A ticket or a counter?

When Bishop Berkeley in The Querist (1737) asked “And whether the true Idea of Money, as such, be not altogether that of a Ticket or Counter?”, he was clearly on to something important. But if one tries to suggest that money is a ticket, one will run into fierce resistance.

On Twitter, George Selgin is sceptical of the idea that money is some kind of ticket or "credit medium". As he writes: "If Wimpy offers a gold coin for a hamburger today, he's not proposing a credit transaction as he would by offering to make compensation on Tuesday!". Selgin also points out that the restaurant is in no way obliged to give Wimpy anything for the coin.

Point taken, Mr Selgin. You are fighting the last war, though. We are not proponents of Chartalism, nor do we do monetary theory by way of metaphors. That is why we refuse to call money a (government) IOU, liability or a ticket. As you said, metaphors are misleading in that they work only until they don’t.

But what if money is a counter? That is what we mean by a recordkeeping device. Our claim is that money is used to record the cumulative trading histories of agents. They participate in the system of their own free will, and decide whom to sell to and who to buy from. The guiding rule of the system is an intertemporal budget balance (sales = purchases), which we could call "the law of reciprocity"[1]. The participants might not understand how the system works, but the system is usually robust enough to survive despite this ignorance.

A similar system is actually behind one of the best-known models in economics, the Arrow-Debreu model. Here is Debreu in his book The Theory of Value (1959):

No theory of money is offered here, and it is assumed that the economy works without the help of a good serving as [a] medium of exchange. Thus the role of prices is as follows. With each commodity is associated a real number, its price. When an economic agent commits himself to accept delivery of a certain quantity of a commodity, the product of that quantity and the price of the commodity is a real number written on the debit side of his account. This number will be called the amount paid by the agent. Similarly a commitment to make delivery results in a real number written on the credit side of his account, and called the amount paid to the agent. The balance of his account, i.e., the net value of all his commitments, guides his decisions in ways which will be specified in later sections.

Debreu was unable to appreciate the possibility that he was offering a theory of money right there. We argue that Fed notes, and Fed and other bank ledgers in general, are used to make the debit and credit entries that Debreu is talking about. In his model, those entries are about (trade) commitments, while in our simplified model we talk about actual trades and thus the balance of the account represents the cumulative net value of an agent's trades so far.

We are not the first ones to suggest that money might serve this kind of purpose. Here is Ostroy in his paper The Informational Efficiency of Monetary Exchange (1973):

As a monetary version of the model of a trading economy, introduce a central receiving station called a monetary authority. Its function is to collect and collate the bits of information individuals have about each others' trading histories. Each will require his trading partner to write a signed statement, a check, indicating the amount by which the partner's purchases exceed his sales. This record is forwarded to the monetary authority who revises individual accounts on the basis of this new information. Sellers, by requiring payment in money, are guaranteeing a steady flow of information such that the monetary authority, and it alone, is able to monitor trading behavior. Of course, there is every incentive to require and deposit this information with the monetary authority; otherwise, one would not receive credit for sales and so have to cut back on purchases.

As far as we know, Ostroy has never expanded on this idea and applied it to our actual banking system. Yet, it seems to us that the banking system plays the role of Ostroy's "monetary authority".

As is obvious from the name of Ostroy’s paper, information plays a crucial role. In New Monetarist thinking[2], perfect monitoring of trading behavior would allow for some kind of credit arrangement, which in turn would make money inessential/redundant. So it is noteworthy that Ostroy above seems to suggest that the use of money provides that kind of monitoring, making credit feasible. 

In Ostroy and Starr (1990)[3], the two writers establish the most efficient way of trading:

Even though each person aims to execute an overall net trade with zero market value, the most efficient way to accomplish this in a sequence of pairwise trades is not to constrain the value of each bilateral commodity transfer to be zero. […] An individual who takes more than he gives at some pairwise meeting is simply executing a part of the overall plan to which the members of the economy have submitted themselves. It is as though the participants are agents in a firm carrying out their assigned tasks in front of each other. The lesson we draw is that in a world of complete information the requirements for enforcing overall budget balance are met, so quid pro quo is an avoidable constraint on the transactions process.

What they do not say clearly, unlike Ostroy in his earlier paper (see quote above), is that money provides the information which makes quid pro quo an avoidable constraint. On the contrary, in the paper they at times seem to suggest that money fulfills the quid pro quo (QPQ) constraint.[4] This allows Selgin to claim that Ostroy & Starr make a distinction between non-QPQ and QPQ transactions, and that money transactions are of the latter kind. Anyway, we are not arguing Ostroy & Starr’s case, but our own.

Per the Ostroy & Starr quote above, our argument with Selgin boils down to a question of whether money allows for the most efficient way to exchange goods or not. Selgin, like New Monetarists, says it doesn’t. It is widely acknowledged in New Monetarist circles (see footnote 2 again) that the formal part of Kocherlakota’s paper Money Is Memory (1998) proves that credit (no QPQ) beats money (QPQ) when full commitment or perfect monitoring is in place. This is a slightly embarrassing result, as it requires an explanation for why the most efficient solution is out of reach and why we thus have to settle (pun intended) for money instead.

Unlike Selgin, we positively glorify money when we argue that it enables exchange in the most efficient way possible by serving as a recordkeeping device. This recordkeeping allows for sufficient monitoring for a multilateral credit arrangement to work. But there is a catch. To fully accept this, one needs to renounce the concept of money as it is commonly understood. In a way, money becomes “nothing”, as it cannot fulfill the QPQ constraint. A buyer, often thought of as paying for the goods by handing over money to the seller, does not pay. In the words of Ostroy & Starr, he “takes more than he gives” in this pairwise meeting. When the seller accepts money as a record of having given up goods without receiving any goods in return, he does not care about the buyer’s trading history, nor can he know it. He only cares about the record that says that he, the seller, gave up a certain nominal amount of goods on this particular occasion. This is because he values the fact that he may receive more, or must deliver less, goods in future to satisfy his lifetime budget constraint.

We would very much like to hear from George Selgin whether he thinks ours[5] is a valid perspective, regardless of whether he thinks it is useful or not. It seems to us that his usual criticism of credit theories of money is not valid in the case of the Accounting View as we present it.

Antti & Oliver

[1] This rule has probably guided trading among humans for tens, if not hundreds, of thousands of years. We do not know how long people have been recording each other's trading histories, but archeological evidence suggests these kinds of records were what led to the invention of writing, and have been around at least since 8,000 B.C.

[2] See, for instance, p. 380-381 in Lagos, Ricardo, Rocheteau, Guillaume and Wright, Randall (2017) . “Liquidity: A New Monetarist Perspective”. Journal of Economic Literature, Vol. 55, No.2 (June), pp. 371–440.

[3] Ostroy, Joseph M. and Starr, Ross M. (1990). “The Transactions Role of Money”. Handbook of Monetary Economics, Vol. I, Edited by B.M. Friedman and F. H. Hahn, p. 11. Elsevier Science Publishers B.V.

[4] Our best guess for the reason behind this contradiction is that whatever intuition Ostroy had about the role of money was not shared by Starr.

[5] For more information on the perspective, see our paper.