Uneasy Money

In some recent posts (here, here, and here), I have expressed my admiration for James Tobin’s wonderful paper “Commercial Banks as Creators of Money.” Although the logic of the paper seems utterly compelling to me, it was, and evidently remains, controversial, because it advances the idea that the distinction between what is and what is not money is not a hard and fast one, but depends on the institutional and regulatory arrangements under which banks and other financial institutions operate.

The context in which I have been discussing Tobin is whether bank deposits created by commercial banks, unlike the liabilities created by other financial institutions, are “hot potatoes” in the sense that any individual can get rid of “unwanted” bank deposits (“unwanted” meaning that the individual’s deposit holdings exceed the amount that he would like to hold relative to his current income and wealth), but can do so only by giving those deposits to another individual in exchange for something else that the first individual would rather have instead. Thus, according to the “hot potato” view of the world, the total amount of bank deposits held by the community as a whole cannot be changed; the community is stuck with whatever amount of deposits that the banks, in the process of making loans, have created. What this means is that it is up to the monetary authority, through its control of reserves, reserve requirements, and the overnight rate on interbank loans (or its lending rate to the banks), to create the right amount of bank deposits.

Tobin argued that the “hot potato” view of bank deposits is incorrect, because the creation of bank deposits by commercial banks is not a mechanical process, as implied by the conventional money-multiplier analysis, in which bank deposits automatically come in to existence as a result of, and in proportion to, the amount of reserves provided to the banking system. Rather, profit-maximizing banks, like other profit-maximizing financial intermediaries, make an economic decision about how much of their liabilities to create (and on what terms to make them available) based on the public’s demand to hold those liabilities and the banks’ costs of backing those liabilities, inasmuch as any financial intermediary issuing a liability must make that liability sufficiently attractive to be willingly held by some member of the public. Like any financial intermediary, banks seeking to lend to the public can lend only if they obtain the funds to be lent from some source: either the equity of the owners or funds provided to the bank by lenders. If a bank cannot induce people to hold an amount deposits created in the process of making loans equal to the amount of loans it has made, it must find another source of capital to finance those loans.  Otherwise, the bank will be unable to sustain the scale of lending it has undertaken.

The amount of lending that banks undertake is governed by strict profit-loss calculations, just as the scale of operations of any profit-making enterprise is governed by a profit-loss calculation.  A bank makes a profit if the spread between its borrowing and lending rates is sufficient to cover its other costs of intermediation. And the scale of its lending depends on the demand of the public to hold its liabilities (which affects how much interest a bank must pay on those liabilities, e.g., demand deposits) and the interest rates that borrowers are willing to pay on the loans they obtain from the bank. Subject to essentially the same constraints on expansion as other financial intermediaries, banks do not simply lend without considering what costs they are incurring or what contribution to profits their loans are making at the margin. Because in equilibrium, the marginal revenue to a bank generated by an incremental loan just equals the marginal cost of making the loan, so that banks can decide to decrease their lending just as easily as they can decide to increase their lending, it makes no sense to think that bank-created money is, like central-bank currency, a hot potato. (Aside to Mike Sproul: Sorry, I know the previous sentence sounds like a nail scratching on a blackboard to you, but only one argument at a time.) As Tobin puts it:

The community cannot get rid of its currency supply; the economy must adjust until it is willingly absorbed. The “hot potato” analogy truly applies. For bank-created money, however, there is an economic mechanism of extinction as well as creation, contraction as well as expansion. If bank deposits are excessive relative to public preferences, they will tend to decline; otherwise banks will lose income. The burden of adaptation is not placed on the rest of the economy.

To me this seems so straightforward that I cannot understand why anyone would disagree. But some very smart people – I am thinking especially of Nick Rowe — are as convinced that it is wrong as I am that it is right. In earlier exchanges we have had about this, Nick has invoked the authority of Leland Yeager in arguing against Tobin, and Lee Kelly, in a comment on my recent post citing the favorable evaluation of Tobin’s paper given by Milton Friedman and Anna Schwartz in their Monetary Statistics of the US:  Estimates, Sources and Methods, a companion volume to their Monetary History of the US, specifically asked for my opinion of Leland Yeager’s paper “What Are Banks?”, a paper devoted entirely to criticism of Tobin’s paper.

