BOOTSTRAPPING QUANTILE REGRESSION-ESTIMATORS

The asymptotic variance matrix of the quantile regression estimator de pends on the density of the error. For both deterministic and random r egressors, the bootstrap distribution is shown to converge weakly to t he limit distribution of the quantile regression estimator in probabil ity. Thus, the confidence intervals constructed by the bootstrap perce ntile method have asymptotically correct coverage probabilities.