PREDICITING BULL AND BEAR IN THE SWEDISH STOCK MARKET

Abstract: Very little if any previous research has been done on the potential predictability of bear and bull regimes in the Swedish stock market. In this study my aim is to predict OMXS30 bull and bear regimes with dynamic binary time series models. After using a nonparametric approach to identify the regimes of bull and bear periods in the market I looked at both an in sample and out of sample test. Based on monthly data I found different predictive variables, with the variables with highest predictive power being, the term spread and market liquidity deviation. Further variables with statistically significant results and predictive power are, the federal fund rate, purchasing manager index and the nominal return. The result can be improved using the dynamic structure in the binary response model. Using multivariate dynamic binary time series I found that the model yield higher returns than the buy and hold strategy.