Volume 17, No.1 (Spring 2014)
J. Huston McCulloch[1]
Ludwig von Mises Memorial Lecture
Austrian Economics Research Conference
Ludwig von Mises Institute
Auburn, Alabama
March 22, 2014
It comes as a great honor to have been asked to present the Ludwig von Mises Lecture at the 2014 Austrian Economics Research Conference. One of the first books that I read on economics many years ago was his Theory of Money and Credit ([1924] 1953). Other works that had a great influence on my view of economics were his Human Action (1963), The Free and Prosperous Commonwealth (also known as Liberalism, 1962), Omnipotent Government (1944), and Theory and History (1957).
Although von Mises would disapprove of my mathematical and statistical proclivities, his works contain several insights that I think are very relevant to modern economics and should be better known. Today I would like to discuss four of these Misesian insights that pertain to macroeconomic issues.
The first of these insights is von Mises’s concept of the historical transmission of the value of money. An immediate consequence of this is a model of the mechanism by which prices adjust to a change in the money supply.
A second insight, which ultimately also arises from von Mises’s historical transmission concept, is that of market equilibrium, as contrasted to the fashionable concept of “rational expectations” equilibrium. Although “rational expectations” (or more accurately, “equilibrium expectations”) is a useful exercise for evaluating the internal consistency of policies, it is a completely unrealistic model of how the economy actually works.
The third Misesian insight I wish to discuss is that of heterogeneous, inconvertible capital. This contrasts sharply (and realistically) with the popular homogenous capital assumption of the “neoclassical growth model.”
And finally, the fourth Misesian insight I will discuss is the nature of the liquidity effect, and how it relates the so-called “Taylor Rule” to the Quantity Theory of Money.
1. THE HISTORICAL TRANSMISSION OF THE VALUE OF MONEY
In the 1924 second edition of The Theory of Money and Credit (trans. 1953), von Mises addressed the criticism that his contemporary Karl Helfferich has raised against using marginal utility theory to explain the level of money prices:
Helfferich is of the opinion that there is an insurmountable obstacle in the way of applying the marginal utility theory to the problem of money; for while the marginal-utility theory attempts to base the exchange value of goods on the degree of their [usefulness] to the individual, the degree of [usefulness] of money to the individual quite obviously depends on its exchange-value, since money can have utility only if it has exchange-value, and the degree of [its usefulness] is determined by the level of the exchange-value. Money is valued subjectively according to the amount of consumable goods that can be obtained in exchange for it, or according to what other goods have to be given in order to obtain the money needed for making payments. The marginal utility of money to any individual, i.e., the marginal utility derivable from the goods that can be obtained with the given quantity of money or that must be surrendered for the required money, presupposes a certain exchange-value of the money; so the latter cannot be derived from the former. ([1924] 1953, pp. 119–120)
In other words, marginal utility theory states that the relative prices are equal to ratios of the corresponding marginal utilities. This works for the relative prices of say good i and good j as follows:
Pi/Pj = MUi/MUj 1
However, Helfferich argued that marginal utility theory is circular for the prices of goods in terms of money, since the nominal price of good i should be determined by the ratio of good i’s marginal utility toMU$, î , the marginal utility of money to be spent on goods other than i:
Pi = MUi/MU$, î
Yet MU$, î is determined by Pj , j ≠ i, while these are determined, in part, by Pi itself! Figure 1 below illustrates what I call this “Vicious Circle of Helfferich”:
Figure 1. The Vicious Circle of Helfferich
|
Von Mises replied to Helfferich’s argument as follows:
Those who have realized the significance of historically transmitted values in the determination of the objective exchange-value of money will not find great difficulty in escaping from this apparently circular argument.... It is true that the subjective valuation of money presupposes an existing objective exchange-value; but the value that has to be presupposed is not the same as the value that has to be explained; what has to be presupposed is yesterday’s exchange-value, and it is quite legitimate to use it in an explanation of that of to-day. The objective exchange-value of money which rules in the market to-day is derived from yesterday’s under the influence of the subjective valuations of the individuals frequenting the market, just as yesterday’s in its turn was derived under the influence of subjective valuations from the objective exchange value possessed by the money the day before yesterday. (120¬–121)
In other words, von Mises has in effect replaced the “Vicious Circle of Helfferich” with what I call the “Benign Helix of Mises,” illustrated in Figure 2 below. Here, the price of good i at time t, represented byPit , is determined not by Pjt j ≠ i, but by Pjt-λ , where λ is the average lag in price information. This in turn is determined by Pit-2λ , etc.
