This paper presents the application of model error correction (VEC), aimed at measuring the cointegration of two or more time series, or prove that the characteristics of the series are significantly similar and that shocks in one of them alters to a similar extent the behavior of the other. For the analysis is taken as an example trajectories cointegration for gross domestic product (GDP) and the level of traditional exports in Colombia, including for quarterly periods between January 2000 and the third quarter of 2013. It finds a vector cointegration explaining a common evolution for these two variables in the long run, which concluded that an increase of 1 % in exports generates domestic product growth of 0.22 %.