from friends and family were also negatively affected, although few households drew income from these sources. Initial labor income losses were more prevalent among households identified as poor by the Identification of Poor Households (IDPoor), Cambodia’s national poverty identification program.13 These households are more likely to rely on wage employment than wealthier households, which in turn are more likely to operate a non-farm household business. In May 2020, 79 percent of IDPoor wage households reported lower wage earnings since the COVID-19 outbreak; for all wage households in Cambodia, the figure was 63 percent. This difference could reflect the fact that poorer workers disproportionately suffered from job losses in the manufacturing and construction sectors. FIGURE 1.3  Reasons why workers in Cambodia stopped working, 2020–21 25 Percent of respondents 20 17 14 15 15 12 11 10 5 0 May 2020 August–September October–November December 2020– March 2021 2020 2020 January 2021 COVID-19 restrictions Seasonality Other Source: World Bank data based on high-frequency phone surveys of households rounds 1–5. Note: The sample includes respondents 18 years and older from the World Bank’s Living Standards Measurement Study Plus. Figures show the share of respondents who stopped working. COVID-19 restrictions include business/government closures, furloughs (temporary layoffs), and inability to reach a farm because of mobility restrictions. “Other” includes business closures for reasons other than COVID-19, including layoffs, illness, or quarantine; the need to care for an ill relative; vacation; retirement; lack of farming inputs; flooding; among others. For May 2020, it is the share of respondents who had worked before the COVID-19 outbreak, but not in the seven days preceding the survey. For August/September 2020–March 2021, it is the share of respondents who had stopped working since the previous round of the HFPS.  IDPoor uses a proxy means test implemented by community members to identify households living in poverty; identified poor households 13 qualify for various social services. CHAPTER 1: CAMBODIA 45 FIGURE 1.4  Work stoppages in Cambodia, by level of education 25 22 23 adults with education level) 19 20 18 Work stoppages (% of 16 14 15 15 15 14 12 11 12 9 10 10 8 5 0 May 2020 August–September October–November December 2020– March 2021 2020 2020 January 2021 Incomplete primary Complete primary Complete lower-secondary/higher Source: World Bank data based on high-frequency phone survey of households rounds 1–5. Note: The sample includes respondents 18 years and older from the World Bank’s Living Standards Measurement Study Plus. Figures show the share of respondents who stopped working by the highest education level completed in the household. For May 2020, it is the share of respondents who had worked before the COVID-19 outbreak but not in the seven days preceding the survey. For August/September 2020 to March 2021, it is the share of respondents who had stopped working since the previous round of the HFPS. By hitting poor workers and households harder than those that were better-off, the initial COVID-19 shock may have increased poverty and inequality. In phase 2 of the pandemic, firms’ operating status and revenues improved, as mobility picked up and businesses reopened. As COVID-19 cases remained low, most restrictions were lifted between mid-2020 and early 2021. The share of permanent and temporary business closures fell below 10 percent in September 2020 and February 2021. At the same time, the share of formal firms reporting lower revenue declined to slightly below 50 percent (Figure 1.2). Business revenues also increasingly stabilized in the informal sector. In March 2021, half of non-farm household enterprises reported no revenue or less revenue than in the previous month, down from 64 percent in August 2020. As the government eased social distancing measures in phase 2, employment began to recover, although it remained below pre-pandemic levels. Most businesses were open by September 2020, but work stoppages FIGURE 1.5  Loss of labor income in Cambodia, 2020–21 100 Percent of households that lost 79 80 62 labor income 60 53 47 45 40 20 0 May 2020 August–September October–November December 2020– March 2021 2020 2020 January 2021 Source: World Bank data based on high-frequency phone survey of households rounds 1–5. Note: The sample includes households from the World Bank’s Living Standards Measurement Study Plus who had earned labor income in the last 12 months. Labor income includes earnings from wage employment, non-farm family businesses, and family farming. Figures show the share of households who reported a reduction in labor income. For May 2020, it is the share of households who experienced a labor income loss since the outbreak of the COVID-19 pandemic. For August/September 2020 to March 2021, it is the share of households who saw declines in labor income since the previous round of the HFPS. 46 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA remained at similar levels throughout 2020 and early 2021. However, COVID-19-related reasons for stopping work largely disappeared (Figure 1.3). Instead, flooding in early October 2020, which affected 14 provinces, including Phnom Penh, disrupted work activities. In December 2020, worker reports of COVID-19-related work stoppages rose again, possibly as a result of a local COVID outbreak in November, reaching levels reported in August. Differences in work stoppages by educational attainment narrowed in phase 2 (Figure 1.4). In August 2020, women were more likely to have stopped working than men; in December 2020, rural residents were twice as likely to have stopped working as urban dwellers. Otherwise, work stoppages did not statistically differ by demographic factors in phase 2. In the formal sector, employment in February 2021 remained 25 percent below the pre-pandemic level, suggesting that employment did not fully recover even after restrictions were relaxed, and that initial business closures and a persistent demand shock might have led to continued workforce reductions.14 Widespread reductions in household labor income and business revenues continued in phase 2 of the pandemic, but a diminution of losses became apparent as economic activity showed signs of recovery. As restrictions were lifted and mobility slowly picked up, fewer households reported labor income losses. About 45 percent of households saw declines in labor income between December 2020/January 2021 and March 2021, down from 47 percent between October/November 2020 and December 2020/January 2021, 53 percent between August/September 2020 and October/November 2020, and 62 percent between May 2020 and August/September 2020 Poorer households did not persistently experience higher labor income losses in phase 2, limiting the negative distributional consequences of the initial COVID-19 shock. In phase 3, strict restrictions imposed after the COVID-19 outbreak in early 2021 abruptly stalled the economic recovery process that had begun in late 2020. Formal business closures shot up to 30 percent in May 2021, a larger share than in phase 1, when social distancing restrictions were first implemented. The share of formal firms reporting a decrease in sales, 88 percent, returned to the level reported in June 2020. The new outbreak and reintroduction of stringent mobility restrictions also disrupted employment, reducing it to 30 percent below pre-pandemic levels in May 2021.15 Formal firms in the travel and tourism sectors—sectors that the pandemic had brought to a standstill and therefore disproportionately affected—had not recovered by phase 3. In May 2021, more than half of formally registered firms in the hospitality sector had permanently or temporarily stopped operating, compared with just 20 percent of firms in the manufacturing and retail or wholesale trade sectors. The hospitality and other services sectors—sectors susceptible to external shocks and subject to government restrictions implemented to curb the spread of COVID-19—also experienced larger employment reductions than other sectors (Figure 1.6). In response to the surging COVID-19 cases and reinstated social distancing measures, which reduced sales revenues, formal firms implemented various labor adjustments in phase 3. Many firms reduced wages Eit − Eit−1  Following Davis, Haltiwanger, and Schuh (1996), hereafter referred to as DHS, employment growth was computed as g it = 14 . 0.5 ( Eit + Eit−1 ) 15  No data on employment disruptions from the perspective of workers is available. The latest round of the HFPS was conducted before the lockdown in Phnom Penh, and adjacent Ta Khmau City required all nonessential businesses to cease operation in response to the large-scale COVID-19 outbreak. CHAPTER 1: CAMBODIA 47 FIGURE 1.6  Change in employment relative to before the COVID-19 pandemic in Cambodia, by sector 0 Percent change in employment relative to baseline −10 –17 −20 –21 –24 –25 –26 −30 –28 −40 –40 −50 –48 January/February 2021 May 2021 Manufacturing Retail/wholesale Hospitality Other services Source: World Bank data based on business pulse survey rounds 3 and 4. Note: The sample includes firms that were formally registered with a government authority at the time of the survey. Baseline is before the pandemic. “Other services” includes agro- industry, construction, and utilities. DHS growth was used to compute the employment changes relative to the baseline (before COVID-19) to address entry and exit. and/or hours or days of work. In May 2021, over a third of formally registered firms had granted their workers leaves of absence, and about 20 percent of firms had cut workers’ hours and/or wages (Figure 1.7). Some 52 percent of firms had implemented at least one form of intensive margin adjustment, such as granting paid or unpaid absences, cutting wages, or cutting hours worked, in the previous 30 days, indicating that even workers who remained employed were still affected by the COVID shock. In contrast, only about 5 percent of formally registered firms reduced the size of their workforce in the month preceding the May 2021 survey. This figure does not indicate whether firms reduced the number of employees immediately after the large- scale COVID-19 outbreak in February triggered social distancing restrictions. In contrast, most firms did not adjust their use of digital technology or expand their work-from-home practices when COVID-19 cases soared in phase 3. In May 2021, over 60 percent of formal firms either did not increase their use of digital technology, at 34 percent, or did not use such technology at all, at 28 percent (Figure 1.8). Although 34 percent of firms increased the intensity of their use of digital technology, only 4 percent of firms adopted it for the first time, suggesting that barriers to adopting new digital technology may be higher than barriers to increasing its use. Work-from-home practices also did not greatly expand over time. In June 2020, 18 percent of formally registered firms allowed at least some of their employees to work-from- home. A year later, remote work had not become more common, even as new COVID-19 cases increased rapidly. In May 2021, about the same proportion of formal firms, or 19 percent, enabled at least some of their employees to work remotely. Although not all jobs and professional activities are suitable for remote work, some firms likely lack the required technological capabilities and equipment to facilitate work-from-home practices. Pandemic-induced changes in employment may affect long-term productivity and inequality. In Cambodia’s formal sector, changes in employment were positively correlated with labor productivity—that is, low-productivity firms were more likely to reduce employment, and high-productivity firms were more 48 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 1.7  Share of firms in Cambodia that granted leaves of absence, cut hours, or cut wages in the 30 days before the survey, May 2021 40 37 30 Percent of firms (%) 20 20 18 10 0 Granted paid or unpaid leave of absence Cut hours Cut wages Source: World Bank business pulse survey round 4. Note: Sample includes firms that were formally registered with a government authority at the time of the survey. FIGURE 1.8  Firm adoption or increase in use of digital technology in Cambodia, May 2021 40 34 34 Percent of firms that started or increased the use of digital technologies (%) 30 28 20 10 4 0 Started Increased Used but Did not use did not increase Source: World Bank business pulse survey round 4. Note: Sample includes firms that were formally registered with a government authority at the time of survey. likely to maintain or increase employment (Figure 1.9). In the short term, employment reductions in low- productivity firms can hurt the workers employed by these firms. If these newly unemployed, low-skilled/ low-paid workers struggle to be reemployed in productive jobs, this pattern of labor reallocation could exacerbate long-term inequality. Workers were more likely to respond to the COVID-19 employment shock by not working than by switching jobs. Between May 2020 and March 2021, less than 10 percent of working adults had recently changed jobs (Figure 1.10). Job switching rates were lowest in October 2020 and March 2021, at no more than 3 percent. Work stoppages exceeded job switching rates in all survey rounds; in some rounds, they were CHAPTER 1: CAMBODIA 49 FIGURE 1.9  Correlation between changes in employment and labor productivity in Cambodia 0 Change in employment relative to baseline −20 −40 −60 −80 4 6 8 10 12 Labor productivity at baseline (log) Source: World Bank data from business pulse survey rounds 3 and 4. Note: The sample includes firms that were formally registered with a government authority at the time of the survey. DHS growth was used to compute the employment changes relative to the baseline (before COVID-19) to address entry and exit. Labor productivity is defined as the log of sales per worker at the baseline. twice as high. Job switching was mostly unrelated to the worker’s educational attainment or the household’s poverty status. At the beginning of the pandemic, in May 2020, nearly two-thirds of workers changing jobs also switched sectors, most commonly leaving the service sector—personal services, wholesale and retail trade—and entering the agricultural sector. Workers continued to shift into agriculture during most of 2020 and early 2021. These findings suggest that workers moved from more productive sectors to less productive ones, which could worsen long-term inequality. FIGURE 1.10  Share of working adults in Cambodia who switched jobs or sectors, May 2020–March 2021 15 Percent of working adults who switched jobs 10 8 6 6 5 2 3 0 May 2020 August–September October–November December 2020– March 2021 2020 2020 January 2021 Switched sectors Switched jobs but not sector Source: World Bank data based on high-frequency phone surveys of households rounds 1–5. Note: The sample includes respondents 18 years and older from the World Bank’s Living Standards Measurement Study Plus. Figures show the share of currently working respondents who changed jobs. Workers who changed jobs are distinguished into switched jobs vs. sectors. A change in sector is defined based on aggregate economic sectors: agriculture, industry, and services. For May 2020, it is the share of respondents who switched jobs since the COVID-19 outbreak. For August/September 2020 to March 2021, it is the share of respondents who switched jobs since the previous round of the HFPS. 50 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Fiscal Support to Firms and Households Size and composition of the fiscal response To mitigate the adverse socioeconomic impacts of the pandemic on both firms and household, the government spent 3 percent of GDP in 2020 and an estimated 4.9 percent in 2021 (Table 1.1). In addition to scaling up healthcare systems to prevent and respond to the COVID-19 outbreak, the government channeled economic interventions to firms—generally formal in nature—and expanded social assistance to workers and households. TABLE 1.1  Select policy instruments Cambodia used to mitigate effects of the COVID-19 pandemic, 2020–21 2020 2021 Estimated Intervention Description Plan Disbursed Plan disbursement Public health and social intervention Health masterplan: outbreak Improved prevention and detection facilities, clinical 0.39 0.11 0.10 2.48 prevention and treatment management and treatment, coordination and supporting system Support to firms Financing through the Provided low-interest loans for working capital and 0.19 0.18 0.00 0.00 Agricultural and Rural investment in agricultural sector through capital Development Bank (ARDB) injections into the ARDB Co-financing through the In conjunction with commercial banks, provided low- 0.19 0.18 0.00 0.00 Small and Medium-Size interest loans for working capital and investment in six Enterprise (SME) Bank SME sectors through the newly established SME Bank Credit guarantee fund Provided capital to establish the Credit Guarantee 0.77 0.73 0.00 0.00 Corporation of Cambodia, which helps bear risk- sharing with businesses SME financing facility Reserved contingent funds to provide financing to SME 1.16 0.11 0.90 0.50 sector Wage subsidy and skill Provided partial wage subsidies of US$40 a month and 0.25 0.23 0.20 0.20 training technical/soft skills training for furloughed workers in the tourism and garment industries Support to households and workers Cash transfer Provided monthly cash grants to poor and vulnerable 1.16 1.12 0.67 1.12 individuals registered in the government’s IDPoor database Cash for work Provided jobs in rural areas through construction, 0.39 0.36 0.54 0.54 upgrade, and maintenance of rural roads, drainage, and small-scale irrigation Food support during Provided food support to local people during lockdown 0.00 0.00 0.08 0.03 lockdown period Total intervention package 4.50 3.02 2.49 4.87 Source: World Bank 2021. Note: Interventions provided in World Bank 2021 were regrouped by the authors. CHAPTER 1: CAMBODIA 51 By the end of 2021, the government had approved 10 rounds of stimulus measures, with more direct spending going to households than firms. The International Monetary Fund (IMF) estimated that by June 2021, both government direct spending and forgone revenue on key fiscal measures announced or taken in response to the pandemic stood at more than 6 percent of Cambodia’s GDP—a larger share of GDP than several higher- income countries in the region, including Indonesia, the Philippines, and Vietnam. Short-term relief measures were at the center of Cambodia’s firm-level policy response; instruments to increase medium-term competitiveness and diversification were more limited. Formal firms most commonly received public support in the form of tax deferrals, wage subsidies, social security exemptions, and rent deferrals. By reducing the costs of business operations, these policy instruments were intended to avert additional business closures and employment losses caused by mobility restrictions and the collapse in demand. However, in addition to biases toward short-term relief, this mix of instruments is prone to mistargeting (income tax deferrals are effective only for firms with profits) and misallocation of resources (wage subsidies encourage employment retention at the expense of workers’ movements toward more productive firms). Policy instruments to support firms generally targeted SMEs and economic sectors that were both growth engines and vulnerable to external shocks, including the garment, footwear, travel, tourism, construction, and agro-processing sectors. A cash transfer program targeting poor and vulnerable households was the largest social assistance instrument implemented during the pandemic. The government launched this program on June 24, 2020, with the aim of mitigating pandemic-induced increases in poverty and inequality. The Cambodian government initially intended to provide the COVID-19 cash transfers for three months, June–September 2020, but extended the relief program as the pandemic continued. As of March 2021, the program had disbursed US$307 million in transfers to support almost 700,000 households (2.7 million individuals). Poor and vulnerable households registered in the IDPoor database were eligible for the cash transfer program. The magnitude of the monthly transfer depended on the level of poverty; the area of residence, such as Phnom Penh or other urban and rural areas; and household size and composition, including the presence of children, seniors, or a family member with a disability.16 The average beneficiary household with five members received a transfer that allowed it to meet 10 percent of its monthly minimum consumption requirements and 20 percent of its monthly food requirements. The government also supported households and workers through public work programs, food support, and unemployment benefits during the pandemic. Coverage and timing The government’s scale-up of social assistance to households during the pandemic was unprecedented in terms of spending, the scale of payments, and the share of the population receiving social assistance. In less than a year, coverage of social assistance nearly doubled. Coverage increased mainly through the new cash transfer relief program launched by the government on June 24, 2020 to mitigate pandemic-induced increases in poverty and inequality.17 In May 2020, 13 percent of all households received support, nearly 16  Each beneficiary household receives (a) 80,000–120,000 Cambodian riels, or KHR (about US$20–$30) a month, depending on area of residence; (b) an additional allocation of KHR 16,000–52,000 (about US$4–$13) for each household member; and (c) KHR 16,000–40,000 (about US$4–$10) per family member under the age of 5, over the age of 60, or with a disability or HIV/AIDS. 17  Without intervention, the pandemic would have increased poverty by 4.7 percentage points in 2020, reversing three years of poverty reduction progress and likely raising the Gini index by 0.4 percentage points (World Bank 2022). 52 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA all of it in the form of direct cash transfers. By March 2021, household access to social assistance increased to 24 percent, and 23 percent of all households had received a COVID-19 relief cash transfer, up from 19 percent in August 2020 (Figure 1.11). The household cash transfer program quickly reached beneficiaries. Two months into the program, 9 in 10 targeted poor households had access to it (Figure 1.11). By March 2021, 95 percent of eligible poor households were covered. After registering for the program—a prerequisite for receiving the transfers— eligible households received payments on a monthly basis. By March 2021, most eligible households had received nine installments of the COVID-19 relief cash transfers, averaging a total of US$366 per household. Nearly all beneficiary households spent their cash transfers on food; more than half used them to purchase other essential items. In contrast, despite various firm-level policy interventions, the share of formal firms receiving government support remained low. By February 2021, just 18 percent of formally registered firms had received any public support, such as fiscal exemption measures, wage subsidies, or credit support. Government assistance had expanded by the next phase of the pandemic, when cases rose rapidly. However, despite a 38 percent increase in coverage, only 25 percent of formal firms had received any type of public support by May 2021. About 43 percent of firms applied for support, yet more than 40 percent of those who applied received no support. Medium-size and large firms were more likely to receive public support than smaller firms. In February 2021, 25 percent of medium-size and large firms received some form of government support, compared with 13 percent of small firms and 10 percent of microenterprises (Figure 1.12). Government support increased among formal firms of all sizes between February and May 2021, but larger firms continued to be more likely to receive support, with 34 percent of large firms and just 12 percent of microenterprises receiving public support in May 2021. The mix of policy instruments—which included fiscal exemptions, such as tax and social security deferrals—is one reason why access to public support differed by firm size. Access FIGURE 1.11  Household access to COVID-19 cash transfer program in Cambodia, 2020–21 100 92 93 95 90 Percent of households that received 80 COVID-19 cash transfers 60 40 21 24 23 19 20 0 August–September October–November December 2020– March 2021 2020 2020 January 2021 All households Eligible IDPoor households Source: World Bank data based on high-frequency phone surveys of households rounds 2–5. Note: “All households” are from the World Bank’s Living Standards Measurement Study Plus. “Eligible IDPoor households” are households that are enrolled in IDPoor. IDPoor households that hold a valid equity card are eligible for the COVID-19 cash transfer. This figure shows the proportion of households that received the COVID-19 cash transfer since the launch of the program in June 2020. CHAPTER 1: CAMBODIA 53 FIGURE 1.12  Formal firms’ access to COVID-19-related public support in Cambodia, by firm size, 2021 40 Percent of firms that received public support (%) 34 34 30 25 25 20 18 13 12 10 10 0 January/February 2021 May 2021 Micro (0–4) Small (5–19) Medium (20–99) Large (100+) Source: World Bank data based on business pulse survey rounds 3 and 4. Note: Sample includes firms that were formally registered with a government authority at the time of survey. to the wage subsidy was also concentrated in medium-size and large formal firms, with 17 percent of large firms receiving this type of assistance in May 2021, whereas no assistance was received by microenterprises during the same period. Access to credit support was higher for SMEs than for large firms, suggesting that some policy instruments reached their intended beneficiaries, but access to this instrument remained low for firms of all sizes. A large share of mostly informal household businesses did not receive policy support. Non-farm household businesses did not benefit from tax deferrals, one of the most important support programs for formal firms with taxable income. If their household was identified as poor under the IDPoor system, household businesses were eligible for the COVID-19 cash transfer, which constituted the main policy support such informal firms could obtain. However, as households operating a non-farm business tended to be better off than other households, only 12 percent received the COVID-19 cash transfer by March 2021—far fewer than the 27 percent of households without a non-farm business that did so (Figure 1.13). A large share of non-farm household businesses thus received neither firm-level nor household-level support. Targeting Notwithstanding gaps in coverage, the government had some success in targeting the most vulnerable firms. Formal firms in the worst-hit sectors and firms with higher sales losses received more public support than other firms. Formal firms in the hospitality sector were most likely to have received public support, followed by firms in other services sector. In the hospitality sector, the share of firms receiving public support rose from 30 percent in February 2021 to 41 percent in May 2021. In the other services sector, which includes tourism, coverage almost doubled, reaching 22 percent in May 2021. In May 2021, 27 percent of formally registered manufacturing firms, including garment and apparel manufacturers, received 54 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 1.13  Household access to COVID-19 cash transfer in Cambodia with and without non-farm business, 2020–21 40 Percent of households that received 30 30 27 25 25 COVID-19 transfers 20 14 11 12 10 7 0 August–September October–November December 2020– March 2021 2020 2020 January 2021 Households with non-farm business Households without non-farm business Source: World Bank data based on high-frequency phone surveys of households rounds 2–5. Note: The sample of households with and without a non-farm business was taken from the World Bank’s Living Standards Measurement Study Plus. Figures show the proportion of households that have received the COVID-19 cash transfer since the launch of the program in June 2020. government support, up from 23 percent in February 2021. Aside from the health sector, these sectors were most severely affected by COVID-19 outbreaks in 2021. The share of firms receiving government support in the retail and wholesale sector nearly doubled between February and May 2021, although it remained low at 9 percent. These results are consistent with the targeting of tax relief and wage subsidies to the garment, footwear, and tourism sectors. There is also evidence of targeting within sectors toward more vulnerable firms. Conditional on sector differences, firms that experienced larger drops in sales were more likely to have received government support (Figure 1.14). FIGURE 1.14  Correlation between access to COVID-19-related support and employment change in Cambodia, 2021 60 Percent of firms that received public support 50 40 30 20 10 0 −100 −50 0 50 Percent change in sales Source: World Bank data based on business pulse survey rounds 3 and 4. Note: Sample includes firms that were formally registered with a government authority at the time of the survey. CHAPTER 1: CAMBODIA 55 The government also successfully reached the poorest households it targeted. Since the launch of the COVID-19 relief program, poorer households have been more likely to receive cash transfers than better-off households. In March 2021, 39 percent of households in the bottom 40 percent received the cash transfer, compared with only 11 percent of households in the top 60 percent. Among eligible poor households, program reach was universally high and mostly unassociated with household job loss experience or labor income reduction. Among the population as a whole, households experiencing work stoppages were more likely to receive the COVID-19 cash transfer, and increasingly so over time. In contrast, households experiencing a reduction in labor income or a reduction in total household income were not more likely to have received the cash transfer. Effectiveness Despite government spending of historical proportions, the adequacy of the policy response remains in question. Most formal firms that received support considered it useful. In January/February 2021, 41 percent of firms considered public support critical for their survival, and 32 percent considered it very useful. Coverage remains very limited, however, as the instrument mix and implementation gaps preclude many firms from accessing support. For example, tax exemptions apply only to firms with taxable income, generally those of a formal nature. Therefore, it is difficult, if not impossible, to use fiscal exemption measures to target the most vulnerable firms. Access to other instruments that are more amenable to targeting, such as investment grants and access to credit, increased only slowly. In May 2021, only 3.5 percent of formal firms reported having gained access to new credit since the start of the pandemic (Figure 1.15). Difficulty obtaining credit could be caused by cumbersome eligibility criteria. In May 2021, 35 percent of surveyed firms reported that they did not benefit from public support because they lacked the right connections. FIGURE 1.15  Types of COVID-19-related public support to firms in Cambodia, 2021 15.2 15 11.5 Percent of firms that received support 10 6.8 5.8 6.0 5.1 5.3 5 3.5 3.3 3.5 2.5 2.9 1.4 1.6 0.6 0.8 0.4 0.8 0.6 1.0 0 January/February 2021 May 2021 Rental deferral Tax deferral Fixed−cost subsidy Access to credit Social security exemptions Wage subsidy Government purchase Grants to support urgent Trade reforms Other of goods and services investment Source: World Bank data based on business pulse survey rounds 3 and 4. Note: Sample includes firms that were formally registered with a government authority at the time of the survey. 56 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA The COVID-19 cash transfers mitigated increases in poverty and inequality, but they did not fully offset the pandemic’s negative impacts on the poor. In the absence of any government intervention, per capita consumption of the lowest quintile would have declined by 29 percent in 2020. Even with the cash transfer, per capita consumption of the lowest quintile decreased by 25 percent. Although the direct cash transfers were not large enough to fully mitigate the adverse impacts of the COVID-19 shock, they reduced increases in poverty and inequality. In 2020, about 460,000 people were projected to have fallen into poverty—about 290,000 fewer than in the absence of the program. The 2020 poverty rate was thus projected to have increased 2.8 percentage points instead of 4.7 percentage points (World Bank 2022). By targeting the poor, the cash transfers mitigated increases in inequality. The Gini index is projected to have increased 0.2 percentage points in 2020 instead of 0.4 percentage points in the absence of the program (World Bank 2022). Although the relief transfer program did not fully offset consumption losses, beneficiary households considered it an important program for their economic well-being. In March 2021, about nine months into the program, 38 percent of beneficiary households reported that the program had been extremely important for their household’s economic well-being, and another 40 percent reported that it had been very important (Figure 1.16). In addition, for 79 percent of beneficiary households, the relief program made a complete or strong difference to their household’s economic well-being. Beneficiary households that experienced a reduction in labor income had similar perceptions. Therefore, despite experiencing considerable declines in per capita consumption, even with cash transfers, poor beneficiary households did not perceive the expanded government social assistance as inadequate. To cope with the initial shock of the COVID-19 crisis, households reduced consumption and adopted other strategies, some with scarring potential. Poorer households were most likely to resort to such coping mechanisms. Reducing food and non-food consumption was the most common coping mechanism. In May 2020, before the launch of the cash transfer program, 65 percent of households reduced their food consumption, and 61 percent reduced their non-food consumption (Figure 1.17, panel a). Low-cost coping mechanisms were the second-most common strategy, adopted by 64 percent of households. These included FIGURE 1.16  Perceptions of beneficiary households in Cambodia of the importance of COVID-19 cash transfers to their economic well-being, 2020–21 100 11 22 Percent of households that received 23 80 COVID-19 cash transfers 44 60 39 40 40 20 45 37 38 0 October–November 2020 December 2020–January 2021 March 2021 Extremely important Very important Moderately important Not so important Not important at all Source: World Bank data based on high-frequency phone surveys of households rounds 3–5. Note: Sample of IDPoor households eligible for the COVID-19 cash transfers that received the relief cash transfer. CHAPTER 1: CAMBODIA 57 FIGURE 1.17  Coping strategies adopted by households in response to the COVID-19 crisis in Cambodia, 2020–21 a. May 2020 b. December 2020–January 2021 64 64 Low-cost strategies Low-cost strategies 83 99 58 59 Potentially scarring strategies Potentially scarring strategies 79 74 65 68 Reduced food consumption Reduced food consumption 89 86 61 59 Reduced non-food consumption Reduced non-food consumption 84 72 6 9 No changes implemented No changes implemented 2 0 0 20 40 60 80 100 0 20 40 60 80 100 Percent of households adopting strategy Percent of households adopting strategy All households Eligible IDPoor All households Eligible IDPoor Source: World Bank data based on high-frequency phone surveys of households, rounds 1 and 4. Note: “All households” are from the World Bank’s Living Standards Measurement Study Plus. “Eligible IDPoor households” are households that are enrolled in IDPoor. IDPoor households that hold a valid equity card are eligible for the COVID-19 cash transfer. Low-cost strategies include engaging in additional income-generating activity; borrowing or receiving aid from friends and/or family; and relying on savings or assistance from the government, a nongovernmental organization, and/or insurance. Potentially scarring strategies include selling productive assets, taking children out of school, taking out loans from formal or informal institutions, using credit for purchases, and delaying payment obligations. engaging in additional income generation; borrowing or receiving aid from family or friends; and relying on savings, assistance from the government or nongovernmental organizations, and/or insurance coverage. Strategies that may leave scars included selling productive assets, taking children out of school, taking out loans from formal/informal financial institutions, using credit for purchases, and delaying payment obligations; this was the third-most common type of strategy, adopted by 58 percent of households. Poor households were about 20 percentage points more likely to resort to any of these coping strategies than the population as a whole. This difference suggests that they were either hit harder by the COVID-19 shock than other households or had a limited financial cushion on the eve of the crisis, let alone the ability to endure months of the pandemic. In May 2020, for instance, 79 percent of poor households eligible for the COVID-19 cash transfer program resorted to potentially scarring coping strategies, compared with 58 percent of all households. For some households, the COVID-19 relief cash transfers alleviated the need to resort to potentially negative coping mechanisms as the pandemic progressed. The share of poor households adopting negative food-based coping strategies did not further increase between May 2020 and December 2020/January 2021 (Figure 1.17, panel b). This finding is consistent with the fact that COVID-19-related reasons for work stoppages diminished throughout 2020, income losses slowed, and income support rose. Reflecting scaled-up public safety nets, adoption of low-cost coping strategies, which included government social assistance, was nearly universal, at 99 percent, among poor households eligible for the cash transfer program in December 2020/January 2021 (Figure 1.17, panel b). Access to much-needed income support thus potentially mitigated the extent to which poor households that were enrolled in the IDPoor program resorted to coping mechanisms with potentially scarring effects. 58 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Lessons Learned The government expanded firm-level interventions in unprecedented ways during the pandemic. By reducing the costs of business operations, these policy instruments were intended to avert additional business closures and employment losses. The few formal firms that benefitted from this support considered it important for their survival. However, many formal firms, as well as the vast majority of informal firms, were unable to access these relief measures. Future assistance targeted at businesses could consider the following actions: ⦁ Setting up several new institutions to support SME development: To facilitate access to finance, the government established a new SME Bank, designed to support SMEs through co-financing and risk-sharing with commercial banks, and the Credit Guarantee Corporation of Cambodia (CGCC), which launched the first guarantee scheme in March 2021. The Agricultural and Rural Development Bank launched a loan program to support SMEs in agro-processing and other agricultural businesses. The Entrepreneurship Development Fund and Khmer Enterprise were launched in 2019 to support local start-ups through an incubation program, grant provision, and the coordination of financing. Achieving full operationalization and deploying support to firms more quickly will require significant institutional expansion.18 Promoting access to banking and other financial services could strengthen the demand for credit in a country where many firms, especially informal firms, remain unbanked. ⦁ Filling in financing gaps in the informal sector with more flexible policy instruments: The new lending policy of the SME Bank expanded eligibility conditions to include nonregistered businesses, increased tenure, and implemented more flexible repayment terms. ⦁ Creating an environment conducive for firms to enhance their capacity to obtain financing: Cambodia’s low level of financial literacy means that a large share of the population has only limited knowledge of, and ability to engage with, formal financial services. In conjunction with addressing supply gaps, the government could consider providing advisory support, including on the adoption of technology-based accounting management systems; facilitating the formalization of firms; and promoting the adoption of fintech products, such as peer-to-peer lending, credit scoring platforms, and digital payment systems. Given uncertainties about the trajectory of the recovery, it is important to improve the targeting and flexibility of firm-level instruments to support the private sector. Among the instruments in the current policy mix, credit support was best able to reach SMEs. Improving the disbursement of credit support programs could help expand government assistance to smaller and less formal firms. In sectors that will endure prolonged impacts, reviews of program relevance and effectiveness should be conducted regularly, so that adjustments can be made. Resources for sectors that have recovered should be redirected toward more productive use. Governments must credibly commit up front to terminating assistance when it is no longer needed, to improve predictability for businesses and avoid the risk of corruption. Reorienting firm-level policy instruments could help firms compete better in a world with heightened uncertainty, rapid technological shifts, and changing trade patterns. Short-term measures aimed at worker  Initial progress in deploying support to firms was slow, likely reflecting both a lack of implementation experience and weaknesses on the 18 demand side. Of the interventions aimed at providing liquidity and credit support, only 55 percent of the target had been implemented as of the end of December 2020 (AMRO 2021). CHAPTER 1: CAMBODIA 59 retention, such as wage subsidies, should be phased out to prioritize instruments that can facilitate the reallocation of workers and other resources toward more productive firms. Such instruments include reskilling, as well as job search and matching programs. To strengthen the economy’s resilience and long- term growth potential, Cambodia needs to continue its two-pronged efforts to diversify and upgrade existing industries. Doing so requires coordinated efforts to support firms’ embrace of new technologies and changing global trade patterns. Such efforts include improving existing capabilities for technology adoption; facilitating market access through export promotion, trade facilitation, or e-commerce adoption, among others; and reforming the investment climate. Financial and technical assistance, as well as the right policy framework, are needed. The government’s scale-up of social assistance during the pandemic was unprecedented, and there is little doubt that it helped Cambodia’s people to buffer the effects of the crisis. By providing poor households with income support, the COVID-19 relief cash transfers prevented almost 300,000 people from falling into poverty in 2020 and reduced the increase in poverty by an estimated 1.9 percentage points. Beneficiaries believe the income support was much needed and made a difference to their economic well-being. Several lessons of what worked well can be drawn from the implementation of the cash transfer program: ⦁ An existing and well-established social registry system of poor and vulnerable households enabled the government to rapidly introduce and scale up new emergency social assistance. Utilizing the IDPoor database for disbursement allowed the government to reach 90 percent of eligible households within two months of launching the COVID-19 relief cash transfer program. In March 2021, only 5 percent of eligible poor households had yet to receive cash transfers. Not having a valid IDPoor card, rather than lack of awareness of the program, appears to explain why some eligible households remained uncovered. Ensuring that eligible households have valid IDPoor cards could help expand the reach of the program, possibly more than raising awareness of the program. ⦁ A simple registration process, payment collection process, and access to a wide network of payment service providers helped to quickly get money into the hands of those in need. Registering with the village chief or commune council was a prerequisite for receiving the cash transfer. Among people who registered for the program, 95 percent found the registration process easy or very easy. Most beneficiaries were notified by the village chief (89 percent) or commune council member (7 percent) to go and collect the COVID-19 cash transfer; all beneficiaries received the full cash transfer through money transfer agents. Thanks to the wide net of these agents, the average travel time to receive the transfers in March 2021was 18 minutes, and fewer than 5 percent of beneficiaries had to travel for more than 30 minutes. Nearly all beneficiaries (99 percent) received the transfer without paying a fee; 98 percent found the process of receiving the COVID-19 relief cash transfer easy or very easy. ⦁ Disbursing transfers in cash was the quickest option in a largely unbanked country and a preferred payment method among poor beneficiaries. In October 2020, only 19 percent of eligible poor households had a mobile money account, and 75 percent had neither a mobile money account nor a bank account. About 9 in 10 households solely used cash to make payments, and most of them were not interested in any other payment method. People uninterested in payment methods other than cash cited difficulty using them as their primary reason. In this largely unbanked country, cash was the best option for ensuring inclusion of vulnerable populations. Disbursing social assistance via digital payments would have impeded—rather than aided—beneficiaries when accessing benefits. 60 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Cash transfers will be important to support recovery, rebuild livelihoods, and build resilience. COVID-19 continues to adversely affect many households; full economic recovery will take longer than initially expected. Even with the cash transfers, the poorest quintile experienced a larger reduction in per capita consumption than any other quintile. Not withdrawing much-needed support prematurely will prevent poor households from falling deeper into poverty and give them an opportunity to stabilize livelihoods. By providing opportunities for employment or the acquisition of assets for economic production, the cash transfer program can also be leveraged to rebuild livelihoods. As more households are negatively affected by the pandemic than received the COVID-19 cash transfers, expanding coverage of social assistance to include not just the poorest, but also the near-poor and vulnerable households, may be needed to prevent further increases in poverty and inequality. The new cash transfer program for low-income households adversely affected by pandemic-related lockdown measures is a positive step in this direction. The government could also consider integrating cash transfers into the possible choices of social protection schemes, with the goal of promoting long-term development. More and better-targeted cash transfers could boost the incomes of the poor and help them to build savings. Leveraging mobile technologies and automating processes can also facilitate the identification and registration of potential beneficiaries. The fiscal space needed to further increase the amount of cash transfers or expand coverage of social assistance is reaching its limits. By lowering revenue collection and increasing government expenditure, the COVID-19 outbreak created a fiscal financing gap. As of February 2022, the government had spent US$593 million on COVID-19 relief cash transfers since the launch of the program. Spending on this program alone amounted to more than 2 percent of GDP in 2020. It will be necessary to weigh decisions about a further scale-up or expansion of social assistance against financing, while evaluating the tax system could identify potential sources of revenue. It would also be beneficial to integrate an equity lens in the development and implementation of the upcoming Revenue Mobilization Strategy 2024–2028. CHAPTER 1: CAMBODIA 61 Annex 1 Annex 1A.  The Household and Firm Surveys in Cambodia Background The COVID-19 pandemic created an urgent need for timely data to help monitor and mitigate its effects on households and firms. The World Bank, in collaboration with the National Institute of Statistics in Cambodia, designed the COVID-19 High Frequency Phone Survey (HFPS) of households to monitor the evolving socioeconomic impacts of the pandemic on households and inform policy responses and interventions. The nationally representative household survey covers aspects of welfare such as employment and income, coping strategies, and public safety nets. The World Bank also designed the COVID-19 business pulse survey (BPS) to monitor the evolving impacts of the pandemic on businesses. This survey asks firms about their operational status, adjustment strategies, and receipt of public support during the COVID-19 pandemic. Timing This chapter draws on evidence from five rounds of the COVID-19 HFPS and four rounds of the COVID-19 BPS. The household survey was conducted over 10 months, between May 2020 and March 2021. Selected respondents, typically the head of household, completed interviews every eight weeks. Round 1 of the HFPS was implemented in May–June 2020 and round 5 in March 2021. The firm survey tracked businesses for 11 months. Round 1 of the BPS was conducted in June 2020 and round 4 in May 2021. For the most part, the timing of household- and firm-level surveys did not overlap. The survey rounds captured the evolving phases of the COVID-19 pandemic. Round 1 was conducted during an initial phase of stringent government measures and very few cases of COVID-19; restrictions were mostly precautionary. Rounds 2–4 of the household survey and rounds 2 and 3 of the firm survey were conducted during a second phase, in which government restrictions were lifted as COVID-19 cases remained low. The last round of both surveys took place during a third phase, in which stringent government measures were reinstated in response to surging COVID-19 cases. Sampling The HFPS consists of two samples: Living Standards Measurement Study Plus (LSMS+) and IDPoor households. The LSMS+ sample was drawn from the nationally representative LSMS+ household survey implemented in October–December 2019. The IDPoor sample was drawn from the list of households identified as poor by Cambodia’s national poverty identification program, IDPoor. With both LSMS+ and IDPoor samples, phone interviews were conducted with about 1,700 households in each survey round (Table 1A.1). The sampling frame for the BPS is based on lists from the Department of Commerce, the National Institute of Statistics, the Yellow Pages, and other online sources. The sample includes firms that were formally registered with a government authority at the time of survey. The survey was designed to cover the most affected provinces. It oversampled large firms, in order to reflect their disproportionate contribution to employment. Sampling was stratified by five regions, Battambang, Kampong Cham, Kandal, Phnom Penh, and Siem Reap; and four firm sizes, micro, small, medium, and large; but not by business sector. In each survey round, phone interviews were conducted with about 500 firms (Table 1A.2). 62 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE 1A.1  Dates and sample sizes of high-frequency phone surveys of households in Cambodia Round Survey dates Sample size Living Standards Measurement Study Plus 1 May 11–26, 2020 700 2 August 17–September 7, 2020 612 3 October 14–November 6, 2020 481 4 December 17–January 12, 2021 410 5 March 1–21, 2021 378 IDPoor 1 June 11–28, 2020 984 2 August 17–September 7, 2020 1,055 3 October 14 –November 6, 2020 1,184 4 December 17, 2020–January 12, 2021 1,277 5 March 1–21, 2021 1,309 Source: World Bank data based on high-frequency phone surveys of households rounds 1 to 5. Note: The Living Standards Measurement Study Plus sample includes all households in Cambodia; the IDPoor sample includes only poor households that are part of Cambodia’s national poverty identification program. TABLE 1A.2  Dates and sample sizes of business pulse survey in Cambodia Round Survey dates Sample size 1 June 18–July 3, 2020 537 2 September 3–29, 2020 518 3 January 22–February 17, 2021 514 4 May 14–31, 2021 513 Source: World Bank data based on business pulse survey rounds 1 to 4. Note: The business pulse survey sample includes firms that were formally registered with a government authority at the time of the survey. CHAPTER 1: CAMBODIA 63 Representativeness By capturing different degrees of formality, the household and firm surveys complement each other, but such sampling limits direct comparisons. The BPS surveys includes only firms that are formally registered with a government authority. The export-oriented manufacturing sector, with its garment, textile, and footwear factories, is the main provider of formal employment, followed by public administration and defense. The HFPS does not distinguish between formal and informal employment in its sampling strategy or survey questionnaire. However, given the widespread informality—88 percent of all pre-pandemic employment was considered informal—the household survey likely captures informal employment. Moreover, it asks respondents about non-farm household businesses and household farming activities, which are often presumed to be informal. Most informal employment is in the agricultural sector, followed by the wholesale and retail trade sector. The two surveys thus capture different types of employment and shed light on different sectors of the economy. Annex 1B.  Measures Taken to Fight COVID-19 in Cambodia TABLE 1B.1  Timeline of actions January 2020–November 2021 Year/date Action 2020 January 27 First case of COVID-19 confirmed in Cambodia March 16 Schools closed nationwide March 27 Restrictions on international travel imposed April 7 Khmer New Year celebrations cancelled April 9–16 Restrictions on domestic travel (between districts and provinces) imposed April 17 Quarantine of factory workers returning from provincial travel mandated May 20 Entry ban on visitors from six countries lifted June 11 New border measures implemented July 21 Phase 1 of school reopening: Some international private schools in Phnom Penh, Battambang, and Siem Reap permitted to reopen in August, provided that they followed highest safety standards August 4 Entry requirements for foreign travelers to Cambodia revised August 11 Water Festival holiday scheduled for October cancelled September 7 Phase 2 of school reopening: Schools from grades 9 to 12 could reopen across Cambodia, while schools of all other grade levels could only reopen in low-risk provinces October 15 All public universities allowed to reopen November 2 Phase 3 of school reopening: All schools nationwide reopened for in-person instruction November 8 All schools in Phnom Penh and Kandal province closed; all social gatherings and activities banned November 28–29 Many entertainment venues and schools closed; all mass gatherings, including meetings and weddings, banned for 15 days in Phnom Penh and Siem Reap December 29 Schools reopen; gatherings, meetings, and weddings in Phnom Penh and Siem Reap allowed (continues) 64 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE 1B.1  Timeline of actions January 2020–November 2021 (continued) Year/date Action 2021 January 11 Schools reopen February 4 Ministry of Health grants emergency use authorization to Sinopharm BIBP vaccine February 20 Large-scale COVID-19 outbreak in Phnom Penh detected April 15 Phnom Penh and Ta Khmau locked down for 14 days April 24 Sihanoukville locked down June 2 “High-risk” business activities in Phnom Penh restricted July 18 Border crossing with Vietnam suspended for one month, with exceptions only for diplomats, civil servants, some students, and patients in need of medical treatment July 29 Municipal and provincial authorities instructed to strictly enforce COVID-19 measures, including 14-day curfews July 29 Provinces that border Thailand (Banteay Meanchey, Oddar Meanchey, Battambang, Pailin, Pursat, Koh Kong, Preah Vihear) and Siem Reap locked down for 14 days November 1 Prime Minister Hun Sen announces that country will fully reopen all sectors Source: Chorn and Stromseth 2021; Ciorciari 2022; Karamba and colleagues 2021; EC 2020; UNICEF, MoEYS, and Save the Children 2021; WHO 2020a, 2020b, 2020c, 2021a, 2021b, 2021c, 2021d. CHAPTER 1: CAMBODIA 65 References AMRO (ASEAN +3 Macroeconomic Research Office). 