54296 POVERTY THE WORLD BANK REDUCTION AND ECONOMIC MANAGEMENT NETWORK (PREM) Economic Premise APRIL 2010 · Number 11 Fewer Jobs or Smaller Paychecks? Labor Market Impacts of the Recent Crisis in Middle-Income Countries Gaurav Khanna, David Newhouse, and Pierella Paci This note1 presents early evidence on the labor market impacts of the recent economic crisis in 41 middle-income countries.2,3 A broader geographic coverage is prevented by the lack of high-frequency labor market data in other middle-income countries and in the low-income countries. Whereas the economic downturn has threatened recent progress in enhancing employment opportunities, the impact has fallen disproportionately on the quality of employment rather than on the number of jobs. Slower growth in earnings accounts for nearly three quarters of the total adjustment for the average country. The bulk of the earnings adjustment was driven by a reduction in working hours, as well as a shift away from the better-paid industrial sector. Evidence of the adjustment's nature and magnitude suggests a policy package that combines (1) income maintenance programs--that is, cash transfers to low-paid poor workers; (2) interventions that facilitate flexible-hours arrangements; and (3) innovative policies that provide workers access to income maintenance mechanisms to compensate for temporary reductions in standard working hours--for example, by granting partial compensation from the unemployment benefit system or by providing paid training opportunities. Workers Were Hit Hard Figure 1. Trends in the Growth Rates of Main Indicators across Time 12 GDP growth Figure 1 shows how recent positive trends in the growth of 10 employment growth the real wage bill4--that is, of the aggregate labor income in 8 real earnings growth real terms--have stalled since the beginning of 2008. The ini- percentage points 6 wage bill growth tial fall in wage bill growth resulted from increased inflation 4 caused by the food and fuel crisis, and continued after the sharp downturn in GDP in the last quarter of 2008.5 Nearly 2 every country was affected because the wage bill growth fell 0 in 25 of the 28 countries for which data are available. ­2 On average, the wage bill growth rate fell by 8 percentage ­4 points; but the magnitude of the slowdown varied consid- ­6 erably across countries (as shown in figure 2). The impact 7 07 07 07 8 08 08 08 9 09 09 00 00 00 20 20 20 20 20 20 20 20 .2 .2 .2 was particularly severe in the wealthier countries and those n. p. c. n. p. c. n. p. ar ar ar De De Ju Se Ju Se Ju Se M M M with fixed exchange-rate regimes.6 In line with the severe GDP downturn, the impact was especially acute in the month and year countries of Eastern Europe and Central Asia, but also in Source: Authors' calculations. 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Figure 2. Crisis-Related Change in Wage Bill Growth, Selected Countries Figure 3. Change in Employment Growth, Selected Countries Latvia Bosnia and Herzegovina Ukraine Latvia Lithuania Lithuania Russian Federation Bulgaria Sri Lanka Dominican Republic Serbia South Africa Mexico Russian Federation Armenia Jamaica Turkey Mauritius Romania Ukraine Poland Egypt, Arab Rep. Ecuador (urban) Serbia Venezuela, R. B. de Moldova country Bulgaria Poland South Africa Mexico Colombia (urban) Kazakhstan Belarus Peru Moldova Chile Indonesia Belarus Thailand Colombia (Urban) country Peru Indonesia Brazil (urban) China (urban) Philippines Brazil (urban) Kazakhstan Venezuela, R. B. de Chile Georgia China (urban) Romania Argentina (urban) Argentina (urban) Macedonia, FYR Armenia Ecuador (urban) 15 10 5 0 ­5 ­10 ­15 ­20 ­25 ­30 ­35 Kyrgyz Republic percentage points Thailand Source: Authors' calculations. Trinidad and Tobago Macedonia, FYR Malaysia Philippines Mexico and Sri Lanka. By contrast, the wage bill continued Albania to grow at an accelerating rate in Argentina, China, and the Tajikistan former Yugoslav Republic of Macedonia, thanks to sharp Montenegro drops in the growth of consumer price indexes and corre- Sri Lanka sponding reductions in the growth of real earnings. Morocco Turkey But the Number of Jobs Changed Little 5 0 ­5 ­10 ­15 ­20 percentage points Employment growth also slowed in more than three quar- Source: Authors' calculations. ters of the countries (figure 3). On average, however, the de- cline was moderate (2.1 percentage points in the 41 countries sampled); and employment continued to grow in The Bulk of the Burden Fell on Earnings 60 percent of the countries. The reduction in employment growth was particularly severe in Eastern Europe and Cen- Workers suffered most