Policy, Research, and Extemal Affairs WORKING PAPERS Agricultural Policies Agriculture and Rural Development Department The World Bank April 1990 WPS 395 How Gambianls Save - and WVhat Their Strategies Imply for International Aid Parker Shipton An interesting suggestion emerges from this analysis of Gam- bian saving strategies: that aid agencies try subsidizing savings, not credit. The Policy, Research, and Extemal Affairs Complex distributes PRE Working Papers to disseminate the findings of wotk in progress and to encourage the exchange ofideas among Bank staffand all others interwted in development issues.These papers carry the names of the authors. reflect only their views, and should be used and cited accord.ngly. The findings, interpretations, and conclusions ate the authors' own. They should not be attributed to the World Bank, its Board of Directors, its management, or any of its member countries. Policy, Research, and Extemal Affairs Agricultural Policies This paper - a product of the Agricultural Policies Division, Agriculture and Rural Development Department-is part of a largereffort in PRE to assess the factors affecting the success (orcausing failure) of credit programs addressed to smaller farmers. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Cicely Spooner, room N8-039, extension 30464 (60 pages with tables). International aid for Gambian farrners is out of neighbors - and remove it from their own balance, concludes Parker Shipton. temptation. Interventions in this small, impoverished, Some of the most imponant ways of saving West African country have been based mainly money (as in lutating credit and saving associa- on credit (and thus on debt). Public and private tions) and other goods (as in grain storage) lending institutions have failed dramatically and combine principles of individual property and debts are mounting, but lenders continue extend- group control - prinriples that cou!d be cmu- ing loans into the countryside. lated in institutional finance, and that underlie some well-known innovative programs in Before suggesting better ways to help Africa, sout'lerm Asia, and Latin America. Gambians alleviate poverty, Shipton answers the question: How do Gambians save? Gambian farmers choosing savings options -weigh trustworthiness and convenience more Saving takes many forms other than just heavily than real or nominal interest rates or money - including livestock, jewelry, store inflationary losses. Some of their favorite ways crops, and resaleable household goods. Sdme of saving are expensive. forms are gender-specific, some secrct. Some farmers keep cash in sealed containers; some There is room for new options. One of these deposit it with trusted local "money-keepers;" is for financial intennediaries to subsidize not some ask for deferred wage or share-contract credit - the standard approach - but savings. payments. Gainbians and others like them need a Gambians do not prefer liquidity. They balance between credit and saving, liquidity and often convert wealth into forms that shelter it illiquidity, individualism and group responsibil- from the daily demands of spouses, kin, and ity. The PRE Working Paper Series disseminates the findings of work under way in the Bank's Policy, Research, and Extemal Affairs Complex. An objective of the series is to get these findings out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced at the PRE Dissemination Center ACKNOWLEDGEMENTS The field research on which this paper is based was funded by the U.S. Agency for International Development, as part of the Economic and Financial Policy Analyses Project of the Ministry of Finance and Trade in The Gambia. The staffs of the Ministry and the AID Banjul office and my colleagues at the Harvard Institute for International Development (HIID) provided valuable help. The paper was revised under funding from the Carter G. Woodson Institute for Afro-American and African Studies, University of Virginia; and from the World Bank. Members of many other organizations cooperated generously in the research in The Gambia. All this help is gratefully acknowledged. Special thanks are due to C. Baldeh, E. Baldeh, A. Jallow, A. Jammeh, F. Khan, J. Newlands, A. Sillah, and M. Trawally for their dedicated assistance there; to B. *ilman for data processing; and to D. Adams, E. Bazalgette, D. Cole, Y. Dem, C. Elias, F. Janneh, F. Le Grand, M. McPherson, R. and P. Mueller, L. Sanneh, H. Sanyang, B. Sidibe, P. Steele, J.D. Von Pischke, and P. Weil for information and advice. This paper does not, however, necessarily represent the views of any of these organizations or individuals. I owe sincere thanks to the many Gambian villagers who have so generously given of their hospitality and time during field study. How Gambians Save -- and What Their Strategies Imply for International Aid by Parker Shipton Table of Contents I. Introduction 1 II. Rural Saving Without Money: Storage Cycles 9 and the Gender Walls of Property III. Saving in Cash: Illiquidity Preference and the 16 Squawk Factor IV. Local Mechanisms Combining Savings and Credit 25 V. Banking 39 VI. Summary and Conclusion: Balancing Credit with 44 Savings Tables 49 Footnotes 52 Bibliograjhy 57 I. INTRODUCTION Juloo, wrope" to a Mandinko, means several things at once. It can refer to a small-scale trader, or to credit or debt. Every Mar.dinko knows the meanings are related. Traders are also lenders, and their loans, while sometimes useful like a rope ladder, also tie down a farmer like a rope around the neck. When rural people in The Gambia speak of iuloo, in any of these uses, they consciously or unconsciously connote slavery. The Mandinko and other peoples of this small and impoverished West African river nation, an ancient trade route winding thinly through southern Senegal, have had occasion in history to learn quite a bit about ropes and involuntary servitude, and about debt. The linked images and overtones are not empty of emotion. In recent decades, international interventions in Gambian agriculture have been based mainly on credit. Each year, with mixed motives, new committees convene in Banjul, the capital, to seek new ways of extending loans into the countryside: representatives of multinational and bilateral agencies, governmental bodies, private voluntary organizations--lending for men or women, for groups or individuals, for cash crops or food crops. The composite record of the lending projects has been disappointing for farmers and nearly all others concerned. Particularly unimpressive have been the records of the cooperatives, which now dominate the rural credit picture with their inefficient yet now rather indispensable operations, and the commercial banks, which deal with a'most no -2- rural Gambians (see Wing 1983, Ramamurthy 1986, Clark 1987, Shipton 1987, Demissie et al. 1989). At the macro as well as micro levels, the country's debts are considerable. As of August, 1988, financial analysts from Harvard University in the Ministry of Finance and Trade calculated that at the current rates of economic growth and repayments, it would take until the year 2,047 to pay off, even if no new debts were contracted--which, of course, more will be.1 Debt costs money, dulls incentives to produce, and constrains political independence. The disappointments and debt spirals in Gambian agriculture have been paralleled across Africa south of the Sahara (see e.g. Hyden 1973, USAID 1973, Lele 1975, Shipton 1985). This paper looks, then, at the saving side of rural finance, internationally the more neglected side. The simple question, "how do rural Gambians save?" has various answers, some neither simple nor obvious. Describing several indigenous and other saving systems, starting with those closest to a rural home and proceeding outward, this paper shows farmers resorting to both individually and socially devised means of saving, many of them convenient but comparatively costly to them. It suggests that the country's development strategies, formulated with outside assistance, have been in imbalance. Farmers need not just credit, but also more and better opportuiities for savings, partly to reduce their dependency on borrowing. A financial policy based on only credit without savings is not only ethically dubious, but also impractical; it is like walking on one leg. The existing saving habits of rural Gambians suggest that aid efforts need other kinds of balance too. Farmers save, and must save, not just in cash but also in kind. Whereas many economists in the past have assumed small farmers to - 3 - have a *liquidity preference', the research suggests the opposite--illiqzuidity preference--to be at least as important. The best mechanisms for rural finance link credit and savings together, and they combine elements of individualism and collective respornsibility. They take account of both achievement incentives and social risks for savers. Where farmers are involved with financiers, not just capital but also information must flow two ways between them if rural people are to benefit. The paper is based mainly on a field study of farmers' financial strategies, using both intensive, open-ended interviewing with a few key informants, and broader, structured surveys, in April and May 1987 and from July to October. 1988. For the surveys, five villages were randomly chosen in Upper River Division, MacCarthy Island Division, and Western Division (on both sides of the river). The ethnic distribution of informants, mainly Mandinka, Fula (Peulh), and Wollof speaking in that order of frequency, approximated that of The Gambia as a whole. While I lived and interviewed mainly in two villages with locally known, trusted, and trained research assistants, other such assistants surveyed three further villages, occasionally visited briefly by me. 138 farmers were visited in 1987; and in 1988, these were revisited with others added, for a total of 167 (counted after information discards). In four of the five villages, 95 to 100 percent of the compounds (sometimes identical to dabada 'work groups', sometimes larger) were visited in the second survey.2 One man and one woman were interviewed in each compound, each several times.3 The ethnic distribution of informants, mainly Mandinka-, Ful_- (or Peulh), and Wolof-speaking in that order of frequency, roughly approximated that of The Gambia as a whole.4 A word about ethnic variation. The Gambia differs from some - 4 - other parts of Africa south of the Sahar& in that members of different ethno- linguistic groups are fairly well mixed in the countryside; while some groups predominate in particular divisions and districts of the country, it is only at the village level, and usually not even there, that anything like ethnic uniformity may be found. InterethnJc c t.tact is daily and pervasive. Some important differences in economic thought and behavior remain, as briefly noted later, but one may speak of general Gambian saving tendencies in a way that would make less sense in many African countries.5 - g and Credit Realistically Conceived "Saving, refers here to any conservation of movable property by an individual or group for future use or disposal. Usually most of a family's savings will be held in non-monetary forms like livestock, grain, machinery, or jewelry; and further wealth will lie in obligations owed by kin or neighbors. Saving, consumption, and productive investment are often not clearly distinguishable on West African farms. Purchasing a draught animal, for instance, can mean all these things simultaneously, as can contributing labor or grain to a ceremony in which other participants are potential part-time helpers on one's farm. In a sense, lending can also be a form of saving, and an effective one, since it removes property from the constant demands of relatives, neighbors, friends, or tax collectors. "Credit" or a "loan' shall refer here to any transfer of goods or services by one person or group to another, or to any of its members, with the expectation of a compensation at _ater time. The loan and the compensation may take the same form or different forms. In The Gambia, virtually everything is lendable - 5 - and at times will be lent. This includes nearly all factors of agricultural productions land, labor, livestock, seeds, fertilizer, pesticides, and farm tools. Craft tools, vehicles, and household goods are also lent. "Credit" to anglophones has a cleaner ring than 3uloo to Mandinka-speakers. Its connotations are positives confidence, trust (it comes from the Latin credere, to believe or entrust), and helpfulness. This is perhaps one of the reasons why credit is such a common term in development plans, and why planning committee recommendations so often seem to boil down to credit. But credit, of course, is only another way of saying debt. Credit is debt: it matters only whether one is the borrower or the lender, and whether one is speaking before or after the loan. If "credit" is replaced with "debt" wherever the word appears in project documents, these begin to look rather different: "The main impediment to farming is lack of dabt"... " In project year one, 2,500 farmers will be issued debt..." Juloo contains something of this disturbing truth, something of a warning. Just as there is no credit without debt. there is none without patronage. At all levels, international, national, and local, cheap loans can be a way of securing supporters, clients, or voters: credit can be as much a political as an economic device. Yet the stock language of international lending agency conversations and reports carries subtle but clear assumptions of condescension. Lenders are misidentified as "donors". Borrowers are called not borrowers but "loanees" (or "target