This paper analyses the main implications from an exercise of growth accounting applied to the agricultural sector in Colombia, using a theoretical model that includes subsidies and certain degree of non-perfect market competition. The results of estimation from the dual and primal approaches show significant differences between TFP measured by these methods, which could demonstrate distorting effects that generates a “dynamic reallocation effect” phenomenon in agricultural activity. Additionally, the single-sample t-test and the model of differences with additional regressors show that TFP for agricultural products that received subsidies and preferential tariffs by the government compared to non-favored is statistically significant and different from cero. The difference between TFP favored and non-favored groups is 3,2 points and is said to be statistically different than the hypothesized value at the 1% significance level.