I recently wrote two posts (this and this) about the Austrian Theory of Business Cycles (ABCT) that could be construed as criticisms of the theory, and regular readers of this blog are probably aware that critical comments about ABCT are not unprecedented on this blog. Nevertheless, I am not at all hostile to ABCT, though I am hostile to the overreach of some ABCT enthusiasts who use ABCT as a justification for their own radically nihilistic political agenda of promoting the collapse of our existing financial and monetary system and the resulting depression in the expectation that the apocalypse would lead us into a libertarian free market paradise. So, even though I don’t consider myself an Austrian economist, I now want to redress the balance by saying something positive about ABCT, because I actually believe that the Austrian theory and approach has something important to teach us about business-cycle theory and macroeconomics.
The idea for writing a positive post about Austrian business-cycle theory actually came to me while I was writing my latest installment on Earl Thompson’s reformulation of macroeconomics. The point of my series on Earl Thompson is to explain how Thompson constructed a macroeconomic model in many ways similar to the Keynesian IS-LM model, but on a consistent and explicitly neoclassical foundation. Moreover, by inquiring deeply into the differences between his reformulated model with IS-LM model, Thompson identified some important conceptual shortcomings in the Keynesian model, perhaps most notably the downward-sloping IS curve, a shortcoming with potentially important policy implications.
Now to be able to construct a macroeconomic model at what Thompson called “a Keynesian level of aggregation” (i.e, a model consisting of just four markets, money, output, capital services and labor services) that could also be reconciled with neoclassical production theory, it was necessary to assume that capital and output are a single homogeneous substance that can either be consumed or used as an input in the production process for new output. One can, as Thompson did, construct a consistent model based on these assumptions, a model that may even yield important and useful insights, but it is not clear to me that these minimal assumptions provide a sufficient basis for constructing a reliable macroeconomic model.
What does this have to do with ABCT? Well, ABCT seeks to provide an explanation of business cycles that is built from the ground up based on how individuals engage in rational goal-oriented action in market transactions. In Austrian theory, understanding how market actions are motivated and coordinated is primarily achieved by understanding how relative prices adjust to the market forces of demand and supply. Market determined prices direct resources toward their most highly valued uses given the available resources and the structure of demand for final outputs, while coordinating the separate plans of individual households and business firms. In this view, total aggregate spending is irrelevant as it is nothing more than the sum total of individual decisions. It is the individual decisions that count; total spending is simply the resultant of all those individual decisions, not the determinant of them. Those decisions are made in light of the incentives and costs faced by the individual decision-makers. Total spending doesn’t figure into their decision-making processes, so what is the point of including it as a variable in the mode?
This mistaken preoccupation of Keynesian macroeconomics with aggregate spending has been the central message of Austrian anti-Keynesianism going back at least to Hayek’s 1931 review of Keynes’s Treatise on Money in which Hayek charged that “Mr. Keynes’s aggregates conceal the most fundamental mechanisms of change.” But the assertion that aggregates are irrelevant to individual decisions is not necessarily valid. Businesses decide on how much they are going to invest based on some forecast of the future demand for their products. Is that forecast of future demand independent of what total spending will be in the future? That is a matter of theoretical judgment, not an issue of methodological malpractice. Interest rates, a quintessential market price, the rate at which one can transform current commodities or money units into future commodities or future money units, are not independent of forecasts about the future purchasing power of the monetary unit. But the purchasing power of the monetary unit is another one of those illegitimate aggregate about which Austrians complain. So although I sympathize with Austrian mistrust of overly aggregated macroeconomic models, I am not sure that I agree with their specific criticisms about the meaningfulness and relevance of particular aggregates.
