Abstract The objective of this paper is to assess the effect of long-term participation in an extension program on peri-urban family farming earnings. For this purpose, a survey of a sample of 156 participants in the Local Development Program (PRODESAL) of the Chilean National Institute for Agricultural Development (INDAP) was conducted in the Metropolitan Region of Santiago. Half of the respondents were long-term participants with at least four yrs in the program, and the rest of the farmers were one-yr beneficiaries. A treatment regression model (TRM) with a binary endogenous variable was used to identify differences among long- and short-term farmer participation in PRODESAL. The first stage of the model is a probit regression that identifies factors influencing long-term participation in PRODESAL. The second stage has the total value of production (TVP) as a dependent variable, which captures agricultural earnings, and one of the explanatory variables is the binary endogenous regressor of the first stage of the model. The results show a positive impact of long-term participation in PRODESAL on TVP. Other variables, such as farm size, fertilization and access to credit, were significant and positively related to TVP as well. Education and social capital are shown to be variables that increase the probability of long-term participation of family farmers in the program.