The Atlanta Fed's macroblog provides commentary on economic topics including monetary policy, macroeconomic developments, financial issues and Southeast regional trends.

Authors for macroblog are Dave Altig and other Atlanta Fed economists.

November 05, 2015

A Closer Look at Changes in the Labor Market

The Atlanta Fed's Center for Human Capital Studies hosted its annual employment conference on October 1–2, 2015, organized once again by Richard Rogerson (Princeton University), Robert Shimer (University of Chicago), and the Atlanta Fed's Melinda Pitts. This macroblog post provides a summary of the papers presented at the conference.

Many measures of labor market performance remain at relatively low levels compared with levels seen before the Great Recession. A key question for policymakers and academic researchers is the extent to which these changes reflect a slow recovery from a large cyclical shock—or do they simply represent the "new normal"? This conference brought together researchers studying several dimensions of these changes in labor-market outcomes. A common theme is that current labor market outcomes largely reflect the ongoing effect of secular trends that predated the Great Recession.

Recent empirical work has highlighted that the U.S. economy, and in particular the labor market, has seen a pronounced downward trend in several measures of "dynamism." Prominent among these measures are decreases in job and worker flows as well as in the entry rate of new establishments. A key challenge is to uncover the driving forces behind these trends and determine whether they reflect a worsening of U.S. economic performance.

Three papers addressed these changes. In "Changing in Business Dynamism: Volatility of vs. Responsiveness to Shocks?," Decker, Haltiwanger, Jarmin, and Miranda pose a key question for assessing whether these declines might reflect positive versus negative forces. Specifically, if lower volatility in firm-level outcomes reflects a change in the volatility in the economic environment in which firms operate, then it might well be a positive development. On the other hand, if the decreased volatility in firm-level measures reflects less responsiveness to changes in the economic environment, then the changes may constitute a negative development. The paper notes that elements of each may be present in different sectors of the economy, but their analysis suggests that lower responsiveness to shocks is an important factor.

In a second paper on the topic, "Dynamism Diminished: The Role of Credit Conditions," Davis and Haltiwanger focus on the decline in the business entry rate and consider one particular driving force: the role of housing wealth in facilitating start-up entrepreneurship. They ask whether cities that had the largest drops in housing wealth also had larger drops in entrepreneurial activity, holding other factors constant. Their analysis finds a strong correlation between the two, suggesting that the loss of housing wealth from the Great Recession has had a significant negative effect on the rate of business startups.

A third paper on the theme of diminished dynamics offered a somewhat different perspective. In their paper "Understanding the Thirty Year Decline in the Start-Up Rate: A General Equilibrium Approach," Karahan, Pugsley, and Sahin offer a more innocuous interpretation of the trend decline in the entry rate. They note that the growth of the U.S. labor force has slowed in the last 30 years, because of the aging of the baby boomers as well as the slowdown in the growth rate of women in the labor force. Standard models of industry equilibrium imply that this will require a slowdown in the rate of growth of firms, achieved through a decrease in the rate of entry. They also note that standard models imply that substantial differences in cohort dynamics in response to such a change will not be evident, and they depict this in the data.

Secular changes in inequality have received much attention in recent years. Two papers examined the nature of these changes. In "Firming Up Inequality," Bloom, Guvenen, Price, Song, and von Wachter use tax return data from the Social Security Administration to examine the underlying sources of increased income inequality since 1978. A key feature of this analysis is that it is based on tax return data for the universe of individuals, making it much more extensive and reliable than estimates based on smaller samples and self-reported measures of income. The authors find that the rise in income inequality is dominated by an increase in income dispersion across firms rather than within firms, which seems to result from an increase in the extent of sorting of workers across firms. The authors suggest that this increase reflects a change in the way firms are organized. The authors also show that executive pay plays essentially no role in the overall rise of inequality.

Lochner and Shin also examine the dynamics of inequality in "Understanding Earnings Dynamics: Identifying and Estimating the Changing Roles of Unobserved Ability, Permanent and Temporary Shocks." This paper focuses on changes in labor earnings among males from 1970 to 2008. Unlike the previous paper that focused on dispersion between and within firms, this paper focuses on permanent versus transitory components of inequality and the extent to which changes in inequality reflect changes in the price of unobserved skill. The paper provides a detailed decomposition of the evolution of these various components over a 40-year period. The decomposition between permanent and transitory components is of central concern since higher transitory variance averages out over time at the individual level. One key finding is that since 1990, the dispersion of permanent shocks has increased, especially for low-income workers.