2021. AMRO’s 2021 Annual Consultation Report on Cambodia. AMRO. Singapore. https://www.amro-asia.org/amros-2021-annual-consultation-report-on-cambodia/. Ciorciari, John. D. 2022. “Pandemic Containment and Authoritarian Spread: Cambodia’s Covid-19 Responses.” Asia Policy 29 (1): 4–9. https://doi.org/10.1353/asp.2022.0003. Chorn, Adrien, and Jonathan Stromseth. 2021. Covid-19 Comes to Cambodia. Order from Chaos (blog), Wednesday, May 19, 2021. Brookings Institution. https://www.brookings.edu/blog/order-from-chaos/2021/05/19/covid-19- comes-to-cambodia/. Davis, Steven J., John C. Haltiwanger, and Scott Schuh. 1996. Job Creation and Destruction. Cambridge, MA: MIT Press. EC (European Commission). 2020. “Cambodia Loses Duty-Free Access to the EU Market over Human Rights Concerns.” Press Release, August 12, 2020, https://ec.europa.eu/commission/presscorner/detail/en/IP_20_1469. Karamba, Wendy, Isabelle Salcher, and Kimsun Tong. 2021. The Socioeconomic Impacts of COVID-19 on Households in Cambodia. Report No. 5: Results from the High-Frequency Phone Survey of Households Round 5 (March 1–21, 2021). World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35868. National Institute of Statistics and Ministry of Planning. 2020. “General Population Census of the Kingdom of Cambodia 2019: National Report on Final Census Results.” National Institute of Statistics and Ministry of Planning. Kingdom of Cambodia. Nit, Buntongyi, Alexander Lourdes Samy, Shu Leed Tan, Sopanha Vory, Youhok Lim, Ryan Rachmad Nugraha, Xu Lin, Attaullah Ahmadi, and Don Eliseo Lucero-Prisno, III. 2021. “Understanding the Slow COVID-19 Trajectory of Cambodia.” Public Health in Practice 2: 100073. UNICEF, MoEYS (Ministry of Education, Youth and Sport), and Save the Children. 2021. Cambodia COVID-19 Joint Education Needs Assessment. Phnom Penh. https://www.unicef.org/cambodia/media/4296/file/Cambodia%20 COVID-19%20Joint%20Education%20Needs%20Assessment.pdf. WEF (World Economic Forum). 2019. The Global Competitiveness Report: 2019. WEF. Geneva. https://www3.weforum. org/docs/WEF_TheGlobalCompetitivenessReport2019.pdf. World Bank. 2019. Cambodia Economic Update, May 2019: Recent Economic Developments and Outlook. World Bank. Phnom Penh. https://openknowledge.worldbank.org/handle/10986/31641. World Bank. 2020. Cambodia Economic Update: Cambodia in the Time of COVID-19. Special Focus: Teacher Accountability and Student Learning Outcomes. World Bank. Washington, DC. https://pubdocs.worldbank.org/en/357291590674539831/ CEU-Report-May2020-Final.pdf. World Bank. 2021. Cambodia Economic Update: Living with COVID. Special Focus: The Impact of the COVID-19 Pandemic on Learning and Earning in Cambodia. Washington, DC: World Bank. http://documents.worldbank.org/curated/ en/099350012062137172/P1773400f35a770af0b4fa0781dffcd517e. World Bank. 2022. Cambodia Poverty Assessment: Toward a More Inclusive and Resilient Cambodia. World Bank. Washington, DC. https://openknowledge.worldbank.org/handle/10986/38344. WHO (World Health Organization). 2020a. COVID-19 Joint WHO-MOH Situation Report 7. WHO. Phnom Penh. WHO. 2020b. COVID-19 Joint WHO-MOH Situation Report 13. WHO. Phnom Penh. WHO. 2020c. COVID-19 Joint WHO-MOH Situation Report 17. WHO. Phnom Penh. WHO. 2021a. COVID-19 Joint WHO-MOH Situation Report 34. WHO. Phnom Penh. WHO. 2021b. COVID-19 Joint WHO-MOH Situation Report 44. WHO. Phnom Penh. WHO. 2021c. COVID-19 Joint WHO-MOH Situation Report 57. WHO. Phnom Penh. WHO. 2021d. COVID-19 Joint WHO-MOH Situation Report 70. WHO. Phnom Penh. CHAPTER 2 Indonesia By Rabia Ali, Aufa Doarest, Ade Febriady, and Bayu Agnimaruto COVID-19 took a heavy economic and human toll in Indonesia. The country suffered a recession in 2020, and economic activity remained below its potential in 2021. The economy shrank 2.1 percent in 2020, after averaging 5 percent annual growth in the five years before the pandemic. Although the recession was less severe than it was in many other countries, the pace of recovery in 2021 was more drawn out. Private consumption in particular was much slower to pick up than it was in other countries, and the negative impacts of the health shock on the labor market have translated into increased poverty. The unemployment rate stood at 6.5 percent in August 2021, up from 5.3 percent in August 2019. The poverty headcount rate rose to 10.2 percent in 2020, up from a record low of 9.2 percent in September 2019, before falling to 9.7 percent by September 2021. According to official statistics, by the end of 2021, over 140,000 Indonesians had died from the virus. This chapter examines the economic toll of the crisis, the effect on jobs, and the manner in which government policies were implemented during the crisis. It examines the evidence on targeting and the effectiveness of government support, drawing lessons for program design and implementation. The analysis is based on data from the Indonesia business pulse survey (BPS) and High-Frequency Monitoring of Households survey, henceforth referred to as the high-frequency phone survey (HFPS). In Indonesia, the World Bank conducted the BPS and the HFPS between May 2020 and October 2021 (see Annex 2A for further details). Timeline of the COVID-19 Pandemic and Government Measures The BPS and HFPS in Indonesia occurred against a backdrop of several waves of COVID-19. Figure 2.1 overlays the timeline of rounds of the BPS and the HFPS with COVID-19 cases in Indonesia, levels of mobility, and restrictions imposed by the government. Official reports of cases in Indonesia remained significantly lower than in peer countries in the region, with daily cases in the first wave gradually peaking in October 2020 at just 4,850 a day.19 The peak numbers of cases in each successive wave were higher, reaching a maximum of 50,000 cases a day in July 2021, during the Delta wave. After the number of cases fell to 19  Testing was not widespread, and there has been widespread speculation about underreporting of cases and deaths. See, for example, Allard (2020) and Mulyanto (2020). 67 68 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 2.1  Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Indonesia, 2020–22 HFPS R1 HFPS R2 HFPS R3 HFPS R4 HFPS R5 HFPS R6 60 100 Inverted mobility index (14-day average) Stringency index (14-day average) 40 Stringency index 75 20 50 Inverted 0 25 Mobility index -20 0 BPS R1 BPS R2 BPS R3 BPS R4 20 COVID-19 cases/deaths (per 100,000 people) 15 COVID-19 cases 10 5 COVID-19 deaths 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Source: Mobility data comes from Google Mobility Reports; stringency data comes from the Oxford Covid-19 Government Response Tracker (OxCGRT); COVID-19 cases and deaths from COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University, April 2022. Note: The mobility index is the average percentage change in time spent at retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. For each day of the week, the baseline is defined as the median value from the five-week period of January 3–February 6, 2020. The figure shows 14-day rolling averages of the mobility index, which is inverted so that higher values indicate lower mobility. The stringency index records the strictness of lockdown style policies; 14-day rolling averages are shown. COVID-19 cases and deaths are smoothed by a seven-day average. Phases of the pandemic are determined by identifying structural breaks in COVID-19 cases and the mobility index (Hale et al. 2021). See the introduction for further details. its lowest point in May 2020, mobility increased between waves, reaching a minimum when cases peaked. Mobility does not seem to be associated with the stringency of mobility restrictions, which was highest around May 2020. The number of cases fluctuated thereafter but remained relatively high throughout 2020 and 2021. Mobility almost reached its pre-COVID level just before the Delta wave hit, after which it declined again. Its lowest point during the Delta wave was still higher than the minimum during earlier waves, despite the severity of the wave. In terms of the timing of the surveys, three time periods stand out. The first two rounds of the HFPS and the first round of the BPS took place immediately or soon after the initial and most stringent decline in mobility and thus capture conditions soon after the pandemic started. Also salient are the March 2021 rounds of both surveys, which were administered during a period of economic rebound, just before the Delta wave hit in July 2021. The August 2021 round of the BPS and the October 2021 round of the HFPS took place during a partial recovery from the Delta wave. CHAPTER 2: INDONESIA 69 Employment Impacts: Shocks and Recovery Reduction in hours worked and wages rather than lay-offs During the pandemic-induced recession, the performance of firms tracked trends in GDP growth. The initial shock at the beginning of the pandemic was large. Nearly a quarter of firms closed by June 2020, and almost two-thirds reported reduced sales (Figure2.2, panel a). Sales in June 2020 declined by 56 percent from June 2019, though for closed firms, sales are considered to have declined by 100 percent. By March 2021, as the economy showed signs of recovery, closures declined to 8  percent. However, the share of firms reporting reduced sales since the previous round of the survey was still high, at 40 percent. (Figure 2.2, panel b). Sales declined by an average of 16 percent between March 2021 and October 2020. The mid-2021 Delta surge interrupted the recovery process that had been underway in the first half of 2021. By August 2021, closures and sales had either remained stable or deteriorated since the previous round of the survey. Most firms coped with the crisis by reducing work hours or wages rather than by laying off workers. Adjustments on the intensive margin (reduced work hours and wages) were far more common than changes on the extensive margin (laying off workers) and tended to vary with the conditions of the economy. The share of firms adjusting employment on the intensive margin fell to about a quarter during the economic rebound in early 2021, before rising again to over a third after the Delta wave. About 10 percent of firms in each BPS round continued to adjust employment on the extensive margin. Trends in employment and income registered in household data are consistent with the employment adjustments reported by firms. Figure 2.3 overlays firm data with employment and income data on wage workers from the HFPS. It shows that the share of primary breadwinners in wage work that were not working in each round declined from a peak of 25 percent in June 2020, stagnating at about 10 percent thereafter. This pattern is broadly consistent with the firm data, which shows that employment adjustments on the extensive margin were less common yet plateaued after initial improvements. The share of breadwinners FIGURE 2.2  Operating status of firms and changes in sales in Indonesia, June 2020–August 2021 a. Status of firms b. Change in sales 80 0 64.2 –10 60 Percentage change in sales (%) 52.3 Percent of firms (%) 46.4 –15.5 40.1 39.4 –20 40 –22.0 24.3 –30 20 14.2 11.5 7.7 –40 0 –50 Firms closed Firms open, sales Firms open, sales were decreased stable orincreased –55.6 –60 Status of firms Average change in sales Jun-20 Mar-21 Aug-21 Jun-20 Mar-21 Aug-21 Source: World Bank data based on business pulse surveys round 1,3 and 4 Note: Estimates of the change in sales in June 2020 are relative to June 2019. The March 2021 estimate is relative to October 2020. The August 2021 estimate is relative to March 2021. 70 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 2.3  Changes in employment and income in Indonesia, May 2020–October 2021 60 50 Share of firms or percent of breadwinners (%) 40 30 20 10 0 May-20 June-20 Jul-20 Oct-20 Mar-21 Jul-21 Aug-21 Oct-21 Firms adjusting extensive margin Firms adjusting intensive margin Wage-earning breadwinners not working Wage-earning breadwinners with reduced incomes Source: World Bank data based on business pulse surveys round 1-4 and high-frequency phone surveys round 1-5. Note: Income is relative to the pre-pandemic level. The share of breadwinners with reduced income is expressed as a percentage of working breadwinners. with reduced incomes fell from 56 percent in May 2020 to 40 percent by March 2021, rising to 44 percent after the Delta wave. These figures are also broadly consistent with firm data on adjustments to employment on the intensive margin. Both employment and income indicators registered a slight deterioration after the Delta wave in July–August 2021. Despite the economic recovery underway before Delta in the first half of 2021, sales of many firms were still below pre-pandemic levels, forcing them to adjust wages and operating hours. This reduction in wages and working hours was reflected in the large share of wage workers that continued to work with reduced incomes in each round of the HFPS. The Delta wave caused a serious health crisis, but the economy was able to partially weather the storm. The number of cases, hospitalizations, and deaths was much higher during the Delta wave than in earlier waves, pushing hospitals beyond their capacity. The economy was not as badly affected as during the initial shock in 2020, partly because of strong commodity prices, exports, and adjustments to the “new normal,” particularly in the manufacturing sector. In contrast, services and private consumption remained very weak. The level of restrictions did not change significantly during 2021; mobility generally remained above 2020 levels and was not correlated with the stringency of restrictions, suggesting that measures were poorly enforced.20 Indonesia entered the pandemic with a similar business environment quality to the average standards in the East Asia and Pacific region, although higher than that of lower-middle income countries. With macroeconomic stability and relatively established institutions, the country provides a fair economic environment for private sector development. Firms’ exit and entry dynamics, trade openness, and domestic competition forces are strong enough to establish market discipline. Yet, the country displays significant gaps relative to its regional peers on labor flexibility and financial depth, as reflected by their lowest and third lowest scores, respectively, among the countries analyzed in this report in these topical sub-indicators of the WEF Global Competitiveness Index (WEF 2019). 20  It is likely that higher vaccination rates in urban areas, coupled with “pandemic fatigue,” played a role in making individuals less wary of face-to-face interactions, keeping mobility relatively high, especially in residential areas and for leisure activities, throughout 2021, despite maintenance of social distancing restrictions at roughly the same level throughout the period of observation. CHAPTER 2: INDONESIA 71 Disproportionate impact of the crisis on micro and small firms and informal workers While impacts were initially widespread across firms, the effects of the recession started to vary across types of firms as the pandemic dragged on. The initial shock in 2020 was felt across the economy, with firms of all sizes equally likely to experience a severe reduction in sales (Figure 2.6). The severity of the decline was not correlated with productivity; by the time the Delta wave hit in mid-2021, sales by micro, small, and medium-size enterprises (MSMEs) exhibited sharper declines than large firms, holding other factors constant. Microenterprises were hit hardest (Figure 2.4). Consistent with their better sales performance, large firms were significantly more likely to have increased wages and operating hours in 2021. Most small firms had lower employment levels than they did in June 2020; only large firms enjoyed net positive adjustments. Before the Delta wave, just over 40 percent of large firms had increased either wages or operating hours. In contrast, less than 20 percent of small and medium-size enterprises (SMEs) and a mere 6 percent of microenterprises did so. Even during the economic rebound of early 2021, MSMEs were more likely to reduce operating hours or wages than they were to increase them. When the Delta wave emerged, firms of all sizes were more likely to reduce than to increase employment. Like microenterprises, household non-farm businesses were hit hard and struggled to recover. In 2019, 69 percent of workers in Indonesia were engaged in informal household enterprises, as either owners or employees. These workers lacked contracts, worker protections, and social insurance. Ninety-three percent of private sector enterprises in Indonesia did not have legal entity status, according to the 2016 Economic Census. The BPS sample is drawn from a census of firms in the formal sector and thus does not capture the situation of household enterprises, most of which are informal. The HFPS of non-farm businesses at the household level does not collect data on sales of non-farm businesses, but it does collect data on income. This data shows that almost 90 percent of primary breadwinners engaged in these businesses reported reduced incomes. After substantial improvement in July 2020, the share of non-farm businesses with reduced incomes held firm, at about 50 percent. Although the data on business incomes is not directly comparable with the sales performance of MSMEs in the BPS, it does align with the BPS finding that the smallest firms were hit hardest. Among breadwinners engaged in these businesses and working with reduced incomes, the average reduction in income was substantial, with losses of more than 40 percent of pre-pandemic income throughout the pandemic. FIGURE 2.4  Characteristics of firms in indonesia that experienced severe sales shocks 30 20.2*** 20 10.7* 11.8* Percent (%) 9.6** 10 0 –10 High value-added Creative economy Micro Small-medium Female dominated service and tourism (relative to manufacturing) (relative to large) (relative to male) June 2019–June 2020 March–August 2021 Source: World Bank calculations based on the results of the business pulse surveys round 1,3 and 4. Note: Firms were ranked in order of magnitude of changes in sales. The lowest quartile of firms (firms with the largest declines in sales) was classified as experiencing a severe shock. Significance levels: *** 1 percent level; ** 5 percent level; * 10 percent level. 72 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Throughout the pandemic, MSMEs and household non-farm businesses were significantly more likely than large firms to cope by adjusting employment at the intensive margin. During the recovery in early 2021, 40  percent of large firms increased wages and working hours, compared with only 5  percent of microenterprises. When the Delta wave occurred, the shares of firms making positive adjustments dropped significantly across firm sizes, but large firms were still more likely than microenterprises and non-farm household businesses to increase hours or wages. Smaller firms were much more likely to lower wages or reduce working hours even before the Delta wave. For example, in March 2021—a period of partial recovery—20 percent of microenterprises and just 10 percent of large firms reported reducing working hours or wages.21 Firms in the creative and tourism sectors, in which low value-added services are concentrated, were also hit hard. As Indonesia’s borders remained closed to foreign tourists and the visa-on-arrival policy was revoked throughout 2021, the impact on the sales of firms in these sectors comes as no surprise. Service sector firms were also more likely than other firms, including those in manufacturing, to report cutting wages. Over 80 percent of breadwinners working in the trade, hotel, and restaurant sectors reported working with reduced incomes just after the pandemic hit (Purnamasari and Ali 2020). The incidence of reduced incomes declined in all sectors through early 2021, rising slightly during the Delta wave. On average, it remained higher in services than in industry. Initially, the magnitude of income losses was larger in industry than in services; the inverse was true by October 2021, when average income losses were 8 percentage points higher in the service sector. Employment and income shocks were more common among female and low-skilled breadwinners. The HFPS data also shed light on the characteristics of workers who were hardest hit. Male breadwinners were significantly less likely than female breadwinners to report that they had stopped working, and the difference grew after the Delta wave. In contrast, the gender gap in the incidence of income reduction fluctuated, with no clear pattern. Breadwinners with tertiary education were significantly less likely to have stopped working than workers with only junior-secondary or lower education, although the difference diminished over time. The protective effect of education was more obvious and larger for income, but it also declined over time (Sari and colleagues, forthcoming). Digital Transformation and the increase in demand for digital-savvy workers A significant share of firms used the Internet or digital platforms to adapt to the negative shocks from COVID-19. Multiple lockdowns in many regions during the COVID-19 pandemic reduced mobility and sales for many firms relying on traditional point of sales models, such as brick and mortar stores. Firms coped with this limited mobility by introducing or increasing sales through platforms and digital media. Indeed, 41 percent of firms in the June 2020 survey started to use or increased their use of the Internet, social media, and other digital methods to carry out business. This adjustment was implemented by firms of all sizes, though unevenly, with 8 in 10 large firms surveyed in June 2020 implementing a digital tool to do business, compared with 3 in 10 micro firms (Figure 2.5). In August 2021, more firms adjusted by adopting digital tools—30 percent more firms than in June 2020. More MSMEs started to catch up and adopt digital platforms, although the share of digitally transformed firms is still lower for MSMEs than large firms. 21  Data on wage adjustments among household non-farm businesses was available only in the October 2021 round of the HFPS. The data showed that just under a third were operating with decreased operating hours, and that 15 percent were operating with increased hours relative to before the pandemic. CHAPTER 2: INDONESIA 73 FIGURE 2.5  Digital adoption by firms in Indonesia between June 2020 and August 2021 100 35 34 40 90 30 30 28 35 Percent of firms (%) 80 Percent of firms (%) 24 30 70 20 60 25 50 20 40 15 30 10 20 10 5 0 0 High value-added Creative Manufacturing Micro Small-medium Large service economy and tourism All Firms Strata Size Jun-20 (share of firms) left hand side Aug-21 (share of firms) left hand side Digital adoption between Jun-20 and Aug-21 (increased share of firms) right hand side Source: World Bank calculations based on results of business pulse surveys round 1 and 4. Lack of digital skills limited the transformation, potentially increasing the demand for workers with digital skills in the future. The BPS data revealed that lack of skills remained one of the main obstacles to implementing digital transition (World Bank 2021a). More firms wanted to start to implement digital transition or deepen their transition but lacked the in-house skills to do so. Therefore, the increased need for digitalization increases demand for workers with good digital skills. Factors affecting the speed of recovery Multiple rounds of the BPS revealed no statistically significant association between recovery patterns and labor productivity. On average, firms with both high and low productivity had similar recovery speeds. However, the fact that more productive firms did not recover more quickly will likely reduce the average productivity in Indonesia post-pandemic. Faster recovery speed was associated with digital transformation and export orientation. Analysis of the BPS results revealed a positive association between early investment in digital equipment and sales trajectory. Firms that invested in digital equipment in the previous COVID wave had sales that were 5.8 percentage points higher than those that did not invest. Investment in these technologies afforded firms direct access to more of the market in Indonesia and resulted in relatively higher sales. Export-oriented firms were also associated with faster recovery speed. Data from the fourth round of the BPS indicates that export-oriented firms had sales that were 11 percentage points higher than firms that did not export. Their better recovery may reflect the fact that they sell to a broader set of markets, diversifying the risk of dependence on the fate of the domestic economy. Fiscal Support to Firms and Households Size and composition of the fiscal response Budget flexibility allowed authorities to quickly ramp up support during the multiple waves of COVID-19. As a response to the Delta variant outbreak in Indonesia mid-2021, the government increased the annual 74 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA COVID-19 fiscal package by 0.3 percentage points in July 2021, to 4.8 percent of 2020 GDP. The package is 1 percentage point of GDP higher than what was spent in 2020 and was divided about equally among health measures, relief to households, and support to firms. Given the focus on pandemic emergencies, spending on other priority programs, such as digitalization and food security, remained low, at 10 percent of the total fiscal package in 2021. Reflecting these priorities and budget reallocations to meet deficit targets, the January to-September spending decreased by 1 percentage point of GDP but remained above pre-pandemic levels. With all of the support available for households and firms, the government of Indonesia allocated a substantial fiscal package to respond to the COVID-19 crisis. Yet, the fiscal package per number of cases remained low. By June 2021, additional spending and forgone revenue stood at 9.3 percent of GDP. This share was higher than the middle-income average (5 percent) and the average for other Association of Southeast Asian Nations (ASEAN) countries—excluding Singapore, at 18.4 percent, and Thailand, at 14.6 percent (Figure 2.6). Average additional spending of East Asia and Pacific (EAP) countries was also slightly higher, at 11 percent. However, after controlling for the total number of cases, Indonesia’s fiscal allocation per positive case was very low (Figure 2.7). Indonesia ranked at the bottom of ASEAN countries, with the Philippines as the worst performer. Singapore and Thailand remained the two countries with the most extensive fiscal package and fiscal package per number of cases. FIGURE 2.6  Size and composition of fiscal FIGURE 2.7  Correlation between size of packages in select countries in 2021 fiscal package and number of COVID-19 cases in East Asia and Pacific 20 18 5 16 IDN 14 Percentage of GDP 4 12 10 Total cases (Million) 8 3 PHL 6 MYS 4 Middle 2 THA 2 Income 0 EAP 1 VNM Thailand Singapore Indonesia Malaysia Vietnam Philippines EAP countries Middle Income Region (average) (average) SGP 0 0 5 10 15 20 25 Additional spending and forgone revenue Additional spending, forgone revenues, and liquidity support (% of GDP) Equity, loans, and guarantees Source: IMF 2021 and COVID-19 Data Repository by the Center for Systems Science and Source: IMF 2021.22 Engineering (CSSE) at Johns Hopkins University. 22  The numbers reported in the figure differ slightly from those used in the introduction and conclusion but are consistent. The latter, produced by the EAP Chief Economist Office of the World Bank, consider spending in 2020 and 2021, not only 2021. For Indonesia, the size of the package for 2020 and 2021 as a share of GDP, including contingent liabilities and equity injections, is 9.7 percent compared to the 10.2 reported for 2020 based on the IMF database. CHAPTER 2: INDONESIA 75 Coverage and timing In April 2020, in response to the first wave of the pandemic, the government announced social and business support programs, which it expanded in August 2020 (Figure 2.8). The budget allocated to social protection increased from 0.8 percent of GDP in 2019 to 1.4 percent of GDP in 2021 (Purnamasari and Ali 2020). COVID-19 support to firms took two main forms: (a) cash flow assistance, such as wage subsidies, tax deductions, cash transfers, and loan payment deferrals, which aimed to help firms cope and protect employment during lockdowns and the economic downturn; and (b) access to credit in the form of interest rate subsidies and credit guarantees. Firm support in Indonesia was comparable to that of peer countries. It benefited large firms disproportionately, mainly through wage subsidies and tax deductions, benefits that many small and informal firms were unable to access. Cash transfers and loan payment deferrals were better targeted at MSMEs. For instance, the government increased the subsidy for the guaranteed microfinance program, Kredit Usaha Rakyat (KUR), for several months during the crisis. With this extra subsidy, the interest rate of the program dropped from 6 percent to 0 percent in 2020. Most support was provided by expanding existing programs, but the government also complemented these measures by introducing new cash programs during the pandemic (Table 2.1). In April 2020, it FIGURE 2.8  Rollout of COVID-19 support programs for households and firms in Indonesia, January 2020–January 2022 PKH and Sembako card KUR, Electricity bill reduction, Tax exemption/subsidy, BST, BLT-DD, Pre-employment card continued with expanded beneficiaries and started increased benefits Jan ‘20 Feb ‘20 Mar ‘20 Apr ‘20 May ‘20 Jun ‘20 BPUM and Wage subsidy (BSU) Credit restructuring/Loan deferment started started Dec ‘20 Nov ‘20 Oct ‘20 Sep ‘20 Aug ‘20 Jul ‘20 Jan ‘21 Feb ‘21 Mar ‘21 Apr ‘21 May ‘21 Jun ‘21 BPUM and Wage subsidy ended Dec ‘21 Nov ‘21 Oct ‘21 Sep ‘21 Aug ‘21 Jul ‘21 BPUM: Bantuan Produktif Usaha Mikro BST: Bantuan Sosial Tunai BLT-DD: Bantuan Langsung Tunai Dana Desa BSU: Bantuan Subsidi Upah Jan ‘22 KUR: Kredit Usaha Rakyat PKH: Program Keluarga Harapan Source: Ministry of Finance–National Economic Recovery (PEN) 2020 and 2021 presentations, Coordinating Ministry of Economic Affair 2021 presentation, Ministry of Cooperatives and MSME 2020 and 2021 announcement. 76 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE 2.1  Major social and business support programs in Indonesia Program Description Bantuan Langsung Tunai Dana Desa Unconditional cash transfer to poor, underprivileged, and/or vulnerable families affected (BLT DD) by COVID-19. This cash transfer was financed through the Village Fund. Bantuan Produktif Usaha Mikro (BPUM) Cash transfer for MSMEs under which eligible and registered MSMEs received assistance in both 2020 and 2021. Bantuan Sosial Tunai (BST) Unconditional monthly cash transfer of Rp 600,000 per household made to poor, underprivileged, and/or vulnerable families affected by COVID-19. Excludes recipients of the Program Keluarga Harapan (PKH) program, basic food cards, and pre-employment cards. Kartu pra-kerja [pre-employment card] Competency development program aimed at job seekers and laid-off workers. Package includes training assistance (Rp 1 million), incentives for completing training–Rp 600,000 (equivalent to US$42.8) per month for four months, and incentives for job surveys–Rp 150,000 (equivalent to US$10.7). Kredit Usaha Raya (KUR) Flagship program that aims to enhance MSMEs’ access to credit by providing credit guarantees and interest subsidies to participating banks. It was the main channel for supporting MSMEs through the pandemic and one of the largest and longest-standing firm support programs in the world. Program Keluarga Harapan [Family This is a conditional cash transfer (CCT) targeted to poor families and household in Hope Program] (PKH) Indonesia. This program targets poor households/families with the following types of family members: pregnant/lactating mothers, children under six, school-age children up to senior-secondary school, persons with disabilities, and seniors. Like other CCT programs, PKH links the benefits to the compliance of the beneficiaries on required conditions, which include mandatory presence in health and education facilities. Sembako card Non-cash food assistance implemented through electronic account mechanism that can be used only to buy food at food vendors (e-warong) in collaboration with banks. Subsidi upah [wage subsidy] Cash transfer available to workers registered in social insurance program Badan Penyelenggara Jaminan Sosial (BPJS). This program aims to maintain, protect, and improve the economic capacity of formal workers navigating the negative impact of COVID19 pandemic. Total benefit is Rp 500,000 (US$35.7) per month for two months. introduced fiscal incentives programs that allowed firms to forgo paying income tax. In August 2020, it announced and implemented Bantuan Produktif Usaha Mikro (BPUM), a cash transfer program for microenterprises, through which it distributed Rp 2.4 million (around US$171.423) to each registered microenterprise in Indonesia in 2020. With a total budget of about Rp 39.8 trillion (around US$2.8 billion), the program reached 12 million microenterprises. In 2021, the BPUM benefit was reduced to Rp 1.2 million per beneficiary (around US$85.7).24 As the first wave of the pandemic eased by the end of 2020, the government cut budget allocations to most relief programs (Figure 2.9). The budget for social assistance was scaled back from 1.4 percent of GDP in 2020 to 1 percent of GDP in 2021, and support for firms was cut from 0.8 percent of GDP to 0.7 percent of GDP. However, spending for interest subsidies for MSMEs (KUR and non-KUR) increased from 0.1 percent to  US$1 = IDR 14.000 23  Indonesia also had a cash-for-work program during the pandemic, Padat Karya Tunai Desa (PKTD). Its budget was small, and the program 24 was implemented using the Village Fund and managed by the village. CHAPTER 2: INDONESIA 77 FIGURE 2.9  Budget allocation and implementation of support programs in Indonesia, 2020 and 2021 300 120 253 250 100 190 % of budget realized Budget (Rp trillion) 200 80 150 60 100 40 44 50 50 29 30 37 29 33 29 20 15 20 21 17 23 9 11 10 0 0 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 Kredit Usaha Bantuan Wage Electricity Pre-employment Program Sembako Bantuan Bantuan Rakyat (KUR) Produktif subsidy bills reduction card Keluarga card Sosial Tunai Tunai - Dana Usaha Mikro Harapan (BST) Desa (BLT-DD) (BPUM) (PKH) Budget Realization Source: Ministry of Finance 2021 and Ministry of Cooperatives and MSME 2021 Note: The BLT DD budget in 2020 was revised from Rp 31.8 trillion (US$2.27 billion) to Rp 23.0 trillion (US$1.64 billion). 0.2 percent of GDP, and the government increased the KUR allocation from Rp 190 trillion (US$13.6 billion) to Rp 263 trillion (US$18.8 billion). Most assistance programs for households met their estimated numerical targets in 2021. The government spent almost all its allocated budget for the COVID-19 response in 2020 and 2021, with realization rates of more than 90 percent for most programs (Figure 2.10). The two exceptions were cash transfers using Village Fund allocations, with a 72 percent implementation rate in 2021, and pre-employment cards, implemented at 91 percent in 2020 and 86 percent in 2021. All but one household support program reached its target FIGURE 2.10  Beneficiaries of social assistance programs and other relief measures in Indonesia, 2020–21 70 60 Percent of households (%) 50 40 30 20 10 0 Sembako card Program Bantuan Bantuan Electrcity bills Pre-employment Cash for work Keluarga Sosial Tunai Tunai Dana reduction card Harapan (PKH) (BST) Desa (BLT-DD) Estimated target since pandemic up to 2021 Received since pandemic up to Oct-21 Source: World Bank data based on high-frequency phone survey round 5 Note: Error bars show 90 percent confidence intervals. 78 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA number of beneficiaries (Figure 2.10). The exception was the pre-employment card, which reached just 10 percent of the population, far below the intended 16 percent. This poor performance was also reflected in the program’s 86 percent realization rate in 2021. The government began to roll out firm assistance slowly. As of June 2020, just 7 percent of surveyed firms reported receiving any government support. This proportion increased significantly in August 2021, by which time two-thirds of firms reported having received assistance. In the early stage of the pandemic, most support targeted formal firms. Only registered firms were eligible for fiscal incentives, for example, and credit support was also only available for formal firms with existing credit lines. The take-up rate started to increase when the government introduced programs such as BPUM. This cash program primarily targeted microenterprises, which represent a significant proportion of the private sector in Indonesia. Targeting This section analyzes the government’s targeting of support and the complementarity between household and business support programs. Targeting accuracy is measured based on the severity of the crisis and the probability of receiving support. The sample is divided into four categories, based on the severity of the impacts of COVID-19, in which group 1 consists of firms/households hit hardest by COVID-19. The analysis examines two rounds of support for businesses and households. Households experiencing smaller income shocks were initially more likely to receive social assistance, but that relationship waned over time. The options were to target beneficiaries based on socioeconomic status or the severity of the shock experienced, in terms of income shock relative to the pre-pandemic situation. The two metrics often differed during the pandemic. A household that was in the 50th percentile of the income distribution before the pandemic could be hit hardest by the crisis and fall into a lower percentile. If the government targeted beneficiaries based on pre-pandemic socioeconomic status, targeting had a lower probability of helping households pushed into poverty by the economic impact of the pandemic. Household support did not correlate with the severity of the income shock. In July 2020 and November 2021, there was no significant difference in the probability of receiving any social assistance between a household with limited or no drop in household income and a household with steep drop in income. The share of vulnerable households that benefited from a program at least once during the pandemic increased through 2021, but more than half of intended beneficiaries reported not receiving support (Figure 2.11). For example, only 46 percent of intended food voucher beneficiaries reported having received the voucher. The same exclusion occurred in other social assistance programs. This mistargeting might reflect the program’s design or problems with implementation; it was a well-documented weakness of Indonesia’s social assistance programs before the pandemic, with errors of exclusion and inclusion occurring along the entire welfare distribution system (Holmemo and colleagues, 2020). During the initial phase of the pandemic, when adverse impacts were felt almost universally, these inclusion errors implied that many affected households benefited from expanded assistance programs even if they were not in the target group. Self-reported benefit levels of social protection programs were on par with the amounts announced by the government. Another way to measure a program’s accuracy is to compare the actual benefit received and the intended benefit of the program. The actual benefits received by the majority of the beneficiaries were consistent with the intended benefit. For example, most of the beneficiaries of PKH reported having received Rp 200,000, which was exactly the amount of planned benefit allocated for this support. CHAPTER 2: INDONESIA 79 FIGURE 2.11  Targeting accuracy of household support in indonesia, by program and income level, March 2021 80 70 60 Percent of households 50 40 30 20 10 0 Bottom 30% Middle 50% Top 20% Bottom 20% Middle 60% Top 20% Bottom 30% Decile 4 Middle 40% Top 20% Bottom 40% Lower middle 20% Upper middle 20% Top 20% Bottom 40% Middle 40% Top 20% Bottom 40% Middle 40% Top 20% Bottom 40% Middle 40% Top 20% Sembako card Program Keluarga Bantuan SosialTunai - BST Bantuan Tunai Dana Electrcity bills Wage subsidy Bantuan Produktif Harapan (PKH) Desa (BLT-DD) reduction Usaha Mikro (BPUM) Source: World Bank data based on high frequency phone survey round 5 Note: Figures are percentages of households that reported receiving a program benefit since the onset of the pandemic. A solid red bar represents incidence in the intended group of beneficiaries. Error bars show 90 percent confidence intervals. PKH is restricted to households with children under 18 and/or households that reported receiving PKH benefits. More severely hit firms were initially less likely to receive assistance, but this association also waned over time. In June 2020, compared with firms with the most severe decline in sales (this category was omitted in Figure 2.12), firms that experienced little or no drop in sales had a higher probability of receiving support, presumably because programs targeted formal firms. With the introduction of the cash programs in August 2020, targeting accuracy improved slightly. During this year, support reached all firms, regardless of the severity of sales drop. There was no significant difference in the probability of receiving assistance among firms with different degrees of income losses. There were several reasons for this low targeting performance. First, the severity of the crisis was never part of the eligibility criteria for program beneficiaries. Second, even if it was part of the design, the government had limited information to identify severely hit firms among negatively impacted firms. Third, the high degree of informality made it more challenging to rely on pre-COVID and current financial reports to signal the crisis severity. In the end, the government could only rely on the self-registration mechanism and support all applicants. Most large firms received some type of government support, whereas the percentage of MSMEs that did so is far smaller. A larger percentage of large firms also received fiscal incentives and credit assistance. Microenterprises were more likely to receive BPUM support and electricity subsidies. This heterogeneity reflected the characteristics of existing support for the private sector. Most types of support, such as fiscal incentives, credit assistance, and wage and social insurance subsidies were only available for formal firms. 80 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 2.12  Probability of a firm receiving assistance in indonesia based on decline in sales, June 2020 and September 2021 Probability of receiving assistancce 7.4*** 4 4.2* 0.8 1.6 (percent) –0.1 Moderate Limited None June 2020 September 2020 Source: World Bank calculations based on results of the business pulse surveys round 1 and 2 Note: Probability of receiving assistance is relative to the most severe decline in sales. Decline in sales is divided into four categories based on quartile: the most severe decline being in the first quartile, moderate in the second quartile, limited in the third quartile, and none in fourth quartile. Bars represent the marginal effect at-means after a logit regression of receiving assistance on decline in sales. Results control for size, strata, and province. Error bars indicate 95 percent confidence intervals. Significance levels: *** 1 percent level; ** 5 percent level; * 10 percent level. Households with non-farm businesses benefitted from business support programs, such as cash transfers for microenterprises (BPUM) and loan deferment (Figure 2.13). Indonesia’s private sector is dominated by micro and small firms, at 98 percent, and largely comprises informal firms, at 93 percent. A significant share of these firms are household businesses, considered farm and non-farm business activities. Household businesses could thus receive both household and business support. Households with a non-farm business had a slightly lower probability of receiving household support than other households; however, being a microenterprise allowed them to receive firm support, including BPUM and loan deferment, as 17 percent of households with non-farm business activity received each kind of support. FIGURE 2.13  Beneficiaries of social assistance programs and other relief measures in Indonesia between March 2020 and October 2021 60 Percent of households with 50 non-farm business (%) 40 30 20 10 0 on ) rd lua T-D sa KH n d P ro nt idy ST a ar me (B Mik ca Ke (BL De cti ap bs i (B rg D) ) Lo UM) tc du fer ko ar su a en ha na P an re ba aH de ge ( ym sa Tu iD ills m Wa an fU plo l Se na sia yb kti em Tu So cit du e- an ctr an o m Pr Pr ntu Ele ntu ra an og Ba Ba ntu Pr Ba Supports for households Supports for firms Source: World Bank data based on high-frequency phone survey round 5 and business pulse survey round 4. Note: Error bars show 90 percent confidence intervals. CHAPTER 2: INDONESIA 81 Effectiveness This section provides evidence on the effectiveness of the support in helping the household and private sectors to navigate the pandemic’s negative effect. The effectiveness of household support was measured by looking at household’s perceptions of the adequacy of support in helping to cover their needs. The effectiveness of business support was measured by looking at its correlation with probability of hiring and firing. Households that received support in April 2021 were divided based on the benefits received between April and November 2021. Seventy-five percent of households that received support in April 2021 were still receiving support in April 2021–November 2021 (group 1). The remaining 25 percent no longer received government support (group 2). These two groups were evaluated based on their total income ability, that is, their own income and support, to fulfill the household demand. More than 40 percent of households within these two groups reported that their income, including their own and support, was mostly or entirely adequate. However, households from group 2 were more likely to report large unmet needs; in other words, their income was not adequate to cover expenditures. In group 2, 11 percent of households reported that their needs were not met at all. In contrast, only 4 percent of households in group 1 claimed similar inadequacy. These findings suggest insufficient support and weak targeting. Some households in need may have been excluded from social assistance between March and October 2021. There was no correlation between receiving government support and the probability of hiring or firing workers (Figure 2.14). There are several possible reasons for the lack of impact on firms’ employment decisions. One is mistargeting—the failure to support firms most in need of it. Another is the size of the support, which may have been too small to influence hiring and firing. Firms hired more workers if sales or sales expectations increased. Hiring and firing were also closely related to mobility and other government restrictions. As the government introduced new restrictions following the Delta outbreak, firms fired workers and reduced the working hours and/or wages of remaining workers. Indonesia’s social registry, Data Terpadu Kesejahteraan Sosial (DTKS), is of limited use in implementing disaster-response programs that require eligibility assessments to include the poor and vulnerable. Most social assistance in Indonesia uses the DTKS for a list of potential beneficiaries. The government is obliged FIGURE 2.14  Correlates of firing and hiring workers in Indonesia in 2020–21 15 10 5.6 5 4.3 Percent (%) 2.1 1.6 1.0 0.2 0 –5 –10 Fiscal assistance Credit assistance Other assistance Firing Hiring Source: World Bank calculations based on results of the business pulse surveys round 1, 2, and 3. Note: Data was collected before Ramadan (April 2021). Bars represent the marginal effect at-means after logit of receiving assistance. Results control for size, strata, and province. Error bars show 95 percent confidence intervals. Significance levels: *** 1 percent level; ** 5 percent level; * 10 percent level 82 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA to update this list twice a year, but multiple reports document the inability of the national and local governments to do so effectively (TNP2K 2020; UN ESCAP n.d.). Mobility restrictions may also have further limited this updating mechanism. The government may need to combine the self-registration process, community targeting, and other scientific methods to update the database and target the potential beneficiaries of the programs. Small experiments in Indonesia showed that statistical methods, such as a proxy means test, are superior to community targeting in identifying the poor based on expenditure per capita (Alatas and colleagues, 2012). However, community targeting is more consistent with the perceived welfare rank of the individual community members. A World Bank report (Holmemo and colleagues, 2020) recommends improvements to the DTKS that include the following: ⦁ Expand coverage to up to 80 percent of the population, to help reach the uninsured informal sector and the “uncovered middle” ⦁ Facilitate dynamic data updating ⦁ Move to an absolute poverty ranking ⦁ Make the DTKS interoperable with other databases, such as tax, electricity, land/asset ownership, or automobile/ motorcycle purchases ⦁ Integrate the DTKS with geographic information systems (GIS), to enable rapid response to shocks and crisis Awareness of government support programs improved over time but remained low for some programs (Figure 2.15). The government’s ability to effectively announce business support remains very limited. In October 2020, only 16 percent of firms were aware of the government’s fiscal incentive; 48 percent of firms were aware of the electricity subsidy, although the figure rose to 80 percent in August 2021. This lack of FIGURE 2.15  Percent of firms in Indonesia that were aware of government support programs for firms, 2020–21 100 88 81 Percent of firms aware of programs (%) 80 65 67 60 53 48 48 37 40 29 28 21 16 20 0 Bantuan Produktif Electricity Wage subsidy Credit Badan Penyelenggara Fiscal Usaha Mikro subsidy assistance Jaminan Sosial (BPJS) incentive (BPUM) subsidy October 2020 March 2021 August 2021 Source: World Bank data based on business pulse surveys round 2, 3, and 4. CHAPTER 2: INDONESIA 83 FIGURE 2.16  Reasons why Indonesian firms did not participate in government support programs, 2020–21 50 Percent of reasons firms did not participate 40 35 32 30 in programs (%) 25 20 20 17 14 12 12 11 10 10 5 3 2 2 1 0 0 Not eligible Did not know Complex Did not need Did not fit Applied yet Did not expect Other why application with firms rejected to get October 2020 March 2021 August 2021 Source: World Bank data based on business pulse surveys round 2, 3, and 4. awareness particularly affected the government support programs for firms that used registration-based schemes (Figure 2.16). Business support should target firms that are viable but distressed. Although equity concerns necessitate a focus on firms that have been hardest hit by the crisis, it is important to avoid keeping zombie firms— those that would not have survived in the absence of the pandemic—alive. Deferrals and restructurings of outstanding loans through the KUR program, for example, provided relief to borrowers who were in good standing before the crisis. Lessons Learned Continued COVID-19 relief packages may have to balance the goals of meeting needs and containing spending and public debt. The government’s unprecedented fiscal response since the start of the pandemic required increased financing of the government budget from the central bank, commercial banks, and domestic investors (World Bank 2021b). The short-term priorities for fiscal policy will remain focused on containing the pandemic, providing relief to households and viable firms, and stimulating the economy. At the same time, the government’s commitment to lowering the deficit to the legal ceiling, 3 percent of GDP, by 2023 implies containing spending as well as accelerating revenue collection. Changes in the government’s COVID-19 response package between 2020 and 2021 reflect an attempt at balancing the two goals. Although the 2021 package, released in response to the mid-year Delta wave, was larger than the 2020 package, it scaled back funding to social assistance programs, even though the pandemic’s adverse impacts on employment and income persisted through 2021. In contrast, support to MSMEs and other business enterprises increased, with the objective of strengthening performance, creating employment, 84 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA and promoting growth. In 2022, the government reduced the size of the response package to a level lower than that of 2020. Additional increases will go toward economic strengthening programs that focus on job creation and empowerment, whereas allocations to social protection will be reduced. Improved targeting and reprioritization of fiscal support could help create fiscal space for more urgent needs. The government’s COVID-19 relief programs disbursed allocated budgets and largely met their numerical targets in 2021. However, distributing support to the worst-affected households and firms remains a challenge. The pandemic brought into sharp focus constraints on the ability of the social assistance system to reach the hardest-hit households in a timely manner. Social assistance could be improved through better program targeting and delivery by updating and expanding the DTKS database consisting of the poorest 40 percent, improving the identification of households in need through universal digital identification, and conducting more regular program monitoring. Firms most adversely affected by the pandemic were not likely to receive government assistance. Support to viable firms could be better targeted to MSMEs, and financing for large firms could come from commercial banks (World Bank 2021b). This reallocation would help reduce short-term fiscal needs, improve the effectiveness of the fiscal response, and stimulate private credit. The pandemic has highlighted both the need for unemployment insurance and the constraints inherent in relying on a traditional contributory scheme linked to formal labor market status. Such a scheme excludes most workers and their families from coverage against short-term risks (Holmemo and colleagues, 2020), a key challenge for the new unemployment insurance, Jaminan Kehilangan Pekerjaan (JKP). The JKP holds the promise of protecting formal sector workers against income and employment shocks, such as those experienced during the pandemic. However, millions of workers in Indonesia do not have work contracts and thus lack the benefits of traditional contributory social insurance schemes or mandated worker protections. One way of improving coverage of informal sector workers would be to channel disaster relief support to them through household-level mechanisms. Introducing new incentives to encourage voluntary participation in social insurance could also help. The long-term goal should be to insure all workers, including the self-employed and workers in the informal sector, against the risk of employment shocks. The persistence of adverse impacts among MSMEs and socioeconomic groups that were vulnerable pre-COVID points to the potential for increased inequality. The pandemic hit smaller firms and non-farm businesses, as well as firms in low valued-added services, hardest. These firms were more likely than other firms to cope by reducing wages or work hours. Smaller firms were also less likely to receive government assistance, with barely a third of them receiving BPUM assistance, even though the program targeted MSMEs. Data from breadwinners indicates that employment and income disruptions were also more common among female and low-skilled workers. The slow take-up rate of firm support programs during the pandemic highlights the need for a comprehensive formalization strategy in Indonesia. High informality and the absence of MSME databases forced the government to rely on self-registration to deliver business support programs. This mechanism is not efficient or optimal in the long term. Because of asymmetric information of applicants and the lack of correlation between crisis severity and labor productivity, use of such a system could result in the government providing support to zombie firms that would have failed regardless of the pandemic. To increase formalization, the government may need to provide incentives for more MSMEs to formalize. More formal firms would improve targeting accuracy, increase program effectiveness, and reduce fiscal pressure. CHAPTER 2: INDONESIA 85 Annex 2 Annex 2A.  