|
Abstract:
|
This paper proposes a Bayesian modeling framework to estimate the time-varying market factor risks of hedge fund investment portfolios-by studying hedge fund strategy indices. The explanatory variables are major market index and factor returns. Several dynamic models are proposed and compared. Time-varying model parameters, namely factor sensitivities and excess returns, are modeled as a "random walk plus noise" process. Investment industry experience provided guidance for selecting relevant factors in regressions models and setting prior (ex-ante) parameter values; Markov Chain Monte Carlo simulations generated posterior (ex-post) probability distributions to facilitate parameter (point and interval) estimates. Results from quantile regressions are also briefly discussed and compared with the results by dynamic regression models. The models presented here can be easily applied to individual hedge fund analysis and a diversified portfolio that incorporates hedge fund investments.
|
