The Wherefores and Whys of Wages


Tom Heintjes: Welcome to another Economy Matters podcast. I'm Tom Heintjes, managing editor of the Atlanta Fed's Economy Matters magazine, and today we're joined by John Robertson, a senior policy adviser at the Atlanta Fed. John, thanks for joining us today.

John Robertson: G'day, Tom.

Heintjes: John is one of the Atlanta Fed folks behind the creation of our Wage Growth Tracker, which does precisely what the name implies: it tracks wages, something that is a very closely watched metric. So I thought it would be illuminating to discuss the Wage Growth Tracker, and some of the data it's been giving off lately. So John, let me start off by asking—there are several tools that observe wage trends. What sets the Wage Growth Tracker apart?

John Robertson

Robertson: That's a good question, Tom. The Wage Growth Tracker differs because it looks at the change in wages of individuals, as opposed to changes in an aggregate measure of the overall wage bill.

Heintjes: Well, you mentioned what goes into it. Can you tell us a little bit about the inputs that go into the Tracker?

Robertson: The underlying data that's used in the Wage Growth Tracker comes from the Current Population Survey—it's the same survey where we get statistics like the unemployment rate or the labor force participation rate. Periodically they ask questions about wages, including the earnings per hour that individuals make. So we are able to take that data and then match it with what the same individual said a year earlier, and then look at the whole distribution of wage changes for those individuals.

Heintjes: That's a very granular look, then.

Robertson: It is. And then to aggregate it up, what we look at is the median person—the median change in wages from one year to the next—and that's basically what the Wage Growth Tracker is.

Heintjes: Right. So, you've talked about what it shows us. What doesn't it show us?

Robertson: Well, because it's measuring the changes in wages of individuals, it's basically controlling for shifts in the composition of the workforce. However, the aggregate wage bill for a firm—or for the whole economy—is affected by shifts in the composition of who's working over time.

So, for example, during the last recession we know that average wages—that is, the average of the aggregate wage bill—did not grow very rapidly, and then when the unemployment [rate] started to come down, it still didn't grow very rapidly. That was a puzzle, and that's really where the Wage Growth Tracker's origin is, is in trying to solve that puzzle, because we thought what might be going on there is that the composition of the workforce might be changing. That's putting downward pressure on the average wage bill, even if individuals are getting wage increases.

Heintjes: Right, I see. You've talked a lot about individuals today, and these individuals in the survey are not identifiable, are they?

Robertson: No, no; everybody who is in the survey has an identification number, and we just use that number to match that same individual from one year to the next.

Heintjes: And that number is assigned by the BLS [Bureau of Labor Statistics]?

Robertson: No, the number is assigned by the Census Bureau, who conducts the survey on behalf of the BLS.

Heintjes: Gotcha. Well, let's talk a bit about what we've actually been seeing in regards to wages, John. Lately, we've seen evidence of wages increasing. What's behind that increase, after such a long period of what's been relative wage stagnation?

Robertson: Well, the unique feature of the Wage Growth Tracker—that is, that it's tracking individual wages—means that it's much more closely aligned with the overall slack or tightness of the labor market than the average wage bill, because the average wage bill is affected by hiring new workers, who might be coming in at a lower wage, and so that puts downward pressure on the average wage bill. And then, for example, more retirees leaving the workforce, who probably had a higher wage, and that also puts downward pressure on the wage bill, whereas what we see with the Wage Growth Tracker is that it correlates very nicely with movements in the unemployment rate.

Heintjes: So how does this period—say, from the last few years until now—compare to historical trends in wages, especially as we've been emerging from a pretty severe recession?

Robertson: We saw the Wage Growth Tracker slow a lot in the wake of the recession, just as the unemployment rate went up a lot. And as the unemployment rate has been coming down over more recent years, the Wage Growth Tracker has been generally moving higher as well—something that we're only now starting to see in these other measures of average wages.

Heintjes: And that touches on that correlation that you just mentioned?

