You are on page 1of 12

Cambridge Journal of Economics 1994, 18, 379-390

Money's marketability premium and the microfoundations of Keynes's theory of money and interest
Tyler Cowen and Randall Kroszner 1. Introduction Attempts to generate cyclical fluctuations and underemployment equilibria generally focus upon some stickiness of prices, wages, or interest rates. Models of costly price adjustment, efficiency wages, and credit rationing, for instance, can generate Keynesian results to varying degrees. In the search for a 'stickiness culprit', however, economists have neglected the rate of return on cash balances.1 Rigidity of the liquidity or marketability return on cash balances connects the real and monetary sectors and makes nominal interest rates sticky. We analyse Keynes's (1936, p. 23) proclamation in chapter 17 of The General Theory that the money rate of interest 'rules the roost" and argue that this is a key feature of Keynes's theory. Keynes's theory of money and interest in The General Theory has often been criticised as unsatisfactory. Hansen (1953), an influential interpreter of The General Theory, believes Keynes's work on capital, interest, and money in chapters 16 and 17 is simply 'another detour which could be omitted without sacrificing the main argument (p. 155) and says more specifically about chapter 17 that'.. . not much would have been lost had it not been written' (p. 159). In contrast, the Post Keynesians, most prominently Davidson (1980), have emphasised the importance of chapter 17 in understanding The General Theory but have not offered a compelling analytical apparatus for interpreting the chapter and its relationship to other parts of the book. We attempt to provide such an apparatus and restore the logical coherence of Keynes's argument. Keynes's assumption of a relatively flat marginal return curve for money can provide a microfoundation for Keynes's macroeconomics without recourse to the usual assumptions about sticky prices and wages or naive expectations. We begin with an overview of the logic of this approach in the next section. The following section explains how the approach arises from Keynes's 'essential properties of interest and money* and uses the analysis of the 'essential properties' to recast Keynes's theory of the trade cycle. In Section 4, we evaluate the applicability of Keynes's argument, and identify conditions under which Keynesian results do not obtain.
Manuscript received 25 November 1991; final version received 26 February 1993. *George Mason University and University of Chicago, respectively. The authors wish to thank Michael Alexeev, Jack High, Daniel Klein, George Selgin, David Shepherd, Gordon Tullock, seminar participants at New York University, and especially two anonymous referees for useful comments. 1 Townsend (1987) offers a brief survey of works that emphasise the zero pecuniary yield of cash balances; these works include the Mundell-Tobin effect and Friedman's (1969) optimum quantity of money. Carmichael and Stebbing (1983) and Fried and Howitt (1983) use this approach as well. 0309-166X/94/040379+12 $08.00/0 1994 Academic Press limited

380

T. Cowen and R. Kroszner

Our interpretation of Keynes's theory of interest and money also helps us to make sense of a number of other sections of The General Theory that, like chapter 17, typically are dismissed as quaint, incoherent, or simply wrong.1 In the fifth section, we demonstrate how Keynes's endorsement of mercantilism and Silvio Gesell's money stamping scheme are consistent with our interpretation. Keynes's views on money stamping are, in fact, crucial to our interpretation. Also in accord with our account is Keynes's repeated advocacy of policies to reduce the interest rate and for state control of investment, usually dismissed as ' "free-wheeling" detour[s] by Keynes in his less responsible moments' (Hansen, 1953, p. 159).2 2. Money's marketability premium, interest rates, and return-bearing assets We first consider economies where money simultaneously serves as a medium of exchange and medium of account, but yields no pecuniary return. Money assets yield a marketability or liquidity return in lieu of explicit interest payments, dividends, or capital gains (e.g. White, 1987). Medium of exchange services are the source of money's non-pecuniary return. Money has afixed,immediately apparent nominal value and can be passed hand-to-hand in convenient fashion. The implicit or non-pecuniary nature of money's marketability premium implies that the services of money do not have a separate or independent price that can be adjusted at the margin. Money's liquidity services cannot be packaged, traded, or priced separately from the purchasing power of the money asset. For a given quantity of real balances, the marginal return on those balances is constant.3 The fixity of the return on a given quantity of real balances requires that transaction technologies are constant in the short run. To the extent that transaction technologies are not fixed, the government or private firms can install more automatic dollar changers in public places, modify public telephones to accept dollar bills, or negotiate treaties with foreign countries to make dollar an internationally accepted means of payment. All of these policies would alter the convenience yield on a given quantity of real balances but we assume that such changes are too costly or too ineffective in the short run. The equilibrium relationship between real balances and the interest rate requires that the rate of interest in money loans (adjusted for risk) equal the marketability or liquidity services that money provides. Money, with its three 'essential properties' discussed below, is the determining force behind die rate of interest (e.g. Keynes, 1936, pp. 181, 213). In particular, the rate of interest cannot fall below the liquidity premium on the prevailing level of real balances.4 At zero inflation, for example, no one will part with the marketability services of money for less than, say, m %. If this m is higher than the return on new capital projects, say k, then no one will invest. Assuming a standard declining marginal efficiency of capital
Our method follows the advice of Thomas Kuhn (1977, p. xii): 'When reading the works of an important thinker, look first for the apparent absurdities in the text and ask yourself how a sensible person could have written them. When you find an answer, . . . when those passages make some sense, then you may find that more central passages, ones you previously thought you understood, have changed their meaning.' 2 Meltzer (1988) stresses the importance of reconciling Keynes's policy discussions in The General Theory with the theoretical sections. 3 Furthermore, if we make the additional assumption that the marginal utility of money curve is flat or relatively flat, then the marginal return on holding money does not change regardless of the level of real balances. We discuss this point further in the following subsection and in Section 4. 4 Interpreters from Hansen (1953, quoted above) to Meltzer (1988, pp. 149, 173, 201) have described Keynes's theory of interest rate determination as ambiguous or even incoherent. We obviously differ from this view.
1

