The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.
- 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
July 31, 2017
Behind the Increase in Prime-Age Labor Force Participation
Prime-age labor force participation has been on a tear recently. Over the last eight quarters, it is up by about 65 basis points (bps) and more than 40 bps in just the last year. When combined with declines in the rate of unemployment, this increase has helped lift the employment-to-population (EPOP) ratio for this key population group by around 120 bps during the last two years.
Placed in the context of an almost 260 bp decline in the prime-age EPOP ratio between 2007 and 2015, this development is significant. Although the unemployment rate is close to what most economists consider full employment, rising labor force participation can indicate that the labor market might still have some room to run before the employment gap is fully closed. (The Congressional Budget Office offers some analysis consistent with this idea.)
So what's behind the increase in prime-age (defined as people between 25 and 54) participation in the last year? Changes in the labor force participation rate (LFPR) either can be the result of changes in the mix of demographic groups in the population with different average rates of participation (for example, across education and race/ethnicity), or they can result from changes in average participation rates within demographic groups. It turns out that most of the increase in the prime-age LFPR has been because of increased LFPR within demographic groups—in particular, prime-age women and especially women without a college degree. Prime-age men have not contributed much to the rise in participation beyond the increased participation associated with a more educated population.
The following chart shows the contribution to the change in the prime-age LFPR over the last year as a result of changes in the relative mix of age-education-race groups (the blue bars) and changes in participation rates within age-education-race groups (the orange bars). It shows the contribution from both sexes combined and from prime-age women and men separately.
Note that the we computed the contributions using six five-year age groups, three education groups (less than high school, high school but no college degree, and college degree), three race/ethnicity groups (Hispanic, non-Hispanic black, and non-Hispanic white/other), and two sexes.
Of the total increase in the prime-age LFPR, most of that was the result of changes in labor force participation behavior within female demographic groups. In fact, changes in LFPR behavior from prime-age men served as a drag on the overall prime-age LFPR. The modestly positive demographic effect on the LFPR for both men and women reflects the higher LFPR for those with a college degree and the relative increase in the share of both prime-age men and women with a college degree.
This development stands in contrast to the drivers of the change in the prime-age LFPR between 2015 and 2016. Of the 24 bp increase in prime-age LFPR between the second quarters of 2015 and 2016, changes in the demographic composition of the population (primarily increased education levels) accounted for all of it rather than changes in average participation rates within demographic groups.
The next chart shows the contribution to the change in the prime-age LFPR between 2016 and 2017 due to changes in the LFPR behavior of women for specific education-race groups.
As the chart shows, the bulk of the demographically adjusted contribution from female labor force participation came from women without a college degree, and the largest contribution across female education-race groups was from Hispanics without a college degree. The increase in labor force participation among women with less education is consistent with evidence of recent improvement in the wage gains for relatively low-wage earners.
Although this simple decomposition doesn't explain why nondegreed women are increasingly finding the labor force to be an attractive option, we can infer some clues by looking at changes in the reasons people give for not participating. In particular, the largest contribution from changes in behavior among prime-age women over the last year came from a decrease in the propensity to be out of the labor force because of poor health or being in the shadow labor force (wanting a job but not looking).
Recently, former Minneapolis Fed President Narayana Kocherlakota has argued that macroeconomists should take more seriously the differences in behavior across demographic groups. The Atlanta Fed's Labor Force Dynamics web page contains more information on the behavioral trends in the reasons people give for not participating in the labor force across demographic groups, and the page was just updated to include data for the second quarter of 2017. Check it out, and we'll keep reporting here on the relative contributions to the labor force of behavioral versus demographic changes—and whether the winning streak for prime-age labor force participation continues.
July 12, 2017
An Update on Labor Force Participation
With the unemployment rate essentially back to prerecession levels, economists have been paying increased attention to the labor force participation rate (LFPR). Many economists, including those at the Congressional Budget Office , believe untapped resources remain on the sidelines of the labor market.
What exactly does "on the sidelines" entail? Discouraged workers are only a small part of the story. To help unravel the rest of the mystery behind the elevated share of people not participating, we at the Atlanta Fed use the microdata from the Current Population Survey to code the activities of persons not in the labor force. We then calculate how changes in each activity contribute to the total change in the LFPR.
