RELATIONSHIPS AND TRADERS IN MADAGASCAR by Marcel Fafchamps and Bart Minten MSSD DISCUSSION PAPER NO. 24 Markets and Structural Studies Division International Food Policy Research Institute 2033 K St. N.W. Washington, D.C. 20006 U.S.A. July 1998 Contact: Carolyn Roper Tel: 202/862-5600 or Fax: 202/467-4439 MSSD Discussion Papers contain preliminary material and research results, and are circulated prior to a full peer review in order to stimulate discussion and critical comment. It is expected that most Discussion Papers will eventually be published in some other form, and that their content may also be revised. Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Section 1. Agricultural Markets in Madagascar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Section 2. Survey Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Section 3. A Brief Characterization of Agricultural Traders . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Section 4. Trade and Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Tables and Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Tables Table 1. Breakdown of the Sample by Size and Occupational Category Table 2. Size of Operation and Gross Margin Table 3. Capital and Manpower Table 4. Human Capital, Family Background, Wealth and Location Table 5. Factors Important for Success as Perceived by Traders Table 6. Family and Business Table 7. Sources of Information on Market Conditions Table 8. Information Sharing Table 9. Presence of Regular Partners and Ease of Search Table 10. Regular Suppliers and Clients Table 11. Trade Credit Table 12. Loss of Trade Credit in Case of Non-Payment Table 13. Difficulty of Finding Suppliers if Lose One Table 14. Frequency of Contractual Problems Table 15. Verification of Quality of Products Table 16. Credibility of Clients Table 17. Conflict Resolution Method Table 18. Risk Sharing and Access to Financial Help Figures Figure 1: "The poor are poor because they are lazy" Figure 2: "The poor are poor because they have nobody to help them" Figure 3: "I have put money aside for difficult times" Figure 4: "If my business fails, I would have to sell everything to survive" Figure 5: "I will help the others if they are in need" Figure 6: "I can count on family and friends in time of financial problems" Figure 7: "If my business prospers, my family and friends will try to live off me" Figure 8: "I'm only proud of what I accomplish without the help of others" Figure 9: "If I had a lot of money, I would invest it in business" Relationships and Traders in Madagascar Marcel Fafchamps† and Bart Minten†† Abstract1 This paper documents the role that personal relationships play in economic exchange. Ori- ginal survey data show that agricultural traders in Madagascar perceive relationships as the most important factor for success in their business. Evidence details the extent to which relationships are used to serve a variety of purposes such as: the circulation of information about prices and market conditions; the provision of trade credit; the prevention and handling of contractual difficulties; the regularity of trade flows; and the mitigation of risk. Of these, the regularity of supply and demand and the sharing of risk appear particularly important. Larger and more pros- perous traders are those with quantitatively and qualitatively better relationships. Family plays littl e role in business beyond assistance at start-up. ________________ † Department of Economics, Stanford University, Stanford, CA 94305-6072. Tel.: (650) 723-3251. fafchamp@leland.stanford.edu. †† International Food Policy Research Institute, 2033 K Street, NW, Washington DC 20006. Tel.: (202) 862-5650. b.minten@cgnet.com. 1 We benefitted from conversations and comments from Jean Claude Randrianarisoa, Eliane Ralison, Manfred Zeller, Gaurav Datt, and from seminar participants at IFPRI. Special thanks go to Ousmane Badiane for obtaining financial support for this research. Weacknowledge financial support from the United States Agency for International Development. 2 Sociologists have long emphasized the crucial role that interpersonal relationships play in the life and professional success of individuals and groups (e.g., Coleman (1988), Granovetter (1985), Putnam, Leonardi and Nanetti (1993)). In recent years, economists too have begun to recognize that economic exchange is influenced by the level of familiarity and trust that exists between agents (e.g., Gambetta (1988), Fukuyama (1995), Greif (1993, 1994), Platteau (1994), Fafchamps (1998), Tadelis (1998)). In a world characterized by imperfect information and enforcement, it has been shown both theoretically and empirically that personalized relationships can facilitate the circulation of information on new technologies (e.g, Barr (1997)) and market opportunities (e.g, Kranton (1996)), the screening of job and credit applicants (e.g, Montgomery (1991), Cornell and Welch (1996)), the sharing of risk (e.g., Fafchamps (1992), Coate and Raval- lion (1993), Lund and Fafchamps (1997)), and the punishment of cheaters (e.g., Kandori (1992), Fafchamps (1998)). Much of this work remains confined to markets such as credit or labor in which moral hazard issues are severe. Applications to markets for commodities have so far been few (see, however, Gabre-Madhin (1997), Kranton (1996), Bernstein (1996)). The present paper fills this lacuna by documenting the role that personal relationships play in the trade of agricul- tural products. This paper presents original evidence on the extent to which relationships are used by agri- cultural traders in Madagascar to serve a variety of purposes such as: the circulation of informa- tion about prices and market conditions; the provision of trade credit; the prevention and han- dling of contractual difficulties; the regularity of trade flows; and the mitigation of risk. Results show that larger and more prosperous traders are those with better relationships. The fact that larger, more successful traders are better connected will hardly surprise anyone who is familiar with African trade patterns (e.g., Bauer (1954), Jones (1959, 1972), Meillassoux (1971), Cohen (1969), Amselle (1977)). It is also in line with the new literature on social capital that identifies networks of relationships as a productive asset from which individuals can derive a return. But it 3 runs somewhat contrary to the expectation of policy makers and international agencies who often implicitly assume that larger traders are more sophisticated and that sophistication is synonymous to arms-length, anonymous exchange. This is important because the common observations that large traders cultivate close rela- tionships with each other is often interpreted as evidence of collusion and price rigging. Although we cannot comment directly on whether or not collusion is present in Madagascar grain markets, our results indicate that there are many other reasons why traders maintain a net- work of personal relations, such as access to information, regular trade flows, trade credit, and risk sharing. Our results also indicate that traders with better networks have higher margins and thus that they derive a return from their social capital. Although these results must be confirmed by a more rigorous analysis, they suggest that large traders make more profit than their smaller competitors not necessarily because they abuse their market power but because their connections make them more efficient. If correct, our analysis implies that governments and politicians should refrain from the temptation of blaming successful traders for the economic difficulties of a country. This has too often been done in the past2 and is still very much practiced, as the plight of ethnic Chinese traders in Indonesia reminded us only recently. Efforts to increase competition should focus on reducing the difference between the margins of large and small traders. This should not be accomplished by destroying the social capital of large traders since this eliminates a productive asset and thereby reduces social welfare. Rather, increased competition should be sought by eliminating market imperfections, increasing the fluidity of trade, and thus reducing the need for network capital. If this proves to be too difficult, more competition must be sought by raising the social capital of small traders -- for instance raised by improving trust and deterring fraud in ________________ 2 "Indo-Pakistani traders [] bore the brunt of Malagasy violence in the 1987 riots. [T]he Indian premises on either sides [of the main street in Tulear] along with most of the central area were gutted" (see Lonely Planet (1994), p.220). 