Jump to ContentJump to Main Navigation
Social Security in Developing Countries$

Ehtisham Ahmad, Jean Drèze, John Hills, and Amartya Sen

Print publication date: 1991

Print ISBN-13: 9780198233008

Published to Oxford Scholarship Online: September 2011

DOI: 10.1093/acprof:oso/9780198233008.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 27 October 2020

Traditional Systems of Social Security and Hunger Insurance: Past Achievements and Modern Challenges

Traditional Systems of Social Security and Hunger Insurance: Past Achievements and Modern Challenges

(p.112) 4 Traditional Systems of Social Security and Hunger Insurance: Past Achievements and Modern Challenges*
Social Security in Developing Countries

Jean-Philippe Platteau

Oxford University Press

Abstract and Keywords

This chapter attempts to assess broadly the performance of traditional systems of social security as they exist or have existed in Third World village societies. It pursues the three following objectives: (1) to identify the main characteristics of the institutions providing social security in the societies under concern; (2) to bring into light the basic principles that their functioning obeys; and (3) to identify the most important problems and limitations to which they are subject. It achieves these objectives adopting the Scott–Popkin controversy as a convenient point of departure for the whole discussion. The chapter further provides a more detailed picture of apparently successful risk-pooling mechanisms as they have been found to prevail in several village societies. It evaluates the adequacy of traditional systems of social security in the light of modern challenges and present circumstances.

Keywords:   social security, Third World, Scott–Popkin controversy, risk-pooling mechanisms

This chapter is an attempt to assess broadly the performance of traditional systems of social security as they exist or have existed in Third World village societies. It does not aim at a full account of social-security institutions as they have been reported in the specialized literature (mainly, the anthropological writings). Rather, as a first step, the three following objectives will be pursued: (1) to identify the main characteristics of the institutions providing social security in the societies under concern; (2) to bring into light the basic principles which their functioning obeys; and (3) to identify the most important problems and limitations to which they are subject. These objectives will be achieved in the first two sections, adopting the Scott-Popkin controversy as a convenient point of departure for the whole discussion. In a second step, this chapter will provide a more detailed picture of apparently successful risk-pooling mechanisms as they have been found to prevail in several village societies. This will be done in Section 3. In a fourth section, the adequacy of traditional systems of social security will be evaluated in the light of modern challenges and present circumstances. Finally, a general conclusion will close the chapter.

1. The Scott-Popkin Controversy

The controversy between James Scott and Samuel Popkin which took place in the late 1970s is of direct relevance to our topic. Until then, the dominant approach to the study of Third World traditional (precolonial) village societies, especially among anthropologists and other social scientists, went basically unchallenged.1 This approach, known today as the ‘moral economy approach’ (from the title of the well-known book published by Scott in 1976), is grounded (p.113) upon the premiss that these societies are essentially geared to providing social and economic security for all their members. In an equally famous book published in 1979, Popkin strongly opposed the orthodox view and insisted that, on the contrary, social-security mechanisms are conspicuously absent in many precapitalist peasant communities. It is therefore useful to look a little more closely at the basic contentions of these two apparently irreconcilable views of traditional village institutions. This will provide a good starting point for the subsequent analysis and assessment of traditional social-security mechanisms.

1.1. The Moral Economy Approach

In the moral economy approach, precapitalist rural communities are viewed as societies in which social rights of minimum subsistence are secured to all members. The risk of hunger is insured against collectively and it is only under exceptionally adverse circumstances (like wars, epidemics, repeated crop failures) that traditional systems of social security may collapse and give way to social anarchy characterized by individual behaviour of the struggle-for-life type.2 A particular person may lose his membership and, thereby, his traditional claims to guaranteed subsistence if he breaks some customary rules of the society to which he belongs. However, since villagers are keen to avoid subsistence crises in an environment fraüght with serious uncertainties, and since it is always difficult to get accepted in other communities, open violation of customary laws is assumed to be rare.

According to this approach, social arrangements and economic institutions in traditional village societies have thus been especially designed to cope with the threat of hunger and other kinds of contingencies. It is in the light of this central objective that the high incidence of redistribution and reciprocity mechanisms (including patronage relations), to which anthropologists have given so much attention, must be accounted for. Market exchange is seen as a marginal form of transaction which is mostly confined to intertribal relations and to the exchange of goods with no special significance for physical or social survival. The idea here is that the functioning of the market could endanger the subsistence of individuals by reducing their food entitlements, and threaten the reproduction capacity of the group as a whole by promoting socioeconomic differentiation and class polarization beyond the control of traditional power structures. As a consequence, market forces cannot be allowed a free play inside the community space.3 In the words of Polanyi:

(p.114) No community intent on protecting the fount of solidarity between its members can allow latent hostility to develop around a matter as vital to animal existence and, therefore, capable of arousing as tense anxieties as food. Hence the universal banning of transactions of a gainful nature in regard to food and foodstuffs in primitive and archaic society. The very widely spread ban on higgling-haggling over victuals automatically removes price-making markets from the realm of early institutions

(1957: 255).

Within the community, a ‘subsistence ethic’ prevails to guarantee subsistence as a ‘moral claim’ or as a ‘social right’ to which every member is entitled. It finds its social expression ‘in the patterns of social control and reciprocity that structure daily conduct’ (Scott 1976: 32, 40). Notice that this ‘subsistence ethic’ is not tantamount to an egalitarian Utopia: ‘village egalitarianism in this sense is conservative not radical; it claims that all should have a place, a living, not that all should be equal’ (ibid. 40).

As has already been pointed out, intervillage or intertribal relations are the privileged domain of market exchange transactions. What must be added here is that these market relations are not of the usual type postulated in the standard Walras-Arrow-Debreu general equilibrium framework. Called ‘negative reciprocity’ by Service (1966), they are akin to the concept of ‘self-interest seeking with guile’ elaborated by Williamson or to that of ‘opportunistic behaviour’ commonly used in transaction-cost economics. Typically, such behaviour is not restrained by any moral code of honesty or fairness: instead, agents are allowed to seek their personal advantage by using any kind of means at their disposal, including ‘lying, stealing, and cheating’ as well as more ‘subtle forms of deceit’ (Williamson 1985: 47). This explains why, in precapitalist societies, trade is usually equated to stealing and pure trading relations are authorized only with outsiders or foreigners (Pospisil 1958: 127; Bourdieu 1980; 196–7). However, when relations with outsiders involve the exchange of useful objects required for subsistence, or when they are needed to cope with temporary food shortages, ‘the potential hostility which could break out into active aggression must be reduced, compensated for, and controlled’ (Belshaw 1965: 14). This is sometimes achieved through complex procedures such as the ‘ports of trade’ characteristic of the early empires (Polanyi et al. 1957; Polanyi 1968); the kula system discovered by Malinowski (1920, 1921, 1922) in the Trobriand islands; the practice of ‘silent trade’ observed in some regions of Equatorial and West Africa (Sahlins 1968; Giri 1983: 22–3); or the settling of farmer clients in the oases and in the desert fringes among several African pastoralist societies (Colson 1979: 23–4; Torry 1979: 520; Fleuret 1986: 227).4

(p.115) For our purpose, the following set of propositions can be drawn from the moral economy approach.

1. In precapitalist rural societies, villagers are constantly exposed to the threat of subsistence crises, which they have a compelling preoccupation to avert. As a result, such societies are largely organized around the problem of food contingencies and other subsistence hazards and they tend to act as guarantors of minimal subsistence for all their members.

2. In pursuance of this objective, traditional village societies aim at achieving a high degree of self-sufficiency in food and other essentials at the village level and even at the household level (unless there are strong scale economies like in the case of hunting).

3. To the extent that complete self-sufficiency cannot be achieved at the household level, or that household food supplies are subject to dangerous fluctuations, self-regulating markets are not considered as a reliable mechanism to ensure that food goes to the people who most need it. Other institutional arrangements (about which more will be said later) have been adopted that are supposedly more effective in providing food security to the local inhabitants. Such arrangements are backed by value systems and moral codes that emphasize the need for co-operation inside the community space.

4. If food exchanges within the community are insufficient to ensure adequate security for all the members, trade links are established with partners located outside the community space (as in the case of exchanges of grain against salt in West Africa). However, market forces are not allowed a free rein in that case either. Indeed, in so far as moral norms of good conduct for market behaviour cannot be shared among trading partners who belong to different communities (villages, clans, tribes, regions, and so on), and since securing access to food is of such paramount importance for survival, socio-political mechanisms are required to control the operation of market forces and ensure that exchange can actually take place in an orderly and predictable way.

1.2. The Political Economy Approach

The main point emphasized in the political economy approach advocated by Popkin (1979) is that traditional village institutions, arrangements, and norms ‘have neither been motivated nor been effective to guarantee the subsistence needs of community members’ (Hayami and Kikuchi 1981: 19). In the words of Popkin, ‘insurance, welfare, and subsistence guarantees within precapitalist villages are limited’ and ‘the calculations of peasants driven by motives of survival in a risky environment led not to subsistence floors and extensive village-wide insurance schemes, but to procedures that generated and enforced inequality within the village’ (Popkin 1979: 32–3). He further argues: ‘It does not follow from individual risk minimization or security maximization that (p.116) villages will function to minimize risk or maximize security for all.’ Indeed, ‘there are conflicts of interest, in addition to common interests, inherent within the village’, and ‘self-interest can lead also to coalitions organized to drive persons from the village or to deprive them of benefits’ (ibid. 44).

In fact, Popkin starts from the observation that household-level strategies for avoiding risk or averting starvation—what is sometimes called ‘self insurance’—are much more common than village- or community-level schemes designed for the sharing, pooling, and shifting of household risks. Thus, for instance, the household-level strategy of scattered fields ‘is a clear example of conflict between individual and group rationality whereby each individual, following a safety-first strategy, ends up with less production than he would if the village as a whole could follow an aggregate safety-first strategy’. As a matter of fact, ‘consolidated fields with higher average output and higher variance from year to year would be a better strategy for peasants to follow if the village could provide insurance for farmers to compensate for the increased variance of consolidated fields’ (Popkin 1979: 49–50, 105). According to Popkin, this would be a clear case where ‘the actions of individually rational peasants in both market and non-market situations do not aggregate to a "rational" village’ (ibid. 31).

The fact that village-level strategies for risk avoidance are not adopted in spite of their higher efficiency is taken by Popkin as sufficient evidence that the moral economy assumptions about the behaviour of peasants are simply incorrect. More specifically, Popkin finds fault with the hypothesis that peasants are either altruistic actors or passive subjects willing to respect social norms of conduct and moral principles of reciprocity. On the contrary, he contends that peasants in precapitalist societies are egoistic and hard-calculating agents intent on deriving maximum personal advantage from all actions in which they get involved. The opportunistic behaviour displayed by many villagers is bound to make collective action ineffective owing the the well-known free-rider problem; insurance or welfare schemes are among the collective goods which may thus be prevented from being produced (ibid. 24–5). As for social norms, they cannot be expected to mitigate the free-rider problem by instilling altruistic preoccupations into the people’s internalized value systems, or by holding their most dangerous opportunistic tendencies in check. Indeed, norms are regarded by Popkin as malleable, continuously renegotiable, and shifting ‘in accord with considerations of power and strategic interaction among individuals’. As a result, norms are often found to be inconsistent or conflictual so that they ‘cannot directly and simply determine actions’ (ibid. 22).

It is interesting to note that in his general approach to problems of collective action, Popkin is much more pessimistic than Mancur Olson. In actual fact, while Olson (1965, 1982) insists that co-operative strategies are much more likely to succeed in small than in large communities (since the costs of (p.117) organizing such strategies, including the costs of detecting and policing free-riding, are comparatively low in the former), Popkin appears to believe that, as a matter of principle, co-operative endeavours are bound to fail even in small (village) communities.

2. Towards a More Balanced Evaluation of Traditional Social Security Systems

2.1. General Considerations

There is little doubt that with the publication of The Rational Peasant, a gust of fresh wind entered into the mollycoddled world of anthropological ideas. Popkin’s book forced a debate on rather moot questions and challenged the idealized view of traditional village societies which so many anthropologists are inclined to accept. In particular, he has strongly emphasized the point recently made again by Robert Wade that ‘in the peasant context, even where all or most cultivators in a village could benefit from joint action, that action will by no means be automatically forthcoming’ (Wade 1988: 188; see also Lipton 1985).

This being said, two basic criticisms can be levelled against the intellectual contribution of Popkin. First, the opposition he sees between his own ideas and those of the moral economists is sometimes overdone in view of the numerous qualifying statements that some of these writers have made. To give only one example, Migdal himself has amply commented on the limitations which ‘mutual suspicion’ among the villagers creates for collective action, including insurance schemes; in the same line, he has emphasized household-level strategies of risk coverage (Migdal 1974: 74–8). The overdrawing of the boundary between the moral economy and the political economy approaches also arises from Popkin’s disregard of the obvious fact that selfish peasants can well adopt apparently altruistic behaviour (Posner 1981: 160; Hayami and Kikuchi 1981: 19; Booth and Sundrum 1985: 217). The problem with his approach in this respect is that he seems to overlook the fact that self-insurance is not the only way of coping with the risk of hunger for people who are irretrievably egoistic. As a matter of fact, selfishness does not preclude people from entering into relationships which can help all the parties concerned, or only the poorest among them, better to insure against that risk. This is particularly evident in the case of patron-client relationships as they have been analysed by Scott himself (1976: 35–44), and in the case of less hierarchical systems of insurance relations as will be shown at a later stage of this analysis.

Second, the view of traditional village societies which Popkin has put forward is no less partial and incomplete than the one advocated by the most radical proponents of the moral economy approach. For one thing, his judgement on the ability of these societies to elaborate village-wide insurance (p.118) schemes or insurance substitutes is much too negative. It is certainly correct to insist that traditional systems of social security and hunger insurance were imperfect and lacking in some important respects. Yet, they were far from being simply absent and Colson (1979: 27) is probably close to the truth when she writes that they were ‘moderately effective’ in the sense that they worked reasonably well in situations of moderate hardship. For another thing, the explanation offered by Popkin to account for their alleged or real failures is not wholly satisfactory, essentially because he has concentrated too much on the self-interested motivation of villagers, taking the extreme view (inferred from the Prisoners’ Dilemma) that rational people cannot achieve rational collective outcomes due to the insuperable difficulties of organization. By so doing, he has overlooked other, non-behavioural factors that are equally or even more important for understanding the conditions of collective action in the specific instance of insurance schemes.5

In Subsections 2.2. and 2.3. below, I will try to substantiate the two claims made in the second aforementioned criticism. This should allow us to assess properly the respective strengths and weaknesses of both the moral economy and the political economy approaches in regard to the issue analysed here.

2.2. Social Control over Essential Sources of Livelihood

2.2.1. The Institutional Build-up of Traditional Village Societies

In reviewing the central ideas of the moral economy approach, I have emphasized that, typically, market principles do not apply to intracommunity transactions in precapitalist village societies. As for intercommunity transactions, they are carried out under a system of trade which is either politically controlled (when the objects of exchange perform an important function for the physical or social survival of the trading partners’ societies), or allowed to operate in the most unbridled way. In both cases, trade is akin to a state of war (either disguised or open). It is therefore no wonder that the two institutional pillars of the market system—the rule of contract and private property—are conspicuously absent in the societies under concern. This is a crucial fact to bear in mind in any attempt to understand and evaluate traditional systems of social security.

Private property means that a person has exclusive and alienable rights over the things that he or she owns. This requires that ‘a sharp distinction be drawn between a thing and its owner’ (Gregory 1982: 18), and that the rights of use and control over given resources or goods be precisely defined and allocated while being at the same time amenable to effective protection through judicial procedures. The rule of contract presupposes that relations can be established (p.119) between agents, that is between parties who are free and independent in the sense of being fully emancipated from all kinds of non-economic (social, ritual, religious, political) constraints. In the words of one economist, ‘the identity of the parties to a transaction is treated as irrelevant’ (Williamson 1985: 69). Moreover, the nature of the agreement must be carefully delimited and remedies must be narrowly prescribed, so that ‘should the initial presentation fail to materialize because of nonperformance, the consequences are relatively predictable from the beginning and are not open-ended’ (Macneil 1978, cited by Williamson). Legal rules, formal documents, and self-liquidating transactions are thus central characteristics of classical contractual relations (Williamson 1985: 69). And the above institutional environment is congenial to the emergence and maintenance of anonymous, general, and abstract dependence relations that are typical of ideal market transactions or what Marx has called commodity exchange relations: indeed, ‘commodity exchange is an exchange of alienable things between transactors who are in a state of reciprocal independence’ (Gregory 1982: 12).

By contrast, traditional precapitalist societies are the domain of non-commodity (gift) exchange defined as ‘an exchange of inalienable things between transactors who are in a state of reciprocal dependence’. ‘Commodity exchange establishes a relationship between the objects exchanged, whereas gift exchange establishes a relationship between the subjects’ (ibid. 12, 19). As sociologists like Mauss and Bourdieu have stressed, the gift economy is a debt economy: on the one hand, the gift creates a debt that has to be repaid at some future (determinate or indeterminate) time, and, on the other hand, through gift-giving the donor aims at establishing a durable personal relationship with the donee (Mauss 1925: 9–42; Bourdieu 1980: 167–231). In Section 3, it will be shown that gift exchange can be at least partly interpreted as an insurance mechanism through which transactors attempt to cover various kinds of contingencies. In the light of this interpretation, Gregory’s observation that ‘the aim of a transactor in such an economy is to acquire as many gift-debtors as he possibly can and not to maximize profit’ (Gregory 1982: 19) can receive a straightforward explanation in terms of risk-spreading considerations.6

In precapitalist societies, considerations of status and fairness, as opposed to contract, played a dominant role. If considerations of status served the main purpose of maintaining social, political, and ritual stratification,7 those based (p.120) on fairness fulfilled the important function of safeguarding the interests of the poor and the unlucky. As a matter of fact, in contrast to modern legal practice which insists on clear-cut decisions and rigid formal rules, considerations of fairness allowed for maximum flexibility and continuous negotiation in which all sorts of peculiar circumstances could be taken into account. They therefore provided an indispensable cushion in a world fraught with so many uncertainties and contingencies that no formal rule or contract could ever have specified all the relevant conditions. In such a world, as some writers have pointed out, rules other than general moral principles such as mutual help or the recognition of everyone’s claim to continuous subsistence would be an ineffective means of policing, and would result in tremendous hardship for certain sections of the population. Only imprecision allows the unpredictable to be approached with appropriate flexibility (Scott 1976: chap. 2; Hayami and Kikuchi 1981: 16; Wolf 1982: 248; Wynne 1980: 41).