Household and Firm Surveys in Indonesia The Indonesia HFPS was designed to be nationally representative, with a panel of about 4,000 households followed in each round. It was conducted by phone interviews that lasted 20–30 minutes per household. Six rounds of the survey were conducted in 2020–2021: four between May and November 2020 and two in 2021 (Table 2A.1). Attrition was low: 80 percent of the households interviewed in round 1 were re-interviewed in round 6, and 77 percent of the balanced-panel households were re-interviewed in all follow-up rounds. The Indonesia BPS used the 2020 Occupational, Employment, and Vacancy Survey (OEVS) as the sampling frame. The BPS is a representative sample of three economic groups: (a) high value–added services (b) creative and tourism industries, and (c) the manufacturing industry. The OEVS sample was constructed using the 2016 Economic Census Listing Directory for services and the 2017 Medium and Large Manufacturing Survey Directory for manufacturing. The BPS followed a panel of 850 formal firms over four rounds, two in 2020 and two in 2021 (Table 2A.2). TABLE 2A.1  Dates and sample sizes of high-frequency phone surveys of households in Indonesia Round Survey dates Sample size 1 May 1–17, 2020 4,338 2 May 26–June 5, 2020 4,119 3 July 20–August 2, 2020 4,067 4 November 3–15, 2021 3,953 5 March 12–24, 2021 3,686 6 October 18–31, 2021 3,471 Source: World Bank data based on high frequency phone surveys. TABLE 2A.2  Dates and sample sizes of business pulse surveys in Indonesia Round Survey dates Sample size 1 June 15–July 23, 2020 850 2 October 16–November 4, 2020 1,381 3 March 5–26, 2021 1,360 4 August 1–October 3, 2021 1,175 4 August 1–October 3, 2021 1,175 Source: World Bank data based on business pulse surveys. Note: The business pulse survey sample includes firms that were formally registered with a government authority at the time of the survey. 86 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA References Alatas, Vivi, Abhijit Banerjee, Rema Hanna, Benjamin A. Olken, and Julia Tobias. 2012. “Targeting the Poor: Evidence from a Field Experiment in Indonesia.” American Economic Review 102 (4): 1206–40. https://www.aeaweb.org/ articles?id=10.1257/aer.102.4.1206. Allard, T. 2020. “Burial Numbers in Jakarta Indicate Coronavirus Toll Is Higher than Officially Reported.” Reuters, May 1, 2020. https://www.reuters.com/article/us-health-coronavirus-indonesia-cases-idUSKBN22D4XI. Hale, T., Angrist, N., Goldszmidt, R. et al. 2021.A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker). Nat Hum Behav 5, 529–538. Holmemo, Camilla; Acosta, Pablo; George, Tina; Palacios, Robert J.; Pinxten, Juul; Sen, Shonali; Tiwari, Sailesh. 2020. Investing in People: Social Protection for Indonesia’s 2045 Vision. World Bank, Jakarta. https://openknowledge. worldbank.org/handle/10986/33767. IMF (International Monetary Fund). 2021. “Database of Fiscal Policy Responses to COVID-19.” https://www.imf.org/ en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response-to-COVID-19. Ministry of Finance. 2021. Multiple presentations on Indonesia’s COVID19 Recovery Program (PEN). Jakarta. https://pen.kemenkeu.go.id/in/home. Mulyanto, R. 2020. “Data Science Project Reveals Indonesia Covid-19 Death Toll Three Times Higher Than Official Tally.” Telegraph, August  17, 2020. https://www.telegraph.co.uk/global-health/science-and-disease/data-science- project-reveals-indonesia-covid-19-death-toll-three/. Purnamasari, Ririn Salwa; Ali, Rabia. 2020. “Indonesia COVID-19 Observatory: High-Frequency Monitoring of Households: Summary of Results from Survey Round One” (English). Indonesia COVID-19 Observatory brief no. 3 Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/115001604470985121/ Indonesia-COVID-19-Observatory-High-Frequency-Monitoring-of-Households-Summary-of-Results-from- Survey-Round-One. Sari, A. Virgi, Ririn Purnamasari, and Ade Febriady. 2022. The Socio-economic Impact of COVID-19 in Indonesia: One Year into the Pandemic. Jakarta: World Bank. TNP2K (The National Team for the Acceleration of Poverty Reduction). 2020. Covid-19 Pandemic and the Momentum to Strengthen the National Targeting System. Jakarta. UN ESCAP (United Nations Economic and Social Commission for Asia and the Pacific). n.d. Facilitating COVID Responses: Data Terpadu Kesejahteran Sosial (DTKS). UN ESCAP, Bangkok. https://www.socialprotection-toolbox. org/practice/facilitating-covid-responses-data-terpadu-kesejahteran-sosial-dtks. World Bank. 2020. Indonesia Economic Prospect: Towards a Secure and Fast Recovery. December 2020. World Bank. Jakarta. https://www.worldbank.org/en/country/indonesia/publication/december-2020-indonesia-economic-prospects. World Bank. 2021a. “COVID-19 Impact on Firms in Indonesia: Insight from Business Pulse Survey (Round 4).” World Bank, Jakarta. https://www.worldbank.org/en/country/indonesia/brief/indonesia-covid-19-observatory. World Bank. 2021b. Indonesia Economic Prospects, December 2021: A Green Horizon-Toward a High Growth and Low Carbon Economy. World Bank, Washington, DC. World Bank. 2021c. “Indonesia High-Frequency Monitoring of COVID-19 Impacts (Round 5). World Bank, Jakarta. https://www.worldbank.org/en/country/indonesia/brief/indonesia-covid-19-observatory. World Economic Forum (WEF) (2019) The Global Competitiveness Report, 2019. https://www3.weforum.org/docs/ WEF_TheGlobalCompetitivenessReport2019.pdf. CHAPTER 3 Malaysia by Smita Kuriakose, Ririn Salwa Purnamasari, Trang Thu Tran, Sarah Waltraut Hebous, and Zainab Ali Ahmad25 The pandemic and resulting prolonged and stringent mobility restrictions have adversely affected many Malaysian households and businesses, particularly vulnerable groups, and small- and medium-size enterprises (SMEs). In the early stages of the pandemic, vulnerable groups and SMEs were disproportionately affected by the crisis. Although gradual recovery occurred across the board in the second half of 2021, progress has been uneven, increasing the risks of greater inequality. This chapter provides a deep dive into the impact of the pandemic on firms and households and their coping mechanisms. It is based on four rounds of a business pulse survey (BPS) and two rounds of a high- frequency phone survey (HFPS). The BPS focuses on formal firms. It captures basic information on their operational status, adjustment responses, and access to government support policy. The four rounds of the BPS were implemented between October 2020 and March 2022. The HFPS looks at the impact of, and recovery from, the COVID-19 pandemic by Malaysian households. These surveys include both individual and household-level questions on topics such as health, employment and income, social safety nets, coping strategies, and education. The first HFPS was conducted between May and June 2021. The second was conducted between October and November 2021 (see Annex 3A for more details). Timeline of the COVID-19 Pandemic and Government Measures Malaysia faced the pandemic with an especially favorable business environment, compared to other countries in the East Asia and Pacific region. With relatively flexible labor practices and deep financial markets, reallocation of factor inputs is facilitated in the country. A generally strong business dynamism is marked by firms’ ease of entry and exit and an elevated degree of domestic competition, despite sub-par trade openness marked by moderately high and complex tariffs (WEF 2019). After keeping the spread of COVID-19 relatively well-contained for most of 2020, Malaysia saw a dramatic rise in cases in mid-2021 (Figure 3.1). It was almost a year until the first 100,000 cases were recorded; between April and August 2021, the total number of cumulative cases surpassed 1.3 million, and the cumulative number of deaths exceeded 14,000 (Ministry of Health 2021). To curb the spread of the pandemic and ease the burden on the health system, the government reimposed several Movement Control Orders starting in January 2021. In June 2021, when Malaysia began 25  The chapter greatly benefited from input provided by Kok Onn Ting. Kyung Min Lee and Jesica Torres contributed to the framework to integrate firm level analysis with household data analysis. 87 88 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 3.1  Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Malaysia, 2020–22 HFPS R1 HFPS R2 80 100 Stringency index (14-day average) Inverted mobility index (14-day average) 60 80 Stringency index 40 60 Inverted 20 Mobility index 40 0 20 –20 0 BPS R1 BPS R2 BPS R3 BPS R4 100 COVID-19 cases/deaths (per 100,000 people) 75 COVID-19 cases 50 COVID-19 25 deaths 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Source: Mobility data comes from Google Mobility Reports; stringency data comes from the Oxford Covid-19 Government Response Tracker (OxCGRT); and COVID-19 cases and deaths from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. Note: The mobility index is the average percentage change in time spent at retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. For each day of the week, the baseline is defined as the median value from the five-week period of January 3–February 6, 2020. The figure shows 14-day rolling averages of the mobility index, which is inverted so that higher values indicate lower mobility. The stringency index records the strictness of lockdown style policies; 14-day rolling averages are shown. COVID-19 cases and deaths are smoothed by a seven-day average. Phases of the pandemic are determined by identifying structural breaks in COVID-19 cases and the mobility index. See introduction for further details. experiencing a surge in Delta variant cases, the government ordered a total nationwide lockdown, which saw most sectors of the economy closed or operating at limited capacity. Malaysia has achieved impressive progress in its vaccine roll-out since June 2021. By mid-March 2022, 98 percent of adult Malaysians had been fully vaccinated, and 65 percent had received their booster (third dose). The government has been gradually easing movement restrictions since late August 2021; in October 2021, it started to allow interstate and overseas travels for fully vaccinated Malaysians. In April 2022, Malaysia entered the endemic phase, with borders reopened for international travelers with minimal testing requirements. Employment Impacts: Shocks and Recovery Reduced employment and labor income Employment at formal firms fell substantially during the pandemic. The negative shocks to employment were persistent, on both the extensive (laying off workers) and intensive (reducing wages and/or hours) CHAPTER 3: MALAYSIA 89 margins. Throughout the three firm survey rounds, almost 40 percent of firms continued to let workers go, and more than three firms in four reduced hours and/or wages (Figure 3.2). These adjustments slowed net job creation and increased employment losses. With restrictions being lifted, the economy is showing signs of recovery, as the share of firms experiencing negative shocks on both the extensive and intensive margins is declining (Figure 3.2). In July 2021, average employment had fallen by 19 percent relative to the pre- pandemic level, but by March 2022, this employment decline reduced to only 3 percent.26 New forms of work appear to have absorbed some of the shock to the formal sector. A year into the pandemic, most workers remained employed, but there was considerable disruption. By November 2021, almost one-quarter of workers who were working pre-pandemic faced income reduction, and 21 percent had stopped working (Figure 3.3). One-fifth of those working in November 2021 had recently entered or regained their employment after being unemployed in mid-2021. These workers were more likely to be young (ages 18–24) and less educated, suggesting their greater vulnerability. At the same time, the high level of movement indicates flexibility and low barriers to entering low-skilled labor markets during economic downturns. Anecdotal evidence suggests that one of the potential drivers of increased entry into informal work in the early phases of the pandemic was a significant increase in gig work, such as food delivery and courier services, which flourished during lockdowns (Malay Mail 2021). The share of new entrants was larger among informally employed workers than among the formally employed, particularly in the early pandemic period. However, the share of formally employed wage workers increased toward the end of 2021, in tandem with the reopening of the economy. The increase was partially the result of government interventions, including incentives for firms to hire new workers through the Penjana Kerjaya 2.0 scheme (Ministry of FIGURE 3.2  Share of firms laying off workers and share of firms reducing hours and/or wages in Malaysia, October 2020–March 2022 80 78 77 75 59 60 Percent of firms 41 40 40 36 37 20 0 Oct 20 Jan–Feb 21 Jul 21 Feb–Mar 22 Laid o workers Reduced hours and/or wages Source: World Bank data based on business pulse surveys  Employment change in this chapter is calculated following Davis and Haltiwanger (1998), to take into account extreme values and 26 differences in initial employment size. The change in employment is the difference between two periods as a percentage of the average of the two periods. The calculation excludes the top and bottom 1 percent (outliers) of change in employment. 90 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 3.3  Employment dynamics of Malaysians who were working before the pandemic, February 2020–October/November 2021 Percent of people working before the pandemic 41% 55% 100% 34% 24% 25% 21% February 2020 May/June 2021 October/November 2021 Working Working and income remained Working and Stopped working the same or rose income fell Source: World Bank data based on high-frequency phone surveys. Finance 2022) and matching contributions for social insurance among self-employed and gig workers, which brought them into the formal workforce through the PenjanaGig and SPS Lindung programs.27 Uneven impact and increase in inequality The pandemic has had a disproportionate impact on certain sectors, such as tourism, retail/wholesale trade, and construction, which are more sensitive than other sectors to mobility restrictions. Reduced output translates into labor demand shocks. Sectors in which firms faced larger sales losses also had larger shares of workers who stopped working (Figure 3.4), most likely because they are the sectors that rely most on in-person interactions. However, movements between sectors were not clearly observed. Employment disruptions widened existing inequalities, as vulnerable and disadvantaged workers were disproportionately affected by the pandemic shocks. Among formal firms, average employment losses in small firms were significantly greater than losses in medium-size and large firms (Figure 3.5). Throughout the pandemic, the self-employed and those who worked in a family business were more severely affected than wage workers, in terms of both work stoppage and income loss (Figure 3.6). These workers tend to be less educated than people who work in the formal sector, suggesting that the pandemic may have exacerbated existing socioeconomic inequalities. The pandemic disproportionally affected workers who were younger, less educated, and earned less than other workers. Throughout 2021, the share of people who were working in February 2020 that subsequently  Under the PenjanaGig scheme, the government subsidizes a matching contribution of more than 70 percent, in order incentivize registration 27 with the Self-Employment Social Security Scheme under SOCSO (World Bank 2021c). CHAPTER 3: MALAYSIA 91 FIGURE 3.4  Correlation between change in sales and share of Malaysians working in 2021, by sector 60 Construction the pandemic who had stopped working in Share of respondents who worked before 50 Transport and travel between March 2020 and June 2021 Retail and wholesale 40 Manufacturing Utilities 30 Agriculture Education, health, and others 20 Professional activities 10 –0 –25 –20 –15 –10 –5 0 Change in sales (percent) Source: World Bank data based on business pulse surveys and high-frequency phone surveys. Note: The size of the bubbles indicates sectoral employment. The dotted line indicates a linear trend line. FIGURE 3.5  Change in employment in FIGURE 3.6  Share of workers in malaysia that Malaysia, October 2020–March 2022, stopped working or suffered income shock, by firm size May/June–October/November 2021 0 70 Change in employment relative to baseline (percent) 60 –5 50 Percent of workers 40 –10 30 20 –15 10 0 May/June October/November May/June October/November –20 2021 2021 2021 2021 Micro (0–4) Small (5–19) Medium (20–99) Large (100+) Work stoppage Income loss Firm size (number of employees) Self-employed/family business Wage worker Source: World Bank data based on business pulse surveys. Source: World Bank data based on high-frequency phone surveys. Note: Error bars indicate 95% confidence intervals. 92 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA stopped working declined only marginally, from 25 percent in May/June to 21 percent in October/November. But young (ages 18–24), less educated, and lower-income workers were more likely than others to have stopped working during the pandemic (Figure 3.7).28 Women were continuously more likely to stop working. Among households with children that reported at least one adult member who had stopped working or reduced working hours to accommodate for childcare, women were at least twice as likely as men to shoulder these additional duties. On the other hand, men were more likely to endure income losses and have changed jobs. These results were driven partly by the sector, which encompasses agriculture, construction, and transport and travel, as well as region, but gaps across subgroups remain even after controlling for both, implying greater vulnerability among disadvantaged individuals.29 Long-term consequences of reallocation shock for firms and workers After the reopening of economic and social activities in late 2021, output and employment gradually recovered. Conditions in the labor market improved. The unemployment rate in the fourth quarter of 2021 dropped to 4.3 percent, from 4.8 percent in the fourth quarter of 2020 (Department of Statistics Malaysia 2022). By November 2021, half of Malaysian households that had fallen into lower income brackets than their pre-pandemic income level had returned to their pre-pandemic income level in June 2021. Among formal firms, sales in February/March 2022 were just one percent lower than pre-pandemic levels—the highest level since March 2020. There is a potential silver lining to firm and household adjustment responses to the COVID-19 crisis: the adoption of digital technologies. For firms, digital adoption was the most important coping strategy at FIGURE 3.7  Share of Malaysians that stopped working during the pandemic 70 Percent of adults who worked 60 50 in February 2020 40 30 20 10 0 Men Total 18–24 25–34 35–54 Women Secondary Primary/None 55 and above Post-secondary Technical/other RM 2,001–RM 4,000 RM 2,000 and below RM 4,001–RM 10,000 More than RM 10,000 Gender Age group Education level Monthly income before the pandemic May/June 2021 October/November 2021 Source: World Bank data based on high-frequency phone surveys. Note: Error bars represent a 90% confidence interval. 28  The minimum retirement age in Malaysia is 60. The large share of work stoppage among workers 55 and above may represent retirement rather than employment loss related to the crisis. 29  The increase in the number of gig workers with limited social protection, the rise in underemployment, a high unemployment rate, and wage stagnation affected many workers, particularly younger and lower-skilled workers, before the pandemic. CHAPTER 3: MALAYSIA 93 the height of the lockdown. In December 2020, nearly 60 percent of surveyed firms cited an increase in the use of digital platforms as the most important coping strategy in response to mobility restrictions. Among households, use of digital transactions also intensified. In 2021, 74 percent of adults engaged in digital transactions; nearly half of them reported increasing their use over time, mostly for online shopping. The use of working-from-home arrangements helped firms to retain their employees. Acceleration of technology adoption during the pandemic created an opportunity to push both firms and workers to be more productive and reallocate resources toward more productive use. Between October 2020 and July 2021, the average share of remote workers in Malaysia increased from 29 percent to 40 percent. Use of working-from- home arrangements is significantly correlated with lower employment losses at the firm level. Evidence from the manufacturing sector suggests that employment reallocation in the formal sector has been consistent with a “creative destruction”30 pattern. That is, firms that were more productive were less likely to shed workers. The pandemic shock might therefore improve aggregate productivity by weeding out weak firms and allowing room for stronger firms to grow (Figure 3.8). Nevertheless, gaps exist in firms’ and workers’ ability to adopt digital technologies. The lack of access to digital infrastructure in poorer regions is inversely correlated with the likelihood of firms to use remote work arrangements. People living in poorer regions tend to have inadequate digital infrastructure, with less than half of households in poorer regions—including Sabah and Kelantan—having Internet access through a fixed broadband connection.31 The share of workers working from home is also lower in these FIGURE 3.8  Correlation between change in employment between October 2020 and March 2022 and pre-pandemic labor productivity in Malaysia 10 Change in employment relative to baseline (percent) 0 –10 –20 –30 –4 –3 –2 –1 0 1 In (baseline labor productivity) Source: World Bank data based on business pulse surveys. Note: Baseline labor productivity is measured by 2019 sales per worker. Figure excludes the top and bottom 1 percent outliers of change in employment and the top and bottom 5 percent outliers of ln(baseline labor productivity). The independent variable is grouped into bins; each dot represents the average value of the independent and dependent variable within that bin. 30  Creative destruction refers to the phenomenon of economic change through the creation of new ways of doing things that endogenously destroy and replace the old ways. In this case, it refers to the exit of unproductive firms and the entry of new, more productive firms. 31  According to the 2019 Household Income and Basic Amenities Survey (HISBA), Sabah had the highest incidence of absolute poverty, at 19.5 percent. Kelantan followed, at 12.4 percent. The national poverty incidence for Malaysia in 2019 was 5.6 percent. 94 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 3.9  Correlation between working from home during July 2021 and household access to fixed internet connection during May–June 2021 in Malaysia, by region Negeri Sembilan 50 Sabah Kuala Lumpur Average work from home rate (%) 40 Selangor/Petaling Jaya Perlis Johor Melaka Pahang Sarawak Kelantan Kedah Pulau Pinang 30 Perak Cyberjaya/Putrajaya 20 Terengganu 30 40 50 60 70 Share of households with broadband Internet access (%) Source: Data on access to the Internet is from the high-frequency phone survey; data on working from home and change in sales is derived from the business pulse survey. Note: Results are conditional on proxies for the magnitude of the shocks (the average change in firm sales) and worker ability to work-from-home (average skill level). That is, the correlation between working-from-home incidence and Internet access was plotted after controlling for average change in sales and average share of high-skilled workers (at least post-secondary education) at the regional level. regions, potentially exacerbating existing income inequalities across regions (Figure 3.9).32 Although the crisis facilitated digitalization, larger firms were more likely to have incorporated digitization into their production processes than smaller firms. Large firms were also more likely to adopt more advanced digital technologies, such as enterprise resource planning software, and to have a large online presence than SMEs. This digital divide is driven partly by gaps in firm and worker capabilities. Nearly half of individual sellers who do not use digital platforms for their businesses cite the lack of knowledge about using them or a lack of awareness of the platforms themselves as reasons for not using digital platforms. A relatively low level of digital skills among adults may constrain increases in digitalization in Malaysia. There is also a risk of increased skill mismatches, which would dampen the potential productivity dividend from reallocation and technology adoption. Among higher-skilled workers, digital adoption shifts labor demand toward people with digital skills. The loss of foreign workers is also likely to pose a problem for recovery. As of August 2021, 88,000 foreign workers had been repatriated under the Recalibration for Illegal Immigrants repatriation program. Some 107,500 foreign workers registered for the program (Zokkepli 2022), yet the COVID-19 crisis exacerbated  These results are consistent with findings from Rahman, Jasmin, and Schmillen (2020) using 2020 Labor Force Survey data. They find 32 that 65 percent of jobs in Malaysia cannot be performed from home, and that workers most at risk are those who were already vulnerable before the crisis because of their low level of education, low level of income, and youth. Jobs in less developed regions of Malaysia are also particularly vulnerable. CHAPTER 3: MALAYSIA 95 pre-pandemic shortcomings in the labor market. The gaps in digital skills and lack of a transparent regulatory framework to protect foreign workers created labor constraints that could slow recovery if left unaddressed. In March 2022, more than 80 percent of surveyed firms cited difficulty finding skilled labor, and 58 percent cited difficulty finding unskilled workers. The asymmetric nature of the shock has the potential to widen inequality, both now and in the future. The share of workers that switched jobs declined to 10 percent in the second half of 2021, down from 19 percent in the first half of the year. However, younger and less educated workers, who were generally more vulnerable before the pandemic, were more likely to switch jobs. Income reduction is also more likely among reemployed workers than those who continued working, at 44 percent versus 28 percent, respectively. Job instability risks heighten vulnerabilities in the future. Entering the labor force during economic downturns could possibly force some, particularly younger workers, into accepting lower-quality and lower-paying jobs. A large body of literature shows that lack of job security is expected to reduce potential earnings and lead to deterioration in worker skills, thereby limiting long-term career development prospects. Fiscal Support to Firms and Households Size and composition of the fiscal response Malaysia’s increased fiscal spending in response to the pandemic was substantial, compared to many other countries in the region, but was lower globally than countries with similar levels of public health policy restrictiveness. To contain the COVID-19 outbreak, the government imposed measures that placed Malaysia in the 85th percentile of countries for which comparable data is available.33 According to the International Monetary Fund (IMF 2021), by June 2021, direct spending and forgone revenue on key fiscal measures announced or taken in response to the pandemic stood at just below 10 percent of GDP. This share was higher than all but four countries in the East Asia region—Singapore, Thailand, the Republic of Korea, and Mongolia—for which data is available.34 Among countries with similar levels of policy restrictiveness, however, Malaysia’s level of spending is slightly below average (Hale and colleagues, 2021). The government’s responses included a diverse set of mitigating measures (Table 3.1 and Table 3B.1). These measures included two large economic stimulus packages consisting of direct and indirect government support— PRIHATIN (RM 250 billion) and PEMULIH (RM 150 billion). In addition to covering significant COVID-19 health care expenditures, the government introduced a range of short-term income and liquidity relief measures to support firms and households. Support for firms ranged from fiscal exemptions, such as deferrals of taxes and labor contribution, to direct grants and loans to increase liquidity or retain workers through schemes, such as the special relief facility fund for SMEs and micro-credit financing schemes. Support for households included direct cash transfers and indirect support through the Employees’ Provident Fund (EPF) withdrawal facilities and loan moratoriums.35 Over the course of the pandemic, the government 33  Figure is based on the latest policy restrictiveness index data from the Oxford Covid-19 Government Response Tracker (OxCGRT). 34  While the IMF estimates of the size of the fiscal package differ from those reported in the introduction and chapter 8 (based on World Bank estimates from the Chief Economist Office), the conclusion remains the same. Among the six countries studied, Malaysia’s fiscal package was the second largest, behind Mongolia. 35  EPF’s special schemes allow for early withdrawal of retirement savings through i-Lestari, i-Sinar, and i-Citra withdrawal programs. The scheme affected 7.4 million EPF members and comprised RM 101 billion in total withdrawals by October 2021. For more information, see: https://www.kwsp.gov.my/ms/-/epf-focused-on-rebuilding-members-retirement-savings-following-exceptional-withdrawal-facilities. 96 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE 3.1  Instruments Malaysia introduced to support firms and households during the pandemic Target group Measures Firms • Tax deferrals and reductions; deferrals of labor contribution • Support for fixed costs of Internet and discounts on electricity • Access to credit, through direct grants for vulnerable sectors, soft loans, micro-credit, and loan guarantees • Wage subsidies • Hiring and training incentives • Grants and loans to promote digital adoption • Measures to reduce the cost of doing business Households • Emergency cash transfers, including Bantuan Prihatin Nasional (BPN), Bantuan Prihatin Rakyat (BPR), and Bantuan Khas COVID-19 (BKC) • Income or employment loss assistance through the Employment Insurance System (EIS) • Income tax relief and exemptions (for electronics, domestic travel expenses, COVID-19 testing) • Food basket programs • Employees’ Provident Fund retirement savings withdrawal facilities • Loan moratoriums • Free Internet and electricity subsidies also introduced additional measures, such as advisory services and financial incentives, to promote digital adoption. Multiple schemes—including the JanaKerja employment creation scheme and the PenjanaKerjaya and JaminKerja hiring and training incentives programs—also supported labor market recovery (Ministry of Finance 2022). Coverage and timing These measures reached a large share of firms and households. By October 2020, more than 90 percent of formal firms had received some form of government support. Coverage decreased marginally over time, but access to support still remained high as of July 2021, at more than 90 percent. On the household side, the government introduced several emergency cash transfer programs during the pandemic, with eligibility criteria that included even the middle class.36 Bantuan Khas COVID-19 (BKC) was extended to poor and vulnerable households earning up to RM 9,000 (US$2,140) and individuals earning up to RM 5,000 (US$1,189) per month. It was complemented by special transfers, such as Bantuan E-Hailing for ride-hailing and taxi drivers, and increased benefit levels within regular cash assistance programs.37 In June through November 2021, 65 percent of households received government cash assistance, up from 53 percent by June 2021. Targeting Government support was predisposed toward larger firms and less likely to go to more vulnerable firms. In all BPS rounds, access to government assistance for micro and small firms was consistently lower than that of 36  Based on the eligibility criteria in Bantuan Prihatin Nasional (BPN), the bottom 40 percent is defined as households earning no more than RM 4,000 and individuals earning no more than RM 2,000 a month. The middle 40 percent is defined as households earning RM 4,001–RM 8,000, and individuals earning RM 2,001–RM 4,000 a month. 37  Pre-pandemic cash assistance programs included Bantuan Orang Tua (BOT), Bantuan Kanak-Kanak (BKK), Bantuan Orang Kurang Upaya (BOKU), and the rebranded Bantuan Keluarga Malaysia (BKM). CHAPTER 3: MALAYSIA 97 FIGURE 3.10  Access to government support in Malaysia, October 2020–March 2022, by firm size 100 97 96 93 92 93 92 93 91 86 88 82 Percent of establishments that receive public support 79 80 75 73 67 60 60 40 20 0 Micro (0–4) Small (5–19) Medium (20–99) Large (100+) Firm size (number of employees) Oct 20 Jan–Feb 21 Jul 21 Feb–Mar 22 Source: World Bank data based on business pulse surveys. medium-size and large firms, and these gaps widened over time (Figure 3.10). Access by microenterprises was 25 percentage points less than that of large firms, and access by small firms was 10 percentage points lower. More vulnerable sectors—as proxied by the average drop in monthly revenue—also had lower than average access to support (Figure 3.11). Within sectors, firms with lower revenue losses were more likely to have received some government support. This result may be driven partly by the set of instruments used.38 Tax deferral, for example, favored larger firms and was not correlated with revenue loss. Other instruments may allow for more targeted selection. For example, although wage subsidies are more likely to go to large firms, they are also more likely to be provided to firms with larger drops in sales (Table 3C.1). Government support to households is pro-poor but could be more shock responsive. By November 2021, 79 percent of households had received some form of government assistance. This includes nearly 90 percent of low-income households with monthly earnings of RM 2,000 and below. However, nearly a quarter of these low-income households did not have access to cash transfers, while more than a third of households earning more than RM 10,000 received them, suggesting scope for improvements in targeting (Figure 3.12). Furthermore, the chance of receiving government assistance was the same for households that experienced economic shocks and those that did not. As a result, while both households that experienced economic shocks, and those that did not, engaged in a large number of coping strategies, affected households were more likely to reduce food and nonfood consumption, borrow from friends and family, use savings, and sell assets  It is not possible to disentangle targeting versus policy effectiveness using a correlation between access to support and the shock to sales. 38 98 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 3.11  Correlation between receipt of government support in Malaysia and change in sales, October 2020–March 2022, by sector 100 Automotive Percent of firms receiving government support Electronics 95 Other Manufacturing Tourism & Transportation 90 Utilities Fin/Real Estate/Biz Services Construction Comrnerce (food & beverage) Agriculture & Mining 85 Other Services 80 Other Comrnerce –20 –15 –10 –5 0 Percentage change in sales Source: World Bank data based on business pulse surveys (Figure 3.13). Some of these practices, such as the selling of assets, may widen socioeconomic inequalities and reduce the potential for catching up with others, unless appropriate policies are immediately put in place. The potential tradeoff between targeting more productive and more vulnerable firms could be mitigated by improving complementarities between firm and worker support. Although firm-level evidence suggests that support may deepen performance gaps between small and large firms, it also suggests that support has been going to more productive firms, reallocating resources toward more productive uses. In the manufacturing sector, the probability of receiving support was correlated with the firm’s labor productivity, FIGURE 3.12  Share of Malaysian households that received cash transfers, May/June–October/November 2021 100 80 Percent per household 60 40 20 0 RM2,000 and below RM2,001–RM4,000 RM4,001–RM10,000 More than RM10,000 Pre-pandemic monthly household income Source: World Bank data based on high-frequency phone surveys. Note: Error bars represent 90% confidence interval. CHAPTER 3: MALAYSIA 99 FIGURE 3.13  Coping strategies by Malaysian households that did and did not experience economic shocks, May/June–October/November 2021 Reduced non-food consumption Relied on personal savings Received assistance from government Reduced food consumption Relied on Employees Provident Fund/Public Services Pension… Took loan moratorium/delayed payments Engaged in additional income generation Borrowed from friends & family Sold assets Received assistance from NGO/other institutions Took out a loan from a financial institution 0 10 20 30 40 50 60 70 80 Percent of households Economic shock No shock Source: World Bank data based on high-frequency phone surveys. Note: Error bars represent 90% confidence interval. measured by average sales per worker before the start of the pandemic. To ensure that workers in less productive (but potentially more vulnerable firms) are not left behind, the government could target these workers through social assistance programs. To date, the government has not provided such support. At the individual level, access to social assistance is higher among formal workers at 78 percent, than among informal wage workers at 69 percent. Effectiveness The perception of the government’s crisis response improved over the course of the pandemic, with the share of people satisfied by the response rising from 57 percent in June 2021 to 84 percent in November 2021. Nevertheless, some concerns remained about the fairness of financial assistance and the government’s credibility in managing the crisis. Firm support has helped to preserve employment. Conditional on firm characteristics, including sales growth and liquidity, having access to wage subsidies is associated with an 8 percent lower likelihood that a firm laid off workers (Figure 3.14). Moreover, firms that received hiring incentives were also significantly more likely to have hired new workers. In contrast, social protection programs could have performed better. Although these programs included the middle 40 percent by income and gig workers during the pandemic, 55 percent of households were still under financial distress in late 2021. Even with an increased share of households receiving government assistance, about two-thirds of low-income households and those that experienced economic shocks 100 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 3.14  Correlation between various factors and the likelihood of laying off workers in Malaysia, January–February 2021 .4 .2 0 –.2 –.4 6 m e) su s to cess y me s or othe ures no s Ele ed Au ics ufa e g & b ion ) l e nspo ce s r s es Me s m e n ur ther rage ge nth e e on itie e an otiv rg rin te/ tatio diu t idi au lici Ot ervic vic r ne on (ra od truct ./r & tra me La l ctu o s i bs m ec r po er Ut ctr a e th r m ev o s s t ow o biz se he c in gr C t es idi a rm w les o tag sta u t O (fo tb liq to he Sa ism or ss Ot ce to ea sh ce pp c er ss A w Ac To mm su ce flo Fin Ac sh Co ss Ca ce ac No Source: World Bank data based on business pulse surveys. Note: Coefficient estimates from regressing laying off workers dummy on sales change, liquidity, access to support policies, and baseline sector and employment size fixed effects. Error bars indicate 95% confidence intervals. self-assessed as having inadequate financial resources to cover their monthly basic needs in October/November 2021 (Figure 3.15). This finding could partly reflect the insufficient levels of assistance received, including cash assistance, for households. Before the pandemic, Malaysia spent only about 1 percent of GDP on social assistance programs; this figure more than doubled, to 2.1 percent of GDP during the pandemic (World Bank 2020, 2021c). Spending on social assistance was higher in the early stages of the pandemic through the BPN program. Later, average spending per recipient for both the emergency measure program, BKC, and regular cash assistance programs, such as the BPR and the Bantuan Keluarga Malaysia (BKM), were lower than the pre-pandemic program, Bantuan Sara Hidup (Household Living Aid) BSH Viewed together, these findings suggest that the scaling back of the benefit levels of support programs may have been premature, particularly for poor and vulnerable households who are still gradually recovering from the shocks. Lessons Learned To some extent, the labor market was quick to adjust to shocks during the crisis, which was also supported by timely and effective government labor market interventions. In the first year of the pandemic, both workers and firms took advantage of and adapted to new opportunities, which included the gig economy, CHAPTER 3: MALAYSIA 101 FIGURE 3.15  Share of households in Malaysia partially or fully unable to cover current monthly basic needs, October–November 2021 100 80 Percent of households 60 40 20 0 RM 2,000 RM RM Above No shock Economic Self-employed/ Family Wage and below 2,001–4,000 4,001–10,000 RM 10,000 shock family business farming worker Pre-pandemic monthly household income Economic shock Employment type Source: World Bank data based on high-frequency phone surveys. Note: Error bars represent 90% confidence interval. digital technologies, and remote work arrangements. At the same time, timely and effective government labor market interventions cushioned the impact of COVID-19 and promoted job creation during the crisis. These interventions included incentives for firms to hire new workers and matching contributions to social insurance for self-employed and gig workers that brought them into the formal workforce. Government assistance to households and individuals has been impressive, given the speed with which programs were deployed and their coverage. Multiple cash transfer programs were provided to target groups throughout the pandemic, with rapid delivery to reach intended beneficiaries. The first emergency cash transfer, the BPN, was disbursed in April 2020, just 10 days after its announcement (World Bank 2021c). This rapid roll-out was made possible by an efficient enrollment system, including automatic enrolment of existing BSH recipients and colleagues, eligible based on their personal income tax records, as well as manual enrollment of new eligible beneficiaries through a simple application process on the Inland Revenue Board’s website. Given tightening fiscal space and continued financial distress among Malaysians, better targeting and consolidation of support instruments are needed. The impact of COVID-19 on the poor and vulnerable would likely have been worse had there been no government intervention. In the short term, retaining current social assistance programs, including the newly introduced relief measures, is important to mitigate an increase in poverty and vulnerability. Challenges in program implementation include the targeting of intended beneficiaries and adaptability to cover households experiencing shocks. Eligibility for social assistance programs could be broadened to prioritize vulnerable groups, particularly young, less educated, self-employed or informally employed workers, and temporarily adjusted to cover households that experienced economic shocks during the crisis, at least until economic growth resumes. Over the medium to longer term, Malaysia would benefit from consolidating the range of social assistance programs, including beneficiary registries, resources, and delivery systems, to provide better coverage and protection to low-income and vulnerable households, thereby helping to promote equitable recovery and outcomes. Malaysia reached a large number of firms that needed assistance but may have spent resources on firms that may not have needed it as much. Many large firms received government support, suggesting that some 102 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA government spending could have been curtailed and better targeted toward smaller firms or firms in the most vulnerable sectors. As Malaysia transitions into the endemic phase of the virus, government support could be reviewed and redirected toward programs that improve growth potential over the longer term. Discussion of the optimal mix of policy instruments is beyond the scope of this chapter. However, data on firm-level access to select support instruments—tax deferrals, wage subsidies, hiring incentives, and grants for digitalization— indicates some direction for reallocation of resources across policy instruments and beneficiaries. Short- term relief measures, such as access to tax deferrals and wage subsidies, are easy to implement and useful in preventing widespread employment losses, but they are prone to mistargeting and misallocation of resources. Income tax deferrals, for example, affect only firms with profits, and wage subsidies encourage employment retention at the expense of the movement of workers toward more productive firms. In contrast, hiring incentives support medium- and long-term recovery in the labor market. Reorienting support away from employment retention and toward measures that support employment activation would therefore make sense. Given the potential of the pandemic to exacerbate skills mismatches, hiring incentives also need to be balanced with well-designed and well-targeted training programs.39 Direct targeting is difficult, but the choice of policy instruments can lead to the self-selection of firms or behaviors with desirable characteristics. When choosing an instrument, policy makers should prioritize instruments that can help more productive firms self-select behaviors, such as taking out loans, or incentivize investments that will contribute to increased productivity in the future, such as technology adoption and innovation. Investment incentives for firms need to be complemented by advisory support/ technical assistance to ensure that firms have adequate capacity to implement the investment. Data on access to digital support shows that most of the vulnerable sectors during this pandemic, such as tourism and transportation, have the lowest level of support for digitalization—an unsurprising finding given the lower skill levels in these sectors. There is also a need to realign policy support toward the current needs of firms, by providing advisory services for small firms to adopt more advanced technologies for production processes and supply chain management, for example. More comprehensive labor market policies are essential to address labor demand and supply constraints. These policies should not only provide short-term relief measures to reduce the adverse impacts of COVID-19, they should also foster employment creation and the matching of workers and jobs over time. Given the changing nature of work during the pandemic, active labor market policies play a crucial role in facilitating job transitions (World Bank 2020). Such policies can be strengthened by integrating the implementation of programs across ministries and agencies and increasing shock responsiveness during crises. These policies should offer a wide range of skill-building initiatives, including upskilling and reskilling training programs, to leverage and facilitate the reallocation of workers across sectors and occupations as needed. 39  Research based on data from the eRezeki platform (which enables citizens, especially low-income groups, to generate additional income by doing digital assignments via this online crowdsourcing platform), for example, suggests that increasing digital competencies increases the rates of job-seeking activity by women, young job seekers, and non-degree holders more than increasing skills and access does. CHAPTER 3: MALAYSIA 103 Annex 3 Annex 3A.  The Household and Firm Surveys in Malaysia The analysis presented in this chapter draws on data from the Malaysia High-Frequency Phone Survey to examine the impact of the pandemic on Malaysian households. The HFPS includes both individual- and household-level questions on topics such as health, employment and income, social safety nets, coping strategies, and education. The analysis is based on two rounds of the HFPS, the first which took place in May-June 2021 and the second which took place in October-November 2021 (Table 3A.1). To examine the impact of the pandemic on firms, this analysis uses data from a business pulse survey (BPS). The BPS focuses on formal firms. It captures basic information on their operational status, adjustment responses, and access to government support policy. The four rounds of the BPS were implemented between October 2020 and March 2022 (Table 3A.2). TABLE 3A.1  Dates and sample sizes of high-frequency phone surveys of households in Malaysia Round Survey dates Sample size 1 May 18–June 16, 2021 2,210 2 October 17–November 1, 2021 1,047 TABLE 3A.2  Dates and sample sizes of business pulse surveys in Malaysia Round Survey dates Sample size 1 October 2–22, 2020 1,500 2 January 15–February 10, 2021 1,500 3 July 8–28, 2021 1,500 4 February 4–March 6, 2022 1,500 Note: The business pulse survey sample includes firms that were formally registered with a government authority at the time of the survey. 104 Annex 3B.  Malaysian Government Assistance Packages During COVID-19 TABLE 3B.1  Type of assistance to firms and households based on various assistance packages throughout the pandemic Feb 20 Mar 20 Apr 20 Jun 20 Sep 20 Nov 20 Jan 21 Mar 21 May 21 Jun 21 Oct 2021 Stimulus PRIHATIN PRIHATIN PENJANA (RM Kita Budget 2021 PERMAI (RM PEMERKASA (RM PEMERKASAPLUS PEMULIH (RM Budget 2022 Package (RM (RM 250 SME (RM 35 billion) PRIHATIN (RM 322.5 billion) 15 billion) 20 billion) (RM 40 billion) 150 billion) (RM 332.1 billion) 20 billion) billion) 10 billion) (RM 10 billion) Direct grant Direct Direct grant Direct grant Direct grant Direct grant Direct grant Direct grant Direct grant Matching grant to taxi drivers, grant for digitalization tour bus drivers for MSME to adopt green Matching grant practices for tourism products Direct grant (domestic) and matching grant to tourism players Cash Cash Cash transfer, Cash Cash transfer, Cash transfer, transfer, transfer, Bantuan transfer, Bantuan Khas Bantuan Bantuan Bantuan Prihatin Rakyat Bantuan COVID-19 Keluarga Prihatin Prihatin (BPR) Prihatin (BKC) Malaysia (BKM) Nasional Nasional 2.0 Rakyat 1.0 (BPN) (BPN 2.0) (BPR 1.0) Wage Wage Wage subsidy Wage Wage subsidy Wage Wage subsidy Wage subsidy Wage subsidy Wage subsidy subsidy subsidy subsidy subsidy for tourism sector Matching Hiring Hiring incentives Hiring incentives, Hiring Hiring grant to incentives and training apprenticeship Incentives Incentives, Human allowances program reskilling and Resource upskilling CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Development training Fund to train programs workers Grant to train workers in digital skills (continues) Loan Loan Loan deferment Loan deferment deferment for bus and taxi deferment for SMEs hire (nonautomatic) (automatic) Soft loan Soft loan Soft loan Soft loan Soft loan Soft loan Soft loan Soft loan Soft loan Investment matching loan for creative industry Soft loan for tourism sector Microcredit Microcredit Microcredit Microcredit Microcredit Microcredit Microcredit Microcredit Microcredit Loan Loan guarantee Loan Loan guarantee Loan Loan guarantee guarantee guarantee to purchase guarantee for bus and delivery vehicle taxi hire purchase Employees’ Employer EPF Provident contribution Withdrawal Fund (EPF) for EPF is by contribution deferred employees reduction EPF withdrawal by employees Human HRDF HRDF HRDF HRDF HRDF HRDF Resources contribution contribution contribution contribution contribution contribution Development exemption exemption exemption exemption exemption exemption Fund (HRDF) contribution exemption (continues) CHAPTER 3: MALAYSIA 105 106 TABLE 3B.1  Type of assistance to firms and households based on various assistance packages throughout the pandemic (Continued) FEB 20 MAR 20 APR 20 JUN 20 SEP 20 NOV 20 JAN 21 MAR 21 MAY 21 JUN 21 OCT 2021 Income tax Income tax Tax and duties Income tax Income tax Income tax Income tax Income tax instalment instalment concessions concessions and concessions concessions and concessions and concessions and deferment deferment and deductions and deductions deductions deductions deductions deductions Service tax exemption for Sales Tax hotels exemption for purchase Income tax of passenger relief for cars individuals purchasing Tourism tax tourism exemption related and service products tax exemption for hotels. Exemption Exemption Exemption and and and reduction rental reduction reduction of properties rental of rental of properties properties Electricity Electricity Electricity Electricity Electricity Electricity discounts discounts discounts discounts discounts discounts for tourism related industry Free Internet CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA from telecom providers Reduction in Foreign Worker’s Levy (continues) Listing fees for Waive fees Free business Waive late local bourse for late registration payment waived submission for select business renewal business groups joining fees for micro- reporting entrepreneurship SMEs late Listing fees for local bourse waived. Direct grants Direct grants Direct grants Direct grants Direct grants Direct grants to promote to promote to promote to promote to promote to promote E-commerce E-commerce E-commerce E-commerce E-commerce E-commerce adoption adoption adoption adoption adoption adoption Transform rural Internet centers into e-commerce hubs Online one- Small contracts stop business to repair public advisory amenities and platform infrastructure. Legal: Proceeds from Legal: Increase Easing COVID-19 SUKUK PRIHATIN the threshold of government worth RM100 indebtedness of procurement Temporary million used to companies by allowing Measures Act: develop vaccines, the variation of To mitigate treatment, and prices potential diagnostics of negative infectious disease. impact to economic activities caused by contractual breaches Source: Government of Malaysia CHAPTER 3: MALAYSIA 107 108 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Annex 3C.  