Mike Munger recently wrote about an elementary misunderstanding of economic competition many legislators and regulators seem to harbor – the idea that improving competition means ensuring there are more firms rather than less. This misunderstanding springs from what Munger calls a confusion between the textbook definition of “perfect competition” (which Munger calls an “idiotic concept”) and real competition, as it occurs in the world outside of the classroom blackboard. Munger cites this example:

Senator Elizabeth Warren recently argued that the Biden administration, through the FTC, should block the acquisition of Discover by Capital One. Her logic has been “perfect competition,” two small firms are better than one medium-sized firm. Yet one need only take a look at the larger industry, where Visa, Mastercard, and American Express control fully 98 percent of credit-clearing transactions, to realize the folly of that approach. If the Capital One-Discover marriage can be consummated, there will be more competition in the industry, not less. The newly formed entity would have the financial power, and the scale of transactions, to force the credit card industry out of its anachronistic ways.

That got me thinking about other cases where “more firms = more competition = better” fails to hold true. Two big examples struck my mind – banking, and mobile operating systems. 

Let’s start with the first. Prior to the Great Depression, Canada had a largely unregulated banking system. From that system, what emerged was a relatively small number of very large banks that operated across the country. In the United States, there was a highly regulated banking system that (among other things) heavily skewed towards a “unit banking” system rather than a branch banking system. That is, banks were geographically limited in how far they could expand (operating across state lines was often a no-go) and were thus limited in size. From this system, what emerged was a system of tens of thousands of fairly small banks across the country. 

From a “more firms = more competition = better” perspective, it might seem like the United States, with its vast number of banks, would be in a better situation than Canada, which was “dominated” by just a few very large banks. But in practice, the opposite was true. Because banks were so numerous and small, it also meant each individual bank was highly undiversified in the assets it held and was all but chained to local economic conditions. Large, highly diversified banks can better absorb economic shocks than small, undiversified banks. This is part of the reason why in the Great Depression the highly regulated United States banking system had over 10,000 bank failures and the lightly regulated Canadian banking system had none at all. 

The second example that came to mind was mobile operating systems. Right now, it’s not at all uncommon to see a certain amount of handwringing over the fact that mobile operating systems is effectively a duopoly between Android and iOS. Wouldn’t it be better if there were more mobile operating systems on the market, because of increased competition? Well, no, not necessarily. If you’re wondering why, then as Munger himself would say, the answer is transaction costs. 

Let’s consider a pretty extreme example. Imagine a genie snapped its fingers and tomorrow there was 10,000 different mobile operating systems on the market. Why wouldn’t that be good for competition? Well, one of the annoying features of the real world that’s left out of the introductory blackboard models is transaction costs. If you want to produce and sell an app that simulates a coin flip to help improve the lives of indecisive people, there’s simply no way you’re going program and format that app for 10,000 different OS’s. The transaction costs are simply too high. The same is true for every programmer and app developer out there. The more operating systems are out there on the market, the more complicated and expensive it is to make your application or program available to everyone on the market. Over the last couple of decades, there have been a wider variety of operating systems. Symbian was one. BlackBerry had their own, called (rather unimaginatively) BlackBerryOS. WebOS was another. Windows Mobile had a good run, too. These have all fallen away, leaving Android and iOS locked against each other. Would it be better if all these defunct mobile operating systems were still out there, providing more and better competition? Maybe so in the imaginary world of perfect competition. But in the real world, a world where transaction costs exist, it’s not at all clear that this would be the case. More operating systems means increasing transaction costs associated with producing everything we use those operating systems for – which could very well make the mobile OS market less rather than more productive. 

What is the “optimal” number of operating systems? I don’t know. Neither do you. The answer can’t be derived from the blackboard, the armchair, or from tea leaves. But the best chance we have to discover the answer is when markets are free enough for actors to engage in real world market competition.