Talk
Male and Female Wage Functions: A Quantile Regression Analysis using LEED and LFS Portuguese Databases
Maria da Conceição Torres Figueiredo (Figueiredo, M. C.); Elsa Fontainha (Elsa Fontainha);
Event Title
3rd LINKED EMPLOYER EMPLOYEE DATA WORKSHOP - LEED 2013
Year (definitive publication)
2013
Language
English
Country
Portugal
More Information
Abstract
The research aims to study the distribution of hourly wages for men and women in Portugal, adopting a quantile regression (QR) approach. Two databases are used for the estimation of the wage functions: the Quadros de Pessoal, Linked Employer-Employee Data (QP-LEED) and the Inquérito ao Emprego, Portuguese Labour Force Survey (IE-LFS). Three basic models are considered to explain the hourly wages for men and women: the first model is estimated with the same specification (independent variables: education, tenure, potential experience, activity sector, job) using alternatively both databases (QP-LEED and IE-LFS); the second, using data from LEED includes additional determinants related to firm (firm size and foreign social capital); and the third, using data from the IE-LFS includes additional independent variables related to the worker's family (marital status and children). The results indicate that: (i) Regardless of the database used, the quantile regression (QR) shows the superiority over OLS approache; (ii) In general, the same model specification estimated using two alternative databases, one administrative (QP-LEED), and another based on a survey (IE-LFS) present convergent results; (iii) Independently from the database used, the equations for men and for women, reveal that the levels of education have a higher impact on wage determination; (iv) In general, the variables related to the firm contribute to the explanation of wages of men and women while the tested variables related to family only contribute to the explanation of men's wages; and (v) the clear different returns to the same characteristics found between men and women, and the pattern of those differences which increase across quantiles strongly recommends that the present study be deepened in the future, with the analysis of the explanation of the differences.
Acknowledgements
--
Keywords
wage functions; quantile regression; Linked Employer-Employee Data; Labour Force Survey; male-female wage differences