David R. Henderson  

Hummel on Graeber

PRINT
Capital One vs. CFPB, again... Thiel, Schmidt, and Consumptio...

Jeff Hummel sent me a critique of David Graeber's book, Debt: The First 5000 Years. Jeff has given me permission to run an edited version. Here it is:

I have read David Graeber's Debt: The First 5000 Years thoroughly and despite Graeber's readability, scholarship, and erudition, it is a very bad book. Its tone is much too polemical. More important, when it gets to the more recent history that I know well, it is riddled with errors and distortions. Beyond that, it suffers from serious conceptual confusions, and in his excellent critique, Robert Murphy has only scratched the surface.

Nonetheless, Graeber's work is emerging as the most challenging defense of the chartilist (i.e., State) theory of money and has evoked mixed reactions even from Arnold Kling and Tyler Cowen. Let me therefore mention that Leonidas Zelmanovitz, a Liberty Fund fellow, has a fine dissertation arguing that chartilist evidence does not undermine Menger's analysis.

Partly inspired by Murphy's comments, here are some of my own thoughts on Graeber's book. To start with, he makes several different historical claims, not all of them compatible:

(1) Credit transactions preceded and dominated spot transactions in early human societies.

(2) Media of account emerged before media of exchange.

(3) Barter was unknown (or at least extremely rare) WITHIN early human societies.

Notice that point (1) is incompatible with either (2) or (3). Early credit transactions must have involved barter (contradicting number 3) or media of exchange (contradicting 2). There is no other logical possibility. Yet because Graeber's peculiar concept of barter excludes a farmer trading a pig for delivery of an ax in two weeks (to use Murphy's example), his claim that barter was non-existent tends to become true by definition. Murphy was not the only one to catch this semantic sleight of hand; it is even exposed by an Amazon commenter on the book.

Graeber's terminological tautology appears to stem from his confusing (a) the limited ability of credit to mitigate the problem of the double coincidence of wants with (b) the substantial ability of multilateral exchange to do so. Multilateral networks are in reality what he is partly describing when he invokes "systems of broad, non-enumerated credits" (more on the "non-enumerated" part below). Consider the standard three-person THEORETICAL trading problem, where A wants only what B is selling, B wants only what C is selling, and C wants only what A is selling. The lack of a double of coincidence of wants can be solved by using one of the goods as a medium of exchange in two bilateral trades OR by conducting a single multilateral trade. Obviously for small groups, where people know and trust each other, the latter is often more likely and convenient. Whether such a multilateral exchange takes place at one moment in time (a spot transaction) or is extended through time (a debt transaction) is of secondary relevance, although the possibility of debt transactions certainly increases the potential scope of multilateral exchange.

Graeber's book quotes examples taken from several econ principles texts of hypothetical barter transactions that he claims have no basis in history. But the authors design these examples merely to illustrate why monetary exchange is more efficient for complex economies. They imply no accuracy with respect to historical details, with the possible exception of the quotations taken from Adam Smith's Wealth of Nations. It would make as much sense for Graeber to ridicule grade-school arithmetic texts because of their fanciful use of apples to illustrate addition. Graeber goes on to set up an additional straw man with his assertion that economists universally believe that the emergence of money necessarily preceded the development of credit. This is utter nonsense, nowhere supported in the economics literature. Every competent economist fully understands that you can have debt transactions without money. They simply recognize that money facilitates credit as it facilitates other exchanges and, therefore, find monetary credit easier and more relevant to explicate.

Indeed, an extended defense of the historical claim that money spontaneously evolved from barter is almost unique to the Austrian School and Carl Menger (whose first name Graeber misspells as "Karl" and whom Graeber mistakenly accuses of "adding various mathematical equations" and coming up with the term "transaction costs," possibly confusing the economist Menger with his son, Karl, who was a mathematician). Menger's story is in no way undermined by the discovery that what preceded monetary exchange was barter debt transactions rather than barter spot transactions. Moreover, Graeber's classification of Sumerian temples as non-States actually reinforces the Menger story. And his blanket assertion that "non-state bureaucracies . . . are off the map of economic theory" is simply absurd, as if he thinks that economists are unaware that the private firms they analyze often have bureaucracies.

