A Simple Theory of the Extended Family System and Market Barriers to the Poor Karla Hoff World Bank khoff@worldbank.org Arijt Sen Jawaharlal Nehru University senarijit@vsnl.net May 2002 _________ We would like to Maitreesh Ghatak and participants at the Santa Fe Institute Conference on Poverty Traps for useful comments, and the Russell Sage Foundation and The Pew Charitable Trusts for financial support. The findings and interpretations expressed in this paper are those of the authors and do not necessarily represent the views of the World Bank. Abstract In many subcultures, especially in countries at an early stage of development, individuals in extended families can commit to participate in informal insurance arrangements, which we call the extended family system. It is commonly observed that individuals in such an extended family system who could have made transfers to kin in cash, instead make them in-kind. We model two aspects of such in-kind transfers: nepotism in the workplace and the sharing of rental housing with extended kin. These practices can create negative externalities for employers and landlords. Specifically, they can exacerbate moral hazard problems that arise because of the inability of principals to monitor agents’ actions. This, in turn, can lead employers and landlords to discriminate on the basis of subculture. Extended family systems that increased welfare before market opportunities existed may thus lower welfare if they persist after a subculture enters into extensive contact with markets. We distinguish two forces that can lead the extended family system to persist even if all individuals would be better off without it. One set of forces tends to keep participants in when, because of the reputation effect, they would be better off out: an individual might want to commit to an employer or landlord not to participate in the extended family system, but if the only means to do that is social assimilation to another subculture, the cost may be too high. Another set of forces—improvements in future income prospects—tends to operate on the best of the kin group, causing them to exit. That, in turn, may trigger the extended family itself to create obstacles to new market opportunities. A Simple Theory of the Extended Family System and Market Barriers to the Poor The question of the relation between the profit motive and social customs is a very old one in economics. In one view, that of standard neoclassical economics, social customs do not play any independent role. While different cultures are characterized by different regularities in behavior, these regularities reflect different conditions and change when it becomes to everyone’s advantage to change them. “Social customs follow the money,” as some sociologists put it. But there is another view, according to which social customs that are harmful to every individual can nonetheless fail to be overturned by self-interest. The line of reasoning is that social customs can affect economic outcomes whenever there is a market failure. (For example, the custom of treating people according to a norm of fairness is important precisely because it is prohibitively costly to specify all possible outcomes of an interaction.) But the very incompleteness of markets means that the “market” for social customs may itself be imperfect. The present essay inscribes itself in the line of this second view.1 The social custom that this paper considers is the extended family system (kin system, for short). By definition, it is “a system of shared rights and obligations encompassing a large number of near and distant relatives” (Wolf 1956). We model it as 1 Studies of dysfunctional social customs include Akerlof 1976 (a caste system), Greif 1994 (collectivist enforcement), Sah and Stiglitz 1989 (opposition to innovativion), Banerjee and Newman 1996 and Kranton 1996 (reliance on personalized exchange instead of market exchange), and Young 1998. a social contract for mutual insurance within an extended family. It arises in subcultures where there is an informal legal system to enforce such a social contract. We think of the kin system as a pre-existing network into which new parts of the economy have to be fitted.2 What are the implications for individual behavior in markets? Can the kin system persist even if it does not increase welfare? Arnott and Stiglitz (1991) show that one possibility is that informal insurance arrangements mitigate moral hazard problems (by exploiting monitoring devices not otherwise available). A second possibility is that they exacerbate moral hazard problems and lower welfare (by giving individuals too much access to insurance, they lower the care they take to avoid accidents). We show in this paper is that the latter concern about the worsening of moral hazard in formal markets applies much more generally. The kin system can exacerbate moral hazard problems in markets that have nothing to do with insurance. In any market where monitoring and utilization are costly, the kin system can lead to lower wages, blocked employment possibilities, or higher prices charged to individuals in subcultures characterized by the kin system. It can thus change the way an entire subculture with a pre-existing kin system becomes integrated with the market. The central argument of this paper is that, in markets with asymmetric information, the kin system provides individuals with ways to exploit the problem of asymmetric information that they would not otherwise have. One important example is nepotism. Once a member of a kin group is in a managerial position, with power to recruit and promote, other members of the kin group exert pressure on him for favors and can enter into side-contracts through which transfers required under the kin system are 2 Arrow (2000, p. 4) suggests that all existing social relations may be usefully looked at in this way, rather than (as is currently fashionable) looking at them as “social capital.” 