The impact of vertical integration on losses from collusion

Highlights

• Upstream collusion can impose losses on an independent downstream supplier (D).

• The losses depend on whether D’s downstream rival is vertically-integrated.

• Vertical integration (VI) increases D’s loss when D is a relatively weak competitor.

• VI can reduce D’s loss under price competition when D is a strong competitor.

Abstract

Upstream collusion that increases the price of an input can harm an independent downstream producer (D). We ask whether this harm is more or less pronounced when D’s downstream rival is a vertically integrated producer. We find that such vertical integration increases D’s loss from collusion when D is not a particularly strong competitor. However, when D is a sufficiently strong competitor, vertical integration can reduce D’s loss from collusion when price competition prevails downstream.

Introduction

Just as the threat of government prosecution can help to deter illegal collusion by producers of key inputs, so can the threat of private litigation by input purchasers.¹ Basso and Ross (2010, p. 896) note “the trend... toward continued and even growing support for private actions” to recover losses from collusion.² The European Commission (EC)’s White Paper proposes “specific measures that would ensure ... all victims of infringements of EC competition law have access to effective redress mechanisms” (EC, 2008, §1.2).³

The incentives for private litigation against colluding input suppliers depend in part on accepted principles regarding the harms caused by collusion. A potential plaintiff, particularly a risk-averse plaintiff, may be reluctant to sue for actual harm suffered if courts do not routinely recognize the type of harm in question and measure the magnitude of the harm with substantial accuracy.⁴ Considerable effort has been devoted to identifying potential sources of harm from collusion by input suppliers, as explained further below. However, the literature does not provide systematic guidance on how vertical integration affects the harm that collusion imposes on input purchasers that compete in imperfectly competitive downstream markets. This guidance is important because colluding upstream suppliers often operate in downstream markets also.

To illustrate, the nation’s sugar refiners colluded in the 1970s to set the price at which they sold refined sugar to candy manufacturers.⁵ Some of the candy manufacturers were vertically-integrated with sugar refiners (e.g., subsidiaries of Borden) whereas other manufacturers were unaffiliated producers. Suppliers of TFT panels were also accused of colluding in the late 1990$\prime$s and early 2000$\prime$s to raise the prices at which they sold the panels to manufacturers of many products, including cellular telephones.⁶ The accused panel suppliers included Samsung and Toshiba, companies that also manufactured cellular telephones and competed against firms like Motorola and Nokia that did not manufacture TFT panels.⁷’⁸

The primary purpose of this research is to determine whether the damage that collusion by upstream input suppliers imposes on an unaffiliated input purchaser is more pronounced when a rival downstream producer is vertically integrated with one of the upstream suppliers.⁹ Our formal analysis focuses on the interaction between two upstream suppliers (U1 and U2) and two downstream producers (D1 and D2) of differentiated retail products. We allow for the possibility that U1 and D1 might be vertically integrated. We find that such vertical integration increases the loss that upstream collusion between U1 and U2 imposes on the unaffiliated downstream producer (D2) when D1 is a relatively strong competitor. However, vertical integration can reduce D2’s loss from collusion under downstream price competition when D2 is a sufficiently strong competitor.

D2’s loss from collusion is always larger under vertical integration (VI) than under vertical separation (VS) in the presence of downstream quantity competition.¹⁰ This conclusion reflects two effects that arise under VI. First, U1 and U2 value D1’s downstream profit and so are inclined to charge D2 a relatively high input price in order to enhance D1’s profit. Second, the relatively low input price ($c^u$) that D1 perceives under VI induces it to act aggressively in its competition with D2.¹¹

These two effects also ensure that D2’s loss from collusion is larger under VI than under VS in the presence of downstream price competition as long as D2’s competitive strength (${\Delta }_2$) is no greater than D1’s competitive strength (${\Delta }_1$).¹² However, D2’s loss from collusion can be smaller under VI than under VS in the presence of downstream price competition when ${\Delta }_2>{\Delta }_1$. Downstream competition is relatively intense under price competition, particularly under VI where D1 perceives a relatively low input price ($c^u$). The intense downstream competition leads U1 and U2 to focus on securing upstream profit from input sales to D2 rather than enhancing D1’s downstream profit when D1 is a relatively weak competitor. U1 and U2 ensure substantial sales of the input to D2 under VI by charging D2 a lower input price under VI than under VS. The lower input price under VI ensures that when ${\Delta }_2$ is sufficiently pronounced, D2’s loss from collusion is smaller under VI than under VS in the presence of downstream price competition.

The literature observes that, in addition to the “direct” harm collusion by input suppliers can impose on input purchasers, such collusion can impose “indirect” harm on other parties. For example, retail customers can be harmed by the price increases retail producers implement in response to the elevated input prices they face due to collusion by input suppliers (e.g. Kosicki, Cahill, 2006, Han, Schinkel, Tuinstra, 2009, Basso, Ross, 2010, Boone, Müller, 2012, Brander, Ross, 2017).¹³ Retail customers of substitute products can also be harmed when the prices of these products rise in response to the higher prices charged by customers of the colluding upstream suppliers (e.g., Inderst et al., 2014).¹⁴

We abstract from these indirect harms that collusion can impose. Instead, like Verboven and Van Dijk (2009) (VV), we focus on the losses that upstream collusion by input suppliers imposes directly on input purchasers. VV identify three effects of upstream price collusion on the profit of retail firms that purchase the input: higher input costs, increased revenue due to higher equilibrium retail prices, and reduced revenue due to foregone sales caused by higher retail prices.¹⁵ The authors observe that the magnitudes of these effects can vary if a downstream producer is vertically integrated. However, their study is not designed to provide systematic guidance on how vertical integration affects losses from collusion.¹⁶

We offer a first step in providing such guidance. We do so in a streamlined setting where, when the upstream suppliers collude, they do so perfectly and costlessly, both in the presence of vertical integration and in its absence. Thus, we do not analyze whether vertical integration facilitates or hinders collusion.¹⁷ Rather, we investigate how vertical integration affects the losses that collusion among input suppliers imposes on input purchasers. Thus, our analysis might be viewed as an inquiry into whether standard calculations of losses from collusion that do not consider the prevailing vertical industry structure tend to understate or overstate the losses that actually arise when some colluding input suppliers are vertically integrated.

The analysis proceeds as follows. Section 2 describes the key elements of our model. Section 3 characterizes outcomes in the absence of collusion. Sections 4 and 5 characterize equilibrium outcomes and losses from upstream collusion in the presence of downstream quantity competition and downstream price competition, respectively. Section 6 discusses extensions of our analysis and provides concluding observations.¹⁸