Bootstrapping Subset Test Statistics in IV Regression

The finite-sample performance of various bootstrap procedures is studied by simulation in a linear regression model containing 2 endogenous regressors. Besides several residual-based bootstrap procedures, we also consider the GMM bootstrap. The test statistics include t-statistics based on k-class estimators and the robust subset quasi-LR (MQLR) statistic. In the simulations, the restricted fully efficient (RFE) bootstrap DGP based on Fuller estimates and the LIML t-statistic performs best of the Wald-type statistics. Unfortunately, the bootstrap only marginally reduces the conservativeness of the subset MQLR statistic. Finally, the GMM bootstrap does not seem to improve upon the asymptotic approximation. An empirical example illustrates the use of these procedures.