Back in 2018, David Andolfatto made a great presentation of what he calls “the recordkeeping aspect of monetary exchange”. Not only do we agree with much of what he says, but we think this is the fundamentally right way to think about monetary exchange. That is why we find it unfortunate that Andolfatto, like Narayana Kocherlakota before him[1], seems to stop short in his description of phenomena from this angle. Similar to what we have done in our paper, we will try again to take this idea to its logical conclusion in the following post. Let us proceed.
Around 25 minutes into the presentation, Andolfatto tells us a story about Adam, Betty and Charlie who specialise in producing dinner, breakfast and lunch, respectively. It’s a really enlightening story, and you should probably watch it to fully understand our sequel. Andolfatto introduces a recordkeeping device, a token[2], and shows how it functions as evidence of its holder’s trading history. So far, so good, although the suggestion that a seller cares about the buyer’s trading history, and demands evidence thereof, does not carry all the way to a wider monetary exchange. There, an individual seller only cares about evidence of his own trading history, which he might need to prove to a dedicated monitor (bank). The traders do not monitor each other, as they do in an informal setting (e.g, the arrangement based partly on trust, which Andolfatto brilliantly explains in the beginning of his story).
What Andolfatto leaves unexplained, is how Adam came into the possession of the token in the first place. This is typical of economists, with money in their models often appearing as a “manna from heaven”. When asked on Twitter about the origin of the token, Andolfatto says Adam might have received it as a “token of appreciation” after having worked for the government. Or he might have found it on the seashore. It doesn’t matter, according to Andolfatto.
Coming from what we could call a “strict accounting school”, we are not happy with his answer. Monetary tokens issued by a government — whether we are talking about coins, treasury notes or stocks of tallies — serve an explicit recordkeeping purpose all the way from the initial transaction, and not just once they start to circulate among the public (from Adam to Betty in the example). This means the transaction where the token is issued results in a credit and a debit entry, as with all transactions. Anything else would be illogical. Let us try to explain, in terms of Andolfatto’s example.
In addition to Adam, Betty and Charlie, we would like to introduce a guy named Joe. One day, in the time before the introduction of tokens, the four of them decide among themselves that Joe should be their president, an office in which he is responsible for looking after the common good of the four. In this small community, the president is the government. As the president is too busy to cook meals for himself, the other three agree to take on an obligation to provide Joe with breakfast, lunch and dinner. Betty provides breakfast, Charlie provides lunch and Adam provides dinner. This is how the three pay their taxes. There is no public debt in this kind of setting — only taxpayers’ tax liabilities.[3]
In the world of Andolfatto’s story it would look something like this (graphics borrowed from his presentation and edited by us):
After having enjoyed tasty meals three times a day for three months, Joe confesses that he has become addicted to Adam’s steaks. From now on, Joe will accept meals only from Adam, three times a day. But no one thinks it fair to place the tax burden (preparing steaks) solely on Adam’s shoulders. The situation calls for public recordkeeping (a.k.a. monetary exchange).[4]
Each time Adam delivers a meal to the president’s office, he will be issued a token as a receipt for having parted with goods without receiving anything in return. It is not just any “token of appreciation”, but part of formal accounting. Joe has a ledger where he makes a credit entry in a column called “Tokens in circulation” every time he issues a token. Having taken Accounting 101 in his youth, he knows he needs to debit an account, too. So he makes a debit entry in a column called “Government purchases”. Naturally, these purchases increase public debt, ceteris paribus.
In the evening of the first day (see picture below), Adam holds 3 tokens (credit balance). Government purchases, and by extension public debt, stands at 3, too (debit balance). The meaning of these balances is as follows: Adam has the right to receive goods worth 3, as he has delivered goods worth 3 to the public earlier. And if Adam has the right to receive goods, someone else must logically have an obligation to deliver goods. In this case, it is the public, although no one in particular, who has that obligation (debt). Notice that this doesn’t mean Adam can demand goods from anyone. The tokens are not claims on a particular subject, be it another member of the public or the government. They are just records of trading histories. Adam’s right is based on what we have called “the law of reciprocity”.[5]
We have chosen to present Adam’s token holdings as a credit balance on a T-account. We do this to focus on the records themselves, and not on the recordkeeping devices. The government records kept by Joe are no doubt a simplified version of actual records in the modern economy, but the principle is the same as it has been for hundreds, even thousands of years.[6] It is these records which Andolfatto seems — depending on how one interprets him — to have overlooked when he suggested that Adam might have received the token as a “token of appreciation” after having worked for the government.
But let’s get back to the story. The next day, Joe decides it is time to raise some taxes to reduce the amount of public debt. As before, the taxpayers pay their taxes by producing meals for someone else, without earning the right to receive anything in return — this is just like it was when all three were producing meals for the president. The records of their trading histories should reflect this fact. As the public debt is 3, Joe places a tax obligation of 1 (meal) on each of the three, and gives them only one day to pay the tax — and prove it. How do they prove it? With the help of tokens, of course. Just like Andolfatto suggested.
For Betty and Charlie, the task is somewhat complicated. Since the president does not consume breakfast or lunch, their payment of tax — that is, supplying a meal without earning the right to a meal in return — will not take place in Joe’s presence. So, proof of that payment must be presented to him by handing over a token. That ensures that private and public trading records can be updated.
