Horizontal integration can be good for you

It has long been recognised that vertical integration can enhance efficiency. It can deal with hold-up problems etc. But horizontal integration has been looked at with much more suspicion.

The conventional view is that anticompetitive mergers increase industry concentration and hence increase market power, harm competition ex post, and therefore need to be carefully reviewed and possibly restricted by regulators. Hence, regulators, such as the Antitrust Division of the Department of Justice or the Federal Trade Commission, have the mandate to prevent situations that “excessively” transfer welfare from consumers to firms via buildups of dominant positions or firms with disproportionate market power, including mergers perceived to be anticompetitive.

Are these policies effective or desirable? We take a dynamic approach and find the answer to be No in both cases.

This quote is from a posting at the Pro-Market blog by Dirk Hackbarth and Bart Taub in which they ask Can Horizontal Mergers Actually Boost Competition?

To reach these conclusions we built a dynamic, noisy collusion model that captures firms’ optimal output strategies prior to a merger. [...] We thus focus only on the desire of firms to collude prior to merging or potentially to merge if collusion fails.

The conventional view fails to account for dynamics. Firms in our dynamic model are forward-looking, aware that they are in a dynamic cartel-like situation, but are unable to directly observe the actions of the rival firm, which would enable them to enforce the cartel. The inability of each firm to observe the other firm’s output reflects the real world: regulators punish firms that directly track and coordinate with each other’s actions for market power purposes.

Hackbarth and Taub go on to explain,

The conventional view fails to account for dynamics. Firms in our dynamic model are forward-looking, aware that they are in a dynamic cartel-like situation, but are unable to directly observe the actions of the rival firm, which would enable them to enforce the cartel. The inability of each firm to observe the other firm’s output reflects the real world: regulators punish firms that directly track and coordinate with each other’s actions for market power purposes.

Because they are blocked from observing each other directly, firms are unable to punish their rival for directly perceived deviations from collusion–that is, for producing too much in order to realize temporarily higher profits at the expense of the other firm. The inability to directly observe and punish deviations therefore requires a tacit collusion arrangement, in which firms attempt to observe each other indirectly–via prices. This indirect observation is imperfect, however, because prices are affected by random influences, in addition to the effects of the firms’ output choices.

Because of the random influences a firm can mistakenly appear to produce too much output. Under the tacit collusion arrangement this triggers a punishment in which the rival firm increases output, thus driving down prices and so harming the deviating firm: there is a price war, resulting in low profits for both firms. It is the fear of this price war that sustains the tacit collusion arrangement in the long run.

The potential to merge weakens those punishments, because it prematurely terminates them under terms that are an improvement over the price war for the firm that is being punished. Instead of the price war, the deviating firm gets a share in the monopoly that the firms form when they merge. Because the potential for punishment is concomitantly reduced, the trepidation about aggressively producing output in contravention of the interests of the cartel arrangement is reduced: there is more competition, resulting in more output and lower prices. Our conclusion is thus the exact opposite of the conventional view that mergers are harmful for society: making mergers more difficult (i.e., costlier for the firms) is actually harmful to society, because it strengthens the ability of firms to punish each other and enforce the cartel.

In addition to this fundamental result, we also show that pre-merger collusion is dynamically stable: episodes of collusion are long-lasting, and price wars are unusual and brief. Because mergers occur in the face of an incipient price war, mergers are therefore rare–pre-merger collusion is the normal state of the firms.

and

Although the monopoly gains stemming from merging harm consumers in the long run, the enhancement of pre-merger competition benefits consumers in the short-run and those benefits dominate the losses to consumers from the later formation of the post-merger monopoly. This is because discounted expected losses from post-merger increases in market power are small if mergers are rare [empirically they are] and hence the contemporaneously pro-competitive effect of the potential for mergers exceeds those losses if firms spend most of their time in pre-merger competition. This gives further impetus to a regulatory policy that is therefore a bit counterintuitive: reduce barriers and costs of merging in order to harness the pro-competitive effects of mergers.

In short, Hackbarth and Taub show, somewhat counter-intuitively, that regulators can increase consumer welfare by facilitating mergers by lowering frictions, such as barriers, costs, and expenses formally or informally placed by merger regulation such as merger guidelines of the Commerce Commission, the US Department of Justice or the European Commission.

Now just wait for the sound of heads exploding at the Commerce Commission!