The Measurement of Wage Discrimination with Imperfect Information: A Finite Mixture Approach

Abstract

Studies on wage discrimination assume that independent observers are able to distinguish a priori which workers are suffering from discrimination. However, this may not be a good assumption when anti-discrimination laws mean that severe penalties can be imposed on discriminatory employers or when unobserved heterogeneity is significant. We develop a wage discrimination model in which workers are not classified a priori. It can be thought of as a generalization of the standard empirical framework, whereas the Oaxaca–Blinder model can be thought of as an extreme case. We propose a finite mixture model to explicitly model unobserved heterogeneity in individual characteristics and estimate the probabilities of being a discriminated or a non-discriminated worker. We illustrate this proposal by estimating wage discrimination in Germany and the UK.

Citation

Prieto-Rodríguez, J., Rodríguez, J.G. and Salas, R. (2020), "The Measurement of Wage Discrimination with Imperfect Information: A Finite Mixture Approach", Rodríguez, J.G. and Bishop, J.A. (Ed.) Inequality, Redistribution and Mobility (Research on Economic Inequality, Vol. 28), Emerald Publishing Limited, Leeds, pp. 187-204. https://doi.org/10.1108/S1049-258520200000028008