Predictability of contract failure in the U.S. financial futures markets
Suh, Jeong Ho
MetadataShow full item record
Derivatives markets have seen a great deal of changes for the last few decades. Many new contracts have been introduced to the public. Many new exchanges have been established worldwide. More than anything else, the number of participants in derivatives markets has increased in a dramatic fashion probably because of the potential capability of risk management and attractive speculative opportunities that exist in those markets. In the midst of this fastgrowing trend, futures exchanges have been very active in introducing new contracts to meet investor's needs for hedging and speculation. However, it is quite surprising that only one quarter to one third of new innovations have achieved economically viable trading volume in the following years, whereas exchanges commonly spend $300,000 to $2,000,000 for development and marketing of a new contract. The main objective of this dissertation is to examine the predictability of contract failure in futures markets. Motivated by the fact that the current literature disregards the time-series predictability of a model, we investigate whether we can predict failure of a contract using already existing information. In order to achieve this goal, we employ a survival analysis model which provides two major benefits. First, a survival analysis model can simultaneously deliver the information about the probability and timing of contract failure. This model is based on the dichotomous classification of the dependent variable, which has not been formally attempted for a general econometric model in this area. Second, a survival analysis model enables us to accommodate both time-series and crosssectional variations in a panel data set. Previous research is only concerned with the cross-sectional dimension of the relationship between contract success and "determinants."