This page had been redirected to a new URL, please update any bookmarks.

Font Size: A A A


May 16, 2013

Labor Costs, Inflation Expectations, and the Affordable Care Act: What Businesses Are Telling Us

The Atlanta Fed’s May survey of businesses showed little overall concern about near-term inflation. Year-ahead unit cost expectations averaged 2 percent, down a tenth from April and on par with business inflation expectations at this time last year.

OK, we’re going to guess this observation doesn’t exactly knock you off your chair. But here’s something we’ve been keeping an eye on that you might find interesting. When we ask firms about what role, if any, labor costs are likely to play in their prices over the next 12 months, an increasing proportion have been telling us they see a potential for upward price pressure coming from labor costs (see the chart).

To investigate further, we posed a special question to our Business Inflation Expectations (BIE) panel regarding their expectations for compensation growth over the next 12 months: “Projecting ahead over the next 12 months, by roughly what percentage do you expect your firm’s average compensation per worker (including benefits) to change?”

We got a pretty large range of responses, but on average, firms told us they expect average compensation growth—including benefits—of 2.8 percent. That’s about a percent higher than the average over the past year (as estimated by either the index of compensation per hour or the employment cost index). But a 2.8 percent rise is also about a percentage point below average compensation growth before the recession. We’re included to read the survey as a confirmation that labor markets are improving and expected to improve further over the coming year. But we’re not inclined to interpret the survey data as an indication that the labor market is nearing full employment.

We’ve also been hearing more lately about the potential for the Affordable Care Act (ACA) to have a significant influence on labor costs and, presumably, to provide some upward price pressure. Indeed, several of our panelists commented on their concern about the influence of the ACA when they completed their May BIE survey. So can we tie any of this expected compensation growth to the ACA, a significant share of which is scheduled to go into effect eight months from now?

Because a disproportionate impact from the ACA will fall on firms that employ 50 or more workers, we separated our panel into firms with 50 or more employees, and those employing fewer than 50 workers. What we see is that average expected compensation growth is the same for the bigger employers and smaller employers. Moreover, the big firms in our sample report the same inflation expectation as the smaller firms.

But the data reveal that the bigger firms are a little more uncertain about their unit cost projections for the year ahead. OK, it’s not a big difference, but it is statistically significant. So while their cost and compensation expectations are not yet being affected by the prospect of the ACA, the act might be influencing their uncertainty about those potential costs.

Photo of Mike BryanBy Mike Bryan, vice president and senior economist,

Photo of Brent MeyerBrent Meyer, economist, and

Photo of Nicholas ParkerNicholas Parker, senior economic research analyst, all in the Atlanta Fed’s research department

May 16, 2013 in Business Inflation Expectations, Economics, Health Care, Inflation Expectations, Labor Markets, Pricing | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Labor Costs, Inflation Expectations, and the Affordable Care Act: What Businesses Are Telling Us:


Maybe we're finally reaching the point where firms can no longer expropriate productivity gains. If you look at the total hourly compensation for non-supervisory workers vs. productivity, the last 40 years have more or less seen the gains made during the Great Compression utterly obliterated. Now that we're back to Gilded-Age levels of income distribution, it may be that we've reached an equilibrium.

Posted by: Valerie Keefe | May 19, 2013 at 12:22 PM

Post a comment

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

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

May 13, 2013

Labor Force Participation and the Unemployment Threshold

On Friday, my colleague Julie Hotchkiss shared in this space the results of her new research (with Fernando Rios-Avila, a Georgia State University colleague) on the recent and prospective behavior of the labor force participation rate (LFPR). The punch line, from my point of view, is this:

Our results suggest that relative to the average LFPR over the years 2010–12, the average LFPR over the years 2015–17 will rise by about a third of a percentage point—again, if the labor market returns to prerecession conditions. (Italics original)

As Julie notes:

[T]he Federal Open Market Committee has substantially raised the stakes on disentangling...movements in labor force introducing into its policy deliberations concepts like unemployment thresholds and qualitative assessments on "substantial" labor market improvement.

Though the meaning of "substantial labor market improvement"—a condition for adjusting the FOMC's current large-scale asset purchase program—is somewhat ambiguous, the unemployment threshold for considering moving the federal funds rate off the near-zero mark is less so. As the Committee indicated in its May press release:

[T]he Committee decided to keep the target range for the federal funds rate at 0 to ¼ percent and currently anticipates that this exceptionally low range for the federal funds rate will be appropriate at least as long as the unemployment rate remains above 6½ percent, inflation between one and two years ahead is projected to be no more than a half percentage point above the Committee's 2 percent longer-run goal, and longer-term inflation expectations continue to be well anchored.

It is widely understood (a sign of the times, no doubt) that changes in the unemployment rate are not entirely independent of what is happening with the participation rate. We have discussed this issue before here in macroblog. But in light of the new research coming from our own shop (and other research cited in Julie's post), it seems like a good time for a refresher.

First, a step back. Multiple upward revisions to the employment situation since the December jobs report—you can follow the trail courtesy of Calculated Risk here, here, here, here, and here—have led to a more robust picture of the labor market than certainly I was thinking. Here is what the record looks like for most of the recovery:

With an assist from the Atlanta Fed Jobs Calculator, we can provide further perspective on these numbers. In particular, under the assumption that the labor force participation rate will remain at its current level of 63.3 percent (among other things held constant), we can map the recent job growth numbers to a rough date when the unemployment rate will reach 6½ percent.

