In the current U.S. economy, the right thing is happening for the wrong reason.
The right thing is the unemployment rate heading down to Buddha-knows-how-low. A recent forecast from Goldman Sachs had the jobless rate hitting 3 percent by the end of 2020. Put aside the political implications of that, if you can, and consider that this would be the lowest rate since before I was born, and I’m north of 60.
The wrong reasons are the trillion-dollar budget deficits that are juicing the macroeconomy at a time when it’s already closing in on full capacity. I’m less worried about overheating, though that’s a risk, than I am about the loss of needed revenue and the lack of perceived fiscal space when we hit the next downturn.
But whatever the cause, let’s look at the impact of such tight labor markets on a very important labor market variable: the employment rates of prime-age workers (25-54), or the share of that age group with jobs. While I don’t mean to imbue one indicator with too much meaning, I think this one arguably tells us about labor demand, labor supply, racial equity, wage trends and even, maybe, whether the next recession is near.
The first figure shows prime-age rate for whites and blacks (see data note below). Looking at the past decade, two things jump out at you. First, prime-age workers got slammed in the downturn and have been slow in climbing back. Recessions are like that, especially for the least advantaged: They take the elevator down and the stairs up.
But also notice the gains of blacks relative to whites. In 2011, the gap between the two rates was over 10 percentage points. Last month, it was below five points, close to the smallest gap on record. To put some real numbers behind that, the increase in the prime-age black employment rate from 67 percent in the depths of recession to 76 percent last month is the equivalent of 1.5 million more jobs.
The lesson is an old but venerable one: The benefits of a full-employment economy disproportionately help the least advantaged.
Turning to wages, it’s also the case, especially in recent years, that the prime-age employment rate correlates better with wage growth than the unemployment rate. The figure below shows the result of a simple statistical model that predicts the growth of nominal mid-level wages (see data note for details). I ran the two versions through 2012 and then predicted the path of wages after that. The model that uses the unemployment rate overshoots the recent trajectory of wage growth, but the one that uses the employment rate better captures the trend.
I feel extremely confident in both the validity and importance of those empirical facts. This next one is more speculative but worth considering, given concerns about whether the Federal Reserve needs to move from tapping the growth brakes to hitting them harder (i.e., raising interest rates more quickly).
This next figure is the prime-age employment rate for all workers, going back to the late 1940s. To be sure, it mixes in disparate trends of men, whose employment rates have trended down over time, and women, whose rates trended up until a few decades ago. But if you squint enough (assisted by the little lines I drew in), you can see a semiregular pattern where the prime-age rate flattens somewhat before recessions (the shaded periods). From a labor supply perspective, this makes some sense, as the flattening employment rate suggests the supply of workers is getting tapped out.
The punchline is this: We don’t see that happening yet in this expansion, suggesting that there’s more room to run. That is, if unemployment keeps falling, as I suspect it will, employment rates for prime-age workers, and especially for those of color, with less education, and in communities that have heretofore been left behind by the expansion, could keep climbing the stairs, as more potential workers move into jobs from unemployment or from the sidelines of the labor force.
Imagine if the prime-age rate could climb all the way back to where it was in 2000, around 82 percent (last month, it was 79.5 percent). Tapping that little wage model I used above, mid-level wage gains could hit almost 4 percent, which, after inflation, would amount to some seriously long-awaited real gains of 1 to 2 percent.
Now, depending on inflationary pressures, the Fed might be in no mood to accommodate such gains, though let the record show that they prevailed in the late 1990s. Productivity was higher then, and that makes all of this a lot easier, but before we get tangled up in a million complications, let us not lose the thread.
The benefits of full employment are real and important to those who need them the most. The climbing prime-age employment rate is a solid proxy for that reality, so before the next elevator drop, let’s allow it to climb as many stairs as it can.
Data note: I seasonally adjusted the employment rates in the first figure, as the BLS data is non-seasonally adjusted. The wage model regresses the annual percent change in the production, nonsupervisory hourly wage on the slack variable with one lag and four lags of the dependent variable.