Sunday, February 23, 2014

Not heterogeneous enough

Both Austrian school economists and Keynesian economists (New or otherwise) worship at the altar of interest rates. Keynesians worry about liquidity preference and effects on investment and lead to such things as the Taylor rule. (A nice brain-teaser about interest rates in New Keynesian models is here.)

With Austrian school economists, it is about time preference and the structure of production, leading to the famous Hayekian triangle, nicely explained here. In Roger Garrison's words: (somewhat of the go-to guy to explain Austrian economics):
One leg of the triangle represents dollar-denominated spending on consumer goods; the other leg represents the time dimension that characterizes the production process (see Figure 5.1). In a fundamental sense, the Hayekian triangles in their various configurations illustrate a trade-off recognized by Carl Menger and emplasized by Eugen von Böhm-Bawerk. At a given point in time and in the absence of resource idleness, investment is made at the expense of consumption. Investment, which entails the commitment of resources to a time-consuming production process, adds to the time dimension of the economy's structure of production. To allow for investment, consumption must fall initially in both nominal and real terms. Once the capital restructuring is complete, the corresponding level of consumption is higher in real terms than its initial level. The nominal level of consumption spending, however, is lower than its initial level because a greater proportion of total spending is devoted to the maintenance of a more time-consuming production structure.        
The relative length's of the triangle's two legs, then, represent the inverse relationship between nominal consumption spending and nominal non-consumption spending—the latter as reflected by the time dimension of the economy's capital structure. 
Hence the triangle.

James Caton has posted a critique of the macroeconomic usefulness of the Hayekian triangle, noting three shortcomings:
  1. Empirically, prices in all stages of production tend to move together.
  2. Related to critique 1, the Hayekian triangle assumes an economy at full employment equilibrium.
  3. The Hayekian triangle is an incoherent representation of the macroeconomy.
The second point is clear enough in Garrison ("in the absence of resource idleness"). The first point is one in favour of monetary theories of the business cycle, since what is the one thing all transactions have in common in a magnetised economy? The use of money. (As an aside, Hayek supported keeping NGDP--i.e. aggregate nominal income--on a stable growth path.) 

The Austrian theory of the business cycle is based on the notion that central banks over-stimulate by keeping interest rates too low, leading to investments which do not accord with actual supply and demand conditions (including time preferences) and then an economic bust as malinvestments get liquidated. 

The concept of "malinvestments" bothers me, as it looks like an intrinsic quality which sits ill with the subjectivism that is allegedly the basis of Austrian economics. What would be a malinvestment in Port-au-Prince may be a great idea in New York. While changes in demand conditions can shift dramatically  the line between profitable and unprofitable businesses. (As Hayek admitted with his concept of "secondary deflation".)

A commenter on the above post defends the Hayekian triangle, writing:
the triangle is similarly percussively effective at demonstrating the heterogeneity of capital investment. It matters where investment is going. 
On the contrary, surely a problem with the Hayekian triangle is that it does not take the heterogeneity of capital seriously enough. Different industries use different forms and amounts of capital with widely differing production times (and levels of risk). Doing so while both providing and seeking income. There is no single interest rate, but a whole series of interest rates depending on asset risk, longevity and production delay. Which means that investment in various industries has a considerable variety of lead times in responses to changes in expectations. 

To put it another way, there is nothing anywhere near close enough to a single "structure of production" to drive economy-wide downturns: indeed, the entire process of Schumpeterian creative destruction is all about discovery processes in varied, diverse and constantly evolving structures of production. So, the structure of capital, being as heterogeneous as it is, is precisely where we should not look to explain economy-wide downturn happening across all industries at the same time. 

Milton Friedman's plucking model of business cycles (original article here [pdf]) fits the empirical evidence better, evidence that is hard to reconcile with the Austrian business cycle theory that the Hayekian triangle is used to elucidate, as the latter sees the cause of each bust as being in the preceding boom. So, that there is no correlation between booms and subsequent busts is very problematic--it fits more with shocks pushing the economy off its growth path rather than some internal pattern of over-indulgence and hangover.

So, back to monetary explanations of the business cycle. (Yes, there can be aggregate supply shocks, but it matters how aggregate demand responds, and that central banks can control.)


  1. I like this post, but think the commenter you reference is mistaken.

    The Hayekian triangle and the point that capital is heterogenous are two different, but related topics. The Hayekian triangle shows how the process of production takes times. The longer the structure, the more removed it is from the ultimate goal of consumption.

    Heterogenous capital shows that once investment occurs along one structure of production, it can be costly to readjust to another. It is not a chunk of capital, K, that can be switched from designing software to laying cement.

    These two ideas combine to lead to Austrian capital and business cycle theory.

    Another point on critique 1- the stages of production move together during the business cycle, but no where near symmetrically. Check out page 191 of Romer's text

    1. Nicely put. There are at least two axes of heterogeneity: specificity of use and lengths of time between industries. That all industries turn down together is a count against ABCT. That, as Romer points out, different stages of production change at different rates looks more like adjustment to changes in income expectations than time preferences mismatches; though that is less clear, since one could argue that that is how the latter manifests.