Competition as a socially desirable dilemma theory vs. experimental evidence

Cartels are inherently instable. Each cartelist is best off if it breaks the cartel, while the remaining firms remain loyal. If firms interact only once, if products are homogenous, if firms compete in price, and if marginal cost is constant, theory even predicts that strategic interaction forces firms to set the market clearing price. For society, this would be welcome news. Without antitrust intervention, the market outcome maximises welfare. The argument becomes even stronger if the opposite market side has a chance to defend itself; if imposing harm on the opposite market side is salient; if it is clear that cartels are at variance with normative expectations prevalent in society. There is an equally long list of reasons, though, why such optimism might be unwarranted: capacity is limited; interaction is repeated, and the end is uncertain; firms might be willing to run a limited risk of being exploited by their competitors, hoping that the investment pays. This paper explores the question both theoretically and experimentally. In the interest of capitalising on a rich body of experimental findings, and on the concept of conditional cooperation in particular, the paper offers a formal model that interprets oligopoly as a linear public good.