Hall and Schulhofer-Wohl analyze changes in match efficiency in the U.S. economy since 2001 in their paper "Measuring Job Finding Rates and Matching Efficiency with Heterogeneous Job Seekers." Standard estimates based on an aggregate matching function that treats all workers as identical imply that matching efficiency has deteriorated dramatically during the Great Recession and its aftermath. The authors show that if one takes into account heterogeneity in matching rates for workers with different observable characteristics, a very different picture emerges. In particular, although a decrease is still evident in the aftermath of the Great Recession, this decrease reflects a continuation of an existing downward trend. The key implication is that lower matching rates currently found in the data reflect a secular trend.

In "The Great Reversal in the Demand for Skill and Cognitive Tasks," Beaudry, Green, and Sand offer a new perspective on secular trends in the labor market. Key to their explanation is that the boom prior to 2000 is associated with investment in the new general-purpose technology associated with information technology. Their theory holds that this technology is put in place during a period of high investment demand and high demand for skilled labor. But once the new technology is in place, it requires much less high-skilled labor to maintain or operate it. In this "de-skilling" phase, high-skilled individuals will move to jobs that are lower in the skill spectrum, thereby displacing individuals with lower skill levels to either move farther down ladder or even out of the labor force. The authors argue that this de-skilling phase began sometime around 2000 and was somewhat obscured prior to the Great Recession. The paper presents a stylized model of this process and presents several pieces of empirical evidence consistent with this dynamic. The key implication is that recent developments in the labor market indicate secular trends.

Autor, Figlio, Karbownik, Roth and Wasserman examine a different trend in U.S. labor market outcomes. In "Family Disadvantage and the Gender Gap in Behavioral and Education Outcomes," the authors examine the growing gap between male and female educational attainment. This gap is particularly large for children from disadvantaged backgrounds. The authors evaluate the hypothesis that the gap reflects differences in the sensitivity of boys and girls to adverse environments. They use data from Florida that allow them to study brother-sister pairs, allowing them to control for family environment. The key finding is that their study supports this hypothesis, though they are unable to identify which specific factors might be at work.

Full papers for most of these presentations are available on the Atlanta Fed's Center for Human Capital Studies website.

By Melinda Pitts, director of the Atlanta Fed's Center for Human Capital Studies, Richard Rogerson of Princeton University, and Robert Shimer of the University of Chicago

November 5, 2015 in Employment, Labor Markets | Permalink


Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

October 19, 2015

Should We Be Concerned about Declines in Labor Force Growth?

For the second month in a row, the October jobs report from the U.S. Bureau of Labor Statistics (BLS) has revealed a decline in the labor force. From August to September, the labor force lost a seasonally adjusted 350,000 participants. And the August number of participants was a seasonally adjusted 41,000 below July's level. Although two months don't necessarily make a trend, observers have noticed the declines in the labor force (here and here, for example), and they deserve some attention.

Economists might be concerned about these labor force declines for two reasons. First, these losses might indicate that the current unemployment rate doesn't accurately reflect a strong labor market. Second, our economy needs labor to make things, perform services, and continue to grow. Let's take a look at the evidence supporting these two concerns.

Concerns about a shadow weak labor market
Two pieces of evidence suggest that the declines in the labor force don't indicate a weak labor market: employment growth and the reasons people cite for being out of the labor force. Employment growth is robust. According to the Atlanta Fed's Jobs Calculator, the labor market needs to create an average of only 112,000 jobs per month to maintain its relatively low unemployment rate of 5.1 percent. During 2015, the economy has created, on average, 198,000 jobs per month.

But we might be concerned if the workers leaving the labor market were entering into the no-man's land of the marginally attached, a term describing those who want a job, are available to work, have looked for work in the previous year, but recently have stopped looking. Some of these people have stopped looking explicitly because they think jobs prospects are poor (called "discouraged workers"). Others have stopped looking for other reasons such as attending school or taking care of family members. If these categories of nonparticipants were absorbing a large share of those leaving the labor force, we could be concerned that they would, at any moment, reenter the labor market and push that unemployment rate right back up again. The chart below tells us that this possibility is unlikely.

The chart decomposes the year-over-year changes in the total number of labor force participants into changes in the population and the negative changes among reasons given for nonparticipation in the labor force. (I use year-over-year changes because the reasons given for not being in the labor force are not seasonally adjusted.) Year-over-year changes in the population have been consistent in their contributions to changes in the labor force, propping it up. The growth in the contribution of those not wanting a job (pulling down labor force growth) has been fairly striking.

The share of people giving other reasons for not being in the labor force (discouraged, not available, etc.)—in addition to making relatively small contributions to changes to the labor force—has mostly been shrinking since April, meaning that they cannot explain the recent slowing of labor force growth. In other words, only a very small part of the growth in nonparticipants has come from those marginal workers who are most likely to reenter the labor force. So the first fear—that this declining labor force growth is producing a false sense of security in a relatively strong labor market—appears unfounded.

