NON-LINEAR MODELING OF THE MOROCCAN STOCK MARKET

Abstract

Since the sixties, debates have been born on the models which determine the evolution of the stock prices. The purpose of this article is to determine how the successive variations of the MASI index are independent of each other or, in other words, whether the latter follow or not a random walk. As a result, test the hypothesis of the efficiency of the Moroccan stock market. In order to be more precise, the choice was made for the smooth transition autoregressive (STAR) models applied to the MASI index for the period between 01/01/2004 and 31/12/2017, with data daily. It follows that they allow us to better model the nonlinear evolution of series of courses with changing regimes. This entails, therefore, the consideration of the transition, which can be slow compared to the series' adjustment dynamics, which characterizes the majority of stock market trends. According to the results, we note that the Moroccan market does not effectively retain the information issued, which implies its inefficiency.