I just re-read Yeager’s paper, and I still find it unpersuasive. There are too many problems with it to cover in a single blog post. But I will note in passing that in the discussion of definitions of money by Friedman and Schwartz that I mentioned in the previous paragraph, Friedman and Schwartz, while indicating agreement with the gist of Tobin’s argument, also criticized Yeager’s approach in his paper “The Essential Properties of a Medium of Exchange,” which foreshadowed much of the discussion in “What Are Banks?” So it is complicated. But I think that the key point is that, by failing to make explicit how he conceived that the price level and nominal income were determined, Tobin left his analysis incomplete, thereby allowing critics to charge that he had left the price level and nominal income undetermined or specified, in Keynesian style, by some ad hoc assumption. This misunderstanding is evident when Yeager makes the following criticism of Tobin and his followers:

Proponents of this view [Tobin et al.] are evidently not attributing “the natural economic limit” to limitation of base money and to a finite money multiplier, for that would be old stuff and not a new view. Those familiar limitations operate on the supply-of-money side, while the New Viewers emphasize limitations on the demand side. They deny crucial differences (to be explained below) between banking and NFI (non-bank financial institutions) systems regarding limits to the scales of their operations.

This paragraph is hopelessly, if not deliberately, confused. Arguing like a lawyer, Yeager attributes an absurd position to Tobin based on the label (New View) that Tobin et al. chose to describe their approach to analyzing the banking system, an approach based on economic incentives rather than the mindless mechanics of the money multiplier. Having thereby attributed an absurd position to Tobin, Yeager proceeds to discredit that absurd position, even though the quotation above in which Tobin distinguishes between currency and “bank-created money” demonstrates that he explicitly rejected the position attributed to him. And contrary to Yeager’s assertion, rejection of the money-multiplier as an analytical tool certainly was a New View, a view that Yeager spends most of his paper arguing against. The money multiplier is a reduced form embodying both demand (for reserves, by the banking system, and for currency, by the public) and supply; it is not a description of the profit-maximizing choices of banks to supply deposits at alternative prices, which is what we normally think of as a supply curve. To treat it as a conventional supply curve or as an analogue to a supply curve is just incoherent. And then, apparently oblivious to the mutual inconsistency of his criticisms, Yeager, in the paragraph’s final sentence, criticizes the New View for denying “crucial differences” between banks and non-bank financial institutions “regarding limits to the scales of their operations,” a denial completely independent of what New Viewers think about base money.

A few paragraphs later, Yeager launches into a diatribe against the New View based on the following thought experiment:

Suppose, then, that a cut in reserve requirements or expansion of the monetary base or shift of the public’s preferences from currency to deposits initially gives the banks more excess reserves.

To Yeager it seems obvious that this thought experiment demonstrates that the banks, flush with excess reserves, start lending and expand the money supply, raising nominal income and prices rise and thereby increasing the demand for money until the demand to hold money eventually equals the enlarged stock of nominal money stock. Yeager somehow thinks that this thought experiment proves that it is the banking system that has generated the inflationary process.

Even applied to the banking system as a whole, something is wrong with the idea that a decline in yields obtainable will check expansion of loans and investments and deposits. That idea overlooks Knut Wicksell’s cumulative process. As money expansion raises nominal incomes and prices, the dollar volume of loans demanded rises also, even at given interest rates. The proposition that the supply of money creates its own demand thus applies not only to cash balances . . . but also to money being newly supplied and demanded on loan. An unconstrained cumulative process can even lead to embodiment of inflationary expectations in interest rates as described by Irving Fisher. The great inflations of history discredit any notion of expansion being limited as marginal revenues fall in relation to marginal costs. The notion rests not only on an illegitimate imputation of a systemwide viewpoint to the individual banker, but also on a more or less tacit assumption of rigid prices.

