the bootstrap percentile intervals are generally not symmetric around the underlying parameter estimate. Further, the bootstrap procedures can also deal with the joint distribution of various QR estimators,which allows the equality of slope parameters to be tested across various quantiles.