Victor Chernozhukov

Chernozhukov's recent work solved an important, long-standing problem in regression, namely inference on coefficients of interest after selection of covariates. He developed machine learning methods for causal inference and treatment effect evaluation with high-dimensional data. He researched inference for identified sets, which has become an important subject in econometrics and applied economics. His work on quasi-Bayesian methods extends the reach of powerful Bayesian simulation methods far beyond the likelihood setting. He has contributed to regression quantile methods in correcting for endogeneity and accounting for misspecification. He has developed asymptotic theory using sequential approximation instead of standard convergence ideas, including multivariate central limit theorems for very many sample averages.