1. Skip to navigation
  2. Skip to content
  3. Skip to sidebar

The Ludwig von Mises Institute

Advancing Austrian Economics, Liberty, and Peace

Advancing the scholarship of liberty in the tradition of the Austrian School

Search Mises.org

Have Anthropologists Overturned Menger?

Mises Daily: Thursday, September 01, 2011 by


Last week the popular blog "naked capitalism" ran an interview with David Graeber, an "economic anthropologist" whose new book allegedly destroys the standard account of the origin of money. If correct, Graeber's views would prove embarrassing to the Austrian School, because it was none other than Carl Menger who developed the first systematic explanation for how people went from barter to a full-blown monetary economy.

As we'll see, Graeber's critique of the standard (Mengerian) view is much weaker than he believes, while his own explanation makes no sense. Furthermore, we have actual case studies of the development of a new money, which fit the Mengerian story. All in all, I see nothing in Graeber's interview to make me question Menger's orthodox approach.

Menger on the Origin of Money

In an earlier article I systematically laid out the Mengerian position, as well as the extensions that Ludwig von Mises provided, in the area of monetary theory. But to make sense of Graeber's challenge, here we should briefly review the basics.

According to Menger, money emerged spontaneously through the self-interested actions of individuals. No single person sat back and conceived of a universal medium of exchange, and no government compulsion was necessary to effect the transition from a condition of barter to a money economy.

Menger pointed out that even in a state of barter, goods would have different degrees of salableness or salability. (Closely related terms would be marketability or liquidity.) The more salable a good, the more easily its owner could exchange it for other goods at an "economic price." For example, someone selling wheat was in a much stronger position than someone selling astronomical instruments. The former commodity was more salable than the latter.

Notice that Menger is not claiming that the owner of a telescope would be unable to sell it. If the seller set his asking price (in terms of other goods) low enough, someone would buy it. The point is that the seller of a telescope (in a state of barter) would only be able to receive its true "economic price" if he devoted a long time to searching for buyers. The seller of wheat, in contrast, would not have to look very hard to find the best deal that he was likely to get for his wares.

Menger argued that this fairly obvious fact of different degrees of salability would set events in motion that would eventually yield the first money. In the beginning, owners of relatively less salable goods exchanged their products not only for those goods that they directly wished to consume but also for goods that they didn't directly value, so long as the goods they received were more salable than the goods given up. In short, astute traders began to engage in indirect exchange. For example, the owner of a telescope who desired fish didn't need to find a fisherman who wanted to look at the stars. Instead, the owner of the telescope could sell it to any person who wanted to stargaze, so long as the goods offered for it would be more likely to tempt fishermen than the telescope.

Over time, Menger argued, the most salable goods were desired by more and more traders because of this advantage. But as more people accepted these goods in exchange, the more salable they became. Eventually, certain goods outstripped all others in this respect, and became universally accepted in exchange by the sellers of all other goods. At this point, money had emerged on the market.

Graeber Doesn't Buy It

We'll quote extensively from the interview to understand Graeber's problem with the conventional (Mengerian) approach and what he offers in its place:

Philip Pilkington: Let's begin. Most economists claim that money was invented to replace the barter system. But you've found something quite different, am I correct?

David Graeber: Yes there's a standard story we're all taught, a "once upon a time" — it's a fairy tale.

It really deserves no other introduction: according to this theory all transactions were by barter. "Tell you what, I'll give you twenty chickens for that cow." Or three arrow-heads for that beaver pelt or what-have-you. This created inconveniences, because maybe your neighbor doesn't need chickens right now, so you have to invent money.

The story goes back at least to Adam Smith and in its own way it's the founding myth of economics. Now, I'm an anthropologist and we anthropologists have long known this is a myth simply because if there were places where everyday transactions took the form of: "I'll give you twenty chickens for that cow," we'd have found one or two by now. After all people have been looking since 1776, when the Wealth of Nations first came out. But if you think about it for just a second, it's hardly surprising that we haven't found anything.

