中国融资融券业务处置效应的实证分析

处置效应阐述了投资者较早卖出盈利而长时间保留亏损的倾向。2010年我国融资融券业务的启动，大大提高了市场的流动性和活跃性，同时也带来了信用交易的投机特性，加强了市场阶段性波动并放大风险，这为我们研究处置效应提供了独特的样本。本文以某大型券商2014年11月至2016年10月30512个账户共3239305条交易记录为样本，采用衡量处置效应的经典指标PGR-PLR以及生存分析中的cox比例风险回归模型检验了两融市场上处置效应的存在性，并进一步利用生存分析方法对影响个体处置效应的相关因素进行了研究。实证结果表明，融资融券市场整体上存在显著的处置效应；就个体层面而言，性别、年龄、投资水平均能影响投资者处置效应的强度。本文发现，与男性投资者相比，女性投资者的处置效应更为强烈。并且，若依据年龄将样本分为青年、中年、老年投资者三组，青年投资者处置效应显著低于中年投资者，而老年投资者处置效应最低。此外，无论以观察期内购买的不同标的物的合约数目还是合约总成交量作为投资者投资水平的代理变量，结果均显示，投资水平越高的投资者处置效应更低。最后，应用weibull风险模型替代cox比例风险模型重新计算，结果稳健。

An Empirical Analysis on Disposition Effect of Chinese Margin Trading

The disposition effect is the tendency to sell winning assets too soon, while holding onto losers too long. Since the start of securities margin business in 2010, the liquidity and activeness of the market has been greatly enhanced, along with the increase in the speculation brought by credit trading and the magnification of periodic volatility, which provides us with a unique sample for studying the disposition effect. In this paper, data from a famous national brokerage firm, which is comprised of 3239305 trading records of 30512 accounts from November 2014 to October 2016, is used. The classic PGR-PLR measure and the Cox proportional hazard model have been utilized to test the existence of disposition effect in the securities margin market, and the survival analysis is used to study the demographical factors influencing disposition effect. Empirical evidence shows that, overall, disposition effect exists in China's margin trading market. The female investor exhibits a stronger disposition effect than the male investor. People of middle age are more disposition-prone than people under the age of 35 while the magnitude of disposition effect of people of age over 50 is the largest. Using the number of different types of contracts that have been bought during the observation period and the overall volumes of contracts respectively as proxies for an investor's investment level, both results demonstrate that the better an investor is at investment, the lower his disposition effect is. Our results are robust when we rerun the model using Weibull hazard function. Our paper helps investors gain insight into their inherent behavioral bias and thus guide them to make rational investment decisions.