groups"), as though merely passive recipients of something grandly provided. The reports tell of "supervising" borrowers, and of needs for "discipline", as though the borrowers were not quite adult. "Moral hazard" seems to refer always to the "loanees', never to the lenders themselves. - 6 - Cultural differences between international lenders and borrowers in The Gambia make their partnerships hard to manage. Assumptions about interest charges, for instance, differ profoundly between northern and West African peoples. Whereas the former conventionally distinguish fair lending from usury in terms of "interest rates" (i.e., increments per unit of time), the latter tend to draw the distinctions more in terms of interest ratios (i.e, of interest to principal, as the amounts are finally paid), the amount of time elapsed tieing less important. Differences between Judeo-Christian, Islamic, and local religious understandings about interest multiply the misunderstandings.6 In superficial survey interviews, farmers anywhere often say they want credit; they do so knowing that aid agencies prefer loans to other kinds of intervention, and that there will always be a chance of debt forgiveness. But needs and expressed wants are not necessarily the same, and credit always seems more useful to the borrower before a loan, when it is still "credit", than after it, when it is debt. Real behavior also gives the lie to credit clamor. In The Gambia, poorer farmers tend to borrow more often, and on worse terms, than richer ones. Farmers tend to borrow more after bad harvests than after good ones, evidence that they prefer at some level to minimize their seasonal indebtedness. Farmers unassisted by governments, cooperatives, or banks are neither without credit (as some reports would have it) nor entirely at the mercy of "moneylenders" (as others would have it). Many and varied loans occu; within and between villages, including, for instance, seasonal crop loans, share contracting arrangements, delayed marriage payments, contributions for schooling or labf migrations with remittances expected later in return, and seed capital loans between small entrepreneurs (see Shipton 1987, 1990a,b). Some loans last orly -7- hours, others up to several generations. Gamb4;n f&rmers live in a credit economy, and most are perpetually involved in a complex web of debts and credits to relatives, neighbors, friends, and merchants. Anyone involved in lending co smallholding farmers should understand, then, that farmers have their own personal hierarchies of creditors, and that the newest, most distant, and least familiar lenders rank at the bottom. (And seen from a 'target grotup", even big guns, when far enough away, become very small.) This understanding puts into perspective pejorative judgments about *moral hazards' and "delinquencyw where farmers fall into arrears or default on institutional loans. Other Basic Principles West African farmers are usually not just farmers. Multiple occupations (sometimes intermittent and often missed in questionnaire surveys) cut subsistence risks, even out income through the year, and put family labor to beneficial use. Real rural people do not live in 'sectors", and these notions, as embodied for instance in the divisions between corridors of large international aid agencies, are not as real or important to Gambian farmers as to some development planners from northern countries. Saved and borrowed resources are exchangeable and substitutable (to economists, "fungible") between agriculture, for instance, and housing, trade, health, or education, in ways lenders cannot easily control and should probably not attempt to control. The social divicions common to the several larbest Gambian (and southern Senegalese) ethnic groups include not just nucle-ted villages with elected headmen, but also patrilineages and divisions into cross-cutting castes, classes, - 8 - age-grades, Islamic worship groups, and voluntary associations.7 These give each individual multiple economic resources and obligations within a village, and some outside. There is no single social unit of analysis like 'the household" to consider within villages, but a nested hierarchy of decision fields, including village wards (where they exist), lineages compounds, work groups, and cooking groups. Decisions about family resource allocation ate often not made unilaterally by "family heads"; in The Gambia, as elsewhere in Africa south of the Sahara, men and women often make their financial decisions separately, or negotiate and compete about joint savings or investments.8 Seasonal and Longer-term Fluctuations Rural financial systems in The Gambia can only be understood in terms of their seasonal nature. The single rainfed agricultural season produces an annual spurt of wealth, trade, and other economic activity nationwide from about December to March, following the harvest of groundnuts and the other main crops. Moct loans of agricultural inputs will be sought around June, at the onset of the rains; food is commonly borrowed as needed between then and the harvest of the first crops (in some areas, early millet in September), though it may be borrowed at any time for ceremonies or other special needs. It is at harvest time that farmers will make most of the loan repayments they make. Some rotating saving and credit associations (or "ROSCAS", locally called osusu groups, and in French, tontines) among farmers suspend their operations during the rainy season, when cash for contributions becomes scarce, and re-commence in the trade season. In rural bank branches, lending may similarly - 9 - fall off during the year as money for lending becomes scarce. The seasonal nature of credit in the Gambia strongly influences local thought about interest rates (Shipton 1990a). II. RURAL SAVING WITHOUT MONEYt STORAGE CYCLES AND THE GENDER WALLS OF PROPERTY The first and most basic point about savings in The Gambia is that most of it does not take the form of money. Its most important material forms include livestock, food, jewelry, tools, and household goods. Some but not all of these kinds of savings can be directly influenced by projects and programs. Livestock Livestock are the most important form of movable property saved in The Gambia. Nearly every rural family keeps some small or large stock.9 Men claim most of the cattle, and nearly all the donkeys and horses; the sheep and goats are at least as commonly owned by women as by men. Like most other Africans who possess livestock, Gambians seek to convert gŁain to small stock, and small stock to large stock (while still keeping at least a few small stock for diversity and liquidity). Large stock, in turn, can be useful to men for marriage and other ceremonies, and thus for obtaining a larger labor force to produce more grain.10 Men value large stock as a form of savings partly because it is indivisible and thus more or less removed from the day-to-day claims their wives- -this is what I shall call the "gender wall of property"--and other kin. Few who can afford to convert some of their small stock to large stock will fail to do - 10 - so.11 It is only in cases of hardship or special ceremonial needs that Gambian farmers are likely to sell or slaughter cattle. Though it shelters wealth from daily demands of kin, livestock is none the less subject to some special family claims limiting its use. Stock inherited by a group of siblings cannot be disposed of by any individual for his own self-enrichment without the consent of the others, because it is believed that this would produce 'bad money": money which will ultimately be wasted or bring a tragedy to the one who sold the family property.12 In traditional Gambian religions it is spirits who are thought to control the outcome of these events; to fully converted Muslims it is Allah. Today many Gambians question the popular beliefs about alienating family wealth, but most prefer to stay on the safe side. It is common practice to lend cattle out to relatives or friends for months or even years. This is done for several reasons. One is to provide needy kin with milk or manure. Another may be to reduce pressure on grazing lands around one's own village. A third is to conceal one's wealth from tax-collectors, neighbors, or others. A fourth is to minimize risks of losing one's entir- herd through an epizootic disease or a theft. Because herds of different owners are usually mixed, and because it is considered impolite to ask how many cattle a farmer has, it is extremely difficult to assess wealth in livestock. Farmers give several reasons for preferring to save their wealth in stock, in addition to the subtler considerations of gender noted above. These include the fact that animals breed, the advantages of milk and manure production, the usefulness of animal traction, the ceremonial values of animals, the status value, and the satisfactions in having herds to watch. Ecologically, however, an unlimited accumulation of livestock may be hazardous in its soil erosion - 11 - effects, particularly as the human population and cultivation pressure on land also grow, and this is probably a good reason for making other saving options easier. Propensities to invest in livestock, like propensities to save cash, seem to vary from one ethnic group to another. In the Gambia the Fula and Jola have acquired reputations as avid cattle-keepers. But all rural Gambians would rather convert other property to livestock than convert livestock to other property. Therefore project planners should not consider cattle as a store of freely usable wealth, expendable for agricultural inputs like fertilizer or able to be liquidated for community projects. Jewelry If cattle are a characteristically male preserve of wealth, gold and silver jewelry are a female preserve, a shelter from the daily demands of husbands and others. Most important are earrings.13 Those provided to a young woman by the time of marriage commonly cost 1,000 dalasis or more. The ornaments may grow in size as a woman ages; gold can be added by a smith as a woman acquires new wealth. Like livestock, jewelry is something rural Gambians will not part with easily, but it constitutes a useful store of wealth for times of crisis. In the occasional uncommon instances where collateral is demanded for loans by traders or distant acquaintances, it can take the form of jewelry.14 In an emergency, and only then, husbands may ask wives to pawn their earrings. Gender walls of property are not wholly impermeable. - 12 - Stored Crops A substantial part of the wealth of each family (or work group) is its stored crops. Stored food usually falls under the control of one member of the family, though other members may complain if they think it is misused. Mien are usually expected to store coarse grains (millets, sorghum, and maize) for family food; women who grow rice store much of it too for family consumption. Though both men and women now grow groundnuts (the main cash crop) in most of The Gambia, they are expected to use some of it in different ways. Women are expected to store a substantial part of their groundnut crops for family food; men are not. Men, however, are expected to store and provide seed for the family, while women are not. On balance, this means that women store greater proportions of their groundnuts for family use than men do, and that men sell greater proportions than women.15 Gambian farmers who sell food crops tend to do so just after harvest. They recognize that withholding crop sales until later in the year, when food becomes scarce and prices rise, can yield them more cash. But few feel they can afford to wait. One reason is that losses in storage can be considerable. Another is that the demands of needy relatives become stronger in the "hungry season", and saving one's crops then for sale may appear anti-social: while maximizing short-term profits, one risks losing a social support network that may have longer-term economic value. A few who can afford to withhold or buy crops at harvest time, and are willing to risk some social disfavor for doing so, serve as a kind of "food pawnbrokers" for other farmers, selling them back similar foods later in the season at higher prices. - 13 - Some villages have central seed storehouses that members of all compounds may use for groundnuts, coarse grains, and other seed. In these, however, sacks or bundles of seeds are usually individualized and may be labelled. The stores are often kept locked, and two or more keyholders may be required to open them. An advantage of this form of storage is the economy of scale in strong construction.16 A cement house with an iron roof, beyond the means of many individual farmers, cuts down rodent damage. Centralization also makes treatment with storage pesticides easy and inexpensive. Another advantage is that the system adds disincentives to consume one's seed before planting, since doing so requires arrangement with the keyholder(s). A disadvantage is that farmers must let others know how much they possess, something they usually try not to do in the case of personal or family stores. Storage Facility Projects Village seed stores appear to have been an effective form of development intervention by various governmental and non-governmental agencies (most notably the Freedom from Hunger Campaign). Although there have been several projects for building and managing seed stores, many villages remain without these. People in these villages are clamoring to get them. Gambian farmers are losing their own resources to rodents, birds, smaller pests, and molds while they continue to accrue new and expensive debts for food and seed. This unfortunate cycle should be broken. No one appears to have