The unmourned death of the double coincidence

When Debt first came out, I actually looked this point up in a number of mainstream textbooks, and found not a single example of the claim that money has its origin in solving the double coincidence of wants problem. I’m sure there is a counterexample here (and I know Smith makes the claim, but we don’t go around citing 18th century anthropologists as exemplars of modern anthro thinking either), but in general, I imagine most economists consider the question of “how did money originate in moneyless societies” to be a question best answered by historians and anthropologists, about which we have little to contribute.

As to the related question of what money is for in modern states, its role as a medium of exchange is both true and super important! It even shows up in our models often when we give agents preference for money holdings (i.e., when money directly enters the utility function, not simple Keynesian liquidity preference).

That is, Graeber is almost certainly correct in his description of the origin of money, but I think that tells us little about how we ought treat money in our models of the present day economy, nor do I think his story contradicts what it taught by the vast majority of economists.

afinetheorem

“It even shows up in our models often when we give agents preference for money holdings”

???? Surely this is money as a store of value. Money as a medium of exchange is only relevant when it changes hands.

Nah, the distinction is between the transactions, precautionary, and investment demand for money. You need some money (i.e. cash) for whatever you’re going to buy, some more in case of a sudden call on your liquidity, plus however much you feel is worth having as opposed to turning into some kind of investment. 1) is medium of exchange, 2) and 3) are store of value.

It’s very true that a lot of monetary economics would be easier to explain if we said “cash” rather than “money”, as it’s all about the distinction between wealth stored in some way and cash immediately available for use.

By the way here’s my take on creation myths: http://worthwhile.typepad.com/worthwhile_canadian_initi/2010/02/creation-myths-and-economic-history.html

I never even understood the phrase “double coincidence of wants” since it is a single coincidence of wants: a coincidence.

I have been trying to tell economists ever since I found the internets that the barter origin of money would have to be total baloney, because field anthropologists found money inside tribal systems that weren’t big enough to need barter.

If you actually read the ethnography and actually think about people interacting necessarily in small, acephalous tribal systems, you are very likely to conclude that money arose as an ornamental device (for necklaces and so on) in status competitions that were game-like rituals, sometimes accumulating (by swaps, or not) the whole year until everybody can see who wins with the most trinkets, at the final big annual celebration and dance. In other words, money originated in visual status competition, and was likely an enabler of the long process of political nucleation into chiefs and kings. It would get coopted into standing for debt in regular transactions only AFTER the tribal system got so big that other people could be strangers. Now I have to quit writing and go crawl under a house to run a gas line.

“From now on, we can treat money primarily as a store of value, and stop worrying about how it works as a medium of exchange.”

How on earth can modern money be a “store of value” if you can not use it as a medium of exchange?

We’re not talking about money as corn, honey or cigarettes — the entire value of modern money results from its ubiquity as a means of exchange.

Perhaps the origin of money has everything to do with its “value storage” properties — I am skeptical of this, but open to be convinced otherwise — but one has to be high on hormonal fumes to argue that today we can “stop worrying about how [money] works as a medium of exchange.”

Economists have seen money arising from barter because the notion of a transaction between individuals is central to microeconomics. It’s economics’ atom.

The problem that Graeber points out is that prior to money exchanges weren’t transactions between individuals. They were moves in ongoing relationships. Symbols of the relationship. Some still are: “and having declared the same by the giving and receiving of a ring.”

I don’t think I understand the central point here: at the personal level, money is a medium of exchange, surely? It may not have started that way, but it has become that, not so?

What am I missing?

Oh, I see.

But I don’t think that’s true. It’s very much part of our culture, to be sure, and it wasn’t, say, 500 years ago, when most transactions were local and with familiar people. But it’s also an idea very easily learned; cultures that didn’t have the idea “get” money very easily when it is introduced. I think the idea of money is akin to the idea language; a powerful abstraction that comes easily to human minds.

It surprises me that Graeber apparently thinks that what he calls “embedded” relations of production are entirely positive: these are exactly the sorts of relations that lead to huge conflicts within and between families and tribes. In some ways limiting the transactions to money makes this easier: at least one has something definite to fight about!

