Linux smartphones exist? And what does that have to do with antitrust?
You are reading Economic Forces, a free weekly newsletter on economics, especially price theory, without the politics. You can support our newsletter by signing up here:
Who is the main competition for Google Android? You, me, and everyone not under a rock will have the same answer: Apple and its operating system, iOS.
Not so fast, according to the Europeans. Yesterday, the European General Court upheld a €4.34 billion (with a B as in Brian) fine for antitrust violations around Android, including tying the Google Search App with Google Play (the app store). I won’t go into the details of the theory of harm. If you’re interested, my colleague Dirk Auer wrote a paper that lays out all the problems with the lower court’s ruling.
On Market Definitions
I’m more interested in one aspect of the ruling: the market definition.1 According to the European General Court, Android’s main competitor is not Apple’s iOS but Linux.
The reason is the European Commission determined that the relevant market for “licensable” mobile operating systems. Apple’s iOS and BlackBerry’s operating system are closed systems that cannot be put on different phones. You can put Android on Samsung or LG, but iOS cannot. It is only on Apple’s phones.
Why do these arbitrary boundaries matter for the ruling, despite common sense pointing in the other direction?
First, it’s important to recognize that, in general, Google and Samsung writing a contract that says Google does X and Samsung does Y isn’t illegal. Businesses write contracts all the time that restrict the other trading partner. That’s the whole point of a contract! We usually don’t think too much about these restrictions. Yes, when I get married, I can’t marry someone else. That’s a restriction in the contract.
It rises to an antitrust issue if, the theory goes, one of the trading partners is a monopolist and thus can “control” the other side of the market. The idea is that some of these restrictions are unfairly pushed on certain businesses and users because they have no other options. Let’s ignore that even monopolists face competition. It’s fair to say that some markets have more competition than others. Since Samsung had no other options for operating systems, it was “forced” to include Google Search.
That’s why the suing party (like the FTC or DOJ in the US) has to show, not just that the party did something wrong but that they are, in some sense, monopolists. As monopolists, they may have the power to engage in anticompetitive behavior.
A crucial aspect of claiming that a company is a monopoly is defining the “relevant market.” To be a monopolist means to be a monopolist of a certain market. It’s inherent in the idea of monopoly, although, as we will see below, it is not inherent in the related idea of market power. Given that the US and the EU laws deal with monopoly, not market power directly, you could say there is no way around defining markets.
Defining markets isn’t just a weird thing lawyers do. Whenever economists draw an “industry” supply or demand curve, we are implicitly or explicitly defining a relevant market. We talk all the time about a tax in this market or a price control in that market.
At the same time, while helpful for clarifying our ideas, economists recognize that defining markets is arbitrary. As I often say, Josh and I have a monopoly in the market for newsletters titled Economic Forces. But that doesn’t give us any market power, where we could raise price above marginal cost, which is the thing economists often care about as the inefficiency of monopoly.
The arbitrariness of market definitions is why many economists are skeptical of their use. This is not just about silly examples like the monopoly of Economic Forces but matters for actual antitrust decisions.
Take merger cases as another example. As Chris Conlon and Julie Holland Mortimer wrote in their comments for the FTC/DOJ merger guidelines:
As an example, an evaluation of the merger between Whole Foods Market, Inc., and Wild Oats Markets should never hinge on whether Wal-Mart or Costco is “in the market.” Instead, it should depend on whether Whole Foods customers are likely to switch to Wild Oats (or vice versa) if prices were to increase, quality were to be reduced, or stores were to close. When this fraction of customers is large, this should be sufficient to determine that a merger will “substantially lessen competition.” It may be the case that in some cities where Costco is present, less substitution is observed between Whole Foods and Wild Oats, but the extent of substitution between the merging parties should be what determines the outcome of the merger investigation. Put simply, if survey or econometric evidence suggests that for example, 30% of Wild Oats customers are likely to switch to Whole Foods, then the status of Costco, and thus market share analysis, becomes irrelevant.
They capture the standard line in industrial organization. Forget market definitions. It’s arbitrary. Let’s look at things like elasticities and diversion ratios. Two goods are “in the same market” if consumers believe them to be highly substitutable, not whether an outside observer puts them in the same bucket based on some criteria. I’m basically in agreement with the IO mainstream on this. (Hey, I’m not always a crazy outsider!) Let’s try to avoid market definitions whenever possible.
This critique of market definitions is not new. Lous Kaplow is the most prolific writer on the topic. But I think we should be especially skeptical of market definitions in tech. The technical specifications may have little in common with how consumers think about them.
Consumers are the ultimate disciplinary force in any market. If consumers are going to switch from Android to iOS when the quality drops or the price rises, that’s what matters for whether Android has market power. The rest is window-dressing.
But courts and agencies continue to use market definitions, so I think it is helpful to understand why.
The best defense explanation of why antitrust still relies on market definitions comes from a paper by David Glasner and Sean Sullivan. (It turns out, in addition to being a super interesting monetary economist, David Glasner also knows a thing or two about antitrust, which we cover in our podcast a bit 👇).
Glasner and Sullivan point out a few reasons we won’t abandon market definitions anytime soon. As they explain, there is more to antitrust lawsuits than just measures of market power and residual demand curves. Courts care about ease of entry or the possibility of firms coordinating to raise prices.
It is difficult, for example, to imagine courts and practitioners analyzing ease of entry without a market concept. What exactly would firms be entering? Similar difficulties beset efforts to assess the danger of anticompetitive coordination without some idea of which firms would need to cooperate for their coordinated action to be able to raise prices.
Second, they point out that courts aren’t going to abandon market definitions. I hate to break it to my economist friends, but the Supreme Court in the US and the ECJ in the EU ultimately decide antitrust law. And I don’t have a strong feeling the Supreme Court justices are up to date on their econometrics. They’re probably still regressing wages on gender and controlling for occupation. Even worse, they probably think its okay to regress prices on concentration.
Finally, remember that economists use market definitions themselves all the time. Economists, when we need to tackle practical problems, first define what is relevant and what is not. You want to predict the impact of a tax.? Who is the tax on? Who are we considering? Who are we not? That’s defining a market.
A market is a construct in the economist’s head to think through a problem. As Glasner and Sullivan say, “markets do not exist independent of a problem or inquiry but must be defined in terms of a problem or inquiry.” The same holds for agencies considering investigations or courts ruling on a matter. They need a way to tackle a problem.
I’m sympathetic to Glasner and Sullivan’s points. As with anything, there are costs and benefits. Sometimes market definitions help. Sometimes they hurt.
I certainly don’t think market definitions have zero use in antitrust. But the EU’s Android case should make use update our priors about how silly market definitions can become in practice and so we should downgrade them a bit.
Reader Derek Moore first asked for a newsletter about market definitions a few months back. Thanks for the idea!