Keynes was an "Austrian" economist

Today's neoclassical mainstream tends to look at the economy as a series of shifts from one equilibrium state to another. Of course, all of the best neoclassicals recognize this as an idealized abstraction, but they largely think of it as an abstraction that largely captures what really occurs in an economy.

Keynes, in common with Austrian economists, saw that while the market process generates a tendency towards equilibrium, it is never, or almost never, in an actual equilibrium state. Market participants do not have before them a neat little supply-and-demand diagram telling them just what the new equilibrium price will be when supply or demand changes. Instead, they must engage in a discovery process searching for the new "correct" price. Keynes's "vision" was based upon his realization that, in such uncertain conditions, buyers and sellers would be likely to make quantity adjustments as well as price adjustments to the new situation. (In contrast to the neoclassical equilibrium-always model, where they instantly arrive at the new, equilibrium price, and then read their quantity to be supplied or demanded from the model.) Furthermore, those quantity adjustments potentially could produce a positive-feedback loop, where one supplier's quantity-adjustment downward prompts other suppliers to similarly reduce their quantity supplied, which would induce the initial supplier who reduces his quantity supplied to reduce it yet again... which could prompt yet further quantity reductions from other suppliers.

There is nothing "unpraxeological" in the process that Keynes posits. No one in his model is acting contrary to their own self-interest, as they see it. No, what has happened is that market actors' expectations have fed off of each other, so that one business's anticipation of lower sales next month causes it to cancel orders from a supplier, who therefore anticipates lower sales himself, and thus cancels orders from his supplier, and so on.

The neoclassical model of markets with perfect competition is a very useful tool of thought, a reality which neither Keynes, nor Mises, nor Hayek, nor Kirzner, etc. ever denied. In the case of a market in which very many buyers and very many sellers of very large numbers of a commodity of which each unit is essentially identical to every other unit--for instance, shares of Apple stock traded on an exchange--that model will likely be a very good approximation of what really occurs in that market. But when those conditions do not hold, for instance, in the market for real estate, where no two buildings are ever really identical, where the neighborhoods in which they are located are never really identical either, and in which transaction volume is very low compared to that of a publicly traded stock, the perfect competition model will be a much less faithful portrayal of what actually happens. Sellers will frequently make offers significantly higher or lower than "the equilibrium price," and buyers will, similarly, frequently accept offers that are "too low" or reject offers they "ought" to have accepted. Keynes's originality consisted chiefly in noting that such "false trading" could feed upon itself and drive a market further from equilibrium. Neither Mises's nor Hayek's nor Kirzner's understanding of the market process rules out such a possibility!

I suggest that Austrians cease trying to write off Keynes as some sort of economic nitwit, and instead acknowledge that his understanding of the market process is very similar to theirs. Having done so, the real contested ground can then be seen more clearly: How are important are quantity adjustments compared to price adjustments in markets in disequilibrium? If quantity adjustments are a significant factor in disequilibrium conditions, how likely is it that they will result in the positive* feedback effects that Keynes postulated? If such positive feedback is an important part of contemporary economies, is Keynes correct in believing that government and/or central bank action can dampen them?

* "Positive" here means that the effects self-amplify, not that they are to be applauded!


  1. Not that I disagree with what you are saying broadly, but you are oversimplifying. Dynamic models (which are seen as preferable) involve movements from one equilibrium to another, and these movements take time. Of course, these movements are "efficient" and at any given point they are in "equilibrium," at least in some sense, but not in the comparative statics sense as you phrased it.

    1. OK, Ryan, how would you modify what I wrote above in light of dynamic models?

    2. To distinguish between neoclassical models and Austrian economics/Keynes, I would emphasize the true uncertainty perspective. If you don't have that, you can just slap a normal distribution on it, look at the search literature, and never leave neoclassical economics. You can easily build a dynamic process from one static equilibrium to another; it's a question of what underlies it.

  2. Enjoyable post.

    Three points:

    (1) "How ... important are quantity adjustments compared to price adjustments in markets in disequilibrium? "

    The empirical evidence shows that they are very important.
    You can find the extensive empirical literature -- which goes back to the work of Gardiner Means in the 1930s -- in this book:

    Frederic S. Lee, Post Keynesian Price Theory, Cambridge University Press, Cambridge, 1998.

    Also, I recommend Nicholas Kaldor, Economics Without Equilibrium. M.E. Sharpe, Armonk, N.Y.,1985.

    (2) Austrian economics is not a stranger to cost of production pricing, as opposed to flexible market-clearing pricing.

    Supposedly, Eugen von Böhm-Bawerk and Friedrich von Wieser both discussed cost of production pricing, and more recent discussion is in George Reisman, Capitalism: A Treatise on Economics. Jameson Books, Ottawa, Ill. and Chicago.

    (see appendix).

    (3) Also, if Mises was so open to the idea of quantity adjustments, then why did he make statements like this:

    “The characteristic feature of the market price is that it equalizes supply and demand. The size of the demand coincides with the size of supply not only in the imaginary construction of the evenly rotating economy. The notion of the plain state of rest as developed by the elementary theory of prices is a faithful description of what comes to pass in the market at every instant. Any deviation of a market price from the height at which supply and demand are equal is – in the unhampered market – self-liquidating.” (Mises 2008: 756–757).

  3. Gene,

    Aren't downward quantity adjustments (of all or most economic agents) literally equal to the definition of recession? I mean, if only prices ever adjusted, then we'd always be at the full employment/output equilibrium and the puzzle of recessions would have been meaningless.

    My intuition is that the real discord between Austrians and Keynesians lies somewhere deep in the Bohm-Bawerkian theory of capital, but I'm only starting to get into these controversies.

  4. Assuming that Keynes’ positive-feedback loop is real, also assuming that large scale capital misallocation is possible and real. Both are not mutually exclusive:

    How does one distinguish one cause from another? Also additionally if both things are possible, it becomes problematic in trying to solve the positive-feedback loop with monetary or/and fiscal policies if those foster the cause capital misallocation (drives prices further from equilibrium as well). How would you know that you cure more bad effects caused by the positive-feedback loop than you create by misallocating capital (maybe only showing up later in time)?

    Of course if you can answer the first question you also can answer the second. Though if you cannot, then you have the problem that your try to cure the positive-feedback loop with those policies is not only a shot in the dark, but might even be outright causing more troubles down the road than you cure, even if you were completely right and the current economic woes were solely created by Keynes’ positive-feedback loop.

    Honestly, are these stupid questions?

  5. By differentiating between short run v long run equilibrium I think it possible to reconcile the Keynesian and neo-classical view without the need for the use of "false trading" mechanisms.

    Because they may not have much control over input costs businesses will likely adjust to changes in demand via changes in quantity rather than changes in price. This will mean a flat or vertical supply curve.

    When demand falls it will respond by reducing output and possibly firing workers. The fired workers may at first be optimistic about their chances of re-employment and not be prepared to take a wage cut. They will cut back on their spending and set-off the kind of negative feedback loop you describe.

    It is possible that at all times all markets are at their equilibrium prices during this process.

    However these equilibrium prices are based on a mis-match between the views of workers (who think they can find work at their old wage) and employees (who will not hire unless workers will accept a wage cut) the equilibrium can be said to be a sub-optimal short-run one.

    Over-time (other things being equal) workers will accept that real wages have really fallen and adjust their expectations, and as they accept jobs at lower rates so an opposite process of accelerating demand will take the economy back to long run equilibrium.

    I'm not saying that frictions in the market won't in reality further slow down the adjustment process , just that its possible to develop a model that doesn't need them.