So let me offer an alternative criticism of excessive aggregation, but in the context of a different kind of example. Suppose I wish to explain a very simple kind of social interaction in which a decision by one person can lead to a kind of chain reaction followed by a rapid succession of subsequent, but formally, independent, decisions. Think of a crowd of people watching a ball game. The spectators are all seated in their seats. Suddenly something important or exciting happens on the court or the field and almost instantaneously everyone is standing. Why? As soon as one person stands he blocks the vision of the person behind him, forcing that person to stand, causing a chain reaction. For some reason, the action on the field causes a few people to stand. If those people did not stand, no one else would have stood. In fact, even if the first people to stand stood for reasons that had nothing to do with what was happening on the field, the effect would have been the same, because everyone else would have stood; once their vision is blocked by people in front of them, spectators have to stand up to to see the action. But this phenomenon of everyone in a crowd standing when something exciting happens on the ball field happens only with a crowd of spectators of some minimum density. Below that density, not everyone will be forced to stand just because a few people near the front get up from their seats.
A similar chain reaction, causing a more serious inefficiency, results when traffic slows down to a crawl on an expressway not because of an obstruction, but just because there is something off to the side of the road that some people are slowing down to look at. The effect only happens, or is at least highly sensitive to, the traffic density on the expressway. If the expressway is sufficiently uncrowded, some attention-attracting sight on the side of the road will cause only a minimal slowdown in the flow of traffic.
The point here is that there is something about certain kinds of social phenomena that is very sensitive to certain kinds of interactions between the individuals in the larger group under consideration. The phenomenon cannot be explained unless you take account of how the individuals are interacting. Just looking at the overall characteristics of the group without taking into account the interactions between the individuals will cause you to miss something essential to the process that you are trying to explain. It seems to me that there is something about business-cycle phenomena that is deeply similar to the crowd-like effects in the two examples I gave in the previous paragraph. Aggregation in economic models is not necessarily bad, even Austrians routinely engaging in aggregation in their business-cycle analyses, rarely, for example, discussing changes in the shape of the yield curve, but simply assuming that the entire yield curve rises or falls with “the interest rate.” The question is always a pragmatic one, is the increased tractability of the analysis that aggregation permits worth the impoverishment of the model, by reducing the scope for interactions between the remaining variables. In this respect, it seems to me that real-business cycle models, especially those of the representative-agent ilk, are, by far, the most impoverished of all. I mean can you imagine, a representative spectator or representative-driver model of either of the social interactions described above?
So my advice, for whatever it’s worth, to Austrians (and non-Austrians) is to try to come up with explanations for why aggregated models suppress some type of interaction between agents that is crucial to the explanation of a phenomenon of interest. That would be an more useful analytical contribution than simply complaining about aggregation in the abstract.
PS Via Mark Thoma I see that Alan Kirman has just posted an article on Vox in which he makes a number of points very similar to those that I make here. For example:
The student then moves on to macroeconomics and is told that the aggregate economy or market behaves just like the average individual she has just studied. She is not told that these general models in fact poorly reflect reality. For the macroeconomist, this is a boon since he can now analyse the aggregate allocations in an economy as though they were the result of the rational choices made by one individual. The student may find this even more difficult to swallow when she is aware that peoples’ preferences, choices and forecasts are often influenced by those of the other participants in the economy. Students take a long time to accept the idea that the economy’s choices can be assimilated to those of one individual.
Uneasy Money 
“The point here is that there is something about certain kinds of social phenomena that is very sensitive to certain kinds of interactions between the individuals in the larger group under consideration.”
I believe what you are talking about is potential energy. It exists where two forces offset set each other to the point where no net force is applied. However, if the strength of one force is increased, weakened, or if the point of application is shifted, you go from a stable predictable system to an unstable system.
Imagine five people all holding hands standing in a circle. Now all five lean backwards on their heels until their arms are pulled taut. Force is exerted by all five people, but their is no net motion – until one person lets go.
“So my advice, for whatever it’s worth, to Austrians (and non-Austrians) is to try to come up with explanations for why aggregated models suppress some type of interaction between agents that is crucial to the explanation of a phenomenon of interest.”
I don’t think aggregated models suppress some interaction. I think in the aggregation process, counterveiling forces are canceled out. And that canceling out of counterveiling forces works well until someone lets go.