It surprises me that Graeber apparently thinks that what he calls “embedded” relations of production are entirely positive …

He doesn’t. When he introduces the term “human economies” he says “By this I mean not that these societies are necessarily in any way more humane (some are quite humane’ others extraordinarily brutal), but only that they are economic systems primarily concerned not with the accumulation of wealth, but with the creation, destruction, and rearranging of human beings.” (p. 130) (The word “embedded” doesn’t occur in Debt, but I assume this is what you mean.)

Lee #15. Aha! Your incident/coincident distinction was new to me.

Graeber’s glee at his claim to be demolishing the economists’ story about where money came from seemed to me mostly to reveal a lack of understanding of what they are trying to do.

It is interesting, as a matter of anthropology and of history, to learn how societies which did not use money for food & shelter organised things. Here I learned many things from the book. But it’s not clear that this of any use in understanding the role money plays in our society. I always thought the point of that little fable was to show that, with money, there can be quite a long chain of people between the ones we help with our labour and those who in turn grow our food etc. Once you have got the idea that this complex organisation would be very, very difficult to keep track of without money, you can move on to learning more detail about how it actually happens today.

Perhaps the fable should come with a footnote explaining that it isn’t historical truth, so as not to confuse the frosh… but the fact that they don’t bother is telling you something about what the field cares about. The mathematicians also aren’t too concerned about whether Pythagoras was cribbing ideas from Mesopotamia: that is a different discipline.

Money allows people to save without purchasing investment goods; but diamonds, which aren’t a medium of exchange in most places, also allow them to do that. Could Keynesian problems appear in an economy in which diamonds are a store of value and money isn’t? I’m inclined to say yes, but I think Nick Rowe would say no.

John: “To try to find some common ground, I’ll agree that if by “barter economy”, you mean “smoothly functioning Arrow-Debreu economy with complete set of time-dated state-contingent markets, but no numeraire”, then Keynesian economics wouldn’t work in such an economy.”

I agree that Keynesian economics wouldn’t work in such an economy. (Again, just in case someone misunderstands, that is not a criticism of Keynesian economics.)

“As an example, consider Krugman’s baby-sitter co-op. There’s no problem here with a double co-incidence of wants, since’s there’s only a single commodity. It’s all about savings.”

I disagree with the second sentence. The commodity is a dated good. Suppose A wants a sitter Sunday, and wants to sit Monday; B wants a sitter Monday, and wants to sit Tuesday; C wants a sitter Tuesday, and wants to sit Wednesday; etc, all the way up to G, who wants a sitter Saturday, and wants to sit Sunday. Without a medium of exchange, you would need to get all 7 people together to do a 7-way barter deal, which is inconvenient. That’s why these babysitting co-ops use monetary exchange.

Now, Krugman’s co-op model is a model in which the only asset is also the medium of exchange, so we cannot distinguish the medium of exchange from the only savings vehicle. We cannot distinguish between an excess demand for saving and an excess demand for the medium of exchange.

Let’s introduce land as an asset. (Not diamonds, Kevin, because diamonds can be newly-mined ;) ). Suppose land is the only asset in a barter economy where people barter their labour. Could an excess demand for land cause Keynesian unemployment?

I would answer “no”. Here’s why:

Either the price of land is flexible, or it is fixed.

If the price of land is flexible, the price of land will rise until the excess demand for land is eliminated, and people go back to demanding consumption goods instead.

If the price of land is fixed, there’s initially an excess supply of labour matched with an excess demand for land. But when people find they cannot sell their labour for land (because everyone else is trying to do the same thing), they swap their labour for someone else’s labour instead. The fact that people cannot sell their labour for land doesn’t stop them making mutually advantageous exchanges of their labour. “Will you give me some land if I sit your kids?” “No”. “OK, will you sit my kids if I sit your kids?” “OK, it’s second-best compared to land, but it’s better than being unemployed and not having a sitter”.

So, whether land prices are fixed or flexible relative to labour, either way there is no Keynesian unemployment.

There was an argument between Keynes and Silvo Gessell on this point. I side with Gessell against Keynes.

“Sook” is “seek”. I am willing to trade my typos for one banana.

I don’t think I want to read Graeber to critique his ideas, truly, I do not.

I think you’re misunderstanding what’s being talked about. The “human economy” goes with the “primitive money” I mentioned above. So:

Wealth in our society is still after all very much about arrangements of human beings

It’s not true that Jamie Dimon is concerned about whether or not I marry your sister.

