DT 08 - 24 Female selection into employment along the earnings distribution

In Uruguay, women’s employment rates have increased over recent decades, mostly driven by the increase of labour supply of women in couples. However, a significant gender employment gap remains, which reflects the need of correcting for sample selection in empirical wage gap studies. Recent literature studying gender wage gaps have highlighted the importance of correcting for selection into employment along the earnings distribution.

In this study, I estimate the evolution of the gender gap in earnings along the earnings distribution, correcting for selection into employment. Based on the Uruguayan household surveys, Encuesta Continua de Hogares, for the period 2009-2019, I apply the three-step quantile selection model proposed by Arellano and Bonhomme (2017) to estimate the selection-corrected hourly earnings distributions. I use a measure of potential out-of-work income as an instrument to correct for selection into employment. Results show that selection patterns vary across marital statuses. Potential earnings gaps are greater than the uncorrected (raw) earnings gap for individuals in couples in all earnings quantiles, albeit maintaining a decreasing trend over the studied period. The difference between both earning distributions is larger for lower earnings quantiles, suggesting the existence of ’sticky floors’. Lastly, when considering married and cohabiting individuals separately, I find that women’s selection into employment is driven by the selection of married women.

Keywords: gender wage gaps, sample selection, quantile regressions, glass ceiling, sticky floors