Keynes's theory of money and interest

381

schedule, as Keynes did, the capital stock will cease to grow when k reaches m. If m were not binding (or zero), the capital would continue to accumulate until satiation, that is, the marginal efficiency of capital reaches zero. The end state of capital satiation could be achieved through either directly altering m (through money stamping or other changes in the transactions technology) or using government control to invest until the satiation point. Keynes endorsed both Gesell's money stamping proposal and government control of investment as means to this end.
The link between monetary and real sectors

The absence of a separate price for monetary services, combined with the fixity of money's return, links the real and monetary sectors. Inflation lowers real returns on cash balances because cash does not offer an interest rate that can vary nominally to offset the effects of inflation. The nominal yield on bonds will not rise sufficiently to offset the effects of inflation because individuals flee from cash into bonds or other return-bearing assets. Thus, the real interest rate falls when inflations occurs. Similarly, deflation increases the rate of return on holding money and thus real interest rates as individuals move out of bonds and into money. The nominal yield on bonds does not fall in proportion to the deflation. These results are commonly known as the Mundell-Tobin effect.1 While our version of the Keynesian story has some similarities with the MundellTobin effect, we use a different assumption about the source of the economic shock and a different equilibration mechanism. First, we assume that the initial shock is real, not nominal. This shock changes the rate of return on capital (or, in Keynes' terms, shifts the marginal efficiency of capital schedule), thereby forcing monetary variables to respond.2 If the rate of return on holding capital decreases, for instance, the demand for real cash balances will increase. Second, changes in the demand for real balances will not restore equilibrium if the non-pecuniary rate of return on money is unresponsive to changes in the quantity of real balances. If the liquidity premium on money does not change, neither can the nominal rate of interest on money loans.3 When we combine these two features, Keynesian results can follow. Decreases in real rates of return on capital create discrepancies between the rate of return on monetary and non-monetary assets. If money's marginal liquidity return is unresponsive to changes in real balances held, a re-equilibration through interest rate changes will not occur in the short run. The real and nominal rates of interest will remain above the real and nominal returns on capital investment. New investment will not be forthcoming until the marketability return on real balances, and thus the nominal rate of interest, adjusts downward. The nominal rate of interest on money loans 'rules the roost" because the rate of return on capital must conform to the rate of interest on money loans. Exogenous
1 Carmichael and Stebbing (1983) provide econometric evidence supporting the real interest rate effects of inflation; see also Summers (1985) and Fried and Howitt (1983). Also of relevance are Wilcox (1983) and Beckerman (1983), who document unusually low and sometimes negative real interest rates for the 1970s, an inflationary decadea result consistent with our interpretation of Keynes. 2 Keynes emphasises 'animal spirits' as the source of the initial shock to the expected rate of return on capital. In contrast, modem real business cycle theorists (e.g. Kydland and Prescott, 1982; King and Plosser, 1983) typically emphasise technological factors as the source of an initial shock to capital's rate of return. 3 Modern theories of credit rationing (e.g. Sriglitz and Weiss, 1981) derive interest rate stickiness from lenders' imperfect information about borrowers: lenders may not raise the real interest rate on loans because they wish to avoid attracting only the riskiest of borrowers. Our framework, in contrast, explains nominal (not real) interest rate stickiness in the downward (not upward) direction.