Nor is the Menger scenario compromised very much by the claim that media of accounts emerged before media of exchange, which has been suggested by many anthropologists and accepted even by Tyler Cowen and Randy Kroszner in their Explorations in the New Monetary Economics. The only serious challenge Graeber raises is his assertion that most media of account arose from legal penalties or were otherwise arbitrarily imposed. However, as Murphy points out, it seems implausible that something could emerge as a medium of account without already being widely enough marketable to provide at least some guidance about its value relative to other goods. The only reason this poses no puzzle to Graeber is that he apparently considers most prices to be significantly arbitrary and is entirely oblivious to their vital role in workable resource allocation.

This leads to Graeber's naive infatuation with the reciprocal altruism of hunter-gatherers and of assorted tribal societies, an arrangement he labels "communism." The term "reciprocal altruism" comes from evolutionary biology, which has an extensive literature that Graeber seems unfamiliar with. He also embraces the bizarre and unsupported idea that economics assumes away altruism and cannot analyze it. Then couple this with Graeber's erroneous insistence that trade requires an equality of value. In short, he truly does not comprehend the subjectivist analysis of Menger or the obvious fact that voluntary exchange can NEVER take place with an equality of value and instead rests on a reverse inequality of value.

You can now appreciate how he derives his hard and fast distinction between "systems of broad, non-enumerated credit" and those where "people borrow things from one another and expect an equivalent of exactly the same value." He views the precise economic calculation that money permits as a sharp and pernicious transformation rather than an important evolution along a spectrum of human interactions that resulted in greater prosperity. This makes him stubbornly impervious to Friedrich Hayek's astute observation: social and economic systems that work well in small groups where nearly everyone knows each other would be devastating for a large, complex society with an extensive division of labor.

I could go on and on. Graeber misrepresents even scholarly work tending to support his conclusions, such as the article on barter by Caroline Humphrey [gated], which is far more subtle and circumspect in its conclusions and even discusses the category of "delayed barter," i.e., barter debt transactions. He attributes to Ludwig Mises mainstream economic ideas that Mises repeatedly, explicitly, and emphatically rejected, indeed some ideas that even many mainstream economists reject. Overall, he continues to caricature the entire economic discipline without any sincere effort to address its analysis. He fails to appreciate the crucial difference between government-issued fiat money (a form of what monetary economists call outside money, with a net wealth effect) and bank-issued redeemable money such as deposits and banknotes (called inside money). And what can we make of a scholar who actually takes seriously the suggestion that the U.S. government initiated the first Gulf War because of Saddam Hussein's decision to sell oil for euros rather than for dollars?

But Graeber does get some important things right. He, like me, advocates the repudiation of government debts, although he unfortunately advocates the same for private debt. And Debt: The First 5000 Years opens with a compelling example from Madagascar of a process Leonard Liggio used to describe as the dominant motif of European colonialism in Asia and Africa. The imperialist powers imposed head or hut taxes payable only in European sanctioned money, that natives could earned by working in a European-owned mine or on a European-owned plantation, thus indirectly extracting coerced labor. This does show that the State, while definitely not necessary for the origin and evolution of money, has had enormous influence on what people use as money. We therefore will never attain an entirely free-market monetary system until all taxation is abolished.


Comments and Sharing


CATEGORIES: Money



COMMENTS (17 to date)
Steve Z writes:

Impressive review, in that sometimes it can be harder to cut through a tangle of fundamental errors than to spot a subtle but sensible one. One quibble: you wrote "slight" when you meant "sleight."

John Hall writes:

Wouldn't it have been like marks or francs back in the first Gulf War?

David R. Henderson writes:

@Steve Z,
Thanks, and I’m sure Jeff thanks you also. Quibble fixed.

Ed writes:
And what can we make of a scholar who actually takes seriously the suggestion that the U.S. government initiated the first Gulf War because of Saddam Hussein's decision to sell oil for euros rather than for dollars?

Did Graeber really make that claim?!

Jeff Hummel writes:

Ed and John,

No, Graeber made the statement with respect to the U.S. invasion of Iraq in 2003. It is on p. 367 of his book. Thanks for the correction, and my apologies for the error.

