CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

AMD study concludes what it was paid to conclude

Here's an exercise: Declare an organization is evil. Count the profits. Figure out how much of the profits were the direct result of evildoing. Publish.

It's much easier to analyze data when you've already determined what you want to conclude.

The author of the "economic study" produced by the ERS Group on behalf of AMD's antitrust lawyers, O'Melveny & Myers, said that his analysis of Intel's profits over the last 10 years assumed from the start that Intel was guilty of anticompetitive behavior, rather than reaching that conclusion based on the data. Michael Williams, director of the ERS Group, calculated that Intel has pocketed $60 billion in ill-gained profits after a thorough exercise that seems to have been undertaken merely to lend some validity to a number for AMD's press releases.

Pressure is certainly building on Intel after the European Commission's decision to send the company a "statement of objections," asking it to explain behavior that the EC's Competition division thought was anticompetitive. But that's hardly a guilty verdict; Intel has 10 weeks to respond.

Rather than wait to see the final outcome, however, AMD seized on the distraction from its recent financial performance and began pouring on the rhetoric. AMD's study (don't pretend ERS stumbled upon this data and brought it to CEO Hector Ruiz's attention) concludes that Intel "extracted" $60 billion in "monopoly profits" during the 10-year period from 1996 to 2006.

Williams said the analysis was based on the work of Merton Miller and Franco Modigliani, who won a Nobel Prize for their method of determining economic value. It involves figuring out what Intel's total profits were for the 10-year period ($141.8 billion), subtracting Intel's cost of capital ($54.2 billion, or what ERS estimates is Intel's cost of securing investment in its microprocessor business) and determining that Intel earned $87.7 billion in "economic profits," a rate of economic return of 16 percent.

ERS thinks that's unusually high. Williams said the average economic rate of return for the semiconductor industry over that period of time was negative 7.7 percent. Rather than look for other reasons why Intel's number might be high, however, Williams immediately assumes that based on the goings-on in Europe and Japan, those profits must have been gained from monopoly behavior.

"Only four companies earned economic returns of 16 percent or more--Microsoft (38.25 percent), UST Inc. (28.54 percent), Coca-Cola Co. (16.58 percent), and Intel (16.01 percent)--and each of these companies has been associated with antitrust determinations," we learn. But remember, "of course, high economic returns by themselves do not demonstrate anticompetitive conduct."

So, ERS "generously" assumes that some of Intel's profits can be attributed to its strengths as a company, whether that's a recognizable brand, deep patent portfolio, or the tendency of its only major competitor to screw up every five years. It decides that five percentage points, or $27.3 billion, is in the clear, arriving at a nice round $60 billion profit figure as the tainted spoils of Intel's predatory conduct.

So basically, the study was an exercise in assigning a number to throw around in AMD's press releases. After all, if you assume that the vast majority of someone's profits stem from their monopoly position, does it really matter whether the final number is $20 billion, $40 billion, or $60 billion?

"By any standard, Intel has earned extraordinary profits," Williams said. "I'm assuming some of those profits are derived from anticompetitive conduct. There's some chance I'm wrong."

Look, Intel dominates the x86 processor market. And it certainly does have to worry about the EC proceedings and the U.S. antitrust trial. But if anybody at AMD actually reads this thing, here's some free PR advice: the horse is dead.

If Intel is in fact guilty of using a dominant position to hurt AMD's business prospects, either AMD or the government needs to prove that in court. No major PC executive or ex-executive has stepped forward to corroborate AMD's claims (other than Rahul Sood, who didn't exactly scream "J'Accuse!" either) turning the whole exercise into a battle of relentless press releases and ridiculous studies that do more to harm goodwill than to build it.

If you have real evidence of Intel's misconduct, present it, or have one of your customers give me a call. Don't insult the intelligence of the public--which would ordinarily like to support an underdog--with studies such as this one.