The Atlanta Fed's macroblog provides commentary on economic topics including monetary policy, macroeconomic developments, financial issues and Southeast regional trends.
- BLS Handbook of Methods
- Bureau of Economic Analysis
- Bureau of Labor Statistics
- Congressional Budget Office
- Economic Data - FRED® II, St. Louis Fed
- Office of Management and Budget
- Statistics: Releases and Historical Data, Board of Governors
- U.S. Census Bureau Economic Programs
- White House Economic Statistics Briefing Room
February 05, 2016
Introducing the Refined Labor Market Spider Chart
In January 2013, Atlanta Fed research director Dave Altig introduced the Atlanta Fed's labor market spider chart in a macroblog post.
In a follow-up post that June, Atlanta Fed colleague Melinda Pitts and I introduced a dedicated page for the spider chart located at the Center for Human Capital Studies (CHCS) webpage. It shows the distribution of 13 labor market indicators relative to their readings just before the 2007–09 recession (December 2007) and the trough of the labor market following that recession (December 2009). The substantial improvement in the labor market during the past three years is quite evident in the spider chart below.
As of December 2012, none of the indicators had yet reached their prerecession levels, and some had a long way to go. Now, many of these indicators are near their prerecession values—and some have blown by them.
To make the spider chart more relevant in an environment with considerably less labor market slack than three years ago, we are introducing a modified version, which you can see here. Below is an example of a chart I created using the menu-bars on the spider chart's web page:
In this chart, I plot the May 2004 and November 2015 percentile ranks of labor market indicators relative to their distributions since March 1994. As with the previous spider chart, indicators such as the unemployment rate, where larger values indicate more labor market slack, have been multiplied by –1. The innermost and outermost rings represent the minimum and maximum values of the variables from March 1994 to January 2016. The three dashed gray rings in between are the 25th, 50th, and 75th percentiles of the distributions. For example, the November 2015 value of 12-month average hourly earnings growth (2.26 percent) is the 23rd percentile of its distribution. This means that 23 percent of the other monthly observations on hourly earnings growth since March 1994 are lower than it is.
I chose May 2004 and November 2015 because they had the last employment situation reports before "liftoffs" of the federal funds rate. November 2015 appears to be stronger than May 2004 for some indicators (job openings, unemployment rate, and initial claims) and weaker for others (hires rate, work part-time for economic reasons, and the 12-month growth rate of the Employment Cost Index).
The average percentile ranks of the variables for these two months are similar, as the chart below depicts:
Also shown in the chart is the Kansas City Fed's Level of Activity Labor Market Conditions Indicator. It is a sum of 24 not equally weighted labor market indicators, standardized over the period from 1992 to the present. In spite of its methodological and source-data differences with the average percentile rank measure plotted above, it tracks quite closely, especially since 2004. However, as shown in the spider chart that I referred to above, there is quite a bit of variation within the indicators that may provide additional information to our analysis of the average trends.
We made a number of other changes to the spider chart to ensure it reflects current labor market issues. These changes are documented in the FAQs and "Indicators" sections of the new spider chart page. Of particular note, users can choose not only the years for which they wish to track information, but also the period of reference that provides the basis of the spider chart. The payroll employment variable is now the three-month average change rather than a level. Temporary help services employment has been dropped, and two measures of 12-month compensation growth and the employment-population ratio (EPOP) for "prime-age workers" (25 to 54 years) have been added.
Some care should be taken when comparing recent labor market data values with those 10 or more years ago as structural changes in the labor market might imply that a "normal" value today is different than a "normal" value in, say, 2004. The variable choices for the refined spider chart were made to mitigate this problem to some extent. For example, we use the prime-age EPOP as a crude adjustment for population aging, putting downward pressure on the labor force participation rate and EPOP over the past 10 years (roughly 2 percentage points). This doesn't entirely resolve the comparability issue since, within the prime-age population, the self-reporting rate of illness or disability as a reason for not wanting a job has increased about 1.5 percentage points since 1998 (see the macroblog posts here and here and the CHCS Labor Force Participation Dynamics webpage). If this increase in disability reporting is partly structural—and a Brookings study by Fed economist Stephanie Aaronson and others concludes it is—some of the decline in the prime-age EPOP since the late 1990s may not be a result of a weaker labor market per se.
Other variables in the spider chart may have had structural changes as well. For example, a study by San Francisco Fed economists Rob Valleta and Catherine van der List concludes that structural factors explain just under half of the rise in the share of workers employed part-time for economic reasons over the 2006 to 2013 period.
To partially account for structural changes in trends, we allow the user to select one of 11 time periods over which the distributions are calculated. The default period is March 1994 to present, which is what was used in the example above, but users can choose a window as short as five years where, presumably, structural changes are less important. A trade-off with using a short window is that a "normal" value may not produce a result close to the median. For example, the median unemployment rate is 5.6 percent since March 1994 and 7.3 percent since February 2011. The latter value is much farther away from the most recent estimates of the natural rate of unemployment from the Congressional Budget Office and the Survey of Professional Forecasters (both 5.0 percent).
