J U LY, 2 0 2 0 COX’S BAZAR PANEL SURVEY: RAPID FOLLOW-UP ROUND 1 IMPACTS OF COVID-19 ON WORK AND WAGES IN COX’S BAZAR This is the first in a series of briefs to disseminate findings from high-frequency rapid follow-ups on the CBPS sample, by the Poverty and Equity GP of the World Bank. CBPS BRIEF 3, POVERTY AND EQUITY GP 2 A rapid phone survey was implemented on a representative sample of recently displaced Rohingya WAGES W ORK households and their host communities in the Cox’s Bazar district of Bangladesh, to track the impacts of the COVID-19 crisis on labor markets, wages, and household coping strategies. This survey built on the Cox’s Bazar Panel Survey (CBPS), which is a multi-topic survey that focused on socio-economic outcomes and access to health services. This first round of the rapid phone survey, was conducted from 21 April-20 May 2020 (a month into the 2-month long COVID-19 lockdown). A sub-sample of 3,005 out of the 5,020 households surveyed at baseline were covered by this survey. In this first tracking survey 3,009 out of the 9,045 adults surveyed in baseline were covered. HOST EMPLOYMENT HOST INCOME LOSSES R E C E N T LY D I S P L A C E D ROHINGYA MORE THAN HALF OF THE EM- DAILY AND WEEKLY WAGE WORKERS PLOYED RESPONDENTS WERE AB- ACTIVE WITHIN THE LOCKDOWNS SIGNIFICANT DROPS IN EMPLOY- SENT FROM WORK OVER THE 7 DAYS FACED HIGHEST INCOME DROPS MENT IN CAMPS SINCE BASELINE, PRIOR TO THE SURVEY AND REPORT- ACROSS THE DISTRICT BUT ENTER- DRIVEN BY MULTIPLE FACTORS, ED REDUCED OR NO EARNINGS. PRISE OWNERS FACED VERY DIFFER- RELATED AND UNRELATED TO THE ENT CONSEQUENCES ACROSS HIGH COVID—19 CRISIS. AND LOW EXPOSURE REGIONS. While employment and unemployment are not significantly different from baseline Among the few Rohingya participating in the Daily and weekly wage laborers who were levels, reported employment masks a large labor force, employment has dropped from active during the lockdown faced 47 per- proportion of temporarily absent workers. 64 percent in 2019 to 23 percent, while un- cent income drops on average across all More than half of the respondents who re- employment has increased sharply from 36 hosts. Monthly salaried workers were most port being employed are not actively work- percent to 77 percent. More than half of the protected with losses ranging from 15-19 ing, i.e. were absent from work in the 7 days working-age male camp respondents have percent across all regions. Non-wage work- prior to the survey. The largely informal not worked since January 2020, suggesting ers however, including business owners, nature of jobs held by majority of this tem- that this trend is driven by pre-COVID 19 fac- were very differently affected across high porarily absent workers, makes it difficult to tors such as the government directive cir- and low exposure hosts. Those in high expo- predict how much of this employment will culated in September 2019 (post-baseline) sure areas reported 12 percent reduction in translate into active jobs post-lockdown. banning cash-for-work program in camps. earnings on average, while in low exposure areas, reported decrease in earnings from baseline rates was 43 percent on average. CBPS BRIEF 3, POVERTY AND EQUITY GP 3 ABOUT THE COX’S BAZAR PANEL SURVEY AND HIGH FREQUENCY FOLLOW UPS The CBPS was designed to be representative Bangladesh’s local economy started experi- of recently arrived Rohingya (displaced after encing impacts of the COVID-19 crisis in early August 2017) and Bangladeshi households re- to mid-March, with the first case being report- siding in host communities in Cox’s Bazar, and ed on 7 March. A full countrywide lockdown the baseline for this survey was completed was in place from 26 March-28 May. This brief in August 2019. To distinguish between host presents findings from the first round of the communities that are more or less affected CBPS high-frequency follow-ups (conducted by the arrival of these Rohingya, the survey’s between April-May, 2020) on how this crisis sampling strategy uses a threshold of 3-hours has impacted key labor market indicators and walking time from a campsite to define two outcomes in Cox’s Bazar across the host and strata for hosts: (i) Host communities with Rohingya population. The findings from the potentially high exposure to the displaced Ro- follow-up are presented as a panel update on hingya, and (ii) Host communities with poten- baseline adults1. tially low exposure. 