Robertson: Exactly. What we also saw are some interesting distinctions that happen between types of workers. The median wage growth for all workers slowed during the recession, but especially for people who are working part-time.

Heintjes: Now, don't get ahead of me—I want to get to types of workers in a minute, so hold that thought. But what, if anything, makes this period insofar as wages different from other periods, especially as we've come out of recessions in the past?

Robertson: Well, the median wage growth—that is, the Wage Growth Tracker—is not as high as it was before the recession. So, on that measure, it says that there is probably still a little bit of slack left in the labor market.

Heintjes: Even with the relatively low unemployment rate?

Robertson: Even with the relatively low unemployment rate. The rate of people who are working part-time involuntarily is still rather high, so that rate is quite high. The labor force participation rate—we think some of the decline we've seen in that was also a consequence of the recession, and as that stabilizes then that suggests that the labor market’s getting tighter, but still not quite there yet.

Heintjes: Right. Well, we've touched on unemployment, and I want to dive a little deeper there. Wage increases, coupled with an unemployment rate at or under 5 percent like we are now, historically has raised some concerns about inflation. But we're not really seeing strong inflation—if anything, the Fed has struggled to meet its 2 percent annual inflation objective. Why do you think these elements have not combined for a more inflationary environment?

Robertson:Well, that's something that the Wage Growth Tracker doesn't capture directly, because if you're thinking of the costs of labor relative to the productivity of labor as a driver of inflation, then what you need to look at is the wage bill. That has not increased as rapidly as the Wage Growth Tracker, because the Wage Growth Tracker is looking at people who are in effect continuously employed, so it's not affected by the flows in and out of the workforce—and that seems to be what’s keeping that the total wage bill down. So you're not getting growth in what economists call "unit labor costs" at an alarmingly high rate that would be kind of a red flag for inflation pressures.

Heintjes: I want to change gears here, and we're going to play a sort of word association game here. I'm going to say some groups to you, and you tell me what your observations are about that group's wages. So are you ready?

Robertson: Fire away.

Heintjes: All right. In terms of age, how have what we call "prime age" workers—those in the prime of their career—how have their wages performed recently?

Robertson: Well, the Wage Growth Tracker pretty much is representative of those workers, because remember—it's the median wage growth of all the workers that we have in the data, and most workers are prime age.

Heintjes: When we say "prime age," we should explain how you define that.

Robertson: Sure, prime age. It's a crazy economist’s term.

Heintjes: Aren't most?

Robertson: Aren't most? [laughter] But it's basically trying to capture people who are the most attached to the labor market, so that typically is people kind of around between 25 and 55. We see high attachment during those periods. In the earlier years, you have lower attachment because people are devoting a lot of their time to being in school, and not the labor market; and beyond 55, things like retirement start to dominate the activities of people. So the prime age is kind of that 25–55 range.

Heintjes: And how have the wages of prime age workers trended lately?

Robertson: Trending along with the unemployment rate—as the unemployment rate's coming down, you're seeing more growth in their wages. The Wage Growth Tracker is capturing that prime age group because it's the median of the distribution of all people who are working. And since most people are prime age, then it's kind of captured. The median worker is a prime age worker.

Heintjes: John, talk a bit about the wage growth of people who are younger than prime age, versus the wage growth of those in the prime age cohort.

Robertson:Well, this is a good example of what the Wage Growth Tracker does and does not capture. In general, the wage growth of younger workers is much higher than a prime age worker’s, because these are people who are relatively new in their careers—they might be switching jobs around and moving to higher-paying jobs, so they get pretty high median wage growth. You just have to keep in mind that their wage level is probably much lower than the median wage level of other workers. They're just getting more growth because they've got more room to grow.

And because we're capturing people who are continuously employed...if you were, say, in college and you worked part-time for a while and then you stopped because your schedule in college was too hectic to be able to keep a job, then you're not even going to be picked up in the wage growth trends. These are people who are pretty attached to the labor market, and you see pretty strong growth—but that's a long-term trend. You see higher median wage growth for younger workers than older workers.

