barter

Barter

Patrick Heady

‘Barter’ is a non-technical English term which anthropologists have applied to a range of transactions that share certain characteristics. Barter typically denotes the direct exchange of goods or services for each other without the medium of money. Within this broad class of exchanges, the term is generally restricted to those in which the prime focus of interest for the exchange partners is in the goods and services themselves rather than the social relationships arising from the exchange: where social relations are the prime focus of interest the transaction is usually referred to as gift exchange. However, as we shall see later on, the boundary between barter and gift exchange can be rather fuzzy.

I shall start by looking at the practical advantages and disadvantages of barter compared with exchanges mediated by money, paying a good deal of attention to the questions of ‘transaction costs’ and how to ensure ‘coincidence of wants’. After that, I shall look more closely at the relation between barter and gift exchange. The discussion will be framed within a more general contrast between exchanges in which the partners emphasize their own material advantage at the expense of building goodwill between them, and exchanges (of which the most pronounced are outright gifts) in which the partners forgo some material advantages in order to strengthen their relationship. I shall look at how far, and when, exchanges can become purely material, and whether these circumstances are more typical for barter or money-mediated exchange. I shall also look at situations in which those concerned wish to mark certain exchanges as containing a social element, and the ways in which the difference between monetary and barter exchange can sometimes be used to make the distinction.

Barter can thus be understood from both economic and social perspectives. The final issue is, then, how these two perspectives relate to each other. Is there anything about barter transactions, or at least certain kinds of barter transaction, which implies that they are less subject to economic principles than money-mediated exchanges? And where there is a distinction between more gift-like and more self-centered transactions, is the element of social relationship ever totally missing from the latter, or does it simply take a different form than in the case of the gift? It is the light which barter, in its various forms, throws on questions like these which gives it its wider significance for anthropological theory.

Transaction costs and the problem of identifying a coincidence of wants

In order to clarify the economic logic of barter, we need a notion of cost that does not need to be expressed in monetary terms. In order to grasp this notion of cost, it makes sense to start by thinking of very simple exchange situations. In fact, we can start by considering a situation in which no exchange takes place at all. Let us imagine a group of people who live in total isolation and produce for themselves all the goods they want, using the natural resources available in their locality. Suppose that, among other things, they produce fruit. Does this fruit have a cost? Clearly it does not have a monetary cost, since no exchange is involved. However, there is a sort of cost involved, in the sense that the amount of time and effort required to prune the fruit trees and to pick and store the fruit is not available for other productive purposes, or for leisure. This cost, the amount of alternative goods that must be forgone in order to obtain the fruit, is referred to by economists as the ‘opportunity cost’ of the fruit. Since there are several other things that could have been done with the time and effort needed to produce the fruit, this opportunity cost could be described in several different ways: in terms of possible improvement to the group’s housing, of more time available for hunting or even in terms of more time available for resting.

What happens when we introduce the possibility of exchange into the setup we have just described? Imagine a situation in which there were two groups of people, who lived in isolation from everyone else. Imagine further that each group of people produced nearly all their needs directly from domestic production, but that each group needed one thing that only the other could provide. Suppose that the first group lived inland and produced fruit but no fish, and that the second group lived on the coast and produced fish but no fruit. Suppose further that the people in each group enjoy eating both fish and fruit. Then we have a simple exchange set-up. It makes sense for each group to produce more of its unique product (fruit or fish) to exchange with the unique product of the other group. In other words it makes sense to barter.

