This archived content originally appeared at, the predecessor site to, and does not carry the sponsorship of the Cato Institute.

Medium of exchange versus unit of account

Scott Sumner and Nick Rowe have been having a friendly debate about which is the more essential function of money: medium of exchange (Rowe) or unit of account (Sumner). I think they are missing an important point: the combination of the three textbook functions of money — medium of exchange, unit of account, and store of value — in one good makes it far more powerful and significant than if the functions are separated.

When the functions are separated, attempts to influence the economy through monetary policy can lose much of their effect. Here’s an example. When I visited Peru in 1999, I noticed an advertising circular in a local newspaper for Ace Hardware. The prices were in U.S. dollars, not Peruvian soles. I asked my hosts about it and they said that at the cash register, customers paid in soles at the exchange rate of the day against the dollar. And this was after Peru had enjoyed five years of inflation at 15 percent or lower. People were still cautious because in the 1980s inflation been very high, peaking at above 7,000 percent in 1990. Even goods priced in soles might in practice be closely indexed to the dollar.

Where the local currency is not the dominant unit of account, changes in the local central bank’s policy will have much weaker effects on the economy than if people are using the local currency both as a medium of exchange and unit of account (which they are only likely to do if it is also a reasonably good store of value). It’s also true that in such a case, changes in the Federal Reserve’s monetary policy that are significant in the United States, such as a 0.25 percentage point change in the target interest rate, are not so significant in the partly dollarized country. The partly dollarized country has looser transmission of monetary policy–both for its own central bank and the Fed–to prices and incomes because of the separation of the textbook functions of money.

If I recall correctly, Larry White stressed the power of the combination of the textbook functions of money in an article quite a few years ago.

As some of you will know, inflation has remained low in Peru for many years now and, from what I understand, local-currency pricing predominates.

  • Larry White

    I agree with Kurt, but what I stressed in that article was that the unit of account naturally follows the commonly accepted medium of exchange. So does a third function, medium of redemption, and all three together create base money which allows advantageously low-cost settlement.

  • vikingvista

    Is the store of value function a relative or absolute concept? If the latter, then how would it be possible for something to be a medium of exchange and *not* a store of value? Why would somebody hold something for trade for any length of time if they didn't anticipate that someone would value it at the end of the duration?

    Likewise for a unit of account. In Peru in 1999, was the Sol really a unit of account for Ace Hardware transactions when all Ace Hardware goods were priced in USD? Sounds more like Soles were brought into Ace Hardware for barter, no different than if locals did the same with chickens or homemade rice cakes (with the only meaningful difference possibly being that chickens and rice cakes are probably not units of account in *any* context).

    It seems like the medium of exchange function subsumes both unit of account and store of value functions. One should only have to refer to a single function of money–medium of exchange.

    • Kurt Schuler

      Store of value is a relative concept: gold has kept its purchasing power better over the long run than the Swiss franc; the Swiss franc has kept its purchasing power better than the U.S. dollar; the U.S. dollar has kept its purchasing power better than the Venezuelan bolivar. Similarly for the unit of account: those that lose purchasing power fast also shrink in the volume of transactions people use them for.

      • vikingvista

        So it never makes sense to ask *if* something is a store of value, since anything that anybody holds is, by the act of holding, proven to be a store of value. The only question is *how good* of a store of value it is compared to other stores of value.

        Likewise, perhaps, for a unit of account. Every trade reveals a price (the ratio of the two items traded). So for any trade, either of the two items traded is necessarily a unit of account for that trade. The only question is how widespread is a particular unit of account amongst trades that occur.

        A store of value or unit of account to just one person (or even 2 or 3 or 10) aren't money, although may be a part of a money's evolution. So an essential property is popularity, in which absolute significance falls on a gradient with no possible precise cutoff. But relative popularities of two things can be compared with more precision.

        But then the word "medium" to me implies a significant level of popularity. So, if something is *exchanged*, it is both held and traded, so it is a store of value and a unit of account. Putting the word "medium" in front of it describes its popularity.

        So a medium of exchange, it seems to me, is necessarily always a store of value and unit of account with significantly widespread use. So "medium of exchange" covers all functions of money, while a store of value or unit of account needn't be so popular as to constitute a money.

        Or am I putting meaning into "medium" that isn't there?

  • ivancarrino

    I think that a good must be, in the first place, a store of value. Then it becomes a medium of exchange and thus a unit of account. However, all three of them are crucial for a good to be called "money".

    Funny story: in Argentina we have 25% annual inflation and for people not to run from the Peso the gvt decided to impose foreign exchange controls. A member of the government, then, justify the policy by saying that controls will continue until the Peso becomes a "store of value"… He forgot to read Menger.

  • Nick Rowe

    Thanks Kurt. Neat example of separation of medium of exchange and medium of account functions. I agree that monetary policy could have very different effects if those functions were separated.

    My argument against "store of value" being part of the definition:

    • Kurt Schuler

      I'm with Ivan Carrino (above) on this question even though I agree with you that the store of value function is not unique to money. If it's not a store of value it probably won't become or won't stay money. Admittedly, it can remain money even if there are other stores of value that retain purchasing power better, but if it's totally lousy as a store of value other things will displace it as money. So, I favor retaining all three of the textbook functions of money in the definition. Incidentally, Nick, Worthwhile Canadian Initiative is the one of the best blog names I have ever seen.

  • Gonzalo R. Moya V.

    Mr. Schuler, I am glad to read that you wrote about my country again. Please allow me to give you an update: "local-currency pricing" as you put it has become mandatory in Peru since 2006 and the dollar has never been of legal tender here (i.e. the central bank does not back it but merely allows people to use it as money "at their own risk"). The rise of the dollar as second currency happened during the hyper-inflation, precisely as store of value since the local one lost that function (as it does in every inflationary process). Although treasury bonds initially had to be issued in dollars for them to be sellable, such is no longer the case, as country-risk has reached the historically low level of 91 basic points. Both the people and commercial banks preferred to have their savings and to make their loans in dollars although doing so in the local currency is more profitable, but again, such is no longer the case as the dollarization ratio fell to 43.9% in August. Finally, there is an evident appreciating trend of the local currency that started since the beginning of the U.S. recession with no signs of fading in the near future, to the extent that the central bank had to intervene with open-market operations in order to smooth the downward path, thus earning net international reserves to over $62 billion in October, roughly one-third of our GDP.

  • Philo

    It seems to me you are simply taking Sumner's side against Rowe: you are saying that "monetary policy" involving the medium of exchange, if this is not also the medium of account, will be ineffective (while, by implication, "monetary policy" involving the medium of account *would* be effective, even if the latter were not also the medium of exchange).

    • Kurt Schuler

      No. In the Peruvian case, monetary policy involving the local unit of account was fairly ineffective because so many prices were indexed to the U.S. dollar. If the U.S. dollar itself had also experienced high inflation, Peruvians would presumably have switched to some other unit of account, rendering Federal Reserve monetary policy likewise ineffective in influencing the unit of account used in Peru. My point was that Sumner and Rowe were both thinking of a situation such as exists in the United States or Canada, where a single currency that combines all the textbook function of money dominates. Where you have multiple currencies and some separation of the textbook functions among them, matters are different.