Decomposing the gender wage differentials using quantile regression: The case of Lebanon

Abstract

Despite the fact that most developing countries have realized substantial achievements in several major areas of women’s well-being through the past few decades, women are still lagging well behind men in wages (World Bank, 2012).As a matter of fact, it is commonly believed that differences in wages are interpreted as differences in productivity-related human capital factors such as education, seniority, and labor market experience. It is therefore important to investigate the reasons behind this gap because gender inequality in the labor market restrains economic growth (International Labour Organization, 2012).

This paper contributes to the existing literature on gender wage differentials in developing countries in several ways. First, it provides additional evidence from Lebanon, which is a small economy that has received little attention in previous studies. Second, we use two different decomposition methods:  i) the traditional Blinder-Oaxaca (1973) decomposition that divides wage differentials into “explained” component due to the observable human capital factors and “unexplained” component due to discrimination, and ii) Machado-Mata (2005) technique using a quantile regression that decomposes the gender inequality across the wage distribution. Third, we correct for the sample selection bias. This is important since a low fraction of women are working full time in developing countries.

The empirical investigation uses data derived from a representative sample of employees in the banking sector for the years 2008 and 2014. Using Blinder-Oaxaca decomposition, the results reveal that the overall male-female gap is estimated at US$ 5592.413. The results show that the explained component, attributed to differences in workers characteristics represents 22% of the total gap. Differences in education are the most responsible effect of wage differentials. The unexplained component that is due to discrimination represents 78% of the total gap. The results of Machado-Mata method indicate that the impact of human capital characteristics in explaining the total gap is larger at the low and high ranges of the distribution indicating higher wage differentials at the extremes of the distribution. The paper offers insights through these results that may be useful to policy makers in developing countries.