The Long U

My recent post comparing the magnitude of economic downturns certainly generated lots of feedback.  Some comments were constructive and inspired edits to the original post, some comments were constructive but didn’t lead me to change anything, and some were, er, less than constructive.

Taken together, the comments did inspire me to think through the issues again, and I realized that there is one significant limitation to my analysis that is worth emphasizing: the Long U problem.

Like many economists, I do not expect the U.S. economy to rebound briskly from its current troubles.  The economy may well return to positive growth in the third or fourth quarter, as many forecasters anticipate, but that doesn’t mean that the suffering is over.

In short, I don’t expect the recovery to be a V, with recent declines offset by a rapid recovery.  Nor, for that matter, do I expect a Japan-like L, in which the economy flattens at its new low level.  Instead, I expect a Long U, in which the economy heals slowly before eventually returning to solid growth.

There’s obviously remarkable uncertainty out there, and we can’t rule out a V, an L, or, for that matter, a renewed decline.  But a Long U strikes me as the most plausible single scenario.  (For simplicity, I include within the Long U the possibility of various W-shaped recoveries since they will have similar cumulative effects over the longer run.)

Why do I believe in the Long U?  The lessons of history. As Carmen Reinhart and Ken Rogoff have amply demonstrated, financial crises have historically led to periods of prolonged economic weakness.  Furthermore, as the IMF recently showed, economic downturns in an individual country tend to last longer if the rest of the world experiences a downturn at the same time.

Unfortunately, both of those negative factors are in play today.  The U.S. economy will need time to heal from a severe financial shock, and we can’t count on much help from the rest of the world in speeding that along.  So even with all the efforts of U.S. policymakers, it seems reasonable to expect a Long U.

So what does this have to do with my effort at comparing economic downturns?  Well, my original analysis focused solely on the cumulative magnitude of the economic declines.  The counter started ticking when the economy started falling, and it stopped ticking when the economy reached bottom.  I still think that’s an interesting metric, but it has two major limitations:

• First, it doesn’t consider the strength or weakness of the recovery.  A sharp V-shaped downturn might create less overall economic suffering than a shallower, but prolonged U-shaped downturn.  In geek-speak, what you should really be interested in is the integral of the losses, not just the point-to-point decline.

• Second, it doesn’t account for the fact that, absent the negative shock, the economy ought to be growing over time.  In other words, losses should be calculated relative to the potential output of the economy — potential GDP — which is growing over time.  

Both of these adjustments would, appropriately, make prolonged economic downturns look worse.  I haven’t had an opportunity to do this version of the calculation (nor do I know where to find data on potential GDP all the way back to 1929).  But here’s my guess at what an analysis like this would show:

• As I previously argued, the Great Depression was an unspeakably bad time for the U.S. economy.  After all, it combined a exceptionally sharp decline in output with a long duration. Even with a Long U, it seems doubtful that our current economic losses will end up remotely comparable to those of the Great Depression.  (As noted in this post, there are, however, strong parallels in how we got into these two messes.)

• What might well change, however, is the comparison between the current downturn and past downturns.  My original calculations suggested that if economic forecasters are correct, the current downturn will end up roughly tied with the downturn of 1957-58 are the sharpest economic decline since 1947.  With a measure of cumulative losses, however, a Long U might well result in the current downturn being noticeably worse than the prior downturns.