The chart below depicts the drivers of the change in the LFPR from the first quarter of 2016 to the first quarter of 2017. (The interactive tool on our website allows you to make comparisons across gender, age group, and time.) The LFPR rose just slightly (about 0.06 percentage points). However, that small change was the net result of much larger countervailing forces. Other things equal, demographic changes during the year would have lowered the LFPR by around 0.14 percentage points. The aging of the population put significant downward pressure on the LFPR (pushing it down 0.24 percentage points), but a more educated workforce helped push up the LFPR (0.10 percentage points). If the age and education mix of the population had not changed, the LFP rate would have risen by about 0.19 percentage points (see the chart).
The following chart further breaks down the behavioral and cyclical components at work. After controlling for shifts in the demographic mix of the population during the year, the largest contributing factor was a decline in the rate of nonparticipation because of family responsibilities.
This is a particularly important explanation for prime-age women (defined as women between 25 and 54 years of age). A smaller share of prime-age women who say they are busy with home and/or family responsibilities accounts for about half of the 0.62 percentage point increase in LFPR that occurred between the first quarter of 2016 and the first quarter of 2017 (see the chart).
To examine factors affecting prime-age men's participation or to learn more about the cyclical and structural factors behind each reason, visit our website.
May 05, 2017
Slide into the Economic Driver's Seat with the Labor Market Sliders
The Atlanta Fed has just launched the Labor Market Sliders, a tool to help explore simple "what if" questions using actual data on employment, the unemployment rate, labor force participation, gross domestic product (GDP) growth, and labor productivity (GDP per worker).
We modeled the Labor Market Sliders after the popular Atlanta Fed Jobs Calculator. In particular, the sliders take the rate of labor productivity growth and the rate of labor force participation as given (not a function of GDP or employment growth) and then asks questions about GDP growth and labor market outcomes. Like the Jobs Calculator, the sliders require that things add up, a very useful feature for all those backyard economic prognosticators (we know you're out there).
Let's look at an example of using the sliders. The Congressional Budget Office (CBO) projects that the labor force participation rate (LFPR) will maintain roughly its current level of 62.9 percent during the next couple of years, as the downward pressure of retiring baby boomers and the upward pressure from robust hiring hold the rate stable. The CBO also projects that labor productivity growth will gradually increase to almost 1 percent over roughly the same period.
Suppose we want to know what GDP growth would be over the next couple of years (other things equal) if labor productivity, which has been sluggish lately, returned to 1 percent, as projected by the CBO. By moving the Labor Productivity slider in the tool to 1 percent and the Months slider to 24, you will see how productivity alone affects GDP growth: it increases to about 2 percent (see the image below). In this experiment, the unemployment rate, average job growth, and LFPR are constrained to current levels.
However, there's more than one way to achieve GDP growth of 2 percent over the next two years. Let's take a look.
Hit the reset button, and productivity, GDP growth, and months revert to their starting values. Then move the Months slider to 24 and the GDP Growth slider to 2 percent. You then see that—at current levels of labor force participation and labor productivity growth—achieving 2 percent GDP growth over the next two years would require the economy to create about 200,000 jobs per months (see the image below), which would push the unemployment rate down to 3.1 percent (a rate not seen since the early 1950s).
Hit the reset button again. Achieving 2 percent GDP growth over the next two years is also realistic with a higher LFPR, some other things equal. First, move the Months slider to 24, then move the Labor Force Participation Rate slider to 63.7 percent. The higher LFPR is consistent with about 2 percent growth in GDP and roughly 200,000 additional jobs added each month (see the image below). (This scenario constrains the unemployment rate and labor productivity growth rate to their current levels.) Of course, we haven't seen the LFPR at 63.7 percent since 2012, but that's another discussion.
What if we wanted something a bit more ambitious, such as averaging 3 percent GDP growth over the next couple of years? Hit the reset button again, and try this scenario. Keep Labor Force Participation Rate at its current level (consistent with the CBO's projection), set Labor Productivity growth to 1 percent (also using the CBO projection as a guide), move the Months slider to 24, and the GDP Growth slider to 3 percent. The Labor Market Sliders allow us to see that the economy would need to add an average of about 240,000 jobs each month for those two years. This scenario, the tight-labor-market method of achieving 3 percent GDP growth, would bring the unemployment rate down to 2.6 percent.
However, suppose the United States were somehow able to recapture productivity growth of around 2 percent, which we experienced in the late 1990s and early 2000s. In that case, 3 percent GDP could be achieved at the current employment growth and unemployment rate.