Heintjes: You mentioned part-time, and I want to touch on that again. What does the Wage Growth Tracker tell us about full-time versus part-time employees?

Robertson:Well, I think this cut is particularly interesting because it correlates well with the difference between the headline unemployment rate measure that we look at and a broader measure that includes people who are working part-time but really want to work full-time. That rate went up much more than the headline unemployment rate. The headline unemployment rate is called the U-3 statistic, and the U-6 statistic is the one that includes these part-time workers.

Lots of people like to work part-time; but many people during hard economic times want to work a full-time schedule but can only find part-time jobs. So it's a broader measure of slack in the labor market. Well, if you looked at the Wage Growth Tracker, you saw exactly the same pattern emerge. During the recession, the wage growth of people who worked full-time slowed a lot, but the wage growth of people who are working part-time slowed by even more.

So there's always a gap, right? It tends to be the case that people who are working full-time have higher median wage growth than people who are working part-time. But that gap got really big during the recession and it's only now closing—kind of in the same way as that gap between the U-3 unemployment rate and the broader U-6 measure has narrowed in recent years.

Heintjes: The last groups I want to ask you about are what you call "job stayers"” and "job switchers"—and I guess that refers to their frequency of changing employers?

Robertson:Yes. If you think about people who stay in the same job versus people who have changed jobs—the people who change jobs are either changing jobs voluntarily, because they see a better opportunity somewhere else, or involuntarily—they might have lost the job they had but they found another job. And what you see is a gap between the wage growth of people who switched jobs and people who stay in the same job.

The pattern is like this: in good times, job switchers have higher median wage growth than job stayers—because you're switching jobs because you see an opportunity to take another job at a higher wage, and that boosts your wage growth. During bad economic times, however—like during recessions—you see the reverse: that the wage growth of job switchers is actually lower than the wage growth of job stayers, because more of that job switching is involuntary. That is, people out of desperation are taking a job even if it might actually be at a lower wage than the one they had.

Heintjes: Right, right—that makes sense. I wanted to ask you about wage growth and how it usually translates into increased consumer spending, which of course is a huge driver of gross domestic product in the U.S. Are we seeing the sort of consumer spending you'd expect to see with the current levels of wage growth?

Robertson:This is another example of what the Wage Growth Tracker is useful for, and what it's not so useful for. In the case of consumer spending, you know that what matters is primarily, do I have a job?

Heintjes: Yeah, sure.

Robertson:And then if I have a job, what kind of wage gains am I getting? The Wage Growth Tracker is picking up a piece of that wage growth part of the story, but not about the employment part, so that's why most people who try to relate consumption to income look at aggregate measures of income as opposed to...

Heintjes: It gives you a better read on the data?

Robertson: Well, it captures both the wage piece and the employment piece. But certainly, the growth that we see in the Wage Growth Tracker is consistent with reasonably robust consumer spending.

Heintjes: Well, John, this been a really great conversation, and very enlightening for me, and I want to thank you for spending some time with us today. I really enjoyed the conversation.

Robertson: Well thanks, Tom. It was fun.

Heintjes: And I want to encourage anyone listening to this podcast to visit the Atlanta Fed’s Wage Growth Tracker, which is behind obviously a great deal of what we've talked about today. John also writes a good deal about his findings in the Wage Growth Tracker on the Atlanta Fed’s macroblog, so I encourage you to look at macroblog regularly to see John's Wage Growth Tracker updates. It will give you a fuller idea of what we've talked about today.

And that brings us to the end of another Economy Matters podcast episode. I appreciate your spending time with us, and I hope you'll join us again next month when we'll be joined by two Atlanta Fed economists who will discuss the market for long-term care insurance in the United States, some of the issues surrounding that market, and why it's so complex. So come back next month; and thanks for listening.