Back when I studied labor economics, the labor force participation rate—lfpr, those in the labor force as a share of the working age population—wasn’t a big deal. It cruised along in predictable ways, stable for men, rising for women.
Today, the lfpr is almost as closely watched as the rates of unemployment or job growth, and for good reason. It’s a key variable in regard to micro, macro, and even politics. That’s right, the boring old lfpr will help or hurt the President’s reelection prospects.
The basic facts of the case go like this. As you see in the figure, the share of the population participating in the labor force went up about a point in the 1990s expansion, fell in the recession/jobless recovery of the early 2000’s, then stabilized at about 66%. Then it tanked in the Great Recession and has yet to reverse course.
So why is this so important?
First, the low lfpr is suppressing the unemployment rate. You’re only counted as unemployed if you’re looking for work, so if people give up looking and leave the labor market, they don’t show up as unemployed (more on this below). If they stayed in the labor market without a job, the jobless rate would be a few points higher than it is right now—closer to 11% than 8% (note, however, that this is an upper bound—surely some of those who left the labor force would have jobs if they were back in).
So, if the job market were to start improving and pull more people back in, we could end up with the curious outcome of faster job growth pushing the unemployment rate up. That’s actually what happened last month as employment in the household survey went up but the labor force went up more, pushing the jobless rate from 8.1% to 8.2%, though one month’s results don’t mean much.
Second, over the longer term, the growth of labor supply, of which the lfpr is a critical component, is an important input into the growth of the overall economy. Based on our aging demographics, labor force growth, and thus GDP growth, will slow somewhat, but if the lfpr stays way down, it will slow even more than expected.
This has led economists to try to figure out how much of the recent decline in the lfpr is cyclical–a function of the weak jobs picture–and how much is structural, i.e., due to longer term trends that won’t be reversed by a more welcoming job market.
The consensus is that about 1/3 to 1/2 of the decline so far is structural, and I’d lean toward the lower value (as does the great Heidi Shierholz, here). That is, given that the lfpr is down about two percentage points off its pre-recession level, I expect it to gain back about 1.3 points based on the analysis that 2/3 of the decline is due to cyclical weakness in the job market.
There’s some confusion on this point. Some analysts note, correctly, that given the fact that long-term unemployment has been so high for so long, a lot more people seem to be leaving the job market now than in earlier periods. Some of these long-term jobless may be older workers who are deciding to retire now instead of later, and if so, that would be a structural change (though, interestingly, one born of a cyclical dynamic).
It is, as the next figure shows, certainly the case that a lot more unemployed have left the labor market in this recession relative to the last one. The figure tracks monthly counts of just this phenomenon: unemployed persons in month t who leave the job market in month t+1, and three years out from the cyclical peak, this variable was up about 30% after the mild 2001 downturn compared to 200% now.
Source: CPS Labor Market Flows
But here’s the thing: some of the same analysts who recognize this connection between long-term unemployment and falling lfpr assume it’s mostly structural—i.e., they assume that if a long-term jobless person drops out, they’re not coming back. And that’s just unknowable at this point. What’s almost certainly true is that those who make that assumption will overestimate the structural component of the lfpr decline.
One last, cute little observation re this question of structural versus cyclical. In reading up on this, I stumbled upon a point made by a researcher that Hispanic men tend to be quite connected to the labor market while Hispanic women tend to be much less connected.
That led me to think of a little natural experiment. If that’s true re connectedness, and if this is mostly a cyclical story, then the men’s lfpr should go down due to the cycle and the women’s should remain stable. If that both go down, it’s probably more a structural thing.
The figure supports the cyclical hypothesis: Hispanic men are way down, women tick up a bit (data are for adults, 20 years and up, not seasonally adjusted, so I compare the lfpr over the prior three months compared to the lfpr in those same three months in 2007).
Source: BLS (see text for description of sample)
I’ll continue to track this as will everyone else who’s paying attention to the nexus between the economy, the labor market, and the election. If the lfpr starts climbing, it could both be a sign that we’re getting back to normal, but it could also make it harder to knock down the unemployment rate.