The obvious point to make is that Yeager’s thought experiment postulates either an increase in the monetary base or a reduction in the demand (via a reduction in reserve requirements or via a shift in the public’s demand from holding currency to holding deposits) for the monetary base. Thus, according to the New View (or an appropriately modified New View), the equilibrium price level and level of nominal income must rise. The rise is the consequence of an excess supply of currency (monetary base) which Tobin explicitly acknowledges is a hot potato. End of story. Yeager’s outraged remonstrations about Wicksellian cumulative processes, inflationary expectations and the lessons of the great inflations of history are simply beside the point. Throughout the paper, Yeager accuses Tobin of being confused about the difference between the incentives for the banking system as a whole and those of the individual banker, but confusion here is Yeager’s.

One might say that the approach that I follow:  to determine the price level in terms of the demand for currency and the stock of currency, while allowing the quantity of bank money and its yield (i.e. the interest rate paid by banks to depositors) to be determined in terms of the demand for deposits and the cost of supplying deposits, leads to exactly the same conclusion one reaches via the traditional money multiplier analysis: the price level and nominal income go up and the quantity of bank deposits increases. But actually there are some subtle differences. According to Yeager, it is the increase in the quantity of bank deposits that generates the increase in prices and nominal income. That would seem to imply that the quantity of bank money should be rising more rapidly than the quantity of liabilities produced by non-bank financial institutions. The money-multiplier analysis also implies no change in interest rates paid to depositors, while, if inflation leads to increased inflationary expectations and increased nominal interest rates, the New View predicts that the interest paid on deposits would rise as well. According to the New View, there would be no change in the relative quantities of bank and non-bank deposit liabilities; according to the money-multiplier analysis, the amount of bank deposits should rise faster than the amount of non-bank deposit liabilities even though non-banks would be increasing the rate of interest paid on their deposit liabilities while banks did not increase the interest paid on their deposit liabilities.

In connection with the “more or less tacit assumption of rigid prices” that he attributes to Tobin, Yeager adds the following footnote:

Basil Moore, who wavers between the new and traditional views, recognizes that if all prices were perfectly and instantly flexible, an unregulated banking system could not reach a stable equilibrium.

This is just wrong; a stable equilibrium is assured by fixing the nominal quantity of currency even with a completely unregulated banking system. What can possibly be meant by “perfect and instant price flexibility” is simply a mystery to me. The only meaning that I can possibly attach to it is that full equilibrium is continuously maintained, with no trading at disequilibrium prices. Why an equilibrium with a fiat currency and a determinate price level is inconsistent with an unregulated banking system is not explained, nor could it be.

The same confusion emerges again in Yeager’s discussion of how the supply of money creates its own demand. Yeager writes:

This process that reconciles the demand for money with the supply is the theme of what J. M. Keynes called “the fundamental proposition of monetary theory” and Milton Friedman called “the most important proposition in monetary theory.” Briefly, everyone can individually hold as much or as little money as he effectively demands, even though the total of all holdings may be exogenously set; for the total flow of spending adjusts in such a way that the demand for nominal money becomes equal to the exogenous supply.

Of course, the adjustment process can work perfectly well for currency and be irrelevant for bank money; currency is a hot potato, while bank money is not, so that the nominal supply of bank money, unlike the nominal stock of currency, adjusts to the nominal demand.

In a footnote to the above passage, Yeager invokes the authority of Harry Johnson:

Harry Johnson has charged opponents of monetarism with confusion over how nominal and real quantities of money are determined and with “a tendency to discuss monetary problems as if nominal and real money balances are the same thing, and as if ordinary value theory could be applied to the behaviour of money.” The Yale theorists [i.e., Tobin et al.] “are . . . alert to this confusion but by-pass it either by assuming stable prices and confining their analysis to the financial sector, or by building models based on the fictional construction of a money whose purchasing power is fixed in real terms, thereby avoiding confusion in the analysis at the expense of creating it with respect to the applicability of the results.