Additional Results TABLE 3C.1  Correlations between firm characteristics and the likelihood of receiving government support (1) (2) (3) (4) Variable Tax deferral Tax deferral Wage subsidy Wage subsidy Medium sized firm (20–99 employees) 0.0905*** 0.0906*** (0.0255) (0.0287) Large sized firm (100+ employees) 0.158*** 0.139*** (0.0270) (0.0304) Percent change in sales −3.04e-05 −0.000470** (0.000166) (0.000185) Constant 0.0840 0.207*** 0.182*** 0.270*** (0.0610) (0.0645) (0.0686) (0.0718) Region, sector fixed effect Yes Yes Yes Yes Observations 1,500 1,348 1,500 1,348 R-squared 0.048 0.029 0.048 0.040 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1 CHAPTER 3: MALAYSIA 109 References Davis, Steven J. and John Haltiwanger. 1998. “Measuring Gross Worker and Job Flows,” NBER Chapters, in: Labor Statistics Measurement Issues, pages 77–122, National Bureau of Economic Research, Cambridge, MA. Department of Statistics Malaysia. 2020. Household Income and Basic Amenities Survey Report 2019. Putrajaya. https://www.dosm.gov.my/v1/index.php?r=column/cthemeByCat&cat=120&bul_id=TU00TmRhQ1N5TUx HVWN0T2VjbXJYZz09&menu_id=amVoWU54UTl0a21NWmdhMjFMMWcyZz09. Department of Statistics Malaysia. 2022. Principal Statistics of Labor Force, Malaysia, Fourth Quarter (Q4) 2021., Putrajaya. EPF (Employees’ Provident Fund) Malaysia. 2021. “EPF Focused on Rebuilding Members’ Retirement Savings Following Exceptional Withdrawal Facilities.” Kwsp.org. October 31, 2021. https://www.kwsp.gov.my/ms/-/epf-focused-on- rebuilding-members-retirement-savings-following-exceptional-withdrawal-facilities. Hale, Thomas, Noam Angrist, Rafael Goldszmidt, Beatriz Kira, Anna Petherick, Toby Phillips, Samuel Webster, Emily Cameron-Blake, Laura Hallas, Saptarshi Majumdar, and Helen Tatlow. 2021. A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker). Nature Human Behaviour. IMF (International Monetary Fund). 2021. Fiscal Monitor Database of Country Fiscal Measures in Response to the COVID-19 Pandemic, Washington DC. https://www.imf.org/en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response- to-COVID-19. Malay Mail. 2021. “On-Demand Delivery Trends in the New Norm.” June 4, 2021. https://www.malaymail.com/news/ life/2021/06/04/on-demand-delivery-trends-in-the-new-norm/1979509. Oxford COVID-19 Government Response Tracker, Blavatnik School of Government, University of Oxford. https:// www.bsg.ox.ac.uk/research/covid-19-government-response-tracker. Rahman, Abdur Amanina, Alyssa Farha Jasmin, and Achim Schmillen. 2020. The Vulnerability of Jobs to COVID-19: The Case of Malaysia. ISEAS Yusof Ishak Institute, Singapore. WEF (World Economic Forum). 2019. The Global Competitiveness Report, 2019. https://www3.weforum.org/docs/ WEF_TheGlobalCompetitivenessReport2019.pdf. World Bank. 2020. Malaysia Economic Monitor, June 2020: Surviving the Storm. World Bank, Washington, DC. https:// openknowledge.worldbank.org/handle/10986/33960. World Bank. 2021a. High-Frequency Phone Survey (Rounds 1 and 2 Forthcoming) on COVID-19 Impact & Recovery among Malaysian Households. World Bank, Washington, DC. http://documents.worldbank.org/curated/en/284621631899411887/High- Frequency-Phone-Survey-Round-1-May-18-June-16-2021. https://documents.worldbank.org/en/publication/ documents-reports/documentdetail/099705206232213729/p17543701d72f00470968401289d32fa325. World Bank. 2021b. Impacts of COVID-19 on Firms in Malaysia: Results from the COVID-19 Business Pulse Survey. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35872. World Bank. 2021c. Malaysia Economic Monitor December 2021: Staying Afloat. World Bank, Washington, DC. https:// www.worldbank.org/en/country/malaysia/publication/malaysia-economic-monitor-december-2021-staying-afloat. Zokkepli, Farik. 2022. “88,034 Foreigners Return Home in Repatriation Program.” Star Malaysia, August 2, 2022. https:// www.thestar.com.my/news/nation/2021/08/02/88034-foreigners-return-home-in-repatriation-programme. CHAPTER 4 Mongolia by Nona Karalashvili and Ikuko Uochi40 This chapter examines the actions taken by the government of Mongolia in response to the COVID-19 pandemic. It is based on the results of multiple surveys of firms and households, conducted by the World Bank. The World Bank collected survey data to determine how the pandemic affected households and businesses. The results of multiple rounds of both firm and household phone surveys reveal how the pandemic affected employment, firm sales and closures, and household incomes, as well as how government assistance helped buffer some of the negative impacts of the pandemic at both firms and households in Mongolia. The follow-up surveys on COVID-19 in the Enterprise Surveys, known as COV-ES, were conducted twice, in August 2020 and February 2021, with a sample that is representative of non-agricultural, non-extractive private firms with five or more employees that participated in the standard Enterprise Survey implemented in 2019. High-frequency phone surveys (HFPS) of households were conducted five times between May 2020 to June 2021, jointly with the National Statistics Office of Mongolia (see Annex 4A for further details). The two sets of surveys captured different periods of the pandemic’s progression in Mongolia. The first rounds of the firm survey in August 2020, and the household survey in May 2020, captured the early stage of the pandemic, when mobility was relatively high. The third round of the household survey in December 2020 took place during the strict nationwide lockdown, when mobility was low. The second round of the firm survey in February 2021, and the fourth round of the household survey in April 2021, were implemented during periods of increased COVID cases and stringent lockdown measures. They therefore reveal the pandemic’s impacts amidst heightened mobility restrictions. The fifth round of the household survey, June 2021, was implemented when the strict lockdown measure had eased, even though the Delta variant wave hit Mongolia hard. Timeline of the COVID-19 Pandemic and Government Measures From the beginning of the pandemic, the government of Mongolia took decisive containment measures, closing borders, closing all schools, and restricting travel and nonessential service activities. As a result, it managed to prevent the arrival of COVID-19 until late 2020 (Figure 4.1). In mid-November 2020, however, after the first locally transmitted cases of COVID-19 were verified, the number of cases surged. In response, the government imposed a strict nationwide lockdown in November–December 2020. After a decline in  This chapter greatly benefited from input provided by Lydia Kim. 40 111 112 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.1  Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Mongolia, 2020–22 HFPS R1 HFPS R2 HFPS R3 HFPS R4 HFPS R5 100 120 Inverted mobility index (14-day average) Stringency index (14-day average) 50 90 0 60 Stringency index –50 30 Inverted Mobility index –100 0 ES R1 ES R2 250 COVID-19 cases/deaths (per 100,000 people) 200 COVID-19 cases 150 COVID-19 100 deaths 50 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Source: Mobility data comes from Google Mobility Reports; stringency data comes from the Oxford Covid-19 Government Response Tracker (OxCGRT); COVID-19 cases and deaths were drawn from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. Note: The mobility index is the average percentage change in time spent at retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. For each day of the week, the baseline is defined as the median value from the five-week period of January 3–February 6, 2020. The figure shows 14-day rolling averages of the mobility index, which is inverted so that higher values indicate lower mobility. The stringency index records the strictness of lockdown style policies; 14-day rolling averages are shown. COVID-19 cases and deaths are smoothed by a seven-day average. Phases of the pandemic are determined by identifying structural breaks in COVID-19 cases and the mobility index. See the introduction for further details. cases following the lockdown, it gradually eased restrictions, but a new wave of COVID-19 infections in April 2021 led to another stringent national lockdown. Impacts on the Economy The COVID-19 crisis brought the worst contraction in the Mongolian economy since the 1990s. As a result of a sharp decline in demand for key commodities and border closures with China, real GDP contracted by 4.4 percent in 2020. In 2021, COVID–related restrictions on cross-border traffic to China disrupted recovery from the Delta wave outbreak, resulting in only 1.4 percent growth. In addition to the mining sector, the services sector—particularly tourism, hospitality, transportation, and trade—was hit hard by domestic containment measures. The labor force survey reflected employment disruptions from the pandemic, recording the lowest labor participation rate in the first half of 2021 over the previous 15 years. CHAPTER 4: MONGOLIA 113 Employment Impacts: Shocks and Recovery  Mongolia entered the pandemic still lagging significantly behind its peer countries in East Asia and Pacific (EAP) on business environment indicators (WEF 2019). The country’s institutional framework is marked by poor public sector performance and lack of responsiveness to change, weak property rights and judicial independence, and limited macroeconomic stability. Restricted and distorted domestic competition and sizeable non-tariff barriers to international trade limit creative disruption and firm incentives to upgrade. In addition, despite the relatively low costs of starting a business, the insolvency framework is not conducive to a swift reallocation of resources from shrinking and exiting firms to expanding ones. Furthermore, labor rigidities are significant, and financial depth ranks the lowest among the countries analyzed, which proves to be particularly binding for small and medium-size enterprises (SMEs). The impact of the COVID-19 pandemic on sales and employment was substantial. Relative to pre- pandemic conditions in December 2019, the average change in monthly sales fell below –40 percent, and the change in the number of permanent full-time workers reached about –20 percent. Despite a considerable shock to sales and employment, firms attempted to avoid laying off workers. A survey conducted in February 2021 revealed that the vast majority of firms resorted to alternatives, such as reducing hours worked (87 percent of firms) or salaries (86 percent). Consistent with firms’ efforts to avoid shedding workers, many measures of shock, such as a disruption to demand or supply, are not correlated with changes in workforce. Only the size of change in sales is correlated with the probability of firing workers (extensive margin), and only the incidence of the shock to firms’ liquidity is correlated with the size of employment adjustments (extensive margin). Pressure on the labor market appears to have eased somewhat by February 2021, but job losses remained significant. Among surviving firms, job contraction far surpassed job expansion in the early phases of the pandemic, with the estimated aggregate number of jobs 13 percent below the December 2019 level. In the post-initial lockdown in August 2020, the number of jobs had fallen 4 percent lower.41 The net change is therefore estimated at –9 percent. By February 2021, churning across jobs, with regard to firing or hiring, declined, with job contraction falling slightly more than job expansion to 9 and 2 percent, respectively, leading to a slightly improved net change of –7 percent. Among the most affected firms, size and sector stand out. As of February 2021, SMEs were more likely than large firms to be among the bottom 15 percent of firms in terms of changes in monthly sales change (Figure 4.2). Across sectors, non-food manufacturers, hotels, and transportation firms were more likely than other sectors to experience sharp drops in sales. The ability to use coping measures, other than shedding workers, varied across sectors and firm size. The non-food manufacturing and construction sectors were more likely than other sectors to cut their workforce. SMEs and large firms were equally likely to have made such deep cuts. Firm closure rates reached staggering levels in February 2021, when COVID cases increased and the government imposed strict mobility measures. About 2 percent of firms closed temporarily, and 7 percent 41  Job contraction and expansion are measured as follows: First, each firm’s share in total employment is calculated using sampling weights. Those weights are then used as each firm’s contribution to the change in the workforce through firing (job contraction) and hiring (job expansion). Aggregate measures of job contraction and expansion constructed through surveys may differ from aggregate estimates of job destruction and creation measured through census data, which may cover a larger set of sectors and better capture firm entry, as well as each firm’s share in total employment. 114 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.2  Characteristics of firms in Mongolia that experienced the largest declines in sales and employment a. Sales 70 60 55 58 50 40 Percent 28 28 29 28 30 19 21 20 20 15 18 12 10 7 1 2 4 0 SME (5–99) Large (100+) Food Other Construction Retail Hotels and Other services manufacturing manufacturing transportation August 2020 Februrary 2021 All firms, August 2020 All firms, Februrary 2021 b. Employment 35 30 30 25 21 20 18 Percent 15 16 15 14 15 13 12 10 10 10 8 7 7 4 5 1 0 SME (5–99) Large (100+) Food Other Construction Retail Hotels and Other services manufacturing manufacturing transportation August 2020 Februrary 2021 All firms, August 2020 All firms, Februrary 2021 Source: Follow-up surveys about COVID-19 via the COV-ES. Note: The sample is restricted to surviving firms at each survey round. Figure displays shares of firms that are in the bottom 15th percentile in terms of changes to their sales and employment. Horizontal lines show shares of such firms among all surviving firms. closed permanently during the early stages of the pandemic in August 2020. By February 2021, 26 percent of firms had closed permanently.42 Only 56 percent of firms in Mongolia were operational by February 2021. Foreign ownership and exporting status are most consistently related to the severity of the recession’s impact. Although the prevalence of such firms is low, foreign-owned or exporting firms were more likely than other firms to permanently close and make employment adjustments at the intensive or extensive margin (Figure 4.3). In particular, foreign-owned or exporting firms were more likely than others to have let some permanent workers go, experienced more negative changes to their workforce, and closed permanently. According to the pre-COVID Enterprise Survey, foreign-owned or exporting firms accounted for 23 percent  COV-ES involved re-contacting all firms that participated in the standard Enterprise Survey that was implemented in 2019. Firms that 42 confirmed being permanently closed were asked about the date of closing and other questions about their experience. CHAPTER 4: MONGOLIA 115 FIGURE 4.3  Differential impact on foreign-owned and exporting firms in Mongolia a. Probability of firing permanent workers b. Change in employment relative to baseline c. Probability of permanent exit since March 2020 100 0 60 80 −5 Probability (percent) Probability (percent) 40 60 −10 Percent 40 −15 20 20 −20 0 −25 0 Domestic, Foreign−owned Domestic, Foreign−owned Domestic, Foreign−owned non−exporter or exporter non−exporter or exporter non−exporter or exporter Source: Follow-up surveys about COVID-19 via the COV-ES. Note: The sample in the first two panels is restricted to surviving firms at each survey round. All panels control for survey rounds, Enterprise Survey stratification size, sector, and location. Differences in panels a, b, and c are statistically significant at the 5 percent, 10 percent, and 10 percent levels, respectively. of employment and had higher sales per worker than other firms, given factors such as firm size, sector, and location. The difficulties foreign-owned or exporting firms had may have long-term consequences for the non-extractive, non-agricultural part of Mongolia’s economy. Mirroring the firm closure dynamics, work disruptions at the household level became most severe when the government imposed a stringent lockdown in November to December 2020. Nearly one of five respondents who was working before the pandemic had stopped working by May 2020 (Figure 4.4). Work stoppages rose sharply after the December 2020 lockdown, declining as mobility restrictions were lifted by early April 202143 and June 2021. The share of work stoppages caused by the direct impacts of the pandemic and related containment measures, such as government-mandated business closures, quarantine, employment adjustment, and movement restrictions, declined from about 62 percent in December 2020 to 25 percent in June 2021. Even so, the pandemic continued to affect a substantial share of workers through other means, including early retirement and increased caregiving responsibilities, the primary reasons for work stoppages in 2021. Twenty-two percent of workers reported that they had retired, and 18 percent reported the need to provide care as reasons for no longer working.44 By the second round of the firm surveys, disruptions appeared to have been temporary for many firms and workers, as labor market pressure eased. According to the household surveys, among respondents who had stopped working by December 2020, three in five reported having a job to return to after the lockdown was eased. Nearly half of respondents who stopped working by December 2020 (round 3) returned to 43  The round 4 survey was conducted from April 19–30, 2021, but the reference period for employment status was the week of April 5, right before the April lockdown, which went into effect April 10, 2021. 44  The share of work stoppages tied to retirement increased significantly in 2021, rising from 2 percent in round 1 (May 2020) to 20 percent in round 4 (April 2021) and 22 percent in round 5 (June 2021). Self-employed individuals working in retail or as herders were more likely to stop working because of retirement, as were poorer individuals. Although no changes in policy motivated this trend, Mongolia’s current pension system may provide strong incentives for early retirement in economic downturns, especially for herders, informal workers, the self- employed, and retired mothers (Dorfman 2018). 116 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.4  Change in employment in Mongolia between May 2020 and June 2021 among respondents who worked before the pandemic Round 1 Round 3 Round 4 Round 5 (May 2020) (December 2020) (April 2021) (June 2021) Working (percent) 81 48 69 63 R3–R4 work stoppages: 4% R1–R3 work stoppages: R4–R5 work 37% stoppages: 13% (percent) 19 52 31 37 Working Not Source: World Bank data based on high-frequency phone surveys. Note: “Work stoppage” refers to people who worked before the pandemic or in a previous round but did not work (permanently or temporarily) the week before the survey round. The sample is restricted to respondents of the employment module, who were the same respondents across all rounds and working pre-pandemic (N = 656). work in April 2021 (round 4). In contrast, in April 2021, two-thirds of respondents who were out of work indicated that they did not have a job to return to, suggesting that work stoppages were largely comprised of workers who continued to be out of work across rounds. In particular, women were more likely than men to experience long-term or permanent work stoppages throughout the pandemic (Figure 4.5). Household surveys reveal relatively high levels of work stoppages in the hospitality, transport, construction, and manufacturing sectors. Because of restrictions on international and domestic travel, the FIGURE 4.5  Employment status of men and women in Mongolia between December 2020 and June 2021 100 Percent of respondents working 80 48 49 58 68 72 73 pre-pandemic 60 40 25 34 11 7 20 14 8 27 25 31 17 14 19 0 Women Men Women Men Women Men Round 3 Round 4 Round 5 (December 2020) (April 2021) (June 2021) Not working, no job to return to Not working, had job to return to Working Source: World Bank data based on high-frequency phone surveys CHAPTER 4: MONGOLIA 117 hospitality and transportation sectors faced severe contractions throughout 2020 and into 2021 and a high level of work stoppages (Figure 4.6, panel a). Closures of construction sites and factories, as well as reduced global demand for key export products, translated into widespread work stoppages for workers in the industry sector in 2020 and early 2021. Agriculture, public administration, and defense were less affected. As urban workers were almost 2.5 times more likely than rural workers to be engaged in the hospitality, transportation, or industry sectors, the sectoral impact of the pandemic contributed to larger shares of work stoppages in urban areas than in rural areas (Figure 4.6, panel b). Distributional impacts and potential implications for inequality Changes in the size of the workforce were initially positively correlated with firms’ pre-COVID labor productivity, but this association faded completely by February 2021. Firms with lower pre-COVID sales per worker were initially more likely to report losses. On average, these firms experienced significant negative changes in employment by August 2020 (Figure 4.7). As the pandemic continued, firms with higher levels of baseline labor productivity also started to experience the negative impacts of the recession, and lower-productivity firms began to recover. It is possible that the size of pre-crisis sales per worker served as a cushion against shocks, delaying—as well as slightly limiting—the impact of large disruptions. Firms that did not have such a cushion were affected earlier but were also forced to make painful adjustments earlier, which, by February 2021, allowed them to achieve similar workforce changes as firms that did have such a cushion. This pattern is indirectly suggested in Figure 4.7, where the red dots, indicating February 2021, are FIGURE 4.6 Work stoppages among respondents working pre-pandemic, May 2020–June 2021 a. By sector b. By urban or rural location 70 Construction 60 60 Percent of respondents working pre-pandemic who Transport, hospitality Manufacturing 50 41 stopped working Retail 40 36 Other industry 30 35 22 28 Prof., education, health 20 23 Agriculture 10 14 Public administration 0 Round 1 Round 3 Round 4 Round 5 0 20 40 60 80 100 (May 2020) (December 2020) (April 2021) (June 2021) Percent of respondents working pre-pandemic Urban Rural Round 1 (May 2020) Round 3 (December 2020) Round 4 (April 2021) Round 5 (June 2021) Source: World Bank data based on high-frequency phone surveys. 118 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.7  Correlation between change in employment and baseline labor productivity in Mongolia 40 Change in employment relative to 20 December 2019 (percent) 0 −20 −40 −60 7 8 9 10 11 12 ln(baseline labor productivity) August 2020 February 2021 Source: Follow-up surveys about COVID-19 via COV-ES. Note: The sample is restricted to surviving firms at the time of each survey round. The figure shows a binned scatterplot with sampling weights. Baseline labor productivity indicates sales per worker from the 2019 Enterprise Survey. generally above the blue dots, indicating August 2020, for firms with lower baseline productivity. Trends are more mixed for higher levels of baseline labor productivity. This pattern of lower-productivity firms potentially being affected first, with the effect becoming more uniform over time as firms affected earlier moved to the phase of recovery, is visible across several indicators, although it is not statistically significant.45 In particular, firms with higher pre-COVID sales per worker generally had slightly better profiles in August 2020 and worse profiles in February 2021 than other firms. More productive firms, as measured by sales per worker pre-COVID, were less likely to close permanently. This negative association between productivity and closures is especially strong and statistically significant for February 2021. The pattern is the same for foreign-owned and exporting firms, which were both more productive pre-COVID and more likely to have permanently closed than other firms. Poorer workers were more likely to stop working during times of higher mobility, while both poor and wealthier workers were more uniformly affected during times of stringent mobility restrictions. In May 2020 (round 1), when mobility returned to pre-pandemic levels after a brief period of lockdown, workers in the fourth and fifth household welfare quintiles were 43 and 63 percent less likely, respectively, than workers in the poorest quintile to stop working (Figure 4.8). These differences persisted even after controlling for differences in education and sectoral participation across the welfare distribution. Lower job formality, less job security, and greater difficulty working from home among poorer workers may have played a part in driving these unequal outcomes in work stoppages.46 Initially, work stoppages were more prevalent among poorer workers; over time, they became more uniform across the welfare distribution, as lockdown measures 45  The pattern does not appear to be fully driven by a higher rate of exit by lower-productivity firms as of February 2021. The pattern in Figure 4.7 persists after limiting the sample to firms that operated during both rounds. 46  The household phone survey does not include data on job formality or security and thus does not allow empirical testing of this statement. CHAPTER 4: MONGOLIA 119 FIGURE 4.8  Work stoppages in Mongolia, May 2020–June 2021, by welfare quintile 80 52 53 54 48 56 working pre-pandemic Percent of respondents 60 54 34 34 35 31 30 31 30 25 23 38 25 40 27 15 10 20 0 Round 1 (May 2020) Round 3 (December 2020) Round 4 (April 2021) Round 5 (June 2021) Q1 (poorest) Q2 Q3 Q4 Q5 (wealthiest) Source: World Bank data based on high-frequency phone surveys. Note: Error bars show 95 percent confidence intervals of difference with respect to the poorest quintile rather than difference from zero. became more stringent and employment impacts more far-reaching. Work stoppages increased significantly in December 2020, as Mongolia experienced a stringent nationwide lockdown. At that time, differences across quintiles were statistically insignificant. Impacts continued to be more or less uniform in round 4, right before the April 2021 lockdown was put in place. Once the April 2021 lockdown measure was eased, work stoppages rose significantly only for the poorest workers in June 2021. Recovery was faster for workers from better-off households. Workers in the top 80 percent of households in the welfare distribution who had stopped working in December 2020 or April 2021 were about 55 percent more likely than workers in the bottom 20 to have a job to return to. Workers in the bottom 20 percent who had stopped working by December 2020 were about 30 percent less likely than the top 80 percent to be working in April 2021. Unequal effects in work stoppages were also apparent along gender lines, with women less likely than men to return to work as a consequence of the pandemic. Summary of impacts on firms and households Firm and household surveys suggest that the pandemic had substantial negative impacts on jobs. Average year-on-year changes in monthly sales exceeded 40 percent, while more than half of formal sector firms fired at least some of their permanent workers in the early phase of the pandemic, and nearly 90 percent of firms reduced total hours worked and salaries as COVID cases rose and stringent mobility restrictions were enforced. The net change in jobs among surviving firms was well below zero for both firm survey rounds, August 2020 and February 2021, and permanent firm closures reached a staggering level of 26 percent by February 2021, after a prolonged period of lockdown and containment measures. Mirroring the dynamics of permanent and temporary closures, work disruptions at the household level were most severe during the strict lockdown in December 2020, when 52 percent of those who worked before the pandemic stopped working, with women affected more than men. Similar cross-sector patterns emerged from the firm and household surveys. The hospitality, transportation, construction, and manufacturing (especially non-food manufacturing) sectors appear to have been most affected. The firm surveys also reveal a heightened impact on foreign-owned or exporting firms. Lower-productivity firms and poorer households experienced negative impacts first, with the impact becoming more uniform later on. Permanent closures of firms appear to be associated with lower pre-COVID 120 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA levels of sales per worker. Among surviving firms, patterns in the changes in workforce suggest that firms with lower pre-COVID labor productivity may have been affected first, with the effect catching up to some of the rest of the firms over time. The data only weakly reveals this pattern, which should therefore be interpreted cautiously. Household surveys more clearly suggest that the distributional impacts of the pandemic on employment followed lockdown measures. Work stoppages were more severe for poorer workers during times of higher mobility restrictions in the earlier months of the pandemic, such as May 2020; they became more uniform across the welfare distribution by December 2020, as lockdown measures became more stringent and employment impacts more far-reaching. Firm and household surveys indicate some signs of early, albeit tentative and unequal, recovery from the initial shock. In particular, by February 2021, surviving firms showed early glimpses of recovery, with the average change in permanent workforce relative to December 2019 inching above zero, though net change in jobs, an aggregate measure, remained in the deep negative territory. This early recovery by firms was mirrored by more gradual, though uncertain, recovery by households, which took hold at later stages of the pandemic. Household data from June 2021 suggests a lower incidence of work stoppages after the April 2021 lockdown measure was eased, although this improvement was not experienced equally by all households. In June 2021, poorer workers suffered much more from work stoppages than wealthier workers and were less likely to have a job to return to. Fiscal Support to Firms and Households Size and composition of the fiscal response To mitigate the adverse impacts of the pandemic on firms and households, the government provided generous economic support. The economic relief package included top-ups of existing social assistance programs, exemptions on social insurance contributions and utility payments, various tax relief measures, deferrals of payments, and other forms of support to firms and households. Mongolia’s economic relief package is one of the largest relative to GDP in the East Asia and Pacific region, with a total cost of over 15 percent of GDP. Actions included the following (Figure 4.9): ⦁ The monthly benefit of the Child Money Program (CMP) was increased from MNT 20,000 (US$7) to MNT100,000 (US$35) per child under 18, starting April 1, 2020. ⦁ A one-time stimulus cash transfer of MNT 300,000 (US$105) was provided to every citizen to compensate for household income loss from the April 2021 lockdown. ⦁ A generous package of tax relief measures was implemented, most of which went into effect between April 1, 2020 and July 1, 2021. Measures included exemptions from social security contributions (SSCs)—full and partial, for eligible employers and employees participating in social insurance schemes, mandatory or otherwise; exemptions from personal income tax for eligible individuals; waivers of penalties for late payments to SSCs and personal income tax; exemptions from corporate income tax for small and medium-size enterprises (SMEs); tax credits for rental income of firms that reduced their rents; exemptions from customs and the value added tax (VAT) for food and medical suppliers; and continuation of tax relief for firms in remote locations. ⦁ The utility bills of households and private firms in certain sectors—covering electricity, heating, water usage, and waste disposal services—were subsidized, within a set limit. CHAPTER 4: MONGOLIA 121 FIGURE 4.9  Components of Mongolia’s COVID-19 responses, 2000 through mid-2021 Top-ups of Child Money Program 4.8 (April 2000–July 2021) One-time cash transfer 3.0 (April 2021) Reduction of social security contributions 2.4 (April 2020–July 2021) Other tax relief 2.1 Exemptions on utility bills 1.6 Other support measures 1.7 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Percent of GDP (2020 in current MNT) Source: World Bank staff estimates and data from the Ministry of Finance of Mongolia. ⦁ Access to credit for relatively small firms was eased through reduced interest rates; improved loan conditions, including a subsidy for collateral; deferral of loan payments; increased credit guarantees; increases in financial leasing services; a reduced required reserve ratio; and new lending through major development projects. ⦁ Government financing was amplified through increased purchases, targeted mostly to relatively small firms; early repayment; and other servicing of government loans. ⦁ A wage subsidy of MNT 200,000 per month, or 50 percent of the monthly minimum wage, was provided for three months (April 1–July 1, 2020) to private firms affected by the pandemic. Employers were banned from laying off workers or reducing their wages. ⦁ Exporting firms were supported through special financing solutions, an Export Guarantee Fund, and reforms aimed at increasing the capacity of customs. The tourism and construction sectors were supported through special development projects. Cashmere producers were supported through subsidies. ⦁ A one-time cash incentive was extended to reward those who received two doses of the COVID-19 vaccine, starting May 2021, which comprised about 0.2 percent of GDP. Coverage and timing Government support reached a significant share of firms. About a third of businesses, or 34 percent, reported receiving at least some form of support in the early stages of the pandemic. Another 12 percent expected to receive at least some form of government support over the next three months. Cumulative coverage increased to 40 percent of firms by February 2021 (Figure 4.10), at which point an additional 24 percent expected support in the next three months. Among firms that had permanently closed by February 2021, 21 percent received government support. Twenty percent of firms reported having received wage subsidies by August 2020, and 15 percent reported receiving some form of fiscal relief. Firms that reported receiving some government support were 122 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.10  Coverage and types of government support in Mongolia among surviving firms 50 40 40 34 30 28 23 Percent 20 22 20 15 13 15 10 3 3 1 0 August 2020 February 2021 Any assistance Cash transfers Deferral of payments Access to new credit Fiscal relief Wage subsidies Source: Follow-up surveys about COVID-19 via COV-ES. asked to specify whether they received each of the five general types of support shown in Figure 4.10, which were not explicitly mapped to the set of policy measures implemented at the time. Few firms reported cash transfers, deferral of payments, or access to new credit. Comparison of the coverage across the two rounds suggests that by August 2020, firms had received an average of about 1.2 types of support measures; however, some firms may have reported the same support measure as representing multiple types of support, so this measure may be overestimated. As of February 2021, cumulative coverage of nearly all types of support measures increased considerably, with an 18 percentage point increase in deferral of payments, which became the third-most prevalent type of government support—22 percent of firms reported having received this type of support—after fiscal relief, at 28 percent, and wage subsidies, at 23 percent. By February 2021, the average firm receiving support had received about 2.5 types of assistance. For households, the Child Money Program (CMP) and the April 2021 one-time transfer were the two most generous programs. The top-up of the CMP was among the first initiatives put into effect and one of the most extensive, because of its broad coverage. The program covers 80 percent of households in the bottom 40 percent (Figure 4.12), making it a convenient tool for rapidly delivering relief measures to the most vulnerable. In April 2020, the CMP benefit increased from 20,000 to 100,000 Mongolian tugrik (MNT) per child (Figure 4.11). On average, these payments were equivalent to nearly 60 percent of the monthly pre-COVID household income of the bottom 20 percent of households. Although the CMP transfers were provided more frequently, they were smaller than the April 2021 universal one-time cash transfer. On average, the monthly CMP benefit amounted to about 32 percent of monthly pre–COVID household income across rounds. The one-time transfer of 300,000 MNT per household member, which was awarded to every citizen in Mongolia, was nearly three times the amount of monthly pre-COVID income for the bottom 20 percent. For more than half of households, that transfer was greater than or equal to their monthly income. Targeting In addition to high coverage of bottom 40 households in transfer programs, coverage of the wealthiest households was also high. The one-time transfer was untargeted and therefore included many households that may not have needed it. Although the targeting of the CMP is pro-poor—as poorer households are more likely to have more children—nearly half of households in the top 20 percent received CMP transfers CHAPTER 4: MONGOLIA 123 FIGURE 4.11  Size of Mongolia’s child money FIGURE 4.12  Percent of households in Mongolia program and one-time cash payment, that received government assistance, by wealth quintile (MNT) by welfare quintile 400 100 350 Percent of monthly household income 80 300 Percent of households 250 60 200 150 40 100 20 50 0 0 Child Money Program One-time cash transfer CMP (Round 1–Round 4) One-time transfer (Round 4) Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Source: World Bank data based on high-frequency phone surveys. (Figure  4.12). As only households with children are eligible to receive the CMP, wealthier households may have received the CMP at the cost of poorer households that did not have children. In the HFPS, about 70 percent of households that received the CMP at any point were considered non-poor in the 2018 household survey. Firms that experienced the worst shocks do not appear to have been more likely to receive support than other firms. A set of simple regressions with standard controls (survey round, firm size, sector, and location) suggests a weak relationship between the shocks firms experienced and the likelihood of receiving government support, with most coefficients of interest estimated to be zero. Among foreign-owned or exporting firms—which were more productive pre-COVID but were also more negatively affected by the recession—there does appear to be a correlation between the hardship experienced and the likelihood of receiving government support, with firms that experienced declines in supply, demand, or sales more likely to have received government support than other firms. For firms, unlike households, supporting the most vulnerable may not be desirable, as better long-term results may be achieved by concentrating resources on the most productive firms. Coverage of government support helped firms with relatively low pre-crisis labor productivity. Firms that had lower baseline sales per worker before COVID were more likely to receive support, especially in the early phase of the pandemic (Figure 4.13). The pattern is similar for foreign-owned and exporting firms. This correlation is only vaguely suggested by the data, however, with no statistically significant support, and should thus be interpreted cautiously. The HFPS shows that poorer beneficiaries were more likely to cash out, using the CMP transfer, especially for food and household utilities, than wealthier households, which were significantly more likely to save the money (Figure 4.14, panel a). This result suggests that government assistance might have been more effective had it targeted the people who most needed the money. 124 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.13  Correlation between baseline labor productivity in Mongolia and probability of receiving government support in August 2020 and February 2021 1.0 Probability of receiving any government support 0.8 0.6 0.4 0.2 0 7 8 9 10 11 12 ln(baseline labor productivity) August 2020 February 2021 Source: Follow-up surveys about COVID-19 via COV-ES. Note: The sample is restricted to surviving firms at each survey round. The figure shows a binned scatterplot with sampling weights. Baseline labor productivity indicates sales per worker from the 2019 Enterprise Survey data. FIGURE 4.14  Use of Mongolia’s child money program and the one-time cash payment by recipients a. Child Money Program b. One-time transfer 100 5 5 5 4 5 100 5 5 4 6 8 8 12 90 90 18 percent of beneficiary households percent of beneficiary households 80 28 80 19 22 12 34 26 26 15 36 70 46 70 7 53 60 5 60 50 5 4 50 4 40 40 74 6 71 72 73 30 64 30 66 66 57 51 49 43 20 20 33 10 10 0 0 National Q1 Q2 Q3 Q4 Q5 National Q1 Q2 Q3 Q4 Q5 Didn't cash for other reasons Saved Didn't cash for other reasons Saved Used for: loan repayment, other Used for: education/healthcare Used for: loan repayment, other Used for: education/healthcare Used for: food, utilities Used for: food, utilities Source: World Bank data based on high-frequency phone surveys. CHAPTER 4: MONGOLIA 125 FIGURE 4.15  Effectiveness of government support to firms in Mongolia Probability of firing permanent workers Overall support, no controls Overall support, with controls Deferral of payments, with controls Change in employment relative to December 2019 Overall support, no controls Overall support, with controls Deferral of payments, with controls −1.0 −0.5 0 0.5 Probability, percentage change Source: Follow-up surveys on COVID-19 via COV-ES. Note: Each bar represents a separate panel regression with firm fixed effects. The first two bars use logit specification and show marginal effects estimated at the mean. Error bars show 95 percent confidence intervals. The regressions are weighted using the pre–COVID Enterprise Survey sampling weights. The sample is restricted to surviving firms at each survey round. For bars that include controls, the variables are average change in monthly sales and whether the firm experienced a decline in supply, demand, or liquidity; was overdue on obligations to financial institutions; temporarily closed; adjusted or converted production or services or increased online business activity; increased contactless delivery of goods or remote work; exported at least 10 percent of its sales during the pandemic; or received any other type of government support (apart from the deferral of payments). The coefficients do not represent causal links because of multiple complications of statistical identification, such as self-selection of firms that applied government support and endogeneity. The April 2021 one-time transfer was more likely to be spent across welfare distribution. On average, 9 out of 10 households in round 4 (April 2021) spent at least some of it, with the large majority spending it on food and utilities (Figure 4.14, panel b). It is possible that the windfall amount of the April 2021 transfer allowed households to spend more promptly and liberally on necessities such as food and health care, or to otherwise make up for negative shocks caused by the pandemic. Effectiveness Deferral of payments may have been effective in helping firms to retain workers. The regression coefficients displayed in Figure 4.15 suggest that firms that received government support were less likely than others to fire permanent workers, although causality or its direction cannot be established. Firms that received government support, especially deferral payments, experienced higher than average changes in the size of their workforce than did firms that did not receive support. However, the estimated effect does not consider the types of shocks that firms experienced or the nature of the actions taken by firms to adjust to these shocks. For households, government assistance—particularly the April 2021 one-time cash transfer—may have been most effective in preventing adverse impacts on household well-being.47 Panel regression results 47  Given relatively low participation in other government relief programs, such as food stamps, social welfare pensions, and other in-kind transfers, the analysis focuses on the one-time cash transfer implemented before the April lockdown. 126 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.16  Panel regression results: Receipt of one-time government cash transfer in Mongolia in April 2021 and likelihood of being food insecure Pooled Q1 Q5 –9.7 –16.6 –5.3 Received one-time transfer 8.7 9.9 7.7 Food price shock 10.8 18.6 9.9 Labor income reduction –40 –20 0 20 40 –40 –20 0 20 40 –40 –20 0 20 40 Severe food insecurity Source: World Bank high-frequency phone surveys, rounds 1, 3, and 4. Note: Figure shows the percentage point change in the likelihood of being food insecure associated with the specified variable, controlling for other included variables. For example, in column 1, receiving the one-time cash transfer is associated with a 9.7 percentage point decrease in the likelihood of being severely food insecure, controlling for experience of shocks and receipt of the CMP. The figure shows 95 percent confidence intervals. show that controlling for income and food price shocks, the one-time cash transfer was associated with a 10 percentage point decrease in the likelihood of experiencing severe food insecurity (Figure 4.16).48 The magnitude of this likelihood is greater for the poorest quintile than for the wealthiest, implying that the one- time cash transfer may have been more consequential for poorer households in averting severe food insecurity. However, this negative relationship between the one-time cash transfer and food insecurity is significant for nearly all quintiles, suggesting that the government’s decision to provide a lump-sum transfer of an unprecedented size before the April 2021 lockdown benefited households across the welfare distribution. Despite the generosity of both the CMP and the one-time cash transfer, these payments may not have fully offset the pandemic’s impacts for most households. In April 2021, more than 70 percent of beneficiaries of the CMP or one-time cash transfer reported that the transfer was not enough to fully mitigate the effects of the pandemic (Figure 4.17, panel a). However, households that received non-CMP cash transfers49 were 10 percentage points, or 56 percent, more likely to report that the assistance completely mitigated the 48  These results would be expected if trends in food insecurity were secular. However, trends in food insecurity have varied, closely following economic conditions. Moreover, households that reported recent income losses were more likely to be food insecure, controlling for the welfare quintile. Taken together, these findings suggest that trends in food insecurity are sensitive to short-term fluctuations and are not secular. 49  The one-time transfer accounted for more than 95 percent of reported non–CMP cash transfer amounts. CHAPTER 4: MONGOLIA 127 FIGURE 4.17  Perceived effectiveness of the child money program and the one-time cash transfer in mitigating the effects of the COVID-19 shock in Mongolia a. Child Money Program b. One-time transfer 100 2 10 3 100 11 21 11 5 8 80 80 Percent of beneficiaries Percent of beneficiaries 71 66 73 60 77 60 75 73 40 40 20 20 28 32 22 25 18 15 0 0 National Bottom 40 Top 60 National Bottom 40 Top 60 Completely mitigated Partially mitigated Completely mitigated Partially mitigated Did not mitigate at all Not a ected by pandemic Did not mitigate at all Not a ected by pandemic Source: World Bank data based on round 4 of the high-frequency phone survey. negative impacts of the pandemic than the CMP (Figure 4.17, panel b).50 Households in the bottom 40 were about 13 percent more likely to report that the CMP fully or partially mitigated pandemic-induced shocks, and about 28 percent more likely to report that the one-time cash transfer fully mitigated the impacts. Greater awareness of the availability of government support and simplified procedures may have helped more firms in need during the pandemic. About 39 percent of firms that did not receive government assistance were unaware of its availability, 31 percent indicated that they did not need the support, 16 percent perceived the procedures for applying for the aid cumbersome and costly, and 10 percent reported not expecting to get the support because they lacked the “right connections.” Among surviving foreign-owned or exporting firms that did not receive government support, 59 percent reported not needing it. Across sectors, fiscal relief was the most frequently cited type of support that firms reported needing. This finding was consistent across SMEs, large firms, firms that were above the median sales per worker before the pandemic, and domestic and non-exporting firms (Figure 4.18). In contrast, among foreign- owned or exporting firms, as well as among majority-women-owned firms, the most-needed support cited most frequently was deferral of payments. Wage subsidies, digital support, and other types of support were rarely cited among the most needed. COVID-19 assistance to households might not have effectively and efficiently reached the most vulnerable households. The CMP has been a flagship program since its launch in 2005; it contributed substantially to poverty reduction between 2010 and 2018 (National Statistics Office of Mongolia and World Bank 2020). Coverage rates are nearly universal, and the top 60 percent of households receive about 50  These households are not mutually exclusive: About 64 percent of households that received the one-time cash transfer in round 4 also received the CMP. 128 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 4.18  Types of government support firms in Mongolia cited as most needed, February 2021 100 3 2 3 3 3 1 2 3 2 2 3 3 15 18 80 8 32 31 23 42 33 15 60 Percent of firms 13 12 13 48 12 12 18 40 52 24 25 22 1 25 9 20 34 24 24 24 19 15 13 0 All firms SME (5–99) Large (100+) Above median Below median Domestic, non-exporter Foreign-owned or exporter Other support Digital support Wage subsidies Fiscal relief Access to new credit Deferral of payments Cash transfers Source: Follow-up surveys on COVID-19 via COV-ES. Note: Median refers to sales per worker from the 2019 Enterprise Survey. 40  percent of total CMP benefits (author’s calculations based on the 2018 Household Socioeconomic Survey). The COVID-19 relief top-up of the CMP is estimated to have cost more than 4 percent of GDP. Careful analysis of the use of the existing social program as a shock response and the estimated impact of the program will be important to assess its cost-effectiveness in response to the pandemic. According to phone surveys, although households largely agreed with the government’s containment and mobility restriction measures, more than half believed that further support was needed for affected workers and businesses, and half of the bottom 40 percent of households reported that they needed additional financial assistance to cope with the COVID-19 situation (Figure 4.19). Lessons Learned The government of Mongolia provided generous COVID-19 support packages to firms and households with a considerable share of the respective populations covered. Mongolia’s economic relief package was one of the largest relative to GDP in the East Asia and Pacific region, totaling over 15 percent of GDP (2020 –1H 2021). At the early phase of the pandemic in August 2020, over a third of firms received some support. By the time COVID cases increased and strict lockdown measures were imposed, the coverage reached 40 percent of firms, with another 24 percent of firms expecting some form of government support in the next three months. For households, the government made a swift decision to utilize the existing social assistance programs and increased the CMP benefit size by a factor of 5 in April 2020. Coverage reached 80 percent of the bottom 40 of the household income distribution. Furthermore, the one-time cash transfer was provided to every citizen ahead of the April 2021 lockdown. CHAPTER 4: MONGOLIA 129 FIGURE 4.19  Household opinions on government responses to the COVID-19 pandemic in Mongolia a. Share of households that believe the government is doing enough to help people who lost their jobs and businesses that had to close b. Most important types of government support 100 100 8 11 9 12 90 16 18 80 10 80 11 11 35 31 33 Percent of households 70 20 Percent of households 60 60 24 22 50 40 40 29 32 38 51 30 20 42 44 20 25 25 0 10 17 Bottom 40 percent Middle 40 percent Top 20 percent 0 Don't know Others Bottom 40 percent Middle 40 percent Top 20 percent Protective equipment (face mask, etc.) Strongly disagree Disagree Neutral Expand health services for basic needs Don't know Agree Expand the government's financial assistance Source: World Bank data based on round 5 of the high-frequency phone survey. Targeting of coverage appears to have been challenging, as firms that were potentially less susceptible to shocks, as well as non-poor households, appear to have received sizeable support. Firm survey data shows a poor association between firms’ experience of shocks and their likelihood of receiving support. However, such associations were found among the foreign-owned or exporting firms. Furthermore, coverage of government support was more likely for firms with lower pre-COVID sales per worker. This pattern highlights the concept that directing government support toward firms that were potentially more susceptible to shocks or at a higher risk of worse impact may not necessarily be desirable, especially beyond the immediate impact of a crisis. For households, where targeting the most vulnerable is indeed desirable, the targeting of a majority of the government support was instead universal; therefore, a sizeable amount of the government support was observed even among the wealthiest. Indeed, nearly half of the top 20 percent of the household income distribution also received CMP transfers, raising a question of utilizing existing universal transfer programs in response to shocks. However, both firm and household data suggest some evidence of the of the government support in preventing larger employment adjustments by firms and averting food insecurities for poor households. While causality is impossible to ascertain, panel analysis of firm data suggests that one type of government support, namely deferral of payments, is associated with better average change of permanent workforce relative to the baseline of December 2019. Similarly, panel analysis of household data points to effectiveness of the one-time April 2021 cash transfers in helping households avert food insecurity during the pandemic. There is room for the government to improve provision of government support measures through better targeting, increased awareness, simplification of procedures, and other measures. The relative lack of targeting suggests the need for improving the focus of support on the most vulnerable households and 130 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA preferably the most viable firms. Firm surveys further suggest the need for an increased awareness, simpler application procedures, and perhaps an increased focus on the type of support that is the most frequently named as most needed by firms – fiscal relief – and the type that seems to have been effective – deferral of payments. The household survey suggests overall agreement of respondents with government COVID-19 responses in general, with a sizeable proportion expecting government support to be better targeted toward the most affected households and workers. This implies that the marginal tughrik may have been better spent on the most vulnerable and affected households through providing more targeted assistance rather than utilizing the existing universal social transfer programs during the pandemic. Mongolia’s macroeconomy and labor market show signs of recovery, but recovery appears to be uneven. After Mongolia was hit hard by the Delta wave in mid-2021, the economy started to rebound, and the labor force participation rate steadily improved, rising from its lowest level in a decade (Figure 4.20). Yet, improvement varied across sectors. According to the latest business registry database, two out of five business entities in the industry and non-public-service sectors remained inactive. The public administrative, health, and education sectors were less affected (Figure 4.21). Recovery may be slowed by continued border restrictions with China and recent hikes in the prices of food, commodities, and energy, which create new risks, particularly for poor households and vulnerable firms. In addition, two years of expansionary fiscal support have increased the debt burden, limiting fiscal space and raising questions about the sustainability of continued government support to firms and households beyond the COVID-19 crisis. Given the limited fiscal space and evolving challenges that are currently facing the economy, directing resources to households that need support most and focusing on reforms that increase competitiveness FIGURE 4.20  Labor force participation and the FIGURE 4.21  Activity status of registered employment-to-population ratio in Mongolia, business entities in Mongolia, 2019–21 2019–21 100 Percent of business entities registered 62 28 24 26 28 30 32 32 80 40 40 Percent of working-age population 60 60 58 40 76 74 56 72 68 72 70 68 60 60 20 54 52 0 Q4 Q4 Q4 Q4 Q4 Q4 Q4 Q4 Q4 2019 2020 2021 2019 2020 2021 2019 2020 2021 50 Industry Public admin, Other service Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 professional 2019 2020 2020 2020 2020 2021 2021 2021 2021 Active Permanently/temporary stopped Employment-to-population ratio Labor force participation rate Source: National Statistics Office of Mongolia. Note: Figures exclude business entities that have not yet begun operations. CHAPTER 4: MONGOLIA 131 of the business environment may be warranted. The support packages that the government provided to firms have included important reforms to the business environment. Economic reforms that facilitate cross- border trade and increase firms’ flexibility to adjust to changing circumstances will likely provide long-term benefits beyond the pandemic. At the household level, removing income support too quickly could slow poverty reduction. At the same time, given rising fiscal pressures, it is reasonable to shift support from universal programs to programs that target the most vulnerable households. Increasing the resilience of vulnerable households to income shocks and continuing to improve the competitiveness of the business environment will enable a smoother recovery from the pandemic and help set Mongolia’s economy on a long-term growth path. 132 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Annex 4 Annex 4A.  