I encourage you to play around and devise your own "what if" scenarios—and use the Labor Market Sliders to make sure they add up.
March 30, 2017
Bad Debt Is Bad for Your Health
The amount of debt held by U.S. households grew steadily during the 2000s, with some leveling off after the recession. However, the level of debt remains elevated relative to the turn of the century, a fact easily seen by examining changes in debt held by individuals from 2000 to 2015 (the blue line in the chart below).
Not only is the amount of debt elevated for U.S. households, but the proportion of delinquent household debt has also fluctuated significantly, as the red line in the above chart depicts.
The amount of debt that is severely delinquent (90 days or more past due) peaked during the last recession and remains above prerecession levels. The Federal Reserve Bank of New York reports these measures of financial health quarterly.
In a recent working paper, we demonstrate a potential causal link between these fluctuations in delinquency and mortality. (A recent Atlanta Fed podcast episode also discussed our findings.) By isolating unanticipated variations in debt and delinquency not caused by worsening health, we show that carrying debt—and delinquent debt in particular—has an adverse effect on mortality rates.
Our results suggest that the decline in the quality of debt portfolios during the Great Recession was associated with an additional 5.7 deaths per 100,000 people, or just over 12,000 additional deaths each year during the worst part of the recession (a calculation based on census population estimates found here). To put this rate in perspective, in 2014 the death rate from homicides was 5.0 per 100,000 people, and motor vehicle accidents caused 10.7 deaths per 100,000 people.
It is well understood that an individual experiencing a large and unexpected decline in health can encounter financial difficulties, and that this sort of event is a major cause of personal bankruptcy. Our findings suggest that significant unexpected financial problems can themselves lead to worse health outcomes. This link between delinquent debt and health outcomes provides more reason for public policy discussions to take seriously the nexus between financial well-being and public health.
March 02, 2017
Gauging Firm Optimism in a Time of Transition
Recent consumer sentiment index measures have hit postrecession highs, but there is evidence of significant differences in respondents' views on the new administration's economic policies. As Richard Curtin, chief economist for the Michigan Survey of Consumers, states:
When asked to describe any recent news that they had heard about the economy, 30% spontaneously mentioned some favorable aspect of Trump's policies, and 29% unfavorably referred to Trump's economic policies. Thus a total of nearly six-in-ten consumers made a positive or negative mention of government policies...never before have these spontaneous references to economic policies had such a large impact on the Sentiment Index: a difference of 37 Index points between those that referred to favorable and unfavorable policies.
It seems clear that government policies are holding sway over consumers' economic outlook. But what about firms? Are they being affected similarly? Are there any firm characteristics that might predict their view? And how might this view change over time?
To begin exploring these questions, we've adopted a series of "optimism" questions to be asked periodically as part of the Atlanta Fed's Business Inflation Expectations Survey's special question series. The optimism questions are based on those that have appeared in the Duke CFO Global Business Outlook survey since 2002, available quarterly. (The next set of results from the CFO survey will appear in March.)
We first put these questions to our business inflation expectations (BIE) panel in November 2016 . The survey period coincided with the week of the U.S. presidential election, allowing us to observe any pre- and post-election changes. We found that firms were more optimistic about their own firm's financial prospects than about the economy as a whole. This finding held for all sectors and firm size categories (chart 1).
In addition, we found no statistical difference in the pre- and post-election measures, as chart 2 shows. (For the stat aficionados among you, we mean that we found no statistical difference at the 95 percent level of confidence.)
We were curious how our firms' optimism might have evolved since the election, so we repeated the questions last month (February 6–10).
Among firms responding in both November and February (approximately 82 percent of respondents), the overall level of optimism increased, on average (chart 3). This increase in optimism is statistically significant and was seen across firms of all sizes and sector types (goods producers and service providers).
The question remains: what is the upshot of this increased optimism? Are firms adjusting their capital investment and employment plans to accommodate this more optimistic outlook? The data should answer these questions in the coming months, but in the meantime, we will continue to monitor the evolution of business optimism.
March 2, 2017 in Books, Business Inflation Expectations, Economic conditions, Economic Growth and Development, Forecasts, Inflation Expectations, Saving, Capital, and Investment, Small Business | Permalink
July 18, 2016
Lockhart Casts a Line into the Murky Waters of Uncertainty
Is uncertainty weighing down business investment? This recent article makes the case.
