Predicting the limit-hit frequency in futures contracts

Abstract

This study demonstrates how the predicted frequency of a limit hit can assist in designing an optimal futures contract with regard to two issues — the cost of trading and the construction of an optimal limit regime. Using a logit function, we present an initial attempt to estimate the probability of a price-limit hit given the following variables: the price-limit regime, the conditional volatility and the contract price level. The estimation procedure is applied to pork bellies and oats futures. The results imply that the logit model can be employed for predicting the expected limit-hit frequency. Our findings indicate that there is an inverse relationship between the conditional probability for a limit hit and the limit size, whereas the volatility and the price level are positively correlated with the conditional probability for a limit hit. Our findings also demonstrate the magnitude of the reduction in the limit-hit frequency resulting from a given increase in the price-limit size.