In this connection, it is worth pondering over the following description of legal practice in Aristotelian antiquity:

The obligation to keep one’s promise or to abide by a contract (whether explicit or implicit) was only a moral rule and not a formal law. It remained subject to the general principle of justice and fairness. A magistrate was thus entitled to release a transactor from his duty to meet his obligations if he was of the opinion that the terms of the contract were unfair. His appreciation of the degree of fairness of a contract was not made by reference to a priori legal criteria; rather, it was based upon his own assessment of the situation in the light of all relevant concrete circumstances. A just distribution—of assets, honours, privileges, statuses—was then considered as the touchstone of social order and peace. The main function of the magistrate was…to restore a ‘balanced’ situation while taking into account peculiar circumstances and customary norms

(Lepage 1985: 112–13, my translation).

It is difficult to accept Popkin’s contention that in precapitalist societies social norms and values are malleable and continuously shifting, at least as a general statement about such societies. By definition, moral principles are general and liable to various kinds of interpretation;8 but from this, as should be evident from the above analysis, it cannot be inferred that they are incapable of influencing or determining human actions. Clearly, Popkin’s conception of precapitalist societies as ‘amoral’ human groups, devoid of any stable value system and completely abandoned to the free play of selfish motives, is no solid ground for analysing behaviour and social interaction among their members. It can only be useful to apprehend precapitalist societies which are in an advanced state of social disintegration due to slave raids, wars, epidemics, persistent famine conditions, or which are going through what ren Lipton (1985) (p.121) has called ‘transition of trust’ due to capitalist penetration, national integration, and other forms of transformation of societal control.

The absence of genuine rights of private property in productive assets is a well-known feature of traditional village societies. It means that no single owner can claim exclusive property in those assets nor use them at discretion in whatever way he likes (he does not have the rights of usus, fructus, and abusus typical of the Western concept of private property). In particular, he is not entitled to dispose of them (to transfer them, to donate them, and so on) by an act of will: assets are not freely alienable and, therefore, they may not be ‘commoditized’. Note also that the non-exclusiveness of property rights typically manifests itself in the existence of various overlapping rights of use and control over the resources concerned. What needs to be stressed is that the foregoing characteristics hold true not only with respect to clan-based societies in which a communal (corporate) land-use system is in force but also with respect to more inegalitarian societies in which essential resources (like land and water) are concentrated in the hands of a small (feudal or quasi-feudal) elite. There is a saying among Kirghiz shepherds which illustrates perfectly the overlapping of the rights of use and control over the land in a strongly stratified society: ‘God has given us the land of the Khan.’ The statement is totally ambiguous because it recognizes at one and the same time that the land belongs to God (the ultimate owner in whom all property is vested according to Muslim belief), to the Khan (the lord), and to the common people.

The absence of exclusive and clear-cut, absolute and definite, private property rights in essential resources usually implies that customary rules and social control by the community play an important role in the allocation and use of these resources. An interesting explanation for this situation—especially when fields are widely scattered—is the need to overcome the complex problems created by pervasive production externalities, like those arising from the fact that croplands are transformed into pasture lands after the harvest (Dahlman 1980: 93–145; Hayami and Kikuchi 1981: 12–16; Wade 1988). In the present context, however, it is more relevant to concentrate on the social-security function which is performed by a mechanism designed to prevent the emergence of genuine private property rights and to restrict free land market transactions. First, in the case of tribal or clan-based village societies, such a mechanism allows for flexible adjustment in the allocation of rights of access to productive resources according to the subsistence needs of all the members. Thus, for example, a greater amount of land may be allotted to the households whose size has increased since the time lands were initially distributed within the village (and vice versa for households which have become smaller). Second, in the case of inegalitarian societies, the livelihood of the poor sections of the population is often protected by the very restrictions that have been imposed on the land use rights of the big landowning families. For instance, landowners are not entitled to evict the tenants working on their farm unless exceptional (p.122) circumstances can be invoked before the customary authorities and complex procedures are followed to settle the issue.

2.2.2. Guaranteed Access to Productive Assets

In many traditional village societies, guaranteed access to vital resources (particularly land) for every resident household is the main method through which people are protected from the risk of hunger. This is especially evident in the case of tribal societies, where land is usually held under corporate tenure with the result that it is subject to strong communal regulation. The broad principle underlying this land tenure system is that each household head belonging to the group is entitled to be allocated a sufficient amount of land to support his family.9Moreover, the allottee has possession and use of the lands as long as they are being cultivated, and his heirs would normally be given the lands that were cultivated at the time of his death (unless the rights of access are subject to a periodic rotation). In a more general way, customary rules and institutional arrangements regulate all important rights and duties of landholdings such as possession, inheritance, transfer, mortgage, size of operation, access to water, woodland or pasture use, and tenancy (Cohen 1980: 353).

Possession of land and other productive assets in such societies is personal and statutory in the sense that access to a portion of the communal resources is mediated through membership in a social group. The relation is reciprocal: on the one hand, group membership is the basis of social rights (particularly, of the most important among them, the right to subsistence), but, on the other hand, maintaining access to a share of the corporate productive assets serves to validate membership in the group (Berry 1984: 91). Members not only include those who claim descent, actual or putative, from other members, but also strangers and migrants who have been accepted as members of, and reside with, the group (Noronha 1985: 182). Therefore, access to land and other vital resources is not necessarily predicated upon kinship or descent-based ties, but may also be grounded upon loyalty and patronage relations which are often associated with ascriptive forms of status or social identity (Smith 1959: chaps. 1, 3, 4; Berry 1984: 91; Meillassoux 1980: 62).

Popkin is basically correct when he points out that the stratum of residents named ‘insiders’ usually enjoy more rights and benefits than the stratum of ‘outsiders’ (Popkin 1979: 43). Clearly, tribal, clan-based, societies are much less egalitarian than they have often been depicted. But this is not a point of contention between Popkin and most moral economists. On the other hand, Popkin makes another valid point when he stresses that, with increasing pressure on the available land, coalitions may be organized by the ‘insiders’ to (p.123) drive the ‘outsiders’ away from the village or to confine them to marginal lands: as a result, ‘the metaphor of corporate village as "collectivity" should be replaced by the metaphor of the corporate village as a "corporation"’ (ibid. 43–4, 46). In his extensive survey of land tenure systems in sub-Saharan Africa, Noronha (1985:182–3) has also reached the conclusion that, as land availability diminishes, ‘the circle of individuals who are entitled to access to land diminishes in two respects: membership is more narrowly defined in that increasingly, only those who can trace actual descent are entitled to land—the stranger being admitted more as a crop sharer or tenant or laborer without any right to land; and the type of land available for allocation to the newly-admitted member becomes increasingly marginal’ (see also Smith 1959: 52–8).

Regarding the latter point, what deserves to be emphasized is that village societies undergoing the process of exclusion and ‘corporatization’ mentioned above have been profoundly transformed through processes of agricultural commercialization and rapid population growth in modern times. As a consequence, they have entered a phase of dynamic disequilibrium which caused a gradual erosion of their traditional mechanisms of social security or a restriction of the collective rights of subsistence to an increasingly small number of rural dwellers or migrants. Among the ‘insiders’, however, it would be wrong to assume that rapid population growth always promotes social and economic polarization. There are numerous instances in which scarce assets have been increasingly divided and subdivided to safeguard the customary rights of access of the ‘insiders’ (whether within or outside the framework of extended families). In many cases, also, teams working on jointly managed resources have agreed to incorporate new members and to share with them the income therefrom, even though their marginal productivity was negligible or perhaps even negative (due, for example, to overcrowding). A vivid illustration of such an income-sharing through work-spreading practice will be presented later.

Guaranteeing access to land (and other vital resources) may not be sufficient to provide for the minimum needs of households in all circumstances. Collective risks or specific risks may threaten from time to time the livelihood of either many or a few inhabitants. To face these uncertainties, villagers must adopt a range of coping strategies to reduce risks, to share and pool them, or to adjust to their consequences once they have materialized. These strategies of risk management may be carried out at the level of the household (in which case the term ‘self-insurance’ is appropriate), or within a wider social group (which always holds true for risk-pooling insurance mechanisms). At this juncture, it is especially worth noting that land allocation arrangements geared to securing minimum subsistence for all the resident households usually comprise an important mechanism of risk reduction through which peasants can counter a good deal of the hazards of their climate and environment. This mechanism consists of granting them access to a variety of lands of different quality, location, and soil characteristics among which production risks can be spread. (p.124) For example, in areas of the Sahel where flood-recession agriculture is the mainstay of livelihood, peasants have possession and use of different plots of land located at varying heights along the slope of the drainage basin of a river. The variance of yields can thus be reduced since there is a spatial spread of farm plots in response to variability in rainfall so as to ensure that at least some of them will always be flooded.10 Another well-known example of spatial diversification of farm plots across heterogeneous agroclimates with non-perfectly correlated production risks is the traditional system of ‘ecological floors’ practised in the Andes and in the Himalayan region. Here, the scattering of agricultural fields takes into account altitudinal and latitudinal variations.11Note that dispersal of herds over a wide variety of pastures with different risk characteristics can be analysed in a way analogous to plot-scattering in agriculture (Colson 1979: 23; Torry 1987: 519).

It has been rightly emphasized that ‘spatial diversification of farm plots is a closer substitute for crop insurance than other informal means of risk adjustment’ (Walker and Jodha 1986: 25; see also Newbery 1989: 284–5). When it is viewed in conjunction with this possibility of holding a diversified portfolio of plots, communal control over the distribution and use of land assets therefore appears to be an effective mechanism of social security in traditional village societies. The security provided by this mechanism is, however, far from complete, particularly when village lands are more or less homogeneous. In these circumstances, other risk-coping devices are called for to cover the production uncertainties arising from weather and other vagaries (such as pest attacks).

Methods of self-insurance play an important role in this respect and they have actually received a lot of attention in the literature. They may pertain either to the production sphere or to the realm of household domestic management. To the former category belong such measures as crop diversification, intercropping, staggered planting, salvage crop planting, sequential decision-taking and adaptive flexibility in cropping patterns, diversification of seed varieties, selection of resistant, low-variance varieties, activity and livestock diversification, and so on.12 As for the second category, it comprises all measures designed for the accumulation of assets that can be depleted in case of crop failure or food shortage. The assets so accumulated may be of very different kinds, running from durable consumption goods to stored foods through cash holdings and livestock.13

(p.125) We have seen above that in agricultural tribal societies guaranteed access to land plots with varying risk and fertility characteristics together with self-insurance strategies normally allow all the member households to obtain a minimum income and to avert starvation over a full weather cycle. Likewise, in traditional village societies where the people’s livelihood crucially depends upon the exploitation of a common property resource subject to significant production hazards, social arrangements must ensure that access to the resource enables all the households to have a maximum chance of earning their livelihood. As a matter of fact, this requires not only that the latter have equal or fair opportunities of access to the resource in question (pasturelands, rivers, forests, and so on), but also that the probabilities of reaping the produce therefrom are more or less equalized among them, at least to the extent necessary for safeguarding their possibilities of survival. Precapitalist village societies can be ingenious in devising subsistence-oriented systems of regulation of access to vital productive resources. A fascinating illustration of both imaginative and organizational capacities is provided by the system of rotating access to the sea in the traditional beachseine fishermen communities of southern Sri Lanka (Alexander 1980: 97–102, and 1982: chap. 7; Amarasinghe 1988, 1989). A similar sea tenure system has also been observed in Turkey (Berkes 1986).

2.3. An Illustration: Beachseining in Sri Lanka

Beachseining is one of the most common fishing techniques used in traditional maritime communities. The beachseine is a large bag-shaped net with coir wings of extensive length. Its name is derived from the fact that it is operated from the shore itself, though it always requires the help of a boat to be put out at sea (so that its two wings can be properly paid out). It is also important to note that, owing to the very size of the fishing gear, beachseining requires a large area of operation; moreover, the sea bottom must be sandy and free from rocks and other obstructions so that the net can be dragged smoothly. These constraints create a situation of scarce fishing grounds in the sense that all the existing nets cannot be laid out as often as their owners would like. In fact, the number of existing nets is typically higher than the maximum number of nets which can be used in a day. However, even if it were not so and each net could be worked once in a day, rules of procedure would still be needed to govern the rights of access to the fishery. Indeed, all the nets cannot be cast simultaneously, and expected fishing incomes are not independent of the timing of the fishing operations on a given day. For one thing, most of the big catches tend to occur in the early morning and, for another thing, fish caught in the morning bring higher prices (because fish brought late to the market cannot easily find willing buyers). The (p.126) problem is further complicated by the fact that the probability distribution of the fish catches may differ significantly from one fishing spot to another.

To give every net equal chances not only in terms of access to the water but also in terms of access to catch and income opportunities, a rotating system has been devised, the rules of which enable the net owners to know when their turn is due. In the village of Gahavälla (Sri Lanka) studied by Alexander, the fishing area is divided into two stations: the harbour side (from which most big catches come) and the rock side. The net cycle begins on the harbour side and, after a net has had the dawn turn on that side, it is entitled to the dawn turn on the rock side on the next day. Subsequently, it may be used on the rock side each day once the net immediately following it in the sequence has been used. The sequence of net use over a period of five days, assuming that four nets can be used on each day on both fishing stations, is shown in Table 4.1.

Table 4.1. Sequence of net use in the village of Gahavälla


Fishing Station




















































Source: Adapted from Alexander (1982: 145).

As can be seen from the table, twelve nets have been used at least once in a period of five days. Furthermore, nets 5–8 have been worked a maximum number of five times during this period; nets 4 and 9 have been used four times; nets 3 and 10 three times; nets 2 and 11 twice; and, finally, nets 1 and 12 only once. If the total number of nets is twelve, net 1 will reappear in the harbour site on the sixth day (where it will have the last turn) and in the rock site on the tenth day (where it will have the first turn). As for net 5, it will have a new dawn turn in the harbour site on the thirteenth day: a complete net cycle lasts twelve days. Over the full net cycle, each net will have been operated eight times (the total number of possible turns per day in all the existing fishing stations), that is for two out of three days on an average.14 It is worth noting (p.127) that for each net, the period of use is strictly continuous: in the above example, once its first turn has come, a net will be worked during eight days in succession and, thereafter, it will be left idle for four days.

Additional difficulties arise from the fact that there are significant inter-seasonal variations in fish catches. Thus, in southern Sri Lanka, a sizeable portion of yearly incomes from fishing are obtained during a flush season which lasts only one month, and the exact timing of which can never be known in advance (Alexander 1982: 147). These difficulties cannot be serious as long as the total number of nets does not exceed by too large a margin the total number of possible turns (or hauls) per day (in all fishing stations): this should be evident from the foregoing example. However, if the total number of beach-seines rises significantly due to population pressure, increasing entanglement with market forces and rising fish prices, or any other reason, as the period of net use is strictly continuous the annual returns of any beachseine will be governed mainly by its position in the net cycle: most incomes will accrue to the nets which happen to have many turns during the flush period.15 This is exactly the situation which was observed in southern Sri Lanka from 1940 onwards: in Alexander’s area of study, for example, only twenty-five of the ninety-nine nets received turns during the flush period in 1970–1 (Alexander 1980: 105–7, 1982: 147, 203–8).

Nevertheless, the problem must be viewed in its right perspective. Indeed, it is over a single year that the customary means of regulating access to the sea tend to make the distribution of the catch more and more unequal as the number of nets increases: over a longer period, returns should be expected to even out since the sequencing of net-laying rights ensures that fishermen do not get their turn in the same month each year and the flush period tends to appear in a given part of the year. Thus, Alexander writes:‘a particular net’s turn is not tied to any point in the year, and over a four year period each net will be used in each month. As the fishermen are unable to predict the flush period with any great accuracy, each net has an equal chance of good catches each year’ (Alexander 1982: 147). Yet, this clarification should not be taken to mean that the multiplication of nets does not eventually threaten the survival of the fishermen households. There is of course the problem that additional nets may have a zero marginal productivity, with the result that a growing fishing population will have to be content with a constant product.16 But, even (p.128) assuming this problem away, there remains the question as to how year-to-year fluctuations in income will be buffered by the fishermen households. Clearly, risk-pooling and intertemporal redistribution mechanisms become increasingly necessary to compensate for the partial failure of the traditional method of risk management.

Two last remarks are in order. First, even though the beachseine represents a heavy capital investment, all fishermen can normally have access to it: indeed, each net is traditionally divided into a number of shares which, like stock-market shares, refer to the enterprise as a whole, not to particular portions of the net. Moreover, if a fisherman is too poor to purchase a share, he may be invited by an incomplete team of shareholders to join a net as a partner, in which case he is given credit till the first bumper catch occurs. The system of share ownership is also highly flexible inasmuch as a share of a net may be held by more than one person, as when the property of a share has to be divided among several siblings on the death of a family head (ibid. 142–3, 149). These are all important aspects of the system since labour cannot be separated from capital (‘ownership of a share carries the obligation to work the net when required’ and there are normally as many workers as shareholders on a given net) and access to the water is preconditioned upon the ownership of a net share .17 Easy access to capital therefore explains why the function of net shares as tickets of entry into the local fishery is compatible with the essential principle that membership in the village community (whether hereditary or acquired in a lifetime) involves a right of access to the community-controlled sea area.

