Overview
Andrew J. Patton is an econometrician at Duke University, known for work on volatility forecast evaluation, copula-based models of financial dependence, and the econometrics of asset returns. His most-cited methodological contribution in this wiki is a set of robustness results for comparing volatility forecasts when the conditional variance is a latent variable.
Key Contributions / Features
- Robust loss functions for volatility forecast comparison (Patton 2011) — Established theoretical conditions under which forecast-comparison rankings are preserved when an imperfect (noisy) proxy replaces the unobservable true conditional variance. The MSE and QLIKE loss functions are robust — they give the same ranking of forecasters whether evaluated against the true variance or any conditionally unbiased proxy — whereas MAE is not. This justifies evaluating volatility forecasts against imperfect but observable proxies (squared returns, realized variance, VIX²). See Volatility Forecast Evaluation.
- Copula methods — Time-varying copula models for multivariate financial dependence and asymmetric (tail) correlations.
Related