Archive for the 'Knut Wicksell' Category

Thinking about Interest and Irving Fisher

In two recent posts I have discussed Keynes’s theory of interest and the natural rate of interest. My goal in both posts was not to give my own view of the correct way to think about what determines interest rates,  but to identify and highlight problems with Keynes’s liquidity-preference theory of interest, and with the concept of a natural rate of interest. The main point that I wanted to make about Keynes’s liquidity-preference theory was that although Keynes thought that he was explaining – or perhaps, explicating — the rate of interest, his theory was nothing more than an explanation of why, typically, the nominal pecuniary yield on holding cash is less than the nominal yield on holding real assets, the difference in yield being attributable to the liquidity services derived from holding a maximally liquid asset rather than holding an imperfectly liquid asset. Unfortunately, Keynes imagined that by identifying and explaining the liquidity premium on cash, he had thereby explained the real yield on holding physical capital assets; he did nothing of the kind, as the marvelous exposition of the theory of own rates of interest in chapter 17 of the General Theory unwittingly demonstrates.

For expository purposes, I followed Keynes in contrasting his liquidity-preference theory with what he called the classical theory of interest, which he identified with Alfred Marshall, in which the rate of interest is supposed to be the rate that equilibrates saving and investment. I criticized Keynes for attributing this theory to Marshall rather than to Irving Fisher, which was, I am now inclined to think, a mistake on my part, because I doubt, based on a quick examination of Fisher’s two great books The Rate of Interest and The Theory of Interest, that he ever asserted that the rate of interest is determined by equilibrating savings and investment. (I actually don’t know if Marshall did or did make such an assertion.) But I think it’s clear that Fisher did not formulate his theory in terms of equating investment and savings via adjustments in the rate of interest rate. Fisher, I think, did agree (but I can’t quote a passage to this effect) that savings and investment are equal in equilibrium, but his analysis of the determination of the rate of interest was not undertaken in terms of equalizing two flows, i.e., savings and investment. Instead the analysis was carried out in terms of individual or household decisions about how much to consume out of current and expected future income, and in terms of decisions by business firms about how much available resources to devote to producing output for current consumption versus producing for future consumption. Fisher showed (in Walrasian fashion) that there are exactly enough equations in his system to solve for all the independent variables, so that his system had a solution. (That Walrasian argument of counting equations and unknowns is mathematically flawed, but later work by my cousin Abraham Wald and subsequently by Arrow, Debreu and McKenzie showed that Fisher’s claim could, under some more or less plausible assumptions, be proved in a mathematically rigorous way.)

Maybe it was Knut Wicksell who in his discussions of the determination of the rate of interest argued that the rate of interest is responsible for equalizing savings and investment, but that was not how Fisher understood what the rate of interest is all about. The Wicksellian notion that the equilibrium rate of interest equalizes savings and investment was thus a misunderstanding of the Fisherian theory, and it would be a worthwhile endeavor to trace the genesis and subsequent development of this misunderstanding to the point that Keynes and his contemporaries could have thought that they were giving an accurate representation of what orthodox theory asserted when they claimed that according to orthodox theory the rate of interest is what ensures equality between savings and investment.

This mistaken doctrine was formalized as the loanable-funds theory of interest – I believe that Dennis Robertson is usually credited with originating this term — in which savings is represented as the supply of loanable funds and investment is represented as the demand for loanable funds, with the rate of interest serving as a sort of price that is determined in Marshallian fashion by the intersection of the two schedules. Somehow it became accepted that the loanable-funds doctrine is the orthodox theory of interest determination, but it is clear from Fisher and from standard expositions of the neoclassical theory of interest which are of course simply extensions of Fisher’s work) that the loanable-funds theory is mistaken and misguided at a very basic level. (At this point, I should credit George Blackford for his comments on my post about Keynes’s theory of the rate of interest for helping me realize that it is not possible to make any sense out of the loanable-funds theory even though I am not sure that we agree on exactly why the loanable funds theory doesn’t make sense. Not that I had espoused the loanable-funds theory, but I did not fully appreciate its incoherence.)