Any confusion Johnson may have detected could be eliminated simply distinguishing between the price-level analysis carried out in terms of the demand for, and the stock of, currency on the one hand, and the analysis, given the price level determined by the demand for and stock of currency, of the quantity (in both real and nominal terms) of bank money and the competitive interest paid by banks on deposits. In his subsequent work on the monetary approach to the balance of payments, Johnson showed himself to be perfectly content to decompose the analysis in this fashion.

There is more to say about Yeager’s paper; perhaps I will come back to it on another occasion. But any further comments are unlikely to be much more favorable than those above.

During my little vacation recently from writing about monetary policy, it seems that there has been quite a dust-up about endogenous money in econo-blogosphere. It all started with a post by Steve Keen, an Australian economist of the post-Keynesian persuasion, in which he expounded at length the greatness of Hyman Minsky, the irrelevance of equilibrium to macroeconomic problems, the endogeneity of the money supply, and the critical importance of debt in explaining macroeconomic fluctuations. In making his argument, Keen used as a foil a paper by Krugman and Eggerston “Debt, Delevereging, and the Liquidity Trap: A Fisher-Minsky-Koo Approach,” which he ridiculed for its excessive attachment to wrong-headed neoclassicism, as exemplified in the DSGE model in which Krugman and Eggerston conducted their analysis. I can’t help but note parenthetically that I was astounded by the following sentence in Keen’s post.

There are so many ways in which neoclassical economists misinterpret non-neoclassical thinkers like Fisher and Minsky that I could write a book on the topic.

No doubt that it would be a fascinating book, but what would be even more fascinating would be an, explanation of how Irving Fisher – yes, that Irving Fisher – could possibly be considered as anything other than a neo-classical economist.

At any rate, this assault did not go unnoticed by Dr. Krugman, who responded with evident annoyance on his blog, focusing in particular on the question whether a useful macroeconomic model requires an explicit model of the banking system, as Keen asserted, or whether a simple assumption that the monetary authority can exercise sufficient control over the banking system makes an explicit model of the banking sector unnecessary, as Krugman, following the analysis of the General Theory, asserted. Sorry, but I can’t resist making another parenthetical observation. Post-Keynesians, following Joan Robinson, rarely miss an opportunity to dismiss the IS-LM model as an inauthentic and misleading transformation of the richer analysis of the General Theory. Yet, the IS-LM model’s assumption of a fixed nominal quantity of money determined by the monetary authority was taken straight from the General Theory, a point made by, among others, Jacques Rueff in his 1948 critique of the General Theory and the liquidity-preference theory of interest, and by G.L.S. Shackle in his writings on Keynes, e.g., The Years of High Theory. Thus, in arguing for an endogenous model of the money supply, it is the anti-IS-LM post-Keynesians who are departing from Keynes’s analysis in the GT.

Krugman’s dismissive response to Keen, focusing on the endogeneity issue, elicited a stinging rejoinder, followed by several further rounds of heated argument. In the meantime, Nick Rowe joined the fray, writing at least three posts on the subject (1, 2, 3) generally siding with Krugman, as did Scott Fullwiler and Randall Wray, two leading lights of what has come to be known as Modern Monetary Theory (MMT), siding with Keen. Further discussion and commentary was provided by Steve Randy Waldman and Scott Sumner, and summaries by Edward Harrison, John Carney, Unlearning Economics, and Business Insider.

