Evan Soltas
Mar 7, 2012

Cyclical Reasoning

Economists need a new definition of "cyclical" in policy.

Can we talk about a word economists use? Cyclical. It's a very handy word, sure. But I think the way economists are using it and its derivatives -- particularly the word countercyclical -- is interfering with clear thinking.

In particular, I hear economists characterizing the the Fed's monetary policy as countercyclical, because  their policy stance is currently running in such a way that it mitigates recession. They're making a big mistake in their reasoning.

The way economists must define the cyclicality of a policy is by the variable to be stabilized and the nearest variable in the transmission mechanism of a policy to it. Nothing else makes sense.

In the case of a Keynesian fiscal stimulus which seeks to stabilize aggregate demand, the goal variable here is nominal output, or NGDP, and the nearest variable in the fiscal policy transmission mechanism is the change in government expenditure (let's ignore tax and transfer policy for this exercise).

[Two notes: (1) Let's ignore tax and transfer policy for this exercise. (2) I use nominal output, or NGDP, as my goal variable, because that is what makes most sense to me. It is the common end variable for fiscal and monetary policy.]

For this simple case, I think we can all agree that countercylical fiscal policy increases spending in times of low NGDP growth and cuts spending in times of high NGDP growth, because the cyclical volatility in this variable runs opposite to the swings in the variable we aim to stabilize.

Somehow, though, when the countercyclical policy terminology was carried over into monetary policy, things got confused. Economists thought of interest rates or the money supply as the relevant variable from which to define what was "countercylical." This is wholly inconsistent with our definition for fiscal policy.

A consistent definition would specify that a fully countercyclical monetary policy stabilizes aggregate demand, and the nearest variable in the monetary policy transmission mechanism is the rate of inflation, and to some extent, the change in real output which comes from induced investment. (Henceforth, I'll assume zero lower bound conditions, such that the central bank picks the rate of inflation.)

This is important because using our consistent (and I think the only coherent) definition of countercyclical policy, stabilizing money supply growth or the rate of inflation are both insufficient, as doing so will not stabilize aggregate demand, or NGDP.

So then what are these policies?

Let's invent a clear test: a policy is countercyclical if and only if it induces opposite cyclical volatility in the nearest variable to fully stabilize the goal variable. A policy is weakly countercyclical if it induces opposite cyclical volatility in the nearest variable but does not fully stabilize the goal variable. A policy is acyclical if it stabilizes the nearest variable. A policy is procyclical if it induces identical cyclical volatility in the nearest variable.

The lesson here is pretty clear. A money supply rule, like the k-percent rule would be fairly called procyclical monetary policy, in that it induces cyclical volatility in inflation which runs with the business cycle (inflation rises in booms, falls in busts) due to changes in money demand. (Pegging the interest rate or exchange rate seems to be to be even more procyclical.) An inflation target which is expected to be hit every quarter is acyclical, as it stabilizes the nearest variable -- notice that this is different from the old and inconsistent definition, because stabilizing inflation would require a significant reduction of interest rates. A longer-term inflation target, like the Fed now has, is weakly countercyclical, with it falling somewhere between acyclical and countercyclical to the extent that the central bank induces opposite volatility in inflation in the short run.

Finally, if we were consistent with our definitions, the only fully countercyclical policy that exists is NGDP targeting of rate or level. (I'd argue level is more countercyclical, as an imperfect central bank is likely to undershoot accidentally in recessions and overshoot in booms, creating "base drift.")