In our June 2013 macroblog post introducing the spider chart, we wrote that we would reevaluate our tools and determine a more appropriate way to monitor the labor market when "the labor market has turned a corner into expansion." The new spider chart is our response to the stronger labor market. We hope users find the tool useful.
January 29, 2016
Shrinking Labor Market Opportunities for the Disabled?
The labor force participation rate (LFPR) among prime-age (25–54 years old) people averaged 80.8 percent in 2015, down 1.8 percentage points (2.6 million people) from 2009, according to the U.S. Bureau of Labor Statistics. According to our calculations from the Current Population Survey (CPS), a drop in LFPR among individuals with disabilities accounts for about a fifth of that decline.
Many people with disabilities are active in the labor force, working or looking for work. But the disabled LFPR has fallen a lot in recent years—it's down from 39.3 percent in 2009 to 34.5 percent in 2015. In other words, for some reason, more prime-age individuals with disabilities have opted out of the labor market.
A rising share of the prime-age population with a disability is not the culprit. In fact, the 2015 average disability rate was 6.4 percent, the same as in 2009. It is possible that the severity rather than incidence of disabilities has increased in recent years; labor market attachment does vary with type of disability, as this report shows.
But we suspect that the relatively large decline in disabled labor market attachment probably has also to do with shifts in employment opportunities for those with a disability. Some insight into this issue can be seen by looking at the change in employment shares in occupations that tend to have relatively low pay.
Workers with a disability tend to make less than nondisabled workers. We estimate the median wage of a worker with a disability in 2009 to have been 76 percent that of a nondisabled worker. In 2015, the relative median wage had declined to 74 percent. This drop is partly related to a relative increase in the share of employment for workers with a disability in low-paying occupations (which we define as jobs in personal care, food services, janitorial services, etc.), as the chart shows. The employment news has not been all bad for workers with a disability, however. There has been a rise in the share of employment of people with disabilities in higher-paying occupations since 2009, although they do tend to earn less than other workers in those types of occupations.
For some workers with disabilities, the financial return to employment versus nonemployment may have become somewhat less attractive in recent years. One factor related to the decision to engage in the labor market is the ability to collect Social Security Disability Insurance (SSDI). SSDI claims rose notably when the unemployment rate was high, which is consistent with the idea that the expected return to labor market activity for some individuals with a disability declined.
Job seekers with a disability have also struggled to find jobs offering the hours they desire. For example, the share of unemployed people finding full-time or voluntary part-time employment within a month, or "the hours-finding rate," is much lower for the prime-age disabled than for the nondisabled, and this share has improved relatively less over the recovery. Between 2009 and 2015, the average disabled hours-finding rate improved 4.7 percentage points, from 9.5 to 14.2 percent. During the same period, the nondisabled hours-finding rate increased 6.3 percentage points, from 13.0 to 19.3 percent.
The incidence of disability among prime-age individuals has not increased in recent years. But the labor market attachment of the disabled has declined, and this decline accounts for about one-fifth of the 1.8 percentage point fall in prime-age labor force participation between 2009 and 2015. Those with disabilities already have a harder time finding well-paying jobs, but that difficulty appears to have increased in that time span.
By John Robertson, a senior policy adviser in the Atlanta Fed's research department, and
January 07, 2016
What Occupational Projections Say about Entry-Level Skill Demand
On December 8, 2015, the U.S. Bureau of Labor Statistics (BLS) released its latest projections of labor force needs facing the U.S. economy from now until 2024.
Every two years, the BLS undertakes an extensive assessment of worker demand based on a number of factors: projected growth in the overall economy, dynamics of economic growth (such as which industries are growing fastest), labor force demographics (for example, the aging of the labor force), and expected changes in the labor force participation rate. Total worker demand includes both the number of workers needed to meet economic growth as well as the number of workers needed to replace current workers expected to retire.
A number of observations about these projections have already been identified. For example: overall employment growth will be slower, health care jobs will continue growing, and computer programmer jobs will lose ground.
In addition to the number of workers that will be in demand in different occupations in the U.S. economy, the BLS reports the skills that are needed for entry into those occupations—skills pertaining to both education levels and on-the-job training. As I perused this report, I was surprised at how much attention the press pays to the growth in high-skilled jobs at the expense of attention paid to those occupations requiring less skill but actually employ the greatest number of workers.
To be clear, the BLS does not project the educational requirements that will be needed for entering each occupation in 2024. It merely reports the most common education, training, and experience requirements needed to enter each occupation in the base year (in this case, 2014). Also it's important to note that these estimates of education needed to enter an occupation do not necessarily (and almost surely do not) match the average education of workers in that occupation at any given time, as those averages will reflect workers of many different ages and experience. The BLS gives a detailed description of how it identifies the entry-level educational and training requirements for each occupation. With those caveats in mind, let's take a look at the current distribution of jobs across the most common educational requirement for entering occupations.
