|Institution:||Università degli studi di Bergamo|
|Keywords:||Markov Processes; Applied Probability; Portfolio Selection; Stopping times.; SECS-S/06 - Metodi mat. dell'economia e Scienze Attuariali e Finanziarie|
|Full text PDF:||http://hdl.handle.net/10446/31899|
This thesis analyses the impact of parametric timing portfolio strategies on the U.S. stock market. In particular, we assume that the log-returns follow a given parametric Lévy process and we describe a methodology to approximate the distributions of stopping times using the underlying Markov transition matrix. We extend the analysis to non-Lévy processes, considering Markov Regime switching model and the log-Student-t model. Therefore, we propose the use of portfolio strategies based on the maximization of the ratio between the expected first passage time to reach a low level of wealth and the expected first passage time to reach a high level of wealth. Finally, we compare the ex-post wealth obtained maximizing the ratio of proper expected stopping times under different distributional assumptions.