Nationalistic bias in collusion prosecution: the case for international antitrust agreements

• Published date: 01 July 2024
• Author: Filomena Garcia,Jose Manuel Paz y Miño,Gustavo Torrens
• Date: 01 July 2024
• DOI: http://doi.org/10.1111/sjoe.12558

Scand. J. of Economics 126(3), 489–528, 2024

DOI: 10.1111/sjoe.12558

Nationalistic bias in collusion prosecution:

the case for international antitrust

agreements

Filomena Garcia†

North Carolina State University, Raleigh NC 27695, USA

f‌igarcia@ncsu.edu

Jose Manuel Paz y Mi˜

no

Universidad de Chile, Santiago, Chile

jpazm@fen.uchile.cl

Gustavo Torrens

Indiana University, Bloomington, IN 47405, USA

gtorrens@indiana.edu

Abstract

We study the incentives of competition authorities to prosecute collusive practices of domestic

and foreign f‌irms in a multimarket contact model between two f‌irms operating in two countries.

In equilibrium, the country of origin of the f‌irms might prefer to delay prosecution to protect

prof‌its in foreign markets. This strategic delay is valuable because prosecution in the country

of origin of the f‌irms activates an information spillover that triggers prosecution in the foreign

country. Prosecution delays, however, are suboptimal under global welfare. With multiple

industries, both countries can be better off under integration or signing an international antitrust

agreement.

Keywords: Antitrust policy; international antitrust agreements; multimarket collusion; strategic

prosecution

JEL classif‌ication:F23; F53; L41; K21

1. Introduction

Globalization has brought new challenges to antitrust policy (Connor, 2004;

Connor and Helmers, 2007; Barnett, 2007). National competition authorities

have reacted to the rise of globalization by devoting special attention

to collusive practices involving multinational companies. The challenges

associated with international antitrust enforcement have also led to proposals

†Also aff‌iliated with REM/UECE ISEG, University of Lisbon.

c

2023 F¨

oreningen f¨

or utgivande av the SJE.

490 Nationalistic bias in collusion prosecution

for increasing international cooperation among competition authorities

(Barnett, 2007). Nevertheless, most of the formal literature that studies

anticompetitive behavior focuses on closed economies and disregards

interactions among national competition authorities.1The aim of this paper

is to provide a formal analysis of the incentives of antitrust authorities to

prosecute domestic and foreign f‌irms involved in anticompetitive behavior

when prosecution in one country generates informational spillovers to other

countries.

We show that, in equilibrium, some countries may delay prosecution of

national f‌irms even in the presence of evidence of collusive behavior in the

domestic market. This strategic delay is potentially damaging for the rest of

the world. Prosecution delays are also not optimal from the point of view of

global welfare. Moreover, we show that in a multi-industry world, countries

can be better off under an integrated competition authority that maximizes

the combined welfare or, equivalently, signing an international agreement in

which they commit to have competition authorities that only target consumer

welfare. Under integration, each country loses the collusion prof‌its in the

foreign market, but it gains the increase in consumer surpluses in domestic

markets.

We start off our analysis in Section 3by developing a two-country model

of collusion and antitrust policy in which two multinational f‌irms operate in

both countries. In each country, there is a competition authority in charge

of enforcing antitrust laws that targets total welfare. In every period, the

f‌irms decide whether to collude or compete. Afterwards, each competition

authority receives information about the behavior of the multinational

f‌irms and decides whether to prosecute the f‌irms or not in a sequential

manner. First, the competition authority in the country of origin of the

f‌irms receives a signal that directly informs about the collusion behavior of

the f‌irms in the domestic market as well as in the foreign market; using

this information, it decides whether to prosecute the f‌irms or not. Second,

the competition authority of the foreign country observes the prosecution

decision of the country of origin, possibly receives an independent signal

about the f‌irms’ behavior, and, f‌inally, decides whether to prosecute the f‌irms

or not.2

1We discuss some interesting exceptions in Section 2.

2The idea is that it tends to be more diff‌icult for a country to independently detect collusive

behavior of foreign f‌irms than domestic f‌irms. As a consequence, competition authorities often

rely on the information that spills from f‌irms being prosecuted in their country of origin. Indeed,

the observation of the collusion cases brought about by foreign antitrust authorities has become

a standard screening procedure (in addition to other forms of investigation) to detect collusion.

See OECD (2013) for a report on best practices in cartel investigations.

c

2023 F¨

oreningen f¨

or utgivande av the SJE.

F.Garcia,J.M.PazyMi

˜

no, and G. Torrens 491

Our baseline model in Section 4considers a scenario where the competition

authority of the foreign country has no independent detection ability and must

fully rely on the prosecution decision of the country of origin to obtain

evidence about collusion in its own market before it can decide whether to

prosecute the f‌irms. The model generates three key results.

First, the foreign country always prosecutes foreign-owned f‌irms as soon

as there is information of prosecution in the country of origin.

Second, the competition authority in the country of origin prosecutes a

domestic f‌irm that is colluding in both the domestic and foreign markets

if the gain in consumer surplus exceeds the reduction in prof‌its (including

f‌ines) from both domestic and foreign markets. The intuition is the following.

Prosecuting collusion in the country of origin always increases domestic

surplus because of the reduction in deadweight loss. However, because of

the informational spillover across competition authorities, prosecution in the

country of origin triggers prosecution in the foreign country, which activates

f‌ines and eliminates the prof‌its from collusion that national f‌irms are making

in the foreign country. Thus, in equilibrium, it could be the case that the

competition authority in the country of origin of the f‌irms chooses not to

prosecute collusion.

Third, even if the country of origin prosecutes the f‌irms, the foreign

country is affected by the antitrust policy of the country of origin. In

particular, when the probability of detecting collusion in the country of

origin is high, f‌irms only collude in the foreign country and they are never

detected. This suggests that multinational companies might be competing in

their home country – usually a developed country with a professional and

well-funded competition authority – but colluding in foreign markets – usually

developing countries that do not have a competition authority capable of

independently detecting collusion. In a variation of our baseline model with

short-run informational spillovers, multinational companies might collude in

their home country until detected and prosecuted, and then start colluding in

foreign markets.

Having characterized the key trade-offs that drive the decisions of f‌irms

and competition authorities when the foreign country has no independent

detection ability, in Section 5we study how the solution of the model changes

when the foreign country can also detect collusion independently. More

specif‌ically, we show that the equilibrium is characterized by f‌inite delays in

prosecution. To the best of our knowledge, this is the f‌irst paper that formally

models and obtains prosecution delays as an equilibrium. The intuition for

why delays in prosecution are f‌inite in this extended model is as follows.

The country of origin’s decision to restrain from prosecuting is less valuable

than in the baseline model, as now foreign collusion prof‌its can no longer be

indef‌initely protected with a lenient prosecution policy. Therefore, the country

of origin is more willing to prosecute the f‌irms as soon as collusion is detected,

c

2023 F¨

oreningen f¨

or utgivande av the SJE.