Testing for jumps based on high-frequency data: a method exploiting microstructure noise

Abstract

This paper tests for jumps of the price process based on noisy high-frequency data. Under the null hypothesis that the price process is continuous, the test statistic converges to a normal distribution, and under the alternative hypothesis that the price has jumps, the statistic converges to infinity. Compared with the test of Aït-Sahalia et al. [Testing for jumps in noisy high frequency data. J. Econom., 2012, 168(2), 207–222], our proposed statistic uses information on the microstructure noise, tends to infinity more rapidly under the alternative hypothesis and has a better power. A simulation confirms the theoretical results and an empirical study illustrates the practical application of the method.