Think about what they're saying here — basically: that a bunch of Neolithic farmers in a village somewhere, or Native Americans or whatever, will be engaging in transactions only through the spot trade. So, if your neighbor doesn't have what you want right now, no big deal. Obviously what would really happen, and this is what anthropologists observe when neighbors do engage in something like exchange with each other, if you want your neighbor's cow, you'd say, "wow, nice cow" and he'd say "you like it? Take it!" — and now you owe him one. Quite often people don't even engage in exchange at all — if they were real Iroquois or other Native Americans, for example, all such things would probably be allocated by women's councils.

So the real question is not how does barter generate some sort of medium of exchange that then becomes money, but rather, how does that broad sense of "I owe you one" turn into a precise system of measurement — that is: money as a unit of account?

By the time the curtain goes up on the historical record in ancient Mesopotamia, around 3200 BC, it's already happened. There's an elaborate system of money of account and complex credit systems. (Money as medium of exchange or as a standardized circulating units of gold, silver, bronze or whatever, only comes much later.)

So really, rather than the standard story — first there's barter, then money, then finally credit comes out of that — if anything its precisely the other way around. Credit and debt comes first, then coinage emerges thousands of years later and then, when you do find "I'll give you twenty chickens for that cow" type of barter systems, it's usually when there used to be cash markets, but for some reason — as in Russia, for example, in 1998 — the currency collapses or disappears.

This is a fascinating perspective; it had never even occurred to me that someone might claim that money existed before barter. Even so, Graeber's position seems untenable, as I'll try to point out below.

No Record of Barter Transactions?

Graeber thinks he has dealt the Mengerian account a death blow by saying, "if there were places where everyday transactions took the form of: 'I'll give you twenty chickens for that cow,' we'd have found one or two by now."

Yet this doesn't follow at all. At least in the Misesian exposition, the original state of pure, direct exchange — where people just exchanged in order to obtain goods that they directly valued — would last very briefly. The advantages of indirect exchange, where people acquired some goods intending to trade them away again in the future, were so obvious and great that the practice would have begun almost immediately.

Now the transition from an economy characterized by frequent indirect exchange, to an economy using money, would involve the snowball effect we discussed in the beginning of this article. Specifically, as people sought to acquire goods that were more marketable than the ones they started out with, the most marketable of goods would see their superiority amplified. I don't recall Menger or Mises ever giving a guess as to how long this transition would take, but it's not fatal to their theory if anthropologists only have evidence of markets based on money (as opposed to markets based on direct exchange, or what the layperson means by "barter").

First of all, let's be more specific about what Graeber thinks he would need to find. One of the standard disadvantages of barter is that it requires far more prices than a system using a single medium of exchange (i.e., a money). For example, if there are just 20 goods in the economy, then a board showing all the relevant price ratios in a system of direct exchange would need (20 x 19) / 2 = 190 unique prices. In contrast, if an economy with 20 goods were using money, then a board at the marketplace would only need to show 20 unique prices. So is Graeber really that surprised to not find evidence of traders dealing with 190 prices for a small economy, as opposed to discovering the advantages of money and then posting money prices?

This leads to a related point: Mises believed that economic calculation (for which money prices are necessary) was a pillar of economic rationality and civilization itself. So again, it's not surprising that Graeber and his colleagues haven't found evidence of civilizations with bustling markets and written records that were still relying on barter pricing.

Finally, we have actual case studies of communities developing money prices from scratch: namely, prisoners who end up using cigarettes as the common medium of exchange.  The classic work here is Radford's 1945 article, "The Economic Organization of a P.O.W. Camp." There is nothing in Radford's account that conflicts with the standard economists' story about the origin of money. The prisoners certainly weren't giving each other things from their Red Cross kits as gifts or as loans. No, they first were trading (in a state of direct exchange) and cigarettes quickly became the money in their community for all of the reasons that economists typically cite.

And to reinforce the point we made earlier, Radford explains that the prices (quoted in cigarettes) of various items were posted on a board. If Graeber and his colleagues stumbled upon the ruins of this P.O.W. camp, they would presumably conclude that there was never a preexisting state of barter, because they only found boards listing prices quoted in terms of cigarettes. There were no boards listing the thousands of pairwise permutations of direct-exchange ratios, and so clearly the Mengerian story must have been wrong — so would go the erroneous reasoning of Graeber.