First, Graeber is on my to do list. Mainly because his comments about the origins of money caught my eye – I know enough about the ancient Middle East to know recognise the story he tells. My understanding is that the history tells us that money (a negotiable record of an obligation, denominated in some standard unit) arose from a combination of long-distance trade and centralised flows of goods. Gold and silver coinage came some millenia later, in a milieu where states where trying to employ mobile resources – ie the soldiers wanted something portable, liquid, and immediately recognisable. But most trade and most state payments did not involve gold or some other hard currency then or now.

The value in tracing the origin of money, and seeing it arising as a marker for debt, is that it draws attention to an overlooked property, either as a medium of exchange or as a store of value. We ask of a debt how good it is – whether we can expect to receive full value, and when. Debts have quantity and velocity, but they also have hardness. Money does too. At one end, we have gold coins (universal, mostly liquid), at the other the notes for Krugman’s baby-sitting coop. And an endless array in between. A large part of the game is trading up from soft to hard (eg, all those who bought mortgage-backed paper traded hard for soft – those who sold did the opposite and banked their hard cash bonuses). Bankers want money to be as hard as possible. Entrepreneurs want money to be so soft as to be squidgy (they want to swap their claim that their idea will generate untold wealth for the actual goods and services needed to realise it). This tells us that there is no neutral, immutable store of value or medium of exchange, there’s a fluctuating triangular relationship.

Clauswitz likens the relationship of battle to military power to cash settlement in commerce – it’s not the usual or most frequent event, it’s the one that tells you whether the perception matches the reality. This is the “how hard is this money?” moment.

Kevin: “And the fact that people cannot sell their labour for money doesn’t stop them either; an unemployed electrician can rewire the house of an unemployed plumber who has no money, on the understanding that the favour will be returned when the electrician’s drains get blocked.”

True. And if barter was always relatively easy like this, we wouldn’t be too worried about Keynesian/deficient demand unemployment. But even though barter does increase in recessions, AFAICT, which confirms the theory, barter is simply too hard in many cases to solve the unemployment problem. In a massively interdependent advanced economy, it would need hundreds or thousands of people coming together to do a great big barter deal. Your plumber and electrician have a double coincidence of wants, and use little apart from their own labour and tools, so coordination is relatively easy, if they both know each other. Think of how many people are involved in building a car.

@35

“Alex, saying that something is a reductio ad absurdum doesn’t make it so. In case it’s not clear, I do think (as did Keynes) that money is important in the real world.”

I didn’t just say it without argument — I argued that hoarding of diamonds (or corn, or just about any storable commodity) would in fact act as a sort of Keynesian stimulus, being equivalent to consumption which then leads to invesment etc., which is the Keynesian solution to aggregate demand slumps.

Why does this not happen in an aggregate demand slump? According to Keynes, because of liquidity preference! And what is this “liquidity” that Keynes talks about?
Why, it is precisely the exchange power of money –and money-like assets — that the original author wants us to forget about.

” The question is whether it’s possible to leave it out of a model intended to convey Keynesian ideas. I think it is.”

Hamlet without the Prince.

“It’s not true that Jamie Dimon is concerned about whether or not I marry your sister.”

But he is very concerned about whether the people who work in the businesses which own him money or which he owns show up for work.

Control over human lives is still wealth.

Consumatopia: “If people primarily want the land because they can trade it for labor, then it doesn’t solve the problem if the price of land (in labor) rises.”

Well, the yield on land will fall, if the price of land rises, and that should reduce the demand for land. But OK, suppose it doesn’t, so there’s still an excess demand for land. Then I switch to the second part of my argument. If everyone wants to buy land, then nobody wants to sell, so they find they can’t buy land. So, as second best, they go back to bartering their labour and get back to full employment that way.

The Raven: “Might this be achievable through computer mediation?”

In principle yes. In practice I have my doubts.

There are a lot of computer-assisted “barter” systems around the world, like LETS for example. (Actually, these are really monetary exchange systems, rather than true barter, even though there is no physical medium of exchange except computer bits and bytes). And these computer-assisted “barter” systems do seem (AFAICT) to become more used whenever there’s a recession (which seems to confirm this Keynesian/Monetarist/coordination failure theory of recessions). And they almost certainly help some people to some extent. But they can only work if lots of people use them, otherwise you can’t find a seller for what you want to buy or a buyer for what you want to sell. There’s another “network externality” here too, just like with any monetary system.

“You’re still misunderstanding Graeber’s terms. The ‘human economy’ is not about owning people.”

