In a thoughtful post, David Glasner comments on Austrian Business Cycle Theory (ABCT) — what I like to call the Mises–Hayek theory of industrial fluctuations —, offering two points of criticism. The critique is actually a response to George Selgin, who advances a case against excessively low interest rates within the context of NGDP targeting. How far the conversation has deviated from this latter topic is a sign of how different the frameworks of Austrians and Market Monetarists actually are. In any case, Glasner’s critique is unique in that it brings new arguments to the table — a move away from the more typical “rational expectations” (Lachmann 1943, p. 23) case or why the bust causes unemployment, but the boom does not. It certainly is worth considering, and I hope I can do it justice by attempting to correct what I see as some rather grievous errors in the analysis.

Glasner presents ABCT as a story principally about interest rates, which is sensible since Selgin had originally emphasized this particular variable in post Glasner was commenting on. It’s worth mentioning that much of the literature on the topic also stresses the role of the interest rate, since the theory relies on the Wicksellian concept of a natural rate of interest (that Austrians, such as F. Fetter and L.v. Mises, would later modify). Typically, the theory is presented as a case where the “market rate of interest” is depressed below the natural, or equilibrium, rate of interest, introducing a large quantity of credit into the economy. This credit is used to fund unsustainable investment projects, and once the credit is cut short all these businesses collapse, causing a business cycle. I can hardly criticize Glasner for singling out interest as the key variable.

Two separate arguments are posed against the Mises–Hayek theory,

1. That, originally, it did not deal with physical constraints to investment, but instead argues that the bust is caused by limits to credit expansion. Glasner suggests that there is no good reason to assume that central banks are limited in their ability to supply money;
2. The cessation of credit expansion and the consequent crisis should cause a fall in the prices of capital goods, rendering failing firms profitable again.

The first — as Robert Murphy has already pointed out — is a misreading of both L.v Mises and F. Hayek. The second point is more reasonable, but: (a) fails to consider some of the implications of ABCT  with regards to the valuation of boom-time demanded capital goods; (b) is a phenomenon which Hayek considered as early as 1937 (Hayek 1937, pp. 174–177).

Before responding to the two criticisms directly, I should present ABCT with a different emphasis — credit expansion. The key driver of intertemporal misallocation isn’t the rate of interest, but changes in relative prices caused by credit being injected by the banking system. Interest is relevant only in that if we assume that prior to credit expansion there exists an equilibrium in the loanable funds market, such that a lowering of the rate will increase the quantity demanded for credit. But, it isn’t the price that decides the distribution of resources throughout the structure of production. Rather, consequent malinvestment is demand driven.

In fact, it’s changes in relative prices that is the major theme of the first comprehensive exposition of the theory, Hayek’s Prices and Production. Following C. Menger and, especially, E.v. Böhm-Bawerk, Hayek divided the structure of production into “stages” based on the “distance” the output is from the final consumer good. Following this notation, the consumer goods stage is the first (although, it is the “last” in the sense of temporal sequence), the one immediately preceding that is the second, then the third, and so on. The basic idea elucidated in the book is that the initial credit injections, loaned first to entrepreneurs, will be used to invest in the second stage, driving up demand for third stage capital goods. Resources will be re-allocated to the third stage, in turn increasing demand for fourth stage capital goods, and the same process will repeat itself. It’s important to remember that Hayek’s exposition assumes that the structure of production starts in a position of equilibrium, so the pattern of resource allocation in the real world need not parallel this early model.

Hayek re-affirms this particular interpretation of his theory several times. In his reply to Sraffa (Hayek [1932]), for instance, he argues that the most important aspect of his theory isn’t the rate of interest and its derivation, but rather the “phantom profits” created by credit expansion. He does so again in his reformulation of the Mises–Hayek theory in his 1939 monograph Profits, Interest, and Investment — it is explicit in his differentiation between the rate of interest and the rate of profit, which otherwise would be equal. It’s worth mentioning that Mises also sees credit, not interest, as the main vehicle in the process of malinvestment. In Human Action, for example, Mises wrote,

If one wants to know whether or not there is credit expansion, one must look at the state of the supply of fiduciary media, not at the arithmetical state of interest rates.

—p. 556.

This takes us right to the causes of liquidation during the subsequent bust. The process of credit expansion causes the re-allocation of resources towards industries deemed profitable. The distributional process of profit and loss is essentially distorted, since what would otherwise be unprofitable is made profitable by “artificial” credit expansion. When the source of profit, or new credit, dries up, the firms which relied on it — mostly indirectly, in the form of demand for their products by later stages of production — become unprofitable, revealing what M. Rothbard (2000 [1963]) called a “cluster of error.”

While I have emphasized the monetary aspect of the Mises–Hayek theory, there is a real resource constraint. It’s implicit in the argument, in that there is a rightward shift in the demand curve for capital goods during the boom, rather than a rightward shift in the supply curve. Hayek uses other graphical expositions to make the same point; my favorite, since it does a good job of illustrating the real resource constraint is from Hayek (1934), p. 154 — reproduced on the right (click to enlarge). The dotted curves, and the area between them and the fully drawn curves, represents the investments that cannot be completed due to the shortage of capital goods (thus, driving their prices up).