Figure 2. The Benign Helix of Mises
Image
|
Von Mises’s concept of the historical transmission of the value of money leads immediately to his theory of the adjustment of prices to an increase in the supply of money:
An increase in a community’s stock of money always means an increase in the amount of money held by a number of economic agents, whether these are the issuers of fiat or credit money or the producers of the substance of which commodity money is made. For these persons, the ratio between the demand for money and the stock of it is altered; they have a relative superfluity of money and a relative shortage of other economic goods. The immediate consequence of both circumstances is that the marginal utility to them of the monetary unit diminishes. This necessarily influences their behavior in the market. They are in a stronger position as buyers. They will now express in the market their demand for the objects they desire more intensively than before; they are able to offer more money for the commodities that they wish to acquire. It will be the obvious result of this that the prices of the goods concerned will rise, and that the objective exchange value of money will fall in comparison. But this rise of prices will by no means be restricted to the market for those goods that are desired by those who originally have the new money at their disposal. In addition, those who have brought these goods to market will have their incomes and their proportionate stocks of money increased and, in their turn, will be in a position to demand more intensively the goods they want, so that these goods will also rise in price. Thus the increase of prices continues, having a diminishing effect, until all commodities, some to a greater and some to a lesser extent, are reached by it ([1924], 1953, p. 139).
In modern terms, von Mises is describing what is known as a partial adjustment mechanism for the price level, with the excess supply of money driving the inflation rate. In McCulloch (1980), I show that under simplifying assumptions, this can be written as
(1) π = γ(log(M/P) - log mD) + πa + ε
Here, π = Ṗ/P is the inflation rate, M is the nominal money supply (appropriately measured), mD is real money demand, πa is anticipated inflation reflecting the fact that agents extrapolate past prices for obvious inflation trends when predicting current and future prices of goods not currently being purchased, and ε is a white noise error term caused by nonmonetary microeconomic shocks. The adjustment coefficient γ = ηm/(λw) depends on η, the elasticity of marginal utility with respect to wealth (also known as the relative rate of risk aversion), real money balances m, real wealth w, and the average lag in price information λ.
In McCulloch (1980), I call this Mises-inspired equation “the Moderate Quantity Theory of Money.” According to what I call the “Extreme Quantity Theory of Money,” prices adjust instantaneously to changes in the money stock, so that inflation simply equals concurrent money growth, net of real income growth, and plus velocity growth and white noise micro shocks. According to the “Moderate Quantity Theory,” however, inflation does not respond at all to concurrent money growth. In fact, if money supply equals money demand, inflation will be driven entirely by inflationary expectations plus micro noise. However, if these non-monetary factors drive prices away from their Quantity Theory equilibrium level, the first term will eventually begin to pull prices back toward their equilibrium level. The long run average inflation rate will therefore ultimately reflect the long run average money growth rate.