Second, it is interesting to note that the rotating system of access to a natural resource has also been used in some traditional agricultural societies. Thus, in traditional village Ceylon, the customary practice was not for households to hold geographically diversified farm plots. Instead, it consisted of a radical form of communal land tenure which, according to Obeyesekere (1967), was based on the egalitarian ideology governing the concept of shares or ‘pangu’: ‘One has shares in the "gama" [a communal land] as a whole, hence one must have access through a period of years to the total area of land, ensuring an equitable distribution of both fertile and infertile land among the respective shareholders’ (quoted from Alexander 1982: 283). In Tamil Nadu, to take another example, there existed a land tenure system known as mirasi under which ‘ownership tended to be not of specific parts of the village but of shares’ and ‘the parts of the village to which the shares corresponded were periodically reallocated in a sort of lottery’. The lottery was conducted during a ceremony called Curray Edoo during which the several portions of land comprising the village corporate tenure were written on tickets and the villagers drew the plots which they were entitled to use until the time of the next Curray Edoo (Haggis (p.129) et al. 1986: 1446–7). What must be emphasized in the same way as has been done with respect to the use of the sea is the incompleteness of such arrangements from the standpoint of insurance: supplementary mechanisms of self-or collective insurance are therefore required to buffer the yearly income fluctuations which arise from shifting access to lands of varying quality.

2.2.4. Guaranteed Access to Work Opportunities

In strongly inegalitarian peasant communities (often called ‘class-based’ societies), social security cannot be ensured through guaranteeing direct access to or effective control of productive resources. Indeed, by the very definition of these societies, possession of the main means of production is concentrated in the hands of a small minority. In these circumstances, the livelihood of the members of the lower strata is often ensured by granting them access to the lands belonging to the higher strata. In the most common case (but surely not in all cases), decisions pertaining to the organization of production on the farm are left to the actual (or direct) cultivators of the soil and the latter’s rights of access are secure as long as the customary obligations imposed on them by the landlords, the chieftains, or the overlords (payment of the rental, services in kind, and so on) are duly honoured. Still more effective social security is provided to the poor by the rich if these obligations are not rigidly fixed, but are adjustable downwards whenever abrupt output shortfalls threaten the survival of the land tillers (a possibility heavily emphasized by Scott). The widespread system of share-cropping is evidently a mode of rent payment which partially insures the cultivators against the risk of bad harvests—‘partially’) because the product net of rent may still fall below the subsistence threshold when the crop failure turns out to be particularly severe. Nevertheless, share-cropping arrangements can prevent extreme hardships for the tillers if (good) landlords offer special rebates or simply cancel off the payment of the share in times of great adversity (see, for example, Neale 1957: 226; Robertson 1987: 42, 189). Moreover, and contrary to what is often believed, the fixed rent system itself does not necessarily put all the risk of crop failures on the shoulders of the tenants. Indeed, it has been observed not infrequently that the rents thus fixed were subject to downward reduction according to crop conditions (see, for example, Ishikawa 1975: 463, Scott 1985: 185).

In other instances, hunger insurance for assetless households takes the form of guaranteed access to work opportunities in the village. A well-known illustration of an institutional arrangement based on this principle is the bawon system found in Java. Under this system:

rice harvesting is a community activity in which all or most community members can participate and receive a certain share of output. When the crop is ripe, a horde of harvesters enter the field and harvest the paddy using the ani-ani to cut the stalks at the neck of the panicles. The harvested stalks are bundled and brought to the farmer’s house, where the harvesters receive a certain share (bawon) such as one sheaf out of (p.130) eight. Typically, harvesting is open to anybody. By tradition, the farmer cannot limit the number of harvesters who participate

(Hayami and Kikuchi 1981: 156).

The bawon system is thus a practice of work-and income-sharing that ensures food security for villagers with no or little control of agricultural resources. Note, however, that it falls short of offering them perfect hunger insurance since it uses a share system of labour payment. The risk of a crop failure is therefore shared between the landowner and the harvesters and there is good reason to believe that the latter are less able than the former to self-insure adequately against environmental hazards. Nowadays, interestingly, the bawon system is still in use even though its coverage is smaller than before and it increasingly takes on milder forms the evolution of which is often analogous to the ‘corporatis’ tendencies at work in tribal societies (see Section 2.2.2 above). A practice that has been frequently observed during the last decades is one in which harvesting is open only to people in the same village, and to outsiders only so long as the number of participants does not exceed a certain limit. A second system sets a limit to the total number of participants without restricting access to villagers. In a third system, participants are limited to those who received specific invitations from farmers (mainly relatives and neighbours); while in a fourth, still more exclusionary arrangement, known as ceblokan, eligible harvesters are only those who performed without pay such tasks as transplanting and weeding (ibid. 158–9; Kikuchi et al. 1984: 117–30; Hart 1986: 685–9). Clearly, the system of ‘shared poverty’ analysed by Geertz (1970) has limitations if only because the landholders do not allow the number of harvesters to rise above the point from which significant losses may occur due to labour overcrowding (marginal productivity of labourers then becomes negative) and to various moral hazard problems (Hayami and Kikuchi 1981: 157). On the other hand, dividing the economic pie into too tiny pieces under the pressure of rapid population growth may rapidly lead to a situation in which access to work can no longer prevent the labourers from sinking below the poverty line.

Mechanisms of output-sharing through work-spreading similar to the bawon system have been observed not only in agricultural communities outside Indonesia—as evidenced by the hunusan system in the coastal region of the Philippines—18 but also in non-agricultural societies such as maritime fishermen communities. Thus, among beachseine fishermen communities of Africa and Asia, a common practice consists of permitting all fishermen from the same community who wish to participate to join in the hauling operations and thereby claim access to a portion of the catch (personal field observations). (p.131) This customary rule explains why the number of participants in these operations may greatly exceed the economic optimum. Furthermore, as the hauling is carried out from the shore, it is a rather simple task which can be performed by disadvantaged people (like old fishermen without any adult son or male adults who fear the open sea). Finally, when the catch is low, net owners usually forgo their capital share (but not their labour share) so that the entire catch is divided (in kind) among those who have participated in the fishing operations. This latter practice has also been often observed in the case of other fishing techniques than beachseining, when nets are used from boats at sea: thus, in Malaysia, Raymond Firth has noted that the distribution of fish catches may become ‘definitely abnormal through special circumstances, such as a very small week’s yield’ (Firth 1966: 248). It is therefore evident that the risk-sharing mechanism underlying the share system of factor payment can be mitigated to the advantage of the labourers whenever their subsistence is at risk. This fact had already been pointed out earlier, in connection with agriculture (see above, this section).

Work being a traditional social right and duty in village societies (Bourdieu 1980: 198–9), it is only when no member of the household is fit for work that pure gifts are extended to the poor. In fishermen communities, for instance, children belonging to families in distress where there is no male adult breadwinner are usually entitled to take a few fish from every catch that is landed on the beach of their native village. In Sri Lanka, this customary practice is known as Raula Kapanawa (shaving the beard) (Amarasinghe 1988). In Senegal, the system is even more formalized since a certain share of the catch—known as the ndawal among the communities of lebou fishermen—must be handed over to the old and the poor persons in the (extended) family (Sow 1986: 12). In a more general way, the flexibility of the schemes of income distribution in the fishing sector often enables the capital owners to make allowances for special individual circumstances. Thus, according to Firth, a boat-owner or captain can give one man a bonus because of his old age or poverty, as a result of which he diminishes in effect either his own share or the shares of all the other crew workers. However, no objection is raised because this practice is ‘admitted by his crew on grounds of equity’ (Firth 1966: 257). In Senegal, an old custom consists of giving half-a-share (instead of a full share) of the catch to crew members who have been unable to go out fishing due to sickness (personal field observation).

Guaranteed access of poor households to productive resources or to work opportunities under share contracts of rent or labour payment does not constitute, as a matter of principle, a perfect way of protecting them against environmental and other uncertainties. However, it should be evident from the foregoing discussion that these arrangements often come close to an adequate, if not perfect, hunger insurance thanks to various mitigating devices whereby the richer, asset-owning households agree to shoulder a larger part of (p.132) the risk of income fluctuations in times of adversity or environmental disaster.19

This being said, to the extent that the above arrangements do not guarantee the poorer households a fixed income capable of sustaining their livelihood in all circumstances, they must be distinguished from what Shigeru Ishikawa has termed the ‘community principle of employment’. According to this principle, landlords feel a responsibility not only to supply work to all available labourers but also to ensure that their subsistence requirements can be met at a level set by prevailing social norms (Ishikawa 1975: 456). The most typical institutional arrangement evolved by precapitalist peasant societies to achieve this double objective is a highly personalized, multifaceted, and enduring employer-employee relationship known as the patron-client relationship.

Patronage relationships are usually understood as ties displaying the following characteristics: (1) they are highly asymmetrical; (2) they usually comprise, or are perceived as having, a strong element of affection which evokes the emotional tie between a father and his son; (3) they are comparatively stable (they typically apply for an indeterminate period of time) but it is only in a polar case that they are hereditary and that the clients are ‘bonded labourers’ or agrestic slaves; (4) they involve multiple facets of the actors concerned and imply a set of reciprocal obligations which stretch over a wide and loosely defined domain (see, for example, Pitt-Rivers 1954: 140; Wolf 1966: 17; Breman 1974: 16–23; Scott 1976: 167–92; Bardhan and Rudra 1978: 384; Bourdieu 1980: 226–9). The obligations incumbent upon the client consist essentially of being at the continuous disposal of his patron in order to help him in any circumstance for which the contribution of his labour is considered useful (participation in the cultivation of the patron’s land, support of his political career through canvassing, involvement in showdown events, and so on). In return, the client receives an insurance against all hazards that may imperil his subsistence, meaning that he can expect from his patron any kind of protection or assistance that he may need to maintain his livelihood. Note that the haziness of the intervention domains of both the patron and the client can be explained in a way that has been suggested earlier in a more general context (see above, Section 2.2): in a world fraught with many uncertainties and contingencies, no precise rule or contract could ever provide for all the circumstances which can affect the economic position of the parties concerned.

An interesting illustration of the polar case of hereditary patron-client relationships is provided by the ancient jajmani system of India. It is as a customary arrangement of labour relations providing hunger insurance to the (p.133) dependent castes that this system deserves our attention in the context of the present analysis. In this perspective, the following features need to be thrown into relief (see Gough 1960/1: 86–9; Epstein 1967: 230–3):

1. Social organization is highly asymmetrical with the members of the higher castes (the jajmans or peasant masters) holding more power and being much better endowed with land resources than the members of the lower castes (the kamans or untouchables). Moreover, the latter’s holdings are too small to ensure their livelihood, at least when adverse weather conditions prevail.

2. Labour relationships are established between two or more specific families of different castes in the same locality. These relationships are based on hereditary ties and are couched in kinship terms.

3. The peasant masters cannot use or dispose of their lands in whatever way they like and they may not, except under very special circumstances, abrogate their customary relationships with the labouring castes.

4. Reciprocal services and duties as well as the rewards associated with them are governed by customary rules. In particular, the rewards for the untouchables’ labour services on the farms of the peasant masters are paid annually in the form of fixed quantities of farming produce. They are therefore independent of the actual amounts of effort applied and of the output resulting therefrom (the contract is not of the risk-sharing type).

5. The annual wages of the untouchable labourers are fixed in such a way as to assure them of a minimum subsistence income in bad harvests. In addition, the hereditary relationships which they maintain with jajman families provide them with ‘benevolent’ masters expected to look after them as fathers do their children.

If we go by the empirical results obtained by Scarlett Epstein in a village of the Mysore area, two further features can be added to the above list. First, in bad years the total product of the village is distributed more or less equally among all households because this is the only way to keep the families of the lower castes (as well, of course, as the higher castes) alive. Second, in good years economic differentiation tends to occur on a large scale and the consumption level of the peasant masters rises much above that of the untouchables (there is no accumulation of capital). This is so because the former have prior claim on the labour of the latter at a time when labour becomes scarce due to peak demand arising from bumper crop conditions (bumper crops require more weeding and more harvesting). Furthermore, as is evident from characteristic 4 above, the untouchables’ annual wages are not adjusted upwards as a consequence of increasing tightness of local labour availabilities (Epstein 1967: 242–3). The untouchables are thereby assured of an adequate supply of food against the contingency of crop failure while the peasant masters are assured of a ready availability of labour when they most need it.

It is because of their weak bargaining position following from their (p.134) vulnerability in bad years—which, in turn, can be traced back to poor land endowments—that the members of the dependent castes are unable to withstand the pressures exercised by the large landowners to obtain their labour services in good years. As a result, they cannot work on their own (small) farms as much as they would like, although they would probably be in a position to earn a decent living from their lands provided they could apply enough labour to them. In fact, the income forsaken by an untouchable in good years—since his wage is kept at the same level as prevails in bad years when labour is not in scarce supply and the productivity of labour efforts is definitely lower—can be interpreted as the sum of two components: a compensation for the fact that the wage he receives from his patron(s) in bad years probably exceeds his marginal productivity and the risk premium he is willing to pay in order to insure against the risk of starvation at times of weather hardship.20

Emerging from the above discussion is the picture that, in ‘class-based’ traditional village societies, assetless households derive whatever social security they enjoy from situations of labour scarcity. Scarcity of labour combined with an unequal distribution of productive assets give rise to mutually advantageous insurance contracts (such as patronage relations). This picture is not complete, however, in so far as control of scarce labour in these societies is a source not only of economic well-being but also of social prestige and political power. As a consequence, to provide for the subsistence of a large number of clients (hereditary servants, small tenants, and so on) may prove to be an effective strategy even though the retinue of rich families may appear to be oversized in purely economic terms. In a fascinating study of the ‘agrarian origins of modern Japan’ (1959), Thomas Smith has shown that in the early Tokugawa period (seventeenth century), large families could not easily get rid of their old patronage relations even when steadily increasing numbers pressed dangerously on the family’s resources, thus reversing the situation of labour scarcity which initially gave rise to these relations. In the words of ren Smith (1959: 19—my italics),‘custom, pride, and public opinion would not permit the large family with ample resources to solve these problems as poor families did: by merely eliminating surplus members, cutting them off summarily from the resources and protection of the group’. The solution adopted by large and rich families at the top of the village hierarchy usually consisted of giving (cultivation rights to) a small piece of land to conjugal groups from one of the outer circles of the extended family (hereditary servants, remote family branches, and so on) so as to enable them to make a (meagre) living. Such newly formed households (called nago) did not become fully autonomous overnight. If labour services (p.135) continued to be periodically performed by them to the profit of the (extended) family headship, there was no relationship between the size of land allotments and amount of labour services (ibid. 24–8). On the other hand, the ancient master (called oyakata) remained the main source of hunger insurance for his nago:

he [the oyakata] helped nago survive the ever-recurring crop failures that were the marks of agricultural backwardness, and which hit everyone periodically but hit the weak most often and hardest. Custom and self-interest, both, obliged the oyakata to open his storehouse at such times, to provide his nago with food and seed until the next harvest lest he be thought pitiless for thus driving them onto the highways in search of sustenance and a more reliable protector. If that happened, it was usually because famine conditions prevailed and the oyakata was himself short of food

(ibid. 27; my italics).

It would therefore be wrong to take a purely economicist view of the contractual arrangements between rich and poor families in ancient village societies, even class-based ones. In particular, labour services, ‘far from being a means of payment for land . . . were clearly part of a far-reaching system of personal obligation’. Moreover, ritual acts expressing personal subordination were a common feature of these societies, in which they usually formed the basis of political power and social standing. In this connection, it is important to note that in Japan again, ‘the participants themselves made no consistent distinction between ceremonial and labour services, often calling both by the same name. . . . This is not because people failed to observe that one had economic value and the other not, but because this difference though evident was insignificant in view of the fact that the two had the same social character’ (ibid. 29–30).

2.3. The Insurance Dilemma in Traditional Village Societies

There are two well-known reasons why insurance (or quasi-insurance) schemes may be costly or may fail to emerge altogether: incentive problems and covariate risks (Rothschild and Stiglitz 1976; Binswanger 1986; Binswanger and Rosenzweig 1986; Newbery 1989). The first set of problems arise when economic agents are assumed to be opportunistic, an assumption clearly made by Popkin to account for the failure of most social security arrangements in precapitalist peasant societies. As for the second problem, Popkin has largely bypassed it. This is all the more curious as he has hinted at it in one long passage (Popkin 1979: 71–2). The fact that he did not push the argument further reflects his predilection for attributing failure to behavioural tendencies of the peasants: indeed, it was with respect to behavioural factors that he thought he could make his case against the moral economy approach.

Let us now look more carefully at each of the above issues in turn.

(p.136) 2.3.1. Incentive Problems and the Solution of the Village Community

Any kind of collective action—including social-security arrangements—is always under serious threat from incentive problems (Olson 1965, 1982). These problems actually arise when information is costly and asymmetrically distributed, and they usually take the form of moral hazard or adverse selection (Newbery and Stiglitz 1981: 165–6; Binswanger and Rosenzweig 1986: 507, 514–5; Newbery 1989: 278–9).21 If the costs at which moral hazard and adverse selection can be controlled are high for the insurer, the premiums charged by him may exceed the amount people are willing to pay to get insured: in the circumstances, insurance fails to emerge.22 Moreover, the more difficult it is to verify information on contingent claims or to monitor the actions of the insured agents, the greater the likelihood of a market failure.

