The main objective of the work is to test the existence of differences in earnings between salaried jobs and self-generated jobs in the labor markets of Peru. To this purpose it uses data from household surveys from 2007 to 2011. Also it uses non-parametric and semi-parametric methods for premium estimates (propensity score matching, double difference and double difference with matching). The use of 3 different methods and the use of several datasets that combines cross-section and panel samples, provide robustness to the results. The findings suggest the existence of segmentation between these two types of jobs. The calculations were made for different population samples: national, by sex and by area of residence. At national level premiums are in the range of 35-53% depending on the method used. Among women and residents of rural areas the premiums are larger.