Pay period and the distributional effect of education on earnings : evidence from recentered influence function

This paper employs Recentered Influence Function (RIF) regressions to examine the distributional effect of education on earnings in East Africa, using data from the Living Standards and Measurement Study (LSMS) for Malawi, Tanzania, and Uganda. Taking into consideration the pay period of the workers, the paper investigates how education affects earnings at various points of the earnings distribution; how education affects earnings inequality; and how much of the gender gap in earnings can be attributed to gender differences in educational attainment. Results show heterogenous effects of education on earnings along the earnings distribution in all pay periods. Generally, the effect is more substantial for workers reporting monthly earnings than their daily and weekly counterparts. The results also show that in each pay period, there is significant wage inequality between workers in the top decile of earnings and those in the bottom four deciles. Education can either increase or reduce wage inequality depending on the period in which the worker is paid, i.e., education is associated with an increase in inequality for workers paid weekly and reduced inequality for those paid daily and monthly. Significant gender gaps in earnings were found for Tanzania and Uganda but not for Malawi. The decomposition results suggest that gender differences in educational attainment significantly explain the gender wage gap for Tanzania and Uganda. The paper recommends, rather than pooling, researchers should obtain separate estimate for each period for more reliable results.