That looks interesting, but then taking the Hotchkiss and Rios-Avila research onboard means the assumption of a constant labor force participation rate may not be justified. So, turning again to the Jobs Calculator, the following table answers this question: If we continue on the 208,000-per-month pace of job creation of the last six months, and the labor force participation rate is X, what would the unemployment rate be by June of next year? For reference, the first row of the table replicates the earlier result under the assumption that the participation rate will maintain its current level; the second row takes into account the Hotchkiss and Rios-Avila research; and the third assumes an even larger bounce back in participation:

It is probably worth noting that the full increase in the Hotchkiss and Rios-Avila estimates happens in the 2015–17 timeframe, raising the interesting possibility that the threshold for considering interest rate increases could occur sometime before the unemployment rate moves back above the threshold.

Also, it is not at all obvious that rising labor force participation would necessarily arrive along with a rising unemployment rate. From 1996 through 1999, for example, the participation rate rose by nearly by 0.7 percentage point (the difference between the rates in the first and third rows in the table above), even as the unemployment rate fell by just over 1½ percentage points. The key was the strong employment growth over that period—almost 260,000 payroll jobs per month on average.

All of that, as should be clear by now, embeds a whole bunch of assumptions, which may make this the most important part of the FOMC's decision criteria:

In determining how long to maintain a highly accommodative stance of monetary policy, the Committee will also consider other information, including additional measures of labor market conditions...

Photo of Dave AltigBy Dave Altig, executive vice president and research director of the Atlanta Fed

May 13, 2013 in Employment, Labor Markets, Monetary Policy | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Labor Force Participation and the Unemployment Threshold:


Post a comment

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

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

May 10, 2013

Behavior’s Place in the Labor Force Participation Rate Debate

It's not often that the mainstream media is interested in the nuances of labor market statistics, so last week’s debate over the meaning of labor force participation rates (LFPR) in the pages of the Washington Post and the Wall Street Journal was music to this labor economist's ears.

Sparked by an article by Ben Casselman in his April 29 Wall Street Journal Outlook column, the ensuing back and forth (here, here, here, and here) between Casselman and the Post’s Jim Tankersley focused on what has become a central preoccupation in assessing the likely course of the labor market: Is the recent decline in the labor force participation rate the result of structural factors (e.g., an aging population) or cyclical ones (such as weak economic conditions)? Almost contemporaneously, Bill McBride declared in his recent Calculated Risk blog, "…most of the [recent] decline in the participation rate was due to changing opposed to economic weakness."

The changing pattern of labor force participation has been a topic of discussion among economists for some time—for example, see my Federal Reserve Bank of Atlanta Economic Review article—and both Tankersley and Casselman agree that the long-run secular decline in participation is a matter worthy of independent concern. But the Federal Open Market Committee has substantially raised the stakes on disentangling longer-run trends from short-run cyclical (and presumably temporary) movements in labor force participation. It’s done this by introducing into its policy deliberations concepts like unemployment thresholds and qualitative assessments on “substantial” labor market improvement.

Casselman, in an October 2012 WSJ article, cites work by my colleagues at the Chicago Fed, who find that while more than two-thirds of the decline in LFPR between 1999 and 2011 is accounted for by changes in the age distribution of the population, "…over the 2008-2011 period...only one-quarter of the...decline of actual LFPR...can be attributed to demographic factors."

This conclusion—that three-quarters of the decline in the LFPR since the beginning of the Great Recession can be attributed to cyclical factors—is supported by other research. Colleagues at the Kansas City Fed and at the Board of Governors concur that the vast majority of the decline in the LFPR since 2008 is the result of cyclical factors. Even economists outside the Federal Reserve System acknowledge the significant role of cyclical factors in the LFPR decline (for example, see the analysis by economists at the Deutsche Bank).

But there is a critical third piece to the LFPR puzzle that most of these studies ignore. In addition to changing demographics (which have, for example, been associated with a rising share of retirement-age individuals in the total population) and cyclical effects (for example, the tendency for participation to fall when wage growth is tepid or job opportunities scarce), there are also behavioral changes afoot—a point Casselman makes in his final installment of the Post/WSJ debate. For example, individuals of near-retirement age may extend their participation as a result of significant, unexpected declines in wealth. Or women with young children—a demographic group typically less likely to participate in the labor market—may increase participation if a partner loses a job during an economic downturn. In both cases, participation rates for these demographic groups would not fall by as much as expected in response to high unemployment rates alone.

Work that I've done with Fernando Rios-Avila, a colleague at Georgia State University, finds that more than 100 percent of the fall in the LFPR since 2008 is accounted for by the condition of the labor market (cyclical factors), but these particularly strong cyclical forces were countered by increased tendencies to participate (behavioral changes). In other words, if individuals hadn't stepped up to the plate and exhibited even stronger labor force participation behavior than before the recession, the LFPR would be even lower than it is.

To illustrate the role that changing behavior played in the LFPR decline during the Great Recession, the chart below illustrates how this decline can be separated into a trend component (demographics), a cyclical component (strength of the labor market), and a behavioral component. The solid black line reflects the actual LFPR in March of each year calculated using the Current Population Survey, which is the survey data used by the U.S. Bureau of Labor Statistics to calculate the monthly labor force statistics. The orange line reflects the trend estimate of the LFPR using only demographic data (such as the age distribution of the population) through 2007, projecting out to 2012. As many others have pointed out, changing demographics—the aging of the baby-boom generations, if you will—explains only about 30 percent (in this example) of the actual post-2007 decline in LFPR.