To see the merits of the Mengerian account — and to understand just how fast economies can evolve from barter to the use of money — the reader should look at Jeff Tucker's whimsical account of children exchanging Halloween candy in his dining room. (Hint: Neither Tucker nor his wife swooped in to provide a unit of account for the children.)

Graeber's Own Story Is Nonsensical

Above I have argued that Graeber's critique of the Mengerian account poses no threat. Now I want to go further and show that Graeber's rival explanation quite simply makes no sense on the face of it.

Remember that Graeber says before people traded goods directly, there first developed a system of money as a unit of account. This was how people allegedly kept track of their complex debt relationships, based on the loans made of various goods (and presumably services).

But hold on a second. Without having a network of antecedent barter pricing, how would these primitive peoples know how many units of the money to assign to each type of good and service?

For example, let's stipulate for the sake of argument that before people ever thought of exchanging goods with each other, they first would grant loans of those goods. So, Neighbor A has a bad harvest and borrows 10 bushels of corn from Neighbor B, on the understanding that he owes this to Neighbor B and must make up for it down the road.

At the same time, Neighbor B's camel dies so he asks Neighbor C if he can borrow one of his spare ones. Neighbor C agrees to this. Finally, Neighbor C needs a pig, which he borrows from Neighbor A.

Thus Neighbor A owes 10 bushels of corn to Neighbor B, but he in turn owes a camel to Neighbor C, who in turn owes a pig to Neighbor A. If I understand Graeber's position, he is arguing that these types of debts were systematized and reduced to a common denominator by the authorities somehow. For example, perhaps the authorities would simplify the above relationships by saying that on net, Neighbor B just owes a certain number of Temple Units to Neighbor A, and a larger number to Neighbor C.

Yet the only way to calculate the exact numbers would be to have the "money" prices (quoted in Temple Units) for bushels of corn, camels, and pigs. Is Graeber suggesting that the authorities came up with the relative prices for all the goods in their economies, without having a single instance of people trading the goods against each other to see what their market values really were?

It's true, I haven't logically ruled out the possibility of Graeber's interpretation. Rather, I think he doesn't know enough about the economic analysis of exchange in order to correctly characterize his anthropological evidence. For example, later in the interview he says,

[In Mesopotamia at] first there were giant bureaucratic temples, then also palace complexes, but they weren't exactly governments and they didn't extract direct taxes — these were considered appropriate only for conquered populations. Rather they were huge industrial complexes with their own land, flocks and factories. This is where money begins as a unit of account; it's used for allocating resources within these complexes.

Interest-bearing loans, in turn, probably originated in deals between the administrators and merchants who carried, say, the woollen goods produced in temple factories (which in the very earliest period were at least partly charitable enterprises, homes for orphans, refugees or disabled people for instance) and traded them to faraway lands for metal, timber, or lapis lazuli. The first markets form on the fringes of these complexes and appear to operate largely on credit, using the temples' units of account.

It's one thing to suggest that civilization started as a centrally planned economy, where temple authorities came up with the prices of all goods and services (quoted in terms of a money that they invented from scratch) before anyone had ever engaged in barter.

Yet it is incomparably more implausible to suggest that these merchants — using the money prices invented by central planners, and yet who had never witnessed a single act of barter — then went to "faraway lands" and managed to trade woolen goods for metal, timber, and lapis lazuli. Does Graeber really expect us to believe that these merchants engaged in long-distance loans? Or does he concede that at least here there were spot trades occurring, and at prices dictated by supply and demand, not by the temple authorities back home?


It is true that Carl Menger's account of the origin of money is largely a form of armchair reasoning. In that respect, anthropologists with economic training have much to offer in filling in the story with actual historical details.

Unfortunately, David Graeber's rejection of the standard account leads me to believe that he doesn't really understand the claims that people like Menger and Mises made. Furthermore, his own explanation flies in the face of not only basic economic logic but also well-documented examples of the emergence of a new money.