I just finished reading a great little paper on Okun’s rule about the relationship between output gaps and the unemployment rate (I’d say it’s uniquely readable for this sort of thing, btw).
Fifty years ago, when economists were a lot more likely to be driven by what the data actually did as opposed to what they wanted it to do, Arthur Okun “reported an empirical regularity: a negative short-run relationship between unemployment and output.” The rule of thumb came to be that when real GDP is one point above or below its potential, unemployment would go down or up by about half-a-percentage-point. So, if the underlying trend in GDP was 2.3% per year and growth some year was 1.3%, we’d expect unemployment to rise by roughly half a point.
I know. The idea that economists choose to spend our time convincing ourselves that output slumps lead to higher unemployment may well lead you to conclude that we have too much time on our hands. And if it were only a few Chicago professors who disbelieved that output gaps increase unemployment, I wouldn’t much bother you with it. But ponder this:
The current policy agenda being applied in virtually every advanced economy with a current output gap is in direct opposition to the reality that there is a negative relationship between such gaps and the unemployment rate.
Okun’s rule may be alive and well in the statistics, but it’s in serious trouble in practice. The pivot to fiscal austerity well before real GDP was back to its potential level is the rule in today’s advanced economies, not the exception.
Almost daily these days someone asks me, in so many words, “Why is that?” This post gets at part of the answer though I’m sure there’s more to it.
But for now, let me focus of one of the underappreciated costs of this policy mistake: the impact of output gaps and the ensuing high unemployment on real wage trends. That is, Okun’s relations go from an aggregate demand shock to jobs to unemployment. But others have carried this through to the next step: the impact on earnings.
I looked at this today in the context of low earnings (10th percentile) for men and women, by constructing a simple statistical model that maps changes in the unemployment rate onto real wage changes. The results, based on a long-term unemployment rate forecast, are in the figure below (thanks to EPI for the wage data). Data from 2012 forward are forecasted.
I’ve got the unemployment rate falling pretty slowly from 8.1% in 2012 to 5.4% in 2018—this is the Okun-rule part: as we slowly close the remaining output gap, the unemployment rate comes down. However, while the earnings of low-wage workers are a lot more closely tied to the unemployment rate than high-wage workers—they’ve generally been doing well regardless—their real wages continue to decline up until the point when the job market is a lot less slack than it is today.
And that’s assuming we make it to 2018 without bumping into the next recession first.
Obviously, these are simplistic calculations of complex variables. But I suspect they’re ballpark, if not optimistic. Extensive research has clearly revealed that absent very tight labor markets, middle and low-wage workers lack the bargaining power to capture the growth to which they’re contributing.
So let’s just be clear about the cost—in real dollars and cents—of our policy mistakes.
Source: EPI, my calculations; forecasted data begin in 2012.