Price‐Fixing Hits Home: An Empirical Study of Price Fixing Conspiracies in the United States
Margaret C. Levenstein and Valerie Y. Suslow
University of Michigan
Proposal
January 15, 2014
Abstract
More than a century after the passage of the Sherman Antitrust Act and twenty years after the adoption
of more aggressive enforcement by the U.S. Justice Department (DOJ), cartels continue to affect US
markets. In this paper, we analyze data that consist of all Section 1, Sherman Act price fixing cases
between 1961 and 2012. The long time span of these data allows us to examine a number of issues:
asymmetric, firm‐level measures of impatience, cartel formation and breakup over the business cycle,
and the impact of changes in enforcement.
More than a century after the passage of the Sherman Antitrust Act and twenty years after the adoption
of more aggressive enforcement by the U.S. Justice Department (DOJ), cartels continue to affect US
markets. Fifty new criminal cases, most of which were price fixing, were filed in 2013; over a billion
dollars in criminal fines were levied.1 In this paper, we study the formation and breakup of all price‐
fixing cartels convicted of operating in the United States in the last half‐century. Understanding how
and why these cartels continue to form and persist is critical to developing better models of firm
behavior and better competition policy.
In this paper, we analyze data that consist of all Section 1, Sherman Act price fixing cases between 1961
and 2012. This analysis builds on prior work, including Posner’s (1970) landmark study of antitrust
enforcement in the United States which catalogued 989 cases involving a horizontal conspiracy between
1890 and 1969. Posner, like others who have since studied U.S. price‐fixing, used the Commerce
Clearing House Trade Regulation Reporter (CCH) to create a picture of the breadth of cartel activity in
the United States. The CCH reports all antitrust convictions by the U.S. Department of Justice (DOJ).
Using CCH data from 1955‐1997, Gallo (2000) reports 688 horizontal per se violations. Bryant and
Eckard (1991) use CCH data to estimate the probability of cartel detection between 1961 and 1988.
Miller (2009) examines U.S. Department of Justice cartel indictments between 1985 and 2005. We have
created a new sample derived from CCH reports and building on data provided by Bryant and Eckard for
the early part of the period. Figure 1 presents cartel duration for a preliminary sample from 1961‐2007.
As is often the case, a large number of cartels break up very quickly, while others endure for decades.
Table 1 presents the sectoral distribution of these cartels. Most were in manufacturing, but virtually all
segments of the economy are represented.
The long time span of these data allows us to examine a number of issues. Approximately half way
through the period of the sample, the Justice Department strengthened its leniency policy. This policy
change increased firms’ incentive to turn in cartels in which they were participating. As we have argued,
the decision to apply for amnesty is an economic decision (Levenstein and Suslow 2011). These data
allow us to ask whether economic factors became more salient in cartel discovery and breakup after this
policy change.
Over this same time period, the DOJ began aggressively pursuing international cartels which had been
considered insignificant in their impact on the US market and out of bounds for prosecution by US
courts. We compare cartels consisting purely of US firms and those that have international membership.
Finally, it is well‐known that different administrations have different priorities for antitrust policy. We
examine changes in the relationship between political and DOJ leadership and cartel enforcement.
One of the few areas of consensus flowing from theoretical research on collusion is that it requires
patience. In general there is a monotonic relationship between the ability of a cartel to sustain and
profit from collusion and the discount rate of its members. There have been very few empirical studies
1 Statement Of William J. Baer, Assistant Attorney General, Antitrust Division, and Ronald T. Hosko, Assistant
Director, Criminal Investigative Division, Federal Bureau of Investigation, before the Subcommittee on Antitrust,
Competition Policy and Consumer Rights, Committee on the Judiciary, United States Senate Hearing on “Cartel
Prosecution: Stopping price fixers and protecting consumers” Presented on November 14, 2013, p. 2.
of the relationship between the discount rate and collusion. Previously, using a sample of international
cartel prosecutions by the DOJ and European Commission (EC), we found that market interest rates had
little impact on cartel stability, but that firm‐level financial weakness, which could raise the discount
rate, was an important determinant of cartel breakup (Levenstein and Suslow 2011).