Threats to economic growth
Labor is an important component in the production process. Short of dramatic technological advancements, both the manufacturing and service sectors need a consistent source of labor to fuel output. Even though the economy appears to be on the right track with respect to job creation, ongoing declines in labor force growth could pose a challenge to economic growth. Additionally, as employers compete for fewer workers, we would expect wages to be bid up. Keep an eye on the Atlanta Fed's wage tracker to see how slowing labor force growth plays out in wages.

October 19, 2015 in Employment, Labor Markets | Permalink


I thought you did a great job with this article. It was clear and readable.

If US labor participation rates are heavily influenced by the departure of baby boomers, won't this same problem be repeated in other countries? Should we expect lower rates of global economic growth in the future?

By the way, the Atlanta Fed does a super job with this site. You definitely have the most interesting research projects.

Posted by: Steve Grunig | October 19, 2015 at 04:12 PM

These results accord with our own analysis of the data and are neatly explained in a way that doesn't require a PhD in math to understand. A third factor that adds to the case that lower labor force participation does not represent hidden slack is the high level of job openings reported by the BLS (admittedly they pulled back in August from a record-high level in July but they remain very high). Labor force participation continues to decline as the number of job openings has risen, which further questions the idea put forward by some that falling participation has a significant cyclical component. We are becoming increasingly concerned about the economy's capacity to meet increased demand due to tightness in the labor market (and disappointing productivity trends).

Posted by: John Ryding | October 20, 2015 at 08:30 AM

Julie This would be more helpful if the labor force was only those 18-65. What is unemployment for this group after deducting those attending college?

Posted by: duke thomas | October 23, 2015 at 02:54 PM

Julie Hotchkiss replies: Thank you for the comments--it's always nice to know folks are reading macroblog! Let me offer some additional information about a couple questions you raised.

Global LFPR: Steve asked about global labor force participation rates. The same mechanisms that link labor force participation and economic growth in the U.S. would be at work in other developed economies. A great place where you can see what's been on going around the world with labor force participation rates (LFPR) can be found on the World Bank website

The graph shows by default the global LFPR. You can add LFPR to the chart for different countries, geographic areas, and countries grouped by income level. (In the box below the chart, simply type what you're interested in--for example, "United States" or "Europe" or "High Income" or "Japan.") If you do this, you'll see that the LFPR in the United States is falling faster than global LFPR, LFPR in the euro area has been flat or seen a slight rise, LFPR is lower among high-income countries and falling, and LFPR has been flat in low-income countries.

Trends among those 25-54: Duke mentioned that it would be illuminating to see what's going on with the participation rates and reasons for nonparticipation among working-age adults. I agree. The typical age range to consider questions related to working age adults is 25-54 (after college but before retirement). The U.S. Bureau of Labor Statistics reports reasons for nonparticipation for the full labor force and a couple of age groups on its website

You'll see that the total number of nonparticipants among 25-54 year olds increased from October 2014 to October 2015 by 287,000 people. The 468,000 increase in people during this period who do not want a job more than accounts for the total increase in nonparticipants in this age group. (The number of nonparticipants wanting a job declined, making up the difference.) This number of nonparticipants among 25-54 year olds who do not want a job, in fact, is a greater share of rising total nonparticipants in that age group (163 percent) than you see among those 55 years of age and older (104 percent).

If you're interested in looking at trends in nonparticipation over time by age group, check out the Atlanta Fed's Labor Force Dynamics web page And thanks again for reading!

Posted by: webmaster | November 19, 2015 at 07:35 AM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

October 05, 2015

Labor Report Silver Lining? ZPOP Ratio Continued to Rise in September

We have received several requests for an update of our ZPOP ratio statistic to incorporate September's data. We have also been asked whether the ZPOP ratio can be constructed from labor force data from the U.S. Bureau of Labor Statistics (BLS).

The ZPOP ratio is an estimate of the share of the civilian population aged 16 years and over whose labor market status is what they say they currently want (assuming that people who work full-time want to do so). A rising ZPOP ratio is consistent with a strengthening labor market. We constructed the ZPOP ratio from the microdata in the BLS's Current Population Survey, but we can also construct a very close approximation from the BLS's Labor Force Statistics data. Here's how (using data that are not seasonally adjusted):

The following chart shows the history of the resulting ZPOP ratio over 20 years, seasonally adjusted.

Unlike the headline U-3 unemployment rate, which remained unchanged from August to September, the seasonally adjusted ZPOP ratio improved slightly (from 92.0 to 92.1 percent). Relative to an estimated 230,000 increase in the population over the month, the improvement in the ZPOP ratio was the result of an increase in the number of people who said they do not currently want a job and a decline in involuntary part-time employment in excess of the decline in total employment.

