Nick Rowe and I, with some valuable commentary from Bill Woolsey, Mike Sproul and Scott Sumner, and perhaps others whom I am not now remembering, have been having an intermittent and (I hope) friendly argument for the past six months or so about the “hot potato” theory of money to which Nick subscribes, and which I deny, at least when it comes to privately produced bank money as opposed to government issued “fiat” money. Our differences were put on display once again in the discussion following my previous post on endogenous money. As I have mentioned many times, my view of how banks operate is derived from one of the best papers I have ever read, James Tobin’s classic 1963 paper “Commercial Banks as Creators of Money,” a paper that in my estimation would, on its own, have amply entitled Tobin to be awarded the Nobel Prize. If you haven’t read the paper, you should not deny yourself that pleasure and profit any longer.
A few months ago, I stumbled across the PDF version of one of the relatively obscure follow-up volumes to the Monetary History of the US that Friedman and Schwartz wrote: Monetary Statistics of the US: Estimates, Sources, Methods. Part one of the book is an extended discussion about the definition of money, presenting various historical definitions of money and approaches to defining money. I think that I read parts of it when I was in graduate school, perhaps when I took Ben Klein’s graduate class monetary theory. As one might expect, Friedman and Schwartz spent a lot of time on discussing a priori versus pragmatic or empirical definitions of money, arguing that definitions based on concepts like “the essential properties of the medium of exchange” (title of a paper written by Leland Yeager) inevitably lead to dead ends, preferring instead definitions, like M2, that turn out to be empirically useful, even if only for a certain period of time, under a certain set of monetary institutions and practices. In rereading a number of sections of part one, I was repeatedly struck by how good and insightful an economist Friedman was. Since I am far from being an unqualified admirer of Friedman’s, it was good to be reminded again that despite his faults, he was a true master of the subject.
At any rate, on pp. 123-24, there is a discussion of definitions based on a concept of “market equilibrium.”
Gramley and Chase, in a highly formal analysis of monetary adjustments in the shortest of short periods (Marshall’s market equilibrium contrasted with his short-run and long-run equilibriua), discuss the definition of money only incidentally. Yet their analysis qualifies for consideration along with the analyses of Pesek and Saving, Newlyn, and Yeager because, like the others, Gramley and Chase believe that far-reaching substantive conclusions about monetary analysis can be derived from rather simple abstract considerations and like, Newlyn and Yeager, they put great stress on whether the decisions of the public can or do affect monetary totals. That “the stock of money” is “an exogenous variable set by central bank policy,” they regard as one of the “time-honored doctrines of traditional monetary analysis.” They contrast this “more conventional view” with the “new view” that “open market operations alter the stock of money balances if, and only if, they alter the quantity of money demanded by the public.
In a footnote to this passage, Friedman and Schwartz add the following comment (p. 124).
In this respect [Gramley and Chase] follow James Tobin, “Commercial Banks as Creators of ‘Money.'” . . . Tobin presents a lucid exposition of commercial banks as financial intermediaries with which we agree fully and which we find most illuminating. His analysis, like that of Pesek and Saving, Newlyn, and Yeager, and as we shall note, Gramley and Chase, demonstrates that emphasis on supply considerations leads to a distinction between high-powered money and other assets but not between any broader total and other assets. Unlike Gramley and Chase, Tobin explicitly eschews drawing any far-reaching conclusions for policy and analysis from his quantitative analysis.
Then on p. 135, Friedman and Schwartz in a critical discussion of the New View, of which Tobin’s paper was a key contribution, observed:
This approach is an appropriate theoretical counterpart to an analysis of changes in income and expenditure along Keynesian lines. That analysis takes the price level as an institutional datum and therefore minimizes the distinction between nominal and real magnitudes. It takes interest rates as essentially the only market variable that reconciles the structure of assets supplied with the structure demanded.
In a footnote to this passage, Friedman and Schwartz add this comment.
It is instructive that economists who adopt this general view write as if the monetary authorities could determine the real and not merely the nominal quantity of high-powered money. For example, William C. Brainard and James Tobin in setting up a financial model to illustrate pitfalls in the building of such models use “the replacement value of . . . physical assets . . . as the numeraire of the system,” yet regard “the supply of reserves” as “one of the quantities the central bank directly controls (“Pitfalls in Financial Model-Building,” AER, May 1968, pp. 101-02). If the nominal level of prices is regarded as an endogenous variable, this is clearly wrong. Hence the writers must be assuming this nominal level of prices to be fixed outsider their system. Keynes’ “wage unit” serves the same role in his analysis and leads him and his followers also to treat the monetary authorities as directly controlling real and not nominal variables.
But there is no logical necessity that requires the New View so elegantly formulated by Tobin to be deployed within a Keynesian framework rather than in a non-Keynesian framework in which some monetary aggregate, like the stock of currency or the monetary base, rather than M1 or M2, is what determines the price level. The stock of currency (or the monetary base) can function as the hot potato that determines (in conjunction with all the other variables affecting the demand for currency or the monetary base) the price level. Denying that bank money is a hot potato doesn’t require you to treat the price level “as an institutional datum.” Friedman, almost, but not quite, figured that one out.