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
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.
April 11, 2017
Going to School on Labor Force Participation
In the aftermath of the Great Recession, labor force attachment declined. However, that pattern has been reversing itself lately. In particular, the labor force participation rate (LFPR) of the prime-age (25 to 54 years old) population, the core segment of the workforce, has been moving higher since late 2015. While this is good news, the prime-age LFPR remains well below prerecession levels, meaning that there are more than two million fewer prime-age people participating in the labor force. What factors have contributed to that decline? Where did those people go?
The Atlanta Fed LFP dynamics web page has an interactive tool that allows users to drill down into the drivers of the change in LFPR. The tool breaks the change in LFPR into two parts. The first part is the effect of shifts in the share of the population in different age groups (we use five-year age groups). The second part is the change attributable to shifts in the rate of nonparticipation. Using a methodology described here, we can drill deeper into the second part to learn more about the reasons for not participating in the labor force.
The U.S. Census Bureau will make the first quarter 2017 microdata on the reasons for nonparticipation available in a few weeks, so the following chart shows a decomposition of the 1.8 percentage point decline in the prime-age LFPR (not seasonally adjusted) between the fourth quarters of 2007 and 2016.
In this chart, "residual" pertains to the part of the total change in the LFPR that is attributable to the simultaneous shifts in both age-group population shares and age-group participation rates. In the present case, the residual is zero.
Because we are examining changes in prime-age participation, and all age groups within prime-age have reasonably similar participation rates, a change in the composition of ages tends to have little impact on the overall prime-age LFPR. Instead, the decline is due to shifts in the nonparticipation rate within age groups (the orange bar). In particular, the decline could indicate an increased likelihood of being in school, having family responsibilities that prevent participation, being in the shadow labor force (wanting a job but not actively looking), and a disability or poor health.
Although all these factors put downward pressure on participation, an important countervailing influence is that the education level of the population has been rising over time, and participation tends to increase with more education. In 2007, 41.0 percent of the prime-age population had a college degree, and they had an 88.3 percent participation rate versus 79.5 percent participation for those without a degree. By the end of 2016, the fraction with a degree had increased to 47.3 percent, and that cohort's participation rate had declined 1 percentage point, to 87.3 percent, versus a drop of 3.5 percentage points, to 76.0 percent, for those without a degree.
To see the importance of rising education on participation, the following chart shows the decomposition of the 1.8 percentage point decline in prime-age LFPR based on education-group population shares (degree and nondegree) instead of age-group shares.
In this chart, "residual" indicates the part of the total change in LFPR due to the simultaneous shifts in both education-group population shares and education-group participation rates.
As the chart shows, the shift in the education distribution of the prime-age population from 2007 to 2016 by itself would have increased the prime-age participation rate by about 0.7 percentage points (the green bar). Conversely, if education levels had not increased then the participation rate would have decreased by even more than it actually did. The nonparticipation effect would be larger for most nonparticipation reasons and especially for reasons of disability or poor health (−0.8 percentage points versus −0.5 percentage points). See the charts and analysis in the "health problems" section of the Labor Force Dynamics web page for more information on health-related nonparticipation by education.
Despite some partial reversal over the last year and a half, the prime-age LFPR is still lower than it had been prior to the recession. However, the decline in participation could have been even larger if the education level of the population had not also increased. Rising education is associated with a lower incidence of nonparticipation than otherwise would be the case, and it's principally associated with less nonparticipation attributable to disability or poor health. While researchers agree on the positive association between education and health, pinning down the specific reasons for this remains somewhat elusive. Factors such as income, informational, and occupational differences—as well as public policy choices—all play a role. Recent research by Nobel laureate Angus Deaton and Anne Case suggests that both education and racial differences are important considerations—emphasizing the sharply rising incidence of health problems among middle-age, white families with lower levels of education—and this Washington Post article highlights rising disability rates in rural America.