382

T. Cowen and R. Kroszner

shocks cause excessive changes in investment because interest rates do not adjust fully in the short run.
Return-bearing exchange media

To understand better the microfoundations for the Keynesian argument we present above, we now consider a scenario in which the Keynesian problems discussed above do not arise. Specifically, we now examine a world in which all exchange media bear a flexible and explicitly pecuniary rate of return; we refer to this as interest on exchange media. It is not our purpose to investigate the feasibility of paying interest on all exchange media. Rather, we examine this case to identify which properties of money drive our results.1 When all exchange media bear interest, the net return on holding exchange media is flexible for a given level of real balances. Changes in the real rates of return on capital goods no longer require price level declines and changes in real balances to achieve equilibrium. Instead, the pecuniary rate of return paid on exchange media can adjust in tandem with the changing rate of return on capital. A negative shock to capital productivity is met with a decline in the pecuniary rate of return on exchange media. The extra 'degree of freedom' to manipulate the total return on money can break the link between the nominal and real sectors discussed above. Furthermore, the real rates of return on exchange media are protected from inflation by a Fisher effect, since changes in the rate of expected inflation or deflation can be reflected in the (flexible) nominal rate of interest paid on the circulating media. No Mundell-Tobin effect is possible because there is no fixed nominal return asset 'taxed' by inflation or 'subsidised' by deflation. The rate of interest on exchange media can change through a number of means. First, competitive profit-maximising financial intermediaries may supply interestbearing cash backed by return-bearing capital goods and change the rate of interest in response to market conditions. Second, the government may pay interest on cash and change the rate of interest to replicate the competitive outcome. More generally, the varying rate of return on interest-bearing assets need not take the form of explicit interest payments. Such alternatives include changes in the transaction technologies (which we assumed away above), or changes in the in-kind services associated with exchange media (e.g. changes in the cash pick-up services offered by banks or changes in the theft protection services offered on American Express travellers' cheques).2 For Keynesian results to be avoided, these implicit rates of return must vary with the real return on capital. All exchange media must have returns that can vary at the margin in the appropriate fashion for Keynesian results to be ruled out. If even a small quantity of non-interest-bearing cash remains, the rate of return on money can still rule the roost. When the real rate of return on capital falls, there is at least one asset (cash) whose rate of return does not automatically fall in tandem. The argument that the rate of return on money can rule the roost does not require that the supply of this asset be large. In addition, universal payment of interest on exchange media avoids Keynesian problems completely only when negative interest on exchange media is possible.
1 Black (1970), Wallace (1983, 1988), White (1987), Hoover (1988), Mott (1989), and Cowen and Kroszner (1987, 1989, 1994) debate the feasibility of paying interest on all exchange media. 2 Interest also could be paid probabilistically, for instance, with randomly announced lottery drawings using note serial numbers to identify winning 'tickets' (see Goodhart, 1986 and McCulloch, 1986).

