Title: Testing For Exogeneity in Instrumental Variable Regressions: A Bootstrap Approach
Authors: Charisios Grivas - Birkbeck University (United Kingdom) [presenting]
Abstract: This paper considers bootstrap versions of a Hausman-type test statistic for econometric models with endogenous regressors whose coefficients are allowed to vary over time both deterministically or stochastically. I compare the finite sample performance of the asymptotic and the bootstrap version of the test by means of Monte Carlo simulations. The bootstrap test statistic appears to have proper size and higher power. More importantly, it is shown that the size and the power of the bootstrap test, are invariant to the choice of the bandwidth parameters and the number of instruments.