股市投资收益与风险直接关系的定量研究

EMPIRICAL STUDY ON THE RELATION OF STOCK RETURN VS.RISK IN SHANGHAI SECURITIES EXCHANGE THROUGH ARCH M GA

科学地定量探测投资收益与投资风险是否具有正比关系，理论和实践上都具有重大意义，而ＡＲＣＨ－Ｍ族正是解决该问题的最佳工具．针对ＡＲＣＨ－Ｍ传统估计方法的不足提出了遗传算法（ＧＡ）的改进，并系统实证估计了上海股市的各种ＡＲＣＨ－Ｍ模型，科学验证了在以周为时间刻度下上海股市中投资收益与投资风险之间确实存在正比关系，同时证明了在此范畴内投资风险的最佳测度是投资收益的标准差而非方差。

The ARCH M model class is the best tool for detecting the proportional relation between return and risk,which is the important basis for modern economics and finance.Due to some serious disadvantages of traditional estimating methods for the ARCH M model,GA is developed and various ARCH M models in SSE are established empirically in this paper.Finally,we conclude that the proportional relation between stock return and risk in S.S.E.does exist and that the best measure for risk is returns standard deviation but not its variance.