March 20, 2017
Working for Yourself, Some of the Time
Self-employment as a person's primary labor market activity has become much less commonplace in the United States (for example, see the analysis here and here ). This is a potentially important development, as less self-employment may indicate a decline in overall labor market mobility, business dynamism, and entrepreneurial activity (for example, see the evidence and arguments outlined here ).
Recessions can be particularly bad for self-employment, with reduced opportunities for potential business entrants as well as greater difficulty in keeping an existing business going (see here for some evidence on this). However, the rate of self-employment has been drifting lower over a long period, suggesting other factors are also playing a role in the decision to enter and exit self-employment.
One especially troubling development is the decline in the rate of self-employment for those in high-skill service providing jobs (management, professional, and technical services)—the people you might expect to be particularly entrepreneurial. For example, for workers aged 25 to 54 years old, the self-employment rate has declined from 13 percent in 1996 to 9 percent in 2016, and for those 55 years of age or older, the rate has dropped from 27 percent to 19 percent (using data from the Current Population Survey).
Not only are people in high-skill service jobs less likely to be self-employed than in the past, those who are self-employed are also less likely to be working full-time. The fraction usually working full-time has decreased from about 79 percent in 1996 to 74 percent in 2016. (The full-time rate for comparable private sector wage and salary earners has remained relatively stable at around 88 percent.) One possible explanation for the decline in hours worked is the last recession's lingering effects, which made it harder to generate enough work to maintain full-time hours. Another possibility is that more of the self-employed are choosing to work part-time.
It turns out that both explanations have played a role. The following chart shows the percent of part-time self-employment in high-skill service jobs. The blue lines are for unincorporated businesses and the green lines are for incorporated businesses. In order to distinguish cyclical and noncyclical effects, the chart shows the part-time rate for those who want to work full-time but aren't because of slack business conditions or their inability to find more work (part-time for economic reasons, or PTER), and those who work part-time for other reasons (part-time for noneconomic reasons, or PTNER).
In the chart, I classify someone as self-employed when that person's main job is working for profit or fees in his or her own business (and hence it does not capture people whose primary employment is a wage and salary job but are also working for themselves on the side). The self-employed could be sole proprietors or own their business in partnership with others, and the business may assume any of several legal forms, including incorporation. The chart pertains to the private sector, excluding agriculture, and part-time is usually working less than 35 hours a week.
On the cyclical side, the PTER rates (the dotted lines) rose during the last recession and have been slowly moving back toward prerecession levels as the economy has strengthened. In contrast, the PTNER rates (the solid lines) have moved higher since the end of the recession, continuing a longer-term trend. Choosing to work part-time has been playing an increasingly important role in reducing full-time self-employment in high-skill jobs. Note that there is not an obvious long-term trend toward greater PTNER for those self-employed in middle- or low-skill jobs (not shown).
Shifting demographics is one important factor contributing to the decline in average hours worked. In particular, the PTNER rate for older self-employed is much higher than for younger self-employed, and older workers are a growing share of part-time self-employed, a fact that reflects the aging of the workforce overall. (For more on the self-employment of older individuals, see here .) The net result is a rise in the fraction of self-employed choosing to work part-time. The higher rate of PTNER for the older self-employed appears to be mostly because of issues specific to retirement, such as working fewer hours to avoid exceeding social security limits on earnings.
The last recession and a relatively tepid economic recovery reduced the hours that some self-employed people have been able to work because of economic conditions. However, there has also been a longer-term reduction in how many hours other self-employed people (especially those in occupations requiring greater education and generating greater hourly earnings) choose to work. This increased propensity to work only part-time in their business is another factor weighing on overall entrepreneurial activity.
February 28, 2017
Can Tight Labor Markets Inhibit Investment Growth?
One of the most vexing developments of the current expansion has been the long and persistent reduction in the pace of business fixed investment (see chart 1).
