A consistent test for the parametric distribution of regression disturbances

This paper proposes a consistent test for the parametric specification of t he distribution of regression disturbances. The test statistic is a goodnes s of fit test which is shown to be asymptotically normally distributed unde r the null hypothesis. The local power property of the proposed test is inv estigated. A parametric bootstrap procedure is proposed to approximate the finite sample null distribution of the test statistic. Limited Monte Carlo experiments are carried out to examine the finite sample performance of the proposed test.