4 commercial relations. Keeping agricultural trade in a state of lawless marginality under the con- stant suspicion of political authorities only favors mistrust and encourages crooks. It does not foster a competitive environment. Competition can also be encouraged by lowering the market imperfections that generate large returns to personal relationships. For example, accurate infor- mation about prices and market conditions can be circulated on the radio to complement the information that circulates informally among traders. The risk faced by traders -- and thus the need to seek insurance through friends and relatives -- can be reduced by increasing road security and deterring theft. These issues deserve more research. The paper is organized as follows. We begin in Section 1 with a brief description of the agricultural markets policies of the Malagasy government since independence. We also provide a rapid survey of the existing literature on agricultural markets in Madagascar, most notably the work of Barrett (1997a, 1997b). We then describe in Section 2 the data used in this paper and the survey methodology used to collect them. A characterization of agricultural traders is provided in Section 3. Section 4 is devoted to a detailed analysis of the different functions performed by rela- tionships. Conclusions and prospects for future work are presented at the end. Section 1. Agricultural Markets in Madagascar After Madagascar obtained independence from France, governments initially increased the intervention of the state in agricultural markets (e.g., Dorosh and Bernier (1994), Shuttleworth (1989), Berg (1989)). By the end of the 1970’s, most trade in agricultural products was in the hands of the state. A reversal of policy took place in the 1980’s with a transition from a state food marketing system to a liberalized market. This transition, however, was very gradual. It begun in 1983 when the state officially abandoned its monopoly on the commerce of agri- cultural products. The initial liberalization measures implied that agricultural trading was open to everybody except in the plains of Marovoay and Lac Aloatra - two main production areas - where two government agencies, FIFABE and SOMALAC, could continue their monopoly. The 5 roles of these two state companies were only redesigned in 1989. In the beginning of the reforms, floor and ceiling prices were maintained in effect. In June 1985 a government decree fixed the floor price of paddy while removing the ceiling price completely. But in reality the government effectively controlled domestic rice trade until 1986. From mid 1983 on they supplied all the big cities with the "riz fokontany", i.e. subsidized rice. In Antananarivo this type of rice represented until 1986 more than 60% of the average household consumption in rice (e.g., Roubaud (1997)). This program continued until October 1988 but its importance declined gradually. In November 1986, the government introduced a buffer stock scheme in response to high seasonal prices during that year and to defend the ceiling price. However, the buffer stock scheme was poorly administered and was ultimately terminated in 1990. In 1991, the government introduced an import tax of 30% on rice to protect local production. This tax was reduced in 1995 to 10%. However, occasionally the government granted tax exoneration for certain companies and shipments to assure a steady food supply. The current situation can be described as one in which private traders have been given free reign to set buying and selling prices and to move agricultural products around the country. The state continues to intervene in agricultural markets through buying and selling operations con- ducted for example by SOMACODIS but these agencies now only represent a very small percen- tage of the total volume of food products transacted domestically. In this respect Madagascar resembles many other African countries that have gone through a similar cycle of government interventionism and retreat (e.g., Berg (1989), Staatz, Dione and Dembele (1989), Gabre-Madhin (1997)). Trade in agricultural products in Madagascar has been analyzed by other authors, most not- ably Barrett (1997a, 1997b) and Berg (1989). Agricultural food products flow mostly from rural areas to urban centers immediately after harvest, and from urban centers to rural areas in the lean period. Although the capital city Antananarivo occasionally draws food products from outside its 6 own province (faritany), most marketed output is consumed within the province where it is pro- duced (e.g., Minten et al. (1997)). Barrett and Dorosh (1996) show that most Malagasy rural households are deficit rice producers and must rely on the market for their subsistence. Food markets are thus important not only for urban dwellers but for rural inhabitants as well. Using surveys of Malagasy grain traders, Barrett (1997a) describes agricultural trade in the country as characterized by extreme disparity between large and small traders. He argues that most traders do not have access to the equipment and credit required to penetrate the more profitable segments of the business. As a result, most trading businesses remain small while a few large traders derive large margins in activities that are secluded from competition. Only in segments where entry is easy is competition fierce. A similar conclusion is reached by Abt Associates (1991) and Kristjanson and Martin (1991). Barrett concludes his work by calling for easier and wider access to credit for traders. Section 2. Survey Methodology Although the work of Barrett and others provides much needed detailed information on agricultural markets in Madagascar, it largely ignores issues of social capital and personal rela- tionships. To fill this lacuna, a survey of agricultural traders was conducted in Madagascar in a joint project between IFPRI (the International Food Policy Research Institute) and the local Min- istry of Scientific Research (FOFIFA). The survey consisted of two rounds. The first round was held between May 1997 and August 1997. The questionnaire in the first round survey consisted mainly of questions dealing with the individual characteristics of the traders and with the struc- ture, conduct, and performance of the trading sector. The second survey round was conducted between September 1997 and November 1997. The same traders were visited and they were asked mostly about the nature of their relationships with other traders, clients, and suppliers. The sample design was constructed so as to be as representative as possible of all the traders involved in the whole food marketing chain from producer to consumer, wherever 7 located. Three main agricultural regions were covered (Fianarantsoa, Majunga, and Antananarivo) and the sampling frame within these regions was set up as follows. Traders were surveyed in three different types of location: (1) Traders operating in big and small urban markets in the main town of every province (fari- tany) and district (fivondronana). These traders are mostly wholesalers, semi-wholesalers, and retailers. (2) Urban traders located outside the regular markets. These often are bigger traders, processors (e.g., rice millers), and wholesalers. (3) Traders operating on rural markets at the level of the rural county (firaisana). These are mostly big and small assemblers and itinerant traders. Rural firaisanas were selected through stratified sampling based on agro-ecological characteristics so as to be representative of the various kind of marketed products and marketing seasons. The survey focused on traders that marketed locally consumed staples such as rice, cas- sava, potatoes, beans, and peanuts. The different forms in which these products are marketed were taken into consideration, i.e., paddy and milled rice, maize and maize flour, etc. Traders involved primarily in export crops, fruits, vegetables, and minor crops were excluded. Most sur- veyed traders -- 67% -- report rice as the agricultural product they trade most intensively. This reflects the importance of rice as the main staple food in the country. Other most actively traded products are beans and lentils (18% of the sample report them as their main traded product), cas- sava (5%), potatoes (5%), peanuts (4%), and maize (2%). A total number of 850 traders were surveyed in the first round, 739 of whom were surveyed again in the second round. To facilitate comparison, the analysis presented here is based on traders that could be located in the two rounds.3 The three provinces of Antananarivo, ________________ 3 Not surprisingly, the category of traders which were hardest to trace during the second survey round are those who are least formal and have the least permanent form of operation. As a result, small itinerant traders tend to be underrepresented in the results reported here. 8 Fianarantsoa, and Majunga are represented more or less equally in the sample. A breakdown of the sample by size and occupational category is given in Table 1. Size categories are defined based on the total value of reported annual sales; occupational categories are based on the occu- pation of the respondent for the main traded crop at harvest time.4 The Table shows that retailers constitute the bulk of the sample. They are divided into retailers with a semi-permanent selling point -- usually a stall in the market itself; and retailers without fixed selling point, that is, those who sell immediately from the roadside. As the Table shows, the latter are typically smaller and less formal.5 In contrast, the largest traders are assem- blers (traders who collect large quantities from the countryside and assemble them for shipment) and wholesalers (traders who operate in bulk). Having described the survey methodology, we are now ready to proceed with the analysis. We begin with a brief characterization of surveyed traders. Section 3. A Brief Characterization of Agricultural Traders Surveyed traders vary widely in the size of their operations. The total sales of the average trader amount to almost $3,300 a month6 but there is enormous variation across traders: the Gini coefficient of total sales computed over the sample is 0.761 (Table 2). Similar results are obtained if we consider purchases or if we restrict ourselves to sales during the month preceding the second round interview. Size is correlated with occupational category: assemblers and wholesalers have the largest monthly turnover at $8,700 and $5,550, respectively; retailers have the smallest -- $1,300 and $400 for those with and without fixed selling point, respectively. These results are similar to those reported by Barrett (1997a) in his study of agricultural markets in ________________ 4 The definition of occupational categories is indeed complicated by the fact that traders who are semi-wholesalers for one product can be retailer for another. Traders may also change occupational category during the year, e.g., they may be assembler during the harvest season but semi-wholesalers the rest of the year. 5 Because their fluid nature makes them harder to trade, they are underrepresented in the second survey round. 6 Sales of listed staple food products over the period April 1996 to March 1997; conversion into US$ using an approximate exchange rate of 5,000 Francs Malgaches for US$1. 