Mises Wire

Home | Wire | David Graeber’s Response to My Article

David Graeber’s Response to My Article


In the comments attached to my article Have Anthropologists Overturned Menger?, David Graeber offered a lengthy response. Below we’ve compiled his comments into a convenient stand-alone post. I’ll respond in turn with a follow-up post.


You know serious scholars, if they are trying to refute an argument set forth in a book, actually take a look at the book, rather than writing an immediate refutation based on an interview about it. If the Von Mises Institute is hard up for cash and can’t afford the book, the relevant chapter can be viewed for free on the “look inside this book” function on Amazon.

This critique goes even further, since it is not a critique even of the position set forth in the interview, but a criticism of several paragraphs from an interview, ignoring most of the arguments made in that interview even on the topic in question. This makes sense only if we assume, either

(a) one of the laziest readers in history, or
(b) someone who is not really concerned with understanding my position at all, but only with being able to make the claim they refuted it

Now, presuming there is someone in the Mises Institute or its readership who actually do have an interest in the argument, let me set out what I am actually saying here:

1) The idea that there is a single “origin” of money is rather dubious in itself – if money is simply a mathematical system whereby one can compare proportional values, then something of that sort must have emerged in innumerable different occasions in human history for different reasons. The standard version of how it emerged, however, that goes back to Adam Smith, is repeated by Jevons, Mengers, etc, is one of the least likely, in fact, which is strongly counter-indicated by all existing evidence.

2) The flaw in the barter theory of the origin of money is that barter presumes SPOT TRANSACTIONS. There is no reason whatsoever to presume that neighbors would limit themselves to spot transactions in dealing with one another. However, if one does not presume spot transactions, then the notorious problem of the “double coincidence of wants” does not occur. You end up with a system of broad, non-enumerated credits, and this is precisely what those who actually did research on communities that do not use money did find.

Here the evidence is simply in and you’d think an honest economist would acknowledge it. As Caroline Humphreys of Cambridge in the definitive anthropological work on barter put it, “No example of a barter economy, pure and simple, has ever been described, let alone the emergence from it of money; all available eth- nography suggests that there never has been such a thing.” (By “barter economy” of course she means a community in which this is how everyday goods are normally distributed amongst neighbors.) Arguing that it might have been a brief period of history which we never happen to have noticed anywhere is a perfect example of special pleading. That argument might stand up if we had no examples at all of moneyless economies, so we had to guess what a moneyless economy would actually look like, but in fact, we have endless examples and they don’t look like what the model predicts. A responsible scholar would try to reassess their theory if it proves non-predictive. Saying “yeah, how do you know they weren’t doing in a time and place for which we can’t have evidence?” strikes me as an act of pure defiance of all normal principles of empirical science.

3) I never argued that the moneyless spot trade (barter) did not exist in human history, only that

a) both the historical and anthropological record confirm it occurred almost exclusively between strangers, or those with whom one has no moral relationship
b) it does occasionally occur on a more widespread level in places where some relatively generalized form of money has existed, but where that money disappears or becomes unavailable.

In neither case can such systems lead to the “origins of money,” in the first case because the scenario of numerous actors exchanging numerous unanticipated products with one another, leading to the “double coincidence of wants” problem, would not arise, in the second, of course, because money has already been invented. The example of the POW camp, incidentally, far from refuting anything I say, is a perfect example of what I propose: that when one sees people interacting in terms of barter on an everyday basis, it’s because you are seeing people who grew up using money in everyday transactions who now suddenly have no access to the stuff.

Let’s take (3a) in greater detail. This is basically what my critic is falling back on – ironically, reproducing a theory of the origins of money which, ironically, seems to have been first proposed by Karl Marx, and later by Karl Bucher. Unable to find any plausible reason why neighbors will be limiting themselves to the spot trade, he seems to be arguing that Mesopotamian merchants _must_ have invented money as a medium of exchange through their own spot transactions in long-distance trade for that system to then be adopted as a unit of account in administrative transactions within Temples. Aside from the peculiar circularity of the argument – “since money can only have been invented by barter, if they were using money, it must have been invented at some time earlier through barter” – it flies in the face of both logic, and again, all empirical evidence we actually have about ancient or “primitive” trade.

