Big tech mergers occur frequently. The vast majority of such mergers were not reviewed by competition authorities, and those that were have been approved. Nonetheless, competition authorities and governments have become increasingly nervous at the perceived concentration in some digital markets, and at the persistent and increasing market power of some firms operating in digital industries. There is also concern that recent mergers were investigated using an inadequate methodology, possibly leading to wrong decisions. As a result, some of the (many) mergers in digital industries may well have favoured the entrenchment of large firms’ market positions.
This paper, available here, explores this possibility, by developing a model and reviewing the main theories of harm that may apply to such mergers.
Section 2 develops a simple model to address the possible anti- and pro-competitive effects of start up acquisitions by digital incumbents.
This model provides some guidance as to what to expect from such acquisitions and as to the instances in which they may be detrimental to welfare. The model assumes that the acquirer and the target are in the same market. Under this setting, a start-up can develop a project with some probability of success – i.e. the merger will consist in the acquisition of an actual, or at least potential competitor. If the start up remains independent, has enough resources and successfully develops its project, it will share duopoly profits with the incumbent. If the incumbent takes over the start up and successfully develops the project, it will be a monopolist with two products (or services, or applications). If the start up project is shelved, or is not successfully developed, the incumbent will remain a single-product monopolist. It follows that acquisitions of start-ups which may become potential competitors may be pro- or anticompetitive. If the start up is unable to develop its project, then the merger is competitively neutral. If the start up will only invest if the project is acquired, then the merger is pro-competitive. But whenever the start up has the ability to pursue its project independently, the merger will be anti-competitive. If the merger is anticompetitive, there are two possible theories of harm. The first is a “killer acquisition” , i.e. the incumbent kills the product. The second theory of harm focuses on product upgrade with suppressed competition, i.e. absent a takeover, the start up will develop the project and, if successful, compete with the incumbent, whereas following a takeover the incumbent will invest in the project but will suppress competition, lowering consumer surplus.
In short, what matters is whether the start up would be able to invest and develop its project. In practice, and unlike what has happened over the last decade when reliance was placed on simple quantitative tools (such as GUPPI, IPR and the like), it looks like a richer qualitative and, if possible, quantitative analysis is necessary to investigate — through the analysis of internal documents, business plans, and financial analyses — the likely scenarios for the target should the merger not go ahead.
Section 2 then develops this model to look at conglomerate mergers, rather than solely at horizontal ones.
It is often argued that that risks for digital incumbents may come from a firm active in another (complementary or independent) market, which may build on its success on such a market (in terms of customer base, data, technology, etc.) to subsequently enter the incumbent’s core market. For example, Instagram might have become a serious threat to Facebook as a social network.
The authors’ simple model can be easily extended to account for such a possibility, and leads to similar results. In short, an incumbent may want to acquire a start-up whose complement product to the incumbent may become a substitute to it. This reflects the fact that “[c]ompetitive threats will typically come from the fringe (…) Buying up promising start-ups that offer fringe products or services may therefore result in early elimination of potential competitive threats—which may be particularly problematic if done systematically.”
Section 2 also models how incumbents will behave in the presence of outside investors.
The authors extend the model to include additional investors who may be interested in acquiring the start up. Under the assumption that monopoly profits are higher than the sum of duopoly profits, at equilibrium the incumbent will always win the bidding competition for the takeover.
This provides us with another natural benchmark to judge the effect of the acquisition by the incumbent, namely a policy whereby not all acquisitions are prohibited, but just the acquisition by the incumbent (i.e. by the firm whose dominant position is threatened by the project). Such a policy would eliminate any anticompetitive effects. In practice, such a policy would imply identifying the likely trajectory of the possible target company, and prohibiting the acquisition by an incumbent whose core products may possibly be placed along such a trajectory. Although in some cases such an exercise may be difficult (unlike pharmaceutical products, development trajectories may be surprising in digital industries), in others such an identification may be relatively straightforward.