But the chart also reveals something that may be underappreciated. Including a measure of labor market conditions in the projection of the LFPR, as well as a depiction of prerecession behavior (the green line), indicates that the LFPR should be much lower than it actually is. The message from this exercise is that the actual LFPR in 2012 was above what would have been projected had each demographic group exhibited the same labor force participation behavior after the recession as before the recession.

As it turns out, women, ethnic minorities, older people, and individuals with small children were much more likely to participate in the labor market after the recession than before it. These workers are often referred to as "added workers," or workers who join the labor force to make up for lost income elsewhere in the household. As I noted above, if these demographic groups had not increased their participation in the labor force, the aggregate LFPR would be much lower than it is.

What the chart tells us is that the cyclical factors affecting labor force participation are even more important than generally imagined. However, it is also true that the inevitable march of time will continue to put powerful downward pressure on labor force participation. Indeed, our research predicts only a modest rise in the LFPR if labor markets rebound to prerecession conditions. Our results suggest that relative to the the average LFPR over the years 2010–12, the average LFPR over the years 2015–17 will rise by about a third of a percentage point—again, if the labor market returns to prerecession conditions. Though higher than today, this level would still leave the LFPR considerably lower than it was before the recession, primarily reflecting the continued downward pressures of aging baby boomers.

Photo of Julie HotchkissBy Julie Hotchkiss, a research economist and policy adviser in the Atlanta Fed’s research department

May 10, 2013 in Economics, Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Behavior’s Place in the Labor Force Participation Rate Debate:


Your analsysis is right. In 2008, we published in the Journal of Applied Economic Sciences a paper that predicted the LFPR fall in 2010 ( We based our prediction on demography. Currently, the same model shows that the LFPR should return to 64% in 2013-2014 ( .

Posted by: kio | May 14, 2013 at 04:41 AM

Post a comment

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

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

April 05, 2013

Labor Market Update: Muddy Waters Continue to Run Deep

Earlier this week, Atlanta Fed President Dennis Lockhart gave a speech in Birmingham, Alabama, focused on labor markets, risks to the outlook, and current monetary policy. One of the things President Lockhart noted was that the picture for the labor market remained muddy. Specifically:

"The fact is that conditions in the broad labor market are quite mixed. While some indicators of labor market health have improved a lot since the recession, others have not improved much at all or have even worsened. As I said, net job creation is picking up. Initial claims for unemployment insurance have fallen. But the official rate of unemployment remains high, many discouraged workers have left the labor force, and there are many people working part-time jobs who want to work full time."

Today's labor report did little to clarify improvement in labor market conditions, with March payrolls estimated to have grown by a much less than expected 88,000 workers and the jobless rate falling one-tenth of a percent, to 7.6 percent, on the back of a decline of 496,000 in the size in the labor force. Updated with today's data, below is the spider chart we have previously offered as one way to simultaneously track and visualize "conditions in the broad labor market."

As a reminder, we've taken the approach of dividing a set of 13 indicators of labor market conditions into four segments:

  • Employer Behavior includes indicators related to the hiring activities of employers.
  • Confidence includes indicators of employer and worker confidence in the labor market.
  • Utilization includes measures related to available labor resources.
  • Leading Indicators shows data that typically provide insight into the future direction of overall labor market activity.

The circle at the perimeter of this chart represents labor market conditions that existed just before the recession. We have dated this as late 2007. The inner circle represents the state of affairs when payroll employment reached its trough in late 2009. The oddly shaped red figure inside the perimeter depicts where each of the indicators was in March 2011 relative to the benchmarks. The purple figure depicts the state of the labor market in March 2012. Finally, the blue figure shows where the indicators were as of March 2013. All of the indicators are scaled so that outward movement represents improvement. The progression of these point-in-time snapshots provides us with a picture of how labor market conditions have evolved over the past four years.

As you can see, substantial improvement has arguably been achieved in the leading indicator series. As a group, these data points are approaching their prerecession levels. Employer hiring behavior and confidence are slowly moving outward but remain quite weak relative to their prerecession benchmarks. Finally, the labor utilization measures are very weak and, notably, have hardly improved at all over the past two years.

In the macroblog post that introduced the spider chart, we noted that there are a couple of immediate issues that arise in using this graphic to interpret market improvement, substantial or otherwise:

First, a variable such as the level of payroll employment will eventually exceed its pre-recession level, and grow consistently over time as the population grows. A variable like "hiring plans"—which is the net percentage of firms in the National Federation of Independent Business survey expecting to hire employees in the next three months—cannot grow without bound....

Second, it is not obvious that 2007:IVQ levels are necessarily the best benchmarks for all (or even any) of the variables we are monitoring [in the spider chart].