LikeLike
Ignoring the interactions simplifies a model. Modeling the interactions would be a lot more work and require more processing.
LikeLike
Even Murray Rothbard aggregated. Rothbard came up with an aggregate called Gross Private Product. That’s GNP less government spending.
There’s also TMS – the true money supply.
LikeLike
Austrian economics and Marxist economics are both Good Things because they are a great place for non-economists to start. They both recognize that capitalist systems are inherently unstable and are a means of exercising social status. They both rely on logical examination of individual and organizational decision making to discern trends and motives. The problem with both is that they are not mathematically sound. Von Mises rejected mathematics outright, and Marx’s mathematical reasoning yields surfaces rather than lines or points. This is why they’re good places to start but awful places to end.
Without mathematics it is extremely difficult to objectively gather and evaluate evidence. It is impossible to be precise or measure error. In short, it is impossible to be scientific. This is why adherents of ABCT should be given as much intellectual quarter as adherents of Aesop’s fables. It’s like reading the Last of the Mohicans and thinking you understand the French and Indian War.
The second problem, and I think it’s what you’re trying to illustrate, is that scaling matters in terms of time, space, and mass. Studying behavior at one scale often doesn’t elucidate the others. Knowing the amount and distribution of deer shit in the woods doesn’t tell me anything about the earth’s orbit around the sun, even though deer shit is clearly part of the solar system.
LikeLike
This is NOT the point Hayek was making:
“This mistaken preoccupation of Keynesian macroeconomics with aggregate spending has been the central message of Austrian anti-Keynesianism going back at least to Hayek’s 1931 review of Keynes’s Treatise on Money in which Hayek charged that “Mr. Keynes’s aggregates conceal the most fundamental mechanisms of change.”
Good lord you make a lot of mistakes when ‘explicating’ Hayek.
LikeLike
Cameron,
“Without mathematics it is extremely difficult to objectively gather and evaluate evidence. It is impossible to be precise or measure error. In short, it is impossible to be scientific. This is why adherents of ABCT should be given as much intellectual quarter as adherents of Aesop’s fables. It’s like reading the Last of the Mohicans and thinking you understand the French and Indian War.”
But there is no mathematical formula for value or liquidity preference. I think you have to approach economics with that in the back of your mind – ultimately choices will be made (both rational and irrational) that do not adhere to a predictable model.
From above:
“Nevertheless, I am not at all hostile to ABCT, though I am hostile to the overreach of some ABCT enthusiasts who use ABCT as a justification for their own radically nihilistic political agenda of promoting the collapse of our existing financial and monetary system and the resulting depression in the expectation that the apocalypse would lead us into a libertarian free market paradise.”
A true libertarian free market paradise would be a system without a legal framework in which all values are set by “markets” without influence of an legal authority. And it would be anarchy. If ABCT is trying to do economics without legal authority, then it is truly lost.
A legal authority exists to provide a framework where both parties in a contract have recourse if either feels that they have been conned or cheated, that one party has misrepresented the value of what they are offering. And so a legal authority exists to make judgements of value.
A legal authority exists to provide a “buyer of last resort” where the present value of an owner’s assets is less than the present value of an owner’s liabilities. And that authority exists to provide liquidity for items whose value have sunk below a market clearing level.
Mathematics will only get you so far. Without a legal authority to underpin value and provide liquidity, you won’t go very far.
LikeLike
About these aggregates: Keynes was inspired by national accounting which actually measures the outcomes of all the individual decisions. And these decisions take place in a monetary economy where, as a matter of accounting logic, total income equals total production equals total expenditure (check the SNA). When I stop spending this will effect the other micro-players – which will show up as smaller aggregates. It’s that simple, we don’t rule our own lives. We rule the lives of others.
LikeLike
This is exactly what Hayek did *repeatedly* when he pointed out that that the ‘models’ of Keynes and the Keynesians failed to include the causal/theoretical price relation structure of heterogeneous production processes involving the choice between production processes which required greater time to produce superior output and production processes which required less time to produce inferior output.
