Friday, May 3, 2019
Comments by Fiona M. Scott Morton
Theodore Nierenberg Professor of Economics
Yale University School of Management
These remarks came to mind as I read Chapter 4 of Jon Baker’s excellent and thought-provoking book. Jon is well known for his work on error costs, the balancing of probabilities and harms from over- and under-enforcement. As we all know, the balance should reflect the probability of a false positive, multiplied by the cost of that false positive to society, set against the same construct for false negatives. For the past 40 years courts have essentially treated this balance as a matter for assumption: false positives are routinely deemed to be more expensive than false negatives.
This posture has led to a steady trend of less enforcement for 40 years. Not surprisingly, waiting that length of time to re-assess the evidence can lead to overshooting of the optimal level of enforcement. Indeed, the amount of competition that has resulted from years of under-enforcement of the antitrust laws is now being studied by academics and governments; the results provide new evidence of both the probabilities and the costs in the error cost framework. We have learned that price increases (quality and innovation reductions) are bigger than we thought, while efficiencies that have been studied turn out to be much smaller than claimed up front. Moreover the large amount of resources spent on experts and litigation has greatly raised false negatives: there are many examples of mergers or unilateral conduct that have been harmful to competition and consumers but not blocked by the agencies. The combination of this evidence leads to the conclusion that tightening enforcement is warranted.
Chapter 4 discusses the structural presumption in a very helpful way. I was inspired by this chapter to go back to first principles to think about mergers. Ordinary models without efficiencies generate the result that fewer firms leads to higher prices (or lower quantities). Even if the market is very unconcentrated, for example, moving from 10 to 9 firms will still cause higher prices when there are no efficiencies. This result has been noted by many people before me and is just a normal feature of the workhorse models of horizontal competition when there are no efficiencies: all firms compete with each other to some degree and therefore elimination of any of them leads to a consumer welfare loss. An interesting enforcement question is as follows, if a 10 to 9 merger raises price by 2% and no efficiencies can be shown, should that merger be allowed? Under current law and jurisprudence, my sense is that it would, on the grounds that surely we can assume some efficiencies because they “ought” to exist.
For example, the HMG assume an efficiency credit of 5% as a default. This is actually quite a strong assumption. Without any evidence on efficiencies and knowing that a major reason to merge is the seeking of profit, there is no reason to assume that efficiencies either a) exist at all, b) are verifiable, and c) are merger-specific. Rather, the consumer and competition is better served when evidence on all these points is introduced in the review of the transaction.
Furthermore, if the assumption that there are efficiencies in the absence of evidence becomes policy, things could go badly for consumers. In a world where buying up a competitor is a choice, a policy that allows price-increasing mergers will cause those mergers to take place. Executives will think to themselves “A 2% increase in price will increase my margin and I will please my shareholders and I will achieve a bonus.” A standard of this type – with no efficiencies required - would cause many firms in the economy to choose to merge in order to reduce competition and raise prices. The result of that choice by many firms across many years (and indeed the US has experienced a lot of consolidation in recent decades) would be significant cumulative real price increases for consumers in the economy.
This discussion points up the importance of finding and proving merger-specific, cognizable efficiencies. Who know about these mergers? The merging firms. Who has the evidence of the efficiencies? The merging firms. Who has the incentive to prove these efficiencies? The merging firms. It is not a good use of resources to set up a process where government lawyers are tasked with determining how the merging firms’ business models match and complement each other, measuring those forces, and proving they are real. A burden of showing efficiencies that falls on the merging parties is likely to cost less and be achieved faster. Some commenters at the conference thought that such a shift would not work because the merging firms would never be able to prove the efficiencies to the standard required -- because proving anything in an antitrust case requires meeting a standard that is very high. But let us remember that this is the standard currently applied to the plaintiff. Is a difference in standards of proof between them justified? An objection that moving the burden of proof will cause the defendant to fail all the time should result in, not a rejection of the idea of moving the burden, but a lowering of the standard of proof for both sides so that the facts of a particular case become relevant and any outcome is possible.
If the merging parties offered information about efficiencies in every merger, authorities could be confident that even in a case of a 10 to 9 merger, the merger benefitted consumers. In a particular 10 to 9 case, for example, economies of scale might be very significant in lowering marginal costs. But if the burden was on the parties to demonstrate that, the agencies would find out. The default would not be that mergers are approved as long as their price increases are small. Rather, the basic model tells us that default is to stay separate until the parties can demonstrate their merger is good for consumers. Firms are always free to grow organically. Indeed, the oft-cited reason that high concentration could be the result of vigorous competition is because of the important impact of organic growth through business stealing that leads to a high market share.