Empirical Methods in Competition Analysis Applying Benford's Law to the Western Australian Petroleum Market

Abstract

In this paper, we develop a robust empirical indicator for screening and detecting statistical anomalies that may provide circumstantial evidence of explicit collusion between companies or a cartelized market. The paper deals with the question of whether in the context of forensic research on cartels a Benford test can be used to improve target selection in order to increase effectiveness and efficiency in identifying companies and markets worthy of closer scrutiny. In various contexts it has been shown that a deviation from the Benford distribution can indicate fraud or data manipulation. We apply Benford’s law to released price data of the Western Australian petroleum market from 2004 to 2012. We show that the deviation from the Benford distribution in relative margin changes is positively correlated with well known indicators of collusion and conclude that anti-competitive behavior can be identified by the Benford method. Moreover, we identify periods, brands and petroleum products showing a concentration of notably higher deviations from Benford’s law.