The Geniza Merchants.ppt

1、The Geniza Merchants,Presentation on the Debate by Max McDevitt Boston University (10/31/16),Overview,The Environment Greif 1989 Greif 1993, 1994 Edwards and Ogilvie 2012 Greif 2012 Goldberg 2012 Thoughts on Economic History,Who were the Maghribi Traders?,Group of Jewish traders in the 10th-12th Cen

2、tury Mediterranean Spread all over northern Africa, Sicily, Syria, and Egypt during the time of interest (11th century) Called the Maghribi because they had common roots in the Maghreb which corresponds to modern day Morocco and Algeria They did not call themselves the Maghribis as a business group.

3、 Rather, Goitein, the prominent Geniza scholar, coined the term. (p. 41, Goldberg 2012) Upper middle class, a desirable profession to be in at the time Greif following, Goitein, calls them the Maghribi Traders,Geniza,A geniza is a storeroom in a synagogue where documents containing the name of God a

4、re placed If a document had the name of God on it, it became sacred and could not be destroyed. These documents would be, if they didnt belong elsewhere, be placed in the local geniza The Geniza refers to the earliest surviving collection of geniza documents: they come from the Ben Ezra synagogue in

5、 Old Cairo (Fustat). “Rediscovered” in the 1890s and procured by St. Johns College and Cambridge University Over 225,000 documents now at Cambridge University Library (p.8 Goldberg 2012) There is a whole branch of academia dedicated to Geniza studies. The most prominent scholars are Shelomo Dov Goit

6、ein and Moshe Gil. Gil advised Avner Greif on his masters thesis,Greif 1989,“Reputation and Coalitions in Medieval Trade: Evidence on the Maghribi Traders” JEH, 1989 Trade in the 11th century Mediterranean was largely free Transaction costs in traveling with goods, monitoring and selling Agency rela

7、tionships reduce transaction costs, increase volume of trade “absent contractual problems, a merchant can decrease his shipping cost by sending goods to an overseas agent rather than traveling with his goods, a large efficiency gain could potentially be achieved by employing overseas agents.” (p. 85

8、7) Agency relationships should arise but,Greif 1989,There is no reason to assume ex-ante that an agent hired to monitor and sell goods will not abscond with those goods or the proceeds ex-post For agency relationships to occur, there needs to be an institutional structure that enables them to functi

9、on. Specifically, their needs to be an institution that allows agents to credibly commit ex-ante to ex-post cooperation This institutional structure can be public or private in nature. The legal system? According to Greif, it was not useful for solving the agency problem because of lack of informati

10、on, lengthy and costly nature of settling cases, etc Simply put the legal system couldnt solve the agency problem. This opened up the need for a private-order institution,Greif 1989,The Geniza provides extensive evidence of agency relations between the Geniza merchants. Thus, the agency problem must

11、 have been fixed somehow “Since agents were employed by the Maghribi traders, and trust prevailed in agency relations, it seems reasonable that they established a private order institution. Where contractual relations are expected to be repeated, reputation may provide the base for such an economic

12、institution.” Greif is 100% correct in his assertion that reputation and trust seem paramount in the Geniza documents,Greif 1989,The 2 person model (p. 867): Merchant and agent Merchant gives wage premium and implicit contract (never re-hire if cheated) Wage premium creates gap between agents expect

13、ed lifetime utility of working for merchant and expected lifetime utility of outside option Agent chooses between short-run gain of cheating the merchant this period and then receiving outside option and long-run gain of wage premium Merchant offers optimal premium defined as lowest premium for whic

14、h the agent will be honest Establishes honesty and cheating results without relying on arbitrary assumptions about types,Greif 1989,The Coalition: Same idea as 2 person model Closed coalition of merchants and agents Pay optimal premium and agree to never hire anyone who cheated another coalition mem

15、ber Note: this reduces the optimal premium because the multilateral punishment strategy reduces the expected value of the outside option to the agent Requires strong and reliable information transfer inside coalition,Greif 1989,Further details: Most agency relations we have evidence of are overlappi

16、ng: Principals used several agents at once Most agency relations we have evidence of are reciprocal: Subha The reciprocal nature of the agency relations further eases the constraint because there is effectively pledged collateral. This again lowers the optimal premium It is key to note that the hone

17、sty of an agent is established not by type, not by social pressure, not by norms, and not by legal threats. The honesty of the agent is entirely derived in a context of self-interested agents fearing loss of coalition membership. The coalition aligns incentives to overcome the agency problem and doe

