The Saga of the First Stock Index Futures Contract: Benchmarks, Models, and Learning

In the presence of imperfect, or perturbed benchmarks, pricing based on relative valuation or arbitrage can be affected adversely. This study highlights persistent pricing irregularities in the Kansas City Value Line (KCVL) stock index futures market. This contract is especially interesting because the underlying benchmark underwent a midstream change in definition from a complex equally weighted geometric index to an equally weighted arithmetic index. Empirical analysis reveals that during a four-year period the market displayed herd-like behavior and priced the KCVL (Geometric) contract using the simple but wrong cost-of-carry benchmark model. A dramatic change occurs around the publication of the Eytan-Harpaz (1986) model when the market turns efficient from the perspective of this new model. The market appears to have succumbed initially to a bad information cascade primarily because arbitrage was difficult to understand and carry out under complex benchmarking. The evidence provides empirical extension to the observational social learning paradigms of Simon (1955), Banerjee (1992), and Bikhchandani, Hirshleifer, and Welch (1992) that lead to behavioral herding and information cascades.