Of these two issues, the second one is potentially the more problematic. The spider chart invites you to think of the inner circle as the starting point and the outer circle as the "goal," or a representation of "normal" labor market conditions. In assessing improvement in variables such as payroll employment, using the prerecession level as a reference point makes some sense as a minimal standard. But for some, "minimal" may be the operative word. If we are interested in questions of how employment is doing relative to some concept of "full employment," it might be appropriate to assess the data relative to some measure of trend. For example, payroll employment is still about 3 million below the prerecession peak, but the labor force, despite the drop in March, is by about 1 million higher than before the recession. When measured relative to the size of the labor market, progress on employment is less impressive than it would appear by just looking at growth in employment itself.

One way to address both of the caveats noted above is to scale variables that involve numbers of jobs or people—such as payroll employment or the number of unemployed—by the size of the population or the labor force. Doing so ensures that variables measured in numbers of jobs or people do not grow without bound. It also helps in assessing progress in these variables relative to a (back-of-the-envelope) measure of the trend in labor resources.

We take this approach in the following chart, which reproduces the spider chart but divides the variables that are counts of people by the size of the labor force. (The labor force has grown more slowly than the population since the end of the recession, but a generally similar picture emerges if the variables are instead deflated by the population.)

Not surprisingly, for most of the indicators, labor market progress is a bit more subdued relative to postrecession growth in the labor force than growth in the indicators alone would suggest. These adjustments definitely would not alter our view that the labor market picture is "quite mixed." President Lockhart's recent comments on CNBC—in which he said he would like to see more positive data before declaring that sustained improvement has taken hold—seem especially prescient in light of today's job numbers:

Photo of John RobertsonBy John Robertson, vice president and senior economist in the Atlanta Fed's research department, and


Photo of Dave AltigDave Altig, executive vice president and research director of the Atlanta Fed


April 5, 2013 in Employment, Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Labor Market Update: Muddy Waters Continue to Run Deep:


Post a comment

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

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

March 28, 2013

The Same-Old, Same-Old Labor Market

In his March 24 Wall Street Journal piece on declining government payrolls, Sudeep Reddy offers up a key observation:

The cuts in the public-sector workforce—at the federal, state and local levels—marked the deepest retrenchment in government employment of civilians since just after World War II... down by about 740,000 jobs since the recession ended in June 2009. At the same time, the private sector has added more than 5.2 million jobs over the course of the recovery.

As the Journal article notes, the story of shrinking government employment combining with private-sector payroll expansion has been remarkably consistent for much of the recovery.

About a year ago, we provided a graphical illustration of postrecession employment patterns using payroll-employment "bubble charts." These charts measure postrecession average monthly employment changes by sector relative to the changes in the prerecession period from December 2001 through October 2007. Not a lot in that chart has changed over the intervening year (just as not a lot had changed in 2012 compared with 2011):

The stability in the employment picture across private industries, both relative to one another and relative to the precrisis pace of job gains, is just as notable as the changing fortunes of private versus public employment. In fact, the charts offer some pretty clear impressions:

  • Virtually all private-sector industries have moved into positive employment-growth territory. That movement includes the construction and financial activities sectors, which have generally lagged the improvement in the rest of the economy. The only broad category still shedding jobs in the private sector has been the information industry—which includes publishing, motion picture production, telecommunications, data processing, and the like—an industry that was also shrinking in terms of employment over the decade leading up to the financial crisis.
  • All of the private-sector bubbles in the charts are now close to, on, or above the 45-degree line, meaning that the average pace of monthly job creation in each sector is near, equal to, or greater than what prevailed during the last recovery.
  • As the Reddy piece emphasizes, government employment has been in decline since early 2010, though the government sector as a whole retains its status as the sector with the largest employment share. (The size of the bubbles in the charts above represents the share of employment in each sector at the end of the period for which the graph is drawn.) But the chart also illustrates another key point of Reddy's article: To date, the decline in government employment has been concentrated in state and, especially, local government jobs. Until recently, job creation by the federal government, which is relatively small in the bigger scheme of things, has not deviated much from its prerecession pattern.

The last point brings us to this observation, from the WSJ article:

How the rest of the private sector responds to a shrinking of the federal government could play a bigger role in determining how the budget fight hits the workforce.
"The private sector in the U.S. is growing so much stronger than anyone had expected," said Bernard Baumohl of the Economic Outlook Group. "This organic growth is going to significantly offset the effect of the sequester in terms of economic output and employment."

It is worth pointing out that the monthly average of 17,000 state, local, and federal government jobs lost since March 2010 has been nearly matched by average monthly increases of better than 14,000 jobs in manufacturing, a sector that persistently shed jobs in the previous recovery. The replacement of private- for public-sector employment has generated 175,000 to 185,000 net jobs per month in both 2011 and 2012. To put one perspective on that figure, at current labor force participation rates (along with some other assumptions and caveats), that pace would be sufficient to reach the Federal Open Market Committee's 6.5 percent unemployment threshold by sometime in spring 2015 (as you can verify yourself with the Atlanta Fed's Jobs Calculator). That calculation raises the stake somewhat on the matter of how "the rest of the private sector responds."

Photo of Dave AltigBy Dave Altig, executive vice president and research director of the Atlanta Fed


March 28, 2013 in Employment, Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference The Same-Old, Same-Old Labor Market:


I'm a little confused about the sectorial changes in government graphic....

Posted by: jay | March 30, 2013 at 01:44 AM

A little bit confusing... not?

Posted by: Economizar | April 01, 2013 at 12:51 PM

Not shure about the sectorial changes in government graphic....