Mises and Hayek hold that in order for the boom to continue, the supply of credit needs to increase at an accelerating pace. There needs to be enough new money to cope with rising prices, otherwise the malinvestments will reveal themselves. It’s also worth noting that they posited that injected credit would eventually find its way to the original factors of production, viz. labor, allowing for an upward force against the price of capital goods. What if there is no check to credit expansion? Writes Mises, “If the credit expansion is not stopped in time, the boom turns into the crack-up boom; the flight into real values begins, and the whole monetary system founders” (Human Action [1998 (1949)], p. 559). Surely, if sustaining the structure of production requires an accelerating rate of fiduciary expansion then there is a limit to credit expansion in terms of confidence in the currency.

This leaves us with Glasner’s second objection. To repeat, he argues that if the price of capital goods falls during the bust, this must mean that investments will return to profitability (sure, their incomes might fall, but so do their costs). A response is tricky, because in a sense Glasner is correct. But, the fall in prices is only a symptom of the much more important phenomenon at work.

The value of capital goods is imputed from the final consumer good, meaning that people find capital goods useful only as a means of achieving the final good somewhere down the road. If there is no final good, the capital good ceases to be a good (Menger [2007 (1871)], pp. 63–67). Recall that during the boom there is an increase in demand for capital goods caused by the credit injections and the structure of production stretches as this demand is imputed farther back to even earlier stages. Without credit expansion, many of the capital goods demanded and produced no longer carry any value, because their raison d’être
vanishes. The firms in the business of manufacturing these goods will need to liquidate, because the price of their output is essentially driven to zero.

In Prices and Production, Hayek talks about fixed and circulating capital, and the spectrum between specific and non-specific goods. Specificity refers to the range of complementarity that a particular good has; more specific goods are useful only towards a much more narrow range of ends. Mises, in Human Action, also argues that the quantity of “fixed” capital will rise during the boom. These types of outputs will be the most inflexible in terms of returning to profitability. Also, production taking place farther away from the consumer good will be subject to more pressure to liquidate, since the structure of production will tend to shorten and narrow.

Nevertheless, Hayek is clear that some of this boom-time structure of production will become “permanent” — it will remain, despite the liquidation. In Prices and Production, he notes that the readjustment will have to consider fixed capital. In his 1937 piece (linked above), “Investment that Raises the Demand for Capital,” he argues that some firms will continue to exist, even with decreases in profitability, meaning that some of the capital goods created during the boom will retain some value. The conclusions drawn in the 1937 article are really an extension of the ideas presented in his seminal 1935 contribution, “The Maintenance of Capital.” But, clearly, the effects of the disappearance of “phantom profits” will lean towards liquidation, more so than simply reduced profitability.

In the final paragraph of his critique, Glasner argues that that the structure of production need not face reorganization “as long as [the] economy is not required … to undergo a contraction, in total spending.” Some might argue that Hayek agrees with Glasner here. Certainly, even as early as Prices and Production Hayek advocated “neutral” money, or equilibrium between the supply of and demand for money. In later life, he would more adamantly declare his agreement with economic historians like Milton Friedman, who advanced the widely-acclaimed thesis that the Great Depression was caused by monetary contraction. If this is this is the case, this is where I part ways with Hayek.

To better illustrate this part of the bust, it might be worthwhile to summon I. Fisher’s 1933 article, “The Debt-Deflation Theory of Great Depressions.” While not perfectly compatible, the Austrian theory of “secondary deflation” is similar. If the principal method of credit injection is through the loanable funds market, the new credit also represents debt — a debt owed by the entrepreneur to the originator bank (or, whoever owns the debt, to account for a modern banking system). The liquidation of a firm through bankruptcy oftentimes means that the full value of the loan is never repaid, essentially decreasing the quantity of money. This is simply excess fiduciary media ceasing to exist. Of course, this represents the contraction in total spending that Glasner fears, but it’s not something monetary authorities can control (or ought to control — replace distortion-causing excess fiduciary media with more of it?). In clearer words, the secondary deflation is simply an inevitable outcome of the malinvestments caused by previous credit expansion.

In Profits, Interest, and Investment Hayek proposes a much more sensible solution to secondary deflation: an increase in savings. A decrease in the demand for consumer goods in the present will liberate resources for the earlier stages of production, thirsting for capital goods. I have argued before that an increase in the demand for money serves this purpose; I call it a “cushion.”

While Glasner’s critique raises several good points, ultimately he fails to really engage the Austrian theory. In large part, it’s mostly a misunderstanding of the implications of ABCT; oftentimes, in Glasner’s defense, confusion can be stimulated by how the theory has been developed and explained in the past. Hopefully, this comment serves to clear the confusion and put the debate back on track.