Sunday, December 16, 2012

Chartalism = the McDonald's coupon theory of money


While chartalism is usually associated with the state and its ability to tax, there's no reason that this must be so. In fact, I like to think of chartalism as stateless. I call this the McDonald's coupon theory of money.

What I'm specifically addressing here is the chartalist idea that paper notes issued by the state have value because the state imposes a liability on its citizens that can only be discharged by the use of this paper.

McDonald's Corporation can do the same thing if it imposes an obligation on the burger-eating community to pay for all Big Macs with McDonald's-issued coupons. One coupon buys one Big Mac, say. In order to pay their "tax" and get a burger, Big Mac fans have to somehow acquire these coupons ahead of time.

McDonald's pays for a portion of its supplies by printing and issuing these coupons as payment. So one way that Big Mac fans can get their hands on coupons is by providing services to McDonald's... say cleaning floors or selling them pickles. Another way fans can get coupons is by approaching someone else who is already a supplier to McDonald's and offering them gold, silver, or some other media in exchange for coupons. An active secondary market would probably develop for these coupons.

McDonald's is required to spend enough coupons into existence so as to meet the very real demand people have to settle their obligation to buy Big Mac's with coupons.

Because McDonald's and Big Mac fans are everywhere, it's likely that these coupons will begin to circulate broadly. For instance, corner stores may begin to accept McDonald's coupons in payment for stuff, as will hotels and other merchants. They'll be willing to do so because they know they can pass the coupon off as change to a Big Mac fan at some later point in time. If not, they themselves may want to use it to buy a Big Mac. McDonald's coupon become ever more liquid, their degree of moneyness increases, and they emerge as a globally-useful medium-of-exchange.

Now burger lovers needn't submit to the McDonald's "tax". Burger King still allows customers to buy a Whopper without having to secure a coupon ahead of time. But if Burger King, Wendy's, A&W, and other burger joints institute a similar coupon-mechanism, then burger lovers will have to submit to some sort of coupon-tax.

In any case, I hope you can see why chartalism needn't be explicitly intertwined with the state. We can imagine worlds without states that have circulating chartal media-of-exchange. This is a world of private coupon monies. Theorizing in this way makes monetary discussion easier since it removes the political aspect of the debate, rendering it purely technical.

While I can imagine chartal worlds, I don't think the real world is particularly chartal. Coupons exist, but they rarely circulate outside of a very proscribed range. Here is one example of a circulating chartal money, in which Zimbabweans started to widely use gasoline coupons during the Zim$ hyperinflation. Chartal monies certainly become more prevalent during times of monetary stress.

Nor do I think modern central bank liabilities are chartal coupon monies. In my hypothetical example, McDonald's spent coupons into existence. Modern governments can't fund themselves by issuing coupons, they can only spend after having taxed or issued bonds. Nor are people obligated to use central bank notes or deposits to discharge government obligations – they can, and typically do, use other media. The upshot of this is that if governments announced that they'd only accept gold in payment of taxes from here on in, central bank liabilities would continue to be a valuable and popular medium of exchange. Not so our hypothetical McDonald's coupons. If McDonald's ceased accepting coupons for Big Macs, they'd be worthless.

PS. Just because I am saying that modern central bank money is not chartal doesn't mean that I think we can ignore the state in understanding how modern money gets valued in the marketplace. I think that acceptance of particular media of exchange by the state helps drive the liquidity premia of those media. See this old post.

44 comments:

  1. JP,

    I think that if you fixed the price (1 Big Mac = 1 coupon), it would not work, because once the liquidity premium emerged, the analogy of Gresham's Law would kick in and remove the coupons from circulation. Both price and liquidity must emerge through market forces (that's my interpretation of the Mises' Regression Theorem, as I explain it in my thesis).

    Now I'm not as strict as Mises and I'm not going to say that this necessarily refutes chartalism, it just makes it very very difficult to pull off, certainly more difficult than chartalists portray it.

