Analysis of Skewness Preference for Horse Track Gamblers in Hong Kong and Japan
Empirical studies of parimutuel wagering markets almost universally find that bettors systematically overbet longshots and underbet favorites. This favorite-longshot bias has been interpreted as evidence of the risk preference of horse track gamblers. More recently the skewness of returns has been explicitly included into the representative bettor's utility function in addition to the variance and the mean of betting returns. The anomalous finding of risk preference in previous studies is consistent with the omission of the skewness of betting returns from the utility function. In this research we thoroughly re-examine the mean-variance-skewness utility model using horse track betting data from Hong Kong and Japan where there is no favorite-longshot bias. Using robust statistical tools, we find evidence in support of risk aversion and skewness preference when the favorites are sufficiently overweighted.
Keywords - Betting biases, Mean-variance-skewness utility model, Skewness preference, Horse track gambling