Posted by: Poupar | January 29, 2014 at 05:02 AM

Post a comment

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

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

March 19, 2013

Being Ahead of the Curve: Not Always a Good Thing

Our friends at the New York Fed have a nifty interactive graphic that compares the unemployment rate, labor force participation rate, and employment-to-population ratio over the last five business cycles. You can even break these indicators down by gender, by age, or by a particular business cycle. (For a deeper dive, check out this post at Liberty Street Economics by Jonathan McCarthy and Simon Potter.) And though it’s not exactly late-breaking news, no matter which of the three indicators you look at, you can’t help but conclude that the most recent recession is an outlier.

The Beveridge curve is a fourth and particularly useful graphical representation of a steady-state economy showing how, in theory, one might expect the vacancy rate to change, given an unemployment rate. It depicts the relationship between job openings and the unemployment rate. (The Atlanta Fed’s magazine, EconSouth, discussed the Beveridge curve.) It, too, has been standing out over the course of the most recent recovery, so much so that we think it warrants at least a second glance. There are a number of ways to estimate a Beveridge curve (see, for example, methods described by Gadi Barlevy of the Chicago Fed here and by the Richmond Fed’s Thomas Lubik here).

We use the method described by Barnichon et al. (2012) to estimate the solid curve used in the first chart below. The square plots represent actual vacancy rate (y-axis) and unemployment rate (x-axis) combinations by month from December 2000, when the Job Openings and Labor Turnover Statistics (JOLTS) data series from the U.S. Bureau of Labor Statistics (BLS) series begins, to January 2013, the most recent month of data available for both series.

Blue squares represent data from December 2000 to December 2009, when the “errors” between actual plots and the curve estimation were below 2 percentage points, and red squares represent data since January 2010, where data suddenly seem to jump higher than the predicted Beveridge curve to the tune of 2 percentage points or greater (see the chart below).

But why?

In June 2012, Regis Barnichon and his coauthors concluded that the unemployment rate’s lackluster performance so far in the recovery was attributable to a shortfall in hires per vacancy. Since then, the vacancy rate has climbed its way back to its June 2008 level of 2.7 percent. However, the unemployment rate has clearly not returned to either its June 2008 level (5.6 percent) or where the Beveridge curve says it should be given this vacancy rate, which one might predict to be 5.5 percent using the methodology of Barnichon et al.

This “ahead of the curve” phenomenon has not gone unnoticed and has prompted some explanations. In a March 6, 2013, article in The New York Times (which also has some cool charts), Catherine Rampell posits that available positions are staying unfilled longer, while interview processes have become lengthier.

The next day, Rampell went into more detail about why we’re going “off the curve” in a New York Times Economix post. She cites skills mismatch and a skills atrophy effect of the long-term unemployed affecting the ability of employers to fill positions (which she explains aren’t full explanations, yet we would expect to see wages for highly coveted positions rise significantly).

Rampell goes back to the explanation many of us continue to hear from business contacts: employers are unwilling to fill vacant positions because of economic and fiscal policy uncertainty. She quotes Stephen Davis of Chicago’s Booth School: “They’re taking longer to fill vacancies because they just feel less need to fill jobs now,” Davis said. “They recognize that in a slack labor market, there is an abundance of viable candidates. If something happens, and if they need to hire quickly, they know they can do that. That’s harder in a tight labor market.”

So maybe as labor markets “tighten up,” or perhaps if the speed by which they tighten up quickens, we’ll get back on the Beveridge curve. Only time, and several BLS releases, will tell.

Photo of Patrick HigginsBy Patrick Higgins, an economist at the Atlanta Fed, and

Photo of Mark CarterMark Carter, a senior economic analyst at the Atlanta Fed

March 19, 2013 in Economics, Employment, Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Being Ahead of the Curve: Not Always a Good Thing:


This post de-emphasizes Rampell's point that vacancies remain open longer because of economic uncertainty and growth expectations. Instead, it stresses the effect of the labor market supply.

Firms' hiring decisions likely are impacted by the number of qualified candidates available. However, most firms probably also hire because of the amount of work on hand, and the amount of work anticipated. The latter point is obscured by the quote by Stephen Davis, which is unfortunate, since it is probably quite important. In fact, macroblog referenced this relationship in a previous post, but unfortunately, decided to ignore it here.

Posted by: Carl | March 21, 2013 at 10:49 AM

Post a comment

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

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

February 15, 2013

Promoting Job Creation: Don't Forget the Old Guys

In a provocative article posted this week, the American Enterprise Institute's James Pethokoukis concludes that the state of entrepreneurship in the United States is, disturbingly, weaker than ever. In particular, Pethokoukis documents a decline in jobs created by establishments less than one year old, a trend that began before the 2001 recession and has continued more or less unabated since. He specifically cites the following symptoms of trouble:

  1. Had small business come out of the recession maintaining just the rate of start-ups generated in 2007, according to McKinsey, the U.S. economy would today have almost 2.5 million more jobs than it does.

  2. There were fewer new firms formed in 2010 and 2011 than during the Great Recession.

  3. The rate of start-up jobs during 2010 and 2011—years that were technically in full recovery—were the lowest on record, according to economist Tim Kane of the Hudson Institute.

That last point appears to be all the more ominous given this observation from Tim Kane:

"...that startups create essentially all net new jobs. Existing employers, it turns out, tend to be net job losers, averaging net losses of 1 million workers per year."

