OPINION: Who Blinked, Intel Or The US? The Best Guess Is Uncle Sam Made A Hasty Retreat From Its Antitrust Case
THE BridgeNews FORUM: One of a series of viewpoints on government regulation.
March 9, 1999
By James V. DeLong

WASHINGTON-The Federal Trade Commission staff and Intel, the California-based computer chipmaker, reached a preliminary settlement of their antitrust case just 24 hours before trial.

The terms are unknown and will remain secret until reviewed by the commission itself. But given the great interest in the case in high-tech circles, there is speculation on the possible contents of the settlement.

The best guess is that Intel gave little ground, and that the interesting issues were left for another day.

The FTC complaint started from the premise that Intel has a monopoly on general purpose microprocessor chips. It then charged that the company tried to maintain this monopoly with conduct designed to discourage other companies from developing microprocessor technologies.

The specific conduct involved Intel's actions in intellectual-property disputes it had with three of its customers.

IN EACH CASE, Intel pressured the customer to grant the chipmaker a license to microprocessor patents developed by the customer, on pain of being cut off from information about Intel technology necessary to the customer's core business.

The evil of this, according to the FTC, was not the direct effect on the customers, but that Intel's actions lessened the incentive for other companies to develop new microprocessor technologies. The FTC says this served to entrench Intel's monopoly.

It would have been a tough case for the FTC to make.

The first hurdle is the definition of monopoly. The charge that Intel has a lock on the microprocessor market rests on a few facts. Intel has about 80 percent of the worldwide market. Entering that market would require about $250 million in development costs and $1.6 billion for factory construction.

In addition, a new entrant into the microprocesor field would need to cultivate the support of both software developers and customers.

BUT THESE facts are not very impressive. An 80 percent market share sounds good to the newspapers, but absent some serious barrier to entry, it says little about real monopoly power-the power to charge a high price.

A company that has large market share because it charges low prices is not a monopolist, especially if other, technically competent companies are circling it waiting for a sign of weakness.

The FTC argued the expense of entry-meaning the couple of billion dollars needed for development and construction-is prohibitive. But in the high-tech industry that's chicken feed.

Intel's revenues exceeded $25 billion last year, and those of the customers it was supposedly bullying averaged $13 billion. For a shot at Intel's market share, lots of companies could raise the money in an afternoon.

THIS LEAVES the FTC with the "network effect" argument. This theory contends that, once a company gets a head start in a field, it cannot be overtaken because software suppliers and customers will all go with the winner. Thus a competitor has no way to break the circle.

Network effect is a neat theory, but like many theories, it is hard to prove in practice.

The most important putative networks in our history-in telephones, electricity and trash collection-were in fact not the result of true network effects. Rather, these industries were transformed into monopolies by government fiat, usually at the behest of private companies.

The most commonly cited examples of so-called network effect, such as the VHS videotape system, or the alleged superiority of the Qwerty keyboard to the Dvorak, have been refuted by careful researchers.

In addition, network theory ignores the ability of real companies to respond according to their strategic interests.

INTEL WAS accused of bullying Digital Equipment ($14 billion in annual sales), Compaq ($25 billion) and Intergraph ($1 billion).

Such large companies are well aware of the risks of being dependent on one supplier for a crucial component. Individually and collectively, they have the clout to offset any network effects if they think the issue crucial, simply by ensuring that no supplier gets a monopoly.

If all these hurdles were not enough for the FTC to overcome, it also had to demonstrate that Intel's actions vis-a-vis its customers had no legitimate business justification, but were meant only to extend its monopoly.

This, too, would have been difficult, given the real disputes between the parties.

WITH ALL these problems facing the FTC, the best guess is that it got mostly fig leaves out of the case. It is difficult to imagine Intel conceding that it is a monopolist in any way, shape or form, since doing so could cause it big problems down the road.

Intel could have easily agreed, however, not to cut off customers without legitimate reason.

This would have allowed the FTC to leave the field with its colors and arms, postponing the fight over the meaning of monopoly and network effects for another case-perhaps against a weaker opponent.