IZA DP No. 193: The Determination of Wages and the Gender Wage Gap: A Survey
substantially revised version published as 'Gender wage gap studies: Consistency and decomposition' in: Empirical Economics, 2008, 35 (1), 63-76
In this paper the extensive empirical literature on the gender wage gap is reviewed with particular attention given to the identification of the key parameters in the specified human capital wage regression models. This aspect has been of great importance in the literature chiefly for two reasons. On the one hand, the main explanatory variables in the wage model, i.e. measures of work experience and time out of work, are endogenous and, hence, applying traditional estimators may lead to inconsistent parameter estimates. On the other hand, empirical evidence on the gender wage gap hinges on the estimates of the main parameters of interest and its economic meaningfulness may be limited by restrictive assumptions imputed on the wage model. The survey shows that econometric methods are still more advanced than their applications, and that in applications consistency often is only achieved at the expense of restrictive assumptions that are dubious from an economic perspective. In short, it seems that current measures of male-female wage differentials are likely to be biased because of the failure to appropriately account for endogeneity and selectivity in the wage regression models.