Pethokoukis makes his case with political commentary that we don't endorse and don't find particularly helpful. But we won't argue with his conclusion that more entrepreneurial start-up activity would be a good thing. Nonetheless, we get a little concerned when the conversation jumps from data on net job creation and the role of start-ups and early life-cycle firms, and moves on to policy conclusions that seem to disproportionately focus on that class of businesses specifically.

Here is the source of our concern: Though it is also tempting to lump all "existing employers” into the basket of net job destroyers, there are existing firms that create jobs, and a few are doing so on a very large scale.

Take 2006, for instance. Based on data from the Commerce Department called Business Dynamics Statistics (BDS), new firms (businesses with a payroll that existed in March 2006 but not in March 2005) had about 3.5 million employees. This is the large net job creation by new firms reported by Kane. However, over the same year, expanding firms more than 10 years old added a whopping 11 million jobs—about three times as many jobs as created by new firms. Of course, some older firms were downsizing or closing—contracting mature firms destroyed an estimated 10 million jobs. So the net number of jobs created by older established firms looks somewhat less impressive than the record of those young start-ups. But in the overall picture, were the 11 million jobs created by the expanding older businesses really less important than 3.5 million created by the newbies?

It turns out that older firms also account for a large fraction of the job creation occurring in fast-growing firms, arguably a better characterization of entrepreneurism than newness. We found some compelling evidence reported in recent research by Akbar Sadeghi, James Spletzer, and David Talan. Using data from the U.S. Bureau of Labor Statistics' Business Employment Dynamics (BED), Sadeghi, Spletzer, and Talen find that older firms (those at least 10 years old) accounted for more than 40 percent of the employment created by high-growth firms (those with at least 20 percent annual employment gains between 2008 and 2011). A similar conclusion about the role of older, fast-growing firms is found in this earlier Kauffman Foundation report based on BDS data looking at the 1 percent of fastest-growing firms in the United States.

The point is not that start-up entrepreneurial activity is unimportant. It is vitally important. But in larger terms, we should recognize that all entrepreneurial activity is important, no matter what the age of the firm in which it occurs. Atlanta Fed President Dennis Lockhart highlighted this point in a speech delivered earlier this week at Instituto de Empresas in Madrid, Spain:

My bank's experience in trying to understand the role of small businesses, small-growth businesses, young businesses, and mature-growth businesses in job creation illustrates a key point, I think. In the pursuit of economic growth and increased employment, there is no silver bullet. Rather, the policy community should be pursuing an effective mix of policy elements (with focus in areas such as new business formation, labor rules, and regulatory efficiency, to name a few) that together catalyze a virtuous circle of innovation, growth, and employment.

Certainly, entrepreneurial risk-taking, whether by large, mature businesses or start-ups aimed at becoming growth companies, is part of the solution.

When it comes to promoting job creation, forgetting to throw mature businesses into the mix with start-ups is surely not the path to finding the best policy solutions.

Dave AltigBy Dave Altig, executive vice president, and

John RobertsonJohn Robertson, vice president and senior economist, both in the Atlanta Fed's research department

February 15, 2013 in Employment, Labor Markets, Small Business | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Promoting Job Creation: Don't Forget the Old Guys:


I think that to a certain extent the decline in new business formation in 2010-2011 as opposed to 2009 is a bit of a statistical illusion for a couple of reasons.

First, a lot of the start ups in 2008 and 2009 reflect arangements that were first agreed to in 2007 or before--there is a delay between the time when the committment to start a new business is made and when the articles of incorporation are actually filed. Moreover, business cycles follow a bit of a learning curve--back in 2008 and most of 2009, people thought the recession was a lot smaller than it actually was, a misunderstanding that was cleared up with better data in 2010.

Second, there is an extent to which small and new businesses are an inferior good. New businesses typically specialize in smaller orders and attempt to undersell large established suppliers. At the height of the recession, customers downsized, switching from large established providers to small/new competitors in an effort to reduce spending. As the owner of one of these small businesses during the recession, it was definitely my experience that while the type of customers changed in 2008-2009, overall business was still up compared to 2007. That ended when, in 2010, customers started up-sizing again and went to the larger industrial suppliers, so that my business wasn't really hit by the recession until the recovery was in full-swing. If this is a common experience, then the data will show that small/new businesses fared better than large established business in 2008-2009, but then took a hit in 2010-2011.

Posted by: Matthew Martin | February 16, 2013 at 03:27 PM

Post a comment

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

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

February 05, 2013

2013 Business Hiring Plans: Employment, Effort, Hours, and Fiscal Uncertainty

How much is fiscal uncertainty holding back hiring? The answer seems to depend on whom you ask. Early in January, the Atlanta Fed spoke to 670 businesses in the Southeast about employment. Conditional on the respondents’ 2013 hiring plans (expand, hold steady, or contract), the following set of charts summarizes the results for how the businesses viewed activity relative to their own interpretation of “normal” along three dimensions: their current employment level, the amount of effort required from their staff per hour, and the average hours worked per employee. These questions were modeled on questions asked in the Atlanta Fed’s December 2012 Business Inflation Expectations Survey. In the following three charts, the green bars represent firms that said they planned to expand employment in 2013. The grey bars represent firms that said they did not plan to change their employment level in 2013, and the red bars represent firms that planned to reduce employment in 2013.