Uncertainty as an obstacle to business decision making and perhaps even a "propagation mechanism" for business cycles is an idea that that has been generating a lot of support in economic research in recent years. Our friend Nick Bloom has a nice summary of that work here.
Last week, the boss here at the Atlanta Fed gave the trout in the Snake River a break and made some observations on the economy to the Rocky Mountain Economic Summit, casting a line in the direction of economic uncertainties. Among his remarks, he noted that:
The minutes of the June FOMC [Federal Open Market Committee] meeting clearly pointed to uncertainty about employment momentum and the outcome of the vote in Britain as factors in the Committee's decision to keep policy unchanged. I supported that decision and gave weight to those two uncertainties in my thinking.
At the same time, I viewed both the implications of the June jobs report and the outcome of the Brexit vote as uncertainties with some resolution over a short time horizon. We've seen, now, that the vote outcome may be followed by a long tail of uncertainty of quite a different character.
But he followed that with something of a caution…
If uncertainty is a real causative factor in economic slowdowns, it needs to be better understood. Policymaking would be aided by better measurement tools. For example, it would help me as a policymaker if we had a firmer grip on the various channels through which uncertainty affects decision-making of economic actors.
I have been thinking about the different kinds of uncertainty we face. Often we policymakers grapple with uncertainty associated with discrete events. The passage of the event to a great extent resolves the uncertainty. The outcome of the Brexit referendum would be known by June 24. The interpretation of the May employment report would come clear, or clearer, with the arrival of the June employment report on July 8. I would contrast these examples of short-term, self-resolving uncertainty with long-term, persistent, chronic uncertainty such as that brought on by the Brexit referendum outcome.
As President Lockhart indicated in his speech, the Federal Reserve Bank of Atlanta conducts business surveys that attempt to measure the uncertainties that businesses face. From July 4 through July 8, we had a survey in the field with a question on how the Brexit referendum was influencing business decisions.
We asked firms to indicate how the outcome of the Brexit vote affected their sales growth outlook. Respondents could select a range of sentiments from "much more certain" to "much more uncertain."
Responses came from 244 firms representing a broad range of sectors and firm sizes, with roughly one-third indicating their sales growth outlook was "somewhat" or "much" more uncertain as a result of the vote (see the chart). Those noting heightened uncertainty were not concentrated in any one sector or firm-size category but represented a rather diverse group.
As President Lockhart noted in his speech, "[w]e had a spirited internal discussion of whether one-third is a big number or not-so-big." Ultimately, we decided that uncovering how these firms planned to act in light of their elevated uncertainty was the important focus.
In an open-ended, follow-up question, we then asked those whose sales growth outlook was more uncertain how their plans might change. We found that the most prevalent changes in planning were a reduction in capital spending and hiring. Many firms mentioned these two topics in tandem, as this rather succinct quote illustrates: "Slower hiring and lower capital spending." Our survey data, then, provide some support for the idea that uncertainties associated with Brexit were, in fact, weighing on firm investment and labor decisions.
Elevated measures of financial market and economic policy uncertainty immediately after the Brexit vote have abated somewhat over subsequent days. Once the "waters clear," as our boss would say, perhaps this will be the case for firms as well.
May 19, 2016
Are People in Middle-Wage Jobs Getting Bigger Raises?
As observed in this Bloomberg article and elsewhere, the Atlanta Fed's Wage Growth Tracker (WGT) reached its highest postrecession level in April. This related piece from Yahoo Finance suggests that the uptick in the WGT represents good news for middle-wage workers. That might be so.
Technically, though, the WGT is the median change in the wages of all continuously employed workers, not the change in wages among middle-income earners. However, we can create versions of the WGT by occupation group that roughly correspond to low-, middle-, and high-wage jobs, which allows us to assess whether middle-wage workers really are experiencing better wage growth. Chart 1 shows median wage growth experienced by each group over time. (Note that the chart shows a 12-month moving average instead of a three-month average, as depicted in the overall WGT on our website.)
Wage growth for all three categories has risen during the past few years. However, the timing of the trough and the speed of recovery vary somewhat. For example, wage growth among low-wage earners stayed low for longer and then recovered relatively more quickly. Wage growth of those in high-wage jobs fell by less but also has recovered by relatively less. In fact, while the median wage growth of low-wage jobs is back to its 2003–07 average, wage growth for those in high-wage jobs sits at about 75 percent of its prerecession average.
