The John M. Olin Center

Paper Abstract

Untitled Document

70. Reid B. Stevens & Jeffery Y. Zhang, Robo-Regulator: Algorithmic Detection of Market Manipulation, 04/2017.

Abstract: In recent years, a tremendous increase of trading activity in physical commodity markets has created concern among policymakers and regulators over the price impact of speculation and manipulation. The former is a regular market activity, where rational market agents act in anticipation of a future shock. The latter is irregular, and involves market agents hoarding assets in order to artificially raise their prices. Detecting manipulation is not a costless enterprise. The objective of this article is to facilitate that task by providing regulators with an algorithm--a robo-regulator--that acts as a fire alarm for detecting manipulative schemes. This algorithm uses a novel approach that exploits a commodity’s production complement in order to test whether the commodity is affected by manipulation. We apply the algorithm to three recent, well-publicized U.S. Senate investigations of commodity price spikes in order to test whether any episode was caused by manipulation. The algorithm shows that the crude oil price increase in the early-2000s was not caused by manipulative schemes, but the wheat futures price spike in the mid-2000s and the aluminum regional price spike in the early-2010s were caused by manipulation.