    ReplyDelete
    Replies
    1. You'll have to explain to me how Gresham's law would kick in. Say that McDonald's sells Big Macs for 1 gram of gold. It switches to selling them for 1 coupon. At first, the market is willing to pay 1 gram of gold for a coupon. Over time, coupons begin to be prized for their liquidity such that they cost 1.05 grams to buy. I'm assuming that Big Macs stay the same and people's tastes stay the same.

      Delete
    2. If Big Mac costs 1 gram of gold, and the coupon appreciates to 1.05, if they still accept the coupons at an exchange rate of 1 BM / 1 coupon, people would trade coupons for gold outside of Mc Donalds and buy BMs with gold. Paying 1 grams of gold for 1BM is preferable to paying 1.05 grams of gold. So the coupons would only circulate outside of Mc Donalds (assuming the liquidity would be sufficient).

      I don't know if 1.05 is a ratio sufficient for GL to kick in, but I hope I got my point through.

      Delete
    3. Sorry, one more addition.

      Since you say that Mc Donalds would impose an obligation to use the coupons, the more probable course of action is that at a sufficiently high liquidity premium of the coupons, people would simply increasingly switch to a competitor of Mc Donalds (say Burger King). It is a self-defeating act, unless of course you're a monopolist (e.g. the state).

      Delete
    4. "...people would trade coupons for gold outside of Mc Donalds and buy BMs with gold. Paying 1 grams of gold for 1BM is preferable to paying 1.05 grams of gold."

      That's the whole point. You can't buy coupons for 1 gram of gold anywhere but outside of McDonald's, so you can only ever pay 1.05 grams for them.

      "...at a sufficiently high liquidity premium of the coupons, people would simply increasingly switch to a competitor of Mc Donalds (say Burger King). "

      Yes! I think that is the better answer. Another answer is that Burger King, which can only sell Whoppers for 1 gram, sees that McDonald's is selling theirs at a premium b/c their coupons have moneyness. Burger King will try to steal the premium away from McDonald's by instituting its own coupon system. This is a similar idea to my old bitcoin post in which I talked about how litecoin and other alt-currencies could steal away BTC's premium.

      Delete
    5. BK would have it in my opinion more difficult to do something like this if already McD did it first. BK would also have to compete not only with McD but also with KFC, Taco Bell and others. First mover advantage, or exactly as you said in your previous post, "liquidity is sticky". But in any case I find it difficult to see how McD could profit from this in the long run. If the purchases of Big Macs were compulsory, that would be much more easier.

      I'd like to remark here however that whenever I see others saying "I can't imagine how this or that would work", my interpretation is that they lack imagination. So maybe I'm missing something.

      Furthermore, there is no price fixing in Bitcoin, and it appears that the production costs follow the market price (and theory supports this), which means there is no point to switch to a competitor in the first place. There could be other reasons why to switch (e.g. 51% attack and so on), but these are not influenced by the liquidity premium.

      Also, money has the characteristics of having the influence of location on its value arbitraged away. This is not the case for Big Macs. And as far as I know, Walmart (which you mention as another option) does not even have any stores in Europe. Bitcoin, on the other hand, has a much narrower location-specific value (see http://bitcoin-analytics.com/#arbitrage ), lower than Big Macs. According to Wikipedia ( http://en.wikipedia.org/wiki/Big_Mac_Index ), Big Macs are also not homogeneous around the world.

      Delete
  2. 1) When Massachusetts first issued paper shillings in 1690, they didn't insist that taxes be paid only in paper shillings. They accepted quite a lot of other things as well, as did all the other colonies.
    2) I don't pay my taxes in Federal Reserve notes, but in privately issued substitute moneys that are denominated in dollars. If I fail to pay, they will accept my house or car instead.
    3) Chartalism says taxes create a demand for money, while the backing theory says money is backed by the issuer's assets, one of which is taxes receivable. Very different theories.
    4) Modern governments can and do fund themselves by issuing coupons (e.g., dollar bills), which they use to pay employees, buy materials, etc.

