Asymptotic and finite sample distribution theory for IV estimators and tests in partially identified structural equations (Q1186050)

The authors consider the identification problem of a structural equation in a simultaneous system. Formulae for the finite sample and asymptotic densities of the instrumental variable estimators are given by the rotation of coordinates under the condition that the coefficient vectors of both endogeneous and exogeneous variables are only partially identified. Standard Wald statistics are formulated and are shown to converge in distribution to random variables which are not distributed as chi-square. Some numerical computations in partially identified models are given, and figures of the probability density functions of the estimators and Wald statistics based on simulations are also obtained.