Why do I say that the loanable-funds theory is mistaken and incoherent? Simply because it is fundamentally inconsistent with the essential properties of general-equilibrium analysis. In general-equilibrium analysis, interest rates emerge not as a separate subset of prices determined in a corresponding subset of markets; they emerge from the intertemporal relationships between and across all asset markets and asset prices. To view the rate of interest as being determined in a separate market for loanable funds as if the rate of interest were not being simultaneously determined in all asset markets is a complete misunderstanding of the theory of intertemporal general equilibrium.

Here’s how Fisher put over a century ago in The Rate of Interest:

We thus need to distinguish between interest in terms of money and interest in terms of goods. The first thought suggested by this fact is that the rate of interest in money is “nominal” and that in goods “real.” But this distinction is not sufficient, for no two forms of goods maintain or are expected to maintain, a constant price ratio toward each other. There are therefore just as many rates of interest in goods as there are forms of goods diverging in value. (p. 84, Fisher’s emphasis).

So a quarter of a century before Sraffa supposedly introduced the idea of own rates of interest in his 1932 review of Hayek’s Prices and Production, Fisher had done so in his first classic treatise on interest, which reproduced the own-rate analysis in his 1896 monograph Appreciation and Interest. While crediting Sraffa for introducing the concept of own rates of interest, Keynes, in chapter 17, simply — and brilliantly extends the basics of Fisher’s own-rate analysis, incorporating the idea of liquidity preference and silently correcting Sraffa insofar as his analysis departed from Fisher’s.

Christopher Bliss in his own classic treatise on the theory of interest, expands upon Fisher’s point.

According to equilibrium theory – according indeed to any theory of economic action which relates firms’ decisions to prospective profit and households’ decisions to budget-constrained searches for the most preferred combination of goods – it is prices which play the fundamental role. This is because prices provide the weights to be attached to the possible amendments to their net supply plans which the actors have implicitly rejected in deciding upon their choices. In an intertemporal economy it is then, naturally, present-value prices which play the fundamental role. Although this argument is mounted here on the basis of a consideration of an economy with forward markets in intertemporal equilibrium, it in no way depends on this particular foundation. As has been remarked, if forward markets are not in operation the economic actors have no choice but to substitute their “guesses” for the firm quotations of the forward markets. This will make a big difference, since full intertemporal equilibrium is not likely to be achieved unless there is a mechanism to check and correct for inconsistency in plans and expectations. But the forces that pull economic decisions one way or another are present-value prices . . . be they guesses or firm quotations. (pp. 55-56)

Changes in time preference therefore cause immediate changes in the present value prices of assets thereby causing corresponding changes in own rates of interest. Changes in own rates of interest constrain the rates of interest charged on money loans; changes in asset valuations and interest rates induce changes in production, consumption plans and the rate at which new assets are produced and capital accumulated. The notion that there is ever a separate market for loanable funds in which the rate of interest is somehow determined, and savings and investment are somehow equilibrated is simply inconsistent with the basic Fisherian theory of the rate of interest.

Just as Nick Rowe argues that there is no single market in which the exchange value of money (medium of account) is determined, because money is exchanged for goods in all markets, there can be no single market in which the rate of interest is determined because the value of every asset depends on the rate of interest at which the expected income or service-flow derived from the asset is discounted. The determination of the rate of interest can’t be confined to a single market.


The Well-Defined, but Nearly Useless, Natural Rate of Interest

Tyler Cowen recently posted a diatribe against the idea monetary policy should be conducted by setting the interest rate target of the central bank at or near the natural rate of interest. Tyler’s post elicited critical responses from Brad DeLong and Paul Krugman among others. I sympathize with Tyler’s impatience with the natural rate of interest as a guide to policy, but I think the scattershot approach he took in listing, seemingly at random, seven complaints against the natural rate of interest was not the best way to register dissatisfaction with the natural rate. Here’s Tyler’s list of seven complaints.

1 Knut Wicksell, inventor of the term “natural rate of interest,” argued that if the central bank set its target rate equal to the natural rate, it would avoid inflation and deflation and tame the business cycle. Wicksell’s argument was criticized by his friend and countryman David Davidson who pointed out that, with rising productivity, price stability would not result without monetary expansion, which would require the monetary authority to reduce its target rate of interest below the natural rate to induce enough investment to be financed by monetary expansion. Thus, when productivity is rising, setting the target rate of interest equal to the natural rate leads not to price stability, but to deflation.