Section 2 also considers whether exclusionary practices provide alternatives to acquiring a potential competitor.
While there are reasons to think that a takeover may be a more convenient way to exclude a potential rival, in particular because it does not entail the dissipation of market profits, there may be reasons for an incumbent to select an exclusionary conduct. In some cases, the start-up might have very optimistic predictions about its future, or may want to have a fair share of the profits that its elimination would save to the incumbent. In other instances, the exclusionary conduct may not be costly in the short-run, e.g. when the incumbent can imitate the start-up’s features.
Further, exclusionary conduct and acquisition may actually be complementary, rather than substitute, strategies. In effect, an exclusionary strategy can allow the incumbent to reduce the acquisition price: by “fighting” the potential entrant, an incumbent reduces the price at which it will take it over.
Section 3 looks at six novel exclusionary theories of harm concerning big tech mergers.
Big tech mergers do not exclusively concern potential competitors. They sometimes involve actual competitors. The purpose of this section is to point to novel theories of harm in the context of big tech that rely on some specific features of digital industries (e.g. network effects; two-sidedness and free services to one side; the prominence of big data).
The six theories of harm are as follow:
- Mergers may be consumer welfare decreasing even if consumers pay a zero price, e.g. a merger between competitors in offerings to single homing customers that reduces the quality of those offerings (e.g. by forcing additional advertising on them or increasing data extraction).
- When a merger enables the firm to combine the installed consumer base through a merger, its motive may be not so much the demand-side efficiency arising from the consolidation of the installed bases, but depriving a competitor from reaching demand-side efficiencies itself and becoming a stronger competitor. While overall consumers may benefit from such a merger under the status quo, they would be even better off if the merger was prohibited leading to an alternative merger. This is particularly relevant if the merger control authority is aware that the relevant counterfactual is not the status quo but a merger between different firms.
- A merger of firms operating two-sided platforms can induce price effects that make it consumer welfare decreasing. If consumers multi-home, there is competition between advertisers in the product market in which advertisers operate as sellers. Since advertisers receive attention if they appear on any of the merging platforms, a merger can help an incumbent seller foreclose a less known seller. For example, a merger between platforms which coordinate their selling of attention is profitable because it reduces product market competition. The ensuing higher seller profits are partly extracted by the intermediary who runs the platforms. The merger is consumer welfare and total welfare decreasing as it preserves the monopoly position of one of the advertisers in the respective product category. This issue arises if the attention market becomes rather concentrated after the merger.
- In the context of big data, a conglomerate merger can be anticompetitive under some conditions. Mergers affect the volume of data available to firms. Data may allow firms to offer better services to consumers, to better price discriminate between consumers, to provide better services to advertisers, and to better extract surplus from advertiser. While it is possible to construct theories of consumer harm based on the combination of data, such theories will only be applicable to very specific cases.
- A conglomerate merger that enables one-stop shopping can be consumer welfare decreasing despite the efficiencies that arise from one-stop shopping. This has been modeled by reference to how a conglomerate merger allows consumers to learn prices and match value about a product in each market with a single search. After a single conglomerate merger there is an equilibrium in which consumers first search out the conglomerate firm (this is driven by the consumption synergy from one-stop shopping). In this equilibrium the conglomerate firm charges lower prices than its single-product competitors, but makes higher profits than the combined profits of its single-product competitors. Compared to the market structure in which all firms remain independent, consumers are worse off with the merger. Therefore, the conglomerate merger inflicts consumer harm.
- A profitable conglomerate merger that allows for bundling of a “free” service may make a more-efficient competitor non-viable and, as a result, be consumer welfare decreasing.
Section IV concludes by making some policy recommendations.
Some big tech mergers may well have adverse competitive effects. Foremost, the risk that the merger removes a potential competitor often deserves careful consideration. This has a number of implications for merger policy.