And this is what Hayek did *repeatedly* when he pointed out that this adjusting structure of relations between alternative production processes interacts causally & in terms of price relation with all sorts of different monies and near monies, ie assets of changing value, risk and liquidity which can be uses as money or money substitutes.
What part of this do you claim Hayek doesn’t spell out in detail?
David writes,
“So my advice, for whatever it’s worth, to Austrians (and non-Austrians) is to try to come up with explanations for why aggregated models suppress some type of interaction between agents that is crucial to the explanation of a phenomenon of interest. That would be an more useful analytical contribution than simply complaining about aggregation in the abstract.”
LikeLike
“The student then moves on to macroeconomics and is told that the aggregate economy or market behaves just like the average individual she has just studied. She is not told that these general models in fact poorly reflect reality. For the macroeconomist, this is a boon since he can now analyse the aggregate allocations in an economy as though they were the result of the rational choices made by one individual.”
And this is a big problem in the field of economics. Trying to apply the theories of the small (microeconomics) to the realities of the big (macroeconomics) is misguided.
A lot more sense can be made out of economics if you break things down in terms of the static (measures of value, balance sheets, and solvency) and the dynamic (income, liquidity, cash flow statements, and bankruptcy).
As a simple example, it is quite rational on a microeconomic level for someone to consume less than they produce, or spend less than they receive in income. On a macroeconomic level it is impossible for everyone to do this at the same time.
So you go from logical micro-economic behavior to illogical macro-economic possibility.
LikeLike
Hayek *explicitly* did this when he repeatedly pointed out that the interaction between heterogeneous production goods and heterogeneous labor inputs was absent from the models of Keynes and the ‘Keynesians’ — to name just one aspect: the substitution of capital goods for labor inputs.
Hayek flags this in the most direct way possible, quoting Mill against Keynes on the “test” of a competent economist.
This core content isn’t hidden, rather, it is ignored. Why?
David writes,
“So my advice, for whatever it’s worth, to Austrians (and non-Austrians) is to try to come up with explanations for why aggregated models suppress some type of interaction between agents that is crucial to the explanation of a phenomenon of interest.”
LikeLike
David,
FWIW, I have for decades used precisely your example of spectators at the stadium to illustrate Hayek’s idea of spontaneous order. (I say “Hayek’s idea,” but of course it goes back!) I guess there are self-proclaimed “Austrians” who rail against “aggregates” as such. And even the usually careful and correct Mises seems to think “scale of prices” is okay but “price level” is somehow not. But an occasional Misesian lapse notwithstanding, I don’t think there is any real Austrian criticism of aggregation as such.
Look again at the quoted snippet from Hayek: “Mr. Keynes’s aggregates conceal the most fundamental mechanisms of change.” Presumably, Hayek is thinking of the sectoral shifts characteristic of ABCT. If you are working with aggregates that conceals such shifts, you can’t even see the fundamental mechanism of change. that’s a fair criticism having nothing to do with any supposed taboo on “aggregation” as such.
Mises’s disciple Alfred Schutz says somewhere, demand and supply curves are “aggregates.” But we can unpack them and see what they are saying about who does what. If you can’t unpack your aggregates, they may be inappropriate tools of analysis in social science. (I said “may be,” not “therefore certainly are.”) That’s another legitimate line of criticism that, again, has nothing to do with any supposed taboo on “aggregation” as such.
I guess some self-proclaimed “Austrians” make the sort of argument you criticize, but I don’t think you’ll find it in Hayek or, like, Roger Garrison or George Selgin. (George may not forgive me for lumping him in with the Austrians, but we all know he has the red A burned into his chest!) In the early part of Prices and Production Hayek notes that aggregates do not act directly on one another, which seems right. But that’s not a taboo on aggregation as such. I think it’s just the admonition that we should be able to infer what statements about our aggregates imply for who does what.