Ken B writes:

Aside from anything else I am at a loss to understand how we can know that there was no barter on the veldt 200,000 years ago. There was barter even in my dormitory. Can someone explain this?

dofthenile writes:

Can someone explain a little more, or suggest a text that will do so, "Friedrich Hayek's astute observation: social and economic systems that work well in small groups where nearly everyone knows each other would be devastating for a large, complex society with an extensive division of labor"?

Ken B writes:

@dofthenile: I can give an illustration.

When I was a child my parents rented a cottage, in a rather small isolated community. Everyone knew everyone else. All weere middle class families, many from the same places in Ontario. All the kids were allowed out all the time, even late at night, and in mixed groups. There was a high level of trust and knowledge. The kids who were from Toronto though had no such freedom when they returned to the big city. I can never recall a locked cottage door either.

That's a social example.

Yancey Ward writes:

The most important observation from Graeber's book is this:

The only serious challenge Graeber raises is his assertion that most media of account arose from legal penalties or were otherwise arbitrarily imposed. However, as Murphy points out, it seems implausible that something could emerge as a medium of account without already being widely enough marketable to provide at least some guidance about its value relative to other goods. The only reason this poses no puzzle to Graeber is that he apparently considers most prices to be significantly arbitrary and is entirely oblivious to their vital role in workable resource allocation.
UnlearningEcon writes:

This review seems like it was made by somebody who has something to contribute, but really just wanted to throw as much as possible at the author to try and discredit him.

To start, Bob Murphy's critique was certainly not 'excellent', firstly because Murphy had not even read the entire interview, let alone the book, and secondly because it basically relied on the argument that, while barter hasn't been shown to arise in a Mengerian fashion, it might have and Graeber can't prove otherwise!!!!!

More importantly, your idea of debt as 'delayed barter' is exactly the type of thing Graeber is after in his book - defining something arbitrarily and then contorting everything around it to make it fit your definition. He tries to set apart his communism, hierarchy and exchange considerations to combat this, and criticises attempts to view absolutely everything as a form of debt, too.

You can't simply call debt 'delayed exchange' if the delay itself is the most important part of the relation. Graeber goes to pains to emphasise that early societies operated with people always a little bit in debt to each other, and this bound people together. This wasn't 'delayed exchange;' it was intertwined with community, trust and what was owed was vague - a pig might be exchanged for marrying one's daughter, for example.

Graeber argues problems only arise when it becomes cold and calculating, and that this only happens when there is a real threat to not paying the debts back. Whether this happens gradually or suddenly is besides the point; your speak of "evolution that resulted in shared prosperity" is the kind of narrative historians continually struggle to combat - "history naturally tended towards this state."

The only reason this poses no puzzle to Graeber is that he apparently considers most prices to be significantly arbitrary and is entirely oblivious to their vital role in workable resource allocation.

Where does he state this? He offers a critical perspective on markets, yes - and people are allowed to do that without referencing Hayek and making caveats about the USSR.

Also, don't mistake 'rejects' for 'oblivious to.' Markets are full of externalities, rents, fraud etc. They work best when things are produced and consumed regularly, and Graeber acknowledges that they are a 'convenient' way to feed people, originating with the desire to do just that for public servants.

One thing you are correct about is Graeber's characterisation of economics and economists as advocates of selfishness and other memes. But this is clearly a commonly shared conception - don't dismiss them from your tower and point them to a paper they don't understand; worry about the impression the spokespeople for your discipline have given, and communicate better with the public yourself.

@Ken B

There was barter even in my dormitory. Can someone explain this?

If people are already familiar with spot exchange then it will take place after their medium of exchange is removed, such as prison etc. But it rarely comes out of nowhere.

Brandon writes:

The debate between anthropologists and economists has been going on for quite some time now. As the two social sciences with the most to contribute to a better understanding of humanity, this is perhaps inevitable, but I still find the debate between Frank Knight and Melville Herskovits in the 1940's to be the most relevant to The Discussion. I am too lazy to find an online source, but if anybody is interested in their debate they can check it out in every edition of Herskovits's Economic Anthropology after the first. The debate can be found in the appendix.

Cornelius writes:

Hummel completely missed the point of the book. I have read Graeber's book, and it is in no way incompatible in the way Hummel believes it is. Graeber's grasp of neoclassical economics is poor, and many of his ideas are fanciful, but Debt does contain some brilliant insights.

