当前位置: 首页 / 学术科研 / 学术活动 / 正文

【经济学院名家讲座·行为经济学系列】2018年秋季学期​第一讲:金星(XingJin)

[发表时间]:2018-09-03 [浏览次数]:

讲座题目:UnderstandingCommon Risk Factors in Variance Swap Rates, Market Return Predictability andVariance Swap Investments When Volatility can Jump

主讲嘉宾:金星(XingJin),英国华威大学商学院金融副教授

讲座时间:2018年9月6日(周四)下午2:00

讲座地点:沙河主教207

嘉宾简介:

金星(XingJin)。1983年在安庆师范学院获得数学学士学位;1986年在西安交大获得应用概率与运筹硕士学位;1991年在中国科学院应用数学研究所获得应用概率与运筹博士学位;2002年在美国马里兰大学Smith商学院获管理科学与统计博士学位。工作经历:从1993到1998在中国科学院系统科学研究所任副研究员和研究员;2002到2006在新加坡国立大学数学系任助理教授和副教授;2006至今在英国华威大学商学院任金融副教授。在主流金融,经济学和运筹学杂志上发表论文十多篇,如,ReviewofFinancial Studies, Management Science, Mathematical Finance, FinanceandStochastics, Journal of Banking and Finance, Journal ofMathematicalEconomics, Journal of Economic Dynamics and Control , Mathematicsof Operations Research, European Journal Operational Research等。2016年获中国金融年会最佳论文二等奖;亚洲金融学会会刋International Review of Finance编委.

内容摘要:

This paper proposes a tractableself-exciting double-jump model for stock return and its variance processes,extending existing two-factor term structure models of variance swap rates inthe literature to a new three-factor model. Various goodness-of-fit tests showthat our three-factor model outperforms the two-factor model in fitting the S&P500 return and its variance swap rates. More importantly, we show that theexpected log market excess return is linearly related to variance rates,suggesting a novel predictive regression model for the market returns. In starkcontrast to the existing literature, our empirical results demonstrate thatvariance swap rates have superior predictive ability compared with variancerisk premiums (VRPs) for predicting market returns both in-sample andout-of-sample for horizons up to two years. Unlike the popular double-jumpmodel in the literature, our new model allows us to solve the optimal varianceswap investment in a semi-closed form which greatly facilitates newunderstanding of volatility trading. Specifically, we find that the investoralways takes a long-short-long strategy in investing variance swaps, and incurssizeable economic loss caused by both model and parameter misspecifications.