This “Moderate Quantity Theory” is a great improvement over the famous “real adjustment” equation of Cagan (1956), which was used by Chow (1966), Goldfeld (1973), and others to estimate the demand for money. According to the real adjustment equation, the rate of change of real money balances is proportionate to the gap between real money demand and real money balances. This turns out to be equivalent to (1), but with the expected inflation term replaced by the rate of money growth. So long as the money supply is constant, this will behave like a price adjustment equation in response to any changes in money demand or microeconomic shocks. However, it unrealistically predicts that the price level will perfectly track any changes in the money supply, just as does the Extreme Quantity Theory.2
The “P-star” model of Hallman, Porter and Small (1991) comes very close to the Mises-inspired equation (1). The only difference is that they did not provide micro foundations for their equation, and that they replace anticipated inflation by several lags of actual inflation. In practice it may be that a distributed lag of past inflation is the best available proxy for inflationary expectations, but it is better to make the inflationary expectations explicit, and only then to consider how best to proxy them empirically.3
2. MARKET EQUILIBRIUM VERSUS “RATIONAL EXPECTATIONS” EQUILIBRIUM
Von Mises’s theory of the historical transmission of the value of money leads directly to his concept of a Market Equilibrium, in which prices convey sufficient information about other agents’ tastes and endowments that agents’ actions are efficiently coordinated, without central planning or indeed any actual knowledge of other agents’ information. In an “evenly rotating economy” in which tastes and endowments are constant, the market will actually find the Walrasian equilibrium, without the help of a “Walrasian auctioneer” who calls out hypothetical price vectors. If tastes and endowments change over time but with a considerable element of continuity, the market will never perfectly reach the Walrasian equilibrium, but at least will be continually moving in its direction.
This theory of Market Equilibrium stands in sharp contrast to the fashionable “Rational Expectations Equilibrium” concept first proposed by Muth (1961), and elaborated on by Robert Lucas, Thomas Sargent, Neil Wallace, and others. According to this model, every agent knows the tastes, endowments, and production possibilities of every other agent, along with the intentions of policy makers, and then computes the equilibrium of the economy, to within random shocks not known to anyone.
As was pointed out by the Austrian economist Fritz Machlup, in an unpublished note written shortly before his death, the term “Rational Expectations” is an abuse of terminology, since “rational” simply means using one’s own information in a logical manner, and does not imply knowledge of other agents’ information, let alone the computationally impossible task of coordinating all this information.
“Equilibrium Expectations” would be a far more appropriate, and far less misleading, term for what Muth called “Rational Expectations.” As such, “Equilibrium Expectations” is a very useful exercise for evaluating the long-run implications of policies such as inflationary finance or monetary stimulation through the Phillips Curve, even if it is an entirely unrealistic model of how the economy actually works.
Proponents of “Rational Expectations” often make the fallacious argument that if economists believe that agents are rational, then they must accept that agents’ expectations are “Rational.” This is the logical fallacy of assigning one term two definitions, and then equating the two definitions. Some even accuse those who deny this equation of irrationality on their own part.
Today, even some of the early proponents of “Rational Expectations” are admitting that it imposes unrealistic informational assumptions on agents, and now instead propose what they call “Bounded Rationality” (e.g. Sargent, 1993). But this itself is yet another misleading misnomer: If a person is only 30 percent rational, then that person is 70 percent mentally incompetent. But if a person is only 30 percent omniscient, then he or she is 70 percent human. “Bounded Omniscience” would therefore be a better term for what these economists really have in mind than “Bounded Rationality.”
3. HETEROGENEOUS, INCONVERTIBLE CAPITAL
Mainstream macroeconomics typically treats “capital” as a homogeneous good whose structure, if any, is only relevant for microeconomic questions. Thus, the popular “neoclassical growth model” typically contains an equation according to which consumption plus investment equals output net of depreciation:
C + ΔK = f(K,L) - δK
The Austrian economists Ludwig von Mises and F.A. Hayek instead stressed that capital (“the produced means of production” as Böhm-Bawerk put it) is to a great extent specific to the intended ultimate product, and thus is both heterogenous and inconvertible. In their “Austrian” theory of the business cycle, they emphasized that goods available at different points in time are economically different goods. Following Böhm-Bawerk, the “more roundabout means of production” are generally more productive, in part because they permit the use of technologies that involve types of capital that would not have time to be constructed for less roundabout production plans. The inconvertibility of this horizon-specific capital plays a central role in the Austrian business cycle theory.
