In this paper, we consider the problem of testing for an autocorrelation change in discretely observed Ornstein-Uhlenbeck processes driven by Lévy processes. For a test, we propose a class of test statistics constructed by an iterated cumulative sums of squares of the difference between two adjacent observations. It is shown that each of the test statistics weakly converges to the supremum of the square of a Brownian bridge. The test statistics are evaluated by some empirical results.

http://link.springer.com/article/10.1007%2Fs11425-012-4511-y

http://link.springer.com/article/10.1007%2Fs11425-012-4511-y

http://link.springer.com/article/10.1007%2Fs11425-012-4511-y