comprehensive knowledge of what storage facilities exist in the Gambian countryside--a nationwide survey would be welcome--but it - 14 - is clear that there are many villages with no central seed stores or cereal banks, and almost none with special storage facilities for women. Crop and seed storage facilities should be created and maintained not just for cash crops, but also for food crops. They should be created and maintained not just for men's crops, but also for women's. Hitherto, emphasis in seed store construction has been placed on groundnuts. Storage facilities are also needed for rice--a crop highly vulnerable to rodents and other pests--and perhaps coarse grains. Seed and crop stores, once constructed, take relatively little monitoring and follow-up. This is a great advantage of this kind of "savings' scheme over credit schemes. Gambian women are accustomed to storing rice and coarse grains individually in their houses or grain stores, for good social reasons. One is that they do not wish other families to know exactly how much they possess at any time. Probably few would be willing to put all their crops in central village stores. Attempts should be made to help them improve their family stores --these seem in many places to have deteriorated over recent decades--as well as to help establish more village stores. The approach should be experimental at first. Gambian women are not widely accustomed to storing their crops in sacks. Sacks will need to be made more easily accessible to women if village storage facilities are created for them. It should be expected that some or even most sacks will be used for other purposes, like bedding: this is one of those "intersectoral leakages", infuriating to project managers but perfectly natural to farmers. "Cereal banks" have been another approach tried in recent years. These have differed from seed stores in that they tend to be larger-scale, designed - 15 - not for single villages but for clusters of several or many, and in that they have included finances resources for buying and selling crops. The buildings constructed seem to have been more successful than the cash components of these projects, which seem to have suffered from the same kinds of graft problems as co-operatives and have soon collapsed. A problem, just as in the co-operatives, seems to have been the absence of a feeling of shared responsibility between members of different, ostensibly cooperating villages. Generally, basing storage facilities on the village unit has proved more successful than basing them on larger aggregates. Linkages between Credit and Storage Evidence from 1988 shows that farmers are willing and able to store higher proportions of their own groundnut seeds when warned that they will not receive them on credit from the co-operatives. They store more for their "strange farmers (seasonal land clients) as well as for themselves. There are two policy implications. One is that the government and co-operatives should inform farmers early in the trade season about credit availability for the following season. The other is that when co-operative groundnut seed credit diminishes, storage facilities for other crops become more necessary, since farmers' storing more groundnuts squeezes other crops out of the stores. In 1988, even after three consecutive good harvests, most of the informal borrowing of money was still for food, mainly rice. Cultural and religious pressures favor credit for immediate consumption over credit for immediate production. But the truth is not as simple as this. Most informal borrowing of money and food occurs during the rains. This is when hunger, weakness, and - 16 - malarial disease coincide with the peak labor season, and when most crops need weeding and rice needs transplanting. So credit for food is also a kind of indirect production credit. Because labor is about the scarcest resource in Gambian farming, this function is not to be underestimated. Those who cannot consume cannot produce. Other Household and Farm Goods In the brief period of relative solvency after harvests, rural Gambians often buy hard goods that they will use and later re-sell, whether the re-selling was part of their original plan or not. These durable goods are an important form of savings. They may include electronic goods like radios or tape recorders, "luxuries' like fancy beds; vehicles like bicycles or carts; or food processing machines like peanut-butter grinders. (In the cities and larger towns, fine clothing is also used and re-sold frequently.) The intimacy of social life in a Gambian village ensures good communications between buyers and sellers of these used goods. III. SAVING IN CASH: ILLIQUIDITY PREFERENCE AND THE SQUAWK FACTOR Cash in rural Africa is an odd commodity surrounded by ambivalent attitudes (Shipton 1989). Nothing in The Gambia is more sought after than money, but nothing is more quickly disposed of. Indeed, money is even seen as something to get rid of, something to convert into longer lasting forms. Several features make money an unstable form of wealth in The Gambia: its nearly universal fungibility, its divisibility, and its portability. These features make money - 17 - contestable. Everyone needs it for scAiething, particularly in the lean season from June to August; and one with money will usually have an almost infinite number of relatives or neighbors with pressing needs. Inflation, of course, is a further reason not to hold onto money. Though few farmers have the means to measure inflation, nearly all are aware of the process. Rural Gambian saving strategies, then, are largely concerned with removing wealth from the form of readily accessible cash, without appearing anti-social. In communities where one has many relatives, as is usual, this is a delicate balancing act, and besides any ethical issues involved, the "squawk factor", the potential for complaints and accusations, must enter every individual savings decision. Nonetheless, individuals do keep some cash reserves (in Mandinka, fangkanto; in Fula and Wolof, fangkanta). Much of the cash that rural Gambians keep, they keep at home; and when asked how they preferred to save money, nearly half said they preferred to keep it for themselves.17 About 25 percent said they preferred to use banks--far more than had actually used them, suggesting problems of access. Other ways of saving mentioned included entrusting money to another individual, post office savings, burial, and mother' means including savings 9clubs (see Table 1). We review below some of the various ways of saving cash. Perhaps the oldest way of removing money from the public eye, in The Gam.bia as in some other parts of Africa, is literally to bury it, in an earthen pot, bottle, or box.18 This old practice is disappearing, no one is proud to use the method--only 2 percent of the men and none of the women named burial as their preferred way to save. Metal locks, now available in every Gambian market town, have made burial seem less necessary. The main advantage left in burying one's wealth is the safeguard - 18 - against house-burning. In those villages where farmers have gradually replaced thatch with metal roofing, this function too is being obviated. Most cash savings in the farming communities today are kept well concealed, usually together with clothes, jewelry, or other valued possessions, in a locked box or trunk, under or beside the owner's bed or sleeping platform. Locking customs deserve some attention; they differ from what foreign visitors sometimes expect. Most Gambian houses constructed as permanent dwellings now have lockable fixtures on their exterior doors, but just about anyone who can afford to do so keeps a lockable box within. When asked about why they lock the boxes inside their horses too, Gambians often speak of their spouses. Where cash is concerned, many marriages include very little free and easy sharing, nor is it usually expected that spouses will share knowled3e of how much cash they possess. Bluffing s.nd negotiation are routine and well-understood. Rural Gambians are sometimes surprised to hear how, in the United States (lately such a world beacon of individualism on the radio), spouses often leave cash in bedroom drawers or handbags hung in pantries, known and accessible to each other, and how many indeed lock nothing away from each other within their houses.19 To rural Gambians such habits can seem like rather unnatural family communism. But owners must protect their cash from themselves as well as from their kin and neighbors. Hence the custom of the box called "kondem" (from the English "condemn*) or ukondanehZ (from the French, condamnee) in local languages, resembling in a western "piggy bank" in function. The farmer engages a carpenter to build a box (usually costing about 5 dalasis) that must be broken to be opened, with a small slot for inserting money. Both younger and older adults of both sexes do this. (Kondem boxes are common not just in the countryside but also - 19 - in the citios, where they are likely to be made with plywood boards.) The boxes, like old burial pots, are evidence of something contrary to the 'liquidity preference" that foreigners sometimes assume. Just as important, or more, is an "illiquidity preference". Rural Deposit-Takers Another widely-used strategy is the money-keeper.20 Some rural Gambians give money to others in their villages to hold for them. They may choose any trusted person--there are no specialized deposit-takers--but often the depositor will choose a relative or other person who is likely to have other cash available for his or her own urgent needs. In the 1988 survey, 46 percent of the women and 20 percent of the men respondents had entrusted money to other individuals in 1987 and 1988.21 More had given deposits than taken them: in the same period, 16 percent of men and 15 percent of the women respondents had received deposits from others. An individual may hold money for several people simultaneously, as these figures suggest; and more occasionally, an individual deposits with more than one keeper at a time. The great majority of depositors of both sexes had made their deposits in January and February, in the season of groundnut harvesting and marketing, and more than half said they had made only one deposit in a year. 63 percent had withdrawn their deposits in Hay and June, reflecting tne importance of land preparation and planting expenses as a purpose of saving. Otherwise the withdrawals were spread about evenly through the year. Only rarely does an individual leave money in deposit with another for more than a year. - 20 - Who were the money-keepers? Mostly they were people from families like those of their depositors, but often rather better-off. 73 percent of the depositors identified their money-keepers as mainly farmers by occupation. 13 percent named shopkeepers or traders, 4 percent named government employees, and 10 percent named others. 71 percent of the depositors (80 percent of the depositing men and 64 percent of the depositing women) left the money with men. So, while roughly equal proportions of men and women take appear to take deposits, more people entrust their money to the men than the women. A village headman (alkalo) or other local notable sometimes holds several deposits simultaneously. Why the bias toward men is unclear. It may be because men are more respected and feared, and considered more likely to pursue or punish a thief. Money-keepers may sense that men's higher prestige gives them more to lose if they should fail to refund the savings. (These are also reasons why some depositors choose village headmen.) Men are likely to have had more formal schooling than women, and to be more comfortable doing arithmetic with large numbers. It may be that, having more cash on average than women, men are locally deemed less likely to dip into the cash. Perhaps men, generally often have more "blood kin" in their respective villages than women, are expected to have more alternate cash sources for their own emergencies (though they also have greater cash obligations). Perhaps, for reasons of pride, men simply did not wish to admit saving with women as much as women did with men. Money keeping is very local. 79 percent of the depositors named as their money-keepers other members of their respective villages, reflecting the importance of both residential nearness and kinship--since these often coincide- - 21 - -in determining trust or confidence. To isolate the role of kinship, 69 percent specified known relatives as their money-keepers, and 45 percent named their parents, spouses, full siblings, or offspring. 