Labour markets in advanced economies react more strongly to economic downturns than those in emerging markets. A key reason is the higher prevalence of self-employment in emerging markets, which helps cushion job losses. This column presents new evidence on how this difference in employment composition shapes labour market responses to economic fluctuations. While self-employment plays a major role in the short run, other structural factors become more important over time. As informality declines, stronger unemployment protections will be needed to support workers.
In 1962, Arthur Okun estimated that a one percentage point decline in output growth raises the US unemployment rate by 0.3 percentage points (Okun 1962). Remarkably stable by macroeconomic standards, this relationship – dubbed ‘Okun’s Law’ or ‘Okun’s coefficient‘ – has become a cornerstone of undergraduate and graduate macroeconomics textbooks (Blanchard and Fischer 1989, Mankiw 2022) and has since been widely estimated for nearly every country (Ball et al. 2017, Ball et al. 2019, An et al. 2022). Several Vox columns have examined distinct features of Okun’s Law. For example, Lim et al. (2019) use a modern labour flows approach and uncover that the law is asymmetric – negative shocks have a bigger impact than positive shocks – while Lim et al. (2016) estimated variations across age and gender groups. Leigh et al. (2013) highlighted that, despite overall stability, the Okun coefficient varies considerably between countries, with limited success explaining this variation. In this column, we specifically explore the stronger cyclicality of employment in advanced economies compared to emerging market and developing economies (EMDEs) seen in Figure 1, by shedding new light on why the Okun relationship varies across these two groups.
Figure 1 Employment rate and GDP around GDP troughs
Notes: Difference in the employment rate and GDP growth relative to trough years. Identification of trough years is from Kose et al. (2020) based on turning points using the algorithm introduced by Harding and Pagan (2002). On the x-axis, -3 to -1 are the years leading up to a GDP trough and years 1 to 3 are the initial years of recovery after a recession ends. Based on a sample including 28 advanced economies and 30 EMDEs, between 1990 and 2019.
Source: Loungani et al. (2025).
A crucial factor explaining differences in Okun’s coefficients between advanced economies and EMDEs is the distinct composition of their labour markets, especially regarding wage employment versus self-employment. In EMDEs, self-employment typically represents more than twice the share of total employment seen in advanced economies and is frequently linked to informal employment arrangements (Figure 2). In our forthcoming paper (Loungani et al. 2025), we show that this difference in labour market composition significantly shapes how employment and unemployment respond to fluctuations in GDP, helping account for much of the observed divergence in Okun coefficients across these country groups.
Figure 2 Prevalence of self-employment
Notes: Circles denotes the average of the self-employment to total employment ratio (based on country averages between 1990-2019 for 58 countries). The midline of the box denotes the median, the top and bottom lines of the box denote the upper and lower quartiles, respectively. Whiskers denote the upper and lower deciles.
Source: Loungani et al. (2025).
Self-employment can shape the evolution of employment over the cycle through a variety of channels. For instance, an incomplete pass-through of productivity shocks in the economy to the self-employment sector can generate a more nuanced response of self-employment to economy-wide fluctuations, or the formal sector may be affected by shocks to the foreign interest rate that freezes the hiring rate, whereas the self-employment sector remains insulated from these shocks (Fernandez and Meza 2015, Leyva and Urrutia 2020). Another strand of literature focuses on the size of the informal sector, showing that the larger the informal sector, the easier workers can transition in or out of informal jobs, helping workers adjust to adverse shocks (Shapiro 2014, Horvath and Yang 2022).
Using a country fixed-effects regression, we estimate average Okun coefficients separately for advanced economies and EMDEs and focus on their differences. We find that a one percentage point reduction in GDP growth raises the unemployment rate by 0.3 percentage points in advanced economies but only by 0.1 percentage points in EMDEs (see Table 1, row 1). Similarly, employment growth is more sensitive to GDP fluctuations in advanced economies compared to EMDEs (see Table 1, row 2).
Table 1 Summary of results
Notes: Summary of estimated elasticities and sample used in Loungani et al. (2025).
Our empirical strategy in attempting to explain the ‘aggregate’ (or ‘headline’) difference between the two groups is as follows:
- First, we estimate Okun’s Law for the self-employment and wage employment sectors for the two groups of countries (see Table, rows 3 and 4 for the estimated elasticities). The similarity in the cyclicality of wage employment is apparent in Figure 3, while self-employment tends to behave differently during recessions in advanced economies and EMDEs.
- Armed with these sectoral Okun coefficients, we then perform a counterfactual ‘calculation’ in which we assume that EMDEs have a similar share of self-employment as in the average advanced economy. This calculation suggests that differences in the sectoral composition of employment can explain the bulk of the difference in the aggregate or headline elasticity noted above.
Figure 3 Wage employment, self-employment, and total employment around GDP troughs
Notes: Difference in wage-, self-, and total employment rate relative to trough years. Identification of trough years is from Kose et al. (2020) based on turning points using the algorithm introduced by Harding and Pagan (2002). On the x-axis, -3 to -1 are the years leading up to a GDP trough and years 1 to 3 are the initial years of recovery after a recession ends. Based on a sample including 28 advanced economies and 30 EMDEs, between 1990 and 2019.
Source: Loungani et al. (2025).
Reaching this conclusion, however, requires some additional qualifications. First, we have to make a distinction between expansions and contractions, as in the Vox column by Lim et al. (2019). Second, we make a distinction between short-run Okun coefficients (reflecting the relationship between unemployment or employment and output over a one-year period) and medium-run Okun coefficients (reflecting the relationship between unemployment and output over a three-year period).
The first qualification allows us to establish that the two groups differ more in their labour market adjustment during recessions than during booms, and it is this fact that generates the difference in the aggregate or headline elasticity that we seek to explain. It is during recessions that the greater prevalence of self-employment in EMDEs plays a sheltering or buffering role, shielding the measured unemployment rate from rising much in the face of an adverse output shock (see Table 1, rows 5-8 for the estimated elasticities during recessions).
The second qualification allows us to establish that self-employment plays a bigger role in the short run than in the medium run. Using counterfactual calculations described above we find that self-employment explains about 70% of the difference in the short-run Okun coefficient but only 40% of the difference in the medium-run Okun coefficient, suggesting that other structural factors such as differences in the industry composition of formal employment and differences in labour market institutions play a bigger role in the medium run than in the short run (Figure 4).
Figure 4 Explanation of the difference between AE and EMDE employment elasticities
Notes: Share of the difference between Advanced Economies (AE) and EMDE employment elasticities explained by sectoral composition of employment or other factors.
Source: Loungani et al. (2025).
Our forensic work explains why previous seemingly straightforward attempts to estimate cross-section relationships between Okun coefficients for countries and various structural characteristics of countries have met with limited success. On the right-hand side, the estimates are flawed by the failure to distinguish recessions from expansions and the horizon (short versus medium). And the empirical relationship is further confounded by the fact that the importance of the right-hand side variables vary by horizon.
Policy implications
Our results should not be taken to suggest that EMDEs should not continue to reduce informality. In our view, the long-run benefits of larger formal sectors (such as a broader and more stable tax base, among many others) outweigh the short-run benefits of buffering unemployment in the short run that is provided by a large informal or self-employment sector. However, as the protection provided by informality wanes, countries should move toward more formal and generous unemployment insurance systems to protect workers better from the consequences of higher unemployment, as advocated by Duval and Loungani (2019).
Source : VOXeu
