I think EU’s anti-trust ruling against Intel is wrong. A few years ago, AMD have a really good chance fighting Intel but it flopped and missed a technology cycle. It’s CPU is slower and less powerful than Intel’s. AMD can’t cry unfair and ask government step in to punish Intel because their own incompetence.
Aug 20th 2009 – The Economist
Discounting that promotes competition is hard to distinguish from predatory pricing
TWO decades before he won the Nobel prize for economics in 1991, Ronald Coase wrote an essay decrying the poor state of research in industrial organisation, the discipline in which he established his reputation. The field, he complained, was devoted to the study of monopoly and antitrust policy. That, he said, made for bad scholarship: an economist faced with a business practice that he cannot fathom, according to Mr Coase, “looks for a monopoly explanation”.
A lot has changed in the 37 years since that lament. The broader research effort for which Mr Coase called has fostered a richer understanding of how firms respond to customers and rivals. Monopoly explanations now compete with theories that see the same behaviour as helpful to consumers. That has made it harder to sort malign from benign business practices. The recent antitrust finding against Intel, a maker of computer chips, is a case in point. After a long investigation, ending in a bulky 524-page verdict, the European Union in May fined Intel €1.06 billion ($1.44 billion) for illegally using its muscle to price AMD, a rival chipmaker, out of the market. Intel rejects the charge of predatory pricing and plans a court appeal. Its lawyers have a block of theory on which to build a defence.
Allegations of predatory pricing have a long history. The Sherman Antitrust Act of 1890, the foundation of America’s competition policy, was partly a response to complaints by small firms that larger rivals wanted to drive them out of business. Trustbusters need to be wary of such claims. Low prices are one of the fruits of competition: penalising business giants for price cuts would be perverse. But in rare circumstances, a big firm with cash in reserve may cut prices below costs in order to starve smaller rivals of revenue. The profits sacrificed in the short term can be recouped by higher prices once competitors are out of the way.
Establishing that a firm is guilty of predation is difficult. If rivals stumble or fail, that may be down to their own inefficiency or poor products, and not because they were preyed upon. Proving that a firm is pricing below its costs is tricky in practice. Even where a reliable price-cost or profit-sacrifice test is feasible, failing it need not imply sinister intent. There are often pro-competitive reasons to forgo short-term profits. Firms with a new product, or a new version of an existing one, may wish to pick a lossmaking price to defray the cost to consumers of switching, or because they expect their own costs to fall as they perfect the production process (video-game consoles are a classic example). Losses would then be a licit investment in future profits.
Predation is even trickier to uncover when goods are sold together. A firm that enjoys fat profits on one good may “bundle” it with another on which margins are lower. If the discount on the bundle is hefty enough, other firms may struggle to offer as enticing a deal. In 2001 the EU blocked a proposed tie-up between GE and Honeywell for fear that the merged firm might use bundled discounts to squeeze rival suppliers. In 2007 a committee of antitrust experts appointed by the American government proposed a test for whether bundling is predatory. First, assume the discount applies solely to the low-margin good. So if each good sells for $10 separately and $16 as a bundle, allocate the $4 discount to the more “competitive” product. Next, apply a price-cost test: if the product costs over $6 to make, the bundle is predatory.
That check seems neat but sound business practices may still fall foul of it. It may be cheaper for a firm to sell the two goods together, because of cost savings on distribution. Firms also often use bundling as a way of charging high-demand users more. Thin margins on sales of printers, for example, can be made up by bundling in more profitable toners. This kind of “metering” is an efficient way of recovering fixed costs such as research.
Another ambiguous tactic is to offer rebates to customers that reach certain sales targets. Bulk buyers generally pay lower unit prices to reflect suppliers’ economies of scale. Rebates can also help align incentives. Suppliers want retailers to promote their products, offer in-store information and keep plentiful stocks. The trouble is, retailers bear all the costs of such sales efforts but reap only some of the benefits. Rebates provide incentives for retailers to drive sales, as profits are bigger once the target is met.
The price of loyalty
The EU reckons that Intel’s use of such rebates was nefarious. It is in the nature of rebates that, just above the target threshold, the price of each additional purchase can be negative. If, say, a firm charges $1 for each sale of up to nine units, and a unit price of 80 cents (a rebate of 20%) for sales of ten items or more, the price of the tenth sale is minus $1, since nine units cost $9 and ten units costs only $8. A dominant firm like Intel can rely on a certain market share (an “assured base”, in the jargon). It could in theory set a rebate threshold above that mark, where smaller rivals may hope to mount an effective challenge but cannot match the negative marginal prices on offer to buyers.
Intel’s conduct was certainly worth investigating. Its rebates kicked in if customers gave the firm between 80% and 100% of their business. The schemes looked like a response to a competitive threat from AMD. Yet the EU’s trustbusters cannot feel too sure of themselves. Intel’s rival is still alive and kicking: AMD has not been excluded from the market (though its investment plans may have been thwarted and potential entrants deterred). Moreover, such a complex case, with a bulky ruling which is still not in the public domain, does not offer much guidance on what sort of rebate schemes might be deemed predatory.
Trustbusters have moved away from the practice that so concerned Mr Coase in the early 1970s—of being too quick to condemn big firms on the basis of crude judgments. But they are unlikely to find a robust and simple rule to put in place of the old presumption that firms with market power are always suspect.