EARLY LABOR MARKET IMPACTS OF COVID-19 IN DEVELOPING COUNTRIES  1 The economic crisis caused by the COVID-19 pandemic sharply reduced mobility and economic activity, disrupting the lives of people around the globe. This brief presents estimates on the crisis’ impact on labor markets in 39 countries based on high-frequency phone survey (HFPS) data collected between April and July 2020. Workers in these countries experienced severe labor market disruptions following the COVID-19 outbreak. 34% of respondents reported stopping work, 20% of wage workers reported lack of payment for work performed, 9% reported job changes due to the pandemic, and 62% reported income loss in their household. Measures of work stoppage and income loss in the HFPS are generally consistent with GDP growth projections in Latin America and the Caribbean but not in Sub-Saharan Africa, indicating that the phone survey data contributes valuable new information about the impacts of the crisis. Ensuring availability of such critical data in the future will require investments into statistical and physical infrastructure as well as human capital to set up Emergency Observatories, which can rapidly deploy phone surveys to inform decision makers. The global coronavirus pandemic (COVID-19) has of the crisis in developing countries due to lack of data. slowed economic activity as governments imple- This brief is an early attempt to measure the initial labor mented lockdown measures, individuals reduced market consequences of the crisis using harmonized their mobility, and firms’ production processes were data from high-frequency phone surveys. The measures disrupted. These broader shifts in the economy affected derived from the HFPS data differ from macroeconomic both the demand for labor and workers’ willingness to projections and preliminary estimates from labor force work. The labor market impacts, like overall economic surveys, and therefore provide additional insights into impacts, likely varied considerably across countries, the immediate initial impacts in developing countries. based on initial economic and labor market conditions and differing policy responses. DATA AND METHODOLOGY 2 The economic slowdown during COVID-19 led to adverse labor market impacts in both developed This brief provides estimates on the early labor and developing countries. Understanding how the market impacts of the COVID-19 pandemic in 39 pandemic affected labor markets in the developing countries using high-frequency phone surveys world in Spring 2020 is crucial as governments and other (HFPS). Countries administered slightly different surveys actors continue to develop responses. Yet there is little and sampling frames also differed across countries. The systematic knowledge about the labor market impacts World Bank’s Data for Goals team harmonized these 1 This brief is based on Melanie Khamis, Daniel Prinz, David Newhouse, Amparo Palacios-Lopez, Utz Pape, Michael Weber, “The Early Labor Market Impacts of COVID-19 in Developing Countries: Evidence from High-Frequency Phone Surveys.” World Bank Policy Research Working Paper 9510, 2020. 2 For a detailed description of the data and methods, including issues of representativeness and weighting, see Melanie Khamis, Daniel Prinz, David Newhouse, Amparo Palacios-Lopez, Utz Pape, Michael Weber, “The Early Labor Market Impacts of COVID-19 in Developing Countries: Evidence from High-Frequency Phone Surveys.” World Bank Policy Research Working Paper 9510, 2020. 1 surveys to the extent possible. We use data from the FIGURE 1 first wave of the HFPS between April and July 2020 Share Stopped Working by Country that were collected in the December 1 vintage of the harmonized data. The data contain 6 countries in A. By Region Europe & Central Asia, 7 in East Asia & Pacific, 12 in Latin America & Caribbean, 2 in Middle East & North BOL 69% Africa, and 12 in Sub-Saharan Africa. 8 countries are TUN 64% low income, 17 countries are lower middle income, KEN 62% 10 countries are upper middle income and 4 countries PER 59% are high income. MMR 57% SLV 54% WORK STOPPAGES WERE COMMON IN HND 52% MANY COUNTRIES DOM 51% ECU 51% During the COVID-19 pandemic, a large share of COL 50% workers stopped working in all countries (Figure 1a). UZB 50% Taking a simple average across countries, 34% of NGA 50% respondents reported stopping work. The average PRY 43% across countries in our data is 21% in the EAP region, GTM 42% 29% in the ECA region, 48% in the LAC region, 45% in the MENA region, and 26% in the SSA region. We SSD 39% note that the set of countries in our data is not repre- MEX 39% sentative of regions). There is significant variation, even CRI 36% within regions. For example, within the LAC region, at CHL 30% the lower end 30% stopped working in Chile and 36% HRV 29% in Costa Rica, while at the higher end 59% stopped MLI 28% working in Peru and 69% in Bolivia. In the SSA region DJI 27% estimated shares are as low as 8% in Madagascar and GHA 26% 11% in Burkina Faso and shares as high as 50% in