Reputation and Intermediaries in Electronic Commerce

• Author: Clayton P. Gillette
• Position: Max E. Greenberg Professor of Contract Law, New York University School of Law

Max E. Greenberg Professor of Contract Law, New York University School of Law. I owe thanks to Alan Schwartz, Robert Scott, Peter Swire, and the participants in the Symposium on Unifying Commercial Law.

I Introduction: Substitutes For Contract Enforcement

It is a privilege to participate in a symposium dedicated to Chancellor William Hawkland, in large part because of the depth and breadth of his understanding of the role of commercial law in creating robust markets for trade. Obviously, the most important manifestation of that understanding is reflected in his work on the Uniform Commercial Code ("U.C.C."). I hope that I will not be considered to have deviated too far from the focal point of his contributions by making more explicit a point that I believe implicitly underlies the development of the U.C.C. and the current critiques of its evolution. The point is that commercial law is necessary to overcome problems of distrust that would otherwise frustrate exchange. Successful systems of commercial transactions are typically assumed to require a background set of legal rules that clearly assign entitlements to traders, and a legal system that enforces those assignments.¹The former reduces transactions costs that may derail otherwise efficient exchanges. The latter reassures parties to non-simultaneous exchanges that they will be able to obtain redress in the event of postcontractual breakdowns. By supplying both law and a means of enforcing it, the state encourages market activity without more invasive governmental intervention such as setting prices or allocating rights to the goods themselves. It is on this basis that some find correlations between national economic growth and the development of robust systems of contract law.²

While this story of exchange explains much of the commercial market, it fails to account for transactions in which contractual parties would predict that state-sponsored enforcement will be underutilized in the event of contractual breakdowns. Even those actors who have available both clear legal entitlements and state enforcement mechanisms will forgo legal recourse when enforcement costs exceed expected recovery.³ This calculus is likely to occur in a variety of circumstances. The amount at stake may be very small, so that even moderate litigation costs deter enforcement; the aggrieved party may have to litigate in a distant jurisdiction, so that the cost of recovery is high; the forum's judicial system may be considered untrustworthy, so that the probability of recovery is low; or resolution of the underlying dispute may depend on conditions that are difficult to verify, so that recovery costs are high and expected recovery is low.

In each of these cases, law will be of little importance to the parties. In response, potential traders may eschew otherwise beneficial contracts; alter the prices at which they sell or buy to obtain ex ante compensation for the risk of unredressable breach; seek lower-cost sources of redress (arbitration or dispute resolution within industry tribunals⁴); or seek lower-cost alternatives to redress. This last device includes monitoring the quality of the goods under the contract, either directly by the buyer or through intermediaries.⁵ Ex ante monitoring, however, is unlikely to be cost-effective for low-value transactions or for those involving distant buyers and sellers. Even if it is possible to find a third-party monitor at a distant location, the lower likelihood that a distant trading partner will be a repeat player indicates that agency costs between the monitor and its principal are likely to be high. Even private enforcement mechanisms such as gossip are unlikely to be suitable for geographically distant transactions if the parties are not members of the same enforcement regime, such as a local trade association.

These issues constrain the development of long-distance trade, especially retail trade, that was initially promised when Internet commerce was initially introduced.⁶ Notwithstanding expectations that the Internet would vaporize boundaries, the need for buyers and sellers to trust each other where compliance with contractual terms cannot readily be verified ex ante or enforced ex post diminishes the likelihood of trade. Thus, even while electronic commerce reduces search costs and transactions costs of putting together willing buyers and sellers, its full potential cannot be realized if parties fear that trading partners will perform opportunistically because remedies for breach or chiseling are unavailable.

One plausible solution is to reduce the risk that one is dealing with a potential breacher by monitoring the reputation of one's trading partner. On the assumption that reputation from prior experience is a valid predictor of future performance, reputation can be particularly useful in the very transactions- those that involve low values and those that involve parties who are geographically distant-in which legal rights to enforce the contract are less likely to be asserted and in which monitoring performance is infeasible.⁷ These are also the transactions that are most susceptible to expansion through electronic commerce because the number of buyers and sellers is likely to expand dramatically as search costs and transactions costs are reduced in the ways made available by the Internet. Moreover, the ease of communications with geographically disparate buyers and sellers makes it easier to register satisfaction or dissatisfaction about one trader with a larger number of potential traders, thus enhancing the likelihood of discovering both positive and negative reputational information.

At the same time, the vast number of traders on the Internet raises the probability that any potential trader searching for a partner will be confronted with an amount and quality of information that frustrates efforts to discern a meaningful reputation. If reputation is to serve as a sorting mechanism that allows potential traders to assess each other's credibility, the information on which reputation is based must itself be trustworthy, and the ease with which Internet users can post information about a trader may reduce reliability.⁸"Cheap talk" raises the specter that information that is provided will be of little practical use unless potential users can distinguish credible information.⁹ The costs of overcoming this problem of informational overload may reduce the feasibility of small value transactions among distant potential traders and thus reduce the promise of electronic commerce.

The legal literature has long understood that reputation can assure quality of contractual performance and substitute for legal enforcement.¹⁰ But the traditional literature implies that the situations in which reputation will matter have significant limits. Most of the literature focuses on situations in which parties have repeated interactions with each other and thus invest in reputations to induce cooperation from their immediate trading partners. The salient features of relational contract do largely depend on vulnerability to retaliatory threats in a subsequent interaction to restrain each party from acting opportunistically in a current interaction.¹¹ Thus, reputation has been assumed to have its most significant effects where parties are engaged in repeat play, rather than where one trader's reputation with a second trader affects subsequent dealings with third parties. Reputation is particularly effective in relational situations because long-term contracts tend to be incompletely contingent; as a consequence, the specific obligations of the parties, and hence the existence of breach, are highly uncertain. Ex post enforcement costs will therefore be high, and ex ante constraints such as reputation can therefore compensate for the risk of underenforcement.

In theory, reputation can have similar effects in transactions among strangers, if there is a sufficiently low-cost mechanism for communicating reliable reputational information to the affected parties. In this paper, however, I suggest that the potential utility of reputational mechanisms in electronic commerce is likely to be hampered by the limited capacity of users to confirm the data on which a reputation is formed. My concern is consistent with, but separate from, the concern that the voluntary nature of reputational systems discourages participation. That critique suggests that participation will be too low to create a representative set of reputational data. As a result, the information that exists will be unreliable. In this paper, I suggest that even where participation rates are high, the quality of information may be of limited utility to potential users of the system.

II Reputational Intermediaries

Even in transactions among strangers, reputational information theoretically can indicate the likelihood of satisfactory performance. In order to achieve that result, however, traders must have both the opportunity and the incentive to communicate information concerning their transactions to potential subsequent traders, and the latter must have reason to trust the information they receive.¹² In order for reputation to be reliable, it must be capable of assimilating both positive and negative information. While traders may self-report about positive experiences others have had with them, negative information will likely depend on one party's reporting the defalcations of another in a manner that is discoverable by subsequent traders. In sum, a trader's reputation can reduce distrust about dealing with a stranger if (a) credible positive and negative information about the trader is available;¹³ (b) there exists a...