The Household and Firm Surveys in Mongolia The World Bank and the National Statistics Office of Mongolia (NSO) jointly implemented five rounds of high-frequency phone surveys (Table 4A.1). These surveys, which drew a subsample of 2,000 households from the nationally representative 2018 Household Socioeconomic Survey, aimed to collect information from the same households across multiple rounds. Sampling weights were constructed to ensure unbiased estimates from the sample. The sample distribution of the phone survey is like that of the 2018 Household Socioeconomic Survey on key household characteristics, such as location, education level of household head, and poverty status. The World Bank conducted a standard Enterprise Survey in Mongolia in 2019. It included a representative sample of 360 formal, non-agricultural, and non-extractive firms with five or more employees. These firms were re-contacted by phone after the onset of the pandemic, first in August 2020 and then again in February 2021 (Table 4A.2). These surveys included questions about the effects of COVID-19. Two sets of sampling weights were included in the surveys, from the standard Enterprise Survey and from the COVID-follow up (COV-ES), with the latter considering firm closures during the pandemic. TABLE 4A.1  Dates and sample sizes of high-frequency phone surveys of households in Mongolia Round Survey dates Sample size 1 May 22–29, 2020 1,333 2 August 31–September 7, 2020 1,212 3 December 3–15, 2020 1,147 4 April 19–30, 2021 1,085 5 June 14–23, 2021 1,046 Source: Original table for this publication TABLE 4A.2  Dates and sample sizes of the the COVID-follow up enterprise surveys in Mongolia Round Survey dates Sample size 1 August 3–15, 2020 314 2 February 1–18, 2021 323 Source: Original table for this publication Note: The Enterprise Survey baseline sample of 360includes firms that were formally registered with a government authority at the time of the survey. CHAPTER 4: MONGOLIA 133 References Dorfman, Mark Charles. 2018. Mongolia: Policy Options for Pension Reform. World Bank, Washington, DC. http:// documents.worldbank.org/curated/en/873841592807081133/Mongolia-Policy-Options-for-Pension-Reform. National Statistics Office of Mongolia and World Bank. 2020a. Mongolia Poverty Update 2018. Washington, DC: World Bank. https://documents1.worldbank.org/curated/en/532121589213323583/pdf/Mongolia-Poverty-Update-2018.pdf. National Statistics Office of Mongolia and World Bank. 2020b. Results of Mongolia COVID-19 Household Response Phone Survey (Round 1). World Bank, Ulaanbaatar. National Statistics Office of Mongolia and World Bank. 2020c. Results of Mongolia COVID-19 Household Response Phone Survey (Round 2). World Bank, Ulaanbaatar. National Statistics Office of Mongolia and World Bank. 2021a. Results of Mongolia COVID-19 Household Response Phone Survey (Round 3). World Bank, Ulaanbaatar. National Statistics Office of Mongolia and World Bank. 2021b. Results of Mongolia COVID-19 Household Response Phone Survey (Round 4). World Bank, Ulaanbaatar. National Statistics Office of Mongolia and World Bank. 2021c. Results of Mongolia COVID-19 Household Response Phone Survey (Round 5). World Bank, Ulaanbaatar. WEF (World Economic Forum). 2019. The Global Competitiveness Report, 2019. https://www3.weforum.org/docs/ WEF_TheGlobalCompetitivenessReport2019.pdf. World Bank. 2020. From Containment to Recovery: World Bank East Asia and Pacific Economic Update. October 2020. World Bank, Washington, DC. https://www.worldbank.org/en/region/eap/publication/east-asia-pacific-economic-update. World Bank. 2021. Mongolia Economic Update, February 2021: From Relief to Recovery. World Bank, Ulaanbaatar. https:// openknowledge.worldbank.org/handle/10986/35161. World Bank. 2022. Mongolia Economic Update, April 2022: Navigating Stronger Headwinds. World Bank, Ulaanbaatar. https://thedocs.worldbank.org/en/doc/af1a0293254ac2e448cafa165c669d88-0070012022/original/MEU-2022- April-ENG.pdf. CHAPTER 5 The Philippines by Edgar Avalos, Irene Jo Estigoy Arzadon, Jaime Frias, Karl Robert Lasmarias Jandoc, Jesica Torres, and Trang Thu Tran51 The COVID-19-induced economic crisis precipitated the worst single-year decline in Gross Domestic Product (GDP) since the 1940s in the Philippines and affected economic growth, poverty, and inequality. GDP fell 9.5 percent in 2020 (PSA 2021b). This decline in growth dealt a heavy blow to the poor and vulnerable and threatened to reverse the gains in poverty and inequality reduction achieved over the past decade. Government estimates show that the incidence of poverty increased to 18.1 percent in 2021 (up from 16.7 percent in 2018), an addition of 2.3 million poor between 2018 and 2021. While the Gini coefficient decreased by 1.6 percentage points, down from 42.3 percent in 2018 to 40.7 percent in 2021, this was mainly caused by the steep deterioration in incomes from higher segments of the distribution. The Philippines implemented drastic measures to contain the spread of COVID-19. In mid-March 2020, the government imposed strict community quarantine measures, allowing only a few essential economic activities to take place. Restrictions eased over time, but containment measures consistently remained some of the strictest in the East Asia and Pacific (EAP) region and the world. This chapter takes stock of the impact of the pandemic on firms and households in the Philippines and examines the government response. It assesses the potential scarring effects of the crisis on employment and inequality in order to inform the design of policy during the recovery phase. To look at the impact of the pandemic on firms, this chapter draws on data from four rounds of the business pulse survey, spanning from July 2020 to March 2022. To examine impacts on households, the chapter uses three rounds of the high- frequency phone survey (HFPS), which span August 2020 to June 2021. The BPS and HFPS were administered by the World Bank, the Philippines Department of Finance, and the National Economic Development Authority (NEDA) For more information, see annex 5A. Both surveys are complemented by data from the Labor Force Surveys (LFS), administrative datasets for the tracking of policy responses, and other publicly available sources, to cover the period from the second quarter of 2020 to the second quarter of 2021. Timeline of the COVID-19 Pandemic and Government Measures Figure 5.1 shows the collection period of the BPS and the HFPS with COVID-19 cases, levels of mobility (mobility index), and restrictions imposed by the government (stringency index) in the Philippines.  This chapter greatly benefited from input provided by Kyung Min Lee and Sharon Faye Alariao Piza. 51 135 136 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 5.1  Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in the Philippines, 2020–22 HFPS R1 HFPS R2 HFPS R3 80 125 Inverted mobility index (14-day average) Stringency index (14-day average) 60 100 Stringency index 40 75 20 50 Inverted Mobility index 0 25 –20 0 BPS R1 BPS R2 BPS R3 BPS R4 40 COVID-19 cases/deaths (per 100,000 people) 30 COVID-19 cases 20 10 COVID-19 deaths 0 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Source: Mobility data comes from Google Mobility Reports; stringency data comes from the Oxford Covid-19 Government Response Tracker (OxCGRT); COVID-19 cases and deaths are derived from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. Note: The mobility index is the average percentage change in time spent at retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. For each day of the week, the baseline is defined as the median value from the five-week period of January 3–February 6, 2020. The figure shows 14-day rolling averages of the mobility index, which is inverted so that higher values indicate lower mobility. The stringency index records the strictness of lockdown style policies; 14-day rolling averages are shown. COVID-19 cases and deaths are smoothed by a seven-day average. Phases of the pandemic are determined by identifying structural breaks in COVID-19 cases and the mobility index. See the introduction for further details. Employment Impacts: Shocks and Recovery At the onset of the crisis, hunger soared and perceptions of quality of life deteriorated. The situation improved gradually throughout 2021. Data from the Social Weather Stations Survey—which collects quarterly information on welfare indicators, including self-rated poverty, hunger incidence, and perception of quality of life among Filipinos—reveals substantial improvements across these indicators from 2010–19, followed by deterioration to unprecedented levels (SWS 2021a, 2021b, 2021c). The incidence of hunger— defined as having experienced involuntary hunger in the three months preceding the survey—declined from 21 percent in the first quarter of 2010 to 9 percent in the last quarter of 2019; it reached 31 percent in September 2020. At the same time, the proportion of people experiencing severe hunger skyrocketed to six times its level in 2019 and three times its level in 2010. The proportion of respondents who reported feeling that their quality of life was better than it was in the previous year increased from 19 percent in 2010 to 39 percent in 2019 before crashing to 6 percent by September 2020. Optimism about the future quality of CHAPTER 5: THE PHILIPPINES 137 life also decreased, from 48 percent of respondents in the fourth quarter of 2019 to 32 percent by September 2020. By mid-2021, the hunger rate had fallen to 14 percent, perceptions about better quality of life had increased to 18 percent, and the proportion of people feeling that their quality of life would be better in the next 12 months had risen to 37 percent, still below pre-pandemic levels (SWS 2021c). Subjective measures of poverty remained persistently high from 2019–20, reflecting unmet aspirations for better living conditions in the Philippines. Self-rated poverty increased from 43 percent in the first quarter of 2010 to 54 percent in the fourth quarter of 2019, before falling to 48 percent in the second quarter of 2021. Self-rated food poverty followed the same trend, increasing from 31 percent in 2010 to 35 percent in 2019 before falling to 32 percent in the second quarter of 2020.52 Self-rated poverty indicators have shown greater volatility across both years and quarters: Between the third and fourth quarter of 2019 alone, self- rated poverty increased 12 percentage points, even though the incidence of hunger fell, optimism improved, and a larger share of respondents reported that their quality of life had improved over the past year. Evidence from the Annual Poverty Indicators Surveys (APIS) documents income declines, particularly among urban households and households headed by women (PSA 2021c). They show that real per capita household income declined 14 percent between 2019 and 2020, with most of the decline occurring between the first and second quarters of 2020. Although all population groups were affected, urban households and households headed by women experienced the largest income declines. Because of their concentration in areas and sectors that were hardest hit by the crisis, households in the third and fourth income deciles suffered larger declines in their incomes than other groups. Households in the poorest quintiles suffered the largest declines in food expenditures. Long-term prospects for school-age children are particularly concerning, as the pandemic reduced educational attainment and learning, particularly among the poor. With the shift toward remote and modular learning because of lockdowns, poor students faced additional disadvantages because many lacked the computers and Internet connectivity needed to use these learning methods. According to the Program for International Student Assessment (PISA) conducted by the Organization for Economic Cooperation and Development (OECD) before the pandemic, fewer than 10 percent of households in the bottom economic, social, and cultural status (ESCS) decile had Internet access or at least one computer, and only 56 percent had at least one smartphone. In the top decile, more than 90 percent of students had all three. In the lowest ESCS decile, only 53 percent of students had a desk at home and just 50 percent had a quiet place to study, compared with 91 percent and 83 percent, respectively, for the top decile (OECD 2019). These factors will lead to substandard learning opportunities and weaker future labor market prospects for these students. The pandemic effects in the Philippines were mediated by a business environment quality that trails slightly behind the average in the East Asia and Pacific region. Despite a stable macroeconomic framework, security problems and corporate governance drag down institutional quality indicators (WEF 2019). While tariffs and non-tariff barriers on imports are reasonably low, domestic competition is limited by a high degree of market dominance and substantial costs of starting new businesses. Labor flexibility is comparable to regional standards, yet credit to private sector, particularly to small and medium-size enterprises (SMEs), still lags behind the levels observed in some of the country’s peers (WEF 2019). 52 Respondents were asked to categorize their household as “Not poor”, “On the line”, or “Poor” based on the type of food eaten by their household. 138 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA The Philippines’ severe containment measures drastically affected private sector performance, leading to large losses in sales revenue and a high rate of temporary business closures. Businesses in the Philippines experienced a large drop in sales and high rates of disruption to operations. By the last quarter of 2021, the Philippines’s output remained below its pre-pandemic level, while other countries in the region, such as Malaysia and Indonesia, surpassed their pre-pandemic output. Much of the shock to sales can be explained by the severity of mobility restrictions. Differences in sales across regions were strongly correlated with changes in average mobility. Lockdown and mobility restrictions were strictest in the National Capital Region, which experienced the largest drop in sales. The business pulse survey respondents cited restrictions on business operations and consumer mobility as the main reasons for sales decline. The first quarter of 2022 saw a strong rebound with the economy expanding 8.3 percent, year on year, thanks to easing of mobility restrictions. Still, in March of 2022, average revenue losses the month before the business pulse survey interview were 59 percent lower than in the same period in 2019. The shock led to sharp reductions in private employment, which slowed as restrictions eased. By December 2020, average employment had declined 37 percent from its pre-pandemic level, and 45 percent of businesses reported laying off workers. As restrictions eased and more businesses were allowed to resume operations, businesses across the size distribution and in every sector started to hire workers. Between the end of 2020 and the second quarter of 2021, the share of businesses reducing operations fell significantly (Figure 5.2, panel a), and the share of both small and large firms increasing operations increased (Figure 5.2, panel b). In every sector, including hospitality, employment losses were lower than at the end of 2020. Strict containment measures produced a severe labor market shock during the early months of the pandemic, with the unemployment rate more than tripling between January 2020 and April 2020, rising FIGURE 5.2  Net job layoffs and net job creation by firms in the Philippines in November/December and May 2021 a. Net layo s b. Net job creation 50 job destruction 30 days prior to interview 50 job creation 30 days prior to interview Adjusted fraction of firms with net Adjusted fraction of firms with net 40 40 30 30 20 20 10 10 0 0 1 5 25 50 100 1 5 25 50 100 Employment in baseline (December 2019; log scale) Employment in baseline (December 2019; log scale) Nov–Dec 2020 May 2021 Nov–Dec 2020 May 2021 Includes controls for sector and region. Computations use sampling weights Includes controls for sector and region. Computations use sampling weights Source: Authors’ calculations using data from business pulse surveys. Note: Computations use sampling weights. Figures are based on activities 30 days before the interview. CHAPTER 5: THE PHILIPPINES 139 from 5.3 percent to 17.6 percent. The labor force participation rate plummeted by 6 percentage points over this period (Figure 5.3). There were also significant shifts in sectors of employment, with the share of workers in agriculture rising from 23 percent in January 2020 to 26 percent in April 2020 and the share of both industry and services declining. The dip in key labor market indicators was followed by a sharp recovery, but by the second quarter of 2021, unemployment remained above its pre-pandemic trend. As mobility restrictions were eased and businesses were permitted to gradually reopen, employment figures started to improve. Labor force participation rapidly increased to pre-pandemic levels in the third quarter of 2020, before dropping by 4 percentage points in the fourth quarter. Unemployment fell to 10 percent in July 2020, then plateaued at 8.8 percent during the fourth quarter of 2020 through February 2021, before falling to 7.1 percent in March 2021, more than a year after the start of the pandemic. Unemployment rose again in April 2021, as Metro Manila and its neighboring provinces reverted to Enhanced Community Quarantine (the strictest category of mobility restrictions involving total lockdowns) as a result of rising COVID-19 cases. Both employment and labor force participation increased in June 2021, recovering to pre-pandemic levels. There is some evidence of potentially permanent scarring effects in the labor market. The recovery was characterized by a reallocation of employment into agriculture and wholesale and retail trade, where earnings have been comparatively low. In the second quarter of 2021, employment in the hospitality industry was still 27 percent below the pre-pandemic trend (Figure 5.4). In contrast, the number of workers in agriculture and the wholesale and retail sector increased by 12 percent and 15 percent, respectively. The pandemic may have exacerbated existing shortcomings in the labor market, as the three sectors already accounted for the largest shares of employment in the Philippines yet had the lowest earnings. The reallocation into low-quality jobs was more pronounced among women than among men. Both self-employment and employment in family firms increased markedly among women as the labor market FIGURE 5.3  Labor force participation and unemployment rates in the Philippines, 2019–22 64 20 18 62 Labor force participation rate (percent) 16 Unemployment rate (percent) 60 14 12 58 10 56 8 54 6 4 52 2 50 0 2019Q1 2019Q2 2019Q3 2019Q4 2020Q1 2020Q2 2020Q3 2020Q4 2021Q1 2021Q2 2021Q3 2021Q4 2022Q1 Labor Force Participation Rate Unemployment Rate Source: Authors’ calculations using data from Labor Force Surveys. Note: Computations use sampling weights. 140 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 5.4  Index of number of workers in the Philippines in select sectors, 2019–21 120 110 Million workers (2019 = 100) 100 90 80 70 60 50 40 2019 2020 2021 Agriculture Manufacturing Wholesale and retail Hospitality Other services Others Source: Authors’ calculations using data from Labor Force Surveys. Note: Computations use sampling weights. Figures are from the second quarter of each year. recovered. Relative to the second quarter of 2019, employment in family firms in the second quarter of 2021 was almost 40 percent higher among women and about 20 percent higher among men (Figure 5.5). The number of self-employed women also increased, but only in agriculture and retail, suggesting that these new entrepreneurs likely started their businesses out of necessity (barriers to entry in agriculture and retail are typically low). The rate of business ownership with employees quickly recovered in 2021, after a drop of more than 40 percent in 2020. A heavier childcare burden, which was borne disproportionally by lower-income workers, also led to lost employment opportunities. Competing childcare needs remained the top employment barrier for FIGURE 5.5  Index of employment in the Philippines, by type of employment and gender, 2019–21 a. Women b. Men 160 140 140 120 120 Million workers (2019 = 100) Million workers (2019 = 100) 100 100 80 80 60 60 40 40 20 20 0 0 2019 2020 2021 2019 2020 2021 Private establishment Self employed Private establishment Self employed Employer Employer Without pay (family operated) Source: Authors’ calculations using data from Labor Force Surveys. Note: Computations use sampling weights. Figures are from the second quarter of each year. CHAPTER 5: THE PHILIPPINES 141 prime-age (25–54) workers in the bottom 60 percent of the income distribution. Data from the HFPS show that at the end of 2020, prime-age workers in the bottom 60 percent of the income distribution were 8 percentage points more likely not to report employment at the time of the interview than workers in the top 40 percent (31 versus 23 percent). The need to care for children and the closing of firms by employers because of pandemic restrictions were the top reasons reported for not working. These factors disproportionately affected workers in the bottom 60 percent, among whom 35 percent of those not working cited the need to care for children, and about 25 percent cited a business closing by their employer. The corresponding figures for respondents in the top 40 percent were 30 percent and 10 percent, respectively. When the labor market recovered, in the second quarter of 2021, workers in the bottom 60 percent still disproportionately reported childcare as the main reason that they were not working (35 percent, compared with 17 percent for workers in the top 40 percent). Among both groups of respondents, a third reported not having worked the week before the interview. The pandemic increased unemployment for skilled labor but also created potential for gains in entrepreneurship. The initial economic shock drove the unemployment rate up about 6 percentage points across the entire skill distribution (Figure 5.6). Employment among workers with elementary education had fully recovered by the second quarter of 2021. In contrast, unemployment among workers with high school or college education was still about 3 percentage points above the pre-pandemic trend. The gap in unemployment rates between workers with a college education and those with only elementary schooling increased from 3 percentage points in the first quarter of 2020 to 5 percentage points in the second quarter of 2021. Among employed college-educated workers, self-employment dramatically increased. In the second quarter of 2021, the number of workers with at least incomplete college in self-employment was about 30 percent higher than the pre-pandemic level. Most of these workers, however, flowed into agriculture and retail, where entry barriers are typically low, which would suggest that self-employment served as a shock absorber. The COVID-19 crisis thus seems to have exacerbated the persistently high unemployment for skilled workers that has characterized the labor market in the Philippines, although it may have increased entrepreneurship activity among this group. FIGURE 5.6  Unemployment rate in the Philippines, by years of formal schooling, 2019–21 12 10 Unemployment rate 8 6 4 2 0 2019-Q1 2019-Q2 2019-Q3 2019-Q4 2020-Q1 2020-Q2 2020-Q3 2020-Q4 2021-Q1 2021-Q2 Elementary (complete/incomplete) High-school (complete/incomplete) Some college and above Source: Authors’ calculations using data from Labor Force Studies. Note: Computations use sampling weights. 142 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Most of the jobs performed by Overseas Filipino Workers (OFWs) were severely hit by the pandemic. Millions of Filipinos work abroad as domestic workers, hotel clerks, office cleaners, and sea workers in the global cruise ship industry. Remittance inflows in 2020 amounted to around 9.6 percent of GDP, making the Philippines one of the most vulnerable countries to a reversal of remittance flows as a result of the pandemic. Households with at least one migrant member received income from remittances that was slightly greater than their income from domestic sources. International travel bans made it impossible for many OFWs who were in the Philippines during the pandemic to return to their host countries. Supporting them while they were unable to work abroad was critical to preventing them and their families from falling into poverty. The COVID-19 pandemic boosted the adoption of digital technology in the private sector, but the degree of adoption was uneven across firms. The average number of new digital services consumed by individuals climbed from 3.9 pre-pandemic to 8.2 in 2021. By the second quarter of 2021, 64 percent of firms reported new or increased use of online sales in response to COVID-19, and 71 percent reported new or increased use of digital promotion. Medium-size and large firms increased their use of technologies significantly more than micro and small firms (Figure 5.7). The type of digitalization also varied across firms. The use of back-end digital technologies, which are likely to lead to productivity growth, lagged behind the use of customer-facing solutions. The share of new adopters of digital technologies was 9 percent for marketing and 8 percent for customer relationship management and supplier relationship management software. New adoption of enterprise resource planning was even lower, at just 7 percent of firms. Micro, small, and medium enterprises (MSMEs) were particularly unlikely to adopt productivity-enhancing technology. This adoption gap is driven by financial constraints (reported by 54 percent of businesses in the BPS), infrastructure constraints (26 percent), and the limited managerial capacity of firms and their employees (20 percent). FIGURE 5.7  Share of firms in the Philippines that report increasing their use of select technologies in the second quarter of 2021 40 35 30 25 Percent of firms 20 15 10 5 0 Micro Small Med/Large Micro Small Med/Large Micro Small Med/Large Micro Small Med/Large (0–4) (5–19) (20+) (0–4) (5–19) (20+) (0–4) (5–19) (20+) (0–4) (5–19) (20+) Online sales or payment Social media data CRM or SRM software ERP software solutions analytics and marketing Before Covid-19 After Covid-19 Source: Authors’ calculations using data from business pulse surveys. Note: Computations use sampling weights. CRM: customer relationship management; SRM: supply relationship management; ERP: enterprise resource planning. CHAPTER 5: THE PHILIPPINES 143 Digital payment transactions increased rapidly during the pandemic. Both the number and the value of account-to-account transfers made through fast-payment infrastructure increased by a factor of six between February 2020 and April 2022. These transactions include payment transactions initiated by mobile phones using the QR Ph, a standardized Quick Response code.53 Fiscal Support to Firms and Households Size and composition of the fiscal response In March 2020, the government launched relief support for households and firms, yet the total support package was modest compared with other countries in the region. Implementing agencies were provided with budgets to improve health systems and grant relief to affected households, workers, and businesses. By September 2020, more funds were appropriated for lending to small businesses; providing relief to agriculture, transport, and tourism; and extending cash transfers to households. The program also supported health and education. However, as shown in the introduction, the fiscal response in the Philippines was relatively modest: the country reduced taxes less than Singapore, Thailand, Indonesia, and Malaysia. It also allocated less to equity, loans, and guarantees.54 Philippine policymakers developed a broad range of policy tools to address problems created by the pandemic shock. Interventions targeted both the supply side (households) and the demand side (firms) of the labor market (Table 5.1). On the supply side, workers received direct cash support, unemployment insurance, and/or death and disability benefits. Health care workers received hazard pay, and the government made efforts to expand capacity in health care. OFWs received cash subsidies and repatriation assistance. The government also sponsored training and job fairs. Some workers benefited from government programs that targeted households, including cash subsidies through the Social Amelioration Program (SAP), as well as subsidies on food and non-food items. For firms, the main line of support came in the form of asset-based programs—equity, loan, and guarantee assistance—although some microenterprises received cash subsidies. Some firms received management and labor training and were granted flexibility in fulfilling payment obligations for utilities, rent, and loans. The government’s free vaccination program paved the way for workers to return to work and eased movement restrictions to open businesses. The government’s relief strategy was articulated in two acts, Bayanihan 1 and Bayanihan 2 (Box 5.1). In addition, the government reallocated a part of the regular funds from the General Appropriations Act for fiscal years 2020 and 2021 for COVID-19 related expenses. Overall, the government disbursed ₱570.04 billion for COVID–related expenses as of September 30, 2021 (DBM 2021), about 3.2 percent of GDP for 2020. 53  To increase the adoption of digital payments, in 2019, the central bank (the Bangko Sentral ng Pilipinas [BSP]) issued a circular that required all participating payment service providers, including banks and non-bank electronic money issuers (EMIs), to adopt QR Ph. The BSP and the payment industry introduced a new stream that enables digital payments between customers and merchants, which should increase adoption of digital payments by micro and small firms. 54  The share of spending and forgone revenues was larger than support in the form of equity, loans, and guarantees, as also observed in other economies. 144 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE 5.1  Labor market responses to the COVID-19 pandemic by the government of the Philippines Target group Intervention Workers • Cash subsidies for households • Wage subsidies for workers • Cash for work • Unemployment insurance • Disability and health benefits • Hazard pay for health care workers • Hiring of additional health care workers • Repatriation of Overseas Filipino Workers • In-kind support for households (food and non-food items) • Loan payment moratorium • Assistance to job seekers • Labor regulation adjustment • Free vaccination Firms • Livelihood seeding programs • Access to loans and credit guarantee • Technology tools and resources • Support for upskilling and reskilling • Training for management capacities • Postponement of utility and rent payments • Loan payment moratorium • Access to market (online platforms, trade fairs) Source: World Bank 2021. Coverage and timing MSMEs represented 99.5 percent of all establishments in the Philippines in 2020. They were the target group for loan assistance from CARES, the Land Bank of the Philippines, and the Development Bank of the Philippines. The Small Business Wage Subsidy (SBWS) was designed with the aim of providing a two-month subsidy to the eligible 3.4 million employees of 1.6 million small businesses affected by the community quarantines. The cost was estimated at about 0.3 percent of GDP. The SBWS required that (a) business beneficiaries continue to maintain the employment status of eligible employees throughout the quarantine, and (b) employees who benefit from the program cannot resign during the quarantine period. The program was meant to preserve jobs at firms experiencing a temporary reduction in business activity by reducing their labor costs while supporting the incomes of workers whose hours were reduced. The SBWS program was designed as a job retention scheme for MSMEs. It was open to small firms that were (a) not part of the Large Taxpayer Services List of the Bureau of Internal Revenue and (b) were either nonessential businesses (category A) that were forced to temporarily close or suspend operations, or quasi- essential businesses (category B) that were permitted to operate with only minimal staff. The wage program was a temporary measure that was paid for up to two months between May 1 and June 28, 2020, so that affected small businesses could retain their employees during the quarantine period. Employers submitted applications through the Social Security System (SSS); payments were made directly to employees. The wage subsidy amounted to ₱5,000–₱8,000 per month per eligible employee, depending on the regional minimum wage. CHAPTER 5: THE PHILIPPINES 145 BOX 5.1  Emergency social transfers in the Philippines during the COVID-19 pandemic To help mitigate the economic impacts of COVID-19, the Philippines passed Bayanihan 1 on March 2020. It lapsed in June 2020. Bayanihan 1 was replaced by Bayanihan 2, which was enacted in September 2020 and lapsed in June 2021. Bayanihan 1 Bayanihan 1 addressed the immediate challenges posed by the spread of COVID-19. Its main component was the provision, through the Emergency Subsidy Program, of ₱5,000–₱8,000 a month per household for two months to help ease difficulties from the initial Enhanced Community Quarantine. Additional outreach and funds were included to reach poor households missed by the initial transfers. Bayanihan 1 also included targeted support for displaced workers, including the following: • The Department of Labor and Employment’s Abot Kamay ang Pagtulong (DOLE-AKAP) for OFWs • Tulong Panghanapbuhay sa Ating Disadvantaged/Displaced Workers (TUPAD), which provided temporary jobs for informal sector workers • The COVID-19 Adjustment Measures Program (CAMP) for displaced formal sector employees • The Social Amelioration Program (SAP), a cash subsidy program to families • The Department of Agriculture’s Rice Farmers Financial Assistance Program (RFAP) • The Department of Trade and Industry’s Livelihood Seeding Program and Negosyo Serbisyo sa Barangay, which provided livelihood kits in case of catastrophic events • Relief assistance from the Department of Social Welfare and Development (DSWD) • Relief assistance from governmental bodies other than the DSWD • Relief assistance from sources other than the government In addition to measures targeting households and workers, Bayanihan 1 contained provisions to support micro, small, and medium-size enterprises (MSMEs) and increase funding for the health sector. Programs assisting firms include the Small Business Wage Subsidy Program, a credit guarantee program for loans to MSMEs through the Philippine Guarantee Corporation (PhilGuarantee), tax relief for firms, and an interest-free and collateral-free loan facility under the COVID-19 Assistance to Restart Enterprises (CARES) Program of the Small Business Corporation. Bayanihan 2 Bayanihan 2 extended many of the provisions of Bayanihan 1, maintaining the programs for displaced workers, AKAP, TUPAD, and CAMP. In addition, to continue support for displaced workers, it contained many sector-specific programs, including transfers to the transport sector, assistance to teaching and nonteaching personnel at schools, and cash and training support for the tourism sector. For agricultural workers and fishers, Bayanihan 2 provided ₱5,000 worth of cash and food assistance. In 2021, to mitigate the impacts of a new Enhanced Community Quarantine in the National Capital Region and surrounding provinces, it provided more funds (up to ₱4,000) to households in affected areas. Bayanihan 2 expanded SBCorp’s lending programs for MSMEs to include cooperatives, tourist-related businesses, hospitals, and affected OFWs. In addition to increasing funding for loans, the program streamlined documentary requirements and loan-processing time, increased the maximum allowable number of borrowers, extended loan terms, reduced interest rates, and facilitated the use of financial technologies. Other government and financial institution, such as the Land Bank of the Philippines and the Development Bank of the Philippines, also implemented loan programs for MSMEs and cooperatives engaged in agriculture and fishing-related activities. 146 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Balance sheet measures, including debt relief, equity, and loan guarantees, were used to support MSMEs. Because of the uncertainty created by the quarantine restriction, banks may have been inclined to ration credit in favor of borrowers that were more creditworthy, subjecting distressed MSMEs to more stringent financing conditions and borrowing costs. Credit guarantees offer a form of backing from the government to secure, either partially or fully, the value of a loan, reducing the risk to lenders and thereby incentivizing them to extend credit when collateral is missing, or during times of uncertainty, when financial instructions may be hesitant to lend. The special credit guarantee facility was meant to prevent a potential MSME liquidity crunch by enabling small business to access liquidity. The guarantee program for MSMEs covered around 50  percent of loans to small businesses. Equity is a longer-term investment instrument, which normally pays minimal dividends in the early life of a company. It provides insolvency risk for the invested capital in target companies. Cash-for-work programs such as the Tulong Panghanapbuhay sa Ating Disadvantaged/Displaced Workers (TUPAD) helped support informal sector workers. In contrast to direct cash transfers for formal workers under the SWBS, TUPAD beneficiaries needed to perform community service, such as street sweeping or maintenance work, for which they could earn up to ₱8,000. TUPAD is a community-based program that existed before the pandemic; systems for identifying qualified beneficiaries were therefore already in place. TUPAD is managed by the same local government unit handling conditional cash transfers for households. Poor and near-poor families received social amelioration and government relief assistance under the Bayanihan Act. SAP provided cash transfers of ₱5,000–₱8,000 per month to about 18 million households, depending on the minimum wage in the region of residence of a given household, for two months. This was equivalent to monthly earnings at about the minimum wage and households’ subsistence expenditures in each region. Beneficiaries are either poor or have low income; most work in the informal sector. Of the 18 million families covered by SAP, 4.3 million were beneficiaries of the Philippines’ conditional cash transfer (CCT) program, the Pantawid Pamilyang Pilipino Program (4Ps). In addition, 13.6 million non-4Ps households received similar payments. SAP coverage increased during the first half of 2020. Between January and June, the average transfer per household under conventional social assistance was ₱7,356 (₱7,789 for the poorest quintile and ₱5,924 for the richest one); average transfers under the 4Ps were ₱8,789 for the poorest quintile and ₱9,786 for the richest one. By the summer of 2020, 74 percent of households surveyed in the HFPS reported receiving some form of public support. The financial assistance program for OFWs affected by the COVID-19 pandemic (the DOLE-AKAP Program) provided one-time cash assistance of ₱10,000. The program covered regular and qualified undocumented OFWs, as well as OFWs who were unable to return to their host country because of COVID-19 related restrictions. The temporary support provided through the program was expected to assist 70,000 qualified OFW beneficiaries, who were also provided with temporary shelter, food, and transportation assistance. By July 2021 the program had provided ₱5.487 billion in support to over 540,000 OFWs (Patinio 2021). The government also implemented reintegration programs for returning OFWs. These programs are overseen by DOLE and implemented primarily by the National Reintegration Center for OFWs, regional offices of DOLE, and the Overseas Workers Welfare Administration (OWWA). These packages are composed of (a) services that mitigate the impact and costs of returning, such as psychosocial counseling, stress CHAPTER 5: THE PHILIPPINES 147 debriefing, values formation, and OFW family circles; and (b) services to better equip return migrants to meet their material needs and economic goals, such as finding paid employment or starting an enterprise. At least half of returning OFWs in 2020 registered for reintegration. According to a survey by the International Organization for Migration, 53 percent of OFWs indicated that they needed financial help to support their basic needs, but only 26 percent had received government support by December 2020 (IOM 2021). Targeting Data from the HFPS and the BPS point to complementarities between support programs targeting households and firms. Considering only households in which the head is a business owner, the share of business owners in the HFPS that report having received support increases to 73 percent (compared with about a quarter in the BPS). This finding suggests that support was targeted to individuals rather than firms, even as support to households declined to about 50 percent in the second quarter of 2021. According to the World Bank COVID-19 Household Survey, the provision of non-cash aid (food and non-food) reached almost all households in the Philippines, regardless of income status. DSWD provided family food packs and non-food items worth ₱1.1 billion to local government units (LGUs) for distribution to vulnerable families. The government’s food pack typically included rice, canned food, and energy drinks, which could sustain a family of five for one to two days. Other in-kind assistance reached 34 percent of households. The 4Ps beneficiaries were slightly more likely to receive this assistance, at 37 percent versus 31 percent for non-4Ps beneficiaries. Among the non-food items received, 65 percent of households received personal protective equipment items, 25 percent received disinfectants, and 11 percent received medicines. While food and non-food aid package distribution was widespread, each package was quite small—equivalent to approximately ₱520 (US$10.4) per household. The targeting of assistance seems to have been mixed, indicating opportunities for directing support toward the most vulnerable. Social assistance and the 4Ps were progressive, with over 30 percent of funds allocated to the poorest quintile and just 6 percent to the richest quintile. In contrast, Bayanihan 1 did not have a progressive structure upon implementation, possibly because policymakers recognized that the effects of the pandemic were not related only to income level or location. The widespread initial impact of the crisis may have warranted broad coverage, but adjustments will be needed to better target assistance programs, as impacts seem to cumulate for the poor, while better-off households experience better recovery of their incomes. The Philippines needs to address the country’s lack of updated social and beneficiary registries and a national ID system, both of which are crucial for rapidly expanding support. When there is accurate and up-to-date information relevant in determining the basic socioeconomic status of households, the government can more readily identify potential beneficiaries. Support should be directed to firms in distress from the COVID-19 shock, which are economically viable (Freund and Garcia Mora 2020). Providing support to nonviable firms risks “zombification”— prolonging the demise of firms that will eventually go bankrupt. Factors such as solvency, vulnerability, and informality will depend on the context and the firm type. Globally, a considerable number of firms that did not experience observable vulnerability from a drop in sales because of COVID-19 received public support, while some firms in need of support received none (Cirera and colleagues, 2021). A related study (Apedo- Amah and colleagues, 2020) reveals that smaller firms, which were affected more severely than larger ones, tended to receive less support, suggesting either eligibility, awareness, or capacity barriers to access support, or poor targeting by the government. 148 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA The BPS shows that larger firms, which reported proportionally lower revenue losses than smaller firms, were more likely to report access to public support, including wage subsidies (Figure 5.8 Correlation between Firm Size and Likelihood of Receiving Financial Support or Wage Subsidy in the Philippines). Similarly, households in the top 40 percent of the income distribution were just as likely to report access to support as households in the bottom 60 percent (Figure 5.9); there was no significant difference in the level of support in the HFPS across income groups. These results may be driven partly by the set of instruments used. Eligibility issues on government support applications may have driven the mistargeting. Most programs providing wage subsidies applied a “no redundancy rule,” which did not allow workers eligible for SAP to receive support from CAMP or TUPAD. SAP covers only low-income households that meet at least one of the following criteria: participate in the 4Ps or conditional cash transfer program, work in the informal economy, or belong to a vulnerable demographic group—for example, seniors, people with disabilities, indigent indigenous people, or homeless people. Some workers in businesses with a small number of employees, such as microenterprises or businesses in the informal sector, were not eligible for wage subsidies. Some firms and workers did not meet the minimum eligibility requirements for government support, which include basic compliance requirements with the tax administration. Under the Small Business Wage Subsidy program, for example, employees registered with the Bureau of Internal Revenue and the Social Security System (SSS) who had paid their tax and payment obligations were prioritized. CARES and other loan assistance programs required that firms register with the Department of Trade and Industry (DTI) or the Securities and Exchange Commission to qualify for debt relief. Reports suggest that some workers and business owners did not register with these agencies, although the government urged employees and employers to do so (Villanueva 2020) FIGURE 5.8  Correlation between firm size and likelihood of receiving financial support or wage subsidy in the Philippines a. Financial support b. Wage subsidy 50 20 Adjusted fraction of firms with Adjusted fraction of firms with 40 access to wage subsidies access to public support 15 30 10 20 5 10 0 0 1 5 25 50 100 1 5 25 50 100 Employment in baseline (December 2019; log scale) Employment in baseline (December 2019; log scale) Nov–Dec 2020 May 2021 Nov–Dec 2020 May 2021 Includes controls for sector and region. Computations use sampling weights Source: Authors’ calculations using data from business pulse surveys. Note: Computations use sampling weights. CHAPTER 5: THE PHILIPPINES 149 FIGURE 5.9  Government support of households in the top 40 and bottom 60 percent in the Philippines a. Percent of households receiving public support b. Average self-reported value of support 80 10000 Fraction of households with access to public support 70 9000 8000 Value of public support (₱ per month) 60 7000 50 6000 40 5000 30 4000 3000 20 2000 10 1000 0 0 Round 2 Round 3 Round 2 Round 3 Bottom 60% Top 40% Bottom 60% Top 40% Source: Authors’ calculations using data from high-frequency phone surveys (HFPS). The round 2 survey was implemented in December 2020, and the round 3 survey in May 2021. Note: Computations use sampling weights. Some mistargeting may have been driven by agencies underspending their budgets for COVID-19 support. A report by the Commission on Audit (COA) revealed that the gap between budget allotments and disbursements of government agencies that handled COVID-19-related funds was 50.9–99.3 percent. Among agencies supervising the funds for the labor market (Table 5.2), the DTI and the Department of Agriculture had the lowest disbursed-to-allotted budget ratio, at 53.1 and 50.9 percent, respectively. The report attributes these gaps to a lack of personnel to organize the response, missing information for locating and reaching out to beneficiaries, and increasing demands for decentralization (COA 2021a, 2021b; Aurelio 2021). Prospective beneficiaries’ fears that they did not meet relief conditions also reduced participation. Firms, especially in the tourism sector, were hesitant to borrow under the CARES program because of TABLE 5.2  Allotments, disbursements, and use of COVID-19 budget by select government agencies in the Philippines, 2020 Allotments Disbursements Disbursed funds as a Government agency (billion ₱) (billion ₱) percent of allotted funds Department of Labor and Employment (DOLE) 28.98 24.59 84.9 Department of Trade and Industry (DTI) 1.30 0.69 53.1 Department of Social Welfare and Development (DSWD) 217.42 185.01 85.1 Department of Agriculture 34.68 17.64 50.9 Department of Finance 101.05 100.33 99.3 Social Security (SSS) 51 45.33 88.9 Source: SSS budget based on data from DBM 2021; for all other figures, consult COA 2021b. 150 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA uncertainty about quarantine restrictions. According to an audit, 60 percent of the CARES budget was allocated to the tourism sector. As a result, of the sector’s 995,745 MSMEs, only 48,010 (4.8 percent) were able to borrow under the program as of June 2021 (Figure 5.10). In the National Capitol Region, Calabarzon, and Central Luzon (which had the largest number of COVID cases), only about 1.5–2.5 percent of MSMEs borrowed under the CARES program. The Mimaropa and CAR regions had the largest shares of borrowing by MSMEs, but even there the program covered only about 7–10 percent of MSMEs. Effectiveness Support to firms appears to have been effective at reducing the risk of bankruptcy, but only 20–25 percent of businesses participating in the various rounds of the BPS reported receiving any public support. In the third round of the BPS, firms that reported receiving at least one type of support were 10 percent less likely than firms that received no support to expect to file for bankruptcy in the next six months. Access to credit mediation and refinancing was negatively associated with the likelihood of having fallen— or expecting to fall—into arrears. Public support was positively associated with increases in both sales and employment in the 30 days before the survey, although these relationships are less precise. The SBWS appears to have been effective in stemming job losses. The International Labor Organization (ILO 2021) conducted an analysis of SBWS policy implementation during May–June 2020 and employment trends in July. It found that the SBWS employment protection policy appeared to have reduced job losses. FIGURE 5.10  Percent of micro, small, and medium-size enterprises in the Philippines that borrowed under the CARES program, by region NCR CALABARZON Central Luzon Central Visayas Western Visayas Davao Region Ilocos Region SOCCSKSARGEN Bicol Region Northern Mindanao Eastern Visayas Zamboanga Peninsula Cagayan Valley MIMAROPA CAR CARAGA BARMM 0% 2% 4% 6% 8% 10% 12% Percent of MSMEs that borrowed under the CARES program Source: World Bank computation using the Philippine Statistics Authority List of Establishments 2021 (PSA 2022) and COA 2021 Performance Audit Report on CARES (COA 2021a). CHAPTER 5: THE PHILIPPINES 151 On average, a higher level of SBWS disbursements (as a share of the regional GDP) correlates with smaller employment losses. A simulation suggests that in the absence of the SBWS, nationwide employment in July 2020 might have fallen by 4.5 percent rather than the reported 3.8 percent. Job losses for women would have been 4.9 percent without the SBWS, instead of the 4.1 percent reported in July 2020. Provision of SAP top-up benefits for 4Ps beneficiaries mitigated food insecurity during the Enhanced Community Quarantine period. The findings of the World Bank’s COVID-19 Household Survey suggest that 4Ps households were less likely to report experiencing food insecurity. Households in the bottom 40 percent of the earnings distribution drove the impact of 4Ps on reducing food insecurity, suggesting that SAP emergency cash subsidies helped very low-income 4Ps households to cope with the shock (Cho and colleagues, 2021). The government provided large-scale cash assistance, but implementation faced challenges. The existence of a well-established social protection program with a solid delivery system proved useful. SAP benefitted many low-income households, and poorer and more vulnerable households were more likely to receive SAP payments. However, the provision of SAP benefits to non-4Ps beneficiaries was delayed, and weak delivery systems meant that many beneficiaries had to spend many hours waiting in line during the pandemic to receive benefits. Moreover, the amounts received by beneficiaries of these cash or cash-for-work schemes were limited, intermittent, and well below the ₱8,393 monthly food threshold for an average family of five in the Philippines in 2021 (PSA 2021a). Results from a macro-micro simulation model suggest that as a result of the 9.5 percent drop in GDP in 2020, the incidence of poverty would have increased to 23.5 percent had the government not implemented emergency COVID assistance measures. Thanks to assistance programs, the incidence of poverty actually declined by 1.6 percentage points, preventing 1.8 million Filipinos from being poor.55 Poverty is projected to decline in the coming years, but it will remain higher than it was before the pandemic because of the record economic slump triggered by the pandemic.56 Lessons Learned The pandemic hit the Philippines harder than it hit most countries in the region—and it had disproportionately severe effects on vulnerable groups. The shock caused severe disruptions to the private sector, sharp reductions in employment, severe declines in income, and unprecedented increases in hunger. In 2020, it caused the worst decline in GDP since the 1940s, changing the long-term trajectory of economic growth, poverty, and inequality. Households at the lower end of the income distribution suffered the largest declines in food expenditure, and households led by women experienced the largest declines in income. Workers at the lower end of the income distribution also disproportionally bore the burden of childcare needs. The pandemic brought some unexpected benefits, however, such as an increase in private sector adoption of digital technology. The extent of digitalization has been uneven, however, with larger firms much more likely to digitalize than smaller firms. Digital adoption has been linked with increased performance of firms, although the BPS points to risks of a widened digital divide between small and large firms. 