Overview
Jangryoul Kim is an economist associated with the Korea Development Institute (formerly affiliated with the Federal Reserve Bank of Minneapolis). His research focuses on Bayesian econometric methods for macroeconomic forecasting, with particular emphasis on mixed-frequency BVAR models.
Key Contributions / Features
- Kim-Miller-Ozanne (2004): Co-developed two BVAR-based approaches for estimating and forecasting U.S. capital gains realizations for CBO revenue projections. The integrated quarterly method — interpolating annual gains to quarterly frequency and incorporating them in a unified BVAR with Kalman filtering — achieved a 36% reduction in 1-year RMSE relative to the CBO historical mean-reversion benchmark (11.92 pp vs. 18.57 pp). Key methodological insight: linear interpolation outperforms more elaborate schemes because annual gains residuals are positively serially correlated (DW = 0.78).
Related