The first chart shows the results for current employment. Regardless of hiring plans over the next 12 months, most firms said they were currently at or below normal employment levels. Those planning on increasing employment over the next 12 months were a bit more likely to say they have already surpassed normal levels of employment than other firms, while those looking to shed employees were very likely to say their employment level is below normal employment levels.


Chart 2 shows that businesses are generally pushing hard along the effort dimension. Firms were quite likely to say that their staff’s effort per hour worked was currently at or above normal,  whether or not they were planning to change employment in 2013.


Chart 3 shows that firms planning to expand were very likely to say that average hours worked were at or above normal (28 percent said hours were above normal, 60 percent about normal), whereas firms planning to contract were more likely to say that hours were at or below normal (48 percent about normal, 39 percent below normal).


Taken together, these results suggest that some firms are approaching the limit of how far they can go along the intensive margins of effort and hours before they have to hire more workers. With effort elevated, as more firms increase average hours worked to above-normal levels, one might expect more hiring to follow.

Each business was also asked how uncertainty about future fiscal policy was affecting its hiring plans. Firms planning to reduce employment tended to cite fiscal uncertainty as having a negative impact on their hiring plans. However, for those firms, hours also tended to be well below normal, so it is unlikely that removing fiscal uncertainty would move many of those firms into expansion mode (although it may help stabilize their outlook).

In contrast, fiscal uncertainty was generally viewed as having less impact by those planning to expand employment and those planning to hold employment levels steady. Presumably, reducing fiscal uncertainty would move some of the firms planning to hold steady into expansion mode, and those planning to expand would do so a bit more. To get some idea of this potential, Chart 4 shows the responses by firms who reported above-normal effort per hour and above-normal average hours worked. About 40 percent of those businesses said that fiscal uncertainty had caused them to scale back their hiring plans.


It is unclear whether eliminating fiscal uncertainty would have a big impact on the hiring plans of these firms. But these results suggest that it sure couldn’t hurt.

John RobertsonBy John Robertson, vice president and senior economist, and

Ellyn TerryEllyn Terry, a senior economic analyst in the Atlanta Fed's research department

February 5, 2013 in Employment, Labor Markets, Small Business | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference 2013 Business Hiring Plans: Employment, Effort, Hours, and Fiscal Uncertainty:


Post a comment

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

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

November 05, 2012

Reading Labor Markets

When the September employment report was released on October 5, the top-line payroll employment gain for the month, as reported in the U.S. Bureau of Labor Statistics' (BLS) establishment survey, logged in at 114,000. Under standard assumptions, a number of this magnitude would be barely enough to absorb the growth of the labor force and keep the unemployment rate constant. In contrast, in that same October 5 report we learned from the BLS household survey that the measured unemployment rate fell from 8.1 percent in August to 7.8 percent in September.

According to Friday's BLS report on the employment situation for October, the top-line payroll employment gain for the month from the establishment survey was 171,000. At that pace—which is also the current average gain for the past three months—the Atlanta Fed jobs calculator suggests the unemployment rate should fall another one-half of a percentage point over the next year. At the same time, according to the BLS household survey, the unemployment rate rose from 7.8 percent in September to 7.9 percent in October.

This is as good an illustration as any to explain why, on November 1, Atlanta Fed President Dennis Lockhart said the following in a speech to the Chattanooga Tennessee Downtown Rotary Club:

In its post-meeting statement on September 13, the FOMC said, "If the outlook for the labor market does not improve substantially, the Committee will continue its purchases of agency mortgage-backed securities, undertake additional asset purchases, and employ its other policy tools as appropriate until such improvement is achieved in a context of price stability."...

For policy purposes, I think it's appropriate to be cautious about relying on a single indicator of labor market trends—for example, the unemployment rate—to determine whether the condition of "substantial improvement" has been met.

As the FOMC went into its September meeting, the official BLS statistics indicated that net U.S. job creation in August was a mere 92,000. That number is below the “all else equal” threshold of about 100,000 jobs required to keep the unemployment rate from rising, and that information is what Fed policymakers had in hand when they met on September 12–13 and decided on the policy action described by President Lockhart.

On Friday, after two revisions, the BLS told us jobs expanded by 192,000 in August, well above the average for the jobs recovery that started in early 2010, 100,000 jobs (more than double) above the initial estimate.

Looking through month-to-month variations is not a lot of help in real-time tea-leaf reading. Here is the 12-month moving average of employment gains, the blue line indicating the way things looked in September, the red line showing the way they look today:

Over time, it remains the case that monthly employment gains are pretty consistently coming in at 150,000 to 160,000 jobs created per month, and that rate has been enough to generate relatively steady declines in the unemployment rate:

That could change, of course, and the last four months of data have generally shown an acceleration in the job-growth trend. But the data definitely were not indicating that trend as it was happening, an unfortunate reality that isn't likely to change. One way to soften the blow of that problem, emphasized in the Lockhart speech, is to keep an eye on as a broad a set of signals as possible:

... let me share a qualitative framework for defining "substantial improvement."

The starting point certainly should be the headline unemployment rate and the payroll jobs number. The interpretation of movements in these two statistics would be enriched and reinforced by a review of additional data elements.