18、s so in a way that strictly improves on bilateral implicit contracts,Greif 1989,Rigid implications of the model: No employment of non-coalition agents. This is due to the fact that the coalition reduces the wage required to keep an agent honest below that in the case of bilateral punishment Agents t

19、hat are known to have cheated should not be reemployed by coalition membersin fact there should be no cheating strictly speaking. Cheating is off the equilibrium path Limited reliance on the legal system. Strictly speaking this is not an implication of the model. Rather, this is an implication of th

20、e assumption that the coalition arose to solve the agency problem that the court was incapable of fixing Note: Greif never argues the model and the equilibrium should hold perfectly in the real world. But, the real world should track closely to the model if the theory is approximately true,Greif 198

21、9,Direct evidence (five letters used in Greif 1989): 2 letters dated 1055: Agent Abun Ben Zedaka accused of embezzling the money of a Maghribi trader. Many other Maghribi traders cancel agency relations with him (p. 868-869) Joseph ben Awkal of Fustat and Samhun ben Daud of Tunisia. “If your handlin

22、g of my business is correct, then I shall send you goods.” establishes a link between future agency relations and past conduct. This is the essence of the reputation mechanism according to Greif Also in this letter: The former believes the latter is cheating him and withholds agency services includi

23、ng paying some of Samhuns creditors in Fustat. Samhun complains his reputation is being ruined,Greif 1989,More Direct Evidence: A letter sent from a merchant in Palermo (Sicily) to another in Alexandria reads, “Had I listened to what people say, I never would have entered into a partnership with you

24、” (p. 871) linkage between past conduct and future opportunities is important The same letter details the merchant from Palermo unnecessarily sharing profits from a side-venture allegedly to calm any suspicion of poor behavior on his part. This merchant also signals a desire to end the relationship.

25、 This is evidence that the reputation mattered because of future dealings with other coalition members. That is, we have multilateral incentives for honesty, not just bilateral,Greif 1989,More Evidence and conclusion: “Evidence of business association between Maghribi traders and non-Maghribi trader

26、s (Jewish of Muslim) is rare.” (p. 877) Legal system was incredibly costly to use (p. 881) “The Maghribi traders overcame contractual problems associated with agency relationships by organizing such relationships through a nonanonymous organizational framework, the coalition.” (p. 881),Greif 1993,“C

27、ontract Enforceability and Economic Institutions in Early Trade: The Maghribi Traders Coalition” AER, 1993 Develops arguments made in 1989 paper in a formal repeated game model Key element in delivering results is the assumption of exogenous forced end of agency relation Multilateral Punishment Stra

28、tegy (MPS) is an equilibrium in the game Merchant prefers to hire non-cheater over cheater because the non-cheater wont be hired in the future by others, under MPS, and thus requires a higher wage to keep him honest The Carrot is still there, but the stick is gone. Strict Preference for hiring non-c

29、heater given other players MPS No additional historical evidence added to that from 1989,Greif 1994,“Cultural Beliefs and the Organization of Society: A Historical and Theoretical Reflection on Collectivist and Individualist Societies” JPE, 1994 Greif argues that culture is fundamental to understand

30、ing institutions in a society and how people behave. This argument dates to the beginning of the institutions school and features prominently in Douglass Norths Structure and Change in Economic History (1981) Greif defines cultural beliefs as “the ideas and thoughts common to several individuals tha

31、t govern interactionand differ from knowledge in that they are not empirically discovered or analytically proved.” (p. 915),Greif 1994,Cultural beliefs are interpreted in the formal model as belief functions/distributions the players have over each others strategies. Cultural beliefs are thus a mean

32、s of restricting the set of rational strategies As noted before, the game has multiple equilibria. Specifically, there exists a different symmetric equilibrium that corresponds to Bilateral Punishment Strategies (BPS) Cultural beliefs become a way to explain different equilibria. What equilibrium pa

33、th of play a given society is on is determined by the “culture” of that society BPS corresponds to individualistic societies and MPS corresponds to collectivist societies,Greif 1994,Path dependence Former equilibria and institutions have impacts on dynamics of cultural beliefs and subsequent institu

34、tional changes BPS and individualistic societies lead to stronger push towards strong legal system strong legal framework for enforcing contracts trade scales up more and more Genoa and other Italian cities and Europe dominate MPS and collectivist societies do not push towards strong legal system up

