Bootstrapping the GMM overidentification test under first-order underidentification

We first show that the standard GMM bootstrap fails to consistently estimate the distribution of the overidentification restrictions test under lack of first-order identification. We then propose a new bootstrap method that is asymptotically valid in this context. The modification consists of adding an additional term that recenters the bootstrap moment conditions in a way as to ensure that the bootstrap Jacobian matrix is zero when evaluated at the GMM estimate.