9 Madagascar. To reflect variation in size, in much of the presentation that follows we divide the sample into three terciles called ’small’, ’medium’, and ’large’, respectively. The data also indi- cates that trade in agricultural food products is a highly seasonal activity: monthly sales in the lean October period only amount to 40% of average annual sales. This is due to the highly sea- sonal nature of agricultural production and the relative lack of in-village storage (e.g., IFPRI (1998)). Most assembly takes place after harvest in April and May, which explains why the difference between annual averages and October sales is widest for assemblers. In contrast, retail where small traders dominate is less seasonal in nature since consumption takes place throughout the year. Next, we investigate whether the profitability of traders varies systematically with size. Gross margins are computed as the difference between total annual sales and purchases. Results provide an order of magnitude of the total payments to labor, management, and equipment, but they are subject to a lot of measurement error.7 On average, gross margins amount to $460 a month -- significantly higher than the average GDP per head of $230 per year (e.g., The World Bank (1997)).8 Assemblers have the highest gross margin -- $1800 a month -- retailers without table the lowest -- $70. In percentage terms, the average gross margin is 14% -- 19% among assemblers, 14% among wholesalers, 10% to 16% among retailers. There does not seem to be a systematic relationship between firm size and gross margin rate. ________________ 7 First, certain traders are hesitant to communicate their effective profit margin to outsiders and seek to disguise the volume of their activities. This generates inconsistencies in reported sales and purchases, e.g., medium size traders report higher average annual purchases than sales in Table 2. Second, very few traders keep an accurate accounting of their sales and purchases over the year (if only for fear of taxation). As a result, they easily forget how much and at which price they actually bought and sold products. Finally, we suspect that certain traders actually do not know how much they sold and bought as trading happens on anad hocbasis. This is particularly true for small traders who seldom make a clear distinction between their production, consumption, and trade in agricultural products. 8 To minimize errors due to inconsistencies between sales and purchases, gross margins were computed as follows. For each product, a gross margin was computed by multiplying quantities sold by the difference between sales and purchase price; adding the result over all products yields a gross margin estimate based on quantities sold. An alternative gross margin estimate was constructed using quantities purchased instead. The gross margin figures reported in Table 2 are the average of the two. 10 Capital and Labor We now examine in detail the capital and labor structure of surveyed firms before turning to the human capital and family background of respondent traders. Working capital comes mostly from own sources, not from credit: 89% of the traders rely exclusively on their own funds for their business activities. The average working capital is US$2,060 -- roughly two thirds of aver- age monthly sales (Table 3). Not only do larger traders have more working capital than small ones, they also appear to rotate it faster: the ratio between working capital and monthly sales indeed falls with firm size, i.e., from 1.7 among small traders to 0.9 among medium size traders and to 0.6 among large traders. Although most of the traders rely on own funds to finance their operations, 81% estimates that these funds are not sufficient and they would like to see their funds increase threefold. Formal credit as a mean to finance trading activities is almost non-existent: it is mentioned by only 1.5% of the traders representing 6.1% of total sales. The minor importance of formal financial institutions is further illustrated by the fact that only 16% of the surveyed traders have a bank account; one trader out of 100 has a bank line of credit. Only 4% of the traders has ever asked for credit from a formal institution. When asked why they do not apply for formal credit, half of the traders respond either that they do not know how to apply (28%) or that the applica- tion procedure is too complicated (19%). The rest either consider the interest rate too high (23%) or or do not possess any collateral (16%). As is often the case in surveys of this type (e.g., Cuevas et al. (1993), Fafchamps et al. (1994), Fafchamps, Pender and Robinson (1995)), we observe a positive relationship between firm size and reliance on formal financial institutions. Informal credit does not appear to compensate for the limited use of formal credit. Only one trader out of ten derives part of its working capital from informal credit sources. Less than 2% of the traders are members of savings mutuals; only 1% are member of a "tontine" (a rotating sav- ings group). The use of trade credit is also very limited, as we shall discuss more in detail in Sec- 11 tion 4. The median self-declared opportunity cost of capital reverts around 20% a year; some respondents declare facing a much higher shadow cost of capital, however. There is no clear rela- tionship between size and the shadow interest rate perceived by traders. The level of the equipment at the disposal of traders is very limited, even among large traders. The only piece of equipment that is nearly universal is the balance, which is owned by 79% of the surveyed traders. Half of the traders own a location that they use for storage -- typi- cally the shop itself or a small warehouse. Less than one trader out of ten -- one out of three among large traders -- owns a vehicle for transportation purposes. The total equipment owned by traders is worth less than one fifth of their working capital; most trader capital is thus tied up in stocks and receivables. Malagasy traders have imperfect access to modern means of communication. The great majority of traders (95%) do not have a telephone for their business; virtually none has a fax machine. Even among bigger traders, only 11.5% declare having a phone. Half of the surveyed traders nevertheless have access to a phone, but few avail themselves of this opportunity in the conduct of their business. The use of fax machines for trade purposes is virtually non-existent. In terms of management experience, surveyed traders have on average spent 6 years trading in agricultural products. The average starting date is 1991, six years before the survey, but significantly later than the onset of agricultural trade liberalization (1983-1987) (see Section 1). The link with the previous state marketing system is minimal: only 2% of respondents ever were employed in the state marketing system. Large traders are slightly more experienced than small traders, but the difference is not very large. The majority of surveyed traders operate all year round and focus most of their attention on trade, with no noticeable difference by firm size. Some 14% of the small traders list agriculture as their main activity while some of the bigger traders declare transformation, transport, or agricul- ture as their main source of income. As a secondary activity, farming remains important: 69% of 12 surveyed traders participate, in one way or another, in agriculture. In addition, 16% of the respondents participate in non-farm activities, 17% obtain a regular salary, and 11% have a regu- lar source of income other than earned income. Only 40% of the respondents (half of the small traders) derive all their income from agricultural trading. Malagasy traders employ very few people other than themselves (Table 3) -- on average, one unpaid family helper, one permanent employee, and a little over one casual worker. Small and medium size traders have almost no external help in their business and they seem to do most of the trading on their own or with the help of family members. Large traders make more use of permanent and temporary employees and may also use the services of collecting agents. Judging from the number of months in a year than different categories of trade workers spend participat- ing to the activities of the firm, temporary employees work about half the year while all other categories work close to full time. While one observes a positive relationship between employ- ment and total sales, the relationship is far from being proportional. In other words, large traders have a much higher volume of activity per worker than small traders. Since they also use less working capital per volume of sales than small traders, they appear to be more efficient overall. Section 4 investigates whether relationships may account for part of the performance differential between small and large traders. Human capital and family background Turning to the personal characteristics and human capital of the owners (Table 4), we see that close to half the surveyed traders are female; the proportion of women is much higher among small than large firms, however. Large firms also tend to have a slightly older owner, but the difference is not large. Surveyed traders are, on average, well educated, having spent on average 9 years in school. At first glance, there appears to be little difference across firm size terciles but this is partly incorrect. Among the small traders, 11 % are not able to read or write. Morover, while 46% of the big traders finished at least secondary school, this percentage is only 15% for 13 the small traders. Small traders are also more likely to identify with traditional religions and identify themselves as non-Christian, the dominant religion on the island. Other religious affiliations are extremely rare. The overwhelming majority of the surveyed traders were born in the country and are ethnically Malagasy. Unlike in other parts of Africa (e.g., Fafchamps (1998)), the ethnic and religious make-up of the trading community is thus fairly homogeneous.9 