In this dissertation, we study three important issues in the investment science which are portfolio selection, asset allocation, and performance evaluation by incorporating the regime switching framework. We focus on the violated assumptions, especially the normality and stationarity of asset price processes in theories related to three issues. To handle these problems, we employ the regime-switching concept.
First, we apply the regime-switching model to solve robust portfolio selection problem. While the robust portfolio optimization takes uncertainties of parameters into account when forming an optimal portfolio, the method still assumes that the true parameters of the model are not time-dependent. Hence, we incorporate the regime switching model to construct the portfolio that is capturing the broad market condition as well as robust against estimation errors.
Second, we construct the regime-switching model to identify regimes in the stock, bond, and commodity markets to improve the solution of the dynamic asset allocation problem in these markets. Thus, we develop a stochastic program to optimize portfolios under the regime switching framework by combining the regime information and scenario generation technique.
Third, we investigate equity hedge fund performances by focusing on the time-varying regimes of the equity market. Because the target rate of return of hedge funds is determined in an absolute sense (i.e., no benchmark), hedge funds manage their investment portfolio dynamically depending on the market condition. In this regard, we separate hedge fund data into several sub-periods with distinct market conditions to analyze and evaluate hedge funds.
The findings in this study will provide a better understanding of three financial issues in general, and the newly suggested models incorporating the ideas related to regime-switching will be valuable for investors concerned about varied market conditions.