Trade and Institutions in the Medieval Mediterranean examines the economic geography of long-distance trade in the eleventh-century Muslim Mediterranean. It relies on commercial letters found in the (Cairo) Geniza and contains two parts. One part presents and analyzes the geographical patterns of trade-related variables, while the other elaborates on the determinants of these patterns focusing, in particular, on the nature and roles of institutions.
This is by far the most comprehensive study of the geography of long-distance trade centered in Egypt on the eve of the European commercial expansion. Its use of content analysis constitutes a novel methodological innovation and the empirical payoff is impressive. Goldberg establishes, for example, that long-distance trade was interregional, but the long-distance traders were heavily involved in the (intra-regional) pre-trade production processes of agro-export commodities such as flax and oil. Examining these and other variables highlights, among other issues, that political conflicts and personal insecurity had a large impact on the geography of trade. In particular, these factors induced, during the second half of the century, a shift in trade from the West to the Syrian coast.
With respect to institutions, the main claim is that the “legal system [was] efficient as a venue for enforcement and redress” in agency relations (p. 162). The Muslim court effectively prevented opportunism while the Jewish court mediated mercantile conflicts. Thus, “the only necessary condition for forming business relationships was shared participation in the Islamic legal system…[which] allowed individuals in different Islamic polities, whatever their confession, to enter into business relationships with one another” (p. 355). The boundaries of the Islamic legal system thus determined “the geographic reach of merchants” (p. 22).
The book contrasts this legal-centralism claim with the reputation view I have advanced including in an article published in this Journal (1989). Although the traders used the legal system for various purposes, a multilateral reputation mechanism was the main institution underpinning trust in agency relations. The traders shared the expectations that information about opportunism in bilateral relations would circulate, that unaffected traders would respond, and their negative responses would impose high cost on the opportunistic agent.
The rest of this review is devoted to a discussion on why Goldberg's conclusions differ from my own. Specifically, it presents why the disparity in our conclusions is much bigger than the disparity in our views of the evidence. This short discussion does not do justice to other issues covered in the book, or to the subtleties of, and distinctions between, the two views. It only demonstrates that the conclusion of legal-centralism does not follow even if one accepts the evidence in the book. To identify which institution predominated requires a methodology the book does not employ.
The main evidence of reliance on legal enforcement is “a wealth of evidence that the Geniza merchants used the legal system to circumscribe business relationships and resolve disputes” (p. 150). With respect to the crucial question of enforcement in agency relations, however, the book concurs my observation that the evidence is weak. “The evidence of content analysis presented in chapter 3 suggests that the incidence of lawsuits was indeed low and that the Geniza merchants avoided court when they could” (p. 155). It should be noted that Arabic written legal documents found in the Geniza reveal extensive use of the Muslim court by Jews in non-trade related legal issues. The lower level of agency and trade-related dispute is thus puzzling if, as the book claims, the traders extensively relied on the court.
The lower level of agency and trade-related dispute might be, as I argued, due to the limited capacity of the legal system. The book only asserts that the court had the information required to prevent opportunism in agency relations. The state “sustained systems of public information that made access to Islamic legal protection by qadis of the various cities a possible and reasonable enforcement threat within long-distance trade” (p. 178).
At the same time, the book also recognizes that merchants regularly concealed the goods they sent to their agents from the authorities. “Their [the Jewish traders’] carefully packed goods, in particular the precious bundles of mixed goods mentioned in chapter 4, typically passed from port to port undisturbed. This allowed Geniza merchants both to hide goods from authorities who might wish to sequester them and often to avoid most customs charges on their most valuable cargo” (pp. 172–73). If merchants hid what they sent to agents from the Muslim authorities, it is unlikely that the merchants could then ask the Muslim court to enforce agency disputes.
Moreover, the book explains that traders hid their goods because the state was predatory. “Geniza merchants appear to have been especially careful to keep their capital as mobile and as invisible from either popular or government notice as was possible for such wealthy men” (p. 179). The book thus qualifies its claim of reliance on the law by noting that the traders faced a trade-off between the risk of predation by the authorities and the benefit of legal enforcement (pp. 180–84).
Goldberg describes a multilateral reputation mechanism that provided an institutional alternative with low predation risk (section 5.2). Although her mechanism and mine differ in details, they rest, to the best of my understanding, on the same principle that the expected response of other traders rendered collective punishment credible. In any case, Goldberg maintains that this reputation mechanism provided the merchants “as a group a powerful enforcement tool to manage commercial services” (p. 147).
Goldberg asserts, however, that multilateral reputation only prevented reneging on a trader's obligation to reciprocate for past agency services. Reciprocal agency was the main way to get agency services but there was no legal remedy if one reneged on the obligation to reciprocate. To support her view that multilateral reputation governed only labor reciprocity Goldberg argues that the legal system could govern all other obligations. “The mercantile letters suggest that Geniza merchants, including Khalluf and Yeshu'a, used a reputation mechanism among themselves to handle these problems of labor reciprocity, while relying on the state and legal system to protect their property rights” (p. 145).
Yet, if legal enforcement was costly, risky, and often avoided, as the book concludes, the claim that it was generally preferable to reputation mechanism is exactly what the book has to substantiate and not to assert. Such substantiation, I argued, has to recognize that expectations and beliefs are inherently unobservable and multiple institutions can operate at the same time and govern the same or different transactions. Accordingly, it is useful to augment the historical method with the social scientific method in which a conjecture's predictive power is a measure of its plausibility and selection among conjectures relies on Occam's Razor.
The book takes the predictions of the multilateral reputation for granted and does not consider them evidence thus violating Occam's Razor. To illustrate, the traders’ group was horizontal in the sense that each trader operated through agents while also providing agency relations. The multilateral reputation conjecture predicts this pattern while the book takes it parametrically. At times, the book confuses predictions with necessary conditions and thereby rejects the reputation view based on evidence that confirms it. To illustrate, the reputation conjecture predicts that agency relations would be mainly among group members. The book considers any out-group relations as refuting it.
Similarly, the social scientific method recognizes that comparison across historical episodes is meaningful if the differences do not directly influence the comparison. For example, I noted that the lack of reciprocal agency among the Genoese traders of the mid-twelfth century suggests a meaningful institutional distinction. The book dismisses comparisons on the ground that the Italian traders had political power. It does not clarify why their power hindered their ability to reciprocate.
In short, a main cause for the disparity between the legal centralism in this book and the reputation view is distinct attitudes toward the social scientific method. The distinct methodological attitudes, in turn, seem to reflect distinct research objectives. The book is rich in details, celebrates complexity, and emphasizes the limit of our knowledge. The multilateral reputation studies sought simplicity and the best-fit comprehensive explanation. The two objectives, however, are complementary. Generalizations are predicated on the details to which they give meaning and context.
The strength of Trade and Institutions in the Medieval Mediterranean is in its novel use of content analysis, its focus on economic geography, and its grand synthesis of, and contributions to, trade history. The book is a major contribution to economic history, the history of the Middle East, and Jewish history.