Unlike much of the African mainland, Malagasy people share a common language which is spoken throughout the island, hence facilitating communication and trade. French is widely used in the administration and in high school instruction. When interviewed about the languages they speak regularly, almost 50% of surveyed traders declare speaking a language other than Malagasy on a regular basis -- usually French. Larger traders are slightly more likely to speak a second language, but the difference is small. Although language is not as much a barrier to exchange than it might be in other countries, regional differences and sensibilities exist and geo- graphical mobility is limited.10 Most respondents trade within the area where they were born or where they grew up. On average, only one trader out of twenty comes from outside the province where he or she operates. The family background of surveyed traders suggests that agricultural trade is an activity undertaken mostly by mature, settled adults with a family and kids to provide for. Most surveyed traders are married but a large part of the smaller traders are either bachelors, widow(er)s, or divorced. The small trader category thus seems more heterogeneous, i.e., made of individuals who are in the beginning of their career and of people that might have entered the sector because of personal problems such as divorce or death in the family. Traders have three children on aver- ________________ 9 The reader should bear in mind that Indo-Pakistani traders, who constitute a small minority of traders, tended to refuse participation to the survey. 10 Although most Malagasy are from mixed Asian and African ancestry, people from coastal areas harbor a historical resentment against people from the central highlands who traditionally ruled the country. Divisions also exist along the lines of former kingdoms that historically divided the island. Remnants of pre-colonial caste distinctions between members of the former royal family, the middle class, and former slaves are said to survive in certain rural areas, but the survey made no attempt to revive these feudal classifications by asking questions about it. 14 age, half of whom are old enough to help with the business. Respondents also have brothers and sisters who, if necessary, could assist in the firm. Parents education does not vary much across firm sizes and thus appears an unlikely determining factor in trade success. In terms of personal wealth, 56% of the surveyed traders own a house but only 5% have a television and 23% a bicycle. 44% and 90 % of the traders possess a radio and a cassette recorder, respectively. In all cases, the percentage is higher for bigger traders who also tend to live in a more expensive home. The value of the house that traders live in is on average commen- surate with the value of their working capital, hence suggesting that their business risk exposure is far from negligible. Finally, there does not appear to be a strong relationship between geo- graphical location and firm size (Table 4). To summarize, Malagasy traders in agricultural food products are characterized by their extreme diversity in terms of volume of operation and their unsophisticated mode of operation (little equipment, few employees). In contrast with the common view that trade in Africa is mostly a secondary activity, most surveyed traders are heavily involved in trade: they have invested in it a large proportion of their total wealth and they derive a significant proportion of their income from it. Given the low level of technical sophistication of the industry as a whole and the relative unimportance of credit, even for large traders, returns to scale are unlikely to be present. As a result, one would expect entry to be easy and competition to be fierce. Some of the evidence presented above seems to support this view, in particular the plethora of small traders, especially in retail, and the fact that small traders are younger and less esta- blished but otherwise share a family background similar to that of successful traders, suggesting a life-cycle explanation for size differences. Some of the facts, however, do not fit a simple free entry, life-cycle explanation for the size distribution of trading firms. Traders in the upper tercile of the firm size distribution use 15 times more working capital and 2.2 times more labor but they sell 44 times more and get 46 more gross margin than traders in the lower tercile. Without doing 15 any complex calculation, it is clear that large traders have a much higher total factor productivity than small traders. What factors could explain this difference? Human capital has been put forth in the recent literature as a major determinant of economic performance (e.g., Mankiw, Romer and Weil (1992)). It is unlikely, however, that the very small difference in schooling observed between small and large traders could account for the difference in total factor productivity. Another possible candidate, one that is receiving increasing attention, is the social network capi- tal of traders, that is, the relationships that they have with others. To explore this possibility, we now investigate the many roles that relationships play in the business of Malagasy traders. Section 4. Trade and Relationships We begin with Table 5 which illustrates the importance of relationships as perceived by traders themselves. The Table shows that relationships are by far the most important factor for the success of a trader. 71% of the respondents regard reputation and relationships as very impor- tant for the success of their business. This proportion is much higher than that for credit, price, or equipment. Access to credit, which is typically presented as a major constraint by small businesses the world over, ranks much lower than relationships: only 11% of the respondents see it as a very important factor in business success; close to 40% of the respondents think it is not important at all. It is sometimes argued that relationships are important among the poor because they need the support of their family and community to deal with the vagaries of life while the rich can afford to behave in a more individualistic fashion (e.g., Platteau (1996)). This is not the case here. Table 5 indeed also shows that the importance given to relationships rises with firm size: while 62% of small firms think relationships are very important, 77% of large traders do. It is therefore not the case that the emphasis on relationships results from the presence in the sample of small, poor traders who life in symbiosis with their community. If anything, larger, richer traders put more emphasis on relationships than the poor, not less. 16 These results beg the question of why relationships are important. To try to answer this question, we examine six possible roles that relationships may play in trade: (1) business training and start-up support; (2) information sharing; (3) regularity of demand and supply; (4) credit; (5) prevention of contractual breaches; and (6) risk sharing. Business training and start-up support Table 6 shows that a quarter of surveyed traders had either a father or a mother in trade. Only 14% of respondents say they are in this business because of family traditions, however. Half the traders were helped by family and friends at start-up and close to half learned the busi- ness with a relative or a friend. The rest learned business on their own. Larger traders seem to have had parents with more experience in trade but otherwise are similar to their smaller counter- parts: if anything, they are more likely to indicate they learned the business on their own -- a finding hardly consistent with the idea that parents in trade is a condition for success. In addition, the bottom of the Table shows that traders have typically outgrown their family base: while on average they have about one relative in trade, they know close to 10 traders personally. Taken together, this evidence suggests that while, for some traders, family relationships were important at start-up for capital and experience, they do not seem to be strong determinants of business suc- cess. If anything, traders who learned the business from their family appear less likely to be suc- cessful. In contrast, non-family types of relationships which are initially unimportant seem to grow over time. Information sharing In contrast with training and start-up support, non-family relationships appear critical for getting access to business-relevant information. Table 7 lists the sources of information on prices, supply, and demand conditions used by surveyed traders. The numbers bring to light the paramount importance of relationships as sources of information: other traders and suppliers and 17 clients are by far the most important source of business information. Public sources such as news- papers, radio, and public services play an extremely marginal role. Larger traders rely more than small traders on messengers, that is, individuals who are sent explicitly to collect information, but even among large traders their role is dwarfed by relationships. Another interesting regularity present in the data is that small traders are more likely than large traders to seek information from other traders instead of getting information from suppliers and clients. One likely explanation is that large traders have a closer relationship with their suppliers and clients and feel they can rely on the information they provide. Small traders, in contrast, probably fear they will be cheated if they trust what suppliers and clients tell them. The frequency with which respondents share information with other traders appears fairly low, however, as shown in Table 8. While more respondents discuss quality, bad clients, and prices with others at least once a year, the great majority of them do not discuss these issues every week.11 Regularity of supply and demand Another possible role that relationships play is in ensuring secure supply and demand. Sur- vey