Business Cycle “Symmetries & Asymmetries”

I liked very much this analysis by Ryan Avent on “what does ‘above potential’ mean” (I provide an extended quote for it to be understandable and highlight crucial points):

In a column at Vox Antonio Fatas and Ilian Mihov describe an interesting new paper of theirs which posits a third phase of the business cycle. Expansions and contractions we all know and love, but that two-way division leaves out important dynamics, they reckon. Better to describe cycles in terms of a recession, in which output falls well below trend, a recovery, in which output returns to trend, and then an expansion, in which the economy grows more or less at trend. It’s a neat approach which allows them, among other things, to isolate the cost of recessions by computing the aggregate GDP loss relative to trend. The cost is big; they peg the loss from the current cycle at $3.4 trillion and counting. Strikingly, about three-fourths of that cost has accumulated during the ongoing “recovery” phase of the cycle, during which output has been growing but remains below trend.

Paul Krugman uses the piece to reflect on the nature of business cycles:

“I’m very much in sympathy with their underlying view about the asymmetry between booms and busts…”

Mr Krugman makes the important point that asymmetric business cycles mean that the focus of monetary policy shouldn’t be “stabilisation” around potential, since there’s little risk of output rising above that level. Instead it should focus overwhelmingly on fighting output shortfalls. But I think something is being lost in the failure to distinguish between the real and the nominal.

We could then speak very clearly about what a recession is: an instance in which demand falls below potential. And perhaps then we could usefully define the opposite of a recession: an instance in which demand rises above potential. If we go on to define demand as all the money spent each year—nominal output—then we find ourselves in a world in which business cycles can and historically have been fairly symmetric. Until recently, anyway. Consider this chart:

Here we have annual nominal GDP growth set against the median NGDP growth rate for the period, which is 6.4% (the mean is 6.7%). There are a few things worth noting. First, over the period as a whole business cycles look reasonably symmetrical; demand growth plummets in downturns and rockets up in recoveries. Second, there are absolutely periods in which demand growth rises as far above long-run trend as it falls below trend in recessions.

May be, but lately have not been as the chart clearly shows. And that is due to the evolution of monetary policy thinking. In the early postwar period we had an era in which policy-makers seemed to intuitively understand the relationships described above, aimed for full employment, and generated pretty good outcomes. Then we had an era in which policy-makers convinced themselves that the negative relationship between demand and unemployment that operates at some points in the cycle applies always and everywhere. They sought to push demand ever higher, leading to output stagnation and very high inflation. And then we had an era in which policy-makers convinced themselves that the goal of monetary policy should be low and stable inflation above all else. In this period, policy-makers sought to counteract low demand, but in such a way that demand coasted back up to, but not above, trend, lest inflation get out of hand. The result was a steady disinflationary trend…. In short, America exchanged a world in which individual cycles were demand-symmetric for one in which cycles are symmetric on average across generations, because policymakers swung from never allowing demand growth to dip too low to never allowing demand growth to rise too high.

I reproduce his chart with some changes. I define the different periods (and the associated Fed Chairmen) according to his characterizations of monetary policy. Instead of taking the mean growth (not the median growth) for the whole period, I show those only for the periods (1948-65 and 1987-05) where they ‘exist’ (i.e., are meaningful), in the sense that during those periods nominal growth was stationary. During the Martin/Burns “Great Inflation” nominal growth trended up (consistent with “pushing demand ever higher”). In the Volcker disinflation nominal growth trended down.

The mean growths for the two periods where they ‘exist’ are statistically the same. What is very different is the standard deviation (volatility) of nominal growth between the two periods. While during McCabe/Martin “wide swings” it was 4.5, during the “Great Moderation” it was just 1.4!

In this regard I quibble with RA´s view that “aiming for full employment generated good outcomes”. Given that mean growth was the same but volatility greatly lower during the “Great Moderation”, the outcome was much better in the later. It must also be noted that “aiming for full employment” was an important factor generating the “Great Inflation” that followed McCabe/Martin , which I have discussed here.

Additionally, what the “Great Moderation” period shows is NOT that “policy-makers sought to counteract low demand, but in such a way that demand coasted back up to, but not above, trend, lest inflation get out of hand”, but that the pursuit of NOMINAL STABILTY was largely successful. There were two ‘mistakes’: a small and quickly reversed one in 1991 and a deeper and more protracted one in 2001.

The “low and stable inflation above all else” is a ‘dogma of the Bernanke Fed. In this case, despite the very large drop, “demand growth is not even coasting back up to trend”! So obviously, the LEVEL of demand is much lower than it otherwise would be AND unemployment much higher/employment much lower.