I. INTRODUCTION Klein and Leffler [1981] and Shapiro [1982; 1983] have shown that a firm that engages in repeat transactions has an incentive to use the market, as distinct from formal legal processes, to assure the quality of its product. The firm charges a price above salvageable costs of production and invests the premium in reputation (trademark) capital that it forfeits, wholly or in part, if it allows quality to fall below the standards it represents to its customers.(1) The quality-assurance hypothesis seems to be supported by stock market event studies, such as those by Jarrell and Peltzman [1985] and Karpoff and Lott [1993], which show that product recalls, airline crashes, deceptive advertising, and fraud result in trademark losses. The quality-assurance hypothesis, however, does not explain the massive size of these losses, which typically are many times the outlays incurred by the firm as a result of the event. The law and economics literature on contracts offers an answer to this and related puzzles, suggesting that trademark capital does more than assure quality: It assures specific performance, that is, fulfillment of the precise terms of the contract.(2) After developing the specific-performance hypothesis (section II), we show that it is supported by the evidence from event studies. Furthermore, it explains other phenomena not covered by the quality-assurance hypothesis, including why competitors of firms whose products are recalled lose wealth, why some damages, such as those in airline crashes, intentionally fail to provide full compensation, and why firms who sell to other firms also invest in trademark capital (section III). We conclude with a summary and a few remarks on the issue of market versus legal forces (section IV). II. THE SPECIFIC-PERFORMANCE HYPOTHESIS The quality-assurance hypothesis focuses on the problems that arise when the physical characteristics of a commodity are costly to determine, even after purchase, and sellers' warranties and quality representations are costly to enforce. Because higher quality usually costs more to produce and can be sold at a higher price, producers have an incentive to represent their goods as being of higher quality than they really are or, having established a reputation, to cut quality and costs. When consumers observe that the quality of the goods they purchase is lower than they had been led to believe, they shift to lower-cost, lower-quality goods whose quality is easier to determine. Higher-quality goods seemingly are driven off the market. Klein and Leffler [1981] hypothesize that a firm that engages in repeat transactions can gain by producing higher-quality goods, charging a price above salvageable costs of production, and investing the premium in firm-specific (trademark) capital. The value of this trademark, also described as reputation or goodwill, is the present value of the expected stream of price premiums. Misrepresentation causes a reduction in this income stream and a corresponding loss in trademark capital.(3) Thus, trademarks assure quality without a third party to enforce the contract.(4) A buyer, however, is concerned about other characteristics of a transaction besides the physical attributes of the product. The expected harm if the seller breaches the contract (for example, by misrepresenting the product, failing to deliver it on time, or failing to provide future service) and the ability of the legal system to assure full compensation also matter. Accordingly, a seller who engages in repeat transactions has an incentive to provide a bond that assures specific performance, rather than just product quality, when a breach results in damages that are hard to measure and the legal system does not enforce contracts costlessly. The next task, therefore, is to examine the limitations of the legal system in enforcing contracts and providing compensation in case of breach. Specific Performance and Penalty Clauses The courts could require contracting parties to provide specific performance. …