The reason is simple: if you have irregular, occasional exchange between strangers, it will not generate a money system – since irregular, occasional exchange will not produce any kind of system at all. If you have regular exchange between strangers, it’s because there are specific goods that each side knows they want or need. One has to bear in mind that under ancient conditions, long-distance trade was extremely dangerous. You don’t cross mountains and deserts, or oceans, risking death in a dozen different ways, so as to show up with a collection of goods you think someone might want, in order to see if they happen to have something you might want. You show up because you know there are people who have always wanted woolens and who have always had lapis lazuli. Logically, what such a situation would lead to is a series of conventional equivalences – so many woolens for so many pieces of lapis lazuli – which are maintained despite contingencies of supply and demand, because all parties need to reduce risk or the trade would simply stop. And once again,what logic would predict is precisely what we find. Even in periods of human history where money and markets did already exist, high-risk long distance trade has often continued to be carried out through a system of conventional equivalents, administered prices, between specific commodities that merchants already know will be available, or in demand, at certain pre-established locations.

Now, could such a system generate something like money of account – that is, the use of one or two relatively desirable commodities to measure the value of other ones, once more items were added to the mix (say, you’re making several stops)? Sure. It is likely that in certain circumstances, something like this did happen – but it would have meant that money, in such cases, was created first as a means to avoid market mechanisms, and that it was not used mainly as a medium of transactions, but rather, primarily as a means of account. One could even make up an imaginary scenario whereby once you start using one divisible/portable/etc commodity as a means of establishing fixed equivalents between other ones, you could start using for minor occasional transactions, to measure negotiated prices for spot trade swaps on the side, in a more market-driven way. All that is possible and likely as not did happen here and there. However there is no reason to assume that such a system would produce a concrete medium of exchange actually used in making these transactions – in fact, given the dangers of ancient trade, insisting that some medium like silver actually be used in all transactions, rather than a credit system, would be completely irrational, since the need to carry around such a money-stuff would make one a far, far, more attractive target to potential thieves. A desert nomad band might not attack a caravan carrying lapis lazuli, especially if the only potential buyers were temples which would probably know all the active merchants and know that you had stolen the stuff (and even if you could trade for them, what are you going to do with a big pile of woolens anyway, you live in a desert?) but they’d definitely go after someone carrying around a universal equivalent. (This is presumably the reason why the great long-distance traders of the Classical World, the Phoenicians, were among the last to adopt coinage – if money was invented as a circulating medium for long-distance trade, they should have been the first.)

The other problem is there is no reason to believe that such mechanism – which would presumably only be used by that tiny proportion of the population who engaged in long distance trade, and who tended to treat such matters as specialized knowledge to be guarded from outsiders – could possibly create a money system used in everyday transactions within a society or any evidence that it might have done so.

4) The actual evidence is that in Mesopotamia – the first case we know anything about – these more widespread pricing systems in fact emerged as a side-effect of non-state bureaucracies. Again, non-state bureaucracies are a phenomenon that no economic model would even have anticipated existing. It’s off the map of economic theory. But look at the historical record and there they are. Sumerian Temples (and even many of the early Palace complexes that imitated them) were not states, did not extract taxes or maintain a monopoly of force, but did contain thousands of people engaged in agriculture, industry, fishing, and herding, people who had to be fed and provisioned, their inputs and outputs measured. All evidence that exists points to money emerging as a series of fixed equivalents between silver – the stuff used to measure fixed equivalents in long distance trade, and conveniently stockpiled in the temples themselves where it was used to make images of gods, etc – and grain, the stuff used to pay the most important rations from temple stockpiles to its workers. Hence a silver shekel was fixed as the amount of silver equivalent to the numbers of bushels of barley that could provide 2 meals a day for a temple worker over the course of a month. It was the Temples that actually had a need to extend a silver system from a unit used to compare the value of a limited number of rare items traded long distance, used almost exclusively by members of the political or administrative elite of the societies in question – to something that could be used to compare the values of everyday items, like planks of wood, jugs of beer, and so forth. The development of local markets within cities, in turn, came as a side effect of these systems, and all evidence shows they too operated primarily through credit. For instance, Sumerians, though they had the technological means to do so, never produced scales accurate enough to weigh out the tiny amounts of silver that would have been required to buy a single cask of beer, or a woolen tunic, or a hammer – the clearest indication that even once money did exist, it was not used as a medium of exchange for minor transactions, but rather as a means of keeping track of transactions made on credit.