Are middle-wage earners experiencing good wage growth? In a relative sense, yes. The 12-month WGT for high-wage earners was 3.1 percent in April compared with 3.2 percent and 3.0 percent for middle- and low-wage workers, respectively. So the typical wage growth of those in middle-wage jobs is trending slightly higher than for high-wage earners, a deviation from the historical picture.
Interestingly, this pattern of wage growth doesn't quite jibe with the relative tightness of the labor market for different types of jobs. As was shown here, the overall WGT appears to broadly reflect the tightness of the labor market (possibly with some lag).
In theory, as the pool of unemployed shrinks, employers will face pressure to increase wages to attract and retain talent. Chart 2 shows the 12-month average unemployment rates for people who were previously working in one of the three wage groups.
Like the relationship between overall WGT and the unemployment rate, wage growth and the unemployment rate within these wage groups are negatively correlated (in other words, when the unemployment rate is high, wage growth is sluggish). The correlation ranges from minus 0.81 for low-wage occupations to minus 0.88 for middle-wage occupations.
However, notice that although the current gap between unemployment rates across the wage spectrum is similar to prerecession averages, the current relative gap in median wage growth is different than in the past. In particular, the wage growth for those in higher-wage jobs has been sluggish compared to middle- and lower-wage occupations.
Nonetheless, it's clear that the labor market is getting tighter. Wage growth overall has moved higher over the past year, driven primarily by those working in low- and middle-wage jobs. Is firming wage growth starting to show up in price inflation? Perhaps.
The consumer price index inflation numbers moved higher again in April, and Atlanta Fed President Dennis Lockhart said on Tuesday that—from a monetary policy perspective—recent inflation readings and signs of better growth in economic activity during the second quarter (as indicated by the Atlanta Fed's GDPNow tracker) are encouraging signs.
April 04, 2016
Which Wage Growth Measure Best Indicates Slack in the Labor Market?
The unemployment rate is close to what most economists think is the level consistent with full employment over the longer run. According to the Federal Open Market Committee's latest Summary of Economic Projections, the unemployment rate is currently only 15 basis points above the natural rate. Yet, average hourly earnings (AHE) for production and nonsupervisory workers in the private sector increased a paltry 2.3 percent in March from a year earlier (as did the AHE of all private workers), and is barely above its average course of 2.1 percent since 2009.In contrast, the Atlanta Fed's Wage Growth Tracker (WGT) suggests that wage growth has been increasing. The February WGT reading was 3.2 percent (the March data will be available later in April), considerably higher than its post-2009 average of 2.3 percent.
Why is there such a large difference between these measures of wage growth? Besides differences in data sources, the primary reason is that they measure fundamentally different things. The WGT is an estimate of the wage growth of continuously employed workers—the same worker's wage is measured in the current month and a year earlier.
In contrast, the AHE measure is an estimate of the change in the typical wage of everyone employed this month relative to everyone employed a year earlier. Most of these workers are continuously employed, but some of those employed in the current month were not employed the prior year, and vice versa. These changes in the composition of employment can have a significant effect.
A recent study by Mary C. Daly, Bart Hobijn, and Benjamin Pyle at the San Francisco Fed shows that while growth in wages tends to be pushed higher by the wage gains of continuously employed workers, the net effect of entry and exit into employment tends to put a drag on the growth in wages. Moreover, the magnitude of the entry/exit drag can be relatively large, varies over time, and differs by the type of entry and exit.
For example, older workers who have retired and left the workforce tend to come from the higher end of the wage distribution, and their absence from the current period wage pool exerts downward pressure on the typical wage. The greater number of baby boomers starting to retire is having an even larger depressing effect on growth in wages than in the past. Because the WGT looks only at continuously employed workers, it is not influenced by these net entry/exit effects.
To the extent that firms adjust the pay for incumbent workers in response to labor market pressures to attract and retain workers, the WGT should reasonably capture changes in the tightness of the labor market.
Economists at the Conference Board modeled the relationship between different wage growth series and measures of labor market slack. One of the slack measures they use is the unemployment gap—the difference between an estimate of the natural rate of unemployment and the actual unemployment rate.To illustrate their findings, the following chart shows the WGT and AHE measures along with the unemployment gap lagged six months (using the Congressional Budget Office estimate of the natural rate).
The WGT appears to move more closely with the lagged unemployment gap than does the growth in AHE, and a comparison of the correlation coefficients confirms the stronger relationship with the WGT. The correlation between the lagged unemployment gap and the change in average hourly earnings is 0.75.
