TheWorld Bank A p r i l PREMnotes 2 0 0 4 n u m b e r 8 4 Gender Minimizing negative distributional effects of public sector downsizing Downsizing has different impacts on men and women. Negative distributional consequences can be minimized using the Downsizing Options Simulation Exercise (DOSE) developed by the World Bank. Downsizing programs are an important part of shows that, in most cases, women suffer many public sector reforms supported by the greater losses in both income and non- World Bank. Although these programs can monetary benefits. reduce budget deficits and address inefficien- cies caused by state-led development strategies, Why is one sex disproportionately laid off? many observers are concerned about the polit- In most programs, laid-off workers are over- ical and social consequences of mass layoffs whelmingly male or female, rather than In most cases, as well as the disproportionate share of the evenly distributed. This is because: losses that some groups of workers may bear. · Female workers tend to be concentrated women suffer This note examines the differing possible in the public sector. impacts of downsizing on male and female · Within the public sector, male and female greater income employees and the consequences for house- workers tend to work in different depart- holds and the economy at large. After dis- ments. and nonmonetary cussing why the distributional consequences · Within departments, male and female work- of downsizing are important, the note intro- ers tend to have different types of jobs, con- losses from duces a simple tool that can be used in the tracts, and levels of seniority. design of downsizing programs to minimize In many developing countries, women are downsizing negative distributional consequences. Al- more likely than men to be employed in the though this note focuses on the differing public sector, even where there may be a effects of downsizing on men and women, sim- larger absolute number of male public sec- ilar concerns apply to other categories of work- tor employees. Female overrepresentation ers (such as different ethnic groups). This tool in the public sector is due to a variety of fac- can easily be adapted to minimize the nega- tors. In some countries, such as Egypt, hir- tive distributional consequences for other ing discrimination makes finding a formal groups as well. private sector job very difficult for women (Assaad and Arntz 2003). Many governments, Downsizing affects women and on the other hand, have explicit or implicit men differently affirmative action policies to increase the Downsizing can have differing effects on male number of women hired. In addition, female and female employees for two reasons. First, workers in developing countries may prefer layoffs are rarely distributed evenly among to work in the public sector rather than in male and female employees. Second, there the private sector. This is because the pub- tend to be systematic differences in the wel- lic sector tends to offer greater sex-specific fare loss experienced by laid-off men and benefits, such as maternity leave, as well as women. Experience in a range of countries higher wages. In fact, the public sector wage from the development economics vice presidency and poverty reduction and economic management network Table 1 Ratios of public sector to private sector wages man, Canagarajah, and Younger 1996). Pre- sumably, the same lack of satisfactory employ- Country Female workers Male workers ment opportunities also contributes to Egypt, civil service 1.63 1.22 individuals dropping out of the labor force. Egypt, state-owned enterprises 1.97 1.68 Ethiopia 2.54 1.74 Again looking at Vietnam, 42 percent of laid- Uganda 1.13 1.05 off women left the labor force, compared to 36 percent of men (Rama 2002). Source: Appleton, Hoddinott, and Krishnan 1999; Assaad 1999. Those men and women who do find new premium--the ratio of public sector wages jobs generally suffer a loss in earnings as to private sector wages--is often higher for they move into the private sector. Because women than for men (table 1). the public sector wage premium is generally Women who are laid Within the public sector, men and women larger for women than for men (see table 1), tend to work in different departments and their loss in earnings also tends to be larger. off from public sector ministries. Men dominate in the military The wage ratios in table 1 do not control and in heavy industry state-owned enter- for productivity differences or level of edu- jobs often have more prises. Women often dominate in education, cation, but studies that control for these fac- health, and light industry enterprises. tors also find evidence of greater wage difficulty than their Because downsizing programs tend to focus discrimination in the private sector, causing on specific parts of the public sector, they women to experience a greater loss in earn- male colleagues in usually involve disproportionate layoffs of ings than men. In Vietnam, even after the one sex. For