1 The results are weighted using adjusted baseline weights that account for non-response and selection into the in- terview based on characteristics measured at baseline. The CBPS is the result of a partnership between the Yale Macmillan Center Program on Refugees, Forced Displacement, and Humanitarian Responses (Yale Macmillan PRFDHR), the Gender & Adolescence: Global Evidence (GAGE) program, the Poverty and Equity Global Practice of the World Bank. CBPS BRIEF 3, POVERTY AND EQUITY GP 4 SEEMINGLY NORMAL EMPLOYMENT RATES WITHIN LOCKDOWN MASK HIGH RATES OF TEMPORARY ABSENCE FROM WORK. Reported employment among hosts, even Figure 1: Labor force indicators for hosts within the lockdown period, remains high (89 between the baseline and follow up percent)2. However, the figure masks a large share of the labor force who are reporting be- 95,3% 88,7% ing employed3 but temporarily absent from work i.e. not actively working. Almost 2 out of 3 hosts who report being employed were in 51,4% fact not actively working in the 7 days before 42,1% the survey. In contrast, during the baseline (Mar-Aug ’19), temporary absence among the employed was less than 1 percent. Unsurpris- 11,3% 4,7% ingly, an overwhelming majority of temporar- ily absent workers attributed the reason to Labor force Employment Unemployment participation rate rate COVID-related restrictions to work. Baseline (Mar-Aug 2019) Follow up (Apr-May 2020) Figure 2: Host labor force work status since jan '20 All hosts Male Female High Exposure Low Exposure Not worked since Jan Worked between Jan-Mar Worked in Mar Employed but did not work in Apr Actively worked in April Host communities also reported increased rates of labor force participation, and this was ac- companied by a 7 percent decline in the employment rate; and a similar rise in the unemploy- ment rates. High exposure hosts demonstrate relatively smaller changes from baseline across all indicators, in comparison to low exposure hosts. 2 Bangladesh’s local economy started experiencing im- 3 The findings from the follow-up are presented as a pan- pacts of the COVID-19 crisis in early to mid-March, with el update on baseline adults. Employment is defined as the first case being reported on 7 March. A full country- the share of the labor force (adults over the age of 15 who wide lockdown was in place from 26 March-28 May. The are either currently employed or not employed but ac- labor module measures outcomes across three periods: (i) tively seeking work over the past 7 days) reporting having During the survey period (7 days prior to survey period in worked at least one hour in the past 7 days or being tem- late April to mid-May); (ii) from 1 March to early April 2020 porarily absent from work. Similarly, unemployment rates for individuals who report being unemployed during the are reported as a % of the labor force that has not worked survey period (when a potential lockdown was under dis- in the past 7 days or been temporarily absent from a job cussion and gradually came into effect); and (iii) from Jan- but has actively looked for work in said recall period. uary to early March 2020 for individuals who report being unemployed 1 March onwards (when the first known cas- es of COVID-19 were identified in Bangladesh). CBPS BRIEF 3, POVERTY AND EQUITY GP 5 T H E P O P U L AT I O N T E M P O R A R I LY A B S E N T F R O M W O R K A R E P R I M A R I LY N O N - W A G E O W N A C C O U N T W O R K E R S A N D D A I LY O R W E E K LY W A G E L A B O R E R S , I N D I C AT I N G U N PA I D A BS E N C E F R O M WO R K . About 1 in 3 hosts who report being employed Figure 3: Share of employment type among but temporarily absent from work report be- temporarily absent hosts ing daily or weekly wage laborers; and half 12% 14% of them are non-wage own account workers. Both of these employment types generate 35% 38% earnings based on work done i.e. either on how many hours or days worked for the for- mer, or how many customers or clients were served for self-employed workers. This in- 53% 49% dicates that despite reporting that they had employment during the lockdown, these tem- porarily absent workers were likely not paid or High Exposure Low Exposure generating income during this time; and that they were unlikely to receive any back pay4. Monthly salaried workers Daily/weekly wage laborers Non-wage workers Table 1: Top 5 Jobs Reported By Actively Working Hosts In Cox’s Bazar HIGH EXPOSURE HOSTS % LOW EXPOSURE HOSTS % Farmer (on own land) 15% Self-run agricultural activities 16% Agricultural day laborer 14% Small businessman (tongs) 11% Small businessman (tongs) 9% Agricultural day laborer 10% Self-run agricultural activities 9% Private sector employee 8% Rickshaw/van driver 5% Hens/duck rearing 5% Other 48% Other 51% I N D I V I D U A L S W I T H A G R I C U LT U R A L A N D H O M E - B A S E D I N C O M E S O U R C E S R E P O R T E D B E I N G M O R E ACT I V E I N T H E L A B O R M A R K E T D U R I N G T H E LO C K D OW N , CO M PA R E D TO THOSE IN SERVICE-SECTOR JOBS. Rates of temporary absence were higher for low exposure hosts (67 percent) than for high expo- sure hosts (53 percent); and for employed men. This could potentially be explained by the nature of jobs that these groups are typically engaged in: population segments which were more de- pendent on agricultural and home-based income generating activities were able to participate in some kind of economic activity during the lockdown, compared to those in service sector jobs that were less accessible during this period. The table below highlights how active income sourc- es during the lockdown (April) are mostly agricultural, not only in the more rural high exposure areas but also among active jobs reported in the relatively more urbanized low exposure regions5. 