79 percent of women depositors used kin as their safe-keepers, as against only 55 percent of depositors. Among the women depositors, 36 percent named their husbands as the money-keepers, whereas among the men depositors, only 5 percent named their wives, raising interesting questions about who trusts whom with money in marriages. Only 24 percent of the male and female depositors identified their respective money- keepers to be unrelated friends. Note, however, that the sample sizes behind these figures are small (see Table 2). Gambians prefer to entrust their money to their elders. Thus they use their parents but not their offspring as money-keepers. That women entrust their money to their husbands more than men to their wives (Table 3) fits the age pattern too, since most husbands are older than their wives, some by many years. How big are the amounts involved? Irn the 1987 survey they varied from 20 dalasis--the lowest reported, at any rate--to over 2,000, the median deposit being about 251 dalasis. Though women are more likely to deposit with money- keepers than men, the men deposit bigger sums. In their last visits with money- keepers before being interviewed in 1988, depositing men had left a median sum of 530 dalasis for safe keeping, as compared with the depositing women's median of 150 dalasis. And 26 percent of the men depositors, as against only 10 percent of the women ones, reported having entrusted amounts of 1,000 or more. - 22 - Terms, Risks, and Incentives in Informal Money-Keeping Interest payments or service charges are unheard of in village money-keeping, and the depositor often expects to receive back the same notes or coins lent. It is generally agreed that the money-keeper should not use the money deposited, but it is not so clear who holds the legal right to the money during the period of the deposit. A Basse court record describes a civil case in which a shepherd entrusted 80 dalasis to a shopkeeper for safe keeping, and the shop's cash was later stolen by thieves. The shepherd sought the full 80 dalasis. The court ruled that the shopkeeper owed the shepherd 40 dalasis, half the amount deposited. Villagers interviewed disagreed among themselves about the correctness of the decision. If moneykeeping can entail some financial risk to the moneykeeper, why does he or she agree to perform the service? The obvious and ostensible reasons include helping the saver, cementing a social bond, and demonstrating trustworthiness to the community by eventual word of mouth. Another possible reason is that the cash, though normally expected not to be used, none he less gives the money-keeper a reserve available for family emergencies: if need arose to taxi a snakebitten child to the hospital or to pay a son's court bail at short notice, cash would be on hand. Since the money belongs to someone else, this emergency reserve is easier to defend from the daily requests of spouses, other kin, or neighbors than money of one's own. A bigger question is why the depositor chooses the money-keeper system, rather than one that might pay interest, as with a bank. Islamic prohibitions against interest, discussed more fully elsewhere (Shipton 1990a), are only part of an answer, since local loans with interest are countenanced in practice. When - 23 - asked what they liked most about the money-keeper system, farmers tended to talk of trustworthiness and safety more than of proximity or convenience.22 Since most deposit-taking goes on within villages and among kin, those who save this way are displaying trust not Just in the individuals to whom they are confiding their savings, but also in the networks and groups to which both they and the keepers belong: the mechanisms to which they would turn for support as a last resort if keepers refused to return the money. An implication for banks in the western Sahel is that if they wish to attract more customers, they may need more and better public relations in the honest sense of the term--not just better interest rates or accessibility, though these would probably interest customers in their services. It may help to recruit some elders where literacy or other qualifications permit, since local practice suggests they are more trusted than juniors.23 But banks and other formal institutions should expect only to supplement rather than supplant the local money-keeping system, since the latter has devoted practitioners. In the 1988 survey, for instance, informants who used the services of "informal" money-keepers were asked whether they were content with this system of saving. 94Z replied yes. Conversely, when asked if there were any problems with it, 98Z replied no.24 A bank might envy such user confidence. Deferred Wages Wage-earners have a related way of saving, which is to ask employers to withhold wages over short times to let them accumulate. (This is a system familiar to researchers hiring local research assistants. While some assistants in the city or countryside constantly request advances, others ask to defer - 24 - payments for months at a time.) Some agricultural wage laborers prefer to be paid only quarterly, and a recent study of Gambian share contracting in farming found that most workers prefer arrangements in which they receive their main compensation seasonally rather than at shorter intervals (Robertson 1987: 221). As in the village money-keeping customs described above, individuals entrust their money to others who they know are more solvent than they are, and who have something to lose in reputations in the case of failure to repay. Amounts Commonly Saved It is nearly impossible to determine the amounts of farmers' personal cash savings. Indirect questioning in the villages suggested that most farmers' liquid cash savings could be counted in the hundreds of dalasis (sometimes held partly in francs CFA) at any given point during the "trade season" from December to March, a few farmers' rising into the thousands; but that most have cash savingc of less than 100 dalasis at any given time uring the rest of the year. Men appear generally to have larger sums at their disposal than women have during the trade season, but women seem usually to save more than men, as a proportion of their wealth and perhaps also in absolute sums, in the lean season. Thus men's cash savings seem to fluctuate more than women's. One reason for this may be women's considerable involvement in non-farming economic activities outside the crop season, when many men are idle. A second may be that since men usually do their land preparation before women do, and may pre-empt the use of animal-drawn tools for their own fields, women often expect to have to hire labor in preparing their own. There are indications that elders tend to save more in cash than juniors. - 25 - Interviews indicated that most cash savings over about 100 dalasis for periods longer than a few weeks are made for specific target expenditures.25 People with thousands of dalasis in savings will usually be those with upcoming ceremonies like weddings to finance, or young men planning to travel to find work. Savingt in these cases will usually include earmarked rontributions from kin. Gambians may vary somewhat by ethnic group in saving habits. It is thought that some Serahuli and others engaged in trade as well as agriculture keep larger stocks of cash in their houses than other rural Gambians (though stereotypes may exaggerate these tendencies). They may use heavier boxes or even chained or cement-anchored safes. Mauritanian and other merchants tend to do the same. Many merchants, however, appear to re-invest cash quickly in merchandise rather than saving large amounts. Small shopkeepers commonly deposit cash savings with larger traders in their villages or in larger towns. IV. LOCAL MECHANISMS COMBINING SAVINGS AND CREDIT Kafo Groups: Village Women's and Men's Associations Normally every Gambian village has one or more groups known known as kafo (or kafoo) in Mandinka and some other Gambian languages.26 Though the term referred in the past to men's or women's age-grade organizations, and usually still does, it can refer more generically to various kinds of interest groups in a village.27 Today they may be organized around a specific activity like a sport or dancing, or loyalty to a political party; and they often ha%e multiple functions. - 26 - Membership in kafo groups is voluntary, but commonly it includes most persons of the eligible gender and age range in a village. As few as under 10 or, in a sizeable town, as many as over 300 members may belong.28 Nearly always, in a village, women's associations have a more comprehensive coverage than men's. Occasionally an individual can belong to kafo groups in more than one village.29 Rafo groups are commonly unnamed, or their names shift by activity (so 'the vegetable garden group" may also be 'the soap-making group" at other times). But the larger ones usually have a number of elected officers, organized in hierarchies sometimes modelled after governments of larger polities, often with presidents, secretaries, treasurers, messengers, and sometimes even 'police". Women's groups are not always wholly autonomous as such; commonly a group has a trusted male patron in its village who may hold the group's money or, with an advantage of literacy in English or Arabic, keep the accounts as a favor. While the head of a men's kafo usually plays a subordinate role to the village headman, the head of a women's kafo may be, by virtue of this position, the most powerful woman in her village. Kafo groups have traditionally served the function of mobilizing labor for free use by members, for hire to other villagers, or for community projects. Today they also variously serve other economic purposes. Some have collective fields for cash or food crops, and collective crop stores; others have fields with collective infrastructure but individuated plot and crop ownership. Some have small restaurants, hotels, or milling machines. Kafo groups save money from members' dues (often less than a dalasi per member, per week) and fines, from group labor hired out, or from dances, concerts, wrestling meets, or market stall sales they organize, among other ways. Some young adult groups have dances, - 27 - concerts, or sales as fund-raisers, members paying less than non-members. Some make charitable donations, which may take the form of hard goods, to hospitals or mosques. organizing rites of passage--naming celebrations, circumcisions, weddings, and funerals--is an important function of kafo groups. These, of course, are also ways of gaining power and prestige, and redistributive events for villagers and their guests. Kafo groups handle money in various fiduciary ways. Some allocate their group labor on credit. Some strong kafos save money in group bank accounts; they may use several senior officers as joint signatories to ensure no individual has private access to the savings. A kafo may lend or give money, food, seeds, or tools to members and their families, to or other villagers, members getting preferential terms. These borrowers sometimes use the food or money for their own ceremonies, sometimes for school fees or other purposes. Cash loans from kafo groups to individual villagers in 1987-8 usually ranged from about 10 to 100 dalasis, and more often than not they carried interest, which varied between about 10 percent and 60 percent (nominal rates) over six months. The interest was used by some to buy consumables like soap for distribution to members.30 Some groups take chattels as loan collateral. A common kind of kafo makes its purpose to serve the needs of members' personal or family emergencies. Such groups can mobilize contributions quickly for hospital trips, court bails, house rebuildings after storms, and the like. In an impoverished country, emergency kafo groups have an advantage over "formal" insurance or savings banking for emergencies, in that they keep the money in local circulation rather than draining it to cities. - 28 - Kafo groups vary substantially in organizational strength and command over resources. Some save thousands of dalasis and made substantial investments in community agricultural projects and purchases, build community centers, or finance political campaigns, while others seem hamstrung by problems including factionalism and graft and are unable to mobilize any capital. Often the women's kafo groups are more vigorous and economically more important than the men's, but any kafo is only as strong as the personalities of its leaders. Recently some private voluntary organizations in The Gambia have experimented with kafo groups as conduits for credit for various purposes, including small enterprises and market gardens. (Some of the kafo groups are new ones of the international organizations' own creation.) The Department of Agriculture has also used kafos as channels for seed and fertilizer loans for group farming ventures. Some kinds of kafo seem to hold promise as a link for connecting poorer farmers with institutional financial resources. A kafo's strength is its multi-functional nature. Since no one wants to be ostracized from a main village kafo--this can mean becoming a pariah--the group can sanction its members (by fines of a few dalasis) for not participating in its work gatherings; it may similarly be able to sanction loan defaulters. As yet, however, little is known about what kinds of kafos can do these things effectively.31 And since most have little experience lending large amounts of money to their members, so institutional advances in this direction should be cautious and gradual. - 29 - Rotating Savings and Credit Associations Contribution clubs for money are a newer idea in rural parts of The Gambia than traditional kafo groups for labor. A form that seems to be spreading outward from the cities and towns is the osusu, or rotating savings and credit association (ROSCA, also known by other terms, e.g. merry-go-round or the French term tontine). This is a local-level financial mechanism with parallels in all coastal countries of West Africa and indeed, as anthropologists and others have found, in patchy distribution all over the world (Geertz 1962, Ardener 1964, March and Taqqu 1986).32 In The C'mbia as elsewhere, it is most commonly organized by women. The basic principle is simple. All members of the osusu make fixed contributions of money at regular intervals (usually daily, weekly, or monthly, though 10-day intervals were also encountered). Each time, one member takes it all. Each member takes a turn until the cycle is completed; then it may recommence. For one whose turn comes early in the cycle, the groups are a credit mechanism; for one whose turn falls late, they are more important as a savings mechanism. (These differences are evened out if the grsup lasts through several cycles.) Since Gambian osusu groups usually involve no interest payments, a member who withdraws early in the cycle gains an interest-free loan, and indeed, if inflation is taker, into account, makes a real net gain. One who withdraws late in the cycle has given an interest-free loan and incurs a net loss; but osusu members appear to feel that the function of enforced savings is valuable enough to them to offset this loss. Gambian osusu groups seem to be especially important among market women in towns, and for junior-grade civil servants in Banjul and in the province, but - 30 - they also appear in farm.villages. Osusus operated in four of five villages studied during the time of research (the exception being a Fula village in Upper River Division), though some villagers everywhere knew the idea and some had belonged to other osusu groups when living outside their villages.33 In the 1987 survey, 17 percent of the women interviewed currently belonged to osusu groups, and only one (1Z) of the men. In the 1988 survey, farmers were asked whether anyone in their compounds had ever belonged to one. 192 percent said one or more women had belonged; 4 percent said one or more men had belonged. 