Nigeria and 62% in Kenya. ZMB 26% ROU 25% Figure 1b shows that upper middle-income countries IDN 23% (41% on average) and lower middle-income countries POL 22% (37%) had the most work stoppage. High income ZWE 21% countries had 26% of respondents on average stop PNG 21% work, followed by low income countries at 19%. (We UGA 20% note that the set of countries in our data is not repre- BGR 20% sentative of country income groups). MNG 19% In the LAC countries, respondents were also asked SLB 15% about whether they were planning to return to work ETH 14% if they stopped working. Figure 1c suggests that the LAO 13% majority of workers who stopped working were plan- MWI 11% ning to return to work, though there is some variation BFA 11% across countries. MDG 8% VNM 3% While variations across regions and countries occurred, a significant proportion of respondents 0 20% 40% 60% 80% stopped work in the early stages of the pandemic. Share Stopped Working Work stoppages tended to be less severe in agriculture EAP ECA LAC MENA SSA than in industry and services. Taking the simple average 2 FIGURE 1 Share Stopped Working by Country B. By Income C. Planning to Return to Work BOL 69% BOL 52% 16% TUN 64% PER 32% 27% KEN 62% SLV 39% 15% PER 59% MMR 57% HND 37% 15% SLV 54% DOM 37% 15% HND 52% ECU 33% 17% DOM 51% ECU 51% COL 26% 24% COL 50% PRY 30% 13% UZB 50% GTM 24% 18% NGA 50% PRY 43% MEX 28% 11% GTM 42% CRI 20% 16% SSD 39% CHL 24% 6% MEX 39% CRI 36% 0 20% 40% 60% 80% CHL 30% Share stopped working HRV 29% Planning to return to work MLI 28% Not planning to return to work DJI 27% GHA 26% across countries, 22% of agricultural workers reported ZMB 26% stopping work, as opposed to 40% for industry and 38% ROU 25% for services. This suggests that the disruptive labor mar- IDN 23% ket impacts were substantial throughout the economy. POL 22% ZWE 21% PNG 21% PARTIAL PAYMENT, JOB CHANGES, AND UGA 20% LOSS OF INCOME WERE ALSO SIGNIFICANT BGR 20% During the pandemic, a substantial share of MNG 19% employees experienced partial or no payments SLB 15% for work performed (Figure 2). This question on ETH 14% partial or no payments for work performed is avail- LAO 13% able mostly in countries in the LAC region. The share MWI 11% reporting partial or no payments in this region ranges BFA 11% from 17% in Chile to 30% in Peru. This indicates that MDG 8% in addition to stopping work, reductions in pay due to reduced economic activity was an important challenge VNM 3% to workers. The workers nominally kept their jobs but 0 20% 40% 60% 80% were not receiving the full payment for the work per- Share Stopped Working formed, either possibly due to some furlough type of arrangements or employers delaying or reducing the High income Low income pay in response to the crisis. Importantly, we cannot Lower middle income Upper middle income measure reduced working hours directly in the HFPS data. 3 FIGURE 2 FIGURE 3 Share of Wage Workers with Partial or No Payments Share Changed Job During the Pandemic IDN 57% SSD 21% PER 30% ZMB 18% ECU 30% ECU 14% MEX 26% BOL 14% SSD 25% PER 14% PRY 24% UGA 13% HND 24% SLB 12% NGA 24% IDN 11% BOL 23% BFA 10% DOM 21% SLV 10% COL 20% GTM 9% CRI 20% VNM 7% GTM 19% LAO 6% SLV 17% HND 6% CHL 17% ZWE 6% VNM 12% MMR 5% ETH 10% PRY 5% MWI 7% MEX 5% ZWE 6% MWI 5% MNG 5% DOM 5% DJI 5% COL 5% PNG 4% 0 10% 20% 30% 40% 50% 60% CRI 4% Share with Partial or No Payment CHL 4% EAP LAC MENA SSA ETH 2% 0 5% 10% 15% 20% 25% Share Changed Job The large disruption in the labor market is also apparent from the high share of workers changing EAP LAC SSA jobs during the pandemic (Figure 3). Where data is available, job changing ranged from 2% to 21% in the SSA region and 4% to 14% in the LAC region. reporting stopping work is likely because other sources This could be an indication that some of the jobs that of income from the respondent or other household workers changed from where affected by the pandemic members, including from family businesses are cut and while the jobs that workers changed to were either new because even respondents who do not stop work, often jobs or some type of self-employment or in sectors that receive partial or no payments or had to change jobs. were differentially affected by the crisis. Labor market disruptions translate into income INITIAL LABOR MARKET IMPACTS losses, but income loss is even more prevalent DIFFERED FROM GDP PROJECTIONS, (Figure 4). Where data is available, we find that house- ESPECIALLY IN SUB-SAHARAN AFRICA holds report losing income. A high share of respondents reported total income loss (62%), as well as loss from Estimates of labor market disruptions line up farming (62%) and non-farming (75%) family businesses, well with macroeconomic estimates of economic and wage incomes (49%). The fact that more respondents impacts in the LAC region, but not in the SSA region report income loss in their household than the share (Figure 5). Using the change in the IMF’s GDP growth 4 FIGURE 4 FIGURE 5 Total Income Loss Stopping Work vs Change in GDP Projection A. Latin