When is upstream collusion profitable?

• Author: Rangrang Bai,Wen Zhou,Dingwei Gu,Zhiyong Yao
• Date: 01 June 2019
• Published date: 01 June 2019
• DOI: http://doi.org/10.1111/1756-2171.12271

RAND Journal of Economics

Vol.50, No. 2, Summer 2019

pp. 326–341

When is upstream collusion profitable?

Dingwei Gu∗

Zhiyong Yao∗

Wen Zhou∗∗

and

Rangrang Bai∗

Motivated by the recent antitrust cases in which Japanese auto parts suppliers colluded to

raise supply prices against their long-term collaborators, the Japanese carmakers, we study the

conditions under which an upstream collusion is profitable even after compensating downstream

direct purchasers.Oligopoly competition in successive industries is shown to give rise to a vertical

externality and a horizontal externality. If a collusive price of intermediate goods better balances

the two externalities, the collusion will raise the joint profit of all firms in the two industries and

is therefore profitable for the upstream after compensation of downstream firms.

1. Introduction

In 2008, the United States Department of Justice (DOJ) and Federal Bureau of Investi-

gation (FBI) started investigating automobile parts suppliers for anticompetitive conduct in the

United States. The perpetrators were mostly Japanese firms, including Furukawa, Yazaki, Denso,

Mitsubishi, Sumitomo, Aisan, etc. They were accused of price fixing and bid rigging for more

than a decade, from year 2000 to 2011, in the sales of automotive wire harnesses, bearings,

air-conditioning systems, windshield washer and wiper systems, compressors and condensers,

radiators, seat belts, etc.1By March 2017, 48 companies and 65 individuals had been charged;

∗Fudan University; Dingwei.gu@gmail.com, yzy@fudan.edu.cn, rrbai@fudan.edu.cn.

∗∗University of Hong Kong; wzhou@hku.hk.

We thank twoanonymous referees and the Editor for very valuable comments and suggestions. We also thank Yongmin

Chen, Jiajia Cong, Larry Qiu, and seminar participants at Fudan University and the University of Hong Kong for helpful

comments and suggestions. This research is supported by the HKU-Fudan IMBA joint research fund, and the National

Science Foundation of China (no. 71372114; 71873036). The usual caveatapplies.

1Executiveso fthe companies held regular meetings and telephone communications in the United States and Japan

to rig bids, set prices, and allocate supplies. They used code names and met in remote locations to keep secrecy,and had

measures to monitor and enforce the collusive agreements (DOJ,2013). Most of the involved parts are standard products

and were used by carmakers in a modular way. Although the costs of these parts accounted for only about 5%–8% in the

final prices of the automobiles, they were needed frequently in large quantities, which facilitated collusion and caused

substantial damages to the carmakers.

326 C2019, The RAND Corporation.

GU ET AL. / 327

all had pleaded guilty and agreed to pay a combined fine of $2.9 billion (DOJ, 2017). Alerted by

the US investigation, antitrust authorities in Japan, the European Union, Canada, South Korea,

Mexico, Australia, and China subsequently carried out their own investigations of Japanese sup-

pliers operating in their countries. All companies pleaded guilty to the respective jurisdictions

and paid various amounts of fines (DG Competition, 2013; DOJ, 2013; NDRC, 2014).

What’s puzzling about this antitrust case is that the direct victims were mostly Japanese

carmakers, including Nissan, Toyota, Honda, Mazda, Mitsubishi, and Fuji Heavy Industries.

Japanese companies, especially those in the automobile industry, are wellknown for maintaining

long-term relationships and close collaborations with their suppliers. Many observers believe that

such tight cooperation contributed greatly to the success of the Japanese automobile industry, as

close collaboration brings many benefits in the form of reduced time of new model development,

shared cost of innovation,quick response to fluctuations in market demand, and so on (Ahmadjian

and Lincoln, 2001). Why would the suppliers conspire against their long-term customers? Why

did the Japanese carmakers not report or complain about the price fixing and bid rigging? Given

that the victims were global heavyweightssuch as Toyota,Honda, and Nissan, it is hard to imagine

that the carmakers were ignorant of the conspiracy or lacked the power to strikeback. Then why

had the practice persisted for so long?

We suggest one answer: that the downstream carmakers may have received compensations

from these suppliers for the elevated supply prices. In fact, side payments between collusive

suppliers and their direct purchasers are quite common, as discussed by Schinkel, Tuinstra,

and R¨

uggeberg (2008). The fundamental question then becomes: if downstream firms must

be compensated for any damage arising from collusions among their suppliers, would these

suppliers even want to collude in the first place? This question is of general importance. In real

life, many cartels exist in supply chains in which the buyers are intermediate firms who have the

power, incentive, and information to report the cartel or sue its members, so it is reasonable to

assume that at least some cartels must compensate their direct purchasers. Standard cartel theory,

however, has focused almost exclusivelyon the incentive-compatibility issue, that is, a cartel must

prevent its members from deviating from the collusive agreement. As the compensation and the

incentive-compatibility constraints are orthogonal in nature, they are likely to lead to different

cartel behavior and consequences. Tobetter predict car tel incidence and inform antitrust policies,

therefore, it is important to fully understand the implications of the compensation constraint in

cartel practices.

In this article, we study the conditions under which upstream collusion is profitable, even

after compensating the downstream firms. Given the feasibility of many forms of side payments

across vertical business partners, the task pins down to identifying the conditions under which an

upstream collusion raises the joint profit of all firms in the two industries. Using a standard,general,

two-stage model of successive oligopoly, weshow that firms’ equilibrium choices give rise to two

externalities: a vertical externality,by which a downstream firm ignores the benefits that its output

quantity brings to upstream firms, and a horizontal externality,by which a downstream firm ignores

the damages that its output brings to other downstream firms. The vertical externality makes firms

produce too little as compared to the monopoly level, whereas the horizontal externality makes

them produce too much. If the horizontal externality dominates, firms overproduce in the oligopoly

equilibrium. By raising the input price and therefore reducing the total output, then, an upstream

collusion moves the outcome closer to the monopoly level and thereby raises the joint profit. We

show that such collusions always reduce consumer surplus and social welfare.

The analytical frameworkhelps us identify a number of parameters that affect the profitability

of such upstream collusion. First, collusion is more likely to be profitable if either industry has

more firms, or if the structure of the two industries is more balanced. Second, a smaller product

differentiation or demand concavity for the final products are conducive to upstream collusion.

Third,upstream collusion is hindered by greater cost convexity in the downstream, but not affected

by the cost structure in the upstream. These findings can inform antitrust authorities. Collusion in

a supply chain is hard to uncover,as the real victims (i.e., consumers) and the conspirators (i.e., the

C

The RAND Corporation 2019.