In this paper, statistical learning techniques with particular focus on LASSO, are proposed to create Counterfactual Series and compare them with the actual series, in order to see if a difference is present and to assess the direction and magnitude. Previous methodology is applied to study some labor economics cases, comparing the performance of the methods. Four cases are studied, in two different categories: the first category; three cases of reduction of maximum standard workweek hours in Portugal (1996), France (2000) and South Korea (2004). the second category; the effect of the decrease of firing cost in Spain (2010). In the comparison the methodology is shown as reliable, especially LASSO; with respect to the empirical application, it is relatively simple to implement and is an improvement on the current techniques. For the empirical applications, Portugal and France depict a reduction in unemployment, whereas South Korea shows an increase in the employed population and a maintenance of a low unemployment. In the Spanish case, the reduction in the cost of firing employees has increased unemployment.
Deja un comentario