Marglin on Keynes on Money

Fall 2021 - A Symposium on: Stephen A. Marglin’s Raising Keynes: A Twenty-First Century Theory

As I have written elsewhere, Stephen Marglin has written an important, even a great book, a masterpiece of its type. Its chief value lies in a demolition, beyond thorough, wholly persuasive, of the neo-Keynesians, new Keynesians and forthright anti-Keynesians who clutter the landscape, and have done so for 70 years at least, twisting Keynes into the opposite of what he was. But the treatment of money in Raising Keynes does not do justice to Keynes.

In the conclusion of Chapter 13 of Raising Keynes, Marglin writes:

“There is no theory of money in the General Theory. Or, rather, there is no coherent theory of money. Given the existence of a central bank, Keynes undoubtedly thought of the money supply as fixed by the monetary authority, and the hurdle rate of interest as responding, perhaps insufficiently, to the money supply.” [1]

Although in his introduction Marglin asserts that no one can know what Keynes “really thought,” here he is beyond doubt. The certainty is based on an absence – or alleged incoherence – of Keynes’s treatment of money, mostly in Chapter 17 of The General Theory of Employment, Interest and Money.  This claim appears several times in Raising Keynes, and it has motivated the moderators of this round-table to assert that Keynes ignored “the role of banks in money creation,” that he understood money as the “creation of a central bank”, and yet he theorized as though money were a commodity – “gold, silver, or cowrie shells.” Marglin also cites Keynes’s statement that all money is Chartalist, which is to say determined by the State, and declares, “Chartalist money is fiat money.” [2]

In matters relating to Keynes, condescension is a risky tactic. The man wrote a great deal. One can often – not always, but quite often – work out the basics of what he thought and meant by reading, if not in one place, then in another. It is true that the treatment of money in The General Theory is compact. A prudent hypothesis might allow for the possibility that Keynes – who both saw himself as and actually was the preeminent monetary theorist of his age – had already made his theory of money plain, notably in the two volumes and almost 800 pages of A reatise on Money, published in 1930, just six years before The General Theory. The first volume of the Treatise is subtitled “the pure theory of money;” the second volume is “the applied theory of money.”

In the opening pages of Volume 1, Keynes does state that all money, “beyond the possibility of dispute,” is Chartalist, meaning that the State exercises the power to declare and define what is money, and to change, from time to time, that definition. [3] But Chartalist money is not (necessarily) fiat money. States may, can, and have declared their money to be commodity money, such as weights of gold, silver or copper. They have declared it to be “representative money,” which is to say notes and tokens convertible into commodities, the common case of the gold standard. Fiat money is unbacked state money – such as continentals or greenbacks – whose modern form dates in the United States only to the end of the gold-exchange standard – partly Keynes’s invention – in 1971.

Turning the pages, we find Chapter 2 of the Treatise, entitled, “Bank Money,” and Chapter 3, “The Analysis of Bank Money.” Bank money is not central-bank money. It is created by private bank lending (or equivalently, investment financed by a draft on the bank). Bank money is unstable, since a bank is strengthened when it is the recipient of deposits created by and transferred from other banks, and weakened when the deposits it creates flow out, draining reserves. (Thus, the strength of a bank is determined by the relative degree of trust it enjoys.) If banks act in tandem, and if the central bank goes along, then reserves may grow everywhere alongside deposits, and there may be no reserve constraint on lending. Hence, given the willingness of the public to take out loans, the quantity of bank money is endogenous to the decisions taken jointly by bankers as a class, and contrary to textbooks, the central bank won’t enjoy full control. If it did, perhaps panics would be more rare than they are. All of this is in the Treatise.

It is therefore not Keynes who neglected banks and the financial instability hypothesis later developed by Hyman Minsky – unmentioned in Raising Keynes – in a book entitled John Maynard KeynesThe disappearing act occurred in the crossing from Cambridge to Cambridge (and Chicago), when The General Theory was taken up by Paul Samuelson, Franco Modigliani, and Milton Friedman. It consists of insisting that Keynes of The General Theory was an entirely different person from Keynes of A Treatise on Money, so that the earlier book can be ignored.

