No Exit

So far, I’ve used three technical terms almost interchangeably: network economics, positive feedback, and increasing returns. I’ve also glossed over another term critical to the economics of the information sector: lock-in.

Research on network economics began with the observation that most—but not all—purchases follow a familiar pattern. A consumer selects a product, pays for it, and uses it. If she uses it often and enjoys it, it’s valuable; if not, it’s less valuable. In either case, no one else’s actions much matter to the product’s value. The minute that she bought it, her relationship with the manufacturer, the seller, and the realm of other potential buyers ended. But for an increasing number of techno­logical products, those relationships persist. For such products, the value of ownership grows with the number of people who own compatible products, effectively creating a community of consumers with interlinked value functions.

These interconnected value functions form a communal “network” of owners—each of whom is a “member” of the network. The larger the network, the more valuable the membership. The “returns” on my investment in membership “increase” as the network grows. The value of my chair, for example, is unaffected by the number of other people who own comparable chairs; the value of my phone increases when new people buy phones. Telephones are thus network goods, while chairs are not.

Of course, in order to extract value from my phone, I not only need to know other people who own phones, but other people who own com­patible, interconnected phones. When I choose a telephone—or analo­gously, a design for a rail car, or a computer, or an operating system, or a word processor—I need to know not only how well the specific phone works, but also how well the associated network works. Because the functioning of the network depends in large part upon its size, mem­bership in a large network is more valuable than membership in a smaller network defined by an essentially comparable product. Thus, it’s rational for me to choose the most popular phone—thereby increasing its popularity.

That choice means that I’ve bought more than just a phone. It means that I’ve bought into someone else’s dream (about phones). In fact, I’ve become an investor in their network, to the tune of at least a phone and likely some training. And I’m not done yet; I may need further invest­ment to add equipment and/or amenities to my phone. I’ve also devel­oped a vested interest in enhancing the network’s value; if I can convince my friends and family to join my network, the value of my investment will increase. And so, I’m likely to become an evangelical marketer spreading the gospel of my network—helping to further a dream that someone else may own.

When I evangelize my new network, I provide the network owner with positive feedback. When I convince my friends not only to join my network, but also to convert their own friends to our network, our orgiastic outpouring of zeal creates a virtuous cycle in which growth leads to more growth and value enhances value. How fast will the value increase? According to Metcalfe’s Law, the network’s value is propor­tional to the number of members, squared.¹⁸ While this “law” is far from binding on most real world networks, it does embody the key insight driving the entire analysis of network industries, namely that the value of membership increases rapidly as the network grows.

Imprecision aside, a virtuous cycle, once started, will continue until so many rational consumers join the largest network that it becomes the only reasonable choice. Whereas I may once have debated the relative merits of competing telephone standards, new consumers will be spared the agony of deliberation and choice: anything other than the dominant network will have become either obsolete or relegated to a niche. The market will have “tipped to a (de facto) standard.” This dynamic, whereby increasing returns lead to positive feedback forming a virtuous cycle of growth that eventually tips the market to a standard is what differentiates network industries from their more conventional counterparts.

The subtle differences among these terms—increasing returns, positive feedback, Metcalfe’s Law, virtuous cycles, and tipping to a standard— will help us understand how Microsoft succeeded while the dot-coms failed. Each describes a somewhat different aspect of a network indus­try’s almost organic growth, and they are collectively termed network effects. Furthermore, because each of these effects enhances the value of my network membership based entirely on events beyond my control, some economists prefer to call them network externalities. Finally, network economics is the economic study of industries exhibiting these effects. So, at least nine different terms describe the same basic growth phenomenon!

Lock-in is different. Rather than describing network growth, it enables that growth. Lock-in was always central to network economics. Waldrop talked about it. David talked about it. Arthur talked about it. Even Liebowitz and Margolis talked about it (though critically, and not by name). But they didn’t talk about it much because, in all honesty, it didn’t seem to be that interesting. Networks grow until they tip to a standard. Consumers are then locked in to a single choice. The exciting part is the growth! Organic growth fueled by increasing returns ran counter to much of classical economic theory. Now there’s excitement.

When Congress passed the key antitrust statutes, it basically told the courts to decide which business practices to prohibit. The Supreme Court introduced the distinction between per se and rule of reason offenses, and it laid the groundwork for differentiating legal from illegal uses of selected business practices. But antitrust law is fluid. From time to time, a group of economists and attorneys approach the Supreme Court armed with new theories about markets, about potential market failures, and about the consequences of those potential failures on competition, com­petitors, and consumers. An important wave of antitrust theorists swept into Washington from the University of Chicago along with the Reagan administration in the early 1980s. This Chicago School, inspired by Robert Bork’s The Antitrust Paradox, revolutionized antitrust law.¹⁹ The Chicagoans believed—correctly—that the courts had banned many benign, and even some beneficial, business practices. They explained that markets were more robust than the then-existing antitrust laws implied, and that they were perfectly capable of policing themselves. Chicago School economists felt that market magic would deter or punish anti­competitive behavior in virtually all markets, and that antitrust enforce­ment was rarely warranted. They gained many converts, including the government agencies that enforce the antitrust laws and the courts that interpret them.

A decade or so later, the Chicago School reforms had become the status quo. A new wave of economists and attorneys arrived to challenge their supremacy.

The corrective counterrevolution began with a chance encounter between some Post-Chicago economists and some Independent Service Organizations (ISOs) who repair copying machines. Together they chal­lenged Kodak, who though best known for its cameras, also makes copying machines.²⁰ Now copying machines—it’s worth noting—are not network products, which is why Kodak could remain in the market with a small share and still compete against much larger players like Xerox. But, of course, Kodak’s machines occasionally needed repairs. Con­sumers with broken Kodak copiers could call Kodak, who would send a technician, supply replacement parts, and charge a high price—or they could call an ISO, who would send a technician, buy the necessary replacement parts from Kodak, and charge a lower price. Then Kodak changed the rules. Kodak announced that it was no longer willing to sell spare parts to ISOs. Consumers who needed service would have no alter­native to Kodak’s high prices. They couldn’t turn to ISOs, because the ISOs couldn’t get parts. And they couldn’t abandon their expensive copiers, because the purchase price and their personnel training had locked them in; their switching costs created a sizable barrier to any com­peting copier manufacturer attempting to win away their business.

Sunk switching costs locked in these consumers, Kodak chose to exploit them, the ISOs sued to protect their businesses, and the Post­Chicagoans supported the ISOs with cutting-edge economic theories.

Over the next few years, the courts came to realize that though the Chicagoans had rightly forced them to reconsider the degree to which markets could police themselves, they had dropped the restraints too far. The Post-Chicagoans taught them to scrutinize monopolists’ behavior more carefully, and to appreciate that natural market forces and natural barriers to entry can often insulate monopolists—and even some oli- gopolists—from the sorts of consumer feedback that are supposed to make markets work. Entrenched incumbents learn to exploit consumers and then work to preserve the barriers enabling the exploitation. The Post-Chicagoans insisted that such barrier preservation—often accom­panied by artificial barrier creation—constituted cheating that market forces would not correct in a timely manner, and that only rigorous market cops could prevent. Six years after the Supreme Court ruled against Kodak, the government sued Microsoft, and showed how a monopolist blessed with the natural barriers of network growth and lock-in can leverage them even further to create artificial barriers capable of devastating the information sector.