In the specific context of traditional village societies, there are two factors which would tend to make insurance or quasi-insurance schemes apparently more difficult and more costly to operate. First, shortfalls of production from normal yields are frequent because, owing to the low level of technology development, production results are strongly exposed to climatic and environmental vagaries. As a result, the transaction costs of providing insurance (administrative and information costs) tend to be high. Second, almost by definition, traditional precapitalist societies are entities lacking elaborate record-keeping systems, formal laws and contracts, and judicial enforcement procedures. There is therefore a serious risk of incentive problems being altogether unmanageable: in so far as punitive procedures are weak or nonexistent (there are no written contracts with enforceable clauses), free-riding would be an almost irresistible temptation that could undermine all kinds of collective action.

However, the above discussion should not be taken to mean that collective action is impossible or highly unlikely in traditional village societies, as Popkin has contended. Indeed, there is an almost absolute need for co-operation in many of these societies because there are not only many uncertainties to be insured against but also numerous production externalities to be internalized. And, as argued by Robert Wade, when the net benefits from collective action are large, co-operation is likely to emerge and to be sustained voluntarily—that is, without selective inducements or punishments bearing decisive weight and without the supply of local leadership being a constraint in initiating and (p.137) sustaining co-operation endeavours (Wade 1988: 183–7, 205–10).23 In fact, the traditional village community can be viewed as an ingenious solution to the incentive problems in the form of a tightly-knit, highly integrated society characterized by intense and multi-stranded interpersonal relationships. Thus, Hayami and Kikuchi (1981: 21) consider it to be a more efficient institution than the market in an economy dominated by pervasive interdependence in production activities and high information costs. And even Popkin is not far from holding this position (although he carefully avoids making it explicit) when he writes that ‘peasant schemes generally tend to involve small groups’ (1979: 48).

The reasons why small communities or social groups help reduce incentive problems are easy to understand. As a matter of fact, the cost of information collection and contract enforcement is reduced to manageable levels when it is common to many transactions and when a lot of information can be held in common due to restricted privacy (Hayami and Kikuchi 1981: 14; Wynne 1980: 48–9; Posner 1981: 146–8; Platteau 1986: 186–7; Platteau and Abraham 1987: 468–9). Historical ties and continued personalized relationships tend to allow further reductions of information costs because ‘performances in past transactions comprise a reliable data set for prediction of future performance’ (Hayami and Kikuchi 1981: 14). Moreover, given the interlinked nature of many transactions and the lack of alternative possibilities (other village communities being themselves tightly-knit social entities with entry barriers), the cost of free-riding or rule-breaking tends to be so high that even implicit contracts or tacit commitments can be considered as more or less self-enforcing.24

The importance of customary rules, moral principles, and community norms must also be viewed against the background of incentive problems since they constitute a powerful means of assuring each participant that co-operation will ensue and that the obligations created (to return a loan, to pay a risk premium, to contribute labour or income to a collective scheme, and so on) will be enforced. It is in this sense that moral systems or codes of conduct— including internalized community values emphasizing altruism and respect of the hierarchy—can be considered as a substitute for more coercive enforcement of law systems or for complex schemes of selective inducements and punishments (Collard 1978; Lipton 1985: 81; Platteau forthcoming; Wade 1988: 203). Many economists—including Popkin himself—are used to supposing that every human action is reducible to a voluntary act of personal choice. Wynne’s reflection that this is hardly a satisfactory approach since (p.138) villagers are always immersed in a universe of community norms (Wynne 1980: 45–7) is worth pondering, and represents a full departure from Popkin’s superficial treatment of value systems (see above Section 2.2). According to Wynne, time- and resource-consuming activities like ceremonial rituals and myth-making were important activities in the course of which members of a given social group were appropriately socialized and identified with common systems of sympathy and belief. By helping to create reciprocal trust and an ‘emotional cement’ among them, these activities performed the function of ‘our contemporary systems of taxation, book-keeping, business, law, and government’ (Wynne 1980: 42–5). It is true that systems of values and beliefs can only be a partial substitute for institutional arrangements which give people the assurance that other participants will co-operate or will be punished if they do not comply with the rules. Some amount of coercion is always needed to back up agreements and to sustain collective action even when consensus has been sufficient to initiate it. In the words of Wade, ‘the rules must be backed by a system of punishment, the existence of which helps to assure any one person that if he follows the rules he will not be suckered, and which at times of crisis can directly deter’ (Wade 1988: 209). What needs to be stressed in the specific context of traditional village communities is that the cost of monitoring deviance and holding opportunism in check may not be too high because the sanctions usually take the form of social opprobrium or ostracism. As a matter of fact, due to their desire for social acceptance by a group and/or their fear of the material consequences of reputation loss, people are very concerned with social reputation and bounds within their community. As a result, ‘reputation in a small agricultural community is not lightly exposed to attack’ (ibid. 193). It is therefore the nature of the sanction system in these communities that helps explain why enforcement of promises and monitoring of rule-breaking are comparatively easy in spite of the fact that formal laws and judicial procedures are not available.

Even so, it would be wrong to assume that incentive problems were fully solved in precapitalist societies and that collective actions could always be undertaken at reasonable cost. Depending on historical circumstances, these societies achieved varying degrees of social cohesion and efficient leadership, and it is not hard to find evidence of societies affected by processes of social disintegration and unable to undertake group actions of any significance. In fact, even under normal circumstances, incentive problems can be kept under control only to a certain extent. The example of plot-scattering mentioned by Popkin (see above, Section 1.2) is perhaps a good illustration of the limits of collective action in many agricultural societies. Assume that in a given village land is of more or less uniform quality, yet, owing to the varying risk characteristics of different land locations, yields fluctuate significantly between farm plots. If there are economies of scale and if incentive problems could be handled at reasonable cost, it would then be a socially more efficient solution to (p.139) allocate compact areas of land among the various household units rather than allowing them to scatter their fields so as to diversify their risks. But this would precisely require that, after production has taken place, income would be redistributed from peasants endowed with lands which performed rather well during the ending agricultural season to those in the opposite situation. And it is probably here, in the sphere of redistribution, that the society would have to incur considerable transaction costs to carry the scheme through.25 However, in the case of insurance schemes, there is another cause of failure which can never be overlooked: it is to this cause that I shall now turn my attention.

2.3.2. Covariate Risks and the Insurance Dilemma

A fundamental theorem of the economic theory of risk and insurance is that the cost of an additional risk depends on its covariance with existing risks: the cost will be higher the stronger the degree of positive covariance, whilst negatively correlated risks will have the effect of reducing the total cost of risk-bearing. Consequently, there are potential gains from trade in risk when incomes or contingencies are uncorrelated while, if all agents face similar risks, risk cannot be reduced much by trading it between the participants (Newbery and Stiglitz 1981: 165; Newbery 1989: 270–2).26 Typically, farmers living in an ecologically uniform area and carrying out activities which are similar from a risk point of view have little to gain by sharing or pooling their risks. In these circumstances, an insurance arrangement would degenerate to a centralized reserve scheme and each farmer could actually self-insure at the same cost; as self-insurance through holding reserves avoids all information and incentive problems, it is actually preferred by the peasants (Binswanger 1986: 78). Therefore, the prevalence of household over village granaries or food stores cannot be taken, as Popkin has done, as decisive evidence that peasants are too opportunistic to make insurance schemes viable. A technical reason alone—the high covariance of yields—accounts for the difficulty of providing insurance against collective risks, that is, risks which affect all the participants simultaneously.

As has been pointed out by several authors, access to credit may provide an important substitute for insurance where insurance is absent or costly (Binswanger 1986: 79; Platteau and Abraham 1987; Eswaran and Kotwal 1989). (p.140) Thus, in situations involving uncertain income streams, consumption credit enables risk-pooling across time if people with good draws accept to lend to those with bad draws: in such a way, agents could stabilize their consumption streams without resorting to costly self-insurance methods like the hoarding of savings as a sort of contingency fund. Again, what needs to be stressed is that such an insurance substitute will not work if contingencies are correlated across the population. As a matter of fact, covariance of income risk leads to covariance of default risk: as a result, financial intermediaries would have to keep high reserve ratios (or require high collateral). Moreover, the incomes of both depositors and borrowers would also be correlated (Binswanger 1986: 79–80; Binswanger and Rosenzweig 1986: 516–17). Under the same conditions, co-operative credit schemes would fail for the same reason and only money-lenders lending out of equity funds (and able to reschedule loan repayments) would stay in business.

There is an obvious way of overcoming the problem of risk covariance: it consists of pooling risks over a wide geographical area so as to cover ecologically heterogeneous zones and economic activities which are complementary from a risk point of view. The difficulty with this solution is that it increases the incentive or information problems at the same time as it reduces the covariance problem: thus, the larger and geographically less concentrated the social group concerned in the insurance scheme, the lower the covariance of their incomes and contingencies is likely to be, but the more serious the moral hazard problem. The other side of this insurance dilemma is that while incentive problems tend to be less severe within restricted social groups where information asymmetries are comparatively small, such groups typically face a wide range of collective risks in so far as they live in localized areas with uniform ecological characteristics (Posner 1981: 156; Binswanger 1986: 78; Platteau and Abraham 1987: 468–9). Unfortunately, it cannot be taken for granted that a manageable trade-off always exists between the incentive and covariance problems: on a priori grounds, it is impossible to decide whether cost-effective insurance arrangements may be found that allow risk-pooling to take place in peasant societies. To put it in another way, one can never be sure that the costs imposed by the need to control moral hazard problems will be offset by the risk-pooling benefits, given covariate yields or incomes.

3. Traditional Risk-Pooling Mechanisms at Work

3.1. The Case of Weather-Unrelated Specific Risks

There is a positive lesson to be learned from the above discussion: risk-pooling or mutual insurance arrangements have more chances to emerge and to be (p.141) maintained successfully when there is a low covariance of risks within a highly integrated social group of limited size.27 As pointed out by Posner (1981: 154–5), the family is the institution most likely to satisfy the latter requirement but, at the same time, it is often too small, even in its extended form, to create an adequate risk pool for insurance purposes and, therefore, to satisfy the former requirement. One anthropologist has however contended that the size of the domestic production and consumption unit in precapitalist agricultural societies is precisely determined in such a way as to insure it adequately against the risk of illness and accident among its productive members (Meillassoux, 1980: 69). Incidentally, it may also be noted with Meillassoux that vertically extended households (that is, households composed of nuclear units of successive generations) can provide insurance for covariate risks such as area-wide crop failures. This possibility arises when older individuals in such households have accumulated assets which can serve as insurance substitutes to younger household members (ibid.; Binswanger et al. 1989: 131; Newbery 1989: 287; Von Braun and Webb 1989: 526).

Be that as it may, what can be asserted with confidence is that members of these societies have frequently formed informal village associations—remaining within or stretching beyond the bounds of the extended family network—to insure themselves against contingencies uncorrelated across the group thus formed. This is hardly surprising in so far as a number of weather-unrelated risks can be considered as specific risks that do not affect all the village households simultaneously. Risk-pooling groups for such uncertainties can be either themselves specific or encompass a number of contingencies. Examples of the former are traditional burial societies or fire associations. Burial societies provide help for the organization of funerals whenever a member family suffers a bereavement: in this case, the occurrence of the event is certain but its timing is uncertain. As for fire associations, they are insurance schemes whereby every participant commits himself to help any victim in the group to extinguish a fire in his fields or in his farmstead and, if needed, to rebuild the house after the disaster. As with most insurance schemes, members of village fire associations are willing to pay a certain cost of moderate amount—since, given the high average frequency of fire occurrence, they are sure of having to contribute their labour from time to time to come to the rescue of unfortunate victims in the group—in order to avoid incurring an uncertain loss of considerable size (if their own property catches fire). Similar to fire insurance societies are sea rescue associations through which small-scale fishermen communities insure collectively against the risk of being lost at sea. As in the previous case, the insurance scheme is implemented in a completely informal and decentralized way. Each fisherman participating in the scheme agrees to come to the rescue (p.142) of a fellow fisherman who has not returned home on the scheduled time, and to bear all the expenses related to the rescue operation (labour time, fuel, and so on). This he does in the expectation that in the future he may also find himself in difficulty out at sea and in need of rescue by other fishing units.

Risk-pooling groups with wider risk coverage are also very common. They often take the form of co-operative labour-exchange pools. The commitment of every member is then to come to the rescue of any fellow member who has fallen into distress owing to an event considered to be a misfortune: illness or accident of the farmer or of other productive members in the family; death or illness of draught animals or of livestock; crop damage due to trampling of fields by wild animals; destruction of the house or the fields by a flood or a fire, and so on. The understanding is that labour (or any other kind of help) provided under the scheme does not give rise to any formal and immediate payment or return (except for some food and drink). This does not mean, however, that it is free since it is actually offered against the conditional promise of reciprocal help at a future indeterminate time. The promise is conditional because today’s helpers will be entitled to claim reciprocity only if they themselves run into difficulties out of bad luck at some time in the future. Some insurance-like reciprocal arrangements have an even wider coverage, for example, the provision of minimal subsistence to widows. Thus, among the Siang Dyaks of central Borneo, ‘If a woman is left with several small children and has no relatives upon whom she can call, she is usually assisted by the others in the village, through gifts of rice and wild pig or by help in the clearing of her field; at least until such time as the children have become old enough to help her’ (Provinse 1955: 149). Again, those who bring assistance to unfortunate women are thereby assured that their own wife and children will be taken care of if they pass away and no close relatives remain alive in the village.

Contributions of the above kind can therefore be viewed as so many claims or lines of credit which are accumulated because they might always be mobilized in an emergency (see Belshaw 1965: 24). They are thus a close substitute or functional equivalent of our modern insurance policies. Furthermore, as with any insurance scheme, the ex post distribution of wealth is equalized since comparatively lucky community members extend more help than they receive while the opposite is true of unlucky members. And when ‘bad luck’ turns out to be an easy excuse for laziness or carelessness, the group reacts by refusing to extend any further help to the opportunist. For instance, we are told that, among the Siang Dyaks,

Occasionally help will be given unfortunate individuals by the other members of the village for which help no [immediate] repayment of any kind is expected by the donors; but such help is only tendered those whose misfortune has come through no misconduct on their own part and who are deserving of it . . . a man’s shiftless ways quite soon spread to all the long houses. . . . No really deserving person who through sickness or other misfortune has come to difficulty will be permitted to suffer or starve among the (p.143) Siangs, but an undeserving person is seldom tolerated longer than is necessary to find out what he is

(Provinse 1955: 147, 150).

As a matter of fact, it is precisely because informal insurance systems can be easily terminated if dissatisfaction arises, and also because being decentralized they do not involve the management of a common fund (which always turns out to be a highly complex and risky affair in village societies), that peasants seem to have a predilection for them all over the world. It is noteworthy that the above kind of informal, risk-pooling, labour-exchange co-operative groups have been observed not only in Third World precapitalist societies but also in contemporary rural societies in the northern hemisphere (Bennett 1968: 291–3).

3.2. The Case of Weather-Induced Specific Risks

To my knowledge, the literature offers only a few instances of insurance schemes in which yield risks are pooled in the context of traditional societies. Nevertheless, these examples are very illuminating and cover most nature-related activities: hunting-gathering, stock-raising, agriculture, marine fishing. In all the cases examined, as could be expected, the covariance of yields is rather low while moral hazard problems are under reasonable control. Moreover, as in the case of risk-pooling labour-exchange groups, the basic principle underlying these schemes is that of a collective disaster-avoidance strategy whereby the participants form a long-term ‘partnership reservoir’ which can be tapped in times of stress. Given the difficulties or costs of storing food (especially in moist climates where most foods get easily infested), the insurance-seeking agents make the following agreement: those who produce a harvest in excess of their consumption needs will give at least part of their surpluses to those who are short of food in exchange for the latter’s commitment to reciprocate if their respective positions are ever reversed (Posner 1981: 153; Carter 1987; Kimbalt 1988; Coate and Ravaillon 1989).

This kind of arrangement is often referred to as ‘generalized reciprocity’ in the anthropological literature and it is assumed to operate in the social framework of kinship groups (Service 1966; Sahlins 1968; Orlove 1977). As a leading anthropologist has remarked: ‘it is scarcity and not sufficiency that make people generous, since everybody is thereby insured against hunger’, and ‘in a community where everyone is likely to find himself in difficulties from time to time…he who is in need today receives help from him who may be in like need tomorrow’ (Evans-Pritchard 1940: 85). In the following pages, some important cases of reciprocity arrangements will be investigated, with special emphasis on their insurance aspects.

3.2.1. A Reinterpretation of the Potlatch System

Observed among the Kwakiutl Indians of the north-west coast of America, the potlatch system has traditionally (p.144) been viewed by anthropologists as a subtle way of forging or reinforcing alliances as well as achieving status and power through delayed and calibrated (though not necessarily balanced) gift exchanges conducted in the course of complex ceremonials (Boas 1920; Drucker 1955; Codere 1956, 1957; Belshaw 1965:24–6). However, Piddocke has contended that the above characterization is only valid as an account of the modern potlatch as it has been influenced by direct contact with Western civilization and modified under the impact of new economic opportunities. In aboriginal times, by contrast, ‘the potlatch had a very real pro-survival or subsistence function, serving to counter the effects of varying resource productivity by promoting exchanges of food from those groups enjoying a temporary surplus to those groups suffering a temporary deficit’ (Piddocke 1965: 244). The analysis of Piddocke (ibid. 244–64) can be summarized in the following way.

l. The basic unit of Kwakiutl society was the numaym which had a social control over given fishing locations and hunting territories. Several numayms formed a tribe.

2. The numayms suffered from time to time from serious food scarcities which threatened their existence. These scarcities were primarily due to considerable year-to-year variations in their environmental conditions, which got reflected in the actual numbers of fish and game available for food.

3. Variations in resource productivities across different numayms were largely independent of each other, because the numayms operated in a variety of heterogeneous micro-environments associated with specific hunting grounds and fishing stations on rivers.