2 fulfilled in-kind. Labor market evidence supports the view that ethnic loyalties introduce new constraints into organizational performance (Collier and Grag, 1999). Another example occurs where an individual in a kin group who becomes rich enough to leave his village and rent a city apartment fulfills his obligations to his kin at least partly through sharing his home.3 Such obligations to one’s extended family create externalities for one’s employer and landlord. More generally, kin groups entail a web of reciprocal obligations that can be a menace in modern, market-based settings because they aggravate the problem of moral hazard. As a result, individuals from a subculture in which the kin system is widespread may be stereotyped and find that they face less favorable terms in the market than individuals from subcultures in which the kin system is not practiced. The idea that the kin system may give rise to barriers in the labor market has long been recognized. As evidence of this point, consider the following passages from pioneer books in development economics, cited in Platteau (2000, pp. 20-22): Where the extended family system exists, any member of the family whose income increases may be besieged by correspondingly increased demands for support from a large number of distant relations… A strong sense of family obligation ... may cause a man to appoint relatives to jobs for which they are unsuited...it may be fear, rather than affection, which drives them to nepotism” (Lewis 1955 p. 114 ) Demands to provide jobs for a wide range of kin, irrespective of their qualifications, and requests for cash donations or gifts of varying amounts from a stream of ‘visitors’ are probably the most frequent claims [on businessmen in Africa.]. (Kennedy 1988, p. 169) [The extended family system] adds to the reluctance of foreign firms to employ members of the local population in positions of trust and responsibility. (Bauer and Yamey 1957, p. 66) 3 In urban apartments in developing countries, it is common for an entire bank of mail boxes to display not a single name. The usual explanation is that people fear unending demands to provide shelter to a large number of distant relations, and so try to conceal where they live. 3 The presumption that these authors attempt to establish is that (i) the extended family system may not be optimal because of its effect on the labor market, and (ii) a dysfunctional kin system may persist indefinitely. But there does not exist a theory of the equilibrium of the interactions between the kin system and the labor market or goods markets, and without a theory it is difficult to make welfare economic evaluations of the system. This paper offers a simple theory. In this theory, each extended family in a subculture takes the market environment as given. This is reasonable, since a single family is only a small part of the subculture. But in response to the informal insurance arrangements of the kin groups in a subculture, market opportunities do change. An individual’s subculture conditions his opportunities on the market. The kin system— which may have had decisive importance for survival in high-risk peasant societies—may be harmful once a market system is created. We then distinguish two sets of forces that may cause a dysfunctional kin system to persist. One set of forces tends to keep participants in when they’d be better off out, because of a reputation effect. An individual might want to commit to an employer or landlord not to participate in the kin system. But if the only means to do that is social assimilation to another subculture, the cost may be too high. Another set of forces are, paradoxically, improvements in future prospects for earning income. Such prospects tend to attract the best of the kin group and cause them to exit. Their exit makes the rest of the kin group worse off. It may thus trigger opposition to the new prospects, which we call status quo bias.4 4 A survey of sources of such bias is Kuran (1988). 4 The idea that a new opportunity that selects the best from the kin group will be vigorously opposed by the group is captured in Carol Stack’s (1974) All Our Kin. This book describes a kin system in a poor African-American community in a US city. Cooperating kin exert various forms of social control to reduce the ability of group members to exit. The following passage reports details of a woman’s story that were substantiated by Stack’s discussions with many other people. Me and Otis could be married, but they all ruined that. Aunt Augusta told Magnolia that he was no good. Magnolia was the fault of it too. They don’t want to see me married! Magnolia knows that it be money getting away from her. I couldn’t spend the time with her and the kids and be giving her the money that I do now. I’d have my husband to look after. I couldn’t go where she wants me to go. I couldn’t come every time she calls me, like if Calvin took sick or the kids took sick, or if she took sick. That’s all the running I do now. I couldn’t do that. You think a man would put up with as many times as I go over their house in a cab, giving half my money to her all the time? That’s the reason why they don’t want me married. You think a man would let Aunt Augusta come into the house and take food out of the icebox from his kids?...