35、per bound on scale of operations because information flows and effectiveness of coalition cant grow too large trade doesnt expand relative decline of Maghribi Far reaching implications,The Debate,Edwards, Jeremy and Ogilvie, Sheilagh, “Contract Enforcement, Institutions and Social Capital: The Maghr

36、ibi Traders Reappraised.“ 2008 working paper Greif, Avner, “Contract Enforcement and Institutions among the Maghribi Traders: Refuting Edwards and Ogilvie” -2008 response, working paper Edwards and Ogilvie have a second working paper from 2009 under the same title Edwards and Ogilvie seem to focus t

37、heir argument over time Strawmen arguments, occasionally, on both sides, butEdwards and Ogilvie narrow in quickly on well formulated arguments while Greif consistently goes after strawmen,Edwards and Ogilvie 2012,“Contract Enforcement, Institutions, and Social Capital: the Maghribi Traders Reapprais

38、ed” EHR, 2012 Thesis: “This articlefinds that the Maghribi traders based their agency agreements on legal as well as informal mechanisms. A subset of traders did form webs of trusted associatesbut these were very different from Greifs hypothesized coalition, in neither being exclusive nor having a c

39、learly defined membership. The Maghribi traders used informal sanctions but also resorted to legal enforcement, in ways strongly resembling European merchants. We find no evidence that the Maghribi traders had more collectivist cultural beliefs than their European counterparts. Their similarities ar

40、e more striking than their differences.” (p. 423),Edwards and Ogilvie 2012,What is the evidence we actually have? Greif relies on fragments of 5 letters in the 1989 paper. This sample doesnt grow in his other papers. The evidence is thin given the size of his claims and their potential implications

41、The Geniza documents are not a random sample. The Geniza documents are not a representative sample “A geniza is not an archive, the artifact of an attempt to organize and preserve texts; it is a repository of purposefully discarded texts an anti-archive, as Goitein, the most important scholar of the

42、 historical Geniza calls it.” (p. 8, Goldberg 2012) Potential geographic biases: Fustat and SE Mediterranean Also, weird depositing history from letters marked to other places Potential content bias: letters between Jews and written in Hebrew,Edwards and Ogilvie 2012,Would we expect to find legal co

43、ntracts or court records in the Geniza? What evidence would we take as implying that a legal system was or was not useful? Primary ways that the Maghribi traders established business associations: subha (“formal friendship”) and legal partnership “Goldberg estimates that, in terms of proportion of t

44、ext, three-quarters of discussion in merchant letters is devoted to mutual service agency and one-quarter to legal partnership.” (p. 433) Even under subha relationships, use of the legal system could and did happen Parallel legal systems: Jewish and Muslim,Edwards and Ogilvie 2012,A thing to note: l

45、etters referring to reputation mattering are not evidence of the coalition existing A remarkable thing about the five examples Greif uses is that many of them can only be viewed as supporting his theory if they are viewed as exceptions to the rule: The first letter he cites is about embezzlement whi

46、ch would be off the equilibrium path of play The letter at the bottom of p. 870 in Greif 1989 is also about cheating happening My personal favorite: “Had I listened to what people say, I never would have entered into a partnership with you”,Edwards and Ogilvie 2012,A specific example: letter from Kh

47、alluf b. Musa in Palermo to Yeshu b. Ismail in Alexandria dating 1060 The sender explains that he split some profits of his own side venture with the other Greif notes that the former wished to end the partnership with the other and that this fact must mean that the profit sharing was done to avoid

48、any chance of being hit by the MPS (p. 871, Greif (1989) Butthey end up litigating over the ending of the partnership. This a) casts doubt on Greifs interpretation and b) shows that the legal systems were being used by the traders (p. 428),Edwards and Ogilvie 2012,Was it closed? “Most scholars regar

49、d the Geniza as under-representing relationships with non-JewsYet despite the selection bias, and all the practical reasons for coreligionists to trade within their own denomination, Goitein concludes that the Geniza letters reveal an astonishing degree of inter-denominational cooperationPartnership

50、s and other close business relationships between Jews and Muslims, or Hindus, or Christians were commonplace.” (p. 431),Edwards and Ogilvie 2012,In conclusion, the legal system had the potential to be used and we do have evidence from the Geniza that it was used. We have a fair amount of evidence showing that MPS was not played in the manner described by Greif. There certainly wasnt a closed coalition. Furthermore, it seems doubtful there was much of a coalition,