LikeLike
Frank, I see the analogy, but I don’t think that is what I am talking about at all.
johnny bakho, I agree that it would require more work and more processing, but I think that there are conceptual issues introduced when there are simultaneous interactions taking place. Think of the difference between solving a two-body problem and n-body problem. There is an analytical solution for the two-body problem, but not for the n-body problem (apologies in advance to all the physicists for venturing into territory where my ignorance is palpable).
JP, Thanks.
Cameron, I think that you can do Austrian economics and use mathematics. Hayek, unlike von Mises, did not object to mathematical reasoning per se. But sometimes adopting mathematical notation involves making assumptions (often implicit assumptions that go largely undetected) that are necessary or at least convenient for arriving at an analytical solution. Like any tool, using mathematics involves costs and benefits. The problem is that the costs are often overlooked. And yes, I agree also that scaling is a problem, though I would probably have chosen a different example to illustrate it.
Greg, Perhaps not, but I think that you are misunderstanding my point, which was mainly supportive of Hayek’s preference for (somewhat) disaggregated models. My criticism was that the argument against aggregation was often made on methodological grounds when the more appropriate argument was analytical or empirical. Hayek made both, types of argument at various times, and I was interpreting his criticism of Keynes as a methodological criticism, not an analytical one. I agree that it might have been analytical, but he certainly made the argument in methodological terms on numerous occasions, including in his Nobel Lecture. I’ll respond to your more substantive observations below.
Frank, Actually a libertarian free market paradise without a legal framework is a self-contradictory idea, because market exchange is mainly an exchange of property rights not an exchange of physical objects, and to imagine the existence of property rights without some legal framework to define and enforce those rights makes no sense. I am not aware of any Austrian economist, even the crackpot Murray Rothbard, who ever thought that there could be a free market economy without a legal system. Rothbard thought that it would be possible to have a legal system but no government which is not quite what you are attributing to Austrian economists in general. Hayek and Mises were certainly not anarchists.
anonymous, I am not so sure that Keynes was inspired by national income accounting. I think that he may have been influenced by it, but I think that his conception of how the economy works was derived independently. The two systems were developed at about the same time and complemented each other, but were independent. That, at any rate, is my impression, but I could be wrong.
Greg, I repeat that I think you are misunderstanding what I wrote, but I would be prepared to claim that Hayek did not spell out in detail how the upper turning point of a business cycle leads to a cumulative contraction of total spending and employment. He actually did recognize a connection between the upper turning point and what he called “a secondary deflation,” but it was not part of a well-developed theoretical explanation of the process. He therefore was unable to provide an appropriate policy recommendation to address the secondary deflation that he, at least sometimes, acknowledged was taking place. This is not to say that he did not provide important insights into the process, because his wonderful paper “Economics and Knowledge” is indispensable for understanding the process. But he did not provide that explanation himself. However, I could be wrong, and if so, I am sure that you will be glad to cite chapter and verse of the primary source or sources that I have overlooked.
Frank, The inconsistency between an individual’s ability to spend more or less than his income and the supposed impossibility for the entire community to spend more than its income in the aggregate is easily derived from the budget constraint that individuals are required to obey. In fact spending can exceed income in the aggregate if individuals draw down their cash balances or other previously accumulated assets.
Greg, You have a habit of making general references to Hayek having repeatedly done this or done that, but if you don’t provide a specific example that I can refer to, you make it hard for me to respond to you in a satisfactory way that would advance the discussion. Just a thought.
Roger, I think that I agree with you, but I am not 100% sure. To rephrase my point, I would say that the permissible level of aggregation depends on the nature of the analytical problem one is trying to solve. It seems to me that for certain types of problems it is perfectly OK to use very macro aggregates, like the total money supply, the price level, total spending, and total employment. So I can’t agree with the 100% Austrian prohibition (if that’s what it is) on those aggregates. But I would impose a 100% prohibition on representative agent models. And certainly there may be important macro-problems that can only be analyzed in the context of a less aggregated model. But the justification for the less aggregation is not methodological, it is an argument that a certain kind of sectoral interaction is important to the understanding of the entire system. And I think that Austrians have been right to emphasize that such interactions can be important. My criticism is that too often they state their preference for disaggregation in methodological terms rather than in analytical terms. I believe that Hayek in several places argued that “statistical aggregates” that are objects of individual choice are illegitimate variables to be used in economic models.
