On Some Hypothesis Testing Problems in Econometrics

University Of Chicago, Chicago IL

Investigators

Abstract

This research project will focus on the development of new ways to test hypotheses in some important models that are workhorses for much empirical work not only in economics, but throughout the social sciences. Indeed, the models considered will include the well known linear regression model. The Principal Investigator and his collaborator, Joseph Romano (Stanford University), will consider the use of randomization methods to test certain hypotheses in these models. The researchers will provide conditions under which tests of hypotheses constructed in this way control the probability of a false rejection in finite samples. The researchers will also provide even weaker conditions under which tests of hypotheses constructed in this way control the probability of a false rejection approximately in large samples. Finally, the researchers will explore the extent to which these tests have desirable power properties, at least in large samples. Due to the widespread use of these models, it is expected that improvements in the ways to tests hypotheses in these models will have broad impact.