How can competition law avoid being anti-competitive?

The rise of large tech companies mean that antitrust law is in vogue again. The US’s Federal Trade Commission disappointed many people by deciding that there was nothing anti-competitive about Amazon’s purchase of Whole Foods, but last month the European Commission ruled that Google’s presentation of Google Shopping search results was a monopolistic practice. It looks like this debate will get bigger and bigger, particularly since after Brexit the UK may need to make new rules about how we regulate large firms to replace Articles 101 and 102 of the Treaty on the Functioning of the European Union, which govern important elements of competition law in EU member states.

Broadly speaking, there are two views of how competition regulators should act when faced with a potentially anti-competitive firm. The first, which is dominant in the United States, sees direct harms to consumers, such as excessively high prices in an uncompetitive market, as being the most and often only reliable measure of whether a firm is acting monopolistically. This position tends to prefer to wait and see whether a big firm which may have market power will use it to harm consumers. 

The second view is that regulators should take action before a firm gets powerful enough to do this, and wants anti-monopolistic interventions in advance of a firm getting big enough to harm consumers. Members of this group would look at a firm like Amazon and, acknowledging that it has not yet hurt consumers by raising prices to extract monopoly rents (profits above normal market rates), worry that it may someday be in a position to do so, its rivals in the market being too weakened to compete back. 

This disagreement has several different elements. How good are courts at judging whether a practice is pro-competitive or anti-competitive? Many behaviours by firms could be either – taking over a supplier could allow a seller to raise prices for consumers monopolistically, or it could allow them to lower prices for consumers by reducing markups across the supply chain. How do you know which a given takeover will do, and is it better to wait and see? 

How good are markets at financing smaller rivals to monopolists? How important is innovation, and do monopoly rents incentivise innovations that can disrupt monopolists? Should a firm’s political power, not just its market power, be something that regulators consider – in other words could a giant Amazon be so influential someday that it is able to bully politicians into protecting it from antitrust lawsuits? (Interestingly, a new paper appears to show that ‘social lobbying’, like wining and dining, but not ‘office lobbying’ of politicians is effective at changing their minds.)

These are all disputed, but probably the biggest divide is whether you view false positives or false negatives as being equally harmful. An influential Frank Easterbook article from 1984 argues that the harms are not symmetrical, and that in fact a mistaken conviction of a firm for monopolistic behaviour that is in fact pro-competitive and efficiency-raising is much worse than a mistaken acquittal of a firm that is acting monopolistically. 

The reason for this is that even very imperfect markets still have some dynamics that work against monopoly firms. As Ben recently argued, when monopolists earn excessive profits (rents), there is an incentive for other firms to think up or apply new innovations that allow them to challenge them and win some of those rents for themselves. Specifically in tech, antitrust actions against IBM, AT&T and Microsoft do not seem to have boosted consumer welfare.

Waiting and seeing might give more information about a practice, or it might encourage investment in innovation by rivals who want to displace a monopolist with their own technological monopoly. Google’s large market share in search is clearly very valuable and has led Microsoft to try to displace it (unsuccessfully) with Bing. If Bing was indeed a superior product to Google, this outcome would have been positive for consumers even if Google had been effectively monopolistic for some time. 

So we have some pressures internal to markets that will push against monopolistic behaviour even if a court has failed to convict on it. Judicial errors of excessive leniency may still be corrected by the market – this is not to say that they always or even often will be, just that this mitigating pressure exists and, clearly, some monopolists do eventually get beaten by rivals.

On the other hand, false positives – judgements against firms that are in fact engaging in pro-competitive, pro-consumer behaviour – have no such self-correcting mechanism. Practices that appear anti-competitive but are in fact pro-consumer, such as aggressive price cuts that are ruled to be ‘predatory’ on other firms, will tend to be abandoned by all firms and the benefits they would have delivered to consumers will be lost. This does not mean that all antitrust convictions are wrong, but it does suggest that the burden of evidence should be more akin to that of a criminal conviction (beyond reasonable doubt) than a civil ruling (“fifty percent plus a feather”).

Even worse, consider what an antitrust conviction actually involves. As Easterbrook pointed out, to determine whether a given business practice like buying up your supplier is pro- or anti-competitive requires both a large amount of knowledge of the businesses and sector you’re looking at and a theory of how the sector would look if this practice was not happening. 

Economists rarely agree except on basic things like rent controls (bad) and free trade (good). In many cases a court will be expected not just to judge what the evidence before it says, but choose from several rival economic theories about how a given market works. Doing so accurately may be difficult. Easterbrook’s whole essay, which proposed five ‘filters’ for eliminating misguided antitrust cases, is still engaging and relevant today, as is Joshua Wright and Geoffrey Manne’s paper relating it to modern tech firms.

These dangers make me think that engaging in pre-emptive antitrust action would be very dangerous. As Rohan points out, Amazon buying Whole Foods is hardly anti-competitive on the face of it as both are tiny players in the US grocery market. We would only be investigating it because of the possibility that it could become big (which it is in other areas of retail), and then also that it would begin acting in a monopolistic, anti-consumer way which it has not yet done in any other area – it is putting a lot of pressure on its suppliers and its rivals, but we should only care if it begins to use its position to hurt its consumers. 

As I argued after the European Commission’s ruling about Google Shopping being anti-competitive, competition between platforms can be a more efficient model than (legally mandated) competition within them. Business models based on freely-provided platforms like Android, and the innovation and investment that goes with them, may be at risk if regulators try to force them to be internally ‘competitive’.

It’s an open question what competition law in the UK will look like after Brexit, if we aren’t just rule-takers from the EU altogether. If we can set our own rules, and want to make Britain friendly to innovative firms, we’ll need to understand what pro-competitive law actually looks like.