In this paper, we continue to pursue the relationship between patience and cartel stability. For the last
twenty‐five years, for which we can obtain firm‐specific financial measures, we replicate the analysis of
our previous research examining the relationship between cartel stability and the solvency of cartel
members. For the earlier period, for which firm‐level financials are not available, we focus on how
changes in industry‐level leverage and liquidity affect cartel stability.
In contrast, there is little consensus on how the business cycle affects cartels. This question has been
the focus of much of the prior empirical research on cartel breakup (see Levenstein and Suslow 2006,
2011 for discussion of this literature). We examine the impact of business cycle fluctuations on cartel
breakup, comparing measures of observable and unobservable fluctuations in demand. Our previous
research suggests that economy‐wide fluctuations in demand have had little impact on cartel breakup in
contemporary cartels (Levenstein and Suslow 2011). That analysis covered a period of remarkable
macroeconomic stability in the United States so may not provide a good setting for addressing this
question. These data, reaching back over a number of business cycles, provide a better basis for analysis.
We also examine the cyclicality of cartel formation. Competition observers have often claimed that
cartels are more likely to form during recessions, but there has been no systematic study of this
question. Admittedly measuring cartel formation is challenging. Using samples from prior research and
NBER business cycle dates, there is no obvious pattern (Table 2). Cartels form during recessions, but
only in proportion to the frequency of recessions during the particular period of study (Levenstein and
Suslow 2012).
Bibliography
Bryant, Peter G. and E. Woodrow Eckard (1991), Price Fixing: The Probability of Getting Caught, Review
of Economics and Statistics 73:531‐36.
Gallo, Joseph C., Joseph L. Craycraft, Kenneth Dau‐Schmidt, and Charles A. Parker (2000). “Department
of Justice Antitrust Enforcement, 1955–1997: An Empirical Study.” Review of Industrial
Organization, 17(1): 75–133.
Levenstein, Margaret C. and Valerie Y. Suslow (2006), What Determines Cartel Success?, Journal of
Economic Literature 44(1):43‐95.
_____ _____ (2011), Breaking Up Is Hard to Do: Determinants of Cartel Duration, Journal of Law and
Economics 54:455‐92.
_____ _____ (2012), Cartels and Collusion ‐ Empirical Evidence. Ross School of Business Paper No. 1182.
Available at SSRN: http://ssrn.com/abstract=2182565.
Miller, Nathan H. (2009), Strategic Leniency and Cartel Enforcement, American Economic
Review 99:750‐68.
Posner, Richard A. (1970). “A Statistical Study of Antitrust Enforcement.” Journal of Law and Economics,
13(2): 365–419.
Suslow, Valerie Y. (2005). “Cartel Contract Duration: Empirical Evidence from Inter‐War International
Cartels.” Industrial and Corporate Change, 14(5): 705–44.
Table 1: US DOJ Cartel Convictions by Industry, 1961‐2007
2‐Digit NAICS
Industry Definition Frequency Percentage
21 Mining, Quarrying, and Oil and Gas Extraction 7 3.68%
22 Utilities 1 0.53%
23 Construction 5 2.63%
31 ‐ 33 Manufacturing 147 77.00%
42 Wholesale Trade 2 1.05%
44‐45 Retail Trade 7 3.68%
48 Transportation and Warehousing 10 5.26%
51 Information 1 0.53%
52 Finance and Insurance 1 0.53%
54 Professional, Scientific and Technical Services 1 0.53%
Administrative, Support, Waste Management and
56 Remediation Services 3 1.58%
62 Health Care and Social Assistance 3 1.58%
81 Other Services (Except Public Administration) 2 1.05%
Table 2: Cartel formation over the business cycle
Study Scope (sample Years Formed during Economy in
L&S 2006 size) recession recession
L&S 2011
U.S. case studies 1857‐1950 67% 63%
(12)
International (81) 1971‐2004 11% 13%
Suslow 2005 International (71) 1918‐1938 55% 43%
Bryant & Eckard U.S. (184) 1932‐85 39% 39%
1991
Number of Cartels
Figure 1:
Cartel Sample Distribution by Duration (Years)
50
45
40
35
30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 18 19 20 21 23 24 27 28
Years