89 percent of the respondents' osusus were in the respondents' home villages, 3 percent were in nearby towns, and 8 percent were in the cities of Banjul and nearby Serrekunda. Nearly all the groups in the villages were female-only, and many of these included only persons of comparable ages internally (many participants variously described the age-ranges as "young", "middle aged", or "old"). The groups in the towns and cities were more likely to include mixed sexes. Each women's osusu group represented in the sample existed within a single village, town, or city; and in the largest village they were composed of members of the same neighborhoods. Most rural village osusus consisted of women roughly the same ages (year of circumcision is sometimes used to determine association membership). Some consisted entirely of kin, others not. One group was based on ethnic identity, its members representing a minority in their village. The overlaying ties of neighborhood, gender, age, kinship, and ethnicity all seem to give the groups the capacity for peer-group pressure that may be needed to ensure members' regular participation. Each group did, however, have a recognized leader or organizer. - 31 - How big were the groups? The osusus in which infoi%ants (1988) themselves had participated contained between 7 and 31 members, with a mean of 13.4, in the osusus within informants' own village: between 7 and 50, with a mean of 24.8, in the town osusus; and 17 and 30 in two urban osusus of Banjul and Serrakunda. (Some other osusu groups of which we have learned in the cities are much smaller). Individual contributions in the osusus ranged from one to 20 dalasis weekly, or equivalent; but in 91Z of the groups each member paid between one dalasi and five dalasis weekly or equivalent, i.e., less than one U.S. dollar a week. A number of the women who belonged to osusus in the villages had some small sources of cash income in addition to seasonal crops, for instance small-scale fried food vending or palm-oil processing. The women in osusus had used their takings for a variety of purposes, including ceremonies (probably the most common use), clothing, livestock purchases, and (in one self-report only) seeds for mango trees.34 A male informant who had recently belonged to an osusu described it as being composed of 12 civil servants, living in three villages, who each contributed 75 dalasis monthly. (Apparently women civil servants in the cities and towns also make osusu contributions on this order of size, sometimes putting in a quarter of their salaries monthly.) Another had belonged to an osusu of young men in an army camp, each contributing 7.50 dalasis monthly. Little is known about how long osusu groups last in The Gambia or elsewhere. The oldest group within the sample was said to have functioned continuously for ten years, and others had worked for five or six. Many operated - 32 - only during the dry season each year, suspending operation during the rains when members were short of money; some others simply reduced the amount of contributions during the rainy season. In a group that suspended its main function seasonally, members continued contributing (a dalasi each) to each others' ceremonies during the rainy season, reflecting the valued social and symbolic roles of both the osusu and the ceremonies. Osusu groups in the Gambia are characteristic of the West African forms in that they do not involve variable individual contributions. The order of rotation is variously pre-set as the cycle begins--in one case, by members' ages --or determined by lottery upon each collection. In either case, the order may be broken when particttlar members encounter special cash needs. Members also make their own personal arrangements with each other to swap turns. The advantages of the osusu as a way of mobilizing capital are several; some of these are also enjoyed by the kafo group: 1. Osusu membership controls the 'squawk factor' in saving, as seen earlier, since gives the saver a socially acceptable excuse for denying gifts or loans to persons who do not belong. 2. The osusu provides financial services to people who may be ineligible for banking, for lack of initial deposit capital or other problems. 3. It normally requires no collateral. 4. It requires little or no paperwork or travel; in these ways it minimizes "transaction costs". 5. It is not intimidating, as financial institutions can be, to illiterate farmers. It need not involve language barriers. 6. It can provides an occasion for social gatherings. - 33 - 7. Since transactions commonly take place in group meetings, thefts and cheating can be difficult. 8. If the group wishes, it can supervise its members' expenditures of the funds, choosing a member to spend the money with or for the recipient each time for a pre-arranged purpose. 9. The osusu can serve as a form of insurance. 10. Because it arises from members' initiative, the osusu commonly benefits from a high level of commitment. 11. Where members belong to several osusu groups, thes.e lzelp to circulate money throuighout a population in useful sums. 12. Savings in an osusu are not subject to government scrutiny, control, or taxation. There are, however, disadvantages: 1. The year-round osusu is not ideally suited to communities depending on rainfed agriculture, because in these all members' needs for money are likely to come at the same season: the "seasonal covariance" problem. By the same token, Osusu groups are not very useful for financing agricultural inputs. 2. Even in the abundant season, one cannot always get the money when one wants it. Most osusus have flexible orders of rotation, but someone else's needs may be greater when one needs the money. 3. The group depends heavily on all members' continued participation. 4. The group canrnot easily accommodate members who migrate in and out of a community. 5. Savings and borrowings in an osusu cannot be kept secret locally. - 34 - Size Considerations in ROSCAS The optimum size of an osusu, or the range of acceptable sizes, may vary with the task, and with the characters of the members or leaders. It might also very by gender, by ,thnic group, and by age: little is yet known on these questions. If the group is too small, these problems arises (1) The potential benefits of joint savings or co-operation are too small to repay the efforts of joining up; (2) each member can simply avoid the other(s) when he or she has not been co-operating. The smaller the group, the easier the avoidance. (3) There is not enough likelihood that the members also belong to other organizations in common, e.g. sport groups, age-sets, etc., through which to apply pressure. (4) The potential damage of a member's quitting or moving away is too great. If the group is too big, these problems arise: (1) Each member will lose only a very small fraction of his or her investment by the non-cooperation of one other member. There is little incentive to pursue him or her, or work to enlist other to apply pressure. (2) Where the group functions on a principle of rotation, it is too infrequent that each member gets his or her share of the group's rewards. (3) Communication between members gets too hard. (4) It is too likely that some members will move away. The optimal sizes or acceptable ranges for solidarity groups seem to vary widely, and may depend on the nature of the activities, but in many societies around the world, 5-8 people seem to be an especially appropriate range for saving and loan projects. Many indigenous Gambian village kafo and osusu groups are rather larger, and the local models should probably be followed in interventions.35 - 35 - While urban savings are beyond the scope of this paper, it is worth noting that most saving methods described above, including the kondaneh boxes, the money-keepers, and rotating savings and credit associations (osusu groups), are also common in the cities of Banjul and Serrekunda. Osusus are common, for instance, among secretaries and messengers of government ministries, and not just among people without access to banks or other "formal" financial institutions.36 Whether to Act Upon ROSCAS In development agencies there is now considerable interest in finding ways of 'plugging in' external funding to rotating credit and savings associations, though there is no evidence in The Gambia and very little elsewhere that this is approach can succeed. Certainly, the largely female composition of these groups and the peer-pressure elements are attractive features to donors. The incorporation of savings with credit is something many lenders are now striving for, and this principle is already woven into the fabric of the osusu. It is not clear, however, that men would not take the groups over if substantial amounts of cash were infused from outside. It is likely, too, that new groups would spring up overnight to receive funds, without the overlapping bonds of kinship and .s:;.er social ties that make osusu groups work. Perhaps the surest way in which institutions can work with osusu groups is simply by teaching about how they work, in schools or other groups. But this should probably not be seen locally as a government initiative, for part of the appeal of the osusu to its members is that it is truly their own. Osusu groups are functioning comparatively well, though the resources they muster are limited; and in the end the best way of helping them might be to leave them alone. - 36 - The osusu does, however, have features worth emulating in the design of institutional credit schemes to serve complementary functions. The social pressure within a group of about a dozen people to repay (and even to use credit wisely) is clearly a desireable feature. At the primary level, at least, the number of farmers who have collective responsibility for repayment should be limited: the members of the group should know each other and have other kinds of influence to bear upon each other. Second, the small group should probably have a single leader, chosen by the members, as the osusu has. Third, the integration of savings and credit--they are the same transactions in the osusu --gives members the pride of accomplishment: they earn what they borrow. They have a financial investment in the group's success. Other General Lessons from Informal Financial Institutions Kafo and osusu groups are found among both men and women, but in the rural areas, they seem more important to women; and the associations that succeed in rural financial matters usually are usually single-sex. There are lessons in this. Gender and age specificity of rural kafo and osusu groups cut down the chances of great disparities between members' solvency or financial abilities. A rough comparability in turn helps prevent some members from easily domineering or exploiting the others, though of course either can still happen.37 Gender and age specificity also make it more likely that members will have other kinds of contacts with each other, and thus ways of devising their own sanctions for non-cooperF.tion. In external interventions, dividing projects for women and men into separate but parallel projects helps keep the control over gender distribution - 37 - in an agency's hands, and helps to prevent men's monopolizing resources. A similar principle may apply to age groups. Dividing project resources between younger and older people's kafos can help ensure a more balanced distribution. (Leaving out one group or another, however, may induce it to try to sabotage a project.) The experiences of private voluntary organizations in The Gambia, as elsewhere in Africa, suggests that credit to women is not necessarily harder to collect than credit to men. Provided women's borrowing groups are appropriately constituted, it may even be easier. For leadership and group membership, insiders can choose each other on character better than outsiders can choose them. They are also, of course, more aware of interpersonal "chemistry' in the groups: potential harmonies or clashes between the personalities. This may take a lifetime of mutual acquaintance. Local leaders such as alkalos and women's kafo heads are usually chosen with much deliberation; it generally makes sense for projects to use these local leaders as contact points, or as pressure points if necessary. The local-level groups most likely to succeed in mobilizing or managing finances are those that exist already for other purposes, such as farm-work groups, youth activity groups, sport groups, etc. The