In contrast, the correlation with the wage growth tracker is higher at 0.93. Moreover, the unemployment gap-AHE relationship appears to be particularly weak since the Great Recession. The correlation since 2009 falls to just 0.08 for the AHE, whereas the WGT correlation is still 0.93.
Our colleagues at the San Francisco Fed concluded their analysis of the effect of flows into and out of the employment on wage growth by suggesting that:
"... wage growth measures that focus on the continuously full-time employed are likely to do a better job of gauging labor market strength, since they are constructed to more clearly capture the wage dynamics associated with improving labor market conditions. The Federal Reserve Bank of Atlanta's Wage Growth Tracker is an example."
That assessment is consistent with the Conference Board study, and suggests that labor markets may be tighter than is commonly believed based on sluggish growth in measures of average wages such as AHE.
February 17, 2016
Are Paychecks Picking Up the Pace?
From the minutes of the January 26–27 meeting of the Federal Open Market Committee, it's clear that many participants saw tightening labor market conditions during 2015:
In their comments on labor market conditions, participants cited strong employment gains, low levels of unemployment in their Districts, reports of shortages of workers in various industries, or firming in wage increases.
Based on the Atlanta Fed's Wage Growth Tracker (WGT), the median annual growth in hourly wage and salary earnings of continuously employed workers in 2015 was 3.1 percent—up from 2.5 percent in 2014 and 2.2 percent in 2013. That is, the typical wage growth of workers employed for at least 12 months appears to be trending higher.
However, wage growth by job type varies considerably. For example, the WGT for part-time workers has been unusually low since 2010. The following chart displays the WGT for workers currently employed in part-time and full-time jobs. For those in part-time jobs, the WGT was 1.9 percent in 2015, versus 3.3 percent for those in full-time jobs. The part-time/full-time wage growth gap has closed somewhat in the last couple of years but is still large relative to its size before the Great Recession. Note that full-time WGT is similar to the overall WGT because most workers captured in the WGT data work full-time (81 percent in 2015).
In addition to hours worked, median wage growth also tends to vary across occupation. The following chart plots the WGT for workers in low-skill jobs, versus those in mid- and high-skill jobs. (We define low-skill jobs as those in occupations related to food preparation and serving; building and grounds cleaning; and maintenance, protection, and personal care services.)
Notably, after lagging during most of the recovery, median wage growth in low-skill occupations increased 2.8 percent in 2015, versus 2.0 percent in 2014 and compared to 3.2 and 2.7 percent for other occupations in 2015 and 2014, respectively.
The improvement in wage growth for low-skill occupations seems mostly attributable to full-time workers; wage growth for people in low-skill jobs working part-time was about half that (1.6 percent versus 3.0 percent) of those working full-time (see the chart).
This pickup in low-skill wage growth fits with some anecdotal reports we've been hearing. Some of our contacts in the Southeast have reported increasing wage pressure for workers in lower-skill occupations within their businesses. One can also see evidence of growing tightness in the market for low-skill jobs in the help-wanted data. As the following chart shows, the ratio of unemployed to online job postings for low-skill jobs is always higher than for middle- and high-skill occupations. But the ratio for low-skill jobs is now well below its prerecession level, and the tightness has increased during the last two years.
The take-away? Wage growth for continuously employed workers appears to have picked up some steam in 2015, and the recent trend in wage growth is positive across a variety of job characteristics. Wage growth for people in lower-skill jobs has increased during the last couple of years, consistent with evidence of increasing tightness in the market for those types of jobs. The largest discrepancy in wage growth appears to be among part-time workers, whose median gain in hourly wages in 2015 still fell well short of those in full-time jobs.
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.
- Behind the Increase in Prime-Age Labor Force Participation
- An Update on Labor Force Participation
- Another Look at the Wage Growth Tracker's Cyclicality
- GDPNow's Second Quarter Forecast: Is It Too High?
- Are Small Loans Hard to Find? Evidence from the Federal Reserve Banks' Small Business Survey
- Slide into the Economic Driver's Seat with the Labor Market Sliders
- The Fed’s Inflation Goal: What Does the Public Know?
- Going to School on Labor Force Participation
- Bad Debt Is Bad for Your Health
- Working for Yourself, Some of the Time
- July 2017
- May 2017
- April 2017
- March 2017
- February 2017
- January 2017
- December 2016
- November 2016
- October 2016
- September 2016
- 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