Von Mises himself made no attempt to quantify this concept, either mathematically or graphically. In his Pure Theory of Capital, Hayek (1975) attempted to illustrate it graphically, but as he confined himself to unrealistically simplistic technologies, he was not very successful.
In fact, this Misesian concept is easily illustrated in terms of the “Production Possiblity Frontier” (PPF) of the early marginalist F.Y. Edgeworth. This tool was adapted to intertemporal production by Irving Fisher in his very Böhm-Bawerkian Theory of Interest (1930).4
Figure 3 below shows a typical PPF for the case of two outputs, X and Y. Points A and B on the PPF represent two output combinations that are feasible given the available inputs. In the elementary exposition of this PPF, there are two inputs, homogeneous capital K and labor L, and two concave, constant returns to scale production functions with differing factor intensities. The resulting PPF bends away from the origin as illustrated.5
Figure 3. A Production Possibility Frontier (PPF) with Two Outputs, X and Y
|
But suppose now that instead of being pre-existing and homogeneous, capital must be produced, and that the types of capital tools and equipment that one would ideally use to produce X are different from those that would be used to produce Y. Production now has at least two stages and involve three points in time: At time t=1, the initial endowment of resources is used to produce specialized tools and equipment that will become available at time t=2. Then at time t=2, these produced means of production are combined with non-produced factors like land and labor to produce X and/or Y at time t=3.
The t=1 PPF over X and Y will still be convex as in Figure 3, but now the t=2 PPF will depend on which combination of X and Y was targeted back at t=1. If point A was targeted at t=1, A will still be feasible at t=2. However, point A will be the only point on the original PPF that is still feasible. Elsewhere, the new t=2 PPF will lie strictly inside the original PPF, as illustrated by the red curve in Figure 3. In particular, point B will no longer be attainable. On the other hand, if point B was targeted back at t=1, only B will still be attainable at t=2, as illustrated by the blue curve in figure 3. In mathematical terms, the elasticity of transformation of the PPF declines as time passes and production is committed to a particular point on the original PPF. (The elasticity of transformation is the negative of the elasticity of substitution, defined along a PPF instead of an isoquant or indifference curve.)
Figure 4.
|
The situation illustrated in Figures 3 and 4 is basically a microeconomic problem, and not a macroeconomic one: If at t=1 one builds umbrella factories instead of sunscreen factories, but at t=2 if it is realized that the demand will be for sunscreen instead of umbrellas, one will have lost the capability to produce as much sunscreen as would have been feasible ex ante. However, there is not too much or too little demand for goods as a whole, just too much or too little demand for umbrellas versus sunscreen.
In an intertemporal context, however, it is possible to have too much or too little demand for present goods as a whole relative to future goods, their relative price being governed by the real interest rate. In order for malinvestment to be an issue, there must have been a third, earlier period when horizon-specific capital investments were made. Figure 5 below illustrates a general PPF for aggregate consumption possibilities C1,C2, and C3 in a world with three periods t1, t2, and t3. At time t1, producers will target an output stream like point A that maximizes the present discounted value of future output, given the real interest rates that are then current.
Figure 5. A Three-Period Intertemporal PPF
|
During the second period t2, the chosen value of C1 will have already been produced and consumed, so that it can no longer be changed. Figure 6 below shows the section through Figure 5 corresponding to the C1 value of point A. Figure 7 shows this section by itself without the C1 axis.
Figure 6. Time t1 PPF with Section at Level of C1 Corresponding to Point A
|
Figure 7. Section of t1 PPF Shown in Figure 6, Without C1 Axis
Image
|
Although any point on the t1 PPF section shown in Figure 7 could have been produced if one had begun back in t1, by t2 it is too late to produce any point on it but A. It is not necessary to follow through with A itself, but any other combination of C2 and C3 will be inside the original PPF, as shown in Figure 8 below.