4) As for the supposed refutation of my example of the village where people loan things to one another, no, the commentator does not get my argument right at all. First of all, we are not dealing with a situation where people borrow things from one another and expect an equivalent of exactly the same value. I suppose the certain Austrian school theories of human nature assume that’s what neighbors in a moneyless economy _would_ do with one another, but again, this just shows a flaw in those theories of human nature, because when tested against the empirical evidence, this is not what one finds. What one finds are a variety of mechanisms of distribution, some open-ended sharing, some relatively centralized allocation (the actual Iroquoian societies that people like Jevons used as examples mainly had women’s councils allocate things and didn’t swap directly between households at all), and some direct exchange, particular with people in that vague middle ground between neighbors and strangers – but that exchange was based not on exact value equivalence – a concept that presumes the prior existence of money in the first place, and is therefore completely illogical to attribute to people unfamiliar with the use of money to buy and sell things – but a broad sense of owing someone a favor of a roughly equivalent sort. This need not be a material object at all, it could be help, or ritual sponsorship, or maybe your son is in love with your neighbor’s daughter, but let’s leave that aside for a moment. Insofar as it is goods, it would generate a system of vague ballpark equivalents. And this is indeed what one finds where there is extensive “gift exchange”: a rank system, whereby certain types of goods are seen as roughly equivalent to others, but not a proportional (i.e., monetary) system where you can say how many of this type of rank B objects is equivalent to one of these rank A objects. The question is what would be the circumstances that would generate a system where one can measure such precise equivalents. Here, again, ethnography provides endless examples of what actually does happen and it’s nothing like the economists predict.

First of all, it is almost invariably the case that such transactions are not presented as “loans” but as gifts. That’s why I presented it the way I did: one person says “hey, nice cow” the other gives it to him, pretending it’s just a gift with no expectation of any sort of return, though of course everyone knows otherwise. (There are many parts of the world where this is still the case, you can’t praise another person’s possessions or they will automatically insist you take it, and then of course, they can later come and claim something of yours.) Now, this can lead to all sorts of results – maybe you don’t ask for anything back because you like to have someone feel beholden to you, maybe the other person gives you back something much more valuable than your cow to embarrass you in public by showing he’s more generous… This is because in such situations people are not, as the economic models assume, only thinking about the goods and how to maximize their advantage in material terms, but mainly interested in the social relations. But even if someone returns something you feel is inadequate, to head off any further claim you might have, you may well make fun of him as cheapskate, but you are extremely unlikely to come up with a mathematical formula for exactly how cheap you think they are.