Suppose that the exchanges took place on the coast, and that both sides were content with a rate of exchange in which the inland group gave three fruit for each fish that they received from the coastal group (we shall discuss the factors determining this rate of exchange later). In a sense we can now talk about a price: the price of one fish is three fruit, and the price of one fruit is a third of a fish; or more precisely we can say that these are the prices prevailing in exchanges that take place on the coast. From the point of view of the coastal group, this price reflects the true opportunity cost of the fruit, since once they have handed over a fish they can immediately start eating the three fruit which they have received in exchange. But this is not so for the people who have come from inland to bring the fruit and carry back the fish. From their viewpoint, the opportunity cost of each fish is not simply the three fruit given in exchange, but also whatever else they could have done with the time and effort needed to make the journey to the coast and back. This additional opportunity cost is the transaction cost. When they get back to their inland village, the trading party will want some reward for their time and effort, and as a result they will not be willing to hand over fish to their neighbours at the rate of one fish for three fruit. Perhaps they will ask for four fruit for each fish. The total cost of a fish in the inland village is then four fruit, three of which represent the purchase price of the fish on the coast, and one of which represents the transaction costs. If the trading journey had taken place in the opposite direction, with coastal people carrying their fish inland to exchange in the inland village, the story would be the same. The purchase price of each fruit (in the inland village) would be one-quarter of a fish, but its sale price (on the coast) would be one-third of a fish. The difference (one-twelfth of a fish) represents the transaction cost.

If the transaction cost were higher, each community would be worse off, since the opportunity cost of fish in the inland village, and of fruit in the coastal village, would both be greater. In the example we have considered, in which the transaction costs were simply due to the time and effort involved in transportation, one could expect the transaction costs to be higher if the villages were further apart. However, there are other factors as well which might increase the transaction costs. These all involve, in one way or another, the problem of ‘coincidence of wants’, the difficulty of bringing together a person who can offer good A and wants good B, with a suitable trading partner who wants good A and can offer good B. These are the problems which money can help with, and their importance explains why most trade in the modern world is carried out with the medium of money.

In order to illustrate these problems, let us develop our example a little bit more. Suppose that the fishing season takes place in the spring, and the fruit-picking season in the autumn, and that neither good is easily conserved. In that case, when the inland people bring their fruit down to the coast, the coastal people will have nothing to offer in exchange. If they are to trade their fruit at all, the inland people will have to offer the fishing people credit, and hope that they will fulfill their side of the bargain by delivering the fish that they owe next spring. If the inland people do not know their coastal trading partners very well, there is always the risk that they may give some of their fruit to unreliable individuals who will fail to deliver the due amount of fish. This risk of default on credit amounts to an additional transaction cost, and therefore will tend to discourage trade between the two villages. It also provides the first situation we have considered in which money exchange would have an advantage over barter. If there were a form of money available which was accepted by both the coastal and the inland people, then it would be possible for the inland people to insist on payment in this currency when they delivered the fruit in the autumn, and use the same currency to purchase the fish they required in the spring. The same would, of course, apply to transactions initiated by the coastal community. The necessity of credit and the risk of default would have been removed, thus reducing the transaction costs for both communities.

Problems of the coincidence of wants also arise when the economy becomes more complex. Suppose that some of the people in the inland community grow apples while others grow pears, and that some of the coastal community fish for herrings while others catch octopus. Suppose that a man from the inland community has a load of pears which he wishes to exchange for octopus. There is no guarantee that the first person he meets when he arrives at the coast will wish to exchange octopuses for pears. Half of the people he meets will offer him herrings, which he does not want, while half of the rest would rather exchange their octopuses for apples than for pears. He will, therefore, have to spend some time, and hence opportunity cost, contacting the right person to make the exchange. This problem is perhaps not very serious in our example. However, when you start to think of more complex economies in which many householders wish to obtain dozens, or even hundreds, of different kinds of good by exchange every week, then it becomes clear that arranging for each person to meet with people who wanted to make a direct exchange between the product he or she had for sale, and the many different products he or she wished to buy, would be a very complex business indeed. However, if there were a currency which all concerned were willing to accept, this problem would be greatly reduced. In that case, a man who arrived at the coastal village with pears to sell would only need to identify a few people who wanted to buy pears, and it would not matter whether or not they wished to sell octopuses, or any of the other goods he wished to buy. All he need do is sell his pears to those few people for money, and use the money to buy octopuses and other things from the people who had them to sell, secure in the knowledge that they would accept cash, even if they did not care for pears.1

Some real-world examples of the choice between monetary exchange and barter

The essential point of the previous section is that money provides advantages in situations where it is not easy to bring together partners who want to make corresponding exchanges of actual goods. If the argument is sound, we would expect to encounter barter in situations where finding a partner with coincident exchange wants is relatively unproblematic, or where the transaction-cost advantage of monetary exchange is counteracted by some other disadvantage. In fact there is a good deal of evidence which is consistent with these expectations.