groups' having more than one function ensures that members have ways to exert pressure on each other when needed. They will usuallv be bounded in membership, for instance by residence in a single village. SimplicIty seems to be a main key to success in local financial organizations. They tend to have clearly delimited membership, and to use strict and regular meeting schedules, standard contributions, standard fines for non- participation, and round numbers. - 38- Probably nothing should be issued on credit without some contribution from the borrowers themselves, whether in the form of capital, labor, or local materials. Such a rule is important for several reasons: (1) It screens out borrowers who are not serious, capable, or organized. (2) It gives borrowers an investment in the enterprise or development, and thus (3) it adds their scrutiny to the financial management of local project organizers, an important factor in remote areas hard to monitor or supervise regularly from the capital city. (4) Demanding a contribution also helps to accustom farmers not to expect free-hand-outs in the future. Projects will succeed only when rural people consider them as substantially their own. Finance Based on Groups and Individuals A general observation emerging from studies like ours on rural financial systems in The Gambia and other West African countries is that group-based saving and credit systems tend to work more smoothly than individual-based ones among the rural poor, and particularly among women.38 One reason, as seen earlier, is that they allow individuals to save without seeming selfish to their neighbors and kin, an important point in most of rural West Africa. Another is that the groups and their leaders can serve as conduits to institutions like banks, which ordinary rural people may perceive as inaccessible, and may reassure the members who feel that approaching institutions is risky. Group-based saving and credit does not mean resources cannot be directed to individual members--often they can and should--but rather, that local groups are used as units of collective responsibility for collections and repayments. - 39 - V. BANKING The commercial banking system is now almost useless for rural people. Gambian farmers perceive banks as remote, intimidating, and not fully trustworthy, and by and large they do not go to them.39 The senior officers of the banks in The Gambia, on the other hand, imagine smallholding farmers farmers as too poor and financially inactive to be of interest, and not particularly trustworthy as borrowers. There are only two up-river branches: outside Banjul and nearby Serrekunda and Bakao, commercial banks are found only in Farafeni and Basse.40 The numbers of rural bank depositors clearly reflect the distances, social and geographic. A 1975-6 survey of rural Gambian villages found that only two (2Z) of 94 informants had bank savings (Dunsmore et al. 1976: 307). In our 1987 survey, four (3Z) of 138 male and female informants stated they had deposits in commercial banks; four others, again 3Z, said they had money in post office savings. Only one of 69 women interviewed had a savings account in either a bank or post office savings. Branch bank information confirmed that most farmer depositors hold amounts below about 250 dalasis, that very few of them are women, and that almost no farmers borrow from banks. In May 1987 the Basse branch of the GCDB (one of only two branches of the bank outside the greater Banjul area) reported that 492 of its 858 savings depositors had less than 250 dalasis in their accounts (i.e., less than the value of one sheep). Interviews with bank staff suggested that most of the roughly 100 farmers (people whose primary source of income is farming) saving at the bank were among those with less than 250 dalasis. Only - 40 - an estimated 10-15 percent of the 858 depositors were women. Only about five farmers were currently borrowing from the branch. Since saving at home or with money-keepers, instead of with financial institutions, costs farmers the opportuiity of earning any interest (lately 152 in the banks), the problems with banking appear serious. When asked in a neutral way about their reasons for not using banks, farmers in the villages surveyed cited several: the distance and prohibitive travel costs, inconvenient opening schedules, requirements of minimum deposits, difficult paperwork, and distrust of clerks. Some cited with bitterness their recent experience with the Agricultural Development Bank, which had taken deposits in the early 1980s and ceased operating without refunding them. While Islam, strictly speaking, forbids interest charges, the fact that many rural Gambians do charge interest in informal lending (and do so with a variety of legal devices, hiyal in Arabic, to dodge the prohibitions--see Shipton 1990a) suggests that the interest- collecting aspect of savings banking is unlikely by itself to stop savers who would otherwise be interested in banking.41 Rural Gambians are generally more concerned about security of savings, and about availability for emergencies, than about interest earned or forgone. They will also save money with institutions or clubs (where they do) in order to gain access to credit when needed. But many are unaware that banks pay interest on savings. And most do not have access to the kind of information or schooling that would let them calculate inflationary losses and interest opportunity costs. Many cannot appreciate fully how quickly they lose purchasing power by hoarding money or depositing it with keepers who pay no interest. - 41 - Though few farmers use banks now, most interviewed responded quite positively to the idea of mobile banks, saying they would like to see these tried, or tried again, in their areas.42 Many said they would prefer to see small branch banks set up near them, since they would have doubts abou- putting their money into something that wuuld roll away. But clearly, inaccessibility has been a major barrier to institutional savings. The commercial banks are the existing organizations with the most managerial competence to handle a system of rural branch banks. Unfortunately they are the organizations likely to be least interested in serving rural people, and so may need added incentives and safeguards to be provided by large international development organizations. Informal discussions with some officers of large agencies suggest potential interest in providing guarantees. If banks are to serve Gambian farmers, and particularly women, it is likely to be by group access through kafo groups and other village-based associations. Private voluntary organizations (PVOs, or non-governmental organizations) may find a useful role in making banks more accessible to women's groups: explaining the pros and cons of banking, and serving as screening mechanisms for the banks, on one hand, and bureaucracy-cutters for the farmers on the other. Brokering the two-way flow of information is a role the PVOs can more easily serve than banking itself, since most lack the funds to keep trained money managers on their staffs. The large agencies, for their part, can provide guarantees to back up new bank experiments, and they can provide workshops for coordinating activities of various PVOs. - 42 - Agricultural Seasonality and Savings Banking If institutional saving facilities set up more rural branches, mobile or stationary, seasonality will put them to a stern test. The predominance of rainfed agriculture in the economy and the covariance of incomes make financial intermediation hard. Most farmers are likely to deposit at the same time (the trade season, December-February) and to withdraw at the same time (June- September). Moreover, most farmers who borrow are likely to do so at the same time of year all are making withdrawals, and not to repay until the following trade season. A problem is that a bank has stronger financial incentives to make the rounds (if using mobile branches) or stay open (if using stationary ones) in the 'trade season", when farmers have significant sums of money to deposit, than during the rest of the year, when farmers are more interested in withdrawals. A bank able to use Gambian capital domestically or overseas during the period of savings, and to import capital during the period of withdrawing and borrowing, might be most able to keep its system afloat. So banks with international links may be best placed to accommodate Gambian seasonality. An Untried Strategy for the Adventurous While the issue of interest rates in institutional saving cannot be treated in detail here (see Shipton 1990a), some possibilities must be noted. If interest rates on commercial bank credit are raised from their present rates of 27 to 28 percent per annum (as of October 1989), interest on savings deposits (12-14z) should be raised too. The low or negative real interest rates (inflation was 12.5Z in 1987/8) may help explain why bank savings have been so unattractive for potential rural depositors.43 - 43 - Subsidized credit has been the norm for financial institutions in The Gambia, as in other poor countries. Consistently, staffers of large aid agencies have sought in this way to dispose of large amounts of money (see Tendler 1975), and statespersons and local politicians, to dispense patronage and win supporters. The approach has brought about familiar institutional problems (see Von Pischke et al. 1983, Adams et al. 1984). Some analysts believe as a conclusion, or as an ideological principle, that subsidies generally should be reduced or eliminated, as has gradually been happening in The Gambia. This "pro- market" approach has indeed become a new orthodoxy in financial agencies, and there is some practical sense in it, but credit subsidies persist across Africa. Apparently no institution in The Zambia (or perhaps even in Africa) has yet been documented to have tried a radical reversal, subsidizing savings rather than credit, as a rural development strategy. Low ceilings on such deposits might discourage abuse by the rich. The farmers and rural poor would become the "moneylenders'. The approach would probably require some cooperation between banks, private voluntary organizations, and (as guarantors at least) large aid agencies. There would be pitfalls to avoid, of course, and transaction costs would be high.44 Whether such a program would be administratively workable, and whether subsidizing it would cost international lenders and donors more than costly conventional credit schemes, are unknown. This bold approach is not for the faint-hearted or insolvent institution, and certainly, any such project should be small and experimental at first. It should be remembered, in establishing more banking opportunities, that not all rural people want them. Many are satisfied with saving in the forms of livestock, jewelry, etc. These preferences should be respected. If it does not - 44 - pay substantial interest to depositors, savings banking does little more than sap wealth from rural communities to towns and cities. And interest, even if redefined as *profit", itself involves religious and moral quandaries in a mainly Muslim country. Rural Gambians are not of a single voice in demanding savings facilities. What they want is more options. VI. SUMMARY AND CONCLUSION: BALANCING CREDIT WITH SAVINGS To conclude, rural Gambians save in diverse ways. The diversity and complexity of rural savings have their own rationalesi and currency instabilities make it sensible to put savings into other forms. Saving, like borrowing and lending, is part and parcel of social life; and any institut4on that becomes involved in Sahelian rural finance is also involved with networks of kin, neighbors, and age-mates. It is untrue, as sometimes said, that rural Gambians do not carry over wealth from year to year. But most of what they do save over periods of years is in non-monetary forms. They choose these forms for many reasons: to avoid inflation or political and regulatory control over their assets; to gain the benefits of milk, manure, and traction that animals provide, or to gain increases through livestock breeding; to consolidate their savings into indivisible forms or forms their relatives may help keep them from spending rashly; to avoid guilt associated with overt interest earnings; and to cut their losses to collapsing banks. Diversifying savings into many forms means reducing risks. It also means keeping the wealth less conspicuous, both to other local people and to tax collectors. There is obviously sense in these strategies, even though some of - 45 - the means of saving that farmers choose provide them sub-optimal earnings in monetary terms. It is also untrue that Gambian farmers are uninterested in monetary savings. The kondaneh boxes, the rural deposit-takers, and the requests for infrequent wage payments all testify that many rural people want ways of saving cash. So too does rural people's liking to convert small notes to large ones (as men try to convert small livestock to large stock). While some economists assume farming people to have a "liquidity preference", these patterns indicate the opposite, an "illiquidity preference". Part of the reason why rural Gambians seem to save little of their wealth in money is that there are social pressures against hoarding during times of general scarcity. This is why group savings approaches appear more likely to succeed than approaches based merely on individuals: they make it possible to save without appearing selfish. This principle, at the root of the osusu, may be worth incorporating into institutional financial systems as well. The positive responses to questions about mobile banking (and small rural branch