About Austrian economics

I find Steve Horwitz, along with George Selgin (prominent advocate of free banking and supporter of a productivity norm [pdf] for monetary policy), the most accessible of contemporary Austrian school economists as they are both clear writers who seek to engage with those who are not of their school and are refreshingly free of the nastiness that so many Austrian school commentators seem prone to. (Little, if any, of what I have to say in the following applies to Selgin and some applies to general tendencies in Austrian commentary rather than Horwitz.)*

Horwitz has written a very useful paper on Hayek and Keynes’ different understanding of capital. Points that strike me reading the paper, and particularly Horwitz’s presentation of the Hayekian/Austrian concept of capital, include:

(1) The Austrian view of capital is over-impressed by differentiation, as when Horwitz writes:

What is central for the Austrian theory is that capital is not homogenous; capital goods are not perfect substitutes for one another. Any given good can only serve in a limited number of production plans, and it is not possible to create any given production plan out of any capital goods. Goods are not infinitely substitutable, and not all goods have the requisite complementarity necessary to be part of any particular production plan. This emphasis on the “heterogeneity” of capital distinguishes Austrian capital theory from many of its predecessors, especially those, most obviously Knight‘s “Crusonia plant” or Solows “shmoo,” that viewed capital as a homogenous fund of resources from which equally useful “ladles” could be applied to any production process

Yes, once resources are allocated to specific capital, they are difficult to shift to other uses. Nevertheless, the ongoing process of allocating resources to creating capital is important in its own right. Capital is not homogeneous (neither is labour; something which seems to figure rather less in Austrian analysis) but there is enough flow of resources in an economy that heterogeneity is not all there is to grapple with.

(2) Thus, there is, in the Austrian approach, a somewhat “frozen” view of capital: that it can be difficult to reallocate, does not make it impossible. As Horwitz notes later in the paper, the loss of value in capital no longer allocated to its original use measures how difficult but not impossible it is. Conversely, heterogeneity of labour would suggest that labour markets also have adjustment delays and constraints, which sits rather poorly with Austrian confidence in fully flexible wages if there were no regulatory interventions.

(3) At the same time, there is a perfectionist view of markets, that there is a “correct” arrangement of capital. As when Horwitz writes:

… the capital structure and the process of monetary calculation that drives it is the fundamental coordinating process of the market economy. Fitting those pieces together as correctly as possible, in response to knowledge and incentives produced by the pleasurable and unpleasureable beeps of profit and loss, is what ensures ongoing economic coordination and growth.

Yes, profit and loss direct resources to more valuable uses, but the notion of a single “correct” outcome glides over a whole lot of issues about incentives, constraints, information flows, etc. Yes, the “pieces” have to fit together, but they are constantly being created and replaced, with a fair bit of “good enough” going on because of various constraints, varied incentives, information asymmetries, etc.

(4) The emphasis on the role of money in economic calculation is somewhat one-sided, as when Horwitz writes:

What guides this process of plan formation, deconstruction, and reconstruction is monetary calculation. In a market economy where capital goods have money prices, those prices enable entrepreneurs to prospectively formulate budgets and retrospectively calculate profits and losses. Budgets based on those prices are what enable entrepreneurs to decide which capital goods will effectively serve as complements in an integrated production plan. After that plan has been executed, profits and losses signal owners of capital whether or not the plan was successful, which enables them to decide whether the uses of capital were, in fact, sufficiently complementary to continue. If not, then the money prices of other capital goods provide the information necessary to engage in another round of calculation and budgeting to see what sorts of capital goods might serve as substitutes for pieces of the failed plan.

Yes, the various “swap values” of money (for goods and services) matter, but what entrepreneurs are really interested in is expected income; that is, price x transactions. It is not that there is no concept of demand/expected income in the Austrian analysis, it is that, at crucial points in the argument, price is emphasized to the extent that transaction levels, and so income expectations, disappear from view. (David Beckworth has a nice post connecting collapsing transaction levels, and so income expectations, as reported by business survey respondents, with the economic downturn in the US; or, to put it another way, it’s the level of transactions,… .) Price is not a perfectly flexible lever; particularly given the different time scales between stages of production, to use Austrian language, and constraints in labour markets.

(5) Following on from (4), the fundamental role of money in an economy is to facilitate transactions, including across time. The typical Austrian emphasis on money-for-calculation, and presumption of state monopolies as over-suppliers of money, leads to an obsession with inflation and the risk of hyperinflation. (Hence my comment that an internet Austrian is someone who has predicted 10 of the last 0 bouts of hyperinflation.) As a matter of historical fact, deflation (and unexpected disinflation) can be much more destructive of economic activity than inflation. Both inflation and deflation interfere with money as a means of economic calculation, but deflation (and unexpected disinflation) also drives down income/economic activity, which makes is much more destructive. But the Austrian view of business cycles is so focused on finding reasons for busts in the previous booms (those money over-supplying central banks) that it, in effect, blames the effects of deflation on previous inflation. Yet inflation and deflation are equally monetary phenomena, one is not causally dominant over the other.

