Boom, bust, and collusion: Does it really matter?

Eric Fruits —  22 January 2019

A recent working paper by Hashmat Khan and Matthew Strathearn attempts to empirically link anticompetitive collusion to the boom and bust cycles of the economy.

The level of collusion is higher during a boom relative to a recession as collusion occurs more frequently when demand is increasing (entering into a collusive arrangement is more profitable and deviating from an existing cartel is less profitable). The model predicts that the number of discovered cartels and hence antitrust filings should be procyclical because the level of collusion is procyclical.

The first sentence—a hypothesis that collusion is more likely during a “boom” than in recession—seems reasonable. At the same time, a case can be made that collusion would be more likely during recession. For example, a reduced risk of entry from competitors would reduce the cost of collusion.

The second sentence, however, seems a stretch. Mainly because it doesn’t recognize the time delay between the collusive activity, the date the collusion is discovered by authorities, and the date the case is filed.

Perhaps, more importantly, it doesn’t acknowledge that many collusive arrangement span months, if not years. That span of time could include times of “boom” and times of recession. Thus, it can be argued that the date of the filing has little (or nothing) to do with the span over which the collusive activity occurred.

I did a very lazy man’s test of my criticisms. I looked at six of the filings cited by Khan and Strathearn for the year 2011, a “boom” year with a high number of horizontal price fixing cases filed.

khanstrathearn

My first suspicion was correct. In these six cases, an average of more than three years passed from the date of the last collusive activity and the date the case was filed. Thus, whether the economy is a boom or bust when the case is filed provides no useful information regarding the state of the economy when the collusion occurred.

Nevertheless, my lazy man’s small sample test provides some interesting—and I hope useful—information regarding Khan and Strathearn’s conclusions.

  1. From July 2001 through September 2009, 24 of the 99 months were in recession. In other words, during this period, there was a 24 percent chance the economy was in recession in any given month.
  2. Five of the six collusive arrangements began when the economy was in recovery. Only one began during a recession. This may seem to support their conclusion that collusive activity is more likely during a recovery. However, even if the arrangements began randomly, there would be a 55 percent chance that that five or more began during a recovery. So, you can’t read too much into the observation that most of the collusive agreements began during a “boom.”
  3. In two of the cases, the collusive activity occurred during a span of time that had no recession. The chances of this happening randomly is less than 1 in 20,000, supporting their conclusion regarding collusive activity and the business cycle.

Khan and Strathearn fall short in linking collusive activity to the business cycle but do a good job of linking antitrust enforcement activities to the business cycle. The information they use from the DOJ website is sufficient to determine when the collusive activity occurred—but it’ll take more vigorous “scrubbing” (their word) of the site to get the relevant data.

The bigger question, however, is the relevance of this research. Naturally, one could argue this line of research indicates that competition authorities should be extra vigilant during a booming economy. Yet, Adam Smith famously noted, “People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices.” This suggests that collusive activity—or the temptation to engage in such activity—is always and everywhere present, regardless of the business cycle.

 

Eric Fruits

Posts

Eric Fruits, Ph.D. is Chief Economist at the International Center for Law & Economics and Economics International Corp. He is the Oregon Association of Realtors Faculty Fellow at Portland State University. He has written peer-reviewed articles on initial public offerings (IPOs), the municipal bond market, real estate markets, and the formation and operation of cartels. His economic analysis has been widely cited and has been published in The Economist and the Wall Street Journal. Dr. Fruits is an antitrust expert who has written articles on price fixing and cartels for the top-tier Journal of Law and Economics. He has assisted in the review of several mergers including Sysco-US Foods, Exxon-Mobil, BP-Arco, and Nestle-Ralston. He has worked on many antitrust lawsuits, including Weyerhaeuser v. Ross-Simmons, a predatory bidding case that was ultimately decided by the United States Supreme Court. As an expert in statistics, he has provided expert opinions and testimony regarding market manipulation, real estate transactions, profit projections, agricultural commodities, and war crimes allegations. His expert testimony has been submitted to state courts, federal courts, and an international court.

One response to Boom, bust, and collusion: Does it really matter?

  1. 

    I’m not sure I’m getting this. When demand is increasing competitors can (always) increase prices without the risk of collusion. In recession when demand is falling, the risk of competitive price reductions (and losses) would create incentives to “stabilize” industry prices at or near industry average costs and at least above variable costs. My study of some of the older and more famous price conspiracies (Addyston Pipe, Trenton Potteries, Electrical Equipment {1961]) revealed that collusion occurred during periods of weak demand; which is not to say that the prices were ever really “fixed” and rigid since the companies cheated and offered secret discounts to favored customers. Thus cartels can be unstable both when demand is increasing (who needs collusion) and when it is decreasing (cheating on agreements to gain an advantage in a shrinking market).