Mark J. Jensen

Overview

Mark J. Jensen is an economist at the Federal Reserve Bank of Atlanta. His research focuses on Bayesian nonparametric methods for financial time series, particularly semiparametric stochastic volatility models. He is co-developer (with John Maheu) of the SV-DPM model, which combines a parametric AR(1) log-volatility process with a Dirichlet Process Mixture for the return innovation distribution.

Key Contributions

• Jensen-Maheu (2008): Semiparametric Bayesian SV — AR(1) log-volatility + nonparametric DPM return distribution; showed that Gaussian SV overestimates $\sigma_v^2$ by 2.7× (CRSP) or ~20× (mixture DGP) due to distributional misspecification; 4-block MCMC with Fleming-Kirby block sampler and Chinese Restaurant Process updates; avg. $k=7.16$ clusters in CRSP daily returns 1980–2006. See Jensen-Maheu (2008) and Stochastic Volatility.

Related

• John M. Maheu
• Stochastic Volatility
• Dirichlet Process Mixture
• Mixture of Normals