banks) suggest a practical path for public or private agencies to follow, and untried adjustments in interest policies suggest room for experimentation. More broadly, planners in institutions wishing to assist Gambians to save resources should consider interventions in crop storage--incidentally an approach low in recurring expenses--or livestock health. The aim should not be to "capture" local savings, but to augment and supplement them. Credit is not the answer to all financial problems in development, and it is perhaps one that both international and national institutions have tended to rush into too readily. The most obvious fact about credit seems ironically to - 46 - be the one most often overlooked; credit indebts. People who have borrowed feel behind; the English phrase "in the hole" expresses a position rural Gambians know too well. As some farmers explained it, one who has a loan to repay is working for someone else, and this drains incentives to work.45 (This is particularly so where the lender is perceived as an alien institution.) That Gambians borrow less from local traders after the good years than after the bad, and that few of the richest farmers interviewed borrowed from these at all, suggest again that rural Gambians basically do not like to be in debt to strangers (especially at the kinds of interest rates that make institutions financially viable). They know the "debt ratchet" principle, the vicious circle whereby interest charges can progressively impoverish farmers and make them increasingly dependent on credit. That the co-operatives have witnessed heavy demand for credit seems to be largely due to the easy lending terms offered, and to a history of lax collection. Gambians seem to feel they need both private lenders and co-operatives for credit, but they are not happy with either. Nor are the motives behind all institutional credit entirely altruistic. In agencies, government, and cooperatives, credit seems to be used at times as a way of moving money to satisfy committees and office superiors, and perhaps as a way of extending political patronage to regional or local power-brokers or a voting public. These things do not mean that credit cannot do its recipients any good, but that it may be continued too readily when it does not. If it is decided to extend easy gifts for political reasons, these should be known as grants and not loans, in order to give true credit programs a chance. - 47 - If institutional interest rates are raised to reflect real costs in institutional credit, in the interests of financial viability. it may be worthwhile at the same time to re-assess critically the role of credit in general. Where interest rates are based on real costs, borrowing is more expensive than saving.46 Farmers who cannot afford to save can even less afford to borrow. The rural financial mechanisms working least well are those designed for credit only. It has been the habit of most large and small development agencies with resources to dispose of, to look for ways to lend, rather than to examine local resources that might be mobilized, or to do both. (There are a few encouraging exceptions.) Rural people consider most of the resulting credit projects to belong to someone else: the capital city, the Government, the whites. By contrast, In The Gambia as elsewhere in Africa, the credit systems that appear to work most smoothly, like the informal osusu groups, appear to be those that incorporate savings and link them with lending. Savings in the osusu are the same as credit (the same transactions functioning as one or the other, for different members at different times). The groups embody several other sound principles: local initiative and capital, character screening by members, peer group pressure as the main means of enforcing co-operation and repayment. Osusus spring up by voluntary initiative in rural towns and villages, as in cities. Although some suspend operation seasonally, some last well over a decade, and as they dissolve, new ones form. These systems rural people consider their own. These observations suggest an incorporation of savings components in credit programs wherever the institution in question is financially sound enough to merit farmers' investment, and wherever it can build upon existing social - 48 - networks.47 Much remains to be learned about how, and whether, financial institutions can emulate the financial principles observed already working in the rural areas. Whether savings and credit are linked or not, however, farmers clearly need some of each, and in a better mixture than they have them now. In the long run, the rope of credit and debt without the box of saving can scarcely serve farmers' interests. Rural Gambiar.s aeem to be saying they need balance in their financial livess between s.ving and credit, between liquidity and illiquidity, and between individual and group action. - 49 - TABLE l: Preferred Means of Saving Money 1988 Men Women Both Keeping for self 47Z 462 47Z Banks* 31Z 19Z 25Z Money-keepers 162 212 18Z P.O. Saving 2Z OZ 12 Burial 2Z OZ 1Z Other 22 14Z 82 TOTALS 1002 1OOZ 1OOZ nm 51 52 103 (* Notes Preferred means stated differ can differ from means actually practiced. Far fewer than the percentages indicated acutally used banks). - 50 - TABLE 2: Relationships of Informal Money-Keepers To Depositors 1988 Money-keepers Depositors Interviewed Male Female Both Kin: Spouse 1 10 11 Parent 4 5 9 Sibling (full) 0 2 2 Offspring 0 0 0 Other kin 6 5 11 Subtotal, kin 11 (552) 22 (79Z) 33 (692) Non-kin 9 6 15 TOTALS 20 (lOOt) 28 (1002) 48 (1002) - 51 - TABLE 3: Genders of Depositors and Money-Keepers 1988 Depositors Money-Keepers Interviewed Male Female Both Male 16 (80X) 4 (20Z) 20 (100?) Female 18 (64Z) 10 (36Z) 28 (1002) Total 34 (71Z) 14 (292) 48 (100Z) - 52 - FOOTNOTES 1. The calculation includes non-agricultural as well as agricultural debts. 2. The fifth village, being too large tc cover, required a more limited but systematic sample of one side. 3. See Shipton 1987, Appendix on research methods, for more details. 4. Our informants in the 1988 survey, by language, were Mandinko (36Z), Fula (28Z), Wolof (21Z), Jola (82), Serahuli (42), and others (3Z). In the 1983 national census, not a wholly reliable source but the best available, the largest ethnic groups in The Gambia were listed as follows: Mandinko (40.4z), Fula (18.8Z), Wolof (14.62), Jola (10.32), Serahuli (8.2Z), and others (7.7z) including Arabic-speaking Mauritanians (Gambia 1986: 8). International seasonal labcr migrations, among other things, make census-taking difficult. Some have also questioned whether the national census may have under-enumerated Fula for political reasons. 5. I hope to discuss in another study some of the interethnic similarities and differences in Gambian economic life. 6. The clash of cultures in the area of interest rates and ratios is discussed more fully in Shipton 1990a. 7. Patrilineages are kin groups related through males. The castes are in-marrying occupational categories whose members, recruited by heredity, may or may not practice the specified occupations. See Diop 1981 on Wolof castes and their economic importance. 8. See the essays in Moock 1986 on other parts of the continent. 9. For livestock counts see Derman et al. 1985: 112; Mills et al. 1988, Sumberg 1988, Sumberg and Gilbert 1988. As elsewhere in Africa south of the Sahara, livestock loans and (among some groups) secrecy about numbers owned makes livestock tallies difficult. 10. Marriage payments among Gambian peoples are paid mainly in cash and manufactured goods, however, not in cattle as among many eastern African groups. 11. Analogies between livestock and cash are common. When I once asked a Mandinko informant why he preferred large to small stock, he asked, "Would you rather have a 10 dalasi note or a 25?' -- an argument he knew was hard to refute without pedantry. Mandinko say their grandparents historically referred to livestock as fangkanta (lit. self-guard), or reserve; they use the same term now for personal cash savings. 12. This applies to other permanent, heritable lineage property like compounds. - 53 - 13. West African literature refers often to earrings as wealth. See, for instance, Camara Laye's autobiographical Dark Child, in which the father is a goldsmith, and Sembene Oumane's novelette and film, The Money Order, in which they are pawned to a usurious Mauritanian merchant in Dakar as a desperation measure, and lost. 14. Only one such transaction was reported in interviews. 15. In a preliminary count of data from the two villages studied in Western Division, women who grew groundnuts harvested an average of 732 as much as 'men who grew them. On average, the men sold 73Z of their crop, stored 202 as seed, and stored 32 as food. The women sold 652 of theii crop, stored 132 as seed, and stored 232 as food. While only 272 of the men who grew groundnuts kept any after harvest for use as food, 92 percent of the women groundnut growers kept some for food. All the men growers and 912 of the women growers kept some for seed. Of course, some groundnuts originally intended as seed get consumed before planting time. 16. Because the volume of a storehouse increases faister than the surface as its size increases, by simple geometric principle, there is an economic advantage in larger storehouses. But, as seen later, there are social disadvantages in stores for more than one village. 17. Survey response percentages in this paper refer to the 1988 survey unless otherwise indicated. Figures given are "valid percentages", i.e., they refer to totals excluding "no answer' or "unintelligible" responses and responses discovered in cross-checks to be inaccurate. Figures are rounded to the nearest percentage point. 18. Saving by burial has long occurred elsewhere in Africa. The late President Jomo Kenyatta of Kenya once reported on the Kikuyu, 'they say it is better to buy a sheep or goat instead of shillings which, if buried in the ground (the only form of saving money the majority of the people know), would rot and lose their value" (1965 [1938]: 66). 19. This is not, of course, to say all Gambian marriages are suspicious or hostile. They vary as anywhere. But ties between spouses often seem more distant than some "blood" kin ties, and cash is more likely than other commodities to be privately stashed. As discussed below, spouses sometimes do specifically entrust money to each other for safe keeping. See Field 1940, Moock 1986, and Shipton 1989 for other discussions of the separation of resources between spouses elsewhere in Africa south of the Sahara. 20. For comparisons frora Senegal and Niger, respectively, see Tuck 1983: 66; Graham et al.1987: Ch. IV, 18-21. 21. The difference between the figures is more than men's superior access to banks as alternatives might account for. It suggests that men have greater confidence in their own saving means, perhaps, or that women are more readily accepted as depositors. - 54 - 22. In the 1988 survey, farmers who used money-keepers were asked what they liked about this system of saving. 62? cited safety or trustworthiness. 18Z said they liked being able to withdraw whenever they chose, 7Z cited proximity of the moneylender, and 13? had other answers. Of course, these reasons mentioned are not mutually exclusive. 23. Whether men should be recruited in preference to women, to reflect local financial practice, or the reverse of each, to pursue political ideals, is an important related ethical question. As so often, foreign agents have here the awkward choice of being sexists or cultural imperialists. 24. One infirmant mentioned the problem of risk. Opinion polling of this sort is only of limited use in most African contexts because informants' finely developed etiquette can produce much second-guessing. Also, of course, this sub- sample was self-selected as people who liked money-keepers. 25. Little research has been done on attitudes toward money notes and coins themselves and their denominations. Gambian coins in the smallest Gambian denominations (in multiples of bututs, hundredths of a dalasi) are sometimes used by themselves, especially by women and children, but men and others who save large amounts like to convert their money into the largest notes possible, to help protect themselves against temptations. Printing more notes in denominations larger than 25 dalasis might not just serve the convenience of bank clerks and the rich (who queue up in urban banks to count out many hundreds of notes for deposits or withdrawals) but also help and encourage cash saving among poorer people, for better or worse. There would be dangers, however, in problems of innumeracy. 26. The Mandinka term Kafo (corresponding roughly to yirde in Fula, morom in Wollof) is sometimes used by non-Mandinko. In Fula, kafo groups are more commonly called compin, from "company". 27. This discussion of village kafo groups has benefitted from shared preliminary findings of surveys by E. Bazalgette, F. Le Grand, and A. Sillah. Separately, they and I have interviewed on the topic in over 40 Gambian villages. Mandinka villages commonly have three age-graded kafo groups for men and two or three for women. For a cross-ethnic comparison of Gambian age-sets and age- grades, see Sonko-Godwin 1986: 36-50; see also Dunsmore et al. 1976: 277-8. Much more is available in the literature on Senegal. 28. E. Bazalgette and F. Le Grand, personal communications. 29. Some women who have married near their natal homes belong to the kafos of their parents' and husbands' villages. 30. One powerful woman villager interviewed had borrowed several hundred dalasis from a kafo and re-lent the money to other villagers at double the interest. 