The slide in investment spending evident in this chart has had a substantial impact on the pace of gross domestic product (GDP) growth in recent years and is also behind the slow pace of capital accumulation that has been a major factor in the slow labor productivity growth postrecession .
The other notable aspect of chart 1 is that employment growth has been robust during most of the recovery, and that growth remains robust. That sustained performance has taken the economy to the point where measures of labor market performance can be reasonably described as "close to a state of full employment."
Continued strong employment growth could sensibly support a relatively bullish story on investment going forward. As the table below shows, "high-pressure" labor markets—defined as periods when the official unemployment rate falls below the Congressional Budget Office's estimate of the "natural unemployment rate"—tend to be associated with strong levels of business fixed investment spending.
That said, we are taking note of some cautionary sounding from a special question about investment constraints on the most recent Federal Reserve Small Business Credit Survey, whose full results will be released in April. (The Small Business Credit Survey is a collaboration among Federal Reserve Banks and collects information from small businesses throughout the country. The 2016 survey was open from mid-September to mid-December, and generated more than 16,000 responses—about 10,000 of which were from employer firms.)
One of the survey's special questions was the following: What factors constrained your investment decisions over the past 12 months? The respondents were allowed to check as many factors as they deemed relevant and, perhaps not surprisingly, the collective answer was "a lot of things," as chart 2 shows.
Though there are a lot of contenders in that chart, it was interesting to us that the modal response (though admittedly by a hair) was an inability to find or retain qualified staff. It gets even more interesting when you focus on stable, growing firms—those that were profitable in 2016, are increasing payrolls and revenues, and have been in business for at least six years (see chart 3).
For this group—by definition, the most dynamic firms in the sample—perceived constraints on talent acquisition and retention is easily the largest issue when it comes to investment spending headwinds, independent of the size of the firm (measured by annual revenues). Indeed, more than 50 percent of the businesses with revenue in excess of $10 million identified the labor market as a problem.
We want to be sufficiently modest about interpreting these survey results. (The survey's full results will be released in April.) We have only asked this question once and therefore have no ability to compare with historical data. We also don't know for sure if firms truly are being constrained by their ability to find or retain qualified staff, or if respondents were simply identifying with that option as an issue with their business in general. But the idea that business investment could be constrained by access to talent is important for thinking about the growth potential of the economy. The possibility that education and workforce development efforts could have spillover effects into investment growth is intriguing.
February 23, 2017
More Ways to Watch Wages
The Atlanta Fed's Wage Growth Tracker slipped to 3.2 percent in January from 3.5 percent in December. The Wage Growth Tracker for women was 3.1 percent in January, down significantly from what we saw in late 2016, when gains topped 4 percent. For men, the January reading was 3.4 percent, very close to its average for the past 12 months. As I noted last month, I did not think the unusually high female wage growth was sustainable, and that proved to be the case. Since 2009, the Wage Growth Tracker for women has averaged about 0.3 percentage points below that for men—the same as the gap in the latest data.
Understanding why the Wage Growth Tracker slowed last month highlights the importance of being able to look beyond the top-line number. To provide Wage Growth Tracker users with more information, we have now added several additional cuts of the data to the Wage Growth Tracker web page. The amount of detail we can provide is limited by sample size considerations, and as a result, the additional data are reported as 12-month moving averages. The new data provide more detailed age, race, education, and geographic comparisons, as well as comparisons across broad categories of occupation, industry, and hours worked. As an example, here is a look at the (12-month average) median wage growth data for those who usually work full-time versus those who usually work part-time.
Have fun with these new tools, and we encourage you to comment and let us know what you think.
February 21, 2017
Unemployment versus Underemployment: Assessing Labor Market Slack
The U-3 unemployment rate has returned to prerecession levels and is close to estimates of its longer-run sustainable level. Yet other indicators of slack, such as the U-6 statistic, which includes people working part-time but wanting to work full-time (often referred to as part-time for economic reasons, or PTER), has not declined as quickly or by as much as the U-3 unemployment rate.