(6) The Austrian theory operates in terms of saving preferences, conceived as time preferences, as when Horwitz writes:

Hayek and the Austrians were working from a classical loanable funds theory of the interest rate. On this view, the interest rate was the price of time, emerging from the supply of loanable funds from savers and the demand for loanable funds by investors.

If interest rates are the price of time, where is risk? It is true that time preferences vary between people and periods of life. But so does risk aversion and assessments of risk. Shifts in risk assessments (and income expectations) seem generally a much more plausible reason for changes in interest rates (and asset prices generally) than some unexplained (what economists call ‘endogenous’) shift in time preference (which can be expected to be much more stable than risk assessments). Saying that risk assessments are subsumed into time preferences would be hand-waving to excuse the downplaying of risk assessments.

It is not that the Austrian picture denies risk, it is that, at a crucial point in the argument, the argument is mounted in terms of time preferences, which gives the prices-will-coordinate-so-well-there-can-be-no-general-glut-of-goods-and-services story more plausibility than it otherwise would have. Put things in terms of risk and then the possibility of a precipitous decline in spending which prices will not successfully compensate for becomes much more plausible. Particularly as sticky prices are typically a form of risk-management.

(7) Horwitz puts the Austrian business cycle theory on its best construction when he writes that Austrians:

look for explanations of recessions or “busts” by asking if the bust was preceded by a boom. A boom is not a necessary condition for a recession but it is sufficient, and Austrians argue that a preceding boom is the most frequent cause of a bust. Key to understanding that boom-bust cycle is how the capital structure gets distorted in the boom, which also helps to understand how Austrians view the bust.

Alas, there is no reliable correlation between the level of inflation in any boom and the scale of the subsequent bust. Despite claims to the contrary at the time, the 1920s was not particularly inflationary: certainly nowhere near enough to explain the depth of the 1930s crash. As George Selgin writes in his productivity norm essay (p.58):

Indeed, the relative inflation of the previous decade is only likely to have played a relatively minor part in explaining the length and severity of the depression, in contrast to its major role in causing the stock-market boom and crash.

Similarly, the contemporary Great Recession, the worst recession in the postwar era, has followed the least inflationary boom in the postwar era. There is simply no correlation between the level of preceding inflation and the level of the consequent bust. As noted above at (5), deflation is a monetary phenomenon in its own right. But if state monetary monopolies fail in specific, predictable ways (propensity to oversupply money), the argument for their abolition is easier than if their failures are far more erratic (sometimes oversupplying, sometimes undersupplying, implying that sometimes they get it right, a habit it might be possible to expand; there still is an argument for free banking, but it becomes more grounded in general supply-and-demand arguments applied to money which sit much less well with Austrian business cycle theory).

(8) Following on from (7), there seems to be such a strong wish to see the market system as naturally self-correcting—or, more precisely, as strongly systematic: in economist-speak, strongly equilibrating—that flaws in the boom have to be invoked to explain the level of the bust. Yet why cannot the bust be a result of something that has little or nothing to do with the pre-existing boom? This being more plausible if shifts in risk assessments drive actions more than shifts in time preferences, as the former are likely to be much less stable than the latter and much more susceptible to shifts in information (and so expectations).

(9) This also gets back to the unbalanced focus on money-as-value-calculation as against money-as-transaction-facilitator. There is too much focus on the “price” story of money and not enough on the “supply and demand” story (that is, the level of transactions, so level of income story—income as measured by the medium in which existing obligations, notably debt, are set). For if the demand for money rises when the supply does not, transactions, and so income, will fall. (And there is nothing like heightened risk assessment and/or income pessimism to encourage folk to hang on to money, or use it to liquidate debt; particularly if they expect money to retain, or even increase, its “swap values”.) But, in keeping with the Austrian approach being overly systematic, there is too strong a presumption that a monopoly provider of money will oversupply. (Noting that monopoly providers in other markets are presumptively under-providers; even given that the supply of paper money, at a certain level of technology, is not cost-constrained makes applying normal monopoly theory problematic.)

(10) The Austrian overly systematic story, particularly the downplaying of the role of risk in action across time and risk management in constraining prices, lead naturally to being on the “no general glut” side of macroeconomics, as outlined in Brad DeLong’s nice post about macroeconomics. The downplaying of risk fits in with the over-systematising. (Risk being inherently “messy”, even somewhat chaotic.) This in turn naturally flows from the a priori analytical approach of the Austrian school.

