Bootstrap inference for fixed-effect models

The maximum likelihood estimator of nonlinear panel-data models with fixed effects is asymptotically biased under rectangular-array asymptotics. The literature has devoted substantial effort to devising methods that correct for this bias as a means to salvage standard inferential procedures. The chief purpose of this presentation is to show that the (recursive, parametric) bootstrap replicates the asymptotic distribution of the (uncorrected) maximum-likelihood estimator and of the likelihood-ratio statistic. This justifies the use of confidence sets and decision rules for hypothesis testing constructed via conventional bootstrap methods. No modification for the presence of bias needs to be made.