Figure 8. The “Austrian Capital Effect”
|
This “Austrian Capital Effect” was first illustrated in this manner in my 1981 article, “Misintermediation and Macroeconomic Fluctuations.” This diagram shows how intertemporal malinvestment imposes real costs on the economy. The problem is not overinvestment or underinvestment per se, since the level of C1 is not necessarily incorrect. Rather it is that producers have invested in the wrong structure of capital, given consumers’ preferences over C2 and C3.6
It should be noted that the interest rate that most directly affects the t1 choice of C2 relative to C3 is the tj forward interest rate linking these two points in time, and not the t1 general level of interest rates per se. Unfortunately, Mises and Hayek did not take the term structure of interest rates into account, as is required to make this nuanced point.
4. THE LIQUIDITY EFFECT
A popular misconception in the mainstream macroeconomic literature is that the “Liquidity Effect” of a monetary expansion is the reduction in real and nominal interest rates required to induce agents to holdthe new money, given the received price level and a real money demand function that has some interest-elasticity.
In fact, Mises would be the first to point out that the immediate effect of an injection of money through the banking system is the reduction in real (and therefore nominal) interest rates required to induce agents to borrow the new money from the banks with the primary intention of spending it, on consumption or investment. As Mises pointed out above, a disequilibrium excess supply of money will persist for some time after the monetary injection, and hence the disequilibrium liquidity effect will persist for the same length of time.
This correspondence between the excess supply of money and the disequilibrium liquidity effect can help us understand how the “Taylor Rule” can, at least in principle, be used to regulate inflation in a fiat money economy like the United States since 1968: If the central bank knows the demand for real money balances, it can use the nominal money supply and the Quantity Theory of Money to regulate the price level and hence inflation. On the other hand, if it knows the equilibrium real interest rate but has no trust in the stability of the money demand function, it can, in principle, still regulate the excess supply of money and hence inflation by manipulating interest rates along the lines of the “Taylor Rule.”
CONCLUSION
Ludwig von Mises’s writings contain many insights that are very relevant for mainstream macroeconomics. These insights can and should be accepted even by those economists who do not share Mises’s strong policy recommendations on the inevitable counter-productivity of intervention, the gold standard, and the precise nature of the business cycle.
At the same time, Austrian economists should strive to integrate these insights into mainstream economics, rather than isolating themselves from it.
REFERENCES
Böhm-Bawerk, Eugen von. 1959. Capital and Interest. South Holland, Ill., Libertarian Press.
Cagan, Phillip. 1956. “The Monetary Dynamics of Hyperinflation.” In Milton Friedman, ed., Studies in the Quantity Theory of Money. Chicago: University of Chicago Press.
Carr, Jack, and Michael R. Darby. 1978. “The Role of Money Supply Shocks in the Short-Run Demand for Money.” UCLA Discussion Paper No. 9B.
Chow, Gregory C. 1966. “On the Long-Run and Short-Run Demand for Money,” Journal of Political Economy 74: 111-31.
Gale, David. 1960. The Theory of Linear Economic Models. New York: McGraw-Hill.
Goldfeld, Stephen M. 1973. “The Demand for Money Revisited,” Brookings Papers on Economic Activity 3: 577–645.
——. 1976. “The Case of the Missing Money,” Brookings Papers on Economic Activity 6: 683–739.
Guo, Kevin, and J. Huston McCulloch. 2013. “Heterogeneous Capital and Misintermediation.” Central University of Finance and Economics, Beijing.
Hallman, Jeffrey, Richard D. Porter, and David H. Small. 1991. “Is the Price Level Tied to the M2 Monetary Aggregate in the Long Run?” American Economic Review 81: 841–858.
Hayek, Friedrich A. von. 1975. The Pure Theory of Capital. Chicago: University of Chicago Press.
McCulloch, J. Huston. 1977. “The Austrian Theory of the Marginal Use and of Ordinal Marginal Utility,” Zeitschrift für Nationalökonomie 37: 249–280.
——. 1980. “The Microfoundations of the Moderate Quantity Theory,” Ohio State University. Available at www.econ.ohio-state.edu/jhm/papers/MQT80.pdf.
