Tactical Asset Allocation and Model Uncertainty : An Exploratory Study for the Swiss Stock Market

This paper uses statistical model selection criteria and Avramov’s (2002)Bayesian model averaging approach to analyze the sample evidence onstock market predictability in the presence of model uncertainty. Basedon Swiss stock market data, our posterior analysis finds that neither thecumulative posterior probabilities nor the posterior probabilities of theindividual forecasting models are constant through time. We also decomposethe variance of predicted future excess returns into three components:uncertainty attributed to forecast errors, parameter uncertainty,and model uncertainty. The empirical results indicate that the respectivecontributions are highly dependent on the time period under considerationand the initial values of the predictive variables. In contrastto Avramov (2002), model uncertainty is generally not more importantthan parameter uncertainty. Finally, we also demonstrate the implicationsof model uncertainty for tactical asset allocation strategies. Ourresults do not indicate any reliable out-of-sample return predictability.Moreover, in contrast to Avramov (2002), the out-of-sample performanceof the Bayesian model averaging approach is not generally superior tothe statistical model selection criteria. Consequently, even properly incorporatingmodel uncertainty into a tactical asset allocation model doesnot seem to help improving the performance of the resulting short-termtactical asset allocation strategies.