(p. 114) This paper proceeds as follows. The next section provides evidence that a subculture can provide an informal legal system that permits people in an extended family to define and enforce a social contract for mutual insurance. Section 2 develops, by means of examples, aspects of the moral hazard problem that the kin system may create in labor and housing markets. Section 3 analyzes status quo bias. 1. Culture as a control mechanism Our view of culture follows closely that of Clifford Geertz, who said that “culture is best seen not as complexes of concrete behavior patterns…but as a set of control mechanisms—plans, recipes, rules, instructions…for the governing of behavior” (quoted 5 in Becker, 1996, p. 16). In some cultures, a use to which such control mechanisms are put is the enforcement of egalitarian insurance arrangements. In cultures where kinbased mutual insurance arrangements are important, including cultures in Africa, the Philippines, and Vietnam, egalitarian norms and ethical values prescribing the right to subsistence are widespread (see, for example, Scott 1976 and Platteau 1991, 1996). Sanctions for those who shirk the obligations of the kin group system come in the form of economic consequences (loss of employment or destruction of property), loss of status or even social ostracism, and violence ((Sahlins 1965: 208-213; Colson 1947: 44-50). Punishments also can take a religious form, as the case below describes: [I]n a celebrated Kenyan court case, a successful Luo man left a will specifying that he was to be buried by his wife in Nairobi rather than by his kin group in his home district. The kin challenged the right of the wife to conduct such a burial. The Kenyan high court overturned the will, siding with the kin group. This decision was important because the kin group’s ultimate sanction against recalcitrant members is to bury them near their ancestors, who can be expected to exact retribution. The Kenyan high court was upholding the social obligations inherent in kin group membership.” (Collier 2001, p. 298). Modern social groups can also create such control mechanisms. In the extended family system described by Stack (1974) in All our kin, African-American households in a low-income community in a US city enter into temporary child exchange as a symbol of mutual trust, and the bonds that develop between adults and children reinforce adults’ incentive to honor their commitments to the extended family (pp. 28-29). This network shares the three features that we take to define a kin system: it is a kin-centered nonmarket insurance arrangement, it is long-lasting, and it exhibits control mechanisms that create trust. 6 2. Market discrimination as a response to the extended family system Building upon the observation that some subcultures are characterized by beliefs and norms that make possible the enforcement of the extended family system, we now develop a simple model of such an arrangement. There is a population of risk-averse agents. Each agent belongs to exactly one extended family, and each extended family is part of exactly one subculture. We define a subculture as a set of shared concepts, beliefs, and commitment devices of an ethnic group exhibiting characteristic appearance or patterns of behavior sufficient to distinguish it from others within a society. It is assumed that there exists a subculture, denoted by s′, with control mechanisms to ensure that commitments for mutual insurance within an extended family are enforced. It is also assumed that in other subcultures, the set of shared concepts and beliefs do no permit such enforcement. In those subcultures, informal insurance arrangements will be greatly restricted and, for simplicity, we make the stark assumption that no informal insurance arrangements extend beyond the boundaries of a household.5 We denote such a subculture by s. Our results do not depend on the precise form that the informal insurance arrangement takes. We therefore present the simplest possible model of an egalitarian insurance arrangement. Consider a representative extended family in subculture s′ with N individuals, indexed by i. The income of the ith individual is the sum of two terms: an average income y i and a shock θi. Shocks have zero mean and are independently distributed across individuals. Because of risk aversion, individuals will gain from risk- 5 In a setting in which agents play a repeated game of indeterminate duration, some insurance arrangements are self-enforcing; see, e.g., Coate and Ravallion 1993, Fafchamps 1992, Kocherlakota 1994, and Ray 2001 [to be added]. But as recent work emphasizes, the scope for such arrangements can be quite limited. 7 sharing. They achieve risk sharing by stipulating that they pool the shocks, θi, and share them equally. Each individual thereby obtains an income yi + θ , N where θ = [1 / N ]∑ θ i . i =1 One’s expected income has not changed but his risk has been reduced. Thus, the ex ante gain from such a scheme is positive and voluntary participation ensured for any risk averse individual. Individuals have identical preferences, but the fact that some individuals have the opportunity to participate in an extended family system (and actually choose to do so), while others do not, means that there are two possible types of individual: A K-agent belongs to an extended family (or kin) system, and an I-agent does not. All members of subculture s′ are K-agents, and all members of subculture s are I-agents (because, by assumption, they are not able to enter into informal insurance arrangements with their extended family). We use this notation because later we will consider the incentives of a K-agent to socially assimilate to subculture s′ in order to become an I-agent. This section has two objectives. We first present two examples to illustrate the kinds of forces that will lead to discrimination between individuals of subculture s and s′. We then consider the incentives of an agent of subculture s′ to socially assimilate to subculture s as a means to commit not to participate in an extended family system. 