Some real situations correspond quite closely to the artificial example discussed above, particularly when, in mountainous environments, neighbouring communities at differing altitudes specialize in different crops, so that the existence of demand for one’s own crop and of the supply of the neighbouring crop are well known to all concerned. Among the examples of this phenomenon are communities in Bulgaria and in Nepal.

Closely related are situations in which one crop, potatoes or grain, can be bartered for a wide range of goods. Here the disadvantage of barter compared with money is partly removed by the fact that demand for the core foodstuff is so nearly universal that its acceptability as a means of payment is assured. In many Indian villages, various specialists used to be paid in grain instead of, or as well as, cash (see Harriss chap. 33 infra).2 Cellarius’s study of barter exchange in a Bulgarian mountain village provides a particularly neat illustration of the importance of transaction costs. The village concerned is rather isolated, and most residents did not possess their own motor transport. Traders drove trucks up from the lowland towns with all kinds of goods, which they often exchanged for potatoes. The arrangement must have suited the traders, because they could use the empty space in their trucks to carry the potatoes back to town for very little cost. There was not much disadvantage for the local people since, in the village, everything that could be bought for cash could also be paid for with potatoes. Logically enough, the only people interviewed by Cellarius who insisted on selling their potatoes for cash, and making all their purchases in cash, were a family with their own motor transport, who could therefore take advantage of the greater flexibility that cash offered for making purchases in the more variegated shops and markets outside the valley.

Another factor which, in recent years, had pushed the inhabitants of Cellarius’s village towards barter had been a dramatic fall in their cash incomes, occasioned by the closure of the local collective farm, which had previously provided many of the village people with money wages. The villagers moved away from the money economy because money was scarce. People can also move to barter when they no longer trust money because of inflation, devaluation or bank defaults. A combination of increased scarcity of money, and a sense of insecurity about whether it could hold its value, seems to have been one of the factors behind the very widespread use of barter by Russian businesses in the late 1990s. The unusual extent of inter-business barter in 1990s Russia may have had another cause as well. Before the collapse of communism, a complex system of state planning organized the distribution of materials and outputs between different enterprises. Even after the formal freeing of state enterprises from central control, many of the interpersonal contacts established through this system were still in place, and could be used to identify complex systems of multi-party exchange which would result in a coincidence of wants and supplies. This may have made inter-enterprise barter in Russia more feasible than it would have been in a system without that country’s heritage of detailed state planning.

A final point is that monetary transactions are often subject to tax. In such situations barter may provide a way of avoiding the attentions of the tax-man. Tax avoidance is an important factor in some contemporary barter networks in both Russia and the United States.

Overall, then, the argument that the prevalence or otherwise of barter can be explained by the relative transaction costs of monetary and non-monetary exchange seems to be well supported by available evidence.

Barter, commodities and gift exchange

Having established when barter is likely to take place, it is time to describe and analyze the barter process itself. Here we meet an interesting dichotomy. According to some descriptions, the transacting partners haggle in order to obtain the very best bargain for themselves at the other’s expense, so that the relations between barter partners are characterized by a degree of implicit hostility. Bronislaw Malinowski’s (1978 [1922]: 95–6, 187–90, 361–4) description of the barter deals that took place as part of Trobriand trading expeditions highlights the haggling involved, and contrasts this with the more dignified gift exchange of kula valuables with which the leaders were involved (see Strathern and Stewart chap. 14 supra). At one point in his study of Stone age economics, Marshall Sahlins (1974: 195) suggests that barter should be considered a kind of negative reciprocity that people engage in with outsiders with whom relationships are anyway hostile. However, in the same book, he (1974: 277–314) also describes barter practices which are quite unlike this, in which the trading partners avoid all bargaining, but instead exchange without argument at generally accepted prices. There are many reports of this kind of barter process, which often involves some of the courtesies that go with gift exchange. Indeed, in real-world barter transactions, this rather decorous way of proceeding seems to be just as typical as the haggling approach (Herskovits 1965: 188–96).

