this research presents empirical evidence of the mexican labor market earnings for males and females, pooling together a set of occupational categories into formal and informal workers using the counterfactual technique of dinardo, fortin and lemiux (1996). the semi-parametric specification allows visualizing the earnings distribution according to the decomposition of subgroups and occupational process using a logistic model. we find that informal self-employment is better paid for men than for their counterpart between 1992 and 2002, meanwhile women seem to be better positioned as formal salaried than men. counterfactuals predict an improvement for both informal women and males, had they decided to become formal workers and with the same attributes in 2002; an opposite trend was found for the initial year. this situation remains at odds when depicting predicted multinomial probabilities, as informal workers tend to develop entrepreneurial activities as long as they acquire more experience.