from the sharp slowdown in monthly tral Asia, where eight countries experienced reductions of 2 earnings growth.7 As evident from figure 1, earnings growth percentage points or more. In contrast, employment growth began to slow down in early 2008, as the food and fuel crisis in the Eastern Asia and the Pacific region was barely af- increased inflation and eroded workers' real earnings. For the fected, declining by only 0.1 percentage points. Differences average country, earnings growth fell 5.4 percentage points. based on currency regimes are also stark: on average, coun- For the average country, this decline amounted to nearly tries with fixed currency regimes saw a decline of 1.7 per- three quarters of the total slowdown in wage bill growth centage points, compared with only 0.4 percentage points (figure 4). Earnings played a markedly smaller role in some in countries with floating rates. countries, however--for example, in South Africa and in the 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Figure 4. Decomposition of the Change in Wage Bill Growth and the Change in Earnings Growth a. Contribution to change in wage bill growth b. Contribution to change in savings growth due to wages 26% percent percent due to earnings due to hours percent due to employment 74% ­20 0 20 40 60 80 100 120 Source: Authors' illustrations. mid-range, middle-income countries. Moreover, the drop in come particularly important. If, as in the current context, the earnings was small, compared with that experienced in pre- reduction in earnings is mainly driven by a reduction in hours vious crises: wages dropped by more than 40 percent in worked, effective interventions include innovative policies Mexico, the Russia Federation, and Indonesia, and by 28 per- that offer workers access to income maintenance mecha- cent in Romania in past crises. nisms to compensate for temporary reduction in standard working hours--for example, granting partial compensation Earnings Declined through a Combination from the unemployment benefit system or providing paid of Reduced Hours and Shifts to training opportunities. In addition, targeted income mainte- Less-Well-Paid Sectors nance programs--that is, cash transfers to low-paid poor workers--also would protect the livelihoods of the most vul- The decline in the number of hours worked was a major fac- nerable households from long-term deterioration. Several tor in the reduction in earnings.8 Hours worked fell by 5 per- policies implemented by countries of the Organisation for cent or more in one third of the 14 countries for which data Economic Co-operation and Development--such as partial are available, and increased only in Uruguay and in the West unemployment insurance, expanded cash transfers to poor Bank and Gaza. workers, and temporary wage subsidies--may be priority in- Earnings growth also fell because of the shift in employ- terventions in those countries where hours and earnings ad- ment away from the traditionally better-paid industrial sec- justments dominated. tor. This shift occurred in most countries; but was especially significant in East Asia, and in the Eastern Europe, and Cen- The Political Economy of Interventions tral Asia region where the share of industrial employment declined by 0.7 and 0.6 percentage points, respectively. This Most developing countries facing a labor market crisis face shift contributed to reduced productivity and earnings. three policy constraints. First, monitoring systems and high- frequency labor market indicators are usually lacking, and de- Why This Matters for Policy Makers cisions are often made against a backdrop of extreme uncertainty about who is hit the hardest and how. Second, The magnitude and nature of the adjustment have important the fiscal space for policy intervention is often narrow and implications for the design of policy interventions. The mode shrinking, and many countries have a limited scope for such of adjustment determines the winners and losers and how responses. Third, policy makers may be confronted by insti- they are affected. For example, employment declines tend to tutional and political economy constraints, such as few pre- concentrate losses among an unlucky few who lose their jobs. existing social insurance mechanisms on which to build, In this case, income replacement programs are effective ways limited administrative capacity, and little maneuvering space of compensating them for the loss. These programs include to pursue economically optimal but politically unpopular re- unemployment insurance or, in countries with low institu- forms. tional capacity, public works. In contrast, when