If unemployment and PTER reflect the same business-cycle effects, then they should move pretty much in lockstep. But as the following chart shows, such uniformity hasn't generally been the case. In the most recent recovery, unemployment started declining in 2010, but PTER started to move substantially lower beginning only in 2013. The upshot is that for each unemployed worker, there are now many more involuntary part-time workers than in the past.
Regarding the above chart, I should note that I adjusted the pre-1994 data to be consistent with the 1994 redesign of the Current Population Survey from the U.S. Bureau of Labor Statistics (see, for example, research from Rob Valletta and Leila Bengali and Anne Polivka and Stephen Miller ). This adjustment amounts to reducing the pre-1994 number of PTER workers by about 20 percent.
The elevated level of PTER workers has been most pronounced for workers in low-skill occupations. As shown in the next chart, PTER workers in low-skill jobs now outnumber unemployed workers who left low-skill jobs. Prior to the most recent recession, low-skill unemployment was always higher than low-skill PTER.
The increase in PTER workers is also mostly in the retail trade industry, as well as the leisure and hospitality industry, where low-skill occupations are concentrated. The PTER-to-unemployment ratio for the goods-producing sector (manufacturing, construction, and mining) has remained essentially unchanged. In those industries, unemployment and PTER move together.
Some researchers, such as our colleagues at the San Francisco Fed Rob Valletta and Catherine van der List, have argued that the increase in the prevalence of involuntary part-time work relative to unemployment suggests the importance of factors other than overall demand for labor. Among these factors are shifting demographics (a greater number of older workers who are less willing to do part-time work) and industry mix (more employment in industries with higher concentrations of part-time jobs). Such factors are almost certainly playing a role.
Recent analysis by Jon Willis at the Kansas City Fed suggests that the elevated levels of PTER in low-skill occupations may reflect that during the last recession, firms reduced the hours of workers in low-skill jobs more than they cut the number of low-skill jobs. In other words, firms still had some work that needed to get done, probably with peak demand at certain times of the day, and those tasks couldn't readily be outsourced or automated.
As the following chart from Willis's research shows, between 2007 and 2010, low-skill (non-PTER) employment actually increased slightly overall, but the mix of employment shifted dramatically toward part-time.
Since the recession, the pace of (non-PTER) low-skill job creation has been modest (about 20,000 jobs per month compared with 60,000 jobs per month in the years preceding the recession). Initially, this trend helped reduce low-skill unemployment more than the incidence of PTER—one reason why the ratio of PTER to unemployment continued to increase.
But the number of PTER workers in low-skill jobs has since been declining as more people have been able to find full-time jobs. At the current pace of job creation and (net) transition rates out of PTER, Willis estimates it would take until 2020 to return to prerecession levels of low-skill PTER. That seems a reasonable guess to me.
February 13, 2017
Does a High-Pressure Labor Market Bring Long-Term Benefits?
Though it ticked up slightly in January , the U.S. unemployment rate is arguably at, or near, its long-run sustainable level. At least that is the apparent judgment of Federal Open Market Committee participants, the Congressional Budget Office (CBO), and others. Not surprisingly, this consensus is leading to some speculation that a combination of policy and the economy's natural momentum may result in unemployment rates moving well below sustainable levels—a circumstance some have referred to as a "high-pressure" economy.
Though lower-than-normal unemployment rates may have benefits, at least in the short-term, it is generally recognized that these circumstances also carry risks. Specifically, if the demand for resources (including labor) expands beyond the economy's capacity to supply them, the risk of undesirable inflation, financial imbalances, and other negative developments may grow—a point that Boston Fed President Eric Rosengren emphasized late last year. In recent history, high-pressure episodes have generally ended with the economy entering a recession; soft landings appear to be elusive.
