Bringing back those missing millions
FRIDAY'S jobs report touched off a round of hand-wringing over the possibility of permanent damage to America's labour force as a result of years of labour-market weakness. Labour-force participation fell in April; there has been virtually no recovery in the employment-population ratio over the past three years, despite steady (though disappointingly slow) employment growth. My colleague captured the nature of the concern in a Friday post:While it has fluctuated considerably, the labour force is only slightly larger now than in December, 2007, when the recession began. Yet in January, 2008, the Congressional Budget Office reckoned it would be some 5m larger by now, or 159.5m (see chart). What happened to those 5m people? Why aren't they showing up as unemployed? Some are discouraged workers or other people who want to work but aren’t counted as unemployed; but I reckon they account for only one third of the missing 5m.So what about the others? Is it early retirement? Disability? Returning to school? Illegal immigrants returning home (or failing to enter the country in the first place)? Or were they never there to start with - the labour force simply isn't growing as quickly as we thought it should, for demographic or other reasons? Whichever it is, it is a troubling sign that our economic potential could be a lot lower than we thought just a few years ago. And that's the real bad news from today’s report.I'm going to disagree. The real bad news in Friday's jobs report is that the conversation has once again focused on the possibility of a decline in potential. The real danger is that policymakers will conclude that this is right and give up on countercyclical policy as a result, when in fact most of the seemingly structural deterioration in the labour market is entirely fixable.I don't disagree that structural unemployment can become a problem, but there is little evidence that it is developing in America. What structural employment there is is mostly of the "cyclical structural" sort; it's associated with underwater homes and unemployment insurance, which will become less of a drag as recovery continues. I don't disagree that hysteresis is a risk—that over time skills and labour-force attachment will erode, making it harder than ever to get the unemployed back into gainful employment. For the moment, at least, I think that these problems are less dire than they look in the American context. The labour force is growing; it's up about 1.3m since hitting a bottom in 2010. And consider this chart, which shows the annual change in employment (red) and the labour force (blue):
Several things stand out. One is a long-run downward trend. That's associated, first, with the end of the great female surge into employment and, second, with the aging of the population. The CBO's labour-force projections include a downshift due to aging that was forecast to intensify from the late 2000s on.Next, however, we notice that labour-force growth is cyclical. The slowdown in growth from the late 1970s to the early 1980s, which corresponds to the last recesssion responsible for a period of double-digit unemployment, was actually a bit larger than what has been observed this time around. In thinking about what the CBO missed in its 2008 projection, the answer is pretty clear: a deep recession. It's not at all clear that the recent decline is permanent, however. Brad Delong muses:A straight-line extrapolation from the 2000 through the 2007 business-cycle peaks says that "natural" labor-force participation today is 66.8% of the adult population. Bumping that up by 0.2% points because comparing the 2000 to 2007 peak overstates the demographic decline and bumping it up a further 0.2% points for decreased desired early retirement as a consequence of the vanishing of $8 trillion in home equity values would give us a current "natural" labor-force participation rate of 67.2%. But let's take 66.8%-67.2% as our range.Actual labor-force participation is lower in depressed times than its natural rate: People drop out of the labor force when they think they cannot find jobs. Since World War II, labor-force participation has fallen about 1/3 as much relative to trend as the employment-to-population ratio has fallen relative to trend. Given the current state of the employment-to-population ratio, we would predict that the current labor-force participation rate would be 1.5% points below its natural rate. That gives us a predicted labor-force participation rate today of 64.3%-64.7%. Instead, our labor-force participation rate is 63.6%.That is a gap of 0.7%-1.1% points of the adult population: people who really ought to be in the labor force right now, but who are not. Are they now part of the "structurally" non-employed who we will never see back at work, barring a high-pressure economy of a kind we see at most once in a generation? Probably.That's disturbing, of course, but it immediately scales down the size of the problem. The gap is not 5m, as a look at the CBO's projections might lead one to assume, but more like 1m-2m. Not coincidentally, the number of Americans on disability insurance—which has come to function as a sort of loose, but more permanent form of unemployment insurance—is up 1.6m from 2007. That, I think, it quite close to the full extent of the long-term unemployment problem. Again, that's a problem, but it's not a catastrophe. Most of the employment gap is susceptible to faster growth. Many of those now on unemployment insurance would have soon left the labour force anyway through retirement. Others could be brought back through a combination of disability-insurance reform and a tight labour market.What about that tight labour market? On Friday, I tweeted that there was no structural unemployment in America that couldn't be eliminated in a late 1990s-style labour market. This prompted a wave a responses from individuals arguing that if America has to count on a once-in-a-lifetime internet boom for such growth, then it rightly should be called structural. This, however, confuses the macro with the micro. There were a number of microeconomic trends underway, including the first stirrings of the internet economy, that made the late 1990s a prosperous era. What made the period a job-rich time was accommodative Fed policy; nominal GDP growth averaged well above 6% from 1997 to 2000. After falling about 2.5% in 2009, by contrast, it rose just over 4% in 2010 and just below 4% in 2011. That's below trend at a time in which the economy ought to be catching up to where it previously was.The Fed allowed such rapid growth in the late 1990s because unusual macro circumstances at the time placed substantial downward pressure on broad prices: oil was dirt cheap, a strong dollar was reducing import costs, and China was a net disinflationary force at the time—its cheap products dominated its impact on commodity prices. Core prices hung just a shade over 2% for much of the period. In the absence of those disinflationary forces, a rate of nominal GDP over 6% would probably correspond to a higher inflation rate. And since the Fed seems to have, if anything, become more intent on maintaining a strict 2% limit on inflation, it has been unwilling to do more, leading to disappointing nominal (and real) GDP growth, and correspondingly disappointing employment and labour-force growth.So let's be clear; the primary evidence for permanent loss of potential is the slow recovery in the size of the labour force, which would appear to be largely due to cyclical variation. We are not seeing a surge in labour costs, or prices generally, indicating that the economy is actually running up against capacity constraints. The Fed could have taken the approach that it would seek above-trend growth, as one normally expects to see after a recession (especially a deep one), and then step on the brakes if it became apparent that potential had been lost, that real growth was falling consistently short of trend while inflation was accelerating. That would have made a lot of sense, given the real economic costs of prolonged high unemployment.That's not what the Fed has done. Instead, it's been happy to accept at- or below-trend growth, despite the fact that the large remaining output gap has quickly translated shocks into worrisome disinflationary pressure. Now the public is increasingly willing to read Fed failure as a loss of potential. If the Fed comes to agree, it may begin to fear that it has less room to boost the economy, it may consequently boost the economy less, and it may therefore ensure that the output gap persists until it does indeed become permanent.This is self-induced paralysis: the fear that trying to do things to fix current problems will generate consequences worse than the present problems, all evidence to the contrary. It's frustrating—galling, even—though I suppose at this point we shouldn't find it surprising.One last point: a country to which tens of millions of people around the world—including highly skilled, ambitious, educated workers—would gladly move is one that never has to worry about slowing labour force growth. If we're going to diagnose America's ills, let's diagnose them correctly.