example, Vietnam's retrench- study controlled for productive character- finding a new job ment program of the early 1990s focused on istics, women in the public sector earned 18 state manufacturing. As a result, 70 percent percent less than men, while in the private of laid-off workers were female. In contrast, sector they earned 26 percent less (Rama Vietnam's more recent round of layoffs is 2002). In their Uganda study, Appleton, aimed at transportation, construction, and Hoddinott, and Krishnan (1999) divided energy--sectors dominated by men. the wage gap into the part explained by pro- Within departments, workers with low ductive characteristics and the unexplained seniority and low skills--characteristics com- remainder, which may represent discrimi- mon among female employees--are often nation. They found that in the public sec- selected for layoffs. In Ghana's downsiz- tor, 59 percent of the wage gap remained ing program of the early 1990s, women unexplained, while a far higher 96 percent accounted for more than half of those laid remains unexplained in the private sector. off because they tended to have fewer years The most direct evidence of the relative of experience in the civil service (Alder- loss in income for men versus women comes man, Canagarajah, and Younger 1996). Sim- from studies of public sector downsizing. ilarly, women predominate among workers The left-hand panel in table 2 shows the on short-term and temporary contracts. Dis- determinants of earnings loss in four pub- crimination may also play a role, because lic sector downsizing programs. In both of assumptions about families' reliance on Ecuador and Ghana, women suffered a a woman's income or about her dedication greater earnings loss even after controlling to the job versus her family. for worker characteristics. In Ecuador, a woman's earnings loss was 30 percentage Why are layoffs harder on women? points higher than that of a comparable Women who are laid off from public sector male colleague (Rama and MacIsaac 1999). jobs often have more difficulty than their In contrast, in Slovenia, women suffered a male colleagues finding a new job. In Viet- smaller loss than men and, in Turkey, their nam, a year after layoffs, only 14 percent of loss was the same. women had found new jobs, compared to As noted earlier, one reason that women 22 percent of men (Rama 2002). The num- are overrepresented among public sector bers are even more extreme in Ghana (Alder- workers is that the private sector generally PREMnote 84 April 2004 provides fewer nonmonetary benefits, such ties or leave the labor force altogether. as maternity leave, daycare facilities, and Research indicates that this is what happened flexible work hours. The right-hand panel in rural areas of Egypt after economic reforms in table 2 shows the determinants of total eliminated guaranteed civil service employ- welfare loss (including loss of benefits). ment for graduates of secondary and higher In Egypt, women's total welfare loss was 85 education institutions. While educated men percent higher than men's, and in Ecuador who were previously eligible for guaranteed and Turkey, it was the same. public sector employment have found jobs in the private sector, their female counter- Do the differences matter? parts have moved into very low-productivity There are at least two reasons to care about subsistence agriculture or have left the labor the disproportionately negative effects on market (Assaad and Arntz 2003). Poorly designed women of public sector downsizing programs. First, the income of female workers is impor- How can negative distributional downsizing programs tant to household welfare. Although male consequences be minimized? workers tend to contribute more to house- Most downsizing programs offer severance can result in an hold income than do females, the relevant pay to make layoffs politically and legally issue is the income lost. The evidence cited acceptable and, often, to encourage workers economy moving above suggests that laid-off female public sec- to volunteer for them. Traditional severance tor employees may spend more time unem- packages are typically based on fairly arbi- from one type of ployed and the gap between the wages earned trary multiples of wages and seniority, with in their public sector job versus their private the resulting amount bearing little relation labor misallocation sector job will generally be larger. Further- to the true losses that workers experience. In more, studies from a diverse set of countries contrast, more tailored packages compen- to another indicate that increasing a woman's share of sate workers in proportion to the welfare loss household income (controlling for total they experience from separation by taking household income) significantly increases into account the determinants of workers' the share of the household budget allocated losses. As table 2 illustrates, wages are not nec- to children's education, health, and nutri- essarily a significant determinant of loss, while tion-related