That said, some have outlined potential labor market benefits to individual workers during high-pressure episodes—including higher labor force attachment, higher wages, and better job matches (see for example, here, here and here ). But could these types of labor market benefits persist and actually improve a worker's ability to also withstand an economic downturn?
To investigate this possibility, I ask the following question: Do high-pressure economies at the state level reduce the probability that a worker enters into unemployment during a subsequent downturn?
The details of my approach, using cross-sectional data from the monthly Current Population Survey, can be found in this appendix .
The following three charts illustrate the moderating impact a high-pressure economy can have on the probability of unemployment during a recession for various demographic groups. Chart 1 shows the impact on different age groups. The data tell us that the probability of unemployment for 18- to 34-year olds is 3.2 percentage points higher during recessions than during expansions, relative to how much higher the probability of unemployment is during recessions for 55- to 64-year olds (the excluded age group). This estimate is an average across all recessions between 1980 and 2015. Those who are 45- to 54-years old have only a modestly higher probability of unemployment (0.4 of a percentage point) during recessions than 55- to 64-year olds.
However, we also see from chart 1 that the effect of the recession on each age group is moderated by the state's high-pressure economy. Specifically, for each average percentage point by which the state's unemployment rate fell below the state's natural rate of unemployment prior to the recession, the probability of unemployment facing 18- to 34-year olds falls by 2.4 percentage points. Simply put, the hotter the state's prerecession economy, the lower the impact of the recession on workers' probability of unemployment.
We see the same impact across education groups in chart 2. Whereas those with some college face a probability of unemployment during a recession that is 0.7 percentage points higher than that of a college graduate, a prerecessionary high-pressure episode just 1 percentage point higher will wipe out the disadvantage that those with some college face during a recession relative to those with a college degree.
Chart 3 shows that black non-Hispanics experience even greater benefits from a high-pressure economy. A high-pressure period just 1 percentage point greater prior to a recession more than erases the average impact of the recession, relative to white non-Hispanics. (Note that these results are averaged across all recessions since 1980 and hence don't say anything about the labor market outcomes during any particular recession.)
The evidence I provide here suggests that a high-pressure economy may have some longer-term benefits in terms of improving labor market outcomes during economic downturns. If this is indeed the case, understanding how and why will be an important step in assessing the risk/reward calculus of high-pressure periods.
February 06, 2017
Examining Changes in Labor Force Participation
The Labor Department announced on Friday that January's unemployment rate was 4.8 percent, only 10 basis points below the level in January 2016. You can be forgiven if looking at a graph of the unemployment rate since 2007 makes you think of a roller coaster, because it showed a very steep climb, followed by a swift decline. From a distance, it may seem like the car's descent stopped about a year ago and has merely been bumping around a bit as it approaches the elevation of the platform.
But the unemployment rate alone does not fully account for improvement in the labor market. During the past three years, the labor force participation (LFP) rate has become a particularly important metric to look at. The overall share of the population that is working or actively seeking work has been essentially flat during this period, which is striking because there is a powerful demographic trend—an aging population—that is pulling it down with tremendous force.
Many factors are behind LFP's relative flatness, some of which undoubtedly relate to the labor market's strength. The opportunities available in the labor market affect an individual's decision to enter or leave the labor force. For example, it can affect when a person chooses to retire, enroll in college, apply for disability insurance, or stay home to care for family instead of looking for employment.
On a quarterly basis we update our web page with analysis of how these reasons for not being in the labor market have changed during the past year, and we also look at the extent to which these changes affect the overall LFP rate. Between the fourth quarter of 2015 and the same period in 2016, the LFP rate rose 0.14 percentage points (not seasonally adjusted). The chart below breaks out this increase and shows how much the various reasons for nonparticipation account for the increase (holding the age composition of the population fixed) versus the downward pressure exerted by an aging population.
Let's briefly look at the relative contributions to the change in labor force participation in more detail:
Aging of the population: During the last year, the aging population was the only significant factor continuing to depress the LFP rate. In line with this factor's contribution from previous years, it accounted for about 0.15 percentage points of the decline in the LFP rate.
