Antitrust doesn't need reforming

Yet again, members of the Open Markets Institute have been promoting themselves as “saviours of the people” in their attempts to reform fundamentally US antitrust law, despite competition law functioning extremely well over the past 20-30 years, requiring perhaps only some minor tinkering. Indeed, in an attempt to defend themselves against the many valid criticisms levied against them, one of their members has written an article in the Journal of European Competition Law & Practice claiming that Open Markets is merely following an approach set out by Louis Brandeis, a former US Supreme Court Justice. Specifically, this “Brandeisian approach” was aimed at breaking up large firms and cartels that had been operating in the US in the early 1900s but had not yet been tackled by the Sherman Act (the US legislation that provided for antitrust enforcement). Regardless of the fact that the issues with which Brandeis was concerned have long since been resolved, the journal article makes some bold claims that are not supported by evidence.

Most strikingly, Open Markets claims that they do not view firms being large as an inherently bad thing – i.e. that big is not always bad.  However, this stands in stark contrast to their continued insistence that Google, Amazon, and Facebook are “dominant” solely because they are large firms, and despite having conducted an assessment of the relevant market, which goes against all standard approaches within antitrust.

Specifically, in order for a firm to be found “dominant”, it is necessary first to define the “relevant market” over which said firm might hold a dominant position – this involves examining the product(s)/service(s) offered by a firm and then determining the extent to which other firms/products/services operate in the same realm as that firm.

For example, suppose that there is concern that a car manufacturer such as BMW might have a dominant position. To determine what products constitute the relevant market for the purpose of assessing whether or not BMW holds a dominant position, one would first consider whether consumers looking at buying a car from BMW would switch to other car manufacturers should BMW impose a 10% increase in price. If consumers would not switch in sufficient numbers, then BMW cars themselves constitute a relevant market as it would be profitable for BMW to conduct the price increase.

In the more likely event that consumers would switch to, say Mercedes, Audi, and Jaguar cars in sufficient numbers that the price increase would not be profitable, one would then group all those cars together under a “hypothetical monopolist” and see whether consumers would switch away from this group of cars were the hypothetical monopolist increase the price of each of the cars in that group by 10%. As before, if the price increase for this group of cars would be profitable, then the relevant market constitutes that group of cars. Otherwise, if the price increase would not be profitable, the group of cars included is widened further. Only once enough cars/manufacturers have been included in this group that a hypothetical monopolist could increase prices profitably has a relevant market been found, and a firm’s market shares be calculated, competitive constraints be examined, and dominance (or lack thereof) be assessed. In this example, stopping too early could give the misleading impression that BMW has a large share of sales, whereas in reality their share of the actual relevant market (probably something along the lines of “all luxury cars”) would be much smaller.

To my knowledge, Open Markets has not conducted such an exercise for the areas in which Facebook, Google, and Amazon operate, and nor do they ever cite such an exercise carried out by someone else (such as a competition authority). Indeed, Open Markets could have cited the the Google Shopping case, in which such an exercise was conducted by the European Commission’s Competition Directorate. However, in conducting this exercise the Commission disregarded evidence that Google Shopping and Amazon compete with each other - in other words, Amazon cannot be dominant as it is constrained by Google Shopping (and vice versa). Hence, the only logical inference as to why Open Markets consider Facebook, Google, Amazon etc to be in a dominant position is that Open Markets considers that “big is bad”.

Moreover, even though the article itself states that “certain industries tend naturally towards monopoly” with the example of “networks” being cited in particular, Open Markets has been rather hypocritical in that regard. For example, even though Open Markets appears to acknowledge that firms / markets with network effects should not be broken up, but instead provided with the appropriate incentives so as not to abuse their dominant position, one member of the Open Markets institute has previously stated that Amazon should be split up.

This recommendation for splitting up Amazon is despite the fact that there are clear network effects associated with Amazon - the more buyers that Amazon attracts to its Marketplace services, the larger the market available to sellers using Marketplace, and so the more attractive Marketplace becomes to sellers. Similarly, the more sellers available on Marketplace means that potential buyers have a wider range of goods from which to choose, making Marketplace an even more attractive shopping option for buyers. This virtuous circle is the very definition of a network effect.

Hence, what possible reason could Open Markets have for this difference in treatment for Amazon (as well as Google and Facebook) as compared to other industries with network effects? Perhaps Open Markets’ own publications and statements belie their claims that they do not view all big firms as inherently evil.

Furthermore, Open Markets seems to want to open up the assessment of antitrust to those that are not specialists (nor indeed, have any experience whatsoever) in competition economics. For example, they suggest that competition assessments should be conducted by authorities as wide-ranging as “the General Services Administration”, “the Office of the United States Trade Representative”, and “the Department of Agriculture”. This can only be a recipe for disaster – competition analysis often requires very detailed and counter-intuitive examinations that a non-specialist is likely to get wrong (or miss entirely). As such, mandating authorities that have no experience or expertise in competition economics to regulate competition can only be detrimental compared to the current situation.

Finally, Open Markets suggests moving away from the “consumer welfare” standard that has served antitrust exceedingly well over the past 50-odd years, but does not suggest what the new focus should be (or how to incorporate any new focus in antitrust assessments). The “consumer welfare” standard means focusing on how consumers are affected by a firm’s behaviour - for example, are consumers being charged too high a price, or receiving too small a range of options. However, Open Markets criticises this approach because they claim it does not take into account the impact that a firm’s behaviour might have on those that provide inputs to that firm (such as workers or suppliers of raw materials), and that a dominant firm can also harm said suppliers.

Nonetheless, this potential impact on suppliers to an allegedly dominant firm can still be incorporated with just a minor adjustment to the current framework.  Specifically, if “producer welfare” is also included in the assessment (i.e. if one also cares about what happens to the provider of a product rather than just those that consume a product), the impact of a dominant firm on its suppliers can be taken into account by considering the firm as a consumer and its suppliers as producers and looking at what happens to combined producer and consumer welfare. It is important to note that this would require including producer welfare in all assessments of antitrust behaviour rather than just those where many suppliers face a single large buyer so as not to apply different standards of proof / evidence to different cases.

However, as noted by Hovenkamp in 2011, including producer welfare in an antitrust assessment can lead to some perverse results - for example, if a monopoly firm were to substantially increase its profits by raising prices to the extent that its increase in profits was far larger than the harm to consumers, incorporating producer welfare in the assessment would mean this excessive pricing behaviour would be viewed as perfectly acceptable. This seems perverse, and presumably Open Markets would be against this alternative (given their apparent dislike of firms making any profits whatsoever). However, Open Markets does not put forward any suggestions for a potential alternative approach.

Overall, therefore, Open Markets’ ideas for how to reform antitrust are both completely misguided and so utterly vague as to render them useless in their entirety.