    ReplyDelete
    Replies
    1. 1) Agreed
      2) Agreed. That's my point here when I said; "they can, and typically do, use other media"
      3) Agreed, though the topic of this post isn't about backing theory.
      4) The US government is prohibited from having the Fed spend on its behalf. It can only pay employees by first raising taxes or issuing bonds. The setup differs in other countries but this is generally how the divide works.

      Delete
    2. "if governments announced that they'd only accept gold in payment of taxes from here on in, central bank liabilities would continue to be a valuable and popular medium of exchange. Not so our hypothetical McDonald's coupons. If McDonald's ceased accepting coupons for Big Macs, they'd be worthless."

      Big macs were the only thing the coupons could be redeemed in, so if McDonald's stopped accepting them for Big Macs there are no channels of reflux left for the coupons. If the government stopped accepting central bank notes for taxes, there are still lots of other channels through which the notes can reflux to the government. But if the government closed all reflux channels like McDonald's did, then the central bank's notes would lose all value just like the coupons did.

      Delete
    3. Yes. C'mon Mike, I read your paper. I know how it works ;)

      MMT-types constantly use the slogan taxes-drive-money. The key point of this post is to illustrate how the chartal view is state-independent. You're seizing on my last paragraph which probably could have been a post on it's own since there's a lot of material in their to unpack.

      Delete
    4. For curious readers, this is the paper Mike and I are talking about. It's also a politically neutral way to think about chartalism.

      http://mpra.ub.uni-muenchen.de/24813/

      Delete
    5. Beat that one to death, did I? Sorry, one of my many faults.

      Delete
  3. In the textbook setup of a strictly independent central bank, taxes only back the government's bonds, not money. If you don't pay taxes, you go to jail, but the value of money isn't determined by that. The central bank sets the value, subject to the constraint that value*quantity can't be greater than the CB's assets.

    But just because money has value, doesn't mean anyone will use it as a unit of account. Why do most Canadians use the Canadian dollar and not the U.S. dollar? The U.S. dollar is more popular globally. But the fact that the Canadian government uses the Canadian dollar makes it convenient for Canadians to use it.

    ReplyDelete
    Replies
    1. The unit of account (I presume you mean the unit in which shopkeepers post prices?) is a tough issue. I sidestepped it in this post for the sake of keeping things short. The state can to some degree force people to post prices in a certain unit, but there is a large degree of market choice. Kurt Schuler gave a good anecdote about this.

      Delete
    2. It's a market choice in the sense that it's not directly dictated by government, but it's heavily influenced by government being the biggest player in the economy. Almost everyone has to transact with the government, if only to pay taxes. These transactions with government don't give money value (Chartalism is wrong if it claims that), but they make the government's money the most convenient unit of account.

      Did you see this post by Marcus Nunes?
      http://thefaintofheart.wordpress.com/2012/10/31/two-kinds-of-money/

      Delete
    3. "but they make the government's money the most convenient unit of account."

      Makes sense to me. And in a stateless world, the biggest player could be Walmart (?), in which case whatever unit Walmart chooses to express prices would make that unit convenient for everyone else.

      Yes, I read Marcus's post. It's a great anecdote. My main complaint with the whole MOA/MOE debate was a poor usage of terminology, specifically the term MOA. I've complained about it here and on David Glasner's blog. If we get the definitions for UOA, MOA, and MOE right it makes debate infinitely easier.

      Delete
    4. I agree with White (http://ideas.repec.org/p/cvs/starer/82-18.html) that

      "A non-exchange medium numeraire commodity would furthermore be subject to greater bid-ask spreads in barter against other commodities, as by hypothesis it is less saleable, than the medium of exchange."

      White also argues about fluctuating exchange rates (which looks similar to your "convencience") create a problem for a non-unit-of-account medium of exchange, but computers have made that largely irrelevant. Merchant systems based on Bitcoin, for example, are typically set to quote prices in USD (or whatever local currency is) and use the realtime exchange rate to determine how much to request in BTC. Payment processors also provide exchange rate guarantees (for a fee), although theoretically if you have enough reserves compared to your cash flow you can do it yourself (open a hedging position on an exchange when you see a payment incoming).