8 A. Two examples of endogenous barriers to markets Example 1. Nepotism In this first example, an agent undertakes a productive activity on behalf of a principal and in doing so, hires an assistant. By nepotism, we will mean favoritism in hiring towards relatives at a cost in the quality of employees. Some nepotism can be prevented by rules barring the hiring of family members. But when employees have obligations to an extended family, it is difficult for an employer to identify nepotism because members of the group may include distant relatives. The model Consider a sector in which the production process requires a senior worker and a junior worker. Both are unionized jobs: the senior worker has to be paid a base salary of w and the junior worker has to be paid w , with w > w. There are many firms and in some of them, the “match quality” between the junior worker and the firm affects the level of output. If the match is good, output is equal to π , and if it is bad, output is equal to π , where π > π. Only the senior worker can ascertain the quality of the match, and so the responsibility of selecting the junior worker is delegated to him. If the senior worker picks an arbitrary individual (or his nephew ) to be the junior worker, then the match quality will be good with probability λ ∈ {0,1}. On the other hand, if the senior worker expends search effort c, he will find a good match with probability 1. Because the senior worker has to be paid a baseline wage of w whether or not he screens, the firm can institute a bonus scheme, where the senior worker is paid w + β when π = π in order to induce him to search for a good match. We now characterize the optimal bonus scheme for an I-agent and a K-agent. 9 We write an individual’s utility function as u(y) – e, where e is the effort expended in searching. Under a bonus scheme β, an I-agent will search if and only if u( w + β) – c ≥ λu( w + β) + [1-λ]u( w ) i.e., he will search if and only if the bonus is at least βI , where βI solves the above with equality. The firm’s expected payoff is thus π - w - βI if it offers that bonus scheme, and λ π + [1-λ]π - w if it does not offer a bonus scheme. We assume that the former is higher than the latter (i.e., the values λ , π - π , and c are small). Thus, if the firm hires an I-agent as the senior worker, the optimal wage contract is to offer him a bonus βI when output is high. Now consider a K-individual as a senior worker. The moment that this individual has the job at w , that determines his transfer, denoted by τ , to his kin group. Here we assume that τ is determined by his permanent wage w and not by any subsequent bonus that he might receive, and that w is sufficiently high in relation to the individual’s ex ante expected income y i that to obtain a job as a senior worker puts the individual near the top of the income distribution in the extended family. That is, recalling Section 2, it is assumed that for one who attains the status of senior worker, the “shock” to one’s own income, [ w − y i ] , exceeds the average income shock in the kin group. That difference is N − 1 τ ≡ [ w − y i ] − θ > 0. N 10 Every K-individual has a nephew in his kin group who, on his own, has a probability µ < 1 of getting a job as a junior worker. If the K-agent who is a senior worker engages in nepotism by giving his nephew the junior job, that is construed by the extended family as making an in-kind transfer of value [1-µ] w . Denote by δ the net amount (net of the in-kind transfer) that the senior worker owes to his kin group: δ ≡ τ - [1-µ] w and we assume that δ ≥ 0. Now consider a K-agent who has been hired as a senior worker. He will obtain utility u( w + β - τ) – c if he searches, and expected utility λu( w + β - δ) + [1-λ]u( w - δ) if he does not search and engages in nepotism. Thus, the minimal bonus that induces him to search is βK such that u( w + βK - τ) – c = λu( w + βK - δ) + [1-λ]u( w - δ) Note that if δ is arbitrarily close to zero, then evaluated at β = βI , the left-hand side is less than the right-hand side. This implies that for δ small, βK > βI. This means that when a K-agent can meet a large part of his kin obligations by nepotism, he has to be given a larger bonus than an I-agent in order to induce him to search for a “good match”. 11 To summarize, the following inequalities hold for δ sufficiently small: (a) Hire an I-manager with bonus βI is preferred to: hire anyone with no bonus scheme (b) Hire an I-manager with bonus βI is preferred to: hire a K-manager with bonus βK It follows that if a K-manager can fulfill a sufficiently large part of his transfer obligations to his kin group by nepotism, no firm will hire a K-individual in lieu of an Iindividual. Example 2. Housing In the preceding example, an agent undertook production activities on behalf of a principal. In this example, an agent rents a durable good from a principal. We focus on housing, but similar problems arise for a principal who rents any durable good for which both utilization and monitoring are costly. We employ the simplest model in which moral hazard is present. There are two locations, the city and the surrounding villages. In the city, there are a set of housing units. A set of agents holds jobs in the city. A requirement of the job is that the jobholder rent a housing