2 Keynes rejected the natural rate as a criterion for monetary policy, because the natural rate is not unique. The natural rate varies with the level of income and employment.

3 Early Keynesians like Hicks, Hansen, and Modigliani rejected the natural rate as well.

4 The meaning of the natural rate has changed; it was once the rate that would result in a stable price level; now it’s the rate that results in a stable rate of inflation.

5 Friedman also rejected the natural rate because there is no guarantee that setting the target rate equal to the natural rate will result in the rate of money growth that Freidman believed was desirable.

6 Sraffa debunked the natural rate in his 1932 review of Hayek’s Prices and Production.

7 It seems implausible that the natural rate is now negative, as many exponents of the natural rate concept now claim, even though the economy is growing and the marginal productivity of capital is positive.

Let me try to tidy all this up a bit.

The first thing you need to know when thinking about the natural rate is that, like so much else in economics, you will become hopelessly confused if you don’t keep the Fisher equation, which decomposes the nominal rate of interest into the real rate of interest and the expected rate of inflation, in clear sight. Once you begin thinking about the natural rate in the context of the Fisher equation, it becomes obvious that the natural rate can be thought of coherently as either a real rate or a nominal rate, but the moment you are unclear about whether you are talking about a real natural rate or a nominal natural rate, you’re finished.

Thus, Wicksell was implicitly thinking about a situation in which expected inflation is zero so that the real and nominal natural rates coincide. If the rate of inflation is correctly expected to be zero, and the increase in productivity is also correctly expected, the increase in the quantity of money required to sustain a constant price level can be induced by the payment of interest on cash balances. Alternatively, if the payment of interest on cash balances is ruled out, the rate of capital accumulation (forced savings) could be increased sufficiently to cause the real natural interest rate under a constant price level to fall below the real natural interest rate under deflation.

In the Sraffa-Hayek episode, as Paul Zimmerman and I have shown in our paper on that topic, Sraffa failed to understand that the multiplicity of own rates of interest in a pure barter economy did not mean that there was not a unique real natural rate toward which arbitrage would force all the individual own rates to converge. At any moment, therefore, there is a unique real natural rate in a barter economy if arbitrage is operating to equalize the cost of borrowing in terms of every commodity. Moreover, even Sraffa did not dispute that, under Wicksell’s definition of the natural rate as the rate consistent with a stable price level, there is a unique natural rate. Sraffa’s quarrel was only with Hayek’s use of the natural rate, inasmuch as Hayek maintained that the natural rate did not imply a stable price level. Of course, Hayek was caught in a contradiction that Sraffa overlooked, because he identified the real natural rate with an equal nominal rate, so that he was implicitly assuming a constant expected price level even as he was arguing that the neutral monetary policy corresponding to setting the market interest rate equal to the natural rate would imply deflation when productivity was increasing.

I am inclined to be critical Milton Friedman about many aspects of his monetary thought, but one of his virtues as a monetary economist was that he consistently emphasized Fisher’s  distinction between real and nominal interest rates. The point that Friedman was making in the passage quoted by Tyler was that the monetary authority is able to peg nominal variables, prices, inflation, exchange rates, but not real variables, like employment, output, or interest rates. Even pegging the nominal natural rate is impossible, because inasmuch as the goal of targeting a nominal natural rate is to stabilize the rate of inflation, targeting the nominal natural rate also means targeting the real natural rate. But targeting the real natural rate is not possible, and trying to do so will just get you into trouble.

So Tyler should not be complaining about the change in the meaning of the natural rate; that change simply reflects the gradual penetration of the Fisher equation into the consciousness of the economics profession. We now realize that, given the real natural rate, there is, for every expected rate of inflation, a corresponding nominal natural rate.

