Saturday, July 15, 2017

Public Service Announcement: The US Labor Market is Still Losing Ground Relative to Trend

We keep hearing how good the labor market is these days. We've created more than 16 million jobs since the financial crisis! Unemployment is the lowest since 2001!  Time to raise rates, since the economy is overheating. Of course, this mostly comes from current and former policy makers, all of whom have a stake in trying to tell us that the Obama/Bernanke, or the Trump/Yellen economy has done quite well. However, how does job creation look like these days in terms of the long-run historical rate of job growth in the US? I plot total nonfarm employment relative to the long-run trend below. It looks a bit worse than I imagined, actually. We are now something like 23% below the long-run trend, but the surprise for me is that even in the past couple years, as the Fed tightens MP due to an economy that is supposedly overheating, we seem to still be moving further away from the long-run trend.

Of course, there are caveats here. Population growth did naturally slow a bit, and the absorption of women into the labor force was a one-time event that was mostly played out by the 2000s; 9/11 exogenously reduced immigration, and thus job growth, and the Boomers have been retiring, etc. Certainly, you could also quibble a bit with the trend. Yet, even if you plot the trend from 1945 to 1995, in recent years we still will not really have been gaining on this slower trend growth. These other events/excuses/caveats are not going to explain the relatively sudden collapse of employment some 20% +/- below trend. And why should exogenous negative shocks to labor supply cause wage growth to slow? I'm afraid I'm losing the plot of these other stories.

I have another explanation: maybe the economy is not really that overheated.

Note: you can follow me on twitter @TradeandMoney


  1. May I ask what you used as a trend?

    Also, do you think fewer job additions could be an indicator of labor tightness? Agree this theory does not reconcile with low wages but would like to know what you think nevertheless.

    1. Hey Anoop -- the trend is the rate of growth between 1942 and 2000 (when the values are 1 in the graph). In fact the picture won't look that different if you used 1945 (a boom) as the start), and say, 1995 (a recession) as the end point.

      My view is that the labor market, while certainly improving, still has some additional slack. I think both the slow wage growth and employment-to-population data (even adjusted for prime-aged workers) point to this conclusion.

  2. It seems a little odd that two of the highest wage growth periods shown on the chart - the late 1960s and the late 1990s - are both below trend. I suspect that using data from the 1940s and early 50s is in some way producing a distorted benchmark.

    1. The 1960s actually look pretty good to me. I agree that the labor market was actually strong in the 1990s, so probably would be better to fix a trend which had the 1990s above trend. It's likely the 1990s look poor here for innocuous reasons -- women had already been absorbed into the labor market, aging, declining birth rates, etc.


Note: Only a member of this blog may post a comment.