RANDOMIZED TESTING FOR JUMP DETECTION
提出一种序贯随机化检验方法，通过向局部平均收益的稳健近似中人为添加随机性来检测资产价格路径中的跳跃，并推导了有限活动跳跃下全局统计量的渐近分布，模拟和实证表现良好。
This article proposes sequential randomized tests to locate the presence of jumps on the paths of efficient asset prices in a continuous-time model. The randomized statistics are generated by artificially adding randomness to the robust approximations of the locally averaged returns of the efficient price. In the case of finite activity jumps, we derive the asymptotic distribution of the maximum of all the local statistics unaffected by jumps, which makes it feasible to control the limiting probability of the global type I error and demonstrate the power of the test. We also present the theoretical results to illustrate the behaviors of the test statistics in the presence of infinite activity jumps. Simulation studies indicate the favorable performance of the proposed test in finite samples, and we also apply the test to the stock price data of Apple and Microsoft.