An Optimal Investment Model with Regime Switching Volatility

Shangzhen Luo,Xudong Zeng
2009-01-01
Abstract:We consider a multi-stock market model. The processes of stock prices are governed by stochastic dierential equations where volatilities are stochastic and regime-switching, i.e., driven by a Brownian motion and a finite-state Markov chain. Both the underlying Brownian motion and the Markov chain are unobservable. Investors can observe the stock prices only. We study its discretized model which is a discrete time hidden Markov process. The objective is to control trading at each time step to maximize an expected utility function of terminal wealth. Exploiting dynamic programming tech- niques, we derive an approximate optimal trading strategy that results in an expected utility function close to the optimal value function. Necessary fil- tering and forecasting techniques are developed to compute the near optimal trading strategy.
What problem does this paper attempt to address?