Overview
Christopher L. Skeels is an econometrician at the Australian National University (Research School of Economics). His work focuses on finite-sample properties of econometric estimators, with particular attention to conditions under which standard procedures are valid.
Key Contributions / Features
- Sample size requirements for SUR (Griffiths-Skeels-Chotikapanich 2001): co-derivation of Theorem 1 (T ≥ M + ρ − η for two-stage FGLS) and Theorem 2 (T ≥ M + ρ for ML/Bayesian); showed the standard textbook requirement T ≥ max(M, k_max + 1) is both incomplete and misleading; derived the rank decomposition Ê = M_V Y + D that underpins the weaker two-stage condition; confirmed results numerically showing SHAZAM ML and Bayesian Gibbs sampler both fail silently for undersized samples
Related