TREND WATCH
Good news continues to roll in on the job front: Seasonally adjusted unemployment matched its post-GFC low of 4.1 percent in February, with the U.S. economy adding 313,000 jobs, well above expectations. So why then aren’t wages rising faster? That’s the question at the heart of a recent San Francisco Fed report, which suggests that the retirement of high-wage Boomers is in effect suppressing average wage growth. But our analysis finds that the quantitative impact of this trend has been negligible: The vast majority of the change in average wages since the Great Recession cannot be attributed to the shifting age profile of the U.S. workforce. In fact, today’s high-earning older workers are working longer than ever, which actually boosts average wages.
If we are close to full employment, why isn’t wage growth picking up speed? According to the CES Establishment Survey, real wage growth has averaged well under 1 percent annually since the Great Recession (+0.68 percent) and doesn’t seem to be accelerating with time. The only years in which growth was greater than 1 percent were the post-crash pick-up years of 2009 and 2010—and then again 2015, when a 2.2 percent spike in real wages (triggered by plummeting energy and food prices) coincided with a broader rise in median household income. (See: “A Good Year for Middle America.”)
The San Francisco Fed is the latest to come up with a hypothesis to explain this puzzle. In a groundbreaking paper published in 2016 and updated last summer, authors Mary C. Daly, et al. suggest that demographic headwinds are to blame. In their view, the disappointing trend in the average wage is an artifact of the economy skewing more toward younger (lower-paid) workers and away from older (higher-paid) workers. The typical worker at any given age is therefore doing “better” than the average. The story has gotten plenty of run in news outlets from The Wall Street Journal to Bloomberg Business to MarketWatch.
If true, this theory would change how economists parse the data. It would be good news for the economy to the extent that unretired workers are actually experiencing higher wage gains than indicated by the average wage number. If these older workers’ higher wages reflect higher marginal product, it might also help explain lagging productivity growth. (Hopeful prediction: Once all those workaholic Boomers are retired, this negative influence will disappear.) It is also a helpful story for Fed hawks who don’t want to back off from tightening.
But is the theory true? In our analysis, we can’t examine the CES numbers directly because (as establishment “payroll” data) they include no personal information, such as worker age. So instead we use annual data from the CPS household survey, which includes a detailed and vetted age breakdown of full-time earnings per worker—and which tracks well with the CES series.
Let’s look at the difference in wages by age. The basic assumption of the Fed theory is that older people get paid more. And, yes, the CPS data bear that assumption out. In 2017, for example, a full-time worker age 55-64 earned $550 per week—versus $248 for a worker age 20-24. So if the workforce is “skewing younger,” that could in theory have a negative impact on average wage growth.
That’s not the whole story, however. In our analysis, we have taken the CPS average full-time wage data, breaking down each year’s growth into two components: the share caused by wage changes in each age bracket, and the share caused by changes in age composition of workers.
When we decompose wage growth this way into a “population effect” and a “wage effect,” we discover two things. Most importantly, demographic shifts have had only a very small impact on average wage growth in either direction. In any given year going back to 2000 (change from 1999), virtually all yearly wage growth, positive or negative, is generated by changes in wages by age, not in changes in worker numbers by age. To the extent demography does have an effect, however, it is true that the effect has been in the direction suggested by the Fed. From 1999 to 2009, demography added 0.16 percent yearly to average wage growth. From 2009 to 2017, on the other hand, demography subtracted 0.06 percent yearly from average wage growth. So there was a 0.24 percent “negative swing” from one period to the next.
What needs some explanation, perhaps, is why the demographic effect is relatively small and (especially) why in no period has it actually pushed wages down to a significant degree.
We can get a bit clearer view by going back to our earlier observation that the big threshold between low and high wages by age was 35. So let’s break down the 35-plus ranks into their various age brackets. To be sure, we see here a steep (Boomer-fed) rise in the 55-64 share from 2001 to 2011 as the large 1946 birth-cohort bow wave moved into this bracket. Starting in 2011, this bow wave has been reaching age 65. The ascent of the 55-64 share has since slowed. It will reach a peak (at 26.2 percent of all workers) in 2020—and thereafter will fall.