In order to understand what is going on here, it is useful to refer to a distinction between commodity and gift exchanges. As defined by Gregory (1982), who draws on the ideas of Marcel Mauss (2002 [1923]) and Karl Marx, commodity transactions involve the exchange of unlike goods and services in order to obtain a material benefit or profit, while the partners in gift exchange present each other with goods and services which are basically alike in order to reinforce the social relationships between them. The commodity perspective is the one adopted by standard microeconomic theory, of the kind which this chapter has drawn on up this point.

Microeconomic theory can also help us understand some aspects of the style of barter transactions. In particular, it can help us to understand why some barter transactions involve adversarial bargaining. As we saw earlier, one of the advantages of monetary exchange over barter is that it reduces transaction costs by replacing the comparatively rare exchange partnerships, in which good A would be directly exchanged for good B, with far more frequent potential partnerships in which good A and good B are both exchanged for money. The relative scarcity of partners available for any particular direct exchange has another disadvantage as well, in that it creates situations in which, although both parties benefit from the transaction, the gain experienced by one party is inversely related to the gain experienced by the other. The argument concerned goes back to F.Y. Edgeworth (1881: 20–30, cited by Pressman 1999: 70–72), who showed that the exchange rate offered in a bartering transaction (or any kind of trade) is fully determined only when each party to the transaction has many potential exchange partners. Where the number of potential partners is small, or simply consists of the two partners actually present, then there will usually be a range of exchange rates that would leave both partners feeling better off than before. This sets limits beyond which one or other party would pull out of the deal altogether. However, within these limits, the particular exchange rate selected will depend on the relative bargaining power and skill of the two parties, in a zero-sum (at least within limits) game in which an increased benefit for one partner represents a decrease in the benefit to the other.

So, within the commodity-exchange framework, we can understand why barter sometimes involves a process of adversarial haggling. By why is this not always the case? One problem with adversarial bargaining is that it is unlikely to leave the exchange partners feeling very friendly to each other. This need not matter if the bargaining takes place in a framework where social rules are guaranteed by some other authority, such as a state, or the local big man who sponsors a particular market. But it would be very serious indeed if the traders came from different and potentially antagonistic ethnic groups. This is often the case when the trade takes place between partners from different ecological zones: highlanders with lowlanders, or between people from different islands. It is in these circumstances that we find the trade taking gift-like forms, such as between trading partners who are ritual friends themselves or who are, as in the case of the barter that takes place in the margins of the kula expeditions, under the protection of expedition leaders who are ritual friends. The transactions have a gift-like element, because they could not take place without the existence of the secure social relationship which the gift-like aspects of the transaction help to ensure.

Nevertheless, however decorously they are conducted, barter transactions between ritual friends do not really fit into Gregory’s category of gift exchange, because the things exchanged are not alike. In a detailed study of gift-like barter between communities at different altitudes in eastern Nepal, Humphrey (1992) argues that the idea that the things exchanged should be of equivalent value provides a basis for a sense of moral commitment between trading partners which resembles that established by gift exchange. This interpretation fits well with the numerous reports of barter exchanges which take place without any bargaining, using locally accepted ‘fair’ prices. The next question is whether the importance of this notion of equivalent value, and the associated instances of gift-like behaviour, interfere with the operation of the commodity-like microeconomic mechanisms described in the first part of this chapter. The answer seems to be that they do not interfere too much. Humphrey (1985) provides data on the variation of exchange rates in relation to the distance between the points of origin of the goods concerned which are broadly in line with the hypothetical analysis given earlier in this chapter, and also reports changing exchange rates in response to major changes in supply conditions. Sahlins commented earlier (1974) on similar data for barter trade in Melanesia. The explanation offered by Sahlins, which is confirmed by Humphrey’s data, is that what counts as a fair price is judged with respect to the range of exchange rates being offered in the locality. If the rate offered by one’s partner is out of line with this, there is no need to engage in adversarial haggling; one simply opens up a new gift-like relationship with a partner whose exchange rate is more reasonable  (Sahlins 1974: 312–14).