So what, in such a context – that is, in the absence of Temple bureaucracies or whatnot – might lead to the creation of a system of pricing, of proportional equivalents between the values of any and all objects? (Remember I noted that money as a system of account probably emerged through different mechanisms at different times and places.) Well, again, anthropology and history both provide one compelling answer, one that again, falls off the radar of just about all economists who have ever written on the subject. That is legal systems. If someone makes an inadequate return you will merely mock him as a cheapskate. If you do so when he is drunk and he responds by poking your eye out, you are much more likely to demand exact compensation. And that is, again, exactly what we find. Anthropology is full of examples of societies without markets or money, but with elaborate systems of penalties for various forms of injuries or slights. And it is when someone has killed your brother, or severed your finger, that one is most likely to stickle, and say, “the law says 27 heifers of the finest quality and if they’re not of the finest quality, this means war!” It’s also the situation where there is most likely to be a need to establish proportional values: if the culprit does not have heifers, but wishes to substitute silver plates, the victim is very likely to insist that the equivalent be exact. (There is a reason the word “pay” comes from a root that means “to pacify.”) Again, this is not hypothetical. This is a description of what actually happens – and not only in the ethnographic record, but the historical one as well. The numismatist Phillip Grierson long ago pointed to the existence of such elaborate systems of equivalents in the Barbarian Law Codes of early Medieval Europe. For example, Welsh and Irish codes contain extremely detailed price schedules where in the Welsh case, the exact value of every object likely to be found in someone’s house were worked out in painstaking detail, from cooking utensils to floorboards – despite the fact that all evidence is that there actually were no markets where such items could be bought and sold. The pricing system existed solely for the payment of damages and compensation, both if someone broke something, and in order to calculate how the damages could be paid. True, Welsh and Irish people of the time probably had heard of the existence of commodity money – the Romans had been nearby in times past – but they had not adopted it. When they did start developing the use of money of account, it was for a completely different set of purposes than economists would have predicted.

I think the participants in this forum should reflect on what they consider the status of economics to be. Is it a science that generates hypotheses about empirical reality that can then be tested against the evidence, and changed or abandoned if they don’t prove to predict what’s empirically there? Or is it a kind of faith, a revealed Truth embodied in the words of great prophets (such as Van Mises) who must, by definition be correct, and whose theories must be defended whatever empirical reality throws at them – even to the extent of generating imaginary unknown periods of history where something like what was originally described “must have” taken place?

Even there, I think the persistence of the barter myth is curious. It originally goes back to Adam Smith. Other elements of Smith’s argument have long since been abandoned by mainstream economists – the labor theory of value being only the most famous example. Why in this one case are there so many economists desperately trying to concoct imaginary times and places where something like this “must” have happened, despite the overwhelming evidence that it didn’t? It seems to me because it goes back precisely to this notion of rationality that Adam Smith too embraced: that human beings are rational, calculating exchangers seeking material advantage, and that therefore it is possible to construct a scientific field that studies such behavior. The problem is that the real world seems to contradict this assumption at every turn. Thus we find that in actual villages, rather than thinking only about getting the best deal in swapping one material good for another with their neighbors, people are much more interested in who they love, who they hate, who they want to bail out of difficulties, who they want to embarrass and humiliate, etc. – not to mention the need to head off feuds. In actual long-distance trade, it turns out they are mainly worrying about danger, adventure (sexual and otherwise), glamour, fame, religious devotion, and all sorts of factors that are not simply getting the most lapis lazuli for the wool. (And that even if they were, it wouldn’t generate circulating money!) Indeed, the only occasions in which people act remotely like the models say they ought to act is when you already have a circulating medium of exchange and already have impersonal markets, and people become accustomed to operating that way – or even recreating something along the same lines, like the POWs, if suddenly the money is taken away. What you seem to be doing is projecting certain types of behavior created by certain social institutions backwards as an explanation of the institution themselves, rather like saying that the game of chess was invented to fulfill people’s preexisting desire to checkmate their opponents’ king – and then justifying it by saying that well, people are competitive, they like to win games, therefore, the desire to checkmate must always have existed.


Contact Robert P. Murphy

Robert P. Murphy is a Senior Fellow with the Mises Institute. He is the author of many books. His latest is Contra Krugman: Smashing the Errors of America's Most Famous KeynesianHis other works include Chaos Theory, Lessons for the Young Economist, and Choice: Cooperation, Enterprise, and Human Action (Independent Institute, 2015) which is a modern distillation of the essentials of Mises's thought for the layperson. Murphy is co-host, with Tom Woods, of the popular podcast Contra Krugman, which is a weekly refutation of Paul Krugman's New York Times column. He is also host of The Bob Murphy Show.

Add Comment

Shield icon wire