unit in the city. (For example, punctual attendance may require that the individual has assured housing near his job.) The agent’s utility function is u(m) - v(n) where m is the numeraire good and n is the number of residents, in addition to the household, who reside in the housing unit. The functions u and v are increasing and 12 strictly concave, and v(.) denotes the agent’s disutility from sharing his apartment with kin outside his household.6 To a landlord, the opportunity cost of renting a housing unit is (i) the opportunity cost, F, of his fixed capital and the depreciation (i.e., wear and tear) when the home is occupied by one household, and (ii) the additional depreciation C(n) when the renting household has n kin members sharing the unit. We take C(n) to be an increasing convex function. While a landlord would like to specify the value of n in the rental contract, we assume that is not possible. For example, the law may bar the landlord from making spot checks of the apartment at night. Thus, a household can only be charged a lump-sum rental rate, denoted by R, independent of how many kin members it shares the unit with. (We address the issue of determining the value of R below.) Access to housing in the city is valuable to someone who lives in a village. For example, it could save a student the cost of commuting to a school in the city, or increase an individual’s probability of finding a city job. Let H denote the monetary value of access to city housing, e.g., a place on the floor of an apartment. Consider a K-agent who rents a housing unit at R, has realized an income yi, and has to make a transfer τ to his extended kin. (In general, τ could be negative or positive: any transfers that the household makes in-kind could be compensated through transfers from the extended family system). Let mi denote own consumption of the numeraire good, and let mk denote transfers (which also could be negative or positive) of the numeraire good to kin. The agent’s problem is 6 For simplicity, we treat n as a continuous variable. 13 Max u(mi) - v(n) {n, mi , mk } subject to yi ≥ mi + mk + R (income constraint) τ ≤ mk + nH (transfer constraint) or, equivalently,7 Max u(y – R - τ + nH ) – v(n) n His choice of n solves u′(mi )H = v′(n) Since there is a benefit to the extended kin to be housed in the city, the left-hand side is strictly positive. Recognize that for v′(0) small enough, a tenant in an extended family system will always share his housing unit: n* > 0.8 Now consider the landlord’s problem in determining R. In a competitive market with perfect foresight, when a landlord rents to a K-agent with income yi, the landlord 7 The constraints always bind at a solution to this problem; otherwise, the individual could lower the transfers while still fulfilling his obligations. 8 We have implicitly made a stark assumption. Only a K-agent makes side-contracts for use of his rented home. We can easily relax that assumption. Our qualitative result requires only that the extended family system increases the “gains from trade” from such a side-contract. That effect could spring from many sources: for example, an explicit contract in which a renter sublets housing could be illegal and therefore unenforceable; the extended family system may create bonds that make sharing one’s apartment less burdensome (v(n) shifts down for a K-agent); or the fact that an extended family system creates multiplex bonds and limited privacy among a wide group of extended kin increases the scope for punishing an individual who harms his host. 14 will recognize that such an agent will share the housing unit with n* kin members, and so in order to break even, the landlord will charge R = F + C(n*). Thus, although the K-tenant does not directly pay the marginal depreciation cost of sharing, he indirectly pays for the cost of increased utilization by having to pay a higher rental rate. In general, when landlords do not have precise information as to whether their potential tenants are K-agents or not, they will use membership in a subculture as an indicator: the rental rate will be conditioned on subculture. If n is the average value of n for an agent in subculture s′, then a landlord will wish to condition the rent on the tenant’s subculture (denoted Rs′ , Rs ), and the equilibrium rental prices will satisfy Rs′ – Rs = C( n ) Although the variable n is chosen by the tenant on the basis of zero marginal costs, rather than true costs, the tenant in a given subculture pays the balance in terms of higher contractual rents, as shown above. In aggregate, renters in the subculture bear the full social cost of the use of housing. But since they face a distortion in the marginal cost of housing, the allocation is inefficient. The break-even condition of landlords requires that Rs′ be varied to satisfy 15 dRs′ = C′d n . The utility of a tenant who is a K-agent is u(y – R - τ + nH ) – v(n). The change in that utility level for the typical agent as n varies is u′[Hd n - C′d n ] - v′d n which, for the tenant who chooses to shelter the average number of kin (n* = n ), is equal to -C′d n < 0. The implication is that the practice of transfers in-kind rather than in cash must generate utility losses for some K-agents in the economy. If price discrimination on the basis of the renter’s subculture is prohibited, then (since the break-even price for s′ exceeds that for s), no landlord will be willing to rent to an individual of subculture s′.9 9 One could express the same idea with a more general model. Suppose individuals in the village who want to live in the city each choose between renting a housing unit from an owner, or not, and that K-agents