55  The simulation assumes inefficiencies in the targeting of beneficiaries. If targeting had been perfect, the decline in the incidence of poverty would have been 3.2 percentage points in 2020. 56  Estimated poverty rates also incorporate the government’s expanded coverage of conditional cash transfers and additional education and health grants (social transfers not related to COVID). 152 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Targeting of government support to the most vulnerable firms could have been better. Of the firms sampled by the BPS, government support reached an average of 21 percent of firms across all rounds of the survey. When breaking this statistic down to ascertain the proportion of surveyed micro, small, medium, and large firms that received government support, the figures indicate that 17, 27, 31, and 43 percent, respectively, reported having received support. In other words, large firms were more than twice as likely to receive government support than large ones. This disparity suggests that funds may have been mistargeted and could have been better channeled to more vulnerable, smaller firms that bore the brunt of the economic downturn. The recovery points to potential long-run increases in inequality and a worsening of the allocation of talent in the labor market. The recovery in the labor market was characterized by a reallocation of workers into agriculture and wholesale and retail trade, where earnings are low. Both self-employment and employment in family firms increased as the labor market recovered, especially among women. The COVID-19 crisis therefore seems to have exacerbated the persistent mismatch between supply and demand for skilled workers that characterizes the labor market in the Philippines. The rate of unemployment among workers with a college education has remained persistently high. Entrepreneurship activity among these workers has increased, but mainly in the low-paying sectors of agriculture and wholesale and retail trade. The government took swift and substantive action to mitigate the economic blow of the pandemic. The suite of policy responses helped firms cushion the negative effects of the pandemic, particularly liquidity constraints. The evidence also suggests that policy reduced bankruptcy risk among firms and poverty among households. However, the government’s support package was still modest compared with other countries in the region. Implementation has been effective in reducing redundant payments, but there is evidence of mistargeting. Support has not always targeted the most vulnerable groups, and larger and more capable firms appear to have received more support than other firms. Improving targeting by adjusting the eligibility conditions of beneficiaries could help improve the ability of government support to reach the most affected groups. Social protection programs should be timely and targeted, reaching the individuals who need them most. They are important to mitigate the adverse impact of the pandemic on livelihoods, health, and education, especially among the poor. Moving swiftly to provide transfers and support to poor households requires strong delivery and implementation capacity. Successful rollout of the Philippine national ID system and use of the foundational ID for social protection delivery would enable the digital identification of recipients, as well as digital transformation of the delivery system. The Philippines must leverage the national ID system to strengthen its targeting system, by consolidating and systemizing beneficiaries’ information in a unified database, harmonizing various social protection programs, and facilitating financial inclusion and the digital distribution of transfers. Policymakers could also promote business environment reforms to support the entry and reallocation of resources. The Philippines has introduced a series of reforms to streamline regulations for registering a business and obtaining a license (through Anti Red Tape Authority), improved customs procedures, and liberalized investment (through the CREATE Act). However, additional reforms to facilitate the entry of businesses through enhanced quality business entry regulations, digitalization of government, and securing locations to do business are still pending. In addition, the government could help facilitate the exit of unproductive firms, through reform of bankruptcy and debt resolution, insolvency procedures, institutional structures, and judicial proceedings. CHAPTER 5: THE PHILIPPINES 153 To support workers, the government should prioritize upskilling and retooling, especially of workers displaced by the pandemic. The Technical Education and Skills Development Authority (TESDA) and other technical and vocational education and training (TVET) institutions should be encouraged to provide programs to address skills gaps and increase the employability of workers. These programs should help workers transition out of heavily affected sectors toward expanding sectors, especially sectors that are part of the digitalization wave. Incentives must be given to encourage continuous learning and other social protection measures to mitigate shocks during the transition, such as unemployment insurance, should be explored. The policy response could also improve market access. The Philippines maintains several barriers to trade in services, as illustrated by its high Services Trade Restrictiveness Index (2019) and high costs of trade. In 2019, the cost of importing a container was US$509, and the cost of exporting one was US$603, the second-highest figure in the region (after Myanmar). The government should consider broadening the scope of support to promote digitally enabled market access and support innovative firms that were the most severely affected by the crisis. Many firms experienced a loss of intangible capital, such as business connections, tacit knowledge of clients, investments in research and development (R&D), and other types of knowledge capital. Supporting innovative firms is critical for promoting growth, given the critical role these firms play in economic recovery. The new administration should monitor signs of economic recovery to avoid premature withdrawal of support and control the budgetary implications of support over the long term. As the pandemic recovery enters a new phase, support should address structural conditions rather than continue to assist already strong or unviable firms. 154 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Annex 5 Annex 5A.  The Household and Firm Surveys in the Philippines The HFPS includes both individual- and household-level questions on topics such as employment and income. The four rounds were implemented in the summer of 2020, at the end of 2020, and at the end of the second quarter of 2021 (Table 5A.1). The BPS was designed to help monitor the impact of COVID-19 on the private sector. The four rounds of measurements were implemented at various stages of the pandemic (Table 5A.2). The four samples from the BPS do not correspond to panel data; the first and fourth round of BPS surveys did not collect data on employment at the baseline, and therefore many of the results on employment changes were computed using the data from rounds 2 and 3. Annex 5B.  Description of Government Programs The Bayanihan to Heal as One Act (Republic Act 11469), known as Bayanihan 1, provided implementing agencies with ₱387.9 billion to improve health systems and grant relief to affected households, workers, and businesses. The Bayanihan to Recover as One Act (Republic Act 11494), known as Bayanihan 2, allocated ₱214.1 billion for lending to small business; providing relief to agriculture, transport, and tourism; and extending cash transfers to households. The program also supported health and education.57 Table 5.2 shows the objective, support type, target, implementing government agency, and amount of disbursed funds in 2020–2021. TABLE 5A.1  Dates and sample sizes of high-frequency phone surveys of households in the Philippines Round Survey dates Sample size 1 August 1–14, 2020 9,448 2 December 10–22, 2020 1,805 January 7–17, 2020 3 May 7–June 11, 2021 2,122 Source: Original table for this publication TABLE 5A.2  Dates and sample sizes of business pulse surveys in the Philippines Round Survey dates Sample size 1 July 7–August 1, 2020 33,661 2 November 25–December 10, 2020 4,311 3 May 11–26, 2021 1,425 4 March 9–21, 2022 3,336 Source: Original table for this publication Note: The business pulse survey sample includes firms that were formally registered with a government authority at the time of the survey. 57  The government also pulled ₱ 88.2 billion from the regular funds from the General Appropriations Act of 2020 and 2021 to cover additional COVID–related expenses. CHAPTER 5: THE PHILIPPINES 155 For workers, most of the budget disbursed was in the form of cash subsidies. Workers in the private sector received ₱58.8 billion in cash subsidies through the Small Business Wage Subsidy (SBWS), the unemployment benefit of the Social Security System, and the COVID-19 Adjustment Measures Program (CAMP). Workers in the informal sector had to perform community service work to receive cash through Tulong Panghanapbuhay sa Ating Disadvantaged/Displaced Workers (TUPAD), which disbursed ₱9.62 billion. Overseas Filipino Workers (OFWs) received cash subsidies and repatriation assistance through the Abot Kamay ang Pagtulong (AKAP) program of the Department of Labor and Employment (DOLE) for a total of ₱5.48 billion. The Department of Agriculture used ₱16.4 billion to implement programs targeting rice farmers via the Financial Subsidy to Rice Farmers (FSRF) and the Rice Farmers Financial Assistance (RFFA), a recovery package for the agri-fishery sector. As part of households, workers also received support from the Social Amelioration Program (SAP), which disbursed ₱200.19 billion. The central bank, Bangko Sentral ng Pilipinas, (BSP) provided financial relief for loan borrowers through a debt payment moratorium program with BSP-supervised financial institutions. Under the COVID-19 Assistance to Restart Enterprises (CARES) program, the government disbursed ₱4.99 billion as equity contributions. The Land Bank of the Philippines and the Development Bank of the Philippines provided ₱42 billion as equity contributions, whereas PhilGuarantee targeted MSMEs, providing about ₱180.67 billion in credit guarantees. Programs under Bayanihan 1 (CAMP, TUPAD, SAP, the Livelihood Seeding Program, FSRF, RFFA, and CARES) were implemented between March and June 2020. Bayanihan 2, which ran from September 2020 to June 2021, extended CAMP, TUPAD, and CARES. The SBWS took effect in April–June 2020. The Social Security System (SSS) Unemployment Insurance is a part of the Social Security Act of 2018, which was in effect before the pandemic. Some programs, such as TUPAD and CAMP, were extended until December 2021 through funding from the General Appropriations Act of 2021. The 2020 Annual Poverty Indicator Survey (APIS) collected information on the coverage and financial assistance provided under select social amelioration programs of Bayanihan 1. Coverage under conventional social assistance programs, particularly programs other than the Pantawid Pamilyang Pilipino Program (4Ps),58, 59 declined between 2019 and 2020, but almost all households benefitted from Bayanihan 1 (Figure 5B.1). Between April and June 2020, ₱130.1 billion was transferred to households under Bayanihan 1. Between January and June 2020, the total amount transferred to households under conventional social assistance programs—4Ps, the Unconditional Cash Transfer Program (UCT) the social pension, student financial assistance programs (STUFAP), Emergency Shelter Assistance (ESA), and government feeding program—was ₱49 billion. The average transfer per household under Bayanihan 1 in April–June was ₱5,258 (₱5,320 for households in the poorest quintile and ₱4,180 for households in the richest quintile). The largest programs under Bayanihan 1 in terms of coverage and budget were SAP and the government relief program. SAP covered 47 percent of households, reaching nearly 60 percent of the poorest groups and 25 percent of the richest quintile, and used about 57 percent of Bayanihan 1 financial transfers. The government relief program covered over 85 percent of households and used about 33 percent of Bayanihan 1 transfers. 58  The recall period of APIS 2020 was April–June 2020 for Bayanihan programs and January–June 2020 for conventional social assistance programs. The amounts are therefore lower than the full-year administrative data for 2020, as additional tranches of different Bayanihan programs, such as SAP, were provided in the second half of 2020. 59  Until 2018, the 4Ps program included only the regular conditional cash transfer program. The modified conditional cash transfer program was added to the 4Ps in 2019. 156 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 5B.1  Social assistance coverage in the Philippines in 2020, by wealth quintile, area, and gender 100 90 80 70 Percent of households 60 50 40 30 20 10 0 Poorest Q2 Q3 Q4 Richest Rural Urban Women Men Quintiles Area Gender Philippines 4Ps UCCT Social Pension Other SA Total SA Bayahinan LM Source: Annual Poverty Indicator Survey (APIS) 2019 and 2020. Note: As the survey collected information on amounts only in 2020, it is difficult to compare changes in financial assistance over time. APIS did not provide information on coverage of various social assistance programs until 2019. TABLE 5B.1  Summary of programs by the government of the Philippines targeting workers and firms Percent Government Billions of ₱ of 2020 Program Objective Support type Period Target agency as of 2021 GDP Small Business Wage Subsidy Reduce impact Cash subsidy April–June 2020 Employees Social Security 45.3 0.25 (SBWS) of lockdowns of small, System/ and quarantines formal sector Department of businesses Finance Unemployment Insurance Support Cash subsidy March 2020– Social Security Social Security 2.6 0.01 formal workers ongoing System System who were members involuntarily separated from employment COVID-19 Adjustment Reduce impact Cash subsidy March 2020– Formal sector Department 10.9 0.06 Measures Program (CAMP) of lockdowns December 2021 workers of Labor and and quarantines Employment (DOLE) Tulong Panghanapbuhay Reduce impact Cash for work March–June Informal DOLE 9.6 0.05 sa Ating Disadvantaged/ of lockdowns 2020, September sector workers Displaced Workers and quarantines 2020–December (underemployed, [Employment Assistance for 2021 self-employed, Disadvantaged/Displaced displaced, Workers] (TUPAD) marginalized) CHAPTER 5: THE PHILIPPINES 157 TABLE 5B.1  Summary of programs by the government of the Philippines targeting workers and firms (Continued) Percent Government Billions of ₱ of 2020 Program Objective Support type Period Target agency as of 2021 GDP Social Amelioration Programs: Reduce impact Cash subsidy, March–June 2020 Households Department 200.2 1.12 Emergency Subsidy of lockdowns food and non- (Tranche 1), of Social Program (ESP), Assistance and quarantines food items September Welfare and to Individuals/Families in 2020–June 2021 Development Crisis (AICS), Food and Non- (Tranche 2) (DSWD) Food Items (FNI), Expanded Pantawid Pamilyang Pilipino Program (4Ps) Abot Kamay ang Pagtulong Smooth Cash subsidy March 2020– Overseas Filipino DOLE 5.4 0.03 [“Help is at hand”] (AKAP) transition December 2021 Workers for returned Overseas Filipino Workers Livelihood Seeding Program Retain firms Cash subsidy March –June 2020 Microenterprises Department 0.2 0.001 through business (Tranche 1), of Trade and development September Industry (DTI) assistance and 2020–June 2021 services (Tranche 2) Financial Subsidy to Rice Reduce impact Cash subsidy March–June 2020 Rice farmers Department of 16.4 0.09 Farmers (FSRF) and Rice of lockdowns (Tranche 1), Agriculture Farmers Financial Assistance and quarantines September (RFFA) 2020–June 2021 (Tranche 2) Agri-Fishery Recovery Retain firms Cash/loan March–June 2020 Micro and small Department of Package (micro and small interest rate (Tranche 1), enterprises in Agriculture enterprises) subsidy September food production– 2020–June 2021 related sectors (Tranche 2) COVID-19 Assistance to Retain Micro, Loan March 2020– MSMEs DTI 4.1 (equity 0.02 Restart Enterprises (CARES) Small and ongoing contribution Medium of national Enterprises government) (MSMEs) Loan programs for Micro, Retain MSMEs Loan September MSMEs Land Bank of the 42 (equity 0.23 Small and Medium 2020–June 2021 Philippines and contribution Enterprises (MSMEs) Development of national Bank of the government) Philippines Credit guarantee support Retain MSMEs Loan May 2020– MSMEs Department of 180.7 (credit 1.01 ongoing Finance guarantee) Debt payment moratorium Support workers Financial relief March 2020– Individuals with Bangko Sentral n.a. n.a. and retain firms June 2021 loans ng Pilipinas (Central Bank of the Philippines) Source: World Bank authors, based on information from Republic of the Philippines 2022a, 2022b; agency circulars and resources (COA 2021a, 2021b; Crismundo 2022; DBM 2021; DOF 2021, 2022; DSWD-DOLE- DTI-DA-DOF-DBM-DILG 2020; Patinio 2021, 2022); De Vera 2022; and Ugo and colleagues, 2021. 158 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA References Apedo-Amah, Marie Christine, Besart Avidu, Xavier , Marcio Jose Vargas Da Cruz, Elwyn Davies, Arti Grover, Leonardo Iacovone, Umut Kilinc, Denis Medvedev, Franklin Okechukwu Maduko, Stavros Poupakis, Jesica Coronado Torres, and Trang Thu. Tran. 2020. “Unmasking the Impact of COVID-19 on Businesses: Firm Level Evidence from Across the World.” Policy Research Working Paper No. 9434. Washington, DC: World Bank. http://documents. worldbank.org/curated/en/399751602248069405/Unmasking-the-Impact-of-COVID-19-on-Businesses-Firm- Level-Evidence-from-Across-the-World. Aurelio, Julie. M. 2021. “COA Flags DOLE for Poor COVID-19 Fund Control.” Philippine Daily Inquirer, August 15, 2021. https://newsinfo.inquirer.net/1473706/dole-flagged-for-poor-covid-19-fund-control. Cho, Yoonyoung, Jorge Avalos, Yasuhiro Kawasoe, Doug Johnson, and Ruth Rodriguez. 2021. “Mitigating the Impact of COVID-19 on the Welfare of Low Income Households in the Philippines: The Role of Social Protection.” COVID-19 Low Income HOPE Survey Note 1. World Bank, Washington, DC. https://documents1.worldbank.org/ curated/en/698921611118950758/pdf/Mitigating-the-Impact-of-COVID-19-on-the-Welfare-of-Low-Income- Households-in-the-Philippines-The-Role-of-Social-Protection.pdf. Cirera, Xavier., Marcio José Vargas Da Cruz, Elwyn Davies, Arti Grover, Leonardo Iacovone, José Ernesto L. Cordova, Denis Medvedev, Franklin Okechukwu Maduko, Gaurav. Nayyar, Santiago Reyes Ortega, and Jesica Torres. 2021. Policies to Support Businesses through the COVID-19 Shock: A Firm Level Perspective. The World Bank Research Observer 36(1):41–66. https://documents.worldbank.org/en/publication/documents-reports/documentdetail/188461610547807479/ policies-to-support-businesses-through-the-covid-19-shock-a-firm-level-perspective. COA (Commission on Audit). 2021a. Performance Audit Report: COVID-19 Assistance to Restart Enterprises (CARES) Program. Manila. https://www.coa.gov.ph/reports/performance-audit-reports/2021-2/covid-19-assistance-to-restart- enterprises-cares-program/. COA. 2021b. 2020 Annual Financial Report: National Government, Vol 1. Manila. https://www.coa.gov.ph/reports/annual- financial-reports/national-government-agencies. Crismundo, Kris. 2022. “SBCorp Released P5.9-B for MSME Loans.” Philippine News Agency, March 11, 2022. https:// www.pna.gov.ph/articles/1169552. DBM (Department of Budget and Management). 2021. “GAA 2020: Status of COVID-19 Releases.” Report, September 30, 2021. Manila. De Vera, Ben O. 2022. “PhilGuarantee: Credit Guarantees up in 2021 as Bank Lending Growth Resumed.” Philippine Daily Inquirer, February  28, 2022. https://business.inquirer.net/342067/philguarantee-credit-guarantees-up-in- 2021-as-bank-lending-growth-resumed. DOF (Department of Finance). 2021. “PhilGuarantee Expands Credit Guarantee Support to MSMEs, Other Key Sectors amid COVID Pandemic.” Press Release. 2/21/21. DOF, Manila. DOF. 2022. “Small Business Wage Subsidy Program.” Manila. https://sites.google.com/dof.gov.ph/small-business-wage- subsidy. DSWD-DOLE-DTI-DA-DOF-DBM-DILG (Department of Social Welfare Development; Department of Labor and Employment; Department of Trade and Industry; Department of Agriculture; Department of Finance; Department of Budget and Management; Department of Interior and Local Government). 2020. Joint Memorandum Circular No. 1. https://dbm.gov.ph/index.php/mimaropa/265-latest-issuances/joint-memorandum-circular/joint-memorandum- circular-2020/1636-joint-memorandum-circular-no-2020-001-dswd-dole-dti-da-dof-dbm-dilg. Freund, Caroline, and Alfonso Garcia Mora. 2020. “Keeping the lights on: Supporting firms and preserving jobs from crisis through recovery” Technical report, World Bank: Private Sector Development Blog. April 30, 2020. https://blogs. worldbank.org/psd/keeping-lights-supporting-firms-and-preserving-jobs-crisis-through-recovery. Gentilini, Ugo, Mohamed Almenfi, B. Alsafi, John D. Blomquist, Pamela Dale, Luciana de la Flor. Giuffra, Vyjayanti Desai, Maria Belen Fontenez, G.A. Galicia Rabadan, Veronia Lopez, A. G. Marin Espinosa, I. V. Mujica Canas, Harish Natarajan, David Newhouse, R. J., Palacios, Ana Patricia Quiroz, Claudia Rodriguez Alas, Gayatri Sabharwal, and Michael Weber. 2021. Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures. Vol. 2: Global Database on Social Protection and Jobs Responses to COVID-19. COVID-19 Living Paper, World Bank Group, Washington, DC. https://socialprotection.org/discover/publications/social-protection-and-jobs-responses- covid-19-real-time-review-country. CHAPTER 5: THE PHILIPPINES 159 ILO (International Labor Organization). 2021. COVID-19 and Employment Protection Policies: A Quantitative Analysis of the Asia-Pacific Region. ILO Research Brief. ILO, Bangkok. https://www.ilo.org/asia/publications/issue-briefs/ WCMS_810842/lang—en/index.htm. IMF (International Monetary Fund). 2021. Database of Country Fiscal Measures in Response to the COVID-19 Pandemic. Washington, DC. IOM (International Organization for Migration). 2021. COVID-19 Impact Assessment on Returned Overseas Filipino Workers. Makati City, Philippines. https://philippines.iom.int/sites/g/files/tmzbdl1651/files/documents/covid-19- impact-assessment-on-returned-overseas-filipino-workers-resized.pdf. Kang, Jong Woo, and Latoja, Ma Concepcion. 2022. COVID-19 and Overseas Filipino Workers: Return Migration and Reintegration into the Home Country – the Philippine Case. ADB Southeast Asia Working Papers. http://dx.doi.org/ 10.22617/WPS220002-2. OECD (Organization for Economic Cooperation and Development). 2019. PISA 2018 Results. Vol. I: What Students Know and Can Do. Programme for International Student Assessment. Paris: OECD Publishing. Patinio, Ferdinand. 2021. “4.7M Displaced Workers Benefit from Aid Programs: DOLE.” Philippine News Agency, December 27, 2021. https://www.pna.gov.ph/articles/1163841. Patinio, Ferdinand. 2022. “50K OFWs to Get AKAP Financial Aid Next Week.” Philippine News Agency, February 5, 2022. https://www.pna.gov.ph/articles/1167119. PSA (Philippine Statistics Authority). 2021a. “Proportion of Poor Filipinos Registered at 23.7  Percent in the First Semester of 2021.” Manila. PSA. 2021b. “Special Release: PSA Releases Annual Revisions of the National Accounts of the Philippines.” Press Release 2021–522. December  17, 2021. https://psa.gov.ph/content/proportion-poor-filipinos-registered-237- percent-first-semester-2021. PSA. 2021c. “2020 Annual Poverty Indicator Survey.” https://psa.gov.ph/sites/default/files/%5BONSrev-cleared%5D%20 2020%20APIS%20Final%20Report_rev1%20wo%20comments_ONSF3_signed.pdf. PSA. 2022. “More than 1.08 million Establishments Operated in 2021 which Generated Total Employment of 8.57 million (2021 Updating of the List of Establishments Preliminary Results).” 2022–07. https://psa.gov.ph/ content/more-108-million-establishments-operated-2021-which-generated-total-employment-857-million. Republic of the Philippines. 2020a. Bayanihan 1, Republic Act No. 11469. Republic of the Philippines. 2020b. Bayanihan 2, Republic Act No. 11494. SWS (Social Weather Stations). 2021a. “Second Quarter 2021 Social Weather Survey: Hunger eases to 13.6% of families in June 2021.” https://sws.org.ph/swsmain/artcldisppage/?artcsyscode=ART-20210728100035. SWS. 2021b. “Second Quarter 2021 Social Weather Survey: 49% of adult Filipinos got worse off in the past 12 months.” https://sws.org.ph/swsmain/artcldisppage/?artcsyscode=ART-20210724050302. SWS. 2021c. “Second Quarter 2021 Social Weather Survey: 37% of adult Filipinos say their Quality-of-Life will improve in the next 12 months.” https://sws.org.ph/swsmain/artcldisppage/?artcsyscode=ART-20210723062922. Villanueva, Joann. 2020. “Small Businesses Urged to Register with BIR.” Philippine News Agency, June 16, 2020. https:// www.pna.gov.ph/articles/1106092. World Bank. 2021. Philippines Economic Update December 2021 Edition: Regaining Lost Ground, Revitalizing the Filipino Workers. Washington, DC. https://openknowledge.worldbank.org/handle/10986/36874. CHAPTER 6 Vietnam by Sarah Hebous, Shawn W. Tan, Trang Thu Tran, Matthew Wai-Poi, and Judy Yang60 This chapter examines the effects of the COVID-19 pandemic on employment in Vietnam, as well as the policy response made to support households and firms during the crisis. It draws on five rounds of the High Frequency Phone Survey (HFPS), spanning June 2020 to March 2021, to examine the impact of the pandemic on households. To look at the impact on firms, the chapter uses four rounds of the business pulse survey, spanning June 2020 to November 2021. The BPS provides information on firms’ operational status, adjustment strategies, and receipt of fiscal support during pandemic. The household survey covers topics such as employment and income, coping strategies, and social safety nets (see Annex 6A for further details). Timeline of the COVID-19 Pandemic and Government Measures In 2020 and early 2021, Vietnam stood out as a country that had managed COVID-19 very well. It had one of the world’s lowest rates of COVID-19 infections, and the decline in its gross domestic product (GDP) was the smallest of any country in the East Asia and Pacific (EAP) region. Strict policies delayed the emergence of COVID-19 in Vietnam for about 18 months (Figure 6.1). Over the course of 2020 and the first half of 2021, cases remained low and manageable, although Vietnamese households and businesses nevertheless experienced adverse shocks to employment, income, and daily activities. The initial containment of the virus can be credited to the swift and early response by the government to stem the threat of spreading infections. Vietnam was the first country in the world to close its international borders in March 2020. It also imposed a month-long nationwide lockdown immediately afterward, in April 2020. The government closed all nonessential businesses and schools and enforced stay-at-home and work-from-home mandates for a large share of the population, introducing checkpoints to restrict travel. Strict measures led to the containment of COVID-19 in Vietnam for about a year and half, until April 2021. In the first half of 2021, the Delta variant spread quickly throughout Southeast Asia, dampening hopes of a return to normalcy. Previous measures to break the chain of infection and control local outbreaks were no longer effective, due to the high transmissibility of the Delta variant. The Delta wave also occurred during a period of low vaccination rates across the country. At the start of this wave, the share of the population that had received at least one dose was just 0.01 percent (Our World in Data). By the end 60  Anh Thi Bao and England Rhys Can provided additional input. Kyung Min Lee and Jesica Torres contributed to the framework to integrate firm level analysis with household data analysis. 161 162 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.1  Number of COVID-19 cases and deaths, inverted mobility index, and stringency index of pandemic-related measures in Vietnam, 2020–22 HFPS R1 HFPS R2 HFPS R3 HFPS R4 HFPS R5 80 125 Inverted mobility index (14-day average) 60 100 Stringency index (14-day average) Stringency index 40 75 20 50 Inverted mobility index 0 25 -20 0 BPS R1 BPS R2 BPS R3 BPS R4 300 COVID-19 cases/deaths (per 100,000 people) COVID-19 cases 225 150 75 COVID-19 deaths 0 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Apr-22 Source: Mobility data comes from Google Mobility Reports; stringency data comes from the Oxford Covid-19 Government Response Tracker (OxCGRT); COVID-19 cases and deaths are derived from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. Note: The mobility index is the average percentage change in time spent at retail and recreation locations, transit stations, parks, and workplaces relative to the pre-pandemic baseline. For each day of the week, the baseline is defined as the median value from the five-week period of January 3–February 6, 2020. The figure shows 14-day rolling averages of the mobility index, which is inverted so that higher values indicate lower mobility. The stringency index records the strictness of lockdown style policies; 14-day rolling averages are shown. COVID-19 cases and deaths are smoothed by a seven-day average. Phases of the pandemic are determined by identifying structural breaks in COVID-19 cases and the mobility index. See the introduction for further details. of July, as the outbreak intensified significantly, the share of the population that had received at least one dose rose to 5.7 percent. Renewed lockdowns and the halting of most economic activities lasted throughout the third quarter of 2021 in most parts of Vietnam. The strictest case was the two-month lockdown of Ho Chi Minh City in August and September 2021. During this time, migrant workers there faced difficult living conditions, particularly as many had lost employment and were separated from their families. Before the pandemic, the surrounding areas of Binh Duong, Dong Nai, and Long An had an estimated 3.5 million migrants (VNExpress 2021b). Because of strict travel restrictions, migrant workers were unable to return home at the height of the fourth wave. The General Statistics Office of Vietnam estimates that 450,000 workers left Ho Chi Minh City and other southern localities between July and September 2021 (VNExpress 2021a). By October 2021, as restrictions eased, Ho Chi Minh City saw the exodus of tens of thousands of migrant laborers. According to state media, 90,000 people left Ho Chi Minh City between October 1 and October 4 alone (Reuters 2021). CHAPTER 6: VIETNAM 163 Another COVID wave emerged in Vietnam in early 2022 and began to surge in February after Tet. However, the government’s response became less restrictive, most likely because of the high rate of vaccinations, low number of deaths, and the realization that a zero-COVID strategy is very costly. In March 2022, Vietnam reopened its borders to international tourism, after two years of border closures. Impacts on the Economy Vietnam performed remarkably well in 2020, achieving one of the highest GDP growth rates in the world, at 2.9 percent. Although this is about 4.2 percentage points lower than the country’s recent performance, Vietnam’s economy grew, in contrast to the world economy, which contracted by at least 4 percent. Only two other countries in East Asia—China and Myanmar—have reported positive GDP growth in 2020. The economy grew 5.6 percent in the first semester of 2021 before then slowing, as the result of a large COVID-19 outbreak caused by the Delta variant. Since early May, manufacturing and service sector activities became increasingly hamstrung by targeted lockdowns to contain community transmission of the virus. In mid-July, mobility restrictions widened, with the southern part of the country, Ho Chi Minh City, and then Hanoi, placed under strict quarantine, which slowed economic activities. At the same time, the economy faced the risk of increased competition in its external markets, as competitors who were ahead in vaccinations restarted production and recaptured some of the market shares they had lost to Vietnam in 2020. The government reacted by continuing its accommodating monetary policy and adopting a new fiscal assistance package of approximately US$1 billion in early July. It is likely that 2022 will be the year of recovery. Firm-level evidence on sales growth in early 2022 shows signs of improvements in demand, and international tourism has reopened. However, due to rising inflation and oil prices, the war in Ukraine, and the high number of COVID cases, a full rebound of tourism will likely take time. Moreover, even with an economic recovery in 2022, scars from COVID-19 will persist, as learning losses have accumulated, family businesses have closed, and many people have lost their jobs. Despite the strides made on economic outcomes, Vietnam still trails regional peers on business environment quality. Macroeconomic conditions are reasonably stable, but institutional development is still restricted in terms of systems of checks and balances, corporate governance, property rights, and corruption (WEF 2019). Domestic competition is still limited, and the country is significantly more closed to foreign trade than regional peers on average. Factor markets are relatively fluid in terms of hiring and firing employees, internal labor mobility, and credit to the private sector. However, small and medium-size enterprises (SMEs) face much tighter financing conditions. Employment Impacts: Shocks and Recovery At the outset of the pandemic, strict policies limiting mobility were the primary cause of adverse business and employment impacts. The negative economic impact of much of the earlier phase of the pandemic was induced by policy. Health impacts remained low early on, whereas the containment measures slowed growth. In April 2020, half of firms were closed during the lockdown, and average sales fell 52 percent (Figure 6.2). The situation improved in June 2020, as restrictions eased, but the vast majority of firms still 164 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.2  Changes in sales in Vietnam between 2020 and 2022 relative to same period in 2019 0 April 2020 June 2020 September–October 2020 Percent change in employment relative to baseline –10 January 2021 September–November 2021 January 2022 –16 –15 −20 −30 –28 −40 –38 –39 −50 –50 Source: World Bank data based on business pulse surveys experienced negative demand shocks. On average, sales fell almost 40 percent in June 2020 relative to the same period in the previous year. Disruptions to firm operations triggered a contraction in employment (Figure 6.3). In June 2020, firms adjusted employment along both the intensive and extensive margin: 15 percent let workers go, 14 percent put workers on leave without pay, and more than 20 percent cut either hours or wages. Net employment fell by more than 12 percent relative to pre-pandemic levels. Firms reported more severe impacts between September and November 2021 than in June 2020. Almost 60 percent of firms reported reduced operations, and 83 percent reported declines in sales relative to 2019. On average, sales fell 39 percent, back to the level in June 2020. As demand fell, the impact on employment also deepened. Between September and November 2021, the share of firms that let workers go did not increase, but hiring declined significantly (see Figure 6.5). Businesses relied more heavily on adjustments along the intensive margins; 30 percent granted unpaid leaves of absence (up from 3 percent in January 2021), 24 percent reduced wages, and 28 percent reduced hours. These results imply a major shock to labor income even for workers who remained employed. These findings are consistent with anecdotal evidence of a mass increase in early withdrawals of social insurance (VNExpress 2021c). With the easing of restrictions before the Delta wave, negative employment shocks persisted but showed signs of recovery. The share of firms that adjusted employment on the intensive margin, by reducing hours or wages, declined from 37 percent during the early days of the pandemic (June–July 2020) to 14 percent before the Delta surge (January–February 2021). The share of firms that let workers go fell from 15 percent to 8 percent. The rate of net job losses declined from 12 percent to about 10 percent in fall 2020 and early 2021. CHAPTER 6: VIETNAM 165 FIGURE 6.3  Share of firms in Vietnam making employment adjustments, June 2020–January 2022 30 28 Percent of firms making employment adjustments 24 23 21 20 20 17 15 14 13 12 11 12 12 10 9 10 8 8 8 8 7 5 5 5 3 3 2 0 Hired Fired Granted Granted Reduced Reduced workers workers leave leave wages hours of absence of absence with pay June 2020 September–October 2020 January 2021 September–November 2021 January 2022 Source: World Bank data based on business pulse surveys Delta wave–induced reversal of economic recovery The intensifying effect of the fourth wave was registered across the labor market. Labor Force Survey data suggests that the unemployment rate in the third quarter of 2021 was 4 percent, an increase of nearly 50 percent from the previous year. This increase in unemployment was the sharpest since the onset of the pandemic, as unemployment rates had remained below 3 percent until the third quarter of 2021. In the fourth quarter, the unemployment rate fell to 3.6 percent, which was still much higher than pre-pandemic levels (Figure 6.4). The increase in underemployment was larger than the increase in unemployment, indicating that job downgrading or switching was more common than job loss. Underemployment rose 1–2 percentage points across all age and gender groups in 2020; unemployment saw a much lower rise. In the third quarter of 2021, the number of employed workers willing and available for other work increased by over 700,000 people over the second quarter, to 1.8 million underemployed. The underemployment rate stood at 4.5 percent in the third quarter of 2021, more than 1.7 percentage points higher than the year before. The labor market underutilization rate61 reached 10.4 percent in the third quarter, nearly twice the rate of the previous year. Similar to other labor market indicators, underemployment and underutilization showed modest signs of recovery in the fourth quarter. In 2019, before the pandemic, the underutilization rate hovered around 4 percent.  The underutilization rate is the combined time-related underemployment and unemployment. In Vietnam, a person is underemployed if 61 he or she worked for less than 35 hours per week and was ready to work additional hours. 166 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.4  Underemployment and unemployment rates in Vietnam, by gender and age group, 2010–20 Underemployment rate Unemployment rate 15-24 6 15-24 4 Female All All 2 50+ 25-49 50+ 0 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 All 15–24 25–49 50+ Female Male Source: Business pulse survey, Vietnam General Statistics Office. As employment was more severely disrupted in 2021, household income losses increased again. During the initial stages of the pandemic, associated with stringent lockdowns, over half of all households reported lower income than the year before, with about one-fifth reporting having lost 50–99  percent of their income, and another fifth reporting losses of 25–49 percent (Figure 6.5). As restrictions eased, these rates of self-reported loss fell at the beginning of 2021 to about 30 percent of households by March. With the advent of the fourth wave and the reimposition of restrictions, the figure jumped to half of all households by December 2021, with the rates of severe loss the worst of the pandemic. On average, nominal monthly income in the third and fourth quarter of 2021 fell to the lowest levels since the start of 2020. FIGURE 6.5  Self-reported changes in household income in Vietnam, 2020 and 2021 100 80 Percent of households 60 40 20 0 Jul/Aug 2020 vs. Aug/Sept 2020 vs. Jan/Feb 2021 vs. Mar 2021 vs. Dec 2021 vs. previous year previous year previous year previous year previous year Lower (≥100%) Lower (50–99%) Lower (25–49%) Lower (<25%) Same Higher Don't know Source: World Bank data based on High-Frequency Phone Surveys. CHAPTER 6: VIETNAM 167 With the decline in infections from the Delta variant, the labor market began to recover in the fourth quarter of 2021, as restrictions eased in many areas, and the number of cases stabilized following higher vaccination rates. Workers returned to the labor force; the official labor force participation rate, published by the General Statistics Office, increased to 65.6 percent—still 3.9 percentage points lower than the same period in the previous year. Labor force participation in the Southeast and Mekong Delta regions contracted most severely in the third quarter of 2021 yet recovered the most rapidly in the following quarter. Employment levels stood at 49.0 million people in the fourth quarter, up from 47.2 million in the previous quarter. The unemployment rate dropped slightly, to 3.6 percent, but remained nearly 1 percentage point higher than in the same period in the previous year. Other indicators of underutilization improved slightly in the fourth quarter of 2021 but had yet to return to pre-pandemic levels. Nevertheless, average earnings began to increase across all sectors. Unequal impacts of COVID-19 In recent decades, inequality in Vietnam has remained low and stable, and strong growth has reduced poverty (World Bank 2022). Despite these favorable pre-pandemic conditions, several signs suggest that COVID may slow poverty reduction: ⦁ Deferral of household plans: Most households adapted throughout the pandemic, but households with more means were better able to adapt than others, and many households deferred goals and ambitions. Among households negatively affected, the poorest were the most likely to defer education and the least likely to use or adopt digital services and technologies. Some trends proliferated across regions, such as differences in educational continuity during lockdowns. Other disparate outcomes during COVID-19 built on pre-existing disparities in food, digital access, health care use, and education. Impacts on nonmonetary aspects were also greater among certain population groups, including women and ethnic minorities. ⦁ Learning losses: School closures adversely affected learning, due to limited access to distance and remedial learning. Among households with school-age children whose schools were closed, only 61 percent had access to online classes, and almost 20 percent lacked access to any distance learning opportunities between September 2020 and March 2021. In locations where online classes were not available or accessible, the Ministry of Education and Training and provincial departments of education and training arranged to broadcast learning sessions over television and/or radio. However, these mediums accounted for only a negligible portion of distance learning. Short Message Service (SMS) and paper-based self-study were the second-most popular methods; both teachers and students generally considered them ineffective. In areas with online learning, exam scores declined, reflecting learning loss (VNExpress 2022). ⦁ Income recovery: The impacts of the crisis on households and firms in Vietnam before the fourth wave have been well-documented (World Bank 2021). Some groups reported more adverse changes to income, and some groups are recovering more rapidly than others. Women, households in the bottom 20 percent of the welfare distribution before COVID-19, households without any formal channels of income, and households in regions particularly hard-hit by COVID-19 reported more protracted declines in their household incomes than other groups. As a result of strict lockdown measures, urban areas were much more affected than rural areas. In addition, large firms had better capacity to cope than smaller firms. 168 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Disproportionate effects on women The economic shock of the pandemic affected women more than men (Figure 6.6). The crisis was unlike past economic crises, because it was caused by mobility restrictions, business shutdowns, school and institutional childcare closures, and stay-at-home orders (Alon and colleagues, 2020). As a result, childcare emerged as a key concern and a requirement for economic recovery, as the COVID-19 pandemic increased demands on women’s time (UN Women 2020). Average earnings dropped significantly, and gender inequality in earnings widened. In the third quarter of 2021, the average nominal monthly income was D 5.2 million (US$229), nearly 15 percent less than in the previous quarter. Earnings recovered marginally in the fourth quarter but remained nearly 11 percent lower than the previous year (in nominal terms). The gender gap in earnings widened during the fourth wave, specifically, in the third and fourth quarters of 2021. In the third quarter, the reduction in earnings was 1.5 percent higher for female workers than for male workers; by the fourth quarter, it was twice as great. Disproportionate effects on services The services sector endured the largest shocks, whereas agriculture was the most resilient sector. Severe lockdowns during the early outbreaks resulted in the most substantial impact on formal firms in the services sector so far. Before the Delta wave, the wholesale and retail sector had the lowest sales losses outside of agriculture. Average sales losses for this sector dropped by more than 40 percent from September–November 2021, relative to 2019. Other services sectors also experienced losses of this magnitude (Figure 6.7). However, this severe demand shock appears to have been transitory. In the business pulse survey conducted in early 2022, sales growth had bounced back sharply for commerce and other services, although it was still below pre-pandemic levels. Given the declining demand in manufacturing, construction, and services, many workers transitioned into agriculture during the fourth wave. In the third quarter of 2021, agricultural employment grew by FIGURE 6.6  Percentage change in average earnings of women and men in Vietnam in third and fourth quarters of 2021 0 –2 Percentage change in earnings –4 –6 –8 –7.5 –10 –8.7 –12 –14 –12.9 –14.6 –16 Female Male Q3 2021 Q4 2021 Source: Vietnam General Statistics Office 2021. CHAPTER 6: VIETNAM 169 FIGURE 6.7  Change in sales in Vietnam, by sector, June 2020–January 2022 –10 Agriculture Manufacturing Commerce Other services Change in sales relative to 2019 (percent) –20 –30 –40 –50 June 2020 January 2021 January 2022 September–October 2020 September–November 2021 Source: World Bank data based on business pulse surveys. 673,000 workers, thanks largely to the return of many labor migrants to their hometowns. This sector employed 14.5 million people, accounting for nearly a third of total employment. Employment in manufacturing and construction fell by 6 percent relative to the previous quarter, to 15.7 million workers. Employment in services stood at 17.1 million workers, a 12 percent decline from the previous quarter. Employment levels in services, manufacturing, and construction increased in the fourth quarter, but they remained lower than the previous year. Agricultural employment fell but remained higher than the previous year. Average earnings declined across all sectors in the third quarter of 2021, with the sharpest reduction in the services sector. Average monthly income in the sector dropped 14.3 percent from the previous quarter. Earnings in the accommodations and food, transport, and storage subsectors fell by over 20 percent. For workers in manufacturing and construction, average monthly income declined 13.5 percent relative to the second quarter. Agricultural workers saw a decrease of 9.2 percent in earnings compared with the previous quarter, only the second time agricultural income had fallen quarter-over-quarter since the start of the pandemic. By the fourth quarter, average earnings started to increase across all sectors, rising most rapidly for workers in manufacturing and construction. However, earnings in all sectors, with the exception of agriculture, remained lower than in the same period of the previous year (Figure 6.8). Disproportionate effects on informal workers Informal workers experienced a slower recovery than formal sector workers, and more jobs became informal. Both formal and informal workers experienced the impact of the fourth wave, but informality was associated with slower income recovery (Figure 6.9). The lockdowns imposed in many cities both prevented 170 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.8  Average nominal monthly income of Vietnamese households in agriculture, industry, and services, fourth quarter 2020 and 2021 8 7.1 7 6.8 6.3 Average nominal monthly household 5.9 6 income (million D) 5 4 3.4 3.4 3 2 1 0 Agriculture Industry Services Q4 2020 Q4 2021 Source: Vietnam General Statistics Office 2022. many types of informal jobs from being carried out and reduced demand for them. Outside of agriculture, the number of workers holding informal jobs stood at 18 million in the third quarter of 2021, a 13 percent decrease from the same period in the previous year, and a 2.3 percentage point decline relative to the third quarter of 2020. Some of these workers were likely to have moved into agriculture, as the informal employment rate, including agriculture, decreased only marginally. The loss of informal employment exposed many workers (especially migrant workers) to significant economic hardships. Informal workers faced difficulties accessing support for various reasons, including lack of proof of previous employment or income, and ineligibility because they did not live in a registered FIGURE 6.9  Quarterly informal employment rate in Vietnam (excluding agriculture), 2020 and 2021 60% 58% 57.1% 57.4% 56.8% 56.0% Employment rate (percent) 55.8% 56% 55.3% 55.1% 54.5% 54% 52% 50% Q1 Q2 Q3 Q4 2020 2021 Source: Vietnam General Statistics Office, 2022. CHAPTER 6: VIETNAM 171 domicile. Data through March 2021 suggests that households without any formal labor market income sources experienced a slower income recovery than other workers (Figure 6.10) Newly available work seems to be concentrated in the informal sector. Informal employment (outside of agriculture) showed signs of slight recovery in the fourth quarter of 2021, when the employment rate fell to 55.1 percent, 0.9 percentage points lower than the previous year but higher than in the third quarter. Informal employment increased 7.4 percent between the third and fourth quarters; formal employment saw only a 3.9 percent increase. Continuing uncertainty about COVID-19 has made it challenging for businesses to take on new formal employees. Disproportionate effects on urban areas Because of the localized nature of the outbreak, the impact on firms was heavily concentrated in urban centers around Ho Chi Minh City. Closure rates were similar across geographical regions in the first three BPS waves; they surged to 35 percent in Ho Chi Minh City in September–November 2021. The high closure rate was driven by southern provinces that were also under lockdown. Declines in sales were steepest in Ho Chi Minh City, followed by Hanoi (Figure 6.11). Sales trends closely tracked one another in the first three survey waves. In contrast, there was a clear divergence during the Delta wave, with drops in sales averaging more than 50 percent in Ho Chi Minh City, and almost 40 percent in Hanoi. The BPS survey conducted in FIGURE 6.10  Index of household income in Vietnam by people with and without formal sources of income, June/July 2020–March 2021 100 90 Income index (100 = June 2020) 80 70 60 50 No Yes Household has at least one formal income source Jun/July 2020 Jul/Aug 2020 Sept/Oct 2020 Jan 2021 Mar 2021 95% Cl Source: World Bank data from High-Frequency Phone Surveys. Note: Data was collected on ranges of income loss (less than 25 percent, 25–49 percent, 50–99 percent, and 100 percent), not actual percentage increases. Low-impact estimates are shown and assume the largest rate of change if income is reported to be increasing, or the smallest rate of change if income is reported to be declining. Estimates assume that income levels remained constant during off-survey periods. 