But it can’t be. The General Theory is by Keynes’s own description a “monetary-production” theory of output-as-a-whole, from which employment and unemployment are derived. The departure from the Treatise consists in the fact that the General Theory takes the theory of money and builds on that foundation a theory of fluctuations in output and employment, which had been missing from the Treatise. In making the connection of money to production, as I’ve written elsewhere and long ago, Keynes is making an unmistakable reference to Einstein and to the concept of space-time; the “General  Theory” consists of bringing money into the analysis of output, from which classical economics had excluded it, just as time stood apart from space in classical mechanics.

Nothing about the analogy, including the very title of Keynes’s book, is accidental. Keynes had met Einstein and had written about him. He knew (as did every literate person in those years) that Einstein’s theory rested on Reimann’s geometry. So when he wrote that the classical economists  “resemble Euclidean geometers in a non-Euclidean world,” he was making a reference that he could assume his readers would not miss. His analogs to Euclid’s “axiom of parallels” were the classical supply curves: the assumption that labor supply responds to a rise in the real wage, and that savings responds to a rise in the interest rate. Remove those axioms, and the constructs of self-contained, supply-and-demand-driven “labor markets” and “capital markets” collapse. In their place Keynes builds an integrated economics, a true macroeconomics of the whole, not based on “micro-foundations,” in which money plays the binding role. 

Money enters the system in several ways. It is an alternative to riskier assets, a safe haven in times of trouble and a safe harbor from which speculative sorties may be launched. People seek the barren but liquid asset and have to be lured away from it by the promise of a higher reward; thus liquidity preference becomes the foundation of the theory of interest rates and the yield curve. Another way in which money enters the system is in consequence of expected = profits in a new venture – which Keynes called the “animal spirits” of the entrepreneurs. This is what motivates the business firm to request loans in the first place, and banks to provide them, creating bank money by funding their newly-issued loans. Profit expectations are of pecuniary return, inherently subjective, inherently volatile, inherently monetary. After all, businesses are in the business of “making money.”

Keynes’s monetary-production theory was obscured by the MIT-Harvard-Chicago formulation, rooted in the Hicks-Hansen IS-LM model, which focused on allocating a fixed money supply – fiat money issued or at least controlled by a Treasury or a central bank – between transactions and asset demands, while treating interest as the result of “loanable funds” – a habit still deeply embedded in mainstream thought – and business as motivated by production as such and by the lure of “real rates of return.”

Banks and money-creation by banks are indeed absent from this model as they are from other textbook favorites, which explains, in part or perhaps in whole, why conventional “Keynesians,” “New Keynesians,” “monetarists,” and theorists of the “real business cycle” are helpless when faced with a crisis of banks and banking.

The actual line of descent from Keynes in money matters runs through Minsky and (in part) also the fervent anti-monetarist Nicholas Kaldor, to Robert Skidelsky, and recently to the advocates of Modern Monetary Theory, for whom the Treatise is an indispensable touchstone. Indeed the word “modern” in MMT is an ironic reference to the only passage in the Treatise that Marglin quotes:

“The right [to determine what constitutes money] is claimed by all modern States and has been so claimed for four thousand years at least.” [4](emphasis added)

Let me try to summarize the supposedly missing theory. Keynes writes that all assets have three properties: a prospective yield (q), a carrying cost (c), and a degree of liquidity (l). The sum of these properties ( c + l), all expressed in proportion to the asset’s price, is the “own-rate of interest” for that asset, or as Keynes put it, “the total return expected from the ownership of an asset over a period.” The peculiarity of money, in any form, is that l is substantial while q is nil and c is small. And the problem for employment arises when the own-rates on other assets fall below l, the liquidity preference for money. At that point the production of capital goods ceases and a persistent state of mass unemployment results.
“Unemployment develops, that is to say, because people want the moon; men cannot be employed when the object of desire (i.e., money) is something which cannot be produced and the demand for which cannot be choked off.” [5]

Marglin’s treatment of liquidity preference in Chapters 11 and 12 of Raising Keynes aims to establish it as a theory of interest rate differentials – of the rates of interest on assets other than money that are necessary to overcome the attraction of money as a liquid asset. But, says Marglin, the theory cannot be a theory of interest per se, because the base rate of interest on money is not necessarily zero. His evidence is a chart of the overnight rate on bank reserves – the federal funds rate – in the United States, which peaked at above 20 percent in 1979 and 1981. The problem here is that fed funds are not money. They are loans made between banks to meet a deficiency of required reserves – a penalty for being overextended, or for having a weak reputation and losing deposits to another bank.

