Master's in Statistics Oral Thesis Defense: "Dynamic Portfolio Optimization with a Noisy Observation of the Hidden Economic Regime"

Abstract: The project aims to maximize the expected utility of a risk-averse investor by allocating her wealth in a risk-free bond, a stock, and a defaultable security, of which the price processes depend on a hidden continuous-time finite-state Markov chain representing the economic regime. In previous works dealing with hidden Markov chain, the volatility component of the stock is often assumed to be constant, or at least not dependent on the hidden states. Otherwise, the unobservability assumption will be violated. However, this assumption is sometimes restrictive. To make the risky asset have time-varying volatility related to the hidden Markov chain in some noisy sense, I proposed a new model using a 2-D Markov chain with finite states, where one dimension is the hidden process, and the other one is observable. By some special constructions in its generator matrix, the observable process can serve as a noisy observation of the hidden state, and the volatility of the stock process can be linked to the observable dimension.

Host: Jose Figueroa-Lopez

(Access Zoom Presentation HERE)