Keynes's theory of money and interest

383

Consider, for example, a negative shock to the marginal efficiency of capital schedule in an economy with low (or zero) inflation. If the shock makes the real rate of return on capital negative, negative interest on money may be necessary to avoid a liquidity trap. Without negative interest payments, a Keynesian underemployment equilibrium might then arise, with all demands channelled into the money market, rather than into investment.1 Negative interest on cash is possible only if money holders and suppliers have made the appropriate contractual arrangements for 'money stamping' (as Gesell advocated) or for deductions from accountholders' balances. Such contingent contracts might resemble the fees paid to banks when account balances fall below a certain level but these fees would instead be triggered by bank losses on loans (i.e. a fall in the economy's real rate of return). Alternatively, accountholders might hold the underlying loans directly through a mutual fund format and suffer losses on portfolio holdings when the real value of productive activity falls.2 3. Keynes's chapter 17: 'The essential properties of interest and money' Keynes's chapter 17 argues that his theory of economic downturns is predicated upon three essential properties of a monetary economy. Below, we interpret these properties as zero elasticity of production, a constant or unresponsive liquidity premium, and a high and positive liquidity premium. We then show how these properties provide the microfoundation for the results described in the previous section, namely that the rate of return on holding real cash balances 'rules the roost' through its effect on the money rate of interest.3 The first property of money is its zero or negligible elasticity of production (Keynes, 1936, p. 230). Zero elasticity refers to private sector production and not the actions of a central bank. The demand for money is not a demand for a commodity that can be produced and generate employment. Keynes notes that if money could be produced 'like a motor car, depressions would be avoided or mitigated'. Countries with a gold-mining
1 We are indebted to an anonymous referee for this point. This referee also notes correctly that this version of the liquidity trap is most likely at low levels of inflation, where the inability of charging negative nominal interest on exchange is most likely to be a binding constraint. 2 On the risk properties of arrangements where account holders can suffer nominal losses, see Goodhart (1986) and Cowen and Kroszner (1989, 1994). In contrast, Keynes considers fixed nominal value to be an essential feature of money, because individuals hold money to avoid the nominal capital losses possible with bonds. 'Lerner (1952), Conard (1959), Kaldor (1960), Robinson (1961), Turvey (1966), Davidson (1978, 1980), Nell (1983), Vicarelli (1984), Cartelier (1988), Rogers (1989), Mott (1989), and Cornell and Lawlor (1991) offer contrasting discussions of chapter 17. An exhaustive survey of this literature is beyond the scope of this paper, but Cornell and Lawlor (1991) are closest to our analysis. These authors defend Keynes's interest rate theory against the 'natural rate' theory of Leijonhufvud and argue that Keynes's chapter 17 is an integral part of his theory of the interest rate mechanism. Cornell and Lawlor, however, consider the slope of money's marginal return curve only in passing (p. 636). Lemer's interpretation emphasises the role of wage and price stickiness. Under dlis view, chapter 17 demonstrates that the propositions of The General Theory hold true only in a world of sticky prices and wages. Davidson combines Keynes's analysis in chapter 17 with the Post Keynesian system but does not consider a constant liquidity premium for money. Conard and Vicarelli concentrate on textual exegesis of Keynes's writings. Turvey interprets chapter 17 as a parable about growth theory: a growing economy also requires proportional increases in the supply of money. Robinson also focuses upon growth theory and the question of why the long-run rate of interest is positive. Mott deals with Keynes's argument in chapter 17 for a unified medium of exchange/medium of account. Kaldor (I960, p. 70) argues that a sticky convenience yield on money requires the necessary condition that money is utilised as the economy's medium of account and cannot vary in price. Canelier (1988) emphasises the link between money's liquidity premium and determinants of the rate of interest. Meltzcr (1988, chapter 6) surveys various interpretations of The General Theory.

384

T. Cowen and R. Kroszner

industry may dampen fluctuations for this reason, but the world as a whole cannot generate much employment through production of the gold money commodity (Keynes, 1936, p. 231). Keynes refers to the second property of money as its zero or negligible elasticity of substitution, which we interpret as a constant or unresponsive liquidity premium. Keynes claims (p. 231) that: [money's] utility is solely derived from its exchange-value, so that the two rise and fall pan passu, with the result that as the exchange value of money rises there is no motive or tendency, as in the case of rent-factors, to substitute some other factor for it. Thus, not only is it impossible to turn more labour on to producing money when its labour-price rises, but money is a bottomless sink for purchasing power, when the demand for it increases, since there is no value for it at which demand is divertedas in the case of other rent-factorsso as to slop over into a demand for other things. Increases in money's exchange value increase money's use value proportionately and do not generate a strong impetus to 'sell' money (i.e. buy goods). As a result, neither falling prices nor an increase in nominal balances will produce an increase in real demand for goods. In other words, the marginal utility of money curve is relatively flat and does not decline sharply.1 As we will see below, this is the key property that allows money's rate of return to rule the roost. Keynes compares money to non-monetary factors, such as land, where increases in the exchange value of the land do not increase its use value. Increases in the exchange value of land thus create an incentive to sell the land, which generates demand for other goods and services. This incentive is not present or is very weak in the case of money. 2 The relatively flat marginal utility curve for money thus implies that drops in the price level are not undone by resulting increases in consumption; Keynes thus denies the operation of what later writers call the real balance effect (Patinkin, 1965). The 'zero elasticity of substitution' terminology simply means that falling prices do not generate significant 'substitutions' out of cash balances and into goods. This can be considered one of Keynes's implicit replies to the Pigou effect. Zero elasticity holds, of course, only to the extent the marginal value of holding additional cash balances does not diminish.3 Money's third essential property postulates a liquidity premium higher than carrying costs (Keynes, 1936, pp. 226, 233, 237, 239): it is an essential difference between money and all (or most) other assets that in the case of money its liquidity-premium much exceeds its carrying cost, whereas in the case of other assets their carrying cost much exceeds their liquidity-premium. (Keynes, 1936, p. 227) We treat this third property of money as following from the conventional nature of money. By convention, individuals use as money that asset whose liquidity premium exceeds its carrying costs. If some other asset were to possess a greater excess of liquidity
Cartelier (1988, pp. 139142) supports our interpretation of Keynes's second property. Kcyncs (1936, p. 241) does note, however, that in earlier stages of history land may have borne the liquidity premium that we now associate with money. 3 An explicit discussion of the Pigou effect can be found on pp. 232-233 of The General Theory. Also see Leijonhufvud (1968, pp. 315-351) who argues that Keynes was well aware of and had addressed Pigou effect-like arguments.
2 1

