A simple in-sample test of futures market efficiency based on rolling regressions

Given the tremendous volatility of commodity prices, models used to test the efficiency of futures markets often contain unstable parameter estimates. This instability may cause the results of hypothesis tests to become sensitive to the sample period. As a result, tests based on constant parameter specifications may be inconclusive. This article proposes an in-sample test of efficiency based on rolling regressions. Using data for crude oil futures contracts traded on the New York Mercantile Exchange between 1985 and 2010, no significant evidence of inefficiency is found.