Indeed, the ability to weigh up the gains and losses from the whole of a particular exchange relationship, rather than simply those arising from the latest exchange within it, makes it possible to cultivate or withdraw from gift relationships on utilitarian grounds, and therefore provides an incentive for each partner to play his or her part in the relationship, even in the absence of any specific reckoning of the returns resulting from any particular gift from one partner to the other. This point is crucial to understanding the operation of the blat system during the communist period in Russia. As described by Ledeneva (1998), blat arose in response to the chronic shortages of the soviet economy, which were themselves the result of official refusal to let prices rise and fall in order to balance supply and demand. This official refusal of the price mechanism was supported by public opinion which, in addition, strongly disapproved of ‘speculators’ who tried to profit from this situation by  selling scarce goods unofficially at prices above the official ones. However, the same disapproval did not apply to people who merely helped others obtain the scarce goods at the ordinary price. Blat was essentially a system by which people with control over a particular scarce good – anything from shoes to building materials, educational opportunities or medical care – enabled their acquaintances to gain access to a good in scarce supply, hoping that in return the beneficiary would enable them to gain access to something they needed later on. Since the goods themselves were either free or paid for to the supplying organization at the official rate, monetary payments were not involved, nor was it generally a matter of a direct exchange of material benefits. Instead, the relationship was maintained in the expectation that over the long term it would provide benefits to both sides. Once prices were freed in the 1990s and goods became more freely available in the shops, the blat system of exchanging favours lost some of its importance (Ledeneva 1998: 179).

Humphrey and Hugh-Jones (1992: 7) have suggested that barter transactions should be considered as a third category of exchange, distinct from either gift or commodity exchange, that should be studied in its own right. However, the material presented above suggests that it might not be a very good idea to construct a distinct sub-discipline of barter studies. One point is that the different examples of barter are quite diverse, and would be hard to reduce to a single ideal type. A second and more fundamental point is that the things which the different examples of barter have in common are also shared with commodity and gift exchanges. The participants are concerned to obtain access to goods without giving too much in return (as would be expected in commodity transactions). But they are also conscious that the way they conduct their transactions conveys messages about their mutual relationship and needs to be planned in a way that will maintain that relationship (as would be expected in gift exchange). Rather than splitting out a new category of barter, it might be better to say that all exchanges have two aspects: first as transfer of goods or services, and second as a sign of the nature of the relationship between the exchange partners. Commodity exchanges are those in which the partners’ attention is focused on the first aspect, and gift exchanges are those in which attention is focused on the second aspect. In many exchanges, including most barter exchanges but also many monetary exchanges, the partners give some attention to both aspects. In the next section, I shall try to show that this dual perspective helps us to understand the phenomenon of multiple spheres of exchange.

Spheres of exchange

One area to which some of the arguments we have considered above might usefully be applied would be the English family. After all, a marriage involves some kind of division of tasks between the two spouses, and the goods these tasks produce – such as meals, shelter and warmth, clean clothes – matter to each. At the same time, the relationship also matters in itself, and the spouses regularly attempt to strengthen it by exchanging gifts. In a traditional, but now relatively rare, form of marriage the husband undertakes paid work while the wife is responsible for managing the housework. A more common arrangement now is for both partners to do paid work, though the husband may still be the main earner, and for there to be a rather more even division of labour within the household as well.