results indicate that finding a supplier or a client is a recurrent problem for respondents: between 40 and 50% of them occasionally face difficulties identifying potential buyers and sell- ers. As Table 9 illustrates, traders who experience lots of difficulties are those who have the smal- lest numbers of regular suppliers and clients. In other words, traders with regular sources of sup- ply and demand are less likely to encounter problems. Relationships thus reduce search costs. Table 10 indicates the existence of a strong relationship between firm size and the emphasis on regular suppliers and clients: while large firms do between 40 and 45% of their business with ________________ 11 Some caution should be used when interpreting the results, however. First, respondents seem to have understood the questions relative to ’other traders’ as meaning ’other traders who operate in a manner similar to yours’, hence excluding suppliers and clients even if they are traders. Second, questions relative to bad clients and prices were only asked to respondents who have regular clients, thereby introducing a potential bias. 18 regulars, this proportion is much smaller among small traders. As a result, larger traders econom- ize on search costs relative to smaller traders and probably have more secure sources of demand and supply. Results indicates that the ties respondents have with their regular suppliers and clients is not based on family or ethnicity: the overwhelming majority of them (90%) describe their ties as business only. This is not entirely surprising given that Madagascar, unlike other developing countries (e.g., Fafchamps (1998)), shares a single language and sense of ethnic homogeneity.12 The emphasis that larger traders place on regular clients and suppliers is con- sistent with their use of suppliers and clients as sources of information about prices and market conditions: the existence of long term relationships between them ensure that the information provided is more accurate that what would be conveyed to an unknown trader. Trade Credit Another reason why traders might value relationships is because they open access to trade credit in the form of payment facilities with suppliers or advances paid by customers. Table 11 reports the proportion of purchases and sales made cash, on credit, and with advance payment. The overwhelming majority of transactions are cash only. On average, respondents operate one sixth of their business with some element of credit. When credit is present, it floats predom- inantly downstream, that is, from seller to buyer. The ratio of payables and receivables over monthly sales shows that respondents are, on average, net givers of credit, not so much because they sell more on credit than they buy -- on the contrary -- but probably because buyers take more time to pay. Relationships play an important role in access to trade credit. Results shows that respon- dents virtually never grant or receive trade credit on the first transaction. The role of relation- ________________ 12 Although Malagasy people do not distinguish themselves according to language or ethnicity, they do pay attention to geographical origin, however. Unfortunately, no questions were asked about the geographical origin of regular clients and suppliers. 19 ships in trade credit is further confirmed by the likely consequences of default. Table 12 shows that, if a respondent were to not pay a supplier, the credit of the respondent with other suppliers would not be affected very much: half of them estimated that not paying would only reduce their chances of getting trade credit with none or at most some of their suppliers. Similar responses were obtained when the question was asked about the respondent’s clients. Taken together, these figures suggest that the reputational sanctions for breach of contract are mild (e.g., Kandori (1992), Fafchamps (1998a), Greif (1993, 1994)). The loss of the relationship, however, is valu- able: as shown in Table 13, the large majority of respondents feel that it would be difficult for them to find a new supplier if they were to lose one -- as would most probably be the case if they failed to pay. These findings are similar to those described by Fafchamps (1996) in the case of Ghana, and they are consistent with theoretical models of trade that emphasize the self- disciplining role of relationships (e.g., Ghosh and Ray (1996), Fafchamps (1998a, 1998b)). Breaches of Contracts and Conflict Resolution Further evidence in favor of the relationship-based models of trade can be found in the manner respondents prevent breaches of contract and handle contractual conflicts. Table 14 presents estimates of the absolute and relative frequencies of contractual disputes among sur- veyed traders. The evidence shows that the incidence of problems is very high. On average, respondents face a quality or late delivery problem in 1 out of every 13 purchases and a late or non-payment problem in 1 out of every 45 sales. These averages, however, hide the fact that the frequency of problems is much higher among firms that contract forward. Among firms that place orders, for instance, a problem with supplies occurs on average in one third of the pur- chases; the proportion is even higher for large traders (Table 14). Among firms that sell on credit, a case of late or non-payment arises in one out of every 20 sales. Since these firms do not sell all their output on credit, this translates into one case of late or non payment in 20% of the credit sales. Fortunately, only one out of every 35 late payment cases turns into non-payment. 20 Evidence collected in Ghana by Fafchamps (1996) suggests that what probably keeps this pro- portion low is the time traders spend chasing late payers. To summarize, the incidence of con- tractual problems is high whenever traders contract forward, which explains why few of them do. Traders’ desire to avoid the contractual problems created by forward contracting singularly complicates exchange and is achieved at considerable cost. First of all, as the Table itself shows, most transactions take place without orders and without credit. This means that virtually all trade in agricultural products in the entire island of Madagascar takes the form of cash-and-carry tran- sactions. This can hardly be regarded as an efficient and convenient way of conducting trade. Very little if any forward looking transactions occur, and if they do, they are based on a strong relationship of trust between buyer and seller. Since traders hardly ever pay by check,13 this implies that search costs are higher than they should be, and that massive amounts of currency constantly circulate in the countryside -- an invitation to theft and a perfect target for an inflation tax. Not surprisingly, many surveyed traders identify security as their number one problem (e.g., IFPRI (1998)). The prevention of problems also has its costs. Table 15 indicates that the overwhelming majority of traders and their clients inspect quality before purchasing. In other words, quality is inspected visuallyeachtime a product changes hands.14 Given the multi-layered nature of agri- cultural trade and thus the large number of transactions involved in getting foodstuffs from pro- ducers to consumers, we see that inspecting quality alone must account for a significant propor- tion of the spread between producer and consumer food prices. The Table also demonstrates that quality inspection is a task that traders hardly ever delegate: although they employ on average 3.3 people to assist with the business, traders nearly ________________ 13 The fact that Malagasy banks -- according to what we have heard -- take two to four weeks to clear checks drawn on another town hardly incite traders to pay by check: doing so would tie up their working capital for weeks on end. 14 Similar practices are described in Ethiopia by Gabre-Madhin (1997). 21 always inspect quality themselves, presumably because conducting the task accurately is critical for business. In other words, so few cases of bad deliveries are reported not because suppliers are truthful but because buyers go to great lengths to ensure they are not cheated. Given the amount of energy they spend on checking quality, it is surprising that bad deliveries occur at all. Traders’ inability or unwillingness to delegate quality inspection also means that their volume of activity is limited by the quantities that the owner can inspect in person. It also implies numerous trips to supply areas, some of which are for nothing since traders do not use telephones, cannot or will not place or take orders, and and must search for buyers or sellers once they are on location. Such a system can be but expensive to run and in such an environment having close relationships with regular clients and suppliers must singularly simplify one’s business -- hence the emphasis put on relationships as a factor of commercial success. Similar difficulties arise in the granting of trade credit. Table 16 shows that the great major- ity of respondents check the credibility of clients before granting payment facilities. Apart from information collected from the client directly or from a personal visit to the client’s shop, respon- dents rely primarily on information received from traders and other sources such as friends and family. There too relationships serve a role as facilitator of the screening of trade credit reci- pients. The relatively small proportion of respondents who cite information collected from traders and other sources and the fact that this proportion diminishes with firm size nevertheless suggest that reputation mechanisms in agricultural trade in Madagascar can be described as embryonic at best. This stands in stark contrast with the intense sharing of information -- and the much higher incidence of trade credit -- that were found by one of the authors in Kenya and Zim- babwe (e.g., Fafchamps et al. (1994), Fafchamps, Pender and Robinson (1995), Fafchamps (1997), Fafchamps (1998b)). There, firms were found to actively share information about bad payers, either informally (Kenya) or via a credit reference bureau (Zimbabwe). The vetting of clients was also widely practiced. Agricultural trade in Madagascar more closely resembles the 22 manufacturing sector in Ghana where little information sharing was uncovered (e.g., Fafchamps (1996)). The reader may wonder why breach of contract is not efficiently deterred by the presence of legal institutions such as lawyers, courts, and the police. Table 17 indicates the conflict resolution methods most likely to be used after a breach of contract. By far the dominant response is to negotiate with the other party and, in some cases, to call upon a third party to serve as mediator. The use of lawyers is extremely rare (only one case was recorded). Respondents are even extremely reluctant to use thethreat of calling the policy or going to court, let alone actually doing it. Again, these results are similar to those observed elsewhere (e.g., Fafchamps (1996), Bigsten et al. (1998)) though they demonstrate an even lower reliance on legal institutions than elsewhere in Sub-Saharan Africa. Not using legal institutions does not mean that conflicts are not resolved, however. In fact, four fifths of disputes with suppliers and clients are resolved and trade resumed. This brings out yet another function that relationships play: to facilitate the resolution of conflicts through negotiation. It is because parties wish to preserve their relationship that they agree to negotiate and to seek a mutually agreeable solution to their dispute, a solution that preserves the relationship itself. In other words, it is because traders value relationships that con- tractual disputes are resolved. Risk Sharing Relationships can also serve the role of insurance mechanism. Business in general and trade in particular are subject to all kinds of risks -- theft, non or late payment, adverse price fluctuation, storage loss, etc -- each of which can easily cripple a small trading business. In a world where trade credit is inexistent or rare, a trader without working capital cannot operate. Consequently, a trader whose working capital is either lost or tied up in bad debt and unsold stocks loses his or her income. The capacity to borrow from others therefore serves a crucial insurance purpose. Table 18 confirms that the overwhelming majority of respondents are 23 involved in helping and being helped by others. Assisting and being assisted can be interpreted as the two sides of the same coin: people help each other because they expect to be helped in return (e.g., Fafchamps (1992), Coate and Ravallion (1993), Lund and Fafchamps (1997)). Interestingly, the Table shows that larger traders are as involved in solidarity networks as their smaller competitors, and that they have in general more friends they can count on in times of trouble. This flies into the face of those who have claimed that solidarity mechanisms tax the rich and that, as a result, the rich are more individualistic (e.g., Platteau (1996)). To investigate these ideas further, respondents were asked to rank a variety of statements about poverty, prosperity, and mutual assistance on a scale going from very true to very false. Results are summarized in Figures 1 to 9. The first two Figures presents respondents’ attitudes toward poverty. Not surprisingly, the poor are less likely than the rich to blame laziness for poverty, but contrary to expectations they do not see poverty as the outcome of lack of assistance either. Small traders are more likely to declare that they have put money aside for difficult times and less likely to sell everything in bad times (Figures 3 and 4). In contrast, small traders are less likely to help others in need (Figure 5) and, in counterpart, less likely than large and medium size traders to receive assistance in times of trouble (Figure 6). If anyone is afraid that prosperity will be taxed away by family and friends, it is the poor: Figure 7 shows that small traders are sys- tematically more likely to believe their family will invite themselves to their home if they succeed in trade. Consequently, small traders are more likely to derive individualistic pride in their business accomplishment (Figure 8). Individualism thus appears more present among small traders than among large ones. As for investment disincentives (Figure 9), they do not appear to be present either: small traders are systematically less likely to invest their profits in business than medium and large size traders -- possibly because they have less access to social insurance through solidarity networks. To summarize, large traders appear less individualistic and more prepared than small 24 traders to help others and get helped in return. The capacity to successfully join networks of soli- darity may well be critical to their long term prosperity as it shelters them from some of the risks of business and enable them to invest more, grow more rapidly. In addition, solidarity relation- ships probably enable traders to borrow money not so much to deal with negative shocks but to take advantage of especially lucrative arbitrage opportunities. Conclusion We have investigated the role that relationships play in the conduct of agricultural trading businesses. We found that relationships play a wide variety of roles such as: (1) business training and start-up support; (2) information sharing; (3) regularity of demand and supply; (4) credit; (5) prevention of contractual breaches; and (6) risk sharing. Of these, the regularity of supply and demand and risk sharing appear particularly important in the sense that large traders enjoy a significantly larger proportion of sales and purchases from regular partners and systematically emphasize values and action consistent with risk sharing. Together with the circulation of infor- mation, the capacity and willingness to get and give trade credit, place and take orders, and sim- plify the inspection of quality are additional benefits traders derive from good relationships. The value of relationships, not legal institutions, appears to be what motivates traders to honor con- tracts and seek the resolution of conflicts through negotiation. These issues are analyzed in further detail in Fafchamps and Minten (1998). The importance of relationships is partly due to the extreme lack of sophistication in busi- ness practices: no payment by check; no invoicing; very little trade credit and placement of ord- ers; visual inspection of quality by the trader or a trusted associate at each transaction; screening of clients through visual inspection of their shop and repeated interaction; and little or no evi- dence of reputation mechanisms to punish opportunistic breaches of contract. More than a decade after the initiation of market reform in Madagascar, these findings are disturbing and serve as a sobering reminder that, without development of supporting institutions, the free market 25 remains nothing but a flea market. What precise institutions are required is not immediately clear from this work, but results suggest two possible lines of attack. One approach consists in foster- ing the faster and more widespread accumulation of social capital. This could, for instance, be achieved by facilitating interaction and trust among traders, for instance by establishing a Chamber of Commerce or by developing of informal clubs and other brotherhoods.15 A second approach would be to limit the need for social capital by reducing market imperfections, e.g., by setting up institutions that facilitate payments (e.g., faster check clearing), expedite inspection of quality (e.g., grading), reduce insecurity (e.g., police), circulate information (e.g., radio programs, credit reference bureau), penalize cheaters (e.g., pursue fraud), or reduce risk (e.g., bank line of credit, futures market). The results presented here suggest that successful traders owe their success not to individu- alism but to relationships. If anything, the evidence indicates that it is those who can create and nurture relationships who prosper as traders. Perhaps this is not original. After all, in the popular psyche, the trader is often portrayed as someone who is jovial and relates well with others. But the role of relationships is often overlooked in standard economic models that emphasize the maximization of profit through the accumulation of capital and the command of labor. There is also a social dimension to success, one that relies on the accumulation of valuable business rela- tionships, of social network capital. Among traders, this accumulation process is one’s passport to prosperity because it gives better access to information and risk sharing and it reduces the costs of search, quality control, and contract enforcement. ________________ 15 See, for instance, the description of the role that brotherhoods play in building up social capital among traders in Geertz, Geertz, and Rosen (1979). The problem with brotherhoods is that they may restrict entry and favor their members at the expense of outsiders (e.g., Taylor (1997), Fafchamps (1998)). 