Abstract: Five pieces of theoretical prediction on the generation of higher order moment risk are refined and thus the corresponding hypotheses are formed after reviewing the existed literature. Employing the idea of modeling time variant higher-order-moment, these five pieces of hypothesis are involved in a unified econometric framework. An empirical analysis is conducted based on such model to find some dominated theoretical explanations. Sampling daily returns from Shanghai and Shenzhen stock market composite index, the results show that there are two hypotheses relevant to volatility feedback effect and bad news revelation effect are significantly supported but the other three are not. Further analysis indicates that these two effects can generate both skewness and kurtosis risk. Therefore, they are implied as the main generation mechanism of higher-order-moment risk. These results get out of the mess of opinions on the generation mechanism of higher-order-moment risk and thus can benefit further exploration of such topic under a unified theoretical framework.

Key words: higher-moment risk, volatility feedback, short-sale constraint, conditional density