Overview
Herman K. van Dijk is an econometrician at the Econometric Institute and Tinbergen Institute at Erasmus University Rotterdam. A pioneer of Bayesian numerical methods in econometrics, he developed importance sampling algorithms for posterior moments (Kloek-Van Dijk 1978; Van Dijk-Kloek 1980; SISAM, Hop-Van Dijk 1992) and contributed foundational work on Bayesian unit root testing, GARCH stationarity, and cointegration.
Key Contributions / Features
- Kloek-Van Dijk (1978): Monte Carlo integration for posterior densities — one of the first practical Bayesian numerical methods in econometrics, using importance sampling via a Student-t candidate.
- Van Dijk-Kloek (1980): Further development of Bayesian Monte Carlo integration.
- Schotman-Van Dijk (1991): Bayesian analysis of the random walk model and unit root testing.
- Kleibergen-Van Dijk (1993): Bayesian stationarity analysis of GARCH(1,1); quasi-strict vs. weak distinction; fat-tail/unit-root posterior entanglement.
- Kleibergen-Van Dijk (1994): Irregular likelihood shapes in cointegrated systems.
- SISAM algorithm (Hop-Van Dijk 1992): Importance sampling for posterior moment computation via multivariate-t approximation to the joint posterior.
- Kleijn-Van Dijk (2002): Bayesian UC model for real exchange rates (I(2) smooth trend + two stochastic cycles); Bayes factors overwhelmingly reject constant PPP equilibrium; half-life of PPP deviations drops from 3–5 years to ~1 year once time-varying mean is accounted for.
Related