Keynes's theory of money and interest

385

premium over carrying costs, in the long run that asset would become money and the third property would still hold. 1 The excess of money's liquidity premium over its carrying costs determines the money rate of interest. Earlier in chapter 17 (pp. 225-226), Keynes argues that the own-rate of interest on an asset is determined by the sum of its physical yield and liquidity premium, minus carrying costs. Following Keynes's (p. 226) assumption that the expected physical 'yield' or 'output' of money is zero, the rate of interest on money loans is driven by arbitrage to the difference between money's liquidity premium and carrying costs.2 Combining essential properties two and three generates downward stickiness in the nominal rate of interest on money loans. If money's liquidity premium is unresponsive and determines the nominal rate of interest on money loans (money's carrying costs are negligible), the nominal rate of interest on money loans cannot fall below this level. Only changes in the rate of inflation or deflation will succeed in changing real interest rates. Money's liquidity properties explain why we pay particular attention to the rate of interest on money loans, rather than the rates of interest on wheat, land, or other non-monetary assets. The difference between liquidity premium and carrying costs is greater for money than for other assets. If not, the asset with a greater liquidity premium would itself become the money asset. It is the highest own-rate of return against which the production of new capital goods must compete. Prospective investors compare the expected rate of return on new capital goods with the rate of interest on money loans, the highest available own-rate of interest (Keynes, 1936, pp. 222-229). A literally fixed liquidity premium is not necessary for Keynes's argument. Keynesian results may still hold if the rate of return on holding money falls, but falls slowly (pp. 223, 224, 229). Keynes (pp. 229-230) argues: it is that asset's rate of interest which declines most slowly as the stock of assets in general increases, which eventually knocks out the profitable production of each of the others . .. the kind of money to which we are accustomed has some special characteristics which lead to its own-rate of interest in terms of itself as standard being more reluctant to fall as output increases than the own-rates of interest of any other assets in terms of themselves . . . To the extent that the established standard of value has these peculiarities [the essential properties of money discussed above], the summary statement, that is the money-rate of interest which is the significant rate of interest, will hold good. If the marginal return on real balances declines slowly after a negative shock to capital productivity, the price level must fall significantly. Only a large increase in the quantity of real balances, which pushes agents down the marginal return of money curve, will restore equilibrium between rates of return on holding money (and thus money loans) and on producing capital goods. Until the price level adjusts and the liquidity premium on money falls, the rate of interest on money loans will remain above the marginal efficiency of capital and investment and output will fall. Keynesian results hold if the price level does not decline because individuals are slow to cut their spending; that is, if individuals are slow to adjust their demand for real balances. Liquidity traps A sticky liquidity premium on money generates results similar to the well-known liquidity trap, although for different reasons. The liquidity trap occurs when individuals are
1 Keynes (1936, p. 240) states: T h e conception of what contributes to "liquidity" is a partly vague one, changing from time to time and depending on social practices and institutions.' On the role of conventions in Keynes, see Lawson (1989, 1990). 2 Likewise, the own-rates of return on wheat (the difference between futures and spot prices, or the 'basis') are determined by the difference between wheat's liquidity premium and carrying costs; see Williams (1987).

386

T. Cowen and R. Kroszner

reluctant to move out of money and into bonds because they fear a capital loss on bonds. Keynes believes that if nominal interest rates are low, individuals expect a forthcoming rise in interest rates and hold money as a form of wealth insurance (the 'speculative demand formoney 5 ). Even Keynes (1936, p. 207), however, thought that this form of the liquidity trap was a theoretical curiosity. The constant or unresponsive liquidity premium discussed above generates a more general liquidity trap without requiring a special (and unsatisfactory) structure of expectations. Even if individual demand for cash balances is neither literally insatiable nor based upon a specific set of interest rate expectations, the stickiness of money's liquidity premium determines the nominal rate of interest on money loans nonetheless. The LM curve will be nearly horizontal over an economically significant range. Friedman (1989) is correct in his long-standing insistence that underemployment equilibrium requires a liquidity trap, although the liquidity trap need not be interpreted narrowly in the traditional fashion.1
Essential properties of money and the trade cycle

