摘要：This paper uses a class of quantile functions to develop a new time series model for studying financial price behaviors through the tail properties of the price instead of the volatilities (variances) of the price. The model takes the updated information into account and characterizes the price behaviors using a tail order measure which helps forecast how volatile the prices will be, and a tail balance measure that helps estimate whether an investment tends to gain or tends to lose. The model parameters can be estimated using the method of maximum likelihood, and two real data sets are analyzed to show the potential usefulness of our proposed model.
关键词：tail property; tail modeling; quantile function
