• counterfactual
• quantile regression
• urban wage premium
• Labor Market
• wage disparity
• Inequality

The aim of this work is to study, empirically, spatial wage disparities in Italy between the period 2000-2016. Using survey data produced by the Bank of Italy, we apply a counterfactual decomposition analysis in the framework of a quantile regression. The methodology follows the works by Machado-Mata (2005) and by Chernozhukov et al. (2017). Starting from the definition of the Mincerian equation for the determination of the wages level, we investigate the impact of the main covariates on salaries. To consider the spatial dimension, we define two different sub-samples for each wave in the dataset on the basis of population density, as a proxy for the level of urbanization. Therefore, for each wave of survey we run a quantile regression for urban and non-urban areas. The first part of the analysis allows us to determine the “rate of return” of each covariate on the wage’s distribution. It enables the identification of the sources of the wage disparities in Italy. The second part of the analysis consists in a counterfactual decomposition procedure to estimate which distribution of wages would have prevailed in each subset if the covariates were distributed as in the other reference subset. The counterfactual analysis allows us to understand whether coefficients or covariates exert the major effect on wage inequality in Italy. The question we wanted to answer is: is it more important who I am or where I come from? Results show that high education accounts for a considerable amount of disparities in wages among urban and non-urban areas, more than other labor-market and regional-specific covariates. The article goes on suggesting policy implications and potential solutions to address the issue. Moreover, we propose an econometric specification to deal with endogeneity problems, that could bias our results.