Testing the adequacy of conventional asymptotics in GMM

摘要

This paper proposes a new test of the null hypothesis that a generalized method of moments model is sufficiently well identified for conventional asymptotics to be reliable. The idea of the test is to compare the volume of two confidence sets-one that is robust to lack of identification and one that is not. Under the null hypothesis, the relative volume of these two sets is O_p(1), but under the alternative, the robust confidence set has high probability of being unbounded.