Keynes made a very different contribution to our understanding of the natural rate. He was that there is no reason to assume that the real natural rate of interest is unique. True, at any moment there is some real natural rate toward which arbitrage is forcing all nominal rates to converge. But that real natural rate is a function of the prevailing economic conditions. Keynes believed that there are multiple equilibria, each corresponding to a different level of employment, and that associated with each of those equilibria there could be a different real natural rate. Nowadays, we are less inclined than was Keynes to call an underemployment situation an equilibrium, but there is still no reason to assume that the real natural rate that serves as an attractor for all nominal rates is independent of the state of the economy. If the real natural rate for an underperforming economy is less than the real natural rate that would be associated with the economy if it were in the neighborhood of an optimal equilibrium, there is no reason why either the real natural rate corresponding to an optimal equilibrium or the real natural rate corresponding to the current sub-optimal state of economy should be the policy rate that the monetary authority chooses as its target.

Finally, what can be said about Tyler’s point that it is implausible to suggest that the real natural rate is negative when the economy is growing (even slowly) and the marginal productivity of capital is positive? Two points.

First, the marginal productivity of gold is very close to zero. If gold is held as bullion, it is being held for expected appreciation over and above the cost of storage. So the ratio of the future price of gold to the spot price of gold should equal one plus the real rate of interest. If you look at futures prices for gold you will see that they are virtually the same as the spot price. However, storing gold is not costless. According to this article on, storage costs for gold range between 0.5 to 1% of the value of gold, implying that expected rate of return to holding gold is now less than -0.5% a year, which means that the marginal productivity of real capital is negative. Sure there are plenty of investments out there that are generating positive returns, but those are inframarginal investments. Those inframarginal investments are generating some net gain in productivity, and overall economic growth is positive, but that doesn’t mean that the return on investment at the margin is positive. At the margin, the yield on real capital seems to be negative.

If, as appears likely, our economy is underperforming, estimates of the real natural rate of interest are not necessarily an appropriate guide for the monetary authority in choosing its target rate of interest. If the aim of monetary policy is to nudge the economy onto a feasible growth path that is above the sub-optimal path along which it is currently moving, it might well be that the appropriate interest-rate target, as long as the economy remains below its optimal growth path, would be less than the natural rate corresponding to the current sub-optimal growth path.

John Cochrane Explains Neo-Fisherism

In a recent post, John Cochrane, responding to an earlier post by Nick Rowe about Neo-Fisherism, has tried to explain why raising interest rates could plausibly cause inflation to rise and reducing interest rates could plausibly cause inflation to fall, even though almost everyone, including central bankers, seems to think that when central banks raise interest rates, inflation falls, and when they reduce interest rates, inflation goes up.

In his explanation, Cochrane concedes that there is an immediate short-term tendency for increased interest rates to reduce inflation and for reduced interest rates to raise inflation, but he also argues that these effects (liquidity effects in Keynesian terminology) are transitory and would be dominated by the Fisher effects if the central bank committed itself to a permanent change in its interest-rate target. Of course, the proviso that the central bank commit itself to a permanent interest-rate peg is a pretty important qualification to the Neo-Fisherian position, because few central banks have ever committed themselves to a permanent interest-rate peg, the most famous attempt (by the Fed after World War II) to peg an interest rate having led to accelerating inflation during the Korean War, thereby forcing the peg to be abandoned, in apparent contradiction of the Neo-Fisherian view.

However, Cochrane does try to reconcile the Neo-Fisherian view with the standard view that raising interest rates reduces inflation and reducing interest rates increases inflation. He suggests that the standard view is strictly a short-run relationship and that the way to target inflation over the long-run is simply to target an interest rate consistent with the desired rate of inflation, and to rely on the Fisher equation to generate the actual and expected rate of inflation corresponding to that nominal rate. Here’s how Cochrane puts it:

We can put the issue more generally as, if the central bank does nothing to interest rates, is the economy stable or unstable following a shock to inflation?

For the next set of graphs, I imagine a shock to inflation, illustrated as the little upward sloping arrow on the left. Usually, the Fed responds by raising interest rates. What if it doesn’t?  A pure neo-Fisherian view would say inflation will come back on its own.


Again, we don’t have to be that pure.

The milder view allows there may be some short run dynamics; the lower real rates might lead to some persistence in inflation. But even if the Fed does nothing, eventually real interest rates have to settle down to their “natural” level, and inflation will come back. Mabye not as fast as it would if the Fed had aggressively tamed it, but eventually.


By contrast, the standard view says that inflation is unstable. If the Fed does not raise rates, inflation will eventually careen off following the shock.