the labor mar- Policy makers also may face two thorny trade-offs. First, ket adjustment occurs through earnings, losses are spread policies that ease the short-term effects on employment may more widely across the workforce and income maintenance destroy incentives for long-term recovery, especially during programs--such as cash transfers or income tax credits--be- a prolonged structural crisis involving significant sectoral re- 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise allocation. Second, limited resources mean that policy mak- group ranges from Malaysia, with $5,000 GDP per capita, ers also may have to choose between support for those peo- to Trinidad and Tobago, with $10,100. For more details on ple who are most directly affected by the crisis (typically the findings, see Khanna, Newhouse, and Paci (forthcoming). urban-based exporters and those involved in construction After controlling for the drop in GDP growth, the differ- and manufacturing) and interventions directed to chroni- ences in labor market slowdowns across regions are minor. cally poor groups.9 7. Thirty-one countries have data on average monthly earnings, and 28 of these countries have data on employ- Notes ment as well. The countries with earnings growth data (in decreasing order of severity) are Latvia, Sri Lanka, Ukraine, 1. This note draws on Khanna, Newhouse, and Paci (forth- Lithuania, Russia, Serbia, Mexico, Turkey, Romania, Arme- coming). nia, República Bolivariana de Venezuela, Poland, Paraguay, 2. The countries included in the analysis (in decreasing Colombia (urban), Belarus, Thailand, Indonesia, the Philip- order of impact on GDP growth) are Armenia, Ukraine, pines, South Africa, Ecuador (urban), Brazil (urban), Bul- Latvia, Lithuania, Russia, Georgia, Turkey, Romania, Thai- garia, Moldova, Peru, Kazakhstan, Chile, West Bank and land, Mexico, Dominican Republic, Bulgaria, Malaysia, Gaza, Uruguay, Argentina (urban), China (urban), and FYR Paraguay, Peru, Kazakhstan, Kyrgyz Republic, Moldova, Macedonia. República Bolivariana de Venezuela, Serbia, Belarus, Brazil 8. Fourteen countries have data on hours worked. (urban), Colombia (urban), Chile, FYR Macedonia, South 9. For a more detailed discussion of how to best deal with Africa, Jordan, Argentina (urban), Uruguay, China (urban), these trade-offs, see the forthcoming Economic Premise titled Poland, the Philippines, Ecuador (urban), Sri Lanka, Mauri- "Coping with Crises: Policies to Protect Employment and tius, Jamaica, the Arab Republic of Egypt, Indonesia, West Earnings." Bank and Gaza, Albania, and Morocco. The following coun- tries have labor market data, but not data on GDP growth: Bosnia and Herzegovina, Montenegro, Tajikistan, and Reference Trinidad and Tobago 3. Results should be interpreted with appropriate cau- Khanna, Gaurav, David Newhouse, and Pierella Paci. Forthcoming. "Fewer tion, given differences in data sources across countries. For Jobs or Smaller Paychecks? Labor Market Impacts of the Financial Crisis in Middle Income Countries." In The 2007­09 Financial Crisis: Labor example, in one quarter of the countries, employment data Market Impacts and Policy Responses, ed. D. Newhouse, P. Paci, and D. cover only formal workers; and earnings data ignore self-em- Robalino. Washington, DC: World Bank. ployed workers in one third of the countries. In some Latin American and Caribbean countries, data cover only urban About the Authors areas. Despite these inconsistencies, the data provide a useful general indication of recent trends. Gaurav Khanna is a consultant with the Poverty Reduction and 4. The real wage bill is defined as the product of total em- Equity Group (PRMPR) in the Poverty Reduction and Eco- ployment and average earnings. nomic Management Network at the World Bank; David New- 5. For more details, see Khanna, Newhouse, and Paci house is a labor economist with the Human Development (forthcoming). Network­Social Protection at the World Bank; and Pierella 6. Countries are classified on the basis of their precrisis Paci is lead economist with the PRMPR. per capita GDP. In the poorest group of countries, the per capita GDP ranges from $230 (Tajikistan) to $1,800 The note was jointly prepared by the Poverty Reduction and (China). The countries at moderate levels of development Equity Group and the Human Development Network Social range from FYR Macedonia, where per capita GDP is Protection Division. $2,000, to Mauritius, where it is $4,700. The wealthiest The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. It is produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at www.worldbank.org/economicpremise.