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
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 07, 2017
Net Exports Continue to Bedevil GDPNow
Real gross domestic product (GDP) grew at an annualized rate of 1.9 percent in the fourth quarter, according to the advance estimate from the U.S. Bureau of Economic Analysis (BEA), 1.0 percentage point below the Atlanta Fed's final GDPNow model projection. This was a sizable miss relative to other forecasts. Both the consensus estimate from the January Wall Street Journal Economic Forecasting Survey and the January 20 staff nowcast from the New York Fed were expecting 2.1 percent growth last quarter.
The miss was also large relative to the historical accuracy of the GDPNow model. As the table below shows, almost all of GDPNow's error for fourth quarter growth was concentrated in real net exports. For the other broad subcomponents, GDPNow was more accurate than usual, as the last two columns of the table show. But net exports subtracted 1.70 percentage points from real GDP growth last quarter, whereas GDPNow forecasted they would only reduce growth by 0.64 percentage points. All but 0.02 percentage points of this error was in the "goods" category as opposed to services.
Three months ago, I wrote a macroblog post showing that nearly all of GDPNow's 0.8 percentage point error for third-quarter growth was concentrated in goods net exports. That analysis explained how GDPNow's goods net exports forecast is a weighted average of two forecasts. One of these forecasts is a "bean counting" model that uses monthly source data on nominal values and price deflators for goods imports and exports. The other is a quarterly econometric model that uses subcomponents of real GDP for prior quarters. In the GDPNow model, the "bean counting" model gets nearly 60 percent of the weight just before the advance GDP release.
To see how this approach matters for the GDP forecast, the following chart shows the "real-time" forecasts of the contribution of goods net exports to growth just before BEA's advance GDP estimate from the two models alongside the advance estimate of the contribution and the final GDPNow forecast.
We see that the "bean counting" forecast has been much more accurate than the quarterly econometric forecast, particularly for the last two quarters of 2016. Not surprisingly given its name, the "bean counting" model was able to largely capture the 0.75 percentage points that soybean exports contributed to third-quarter real GDP growth and the just over 0.5 percentage points they likely subtracted from fourth-quarter growth. The econometric model was not.
The final forecasts of goods net exports from the "bean counting" model have also been more accurate than GDPNow since forecasts were first posted online in mid-2014. Does this imply that an alternative "bean counting" version of GDPNow would be preferable? The answer is less obvious than you might think. Not putting any weight on the quarterly econometric model for any GDP subcomponents yields an average error for GDP growth (without regard to sign) of 0.635 percentage points, and the same statistic for GDPNow is 0.589 percentage points. This is despite the fact that the "bean counting" approach has been more accurate than GDPNow in its forecasts of net exports and about as accurate, on balance, for the other GDP subcomponents.
The final forecast of real GDP growth last quarter of this alternative "bean counting" model was 2.8 percent—only slightly more accurate than GDPNow. (For each GDP subcomponent, I include the "bean counting" and quarterly econometric model forecasts in this excel spreadsheet.)
However, if variants like the aforementioned "bean counting" approach continue to outperform the GDPNow model in one or more dimensions, we may consider regularly reporting their forecasts along with the GDPNow forecast.
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!
- What Does the Current Slope of the Yield Curve Tell Us?
- Does Loyalty Pay Off?
- Immigration and Hispanics' Educational Attainment
- Are Tariff Worries Cutting into Business Investment?
- Improving Labor Market Fortunes for Workers with the Least Schooling
- Part-Time Workers Are Less Likely to Get a Pay Raise
- Learning about an ML-Driven Economy
- Hitting a Cyclical High: The Wage Growth Premium from Changing Jobs
- Thoughts on a Long-Run Monetary Policy Framework, Part 4: Flexible Price-Level Targeting in the Big Picture
- Thoughts on a Long-Run Monetary Policy Framework, Part 3: An Example of Flexible Price-Level Targeting
- August 2018
- July 2018
- June 2018
- May 2018
- April 2018
- March 2018
- February 2018
- January 2018
- November 2017
- October 2017
- 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