Testing underidentification in linear models, with applications to dynamic panel and asset pricing models

This paper develops the links between overidentification tests, underidentification score tests and Cragg Donald (1993, 1997) Kleibergen Paap (2006) rank in linear instrumental variable (IV) models. For structural model y=Xβ+u, with endogenous explanatory variables partitioned as X=x1X2, this general framework shows that standard are for an auxiliary model, x1=X2δ+ε, estimated by IV estimation methods using same instruments original model. simple structure makes it possible to establish valid robust models where these have not been proposed or used before, like clustered dynamic panel data GMM. The also applies of parameter matrices. Invariant based on LIML continuously updated GMM estimators both first-stage parameters. insight leads proposal new two-step invariant asymptotically efficient estimators, a iterated estimator that, if converges, converges estimator.