AMD's War with Intel Becomes a Street Brawl

The war between Intel and AMD this week became almost entirely rhetorical, following the European Commission's action last week, charging Intel with abuse of its dominant power status on that continent. While US antitrust law holds companies to a higher standard of conduct once they have attained monopoly power through non-illegal means, EU law sets the bar somewhat lower, where the test is dominant power.

But just what is dominant power, legally speaking? A Wall Street Journal editorial last Tuesday raised the question. It's easy to call Intel's 80% market share there "dominant;" but the article asked, why should a company expect to compete its little heart out using any means necessary, until it reaches 80% or some such point, after which time it can no longer be allowed to compete the same way?

"This leaves companies in the absurd position of being free to compete as hard as possible until they reach a certain market share - at which point their hitherto legal behavior becomes unlawful."

That got AMD's blood boiling. In a response statement, AMD's Executive Vice President for Legal Affairs Thomas McCoy said, "Here's where the history of American antitrust law comes in. In the 2004 Trinko decision, Justice Scalia made a careful distinction: Mere monopoly status is not illegal. But the use of anticompetitive conduct to gain or to maintain a monopoly is illegal, because such practices block the dynamic potential of competition.

This is the distinction employed by the European authorities in their statement of objections against Intel. They did not base their case merely on the size and success of Intel. Rather, the authorities concluded that Intel waged a sustained campaign to leverage its monopoly status to coerce computer makers into boycotting AMD.

"Thus, as the European Commission explained, Intel's conduct is 'bad for competition and consumers,"' McCoy continued. "And that's exactly the kind of conclusion that justified the century of landmark U.S. antitrust decisions spanning the decades from Standard Oil, through Alcoa and AT&T, to Microsoft."

It's important to note here that the European Commission has not actually reached a conclusion - not in the legal sense - about Intel's conduct. It can only reach a conclusion after it has given Intel a chance to make its case heard at an oral hearing, which Intel has indicated it wishes to do.

Equally important is the fact that the EC, time and again with various antitrust and competitiveness cases, has chosen to distinguish itself from US antitrust law by saying it does not borrow EU statutes from US code. Much of the EC's incentive for pursuing Microsoft over the past few years is believed to have been sparked by that company's settlement with the US Justice Dept., after a district court judgment that would have seen the company broken in two, fell apart.

But AMD's rhetoric didn't end there this week. Yesterday, the company released what it claimed to be an economic study concluding that Intel "extracted monopoly profits" from the sale of microprocessors worldwide, during the ten-year period between 1996 and 2006, estimated at $60 billion.

AMD released a summary of a report from Dr. Michael Williams of ERS Group, which includes the formula for how the consulting economics firm arrived at a $60 million residual figure. Using publicly available data, the firm calculated that Intel reaped $141.8 billion in profits during the 10-year period in question. It then tries to separate the profits attributable to financing from those attributable to sales.

First, the cost of capital is subtracted. This is the rate of return on profits that are attributable to appreciation of capital, and is the minimum acceptable rate that shareholders would require for them to say they're satisfied, or that their investment is breaking even. Again, using publicly available data, the firm determined cost of capital at $54.2 billion.

What remains is economic profit, also known as "economic value added." This is the residual amount of earnings when investments reap more than the cost of capital, a kind of necessary surplus for a successful company. This is what a company earns when it's doing better than "just getting by." By subtraction, Intel's 10-year EVA is calculated at $87.7 billion. The firm points out that's a 16% rate of economic return, in an environment where the standard rate for publicly held companies is often 1%.

Here is where the careful calculation starts to show signs of fuzz. Next, the firm makes an assumption that of that remaining EVA, $27.3 billion - about 19% - is attributable to what it calls "portion of economic profits attributed to assumed advantages," which is also called later "legitimate advantages." It then registers this portion as 5% of its economic return. Based on the numbers given, that's obviously wrong. Nonetheless, the $27.3 billion is subtracted from the EVA, resulting in a $60.4 billion figure attributable to "monopoly profits."

The theory here being, whatever a company earns over and above what investors would expect for it to earn, plus or minus something or rather, must be attributable to that company being a monopoly. The firm then went on to project that if Intel were no longer allowed to earn this over-and-above amount, by 2011, consumers would each pay about $15 less for microprocessors.

This is assuming someone is capable of enacting a law that caps a company's profits at plus-or-minus-something-or-rather over and above its cost of capital - which could be about as unprecedented as the assumption that all windfall profits are monopolistic in nature.

"Intel's monopoly profits of $60 billion directly contradict Intel's claim that its business practices have resulted in lower prices," Thomas McCoy stated anyway yesterday. "In fact this study shows that billions of dollars have moved straight from consumers' pockets to Intel's monopoly coffers."

Whether the aforementioned, supposedly Nobel prize-winning, formula unveiled yesterday has any impact on the EC's decision making has yet to be seen. However, perhaps it can be suggested that all those sources yesterday and today who ran the $60 billion figure without first checking the math, should perhaps be given an AMD-based computer to help them out.