TESTING FOR A CHANGE OF THE INNOVATION DISTRIBUTION IN AN ARCH MODEL
In this paper, we develop a Cramér-von Mises (C.V.M.)-type test for a change-point detection in residuals of an ARCH model. Under the null hypothesis, we prove that our test statistic is asymptotically a random variable whose distribution is constructed using a Kiefer process and is tabulated in Blum et al. [2]. In a small simulation study, we show its good performance for finite samples.
ARCH model, change-point, Cramér-von Mises-type test, Kiefer process, weak convergence.
