Sangjoon Kim

Overview

Sangjoon Kim is an economist who, at the time of Kim-Shephard-Chib (1998), worked at Salomon Brothers Asia after completing his PhD at Princeton University under John Y. Campbell. His primary contribution is as a co-developer of the mixture-of-normals multi-move MCMC sampler for Bayesian stochastic volatility inference.

Key Contributions

• Kim-Shephard-Chib (1998): Co-developed the MIX sampler — log-squaring the SV observation equation, approximating the $\log\chi^2(1)$ noise by a fixed 7-component Gaussian mixture (Table 4), drawing mixture indicators and the full latent log-variance path via the Carter-Kohn (1994) simulation smoother in one Gibbs block; importance reweighting corrects the approximation error to exact Bayesian inference; inefficiency factors 1–3 vs. 20–200 for single-move Gibbs.

Related

• Stochastic Volatility
• Neil Shephard
• Siddhartha Chib
• Kim-Shephard-Chib (1998)
• Gibbs Sampler
• Kalman Filter