HLS Private Law Workshop; Lisa Bernstein, Revisiting the Maghribi Traders (Again)

Post by Patrick Goold

Avner Greif’s study of Maghribi Jewish traders in the eleventh century is a seminal work in the literature on private ordering. In a couple of highly influential articles (here and here), Greif documented how this group of merchants created elaborate trading networks across the Islamic Mediterranean. Greif argued that the Maghribi formed a close-knit “coalition” that could eschew enforcement of agreements by lawsuits and the threat of liability and instead rely on reputation-based community enforcement. However, in recent years, historians of the period, including  Jessica Goldberg, as well as Jeremy Edwards and Sheilagh Ogilvie, have questioned Greif’s thesis, suggesting that there is little evidence of Maghribi traders boycotting members for misconduct, and some evidence that they relied on courts.

At this week’s HLS Private Law Workshop, Lisa Bernstein presented a draft essay that revisits this topic (Revisiting the Maghribi Traders (Again): A Social Network and Relational Contracting Perspective). Bernstein proposes to revise Greif’s analysis by swapping out the notion of “coalition” on which he relied for social network analysis and relational contract theory. While this is an alternative account of the Maghribi activities, it nonetheless supports Greif’s central thesis.

Although Maghribi merchants sometimes formed partnerships with one another, they more commonly used each other as reciprocal agents under a legally unenforceable agreement known as a Suhba. Under a Suhba, a merchant who asked his agent to perform a task (for example, travel to a foreign city to sell the merchant’s flax) would become obligated to perform a task of equal value (for example, introduce the agent to other important merchants). This system enabled the traders to diversify their trading portfolios and reach many markets across the Mediterranean without needing to travel with their goods. The center of this trading activity was Fustat, today part of Old Cairo, and it is a cache of documents in the Cairo Geniza that serve as the main historical record of the merchants’ activities.

Bernstein argues that the Maghribi traders were organized as a “semi-closed bridge and cluster network with small-world properties.” Within trading centers (cities like Fustat), most trade was conducted in the open with witnesses. Meanwhile, business and social interactions resulted in a dense network of ties that enabled reputation information to spread easily. These “network clusters” were then “bridged” by a number of social institutions and organizations. Postal routes between trading centers enabled information about reputation to flow between clusters. A handful of dominant traders also had personal and family ties spreading across a number of cities. In addition, an institutional functionary known as the “merchant’s representative” had an incentive to insure accurate information about dealings was transmitted between merchants. The merchant’s representative was a trader from a foreign city who established himself in a trade outpost. The representative’s stature in his new city depended on his ability to entice foreign merchants to do business there, which in turn depended on his ability to ensure that traders in the city kept their obligations. This structure of these bridges and clusters enabled reputational information to flow across the Islamic Mediterranean in such a way that network governance could potentially play a major role in supporting Maghribi trade.   

Once the Maghribi group is understood as a bridge and cluster network, Bernstein argues, the facts cited by historians to undermine the Greif thesis are revealed to be consistent with, and indeed to support, the claim that reputation-based enforcement facilitated trade. For example, critics argue that the Maghribi group was not completely “closed,” and that the Maghribi often traded with non-Marghribis. It follows, they say, that the Maghribis must have relied on some alternative enforcement mechanism with respect to these merchants. However, Bernstein points out that this argument underestimates the effect of networks even on non-members of the network. Muslim and Jewish non-Maghribi traders still needed to do business with their Mahgribi counterparts. If a non-Maghribi cheated a Maghribi in a transaction, that information would spread along the network and thus impair the non-Maghribi’s ability to do business with the Maghribi in the future. Critics also argue there is no evidence of market-wide boycotts. However, Bernstein argues that market-wide action was not necessary to ensure traders complied with their obligations. So long as information about an agent’s misdeeds was communicated to that agent’s personal trading network, then the reputational consequences of misbehavior would significantly impair his future business. Accordingly, sufficiently strong incentives for cooperation could be created even in the absence of complete boycotts across the Maghribi group.

While Bernstein’s work support’s Greif’s claim, it also highlights an interesting methodological limitation facing any attempt to understand the Maghribi traders.  The cache of documents found in the Cairo Geniza, from which we know about the traders’ activities, is systematically biased. The bulk of the Geniza documents are letters sent from small merchants-agents in trading outposts to two very influential merchants in Fustat. Given the prominence of these merchants, it is highly unlikely that agents in trading outposts would try to cheat them in their business transactions. However, the same agents may certainly have behaved differently with other, less influential merchants, but we have little record of these interactions. The historical records, therefore, only provide data on a small class of commercial arrangements wherein network governance was likely to be effective.

Despite the qualifications, the paper provides broader insights into when private ordering is likely to be a successful alternative to reliance on judicial enforcement of contracts. The success of private ordering in this regard is often believed to be related to the presence of certain market characteristics: geographical concentration, social or ethnic homogeneity, and repeat dealing among a relatively small group of traders, in particular. The study of the Maghribi demonstrate how private ordering can be successful across large geographical distances without strong social or ethnic ties. The case study suggests that the strength, structure, and density of social ties are the crucial determinants of network governance, and not the type of ties.  In turn, this suggests that private ordering may be able to support exchange in a broader range of contexts than is commonly appreciated.