The stickiness of the rate of interest on money loans can generate contractions in output and employment when the marginal efficiency of capital fluctuates. Keynes (1936, p. 313) notes that: . . . the essential character of the Trade Cycle and, especially, the regularity oftime-sequenceand of duration which justifies us in calling it a cycle, is mainly due to the way in which the marginal efficiency of capital fluctuates. The Trade Cycle is best regarded, I think, as being occasioned by a cyclical change in the marginal efficiency of capital . . . A decline or collapse in the marginal efficiency of capital will imply a decrease in capital investment, because the purchase of new capital goods cannot compete with the rate of return on money loans. At this point, it is no longer profitable to produce additional quantities of capital goods, unless the rate of interest on money falls pan passu (Keynes, p. 228). A fall in the money rate of interest, however is prevented by the unresponsiveness of money's liquidity premium. 2 For this reason, Keynes notes repeatedly that the rate of interest is too high (e.g. pp. 324, 326-329). The undesirability of shifts from investment demand to money demand rests upon the presupposition that private and social returns to holding real money balances differ (Meltzer, 1988). Although physical output declines in the downturn, the private value of real balance holdings increases. The community as a whole is worse off, however, because the social return on additional real money holdings is less than the private return. Keynes (p. 155) is well known for his insistence that the community as a whole cannot increase the liquidity services it enjoys. The social value of increases in real balances is either absent or too small to offset the value lost through the fell in investment and output. 4. Why a sticky rate of return on money? A rapidly diminishing marginal return on cash balances would take much of the force out of Keynes's argument. The rate of return on money would no longer determine the
Arguments presented in Darity (1988) are consistent with out interpretation of the liquidity trap. Keynes (1936, pp. 315-316) admits that liquidity preference may be actually increasing at this point because the future is becoming more uncertain, but that the fall in the marginal efficiency of capital is the primary force behind the downturn. Considering how animal spirits may affect money demand thus strengthens the basic dilemma.
2 1

Keynes's theory of money and interest

387

nominal rate of interest on money loans because money's marginal return would be responsive to the rate of return on other assets. Equilibrating responses to changes in the marginal efficiency of capital could consist partially of price level declines (i.e. real cash balance adjustments), in lieu of adjustments in the quantity of investment. Thus, if the marginal efficiency of capital falls, then only a small decline in the price level is required to restore equilibrium. The empirical evidence on money demand, while subject to ongoing debate, appears to support the proposition that velocity adjusts to shocks very slowly. In a money demand regression, the coefficient on lagged money demand is large and often close to one Qudd and Scadding, 1982, pp. 996, 1016-1020). Money demand responses to changes in economic conditions appear to be slow and uncertain. Slow adjustments in money demand are especially likely if the marginal utility of money curve is relatively flat. If adjusting money demand fully involves even a small cost of adjustment or calculation, individuals may vary money demand sluggishly with a relatively flat return curve. Rational behaviour would dictate simple rules of thumb or perhaps complete inaction in response to changing constraints. 1 Within this framework, payment of interest on cash can be seen as an attempt to place the burden of adjustment upon well-informed profit-maximising intermediaries, who adjust price, rather than upon ill-informed money customers, who must adjust quantity. Note that even if velocity adjusts rapidly, an increased demand to hold money balances may not be a costless equilibrating response to a decrease in the marginal efficiency of capital. Although money's non-pecuniary return could not decrease the quantity of new investment directly, the structure of money's liquidity premium would force the economy through deflation each time there was a negative shock to the marginal efficiency of capital. This source of Keynesian problems, however, does require additional sources of wages and price stickiness in the economy.

5. Other sections of The General Theory: money stamping and mercantilism


Keynes's emphasis of the zero non-pecuniary return on money is brought out by his discussion (1936, p. 234) of'money stamping". Keynes (p. 234) notes that, those reformers, who look for a remedy by creating artificial carrying-costs for money through the device of requiring legal-tender currency to be periodically stamped at a prescribed cost in order to retain its quality as money, or in analogous ways, have been on the right track .. .

In his later discussion of the monetary reformer Silvio Gesell, Keynes (p. 358) notes that the burden of these carrying costs should be adjusted to account for the value of the note, the length of time the note is held, and the discrepancy between the rate of return on money and the marginal efficiency of capital in disequilibrium. Money stamping would create a medium of exchange whose rate of return could be varied in accordance with real rates of return. The rate of return to holding cash balances could, in principle, be continually adjusted to maintain market equilibrium. Decreases in
1 Yeager (1976) argues that lags in money demand adjustment do not contradict maximising behaviour. In a world where individuals adjusted their money demands instantaneously, individuals would also have no reason to hold money. The 'money in the utility function' approach (Patinkin, 1965; Friedman, 1969) and transactions and precautionary money demand models (McCalhim, 1983) generate a downward sloping demand for real balances, but offer no necessary implications about the steepness of this curve.