      Delete
    5. While I agree that bitcoin transactions certainly have low physical transaction costs (9 minutes to clear, 0.01 BTC transaction fee), there are still only a few transaction one can make with it. So one has to spend time searching out for potential transactors. Search costs cancel out the low physical transaction cost advantage. So on larger scale, bitcoin is still not extremely saleable and therefore suffers from a pretty wide "global" bid-ask spread.

      Delete
    6. JP,

      you can sell Bitcoin 24/7 from anywhere in the world at negligible costs against fiat monies in almost any practically relevant quantity at a price close to the market price. Chances are that if you live in a big city, you can also find someone nearby willing to buy your Bitcoins for cash. This is unlike anything that existed before. It creates liquidity and mitigates the "double coincidence of wants" issue somewhat. It's a bit unusual because typically this happens the other way around (first there are a lot of people demanding it, then you have highly liquid markets), but because of the transaction costs the highly liquid markets (Menger calls them "organised markets") formed before Bitcoin is widespread.

      I think this confuses the typical economist when analysing Bitcoin. They look at the "hard" factors of liquidity that are normally relevant, and miss all the "soft" factors that readily compensate for the "hard" ones in case of Bitcoin.

      Delete
    7. I can trade my charts (financialgraphart.com) on my street for cash, and transaction costs are 0. No commissions, no time wasted in passing the chart from hand to hand. But finding someone to accept my charts is the tricky thing. Same with bitcoin. Sure, I can pass it on at 0 cost, no commissions and whatnot, but finding people willing to accept it is the tough thing.

      Delete
    8. JP,

      I doubt that your charts are divisible and homogeneous, that their price is close to uniform around the world, and that you can sell them anytime you want in almost any quantity you want from anywhere you want.

      One of your charts, for example, lists price as $38 and shipping costs $19. That's no transaction costs? $36 for international shipping is uniform around the world? And if you sell it on the street, you need to print it or at least provide a method for the buyer to get it (digital medium or download). Since people can apparently download the image for free from your website, presumably you're talking about a poster. You have no transaction costs with the poster? Manufacturing and storage costs?

      Say you're on a vacation in Tokyo and are short on cash. How quickly can you liquidate your stash of posters into yen? Are you going to lug around the posters on your trip?

      Delete
    9. JP,

      check out these two articles for more "soft" details on liquidity:

      http://thebluemarket.wordpress.com/2012/10/18/bitcoin-dollars-and-pot-banging-protests-in-argentina/

      http://www.businessweek.com/articles/2012-11-29/dollar-less-iranians-discover-virtual-currency

      You can sell Bitcoins without problems in Argentina or Iran, anytime and close to the market price. Can you say the same about your posters?

      Delete
    10. I think you missed my point. I was merely saying that having no physical transactions/fees/commissions costs does not equate to being "liquid".

      Let's defer this conversation to my next bitcoin post or one of yours. It's getting pretty far from the main point which is coupon money/chartalism.

      Delete
    11. I agree that low transaction costs do not mean liquidity. However you appear to miss my point that low transaction costs have a positive effect on liquidity, because it gives an advantage in indirect exchange and through network effect is self-reinforcing. On the other hand, things which have high transaction costs are unlikely to become liquid.

      A medium of exchange with very low transaction costs in the narrower sense can be better medium of exchange even in cases where it has a lower liquidity than the alternative. See e.g. http://www.thebitcointrader.com/2011/10/send-money-to-china-with-bitcoin-and.html

      If it did not, then it would be impossible for one money to replace other money (I actually cought Hoppe in contradicting himself, because he simultaneously claimed that liquidity always determines what people use for money, and then claiming that one medium of exchange can out-compete another one).

      Delete
  4. I've seen debts settled in Canadian Tire money.

