Abstract
Empirical investigations have indicated that there is presence of non-linearity in Botswana stock market. This paper therefore, examines whether the Botswana stock market displays a regime switching behavior using both the Logistic smooth transition autoregressive models (LSTAR) and Exponential smooth transition autoregressive models (ESTAR). By analyzing the residual properties, the results actually show that Logistic smooth transition autoregressive models (LSTAR) is more appropriate to model the series against linear model.