Finally, the chart below shows the performance of the seasonally adjusted ZPOP ratio relative to the comparable employment-to-population (EPOP ratio) and the EPOP ratio for those aged 25–54. The relatively greater recovery in the ZPOP ratio since 2009 is primarily because the EPOP ratios do not adjust for the share of the population who say they do not currently want a job.


October 5, 2015 in Employment, Labor Markets, Unemployment | Permalink


The relatively greater recovery in the ZPOP ratio since 2009 is primarily because the EPOP ratios do not adjust for the share of the population who say they do not currently want a job.
From the post.

I assume, historically, for EPOP, that persons who said they did not want a job were included, i.e. epop has always been calculated that way but still the epop was higher previously than now.
What is the difference in the do not want a job numbers since 2009(second chart) and 1995(first chart). Has it increased. And are there reasons given for not wanting a job.
This comments box is very small and makes for difficult typing.

Posted by: am | October 15, 2015 at 04:28 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

September 22, 2015

The ZPOP Ratio: A Simple Take on a Complicated Labor Market

In her press conference following the latest FOMC meeting, Federal Open Market Committee (FOMC) Chair Janet Yellen emphasized that she still sees cyclical weakness in the labor market, even as the headline unemployment rate has moved close to FOMC participants' median estimate of its longer-run normal level.

She also noted that FOMC participants look at many different indicators of labor utilization, because the headline unemployment rate (commonly known as the U-3 rate) is overstating the health of the labor market. One alternative measure that has received some attention is the employment-to-population (EPOP) ratio. However, a well-recognized problem with the EPOP ratio is that because it defines utilization as employment, trends in demographic and behavioral labor force participation can affect it.

This problem is partially addressed by looking at the EPOP ratio for the prime-age population, or by making adjustments for demographic changes as suggested by Kapon and Tracy at the New York Fed and further analyzed by our Atlanta Fed colleague Pat Higgins. Here, we propose an alternative approach that uses a broader definition of utilization that makes it less affected by labor supply trends.

The Current Population Survey does not ask the question "are your labor services being fully utilized?" Therefore, we have to use our judgment to classify someone as fully utilized. The figure below shows the choices we make. We assume that everyone who says they are working fewer hours than they want is underutilized (the red boxes). This includes those in the labor force but unemployed, those not in the labor force but wanting a job, and those working part-time but wanting full-time hours (similar to the treatment of underutilization in the broad U-6 unemployment rate measure).

Everyone working full-time, working part-time for a noneconomic reason, and those who say they don't want a job are considered fully utilized (the green boxes). Of course, this takes the "don't want a job" classification at face value. For example, someone who is retired is counted as fully utilized, irrespective of the (unknown) reason they chose to retire.

diagram of choices made in Current Population Survey

As shown in the Chart 1 below, the share of the population 16 years or older that is fully utilized—what we call the utilization-to-population (ZPOP) ratio—is currently about 1.5 percentage points below its prerecession level, after having fallen by 6 percentage points during the recession.

Chart 1: Z-Pop: The Share of the Population Fully Utilized

Notice that because the ZPOP ratio treats those who are not employed and don't want a job as fully utilized, it is less affected by demographic and behavioral trends in labor force participation than the EPOP ratio. (You can learn more on our website about how demographic and behavioral trends are affecting labor force participation.) When compared with the EPOP ratio, the ZPOP ratio paints a somewhat rosier picture of labor market conditions (see chart 2).

Chart 2: Z-Pop and the Employment-to-Population Ratio

In sum, the utilization-to-population (ZPOP) ratio is the share of the working-age population that is working full time, is voluntarily working part-time, or doesn't want to work any hours. According to this measure, about 91 percent of the working-age population is considered fully utilized. The remaining 9 percent are "underutilized" and are a roughly even mixture of the unemployed, those not in the labor force but wanting to work, and those working part-time but wanting full-time hours.

The headline U-3 unemployment rate is very close to its prerecession level but is thought to overstate the health of the labor market. At the same time, we think that the EPOP ratio overstates the amount of remaining labor market slack. The ZPOP ratio is in the middle; approaching its prerecession level but still with some way to go.

September 22, 2015 in Employment, Labor Markets, Unemployment | Permalink


Have you, recently or otherwise, discussed the actual data regarding falling labor force participation? The popular press and politicians say it's because geezers are "retiring". Yet the data, and a smattering of news pieces over the last few years, say quite otherwise. We geezers continue to work. No pensions, anymore.

Posted by: | September 23, 2015 at 09:18 AM

Alas, I can't edit that comment, and your link didn't show clearly. So, here's the actual BLS data on LPR:

While the geezer rate is lower then younger, the rate itself has been rising, and is predicted to continue rising.

Posted by: | September 23, 2015 at 09:29 AM

I've seen a bit of talk, recently I think through NBER, about the structural nature of the low/falling LFPR. I wonder whether prevailing economic conditions (specifically the general rate for compensation for one's low- or semi-skilled time) affects the way people answer whether or not they want a job in the first place.

