Abstract
This contribution studies the out-of-sample performance of trading strategies applying 2-State-Markov-Switching models. Thereby, different probability thresholds are considered where the investor decides when to go in, respectively, out of the stock market. Furthermore, the investor may decide to invest in a risk free asset when a bear-market is expected to occur in the forecast period. In this study, the US-stock index S&P 500 is employed where the challenging period from January 2008-December 2010 is used for the out-of-sample experiment. The optimal trading strategies, given that the investor does not decide to invest in the risk free asset, suggest low probability thresholds of 0.075 and 0.050 concerning bull-market probabilities. Therefore, the investor is invested 69% respectively 88% of the investment horizon in the stock market. The optimal trading strategies exhibit total gross gains of 34.35% respectively 6.01% during the out-of-sample period, whereas the S&P 500 stock market return was -8.07% during the period under consideration.