How Far Can Screens Go in Distinguishing Explicit from Tacit Collusion? New Evidence on the Libor Setting

Rosa Abrantes-Metz, Albert Metz, Mar 13, 2012

A class of empirical analysis known as “screening” uses commonly available data such as prices, costs, market shares, bids, transaction quotes, spreads, and volumes to identify patterns that are anomalous or highly improbable under ordinary competitive conditions. Screens can signal the possibility of cheating in a market or industry, as well as identify who may be engaging in that behavior. Surveys of screening methodologies and their multiple applications can be found in Abrantes-Metz & Bajari and in Harrington. The use of these methods in litigation is detailed in the American Bar Association’s 2010 volume Proof of Conspiracy under Antitrust Federal Laws.

Recently, large-scale investigations have been launched around the world on allegations of possible collusion and manipulation of the London Interbank Offered Rate (“Libor”). These investigations followed empirical research that highlighted anomalous patterns in the Libor data, beginning with two articles published in the Wall Street Journal in April and May of 2008. These were quickly followed by a paper we co-authored with Michael Kraten and Gim Seow in August 2008, which applied empirical screens to identify several unexpected patterns of the Libor and the underlying banks’ quotes for the period January 2007 through May 2008. Abrantes-Metz, Judge, & Villas-Boas presented another analysis in 2010, providing additional evidence of unusual patterns beginning earlier in 2006, which in turn was followed by additional research by other authors.

The Libor is determined from sealed daily quotes submitted by sixteen member banks. Empirical research has shown that for a period of nearly a year, the Libor was essentially constant. This is the first anomaly. The second anomaly, which is less well understood, is the virtual unanimity of individual quotes submitted by the member banks. These anomalies lead us to ask whether coordination of some type may have been involved.

While screens can highlight such anomalous patterns, it is unclear whether they can differentiate between the various possible causes of those patterns. In principle, this unanimity in quotes across banks could simply reflect a non-cooperative outcome, or it could be the result of collusion. But whether that collusion was explicit, or a form of tacit or strategic collusion, is not immediately obvious.

Though distinguishing explicit from tacit collusion may be very difficult through screening, this is the challenge we take up in this article. We explore, in the context of the Libor, whether screens can move one step further and distinguish illegal (explicit) from legal (tacit) collusion. While we have always argued that a purely empirical analysis of market outcomes can never be the final proof of illegal behavior, under particular circumstances screens can indeed provide additional evidence to assess the more likely form of collusion.

Links to Full Content