4. The potlatch exchange always took place between the numayms represented by their chiefs, and it ensured ‘a continual movement of food from those groups enjoying a temporary abundance to those groups suffering privation’ (ibid. 249). In fact, food was not the only item offered in the course of potlatch transactions: durable consumer or capital goods (mainly blankets, copper, and canoes) were also important objects of exchange. However, it has to be emphasized that durables could be converted into food outside the potlatch sphere proper.

5. The giving numaym obtained social prestige in return for (and in proportion to) the goods supplied to the receiving numaym(s), thereby enhancing its rank position in the rivalrous competition for numaym leadership within the tribe or the region.

6. Since variations in the productivity of their resource base tended to be intermittent and short term, the direction of gift-giving was always liable to be reversed so that the long-term relative ranking of the numayms tended to remain more or less constant.

3.2.2. Insurance-motivated gift exchanges in Southern Africa

The hxaro system practised by the Kung San hunter-gatherers of the Kalahari (north-west (p.145) Botswana, north-east Namibia, and south-east Angola) is based on a network of kinship ties (real or fictive blood kin) which spread the risk of drought-induced hunger over an area wide enough to avoid the covariance problem. This is possible because ‘localized rainfall and biotic conditions in this semiarid territory vary resource productivity even within areas no larger than 50 square miles, checkering San country with pockets of dearth and sufficiency in relatively dry years’ (Torry 1987: 325; see also Wiessner 1982: 65). Moreover, the Kung are a highly mobile people who can travel over a very large region of several thousand square kilometres (Cashdan 1985: 457). As we have seen above, the situation among the Kwakiutl Indians was different in this respect. Geographical mobility was much more restricted but, fortunately, environmental variations were considerable even within a comparatively small stretch of territory with the result that there was a low correlation of incomes between the various social groups.

The hxaro system can be described as a structured means of hunger insurance or, more precisely, as ‘a social method of pooling risk through storage of social obligations’: according to this practice, ‘a person creates relationships of mutual reciprocity with others in the population and thereby spreads losses over a unit much larger and more varied than the local band’ (Wiessner 1982: 65, 77).28 Therefore, ‘hxaro partnerships are not economic contracts with set terms, but rather bonds of mutual help’, the loose terms of which allow the coverage of a wide variety of needs and contingencies (ibid. 68). The insurance motivation of these relationships is evident from the fact that ‘a person who has given assistance has no desire for an immediate and fixed return that would even the relationship and make it possible for it to be cancelled. Rather, the aim is to store the debt until the situation of have and have not is reversed’ (ibid. 67).

Hxaro transactions obey a number of rules or procedures that can be inferred from Wiessner’s careful empirical study of the Kung San (see also Torry 1987: 325–8). To begin with, there are two procedures to control problems of moral hazard. In the first place, hxaro relationships must be built up and maintained with great care using a well-established ceremonial. Thus, a person initiates a hxaro by giving a gift to a prospective partner, and the latter can express his willingness to enter the proposed relationship only by returning a gift of similar worth. However, it is only after a trial period of at least one year, during which several gifts of non-food items flow between the partners, that the relationship is eventually considered firm. Thereafter, gift exchanges which are both balanced and delayed (it would be an insult to return a gift immediately) must continually take place to inform the partners that the relationship is still intact, a practice that is all the more important the longer the distance between the (p.146) camps in which they live, if they live far apart. These continual flows of ceremonial gifts serve the important function of cementing strong bonds of friendship backed by a powerful ideology or moral code of generosity and equality. Note that, once a relationship is firm, a person is said to ‘haľ the partner in his heart, meaning literally to hold and figuratively to be responsible for that person: ‘from then on, each partner can call on the other in times of need’ (Wiessner 1982: 66–7).

In the second place, hxaro relationships are established among consanguineous relatives or persons who have been remembered over the years as kindred members: only the hearts of these people are considered to be known enough to do hxaro with them appropriately. Furthermore, conflicts are more easily amenable to a settlement when family ties exist between the contending partners. Indeed, ‘if a quarrel does arise over hxaro between consanguines, the Kung feel that their common relatives will unite and try to resolve the conflict, while if one arose between affinal relatives, that each person’s respective kin would side with their relative and a serious fight would arise between the two kindreds’ (ibid. 66). In other words, choosing one’s hxaro partners in the family network has the effect of reducing transaction costs and of keeping moral hazard problems within manageable limits. Yet, owing to long distances and asymmetrical information, transaction costs can never be negligible and ‘considerable time is spent in trying to establish who has and who is in need’ (ibid. 68). Wiessner has indeed observed that families tend to be discreet about the food they have gathered or could gather in the bush. Moreover, ‘it is not unusual to hear a person with a full belly complain that he has not had anything to eat and is "dying of hunger"’ (ibid.)

Another principle underlying hxaro relations is that the Kung San ought to hold a diversified portfolio of hxaro partners. In the specific context of the hunter-gatherers of the Kalahari, this strategy amounts to spreading one’s partners geographically so that the risk of hunger is well distributed over the population of partners. As a matter of fact, individual hxaro networks cross many camps but partnership density and frequencies of exchange fall off with distance: approximately 70 per cent of the Kung’s partners live within a radius of 50 km from their house, while the remaining 30 per cent reside in more distant areas between 50 and 200 km away. Furthermore, there are very few Kung who do not have at least one partner between 150 and 200 km away (ibid. 75, table 3.2 and 76). In times of hardship, a family begins by visiting its hxaro partners who live in adjacent areas; and it is only if their economic fortunes turn out to be affected by the same environmental failure that it will trek further afield to seek food help in more distant bands. The key role of hxaro is precisely to provide an individual with an alternative residence when he can no more make a living from the resources located in his own area of landrights. The rules of hxaro assistance have been described by Wiessner as follows:

(p.147) Hxaro relationships with people in different areas allow individuals and their families to make extended visits to a partner’s camp lasting from two weeks to two years. While living there, they will have access to the resources of the area and the partner will integrate them into the reciprocal relations within the camp. For the first few days the visitors will be supported, but after that they will be expected to hunt and gather their own living and share any surplus with their host and others as they would in their own camp

(ibid. 68).

Third, the rules governing the hxaro system allow for a good deal of flexibility through indirect hxaro relationships. Indeed, hxaro networks extend the bond of a partnership in a given camp to the bandmates with whom the partner has reciprocal relations. Of course, the latter do not feel any direct obligation to the hxaro partner of their fellow camp resident (usually a close relative). Nevertheless, they are willing to include him in daily interactions of reciprocity within the camp when he comes to visit. And, conversely, they know his assistance may possibly be called on for their own benefit through the intermediary of his (direct) hxaro partner in the camp. In so far as adjacent areas of landrights within a particular camp are known for localized plant and animal foods, additional local contacts serve to open up access to a wider range of resources nearby and indirect hxaro relationships help further spread the risk of hunger for any individual (ibid. 62, 69–70, 76).

Finally, there is a good deal of stability in the hxaro system inasmuch as hxaro relationships tend to be reproduced from generation to generation when they have worked to the satisfaction of both parties: ‘Hxaro relationships are passed on in a family in a way which permits those who reciprocated well to be continued and those who did not to be quietly dropped and forgotten and still new ones added to meet changing needs’ (ibid. 77). Note also that camp membership is rather stable with the result that ‘when persons form relationships of reciprocity with other adults, they do so with a good idea of where and with whom their partners will be living’ (ibid. 64).

To the extent that misfortune does not strike all equally and that the hxaro system works effectively as a risk-pooling mechanism, income is redistributed from the lucky to the unlucky households. This goes a long way towards explaining why the Kung San society has been able to maintain its strongly egalitarian pattern of wealth distribution (ibid. 67, 69).

Unlike the Kalahari Kung San, the Nata River Basarwa (Bushmen) inhabit a relatively well-watered area and today they depend primarily on cultivated crops and domestic livestock for their livelihood. However, like their desert-dwelling neighbours, they still rely to a considerable extent on reciprocal exchanges as a way of ensuring their long-term food security (Cashdan 1985: 454–74).

To understand the insurance function of reciprocity among the Basarwa, a number of their production conditions must be highlighted. First, it is important to bear in mind that most of them are rather poor people who do not (p.148) own much livestock of their own, but receive cattle on credit from some wealthy owner (the so-called mafisa arrangement). In return for the labour expended on the livestock care, they are usually given a calf yearly and, moreover, they get the milk and the draft power of the ‘loaned’ animals. This last aspect is essential since cattle are indispensable to plough a field of reasonable size (ibid. 462; Schapera 1970: 115; Robertson 1987: 156–7), Second, the quantity of grain harvested per household depends crucially on the size of the area planted, while the latter is positively correlated with the number of years the field has been ploughed. This last fact results from the time-consuming nature of clearing and fencing tasks (Cashdan 1985: 466–7). Third, because the cattle owners frequently take their livestock away after a few years, the Basarwa are forced to move their compounds periodically to join the cattle post of a new employer (ibid. 468). Besides, as they do not have contracts of specified duration, they never know in advance when their next move will take place.

Unfortunately, Cashdan is rather obscure on certain important aspects of the role of reciprocity as a hunger insurance mechanism among the Basarwa (particularly with respect to its ability to buffer environmental variations). The following is an attempt to reconstruct the logic of the system from the information and data she provides.

As a result of their frequent changes of residence, the Basarwa face important periodic variations in food production even irrespective of weather and other environmental hazards. These variations are at least partly unpredictable, in so far as the Basarwa do not know the time when they will have to stop cultivating a comparatively large field in a given location to start with a small field in a new location. For the same reason (great mobility) and because of the high cost of food transportation, they do not engage in large-scale food storage to cope with the risk of having to move. At the same time, however, one must expect the variance of farm size, and consequently of farm productivity, to be large within a given location since the period of residence varies greatly across the different resident households. Therefore, the comparatively big farmers can share their surplus with the small ones who have recently settled in the area. Of course, the probability that the latter will face a food deficit is greater when rainfall is poor and drought sets in. Yet, even in conditions of collective environmental hardship, the sharing mechanism ought to be workable. Indeed, for one thing, their resource being larger, big farmers are likely to achieve more satisfactory (absolute) levels of agricultural output than small ones; and, for another thing, since bigger farmers have stayed longer in the area, they would presumably have stored some food surplus during the previous years which they can make available to those who did not have sufficient time to self-insure adequately.

Of course, for the system to be practicable, the bigger farmers must have an incentive to give their surplus to those in deficit. That this is precisely the case follows from the continuous movement of each household across the size- (p.149) distribution of farms. In the words of Cashdan: ‘because people who have stayed the longest at a location will change from year to year, the Basarwa households that plough the largest fields in any one year are not likely to remain forever in that category’ (ibid. 468). As a consequence, the same household will not be always on the ‘giving’ side of the gift transactions, and the sharing of food surpluses can be grounded on the mutuality principle. It should be noted that this mutual insurance system can work satisfactorily only if the reciprocity network stretches across a wide range of locations within the Basarwa migration area (that is, along the Nata River). In this context of high mobility over rather long distances, ethnic identity serves the essential function of cementing ties among the migrant households so that incentive problems can be avoided or minimized.

To sum up, there are two important causes of variation in food production among the Basarwa: erratic rainfall patterns and frequent changes of residence arising from stock-raising activities. The solution adopted to cope with these risks is a combination of limited food storage and reciprocal sharing of food surpluses. Revealingly, the hoarding of food is much more important and reciprocal exchanges much less prevalent among other ethnic groups who happen to own moderately large herds and can therefore settle in a given location on a more or less permanent basis. In their case, owing to the high covariance of farm yields and the constancy of the size-distribution of farms, mutual insurance arrangements cannot be used as a means of coping with environmental hazards.

3.2.3. Hunger Insurance through Reciprocal Credit in a Traditional Fishing Community

The economic organization of traditional maritime communities is particularly interesting not only because fishermen are subject to considerable day-to-day fluctuations of their catches but also because the correlation of the outcomes of their fishing trips is notably low.29 Thus, from an unpublished in-depth study of a small sample of households in a South Indian fishing village (the village of Poovar in South Kerala) I found that the coefficient of variation of daily catch values greatly exceeded unity in most cases (either on a monthly or on a yearly basis). On the other hand, the covariance of fishing incomes between various pairs of fishermen turned out to be slightly positive (up to a maximum value of 0.15), close to zero, or even negative (data pertaining to 1981–2).

The same study also reached the conclusion that the behaviour of Poovar fishermen is clearly motivated by their desire to stabilize food consumption in the face of highly uncertain income streams. Thus, the elasticities of current (daily) consumption expenditures with respect to income appear to be in the low range of 0.0004 0.008 (after correcting for autocorrelation through the (p.150) Cochrane-Orcutt method). In actual fact, most of the (comparatively slight) day-to-day variations of recurrent consumption expenditures appear to be accountable in terms of seasonal movements of income and daily fluctuations of the price of necessities. As most of these fishermen live close to the margin of subsistence (Poovar is a poor village due to adverse ecological conditions), these price fluctuations tend to get translated into proportionate variations of nominal (recurring) consumption expenditures.

In order to buffer abrupt variations in their daily catches, the fishermen of Poovar resort to several devices or strategies: adjustment in the timing of nonrecurrent (consumption or capital) expenditures which are not subject to strict timing constraints; movements in cash holdings; and changes in the household’s net indebtedness position through the use of consumption credit facilities. The former two strategies can be considered as self-insurance mechanisms. The latter really belongs to the category of mutual insurance arrangements and, as such, deserves our attention in the context of the present discussion.

What needs to be stressed first is that financial intermediaries specializing in day-to-day consumption credit are not likely to emerge in this situation and to develop the corresponding market. As a matter of fact, there are good grounds to believe that such agents would not be interested in transactions which are so frequent and involve such tiny sums that administrative costs per unit of money lent would be prohibitively high. In these circumstances, credit-givers have every chance to go bankrupt whether or not they choose to transfer the burden of their high operating expenses to their customers. This is obvious if they try to bear this burden themselves. On the other hand, if they pass it on to the borrowers, the latter would be asked to pay interest charges so high that they would probably prove unable to meet the financial obligations (payment of interests and amortization of the principal) they have willingly contracted: in the first instance, the demand side of the market does not exist, while in the second instance the supply side vanishes because the providers of credit cannot stay in business.

The fishermen of Poovar have, however, responded to these difficulties and overcome a potential market failure by evolving a decentralized system of consumption loans through which current receipts on the beach are regularly redistributed so as to provide a time-pattern of expenditures much more even than that of income. As was argued above, consumption credit can serve the role of partial insurance due to the intertemporal nature of risk-bearing provided that the uncertain income streams are uncorrelated across the population concerned. What is remarkable about the system of (subsistence) consumption credit in Poovar is that it really functions like a (partial) insurance market and thereby gives fishermen the advantage of maximum intertemporal flexibility in their consumption choices. In fact, this system bears close resemblance to the gift-exchange system of the anthropologists, since it is (p.151) based on ‘the ethical imperatives of reciprocal obligations’ (Belshaw 1965: 11). Hence the name ‘reciprocal credit’ which has been given it elsewhere (Platteau and Abraham 1987: 467).

The insurance function of reciprocal credit is evident from certain specific aspects of the implicit contract between the lender and the borrower and from the particular role of the community or the larger social group. Thus, it is noteworthy that by accepting a reciprocal loan a fisherman implicitly recognizes that he will be concerned with the future economic fortune of his creditor. Such a commitment implies that in the case that the creditor falls into distress, the borrower will not only have to return his debt immediately, but also that he must be ready to come to the help of his benefactor even if he has already succeeded in paying back his initial loan. Conversely, if the debtor again finds himself on the brink of a subsistence crisis, the creditor is expected to come to his rescue irrespective of whether or not he has cleared his first debt. A link of solidarity is thereby forged between the two parties which arises from their mutual desire to insure themselves against the risk of a catch failure (ibid. 467–8).

It is clear from this account that there is typically no stipulated date for the repayment of reciprocal loans: the tacit rule is that the borrower will clear his debt as soon as he is in a position to do so. Moreover, the fact that reciprocal loans are given free of any formal interest does not mean that they carry no price: the price a borrower has to pay consists of a conditional promise to help his creditor should their economic fortunes be reversed in the future. In other words, ‘the borrower does not end his relation with the lender by repaying his debt as would happen with standard (contractual) types of credit’ (ibid. 470). Conversely, the interest forgone by the credit-giver can be interpreted as the functional equivalent of the risk premium he is willing to pay in order to insure himself against future (production) contingencies.

This system of reciprocal credit is clearly decentralized since lenders and borrowers enter into direct contact with each other without using the services of any specialized agent. A crucial feature of this mechanism is that the fishermen are well informed about their respective economic situations, as a result of which there is no important information gap that a financial intermediary could bridge. This characteristic follows from the fact that the fishermen involved in a given reciprocal credit network live and work very close to each other, all the more so if they have family ties, belong to the same crew, or co-operate in some way or other in the spheres of production and/or marketing (usually through their wives in the latter case). To the extent that incentive problems arise from costly and asymmetrical information, they are not likely to pose a major threat to the functioning of this mutual insurance scheme.

It would, however, be incorrect to assume that things work so well that no role is left for a monitoring or rule-enforcing structure, and that the market for reciprocal credit reduces itself to a number of dyadic relationships between (p.152) pairs of fishermen who are willing to insure each other against the risk of falling below minimum earnings. There are, in fact, two essential functions which the community of Poovar—or, more exactly, any social group corresponding to a network of reciprocal credit relationships—performs to make the operation of the consumption credit market more effective. The first of these functions is to remedy situations where, owing to insufficient diversification of their financial claims, creditors are unable to obtain help from their debtors when they find themselves under stress. In such circumstances, the community may seek third parties ready to advance loans either to the embarrassed creditors or to the insolvent debtors so that they can clear their former debts.

