This group conducts research into theoretical econometrics. Its members are particularly interested in the development of statistical tests, questions of inference and the decomposition of distribution differences. One specific project is presented below.
Uniformly valid inference
This project aims to develop tests that are uniformly valid in situations where “standard” tests, like the Student test, do not work. These situations are characterised by a discontinuity in the asymptotic distribution of standard tests in a nuisance parameter, that is, a non-specified parameter under the null hypothesis. One comes across this problem when, for example, one hopes to make inference for a sub-vector and a nuisance parameter might lie at the boundary, or when a parameter is potentially weakly identified.