A simple solution might go like this: Betty serves two breakfasts to Adam — according to his tastes, we assume — and receives two tokens as a record of it (red entries in the picture below). On the same day, she receives a lunch from Charlie, and hands over one token to him (blue entries). Now each person has a (net) credit balance of 1 in token format. (We have “frozen” Adam’s steak deliveries to the president in order to simplify the example.)
When taxes fall due, Adam walks into the office of the president, gives Joe his token, and says: “See, I have already paid my tax by producing you a dinner.” “Fair enough”, says Joe. He debits the column “Tokens in circulation” and credits a new column called “Taxes paid”, both with 1. (Public debt = Government purchases - Taxes paid = 2.)
Then Betty and Charlie bring their tokens to Joe and tell him that they have provided two meals to Adam (directly or indirectly), who earlier provided three meals to the government. Joe debits the column “Tokens in circulation” and credits the column “Taxes paid” with one each for Betty and Charlie. (Public debt = Government purchases - Taxes paid = 0)
The tokens entered circulation as records of a transaction — Adam having given three steaks to the government (Joe) — and now they exited circulation as proof of three tax payments: Adam’s meal to Joe, Betty’s meal to Adam and Charlie’s meal to Betty. Public debt was reduced to zero and no tokens remain in circulation[7]. This is, of course, an unrealistic and even undesirable situation, assuming there are trades that the private sector would like to record by using government-issued tokens.[8] Even in our example, normally, when taxes fell due, the government would already have issued new tokens to Adam against the continuing steak deliveries, so that there would always be some tokens in circulation (we could perhaps call it a “float”).
As we can see, Tokens in circulation are a mirror image of Public debt, so they, too, function as a record of public debt.[9] In this specific sense, this type of money is debt. But this is not to say the tokens are “government IOUs”. The government owes nothing. The people do. And what we say is only true of government-issued tokens. Tokens issued by a central bank are not debt, not in this or in any other sense.
If you find it hard to get your head around our proposition, let us look at this another way. The outcome is the same, but this alternative reasoning[10] might be easier to grasp in the particular case of government-issued tokens.
When Joe issues the three tokens to Adam, we can view them as receipts for tax payments (the steaks; goods provided to the government). Remember that there was no tax obligation yet on Adam, so that these are actually receipts for payment of taxes he or others might owe in the future. The tokens could even include a text, something like “Has paid a tax of 1 meal”. Now, when taxes fall due, the taxpayers either have to provide goods to the government, or they have to prove that they have done so earlier. In our example, when taxes fall due, all three taxpayers are able to prove that they have already provided meals to the government and so have paid their taxes. (And they all lived happily ever after.)
With direct work for the government, one could rely on implicit recordkeeping to settle a tax obligation. But decentralized settlement, as with Betty and Charlie, requires some kind of explicit recordkeeping. The means of payment itself, though, is the same throughout (meals).
Let us try to sum this all up:
As we said in the beginning, we do not like the “manna from heaven” (or from the seashore) explanation. It doesn’t square with historical issuance and use of tokens. It makes a token, a recordkeeping device, look like a thing that is only meant to circulate as a “means of payment”, never to be retired.
Are we arguing about a small detail in Andolfatto’s presentation? We are not sure. We fully agree with Andolfatto if he says that Ostroy and Kocherlakota realized something fundamentally important about monetary exchange. At the same time, we feel we should be able to move forward. There is much unfinished business. Ostroy explored the idea in 1973 and in 1998 Kocherlakota picked it up, but only really produced one substantial paper as a result. Why did things stop there? Why are there no substantial papers on the subject in the subsequent 23 years? Or are we not just aware of them?
Antti & Oliver
[2] This reminds us of Ostroy & Starr’s (1990) story about two elderly gentlemen and a green stone, which the gentlemen used as a reminder for whose turn it was to provide a dinner to the other.
[3] Their tax payments are inputs in the production of a public good (output), which in this case is whatever the president does.
[4] Historically, this problem has been solved not only through recordkeeping with tokens, but also centrally. Joe could ask Betty and Charlie to deliver a meal (a tax) each to the President’s office, and, if Adam accepts, trade the two meals from Betty and Charlie against three meals from Adam (one of Adam’s meals would be a tax, the other two he would get compensated for). But tokens allow basically the same trade to be executed de-centrally, with increased freedom of choice, as we shall see.
[5] We have modelled this as agents seeking to balance their budgets of goods given and goods received over a lifetime. Importantly, it posits that, with a view only to their personal budget lines, there must always be those who give and those who receive for the system to function, without having to define who takes on which role in any given transaction. Creditors and debtors never meet -- only people do. This is the behavioural logic that powers multilateral credit. The “manna from heaven” story ignores this logic.
[7] As the government can create new tokens at will (assuming there is no “brake” set by legislators), the three tokens in its possession are not worth more to it than their intrinsic/material value.
[8] This is not a problem in our current system, where there are few if any government-issued tokens circulating among the public.
[9] This is true of all Treasury “liabilities”. They are records of public debt. To this day, coins are treated the same way as government bonds when they end up on the central bank ledger: they are treated as records of public debt and recorded on the debit side.
[10] This alternative explanation is actually partly based on a “true story”, from the time when tally sticks were used in England. See above-mentioned W.T. Baxter, 1989, “Early Accounting: The Tally and Checkerboard”, p. 54-55.
No comments:
Post a Comment