    ReplyDelete
    Replies
    1. Canadian Tire money... great example.

      I wrote something about CT$ a few year's back at Nick Rowe's blog. Reading through Nick's old post I see a lot of similarities with this post.

      Delete
  5. FYI, this is my critique of MMT, more specifically how it fails to understand the workings of interest rates and central bank actions.

    ReplyDelete
    Replies
    1. Thanks for the link. I agree with you. I always did find the dichotomization between government and non-government sectors to be a somewhat arbitrary basis around which to build a grand unifying theory.

      Delete
  6. Maybe I'm missing something, but it seems to me as though this does not essentially differ from the Mengerian or Austrian account. For one, you essentially have these coupons backed by burgers (an interesting thing to back a currency, but possible, I suppose), as you can - at will - return to McDonald's and get a Big Mac with each one. Additionally, the question is to ask why a market participant would demand payment in a coupon that no one else accepts and they created themselves - I find that to be unlikely and, frankly, irrational. (The state gets around this problem with the legal tender laws that force people to accept, in addition to other monies, their currencies - which require some backing, at least initially, to have a value that makes sense. Thus, gold and silver backed what are now fiat currencies in the world.)

    ReplyDelete
    Replies
    1. Matt, Mengerian ≠ Austrian.

      But you're right that what I've explained is essentially Mengerian.

      Mises, on the other hand, would say that if McDonald's ceased converting coupons into burgers, they would still circulate as "money" because people would use the value of yesterday's purchasing power to compute today's utility of coupons. In this case, money would be "the ghost of burgers", not the "ghost of gold."

      "...why a market participant would demand payment in a coupon that no one else accepts and they created themselves."

      I don't think its likely either. The point of this post was to put chartal money into an imaginary world without a state in order to decompose the essence of chartal money. At its core, chartal money is a highly liquid coupon, not a state-issued token.

      Delete
    2. You think Mises would argue that these coupons would still circulate as money? Sure the utility gained at the time of the announcement ceasing redemption is based on X-1, where the announcement is X, but the purchasing power at X would be 0, resulting in a 0 utility at X+1, which can be a moment after the announcement, not a day.

      Delete
    3. I don't see how Mises couldn't adopt the argument that irredeemable coupon money continues to circulate. It's implicit in his regression theorem. At X, purchasing power is not zero since everyone is willing to accept coupon money based on yesterday's purchasing power. But I'm not a big fan of the regression theorem, I think expectations are largely forward looking, not backward looking.

      Delete
    4. You are right. That makes no sense. I picked up my copy of Human Action and as I suspected, Mises would (like anyone else) realize that the coupons were worthless and not argue that they would continue to circulate as money.

      For as you noted to me in earlier replies, which I don't think I appreciated fully, the regression theorem is a very specific explanation for a very specific phenomenon: "the regression theorem aims at interpreting the first emergence of a monetary demand for a good which previously had been demanded exclusively for industrial purposes"

      That specifically excludes situations like the one being discussed here, so I see no reason to handcuff Mises into applying a theory he deliberately emphasized was only useful in explaining this particular type of money.

      Delete
    5. I think it is an unfair reading of Mises to say "yesterday's purchasing power" and imply that he believes expectations are backward looking. The entire essence of Human Action is about the future. And making uncertain plans for that future etc. But specifically in regards to this issue we have stuff like:

      "He who considers acquiring money is, of course, first of all interested in its future purchasing power...But he cannot form a judgement about the future purchasing power of money otherwise than by looking at its configuration in the immediate past."

      Which seems pretty reasonable to me.

      Delete
    6. Perhaps I am being unfair since there is a significant change in Mises's thinking from his initial Theory of Money and Credit to his subsequent Human Action, and I am dealing primarily with his ideas in the former. Gertchev talks about this evolution in Dehomogenizing Mises's Monetary Theory (2004) and I agree that it's important to be aware of the changes in Mises's thinking. In the future I'll try to speak in terms of his latter monetary theory.