LikeLike
David,
“Frank, The inconsistency between an individual’s ability to spend more or less than his income and the supposed impossibility for the entire community to spend more than its income in the aggregate is easily derived from the budget constraint that individuals are required to obey.”
When I was talking about everyone trying to run a surplus, I didn’t mean a community, I really meant everyone – government, banks, individuals, companies, etc. This is not a supposed impossibility. In a capitalist system, it is a literal impossibility. Microeconomic good sense (running a surplus) becomes macroeconomic impossibility for everyone to do simultaneously.
“In fact spending can exceed income in the aggregate if individuals draw down their cash balances or other previously accumulated assets.”
Drawing down cash balances implies a a change from previous years of cash accumulation (high liquidity preference) to spending (low liquidity preference). In a capitalist system this will not add to aggregate demand if it is matched with a reduction in the value of all outstanding liabilities. If liquidity preference falls and credit contracts by the same amount, there will be no net increase in aggregate demand.
Capitalist System: For someone to run a surplus, someone else must run a deficit. For someone to have more income than they spend, someone else must have sold a liability (debt or equity) to be able to spend more than they have in income.
Fiat System: Money is spent into existence without creating the corresponding liability (usually by a government). And so under this kind of system, all parties can run surpluses indefinitely, because the initial money spender does not create a liability that must be purchased by someone else.
In a fiat money system, there is no budget constraint that anyone is required to obey. The budget constraint comes from either a capitalist system of contractual obligations (liabilities) or a physical constraint on monetary expansion (gold standard for instance). In a true fiat system, neither exist.
LikeLike
David,
I doubt you can come up with any damning quotes from Hayek on aggregation. If you can, well then so much the worse for Hayek. But I bet it’s not there in dear anything dear old Hayek ever wrote. Nor do I think it’s a matter of the “level” of aggregation per se. It’s a question, again, of whether you can unpack your aggregates.
LikeLike
David,
“I am not aware of any Austrian economist, even the crackpot Murray Rothbard, who ever thought that there could be a free market economy without a legal system.”
Instead Austrians try to do economics without acknowledging that a legal system exists in the first place. How many Austrians are claiming that “the U. S. federal government is going bankrupt” (Ron Paul and Alan Simpson – please raise your hands)? There are plenty of good reasons to reduce the federal debt, but threat of bankruptcy or insolvency is not one of them.
LikeLike
Frank, You said:
“Drawing down cash balances implies a a change from previous years of cash accumulation (high liquidity preference) to spending (low liquidity preference). In a capitalist system this will not add to aggregate demand if it is matched with a reduction in the value of all outstanding liabilities. If liquidity preference falls and credit contracts by the same amount, there will be no net increase in aggregate demand.”
Sounds to me as if you are making some assumptions that are not necessarily bound to be satisfied.
Roger, I haven’t done a search of his writings on this point, but I am pretty sure that there are quotes that could be found. But for now at least we’ll have to leave the question unresolved.
Frank, I do not think that you are speaking from a position of exhaustive knowledge of Austrian economics, especially if you are citing Ron Paul and Alan Simpson (it would not surprise me if he has never heard of Austrian economics) as its authoritative spokespersons.
LikeLike
David,
That is why I said “if”…
“If liquidity preference falls and credit contracts by the same amount, there will be no net increase in aggregate demand.”
Your statement:
“In fact spending can exceed income in the aggregate if individuals draw down their cash balances or other previously accumulated assets.”
This will only be true if individuals draw down cash balances to purchase goods. If they draw down cash balances to reduce outstanding liabilities, then this will not be the case.
LikeLike
Frank, OK I think we agree.
LikeLike