A New Merger Tool Protects Consumers from Limits of the Cournot Effect

In theory, mergers of complements can benefit firms and consumers through the Cournot effect. The theoretical model on competition which 19th century French mathematician Antoine Augustin Cournot used to derive this effect presumes that there are two monopolist suppliers of complementary inputs (for example, copper and zinc). The suppliers each set the unit price for their respective input (for example, to a downstream manufacturer of brass) at a point that maximizes their own profit. Each supplier disregards the negative externality it exerts on the other: the higher its input price, the higher a manufacturer’s consumer price, the lower the quantity sold and, consequently, the lower the profit of the other supplier. If the suppliers merge (in what is called a merger of complements), they internalize the negative externality and reduce input prices to increase sales. All thereby benefit: the merging parties, the downstream firm, and consumers. This positive merger outcome is the Cournot effect.

The Cournot effect is widely accepted and thought to apply generally to mergers of complements. In actuality, it applies only to very particular circumstances.

The limitation of Cournot’s model is that Cournot modeled suppliers as monopolists. Such monopolists set high prices. But what about scenarios where competition led the merging parties to set much lower prices pre-merger? In these situations, the merged entity finds it unprofitable to decrease its already low prices because it earns little profit on any additional units sold. Instead, a merged entity that faces sufficiently strong competition wants to increase prices: it may use the merger to pursue strategies aimed at weakening competitive constraints and raising prices. There are various such anti-competitive strategies, in particular tying or bundling.

The conclusion that mergers of complements are more likely to raise subsequent input prices when the pre-merger market is more competitive may seem surprising as it differs from  intuitions based on another type of merger: horizontal mergers (where two direct competitors merge). Indeed, Valletti and Zenger find that horizontal 'mergers are more likely to cause competitive concerns the higher firms’ pricing power is to begin with.'

Given the unsettled literature, how does one identify, in practice, whether the Cournot effect materializes in the case of mergers of complements? A new merger tool I developed resolves this problem by identifying if suppliers are sufficiently constrained pre-merger to rule out a price decrease post-merger. Simplistically, the tool analyzes the suppliers’ margins relative to a downstream firm’s margin to reveal the necessary information. If suppliers earn fewer dollars of profit per unit sold relative to the downstream firm, suppliers are sufficiently constrained. One can then dismiss the claimed benefits of the Cournot effect as inconsistent with margins—ie with observable data—and one should investigate possible anticompetitive strategies.

For example, imagine that two of Apple’s suppliers were to merge. Data shows that all of Apple’s suppliers combined earn less profit per iPhone sold (and therefore a lower margin) than Apple. Pre-merger, suppose the two merging parties each earn $5 of profit per iPhone. Would the merged entity really want to lower its prices (and thereby its margin) in exchange for a higher sales volume? No. Observable margin data indicates that such a merged entity would want to obtain higher prices.

Cournot’s effect is similar to a broader and intuitive argument in economics that firms reduce negative externalities after they internalize them. However, this is not necessarily so. In mergers of complements, it is so only in very specific circumstances. We should not be too quick to accept intuitive arguments. Instead, we should endeavor to test them.

Empirically testing merger theories is crucial because many antitrust authorities use theories from industrial organization (IO) to help decide cases of mergers and acquisitions. However, proponents and opponents of any given case can typically submit IO analyses (often from academic economists) with opposing conclusions and antitrust implications to defend their positions. As Paul Krugman wrote about IO when one of its luminaries, Jean Tirole, received the Nobel Prize in Economics in 2014, 'there came a [… time …] when a smart grad student could produce a model to justify anything.' In the empiricist spirit of George Stigler, we must be vigilant of dominant firms hiring economic experts that excel at selecting those particular modeling assumptions that lead to desired results. Every model uses assumptions. The key question is if a model uses assumptions that lead to distorted or even opposing predictions relative to the true merger effects.

In antitrust cases in the US, there are frequently wide disparities between the views of the Federal Trade Commission, the Department of Justice, and the courts. To reduce those disparities and identify which theories are either consistent or inconsistent with a given case, we need more practical tools that do not rely solely on theoretical assumptions. With regard to Krugman’s critique, we need theories that provide testable conditions that data can either support or, crucially, falsify. Ideally, these new merger tools would be simple, transparent, and require little but observable data. Such tools would serve to counteract convenient theories cooked up by vested interests. I propose one such tool in this article.

Alessandro S. Kadner-Graziano is a PhD researcher in economics at the University of Bayreuth, Germany.

This post first appeared on Chicago Booth's ProMarket blog (here).