Does this mean that the economic basis of British marriage is shifting from cash purchase of the wife’s services by the husband, to the bartering of services between the spouses? Given the way I have developed the concept of barter in the previous section, this would be a fairly reasonable statement to make, but it still sounds distinctly odd. Part of the reason may be that the term ‘purchase’ suggests a commodity transaction, which appears to be contradicted by the exchanges of gifts that also go on within the traditional marriage. The arrangements of a traditional marriage thus challenge an implicit assumption of the discussion up to now, namely that monetary payments are exclusively associated with commodity exchange. But the real reason why the statement sounds odd is probably deeper, namely that we do not think of marriage primarily as an exchange of services between the two partners. Instead we see it as the basis of a family and feel that the partners should each subordinate their own interests to that of the family as a whole, and in particular to the well-being of their children. The attitude of mind that this calls for is simply not the same as the attitude of mind involved in commodity transactions (whether by cash or barter), or even gift relationships outside the family home.³

In that sense the family constitutes a different sphere of exchange from the external world of the job market and the shopping centre. Although material self-interest certainly does play a role in intra-family transactions (see Becker 1981), it is under tighter constraints than apply in the economy outside the household, and the resulting difference in exchange rules is reinforced by our feeling that it is inappropriate to apply the same concepts to exchanges in the two spheres.

This point about distinct spheres of exchange was first made by Paul Bohannan (1955) in a classic article about exchange and investment among the Tiv people of northern Nigeria (see Hart chap. 10 supra). He also noted that the Tiv conceptualise exchanges of services between kin entirely separately from exchanges between non-kin. As in the British case, exchanges between kin are meant ideally to be a direct expression of mutual commitment rather than a matter of calculation, though the range of Tiv kin involved is probably rather wider than in the British case. As in the British case, the distinction is reinforced by the use of a different vocabulary to talk about kinship and non-kinship exchanges.

However, among the Tiv, exchanges between non-kin were also divided into three distinct spheres, also distinguished by different vocabularies and conceptual systems: one for minor goods, a second for prestigious goods and the third for the exchange of women in marriage. Although it was possible to use success in a lower sphere to buy access to exchanges in one of the higher spheres, this had not been easy until a few decades before Bohannan’s fieldwork. It was fairly clear that the restrictions were part of the Tiv political system, reinforcing the power of lineage elders and enabling them to control much of the social and reproductive destiny of their clan members. Thus, consistently with the dual nature of exchange as both material transactions and signs of relationships, these restrictions on legitimate exchange transactions reflected both an ordering of ideas and a way of controlling access to resources.

Why this theme belongs in a discussion of barter is that Bohannan himself believed that the system was threatened by the spread of Western money  in Tiv land. He argued that the division into three spheres was dependent on rules about what could be bartered for what, which broke down with the appearance of Western money that could be used to purchase goods in all three spheres.

However, despite the Tiv experience, the general thesis that modern money tends to destroy the distinction between spheres of exchange seems to have rather little empirical support. In the British case, as we have just seen, money is important in both the family and the commercial sphere, but subject to different rules in both. (Indeed it is also crucial to the third sphere of exchange in modern Britain: that between the state and the citizen. Here again it is subject to different rules than in the other two spheres.) In a general review of this issue in their Money and the morality of exchange, Maurice Bloch and Johnathan Parry (1989) concluded that the distinction between a sphere of self-interested exchange and a superior sphere of more prestigious exchange subject to moral rules was a feature of all societies. However, they did not find any general tendency for money to be particularly associated with the potentially immoral sphere of self-interested exchange.

Conclusion

Microeconomic ideas about the costs of transactions are rather effective in identifying the situations in which barter is preferred to monetary exchange, and also help to explain the rates of exchange between goods even when barter takes forms which resemble gift exchange. But in order to understand why a particular series of barter exchanges is more commodity-like or more gift-like we need to take account of the socio-political context of the exchange, and of the dual aspect of exchange episodes as material transactions and as signals of the nature of the personal relationships involved. This part of the analysis, however, applies to money-mediated exchange just as much as to barter, and the phenomenon of distinct spheres of exchange, which some writers associate particularly with barter-based economies, turns out to be more general. Because barter transactions resist oversimplified analysis, they provide a useful way of testing and extending the concepts which we apply to exchange in general, but they do not call for a distinct body of theory unique to barter.

Notes

1. For more technical treatments of the topics discussed in this section, see Anderlini and Sabourian (1992) and Dutta (2000).

2. See Fuller (1989) for a critical discussion of the literature on this topic.

3. See Carrier (1995: 31–5) for a fuller discussion of this point.