A second useful function performed by the local community consists precisely of controlling the moral hazard problems that are still liable to arise either because, good though it may seem, information is not perfect, or because coercion turns out to be necessary to enforce some promises. Thus, for example, in the kind of problematic situations mentioned above, the community may first check whether the failing debtor is really unable to repay his debt to a creditor in distress and, if not, it will bring social pressure on him to meet his outstanding obligations (ibid. 468). The simple threat of social sanctions is usually sufficient to make the potential free-riders comply. Indeed, fishermen normally have grave fears of losing their social reputation, not only because they are eager to avoid being involved in conflicts with a high degree of public visibility, but also (and mainly) because harmful material consequences may follow from such loss in terms of employment or credit opportunities forgone. As a matter of fact, there is no doubt that owner-employers do not like to take on ‘unreliable’ workers in their crew, and that credit-givers are reluctant to advance money to persons who have been potential or actual defaulters in the past. Moreover, the fear of losing these opportunities is all the more effective as a check on opportunistic behaviour because debtors cannot easily run away from their native village (due to lack of alternative economic possibilities), with the consequence that any free-riding or violation of the implicit social code entails heavy social and economic costs (ibid. 468–9). Be that as it may, one can say that in Poovar as in the villages studied by Robert Wade, the problem appears to be ‘less detection [of free-riding] than resolving conflicts between those seen to be cheating and those who see themselves as harmed by their cheating’ (Wade 1988: 193).

Quantitative evidence shows that, in Poovar, reciprocal credit transactions play a major role in day-to-day adjustments to income fluctuations, especially among poorer fishermen. In other words, together with adjustments in the timing of non-recurrent expenditures (whenever this is possible) and, above all, with movements in cash holdings, reciprocal credit turns out to be an important way of stabilizing subsistence consumption throughout the year. Table 4.2 below provides a striking illustration of the crucial role which consumption credit can play in smoothing the impact of variations of fish (p.153) (p.154) catches. It presents the daily variations in wealth (see column 1, obtained by subtracting all consumption expenditures from current income) and in financial assets (see column 2, obtained by adding loans given and debts repaid and subtracting loans recovered and debts contracted) for an assetless household of Poovar during the month of October 1981. In addition, the table displays the amounts of non-recurrent expenditures incurred on each day of the same month (note that these amounts have been subtracted from income in calculating the figures shown in the first column). Since the household concerned does not incur capital expenditures (it does not own capital assets), daily movements in cash holdings can be directly obtained from the table by subtracting column 2 from column 1. Finally, it is noticeable that all credit transactions have involved small amounts (varying from a minimum of Rs 5 to a maximum of Rs 25).

Table 4.2. Daily Movements in the total wealth position and in the financial assets owned by a crew member household of Poovar (October 1981)


Variations in wealth (current savings)

Variations in financial assets

Amount of non-recurrent expenditures





− 23.75

− 25



+ 4.25

+ 10



+ 7.75

+ 10



− 20.25

− 20



− 19.75

− 20



− 1.0




+ 14.0

+ 10



− 8.75

− 10



− 16.50

− 10a



+ 5.0

+ 5a



− 20.25

− 20



− 1.5

− 5



− 11.0

− 15



+ 2.0




− 11.5

− 10



− 9.0

− 10



+ 3.75

+ 5a



− 17.0

− 20



− 16.5

− 10



+ 4.0




− 15.75

− 10



− 15.75

− 15



− 1.5

− 5a



+ 10.75

+ 15



− 19.25

− 20



− 0.5




− 9.0

− 10



− 5.25

− 5



− 3.0

− 5a



− 3.5




− 8.75

− 10



Rs− 207.5

− 200


(a) This is the net outcome of simultaneous transactions running into opposite directions, typically a borrowal coupled with the repayment of a debt incurred in the past.

(b) Sundays.

Source: Adapted from original field data collected by Anita Abraham during the years 1980–81.

A number of interesting observations can be made on the basis of this empirical material. First, there is no systematic pattern by which nonrecurrent expenditures are adjusted in accordance with income fluctuations: this results from the fact that not all of them have a flexible timing. Second, daily variations in financial assets are not only frequent but they always have the same sign as the daily wealth fluctuations: the household’s net indebtedness position worsens when current expenditures exceed income (when the household dissaves) and it improves in the opposite case (when the household saves part of its income). Third, movements in cash holdings tend to occur when wealth variations are of comparatively low amounts: the household replenishes its cash holdings when it has a small surplus (saving) and depletes them when it has a small deficit (dissaving). Conversely, the household tends to give reciprocal loans or to return reciprocal debts when it has a sizeable enough surplus, while it tries to borrow in the reciprocal credit market or to recover past loans when it has a sizeable enough deficit. Fourth, over the whole month, the household’s wealth variations have been translated into corresponding variations of its financial assets, which means that the level of its cash holdings has not varied significantly between the beginning and the end of the month. During the month under consideration, the household’s net dissaving amounted to Rs 207.5, while its net indebtedness increased by Rs 200.

3.3. Concluding Comments on the Scott-Popkin Controversy

In the next section, an attempt will be made to assess the performance of traditional social-security systems and to highlight a number of dynamic issues related to them. In the process, the main points emerging from the above lengthy discussion will be summarized. Before embarking on this next step of the analysis, it is useful to make a few concluding comments on the Scott-Popkin controversy. Perhaps the most important conclusion in this respect is that both authors have somewhat gone astray by seeing the problem of the (p.155) ‘moral economy’ as concerned only with the motivations of people in traditional village societies. Thus, Scott made the mistake of confusing social-security arrangements with altruistic behaviour or, at least, he seems to have read in such arrangements the prevalence of a pervasive ‘subsistence ethics’ in these societies. As for Popkin, he made the mistake of thinking that he had refuted Scott’s thesis simply by arguing that people are basically selfish and that norms and values do not influence human behaviour. The fact of the matter is that the ‘moral economy’ of traditional rural communities can be largely interpreted as a ‘social-security’ economy in that people belonging to these communities are often eager to find collective methods to protect themselves against major contingencies and production hazards. In so far as these methods or mechanisms have proven to be workable, their success ought to be ascribed both to self-interested behaviour on the part of individuals and to ruling customs and norms that are designed to ensure their continuity and to control major incentive problems. On the other hand, the very fact that collective actions for social protection and hunger insurance happen to fail, or that some socially efficient solutions are not adopted, does not necessarily mean that selfish behaviour tends to ruin or block important attempts at better protecting people against major risks. Covariate risks and incomes seriously limit collective hunger insurance possibilities. Human institutions are determined by material constraints as well as by behavioural characteristics, which is not to say that the latter determinants should be downplayed.

4. Traditional Social-Security Systems: Past Achievements and Modern Challenges

Any inquiry into the concrete socio-economic organization of traditional (precapitalist) rural societies can only lead to more scientific humility, owing to the bewildering variety of the organizational forms encountered. The dangers of drawing general conclusions from a limited number of particular case-studies must always be borne in mind. In the present chapter, I have tried to reduce these dangers by considering a wide range of situations while striving to identify and assess traditional arrangements in the field of social security and hunger insurance. A number of important lessons appear to emerge from this analysis.

To start with, and given due consideration to their whole institutional buildup, traditional village communities can be portrayed as societies which are essentially geared to securing minimum subsistence for all their members. The variety of rules governing the people’s right of access to food is large and the particular approaches to social protection adopted by a given society are partly determined by historical circumstances (including the history of class or caste relations) and partly by environmental factors (natural constraints, population (p.156) density, characteristics of production risks, available technology, and so on). Two main types of social institutions aimed at the reduction of the risk of hunger can be distinguished. In the first place, there are all the collective rules and mechanisms that govern the distribution of available productive assets or work opportunities in such a way that every household in the community has a reliable guarantee of survival over the full weather cycle. At one extreme, one finds kin-based societies (for example, descent groups of a corporate kind such as the many African patrilineage systems) in which a relatively egalitarian access to natural resources is guaranteed to each member. And, at the other extreme, there are the strongly differentiated agrarian societies where members of the lower strata usually enjoy guaranteed access to employment on the lands of the rich. It is noteworthy that, in all cases, traditional social-security mechanisms entail major departures from private ownership and market exchange. In one way or another, social controls are always exercised over the access to, and the use of, key productive resources. Of course, this is mostly evident in the case of village communal resources which play a vital role in ensuring the livelihood of the weaker sections of the population.

The second important type of collective methods of risk management in peasant societies consists of reciprocity networks or gift-exchange arrangements through which villagers pool their risks in order to reduce the total cost they represent. These mutual insurance schemes can operate through the exchange of labour, grain, durables, or cash, and it is rather obvious that individuals need not be altruistic to participate in them (even though that could help reduce moral hazard problems). Reciprocity networks are a special kind of economic transaction in which, as anthropologists have untiredly emphasized, economic and social considerations are deeply intermingled. Interestingly, once gift exchange is interpreted as an insurance arrangement (at least in its form of ‘generalized reciprocity’), this very characteristic appears to perform an essential function, namely that of tying the transactors together in such a way that they become mutually responsible for whatever misfortune may strike them.

Even though empirical evidence is scanty (but not altogether absent), the case can reasonably be made that, barring exceptionally unfavourable circumstances (such as repeated crop failures or crop diseases affecting entire communities), traditional methods for controlling the risk of falling into distress have usually enabled the people to counter natural and other hazards in a rather effective way. The well-established fact that people were very keen to avoid being expelled from their social group and that expulsion was viewed as the gravest possible sanction in any traditional village society suggests that the social protection it afforded its members was of utmost significance. This being said, it is also important to emphasize that traditional ways of coping with the risk of hunger were far from being perfect or complete social protection devices. Thus, one of their most glaring limitations arose from the fact that (p.157) subsistence crises almost always had to be confronted within the narrow boundaries of a restricted local or regional economic space. There can be little quarrel with the claim that, thanks to the development of extended communication networks, efficient transportation means, and extensive redistribution mechanisms, the modern world is better equipped to prevent the kind of major collective disasters that traditional societies could hardly counter. In other words, at least with respect to such disasters, traditional arrangements were a second-best optimum given the absence of generalized food exchanges and the low level of development of the communication infrastructure.

There are other limitations to traditional practices of hunger insurance. Take the first type of institution mentioned above. A basic prerequisite for the effectiveness of asset-sharing or work-spreading mechanisms as hunger insurance devices is that productive resources or employment opportunities be sufficiently numerous in relation to the population. If population pressure becomes too severe, sharing methods may cease to guarantee the livelihood of most of the people. The main problem here is that, beyond a certain point, diminishing incomes resulting from population growth foster centrifugal tendencies even within rather egalitarian societies, and cause struggles over the distribution of productive resources and employment opportunities to intensify. Shifts in the institutions governing income and resource distribution may then occur which are likely to reduce the economic security of many individuals or social groups. Changes in property rights immediately come to mind. Indeed, it is a well-known fact that, under the impact of population growth (and increased commercialization of agriculture), markets tend to emerge to allocate the scarce goods or production factors available (Boserup 1965, 1981; Rosenzweig et al. 1988; Binswanger et al. 1989). In China, such a situation already arose by the middle of the eighteenth century, when reserves of suitable land started to run out, causing the gradual emergence of private rights in land. The state even encouraged reclamation of forest areas by granting legal ownership titles to cultivators once they had paid taxes on the land (Jones 1981: 216). In India, genuine property rights in the soil came much later, during the period of British rule. Yet, as recent historical research has tended to show, quasi-markets of land revenue-collecting rights developed during the Mughal period (such rights became inheritable, saleable, and leasable, possibly with prior approval of state or village authorities), sometimes several centuries before British rule came to be established (Habib 1963: 154–80; Fukuzawa 1982: 250–1; Kumar 1983: 210).30 In most of Africa, however, the emergence of private property rights in land is a very recent phenomenon (Noronha 1985; Pingali et al. 1987; Platteau 1990 and forthcoming).

Social differentiation, polarization, and marginalization processes are the normal consequences of the above evolution (see, for example, Ishikawa 1975; (p.158) Scott 1972, 1976, 1985; Hayami and Kikuchi 1981; Watts 1983, 1984). The gradual deprivation of ‘outsiders’ of their customary rights of access to village lands, and of tenants or share-croppers of their traditional land-use rights, are oft-quoted examples of such processes. Even the archetypical ‘poverty-sharing’ villages of Java do not escape this general tendency, as attested by the fact that exclusionary labour arrangements have multiplied since the late 1960s (Collier et al 1974; White 1976, 1979; Hart 1986). In south Sri Lanka, the system of income-sharing through work-spreading that resulted from guaranteed rotating access to the sea collapsed during the 1950s in some villages and even before in other villages. As a matter of fact, the average share of each net in the total catch had become so small due to the overmultiplication of nets that ‘ownership of a single share was insufficient to sustain life’ (Alexander 1982: 206). A cumulative kind of involution resembling a ‘tragedy of the commons’ was initiated from then on, since to protect their income existing participants had to construct additional nets. This was with a view to preventing the number of their turns per unit of time and their chance of participating in the flush period from declining too much. Many fishermen who could not afford to finance new shares were forced to sell their existing shares to more affluent participants and to leave the fishery (ibid, chaps. 9, 10, 11). In one village, a social upheaval followed during which many small shareholders organized themselves to urge the government to limit the number of seines allowed to operate in their fishing zone (Amarasinghe 1989).

Population growth is not the only factor which may cause the erosion of traditional social-security mechanisms. As ‘moral economists’ and Marxist scholars have emphasized, market penetration and the gradual rise of a modern State system are equally powerful factors. In the above example of beachseining in south Sri Lanka, these factors have also been at work, especially since the Second World War when a dramatic increase in fish prices occurred which resulted from important developments in fish marketing and from growing integration of fisheries into the national economy. The new economic opportunities thus created attracted a class of wealthy capitalists and traders who usually belonged to the landed elite and had tight connections with the political establishment and strong positions in local state institutions (Amarasinghe 1988: 168–9). In a very short time, this elite gained control of most of the beachseine fisheries.

The problem is essentially that by opening up new and varied avenues for social and economic mobility, the growth of the market tends to encourage the overt expression of (perhaps latent) individualistic propensities among the people, to dissolve old co-operative ties and to disentangle the individuals’ interests from those of the social group (Smith 1959: chaps. 10, 11, 12; Robertson 1987: 196–7). Deeper and deeper penetration of market values has the effect of loosening the web of traditional social relations: people become more and more free of group pressure to conform and less and less concerned (p.159) about the well-being of the extended family or social group. Furthermore, the rise of a centralized modern State alongside the expansion of the domain of market forces tends to undermine old systems of authority. A ‘transition of trust’ problem arises both because community norms are weakened in response to market incentives that push personal advantage considerations to the foreground, and because locally based authority or consensus systems are gradually eroded by codified laws and new political structures (Lipton 1985).

Another difficulty with the first type of institution for social protection concerns inegalitarian societies. Indeed, one may justifiably wonder if the price paid by assetless households to guarantee their livelihood by entering into patronage relations is not too high. This question deserves serious pondering, especially so when the (feudal) landlord class appears to hold enormous power and privileges. Thus, Jones has noted that in India ‘the caste system might be said to have provided job security and a form of insurance, but the price was high’ (Jones 1981: 193). It is clear that social and economic inequality is particularly unbearable if the landlord, warlord, or bureaucratic classes are rapacious and oppress the peasantry without regard for its ability to make ends meet. Thus, in Mughal India: ‘The peasants were left in destitution by taxes which were collected whatever the state of the harvest. They received no real help in the face of natural disasters, and these were frequent (there was for example a great run of famines between 1540 and 1670 when the empire was at peace‘) (ibid.: 197; see also Popkin 1979 with reference to nineteenth-century Vietnam). Therefore, as underlined in a recent critique of the moral economy approach, ‘the possibility that there were "immoral" societies introduces an important qualification into Scott’s characterization of pre-colonial peasant life’ (Haggis et al. 1986: 1450).31

Turning now to traditional risk-pooling mechanisms, one should note two fundamental limitations which are in fact typical of all insurance schemes: the incentive and the covariance problems. What needs special emphasis here is that in traditional societies lacking modern record-keeping systems and nationwide law enforcement procedures, these two problems are interdependent, giving rise to an ‘insurance dilemma’: a reduction of the incentive problem is likely to make the covariance problem more intractable, and, vice versa, the incentive problem becomes harder to control when a wider geographical and social area is covered to diminish the correlation of risk across the various participant households. As a consequence, the scope for effective risk-pooling can, in many instances, be quite small.

Furthermore, incentive problems tend to become much more difficult to monitor when the prospects for shared values and trust or for an effective local (p.160) authority structure are undermined by the rise of a centralized modern State (colonial or post-colonial) and by the penetration of capitalism into the economic domain of the village (see above). If problems of societal control are serious during an inevitable transition phase, it is due to the fact that ‘the new authority has not yet acquired the information, the power to reward or penalise, or the powers of moral suasion, permitting it to replace the. old authority’s lost powers’ (Lipton 1985: 51). A void is thereby created which is likely to result in increased transaction costs of detection (of frauds), negotiation, and enforcement.

Finally, there exists an additional difficulty arising from the specific form of risk-pooling arrangements in these societies. Contrary to standard insurance schemes, reciprocity networks tend to discourage capital accumulation. As a matter of fact, while an insurance premium is fixed, and there is therefore no disincentive to accumulate, with a norm of ‘sharing’ and generalized reciprocity, ‘the amount of the "premium" is directly related to one’s ability to give, and individuals have less incentive to produce more than average’ (Cashdan 1985: 456). In the words of Polly Wiessner: ‘Limiting work effort over the long run can result in a lower-than-possible mean income in exchange for reaffirming a strong hold in social relations necessary for reducing the variance around the mean’ (Wiessner 1982: 79). Thus, the limits on time which the Kung of the Kalahari desert put into the search for food is partially the result of their fear of being exploited in reciprocal relations whose terms are that the one who has gives to the one who needs. Conversely, spending time gathering more information about the economic position of partners and trying to collect food from them may be considered a better strategy than putting greater effort into hunting or growing food (ibid. 68, 78–9).

