Evan Soltas
Feb 14, 2012

Getting Hysterical

A 'supply-and-demand side' policy is the only cure for structural unemployment


Two of the most dangerous elements of this recession, in the long term, are the increase in structural unemployment and the shrinking of the labor force.

In combination, I'm worried they could constrict U.S. output growth for years, even decades, to come. It may be why the CBO's forecasts imply the slowest annual growth in real potential output--1.5 percent--and scarcely better for years to come. For the record, this growth averaged roughly 3.5 percent before 2000. This is a really big problem, one which, if this country could have a mature political conversation about economic policy, we'd be having a conversation about the proper role of government and how we can grow faster and more sustainably, which produces more opportunities for everyone. 1.5 percent means that we can do no better than 1.5 percent in the long run. Yikes: we don't want that--that would be the slowest potential growth in the postwar era.

I'd like to think I'm heading towards a "big idea" about productivity growth and its impact on the potential growth rate. Let's call it the "paradox of productivity growth" for now, as I envision it as analogous to Keynes' "paradox of thrift."

Here goes: productivity growth, in the long run, is absolutely essential to growth, as is saving. But when everyone decides to increase their productivity at the same time, particularly when demand is weak, productivity growth may actually fall in the short run. This happens because the productivity gains do not cause firms to increase output--demand is weak, remember--instead, they reduce labor demand in the short run. Firms lay people off. These people spend less. Firms can meet that demand by producing less, potentially offsetting the effect of the layoffs on productivity. (Indeed, if we look at the data, this appears to be borne out on first examination, as either during or immediately before almost every recession, we saw productivity growth fall or even go negative.) Demand has to recover for output to increase and thus for the effect of the layoffs to appear in productivity growth, which is why we see soaring productivity growth immediately after most recessions.

Part two of the thought: what if demand doesn't recover quickly? This is not unlikely: we know that the self-correcting mechanism which closes recessionary gaps involves a leftward shift of the aggregate supply curve--i.e. prices, particularly labor, must fall, which is extraordinarily slow. All of these people put out of work, waiting for the supply curve to shift, are losing employable skills--i.e. their potential productivity falls. Those losses may endure with us for a long time, given how hysteresis works. So even though the "employed productivity" of the economy has risen, the "total productivity" of the labor force may actually be falling. As this happens, the potential growth rate of the economy falls.

So let's recap where we are now: an unfortunately coordinated burst of productivity just caused productivity to fall in the short run and perhaps even in the longer run. Instead of increasing the long-run growth rate, we just saw how productivity growth (as Keynes discovered with saving) can actually depress growth in the short and longer runs. All of this, by the way, puts the lie to the Schumpeterian ideal of the recession as a purge through "creative destruction."

One more place I want to go in today's post: what do we do about this? That is, how do we prevent the adverse shift in the labor picture from kneecapping output growth? I'd like to look at that graph of the estimated natural rate of unemployment again, up top. When do you see it rising? Falling? Falling Faster? (Answer key: 1950s through 1970s, 1980s, 1990s.)

There's a quick explanation for this. In the first period, we saw demand-side policy to the exclusion of anything else. That pretty much failed, and in particular, we saw a steadily climbing natural rate for unemployment as the incentive to work was progressively undermined. In the second, we saw a 180-degree tack to supply-side--better in terms of bringing down the natural rate of unemployment. In the third, however, is when things really got moving, and the natural rate fell faster than we should think possible. The structural unemployment, which is supposedly hard to eliminate, almost fell faster in the 90s than our unemployment rate is falling now. What distinguishes this era in my mind? (Besides Rugrats and The Fresh Prince of Bel Air, of course.) It is the apparent macroeconomic consensus about the importance of both supply-side and demand-side economics. It worked out that when the demand-siders saw to it that demand kept growing, and the supply-siders saw to it that we incentivized work, saving, and investment, we brought down structural unemployment at the fastest rate in our modern economic history.

The irony of the rising structural unemployment rate in the context of Paul Krugman's so-called "dark age of macroeconomics" is not lost on me.