First, relevant mergers should be subject to scrutiny. The vast majority of acquisitions by large digital platforms have not been investigated simply because they did not meet the turnover thresholds that would trigger notification in most jurisdictions. For this purpose, notification thresholds based on the acquisition price may provide a useful complementary screening device. Another possibility is to use a “share of supply” or ‘market share’ criterion as in the UK, Portugal, and Spain, whereby a merger should be notified if the market share of the combined entity were above a certain threshold. However, none of these criteria allow competition agencies to investigate acquisitions by big tech firms of some young start-ups that have neither substantial turnover, built up a substantial user base nor given other clear indications that they are likely to succeed. Given this, obliging big tech firms with a special status to notify all of their acquisitions deserves careful consideration.
Second, competition agencies should have the means to stop anticompetitive mergers. This ultimately concerns the question of where to place the burden of proof, and what the standard of proof is. As a rule, the authors think that a horizontal merger always causes competitive harms that can be neutralised only when efficiency gains are strong enough. It follows that the burden of proof that such a merger is procompetitive should be placed on the merging firms. The balancing of efficiency gains with competitive harms needs some rethinking too, in particular to balance the (un)likelihood of a target becoming a competitor with the magnitude of the anticompetitive effects that may flow from the merger.
This is a paper that covers many of the concerns created by the shopping spree in which digital incumbents have engaged in recent years. However, to me it feels a bit incomplete – perhaps unsurprisingly, given that we are dealing with a working paper.
As I understood the paper, all sections have in common the assumption that acquisitions by digital titans are susceptible of harming competition. This can be modeled for horizontal mergers, and other authors have developed theories of harm that may be relevant in reviewing such mergers. Since there is a potential for harm, it must be ensured that merger control is effective. So far so good, and each of these arguments is interesting on its own. However, I do not see how each section contributes to the others, which leads to the paper feeling a bit disjointed.
In addition, I take some issues with the model adopted in section 2, which simplicity is almost the opposite of the complexity inherent to the theories of harm discussed in section 3. It is true that the authors openly acknowledge that the model is quite simple. The question then is whether this is a realistic or useful model. As to realism, most mergers by digital titans are not really horizontal as far as I am aware – i.e. they do not purchase competitors – even if they can be said to focus on incumbent’s core areas. This can be thought to be indicative of the existence of competition issues, since one reason why there are no horizontal mergers may well be because the incumbents are so entrenched that they face no realistic competition – which deters market entry, on the one hand, and removes incentives to purchase whatever direct competitors may exist, on the other. However, if this is the case then the usefulness of the model is rather limited, since the problem is not with horizontal mergers and acquisitions at all. More importantly, whether a merger is anticompetitive depends on the counterfactual, and it is quite strange to assume that the counterfactual depends on whether the start up is able to be a viable concern on its own or not – in effect, this feels like a shortcut to the conclusion that all horizontal mergers are inherently anticompetitive.
I also have some difficulty in following the authors’ extensions of the model. For example, of course I would expect the model to apply to conglomerate mergers as well as to horizontal mergers if I am assuming that the target’s product can become a substitute to the monopolists’ core offering. After all, in such a scenario what the authors are actually saying is that a conglomerate transaction is really a horizontal one. However, unless we are able to identify the precise cases where a conglomerate merger eliminates a potential competitor (unlikely, unless agencies are better at this than acquirers), then the model does not tell us much. A similar difficulty besets the extension to alternative target purchasers – again, unless and until we are able to accurately identify ‘potential’ competitors, the model is unlikely to be of much use.
Ultimately, I think this paper is most valuable as an overview of where the debate on acquisitions by digital incumbents currently stands – particularly in terms of applicable theories of harm and potential institutional adjustments. I very much enjoyed the last section of the paper, which provides a succinct discussion of these matters. Nonetheless, I am surprised with the authors’ (and actually, all the literature’s) focus on individual mergers. As seems to follow from the above, and the absence of intervention by the authorities thus far, the likelihood of any particular merger being anticompetitive is low. The concern would seem to be with the systematic elimination of potential competitors or disruptors, and this would seem to call for a systemic reaction – or antitrust enforcement.