The main purpose of this document is to explain the business cycles in Ecuador by comparing the classical model of regression used in econometrics to Markov Switching Models which include a non ob- servable Markovian process that governs changes in regime or states. This project runs this model using fixed transition probabilities with two states, recession and recovery/growth period. Then, as an extension of this methodology, it runs a Markov Switching Model only with the constant terms and afterwards with the estimated smoothed probabilities, it runs a linear regression transforming the scale of the predictor variable, in order to see how the explanatory variables affect the probability of remaining at the same state or moving to the other.