How About: Keynes's Approach to Money: What Can Be Recovered
"Keynes's Approach to Money: What can be recovered?"
L. Randall Wray, University of Missouri—Kansas City; Levy Economics Institute
This paper will first take a retrospective look at Keynes’s General Theory approach to money. We next turn to the neoclassical synthesis approach to money to determine what was retained, and what was shed, from Keynes’s approach. Finally, we examine what needs to be recovered to create a coherent and useful approach to money that synthesizes Keynes’s early insights with more recent developments in monetary theory.
Keynes’s Approach to Money in the General Theory
Elsewhere I have argued that Keynes adopted two contradictory, approaches to money in the General Theory. (Wray 2006)The first approach is the more familiar ‘supply and demand’ equilibrium model of Chapters 13 and 15, incorporated within conventional macroeconomics textbooks. The second is presented in Chapter 17, where Keynes drops money supply and demand in favor of a liquidity preference approach to asset prices. I argued that the supply and demand approach suffers from problems of interdependency, and is not consistent with his general approach, while the liquidity preference approach of Chapter 17 is critical to our understanding of the role money plays in causing unemployment.In this section I will briefly summarize only the approach taken in Chapter 17.
In Chapter 17, Keynes presents a general theory of asset prices. The expected nominal return to holding any asset is q-c+l+a, where q is the asset’s expected yield, c is carrying costs, l is liquidity, and a is expected price appreciation. This total return can be used to calculate a marginal efficiency for each asset, including money.The composition of returns varies by asset, with most of the return to illiquid assets (i.e. capital) consisting of q-c, while most of the return to liquid assets consists of the l. If a producible asset’s return exceeds that on money, it is produced up to the point that its marginal efficiency falls back into line with money’s return that rules the roost. If an asset that is not producible has a higher marginal efficiency, its price is pushed up and its return falls back in line. Finally, changing expectations about the future have differential impacts on the marginal efficiencies of different kinds of assets. Increased confidence raises the qs on capital while lowering the subjective values assigned to liquid positions (hence, the l falls), so that the marginal efficiency of capital rises relative to that of liquid assets. Capital assets will be produced (investment rises, inducing the ‘multiplier’ impact) and the full range of asset prices adjusts. Thus, expectations about the future go into determining the equilibrium level of output and employment—defined as a position in which firms hire just the amount of labor required to produce the amount of output they expect to sell.
Two important conclusions follow from the Chapter 17 approach. First, it is the existence of money that prevents the economy from coming to equilibrium (state of rest) at less than full employment; and, second, Keynes does not need an exogenously-fixed money supply to explain the determination of interest rates. The first is important to make the theoretical case against neoclassical theory, while the second implies that unemployment is not due to misguided monetary policy. The supply/demand approach taken in Chapter 13&15 would seem to imply that the central bank could eliminate unemployment by increasing the money supply, pushing the interest rate down, and inducing interest-sensitive spending. However, according to Keynes, the problem is more fundamental and cannot be resolved through monetary policy alone. Keynes argued the existence of money is the cause of unemployment, because “in the absence of money…the rates of interest would only reach equilibrium when there is full employment.” [Keynes, 1964, p. 235] Here he is referring to the spectrum of own rates, equalized in the absence of money only at full employment. However, money sets a standard that is often too high for full employment. Further, he cautioned that an “increase of the money supply” is not necessarily a solution, as “there may be several slips between the cup and the lip”. [ibid., p. 173] If liquidity preference is rising faster (or, the marginal efficiencies of producibles are falling faster) than the money supply, it will not be possible to stimulate employment. He was thus “somewhat skeptical of the success of a merely monetary policy directed towards influencing the rate of interest….since it seems likely that the fluctuations in the market estimation of the marginal efficiency of different types of capital…will be too great to be offset by any practicable changes in the rate of interest”. [ibid., p. 164]
What is the shape of the money supply function that is consistent with the Chapter 17 exposition? The forces that equalize marginal efficiencies do not depend on any particular money supply function—all that matters is the way in which financing terms affect forward-looking marginal efficiencies at the time individual decisions are taken, when the marginal efficiency of a particular capital asset is weighed against the marginal efficiency of money. Whether a higher scale of activity will affect interest rates is neither known nor considered by the individual taking the ex ante decision to invest. He will consider his finance costs but it would be illegitimate to presume that effects on interest rates resulting from greater aggregate spending matter in this decision. Thus, whether the money supply curve is vertical, as in the textbooks, or horizontal, as in a popular version of endogenous money (see below) does not matter. All that is necessary is that money have “peculiar characteristics” that cause its marginal efficiency to “rule the roost”. Interdependence of money supply and money demand functions does not pose a problem for the Chapter 17 approach—where equilibrium is defined not as money demand equals money supply, but as equality among own rates of interest. As I will argue below, this is important because few economists today believe that the money supply can be exogenous.
