基于Hawkes过程的国际原油市场与中国股票市场大幅波动联动性研究

考虑国际原油市场与中国股票市场之间的大幅波动存在联动性，本文采用二维标值Hawkes模型对2007年至2019年8月的布伦特原油期货和沪深300指数的日度数据中大幅波动的单市场延续和跨市场传染的传播特性进行建模。实证结果表明：(1)Hawkes过程可以较好地拟合国际原油和中国股市之间大幅波动的自激发和互激发效应，并捕捉资产收益率在时间和空间上的聚集性、持续性和溢出性；(2)原油市场和中国股市的大幅波动均存在较强的自激发效应；(3)原油市场和中国股市之间大幅波动的互激发效应具有统计意义上的显著性，但在实际影响方面相比于自激发效应更为微弱。本研究对股票市场建设、资产配置和风险防范均有一定意义。

Research on Linkage Effect of Large Fluctuations Between International Crude Oil Market and Chinese Stock Market Based on Hawkes Process

In this essay, large volatility mutually exciting effect between international crude oil market and Chinese stock market are discussed under the circumstances that the two markets have shown greater synchronicity in terms of volatile events over the past ten years. For example, on August 24, 2015, the global stock market crashed and fell into a precipitous spiral, with the CSI300 index plunging 8.75%. Crude oil prices were hit hard by market concerns that a slower recovery in crude oil demand would exacerbate the global supply glut. Brent crude oil futures plunged 6.56% from the previous trading day. It is of great significance to research the propagation characteristics of large fluctuations between international crude oil market and Chinese stock market from the perspectives of stock market construction, asset allocation and risk prevention. There are some defects in the three traditional research methods on the volatility spill-over effect: Multivariate GARCH Model, Extreme Value Theory and Copula. Hawkes process is a path-dependent random point process, whose core idea is that the occurrence of any event will increase the frequency of subsequent events, but this effect will decay in some form over time. So Hawkes process is a more compatible choice for modeling large volatility mutually exciting effect. The 2-dimensional marked Hawkes process is used to model the self- and mutual-excitation of large fluctuations of Brent crude oil futures and CSI300 index from 2007 to August 15, 2019, after the adjustment of statutory holidays and bilateral exchange rate. It is found that: (1) The Hawkes process can properly describe the single market continuation and cross-market contagion between international crude oil market and Chinese stock market, and capture the aggregation, persistence and spillover of asset returns in time and space. (2) There is a strong self-excitation in both crude oil market and Chinese stock market, respectively. (3) The mutual-exciting effect between the two markets is statistically significant, but weaker in actual impact than the self-excitation.