Much like an improving economy attracts more people to the unemployment rolls, an improving wage floor may attract more people to want a job; the ZPOP might be improved by devising an additional variable to account for such a response.

Posted by: Bryan F | October 06, 2015 at 10:25 AM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

September 01, 2015

Should I Stay or Should I Go Now?

A recent article by Jason Faberman and Alejandro Justiniano at the Chicago Fed shows that there is a strong relationship between quit rates—as a proxy for the pace of job switching—and wage growth. Movements in the quit rate and wage growth are both procyclical. A tighter (weaker) labor market implies workers are more (less) likely to find better employment matches, and employers are more (less) willing to offer higher wages to attract new workers and retain existing workers.

To get some idea of the different wage outcomes of job switching versus job staying, we can use microdata underlying the Atlanta Fed's Wage Growth Tracker from the Current Population Survey. The following chart plots the quarterly private-sector quit rate (orange line) from the Job Openings and Labor Turnover Survey using Davis, Faberman, and Haltiwanger (published in 2012 in the Journal of Monetary Economics) estimates before 2001. Also shown is the median year-over-year wage growth of private-sector wage and salary earners who switched jobs (blue line) or stayed in the same job (green line). Job stayers are approximated by the restriction that they are in the same broad industry and occupation as 12 months earlier and have been with the same employer for each of the last four months. Job switchers do not satisfy these restrictions but were employed in the current month and 12 months earlier.

Private Sector Quit Rate and Wage Growth

The correlation between the quit rate and median wage growth is strongly positive and is slightly higher for job switchers (0.91) than for job stayers (0.88). In most periods, the median wage growth of job switchers is higher than for job stayers. This difference is consistent with the notion that job switching tends to involve moving to a better-paying job. However, during periods when the quit rate is slowing, median wage growth slows for both job stayers and switchers (reflecting the correlation between quits and wages), and the wage-growth premium from job switching tends to vanish.

Since the end of the last recession, the quit rate has been rising and a wage-growth premium for job switching has emerged again. Interestingly, during the last year, the wage growth of job stayers appears to have strengthened as well, consistent with a general tightening of the labor market.

September 1, 2015 in Employment, Labor Markets, Wage Growth | Permalink


Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

August 21, 2015

No Wage Change?

Even when prevailing market wages are lower, businesses can find it difficult to reduce wages for their current employees. This phenomenon, often referred to as "downward nominal wage rigidity," can result in rising average wages for incumbent workers despite high unemployment levels. Some economic models predict that a period of subdued wage growth can follow, even as the labor market recovers—a kind of delayed wage-adjustment effect.

In her 2014 Jackson Hole speech, Fed Chair Janet Yellen suggested this effect may explain sluggish growth in average wages in recent years, despite significant declines in the rate of unemployment.

This macroblog post looks at evidence of wage rigidity, particularly a spike in the frequency of zero wage changes relative to wage declines. A comparison is made between hourly and weekly wages and between incumbent workers (job stayers) and those who have changed employers (job switchers).

Chart 1 shows the fractions of job stayers reporting the same or a lower hourly or weekly wage than 12 months earlier. These measures are constructed from the Current Population Survey microdata in the Atlanta Fed's Wage Growth Tracker. They include workers who are paid hourly (accounting for about 60 percent of all wage and salary earners). The measures exclude those who usually receive overtime and other supplemental pay and those with imputed or top-coded (redacted) wages. Weekly wage is defined as the hourly wage times the usual number of hours per week worked at that rate. The data are aggregated to an annual frequency (except for 2015, where the first six months of the year are covered).

Job stayers cannot be exactly identified in the data and are approximated by those who are in the same occupation and industry as they were 12 months earlier and the same job as they were in the prior month. Consistent with other studies (see, for example, the work of our colleagues at the San Francisco Fed), we find that the incidence of unchanged hourly wages among job stayers is substantial (although some of this is probably the result of rounding errors in self-reported wages). The measured share of unchanged hourly wages rose disproportionately between 2008 and 2010, and it has remained elevated since. Zero hourly wage changes (the green line in chart 1) have become almost as common as declines in hourly wages (the blue line in chart 1).


Chart 1 also suggests that weekly wages for job stayers show a pattern over time broadly similar to hourly wages. But the fraction of unchanged weekly wages (the purple line in chart 1) is lower. Each year, about 60 percent of those with no change in their hourly wage had no change in their weekly wage (or hours) either. Also, there are relatively more declines in weekly wages (the orange line in chart 1) than in hourly wages—mostly the result of reduced hours worked. On average, a reduction in weekly wages is associated with a four-hour decline in hours worked per week. About 90 percent of those with lower hourly wages also had lower weekly wages, and 20 percent of those with no change in their hourly wage had a lower weekly wage (working fewer hours).

