³ÉÈË´óƬ

Edwards and Ogilvie (2008) dispute the empirical basis of the view (Greif, e.g., 1989, 1994, 2006) that a multilateral reputation mechanism mitigated agency problems among the eleventh-century Maghribi traders. They allege, based on anecdotal evidence and an interpretation of the secondary literature, that the relations among merchants and agents were founded in law. This paper refutes this assertion using comprehensive quantitative analyses of all available primary sources and a careful review of the documents and the literature Edwards and Ogilvie cite. Their assertion that the legal system had a major role in supporting trade is based on unrepresentative and irrelevant examples, an inaccurate description of the literature, and a consistent misreading of the few sources Edwards and Ogilvie consulted. The claim that merchants' relations with their overseas agents were law-based in this historical era is wrong. Among the recent and new quantitative findings reported here is that: (1) less than one percent of the documents' content is devoted to legal activity on any matter. (2) The legal system was mainly used for mandatory, non-trade related matters. (Edwards and Ogilvie regularly present mandatory legal actions in non-trade related legal cases as evidence for voluntary legal actions in matters pertaining to trade.) (3) The documents reflect thousands of agency relations but there are less than six court documents possibly reflecting its use in agency disputes. (4) A ten percent random sample of all the documents finds no trade-related legal actions among Maghribis beyond those in the court documents. (5) ³ÉÈË´óƬ 75 percent of agency relations were not based on a legal contract. The paper also reaffirms the accuracy of Greif's documentary examples and sheds light on the roles of the legal system and reputation mechanism during this period. The empirical basis for the multilateral reputation view is stronger than originally perceived.