Retirement: Retirement rates ticked down over the year, resuming a trend that had stalled in the past few years. Later retirement was the largest influence on LFP in the past year and completely offset the effect of aging population, boosting the rate by 0.15 points.
Shadow labor force: The share of the population not technically counted as "unemployed" because they are not actively searching but say they want a job fell slightly over the past year. This decline boosted the LFP rate by 0.04 percentage points. (A decline in this category is usually associated with a strengthening labor market.)
Health problems: The share of the population who said they are too chronically ill or disabled to work declined for the second year in a row, reversing the trend of the prior eight years. This decline put upward pressure on LFP (0.04 percentage points) and could partly be a reflection of a stronger job market with more opportunities for those with disabilities (see this report from the U.S. Bureau of Labor Statistics for more information).
Rising education: The share of the population not in the labor market because they are in school increased slightly, lowering the LFP rate by 0.03 percentage points. School enrollments rates rose for decades and accelerated during the last recession. The small contribution of schooling to the change in the LFP rate during the past year likely brings it closer to alignment with the long-term trend.
Family responsibilities: The share of the population not participating in the labor force because of family responsibilities declined during the last year, boosting the LFP rate by 0.13 percentage points.
An interactive chart on our website allows users to choose their own time period for comparison for all those 16 years old and above, those 25–54 years old, as well as for men and women separately. You can see how various factors have contributed to that roller coaster effect—strap yourself in!
January 23, 2017
Wage Growth Tracker: Every Which Way (and Up)
As measured by the Atlanta Fed's Wage Growth Tracker, the typical wage increase of a U.S. worker averaged 3.5 percent in 2016. This is up from 3.1 percent in 2015 and almost twice the low of 1.8 percent recorded in 2010. As noted in previous macroblog posts, the Wage Growth Tracker correlates tightly to the unemployment rate. As median wage growth has risen, the unemployment rate declined from an average of 9.6 percent in 2010, to 5.3 percent in 2015, and to 4.8 percent in 2016.
What does this correlation suggest about the Wage Growth Tracker in 2017? Let's start with a forecast of unemployment. Based on the latest Summary of Economic Projections, the central view of Federal Open Market Committee participants is that the unemployment rate will end this year at around 4.5 percent, about 30 basis points below the median participant's estimate of the unemployment rate that is sustainable over the longer run.
With a modest further decline in the unemployment rate, other things equal, we might then also expect to see a modest uptick in the Wage Growth Tracker in 2017. But I think the emphasis here should be on the word modest. Speaking for myself, sustained Wage Growth Tracker readings much above 4 percent in 2017 would begin to worry me, especially without a compensating pickup in the growth of labor productivity, which has been stuck in the 0 to 1 percent range in recent years. Significantly higher wage growth—reflecting a tightening labor market more than larger gains in worker productivity—could make the inflation outlook a bit less sanguine than we currently think. (This macroblog post discussed the connection among productivity growth, wage growth, and inflation.)
Thus far, many firms appear to have been able to keep their labor costs relatively low by replacing or expanding staff with lower-paid workers. (Our colleagues at the San Francisco Fed have written about how changes in the composition of workers can mute changes in total labor costs.) However, it's not clear how long that approach can be sustained. Indeed, it's noteworthy that average wage costs appear to have accelerated recently. For instance, U.S. Bureau of Labor Statistics data indicate that average hourly earnings in the private sector increased over the year by 2.9 percent in December—the fastest pace since 2009.
We haven't been hearing reports from firms where the typical worker's wage increase in 2017 is expected to be above 4 percent. However, we did get readings for the Wage Growth Tracker pretty close to 4 percent in October and November of last year. As the following chart shows, a sharp increase in women's median wage growth (hitting 4.3 percent in October 2016) drove the overall increase. In contrast, the median wage increase for men was 3.5 percent.