31. See March and Taqqu 1986 for insights gained in other countries. - 55 - 32. Literature concerning osusus and related financial mechanisms in West Africa appear in Rural Africana no. 2, Fall 1978. 33. Tuck reports that in her sample of 54 Senegalese households in 11 rural villages, 342 of the households had members in rotating savings and credit associations. She finds that "only women belong". (1983:68) 34. Tuck reports that in her Senegalese sample, 701 used osusu funds for ceremonies, 20X for buying clothes, and 1OZ for buying "general merchandise" (1983:69). 35. The optimum size, or the range of acceptable sizes, may vary with the task, and with the characters of the members or leaders. It might also vary by gender, by ethnic group, and by age: little is yet known on these questions. 36. In 1988, lower ranking Gambian workers in both the Ministry of Finance and Trade and the USAID mission in Banjul had unsupervised osusu groups. Why salaried workers with access to banks use osusu groups too (or instead) is a useful question for further research. It should also be asked the other way around. Convenience, diversification to avoid risks, and secrecy about one's total wealth are likely reasons. 37. That mixed-sex kafo and osusu financial groups are more common in the cities and large towns appears to result partly from the fact that salaries put some men and women, particularly single ones, into comparable financial circumstances. 38. For some case studies see Mann et al. 1989. Notable institutional experiments in group-based lending have included "solidarity groups" of Accion International in Latin America and the equally famous Grameen Bank in Bangladesh. 39. In 1986, 48Z of the money supply of The Gambia was reported to be outside of the banking system (Ramamurthy 1986: 25). 40. The postal savings system is marginally more accessible but still hard to withdraw from, and useless as a source of credit. 41. A few, including Mauritanian shopkeepers, said that they did not bank because to do so would be to charge interest, an act discouraged in Islam. However, since some of these merchants do charge interest in informal lending, and since Mauritanians in Senegal and The Gambia have recently been victim to persecutions requiring temporary flight, the stated reason is probably not to be taken at face value. 42. In the 1987 survey, when farmers were asked whether they like the idea of mobile banks, 60Z replied yes, 112 replied no, and 292 were uncommitted, some of these lacking familiarity with the idea. The Agricultural Development Bank had tried mobile banks in the early 1980s, but it was unable to sustain them before it failed. 43. The inflation and current interest rate figures are from the Ministry of Finance and Trade. - 56 - 44. These might be minimized by a group-based approach. 45. This is doubtless what Shakespeare meant by "borrowing dulls the edge of husbandry" (Hamlet I, 3: 75). 46. Except, of course, where loans are not collected. Some of these factors may help explain why Gambians have tended to use institutions for borrowing more than for saving. 47. Clark (1987) concludes that the Gambian co-operative system is now too precarious to merit farmers' investment, and therefore to deserve any major savings component in credit at present. 57 Bibliography Adams, Dale W., Douglas H. Graham, and J.D. Von Pischke, eds. 1984. Undermining Rural Development with Cheap Credit. Boulder, Colorado: Westview Press. Ardener, Shirley. 1964. "The Comparative Study of Rotating Credit Associations". Journal of the Royal Anthropological Institute, Vol. 4, no. 2, pages 201- 229. Clark, Randolph H. 1987. A Study of Agricultural Credit Operations of the Cooperative Movement in The Gambia. Report to the United Nations Food and Agriculture Programme, TCP/GAM/2303 (Ma). Colvin, Lucie, Cheick Ba, Boubacar Barry, Jacques Faye, Alice Hamer, Moussa Soumah, and Fatou Sow. 1981. The Uprooted of the Western Sahel: Migrants' Quest for Cash in the Senejambia. New York: Praeger. Cuevas, Carlos. 1989. Rural Household Finance in the Kindia and Nzerekore Regions of Guinea. Report to the World Council of Credit Unions. Demissie, Asafa, Lyle Brenneman, and Jeffrey Nash. 1989. "Study of the Operations and Management of the Gambian Cooperative Union". Report to the Harvard Institute for International Development and the Ministry of Finance and Trade, The Gambia. Derman, William, Frank Casey, Sarah Lynch, Cynthia Moore, and Charles Steedman. 1985. Rural Development in the Gambia River Basin. Ann Arbor: Center for Research on Economic Development. University of Michigan. Diop, Abdoulaye-Bara. 1981. La Societe Wolof: Tradition et Changement. Paris: Karthala. Dunsmore, J.R., et al. 1976. The Agricultural Development of The Gambia. Land Resource Study 22. Surbiton, England: Land Resources Division, Ministry of Overseas Development. Field, M.J. 1940. The Social Organization of the Ga People. London: Crown Agents for the Colonies. Gambia. 1986. Population and Housing Census, 1983. Provisional Report. Central Statistics Department, Ministry of Economic Planning and Industrial Development. Gamble, David. 1979. A General Bibliography of The Gambia. Boston, Mass.: G.K. Hall. Geertz, Clifford. 1962. "The! Rotstirng Credit Association: A Middle Rung in Development". Economic Development and Cultural Change, 10(3): 241-63. 58 Graham, Douglas H., Carlos E. Cuevas, and Kifle Negash. n.d. Rural Finance in Niger. Report to the U.S. Agency for International Development. Haswell, Margaret R. 1975. The Nature of Poverty: A Case-History of the First Quarter-Century After World War II. New York: St. Martin's Press. Hopkins, Nicholas S. 1971. Maninka Social Organization. In Carleton T. Hodge, ed., Papers on the Manding. Bloomington: Indiana University, Research Center for the Language Sciences; The Hague: Mouton. Pages 99-128. Hyden, Goran. 1973. Efficiency Versus Distribution in East African Co-operatives. Nairobi: East African Literature Bureau. Kenyatta, Jomo. 1965 (1938]. Facing Mount Kenya. London: Heinemann. Kropp, Erhard, Michael T. Marx, Ballurkar Pramod, Benjamin Quinones, and Hans Dieter Seibel, eds. 1989. Linking Self-Help Groups and Banks in Developing Countries. Erschborn, West Germany: Deutsche Gesellschaft fuer Technische Zusammanarbeit. Laye, Camara. 1955. The Dark Child. London: Collins. Lele, Uma. 1975. The Design of Rural Development: Lessons from Africa. Baltimore: Johns Hopkins University Press. Linares de Sapir, Olga. 1970. "Agriculture and Diola Society". In Peter F.M. McLoughlin, ed., African Food Production Systems. Baltimore: Johns Hopkins University Press. Pages 193-227. Mann, Charles K., Merilee S. Grindle, and Parker Shipton, eds. 1989. Seeking Solutions: Framework and Cases for Small Enterprise Development Programs. West Hartford, Connecticut: Kumarian Press. March, Kathryn, and Rachelle Taqqu. 1986. Women's Informal Associations in Developing Countries. Boulder, Colorado: Westview. Mills, Bradford F., Mohamed B. Kabay, and Duncan Boughton. 19b8. "Soil Fertility Management Strategies in Three Villages of Eastern Gambia". Unpublished report, Gambia Agricultural Research and Diversification Project, Banjul. Moock, Joyce Lewinger., ed. 1986. Understanding Africa's Rural Households and Farming Systems. Boulder, Colorado: Westview. Ousmane, Sembene. 1983 (1965]. The Money Order. London: Heinemann. Ramamurthy, G.V. 1986. "Agricultural Credit Policy and Structure: The Gambia". Report to the United Nations Food and Agriculture Organization, Technical Cooperation Program. Robertson, A.F. 1987. The Dynamics of Productive Relationships: African Share Contracts in Comparative Perspective. Cambridge: Cambridge University Press. 59 Shipton, Parker. 1985. Land, Credit and Crop Transitions in Kenya: The Luo Response to Directed Development in Nyanza Province. Ph.D. thesis, Cambridge University. Ann Arbor, Michigan: University Microfilms International. Shipton, Parker. 1987. "Borrowers and Lenders in The Gambia: Preliminary Report on a Study of 'Informal' Financial Systems in Some Sahelian Farming Communities". Report to the Ministry of Finance and Trade, The Gambia. Shipton, Parker. 1989. Bitter Money: Cultural Economy and Some African Meanings of Forbidden Commodities. American Ethnological Society Monograph 1. Washington, D.C.: American Anthropological Association. Shipton, Parker. 1990a. "Time and Money in the Western Sahel: A Clash of Cultures in Gambian Local Rural Finance". ,Forthcoming in Michael Roemer and Christine Jones, eds., Markets in Developing Countries. Washington D.C. and San Francisco: Institute for Contemporary Studies, for the International Center for Economic Growth. Shipton, Parker. 1990b. "Borrowing and Lending in The Gambia: Social Distance and Local Credit Systems." Forthcoming in Lawrence White, ed., Financial Sector Reform in Africa [tentative title]. Washington D.C. and San Francisco: Institute for Contemporary Studies. Sonko-Godwin, Patience. 1986. "Ethnic Groups of the Senegambia Region". Banjul: Government Printer. Sumberg, J. 1988. "Notes on Cattle, Sheep and Goats". Report, Department of Animal Health and Production, Ministry of Agriculture, The Gambia. Sumberg, J., and E. Gilbert. 1988. "Draft Animals and Crop Production in The Gambia". Report, Department of Animal Health and Production, Ministry of Agriculture, The Gambia. Tendler, Judith. 197i. Inside Foreign Aid. Baltimore: Johns Hopkins University Press. Tuck, Laura. 1983. "Fcrmal and Informal Financial Markets in Rural Senegal". Report to the U.S. Agency for International Development. United States Agency for International Development. 1973. Spring Review of Small Farmer Credit. Vol. VI: Small Farmer Credit in Africa. Washington, D.C.: USAID. Von Pischke, J.D., Dale W. Adams, and Gordon Donald, eds. 1983. Rural Financial Markets in Developing Counties. Baltimore: Johns Hopkins University Press, for the Economic Development Institute of the World Bank. 60 Weil, Peter. 1971. 'Political Structure and Process among the Gambia Mandinkas The Village Parapolitical System". In Carleton T. Hodge, ed., Papers on the Manding. Bloomington: Indiana University, Researchi Center for the Language Sciences; The Hague; Mouton. Pages 249-272. Wing, Michael. 1983. "The Credit System of the Gambia". Report to the U.S. Agency for International Development. World Bank. 1985. The Gambia: Financial Sector Review. Washington, D.C.: World Bank. PRE Worldng Paper Series Contact ita AeAhor DatM for paer WPS373 Are Better-off Households More Lawrence Haddad March 1990 J. Sweeney Unequal or Less Unequal? Ravi Kanbur 31021 WPS374 Two Sources of Bias in Standard Samuel Laird February 1990 J. Epps Partial Equilibrium Trade Models Alexander J. Yeats 33710 WPS375 Regional Disparities, Targeting, and Gaurav Datt March 1990 C. Spooner Poverty in India Martin Ravallion 30464 WPS376 The World Economy in the Colin I. Bradford, Jr. March 1990 C. Evangelista Mid-1990s: Alternative Patterns of 32645 Trade and Growth WPS377 After the Cold War: Security for John Stremlau April 1990 C. Evangelista Development , 32645 WPS378 How Does the Debt Crisis Affect Patricio Arrau March 1990 S. King-Watson Investment and Growth? A Neoclassi- 31047 cal Growth Model Applied to Mexico WPS379 Some Implications of Policy Games Miguel A. Kiguel March 1990 R. Luz for High Inflation Economies Nissan Liviatan 39059 WPS380 Techniques for Railway Lee W. Huff March 1990 S. Shive Restructuring Louis S. Thompson 33761 WPS381 Trade in Banking Services: Alan Gelb March 1990 W. Pitayatona- Issues for Multilateral Negotiations Silvia Sagari karn 37666 WPS382 The Indonesian Vegetable Oils Donald F. Larson March 1990 D. Gustafson Sector: Modeling the Impact of 33714 Policy Changes WPS383 On the Relevance of World Yair Mundlak March 1990 D. Gustafson Agricultural Prices Donald F. Larson 33714 WPS384 A Review of the Use of the Rational Christopher L. Gilbert Expectations: Hypothesis in Models of Primary Commodity Prices WPS385 The Principles of Targeting Timothy Besley March 1990 J. Sweeney Ravi Kanbur 31021 WPS386 Argentina's Labor Markets in an Era Luis A. Riveros March 1990 R. Luz of Adjustment Carlos E. Sanchez 39059 WPS387 Productivity and Externalities: Jaime de Melo March 1990 M. Ameal Models of Export-Led Growth Sherman Robinson 37947 PRE Workiog Paper Series Contact IIa Autbhr Aor p WPS388 The Distortionary Effects of Tariff Faezeh Foroutan March 1990 S. Fallon Exemptions in Argentina 38009 WPS389 Monetary Cooperation in the CFA Patrick Honohan March 1990 Wilai Pitayato- Zone nakam 37666 WPS390 Price and Monetary Convergence Patrick Honohan March 1990 Wilai Pitayato- in Currency Unions: The Franc and nakam Rand Zones 37666 WPS391 Wealth Effects of Voluntary Debt Daniel Oks April 1990 S. King-Watson Reduction in Latin America 31047 WPS392 Inntitutional Development in World Samuel Paul Bank Projects: A Cross-Sectional Review WPS393 Debt-for-Nature Swaps Michael Occhiolini March 1990 S. King-Watson 31047 WPS394 Threshold Effects in International Mark M. Spiegel Lending WPS395 How Gambians Save - and What Parker Shipion April 1990 C. Spooner Their Strategies Imply for International 30464 Aid WPS396 Strategic Trade Policy: How New? Max Corden How Sensible? WPS397 Antidumping Regulations or Patrick A. Messerlin Procartel Law? The EC Chemical Cases WPS398 Agricultural Extension for Women Katrine A. Saito Farmers in Africa C. Jean Weidemann WPS399 Macroeconomic Adjustment, Andr6s Solimano Stabilization, and Growth in Reforming Socialist Economies: Analytical and Policy Issues WPS400 Macroeconomic Constraints for Andr6s Solimano Medium Term Growth: A Model for Chile WPS401 Policing Unfair Imports: The U.S. J. Michael Finger March 1990 N. Artis Example Tracy Murray 38010