Wednesday, 20 September 2017

Blog

Austrian business cycle theory and rational expectations

February 24, 2016
in Blog

Although Austrian business cycle theory (ABCT) is a powerful price-theoretic explanation for monetary-induced booms and busts, it is not without critics.  Indeed, many find ABCT implausible for two reasons.  The first is that it seems to rely on individuals making systematic and economy-wide errors.  This seems inconsistent with rational decision making, especially the theory of rational expectations.  The second is that the recalculation process following the bust cannot be understood in an equilibrium context, differing from the vast majority of accepted models.  Today I tackle the first critique; I will deal with the second in my next post.

As Will Luther and I argue, it is true that ABCT involves agents making what appears to be systematic and economy-wide mistakes.  Since rational expectations means that individuals do not systematically err, we must formulate a theory of individuals’ decisions in a context where they may be making errors in a narrow sense, but in a broader sense are behaving optimally, if we indeed want to show ABCT is consistent with rational expectations.

This can be done in two steps.  First, one should note that information—in this context, especially information on the optimal length of production plans and central bank strategy—is costly to acquire.  It follows that individuals in the market will acquire information up to the point where the marginal benefit of acquiring information equals the marginal cost.  But since, as for the vast majority of other goods, there are rising marginal costs and falling marginal benefits to acquiring information, individuals’ optimal stock of information will almost certainly leave them less-than-fully informed. Mistakes are perfectly consistent with rational expectations in this sense.

Second, one should note that production behavior does not consist of an infinitely continuous series of spot decisions.  Production involves making a plan—a prespecified ‘recipe’ for production that has built-in contingencies, precisely because deciding on contingencies in the heat of the moment, when things have gone wrong, is intolerably costly.  Plans must be made with many things as basic data, which are given, and must specify under what conditions the plan will be revised.  If something does not occur according to plan, it might not be the case that firms perfectly adjust, because again, plan changes are costly to undertake.  The already decided-upon ‘algorithm’ for plan updating will probably be used, even if this is suboptimal in the narrow sense, because when the full costs of plan are taken into account, this is the best strategy for producers.

Putting these two together, one can see why individuals systematically err when monetary policy becomes unanticipatedly loose.  As Nobel laureate Robert Lucas noted, producers in markets face a ‘signal extraction’ problem when they observed increased demand for their product: is the increased demand real, or due to unexpectedly easy money?  Since information is costly to acquire, and the relevant context for adjustment is not momentary decisions, but entire production plans, there will be less-than-full adjustment to newly easy monetary conditions. Again, producers will rely on their built-in updating procedures, because to do otherwise is too costly.

Admittedly, the magnitude of production errors will probably be smaller in a world of rational expectations, since individuals will recognize that they live in a world of signal extraction problems and, to the extent it is cost effective, will adjust somewhat.  But if the monetary disturbance is large enough, and especially if there are other policies in place that incentivize investment in specific lines of production using easy credit (housing boom, anyone?), ABCT dynamics are still consistent with the ‘tight’ model conventions of equilibrium-always and rational expectations.

One Comment

  1. AtulHost March 5, 2016 2:51 AM

    For the vast majority of other goods, there are rising marginal costs and falling marginal benefits to acquiring information, individuals’ optimal stock of information will almost certainly leave them less-than-fully informed. Mistakes are perfectly consistent with rational expectations in this sense.