Overview
William J. McCausland is an econometrician in the Department of Economics at the Université de Montréal, and a member of CIREQ. His research applies Bayesian methods to time series in economics and finance, consumer and firm demand, discrete choice, and behaviour in repeated-game experiments, with a strong emphasis on efficient statistical computing.
Key Contributions / Features
- Geweke-McCausland (2001): Bayesian predictive specification analysis in econometrics — selecting a low-dimensional vector of interest ψ(y) and checking the observed value against its (post)predictive distribution; stock-return application showing a Gaussian i.i.d. model fails completely and a t-GARCH model over-generates kurtosis once the degrees-of-freedom parameter is integrated out. See Geweke-McCausland (2001).
- Efficient state-space computation: developer of precision-based simulation-smoothing methods for state-space models, including the HESSIAN method (Highly Efficient State Smoothing, In A Nutshell) and comparative analysis of simulation smoothers (with Miller and Pelletier).
Related