I got that. But I do not see the bright line between Graeber’s human economies and his commercial economies, at least in CB’s summary. The reality of an economy where “the principal function of exchange is to maintain that system of relations and to effect ‘moves’ within it” can be oppressive. Conversely, sometimes a commercial economy can create distance between differing people that makes their lives easier; one can work for a boss even if one disagrees with their politics.

But I do not see the bright line between Graeber’s human economies and his commercial economies

The debts aren’t necessarily debts as we think of them; you can give a bunch of shells to someone representing this or that societal obligation or grievance settlement, but they’re not a trading token. Graeber makes it clear that “human economies” could be – can be I guess – no less brutal than any other system.

The Raven: “Yes, and if they take off, they will be subject to multiple sorts of fraud. I wrote about this a year or so back:

Interesting. I hadn’t thought of fraud. But now you mention it, I was naive not to have thought of that practical difficulty. Especially since there’s a whole line of enquiry that sees monetary exchange as a substitute for trust.

Consumatopia: “It shows that Keynesianism assumes that barter, as you’ve defined it, is difficult.

Agreed.

“Keynesian unemployment might still be possible if we assume a weaker yet still incredibly impressive barter economy—in which we can always sort out coincident wants for things we have already produced, or perhaps services we provide immediately, but exchanging promises to perform services in the future is still difficult.”

Maybe. But the unemployed would still want to barter for whatever was available today, and then hope to re-barter those things in the future. They would barter their labour for newly-produced investment goods.

re Gavin

This, as usual, conflates “money” and “coins”. Money was around for at least a millenium before coins, And has stayed around for the bulk of transactions after coins. Money is a transferable obligation, denoted in some standard unit. One early example are the clay tablets used by Assyrian traders – each records the value, and then a series of holders. The last holder settles the account and the tablet is broken. Likewise the English treasury used tally sticks. Soldiers were paid with tablets or tallysticks for centuries, as were other creditors. Coins came in because mercenaries in the Aegean wanted something they could take home or spend somewhere else without the bother of establishing local relationships. But the bulk of transactions still did not – and do not – use coins (or notes). This tells us something important about money. The standard unit (mina, dollar etc) is and always has been quite distinct from physical money, in the same way that a measure of length is distinct from any particular thing measured. And the amounts in circulation are set by what people believe they can get, not by any issuing authority.

At various points in the book, I felt that the distinction between ‘human economies’ and ‘commercial economies’ was strikingly reminiscent of C19 attempts at characterising the differences between modernity and pre-modern; Toennies’ Gemeinschaft v. Gesellschaft distinction, for example. What’s distinctive about Graeber’s approach is that he pushes the transition back to a much earlier date. The contrast doesn’t necessarily set up an idealised community with embedded social relations against present-day alienation, but that’s often the way it’s taken.

I know them both, and always liked the idea of the ‘original affluent society’ – though actually it was probably a bad idea to leave the oceans in the first place…

Surprised Nick Rowe hasn’t yet linked to his WWCI post responding to this one:
worthwhile.typepad.com/worthwhile_canadian_initi/2012/02/money-as-store-of-wealth.html

I found Graber’s point regarding debt preceding money to be similar to Kocherlakota’s “money is memory”. Barter is hard to use, but if we have memory we can just note that someone is obligated due to such and such and settle things later. At large scales human memory isn’t sufficient, and money can substitute.

The one thing that money is not a medium of is exchange. There is no underlying exchange that is channelled via money. The accounting function of money* makes discrete bilateral transactions largely irrelevant.

That’s awkward for the idea of property relations as based on discrete bilateral transactions generating entitlements that the state then meddles in, and for the predominant (right wing) economics that understands basic concepts like efficiency or preference in terms of the operation of those discrete bilateral transactions.