388

T. Cowen and R. Kroszner

the marginal efficiency of capital, for instance, would be accompanied with an increase in the money tax. Individuals would not forsake investment for the haven of cash balances. Keynes's embrace of Gesell's idea indicates his desire to circumvent the potential stickiness of money's rate of return.' Without change in the level of real balances, stamping money can change the rate of return on holding any given quantity of real balances. In contrast, with real balances in the utility function and a diminishing marginal utility of real balances, only a change in the level of real money holdings, ctteris paribus, can alter the marginal utility of real cash balances. A first-best money stamping scheme, by providing an instrument for varying the return on a given level of real money holdings, could offset the linkage between real and monetary sectors which Keynes identified. Furthermore, stamping also allows the application of a negative nominal return to cash holdings, when necessary (i.e. when the real rate of return on capital becomes negative). Keynes's discussion of mercantilism (1936, pp. 333-351) also draws upon insights related to his theory of the essential properties of money. The mercantilists claimed that increasing the inflow of specie would lower interest rates and stimulate investment, an argument which Keynes endorses. Keynes (p. 336) notes that this result holds if the state of liquidity preference is 'somewhat stable', a condition analogous to a constant marketability premium. The inflow and outflow of specie produce exogenous changes in the price level which affect the real rate of return on holding money. When specie flows in, the equilibration of the demands for money and return-bearing assets necessitates a fall in the real rate of interest. Keynes argues that the mercantilist 'fear of goods' and desire for a positive trade balance may not have been irrational, because of real interest rate effects. 6. Concluding remarks We have argued that stickiness in the nominal rate of interest can drive or exacerbate business cycle fluctuations, with this stickiness following from the nature of money's, liquidity premium. Keynes's discussion of the 'essential properties of money' can provide microfoundations for this stickiness. Our interpretation also integrates and renders coherent aspects of The General Theory which frequently have been dismissed: Keynes's theory of interest rate determination, and his endorsements of government control of investment, money stamping, and mercantilism. The absence of pecuniary interest on cash balances is the relevant imperfection that allows these results to be derived. Keynesian results of the sort described in this paper would be less likely if the total return on real money balances were flexible through payment of pecuniary interest. In the absence of such fundamental changes in monetary institutions, however, Keynesian results possesses at least one set of microfoundations. Bibliography
Beckerman, P. 1983. Non-positive market-clearing real rates of interest, Journal of Post Keynesian Economics, vol. 6, Fall, 5363 Black, F. 1970. Banking and interest rates in a world without money, Journal of Bank Research, vol.2, Autumn, 9-20
1 One can interpret the money stamping suggestion as an attempt to make money more like other assets by imposing carrying costs. In contrast, Cowen and Kroszner (1994) examine the possibility of making other assets more like money by the repeal of legal restrictions onfinnnrinlintermediation.

