Yesterday I presented an argument that’s sometimes heard for granting companies that create technological platforms monopoly rights in those platforms. Today I’m going to start to explore what’s wrong with that argument. Today, I’ll discuss network effects, the idea that as new users are added to a network, its value grows faster than the number of users.
My claim is that the more people who use a technological platform, the more valuable that platform will be per user. One classic statement of this idea is Metcalfe’s law, which is summarized well in this article:
Metcalfe was ideally situated to watch and analyze the growth of networks and their profitability. In the 1970s, first in his Harvard Ph.D. thesis and then at the legendary Xerox Palo Alto Research Center, Metcalfe developed the Ethernet protocol, which has come to dominate telecommunications networks. In the 1980s, he went on to found the highly successful networking company 3Com Corp., in Marlborough, Mass. In 1990 he became the publisher of the trade periodical InfoWorld and an influential high-tech columnist. More recently, he has been a venture capitalist.
The foundation of his eponymous law is the observation that in a communications network with n members, each can make (n–1) connections with other participants. If all those connections are equally valuable–and this is the big “if” as far as we are concerned–the total value of the network is proportional to n(n–1), that is, roughly, n 2. So if, for example, a network has 10 members, there are 90 different possible connections that one member can make to another. If the network doubles in size, to 20, the number of connections doesn’t merely double, to 180, it grows to 380–it roughly quadruples, in other words.
The article argues that n2 is probably too high, which I think is probably true. But what nobody disputes is that network effects exist: that a network or platform with 2 million users is going to be more than twice as valuable as an Internet with a million. Or to put it another way, two networks with a million users each will be worth less than a single network with 2 million users. This has the obvious implication that, all else being equal, public policy should encourage the creation of a small number of comprehensive networks as opposed to many small, fragmented networks.
Of course, all else isn’t necessarily equal, and I’ll deal with some of the complications in future posts. But for now I want to offer an analogy that I think helps to flesh out why network effects exist and how they work. I alluded to it a couple of weeks ago: there are close parallels between the argument for interoperability between networks and the classical case for free trade between nations. Just as there are gains to trade when people from different countries can exchange goods and services, so are there gains to interoperability when the users of different networks are able to freely exchange information.
The mercantilists of the 17th and 18th Century tended to believe that the competition for trade was a relatively zero-sum game–for example, that exporting manufactured goods was an economic gain, but importing manufactured goods was an economic loss. If a country’s trade was too open, the mercantilists believed, other nations could impose strategic tariffs that would lead to them having a more favorable balance of trade. Hence, constant vigilance was required to encourage “good” kinds of trade and discourage “bad” ones.
In The Wealth of Nations, Adam Smith demonstrated that this was nonsense. Trade was always mutually beneficial, he wrote, otherwise one party wouldn’t engage in it. And therefore, trade restrictions almost always harmed both countries, by reducing profitable trading opportunities. In a nutshell, what Adam Smith explained is that free trade is perhaps the world’s most important network effect. Two small economies, separated by trade barriers, will be less productive than a single, integrated economy with thriving trade links between them. 200 countries, separated by a bewildering array of trade barriers, is less efficient than a single unconstrained world market would be.
The classical free traders identified two general reasons for these network effects: comparative advantage and specialization. Comparative advantage is the idea that more wealth is created when each nation focuses on producing those goods for which it is most productive, relative to other countries (that is, those goods for which it has a comparative advantage) and trading with other nations for the goods for which other nations have a comparative advantage. In David Ricardo’s famous example, Portugal and England both become wealthier if Portugal focuses on producing wine, while England focuses on producing cloth. It’s inefficient for nations to spend resources producing things at home if it can get them more cheaply by producing something else and using that to trade for the goods it wants.
Specialization increases productivity because many types of productive activities have increasing returns to scale: a large factory can often produce goods at lower unit cost than a small factory. And a craftsman who focuses on one particular task can become more skilled at it than a jack-of-all trades. Free trade promotes specialization by increasing the size of the potential market for goods and services, thereby permitting larger, more specialized firms to turn a profit.
Both of these arguments for free trade apply equally well to interoperability. In fact, gains from specialization due to platform interoperability are arguably a lot more powerful than gains from specialization due to free trade. That’s because many of the goods and services exchanged on technological platform have high fixed cost and very low (or zero) marginal costs. The Google search engine, for example, costs quite a bit of money to develop and fine-tune. But once it’s been built, the marginal cost of providing access to an additional user is very close to zero. Likewise, Amazon.com has managed to succeed due to the incredible gains to scale made possible by having tens of millions of customers. And whoever wrote the Wikipedia entry on network effects probably had to spend some time on it, but now that it’s finished, everyone in the world can read it at close to zero cost–the more readers Wikipedia has, the greater the social surplus generated by each new article.
If instead of having a single global Internet we had half a dozen proprietary Internets that operated using different protocols, it’s likely that many of these gains to specialization would be lost. Amazon would have to maintain six different web sites, each maintained by a different set of programmers. Google wouldn’t work as well, not only because they’d have to build 6 different web sites, but also because they’d have only 1/6 as many links with which to determine the PageRank of a given website. And Wikipedia would be a shadow of its present self, because there would have to be six different Wikipedias, each with 1/6 as many volunteers, and (therefore) 1/6 as much content.
Comparative advantage might seem like it’s not as applicable to platform interoperability, but consider the Firefox web browser. It has emerged as the primary competitor to Microsoft Internet Explorer on the WIndows platform. Its market share recently broke into the double digits, and it’s continuing to grow. Windows-based Firefox users clearly benefit from what they perceive as a better browser, and even non-Firefox users probably will benefit as Microsoft spends more money in the hopes of avoiding further market share erosion.
Firefox has always been a cross-platform product, but a substantial fraction of the programmers who built Firefox are users of open source operating systems like Linux. If each operating system had its own Internet, then Firefox could easily have been built as a Linux-only application, depriving Windows users of valuable. And of course, it works in the other direction too: most of the world’s non-programmers are using Windows, which means that a Linux-only Internet would be missing a lot of useful content produced and published by Windows-oriented users. Having Windows and Linux users on the same network benefits the users of both operating systems.
Samba, which I discussed last month is another example of the same process. It was created by Linux users who wanted their computers to be able to work in a Windows world, but it had reciprocal benefits for Windows users, who have another option when choosing a file server.
In short, the users of different platforms often have complementary skills, and interoperability between platforms ensures that products produced for one platform is available to users on others. The XBox Media Center is a good example of the type of media software that might be more widely available on Windows if Linux programmers weren’t legally prohibited from tinkering with DRMed media formats.
Interoperability restrictions like DRM, then, have much the same impact in technological platforms that trade barriers have in the international trading system. They reduce the efficiency of the system as a whole by preventing people on different platforms from cooperating as fully as they could if their respective platforms were more compatible. If we’re going to use the legal system to close down proprietary platforms, we need to obtain benefits from doing so that exceed the gains from interoperability that are foregone. I’ll argue in subsequent posts that the benefits of legally-enforced platform monopolies are insufficient to offset these opportunity costs.