Money in the Neoclassical Synthesis
There is little need to go through the neoclassical synthesis approach to money in great detail, because it is too well-known. However it should be noted that it is closely based on the Chapter 13/15 approach of the General Theory. Three markets are treated: a goods market, a money market, and a bond market. Equilibrium is defined as the income and interest rate combination at which saving equals investment, money demand equals money supply, and bond demand equals bond supply. By Walras’s Law, the bond market is usually dropped in favor of a two equation IS-LM model. As in Chapter 13/15 of the GT, money demand consists of three motives and is both interest rate and income elastic. The IS-LM model adopts an exogenous money supply fixed by policy (the usual deposit multiplier approach is often invoked). Indeed, derivation of the LM curve requires independence of money supply and money demand functions.
It is well-known that Hicks had created a version of the IS-LM model to compare the “classical” approach to Keynes’s approach. In his March 31, 1937 letter to Hicks, Keynes raised two important objections to this framework. First, he argued that “From my point of view it is important to insist that my remark is to the effect that an increase in the inducement to invest need not raise the rate of interest. I should agree that, unless the monetary policy is appropriate, it is quite likely to.” [Keynes, CW 14, p. 80] He argued that while an increase of planned spending—as well greater actual spending—raises money demand, it will raise interest rates only if money supply does not increase. When Hicks clarified the presentation in a model, Keynes did not like the resulting implication that a shift of the IS curve will change the equilibrium interest rate so long as the LM curve (or, money supply) is independent of spending. This seems to be more evidence that Keynes’s approach in Chapters 13 and 15 did not adequately represent the general argument he was trying to make because he would not accept the assumption that spending increases while the money supply is necessarily fixed.
Keynes’s second point was that the Hicks treatment places too much emphasis on current income, while “in the case of the inducement to invest, expected income for the period of the investment is the relevant variable”. (ibid.) Keynes said he is willing to take both liquidity preference (money demand) and saving as a function of current income, but investment is forward-looking, a function of expected future income as that will go into determining future effective demand. The usual investment function that lies behind the IS curve is far too stable to represent what Keynes had in mind, and what he had presented in Chapter 17. The IS curve makes investment and saving functions of income and interest rates, but Keynes argued that while that might be appropriate for saving, it misrepresents investment, which is a function of expected future profit income.
More importantly, Keynes noted in his letter that he was preparing a lecture on the rate of interest, which became a response to the views of Hicks as well as those held by the Swedish school (Ohlin) and Robertson. There he detailed his objections to the loanable funds approach, as well as to Robertson’s attempts to add saving plus bank loans to obtain a hybrid source of finance. Keynes argued that saving is not equivalent to finance, indeed, “Saving has no special efficacy, as compared with consumption in releasing cash and restoring liquidity…There is, therefore, just as much reason for adding current consumption to the rate of increase of new bank money in reckoning the flow of cash becoming available to provide new ‘finance’, as there is for adding current saving.” [ibid., p. 233] Indeed, “Increased investment will always be accompanied by increased saving, but it can never be preceded by it. Dishoarding and credit expansion provides not an alternative to increased saving, but a necessary preparation for it. It is the parent, not the twin, of increased saving.” [ibid.,, p. 281] Investment, itself, cannot pressure interest rates because it returns to the “revolving fund of finance”, creating equivalent saving. [ibid., p. 208]
Further, “unless the banking system is prepared to augment the supply of money, lack of finance may prove an important obstacle to more than a certain amount of investment decisions being on the tapis at the same time. But ‘finance’ has nothing to do with saving.” [ibid., p. 247] If banks do relax to satisfy the finance motive, interest rates will not rise as the scale of activity increases. In other words, while Keynes adopts equality between investment and saving (as in the ISLM framework), he rejects the notion that saving “finances” investment. Rather, only a reduction of hoarding or expansion of credit can “finance” rising investment.Finally, he noted, “to the extent that the overdraft system is employed and unused overdrafts ignored by the banking system, there is no superimposed pressure resulting from planned activity over and above the pressure resulting from actual activity.” [ibid., p. 222-3] In an “overdraft” system, an increase of investment spending (or any other type of spending that has access to overdrafts) is associated with expansion of bank loans, and hence, with expansion of the money supply (as bank deposits grow with lending). Here Keynes has gone further, arguing that even “planned spending” can lead to increased lending, if there is an attempt to accumulate money in advance of spending it (the finance motive). What is important is Keynes’s recognition that the “goods market” and “money market” are not independent—rising spending can be (indeed, normally is expected to be) met by rising money supply, so that there is no pressure on interest rates.