4 The survey followed a standard labor module and asked faced by these temporarily absent workers from this sur- about wage incomes in the employment reported by these vey round. temporary workers during the last seven days that they 5 According to baseline findings, 41 percent of hosts in worked. However, it did not ask if these workers were also high-exposure areas rely on agriculture for their liveli- paid during this absence from work or if they expected to hoods, compared to 30 percent for hosts in low-exposure be paid for this period when they were able to return to areas. Source: CBPS Brief III: Insights from The Labor Mod- work. This precludes a calculation of the income losses ule on Work and Wages in Cox’s Bazar CBPS BRIEF 3, POVERTY AND EQUITY GP 6 THOSE WHO REMAINED ACTIVE DURING THE LOCKDOWN REPORT REDUCED EARNINGS, WITH LOW EXPOSURE REGIONS BEING MORE AFFECTED ACROSS ALL EMPLOYMENT TYPES. Compared to pre-crisis baseline rates, people Figure 4: Comparison of precrisis and who worked during the lockdown reported lockdown earning levels for different employment types reduced earnings, across all employment types, particularly in low exposure areas. Dai- ly and weekly wage laborers faced much high- -15% -19% er losses (49 percent) in income compared to -12% other workers across all hosts. Monthly sala- -43% ried wage workers have relatively been more -49% -47% protected in terms of income losses, facing Weekly Monthly Nonwage Weekly Monthly Nonwage 15-19 percent reductions across high and low wage salaries earnings wage salaries earnings labor (last labor (last month) exposure hosts. earnings month) earnings High exposure Low exposure INCOME LOSSES DURING THE LOCKDOWN Baseline rates Crisis impacted rates (April-May, 2020) FOR ACTIVE NON-WAGE OWN ACCOUNT WORKERS AND BUSINESS OWNERS VARY ACROSS HIGH AND LOW EXPOSURE AREAS. occupations6, this further highlights the differ- ential impact that the lockdowns have had on Among non-wage workers, high exposure the economic sectors, with agricultural work hosts faced much lower income losses (15 per- facing lower disruptions than other activities. cent) than their counterparts in low exposure areas (43 percent). With high exposure hosts INCREASED UNEMPLOYMENT: MEN IN HIGH EXPOSURE AREAS AND more dependent on agriculture and low ex- WOMEN IN LOW EXPOSURE AREAS posure hosts on industrial and service sector F A C E PA R T I C U L A R LY H I G H R AT E S O F U N E M P L O Y M E N T. Figure 5: Increasing unemployment rates The unemployment rate for men doubled on by gender and exposure area average across both high and low exposure 16,1% areas. Unemployment rates for women in high-exposure areas have not changed sig- 12,3% nificantly, unlike for women in low exposure 9,2% areas, where unemployment has increased by 6,8% 6,7% almost eight times the previous rate (Figure 5). 5,3% 4,7% 2,3% High Exposure Low Exposure High Exposure Low Exposure Male Female 6 Cox’s Bazar Panel Survey Brief 3: Work and wages in Cox’s Baseline (Mar-Aug 2019) Follow up (Apr-May 2020) Bazar. World Bank. CBPS BRIEF 3, POVERTY AND EQUITY GP 7 NEW LABOR FORCE ENTRANTS, MOSTLY WOMEN AND SECONDARY HOUSEHOLD MEMBERS, ARE DRIVING UP PARTICIPATION AND UNEMPLOYMENT RATES. More than a quarter (27 percent) of the cur- Given the unique context of the COVID-19 rent labor force were non-participants during lockdown, it is likely that reported rates of em- the baseline. This increase is largely driven ployment in this rapid phone survey – which by low-exposure hosts (30 percent), where derive in large part from a high rate of tempo- the share of new entrants into the labor force rary absence - are underestimating the mag- is twice that among high-exposure hosts (15 nitude of job losses which may be realized percent). 