Now this really confuses me. What does a shock to inflation mean? From the context, Cochrane seems to be thinking that something happens to raise the rate of inflation in the short run, but the persistence of increased inflation somehow depends on an underlying assumption about whether the economy is stable or unstable. Cochrane doesn’t tell us what kind of shock to inflation he is talking about, and I can imagine only two possibilities, either a nominal shock or a real shock.

Let’s say it’s a nominal shock. What kind of nominal shock might Cochrane have in mind? An increase in the money supply? Well, presumably an increase in the money supply would cause an increase in the price level, and a temporary increase in the rate of inflation, but if the increase in the money supply is a once-and-for-all increase, the system must revert, after a temporary increase, back to the old rate of inflation. Or maybe, Cochrane is thinking of a permanent increase in the rate of growth in the money supply. But in that case, why would the rate of inflation come back on its own as Cochrane suggests it would? Well, maybe it’s not the money supply but money demand that’s changing. But again, one would normally assume that an appropriate change in central-bank policy could cope with such a scenario and stabilize the rate of inflation.

Alright, then, let’s say it’s a real shock. Suppose some real event happens that raises the rate of inflation. Well, like what? A supply shock? That raises the rate of inflation, but since when is the standard view that the appropriate response by the central bank to a negative supply shock is to raise the interest-rate target? Perhaps Cochrane is talking about a real shock that reduces the real rate of interest. Well, in that case, the rate of inflation would certainly rise if the central bank maintained its nominal-interest-rate target, but the increase in inflation would not be temporary unless the real shock was temporary. If the real shock is temporary, it is not clear why the standard view would recommend that the central bank raise its target rate of interest. So, I am sorry, but I am still confused.

Now, the standard view that Cochrane is disputing is actually derived from Wicksell, and Wicksell’s cycle theory is in fact based on the assumption that the central bank keeps its target interest rate fixed while the natural rate fluctuates. (This, by the way, was also Hayek’s assumption in his first exposition of his theory in Monetary Theory and the Trade Cycle.) When the natural rate rises above the central bank’s target rate, a cumulative inflationary process starts, because borrowing from the banking system to finance investment is profitable as long as the expected return on investment exceeds the interest rate on loans charged by the banks. (This is where Hayek departed from Wicksell, focusing on Cantillon Effects instead of price-level effects.) Cochrane avoids that messy scenario, as far as I can tell, by assuming that the initial position is one in which the Fisher equation holds with the nominal rate equal to the real plus the expected rate of inflation and with expected inflation equal to actual inflation, and then positing an (as far as I can tell) unexplained inflation shock, with no change to the real rate (meaning, in Cochrane’s terminology, that the economy is stable). If the unexplained inflation shock goes away, the system must return to its initial equilibrium with expected inflation equal to actual inflation and the nominal rate equal to the real rate plus inflation.

In contrast, the Wicksellian assumption is that the real rate fluctuates with the nominal rate and expected inflation unchanged. Unless the central bank raises the nominal rate, the difference between the profit rate anticipated by entrepreneurs and the rate at which they can borrow causes the rate of inflation to increase. So it does not seem to me that Cochrane has in any way reconciled the Neo-Fisherian view with the standard view (or at least the Wicksellian version of the standard view).

PS I would just note that I have explained in my paper on Ricardo and Thornton why the Wicksellian analysis (anticipated almost a century before Wicksell by Henry Thornton) is defective (basically because he failed to take into account the law of reflux), but Cochrane, as far as I can tell, seems to be making a completely different point in his discussion.

Hawtrey’s Good and Bad Trade: Part II

Here I am again back at you finally with another installment in my series on Hawtrey’s Good and Bad Trade. In my first installment I provided some background on Hawtrey and a quick overview of the book, including a mention of the interesting fact (brought to my attention by David Laidler) that Hawtrey used the term “effective demand” in pretty much the same way that Keynes, some 20 years later, would use it in the General Theory.