If job stayers show a relatively high incidence of no wage change, we might expect a different story for job switchers, since they are establishing a new wage contract with a new employer. Chart 2 shows the fraction of job switchers reporting the same or a lower hourly or weekly wage than 12 months earlier. Job switchers are approximated by workers who are in a different industry than a year earlier.


Not surprisingly, a smaller share of workers experience no change in their hourly or weekly wage when switching jobs. But the pattern of zero wage change for job switchers over time is generally similar to that of job stayers. It is also true that a decline in hourly and weekly wages is more likely for job switchers than for job stayers, with a significant temporary spike in the relative frequency of wage declines for job switchers during the last recession.

Taken at face value, this analysis suggests the presence of some amount of wage rigidity. Also, rigidity increases during recessions and has remained quite elevated since the end of the last recession—especially for job stayers. The question then becomes whether this phenomenon has important macroeconomic consequences. A prediction of most models in which wage stickiness has allocative effects is that it causes firms to increase layoffs when faced with a decline in aggregate demand. Interestingly, during the last recession—when wage stickiness appears to have increased substantially—the rate of layoffs was not unusually high relative to earlier recessions. What was atypical was the size of the decline in the rate of job creation, and this decline contributed to unusually long unemployment spells. As noted by Elsby, Shin, and Solon (2014), it is not clear that an increase in wage rigidity would constrain the hiring of new workers more than it constrains the retention of existing workers.

On the other hand, persistently high wage rigidity in the wake of the Great Recession is consistent with the relatively sluggish pace of wage increases seen in most measures of aggregate wage growth via the "bending" of the short-run Phillips curve (as described by Daly and Hobijn (2014)). Interestingly, the Atlanta Fed's Wage Growth Tracker is an exception. It has indicated somewhat stronger wage growth during the last year than other measures. It will be interesting to see if that trend continues in coming months.

Photo of John Robertson
By John Robertson, a senior policy adviser in the Atlanta Fed's research department

August 21, 2015 in Employment, Labor Markets, Wage Growth | Permalink


Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

July 15, 2015

Have Changing Job and Worker Characteristics Restrained Wage Growth?

In the wake of the Great Recession, nominal wage growth has been subdued. But it is unclear how much of this relatively low wage growth reflects protracted weakness in the labor market versus other factors, such as changes in the composition of the workforce and jobs over time. Wage growth tends to vary across personal and job characteristics, so it stands to reason that changes in the composition of the workforce, alongside demographic and work characteristics, could be an important explanation of overall movements in wage growth.

In this post, we explore the impact of the changing mixture of worker characteristics (by age and education) and types of jobs (by industry and occupation) on the Atlanta's Fed Wage Growth Tracker. We find that composition effects do not account for the low median wage growth experienced in recent years. Holding worker and job characteristics fixed at their 1997 shares raises the median wage growth in 2014 by only about 0.2 percentage point. Our results are consistent with the analysis in a previous macroblog post, which found that changing industry-employment shares could not explain much of the sluggish growth in the average hourly earnings data from the payroll survey.

Median wage growth, composition change by worker characteristics
In terms of demographics, we consider two features: a worker's age and education. As shown in this earlier macroblog post, younger workers tend to experience higher median wage growth than do older workers. Although older workers tend to be paid more based on experience, they are also more likely to be near the top of the wage distribution for their job, so the median older worker experiences less wage growth. The difference is quite large. In 2014, the median wage growth of workers over age 54 was around 1.2 percentage points lower than the overall median.

A person's education can also affect his or her wage growth. Workers with a high school diploma or less tend to have lower median wage growth. In 2014, the median wage growth of less-educated workers was about 0.1 percentage point lower than the overall median, reflecting that these workers are more likely to be earning minimum wage, which does not change very frequently.

In addition, the employment shares by age and education have changed over time. The proportion of workers in the Atlanta Fed's Wage Growth Tracker data who are over age 54 has more than doubled from 12 percent in 1997 to 25 percent in 2014. During the same period, the share of workers without a college degree has declined from 63 percent to 49 percent (see the charts).

Education and Age Distribution Over Time

Wage growth, composition change by job characteristics
In terms of job characteristics, we consider two features: the worker's industry (where they work) and their occupation (what they do). Before 2011, workers in service-producing industries experienced slightly higher (about 0.1 percentage point) median wage growth than all workers. But since then, the trends have flipped. In recent years, median wage growth of individuals working in service-producing industries has been slightly below the median wage growth of all workers.

Nonetheless, workers in professional occupations such as managerial, legal, scientific, and engineering jobs tend to experience relatively higher median wage growth. In 2014, the median wage growth of workers in these professional jobs was 0.2 percentage point higher than the median wage growth for all workers.

