Abstract
In this paper we are interested in the identification and the dating of the turning points of the economic activity in Tunisia, by using the Markov switching approach innovated by Hamilton (1989) for the analysis of the business cycles. We identify the changes of regimes in the stochastic process of the economic growth, by using monthly data relating to the growth of the industrial production covering the period 1994-2004. We base our research on the smoothing probabilities determined by estimating the model of two states Markov switching regimes. The results obtained enable us to identify seven turning points during this period. And, using this approach of two states, we can detect the recession of september 2001.