The chart below tells us that, together, the typical entry-level requirement of a high school degree or less corresponds to nearly 64 percent of all jobs in the U.S. economy in 2014, while those typically requiring at least a bachelor's degree for entry represent 25.6 percent of jobs. The projected distribution of jobs in 2024 looks nearly identical: entry-level requirements (based on 2014 assessments) for 63 percent of all jobs requiring a high school degree or less and 26.2 percent requiring a bachelor's degree or more.
To be sure, the growth in higher-skill jobs far outpaces that for low- or middle-skill jobs. The number of jobs requiring a bachelor's degree or more for entry (in 2014) is expected to grow by 34 percent, whereas the number of jobs requiring less than a bachelor's degree is expected to grow by only 6 percent. This difference in growth rates reflects, in part, an expected continuation of the phenomenon of declining middle-skill jobs that my Atlanta Fed colleagues (and others) have discussed previously. Although labeled "middle-skill," entry into these occupations (such as office support and many manufacturing occupations) is not likely to require more than a high school degree.
However, even though the growth in low- and middle-skill jobs is expected to be slower than in higher-skill jobs, the total number of job openings based on predicted growth and replacement needs between 2014 and 2024 is expected to be nearly 32 million for jobs requiring less than a bachelor's degree for entry (based on 2014 assessments), with 30 million of those requiring only a high school degree or less. The total number of job openings requiring at least a bachelor's degree is expected to be about 12 million. In other words, the number of jobs requiring a high school degree or less in order to enter is twice as large as the number of jobs that require a college degree to enter.
The other side of this story, however, is that those jobs typically requiring less education at entry don't pay nearly as much as jobs requiring higher levels of education. The dollar figures in parentheses on the chart reflect the median annual salary of jobs with the different entry-level educational requirements. What we see is that while the majority of U.S. jobs require a high school degree or less at entry, those jobs pay less than half of what a job requiring at least a college degree pays.
So let's say a worker wants good job prospects (with a large number of job openings over the next decade), doesn't want to go to college, and wants to optimize chances for the highest salary possible. What is this worker to do? Fortunately, some of my colleagues at the Atlanta Fed, Cleveland Fed, and Philadelphia Fed have produced a report identifying what they call "opportunity occupations," which they define as those paying salaries higher than the geographically-adjusted national median for at least 70 percent of adults who have less than a college education. Some jobs among their top opportunity occupations are nurses, bookkeepers, first-line supervisors of retail workers, truck drivers, computer user support specialists, police officers, and electricians and workers in several other construction trades. Their report also identifies the U.S. metropolitan areas possessing a high share of opportunity occupations.Even though the share of jobs in the U.S. economy requiring less than a college degree at entry is getting smaller (very slowly), the largest number of jobs in the economy is, by far, jobs requiring less than a college degree at entry, and those jobs offer a wide range of options that pay above the national median wage.
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.
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 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):
- Take total employment, and add those not in the labor force who do not currently want a job. Then subtract those who were at work from one to 34 hours for economic reasons. The ZPOP ratio is this figure as a percentage of the civilian population 16 years and over.
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.
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.
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.
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).
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 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.
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.
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.
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).
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.
- Introducing the Refined Labor Market Spider Chart
- Shrinking Labor Market Opportunities for the Disabled?
- Are Long-Term Inflation Expectations Declining? Not So Fast, Says Atlanta Fed
- What Occupational Projections Say about Entry-Level Skill Demand
- A Closer Look at Changes in the Labor Market
- Should We Be Concerned about Declines in Labor Force Growth?
- Labor Report Silver Lining? ZPOP Ratio Continued to Rise in September
- The ZPOP Ratio: A Simple Take on a Complicated Labor Market
- What Do U.S. Businesses Know that New Zealand Businesses Don't? A Lot (Apparently).
- 5-Year Deflation Probability Moves Off Zero
- February 2016
- January 2016
- November 2015
- October 2015
- September 2015
- August 2015
- July 2015
- June 2015
- May 2015
- April 2015
- Business Cycles
- Business Inflation Expectations
- Capital and Investment
- Capital Markets
- Data Releases
- Economic conditions
- Economic Growth and Development
- Exchange Rates and the Dollar
- Fed Funds Futures
- Federal Debt and Deficits
- Federal Reserve and Monetary Policy
- Financial System
- Fiscal Policy
- Health Care
- Inflation Expectations
- Interest Rates
- Labor Markets
- Latin America/South America
- Monetary Policy
- Money Markets
- Real Estate
- Saving, Capital, and Investment
- Small Business
- Social Security
- This, That, and the Other
- Trade Deficit
- Wage Growth