The share of workers in service-producing industries and in professional jobs has increased moderately over time. In 1997, 77 percent of workers in the data were employed in service-producing industries. In 2014, the share had increased to 82 percent. During the same period, the share of workers in professional occupations rose from 36 percent to 41 percent.

Composition effects on median wage growth
Individually, an aging workforce is putting downward pressure on wage growth, whereas rising education levels are adding upward pressure. The rising share of workers in professional occupations is also pushing wages up somewhat, although the impact of the rising share of workers in service-producing industries is ambiguous. But how large are these effects when combined?

To get an idea, we conducted two counterfactual experiments. First, we held fixed the age and education distributions at their 1997 levels (the first year in our Wage Growth Tracker data). Second, we held fixed the age, education, industry, and occupation characteristics at their 1997 levels. We used three age groups (16–24, 25–54, and 55-plus years of age), two education groups (college degree and no college degree), two industry groups (service- or goods-producing industries), and two occupation groups (professional and other occupations).

The blue line in the next chart is the median wage growth over time with no adjustments for changes in composition. For example, for 2014, the chart shows median wage growth of workers in the data set with earnings in January 2014 and January 2013, February 2014 and February 2013, etc. This depiction is the Atlanta Fed Wage Growth Tracker, but at an annual frequency. The other two lines show the results of the experiment: demographically adjusted (green) and both demographically and job adjusted (orange).

These experiments suggest that—for our data set, at least—the impact on the median of the wage growth distribution from shifts in the composition of the workforce and jobs over time has increased in recent years, but the impact is not especially large. For example, the unadjusted median wage growth for 2014 is 2.5 percent. Holding fixed all four characteristics at their 1997 levels would have raised this by only 0.2 percentage point. Shifting worker and job characteristics are not a primary explanation of low median wage growth since 2009.


July 15, 2015 in Employment, Labor Markets, Wage Growth | Permalink


Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

June 19, 2015

Will the Elevated Share of Part-Time Workers Last?

There seems to be mounting evidence that at least part of the elevated share of part-time employment in the economy is here to stay. We have some insights to offer based on a recent survey of our business contacts.

Why are we interested? A higher part-time share of employment isn't necessarily a bad thing, if people are doing so voluntarily. Unfortunately, the elevated share is concentrated among people who would prefer to be working full-time. Using the average rate of decline over the past five years, the part-time for economic reasons (PTER) share of employment is projected to reach its prerecession average in about 10 years.

This is significantly slower than the decline in the unemployment rate, whose trajectory suggests a much sooner arrival—in around a year. The deviation raises an important policy question for measuring the amount of slack there is beyond what the unemployment rate suggests, and ultimately the extent to which policy can effectively reduce it.

What are the drivers? Data versus anecdotes
Researchers (here, here, and here) have pointed to factors such as industry shifts in the economy, changing workforce demographics, rising health care costs, and the Affordable Care Act as potentially important drivers of this shift. But we can glean only so much information from data. When a gap develops, we generally turn to our business contacts who are participating members in our Regional Economic Information Network (REIN) to fill in the missing information.

According to our contacts, the relative cost of full-time employees remains the most important reason for having a higher share of part-time employees than before the recession, which is the same response we received in last summer's survey on the same topic. Lack of strong enough sales growth to justify conversion of part-time to full-time workers came in as a close second.

The importance rating for each of the factors was notably similar to last year's survey, with one exception. Technology was rated as somewhat important, reflecting an uptick from the average response we received last year. We've certainly heard anecdotally that scheduling software has enabled firms to better manage their part-time staff, and it seems that this factor has gained in importance over the past year.

The chart below summarizes the reasons our business contacts gave in the July 2014 and the May 2015 surveys. The question was asked only of those who currently have a higher share of part-time workers than they did before the recession. The chart shows the results for all respondents, whether they responded to one or both surveys. When we limited our analysis to only those who responded to both surveys, the results were the same.

Will the elevated share persist?
The results suggest that a return to prerecession levels is unlikely to occur in the near term.

The chart below shows employers' predictions for part-time employment at their firms, relative to before the recession. About 27 percent of respondents believe that in two years, their firms will be more reliant on part-time work compared to before the recession. About 7 percent do not currently have an elevated share of part-time employees but believe they will in two years. About two-thirds believe their share of part-time will be roughly the same as before, while only 8 percent believe they will have less reliance on part-time workers compared to before the recession.

The majority of our contacts believe their share of part-time employment will normalize over the next two years, but some believe it will stay elevated. Still, 2017 does not mean the shift will be permanent. In fact, firms cited a balance of cyclical and structural factors for the higher reliance on part-time. Low sales growth and an ample supply of workers willing to take part-time jobs could both be viewed as cyclical factors that will dissipate as the economy further improves.