When Paul Volcker raised the fed funds rate, some could, indeed, invest lucratively in Treasury bills. But the general public did not get a pass-through on its demand deposits, nor on cash. For most ordinary people and businesses the prospective yield on money remained zero, as before, and those 20 percent rates might as well have been paid on the far side of the moon. Despite this, for many people the appeal of cash increased, as bond and stock prices suffered a huge shock, spreading fear and panic. Investment collapsed and unemployment soared, just as Keynes would have predicted.

Liquidity preference is thus more than the inverse of the demand (in equilibrium) for yield-bearing assets and the theory of liquidity preference is more than an attempt to provide a theory of interest differentials between long-dated and short-dated assets in normal times. Liquidity is the distinctive property of money and liquidity preference is the keystone of the theory of employment.  Impenetrable or not, Chapter 17 in The General Theory is the key to Keynes.

In Marglin’s theory of money, banks “create money to fulfill the transactions demand” [6] subject to the fractional reserve requirements imposed (as Keynes wrote) by law or custom. Marglin writes that banks stay fully loaned-up so far as they can – any less would not be profit-maximizing – so the reserve ratio is a hard constraint on their lending. He illustrates this with a Venetian example, involving “bank reserves of 200 gold ducats.” [7] This is telling.

Marglin’s fractional reserve story suffices, so long as the qualified reserves are fixed and different from bank money itself. In his Venetian example, this is the case: banks must hold gold ducats to a certain percentage of their lending, and as the supply of coins is fixed, so too is the maximum volume of loans. [8] In a world of fiat money, a similar story could be told of vault cash, the stuff of runs, already in Keynes’ day a small proportion of bank money in the United States. But in the modern world, a qualified bank reserve may be nothing more than a deposit at the banker’s bank, taking the same ethereal form as bank money. Total reserves are not constrained to a fixed value, but will vary with bank behavior and with central bank action, such as open market operations.

Thus as Keynes understood, a required reserve ratio may or may not constrain bank lending; the central bank may or may not conduct the orchestra and control the tempo. In the late 1970s, achieving “monetary control” by changing central bank procedures was a monetarist-inflected legislative objective in the US Congress, in which I was unfortunately involved. [9] Effective control was never achieved and the quest was dropped in the debt crises of the early 1980s. In recent years, we find that adding reserves, by buying assets from banks, also does not constitute an effective stimulus to new loans; as Minsky observed, banks do not lend reserves and do not require reserves in order to lend. Textbook-trained economists believed otherwise, that a massive program of asset purchases, known as “quantitative easing” would bring strong growth – but it did not happen. Instead, the excess reserves piled up. In 2008 the Federal Reserve started paying interest on bank reserves, and in 2020 it abolished the required reserve ratio altogether. The life of the banker is now easier than ever: he gets a check for doing nothing, and the rest is both optional and unconstrained.

Keynes is also quite clear about transactions demand for money; it exists to cover the mismatch in time between earnings and outlays, and so forms a basis for deposit balances in banks as a set of institutions for the clearing of payments. Yet this is not the key function of bank money. Banks in well-functioning economies mainly make business loans, and create the corresponding deposits, to advance funds for the manufacture and acquisition of capital assets for business purposes. In this way, they fund the employment of workers producing capital goods, and so make possible, so long as profit expectations hold up, full employment in a growing capitalist system. Yet it is not in any sense the primordial goal of the banker to pursue and achieve these social objectives. The unqualified and unregulated goal of the banker is merely to “make money” – like anyone else. Bankers make money by playing with money, a medium they themselves create. They enjoy the advantage of a medium unburdened by objective constraints – unlike engineers who must design bridges that do not fall down, doctors whose patients must recover more often than not, or even lawyers who must prevail against the opposition of other lawyers. They are in no sense reliable as job creators.