26 References Abt Associates,The Rural Marketing System in Madagascar, USAID, Antananarivo, 1991. Amselle, J.,Les Négociants de la Savanne, Editions Anthropos, Paris, 1977. Barr, A. M., Social Capital and Technical Information Flows in the Ghanaian Manufacturing Sector, Center for the Study of African Economies, Oxford University, Oxford, May 1997. (mimeograph). Barrett, C. B. and Dorosh, P. A., ‘‘Nonparametric Evidence from Rice in Madagascar,’’ American Journal of Agricultural Economics, 78: 656-669, August 1996. Barrett, C. B., ‘‘Food Marketing Liberalization and Trader Entry: Evidence from Madagascar,’’ World Development, 25(5): 763-777, May 1997a. Barrett, C. B., ‘‘Liberalization and Food Price Distributions: ARCH-M Evidence from Madagascar,’’Food Policy, 22(2): 155-173, 1997b. Bauer, P. T.,West African Trade: A Study of Competition, Oligopoly and Monopoly in a Changing Economy, Cambridge U.P., Cambridge, 1954. Berg, E., ‘‘The Liberalization of Rice Marketing in Madagascar,’’World Development, 17, no.5: 719-728, 1989. Bernstein, L., ‘‘Merchant Law in a Merchant Court: Rethinking the Code’s Search for Immanent Business Norms,’’University of Pennsylvania Law Review, 144(5): 1765-1821, May 1996. Bigsten, A., Collier, P., Dercon, S., Fafchamps, M., Gauthier, B., Gunning, J. W., Isaksson, A., Oduro, A., Oostendorp, R., Patillo, C., Soderbom, M., Teal, F., and Zeufack, A.,Contract Flexibility and Conflict Resolution: Evidence from African Manufacturing, Department of Economics, Stanford University, Stanford, March 1998. (mimeograph). Coate, S. and Ravallion, M., ‘‘Reciprocity Without Commitment: Characterization and Performance of Informal Insurance Arrangements,’’J. Dev. Econ., 40: 1-24, 1993. Cohen, A.,Custom and Politics in Urban Africa: a Study of Hausa Migrants in Yoruba Towns, University of California Press, Berkeley, 1969. Coleman, J. S., ‘‘Social Capital in the Creation of Human Capital,’’Amer. J. Sociol., 94(Supplement): S95-S120, 1988. Cornell, B. and Welch, I., ‘‘Culture, Information and Screening Discrimination,’’J. Polit. Econ., 104(3): 542-571, 1996. Cuevas, C., Hanson, R., Fafchamps, M., Moll, P., and Srivastava, P.,Case Studies of Enterprise Finance in Ghana, RPED, The World Bank, Washington D.C., March 1993. (draft). Dorosh, P. and Bernier, R., ‘‘Staggered Reforms and Limited Success: Structural Adjustment in Madagascar,’’Adjusting to Policy Failure in African Economies, p. 332-365, David Sahn (ed.), Food Systems and Agrarian Change Series, Cornell University Press, Ithaca and London, 1994. Fafchamps, M., ‘‘Solidarity Networks in Pre-Industrial Societies: Rational Peasants with a Moral Economy,’’Econ. Devel. Cult. Change, 41(1): 147-174, October 1992. 27 Fafchamps, M., Biggs, T., Conning, J., and Srivastava, P.,Enterprise Finance in Kenya, Regional Program on Enterprise Development, Africa Region, The World Bank, Washington, D.C., June 1994. Fafchamps, M., Pender, J., and Robinson, E.,Enterprise Finance in Zimbabwe, Regional Program for Enterprise Development, Africa Division, The World Bank, Washington, D.C., April 1995. Fafchamps, M., ‘‘The Enforcement of Commercial Contracts in Ghana,’’World Development, 24(3): 427-448, March 1996. Fafchamps, M., ‘‘Trade Credit in Zimbabwean Manufacturing,’’World Development, 25(3): 795-815, 1997. Fafchamps, M.,Market Emergence, Trust and Reputation, Stanford University, Stanford, February 1998a. (mimeograph). Fafchamps, M.,Ethnicity and Markets: Supplier Credit in African Manufacturing, Stanford University, Stanford, February 1998b. (mimeograph). Fafchamps, M. and Minten, B.,Returns to Social Capital Among Agricultural Traders: Evidence from Madagascar, Department of Economics, Stanford University, Stanford, June 1998. (mimeograph). Fukuyama, F.,Trust: The Social Virtues and the Creation of Prosperity, The Free Press Paperbacks, New York, 1995. Gabre-Madhin, E.,Grain Markets in Ethiopia, Food Research Institute, Stanford University, Stanford, August 1997. (mimeograph). Gambetta, D.,Trust: Making and Breaking Cooperative Relations, Basil Blackwell, New York, 1988. Geertz, C., Geertz, H., and Rosen, L.,Meaning and Order in Moroccan Society, Cambridge U. P., Cambridge, 1979. Ghosh, P. and Ray, D., ‘‘Cooperation in Community Interaction Without Information Flows,’’ Review of Economic Studies, 63: 491-519, 1996. Granovetter, M., ‘‘Economic Action and Social Structure: The Problem of Embeddedness,’’ Amer. J. Sociology, 91(3): 481-510, 1985. Greif, A., ‘‘Contract Enforceability and Economic Institutions in Early Trade: The Maghribi Traders’ Coalition,’’Amer. Econ. Rev., 83(3): 525-548, June 1993. Greif, A., ‘‘Cultural Beliefs and the Organization of Society: A Historical and Theoretical Reflection on Collectivist and Individualist Societies,’’J. Polit. Econ., 102(5): 912-950, 1994. IFPRI, Une Analyse Descriptive du Marché des Intrants et des Produits Agricoles et du Comportement des Ménages Ruraux à Madagascar, Ministère de la Recherche Scientifique de Madagascar et International Food Policy Research Institute, Washington D.C., April 1998. Special Report Number Two, Submitted to USAID/Madagascar. Jones, W. O.,Manioc in Africa, Stanford University Press, Stanford, 1959. Jones, W. O.,Marketing Staple Food Crops in Tropical Africa, Cornell University Press, Ithaca, 1972. 28 Kandori, M., ‘‘Social Norms and Community Enforcement,’’Review Econ. Stud., 59: 63-80, 1992. Kranton, R. E., ‘‘Reciprocal Exchange: A Self-Sustaining System,’’Amer. Econ. Rev., 86(4): 830-851, September 1996. Kristjanson, P. and Martin, J., ‘‘More Than Free Markets are Needed: The Case of a Peasant Millionaire in Madagascar,’’Choices, 6(4): 19-21, FourthQuarter 1991. Lonely Planet,Madagascar and Comoros, Lonely Planet Puliblications, New York, 1994. Lund, S. and Fafchamps, M.,Risk Sharing Networks in Rural Philippines, Department of Economics, Stanford University, Stanford, April 1997. (mimeograph). Mankiw, N. G., Romer, D., and Weil, D. N., ‘‘A Contribution to the Empirics of Economic Growth,’’ Quarterly J. Econ., CVII: 407-437, 1992. Meillassoux, C.,The Development of Indigenous Trade and Markets in West Africa, Oxford University Press, Oxford, 1971. Minten, B., Randrianarisoa, C., Zeller, M., and Badiane, O., ‘‘Accessibilité au Marché des Produits Agricoles et Prix aux Producteurs dans les Villages Ruraux à Madagascar,’’ Cahier de la Recherche sur les Politiques Alimentaires, 5, IFPRI-FOFIFA, Antananarivo, Decembre 1997. Montgomery, J. D., ‘‘Social Networks and Labor-Market Outcomes: Toward an Economic Analysis,’’ Amer. Econ. Rev., 81(5): 1408-1418, December 1991. Platteau, J., ‘‘Behind the Market Stage Where Real Societies Exist: Part II - The Role of Moral Norms,’’ J. Development Studies, 30(4): 753-815, July 1994. Platteau, J.,Traditional Sharing Norms as an Obstacle to Economic Growth in Tribal Societies, CRED, Facult(e’s Universitaires Note-Dame de la Paix, Namur, Belgium, June 1996. Cahiers No. 173. Putnam, R. D., Leonardi, R., and Nanetti, R. Y.,Making Democracy Work: Civic Institutions in Modern Italy, Princeton University Press, Princeton, 1993. Roubaud, F., ‘‘La Question Rizicole à Madagascar: Les Résultats d’une Décenie de Libéralisation,’’Economie de Madagascar, 2: 37-62, 1997. Shuttleworth, G., ‘‘Policies in Transition: Lessons from Madagascar,’’World Development, 17(3): 397-408, March 1989. Staatz, J. M., Dione, J., and Dembele, N. N., ‘‘Cereals Market Liberalization in Mali,’’World Development, 17, no.5: 703-718, 1989. Tadelis, S.,What’s in a Name? Reputation as a Tradable Asset, Department of Economics, Stanford University, Stanford, March 1998. (mimeograph). Taylor, C. R.,The Old-Boy Network and the Young-Gun Effect, Department of Economics, Texas A&M University, College Station, TX, February 1997. (mimeograph). The World Bank,World Development Report 1997, Oxford University Press, New York, 1997. Table 1. Breakdown of the Sample by Size and Occupational Category Size (1): %Total%Large%Medium%SmallOccupation: 10.4%7521.4%528.7%220.4%1Assemblers 19.9%14432.5%7915.4%3911.5%26Wholesalers 11.3%8215.2%3712.6%325.7%13Semi-wholesalers 45.4%32929.6%7253.9%13752.9%120Retailers with a fixed selling point 13.0%941.2%39.4%2429.5%67Retailers without fixed selling point 724243254227Total (1) Size categories are based on total sales. Table 2. Size of Operation and Gross Margin All figures given in US$. All figures subject to considerable measurement error. No. ofGini Firm size: observ.coef.TotalLargeMediumSmall1. Size of operation 7240.76132788635908196Value of monthly sales 3/96-3/97 7170.747281268961198173Value of monthly purchases 3/96-3/97 7390.75012542421904395Value of monthly sales 10/97 7390.74810372015740321Value of monthly purchases 10/97 2. Margin 6850.702463119819326Gross margin per month 14.1%13.9%21.3%13.4%Margin rate = gross margin/sales (1) (1) Computed as average gross margin/average sales. Table 3. Capital and Manpower No. ofSize: observ.TotalLargeMediumSmall1. Capital and equipment 71420614949829331Working capital in $ 72711.0%12.4%13.8%5.7%% with outside funding 54743.4%36.0%31.2%66.0%meanOpportunity cost of funds 54720.0%30.0%20.0%20.0%median in percent per year (1): 72426.262.99.55.1Storage capacity in MT 73939988516926Equipment value in $ 7390.10.30.10.0No. of vehicles 2. Communication 7295.1%11.5%2.0%1.8%% with telephone 72956.5%58.0%52.4%59.0%% with access to telephone 72916.2%32.9%7.1%8.8%% using telephone 7290.5%0.8%0.4%0.4%% with fax machine 72921.8%35.8%18.9%10.6%% with access to fax 7290.8%2.0%0.0%0.4%% using fax machine 3. Management 7266.17.65.94.7Year since business has started 73987.3%87.2%93.7%85.0%% Full time traders 72983.4%84.0%85.8%80.6%% All year traders 4. Manpower 7291.01.41.10.6No. unpaid family help 7291.02.30.40.2No. permanent employees 7291.33.10.50.2No. temporary employees 7294.37.83.02.0Total manpower (2) 72911.011.211.110.7Months of owner's time 72910.914.611.86.0Months of family help's time 72911.327.54.31.8Months of permanent empl. 7297.319.31.90.6Months of temporary empl. 72940.572.729.119.1Total months 7390.20.60.00.0No. of collectors (1) Obtained from the answer to the question "How much could you pay back in 6 months if you could borrow [the equivalent of US$20]?" and expressed in percent per year. (2) Owner/manager counted as 1. Table 4. Human Capital, Family Background, Wealth, and Location No. ofSize: observ.TotalLargeMediumSmall1. Characteristics of owner 71145.7%32.4%47.0%59.3%% Female owners 70937.640.136.935.6Age 7359.110.18.88.1Years of schooling 7119.1%4.6%6.8%16.7%% Non-christian 7391.2%2.5%0.4%0.9%% Foreign 7291.471.621.421.35No. of languages spoken 2. Family of owner 73976.7%87.2%80.3%65.6%% Married 7113.33.63.23.0No. of children 7390.91.00.90.7No. sons aged 15 & above 7390.80.90.80.7No. daughters aged 15 & above 7392.52.62.42.5No. brothers aged 15 & above 7392.42.52.42.4No. sisters aged 15 & above 5763.03.42.92.8Years of schooling of father 5762.62.92.42.4Years