172 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.11  Average change in sales in Vietnam, by region, June 2020–January 2022 –10 –20 Change in sales relative to 2019 (percent) –30 –40 –50 June 2020 January 2021 January 2022 September–October 2020 September–November 2021 Hanoi Ho Chi Minh City Other regions Source: World Bank data base on business pulse surveys. early 2022 showed that recovery had taken hold, but that the sales impact was still larger in Hanoi and Ho Chi Minh City than elsewhere in the country. Self-reported labor market impacts were concentrated in the Southeast and Mekong River Delta regions during the fourth wave. Nearly 60 percent of workers in the Southeast region and 45 percent in the Mekong River Delta reported negative labor market impacts in the third quarter. The Southeast region includes Ho Chi Minh City, which experienced one of Vietnam’s largest outbreaks in the third quarter. Although the share of workers in these regions experiencing negative labor market impacts declined in the fourth quarter, they remain the most affected of all regions in Vietnam. The adverse labor market effects of the fourth wave lingered even after restrictions were relaxed, with over half of workers in the Southeast still facing labor market adjustments in the fourth quarter. Changes in average earnings also differed across regions. Average earnings in the Southeast dropped 25 percent in the third quarter compared with the same quarter in the previous year, but experienced the largest rebound in the fourth quarter. Ho Chi Minh City saw average earnings rise nearly 13 percent in the fourth quarter relative to the previous quarter, compared with just 3 percent in Hanoi. The region with the second-largest income reduction in the third quarter (12 percent relative to the same quarter in the previous year) was the Mekong Delta region. In the fourth quarter, average incomes in the Mekong Delta continued to fall. CHAPTER 6: VIETNAM 173 Opportunities presented by widespread adoption of digital technologies The pandemic may have accelerated the adoption of potentially productivity-enhancing technology and improved resource allocation. In Vietnam and globally, the increase in adoption of digital technologies was unprecedented. Between June 2020 and January 2022, the share of formal firms in Vietnam that increased their use of digital platforms increased from less than half to about three-quarters of firms, and the share of firms that invested in digital solutions increased by a factor of seven (Figure 6.12). Firms, including informal enterprises, also embraced digitalization. As use of digital technologies grew, a significant share of firms leveraged remote work arrangements to overcome movement restrictions. Non-farm family businesses reported a large share of sales through digital platforms. Evidence suggests that firms that were more productive were better able to cope, as manufacturing firms with higher pre- pandemic sales per worker suffered smaller employment losses than other firms, implying a potential “creative destruction” effect (Figure 6.13). Larger firms were more likely than smaller firms to adopt or expand the use of technologies, especially technologies for more sophisticated functions (Figure 6.14). Between June 2020 and January 2021, the share of large firms that started using or increased their use of digital platforms was 10–15 percentage points greater than that of SMEs. By January 2021, almost 90 percent of large firms and over 80 percent of SMEs had started to use or increased their use of digital platforms for front-end purposes. In contrast, the adoption gap remained wide for back-end purposes; 90 percent of large firms—but just 60 percent of SMEs—started or increased their use by January 2021. The differences are likely driven by resource and capability gaps. Widening performance gaps can increase concentration, potentially reducing competition and productivity gains in the long term. FIGURE 6.12  Adjustment mechanisms adopted by firms in Vietnam, June 2020–January 2022 80 78 75 73 62 60 47 45 Percent of firms 40 34 33 30 21 21 20 18 13 15 12 12 7 5 0 Increased use of Invested in digital Repackaged Used work-from-home digital platforms solutions product mix June 2020 September–October 2020 January 2021 September–November 2021 January 2022 Source: World Bank data based on business pulse surveys. 174 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.13  Correlation between changes in employment and labor productivity in vietnam’s manufacturing sector 20 Change in employment relative to baseline (percent) 0 –20 –40 4 5 6 7 8 9 ln(baseline labor productivity) Source: World Bank business pulse surveys. Note: Figure is a binned scatterplot. The change in employment is calculated as the difference in the share of average employment over two periods. Baseline labor productivity is 2019 sales per worker. FIGURE 6.14  Adoption of digital platforms in Vietnam, by firm size, June 2020–January 2021 a. For back-end purposes b. For front-end purposes 100 100 Percent of firms that stared/increased use of digital platforms 80 80 60 60 40 40 20 20 June 2020 September–October 2020 January 2021 June 2020 September–October 2020 January 2021 Small or medium Large Source: World Bank data based on business pulse surveys. Note: Back-end purposes include business administration, production planning, and supply chain management. Front-end purposes include marketing, sales, payment methods, and service delivery. CHAPTER 6: VIETNAM 175 Fiscal Support to Firms and Households Size and composition of the fiscal response Government support in Vietnam was more modest than any country in the region other than Papua New Guinea (Figure 6.15). Estimates by the World Bank Chief Economist’s Office suggest that the overall fiscal package in response to the COVID-19 pandemic stood at a less than 2.1 percent of GDP in 2020, and 3.4 percent of GDP in 2021. Firms received the largest share of support. Earlier analysis by the World Bank (2021) suggests that Vietnam had the second-widest gap between household losses and support in the region. Similar estimates are available from the IMF, with data as of June 2021 showing the size of Vietnam’s package relative to other countries in the region. The government implemented two major support packages at the national level, the first in the second quarter of 2020 (at the outset of the crisis) and the second in the third quarter of 2021 (after the arrival of the Delta variant). In addition, provincial governments provided their own support as needed in between periods of active national support. Table 6.1 summarizes the national measures used to support businesses and households. The mix of instruments was not particularly diverse; on the firm side, support consisted of fiscal support, support for fees and other fixed costs, and credit. Support was more easily accessible by, or available only to, formal firms. For households, support in the form of cash transfers was limited. Coverage and timing For firms, coverage was initially low, though it significantly increased as the crisis deepened during the Delta wave (Figure 6.16). More than half of formal firms reported having benefitted from government policies by November 2021, and close to 70 percent had benefited by early 2022, up from less than 20 percent in FIGURE 6.15  Government spending and forgone fiscal revenue in response to COVID-19 in select countries in Asia 25 23.1 Government spending and forgoing fiscal revenues as percent of GDP 20 15.6 14.6 14.6 15 10 8.8 6.1 6.4 5.4 5 3.6 2.2 1.0 0 Papua New Guinea Vietnam Philippines Indonesia China Cambodia Malaysia South Korea Mongolia Thailand Singapore Source: IMF (International Monetary Fund) 2021. Note: Figures reflect data as of June 2021. 176 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA TABLE 6.1  National policy instruments Vietnam used to support firms and households during the pandemic Target group Instruments Formal businesses • Fiscal support (tax deferment, reduction in corporate income tax and labor contribution) • Support for fixed costs (such as land rent) • Credit support (for example, loans to pay employee salaries, loans to SMEs) Households • Cash support • Electricity subsidies • Free food • Unemployment benefits (second package) June 2020. Household businesses also experienced an increase in support and a change in conditionality, having received a lump sum of D 3 million instead of D 1 million per month for a maximum of three months. Tax deferrals and reductions drove the increase in access to support for formal firms. The mix of instruments remained skewed toward fiscal support, with very limited access to credit (Figure 6.17). Between early and early 2022, the share of firms with access to tax deferrals or corporate income tax reductions increased by more than 150 percent, from about 17 percent to about 42 percent. Access to credit remained low, with only 7.5 percent of firms having access to lower interest loans, and 5 percent having access to wage subsidies (loans for employee salaries) by November 2021. For both households and individuals, the first relief package provided benefits to all existing beneficiaries of government social assistance programs. The first package provided aid to about 1.2 million workers, with an average benefit of D 1 million (Table 6.2). In the second package, the state spent D 14,902 billion, FIGURE 6.16  Access to support policies in Vietnam, June 2020–January 2022 80 68 60 57 Percent of firms 40 36 29 19 20 0 June 2020 September–October 2020 January 2021 September–November 2021 January 2022 Source: World Bank data based on business pulse surveys. CHAPTER 6: VIETNAM 177 FIGURE 6.17  Composition of support to firms in Vietnam, January 2021–January 2022 17.6 Corporate income tax reduction 34.1 41.8 17.2 Tax deferral 34.0 39.1 4.4 Lower-interest rate loans (SMEs) 5.3 7.5 0.3 Payment susp. (retirem fund) 5.0 7.3 3.8 Land rental payment deferral 4.5 7.1 1.0 Trade union fee deferral 3.5 7.0 0.7 Loans for employee salary 1.8 5.2 1.4 Reduction of land rents 2.7 5.1 2.6 Other 2.9 4.0 0.2 Simplified customs procedures 2.0 2.6 0.7 Tax dere al (dom. auto industry) 1.2 1.2 0.1 Reduction of road user fees 0.8 1.2 0.2 Industry fee reduction 0.2 0.2 0 10 20 30 40 Percent of firms January 2021 September–November 2021 January 2022 Source: World Bank data from business pulse surveys. Note: Industry fee reduction refers to reduction of fees for registration for use of foreign barcodes and industry property fees. a. With respect to support to households, in the first round of relief, existing beneficiaries of the social protection system and select other groups received top-ups in the form of cash to their regular benefits. Implementation went smoothly, as it was universally applied to all existing beneficiaries, as disbursement channels were already in place. TABLE 6.2  Social assistance vietnam provided in first COVID-19 relief package (percent of households) 2020 February–June/July 2020 Received support Received cash Received Classified Received support from any national support for Applied for assistance from as poor in for purchasing or international existing vulnerable new COVID-19 new COVID-19 Group commune health insurance organization groups relief programs relief programs All 6.6 38.3 7.7 19.8 10.2 1.2 Urban 3.7 31.5 4.6 14.2 13.7 1.8 Rural 8.1 41.8 9.4 22.7 8.4 0.9 Top 60 2.4 31.7 4.7 13.1 11.4 1.4 Bottom 40 13.3 48.7 12.5 30.4 8.3 0.9 Kinh majority 4.8 33.3 6.9 18.5 10.8 1.3 Ethnic minority 16.8 66.1 12.5 26.7 6.8 0.7 Source: World Bank data from the High-Frequency Phone Survey. Note: Existing targeted social assistance programs included cash support for poor and near-poor households, social assistance beneficiaries, and “national devotees” (people who contributed during the “revolution and war times”). 178 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA supporting 11.25 million individuals with an average benefit of about D 1.3 million. The second package reached more workers than the first package, but cash support amounts were similar in value. The second package added and removed beneficiaries. Beneficiaries of existing social assistance programs and “national devotees” (people who contributed during what has been referred to as the “revolution and war times”) were removed from the second allocation. New groups—including contracted employees taken into quarantine, children infected or quarantined, and employees in certain occupations— received cash support. The second package also provided support, including a food allowance, for people infected with or quarantined due to COVID. These changes reduced aid to the preexisting vulnerable populations and expanded assistance to formal workers without a permanent contract (who are less likely to be poor). Targeting Disproportionate Benefits to the formal sector Evidence suggests that support to firms did not initially reach the most vulnerable firms but that targeting improved slightly over time. Access to support in the early months (June–July 2020) was not correlated with the probability that firms experienced a drop in sales. By November 2021, firms that received support experienced, on average, an 8  percentage point larger drop in sales than firms that did not. However, with increased access in early 2022, targeting appears to have worsened again with firms that received support experiencing similar sales losses on average to those that did not (Figure 6.18). During the first three survey rounds, large firms were most likely to receive support. By November 2021, medium-size firms were as likely as large firms to receive support (65 percent of both size firms), though FIGURE 6.18  Average change in sales by Vietnamese firms that did and did not receive public support, June 2020–January 2022 0 Average percentage change in sales in 30 days –10 –15 –15 –16 –17 before interview –20 –27 –30 –32 –37 –36 –40 –39 –42 June 2020 January 2021 January 2022 September–October 2020 September–November 2021 Did not receive public support Received public support Source: World Bank data based on business pulse surveys. CHAPTER 6: VIETNAM 179 access by small firms remained lower at 55 percent. However, the gap in access widened again in early 2022 (Figure 6.19). Results from the BPS suggest that the higher a manufacturing firm’s sales per worker at the outset, the more likely it was to have access to government assistance during the pandemic (Figure 6.20). This relationship suggests that support may have gone to more productive firms, helping to improve static efficiency. The result partly reflects the fact that most access has been in the form of income tax deferrals and reductions—instruments that would benefit only firms that earned profits. The government lacks the tools to support informal businesses, and it has had difficulty reaching informal workers. Due to policy instruments that rely largely on fiscal exemptions, firm support is not able to reach informal businesses. At the same time, social assistance appears not to have reached workers in informal businesses. Support for informal workers who lost their jobs was implemented on a relatively small scale and for only a short period early in the pandemic. It is estimated to have had only minimal welfare impacts. According to World Bank monitoring surveys conducted in December 2021, formal workers who applied for government cash support were nearly twice as likely to receive assistance than informal workers in 2021. Effectiveness Throughout the pandemic, the public has been supportive of the government’s COVID policies and responses, although perceptions of policy relief to households have typically been less positive than perceptions of FIGURE 6.19  Percent of firms receiving COVID–related government support in Vietnam, by firm size, June 2020–January 2022 80 60 Percent of firms (%) 40 20 June 2020 September– January 2021 September– January 2022 October 2020 November 2021 Average ln(baseline labor productivity) Small Medium-size Large Source: World Bank data based on business pulse surveys. 180 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 6.20  Correlation between probability of receiving support and labor productivity in Vietnam 70 Average probability of receiving public support 60 50 40 30 20 4 5 6 7 8 9 Average ln(baseline labor productivity) Source: World Bank data based on business pulse surveys. Note: Figure is a binned scatterplot of manufacturing sector only, controlling for firm size. Baseline labor productivity is 2019 sales per worker. general pandemic management or vaccination (Figure 6.21). In 2021, for example, 94 percent of households were satisfied with the vaccination response, but just 68 percent were satisfied with the relief response. Public approval was stronger during the first round of policies than the second. Business closures, quarantines, vaccination, lockdowns, limited domestic and international travel, and prolonged school closures were considered useful. In contrast, the public was dissatisfied with the firm and household relief packages, including cash support, subsidies, and tax deferrals. FIGURE 6.21  Public perceptions of government response to COVID-19 in Vietnam 2021 Response post-Delta (Q4 2021) Relief Vaccination Lockdown 2020 COVID-19 relief (Q3 2020) Food is easily accessible and a ordable Enough is being done to help those who have lost their jobs or had to close their business The D 62 trillion program reached the most in need 0 20 40 60 80 100 Percent of households satisfied with government response Source: World Bank Data based on High-Frequency Phone Surveys. Note: Fieldwork for round 3 of the HFPS occurred during the entire month of September 2020. The share of respondents that did not feel comfortable traveling domestically ranged from almost 30 percent in week 1 of fieldwork near the Da Nang outbreak to 10 percent by the end of the month. Round 6 was conducted during a surge in COVID-19 cases. As responses can change very quickly, results should be interpreted with caution CHAPTER 6: VIETNAM 181 Implementation challenges resulted in poor perceptions of relief policies. Interviews were conducted to learn about implementation challenges during the first round of national relief (World Bank 2021). The challenges summarized in this section come from interviews in Da Nang with local officials and representatives of the Department of Labor, Invalids, and Social Affairs. Da Nang is a high-capacity city that was able to implement and target relatively well, as it supplemented the national relief packages with its own relief package. Lower-capacity provinces very likely experienced greater challenges than described here. Da Nang faced challenges verifying the eligibility of applicants in the new target groups. Employers had to verify that employees had been laid off. Obtaining this verification was sometimes difficult, as some employers did not assist with applications or provide the necessary documentation. Eligibility was based on the location of residence rather than work, putting migrant workers at a disadvantage. Informality posed additional challenges to verification. In most cases, informal workers without a labor contract could not verify their employment, whereas household businesses were required to confirm their taxable turnover. If they did not keep receipts, they had to prove turnover through the Tax Division. In the first round, eligibility criteria for new target groups was complicated, and occupations approved for relief were overly specific. For example, employees working in enterprises and educational institutions had to have lost their jobs and social insurance on or after April 1, 2020. Many establishments had already stopped operating before that date, and other businesses tried to maintain social insurance for employees after it. As a result, many groups of employees were adversely affected yet ineligible. Required documentation was cumbersome and unclear for applicants, discouraging many from applying. Challenges were expected, because residents were unfamiliar with new registration and verification processes. Many who submitted applications lacked supporting documentation, especially paperwork for identification, such as an ID card or household registration. Confusion and long wait times also led some households to reapply while they waited for confirmation on their application status, causing duplications. The inability to quickly verify and enroll applicants foreshadows larger social assistance challenges in Vietnam unless it modernizes. The large volume of applications to process and check overwhelmed commune/ward and village-level officials, many of whom were already engaged in carrying out tasks to contain the spread of COVID-19. Their responsibilities included communicating policy, guiding residents on registration and collecting supporting documents, checking applications, making lists of beneficiaries, organizing review and assessment, submitting completed applications to the district level, notifying applicants of results, and resolving feedback and complaints from people and communities. Access to business relief policies improved over time but faced similar challenges. Initially, a major concern was the lack of communication to potential beneficiaries: 34 percent of firms were not aware of support policies, and 25 percent considered them too difficult to apply for. In contrast, by late 2021, only 9 percent of firms that did not receive support cited lack of awareness as the reason. As more firms needed support, complaints of difficult procedures also increased: In November 2021, a quarter of firms without access to government support cited difficulty applying as the main reason for not doing so. The most important reason why firms lacked access was ineligibility, suggesting overly stringent criteria. To improve implementation, firms cited simplifying and improving eligibility conditions as the most important changes, after reducing processing time, providing more information, and even increasing the amount of support. 182 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Lessons Learned Initially, Vietnam stood out as a country that managed the pandemic health risks extremely well and preserved economic growth. Yet, the emergence of the Delta and Omicron variants changed its performance. Two years into the pandemic, as other countries are returning to normal, Vietnam, as well as other Asian countries, is experiencing record-setting caseloads. Fortunately, the new variants are less lethal, and the population is highly vaccinated. There may still be longer-term household and business scarring, as learning losses accumulate, many workers have downgraded or lost their jobs, and some family businesses have permanently closed. Vietnam’s experience shows that monetary and nonmonetary gaps caused by COVID-19 can widen even in a country that managed the pandemic extremely well at the outset. Existing disparities have widened; left unchecked, they will likely lead to increased inequality and slower growth. Lost education is unlikely to be recovered, with consequences for lifetime wages; sold assets cannot produce future income. Employment scarring is also associated with lower lifetime earnings. Reducing future disparities will require forward-looking policies and better support systems. The uneven recovery from COVID-19 and the heightened risks from new geopolitical shocks call for stronger actions. Vietnam needs to modernize its social assistance system to prepare for the next crisis. Without a social registry, its options in the short term are either to do relatively little to respond to new households in need, or to cover much of the population somewhat indiscriminately. Crises will emerge that will require a more sophisticated and digital response. Guarding against risks is essential to prevent households from falling back into poverty. Developing an inclusive and responsive social protection system is at the core of this objective. COVID-19 has highlighted challenges in delivering assistance to new vulnerable groups; without modernization, the same implementation challenges experienced during COVID-19 will mar the response in future crises. Digitization and modernization can reduce staff burden and errors, speed disbursements, and alleviate capacity constraints. The Philippines and Thailand were much harder hit by COVID than Vietnam, but they were able to respond quickly and widely because of their strong safety net systems, preparedness, and investments in data. Policy instruments to support businesses may need to move away from liquidity relief to measures aimed at increasing the productivity and resilience of the private sector. The BPS conducted in February 2021 found that 64 percent of firms cited simplification of administrative procedures as the most important reform. Improving the business environment remains a key priority for the private sector in Vietnam. Global evidence suggests that economies respond more sluggishly to crises in the presence of regulatory barriers to business entry and expansion (Barrero, Bloom, and Davis 2020). Given the severe demand shocks during this crisis and new opportunities brought about from US–China trade tension and implementation of free trade agreements, such as the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CTPPT) and the Regional Comprehensive Economic Partnership (RCEP), it is also important to consider policies that support firms’ access to new markets and customers. Most support measures in Vietnam continue to focus on resolving the short-term liquidity crisis. More advanced economies in the region, such as Malaysia, have already implemented a range of other policies to help firms increase their capabilities and improve matching between firms and workers. They include measures such as subsidies for skills training, hiring incentives, and technology adoption. New support packages will need to consider similar measures to prepare firms in the recovery phase and make Vietnam more resilient to future shocks. CHAPTER 6: VIETNAM 183 Annex 6 Annex 6A.  The Household and Firm Surveys in Vietnam The analysis from this chapter draws on five rounds of the High Frequency Phone Survey (HFPS) in Vietnam (Table 6A.1). The HFPS was implemented by the World Bank with the objective to collect household data to monitor and assess the socioeconomic impacts of the COVID-19 pandemic on households in Vietnam. The HFPS used a nationally representative household survey from 2018 as the sampling frame, and sampling weights were constructed to ensure unbiased estimates from the sample. The HFPS in Vietnam covered topics such as health, employment, coping strategies, safety nets, and food insecurity, among others. This chapter also draws on four rounds of the business pulse survey (BPS), conducted by the General Statistics Office of Vietnam (GSO) and the World Bank (Table 6A.2). The BPS aimed to monitor the impact of COVID-19 on the private sector in Vietnam. Firms across different provinces were interviewed by phone between June 2020 and November 2021, and each round is representative at three different firm size categories and four broad sectors (agriculture, manufacturing, wholesale and retail, and other services). TABLE 6A.1  Dates and sample sizes of high-frequency phone surveys of households in Vietnam Round Survey dates Sample size 1 June 5–July 8, 2020 6,213 2 July 27–August 12, 2020 3,935 3 September 9–October 1, 2020 4,560 4 January 2–15, 2021 3,945 5 March 13–31, 2021 3,922 Source: Original text for this publication. TABLE 6A.2  Dates and sample sizes of business pulse surveys in Vietnam Round Survey dates Sample size 1 June 12–July 7, 2020 499 2 September 6–October 26, 2020 501 3 January 27–March 24, 2021 489 4 September 5–November 12, 2021 458 Source: Original text for this publication. Note: The business pulse survey sample includes firms that were formally registered with a government authority at the time of the survey. 184 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA References Alon, Titan M., Matthias Doepkes, Jane Olmstead-Rumsey, and Michele Tertilt. 2020. “The Impact of COVID-19 on Gender Equality.” NBERS Working Paper W26947, National Bureau of Economic Research, Cambridge, MA. Barrero, Jose Maria, Nicholas Bloom, and Steven J. Davis. 2020. “COVID-19 Is Also a Reallocation Shock.” NBER Working Paper 27137, National Bureau of Economic Research, Cambridge, MA. IMF (International Monetary Fund). 2021. Fiscal Monitor Database of Country Fiscal Measures in Response to the COVID-19 Pandemic, Washington DC. https://www.imf.org/en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response- to-COVID-1. Lakner, Christoph, Daniel Gerszon Mahler, Mario Negre, and Espen Beer. 2020. “How Much Does Reducing Inequality Matter for Global Poverty?” Global Poverty Monitoring Technical Note 13, World Bank, Washington, DC. Reuters. 2021. “‘We Are Tired’:Workers Flee Vietnam’s Largest City as Long Lockdown Eases.” October  4, 2021. https://www.reuters.com/world/the-great-reboot/we-are-tired-workers-flee-vietnams-largest-city-long-lockdown- eases-2021-10-04/. UN (United Nations) Women. 2020. Whose Time to Care? Unpaid Care and Domestic Work During COVID-19. UN Women, New York. https://data.unwomen.org/sites/default/files/inline-files/Whose-time-to-care-brief_0.pdf. Vietnam General Statistics Office. 2021a. Report on the COVID-19 Impacts on Labour and Employment Situation in the Third Quarter of 2021. Hanoi. https://www.gso.gov.vn/en/data-and-statistics/2021/10/report-on-the-covid-19-impacts- on-labour-and-employment-situation-in-the-third-quarter-of-2021/. VNExpress. 2021a. “HCMC Industrial Workforce Shrinks by Half after Covid Lockdown Goes.” October 5, 2021. https://e.vnexpress.net/news/news/hcmc-industrial-workforce-shrinks-by-half-after-covid-lockdown-goes- 4367289.html. VNExpress. 2021b. “1.3 Mln Migrants Return Home During Pandemic.” October 1, 2021. https://e.vnexpress.net/news/ news/1-3-mln-migrants-return-home-during-pandemic-4370813.html. VNExpress. 2021c. “More than 700,000 People Withdraw Social Insurance Once.” November  21, 2021. https:// vnexpress.net/hon-700-000-nguoi-rut-bao-hiem-xa-hoi-mot-lan-4391910.html. VNExpress. 2022. “Knowledge Gap after Studying Online.” January 26, 2022. https://vnexpress.net/lo-hong-kien-thuc- sau-thoi-gian-hoc-online-4421338.html. WEF (World Economic Forum). 2019. The Global Competitiveness Report, 2019. https://www3.weforum.org/docs/WEF_TheGlobalCompetitivenessReport2019.pdf. World Bank. 2021. A Year Deferred: Early Experiences and Lessons from COVID-19 in Vietnam. World Bank, Washington, DC. https://documents1.worldbank.org/curated/en/826921633442977133/pdf/A-Year-Deferred-Early-Experiences- and-Lessons-from-Covid-19-in-Vietnam.pdf. CHAPTER 7 Lessons Learned by Pablo Fleiss Weinberger, Alvaro Gonzalez, Lydia Kim, and Maria Ana Lugo The country chapters included in this report reveal a sobering picture of the impact of the pandemic during its first two years. The pandemic reversed welfare gains achieved over 30 or more years in the region, and its after-effects will likely linger for years to come. The number of poor in the region (excluding China) is estimated to have increased for the first time since the early 2000s, and inequality is expected to increase in several of the countries in the region (World Bank, 2022a). The experiences of the six countries studied show that the effect of the pandemic on households depended partly on their pre-pandemic characteristics. The ability of households to recover their pre- pandemic economic position—which most have yet to do—depends on their initial conditions, including their sector of employment, where they live, and the amount of human and physical capital they have. Recovery has been uneven across households, and worrisome signs are emerging that income or consumption inequality increased. Less educated workers, women, and people who were poor before the pandemic were more likely than other workers to experience work stoppages or to return to work when conditions improved. There is suggestive evidence that, at least in some countries, the employment recovery came together with increased levels of informality (see chapter 4 for more information on Malaysia and chapter 5 for the Philippines). This pattern, if it indeed exists, may stem from the reallocation of workers into lower-paid, lower-productivity sectors, which may have a long-term impact on the ability of economies to recover in an inclusive manner. The pandemic’s impacts on human capital are also of concern. COVID-19-related school closings and the move to distance or hybrid learning modalities have not only led to widespread learning losses but have also exacerbated inequities in accessing quality education (World Bank 2021a). Poorer children have had greater difficulty engaging in online or other interactive forms of learning during school closures due to limited access to technology. Extensive learning losses are also expected to have an adverse effect on the future earnings of current students,62 particularly those of poorer students for whom learning losses are likely to be greater. The severity of the impact of the COVID-19 recession depended partly on firm-level characteristics. The economic slowdown hit smaller firms and firms in particular sectors, such as tourism and retail, hardest. Firms that were less able to adopt digital technologies—because of financial, technical, or business constraints—did not adjust as well to the shock. 62  World Bank 2021b estimates that learning losses accrued over the pandemic in the EAP could lead to a 3.8 percent average reduction in future earnings every year. 185 186 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Some signs of recovery have emerged. In the countries studied in this report, four out of five firms that closed temporarily during the early months of the pandemic are now operating again. Although firm sales were still 25 percent below pre-pandemic levels by mid-2021,63 the decline represents a significant improvement over the 45 percent decline in the spring of 2020. Despite disruptions in global production networks, exporters are recovering more rapidly than other firms, although their sales also remain below pre-pandemic levels. Government support to households and firms was critical to cushion the negative impacts of the pandemic. In Cambodia, cash transfers provided valuable income support to poor households, which, according to recent estimates, mitigated 40 percent of the increase in poverty and inequality (World Bank 2022b). Similarly, simulations for the Philippines and Mongolia suggest that poverty would have increased significantly more in the absence of the emergency assistance implemented (see chapter 6 for a discussion of the Philippines. See also World Bank 2021b for more information on Mongolia). Instead, in countries where household support was small, so were the potential impacts on poverty reduction (World Bank 2021c). The question going forward is how government policies can support growth and put households, individuals, and firms on a sustainable path to recovery. As mobility restrictions wind down, economic activity begins to pick up, and countries embark on a new normal, that path may require a look back before moving forward. Although a comprehensive and detailed assessment of what worked and what did not in supporting workers, households, and firms to recover from the pandemic-induced recession is unlikely at this stage, a few lessons have emerged. This chapter brings together the analysis presented in each of the six-country chapters by providing a comparative perspective of the scale and composition of government responses and reflecting on lessons learned from these experiences. It provides answers to the following questions: ⦁ Why did some countries spend more on COVID-19 relief than others? ⦁ How different was the balance between supporting firms and households across the six countries, and what factors may explain these differences? ⦁ Was the support to firms and households timely, well-targeted, and adequate? What are the emerging lessons from these experiences? Where should governments focus their attention as they emerge from the economic shock of the pandemic? How Did Governments Respond to the COVID-19 Economic Slowdown? In March 2020, immediately after COVID-19 expanded throughout the world, almost all countries implemented economic packages aimed at addressing the economic slowdown (Gourinchas 2020). Policymakers were concerned that without proper macroeconomic support, countries could fall into a long, severe, and costly economic recession. Consequently, governments and central banks responded to the pandemic and the economic crisis, using both fiscal and monetary tools on an unprecedented scale (Benmelech  There is no sales loss data for Cambodia. 63 CHAPTER 7: LESSONS LEARNED 187 FIGURE 7.1  Correlation between per capita GDP and size of announced fiscal spending in 2020 and 2021 50 Italy Japan 45 Mauritius Germany 40 Fiscal spending as percent of GDP 35 United Kingdom 30 Mongolia United States 25 20 R2 = 0.31 Average of high- 15 income countries Malaysia Cambodia Philippines 10 Indonesia Vietnam Average of non-high- 5 income countries 0 6 7 8 9 10 11 12 GDP per capita in 2020 (Current US$) in Log. Source: IMF Database of Country Fiscal Measures in Response to the COVID-19 Pandemic, World Bank staff estimates for the six East Asia and Pacific countries in red, and World Development Indicators for GDP per capita in 2020. and Tzur-Ilan 2020). Globally, countries spent almost US$18 trillion on fiscal measures in response to the pandemic, which amounted to about 17.1 percent of the world’s GDP (IMF 2021).64 The breadth and scope of these fiscal measures varied widely across countries. Although the type of policy tools used was broadly similar, there were stark differences in the magnitude and composition of fiscal stimulus programs, determined by a myriad of political, financial, social, and economic factors (Aizenman and colleagues, 2021). The analysis that follows draws from the literature on the economic consequences of COVID-19 to answer questions about which factors helped determine the fiscal responses across countries and how the fiscal responses of the six countries studied in this report compared with responses elsewhere. It examines income levels, fiscal space, and the structure of employed and unemployed workers to determine their relationship to the size and composition of the fiscal support program. Countries’ ability to respond depends on their level of development. Several studies find that the size of fiscal stimuli is positively correlated with countries’ income levels. On average, high-income countries implemented substantially larger fiscal stimuli than other countries, even when controlling for the number of COVID-19 cases (Aizenman and colleagues, 2021; Alberola and colleagues, 2021; Benmelech and Tzur-Ilan 2020; Felipe and colleagues, 2020) (Figure 7.1). In the average high-income country (defined as a country with per capita gross national income of at least US$12,696), the size of the package was almost three times larger, as a percentage of GDP, than it was in non-high-income countries. For the six countries studied, data on the fiscal response in 2020 and 2021 is based on estimates from the World Bank’s Office of the Chief Economist for East Asia and the Pacific.65 The advantage of using 64  https://www.imf.org/en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response-to-COVID-19 65  Estimates are based on the latest government announcements on intended spending, as of February 2022, in response to the pandemic- induced recession. Where disbursement data was not available, planned budget figures were used. 188 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA these estimates, rather than estimates by the International Monetary Fund (IMF), is that they disaggregate support to firms and households. The analysis shows that five of the six East Asian countries (all but Vietnam) announced spending packages that were larger, as a share of GDP, than predicted based on their income levels. However, the size of the fiscal packages was only weakly related to these countries’ level of development. Mongolia and Indonesia, for example, have similar GDP levels, but Mongolia’s spending package was three times larger than Indonesia’s and larger than that of much wealthier Malaysia. Cambodia is the poorest country in the group, but it approved a larger package than Indonesia or Vietnam (Figure 7.1). Countries’ fiscal space—the room they have for temporarily increasing their budget deficits without jeopardizing their access to markets or the sustainability of their debt—well as their ability to access credit markets may help determine the size of COVID-19-related support packages. Worldwide, Benmelech and Tzur-Ilan (2020) find that of all of the macro variables included in their empirical models, a country’s credit rating is the only one that is consistently associated with the size of fiscal policy during the early stages of the COVID-19 crisis (World Bank 2022c). Other papers (Apeti and colleagues, 2021; Lee, 2020) find that public debt levels, as a share of GDP, were not correlated with the size of fiscal stimuli. Given the initial shock and the expectation that the crisis would be temporary, ignoring debt levels in responding to the crisis made sense, as argued at the time by major multilateral global financial institutions (Aizenman and colleagues, 2021, pp. 15–16). Across the six countries studied, fiscal space was not a strong determinant of the size of the fiscal response. Table 7.1 displays three dimensions of fiscal space: public debt, the fiscal deficit, and market access (represented by two variables, credit ratings, and sovereign yields). Mongolia holds the highest level of government debt and borrowing costs, but it also has the largest fiscal package. It also has the largest fiscal surplus. Malaysia has more debt than Vietnam, a similar fiscal deficit, and higher borrowing costs, but fiscal programs in Malaysia were more than twice as large as in Vietnam (as a percent of GDP). Revenues from taxes and social contributions appear to be more related to the size of the fiscal response in several EAP countries. In three of the six countries studied—Indonesia, Malaysia, and Mongolia—the levels of revenue collection in 2019 were surprisingly similar to the size of the fiscal response in 2020 and TABLE 7.1  Fiscal space variables, 2019 Total revenue from Government Government Sovereign Yield social contributions COVID-19-related fiscal debt (percent overall balance debt rating (percentage and taxes (percent support in 2020 and Country of GDP) (percent of GDP) Moody’s/S&P points) of GDP) 2021 (percent of GDP) Mongolia 66.1 0.9 B3/B 4.4 28.8 30.6 Malaysia 57.2 −3.2 A2/A 3.1 11.9 12.4 Indonesia 30.4 −2.2 Baa2/BBB 3.6 9.8 9.7 Philippines 38.6 −1.9 Baa2/BBB+ 2.7 14.5 9.3 Cambodia 28.5 0.4 B2 — 20.0 9.3 Vietnam 42.9 −3.3 Ba3/BB 2.6 14.6 5.5 Sources: Data from the IMF Fiscal Monitor Database; JPMorgan; ICTD/UNU-WIDER Government Revenue Dataset. Note: — = Not available. CHAPTER 7: LESSONS LEARNED 189 TABLE 7.2  Structural indicators in the labor market and dependency ratios, 2019 Employment (percent of total) COVID-19-related Share of fiscal support in Wage and Informal population 2020 and 2021 Country salaried workers Self-employment employment 65 and over (percent of GDP) Mongolia 51.5 48.5 44.1 4.2 30.6 Malaysia 72.6 27.4 — 6.9 12.4 Indonesia 48.3 51.8 80.5 6.1 9.7 Philippines 63.8 36.2 — 5.3 9.3 Cambodia 53.0 47.0 89.4 4.7 9.3 Vietnam 45.7 54.3 69.7 7.6 5.5 Sources: Employment data is from the International Labor Organization (ILO); other data are from the World Bank’s World Development Indicators. Vulnerable employment includes family workers and own-account workers. Informal employment includes own-account workers of informal sector enterprises or for own consumption, contributing family workers, and employees holding informal jobs. Note: — = Not available. 2021. In contrast, Vietnam’s fiscal package was a third of the 2019 revenue collection, Cambodia’s was a half, and the Philippines’ package was 50 percent lower than the 2019 collection. The size of the fiscal package may be associated with the degree of the country’s informality. Worldwide— and in all the countries studied in this report except Mongolia—the size of fiscal packages is positively correlated with the share of salaried and wage workers to total employment, and inversely correlated with self-employment, where levels of informality are typically high (Felipe and colleagues, 2020; World Bank, 2022c). Once Mongolia is excluded, the correlation between the three first variables in Table 7.2 and the size of the fiscal packages is statistically significant. This correlation may be related to governments’ ability to disburse aid through established systems. The informal sector, workers, and firms are difficult to reach since they are not registered in most government systems. The composition of spending packages is also influenced by countries’ income levels. Hosny (2021) finds that high-income countries were more inclined to propose “below-the-line” measures (equity injections, asset purchases, loans), which do not have an immediate impact on the fiscal budget. 66 In contrast, non-high-income countries predominantly increased health-related fiscal spending, such as investments in public health infrastructure and pandemic preparedness, to compensate for their weaker initial health infrastructure and preparedness levels. For high-income countries, the spending allocated to above-the-line measures and to liquidity support was similar (Figure 7.2). In contrast, non- high-income countries favored “above-the-line” (budgetary) measures by a ratio of 2-to-1. Even so, as a share of GDP, high-income countries spent more than twice as much as non-high-income countries on above-the-line measures. 66  On-budget “above-the-line” measures include additional spending (on health, unemployment benefits, transfers) or forgone revenues (tax cuts and credits) provided through standard budget channels. Liquidity support measures include off-budget “below-the-line” measures (equity injections, asset purchases, loans, including through extra-budgetary funds) and contingent liabilities, including government guarantees and other quasi-fiscal operations (Hosny 2021). 190 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 7.2  Fiscal spending on COVID-19 relief to households in six countries in East Asia in 2020 and 2021 14 Revenue measures to households Income support to households 12 19 Percent of GDP 8 6 4 2 0 Mongolia Indonesia Cambodia Malaysia Philippines Vietnam Source: World Bank staff estimates based on the latest data from governments as of February 2022. Note: Where disbursement data was not available, planned budget data was used. Figures show total spending in 2020 and 2021. Spending in 2022 is not included. Consistent with these findings, four of the six countries studied spent more on above-the-line measures than on liquidity support (the exceptions were Malaysia and the Philippines). On average, the magnitude of the total spending announcements and the relation between both components were in line with the world average. As expected, liquidity support in Malaysia, an upper-middle-income country and the wealthiest of the six, was the largest of the six countries. Mongolia allocated almost four times as much to above-the- line measures as it did to liquidity support. Yet, it is important to note that in both Malaysia and Mongolia (unlike in the Philippines), most of the liquidity support (4.9 and 4.1 percent of GDP, respectively) was in the form of contingent liabilities. Contingent liabilities are largely aimed at providing liquidity support to state- owned enterprises and covering local government liabilities rather than helping the liquidity constraints of private sector firms. Therefore, in the section that follows on support to workers and households, these will be excluded from the analysis on government liquidity support to firms. In comparison, below-the-line measures—which will be included in the analysis that follows—include government guarantees granted to banks, firms, or households. These are also contingent liabilities as well, but their aim is to alleviate the liquidity constraints of private sector firms and are therefore relevant to the analysis below. Government support to workers and households Support to households relied on the systems and infrastructure in place. The effectiveness, efficiency, and timeliness of government support measures to households depended on several factors, including the existence of well-functioning social protection programs, up-to-date national identification systems, and robust delivery systems before the pandemic. Countries with well-established social protection systems had greater success in delivering timely assistance to intended beneficiaries, even if not all labor income losses CHAPTER 7: LESSONS LEARNED 191 were offset, especially among non-poor, yet vulnerable households. Given mounting fiscal pressures, better targeting assistance to populations in greatest need will be important in all countries. Size and coverage of household support The size of COVID-19 relief measures for households varied considerably. In all six countries, income support—as opposed to revenue measures—made up the bulk of assistance to households (Figure 7.2). Mongolia stood out as a clear outlier, with spending in 2020–2021 amounting to 12.5 percent of GDP, a level that was relatively high for the extensiveness of labor income shocks in the country (Figure 7.3). The solid foundations of the Child Money Program (CMP), Mongolia’s flagship quasi-universal cash transfer program, and well-established systems of delivery meant that the government was able to quickly disburse relief assistance to households by topping up monthly transfers. Although Mongolia used a diverse mix of policy instruments to support households, cash transfers made up more than half of all spending on households. Beneficiaries consequently received the largest average cash transfers among the six countries. The CMP benefit, which has been provided monthly to households since mid-2020, is equivalent to about a third of monthly pre-pandemic household income, and a universal transfer distributed in April 2021 was greater than the monthly income of more than half of all households. Despite a low number of COVID-19 FIGURE 7.3  Spending on income support to households and income losses in 2020 and 2021 12 Lower losses, higher spending Higher losses & spending Mongolia Spending on household income support (% of GDP) 9 6 Indonesia 3 Cambodia Malaysia Lower losses & spending Higher losses, lower spending Philippines Vietnam 0 40 50 60 70 Income loss relative to pre-pandemic (% of households) Source: Data from the high-frequency phone surveys. Note: The threshold distinguishing low-high spending is set at 2 percent, representing the average global COVID-19-related social protection spending (Gentilini 2022). The threshold for income losses is set at 54 percent, the average income losses observed in High-Frequency Phone Surveys across EAP countries. Labor income losses are defined as a reduction in wages, earnings from agriculture, or non-farm business income relative to before the pandemic. Reductions in total household income are used for Malaysia and Vietnam because of data limitations. 192 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA cases initially and mild impacts on employment and household incomes compared with other countries in the region, Mongolia mounted a generous response early on (Figure 7.3). The move is consistent with its recent history of providing large transfers to households, especially during economic downturns (World Bank 2022d). Other countries with emerging social protection systems, such as Indonesia and Malaysia, also deployed a diverse set of policy tools to provide relief to households during the pandemic. However, overall spending on household support was significantly lower than in Mongolia, at 3.9 percent in Indonesia and 2.9 percent in Malaysia. Support to households was comparatively high in Malaysia, given the share of households that experienced income shocks; both the incidence of household labor income losses and spending on household support were significant in Indonesia (Figure 7.4). Like Mongolia, Indonesia was able to mount a quick initial response by horizontally and vertically expanding existing social assistance programs that targeted the poor and vulnerable. Those expansions were followed by additional cash transfers, electricity subsidies, and revenue measures that aimed to fill coverage gaps and reach households affected by the pandemic that were not poor. Electricity subsidies had the broadest coverage, reaching almost 60 percent of all households. In Malaysia, which also expanded coverage through a mix of existing and new programs, the availability of comprehensive, reliable information systems and advanced digital infrastructure was crucial in enabling a rapid and far-reaching response. Spending on relief measures for households in the Philippines and countries with nascent or less developed social protection systems—namely, Cambodia and Vietnam—was limited. Vietnam and the Philippines recorded the lowest household income support spending as a share of GDP among the six countries (Figure 7.3). Considering the severity of the economic and employment impacts that the Philippines faced in 2020 and the less-than-full recovery in 2021, spending was particularly modest (Figure 7.3). Despite high FIGURE 7.4  Share of households that received government assistance for COVID-19 relief in six countries in East Asia 100 80 Percent of households 60 40 20 0 Mongolia Malaysia Indonesia Philippines Cambodia Vietnam May-June 2020 January-March 2021 July-September 2020 April-June 2021 October-December 2020 October-December 2021 Source: Data from high-frequency phone surveys. Note: Round 4 (April–June 2021) was used for Mongolia; round 3 (October–December 2020) was used for Cambodia; round 2 (July–September 2020) was used for Vietnam. CHAPTER 7: LESSONS LEARNED 193 coverage in 2020 (Figure 7.4), total benefits under the Bayanihan 1 program, which were provided for two months in the summer of 2020, amounted to about a fifth of pre-COVID household monthly income. Under the subsequent Bayanihan 2 program, coverage was reduced significantly, as were direct cash transfers to households. In Vietnam, the impacts of the recession on the economy were modest, but household income losses— albeit less prevalent than in other countries—were considerable. Household income losses affected 45 percent of households. Income losses were also associated with stringent mobility restrictions, which had particularly severe effects on migrant and informal workers. Coverage of assistance was limited because eligibility criteria were narrow, the institutions to verify eligibility were weak, and benefit levels were low. In addition, in Vietnam and the Philippines, cash assistance to households and workers was short-lived, and the lack of well-established identification and/or delivery systems led to implementation challenges that prevented the timely provision of transfers. Spending in Cambodia was higher than in the Philippines or Vietnam, but it was primarily through the existing IDPoor program, which provides cash transfers to registered poor and vulnerable households. As a result, coverage was low, with about a quarter of households receiving government support (Figure 7.3). These cash transfers in Cambodia—although longer-lived than in Vietnam or the Philippines—were small, making up less than 10 percent of household monthly income pre-COVID.67 Although they covered about half of the people in the poorest quintile, the relative loss of income was the largest among this consumption quintile. Cambodia swiftly delivered cash support to beneficiaries thanks to a well-established network of payment service providers. Phasing-out of support In 2021, many countries scaled back government support to households, despite the onset of new and more severe surges and ongoing lockdown measures. Because of mounting fiscal pressures, Indonesia and the Philippines cut back spending on social assistance in 2021, and Indonesia and the Philippines reduced coverage by more than 20 percentage points (Figure 7.4).68 Many households in both countries continued to struggle to regain their pre-pandemic economic status in 2021, with more than half of households in the two countries reporting lower levels of labor income in 2021 than before the pandemic. In Malaysia, about two-thirds of households that experienced economic shocks reported having insufficient financial resources to cover their monthly basic needs in late 2021. Cambodia and Vietnam, which faced major COVID-19 outbreaks later in 2021, maintained spending in 2021, but coverage remained narrow and declined in Vietnam. Mongolia was the only country of the six that appreciably increased spending on household support in 2021—most notably through a one-time universal cash transfer—and continued to provide widespread and generous cash transfers to households in 2022. Despite the generosity of cash transfers to households, in 2021, more than 70 percent of recipient households in Mongolia reported that the transfers were not sufficient to fully mitigate the impacts of the pandemic, while at the same time, precautionary savings shot up. Although fiscal burdens are mounting in many countries, scaling back government assistance despite continuing widespread income losses runs the risk of increasing poverty or curtailing efforts to reduce poverty.  