| Success of the cash settlement futures contract depends, in part, on the accuracy and precision of its settlement index. Providing (at contract expiry) an estimate of the commercial value of the contract's underlying asset is the principal purpose of the settlement index. The better the estimate, the more able the contract is to perform, what are usually described as, it's most important functions, hedging and price discovery. The general focus of this research on problems confronting the designers of cash settlement provisions. More specifically, the primary concern is to identify and construct the optimal settlement index.;Reliance on external pricing agents, since they may have an incentive to distort or manipulate the outcome of the settlement index, subjects the index to the possibility of intentional bias. The decision to misreport cash price observations may also be influenced by strategic considerations. For example, an individual reporting agent may attempt to anticipate the prices other reporting agents are likely to report, thereby being more able distort the index. Moreover, dependence on sample data injects sampling error as well as the likelihood of non-sampling error, both are examples of unintentional bias.;It is difficult to know exactly what the exchange's index design criteria is; however, for this research it is assumed the exchange is interested in choosing the settlement index (i.e., estimator) with minimum mean squared error (MSE). (As it turned out, results showed this assumption to be well justified.) The basic problem analyzed in this research is one of selecting the settlement index with minimum MSE, admitting the possibility that the available may be subject to both intentional and unintentional bias.;Results indicate that use of a modified trimmed mean or the bootstrap procedure may serve to satisfy the exchange's index design objective. In fact, under quite general conditions, both procedures perform better than the estimator now employed by both the CME and CBOT.;When the futures exchange has access to ideal cash price information, index design problems are likely to be nonexistent. However, the further cash price information diverges from the ideal the more difficult it is to construct a reliable index. Furthermore, divergence from the ideal is perhaps greatest in that situation where the exchange is forced to solicit a sample of cash price observations from external pricing agents. Both the Chicago Mercantile (CME) Exchange and the Chicago Board of Trade (CBOT), for purposes of settling the Eurodollar and Municipal Bond contracts, respectively, obtain cash price information in this way.;Finally, it is shown, through the use of a Bayesian - Nash price reporting model, the incentive to misreport cash prices is a function of index design. Methods for minimizing this incentive are also identified. |
