不完全市场中的期权定价：一种基于动态经验投影定价核的期权定价方法

经典的资产定价理论认为，不完全市场中存在多个 定价核，导致期权存在多个可能的理论价格，与现实中期权只有一个价格不符. 考虑到期权定价时投影定价核具有与原始定价核 完全相同的意义，并且投影定价核的信息隐含于相关历史信息中，动态的经验定价核隐含了投资者风险偏好的时变性，本文建立了一种基于动态经验投影定价核的期权定价方法. 通过对Rosenberg-Engle的经验定价核的动态性和标的资产未来收益率估计方法进行扩展，本文还给出了动态投影经验定价核的估计方法. 这种方法能适用于单只期权单只标的资产、多只期权单只标的资产和若干只期权单只标的资产等情形. 本文基于此方法为沪深交易所的部分具有欧式性质的权证进行了实证研究，结果表明：考虑了投资者时变偏好的动态经验投影定价核期权定价方法的定价效果远好于 Black-Scholes-Merton 和常弹性方差（CEV）期权定价模型.

Option pricing based on dynamic empirical projected pricing kernels in incomplete asset markets

Classic asset pricing theory tells us that there are more than one theoretic price since there exit more than one equivalent martingale measure, which conflicts the phenomenon that there is only one price in real financial market. Since the projected pricing kernels are the same to original pricing kernels in options pricing, the real projected pricing kernels are indicated in history data, and the dynamics of pricing kernels imply the time-varying risk preference of investors, an option pricing method based on dynamic empirical projected pricing kernels is suggested in this paper. With the modification of dynamics estimation and the future return of underlying asset in Rosenberg and Engle, the estimation of dynamic empirical projected pricing kernels is presented, which is appropriate in the circumstance of one, few or many options listed on the same underlying asset. The empirical studies show that the dynamic empirical pricing kernels option pricing model which can catch the time-varying risk preference of investors performs quite better than Black-Scholes-Merton and constant elasticity of variance option pricing model.