Faced with the growing power of tech giants, a number of antitrust analysts advocate for a reversal of the burden of proof in merger review: let the acquirer prove in court that the merger is pro-competitive. I argue that such reform would likely have a significant chilling effect on mergers (which is not surprising) and in turn would likely reduce the rate of innovation (which is more controversial) and ultimately would lower consumer welfare. I respond to a series of objections to this view. I also suggest that a combination of ex ante and ex post regulation provides the best approach to rein in the high-tech giants’ excessive market power.
By Luis Cabral1
I. INTRODUCTION AND SUMMARY
There is general consensus among analysts, politicians and the public at large that something needs to be done about the power of, and abuse of power by, high-tech giants such as Google, Amazon, Facebook, and Apple (“GAFA”). The consensus breaks down when it comes down to specific proposals, but very few — if any — would agree that going on with business as usual is an acceptable solution.
A number of antitrust analysts — from both sides of the Atlantic and beyond — advocate for a stronger merger policy in the digital space. The so-called Stigler report (Scott-Morton, 2019), for example, calls for a reversal of the burden of proof in merger review: let the tech giants prove in court that their proposed acquisition is pro-competitive.2 Other industry experts advocate similar proposals.
I would like to present a cautionary note, namely that the kind of merger reform that is being widely proposed would likely have a significant chilling effect on mergers (which is not surprising) and in turn would likely reduce the rate of innovation (which is more controversial) and ultimately would lower consumer welfare.
Before getting to my main argument, however, I want to make two points. First, in a recent anonymous referee report I was told that:
I have seen enough pieces coming from Google & Co calling for a cautionary note. I see little reason for adding to this advocacy. In fact, I would argue that a message like this one can be misleading for readers who know little about competition policy.
I have no fears regarding the latter possibility, as my readers are surely well versed in competition policy, probably more than I am. But I do resent the suggestion — indirect as it may be — that my work is a work of advocacy. To be crystal clear: I have never worked, directly or indirectly, for any of the giant tech firms that I write about in this paper (or for any of their competitors, for that matter).
The second preliminary note is that, while mine is clearly the minority view when it comes to mergers, overall there is more consensus than disagreement when it comes to tech giants and public policy. Like most, I am very concerned about the extent of market power and abuse of dominant positions among GAFA (or GAFAM or any extension thereafter). Like most, I believe U.S. merger policy has largely been characterized by under-enforcement (think healthcare, airlines, etc.). Like most, I agree that many of the mergers we observe are preemptive in nature.
All of the above notwithstanding, I believe the proposed treatment of mergers and acquisitions in the digital space, especially when it comes to GAFA-like tech giants, is unwarranted and would cause more harm than good. To understand this point, it’s important to understand that the “digital” space differs in several important ways from the typical industry that merger policy was designed for. First, in the digital space it is difficult to predict business models and how they evolve. As a result, the vast majority of acquisitions are motivated by option value, not by preemption value. Second, it is difficult to protect and transact intellectual property, and as a result the threat of imitation looms constantly. Third, and related to the previous point, acquisitions play an important role in transferring technology. Fourth, there are frequently significant synergies: the value of innovation is higher with a large firm than with an entrant, for reasons that have little to do with market power. Finally, and perhaps most important, the prospect of acquisition serves as an important incentive for startup innovation. In Cabral (2020) I develop each of these points in detail.3
II. CRITICISM
I am well aware that my view on mergers in the digital space places me clearly in the minority camp. I am also aware that it raises multiple objections. I know that because I have read or heard multiple ones. In what follows, I list those I can recall and address them the best that I can. Specifically, the following are arguments — weak arguments, in my opinion — in support of the idea of reversing the burden of proof in merger approval.
