Transaction Costs and Customary Law in African Economies — Epoche C2
The question, and why 'bad law' is the wrong answer A wholesaler in a West African market town who is asked for goods on thirty days' credit by a trader from another region will usually refuse, and the refusal is not because the two disagree about the rule. Customary law and statute concur that debts are to be paid; both parties know it; neither disputes it. The trade nonetheless does not happen. Any diagnosis of market dysfunction that locates the problem in the content of the applicable rules has already failed to describe this case, which is the ordinary case. What is missing is not a rule but the machinery that makes any rule worth relying on, and the price of that machinery is what the literature calls transaction costs. The essay's claim, which the sections below try to make demonstrable rather than merely plausible, is that a great deal of what is reported as the inefficiency of African customary legal systems is the signature of high transaction costs operating on rules that are perfectly serviceable; that this can be shown by writing down the condition under which an unsecured promise is self-enforcing and inspecting which of its terms collapse when trade extends beyond a community; and that the implication for intervention is a shift from the content of law to the institutions that supply information about who has kept his word. Two papers of Coase's are involved and they do different work, which is worth separating because the essay as first published leaned on the wrong one. 'The Nature of the Firm' (1937) asks why, if the price mechanism allocates resources, so much economic activity is organised inside firms where it does not operate. Coase's answer is that using the market has costs — discovering what the relevant prices are, and negotiating and concluding a separate contract for each transaction — and that a firm substitutes a standing authority relation for that series of contracts, expanding until the cost of organising one more transaction internally equals the cost of carrying it out on the market. That establishes the existence of transaction costs and the idea of a margin at which institutional form is chosen. The step this essay actually needs is in 'The Problem of Social Cost' (1960). There Coase considers a world in which transaction costs are zero and shows that in such a world the allocation of resources is independent of the initial assignment of legal rights, since the parties will bargain their way to the efficient outcome whoever holds the entitlement. The result is generally known as the Coase theorem, a name Stigler gave it, and it is routinely cited for the proposition that law does not matter. Coase's own use of it is the opposite. The frictionless case is introduced as a benchmark in order to be discarded — he calls the assumption very unrealistic — and the substantive argument of the paper is that because transaction costs are always positive, the choice among actual institutional arrangements is what determines outcomes, and the comparison must be between arrangements that really exist rather than between a real one and an ideal. That is precisely the licence needed for the claim that intervention should target institutional foundations rather than legal content. Three costs, and which one binds Transaction costs are conventionally divided into three, and the division is useful here because the three behave very differently as trade extends beyond a community. Search and information costs. In a system in which the record of who is reliable is held in memory and conversation rather than in files, information about a counterparty's history is available at nearly zero cost within a network and at very high cost outside it. Establishing whether an unknown trader from another region has previously defaulted may require a chain of intermediaries, each of whom must be given a reason to answer honestly. Bargaining and decision costs. Customary dispute resolution characteristically aims at reconciliation rather than at determining who was right, and characteristically involves several elders or a chief. Within a community that is an advantage, since the parties must go on living together. Across communities it is a liability: the forum's jurisdiction is contestable, the outcome is not predictable from the terms of the agreement, and the process takes time proportional to the number of people whose assent is needed. Policing and enforcement costs. Enforcement rests on social sanction, reputation and the withdrawal of future dealing. This is the item that fails most dramatically outside the community, and the next section shows exactly how far it falls. Which of the three binds is an empirical question, and there is a body of survey evidence bearing on it. Marcel Fafchamps's Market Institutions in Sub-Saharan Africa (2004) reports enterprise and trader surveys carried out in several African countries, including Ghana, Kenya, Zimbabwe, Benin, Malawi and Madagascar, asking firms how they select customers and suppliers, how they extend credit, and what they do when an agreement is breached. The pattern is consistent across sites: courts are used in only a small minority of disputes and are widely regarded as too slow and too expensive to be worth invoking for ordinary commercial claims; new customers are required to pay cash; credit is extended as a relationship lengthens, and is concentrated on counterparties with whom a firm has dealt repeatedly. Firms behave, in other words, as if the binding constraint were the third item on the list and the first, and they respond by manufacturing the conditions under which reputation works — which is to say by restricting whom they trade with. The enforcement constraint, written down The behaviour Fafchamps documents follows from a condition that can be stated in one line, and stating it is what turns the essay's thesis from an assertion into something with consequences that can be checked. Take a promise that nothing but re