运用Monte Carlo模拟分析的方法,通过对现存几种日内跳跃检验方法进行比较后发现.就检测能力、错判率和跳跃方差偏差而言,ABD跳跃检验方法表现最优。利用ABD跳跃检验方法,通过对上证综指高频数据进行分析后发现,上证综指平均大约每四天发生一次跳跃。跳跃方差贡献约为19%,并且跳跃行为存在非对称性;同时发现跳跃行为是造成上证综指收益率尖峰厚尾特征的重要因素之一。
This paper presents a comparative analysis of the existing detections for intraday jumps by using a Monte Carlo simulation. We find that the procedure of ABD test has the best performance in testing capability, rate of wrong judgment and jump variance deviation. Then we use the ABD test procedure to analyze the jumps of the Shanghai Composite Index in high frequency data. The results reveal that the jumps occur in the Shanghai Composite Index about every four days in average, the contribution of jump variance to volatility is about 19%, and the jump behavior is asymmetric. Meanwhile, we find that the jump behavior is one of the important factors causing the Shanghai Composite Index to yield leptokurtic and heavy-tailed features.