17 percent of these new entrants are once the lockdown lifts and workers attempt unemployed, with a gender ratio that is consis- to rejoin their jobs. Whether these expecta- tent with the unemployment rates. Two-thirds tions will translate to re-employment of these of new host labor force entrants are women, largely informal wage workers post-lockdown again largely driven by low-exposure regions will depend on a host of factors such as which (70 percent); and close to 60 percent of these sectors in the local economy are prioritized entrants are secondary household members i.e. in the partial reopening of the economy, how the spouses or children of the household heads. localized quarantines of neighborhoods and areas translates to job accessibility and mobil- ity, and the overall economic outlook for the REPORTED RATES OF EMPLOYMENT AND major sectors of employment in the economy. UNEMPLOYMENT FOR HOSTS MAY NOT REFLECT THE EXTENT OF JOB LOSSES POST- LOCKDOWN. THE DISPLACED ROHINGYA thirds of the drop offs occurring before March 1, in EMPLOYMENT DECLINED SIGNIFICANTLY the pre-COVID-19. Only a third of these people lost WITH A SHARP RISE IN UNEMPLOYMENT jobs due to pre-COVID regulations and changes COMPARED TO BASELINE, BUT MOST OF in camps, the rest were unemployed during the THIS DECLINE PRECEDES COVID-19. baseline as well. During the baseline (March-August, 2019), one out of three Rohingya adults participated in the labor Figure 6: Comparison of labor force indica- force, with two third of them employed (or roughly tors for camps between the baseline and two-ninths of the working age population). Current follow up employment is one third of baseline rates while 77,3% unemployment has doubled between August 2019 64,0% to May 2020. 32,6% 37,4% 36,0% 22,7% However, work status indicators7 for the camp population show that unlike for hosts, the majority Labor force Employment Unemployment of the changes in employment occurred after the participation rate rate baseline and preceding the COVID-19 crisis. Two Baseline (Mar-Aug 2019) Follow up (Apr-May 2020) 7 The work statuses were derived in the survey in order ed not being currently employed, and if they had worked of most recent period worked in i.e. a respondent would between January to March, if they reported not being em- only be asked if they had worked in March if they report- ployed in March either. CBPS BRIEF 3, POVERTY AND EQUITY GP 8 THE FEMALE LABOR FORCE REPORTS HIGHER RATES OF ACTIVE EMPLOYMENT THAN MEN, WHO SEEM TO HAVE FACED PROGRESSIVE JOB LOSSES SINCE JANUARY 2020. 22 percent of the female labor force in camps re- ly since January, 2020. Job losses in camps in ports actively working for pay in the 7 days pri- March, 2020 are possibly related to reduced op- or to the survey (April-May 2020) as opposed to erations in camps due to COVID-19 which came the male labor force who not only report lower into effect officially at the same time as the gov- shares of active employment (15 percent)8 but ernment lockdown at the end of March9. also higher losses in employment progressive- Figure 7: Share of camp respondents by work status since january, 2020 Not worked since Jan Camp - female Worked between Jan-Mar Camp - male Worked in Mar Employed but did not work in Apr All Rohingya Actively worked in April HIGHER SHARES OF ACTIVE EMPLOYMENT IS RELATED TO SIMILAR FACTORS AS F O R H O S T S : E N G A G E M E N T I N H O M E - B A S E D , S M A L L S C A L E A G R I C U LT U R A L O R S E L F R E L I A N C E A C T I V I T I E S A S O P P O S E D T O O U T- O F - H O M E W A G E L A B O R . Figure 8: Share of wage and non-wage Baseline findings showed that women in camps employment among those who are active or are more likely to be working on their own ac- temporarily absent count (non-wage) and largely on home-based activities such as vegetable gardening, hens/duck 15% rearing and tailoring, while men were more likely 30% 31% to be engaged in a diverse set of non-agricultural wage labor activities outside the house10 . 62% Male workers in camps are also more dependent on wage labor, compared to women. The share of wage and non-wage work among the tempo- 85% rarily absent and unemployed camp population 71% 69% suggests that job loss has been high for those who were previously engaged in wage work. The 38% ban on cash for works programs and reduced NGO operations due to COVID-19 were therefore more likely to affect male employment rates. On Camp Male Camp Female Camp Male Camp Female the other hand, women in camps, are largely en- Actively working Temporarily absent gaged in self-reliance activities for consumption or Unemployed and income generation, which are often also fa- Wage Non-wage cilitated by humanitarian programs 8 Bangladesh’s local economy started experiencing impacts of the 9 Since 25 March, all non-critical operations in camps have been COVID-19 crisis in early to mid-March, with the first case being re- suspended or reprogrammed including Complementary Food ported on 7 March. The Government lockdown was initiated on 26 Voucher, Farmers’ Markets, Self-reliance Support, Livelihood Sup- March, which continued throughout the months of April-May. The port, and Shelter/NFI activities – which were the main sources of survey was conducted in mid-April and assumes 1 March as the income for the Rohingya population in camps. cut-off date to segregate pre- and post-COVID-19 impacted time 10 Cox’s Bazar Panel Survey Brief 3: Work and Wages in Cox’s Bazar periods. Labor trends prior to 1 March are thus indicative of mar- (page 7). ket volatility attributed to factors unrelated to this crisis.