In this post, I want to discuss what I consider the highlights of the first six chapters. The first chapter is a general introduction to the entire volume laying out the basic premise of the book, which is that the business cycle, understood as recurring fluctuations in the level of employment, is the result of monetary disturbances that lead to alternating phases of expansion and contraction. It is relatively easy for workers to find employment in expansions, but more difficult to do so in contractions. From the standpoint of the theory of economic equilibrium, the close correlation between employment and nominal income over the business cycle is somewhat paradoxical, because, according to the equilibrium theory, the allocation of resources is governed by relative, not absolute, prices. In the theory of equilibrium, a proportional increase or decrease in all prices should have no effect on employment. To explain the paradox, Hawtrey relies on the rigidity of some prices, and especially wages, an empirical fact that, Hawtrey believed, was an essential aspect of any economic system, and a necessary condition for the cyclicality of output and employment.

In Hawtrey’s view, economic expansions and contractions are caused by variations in effective demand, which he defines as total money income. (For reasons I discussed about a year and a half ago, I prefer to define “effective demand” as total money expenditure.) What determines effective demand, according to Hawtrey, is the relationship between the amount of money people are holding and the amount that they would, on average over time, like to hold. The way to think about the amount of money that people would like to hold is to imagine that there is some proportion of their annual income that people aim to hold in the form of cash.

The relationship between the amount of cash being held and the amount that people would like to hold depends on the nature of the monetary system. Hawtrey considers two types of monetary system: one type (discussed in chapter 2) is a pure fiat money system in which all money is issued by government; the other (discussed in chapter 3) is a credit system in which money is also created by banks by promising to redeem, on demand, their obligations (either deposits or negotiable banknotes) for fiat money. Credit money is issued by banks in exchange for a variety of assets, usually the untraded IOUs of borrowers.

In a pure fiat money system, effective demand depends chiefly on the amount of fiat money that people want to hold and on the amount of fiat money created by the government, fiat money being the only money available. A pure fiat money system, Hawtrey understood, was just the sort of system in which the propositions of the quantity theory of money would obtain at least in the medium to long run.

[I]f the adjustment [to a reduction in the quantity of money] could be made entirely by a suitable diminution of wages and salaries, accompanied by a corresponding diminution of prices, the commercial community could be placed forthwith in a new position of equilibrium, in which the output would continue unchanged, and distribution would only be modified by the apportionment of a somewhat larger share of the national product to the possessors of interest, rent, and other kinds of fixed incomes. In fact, the change in the circulating medium is merely a change in the machinery of distribution, and a change, moreover, which, once made, does not impair the effectiveness of that machinery. If the habits of the community are adapted without delay to the change, the production of wealth will continue unabated. If customary prices resist the change, the adjustment, which is bound to come sooner or later, will only be forced upon the people by the pressure of distress. (p. 41)

In a fiat money system, if the public have less money than they would like to hold their only recourse is to attempt to reduce their expenditures relative to their receipts, either offering more in exchange, which tends to depress prices or reducing their purchases, making it that much more difficult for anyone to increase sales except by reducing prices. The problem is that in a fiat system the amount of money is what it is, so that if one person manages to increase his holdings of money by increasing sales relative to purchases, his increase in cash balances must have be gained at the expense of someone else. With a fixed amount of fiat money in existence, the public as a whole cannot increase their holdings of cash, so equilibrium can be restored only by reducing the quantity of money demanded. But the reduction in the amount of money that people want to hold cannot occur unless income in money terms goes down. Money income can go down only if total output in real terms, or if the price level, falls. With nominal income down, people, wanting to hold some particular share of their nominal income in the form of money, will be content with a smaller cash balance than they were before, and will stop trying to increase their cash balances by cutting their expenditure. Because some prices — and especially wages — tend to be sticky, Hawtrey felt that it was inevitable that the adjustment to reduction in the amount of fiat money would cause both real income and prices to fall.

Although Hawtrey correctly perceived that the simple quantity theory would not, even in theory, hold precisely for a credit system, his analysis of the credit system was incomplete inasmuch as he did not fully take into account the factors governing the public’s choice between holding credit money as opposed to fiat money or the incentives of the banking system to create credit money. That theory was not worked out till James Tobin did so 50 years later (another important anniversary worthy of note), though John Fullarton made an impressive start in his great work on the subject in 1844, a work Hawtrey must have been familiar with, but, to my knowledge, never discussed in detail.