Recall from above that Keynes blamed money for the existence of unemployment. An ISLM model with Pigou (or real balance) and Keynes effects that make spending a function of the real money supply will achieve equilibrium only at full employment. To be sure, a large number of “rigidities” has been proposed to explain persistence of unemployment—ranging from wage and price stickiness to coordination failures. Still, without rigidities, full employment would be attained. By contrast, Keynes argued that equilibrium—defined throughout the General Theory as the point at which effective demand equals income, and in Chapter 17 as the point at which the marginal efficiency of capital equals the marginal efficiency of money—can occur at any level of employment. In the ISLM model, the equilibrium (real) rate of interest associated with the intersection of the curves is consistent with the full employment level of income (again, in the absence of frictions). In the General Theory there is a different equilibrium interest rate associated with each level of income and employment, only one of which represents full employment. This is because, as described above, Keynes argues that in the absence of money the own rates could come to equality only at full employment, while in the presence of money, its own rate sets a standard that is usually too high to allow production to proceed on the necessary scale to achieve full employment. It is money, not rigidities, that causes unemployment in Keynes’s view.
While the neoclassical synthesis approach to money is based on Keynes’s presentation in Chapters 13 and 14, it is not consistent with his general argument, thus, reaches different conclusions. First, ISLM equilibrium is achieved at a unique interest rate/income combination. Moreover, the goods and money markets are dichotomized such that it is possible to increase spending while holding money supply constant; except for extreme values of parameters, increased spending will raise the equilibrium level of interest rates, while increased money supply will lower the equilibrium interest rate. The money supply is exogenous, under control of the authorities. In the absence of rigidities, involuntary unemployment is eliminated through effects on prices that in turn affect spending through Pigou and Keynes effects. While the ISLM model does not adopt a loanable funds approach, it is relatively silent on the finance process involved when spending rises with a constant money supply. This requires velocity to increase, induced by rising interest rates. However, many or even most practitioners of the neoclassical synthesis have at least implicitly adopted a loanable funds approach, especially in discussing the long run and in growth models where investment is constrained by saving propensities.
What Needs to be Recovered from Keynes?
- Liquidity Preference and Endogenous Money
Even orthodoxy now rejectscentral bank control of the quantity of money, and some orthodox approaches explicitly assume that the money supply—broadly defined to include bank deposits—expands as spending grows, rejecting exogeneity in favor of endogeneity. How do we reconcile liquidity preference theory with the reality that central banks today operate with a short-term rate target? Even if we accept complete discretionary control over the overnight rate, as well as substantial influence over other interest rates on instruments such as government bonds, this still leaves a role for liquidity preference in determining all other own rates. Keynes never argued for mono-causality, rather, he singled out the role played by liquidity preference because he believed that to be the ultimate barrier to operation of the economy at full employment. The authorities can lower interest rate targets in response to unemployment, but if marginal efficiencies of producibles (assets produced by labor) are too low, even enlightened policy would not generate full employment.
We can distinguish between money as a stock, desired due to liquidity preference, and money as a flow used to finance spending. (Wray, 1990) Rising spendingwill normally lead to an increase of the money supply (defined as an increase of bank liabilities as banks make loans). Whether the loan interest rate rises depends on numerous factors, including expected policy and liquidity preference of banks—but neither a completely elastic (horizontal) supply of loans nor a completely inelastic (vertical) loan supply curve is likely. On the other hand, rising liquidity preferenceis associated with a reduction of planned spending as marginal efficiencies of producible assets fall relatively to the return to liquid assets. For this reason, money demand (defined as demand for loans to spend) could even fall when liquidity preference rises. Money supply will not normally rise to meet rising liquidity preference; instead, asset prices adjust across the full spectrum of assets until wealth holders are satisfied to hold the existing set of assets. Hence, endogenous money is reconciled with liquidity preference, with money demand here defined not as a desire for a hoard of money in advance of spending but as a flow demand for finance of spending.
Others, including most prominently Moore (1988), have adopted a “horizontalist” approach. There are two aspects to the argument. First, the supply of reserves is horizontal at the interest rate target chosen by the central bank. For several reasons (some of which are explored below), central banks cannot control the quantity of reserves. In any event, it is now accepted by most theorists and practitioners that central banks should and do operate with interest rate targets. This means that the supply of reserves accommodates the demand so that the interest rate target can be hit (termed “accommodationist”). The second part of the argument is that private banks also accommodate the demand for loans and deposits, “horizontally”. To put it as simply as possible, banks take the overnight interest rate target as a measure of the marginal cost of funds, then “mark-down” to set the deposit rate of interest and “mark-up” to set the loan rate. A distinction is made between “retail” and “wholesale” markets and interest rate determination in each, but that is not important to our argument. What is critical, however, is the recognition that the demand for loans and for deposits is linked to planned spending. Thus, the “money supply” expands “horizontally” as spending rises, with no pressure on interest rates.