Marglin rightly concludes that we are led to a world in which government spending must be the linchpin of full employment. At least, that is true in the Anglo-American world of money-market capitalism, in which banks set the terms of their own existence and the rules governing their own conduct. It was not altogether true in the United States of Franklin Roosevelt’s New Deal and the public-sector Reconstruction Finance Corporation of Jesse Jones. It was not at all true in Western Europe [10] during the years of post-war reconstruction, when ostensibly private banks were led by the nose to serve public purposes given by public authority. It is not true, even today, in those economies where public policy still pursues larger goals of social construction and economic development – what I have called, in honor of my father, the Galbraithian states. But it has become true in the headquarters-nations of global finance, which are left with the tools of “stimulus” to do the otherwise-neglected task of maintaining demand and employment.

Keynes wrote The General Theory at a moment, not unlike the present, when private bankers had abandoned even the pretense of a social role. Yet Marglin’s remark that the complementary role of public spending had eluded him until “some years later” Abba Lerner persuaded him is strange. [11] We Can Conquer Unemployment and Keynes’s companion essay, Can Lloyd George Do It? had already been the platform of the Liberal Party’s 1929 campaign. The fiscal multiplier was already in mind by then and had been fully-worked out by the early 1930s. In 1933 Keynes had already written his famous letter to Franklin Roosevelt:

“Thus as the prime mover in the first stage of the technique of recovery I lay overwhelming emphasis on the increase of national purchasing power resulting from governmental expenditure which is financed by Loans and not by taxing present incomes. Nothing else counts in comparison with this.”.

Once again, reading around is helpful.

( Just Money -   

James K. Galbraith holds the Lloyd M. Bentsen, Jr. Chair in Government/Business Relations at the Lyndon B. Johnson School of Public Affairs and a professorship in Government at The University of Texas at Austin. He is a member of the Society of Kings Economists who began his study of economics at the University of Cambridge in 1974. He then worked on monetary control and monetary policy oversight for the Committee on Banking, Finance and Urban Affairs of the US House of Representatives, where he drafted early versions of what became the monetary policy provisions of the Humphrey-Hawkins Full Employment and Balanced Growth Act of 1978, commonly known as the “dual mandate.” He is a recipient of the Leontief Prize and of the Veblen/Commons Award. He directs the University of Texas Inequality Project. In 2010 he was elected to the seat formerly held by Paul Samuelson at the Accademia Nazionale dei Lincei.

[1] Stephen A. Marglin, Raising Keynes: A Twenty-First General Theory (Cambridge: Harvard University Press, 2020), 522. BACK TO POST

[2] Ibid, 504. BACK TO POST

[3] John Maynard Keynes, A Treatise on Money (Cambridge: Cambridge University Press, Vol. I, 1930), 5. BACK TO POST

[4] Marglin, Raising Keynes, 504. BACK TO POST

[5] John Maynard Keynes, The General Theory of Employment, Interest, and Money (London: Palgrave Macmillan, 1936), 235. BACK TO POST

[6] Marglin, Raising Keynes, 498. BACK TO POST

[7] Ibid, 507. BACK TO POST

[8] As David Graeber relates in Debt: the First 5000 years (p. 338) the penalties for failing to keep adequate specie reserves could be quite bracing.  In 1360 Francesch Castello, a banker in Barcelona, was beheaded in front of his own bank for such a breach.  Standards declined in the seven centuries thereafter. BACK TO POST

[9] I was a staff economist for the Committee on Banking, Finance and Urban Affairs of the US House of Representatives under Chairman Henry Reuss (D-WI); the bill in question was the Depository Institutions Deregulation and Monetary Control Act (DIDMCA) of 1980,  far better remembered for having set in motion the savings-and-loan calamities to come than for the futile exercise of monetary control. BACK TO POST

[10] The study linked, “Monetary Policy, Selective Credit Policy and Industrial Policy in France, Britain, West Germany, and Sweden,”  was conducted by this author over the summer of 1980 in conjunction with John Zysman and Stephen Cohen of Berkeley, Andrew Martin of Harvard, Richard Medley (later founder of Medley Global Advisers) and Catharine Hill, later President of Vassar. BACK TO POST

[11] Ibid, 501. BACK TO POST Return to Symposium Prompt

James K. Galbraith