采用修正的已实现门阀的多次幂变差和C_TZ统计量,利用上证综指2000年1月4日至2008年12月31日每5分钟的高频数据估计了中国股市波动率的跳跃成分,通过构建AHAR模型、AHAR—CJ模型、AHAR—C_TCJ模型及其各自的标准差形式和对数形式模型,实证研究了跳跃对中国股市波动率预测的影响。结果表明:跳跃对中国股市未来的日、周和月的波动率预测都存在显著的正向影响;加入修正的已实现门阀多次幂变差估计的跳跃成分后,标准差形式和对数形式的AHAR—C_TCJ模型能显著提高对日、周和月波动率的预测精度。
The authors utilize C_TMPV and C_TZ test statistics to estimate the jumps of volatility by employing the high-frequency data from SSEC, and AHAR model, HAR-CJ model and AHAR-C_TCJ model to analyze the impact of jumps on volatility forecast in Chinese stock markets. The authors find that the jumps have positive and significant impacts on daily, weekly and monthly volatility forecast in Chinese stock markets. In addition, the square-root and logarithmic AHAR-C_TCJ model based on C_TMPV, provides a significantly superior forecasting ability for daily, weekly and monthly volatility, especially on the day following the occurrence of a jump.