Keynes's theory of money and interest

389

Carmichael, J. and Stebbing, P. 1983. Fisher's paradox and the theory of interest, American Economic Review, vol. 73, September, 619-630 Cartelier, J. 1988. Keynes's General Theory: foundations for a heterodox political economy? in Barrere, A. (ed.), 77K Foundations of Keynesian Analysis, New York, St Martins Press Conard, J. 1959. An Introduction to the Theory of Interest, Berkeley, University of California Press Cottrell, A. and Lawlor, M. 1991. 'Natural Rate' mutations: Keynes, Leijonhufvud and the Wicksell connection, History of Political Economy, vol. 23, Winter, 625-643 Cowen, T. and Kroszner, R. 1987. The development of the new monetary economics, Journal of Political Economy, vol. 95, June, 567-590 Cowen, T. and Kroszner, R. 1989. Mutual fund banking: a market approach, Coto Journal, vol. 10, Fall, 223-237 Cowen, T. and Kroszner, R. 1994. Explorations in the New Monetary Economics, Oxford, Basil Blackwell Darity, W. J. 1988. Review of Pascal Bridel's Cambridge Monetary Thought, History of Political Economy, vol. 20, Winter, 690-694 Davidson, P. 1978. Money and the Real World, 2nd edn, London,Macmillan Davidson, P. 1980. The dual-faceted nature of the keynesian revolution: money and money wages in unemployment and production flow prices, Journal of Post Keynesian Economics, vol. 3, 291-307 Fried, J. and Howitt, P. 1983. The effects of inflation on real interest rates, American Economic Review, vol. 73, December, 968-979 Friedman, M. 1969. The Optimum Quantity of Money and Other Essays. Chicago, Aldine Friedman, M. 1989. The quantity theory of money, in Money: The New Palgrave, New York, Norton Goodhart, C. 1986. How can non-interest-bearing assets co-exist with safe interest-bearing assets?, British Journal of Economic Issues, vol. 8, August, 1-12 Hansen, A. 1953. A Guide to Keynes, New York, McGraw-Hill Hoover, K. D. 1988. Money, prices and finance in the new monetary economics, Oxford Economic Papers, vol. 40, 150-167 Judd, J. P. and Scadding, J. L. 1982. The search for a stable money demand function, Journal of Economic Literature, vol. 20, no. 3, 993-1023 Kaldor, N. 1960. Keynes's theory of the own-rates of interest, in Essays on Economic Stability and Growth, Glencoe, D, Free Press Keynes, J. M. 1936. General Theory of Employment, Interest, and Money, London, Macmillan Kuhn, T. 1977. The Essential Tension, Cambridge, MA, MIT Press Kydland, F. and Prescott, E. 1982. Time to build and aggregate fluctuations, Econometrica, vol. 50, November, 1345-70 Laidler, D. E. W. 1993. The Demand for Money: Theories, Evidence and Problems, 4th edn, New York, Harper Collins Lawson, T. 1989. 'Keynes and Conventions', unpublished manuscript, University of Cambridge Lawson, T. 1990. Keynes and the analysis of rational behaviour, in O'Donnell, R. M. (ed.), Keynes as Philosopher-Economist, London, Macmillan Leijonhufvud, A. 1968. On Keynesian Economics and the Economics of Keynes, New York, Oxford University Press Lemer, A. 1952. The essential properties of interest and money, Quarterly Journal of Economics, vol. 66, 172-193 Long, J. andPlosser, C. 1983. Real business cycles, Journal of Political Economy, vol. 91, February, 39-69 McCallum, B. 1983. The role of overlapping generations models in monetary economics, Carnegie-Rochester Conference Series on Public Policy, Spring, 9 14 McCulloch, J. H. 1986. Beyond the historical gold standard, in Campbell, C. and Dogan, W. (eds), Alternative Monetary Regimes, Baltimore, MD, John Hopkins University Press Meltzer, A. 1988. Keynes's Monetary Theory: A Different Interpretation, New York, Basil Blackwell Mort, T. 1989. A post-Keynesian perspective on a 'Cashless competitive payments system,' Journal of Post Keynesian Economics, vol. 11, Spring, 360-369

390

T. Cowen and R. Kroszner

Nell, E. 1983. Keynes after Sraffa: the essential properties of Keynes's theory of interest and money, in Distribution, Effective Demand and International Economic Relations, New York, St Martin's Press Patinkin, D. 1965. Money, Interest, and Prices, New York, Harper and Row Robinson, J. 1961. Own-rates of interest, Economic Journal, vol. 71, March, 253-271 Rogers, C. 1989. Money, Interest, and Capital, New York, Cambridge University Press Stiglitz, J. and Weiss, A. 1981. Credit rationing in markets with imperfect information, American Economic Review, vol. 71, June, 393-410 Summers, L. H. 1983 The nonadjustment of nominal interest rates: a study of the Fisher effect, in Tobin, J. (ed.), Macroeconomics, Prices, and Quantities, Washington, DC, Brookings Institution Townsend, R. 1987. Asset-return anomalies in a monetary economy, Journal of Economic Theory, vol. 41, April, 219-247 Turvey, R. 1966. Does the rate of interest rule the roost? in Hahn, F. and Brechling, F. (eds), The Theory of Interest Rates, London, St Martins Press Vicarelli, F. 1984. Keynes: The Instability of Capitalism, Philadelphia, University of Pennsylvania Press Wallace, N . 1983. A legal restrictions theory of the demand for money and monetary policy, Federal Reserve Bank of Minneapolis Quarterly Bulletin, vol. 7, 1-7 Wallace, N . 1988. A suggestion for oversimplifying the theory of money, Economic Journal, Conference Volume, 25-36 White, L. H. 1987. Accounting for non-interest bearing currency, Journal of Money, Credit and Banking, vol. 19, November, 448-456 Wilcox, J. A. 1983. Why real interest rates were so low in the 1970s, American Economic Review, vol. 73, 44-53 Williams, J. 1987. The Economic Function of Futures Markets, Cambridge, Cambridge University Press Yeager, L. B. 1976. Bootstrap inflation, Journal of Finance, vol. 31, March, 103-112

You might also like