Disincentive effects resulting from specific forms of hunger insurance and social protection may not represent major problems in traditional village societies. Indeed, these societies obey a reproduction logic designed to ensure their continued subsistence as social groups, and not a logic of capital accumulation. In this context, individual accumulation of capital is normally kept under control by the society lest it lead to economic differentiation and thereby create tensions in the social and political fabric. In other words, there is great apprehension that accumulation of private productive assets may disturb the fragile social equilibrium upon which the group’s collective subsistence crucially depends. Therefore, only certain forms and a certain extent of private (and, of course, collective) capital accumulation are allowed and these often serve the function of diversifying risks (such as the accumulation of cattle in many semi-arid countries, particularly for individuals belonging to the upper strata), or enhancing the group’s social security (such as symbolic investments aimed at reinforcing co-operation ties among the members).

By contrast, in contemporary Third World countries confronted with the challenge of rapid economic growth (not only to absorb increasing numbers (p.161) but also to raise the people’s current standard of living), the disincentive effects of traditional risk-pooling arrangements on capital accumulation and technical change may be strongly felt and generate serious tensions among as well as within individuals. Thus, Wiessner points out that, having developed increasing interest in material possessions, ‘many Kung are really torn between the desire to accumulate goods and the desire to remain within a secure system of mutual help’ (ibid. 82; see also Ortiz 1967: 210–13).

The same kind of difficulties—formulated in terms of static or dynamic efficiency considerations—could actually be mentioned with regard to all traditional arrangements aimed at stabilizing access to key productive resources and at preventing the emergence of markets in order to insure community members against the risk of hunger. For example, David Newbery writes (1989: 286): ‘If land were communally held and the harvest were divided among all those eligible, risks would be evenly shared but there would be little individual incentive to increase output and accumulate wealth.’ Such conflicts between institutions that provide insurance benefits and those that encourage efficiency are in fact quite numerous. The question as to how property and use rights in vital resources (for example, land) can be specified and protected without involving too high a price in terms of lost opportunities for economic growth is probably one of the most difficult policy issues on the agenda of development economists. This is so because efficiency, equity, and insurance considerations need to be taken into account in a simultaneous way.

5. Conclusion

For policy purposes, four important conclusions emerge from the preceding analysis.

First, social-security systems did exist in numerous forms in precapitalist village societies. At the same time, they were far from being perfect hunger insurance mechanisms and, in many instances, they could probably be considered as second best optima given the many constraints confronting these societies: high weather-dependence of many income sources due to low technology development, shortages of roads and other communication means, high income and risk correlation across households, lack of record-keeping systems, and so on.

Second, many traditional village support systems entail high costs in terms of dynamic efficiency, because of their disincentive effects on work and investment efforts. If they were all to be maintained in the future, social security would be bought at the price of capital accumulation and economic growth.32 This is true not only with respect to systems of rights of access and to (p.162) risk-pooling arrangements, but also with respect to law practices. Indeed, ancient laws were built on the notions of fairness and social status, and they took explicit account of personal situations and particular circumstances while defining responsibilities, interpreting agreements or commitments, and assigning liabilities. However, if highly flexible legal rules may be a good way of protecting the weaker sections of the population against serious contingencies, they are also liable to hinder economic growth. In actual fact, economic transactions are not likely to multiply, and division of labour is not likely to increase and to give rise to more or less continuous productivity gains, if business dealings appear too risky to willing transactors because of the lack of predictability of their consequences. As the history of advanced countries has taught us, explicit contracts, formal guarantees, and carefully delimited agreements which bind the parties without regard to their personal situations serve the function of minimizing the risks involved in many economic transactions (see, for example, North and Thomas 1973; North 1981).

Third, there has been a gradual erosion or weakening of many traditional systems of social security due to the joint impact of market penetration, population growth, and the rise of a modern State system. Nevertheless, contrary to what James Scott and other ‘moral economists’ have suggested, this is far from being a systematic or universal tendency. In a number of cases, such systems seem to have been strengthened or intensified. Thus, for example, when the labour market is tight, when productive operations must be done in a timely way (a constraint from which new agricultural technologies are not exempt) or require a large dose of husbandry skill (as in the case of farm animals or mechanical irrigation devices), employers tend to put a high premium on the ready availability and reliability of labour. Patronage may then be conceived of by the employers as a proper way to establish trust and to maintain stable relations with their workers. The latter may thereby get a guarantee of more or less continuous assistance in times of trouble through which the former aim at cementing labour-tying arrangements and securing long-term commitments of labour (Bardhan 1979, 1980, 1983; Bardhan and Rudra 1980; Platteau 1984; Binswanger and Rosenzweig 1984). The rise of industrial paternalism in nineteenth-century European factories could perhaps be interpreted along these lines. To take another example, the increased risk of being lost out at sea arising from the mechanization of fishing techniques has prompted the fishermen in many Third World countries to evolve or to strengthen informal, decentralized, mutual insurance schemes for sea rescue operations. Often, these schemes prove to be superior to more centralized (p.163) mechanisms involving the management of a common fund by a rather bureau-cratized structure. That social-security mechanisms may still play an important role today is evidenced by the experience of Sub-Saharan Africa where powerful ‘horizontal’ solidarity networks allow many people to survive in a context of chronic economic crisis (Bayart 1989: 270; for Asia, and India in particular, see Das Gupta 1987).

Fourth, in the light of the above three points but with the foregoing qualifying statement in mind, one can agree with Gilbert that ‘the continued usefulness of traditional systems as the major source of social protection’ has become ‘highly problematic’ (Gilbert 1976: 365). One can think of a number of circumstances in which these systems need to be replaced by new social-security institutions. Such circumstances arise when traditional methods or arrangements have been largely undermined under the impact of new forces (see the preceding point); when they are rather ineffective in achieving their aims; when they are inadequate to cope with new types of insecurities emerging in the wake of development (such as those resulting from the use of new agricultural technologies which are embodied in costly inputs acquired through the market); or when they are not compatible with an acceptable measure of economic growth in the rural sector. What alternative social-security systems could be is a complex issue that clearly lies beyond the scope of the present analysis. While market forces and institutions may help in reducing the risk of hunger, particularly in so far as they increase self-reliance by diversifying sources of income and supply as well as market outlets,33 they also open the way for new sources of vulnerability. The need for carefully thought public policy measures, discussed at length in other contributions to this book, is an inescapable aspect of the challenge of social security in the modern world.

(p.164) References

Bibliography references:

Alexander, P. (1980), ‘Sea Tenure in Southern Sri Lanka’, in Maritime Adaptations: Essays on Contemporary Fishing Communities (ed. A. Spoehr). University of Pittsburgh Press, 91-lll.

——— (1982), Sri Lankan Fishermen: Rural Capitalism and Peasant Society. Canberra: Australian National University Press.

Amarasinghe, O. (1988), ‘The Impact of Market Penetration, Technological Change, and State Intervention on Production Relations in Maritime Fishermen Communities: A Case Study of South Sri Lanka’. Ph.D. thesis, Faculté des Sciences Economiques et Sociales, Namur.

——— (1989), ‘Technical Change, Transformation of Risks and Patronage Relations in a Fishing Community of South Sri Lanka’, Development and Change, 20/4.

Bardhan, P. (1979), ‘Wages and Unemployment in a Poor Agrarian Economy: A Theoretical and Empirical Analysis’, Journal of Political Economy, 87/3.

——— (1980), ‘Interlocking Factor Markets and Agrarian Development: A Review of Issues’, Oxford Economic Papers, 32/1.

——— (1983), ‘Labor-Tying in a Poor Agrarian Economy: A Theoretical and Empirical Analysis’, Quarterly Journal of Economics, 98/3.

——— and Rudra, A. (1978), ‘Interlinkage of Land, Labour and Credit Relations: An Analysis of Village Survey Data in East India’, Economic and Political Weekly, 13/6–7.

——— (1980), ‘Types of Labour Attachment in Agriculture: Results of a Survey in West Bengal’, Economic and Political Weekly, 15/35.

Bayart, J. F. (1989), L’Etat en Afrique—La politique du ventre. Paris: Fayard.

Beals, R. L. (1970), ‘Gifting, Reciprocity, Savings, and Credit in Peasant Oaxaca’, Southwestern Journal of Anthropology, 26.

Belshaw, C. (1965), Traditional Exchange and Modern Markets. Englewood Cliffs, NJ: Prentice-Hall.

Bennett, J. W. (1968), ‘Reciprocal Economic Exchanges Among North American Agricultural Operators’, Southwestern Journal of Anthropology, 24.

Berkes, F. (1986), ‘Local Level Management and the Commons Problem’, Marine Policy, 10.

Berry, S. (1984), ‘The Food Crisis and Agrarian Change in Africa: A Review Essay’, African Studies Review, 27/2.

Binswanger, H. P. (1986), ‘Risk Aversion, Collateral Requirements, and the Markets for Credit and Insurance in Rural Areas’, in P. Hazell, C. Pomareda, and A. Valdès, (eds.), Crop Insurance for Agricultural Development: Issues and Experiences. Baltimore and London: Johns Hopkins University Press.

——— McIntire, J., and Udry, C. (1989), ‘Production Relations in Semi-arid African Agriculture’, in The Economic Theory of Agrarian Institutions (ed. P. Bardhan). Oxford: Clarendon Press.

——— and Rosenzweig, M. (1984), ‘Contractual Arrangements, Employment, and Wages in Rural Labor Markets: A Critical Review’, in H. P. Binswanger and M. Rosenzweig, Contractual Arrangements, Employment, and Wages in Rural Labor Markets in Asia. New Haven, Conn., and London: Yale University Press.

(p.165) ——— (1986), ‘Behavioural and Material Determinants of Production Relations in Agriculture’, Journal of Development Studies, 22/3.

Boas, F. (1920), ‘The Social Organization of the Kwakiutl’, American Anthropologist, 22.

Booth, A. and Sundrum, R. M. (1985), Labour Absorption in Agriculture. Oxford, New York, and Melbourne: Oxford University Press.

Boserup, E. (1965), The Conditions of Agricultural Growth: The Economics of Agrarian Change under Population Pressure. London: George Allen & Unwin.

——— (1981), Population and Technology. Oxford: Basil Blackwell.

Bourdieu, P. (1980), Le Sens pratique. Paris: Editions de Minuit.

Breman, J. (1974), Patronage and Exploitation: Changing Agrarian Relations in South Gujarat. Berkeley: University of California Press.

Carter, M. (1987), ‘Risk Sharing and Incentives in the Decollectivization of Agriculture’, Oxford Economic Papers, 39/3.

Cashdan, E. (1985), ‘Coping with Risk: Reciprocity Among the Basarwa of Northern Botswana’, Man, 20/3.

Chambers, R. (ed.) (1989), Vulnerability: How the Poor Cope? IDS Bulletin, Special Issue, 20/2.

Coate, S. and Ravallion, M. (1989), ‘Reciprocity without Commitment: Characterization and Performance of Non-cooperative Risk-Sharing’, mimeo.

Codere, H. S. (1956), ‘The Amicable Side of Kwakiutl Life: The Potlatch and the Play-Potlatch’, American Anthropologist, 58.

——— (1957), ‘Kwakiutl Society: Rank Without Class’, American Anthropologist, 59.

Cohen, J. (1980), ‘Land Tenure and Rural Development in Africa’, in R. H. Bates and M. F. Lofchie, (eds.), Agricultural Development in Africa: Issues of Public Policy. New York: Praeger.

Collard, D. (1978), Altruism and Economy. Oxford: Martin Robertson.

Collier, W. L., Soentoro, G., Wiradi, and Makali, (1974), ‘Agricultural Technology and Institutional Change in Java’, Food Research Institute Studies, 13/2.

Colson, E. (1979), ‘In Good Years and in Bad: Food Strategies of Self-Reliant Societies’, Journal of Anthropological Research,35/1.

Corbett, J. (1988), ‘Famine and Household Coping Strategies’, World Development, 16/9.

Dahlman, C. (1980), The Open Field System and Beyond: A Property Rights Analysis of an Economic Institution. Cambridge: Cambridge University Press.

Dalton, G. (1962), ‘Traditional Production in Primitive African Economies’, Quarterly Journal of Economics, 76/3.

——— (1971), Economic Anthropology and Development: Essays on Tribal and Peasant Economies. New York: Basic Books.

Das Gupta, M. (1987), ‘Informal Security Mechanisms and Population Retention in Rural India’, Economic Development and Cultural Change, 36/1, 101–20.

Dirks, R. (1980), ‘Social Responses during Severe Food Shortages and Famine’, Current Anthropology, 21/1.

Drèze, J. P., and Sen, A. K. (1989), Hunger and Public Action. Oxford: Clarendon Press.

Drucker, P. (1955), Indians of the Northwest Coast. New York: McGraw-Hill.

Epstein, T. S. (1967), ‘Productive Efficiency and Customary Systems of Rewards in Rural South India’, in R. Firth (ed.)? Themes in Economic Anthropology. London: Tavistock Publications.

(p.166) Eswaran, M., and Kotwal, A. (1989), ‘Credit as Insurance in Agrarian Economies’, Journal of Development Economics, 31/1.

Evans-Pritchard, E. E. (1940), The Nuer: A Description of the Modes of Livelihood and Political Institutions of a Nilotic People. Oxford: Clarendon Press.

Firth, R. (1966), Malay Fishermen: Their Peasant Economy. London: Routledge & Kegan Paul.

Fleuret, A. (1986), ‘Indigenous Responses to Drought in sub-Saharan Africa’, Disasters, Mar.

Fukuzawa, H. (1982), ‘Agrarian Relations and Land Revenue: The Medieval Deccan and Maharashtra’, in Cambridge Economic History of India, i. Cambridge: Cambridge University Press.

Fuller, C. J. (1988), ‘Misconceiving the Grain Heap: A Critique of the Concept of the Indian Jajmani System’, mimeo, London School of Economics.

Furniss, E. (1930), The Position of the Laborer in a System of Nationalism. Boston: Houghton Mifflin Co.

Geertz, C. (1970), Agricultural Involution: The Process of Ecological Change in Indonesia. Berkeley and Los Angeles: University of California Press.

Gilbert, N. (1976), ‘Alternative Forms of Social Protection for Developing Countries’, Social Service Review, Sept.

Giri, J. (1983), Le Sahel demain. Paris: Karthala.

Gough, K. (1960/1), ‘The Hindu Jajmani System’, Economic Development and Cultural Change, 9, pt. 1.

Gregory, C. A. (1982), Gifts and Commodities. London and New York: Academic Press.

Griffin, K. (1974), The Political Economy of Agrarian Change. London: Macmillan.

Habib, I. (1963), The Agrarian System of Mughal India (1556–1707). Bombay: Asia Publishing House.

Haggis, J., Jarrett, S., Taylor, D., and Mayer, P. (1986), ‘By the Teeth: A Critical Examination of James Scott’s The Moral Economy of the Peasant’, World Development, 14/12.

Hart, G. (1986), ‘Exclusionary Labour Arrangements: Interpreting Evidence on Employment Trends in Rural Java’, Journal of Development Studies, 22/4.

Hayami, Y., and Kikuchi, M. (1981), Asian Village Economy at the Crossroads: An Economic Approach to Institutional Change. Tokyo: University of Tokyo Press.

Ishikawa, S. (1975), ‘Peasant Families and the Agrarian Community in the Process of Economic Development’, in Agriculture in Development Theory (ed. L. Reynolds). New Haven, Conn., and London: Yale University Press.

Jones, E. L. (1981), The European Miracle: Environments, Economies and Geopolitics in the History of Europe and Asia. Cambridge: Cambridge University Press.

Kikuchi, M., Hafid, A., and Hayami, Y. (1984), ‘Changes in Rice Harvesting Contracts and Wages in Java’, in H. P. Binswanger and M. R. Rosenzweig, (eds.), Contractual Arrangements, Employment, and Wages in Rural Labor Markets in Asia. New Haven, Conn., and London: Yale University Press.

Kimball, M. (1988), ‘Farmers’ Cooperatives as Behavior Toward Risk’, American Economic Review, 78/1.

(p.167) Kumar, D. (1983), ‘Agrarian Relations: South India’, in Cambridge Economic History of India, ii. Cambridge: Cambridge University Press.

Lepage, H. (1985), Pourquoi la propriété? Paris: Hachette.

Lipton, M. (1985), ‘The Prisoners’ Dilemma and Coase’s Theorem: A Case for Democracy in Less Developed Countries?’ in R. C. O. Matthews (ed.), Economy and Democracy. London: Macmillan.

Longhurst, R. (1986), ‘Household Food Strategies in Response to Seasonality and Famine’, IDS Bulletin, 17/3.

Macneil, I. R. (1978), ‘Contracts: Adjustments of Long-Term Economic Relations Under Classical, Neoclassical, and Relational Contract Law’, Northwestern University Law Review, 72.

Malinowski, B. (1920), ‘Kula: Circulating Exchange of Valuables in the Archipelagoes of Eastern New Guinea’, Man, July.

——— (1921), ‘The Primitive Economics of the Trobriand Islanders’, Economic Journal, 31/1.

——— (1922), Argonauts of the Western Pacific. London: Routledge & Kegan Paul.

Mauss, M. (1925), The Gift. London: Routledge & Kegan Paul.

Mayer, E. (1985), ‘Production Zones’, in S. Masuda, I. Shimada, and C. Morris, (eds.), Andean Ecology and Civilization, Tokyo: University of Tokyo Press.

Meillassoux, C. (1980), Femmes, greniers et capitaux. Paris: François Maspéro.

Migdal, J. (1974), Peasants, Politics, and Revolution. Princeton, NJ.: Princeton University Press.