have the additional option of entering into a side-contract with a renter in the kin group to share. For subculture s′, the rental price of housing is high and only a few people can afford apartments in the city. For subculture s, the rental price is low and many people can afford apartments in the city. In the first case, the need for insurance (lest one be one of those who can’t afford a city apartment.) is high; in the second case, it is low. There may be simple examples of utility functions, income distributions, and C(n) such that, for some parameter values, the extended kin system increases ex ante welfare, while for others it lowers it. 16 B. A poverty trap A premise of our analysis is that an individual’s participation in an extended family system is not observable by the market, but the subculture to which an individual belongs is easy to observe. In this section we argue that a subculture can maintain the social custom of the extended family system even when the consequence of that custom is to lower everyone’s welfare and income. Consider an individual with a given, exogenous endowment. If his subculture is s′, he chooses whether or not to participate in an extended family system. Let U sK' denote his expected utility if he participates, and let U sI' denote his expected utility if he does not. We now distinguish two cases. In the first case, an individual can choose his behavior within a subculture, but cannot “choose” his subculture. That is, he cannot socially assimilate to another culture by adopting that culture’s religion, language, or patterns of behavior. If U sI > U sK' > U sI' then the individual in subculture s′ will choose to participate in an extended family system, even though, could he have shed his reputation as someone from a subculture in which collective beliefs and private incentives lead to the extended family system, he would have been better off. The extended family system was optimal once, is no longer optimal, but is maintained because individuals cannot commit to being from a subculture in which the extended family system does not exist. And not being able to commit, they are better off belonging to an extended family system than not. And thus the reputation 17 that the subculture has for sharing is inherited by each new generation.10 A population’s vulnerability to risk leads to a social custom, the extended family system, and that custom may outlast its usefulness and perpetuate poverty when valuable market opportunities emerge. Next consider the case more supportive of change. Suppose that by social assimilation to a subculture s, or by migration, an individual can commit not to participate in the extended family system. Platteau (2000) argues that this process has occurred all over the world, but that in some highly traditional societies, principles of equity are so adverse to change [that] a single individual, even when endowed with special qualities and powerful psychological resources, cannot successfully defy the conventions of the society. He will unavoidably…be squashed by various forms of opposition, especially when his economic success depends on his behavior as a hard-nosed businessman in dealing with fellow tribesmen. To break through, he needs the protection afforded by the deviant actions of a sufficient number of other innovators in his locality. Rising economic opportunities alone will usually not suffice to generate dynamic entrepreneurs in the absence of a critical mass of cultural energies harnessed towards countering social resistance… Experience shows that a critical mass of return migrants who have been exposed to outside opportunities for a sufficiently long time are likely to form a core dissident group able to call customary norms and values into question. They may then act as role models dragging other villagers into opposition… (Platteau, 2000, emphasis added)11 We model this idea of threshold effects by assuming that there is a cost E(.) of publicly exiting the extended family system and the subculture that supports it, and that E is decreasing in the fraction ρ of the subculture that exits the extended family system. If for all individuals in a subculture 10 Formal models of the persistence of behaviors that lead to a group reputation that makes everyone worse off are Stiglitz 1974 and Tirole 1996. 11 We quote this passage at length as a corrective to the view, implicit in some recent work on informal insurance arrangements, that the central problem raised by informal insurance arrangements is that of enforcement. While this is true in some settings, in traditional peasant societies the obstacle to efficiency 18 U sK' > U sI − E (0) then the outcome is that all individuals who inherited subculture s′ will choose to live by its values. That is an equilibrium because, since no one departs from traditional values, the cost of doing so is very high. If U sI − E (1) > U sK' then it is also an equilibrium for individuals who have inherited a subculture s′ to abandon its values. That is an equilibrium because when all individuals innovate in this way, the cost of innovation is low, and that gives each individual an incentive to innovate. It is clear that both inequalities can hold in a given setting; both outcomes can be equilibria; but the second outcome is better for every individual than the first. In sum, a subculture may not have a switching capacity, in the sense that s′ → s when all would gain by that switch. We have identified two independent reasons for that. The first is that reputation occurs at the group level, not at the individual extended family system level, and so no one internalizes the gain to behaviors that change