Running the other way, however, is the growing tide of workers moving into the 35-plus ranks. From 2000 through 2011, the baby-bust Gen-X cohorts reaching age 35 were shrinking every year. Since 2011, however—the year in which the smallest 1976-born “buster” cohort reached 35—the incoming 35-year-olds have been growing every year. Today, these incoming 35-year-olds are Millennials, whose biggest-ever cohort will reach 35 in 2025.
So just as the “baby boom” started to hit 65, the “echo boom” has started to hit 35. This rising age-35 inflow has powerfully mitigated the negative wage impact of the Boomer retirement. Also at play have been behavioral factors. Starting in the late 1990s with late-wave Silent—and even more with Boomers—the average retirement age has been rising. At age 65-74, the LFP (labor force participation rate) has climbed from 12.5% in 1996 to 19.2% in 2016. At age 75+, it has climbed from 4.7% to 8.4%. In effect, rising elderly LFP rates have pushed up average wages for all U.S. workers.
What about the future? Assuming constant employment rates and wages for each age bracket, today’s barely perceptible negative age effect will gradually decline from -0.05 percent in 2018 to zero by 2025. Thereafter, the age effect will turn positive again and gradually increase with the aging of the “echo boom” Millennials. It will hit +0.07 percent in 2037.
It’s clear that the aging of Boomers has had only a minimal quantitative effect on average wage growth in recent years. And, to extent that it did slow down wage growth over the last decade, that effect is now ebbing and (within a decade) will disappear entirely.
TAKEAWAYS
- A shifting workforce age gradient cannot explain why real wage growth has stagnated. The latest employment indicators appear to show an economy operating near full capacity. Why then aren’t wages rising faster? Researchers at the San Francisco Fed believe they’ve found the answer: The massive tide of high-wage Boomers retiring from the workforce, and the flood of low-wage Millennials taking their place, has in effect suppressed average wage growth. Our analysis shows, however, that the quantitative effect of this trend has been negligible. Quite simply, most changes in average wages have been caused not by population shifts within the workforce, but changes in earnings by age bracket.
- The rising average age of the workforce underscores the problem with the Fed’s theory. One easy way to assess historical age shifts in the workforce is to look at the average age of all full-time CPS workers. From the mid-1990s to 2009, this figure rose steadily as Boomers were aging within the workforce and as a small generation (Xers) was entering the workforce. This growth slowed down after 2009 and stopped entirely from 2013 to 2017. That’s been the impact of Boomer retirement—not to reduce the average worker age, but simply to flatten it out for a while. Population projections show, moreover, that the average age will soon start rising again.
- The Fed’s explanation falls short on more than just demography. Behavioral shifts also explain why Boomer retirement has not perceptibly suppressed wage growth. For one thing, high-wage older workers aren’t simply dropping out of the workforce once they hit age 65. Increasingly, they’re working page age 65. Altogether, there are over one million 65+ workers today who would not have been working at the LFP rates of twenty years ago. What’s more, wages have risen much faster for the oldest age brackets. Over the last twenty years, real wages for workers 65+ have risen as much as 3 percent annually. No bracket under age 65 has exceeded 1 percent.
- To be sure, demography does explain the slight negative population effect on wages since the Great Recession. In the decade prior to 2009, the large Boom Generation was maturing in midlife while the small Silent Generation was retiring. Starting in 2011 (when the 1946 birth cohort reached age 65), Boomers began to retire. The large positive population effect in 2009 was due primarily to layoffs among younger workers—who, during the ensuing recovery, tended to dominate the number of new employees and thus suppress wage growth. So the timing of the Great Recession may exaggerate slightly the demographic contrast between these two periods. But the basic contrast is clear enough.