expenditures (World Bank 2001). being female and having a lower level of edu- Second, disproportionate layoffs of cation are often as important in determin- women, coupled with private sector discrim- ing loss as seniority. Tailored packages use a ination, can result in an economy moving compensation formula that takes into from one type of labor misallocation (an over- account at least some of these significant sized public sector) to another type of labor determinants. misallocation, in which educated female work- When compensating women directly for ers are forced into low-productivity activi- their greater losses from separation is not Table 2 Determinants of welfare loss from downsizing Total welfare loss Earnings loss (earnings and benefits) Ecuador, Ghana, Turkey, Ecuador Egypt, Turkey, Worker central civil cement and central civil cement and characteristic bank service Slovenia oil workers bank service oil workers Female + + -- 0 0 + 0 Public sector wage 0 0 n.a. + 0 n.a. 0 Seniority + 0 + 0 + ? 0 Education -- 0 0 -- -- + -- Note: Statistical significance is indicated by + or ­, while 0 indicates a nonsignificant coefficient and ? indicates a change in sign across specifications or groups of workers. When a variable was not included in the analysis, n.a. is reported. Source: Adapted from Rama 1999. PREMnote 84 April 2004 politically feasible, including a substantial from Egypt." Research Project Report. lump-sum component in the severance Economic Research Forum for the Arab package offered to all workers may be a Countries, Iran, and Turkey. Cairo, Egypt. solution. Simulations for Vietnam, using Alderman, Harold, Sudarshan Canagarajah, the Downsizing Options Simulation Exer- and Stephen Younger. 1996. "A Com- cise (DOSE) tool developed by the World parison of Ghanaian Civil Servants' Earn- Bank, suggest that, in comparison to pack- ings before and after Retrenchment." The DOSE tool ages based only on wages and seniority, a Journal of African Economies 4 (2): 259­88. severance package involving a large lump- Rama, Martín. 1999. "Public Sector Down- helps policymakers sum payment is less likely to penalize women sizing: An Introduction." The World Bank (Rama 2002). Economic Review 13 (1): 1­22. minimize negative DOSE helps policymakers examine the ------. 2002. "The Gender Implications of distributional consequences of both tradi- Public Sector Downsizing: The Reform distributional tional and tailored severance packages by Program of Vietnam." The World Bank simulating the acceptance rates that would Research Observer 17 (2): 167­89. consequences be achieved by various packages (assuming Rama, Martín, and Donna MacIsaac. 1999. layoffs are voluntary). The classification of "Earnings and Welfare after Downsizing: "stayers" and "leavers" helps identify which Central Bank Employees in Ecuador." The groups of workers may be overcompensated World Bank Economic Review 13 (1): or undercompensated by the packages, 89­116. allowing policymakers to minimize the neg- World Bank. 2001. Engendering Development-- ative distributional consequences of down- Through Gender Equality in Rights, Resources, sizing operations. Predicting which workers and Voice. A Policy Research Report. New will leave also allows policymakers to tailor York: Oxford University Press. retraining and reemployment programs to the needs of specific groups. This note was written by Susan Razzaz (Senior Economist, Social and Economic Development Further reading Group, Middle East and North Africa Region). The Appleton, Simon, John Hoddinott, and author is grateful to Karen Mason, Nayantara Pramila Krishnan. 1999. "The Gender Mukerji, Gurushri Swamy, and Alexandra van Wage Gap in Three African Countries." Selm for useful comments. The DOSE tool is avail- Economic Development and Cultural Change able at http://gender.worldbank.org/economics/ 47: 289­312. admin.htm. For more information about DOSE, Assaad, Ragui. 1999. "Matching Severance contact Martin Rama at mrama@worldbank.org. Payments with Worker Losses in the If you are interested in similar topics, consider Egyptian Public Sector." The World Bank joining the Gender and Development Thematic Economic Review 13 (1): 117­54. Group (contact Wendy Wakeman, x33994) or the Assaad, Ragui, and Melanie Arntz. 2003. Administrative and Civil Service Reform The- "Constrained Geographical Mobility and matic (contact Ranjana Mukherjee, x34301). Gendered Labor Market Outcomes For more information, click on Thematic Groups under Structural Adjustment: Evidence on PREMnet. This note series is intended to summarize good practices and key policy findings on PREM-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank. PREMnotes are widely dis- tributed to Bank staff and are also available on the PREM Website (http://prem). If you are interested in writing a PREMnote, email your idea to Madjiguene Seck. For additional copies of this PREMnote please contact the PREM Advisory Service at x87736. PREMnotes are laid out by Suzanne Luft. Prepared for World Bank staff