Off-campus UMass Amherst users: To download dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.
Non-UMass Amherst users, please click the view more button below to purchase a copy of this dissertation from Proquest.
(Some titles may also be available free of charge in our Open Access Dissertation Collection, so please check there first.)
Predictability and dynamic asset allocation strategies
Recent evidence of predictability in asset returns has led to an increased interest in dynamic asset allocation models and trading strategies. Earlier conclusions that technical trading rules are not useful (Fama and Blume, 1966) and gains from market timing are unlikely (Sharpe, 1975) are now being questioned. While today it is widely accepted that asset returns contain predictable components, the implications of predictability remain controversial. Many recent studies including Brock, Lakonishok, and LeBaron (1992), Bessimber and Chan (1995, 1998), Knez and Ready (1996), Kandel and Stambaugh (1994), Pesaran and Timmerman (1995) examine trading strategies and the economic significance of predictability in historical return data. In contrast, this dissertation employs a Monte Carlo simulation method to determine the level of predictability required for excess portfolio returns to a variety of strategies. Previous research (Samuelson 1991) shows that portfolio decisions depend on risk tolerance, the return process, and the investment horizon. The results in this dissertation show how several additional factors impact previous results. These factors include the frequency with which the portfolio is rebalanced, the predictability of the process, and trading costs. It is assumed that a utility maximizing strategy is followed by an informed investor who knows the parameters governing the process for risky asset returns. The performance of this informed strategy is therefore driven solely by the level of predictability exhibited by the return processes. Several naive strategies that follow simple linear investment rules are also examined. In the absence of trading costs, the returns the to particular naive strategy that is consistent with the return process are highest when the portfolios are rebalanced frequently. However, the naive strategy performance is adversely affected by trading costs. The informed strategy exhibits optimal performance when rebalanced frequently even when trading costs are included. The risk-adjusted return performance of the informed strategy is positive at very low levels of predictability and increases 0.63% with every 1.0% increase in predictability.
Wilkens, Kathryn Alison, "Predictability and dynamic asset allocation strategies" (1998). Doctoral Dissertations Available from Proquest. AAI9909232.