We derive adjusted signed likelihood ratio statistics for a general class ofextreme value regression models. The adjustments reduce the error in thestandard normal approximation to the distribution of the signed likelihoodratio statistic. We use Monte Carlo simulations to compare the finite-sampleperformance of the different tests. Our simulations suggest that the signedlikelihood ratio test tends to be liberal when the sample size is not large,and that the adjustments are effective in shrinking the size distortion. Tworeal data applications are presented and discussed.
展开▼