Meanwhile, higher compensation costs of full-time relative to part-time employees and the role of technology that enables companies to more easily manage their workforce can be considered structural factors influencing the behavior of firms. Firms that currently have a higher share of part-time employees gave about equal weight to these forces, suggesting that, as other research has found, both cyclical and structural factors are important explanations for the slow decline in the part-time share of employment.

June 19, 2015 in Business Cycles, Employment, Labor Markets, Unemployment | Permalink


Current technologies are a great enabler, this may not have been the case in the past. But one of the reasons, which needs further study is the fall out of M&A and the impact on payrolls, which makes very little allowance for full-time additions thereafter. The full time additions have been more in the retail space or service area, followed by technology, while we have seen dwindling fortunes in the Oil & Natural Gas sector, the last one has seen a switch to part-time.

Posted by: Procyon Mukherjee | June 21, 2015 at 11:26 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

June 05, 2015

Atlanta Fed's Wage Growth Measure Increased Again in April

A measure of 12-month wage growth constructed here at the Atlanta Fed increased by 3.3 percent in April. This rate is up from 3.1 percent in March and at its highest level since March 2009 (see the chart).


As mentioned in an earlier macroblog post, this measure behaves broadly like the wage and salary component of the Employment Cost index (ECI). The ECI data pertain to the last month in the quarter and are published with about a four-week lag. In contrast, the Atlanta Fed measure uses individuals' hourly wage data, 12 months apart, from the Current Population Survey (CPS). The data come from publicly available CPS microdata produced by the U.S. Bureau of Labor Statistics (BLS) and are typically released two or three weeks after the monthly BLS labor report.

Timeliness is one thing, but is it useful? It turns out there is a relatively strong correlation between this wage growth measure and the employment rate (100 minus the unemployment rate) lagged by 12 months (see the chart).


At least in terms of this measure of wage growth, it seems that improvement in labor utilization is translating into rising wage growth. This development is something our boss, Atlanta Fed President Dennis Lockhart, has been looking for. We expect to be able to update this wage growth measure with the May CPS data in a few weeks.

Photo of John Robertson
By John Robertson, a senior policy adviser in the Atlanta Fed’s research department

June 5, 2015 in Employment, Labor Markets, Wage Growth | Permalink


"It turns out there is a relatively strong correlation between this wage growth measure and the employment rate."

Didn't Keynes say this nearly 80 years ago? Of course Keynes is "old school" so his observations are necessarily outdated.

Posted by: Paul Mathis | June 05, 2015 at 05:49 PM

It seems that the improvement in the labor market appears to lag the improvement in wages. You would think that it should be reversed, an increase in employment leads to improved wages

Posted by: Ayelet | June 07, 2015 at 10:18 AM

@Ayelet - I think it shows the increase in wages lags the improvement in the employment rate which is what is expected by theory and gives confidence that the continuing improvement in employment should lead to more increases in wages. Given this data would be interesting to see include labor productivity in the analysis as a perceived problem in both the US and UK recovery has been the lack of productivity gains leading to concerns that the long-term potential growth rate has been reduced

Posted by: Oliver Bunnin | June 09, 2015 at 03:47 AM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

May 01, 2015

Signs of Strengthening Wage Growth?

The average hourly earnings measure for the private sector, reported in the U.S. Bureau of Labor Statistics's Establishment Survey, increased by a meager 2.1 percent in the first quarter (year over year). This increase was barely above the 2.0 percent pace observed in the fourth quarter of last year. However, Thursday's Employment Cost Index report showed a more sizable uptick in the wage and salary growth picture. Year-over-year growth in the first quarter was 2.5 percent, up from 2.1 percent in the fourth quarter of 2014. Another wage measure that we discussed in a February macroblog post also moved notably higher in the first quarter. That measure, which is derived from earnings data in the Current Population Survey, increased from 2.8 percent in the fourth quarter of 2014 to 3.2 percent in the first quarter of this year (see the chart).

This Wall Street Journal article (subscription required) also notes that anecdotal signs suggest a turnaround in wage growth, especially among lower-wage occupations. Overall, we take the evidence to suggest some emerging momentum in wage growth. Rising wage growth is an encouraging sign and is consistent with a tightening labor market.

Photo of John Robertson
By John Robertson, senior policy adviser, and
Photo of Ellie Terry
Ellie Terry, an economic policy analysis specialist, both of the Atlanta Fed's research department

May 1, 2015 in Employment, Labor Markets, Wage Growth | Permalink


Great chart, thank you. Where can I find the data for CPS median hourly wage (blue line) ?

Posted by: TM | May 05, 2015 at 09:50 AM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search

Recent Posts

November 2015

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          



Powered by TypePad