David R. Henderson writes:

@Cornelius,
Your statement “Hummel completely missed the point of the book” would be more persuasive if you pointed out, say, two important things he got wrong besides the one he admitted in his comment above.

Ed writes:

@Jeff

Thanks for the clarification. Is it possible for you to publish the critique please? I'm very interested to read it. I haven't read Graeber's work, but I've read Bob Murphy's critique and counter-critique to Graeber's response on Murphy - I'd be interested to see if you and Murphy arrive at the same criticisms independently.

Jeff Hummel writes:

@ Unlearning Econ

Here is how I interpret Graeber on stages of social evolution:

1. Reciprocal altruism, which he calls communism, and is the arrangement you describe. It is something evolutionary biologist have found among many animanl species. I realize that I may have been unclear about distinguishing this from barter debt transactions. I agree fully that reciprocal altruism does not constitute genuine debt or barter. I don't even include the interim development of gift exchange as a true debt transaction.

2. Genuine debts requiring specified repayments, somewhat enforceable, but with no medium of excahnge; what the anthropolgist Humphrey labels "delayed barter." This doesn't require a general medium of account, atlhough one may emerge.

3. Monetary exchange in both spot and debt transactions.

Obviously later forms of interaction do not entirely displace earler ones, as we still observe reciprocal altruism in most families. But are you arguing that I got Graeber's sequence wrong in some fundamental way?

As an anarchist, I certainly don't think these developments in any way necessitate the State, whereas I understand Graeber to be arguing that the State is required at least for step 3 (and possibly for step2?). Indeed, that is one of my basic disagreements with Graeber. The unfortunate historical fact that these developments were often accompanied by the emergence of the State is of no more causal significance than the fact the most modern technological development has occured within State-ruled societies.

@ Ed

What appears beginning in the second paragraph of David's post is the full extent of what I've written about Graeber's book so far.

Cornelius writes:

@David,

As UnlearningEcon has already said, Graeber's conception of "delayed barter" and debt are essentially the same thing. When someone gifts me a chicken, I am indebted to them, and shall have to repay the favour in the future. Furthermore, Graeber never claims barter did not exist in early human societies - it existed among strangers, who did not have repeated interactions, which actually makes sense within a rational choice framework. Just because Hummel does not view this as a "true" debt system, does not make it so.

I also am unsure the extent to which economists do NOT believe in the traditional textbook barter story, unless they have indeed studied economic history - I certainly believed the traditional barter story until I engaged with anthropological literature. At any rate, it is a story we teach our students, and it is wrong to do so.

I can appreciate Hummel's criticisms, but they failed to address Graeber's overall argument - granted, some of Graeber's anecdotes and data are just plain silly, but to critique his overall argument regarding credit and bullion cycles, one needs to undertake a systemic critique of the entire text, which I am not convinced Hummel has accomplished. It is like claiming the Wealth of Nations is incorrect because it contains a few factual errors (which, to be fair, Graeber does himself!).

I agree with Hummel, that it is difficult to establish a causal link between the early state and the development of coinage. However, Graeber's theoretical story is compelling: to maintain a large standing army requires paying that army, and bullion is a convenient and low-cost way to do so. For that bullion to be redeemable requires markets, which can partly be engineered via taxation. Hence state-->coinage-->markets.

Graeber furthermore never claims that "pre-civilization" societies required altruistic exchange; any gift exchange that happened was self-interested and hence required repeated interaction for delayed barter to set in. Furthermore, much of the anthropological material Graeber discusses is grounded in symbolic relations between people, not economic relations per se, which Hummel appears to miss out on.

As I said, Graeber has a poor understanding of neoclassical economics, and his academic sins will find him out. However, Hummel seems to miss the point of Debt.

Ken B writes:

@Unlearning Econ:

If people are already familiar with spot exchange then it will take place after their medium of exchange is removed, such as prison etc. But it rarely comes out of nowhere.

1) I confess I am again at a loss to know how one can know this. We had small bands on the veldt for a very long time. And chimpanzees do some bartering.

2) Doesn't Gruber need never not just rarely?

3) In some species males exchange food for sex. Why is that not a spot trade?

Comments for this entry have been closed
Return to top