Labor force changes and productivity growth

You can always blame the Baby Boom

This is a quick post as a follow-up to my last one that speculated about whether the productivity slowdown in Europe was related to structural changes in labor force participation. Basically, what you see is that productivity growth was particularly low after increases in labor force participation. This perhaps suggests some kind of sorting issue, where as you drag lower-productivity workers into the work force, you lower overall productivity.

For the US, the same evidence didn’t look as clear. But I was playing around with some data on the growth rate of productivity in the US as part of working on a paper, and came up with this figure.

The dark line is a smoothed rate of growth for total factor productivity (TFP) based on the standard BLS measure. I smoothed it out to see past the higher-frequency fluctuations on an annual basis. The smoothing is idiosyncratic to me - I played around until I had something that I thought looked good. To convince you that it isn’t entirely made up, the gray line in the back is a five-year moving average of TFP growth, and you can see all I’m doing is picking up on the big trends.

You can see the productivity slowdown from the 1960s to the 1980s, the “OG” productivity slowdown for economists. Then we get the (presumably) IT-related productivity growth acceleration in the 1990s, and then the “New School” productivity slowdown running out until recently. Data cuts off in 2015 for reasons related to the paper.

You’ll also see that I’ve added a few milestones in US demographics. The OG slowdown starts right as the Boomers enter the labor force. In terms of the story I told regarding Europe, this is equivalent to flooding the labor market with a bunch of relatively inexperienced workers. Once they get some experience, say around the time the oldest ones start turning 35, you see a pick-up in productivity growth. That pickup in productivity growth peaks right as the first wave of Boomers hit 55, which happens to be around when you’d predict that workers are at their lifetime peak of earnings. If you think those earnings reflect some information about actual productivity, then the Boomers hitting 55 are at their peak personal productivity level. As they start leaving the average productivity falls.

Why didn’t this show up in the US data in the prior post? Haven’t noodled that out yet. I think it is because of the way the leads and lags work. Like, productivity growth is rising from 80-85, but lagged labor force participation (75-80) would still have been rising as well. I’ve played with the lags/leads in the European data and I can’t shake the overall negative relationship of labor force participation and productivity growth.

Is this informative for the European situation? Maybe? The changes in labor force participation there are less obviously a pure demographic effect, as with the Baby Boom. It seems more like a structural shift of people from out to into the labor force, and in that case the sorting by productivity level may be more severe? The most optimistic take from the above figure is that a surge of workers into the labor force may lower productivity growth for a while as they stumble around and learn on the job, but that after a few years/decades this pays off with a big surge in productivity growth as they apply that experience.

There is a lot more work to be done to make a convincing case that the Baby Boom labor force participation was the (or even a) driving factor in productivity growth trends. But it is worth doing the work, given the prima facie evidence. And I am amusing myself blaming my mom and aunts and uncles for everything.