a group’s reputation in the market. The second is that if each individual is free to “switch” but at some cost, the cost is likely to depend on the number of individuals who exit; there may be multiple patterns of behavior that give rise to incentives that make those behaviors an equilibrium. may not be lack of enforcement but instead may be too much enforcement—the power of the group as a 19 3. Status quo bias In a setting where the extended family system exists alongside the market, it is useful to distinguish two sets of forces that impinge on such a system. One set of forces, discussed above, tends to keep participants in (when they’d be better off out) because of reputation or the force of tradition: an individual might like to be able to commit to an employer or a landlord not to participate in an extended system, but if the only means to do that is social assimilation to another culture, it is costly. Another set of forces – improvements in future opportunities to earn income – tends to attract the best of the kin group and cause them to exit. Their exit would make the rest of the kin group worse off. It may thus trigger opposition to the new opportunities from the extended family, which we call status quo bias. In this section we sketch the nature of the argument to be developed in a future draft. One view might be that an improvement in outside opportunities that increases the gains to exit from the extended family system would automatically be followed by greater exit from the extended family system and thus a reduced social cost of that system. But new opportunities may also lead to more differentiation. Some individuals’ future prospects improve, whereas those of others do not. Those whose prospects improve may choose to socially assimilate to culture s in order to fully benefit from new market opportunities. The exit of the gainers creates losses (in reduced risk sharing) for the others. If the identity of the gainers is not known ex ante, an opportunity that benefits more than half of the members ex post may be opposed by more than half of the members whole to impose on all individuals an informal insurance arrangement that distorts their incentives. 20 ex ante.12 A majority of the group may thus gain ex ante by measures that discourage individuals from trying to take advantage of the new opportunities. The extended family system may itself become an instrument of stagnation—an instrument that it would be in no one’s interest to wield (ex ante or ex post) absent the tight network of mutual dependence that the extended family system creates. This can be shown as follows. Consider a new opportunity (e.g. , education), where there is individual uncertainty about who will benefit. Each individual has the same ex ante probability p of gaining an amount g from education (and a probability 1-p of enjoying neither direct gain nor loss). A consequence of taking advantage of the new opportunity is that the individual chooses to exit the extended family system.13 Suppose that before individual uncertainty is resolved, i.e. at an initial point t = 0, individuals through some collective choice process reach a decision on whether or not to try to block access to the new opportunity.14 If they do not block access to the new opportunity, then (at t = 1) individuals explore the new opportunity and learn whether they are among the gainers. The gainers then exit the extended family system. Let f(θ1θ2…θN )denote the joint probability distribution of θi , and order individuals, i = 1, 2, 3, …N so that the last pN individuals are the gainers from the new opportunity. Before the identities of the gainers are known (at t = 0), an individual will prefer to block access to the new opportunity if 12 An elegant formulation of this point in the context of trade reform is Fernandez and Rodrik (1991). 13 Consider two examples. First, education leads to opportunities for managerial jobs, but only if the individual has assimilated to a culture s, thereby signaling that he is not part of an extended family system. Second, exposure to worldly influences that conflict with the traditional values of a subculture may itself lead to social assimilation to culture s. 21 ∫ u ( yi + θ ) fdθ 1 ...dθ N > p ∫ u ( y i + g + θ i ) fdθ i + [1 − p ]∫ u ( y i + [1− p ] N 1 ∑θ i [1 − p]N i =1 ) f dθ1 ...dθ [1− p ] N Recalling Section 2, the left-hand side of this inequality represents the status quo: it is the expected utility of an individual i with expected income y i who participates in an extended family system with a total of N individuals, each of whom faces a risk θ, which may differ across individuals. Compared to the status quo, the first term on the righthand side represents a gain in mean income, obtained with probability p, and an increase in risk-bearing. The second term represents no gain in mean income but an increase in risk-bearing, which results with probability 1-p because of the exit of the gainers from the extended family system. The status quo (the left-hand side) can dominate the new opportunity for all individuals if the cost of increased risk after the partial breakdown of the extended family system is sufficiently high. In that case, at t = 0, all will oppose the new opportunity. This can hold even if p > ½, so that ex post (at t = 1, when the identities of the gainers are known), a majority of individuals benefit from the new opportunity. The problem is the unavailability of “ability” insurance: those who do not benefit from the new opportunity become worse off because they will have to share their income risk with a smaller pool. 4. Conclusion In this paper we developed a simple model of the extended family system as a nonmarket insurance arrangement that is enforceable in some subcultures, but not in others. By 14 We do not model how they will oppose it. Our maintained assumption is simply that, within an extended family, an action is more likely to be taken , the larger is the number of individuals who gain by it. 22 widening the scope for side-contracting among individuals in an extended family, it can exacerbate problems of moral hazard in all markets with asymmetric information. Thus, individuals from different subcultures may face distinctively different market opportunities. The paper offers a new perspective on the difficulties of the transition to the market of a subculture characterized by the extended family system. 