of schooling of mother 3. Wealth 73956%61%45%61%% who own house 739198036661295894Value of house in $ 7395%12%1%3%% who own pers. vehicle 4. Location 73916%14%21%12%% who operate in capital city 73931%36%29%31%% who operate in other city 73953%50%50%57%% who operate in rural areas 73218%18%23%12%% who operate in Tana/Hauts Plateaux 73220%26%25%9%% who operate in Vakinantaratra 73225%33%17%26%% who operate in Fianar/Hauts Plateaux 73211%10%15%9%% who operate in Fianar/Cote et falaise 73212%6%13%17%% who operate in Majunga/Plaine 73213%7%7%28%% who operate in Majunga/Hauts Plateaux Table 5. Factors Important for Success As Perceived by Traders To facilitate comparison, cumulative percentages of answers are reported Size: TotalLargeMediumSmall A. Personal reputation and relationships 5%3%6%7%Not important 15%9%19%17%A little important 29%23%27%38%Important 100%100%100%100%Very important B. Access to Credit 39%28%28%64%Not important 70%65%63%84%A little important 89%83%89%96%Important 100%100%100%100%Very important C. Granting Credit 50%40%46%63%Not important 82%75%82%90%A little important 97%94%98%98%Important 100%100%100%100%Very important D. Purchase Price 5%5%2%7%Not important 26%33%19%27%A little important 70%72%67%72%Important 100%100%100%100%Very important E. Sale Price 2%1%1%1%Not important 17%18%11%21%A little important 65%63%59%72%Important 100%100%100%100%Very important F. Transport Equipment 32%27%31%37%Not important 49%46%44%56%A little important 73%68%69%84%Important 100%100%100%100%Very important 729243254227No. observations Table 6. Family and Business Size: TotalLargeMediumSmallA. Family in Trade 25.8%26.3%24.0%27.8%% with parent in trade 4.15.13.83.4No. years father in trade 4.15.13.34.1No. years mother in trade 18.0%17.7%14.2%22.9%% with parent in agricultural trade 2.64.01.92.2No. years father in agr. trade 2.84.01.82.7No. years mother in agr. trade B. Help at Startup 53.2%56.8%48.8%54.2%% helped at startup by family/friends 30.7%25.1%27.2%39.2%Learned working with parents/relative 14.8%14.0%15.8%15.4%Learned working with friend/partner 2.2%2.9%1.2%0.9%Learned as employee of trader 52.2%58.0%55.9%44.5%Learned alone C. Contacts 1.92.11.71.8No. relatives with wage job 0.90.90.90.8No. relatives in trade 0.70.80.70.7No. relatives in agricultural trade 8.810.010.36.3No. traders known personally 739243254227Number of observations Table 7. Sources of Information on Market Conditions Table reports the main source of information on the following: Firm size: TotalLargeMediumSmallA. Prices: 59.9%39.9%60.6%81.1%Other traders 28.3%37.4%31.1%15.0%Suppliers and clients 11.5%22.6%7.9%3.5%Messengers 0.3%0.0%0.4%0.4%Public sources B. Supply conditions: 23.2%18.9%19.7%32.2%Other traders 70.2%68.3%76.4%64.8%Suppliers and clients 5.9%12.3%3.5%1.8%Messengers 0.7%0.4%0.4%1.3%Public sources C. Demand conditions: 16.5%10.3%10.6%30.0%Other traders 77.5%79.0%85.8%67.8%Suppliers and clients 3.7%8.6%1.6%0.9%Messengers 2.3%2.1%2.0%1.3%Public sources 729243254227Number of observations Table 8. Information Sharing To facilitate comparison, cumulative percentages of answers are reported. No. ofSize: observ.TotalLargeMediumSmall 1. Discuss product quality with other traders: 7252%2%2%2%At least once a day 72513%8%11%19%At least once a week 72525%20%27%28%At least once a month 72578%73%73%87%At least once a year 725100%100%100%100%Never 2. Discuss bad paying clients with other traders (1): 3391%2%1%0%At least once a day 3393%4%3%0%At least once a week 33913%14%18%2%At least once a month 33977%79%71%81%At least once a year 339100%100%100%100%Never 3. Discuss prices with other traders (1): 3394%4%3%2%At least once a day 33918%14%16%40%At least once a week 33932%28%31%47%At least once a month 33980%80%76%87%At least once a year 339100%100%100%100%Never (1) Asked to traders with regular clients only. Table 9. Presence of Regular Partners and Ease of Search Ever fail to find a supplier: TotalOftenOccas.NeverRegular suppliers: 51.2%42.9%59.8%49.0%% with regular suppliers 3.61.53.34.4No. of regular suppliers 72984241404Number of observations 100.0%11.5%33.1%55.4%Percentage of sample Ever fail to find a client: TotalOftenOccas.NeverRegular clients: 71.2%47.0%75.0%76.0%% with regular clients 5.82.85.96.5No. of regular clients 729116162451Number of observations 100.0%15.9%22.2%61.9%Percentage of sample Table 10. Regular Suppliers and Clients Size: TotalLargeMediumSmall A. Regular suppliers: 51.2%62.6%59.4%33.0%% with regular suppliers 3.66.23.41.4No. of regular suppliers 36.7%45.6%42.9%22.8%% purchases from regular suppliers 4.14.74.13.1No. years known reg. suppliers (1) B. Regular clients: 71.2%88.9%71.3%52.0%% with regular clients 5.88.35.83.0No. of regular clients 26.8%39.9%26.1%13.3%% purchases to regulars 3.84.24.02.3No. years known reg. clients (2) (1) Computed for the respondents with regular suppliers only. (2) Computed for the respondents with regular clients only. Table 11. Trade Credit Size: TotalLargeMediumSmallA. Credit from and to suppliers: 82.3%79.4%76.9%90.8%% purchases cash 15.8%17.2%21.1%9.1%% purchases on credit 1.8%3.3%2.0%0.1%% purchases advance payment 6.2%5.7%7.7%2.7%ratio payables/monthly sales B. Credit to and from clients: 85.8%76.4%86.1%94.8%% sales cash 13.6%22.4%13.3%5.2%% sales on credit 0.6%1.2%0.7%0.0%% sales advance payment 11.6%16.1%9.8%6.6%ratio receivables/monthly sales 739243254227Number of observations Table 12. Loss of Trade Credit in Case of Non-Payment Non-payment by clientto supplier 21%11%No loss of supplier credit 59%40%Loss of credit from some other suppliers 15%31%Loss of credit from most other suppliers 5%17%Loss of credit from all other suppliers 344194No. observations (1) (1) Computed for the respondents with regular suppliers only. Table 13. Difficulty of Finding Suppliers If Lose One Size: TotalLargeMediumSmall 8%10%8%6%Very easy 16%20%18%3%Fairly easy 44%41%43%56%Fairly difficult 31%30%31%36%Very difficult 194718736Number of observations Computed for the respondents with regular suppliers only. Table 14. Frequency of Contractual Problems Size: TotalLargeMediumSmall1. With suppliers: 7.811.55.56.5No. transactions per month 0.280.480.300.07No. cases deficient quality per month 0.070.100.080.03No. cases late deliveries per month 14.8%19.2%17.8%7.4%% traders who place orders Average incidence of problems: 31.7%40.7%28.3%17.8%Among firms that place orders 3.5%3.9%4.4%2.4%Among firms that do not place orders 7.7%10.7%8.3%3.6%Over all firms 2. With clients: 323261325386No. transactions per month 0.681.120.770.14No. cases of late payment per month 0.020.040.030.00No. cases of non payment per month 13.6%22.4%13.3%5.2%% sales on credit Average incidence of problems: 4.5%5.4%4.5%2.1%Among firms that sell on credit 0.3%0.5%0.3%0.3%Among firms that do not sell on credit 2.2%3.7%2.3%0.7%Over all firms Table 15. Verification of Quality of Products Size: TotalLargeMediumSmall1. Quality inspection by respondent 84%78%83%92%% always inspect quality before purchas 94%89%93%99%% owner inspects quality 5%6%7%0%% family helper inspects quality 2%5%0%0%% employee or agent inspects quality 2. Quality inspection by clients 86%82%86%90%% client always inspect quality 3. Action taken by respondent if supplies are of bad quality: 55%46%49%69%None/quality is the buyer's problem 28%36%31%18%Supplier provides a refund/replacement 17%19%21%13%Other 4. Action taken by client if supplies are of bad quality: 62%52%58%77%None/quality is the buyer's problem 21%26%25%13%Supplier provides a refund/replacement 17%22%17%10%Other Table 16. Credibility of Clients Questions were asked only to respondents who ever grant credit to clients. Size: TotalLargeMediumSmall 1. Respondent verifies credibility of client before sale (1): 6%5%7%6%Never 9%9%9%8%Seldom 29%30%24%38%Sometimes 47%46%37%72%Often 100%100%100%100%Always 2. Sources of information consulted before granting credit: 96%94%97%98%% get information from client directly 27%33%28%9%% visit client's shop 24%25%15%38%% obtain information from other traders 1%2%0%2%% get information from client's bank 12%8%11%23%% get information from other sources (1) Cumulative percentages reported to facilitate comparison. Table 17. Conflict Resolution Method ClientSupplier 1. Conflict resolution method 93.6%91.3%Direct negotiation 9.1%3.8%Mediator 0.5%0.0%Lawyer 3.6%0.0%Threat going to police 0.9%0.4%Threat going to court 2. Outcome of conflict 80.9%81.3%Problem resolved 75.0%87.5%Still trading with party 220160Number of observations Table 18. Risk Sharing and Access to Financial Help Size: TotalLargeMediumSmall 76.3%80.2%77.2%72.2%% who has ever helped others 75.0%74.5%75.6%76.2%% who has ever been helped by others 2.32.72.51.7No. people who can help 739243254227Number of observations Figure 1: "The poor are poor because they are lazy" 0 20 40 60 80 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 2: "The poor are poor because they have nobody to help them" 0 20 40 60 80 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 3: "I have put money aside for difficult times" 0 20 40 60 80 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 4: "If my business fails, I would have to sell everything to survive" 0 10 20 30 40 50 60 70 80 90 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 5: "I will help the others if they are in need" 0 20 40 60 80 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 6: "I can count on family and friends in time of financial problems" 0 10 20 30 40 50 60 70 80 90 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 7: "If my business prospers, my family and friends will try to live off me" 0 10 20 30 40 50 60 70 80 90 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 8: "I'm only proud of what I accomplish without the help of others" 0 20 40 60 80 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large Figure 9: "If I had a lot of money, I would invest it in business" 0 10 20 30 40 50 60 70 80 90 100 Completely true A little bit true Nor true nor false A little bit wrong Completely wrong Small Medium Large