The jump in the relative wage growth of women came as a bit of a surprise. Female wage growth had been generally running below that of men since 2010, and analysis by my colleague Ellie Terry showed that gender-specific factors that are unlikely to change very rapidly explain a fair amount of that lag. Therefore, we suspected that the divergence in wage growth might not be sustainable—a suspicion that proved to be true. Median wage growth for women slowed to 3.5 percent in December, the same growth rate men saw.
Readers who can't get enough Wage Growth Tracker data will be delighted to note that in 2017 we plan on making further enhancements to the tool. These enhancements will include finer cuts by age, education, industry, and hours worked, as well as new cuts by occupation, race, and location. You can stay informed on all Wage Growth Tracker updates by subscribing to our RSS feed or email updates .
November 28, 2016
Does Lower Pay Mean Smaller Raises?
I've been asked a few questions about the relative wage growth of low-wage versus high-wage individuals that are measured by the Atlanta Fed's Wage Growth Tracker. Do individuals who were relatively lower (or higher) paid also tend to experience lower (or higher) wage growth? If they do, then wage inequality would increase pretty rapidly as low-wage earners get left further and further behind.
The short answer is no. As chart 1 shows, median wage growth is highest for the workers whose pay was relatively low (in the bottom 25 percent of the wage distribution), and lowest for those who were the highest-paid (in the top 25 percent of the wage distribution). Median wage growth is reasonably similar for those whose pay was in the middle 50 percent of the wage distribution.
To understand what's going on, let's look at the construction of a Wage Growth Tracker sample. In simple terms, a person's wage is observed in one month, and then again 12 months later. But relatively low-wage workers are less likely to remain employed (and hence more likely not to have a wage when observed a second time) than other workers. Almost half of workers who are not employed 12 months later come from the lowest 25 percent of the wage distribution. For workers in a relatively low-wage job, a greater share who might otherwise have experienced a declining wage left their employment, resulting in a larger share of wage increases among those who remained employed.
In contrast, relatively high wage earners in the Wage Growth Tracker sample have a remarkably low median wage growth—zero in recent years. They also have a much greater chance of experiencing a wage decline than other workers (see chart 2).
However, getting a complete picture for high-wage individuals in the Current Population Survey is limited by the fact that observations are top-coded (or censored to preserve identifiable individuals' anonymity). For example, weekly earnings higher than $2,885 are currently simply recorded as $2,885. If a person in this circumstance gets a wage increase, it will still be reported as just $2,885, which would make it seem as if wages didn't increase, even if they did.
Top-coding itself has only a relatively small effect on the median wage growth for the whole sample because top-coded earnings aren't that common. But they are a reasonably large share of the upper part of the wage distribution, which makes the median wage growth pretty unrepresentative for people who were relatively high wage earners. In principle, one could try to surmount this problem by estimating the earnings for top-coded workers, but my experience has been that doing so is likely to add more noise than insight.
What about examining a worker's current wage instead of their prior wage? Is the median wage growth also higher for workers who are currently in the lowest part of the wage distribution? No. In fact, they are more likely than others not to have received a pay raise or even to have had the rate of pay reduced. Conversely, someone who is currently in the upper part of the wage distribution is more likely to have received a larger pay raise than other workers. Some workers move up the wage distribution—but not all.
The bottom line is that the point of reference matters a lot when looking at the tails of the wage distribution, and top-coding limits the ability to learn much about the wage growth of high wage earners. But for the middle part of the wage distribution, it doesn't matter so much. The median wage growth of the overall sample is pretty representative of the typical wage growth experience of workers in the heart the wage distribution.
- 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
- Gauging Firm Optimism in a Time of Transition
- Can Tight Labor Markets Inhibit Investment Growth?
- More Ways to Watch Wages
- May 2017
- April 2017
- March 2017
- February 2017
- January 2017
- December 2016
- November 2016
- October 2016
- September 2016
- August 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