The barter and double coincidence of wants stuff treats use of money as an improvement over (and on) barter. Rather than saying ‘barter, our paradigmatic example of discrete bilateral transactions, was unimportant and like many other clumsy ways of trying to redistribute goods, is not something you’d want to bother trying when unilateral transactions can be tracked by keeping account of a numeraire’ they say ‘isn’t it great that we have coin for people to take to market, so we don’t have to rely on barter’, conveniently holding fixed the assumption of DBTs and ignoring the role of state-like institutions in issuing, policing and guaranteeing, currency as well as pruning it by means of taxation,

*The accounting function is the fundamental one. There’s the decentralised kind of accounting done with tokens like coin, of course, which does make those tokens a ‘store of value’ so long as they are embedded in context. (And this has nothing to do with being a good investment – it’s a matter of short-term storage, of continuity). On one of these threads someone mentioned clay tablets from some early mesopotamian civilisation. Those seem to lie on the continuum between (1) the totally decentralised ‘Hayekian’ accounting done by means of transferring anonymous rationed tokens (coin) and (2) a ledger of inputs to and outputs from a central fund. Note that asking for money payment is not significantly different in most ways from asking for a signature on a -cheque- chit to be sent into the -clearing bank- Central Bureau of Accounts.

(And since this footnote seems to have grown back into the comment body:) So why should we wish to make our social accounting system resemble a lawless frontier territory in which only bilateral transactions involving self-documenting accounts or commodity-money like gold are possible? Discarding possessive individualist state-of-nature theory and nakedly ideological aims, we are left with some pretty pissweak tea:

(1) Hayekian magic price ‘signals’, which are just a fetish, the term applied to a phenomenon only very contingently implicated in the functions performed by actual people doing market research, marketing, technical improvements in production, assessments and forecasts of scarcity, etc etc (much of which is made much less tractable thanks to the convoluted feedback loops of the Keynesian beauty contest), and on the consumption side, rationing/budgeting. And note too that this weary trope, hangover from the Cold War era in which ‘bureaucracy’ meant real bureaux stuffed with reams of foolscap, will have to meet a standard of efficiency a little more salient and far more stringent than the Pareto kind, which is artifically tailored to require all and only discrete bilateral exchanges and is thus trivial in econ 101 world (cf efficient markets).

(2) Arrow/Nozick/Steiner(?) ‘compossibility’ requirements (which are misnamed since the modality involved is much more complex) which (a) so far as Arrow-type axiomatic approaches are concerned, in a world not constructed from economists’ equations would have a hard time generating any policy recommendation (and ‘but if this logically possible circularity happens, something will have to give’ is a bloody odd kind of criticism coming from the Keynsian beauty contest merchants), and (b) – probably a trifle more devastatingly in this context – similar concerns about valid claims being unfulfillable are rampantly and routinely violated by the very existence of monetary debt. Also (c) Sen on the Paretian liberal.

I think there are two possible claims being made here regarding the role of money as a medium of exchange. One possible claim which people are reacting to is that money never was and is not now useful as a device to overcome the double coincidence of wants in actual economies. This seems self-evidently preposterous, and John Quiggin at the very least doesn’t distance himself from this view. But I read the post as primarily arguing that, while money may in fact serve as a medium of exchange, it is not essential for economists to take this into account when using monetary models to analyze economic fluctuations and policy responses. This strikes me as a very defensible (although controversial) point of view.

Regarding the theoretical function of money as a medium of exchange, I believe the current professional consensus is that the existence of the double coincidence of wants problem is not sufficient to generate a need for a medium of exchange. One also needs an economic environment with limited commitment and imperfect memory. The intuition for limited commitment is that, even if I don’t have anything that you want when we meet to trade at a specific time, if I can commit to giving you something in return in the future then money is inessential. If I can’t commit to doing so BUT the entire history of my transactions (and every other agent’s in the economy) is public knowledge (unlimited memory), then we can still support an equilibrium where the exchange is made because everyone in the society can condition their responses to future encounters on how I behaved in previous encounters, giving me the incentive to play fair.

Only when there is a double coincidence of wants problem, limited commitment AND limited (collective) memory of past transactions is there an unambiguous role for money as a medium of exchange. This point is made very nicely in this article by Narayana Kocherlakota (now president of the Minneapolis Federal Reserve) although others pointed out pieces of the puzzle much earlier.

http://minneapolisfed.org/research/sr/sr218.pdf

This insight is as mainstream as it gets, although perhaps it hasn’t made it into the textbooks. Smaller scale societies and/or groups of people such as merchant guilds have much fewer problems of limited memory, and thus can rely on debt contracts that are self-enforcing in the sense that default is costly due to reputation effects. Money was perhaps not necessary as a medium of exchange in those societies, but it certainly is a necessity for transacting anonymously in an environment with a large number of agent’s whose history is opaque. It would have been nice if Graeber had gotten out of the elementary textbooks and into some of the modern economic literature before making his claims contra economic theory.