In an economic crisis, whose job do employers put on the chopping block first? Many gender equality advocates and policymakers are concerned that “women are at risk of being hired last and dismissed first” during crises. This concern is fuelled by evidence showing that employers often discriminate against women even during less volatile times, that women often bear the brunt of coping with economic shocks, and that, in many countries, gender norms prioritize men’s employment over women’s. Despite a lot of rhetoric, existing studies of the labor market consequences of macroeconomic crises have yielded ambiguous conclusions about the differential impact across genders. Might claims about women’s vulnerability be exaggerated?
Most studies that look at the distributional impact of crises rely on household and labor force data. However, these data cannot distinguish between two mechanisms that could account for gender differences in employment adjustment. First, differences in vulnerability could be the result of sorting by gender into firms and occupations that differ in their vulnerability to crises. In this case, the effect of gender is indirect; women may take jobs that are relatively more or less vulnerable. Second, there could be differential treatment of men and women workers within the same firm. Faced with the need to adjust, do employers treat women differently, either by firing them first or cutting their wages more? It is this second mechanism that underpins concerns about discrimination. To distinguish between these mechanisms, we need to compare the employment prospects and wage trajectories of men and women both across and within firms—which means we need firm-level data.
To complicate matters a little further, the sorting and differential treatment effects could potentially work in opposite directions. For example, women may face greater differential treatment effects, but sort into jobs that are less vulnerable. If so, aggregate employment numbers may obscure the true extent of women’s vulnerability within a firm.
In a recent paper  I coauthored with Bob Rijkers and Andrew Waxman, we use Indonesia’s census of manufacturing firms to analyze employer responses to the 1997-98 East Asian crisis and to identify the importance of these two mechanisms. The results show that women indeed faced disproportionate reductions in employment and wages within firms. However, these differential treatment effects were relatively small. They were also more than offset in the aggregate by sorting: women were more likely to be employed in firms that suffered less in the crisis, with the net result that women fared better than men. Even though 45% of all manufacturing jobs were held by women, women’s net job losses accounted for only a third of all the jobs lost during the crisis.
The reason for these favorable sorting effects is that women’s employment was concentrated in labor-intensive export industries that were more resilient to the crisis. Sorting by gender also occurred along occupational lines; men suffered more because they were more likely to be working in white-collar jobs, which were hit harder. On the other hand, female blue-collar workers appear to have experienced slightly larger wage losses than their male colleagues because they were employed in firms that reduced wages more. Interested readers can find detailed descriptions of how the relative importance of these mechanisms varies across different types of firms and workers in different occupations in our paper .
Existing studies of the impact of crises on the labor market that rely on labor force and household-level data focus on labor supply. However, crisis-induced changes in labor market outcomes are predominantly driven by changes in labor demand; labor supply responses are typically muted compared to the large contractions in labor demand that are a defining feature of crises. This analysis illustrates how firm level data can benefit our understanding of who is at risk of losing their job and why. Indeed, perhaps the most important lesson from this analysis is that employers’ vulnerability and responses to major shocks are important, but often overlooked, determinants of the distributional impact of crises. In the case of Indonesia in the East Asian crisis, women were more vulnerable within a firm, but where you worked mattered more than your gender.
In the current financial crisis—as in the East Asian crisis—it's also important to take into account both sorting and differential treatment before jumping to any conclusions. In the current crisis, the sorting channel is also likely to be important. In countries where the construction sector has been hardest hit, the sorting effect will mitigate the impact on women. However, in countries with large exports of light manufacturing, the sorting by gender is likely to render the women more vulnerable.
Hallward-Driemeier, Mary, Bob Rijkers and Andrew Waxman. (2011) “Ladies First? Firm-level Evidence on the Labor Impacts of the East Asia Crisis .”
World Bank Group staff can also join the upcoming FPD Academy talk by Mary Hallward-Driemeier—details will be posted on the FPD Academy page  when they become available.