A. Resource Asymmetry
The tech giants have considerably more resources than the regulatory agencies. Therefore, it makes sense for them to carry the weight of an investigation, which you effectively do by reversing the burden of proof. My response is that there are many ways of solving the resource issue other than reversing the burden of proof. It would be considerably more efficient to charge the merging notifiers a fee, possibly a fee proportional to the estimated transaction value. Alternatively, the agency in charge could choose one or two or three experts to work on the case and ask the merging parties to cover the costs. Or there could be a rule such that all experts are paid by the merging parties, but for each expert who testifies “in favor” of the merging parties the agency has the option of appointing a “counter expert” of their choice.
B. Status-quo Bias
The U.S. legal system places a disproportionate weight on defendants, thus creating a high bar for agencies intending to block a merger. I agree that reversing the burden of proof would significantly tilt the system in the direction of making it easier to block a merger. However, if the U.S. legal system is biased in favor of civil case defendants, then we need to reform the U.S. legal system, not the merger approval process. A merger should be approved if, on balance, the positive effects outweigh the negative ones. There should be no room for “beyond reasonable doubt” criteria, which should be reserved for criminal cases.
Although the proposal of placing the burden of proof on the acquirer seems to have been well received in Europe and the UK, it’s also not clear how the above argument (disproportionate weight on defendants) applies. And speaking of multiple jurisdictions, one must not forget that a tech giant merger must be approved simultaneously by multiple agencies (i.e. we’re talking about “global” mergers). Given that, is the current system really tilted in favor of would be acquirers?
C. Killer Acquisitions
Cunningham, Ederer and Ma (2020) provide compelling evidence that many mergers are preemptive in nature, thus anticompetitive.4 I have great admiration for this paper and its authors. Their theoretical model is sensible and the empirical analysis tight. However, I have read more than one academic paper and op ed piece that mentions the high-tech giants and the Cunningham-Ederer-Ma paper in the same paragraph. My point is that this amounts to mixing apples and oranges. The digital giants are simply not like pharma (the industry of focus in the Cunningham-Ederer-Ma paper).
D. Kill Zone
Kamepalli, Krishna, Rajan, and Zingales (2019) provide a theoretical analysis and suggestive empirical evidence for the existence of “kill zones” in the digital space, the idea that acquisitions by big tech have a discouraging effect on innovation.5 I do not have a formal empirical study to counteract their argument, but my experience from observing (and talking to) startups is that the prospect of acquisition is frequently an integral part of their “business model.” In fact, the startup’s main fear is not to be acquired by a tech giant but rather to be imitated by a tech giant, what Cabral (2019) refers to as the shadow-of-Google effect.6
E. Data Availability
The merging parties have considerably more and better data than the agencies. There is no question about the asymmetry in the amount of available data. But there are better ways of addressing this — for example, requiring the merging parties to provide the data requested by the agencies. The argument could be made that the merging parties could then be very selective in the data they disclose, but this problem applies just as well when the merging parties are asked to make their pro-competitive case: they would certainly make a highly selective use of their private information. Which leads me to the next point.
F. Asymmetric Information
The merging parties have better information about the business rationale for the merger than the agencies. This is in general true, but one of the points I try to make is precisely that, when it comes to digital high tech, it’s more a case of uncertainty than of asymmetric information. Moreover, at a theoretical level it is not clear what the effect of asymmetric information is on the optimal sequence of events in merger review. Games with asymmetric information where the informed party moves first frequently result in information unraveling, whereby the information received from the informed party is essentially meaningless.
G. Zero Enforcement
Of the hundreds of acquisitions by GAFAM since 2000, not one has been blocked. Clearly, this is a case of merger review under-enforcement, in fact, zero enforcement. Not quite. First, enforcement refers to the effort by merger authorities (which did review a number of cases), not necessarily the outcome. Second, under or over enforcement is a relative concept, relative to the optimal level. Third, and more important, my note of caution refers to the proposed merger reform, not the current setting; and the proposed reforms would swing the pendulum to a point of almost surely over-enforcement.