In such a banking system there is no necessary connexion between the total of the deposits and the amount of coin which has been paid to the banks. A banker may at any time grant a customer a loan by simply adding to the balance standing to the customer’s credit in the books of the bank. No cash passes, but the customer acquires the right, during the currency of the loan, to draw cheques on the bank up to the amount lent. When the period of the loan expires, if the customer has a large enough balance to his credit, the loan can be repaid without any cash being employed, the amount of the loan being simply deducted from the balance. So long as the loan is outstanding it represents a clear addition to the available stock of “money,” in the sense of purchasing power. It is “money” in the the sense which will play, in a community possessing banks, the same part as money in the stricter sense of legal tender currency would play in the fictitious bankless community whose commercial conditions we previously have been considering. This is the most distinctive feature of the banking system, that between the stock of legal tender currency and the trading community there is interposed an intermediary, the banker, who can, if he wishes, create money out of nothing. (PP. 56-57)

This formulation is incomplete, inasmuch as it leaves the decision of the banker about how much money to create unconstrained by the usual forces of marginal revenue and marginal cost that supposedly determine the decisions of other profit-seeking businessmen. Hawtrey is not oblivious to the problem, but does not advance the analysis as far as he might have.

We have now to find out how this functionary uses his power and under what limitations he works. Something has already been said of the contingencies for which he must provide. Whenever he grants a loan and thereby creates money, he must expect a certain portion of this money to be applied sooner or later, to purposes for which legal tender currency is necessary. Sums will be drawn out from time to time to be spent either in wages or in small purchases, and the currency so applied will take a little time to find its way back to the banks. Large purchases will be paid for by cheque, involving a mere transfer of credit from one banking account to another, but the recipient of the cheque may wish to apply it ot the payment of wages, etc. Thus the principal limitation upon the banker’s freedom to create money is that he must have a reserve to meet the fresh demands for cash to which the creation of new money may lead. (Id.)

This is a very narrow view, apparently assuming that there is but one banker and that the only drain on the reserves of the banker is the withdrawal of currency by depositors. The possibility that recipients of cheques drawn on one bank may prefer to hold those funds in a different bank so that the bank must pay a competitive rate of interest on its deposits to induce its deposits to be held rather than those of another bank is not considered.

In trade a seller encourages or discourages buyers by lowering or raising his prices. So a banker encourages or discourages borrowers by lowering or raising the rate of interest. (p.58)

Again, Hawtrey only saw half the picture. The banker is setting two rates: the rate that he charges borrowers and the rate that he pays to depositors. It is the spread between those two rates that determines the marginal revenue from creating another dollar of deposits. Given that marginal revenue, the banker must form some estimate of the likely cost associated with creating another dollar of deposits (an estimate that depends to a large degree on expectations that may or may not be turn out to be correct), and it is the comparison between the marginal revenue from creating additional deposits with the expected cost of creating additional deposits that determines whether a bank wants to expand or contract its deposits.

Of course, the incomplete analysis of the decision making of the banker is not just Hawtrey’s, it is characteristic of all Wicksellian natural-rate theories. However, in contrast to other versions of the natural-rate genre, Hawtrey managed to avoid the logical gap in those theories: the failure to see that it is the spread between the lending and the deposit rates, not the difference between the lending rate and the natural rate, that determines whether banks are trying to expand or contract. But that is a point that I will have to come back to in the next installment in this series in which I will try to follow through the main steps of Hawtrey’s argument about how a banking system adjusts to a reduction in the quantity of fiat money (aka legal tender currency or base money) is reduced. That analysis, which hinges on the role of merchants and traders whose holding of inventories of goods is financed by borrowing from the banks, was a critical intellectual innovation of Hawtrey’s and was the key to his avoidance of the Wicksellian explanatory gap.

About Me

David Glasner
Washington, DC

I am an economist in the Washington DC area. My research and writing has been mostly on monetary economics and policy and the history of economics. In my book Free Banking and Monetary Reform, I argued for a non-Monetarist non-Keynesian approach to monetary policy, based on a theory of a competitive supply of money. Over the years, I have become increasingly impressed by the similarities between my approach and that of R. G. Hawtrey and hope to bring Hawtrey's unduly neglected contributions to the attention of a wider audience.


Enter your email address to follow this blog and receive notifications of new posts by email.

Join 2,198 other followers

Follow Uneasy Money on