23 References Akerlof, G., 1976, “The economics of caste and of the rat race and other woeful tales,” Quarterly Journal of Economics 90, 599-617. Arnott, R. and J.E. Stiglitz, 1991, “Moral hazard and nonmarket institutions: Dysfunctional crowding out or peer monitoring?” American Economic Review 81, 179-90. Arrow, Kenneth, 2000, “Observations on Social Capital,” in Social Capital: A Multifaceted Perspective, eds. P. Dasgupta and I. Serageldin, pp. 3-5. Banerjee, A. and A. Newman, 1996, A dual-economy model of modernization and development, Review of Economic Studies, forthcoming. Bardhan, P. and A. Rudra, 1986, “Labor mobility and the boundaries of the village moral economy,” Journal of Peasant Studies. Basu, K., E. Jones, and E. Schlicht, 1987, “The Growth and Decay of Custom: The Role of the New Institutional Economics in Economic History,” Explorations in Economic History 24, pp. 1-21. Bauer, P.T. and B.S. Yamey, 1957, The economics of under-developed countries, Cambridge, UK: Cambridge University Press. Becker, Gary S. 1996, Accounting for Tastes, Cambridge MA: Harvard University Press. Besley, Timothy, 1995, “Nonmarket institutions for credit and risk sharing in lowincome countries,” Journal of Economic Perspectives, 9, 115-127. Coate, S. and M. Ravallion, 1993, “Reciprocity without commitment: Characterization and performance of informal insurance arrangements,” Journal of Development Economics 40, 1-24. Collier, Paul, 2001, Comment on “Distributive Conflicts, Collective Action, and Institutional Economics,” by Pranab Bardhan, in Frontiers of Development Economics: The Future in Perspective, eds. G. Meier and J. Stiglitz, pp. 296-299. Collier, Paul and Ashish Garg, 1999, “On kin groups and wages in the Ghanaian Labour Market,” Oxford Bulletin of Economics and Statistics 61 (2) pp. 13357. Colson, E., 1947, Tradition and contract: The problem of order (Chicago, Aldine). 24 Fafchamps, M., 1991, “Solidarity networks in preindustrial societies: Rational peasants with a moral economy,” Economic Development and Cultural Change, 147-173. Fernandez, R and D. Rodrik 1991, “Resistance to Reform: Status Quo bias in the presence of individual –specific uncertainty,” American Economic Review 81 (5) 1146-1155. Greif, A., 1994, “Cultural beliefs and the organization of society: A historical and theoretical reflection on collectivist and individualist societies,” Journal of Political Economy, 102,12-50. Kennedy, P., 1998. African capitalism—The struggle for ascendancy, Cambridge, Cambridge University Press. Kocherlakota, N.R. 199- “Efficient bilateral risk sharing without commitment,” Review of Economic Studies. Kranton, R., 1996, “Reciprocal exchange: A self-sustaining system,” American Economic Review 86, 830-51. Kuran, Timur, 1988, “The tenacious Past: Theories of personal and collective conservatism” Journal of Economic Behavior and Organization Sept. 10 pp. 14371. Lewis, W.A. 1955, The Theory of Economic Growth, London: George Allen & Unwin. Platteau, J.-Ph, 1991, “Traditional systems of social security and hunger insurance,” in: E. Ahmad, J. Dreze, J. Hills, and A. Sen, eds., Social security in developing countries, Oxford: Clarendon Press, 112-170. Platteau, J.-Ph, 1996, “Traditional sharing norms as an obstacle to economic growth in tribal societies,” CRED working paper no. 173, Facultes universitaires Notre-Dame de la Paix, Namur, Belgium. Platteau, Jean-Philippe, 2000, “Institutions, Social Norms, and Economic Development,” Harword Publishers, forthcoming. Ray [to be added] Sah, Raaj Kumar, and Joseph E. Stiglitz, 1989, “Sources of Technological Divergence between Developed and Less Developed Countries,” In Guillermo Calvo, Ronald Findlay, Pentti Kouri, and Jorge Braga de Macedo, eds., Debt, Stabilizations and Development: Essays in Memory of Carlos Diaz-Alejandro. Cambridge, Mass.: Basil Blackwell. 25 Sahlins, M.D., 1965, “On the sociology of primitive exchange,” in: B. Banton, ed., The relevance of models for social anthropology (Tavistock Publications, London) 139-236. Scott, J.C., 1976, “The Moral Economy of the Peasant — Rebellion and Subsistence in Southeast Asia,” New Haven & London: Yale University Press. Stack, C. 1974, All our kin: Strategies for survival in a black community, New York: Harper & Row. Stiglitz, Joseph E., 1974, “Theories of Discrimination and Economic Policy.” In G. von Furstenberg, Ann R. Horowitz, and Bennett Harrison, eds., Patterns of Racial Discrimination, 5–26. Lexington, Mass.: D. C. Heath, Lexington Books. Tirole, Jean, 1996, “A Theory of Collective Reputations (with Applications to the Persistence of Corruption and to Firm Quality).” Review of Economic Studies 63 (1) pp: 1–22. Wolf, Charles, 1955, “Institutions and Economic Development,” American Economic Review, 45 (5). Pp. 867-883. Young, H.P., 1998, Individual Strategy and Social Structure: An Evolutionary Theory of Institutions, Princeton NJ: Princeton University Press. 26