In reading through the voluminous posts, I found myself pulled in both directions. Some readers may recall that I got into a bit of a controversy with Nick Rowe some months back over the endogeneity issue, when Nick asserted that any increase in the quantity of bank money is a hot potato. Thus, if banks create more money than the public want to hold, the disequilibrium cannot be eliminated by a withdrawal of the excess money, rather the money must be passed from hand to hand, generating additional money income until the resulting increase in the demand to hold money eliminates the disequilibrium between the demand for money and the amount in existence. I argued that Nick had this all wrong, because banks can destroy, as well as create, money. Citing James Tobin’s classic article “Commercial Banks as Creators of Money,” I argued that responding to the interest-rate spreads between various lending and deposit rates, profit-maximizing banks have economic incentives to create only as much money as the public is willing to hold, no more and no less. Any disequilibrium between the amount of money in existence and the amount the public wants to hold can be eliminated either by a change (positive or negative) in the quantity of money or by a change in the deposit rates necessary to induce the pubic to hold the amount of money in existence.

The idea stressed by Keen, Fullwiler and Wray, that banks don’t lend out deposits and hold reserves against their deposits, but create deposits in the course of lending and hold reserves only insofar as reserves offer some pecuniary or non-pecuniary yield is an idea to which I fully subscribe. They think that the money multiplier is a nonsensical concept, and so do I. I was actually encouraged to see that Nick Rowe now appears willing to accept that this is the right way to think about how banks operate, and that because banks are committed to convert their liabilities into currency on demand, they cannot create more liabilities than the public is willing to hold unless they are prepared to suffer losses as a consequence.

But Keen, Fullwiler and Wray go a step further, which is to say that, since banks can create money out of thin air, there is no limit to their ability to create money. I don’t understand this point. Do they mean that banks are in a perpetual state of disequilibrium? I understand that they are uncomfortable with any notion of equilibrium, but all other profit-maximizing firms can be said to be subject to some limit, not necessarily a physical or quantitative limit, but an economic limit to their expansion. Tobin, in his classic article, was very clear that banks do not have an incentive to create unlimited quantities of deposits. At any moment, a bank must perceive that there is a point beyond which it would be unprofitable to expand (by making additional loans and creating additional deposits) its balance sheet further.

Fullwiler argues at length that it makes no sense to speak about reserves or currency as setting any sort of constraint on the expansion of the banking system, ridiculing the notion that any bank is prevented from expanding by an inability to obtain additional reserves or additional currency should it want to do so. But banks are not constrained by any quantitative limit; they are constrained by the economic environment in which they operate and the incentives associated with the goal of maximizing profit. And that goal depends critically on the current and expected future price level, and on current lending and deposit rates. The current and expected future price level are controlled (or, at least, one may coherently hypothesize that they are controlled) by the central bank which controls the quantity of currency and the monetary base. Fullwiler denies that the central bank can control the quantity of currency or the monetary base, because the central bank is obligated to accommodate any demand for currency and to provide sufficient reserves to ensure that the payment system does not break down. But in any highly organized, efficiently managed market, transactors are able to buy and sell as much as they want to at the prevailing market price.  So the mere fact that there are no frustrated demands for currency or reserves cannot prove that the central bank does not have the power to affect the value of currency. That would be like saying that the government could not affect the value of a domestically produced, internationally traded, commodity by applying a tariff on imports, but could do so only by imposing an import quota. Applying a tariff and imposing a quota are, in principle (with full knowledge of the relevant supply and demand curves), equivalent methods of raising the price of a commodity. However, in the absence of the requisite knowledge, if fluctuations in price would be more disruptive than fluctuations in quantity, the tariff is a better way to raise the price of the commodity than a numerical quota on imports.

So while I think that bank money is endogenous, I don’t believe that the quantity of base money or currency is endogenous in the sense that the central bank is powerless to control the price level. The central bank may not be trying to target a particular quantity of currency or of the monetary base, but it can target a price level by varying its lending rate or by taking steps to vary the interbank overnight rate on bank reserves. This, it seems to me, is not very different from trying to control the domestic value of an imported commodity by setting a tariff on imports rather than controlling the quantity of imports directly.  Endogeneity of bank money does not necessarily mean that a central bank cannot control the price level.  If it can, I am not so sure that the post-Keynesian, MMT critique of more conventional macroeconomics is quite as powerful as they seem to think.