Gender Wage Gap in Canada: An Analysis using Counterfactual Distributions Regression

Abstract

In this paper, I estimate and decompose the gender wage gap in Canada using a counterfactual distribution approach based on distribution regression methods developed by Chernozhukov et al (2013). Using the Canadian Labour Force survey of October and April 2018, I show that men earn higher wages than women throughout the distribution. The difference in log hourly wage increases when moving up the distribution and drops at the end of the distribution, which the discrimination is higher between 20th and 70th quantiles. In the decomposition, the structural effect constitutes the majority of the gap while the composition effect reduces the gap. When controlled for industry and occupation, the gap is completely due to discrimination. This contrasts with the positive sign of the explained portion in the canonical Blinder-Oaxaca decomposition. Overall, the result of this paper provides new empirical evidence to support discrimination is the cause of gender wage gap.