      It remains the case though that "money" only has value according to Mises because of its expected future purchasing power is conditioned by its earlier purchasing power. Thus gold backed money continues to have value even when it has lost all connection with gold. Yesterday is important in forming expectations. By analogy, McDonald's coupons can continue to have exchange value, even if they are no longer redeemable in Big Macs, since people are able to form future expectations about these coupon's value by using yesterday's purchasing power. At least, using Mises's logic.

      Delete
    7. According to Mises's logic people would use the immediate past's configuration to guide their future expectation of the purchasing power of a money. McDonalds announces coupons are now worthless. The moment after this occurs, people look back at this new configuration and significantly decrease their value for the coupons. I don't see anything in Mises's work that says this is impossible and the coupons must circulate because people would ignore the most recent development and only care about the configuration from 24 hours ago.

      Instead, I have an overwhelming impression that people are constantly doing their very best to anticipate the uncertain future and prices are a reflection of the immediate past which can serve as a guide in anticipating the future.

      All prices reflect the past, not the future. I don't think one sentence speaking in general terms of "using yesterday's purchasing power" to help determine one's future expectation of the purchasing power of a money, should handcuff Mises into the view you are attributing to him.

      Clearly the end of redemption would significantly alter the price configuration of the McDonald's coupons, and those newly created prices, would, like all prices, represent the immediate past, which would inform people's newly decreased valuation of the money.

      Delete
    8. Let's put it this way. What Mises says about fiat money goes like this. The central bank issues money convertible into x. One day the CB ceases to allow convertibility yet its paper continues to circulate as "fiat" money. Why? Because people look to yesterday's purchasing power, the day on which it was still convertible into x. I don't think that my summarization is controversial.

      Say McDonald's issues money convertible into x that circulates widely. One day McDonald's ceases to allow convertibility into x yet McDonald's paper continues to circulate as "fiat money". Why? Because people look to yesterday's purchasing power of McDonald's paper, the day on which it was still convertible into x.

      It makes no difference whether x is gold or Big Macs. The principle is the same - so it seems reasonable to me to attribute to Mises the view that fiat McDonald's money might circulate as the "ghost of Big Macs".

      Delete
    9. According to my reading of Mises and Rothbard, they say that once a medium of exchange loses its non-monetary uses, it may remain a medium of exchange. They do not say that it must remain so, i.e. they do not argue that all media of exchange are sustainable.

      My addition is that whether they remain so or not depends on the status of competing media of exchange. If it already is the most liquid medium of exchange, all other things being equal, it will remain so. If it isn't the most liquid one, it probably won't. But if there are other distinguishing factors of competitors (e.g. better store of value or better transaction costs) it still might lose even if it has an initial advantage in liquidity. Gresham's Law also plays a role.

      Delete
  7. This comment has been removed by the author.

    ReplyDelete
  8. Warren Mosler has an excellent article on how coupons can become money, and illustrate the verity of chartalism in practice, I recommend it, over on HuffPo.
    The UMKC Buckaroo: A Currency Model for World Prosperity
    http://www.huffingtonpost.com/warren-mosler/the-umkc-buckaroo-a-curre_b_970447.html

    ReplyDelete
    Replies
    1. Yes, that's a good example of coupons gaining moneyness. LETS systems too. Chartal coupon monies do exist, though they're in the minority.

      Delete
    2. Not entirely. Based on Mosler's description, buckaroos are claims on a commodity. Specifically, one buckaroo is a claim on 1/20th of semester grades. This gives them a starting market price and liquidity. Whether that is consistent with chartalism or not I don't know, but it's not an example of fiat money. In order for them to start being usable as a medium of exchange, semester grades already must have existed and there must have been a demand for them.

      But I would say that Mosler understands the issue of pre-existing demand, based on his confusion about Bitcoin: http://www.youtube.com/watch?v=m_NePQAODv0

      Delete
    3. The "fiat" (command) is that they are worth that claim on grades, by the "gov" - in this case the Uni.

      Delete