America & Caribbean PER 83% 80% MWI 80% NGA 80% 70% BOL GHA 78% 60% Share stopped working PER SLV ECU 77% ECU HND 50% COL DOM VNM 74% PRY GTM COL 73% 40% MEX CRI GTM 73% 30% BOL 71% CHL 20% SLV 69% CRI 68% 10% PRY 66% 0 HND 66% –20 –18 –16 –14 –12 –10 –8 –6 –4 TJK 65% Change in WEO Projection MEX 62% B. Sub-Saharan Africa DOM 62% 70% UZB 60% KEN ZMB 59% 60% CHL 58% NGA Share stopped working ETH 55% 50% MMR 54% 40% LAO 46% SSD POL 45% 30% MLI ROU 38% ZWE ZMB GHA 20% BGR 34% UGA ETH BFA HRV 28% 10% MWI 0 10% 20% 30% 40% 50% 60% 70% 80% 90% MDG Share with Total Income Loss 0 –14 –12 –10 –8 –6 –4 EAP ECA LAC SSA Change in WEO Projection projection between the October 2019 and the October Overall, labor markets were severely disrupted by the 2020 World Economic Outlook (WEO), it appears that pandemic in a wide swath of countries. Work was countries where the IMF downgraded the projection severely reduced, and the loss of employment and the more have a higher share of respondents reporting overall economic impacts of the pandemic led to sub- stopping work in the LAC region (Figure 5a, R2 = 0.38). stantial income loss. Further disruption was apparent In the SSA region, this relationship is weak, and coun- through partial or no payment of wage workers, job tries with smaller declines in GDP growth tended to changes, and income loss. Macroeconomic projections have greater incidence of work stoppage. (Figure 5b, do not capture the full impact of households, partic- R2 = 0.08). This suggests that phone survey data is pick- ularly in Sub-Saharan Africa. The data from phone ing up labor market and economic impacts earlier and surveys therefore contribute valuable information on that are not typically incorporated into macroeconomic how households in a broad cross-section of developing projections. One possible reason for the differences countries were affected by this severe shock. across regions, may be because of the prevalence of informal arrangements in the SSA region. 5 Photo credit: World Bank / Sambrian Mbaabu ENABLING NATIONAL STATISTICS OFFICES for such rapid deployment when crises hit countries in TO DEPLOY RAPID RESPONSE PHONE the future, National Statistics Offices can invest now SURVEYS IN THE EVENT OF A CRISIS CAN to improve the speed of deployment and quality of PROVIDE TIMELY EVIDENCE FOR CRITICAL data. Investments in statistical infrastructure (e.g., the DECISION MAKING preparation of representative sampling frames for phone surveys), physical infrastructure (e.g., setup of phone The rapid deployment of phone surveys to measure the centers) as well as human capital (e.g., establishment socio-economic impacts of COVID-19 was only possible of capable units designing, implementing and dissemi- because of an extraordinary effort around the globe. The nating results from phone surveys) will be needed. The collected data prove to be vital for decision makers to establishment of such Emergency Observatories can be understand the impact of the crisis. To be better prepared a game changer for policy making in the future. This note was prepared by Melanie Khamis (World Bank and Wesleyan University), Daniel Prinz (World Bank and Harvard University), David Newhouse (World Bank), Amparo Palacios-Lopez (World Bank), Utz Pape (World Bank) and Michael Weber (World Bank) as part of the World Bank’s JobsWatch Covid-19: Monitoring Labor Market Impacts and Policy Responses to The Pandemic in the Developing World (P174663; Michael Weber and David Newhouse, Task Team Leaders). This note is a joint product of the Jobs Group and the Poverty and Equity Global Practice. The authors would like to thank Sukti Dasgupta (ILO), Sangheon Lee (ILO), Truman Packard, and Nobuo Yoshida for helpful comments, and Benu Bidani, Ambar Narayan, Michal Rutkowski, Carolina Sanchez-Paramo, and Ian Walker for their guidance. The team is also grateful to the Poverty and Equity Global Practice and the Data for Goals group for collecting, harmonizing, and sharing the phone survey data, and to Denis Medvedev and Leonardo Iacovone for providing aggregate indicators from firm surveys. Aggregate indicators from the high frequency phone surveys are available at the High Frequency Phone Survey dashboard at: https://www.worldbank.org/en/data/interactive/2020/11/11/covid-19-high-frequency-monitoring-dashboard The production and publication of this report has been made possible through financial support from the World Bank’s Jobs Umbrella Multi-donor Trust Fund (MDTF), which is supported by the UK’s Foreign, Commonwealth & Development Office/UK AID, the Governments of Norway, Germany, Austria, the Austrian Development Agency, Italy, and the Swedish International Development Cooperation Agency; and from the Korean Trust Fund for Economic and Peace Building Transitions. All Jobs Group’s publications are available for free and can be accessed through the World Bank or the Jobs and Development Partnership website. Please send all queries or feedback to Jobs Group. Join the conversation on Twitter: @WBG_Jobs #Jobs4Dev. 6