——. 1981. “Misintermediation and Macroeconomic Fluctuations,” Journal of Monetary Economics 8: 103–115.
Mises, Ludwig von. 1960. Epistemological Problems in Economics. Trans. George Reisman. Reprinted by Ludwig von Mises Institute. Available at www.mises.org/epofe.asp.
——. 1962. The Free and Prosperous Commonwealth: An Exposition of the Ideas of Classical Liberalism. Princeton: Van Nostrand.
——. 1963. Human Action: A Treatise on Economics, 2nd ed. New Haven: Yale University Press.
——. 1944. Omnipotent Govenment: The Rise of the Total State and Total War. New Haven: Yale University Press.
——. 1957. Theory and History: An Interpretation of Social and Economic Evolution. New Haven: Yale University Press.
——. 1953. The Theory of Money and Credit, new enlarged edition. New Haven: Yale University Press, 1953. Parts 1–3 trans. H.E. Batson from the 1924 2nd. German edition. First German edition, 1912.
Muth, John. 1961. “Rational Expectations and the Theory of Price Movements,” Econometrica 29: 315–335.
Sargent, Thomas. 1993. Bounded Rationality in Economics. Clarendon Paperback.
- 1With discretely divisible goods, the Austrian version of utility theory actually admits the possibility of intrinsically ordinal marginal utilities which cannot be manipulated arithmetically. However, they will still place quantifiable bounds on marginal rates of substitution and therefore on prices. In this paper we assume that goods are continuously divisible, which in turn implies that utility is essentially cardinal and that marginal rates of substitution are exactly determined by ratios of (essentially cardinal) marginal utilities. See McCulloch (1977) for further clarification.
- 2This insight is due to my Boston College colleague Harold Peterson.
- 3Hallman et al. define their P* in terms of the M2 monetary aggregate, though this is not an essential part of their adjustment mechanism. The “nominal adjustment” equation of Goldfeld (1976) makes sense for a small open economy with a fixed exchange rate and an endogenous money supply determined by the specie flow mechanism. However, it makes little sense when the money supply is exogenous or determined by seigniorage considerations. Carr and Darby (1978) propose an equation like (1), but with expected inflation replaced by expected money growth. This term would only make sense in an “Extreme Quantity Theory” context in which inflation was driven contemporaneously by money growth.
- 4Fisher in fact dedicated his 1930 book to Böhm-Bawerk and to Böhm-Bawerk’s precursor John Rae.
- 5The Austro-Hungarian mathematician John von Neumann provided a very Menger-like proof of both diminishing marginal products and of the general convexity of the production possibility set for a very general constrained technology, as discussed in Gale (1960). In each stage of production, there is a nonnegative vector x of inputs, a nonnegative vector y of production activities, and a vector z of outputs. Activities are constrained by the vector inequality Ay ≤ x, where A is a matrix of nonnegative input requirements, while outputs one period later are given by x = By, where B is a matrix of nonnegative output coefficients. The Production Possibilities Set is then convex. Its PPF is piecewise linear in the case of two outputs, but in a complex economy with a vast number of inputs and activities, will be nearly smooth as in the illustrations. Two inputs are complements in Menger’s sense when they are both required by a single activity. Two inputs are substitutes when they are required by different activities that produce the same output. Complements tend to increase one another’s marginal product, while substitutes tend to decrease the other’s marginal product, so that two inputs may be said to be net complements or net substitutes depending on their net effect on one another’s marginal product. Outputs are joint if they are produced by the same activity or activities. This very Mengerian von Neumann model is differentiated from the oversimplified Input/Output model of Leontief by the separation of outputs from inputs and the passage of time. In the Leontief I/O model, for example, a chicken could be used to lay the very egg from which it itself hatches, while this is impossible in the von Neumann model. I have no direct knowledge that von Neumann was inspired by Menger, but it seems very likely that he was.
- 6Guo and McCulloch (2013) illustrate the Austrian capital effect with a simple agricultural economy that can choose between using labor to plant crops immediately or to build plows that will increase future productivity but delay consumption.