H. Scrambled Eggs
The “scrambled eggs” metaphor reminds us that, in the digital world, ex post divestiture is difficult if not impossible: you cannot unscramble the eggs. If merger approval is an irreversible decision, then we should err on the side of caution. This is a valid argument. However, it’s important to understand that divestiture is not the only ex post remedy. In fact, it is frequently the worse possible remedy (because, for example, of network effects). There are many alternatives, including in particular behavioral regulation.
I. Theoretical Analysis Shows that Digital Mergers can be Preemptive
Two recent papers provide different narratives leading to a pre-emptive view of acquisitions in the high-tech space: Fumagalli et al (2020)7 and Katz (2020).8 As often happens in economic analysis, the results depend on the assumptions you make. I believe that, when it comes to the digital space, more often than not my perspective on mergers is more relevant than the alternatives. This is not to say that there are no preemptive mergers or that I propose we abdicate from merger review. I am only providing a word of caution with respect to a wholesale reform as the reversion of the burden of proof would imply.
III. REGULATION
So as not to appear overly negative, and consistently with my initial claim that something needs to be done, I next propose a possible plan of action. Competition policy (and antitrust) is typically thought of as a set of three bins: collusion, mergers, and abuse of dominant position. When it comes to GAFAM, collusion has little bite; and mergers, for the reasons I presented earlier, should not have much bite. We are thus left with addressing situations of abuse of dominant position. I next address this approach.
A reasonable system of digital firm regulation requires that we address at least three points: First, what firms are subject to regulation. A good start would be “whatever size thresholds include GAFAM.” One might add other thresholds so as to include, for example, platforms and gatekeepers that are effectively “essential facilities” even if their size is not even close to GAFAM.
The second point to address is the definition of what behaviors are considered anti-competitive. I think an examination of the EU and U.S. cases involving GAFAM in the past 20 or 30 years provides a good starting point: self-preferencing would likely be on top of the list, but there are certainly other issues.
Finally, the third decision required for a robust regulatory system is the balance between ex ante and ex post regulation. My proposal here would be to have a (possibly evolving) list of presumed anti-competitive behaviors (e.g. self-preferencing). While these would not be considered per se illegal, the anti-competitive presumption would signal that the firms would need to have a clear efficiency and/or pro-competitive justification so as to prevent ex post regulation.
IV. CONCLUSION
I agree with most industrial organization economists that the current state of affairs in the high-tech sector is unacceptable. However, policy experts are barking at the wrong tree when they insist on drastically increasing the bar for high-tech mergers. If technical change in the 1970s and 1980s moved us away from sectoral regulation, the nature of high-tech platforms all but requires that we seriously consider this option as the best means to preserve the benefits from innovation and network effects – and make sure those benefits get to the consumer.
1 Paganelli-Bull Professor of Economics and International Business; Chair, Department of Economics, Leonard N. Stern School of Business, New York University.
2 Fiona Scott Morton (Chair) (2019), Report of the Committee for the Study of Digital Platforms, Stigler Center for the Study of the Economy and the State. See https://research.chicagobooth.edu/stigler/media/news/committee-on-digital-platforms-final-report.
3 Luís Cabral (2020), “Merger Policy in Digital Markets,” forthcoming in Information Economics and Policy.
4 Colleen Cunningham, Florian Ederer, Song Ma (2020), “Killer Acquisitions,” forthcoming in Journal of Political Economy.
5 Sai Krishna Kamepalli, Raghuram G. Rajan, and Luigi Zingales (2020), “Kill Zone,” Becker Friedman Institute for Economics Working Paper No. 2020-19.
6 Luís Cabral (2019), “Standing on the Shoulders of Dwarfs: Dominant Firms and Innovation Incentives,” CEPR DP 13115.
7 Chiara Fumagalli, Massimo Motta, Emanuele Tarantino (2020), “Shelving or Developing? The Acquisition of Potential Competitors under Financial Constraints,” forthcoming in Information Economics and Policy.
8 Michael L. Katz (2020), “Big-Tech Mergers: Schumpeterian Competition, Innovation, and the Treatment of Emerging Competitors,” forthcoming in Information Economics and Policy.