This estimate is based on average household disposable income from the 2019/20 Cambodia Socio-Economic Survey. 67  Malaysia also scaled back spending in 2021. However, as there were no high-frequency phone household surveys in 2020, it is not possible 68 to determine whether the observed coverage in 2021 was also lower than in 2020, as it was in Indonesia and the Philippines. 194 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA Targeting Assistance to households was generally pro-poor. Shocks to household income and livelihoods are more likely to be debilitating for the poor and vulnerable, who have fewer coping strategies at their disposal (such as relying on savings) than wealthier people and may need to rely on strategies that can be scarring (such as selling productive assets or decreasing essential food expenditures). Targeting people at the bottom of the welfare distribution helps ensure that poverty and inequality do not increase. Among the six countries studied, strong reliance on existing social protection programs that target the poor and vulnerable has meant that COVID-19-related social assistance to households in the region has generally been pro-poor (Figure 7.5). Cambodia had success in efficiently targeting poorer households thanks to its reliance on the IDPoor program. Still, its narrow coverage meant that two in three households in the bottom 40 received no assistance (Figure 7.5. panel a). At the other end of the spectrum, coverage was widespread in the Philippines, but the bottom 40 was marginally more likely to receive government support than the top 60 as a result of targeting inefficiencies in the Bayanihan 1 program. Despite better targeting of poorer households, coverage of the top 60 was also high in Mongolia, Malaysia, and Indonesia, because of the provision of utility subsidies, which complemented social assistance programs. FIGURE 7.5  Percent of households that received social assistance in the past 30 days in six countries in East Asia b. By whether household experienced labor a. By pre-pandemic household welfare class income loss relative to pre-pandemic 100 100 80 80 Percent of households Percent of households 60 60 40 40 20 20 0 0 Mongolia Malaysia* Philippines Indonesia Cambodia Vietnam Mongolia Malaysia Philippines Indonesia Cambodia Vietnam Bottom 40 Top 60 Labor income loss No labor income loss Source: Data from High-Frequency Phone Surveys. Note: In panel a, self-reported income brackets are used instead of the welfare distribution to define the bottom 40 and top 60 for Malaysia. The bottom 40 for Malaysia thus represents households with a monthly income of RM 0–4,000 before the pandemic, and the top 60 represents households with income greater than RM 4,000. In panel b, labor income losses are defined as having experienced a reduction in wage, agriculture, or non-farm business income relative to before the pandemic or having experienced a work stoppage. Averages are across rounds for each country. Error bars represent 95% confidence intervals. Because of data limitations, total household income is used for Malaysia in round 2 and for Vietnam in all rounds. CHAPTER 7: LESSONS LEARNED 195 In Indonesia, for instance, electricity subsidies had the largest coverage of all programs deployed. Although they were progressive (reaching two-thirds of the bottom 40 households), they also benefited a third of the top 20 households. Pandemic-induced income losses and work stoppages were extensive across the welfare distribution, resulting in significant churning across the distribution. In many countries, workers that were hardest hit by the pandemic (workers in the manufacturing, retail, and tourism sectors) were not necessarily those who were typically poor, such as agricultural workers. As a result, social assistance programs in some countries in the region did not do a good job of targeting households that faced labor income losses as a result of the pandemic (Figure 7.5, panel b). Some countries also encountered a “missing middle” challenge. Households in the middle of the distribution were not adequately supported by social assistance programs targeted to households at the bottom of the distribution, or by social insurance programs or support to formal firms that cater to formal workers at the top end of the distribution. Many of these households are engaged in non-farm businesses, which sustained heavy losses as a result of lockdown measures but missed out on government support measures to firms because of their informality or size. In Cambodia, for instance, households with non- farm businesses—nearly two-thirds of which are in the middle 60—were not eligible for tax exemptions, one of the main support measures for registered firms. Cambodian households with non-farm businesses were 26 percent more likely than households without businesses to experience labor income losses yet were 53 percent less likely to receive any type of social assistance. Other countries had greater success in reaching middle-class households and informal household enterprises. In Indonesia, where 69 percent of employment in 2019 was in non-farm household businesses, and over 90 percent of private sector enterprises are not legally registered, a mix of household support and firm support was used to reach the middle class and informal household businesses. Households with non-farm businesses benefitted from electricity subsidies targeted toward households as well as from firm support measures such as loan deferments and cash transfers for microenterprises. Lessons learned from support to households What worked? Many countries in the region mounted impressive and unprecedented emergency responses to help households and workers cope during the pandemic. The East Asia and Pacific region recorded the highest mean country coverage rates of cash transfers globally: across EAP countries, on average, the share of people that were direct beneficiaries of cash transfer programs was 50 percent (Gentilini 2022), which is higher than the rates found in other regions, including Eastern Europe and Latin America. Coverage rates were particularly high among high-income countries, reaching, on average, three-quarters of the population in Japan or Korea, for instance, but the average rate was also above those found in other regions when considering only developing (non-high income) countries (Figure 7.6). At the same time, the region presents the second lowest mean benefit amount relative to the countries’ median income. Coverage, more than adequacy, seems to have been the focus of the assistance to households. This support was provided to both protect households and support consumption as a contributor to welfare and growth. The dominance of cash transfers over in-kind assistance suggests that countries in the region did not suffer from supply shortages or market breakdowns in food, despite the lockdowns. 196 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA FIGURE 7.6  Coverage and adequacy of cash transfers by region a. Share of population covered b. Average benefit relative to median income 80 90 Share of population that is direct beneficiary (%) Share of population that is direct beneficiary (%) 70 80 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 EAP NAC LAC MNA ECA SAR AFR AFR SAR ECA NAC MNA EAP LAC Coverage High income Non-high income Adequacy High income Non-high income Source: Figure based on Gentilini, 2022. Countries with more established social protection programs and delivery systems, such as Mongolia, Malaysia, Indonesia, and the Philippines, were quick to supplement benefits and expand coverage. Countries with more nascent social protection systems, such as Cambodia, were also able to quickly deploy additional assistance to the chronically poor, thanks to a preexisting system of beneficiaries; however, increasing coverage proved to be a challenge. In many countries, governments used a diverse mix of policy tools to reach affected households beyond those traditionally targeted by social protection programs. Evidence from several countries suggests that households that received assistance were less likely to experience food insecurity than households that did not. Cho and colleagues, (2021) found that among low-income households in the Philippines, beneficiaries of the 4P program (the country’s flagship cash transfer program)—who were more likely to receive emergency support through the Social Amelioration Program (SAP)—were better able to cope with food insecurity than non-4P recipients. In Mongolia, the universal cash transfer distributed in 2021 helped households, particularly poorer households, to avert severe food insecurity (see chapter 4). In some cases, the scale-up of social assistance also included innovations in delivery systems, which will have long-lasting effects on their effectiveness. The successful implementation of new programs and expansions in social insurance to reach informal workers in countries like Malaysia, Indonesia, and Mongolia illustrated that such programs could be deployed in the future during times of crisis. Malaysia’s enrollment mechanisms, which benefitted from advanced information technology (IT) systems and digital infrastructure, allowed it to rapidly roll out emergency cash assistance. The use of digital solutions to deliver assistance can improve the speed, transparency, and accountability of cash transfers more broadly in the future. What did not work? The pandemic exposed the limits of the social protection systems in place. While the scale and timeliness of government support to households in the region were unprecedented, gaps in coverage were evident CHAPTER 7: LESSONS LEARNED 197 in some countries, and the duration and size of government support were often not sufficient to offset persisting labor income losses. Many countries with limited social assistance programs struggled to respond quickly and efficiently to the pandemic. Implementation challenges were widespread in the Philippines and Vietnam, delaying the timely provision of benefits and revealing the shortcomings of national systems of delivery and identification. Countries without up-to-date or comprehensive social registries and national identification systems, such as Cambodia or Vietnam, were not able to effectively identify affected workers and households; therefore, coverage remained narrow. In these countries, reaching non-poor, yet vulnerable households and informal workers—populations typically unaddressed by social protection programs—also proved to be a challenge. Updating social registries amid a crisis proved problematic in countries such as Indonesia, particularly with mobility restrictions in place. The expansion of existing social assistance programs was largely pro- poor, but even countries with strong social safety nets were unable to identify households that were more likely to experience income losses or most vulnerable to the pandemic-induced economic shock. The combination of high levels of informality and narrow social programs meant that a large share of households engaged in informal businesses that suffered sizeable losses, were not covered by either social assistance, social insurance, or support to formal firms. Moreover, assistance to households and workers was short-lived, and plans for extensions were often unclear. In many countries, social assistance was rolled back in 2021, despite persisting household labor income losses. Consistent with global trends, social assistance, particularly cash transfers to households, was concentrated in the early months of the pandemic. As mentioned, while coverage of households in the region was the highest globally, benefit levels as a share of median income were among the lowest (Gentilini 2022), suggesting that large scale-ups in assistance may not have been adequate to mitigate household income losses. Public work programs could have played a more prominent role during the pandemic. Cash-for-work can assist informal workers in times of crisis (Packard and Weber 2020). However, these programs were often suspended in the context of social distance protocols. But public works can be helpful if adapted to the situation. In 2020, several African countries adapted their programs to support digital infrastructure by supporting the geolocalization of slums and identifying solid waste areas from aerial imagery (Deparday and colleagues, 2022). Digital public works can also include services such as digitizing or classifying documents for public records and tracking contacts, among others (Weber 2020). Within the countries included in this report, three of the six had public works programs as part of their emergency package, i.e., Cambodia, Indonesia, and the Philippines. In the Philippines, out-of-work individuals could be employed for 10 days to assist with disinfection and sanitation. Yet, as described in the respective country chapters, typically, cash-for-work schemes were limited, intermittent, and provided relatively low benefits. What should governments focus on as economies emerge from the economic shock of the pandemic? One clear message that emerges from the country chapters is the need to strengthen social protection systems. Doing so may require that some countries commit to increasing fiscal spending relative to what they had before the pandemic struck. Strengthening and rethinking unemployment insurance and social insurance more broadly, as was done in Malaysia and Mongolia, in the context of high informality will also 198 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA be important to protect the unemployed and fill in gaps in the coverage of precarious workers (Gentilini 2022). Malaysia provided incentives to firms and the self-employed for firms to hire new formal workers and matched social contributions for self-employed and gig workers. Mongolia gave exemptions from social security contributions and waivers for late fees, lowering the cost of formality on a temporary basis. Improved systems of providing unemployment insurance will strengthen the establishment of automatic stabilizers for future crisis situations.69 Narrow or static targeting, following Cambodia’s example, will invariably result in the exclusion of non-poor, yet vulnerable people that may be in need in times of severe shock. Developing identification systems or up-to-date social registries that allow dynamic targeting mechanisms of those in greatest need of social assistance can help to reach these populations. Ideally, registries are updated yearly; in practice, many countries update them every three to five years. (O’Keefe, forthcoming; World Bank 2021a; World Bank 2022c). This requires combining multiple registration methods, including self-registration and community targeting, and interoperability with other administrative databases, such as national ID, personal income tax, and utilities, as Malaysia did. Monitoring and evaluation are also central. Preparing for the next crisis means not only strengthening data and monitoring systems for the delivery of assistance but also allowing for the evaluation of the effectiveness and impacts of the assistance provided. Some of the questions posed here in terms of what worked, where, and how efficient the different solutions were will be better answered in time, if adequate information is collected. Some instruments were effective in reaching households during the crisis but need to be scaled back. Beyond the impact they have on fiscal accounts (especially in the context of increasing fuel prices), utility subsidies tend to be regressive, and they create incentives that are inconsistent with countries’ efforts to meet their national climate development targets. Energy subsidies could be removed gradually, starting with those which benefit the non-poor, and safety nets expanded and strengthened to provide targeted assistance to people in need (Amaglobeli and colleagues). Mongolia’s generous support to households proved essential to help households cope with work stoppages and lower earnings. However, the program is fiscally unsustainable, and a more progressive design of the benefit structure could help target the support to the most vulnerable households. Government support to firms Governments quickly implemented fiscal measures to help firms. Much of the support was directed at helping enterprises—especially small and medium-sized enterprises (SMEs) in many cases—address the liquidity shortfalls they faced or expected to face. Measures included (a) encouraging and supporting the efforts of financial institutions by extending repayment obligations and offering additional credit, often through credit guarantees; and (b) providing wage subsidies and tax, utility, debt, and rent payment deferments. The fiscal packages launched by governments in response to the pandemic included many components. Government support for firms included above-the-line measures, such as increases in government expenditures and reductions in tax revenues, and below-the-line measures, including loans, asset purchases, equity stakes,  In February 2022, the Indonesian government announced the initial implementation of its unemployment insurance program. Although it 69 has a narrow base because of high levels of informality, it could become an important automatic stabilizer for future crises. CHAPTER 7: LESSONS LEARNED 199 FIGURE 7.7  COVID-19-related government support for firms in 2020 and 2021 by type of assistance Percent of 2020 and 2021 GDP 8 6 5.8 4 0.1 1.8 2.7 2 1.1 1.2 0.6 3.0 2.4 0.7 1.6 0.6 1.2 0.9 0 0.3 Cambodia Indonesia Vietnam Malaysia Mongolia Philippines Below-the-line measures (S8) Revenue measures to firms (S6) Spending on income support to firms (S3) Source: World Bank staff estimates based on the latest data from governments as of February 2022. Note: Where disbursement data was not available, planned budget data was used. Figures show total spending in 2020 and 2021. Spending in 2022 is not included. and guarantees to firms and banks. The economic impact of these measures depends on the extent to which they are taken up and how the funds are spent. For the purposes of the present analysis, fiscal support in the form of above-the-line measures was divided into two groups: spending on income support to firms and spending on revenue measures to firms. Below-the-line measures were not disaggregated. Note that ‘below-the-line measures’ exclude contingent liabilities from Figure 7.2 these are not aimed at relieving the liquidity constraints of private sector firms but instead typically of state-owned enterprises and local governments. Additionally, the 2022 data is not yet complete and therefore was not used for this analysis. The six East Asian countries varied greatly in the magnitude and composition of public sector support to firms. The Philippines announced spending of 7.4 percent of GDP in 2020 and 2021 to support firms, including 0.9 percent of GDP in direct income support and 0.7 percent in revenue measures (Figure 7.7).70 One way to understand the factors that influenced the composition of packages is to suppose that governments responded to firms’ biggest challenges. The need for working capital was one clear, apparent, and immediate challenge. The slump in cash inflows coupled with unchanging demands for cash outflows (wages, rent, utilities, taxes, and loan payments) led to short-term working capital shortages. Indonesia, Malaysia, Mongolia, and the Philippines addressed working capital constraints with direct transfers to firms intended to keep businesses temporarily solvent. These transfers were not sizable in comparison to other measures. All six countries, except for Malaysia, spent less on direct cash transfers to firms than on below-the- line measures. Below-the-line measures provide liquidity at subsidized conditions to cover banks’ increased risk of lending. Governments may have perceived that the effects of high uncertainty about the course of 70  Public investment, other above-the-line measures, and contingent liabilities were excluded, because it was not possible to confirm across all countries that these measures were used exclusively to support firms. 200 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA the pandemic threatened a collapse in asset values that could lead to a sharp deterioration in financing conditions. Uncertainty over whether the financial system would be ready to provide firms with funding through credit lines or loans was also likely high. There was a clear risk that firms that had run into liquidity problems could see them turn into solvency problems. Governments responded with measures that included targeted temporary loan guarantees and other below-the-line measures. Most governments introduced temporary credit guarantees to ensure that banks could provide the liquidity firms needed during the crisis. What governments perceived to be the greatest constraints to firms factored into what they put into their support packages to firms, but demand for these packages also probably mattered. Differences in demand for policy support may be one reason these support packages varied across countries. Demand for tax reductions and deferments and increased liquidity are correlated with higher per capita income, driven by a higher degree of formalization and larger average firm size. In economies with many informal and small firms, demand for cash transfers is more likely to be favored because this kind of support better serves these types of firms, many of which do not pay taxes and are much less likely to have access to credit. Business demand for additional lending depended on whether economic prospects remained uncertain. If it was, it is unlikely that firms would have wanted to incur additional debt unless it could have been obtained on greatly reduced terms. Mongolia offered more sizeable support programs in these two areas; however, eligible firms were formal and larger firms. That meant that nearly 15 percent of firms, which are informal and account for 9.2–15.7 percent of GDP (National Statistics Office 2020), and 77 percent of all registered firms, which are small or microenterprises (ADB 2020), were uncovered by government support. The Philippines may have spent too much on below-the-line measures and other ways to alleviate liquidity constraints. Direct transfers to firms and households may have been more effective, given its large informal sector and the predominance of small and micro firms.71 Vietnam’s use of revenue measures to support firms may have left out firms that do not pay government fees or taxes because they are informal. Providing informal sector entrepreneurs and their workers with household-targeted cash transfers may have been a more effective strategy. As the predominance of small firms is highly correlated with per capita income, Cambodia and Vietnam (the lower-income countries in the group) could have been expected to provide direct support to firms. In fact, neither did so. Coverage and targeting Policy coverage was limited and varied across the six economies. Of all firms, less than half received any type of public support. This share ranged from 1 in 5 firms in Cambodia and the Philippines to 9 in 10 firms in Malaysia (Table 7.3). Across the world, lower-income countries were less successful in extending support to 71  It is difficult to obtain figures on the number or proportion of firms that are informal. It is stipulated that these figures are positively correlated with the proportion of output produced by the informal firms. If that is the case, a compiled database by Elgin, C., M. A. Kose, F. Ohnsorge, and S. Yu., “Understanding Informality.” (2021), is used to proxy the proportion of informal firms that exist across the six East Asian countries. Using a Dynamic General Equilibrium model, Elgin and colleagues (2021) estimate the amount of output that is produced by the informal sector (as a percent of official GDP). According to this metric for the latest year in which all six countries are included in the model, the Philippines is one of the most informal economies of the six, with 34.8 of GDP produced by the informal sector, second only to Cambodia, with 37.4 percent. The rest of the country estimates in descending order are as follows: Malaysia (26.9 percent), Indonesia (15.5 percent), Mongolia (15.4 percent) and Vietnam (11.9 percent). Readers may note that in some cases, the ranking based on this measure of output does not always coincide with the size of informal employment measures (see table I.1 in the introductory chapter,). Differences may be attributed to the fact that output-based measures consider the intensity of employment and productivity, while informal employment indicates participation. CHAPTER 7: LESSONS LEARNED 201 TABLE 7.3  Percent of surveyed firms reached by government support, by survey wave Country May–Aug 2020 Sep–Dec 2020 Jan–Apr 2021 May–Oct 2021 Sep–Dec 2021 Average Malaysia — 91 89 86 — 88 Mongolia 47 — 66 — — 56 Indonesia 7 47 61 — — 38 Vietnam 19 29 42 — 60 38 Cambodia — — 22 — — 22 Philippines 24 16 — 24 21 — — — — Average 24 46 56 55 60 46 Source: World Bank data from business pulse surveys. Note: — = Not available. firms than higher-income countries. Among the six countries studied, Malaysia and Mongolia, the wealthiest countries in the group, performed best in reaching firms. The share of firms receiving support ranged from 17 percent for micro firms in the Philippines to 94 percent for large firms in Malaysia. Malaysia provided the best coverage of firms, reaching 88 percent of firms surveyed. Cambodia and the Philippines were the least successful, covering 22 percent and 21 percent of surveyed firms, respectively. In nearly all countries, smaller firms were less likely than larger firms to receive support. The businesses most affected by the shock—smaller firms—therefore appear to have been the least likely to receive government support (Table 7.4). Most governments targeted the tourism sector, and all six countries had special support programs for SMEs. All but Mongolia provided government support to the tourism industry, including airlines in Cambodia, the Philippines, and Vietnam. Mongolia extended government support to the agricultural and TABLE 7.4  Percent of surveyed firms reached by government support in six countries in East Asia, by firm size Country Micro Small Medium Large Average Malaysia 71 86 94 94 88 Mongolia — 59 51 54 56 Indonesia 33 39 58 74 38 Vietnam 38 36 43 47 38 Cambodia 10 15 31 30 22 Philippines 17 27 31 43 21 Average 32 42 60 75 46 Source: Data from business pulse surveys. Note: — = Not available. 202 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA livestock (cashmere) sectors. Given the loss of its largest export market, China, Mongolia also provided support to exporters. Cambodia and Vietnam supported exporters in the garment, textile, footwear, and handbag industries. Vietnam extended support to the agricultural, forestry, fishery product processing, aquatic products, and automobile industries (Table 7.5). Examining the relationship between the likelihood of receiving government support and the extent of sales losses is one way to assess the efficacy of sector-level targeting. If sector-level targeting was effective, firms that experienced more severe sales losses in targeted sectors should have received government support. The effectiveness of targeting firms most affected by the crisis varied significantly. Within sectors, however, funds did not go to the most affected firms in all countries. In Mongolia, the Philippines, and Vietnam, firms that suffered greater losses in sales were more likely to receive government support. In contrast, in Indonesia and Malaysia, firms with smaller losses in sales were more likely to have done so. Phasing of support In addition to the size, composition, and targeting of government support packages, the phasing of support was important. The initial phase of support was determined by what governments perceived to be the most difficult challenges firms faced in the first stages of the pandemic. In the initial phase of lockdown, direct income support to businesses in the most affected sectors is likely to have been most effective. As the pandemic-induced recession continued, measures to keep TABLE 7.5  Announcements of support to sectors by governments in six countries in East Asia Country Measure announced Cambodia • Reductions in taxes and fees for hotels, guesthouses, restaurants, travel agencies, airlines, garment/footwear, and bag manufacturing sectors Indonesia • Spending on income support to the tourism and creative industries sectors Malaysia • Subsidized financing for tourism sector, including taxis • Direct income support to registered tour agencies and the creative arts industry • Income support to shopping malls, travel agencies, convention centers, theme parks, and hotel operators in the form of a 10 percent discount on electricity bills Mongolia • Subsidized long-term financing for non-mining exports, including businesses in the manufacturing, cashmere, and services sectors, and entities with more than 200 employees in the trade sector • Income support to the agricultural sector through discounts on agricultural equipment Philippines • Subsidized long-term financing for the agricultural and fisheries sectors and the tourism industry • Waiving of fees for domestic air carriers Vietnam • Loan guarantees and equity support to the airline industry • Discounts on fees for take-off and landing and flight control services for domestic flights • Preferential import and export tariffs for businesses in the footwear, textiles, agricultural, forestry and fishery product processing, aquatic products, mechanics, supporting industries, and automobile industry • Reduced regulations, charges, and fees in the construction, travel, and water resources sectors • Reduction in registration fee for domestically manufactured/assembled cars until end of 2020 Source: ADB (Asian Development Bank) 2022 Note: Table excluded government support to the financial and health sectors. CHAPTER 7: LESSONS LEARNED 203 firms solvent (through interventions in the financial system) may have made sense. Financial sector policy measures focused on providing liquidity to financial institutions and markets, for example, lowering reserve requirements, purchasing financial assets; and maintaining operational and business continuity, such as granting extensions of deadlines on supervisory reporting. Policy measures also focused on facilitating the flow of credit and supporting borrowers that faced short-term repayment difficulties either directly (by lowering interest rates, introducing debt repayment moratoria, facilitating loan restructuring, and offering government guarantees and loans to affected sectors) or by providing regulatory relief, encouraging banks to use available capital and liquidity buffers, and allowing for the flexible treatment of non-performing loans and asset classification. With the data available, it is difficult to assess the effectiveness of each of the disparate financial sector interventions. As the paragraph above indicates, there were a series of interventions implemented through the financial sector. However, the data used to understand the effectiveness of financial sector interventions is aggregated under the broader category of liquidity support, making it difficult to pinpoint the effectiveness of each financial sector instrument. For a more complete analysis of financial sector support in East Asia and around the world, where all financial instruments are discussed, see Feyen and colleagues, (2020), There is no clear sequencing of firm support in the six countries studied (Figure 7.8). In 2020, four of the six countries announced spending on direct income support to firms. In 2021, two countries, Malaysia and the Philippines, reduced or eliminated this type of support (In the Philippines, the government stopped providing firms with income support after 2020). In contrast, Indonesia and Mongolia modestly increased direct income support in 2021. In Indonesia and Vietnam, liquidity support increased slightly in 2021. Mongolia was the only country to stop using below-the-line measures in 2021. After spending 1.8 percent of its GDP on government support to firms, Cambodia eliminated below-the-line measures in 2021. Hence, FIGURE 7.8  COVID-19-related government support for firms in six countries in East Asia in 2020 and 2021, by category of support 4 Percent of 2020 and 2021 GDP 3 3.0 2.7 0.8 2 2.8 0.1 0.6 1.2 1 0.5 0.2 1.9 1.0 1.4 1.6 0.4 1.0 0.8 0.9 0.7 0.6 0.4 0.6 0.5 0 0.1 0.2 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 Vietnam Indonesia Cambodia Malaysia Philippines Mongolia Spending on income support to firms (S3) Revenue measures to firms (S6) Below-the-line measures (S8) Source: World Bank staff estimates based on the latest data from governments as of February 2022. Note: Authors’ estimates based on the latest announcement from the government as of February 2022. Where disbursement data was not available, planned budget data was used. 204 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA there does not seem to be a pattern in how spending was phased in. Although most countries provided less support to firms in 2021 than in 2020, provision of income support first and liquidity later is not apparent. Lessons learned from support to firms What worked? Digital solutions helped firms across all six economies to weather the economic shock of the pandemic. Firms either intensified the use of digital technologies or adopted them for the first time. This strategy helped these firms keep revenue streams flowing and manage suppliers and budgets, but it meant that firms that were less technologically savvy fell farther behind. The East Asian economies that prepared themselves to leverage digital technologies before the pandemic performed relatively well compared to other emerging economies. Economies with strong information and communications technology (ICT) infrastructure also fared better during the pandemic than economies with weak ICT infrastructure. Among the six countries studied, Malaysia has the largest share of the population with access to the Internet, followed by Vietnam and Mongolia. At the other end of the spectrum are Cambodia, the Philippines, and Indonesia, where digital solutions were out of reach for most people because of inadequate infrastructure, financing, or human capital. Addressing these constraints to digital adoption would help firms to better weather difficult times. The success of firms that adopted digital solutions suggests that many more firms would have benefited from these solutions if access to credit to finance their adoption had been more accessible. In addition, the relatively successful recoveries in Vietnam and Malaysia suggest that diversified trade is important. Efforts to provide income support directly to firms and liquidity to financial markets may have paid off. The combination of a sudden drop in revenues coupled with continued financial commitments presents increased solvency risk. A wave of corporate insolvencies would have had serious implications for growth and financial stability. Firms in economies with higher incomes and smaller informal sectors were better served by support instruments that were less appropriate for firms operating in poorer economies with larger informal sectors. Below-the-line measures seemed to be more effective instruments of firm support where firm formalization is higher, and firms are bigger. The reach of this type of support seemed to be more extensive. In comparison, in economies with many more informal and smaller firms, direct income support, in the form of cash transfers, reached more firms. This kind of support better serves these types of firms, many of which do not pay taxes and are much less likely to have access to credit. Wage subsidies were successful in mitigating liquidity constraints, reducing layoffs, and thus upholding the employee-employer relationship during the crisis. The Philippines’ small business wage subsidy was effective in stemming job losses when strict lockdown measures were in place (ILO 2021). These subsidies cannot be maintained indefinitely; moreover, unless they are directed to specific groups or sectors, they may be regressive if used without complementary measures. What did not work? The six country experiences reveal significant gaps in the reach of policy support to businesses. Smaller firms in sectors such as hospitality faced some of the largest declines in sales—and the most limited access to policy support. Although governments recognized the need to target assistance to the most affected sectors and CHAPTER 7: LESSONS LEARNED 205 smaller firms, they failed to reach most of these firms. In Cambodia, the Philippines, and Vietnam, most smaller firms were left out of government support. Given the need to act quickly in the early stages of the pandemic, governments tended to adopt a targeting strategy that minimized exclusion rather than carefully targeting beneficiaries, which would have taken time. As a result, many firms that did not experience severe adverse effects of the COVID-19 shock received government support—and many firms that suffered badly did not. While recognizing that targeting was generally ineffective in reaching the smallest firms, the effort to reach firms by sectors most negatively affected by the pandemic-induced recession was a good second- best solution. In the case of the six EAP countries, the firms most hurt by the pandemic did have a higher chance of receiving government support when support targeted specific sectors. What should governments focus on as economies emerge from the economic shock of the pandemic? It is important to identify which economies have well-formulated and established procedures to address insolvencies in an orderly fashion. Indonesia and Malaysia have established special, out-of-court debt restructuring mechanisms to tackle corporate distress. They are the most effective solution to bridge the gap between court capacity to handle cases and the number of firms that need financial restructuring (Garrido 2012). In contrast, Cambodia and Mongolia have minimal insolvency practices.72 This recession did not result in a massive wave of enterprise bankruptcies, at least in part because of the support governments provided to businesses. Governments swiftly served as financiers of last resort through large financial support measures, such as loan and guarantee programs and equity injections in firms. These measures prevented bankruptcies and attenuated the recession by increasing firms’ liquidity, reducing risk premiums, and boosting confidence (IMF, 2022).73 At the same time, these same support measures may have kept some businesses alive artificially; continued government support may risk keeping firms that should have long exited the market alive. Governments should avoid supporting such firms, but they also want to protect firms that could be viable under normal circumstances. Firms that were most severely affected— namely, young, small firms in sectors that require that workers show up at the firm’s premises and/or have customers that require physical proximity—may need government support for a while longer. Some firms would have had difficulties regardless of the economic stress added by the pandemic. Therefore, government support programs run the risk of supporting firms that are not viable in the long run even without the COVID-19 shock. For this reason, it is important that support schemes are temporary in nature. Eligibility for aid may need to be limited to firms that can provide evidence that their business was not in the red prior to the outbreak of the COVID-19 pandemic. In line with these two observations, the European Commission adopted a temporary framework for state aid schemes aimed at ensuring firms’ access to liquidity and finance, as well as preserving employment. This framework provided some limiting principles, establishing the temporary nature and eligibility criteria that aimed to target aid to viable firms. For instance, firms that were already having difficulties on December 31, 2019, and hence before the crisis, could not have access to most measures; credit guarantees for loans beyond a certain Euro limit cannot apply  Mongolia enacted its first insolvency law in 1997, but not a single reorganization case has been successfully completed since then. 72  See also Battersby; B., R.A. Espinoza, J. Harris, G. Hee Hong, S.V. Lizarazo Ruiz, P. Mauro, and A. Sayegh (2022) The State as “Financier 73 of Last Resort” International Monetary Fund Staff Discussion Notes No. 2022/003, https://www.imf.org/en/Publications/Staff-Discussion- Notes/Issues/2022/10/11/The-State-as-Financier-of-Last-Resort-523706 206 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA to more than 90 percent of the loan, and the loan principal would normally not go beyond certain amounts (25 percent of yearly turnover, or twice the yearly wage bill). In addition, wage subsidies given to workers who would have otherwise been laid off because of the crisis could not exceed 80 percent of the monthly gross salary. As countries recover, active labor market policies—including hiring incentives and well-targeted training programs—will become more relevant. These types of policies can support a transition back to more productive employment. In a context of rapid technological change and automation, policies to reskill and retrain workers and improve matching of workers to jobs will also play crucial roles. Going forward, governments need to consider alternative mechanisms to manage these risks. In the long term, wage subsidies are inefficient, because they hinder the reallocation of workers from less productive to more productive firms and sectors. Creating unemployment insurance and social safety nets that can quickly adapt to shocks would be a better way to ensure that workers do not bear the brunt of recessions. Other instruments that can be quickly deployed—such as forbearance of payment of bills or short-term lending support through micro-lending institutions—are also likely to be helpful. A critical lesson is the importance of preparing for the next economic shock. Investments in robust social safety, education, and health systems clearly helped some of these economies. Even so, better data could have saved governments money. Millions of dollars went to households and businesses that did not need government support the most, while some households and businesses that did need it went without. The COVID-19 crisis has revealed how robust and credible monitoring and evaluation systems are vital to ensure that policies are effective and support the people and businesses most in need. As countries move on to the next phase of recovery, government budgets may need to be contained. There could even be sharp tradeoffs between supporting recovery and promoting growth. Knowing more about which programs work will help governments assess how and when to curtail assistance. Such knowledge will increase their effectiveness, not only in crises, but also in normal times. Without it, support to households and firms can be both wasteful and unsuccessful. With respect to growth, given the uneven nature of the impact that the pandemic-induced recession had on poorer households and smaller firms, more than ever before, there may be a need to ensure that growth is as inclusive as possible and to compensate for when it is not. Research Agenda Research is ongoing on the economic impacts of the pandemic-induced recession. Initial analyses suggest that the impacts of the first two years of the pandemic on inequality and poverty are likely to be severe. As new household surveys become available, more will be known about the extent of the shock to households, the mechanisms that made households more or less resilient, and measures that might have helped them recover more quickly. A few examples of areas that merit further exploration include the following: ⦁ In some countries, the recovery of employment in 2021 was driven by the growth of more informal, lower-productivity jobs. This could mean that the output recovery results in fewer formal jobs created than in previous decades of sustained growth. As new data becomes available and countries can achieve a sustained recovery, it will be important to ascertain whether there is a structural change where growth does not deliver more formal employment as it has in the past. CHAPTER 7: LESSONS LEARNED 207 ⦁ Some of the recent gains of women in the labor market may have been reversed, as women were forced to take on care responsibilities. As economies recover, it will be important to monitor whether this change is short lived, or if measures are needed to support increased female labor force participation. ⦁ Countries in EAP, as well as around the world, used wage subsidies to preserve the employee-employer relationship at the time of the initial shock, to varying degrees. Taking advantage of these variations, the question remains regarding its effectiveness; that is, to what extent countries that spent more in wage subsidies also saw lower declines in employment. ⦁ The pandemic likely affected countries’ long-term prospects by reducing the accumulation of human capital among the young. In the Philippines, schools remained closed for nearly two years; in Cambodia and Indonesia, many schools remained closed for months. World Bank estimates suggest that school closures could result in a loss of up to two-thirds of a year of learning-adjusted years of schooling in the EAP region (World Bank 2021). Because of unequal access to technology, children from poor and vulnerable households are less likely to have engaged in remote learning activities and are therefore far more likely to have experienced learning losses than other children. Understanding how long-lasting these losses are and what can be done to recover them should be a priority. ⦁ International trade plunged in 2020 but rebounded in 2021. The heterogeneity of impacts and changes in trade flows across products, sources, and destinations creates high uncertainty and imposes adjustment costs on firms. It is yet unclear why some countries, sectors, and firms were able to recover quickly, even when exposed to plunging trade while others have yet to recover. ⦁ Much has been written about the role of the investment climate (or business environment) in contributing to a more efficient economy. This recession and its nascent recovery may reveal the role of the investment climate as a shock absorber and facilitator of recovery. ⦁ Finally, there is much evidence of high levels of adoption of digital technologies and teleworking as a result of the COVID-19 pandemic. Optimism over increased adoption of new technologies is based on an implicit assumption that there was socially suboptimal adoption of digital technologies prior to the pandemic. But there is room for a less optimistic discussion about the adoption of these new technologies. Firms could have had the socially appropriate level of investment in digital pre-crisis, but the pandemic has forced them to spend inefficiently large amounts on digital platforms—in other words, added an extra burden to firms which, when public support is withdrawn, may raise questions of sustainability. Ultimately, the best measure of technological adoption is productivity growth. It is possible that COVID has changed the direction of technology shifts (namely, toward specific kinds of digital), but overall diffusion and productivity growth has fallen. Whether productivity will sustainably increase as a result of these new technologies has yet to be understood. Conclusion The COVID-19 pandemic severely tested public sector institutions. It forced nearly all households, businesses, governments, and individuals around the globe to participate in experiments that never would have been undertaken in normal times, including mobility restrictions, near universal lockdowns, and working from home. There is much to learn about these unexpected changes to daily life. 208 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA The longer-term consequences of the pandemic remain unclear. The myriad impacts of working from home on productivity, organizational planning, real estate markets, infrastructure, and urban planning could be consequential. The pandemic may have permanently affected travel, tourism, and the hospitality industries and the way services are delivered. It may have changed the nature of trade; global value chains; and onshoring, off-shoring, and near-shoring. Of greatest concern are the long-term impacts of the pandemic and the recession that ensued. More specifically, the economic sacrifices during the downturn seemed unequal, both for households and firms. This report endeavors to show the reader how this was so. The open question is whether the recovery will be unequal. Will it make up for the losses of those that were hurt the most or add to the gains of those that lost the least? Government policies and the economic forces that the pandemic has unleashed will have a lot to do with how the recovery takes shape. Addressing some of the structural conditions that contributed to the vulnerability of the countries in this study can help prepare economies for stronger growth and soften the impacts of current and future recessions. For the moment, it will be important for policy makers around the world to remain vigilant and focused on making sure that the recovery is as inclusive as the recession was painful for those who could least afford to lose economic ground. CHAPTER 7: LESSONS LEARNED 209 References ADB (Asian Development Bank). 2020. “Sustainable Recovery Options for Mongolia’s Micro, Small, and Medium-Sized Enterprises.” ADB Brief, December 2020. https://www.adb.org/publications/msmes-informal-sector-mongolia- covid-19-impacts. ADB. 2022. COVID-19 Policy Database. (accessed June 27, 2022), https://data.adb.org/dataset/adb-covid-19-policy- database. Aizenman, Joshua., Yothin Jinjarak, Hien Thi Kim Nguyen, and Ilan Noy. 2021. “The Political Economy of the COVID-19 Fiscal Stimulus Packages of 2020.” NBER Working Paper No. 29360. National Bureau of Economic Research, Munich. Alberola, Enrique, Yavuz Arslan, Gong Cheng, and Richhild Moessner. 2021. “Fiscal response to the COVID-19 crisis in advanced and emerging market economies.” Pacific Economic Review 26( 4), 459– 468. https://doi.org/10.1111/1468- 0106.12370. Amaglobeli, David, Emine Hanedar, Gee Gee Hong, and Celine Thévenot. 2022. “Response to High Food, Energy Prices Should Focus on Most Vulnerable.” IMF(blog). June 7, 2022. https://blogs.imf.org/2022/06/07/response-to-high- food-energy-prices-should-focus-on-most-vulnerable/. Apeti, Ablam Estel, Combes, Jean-Louis, Xavier Debrun, and Alexandru Minea. 2021. “Did Fiscal Space Influence Covid-19’s Fiscal Response?” COVID Economics (74) March  30, 2021. Center for Economic Policy Research. https://cepr.org/publications/covid-economics-issue-74#392514_392950_390751. Benmelech, Efraim, and Nitzan Tzur-Ilan. 2020. The Determinants of Fiscal and Monetary Policies During the Covid-19 Crisis. NBER Working Paper No. 27461. National Bureau of Economic Research. Cambridge, MA. Chen, Can, Yu Shi, Ping Zhang, and Chengri Ding. 2021. “A Cross-Country Comparison of Fiscal Policy Responses to the COVID-19 Global Pandemic.” Journal of Comparative Policy Analysis: Research and Practic 23(2). https:// doi.org/10.1080/13876988.2021.1878885. Cho, Yoonyoung, Jorge Avalos, Yasuhiro Kawasoe, Doug Johnson, and Ruth Rodriguez. 2021. Mitigating the Impact of COVID-19 on the Welfare of Low Income Households in the Philippines: The Role of Social Protection. Washington: DC: World Bank. https://documents1.worldbank.org/curated/en/698921611118950758/pdf/Mitigating-the-Impact-of- COVID-19-on-the-Welfare-of-Low-Income-Households-in-the-Philippines-The-Role-of-Social-Protection.pdf. Elgin, Ceyhun, M. Ayhan Kose, Franziska Ohnsorge, Shu Yu. 2021. “Understanding Informality.” CERP Discussion Paper 16497. Center for Economic Policy Research, London. https://cepr.org/publications/dp16497. Felipe, Jesus, Fullwiler, Scott., Bajaro, Donna F., Yusoph, Al-Habbyel, Askin, Simon A., and Cruz, Martin. A. 2020. “An Analysis of the Worldwide Response to the COVID-19 Pandemic:What and How much?” Asian Development Bank Economics Working Paper Series No. 626. ADB, Manila. Feyen, Erik, Alonso Gispert, Tatiana Kliatskova, and David S. Mare. 2020. “Taking Stock of the Financial Sector Policy Response to COVID-19 around the World.” Policy Research Working Paper Series No. 9497. Washington, DC: World Bank. http://documents.worldbank.org/curated/en/143061607958551050/Taking-Stock-of-the-Financial- Sector-Policy-Response-to-COVID-19-around-the-World. Gentilini, Ugo. 2022. Cash Transfers in Pandemic Times : Evidence, Practices, and Implications from the Largest Scale Up in History. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/37700. Gourinchas, Pierre-Olivier. 2020. “Flattening the Pandemic and Recession Curves.” In Mitigating the COVID Economic Crisis: Act Fast and Do Whatever It Takes, edited by R Baldwin and B Weder di Mauro. A VoxEU.org Book, pp. 31–40. https://www.researchgate.net/publication/342961959_Mitigating_the_COVID_Economic_Crisis_Act_Fast_and_ Do_Whatever_It_Takes_A_VoxEUorg_Book. Hosny, Amr. 2021. “The Sooner (and the Smarter), the Better: COVID-19 Containment Measures and Fiscal Responses.” Working Paper No. 2021/065. IMF, Washington, DC. https://www.imf.org/-/media/Files/Publications/WP/2021/ English/wpiea2021065-print-pdf.ashx. Lee, Daniel;. Jeong-Dai. 2020. “An Assessment of Fiscal Space for COVID-19 Response and Recovery in Asia-Pacific Developing Countries.” United Nations Economic and Social Commission for Asia and the Pacific ( UN ESCAP). MPFD Policy Briefs 116. .UN ESCAP. O’Keefe, Phillip. (forthcoming). Diverse Paths: The Dynamic Evolution of Social Protection in Asia and the Pacific. Washington, DC: World Bank. 210 CRISIS AND RECOVERY: LEARNING FROM COVID-19’S ECONOMIC IMPACTS AND POLICY RESPONSES IN EAST ASIA World Bank. 2021a. Long COVID: EAP Economic Update, October 2021. World Bank, Washington, DC. https:// openknowledge.worldbank.org/bitstream/handle/10986/36295/211799sum.pdf. World Bank. 2021b. From Relief to Recovery: Mongolia Economic Update, February 2021. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35161. World Bank. 2021c. A Year Deferred - Early Experiences and Lessons from Covid-19 in Vietnam. Washington, DC: World Bank. https://documents.worldbank.org/en/publication/documents-reports/documentdetail/099110006102228289/ p1647370e8e7f20660b1af0a203f9d4d5e7. World Bank. 2022a. EAP Economic Update April 2022: Braving the Storms. World Bank, Washington, DC. https:// openknowledge.worldbank.org/bitstream/handle/10986/37097/9781464818585.pdf. World Bank. 2022b. Cambodia Poverty Assessment: Toward a More Inclusive and Resilient Cambodia. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/38344?show=full. World Bank. 2022c. Poverty and Shared Prosperity Report 2022: Correcting Course. Washington, DC: World Bank. World Bank. 2022d. Mongolia Economic Update: Navigating Stronger Headwinds–April 2022. World Bank, Washington, DC. https://thedocs.worldbank.org/en/doc/af1a0293254ac2e448cafa165c669d88-0070012022/ original/MEU-2022-April-ENG.pdf.