Neale, W. C. (1957), ‘Reciprocity and Redistribution in the Indian Village: Sequel to Some Notable Discussions’, in Polanyi et al. (1957).

Newbery, D. (1989), ‘Agricultural Institutions for Insurance and Stabilization’, in P. Bardhan (ed.), The Economic Theory of Agrarian Institutions. Oxford: Clarendon Press.

——— and Stiglitz, J. (1981), The Theory of Commodity Price Stabilization: A Study in the Economics of Risk. Oxford: Oxford University Press.

Noronha, R. (1985), ‘A Review of the Literature on Land Tenure Systems in sub-Saharan Africa’, Research Unit of the Agriculture and Rural Development Department, World Bank Report ARU 43, Washington, DC.

North, D. C. (1981), Structure and Change in Economic History. New York and London: W. W. Norton and Co.

——— and Thomas, R. P. (1973), TheRise of the Western World: A New Economic History. Cambridge: Cambridge University Press.

Obeyesekere, G. (1967),Land Tenure in Village Ceylon. London: Cambridge University Press.

Olson, M. (1965), The Logic of Collective Action. Cambridge, Mass.: Harvard University Press.

Olson, M. (1982), The Rise and Decline of Nations. New Haven, Conn., and London: Yale University Press.

Orlove, B. (1977), ‘Inequality Among Peasants: The Forms and Uses of Reciprocal Exchange in Andean Peru’, in R. Halperin and J. Dow, Peasant Livelihood: Studies in Economic Anthropology and Cultural Ecology. New York: St Martin’s Press.

Ortiz, S. (1967), ‘The Structure of Decision-making among Indians of Colombia’, in R. Firth (ed.), Themes in Economic Anthropology. London: Tavistock Publications.

(p.168) Piddocke, S. (1965), ‘The Potlatch System of the Southern Kwakiutl: A New Perspective’, Southwestern Journal of Anthropology, 21.

Pingali, P., Bigot, Y., and Binswanger, H. (1987), Agricultural Mechanization and the Evolution of Farming Systems in Sub-Saharan Africa. Baltimore: Johns Hopkins University Press.

Pitt-Rivers, J. (1954), The People of the Siena. London: Weidenfeld & Nicolson.

Platteau, J. P. (1984), ‘The Drive Towards Mechanization of Small-Scale Fisheries in Kerala: A Study of the Transformation Process of Traditional Village Societies’, Development and Change, 15/1.

——— (1986), ‘La fonction euphémisante et mystificatrice de ľaide’, in X. Dijon and J. Burton, (eds.), Dis-moi qui tu aides…. Namur: Presses Universitaires de Namur.

——— (1988), ‘A Two-season Model of Hunger Insurance through Patronage: The Jajmani System Put to the Test’, Cahiers de la Faculté des Sciences Economiques et Sociales, Facultés Universitaires of Namur.

——— (1990), ‘The Food Crisis in Africa: A Comparative Structural Analysis’, in J. Drèze and A. K. Sen (eds.), The Political Economy of Hunger. Oxford: Oxford University Press.

——— (forthcoming), Land Reform and Structural Adjustment in Sub-Saharan Africa: Controversies and Guidelines, Rome: FAO.

——— and Abraham, A. (1987), ‘An Inquiry into Quasi-Credit Contracts: The Role of Reciprocal Credit and Interlinked Deals in Small-Scale Fishing Communities’, Journal of Development Studies, 23/4.

——— and Baland, J. M. (1989), ‘Income-Sharing through Work-Spreading Arrangements: An Economic Analysis with Special Reference to Small-Scale Fisheries’, Cahiers de la Faculté des Sciences Economiques et Sociales, Facultés Universitaires of Namur.

——— and Nugent, J. B. (1989), ‘Contractual Relationships and their Rationale in Marine Fishing’, Cahiers de la Faculté des Sciences Economiques et Sociales, Facultés Universitaires of Namur.

Polanyi, K. (1944), The Great Transformation. New York: Holt, Rinehart, & Winston.

——— (1957), ‘The Economy as Instituted Process’, in Polanyi et al. (1957).

——— (1968), ‘Ports of Trade in Early Societies’, in G. Dalton (ed.), Primitive, Archaic and Modern Economies. New York: Doubleday.

——— (1977), The Livelihood of Man. New York and London: Academic Press.

——— Arensberg, C. W., and Pearson, H. W. (eds.), (1957), Trade and Market in the Early Empires. New York: Free Press; London: Collier-Macmillan.

Popkin, S. L. (1979), The Rational Peasant: The Political Economy of Rural Society in Vietnam. Berkeley and Los Angeles: University of California Press.

Posner, C. (1981), The Economics of Justice. Cambridge, Mass., and London: Harvard University Press.

Pospisil, L. (1958), Kapauka Papuans and their Law, Yale University Publications in Anthropology, No. 54. New Haven, Conn.: Yale University Press.

Provinse, J. H. (1955), ‘Cooperative Ricefield Cultivation Among the Siang Dyaks of Central Borneo’, in E. A. Hoebel (ed.), Readings in Anthropology. New York: McGraw-Hill.

Robertson, A. F. (1987), The Dynamics of Productive Relationships: African Share Contracts in Comparative Perspective. Cambridge: Cambridge University Press.

(p.169) Rosenzweig, M. R., Binswanger, H. P., and Mclntire, J. (1988), ‘From Land Abundance to Land Scarcity: The Effects of Population Growth on Production Relations in Agrarian Economies’, in R. D. Lee et al., (eds.), Population, Food, and Rural Development. Oxford: Clarendon Press.

Rothschild, M. and Stiglitz, J. E. (1976), ‘Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information’, Quarterly Journal of Economics, 90/4.

Sahlins, M. (1963), ‘On the Sociology of Primitive Exchange’, in The Relevance of Models for Social Anthropology (ed. M. Banton). London: Tavistock Publications.

——— (1968), Tribesmen. Englewood Cliffs, NJ: Prentice-Hall.

——— (1972), Stone Age Economics. Chicago: Aldine.

Schapera, I. (1970), Tribal Innovators: Tswana Chiefs and Social Change, 1795–1940. London: Athlone Press.

Scott, J. (1972), ‘The Erosion of Patron-Client Bonds and Social Change in Rural Southeast Asia’, Journal of Asian Studies, 32/1.

——— (1976), The Moral Economy of the Peasant. New Haven and London: Yale University Press.

——— (1985), Weapons of the Weak: Everyday Forms of Peasant Resistance. New Haven and London: Yale University Press.

Service, E. R. (1962), Primitive Social Organization: An Evolutionary Perspective. New York: Random House.

——— (1966), The Hunters. Englewood Cliffs, NJ: Prentice-Hall.

Smith, T. C. (1959), The Agrarian Origins of Modern Japan. Stanford: Stanford University Press.

Sow, F. (1986), ‘ĽÉconomie du poisson sur la petite côte (Senegal): Le Role des femmes’, Études Scientifiques, (Université de Dakar), Mars.

Streeten, P. (1987), What Price Food? London: Macmillan Press.

Torry, W. I. (1979), ‘Anthropological Studies in Hazardous Environments: Past Trends and New Horizons’, Current Anthropology, 20/3.

——— (1987), ‘Evolution of Food Rationing Systems with Reference to African Group Farms in the Context of Drought’, in M. H. Glantz (ed.), Drought and Hunger in Africa. Cambridge: Cambridge University Press.

Townsend, W. (1786), Dissertation on the Poor Laws.

Turnbull, C. (1984), The Mountain People. London: Triad/Paladin Books.

Von Braun, J., and Webb, P. J. R. (1989), ‘The Impact of New Crop Technology on the Agricultural Division of Labor in a West African Setting’, Economic Development and Cultural Change, 37/3, 513–34.

Wade, R. (1988), Village Republics: Economic Conditions for Collective Action in South India. Cambridge: Cambridge University Press.

Walker, T. S. and Jodha, N. S. (1986), ‘How Small Farm Households Adapt to Risk?’, in P. Hazell, C. Pomareda, and A. Valdès, (eds.), Crop Insurance for Agricultural Development: Issues and Experiences. Baltimore and London: Johns Hopkins University Press.

Watts, M. (1983), Silent Violence. Berkeley: University of California Press.

——— (1984), ‘The Demise of the Moral Economy’, in E. Scott (ed.), Life Before the Drought. Boston: Allen & Unwin.

White, B. (1976), ‘Population, Involution and Employment in Rural Java’, Development and Change, 7/3.

(p.170) White, B. (1979), ‘Political Aspects of Poverty, Income Distribution and their Measurement: Some Examples from Rural Java’, Development and Change, 10/1 .

Wiessner, P. (1982), ‘Risk, Reciprocity and Social Influences on !Kung San Economics’, in E. Leacock and R. B. Lee (eds.), Politics and History in Band Societies. Cambridge: Cambridge University Press.

Williamson, O. (1985), The Economic Institutions of Capitalism. New York: The Free Press.

Wolf, E. R. (1955), ‘Types of Latin American Peasantry’, American Anthropology, 57.

——— (1966), ‘Kinship, Friendship, and Patron-Client Relations’, in The Social Anthropology of Complex Societies (ed. M. Banton). New York: Frederick A. Praeger.

——— (1982), Europe and the People Without History. Berkeley: University of California Press.

Wynne, E. (1980), Social Security: A Reciprocity System Under Pressure. Boulder, Colo.: West view Press.


(*) I am grateful to all the persons who have offered me valuable comments on this chapter. I am especially indebted to Michael Lipton, Barbara Harriss, and Jean Drèze for their detailed reactions, as well as to Jean-Marie Baland, Pranab Bardhan, Frances Stewart, and Bina Agarwal for their useful punctual remarks.

(1) See Wolf (1955); Polanyi (1944, 1957, 1977); Service (1962, 1966); Dalton (1962, 1971); Belshaw (1965); Sahlins (1963, 1968, 1972); Migdal (1974); Scott (1972, 1976), among others.

(2) For a particularly striking illustration of this possibility, see Turnbull (1984). See also Dirks (1980: 26–31).

(3) In the urban areas of ancient empires, food trade was also regulated because ‘the authorities knew that as far as grain was concerned they might have a revolt on their hands among those who could not meet a free market price in times of scarcity’ (Jones 1981: 87).

(4) In the ‘ports of trade’ system, commercial transactions are carried out in special trading centres under the aegis of representatives of policial entities. In the kula system, potential enemies involved in the trading relationships neutralize each other through a complex network of ceremonial gift-giving with strictly calculated reciprocity. In the ‘silent trade’ system, the trading agents never meet physically and the bargaining is effected through repeated journeys to and from an agreed place where each party proceeds in turn.

(5) Incidentally, Popkin’s characterization of insurance schemes as a collective good is not very helpful. As will be evident at a later stage of the analysis, problems of insurance are best analysed in terms of the economic theory of information.

(6) Gregory and Bourdieu have not adopted this line of enquiry since they have chosen to look at gift exchange as a means through which relations of domination and control are established. This is not necessarily contradictory with my own interpretation, in so far as a person who is well insured is in a good position to accumulate wealth and power in the long run.

(7) Thus, it has been noted that ‘Indian legal procedure considered that contesting parties to a dispute were not discrete individuals but were connected to others by complex and multiple social, political, and ritual ties. It recognized the existence of socially, politically, and ritually unequal corporate lineages and castes, and understood cases of conflict as moments in ongoing relations among such groups’ (Wolf 1982: 248).

(8) If this were what Popkin meant by ‘malleability’, I would not find fault with his statement. However, this is clearly not the case since he precisely expressed the view that social norms cannot influence human actions owing to their volatility.

(9) In stratified tribal societies, however, members belonging to the upper social strata (for instance, members who have a long period of residence in the village or can trace their ancestry back to the founders of the village) are entitled to parcels of greater size and higher fertility (Noronha 1985: 182).

(10) In the normal situation, some plots will be over-flooded and others will remain beyond the reach of the rising level of water.

(11) See, in particular, the special issue of Mountain Research and Development, 5/1, 1985, and Mayer (1985).

(12) For a short survey of these risk reduction strategies, see Longhurst (1986). See also Corbett (1988), Chambers (1989), Drèze and Sen (1989), and the literature cited there.

(13) An interesting example of livestock accumulation as a hedge against critical emergencies (and as forced savings for future investment) is that of pig-raising in Mexico studied by Ralph Beals.Pig-raising is an accretionary type of savings since the pig has to be fed daily and his fattening makes his value increase gradually. According to the author, it is preferred to hoarding cash because it is a less liquid asset: having cash at hand would be too dangerous because it would induce people to spend it on consumption (Beals 1970: 237–8).

(14) The intensity of use of each net is therefore equal to (n.x/N), where N is the total number of nets; n is the number of fishing spots available; and x is the number of hauls per day which can be made in any fishing spot.

(15) In the earlier example, if the number of nets is doubled (from 12 to 24), a net will be used only 1 out of 3 days and if it is quadrupled, it will be used 1 out of 6 days. In the latter case, the net will be idle during 40 days in succession after each 8-day period of use, thus implying that it may not be used once during the whole flush season (which lasts approximately one month).

(16) An interesting issue is the following: why did the growing fishing population choose to multiply the number of nets beyond any reasonable level instead of sharing employment in a number of nets corresponding roughly to the total number of possible turns per day? This would indeed have avoided a tremendous waste of capital. An answer to this puzzling question as well as a more elaborate treatment of the system discovered by Alexander is provided in Platteau and Baland (1989).

(17) Up to recent times, there was no clear distinction between share-owners and wage labourers since many ‘hired men’ used to be ‘the potential heirs of the share owner’ (Alexander 1982: 151).

(18) According to Hayami and Kikuchi (1981: 80): ‘Hunusan is a form of contract by which, when a farmer specifies a day of harvesting, anyone can participate in harvesting and threshing, and the harvesters receive a certain share of the output.’

(19) This is particularly obvious in the case of the mixed contracts commonly in use in traditional maritime fishing communities: here, before the catch proceeds are shared between the equipment owners and the crew workers according to some predetermined ratio, the latter are entitled to take a few fish for domestic consumption. If no surplus is left over after the fish for self-consumption has been removed, no sale takes place and the owners get nothing except their own handful of fish for home use (for more details, see Platteau and Nugent 1989).

(20) Elsewhere (Platteau 1988), I have shown, however, that an attempt at formalizing the jajmani system as it has been described by Epstein brings into light a number of rather odd features. In particular, the system of equations thus obtained appears to be overdetermined and it is only through a special assumption that the model can be made to incorporate a true maximization procedure (on the part of the peasant caste).

(21) In the economic theory of insurance, moral hazard arises ‘when an agent who obtains insurance has an incentive to take less care to avoid the contingencies which give rise to claims’. As for adverse selection, it occurs ‘when the insurance company cannot distinguish between agents who have differing probabilities of claims, and hence must offer all the same contract’, with the result that the contract only appeals to (and adversely selects) those belonging to a comparatively high-risk category (Newbery 1989: 278). Both problems arise from ‘asymmetric information’ regarding the relevant personal characteristics of the agent seeking insurance.

(22) It is true that coinsurance clauses may also reduce the incentive problems. Yet, as they cause insurance to be incomplete, they also reduce the potential utility gain to the insured (Binswanger 1986: 78; Newbery 1989: 278).

(23) Wade has rightly noted that, for Popkin, the temptations to free-ride with respect to leadership are typically so strong that insufficient leadership is normally available within peasant communities. As a result, leadership would have to come from outside the local community to make the villagers co-operate (Wade 1988: 207).

(24) It may be recalled in this context that the comparative ease of holding free-riding in check in small communities was one of the strong points made by Olson (1965, 1982).

(25) In a valuable study of the ancient English open field system of agriculture, Carl Dahlman has offered another, more positive, justification of the scattering of plots than the one given in the text: for him, the main advantage of plot scattering lies in the fact that it creates an incentive for the farmers ‘to participate in the collective decision making and control necessary to regulate the use of the large grazing areas in both the commons and in the arable fields’ (Dahlman 1980: 125).

(26) However, this need not be so if the participants have different degrees of risk aversion. Those with high risk aversion could dispose of risky assets to those with low risk aversion and purchase less risky, less profitable assets from them, with both sides obtaining consumer’s surplus from the exchange transaction (I am indebted to M. Lipton for having drawn my attention to this point). Note, however, that trade in assets is only one of the ways in which people with different degrees of risk aversion can achieve mutual gains.

(27) Note, incidentally, that the latter aspect of this proposition has been emphasized by Popkin himself: There was, in fact, cooperation among small groups of peasants to stabilize production and provide insurance’ (Popkin 1979: 96).

(28) Wiessner has observed that the storing of food is widespread among the Kung households. However, from the high incidence of hxaro relationships one can infer that self-insurance offers only an imperfect hedge against environmental hazards.

(29) The reasons accounting for these two phenomena have been spelt out in Platteau and Abraham (1987: 463–6).

(30) For a summary of the findings of these studies, see Fuller (1988: 25–9).

(31) Note that I have not dealt with the situation prevailing in the urban areas of precolonial empires or kingdoms. The contention that the right to subsistence ‘certainly did not exist everywhere’ (Haggis et al. 1986: 1449) is probably more valid for urban than for rural areas. A lot more research needs however to be done on this issue before any firm conclusion can be reached.

(32) In Europe, it was precisely with the advent of capitalism that the old values and institutions (like the Poor Laws in England) guaranteeing food security for everybody came under the strong attacks of the heralders of the new ‘laissez-faire’ order. Interestingly, incentive considerations were the main argument which the latter put forward: only poverty can incite (force) the lower classes to work and, according to Townsend (1786), hunger is actually the most effective pressure, because it is ‘peaceful, silent and continuous’ (quoted from Polanyi 1944: 113–14; see also Furniss 1930: 118).

(33) For a useful distinction between the concepts of self-sufficiency and self-reliance, see Streeten (1987: 39–40).