(12) An unfortunate tendency from the last feature is that internet “Austrians” (though not Horwitz or Selgin) are often abusive, indifferent to evidence and discount experience. No contrary evidence is permitted, as it is all redefined so it is not contrary. (A technique I am familiar with from natural law analysis, where contrary evidence is just defined as perverse and so does not count: the conclusion gets to choose the ambit of its premises—what philosopher Anthony Flew called “the no true Scotsman” fallacy.) The analysis thus subsumes all evidence into its a prior analysis. (Again, anyone who has had to deal with the natural law “marriage is by definition between a man and a woman” assertion parading as an argument is familiar with the pattern.) The analysis is therefore so “self-evident” that only stupidity or malice is left to explain disagreement.

Ironically, Catholic natural law philosopher Ed Feser has posted a critique of noted Austrian school economist Murray Rothbard, and a response to an attempted defense thereof, that targets nicely the “I have the answers, because I can just deduce them” outlook that is a persistent feature of Austrian commentary and is not an impressive, persuasive or encouraging feature thereof. I was, at first, terribly excited by the “reasoning from first principles” at the beginning of of Ludwig von Mises‘ magnum opus Human Action. Alas, experience and reflection since has made me much more sceptical about this sort of approach.

For example, the single most useful analytical tool developed in C20th economics was Ronald Coase’s identification of transaction costs. How did he do this? By asking a great question (why do firms exist?) and then finding the answer by asking people in business how they make the decision whether to produce in-house or buy on the open market. Using an analytical framework is not a matter of “looking for footnotes” because you already have the answers. It is a matter of intelligent engagement with reality so that part of what developing understanding means is the slow development of an appropriate analytical framework. In that sense, a good analytical framework represents, not merely a work in progress, but distilled continuing engagement with reality; as with Ronald Coase and transaction costs.

So this is not an objection to analytical frameworks per se: far from it. The analytical frameworks of economics and of law provide powerful advantages in examining events. Not because they are deduced from first principles but because they have been built up by an interactive engagement with reality.

(13) The Austrian downplaying of risk and the unbalanced focus on money-as-price-calculation, as against the level of transactions, as well as the notion that there is a “correct” allocation of capital, leads to one of my pet dislikes, the concept of malinvestment. A concept I strongly disagree with, for reasons I explain here and here. There is not any useful general concept of malinvestment that is independent of the level of economic activity. Sure, businesses fail but they do so all the time, even in the height of booms, and this discovery process is much more about exploring boundaries of what is or is not profitable (boundaries which shift as the level of economic activity shifts) than displaying some inherent characteristic.

Yes, one can get inappropriate construction (e.g. empty housing estates in post-bust Ireland or various US cities) but they were the result of very specific forms of perverse incentives, not indicative of some general phenomenon, even in housing construction (even if you add in various complications).

Expectations about future income obviously will affect what people invest in, but a monetary authority that drives such expectations up is less dangerous than one that drives them down. (See previous comment about deflation and unexpected disinflation being worse than inflation.)

In conclusion
That I am critical of the Hayekian-Austrian view Horwitz is presenting does not mean accepting Keynes’s position. I agree with David Glasner, both were wrong on crucial points. Keynes’ “animal spirits” for example, is much more usefully thought of as how people frame (Knightian) uncertainty. And, while monetary policy can certainly be incompetent, it is never impotent. Reviewing a recent book on the Hayek-Keynes debate, economist Tyler Cowen provides an excellent survey (pdf) of errors by Hayek and Keynes. (Though David Glasner demurs on Cowen’s characterization of what Scott Sumner is doing.)

In a monetary exchange economy, recessions (and depressions) are always and everywhere a monetary phenomenon but not, contra Austrian analysis, a particular pattern of monetary phenomena. For example, it was the deflationary policies of (pdf) the Fed and the Bank of France that caused the 1929-32 Depression not some mythic inflationary boom. Just as the European Central Bank has caused the current euro crisis and the Fed turned a financial crisis into a Great Recessions: on all these occasions, the sin being the under-provision of money by monopoly providers, not its over-provision.

Steve Horwitz has written an admirably clear paper that has enabled me to identify and clarify my disagreements with the Austrian school. Such clarity is especially useful for those of us using downloadable papers and blogs as an ongoing economics education. He is to be commended for his care and clarity in exposition. Especially since risk aversion encourages many an academic to be obscurely indeterminate, a temptation he nobly resists.

* Lars Christensen tells me that George Selgin may nowadays object to being called an Austrian; that would account for much.