Donald Andrews, Distinguished Fellow 2026
Donald W. K. Andrews, Tjalling C. Koopmans Professor of Economics at Yale University, has made fundamental contributions to econometric theory. His work has had a lasting influence on a wide range of topics in modern econometrics, including structural change, weak identification, unit roots, instrumental variables, generalized method of moments, subsampling and bootstrap methods, semiparametric inference, and the use of empirical process methods in econometrics.
Andrews’s scholarship has repeatedly changed how econometric problems are understood, and he has pointed out where conventional methods relying on large-sample normal approximations are not appropriate. His work on structural change and parameter instability is a leading example. In settings where the timing of a break is unknown, standard methods often fail to provide a satisfactory basis for inference. Andrews developed a general framework for testing for structural change with unknown breakpoints, and that work became central to the subsequent literature. His work helped establish how to conduct inference when parameters may shift over time and when certain parameters, such as the unknown date of a structural break, are not identified under the null hypothesis of no change, so that standard asymptotic approximations for test statistics do not apply.
A second major area of contribution concerns inference under weak identification. Andrews was among the economists who did the most to clarify the consequences of weak instruments and weakly identified models for econometric practice. His work showed why conventional asymptotic approximations can be misleading in such settings and developed methods that remain valid when standard procedures do not. These contributions thus give empirical researchers stronger tools for conducting inference in models where identification is tenuous.
His work on generalized method of moments, bootstrap and subsampling procedures, and semiparametric econometrics was influential. Andrews has repeatedly helped determine when conventional, normal-distribution-based asymptotic approximations are reliable, when resampling methods can improve inference, and how robust procedures can be constructed in settings where parametric assumptions are too strong. His research has often provided the theoretical foundations on which later methodological work built, and upon which empirical applications have depended.
Andrews has also played an important role in bringing sophisticated probabilistic and mathematical statistics tools into econometrics. His work on empirical process theory provided general conditions under which estimators and test statistics in semiparametric and nonparametric models have well-behaved limiting distributions, extending the reach of formal inference to model classes where earlier results had been ad hoc or unavailable.
The range and quality of Andrews’s work have long been recognized by the profession. He is a Fellow of the Econometric Society, a Fellow of the American Academy of Arts and Sciences, and a Fellow of the Journal of Econometrics. He has received major recognition for his research contributions, including the Plurima Scripsit award in Econometric Theory. At Yale, his teaching has been recognized with multiple awards, and he has advised leading econometricians over decades.
In addition to his research, Andrews has made substantial contributions to the profession as Director of the Cowles Foundation, where he helped sustain one of the principal institutional centers of econometric research.