I added the emphasis there, as I think the point bears highlighting. President Lockhart goes on to give examples of what he would look for in determining whether the substantial improvement threshold has been met. Things like reductions in the numbers of marginally attached and discouraged workers, growing labor force participation rates, declining numbers of people who want full-time work but have to settle for part-time, and positive forward indicators like falling initial claims for unemployment insurance.

The Calculated Risk blog continues to be a one-stop shop for a lot of this information—here and here, for example—and overall nothing much overturns the picture of steady, but slow, progress. That would suggest the acceleration of the past several months is probably not a new trend, but a continuation of the same-old same-old. But then again, the track record painfully demonstrates how hard that is to know in real time.

Dave AltigBy Dave Altig, executive vice president and research director at the Atlanta Fed


November 5, 2012 in Employment, Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Reading Labor Markets:


Should not the "starting point" be % of population employed? This number has been dropping since 2008 and at its lowest in 30 years - a better metric of the labor market than a headline seasonally adjusted number.

The 2nd point should be the quality vs quantity of jobs. Adding minimum wage jobs versus higher paying skilled jobs should be recognized.

Who's gaining jobs & why?
The 55-64 year old age group has had the highest employment gains - while the 20-30 somethings have lost 2 mil jobs in the past 4 years.
ZIRP policies have reduced fixed income returns to near zero. The 55+ups HAVE to get jobs as their incomes from interest have evaporated. 55+ups have the experience to gain (any paying level) jobs over the 20/30 somethings.

New household formation will continue to suffer as the (many) younger gen is unable to earn and save for a downpayment on a house.

Posted by: Barclay | November 10, 2012 at 10:02 AM

Post a comment

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

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

October 10, 2012

Divergent Jobs Reports: Will the Real State of the Labor Market Please Stand Up?

The September employment report from the U.S. Bureau of Labor Statistics (BLS) was predestined to create a significant amount of buzz. But the confluence of headline jobs growth at a modest clip of 114,000 and a surprisingly large 0.3 percentage point reduction in the unemployment rate has made the report more buzz-worthy than we (here at macroblog) expected. Although some of the commentary has been more heat than light, there have been some particularly good reminders of the difference between the establishment survey data, from which the headline jobs figure is derived, and the household survey data, from which come the unemployment statistics. The discussions on Greg Mankiw's blog and by Catherine Rampell (at The New York Times's Economix blog) are especially useful. Or, perhaps even better, you can go to the source at the BLS.

It's important to remember that both surveys are subject to error and, because of its much smaller sample size, the household survey can be subject to particularly sizeable swings. Specifically, the standard error of the household survey's monthly change in employment is 436,000(!). Based on the most extreme assumptions about flows in and out of unemployment and in and out of the labor force, understating or overstating actual employment by 436,000 would imply a measured unemployment rate ranging from 7.5 percent to 8.1 percent. (The BLS estimate of the standard error for unemployment puts a range on September's number of 7.6 percent to 8 percent.)

In his post, Greg Mankiw makes reference to a Brookings Institution paper by George Perry from a few years back that offers what is probably good advice: since both the payroll and household surveys are subject to error, and since the errors in each are likely unrelated to one another, the clearest picture about what is happening to employment in real time can be gleaned by combining information from both.

In fact, in a directional sense, both the household and payroll surveys are giving the same signals. In the table below, we compare the recent trends in monthly job gains measured in both surveys. The coverage in the two reports is slightly different. Unlike the payroll count, the household survey includes the self-employed and counts multiple jobs held by a single person as a single instance of employment. Because of this, the BLS also reports an adjusted version of the household survey, called the payroll concept adjusted employment measure. This payroll-consistent measure is designed to control for definitional differences across the household and establishment reports and also makes statistical adjustments for changes to the population controls in various years. So we include the data from this measure in the last column of the table.


Overall, all three measures suggest a weaker trend over the last six months than over the last nine months. All three measures also indicate that things were somewhat stronger on average in the last three months than in the prior three months. The bottom line in our view is that, though the employment levels can be quite different across the three measures, all suggest that the jobs picture has improved somewhat in the past three months.

The suggestion in George Perry's Brookings paper—combining the household and establishment data—can be implemented by constructing a weighted average of the two surveys. In our variation we put weights in proportion to the inverse of the sampling variability of the payroll and household surveys, which would roughly imply an 80 percent weight on the establishment measure and 20 percent on the payroll-consistent household measure. The estimates using these weights are reported in the last column of the table above. Because the component employment measures display directionally similar trends in recent months, the weighted average does as well. 

In a speech given a few weeks ago, Atlanta Fed President Dennis Lockhart, our boss here, offered the opinion that

Taking a two-year view, the trend rate of gains in employment has been roughly 150,000 per month. This pace would be sufficient, at current levels of participation in the workforce, to sustain a steady, gradual reduction of the unemployment rate.

As the September employment reports show, predicting the unemployment rate month to month can be tricky business, and there may be better ways than just extrapolating from the jobs data. But thus far we are inclined to think that slow but steady progress on the jobs front is still the best story.

David Altig By Dave Altig, executive vice president and research director at the Atlanta Fed

October 10, 2012 in Employment, Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Divergent Jobs Reports: Will the Real State of the Labor Market Please Stand Up?:


The problem is the focus on seasonally adjusted data. The NSA data trends are very clear and consistent. See and

Posted by: Lee Adler / The Wall Street Examiner | October 10, 2012 at 05:58 PM

Post a comment

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

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