A Subtle View of Labor Market Improvement

In a speech delivered Tuesday to the Japan Society in New York City, Federal Reserve Bank of New York President William Dudley offered his view on how he might assess the appropriate pace of the Federal Open Market Committee's (FOMC) current $85 billion per month asset purchase program:

Let me give a few examples of how my own thinking may evolve. In terms of our asset purchase program, I believe we should be prepared to adjust the total amount of purchases to that needed to deliver a substantial improvement in the labor market outlook in the context of price stability. In doing this, we might adjust the pace of purchases up or down as the labor market and inflation outlook changes in a material way. For me, the base case forecast is not the sole consideration—how confident we are about that outcome is also important.

Because the outlook is uncertain, I cannot be sure which way—up or down—the next change will be. But at some point, I expect to see sufficient evidence to make me more confident about the prospect for substantial improvement in the labor market outlook. At that time, in my view, it will be appropriate to reduce the pace at which we are adding accommodation through asset purchases. Over the coming months, how well the economy fights its way through the significant fiscal drag currently in force will be an important aspect of this judgment.

My own boss, Atlanta Fed President Dennis Lockhart, expressed a similar view in a speech to the Birmingham Alabama Kiwanis Club last month:

The key word in the phrase "substantial improvement in the outlook for the labor market" is outlook. For my part, a critical consideration in judging how much longer asset purchases should continue will be confidence in the positive outlook. Confidence that is solidly grounded in improving economic data, accumulated over a sufficient span of time, will help me conclude that the work of the large-scale asset purchase program, as a temporary supplement to conventional interest-rate policy, is complete.

And there is this, from the minutes of the latest meeting of the FOMC (emphasis added):

Participants also touched on the conditions under which it might be appropriate to change the pace of asset purchases. Most observed that the outlook for the labor market had shown progress since the program was started in September, but many of these participants indicated that continued progress, more confidence in the outlook, or diminished downside risks would be required before slowing the pace of purchases would become appropriate.

Neither President Dudley nor President Lockhart (nor the minutes) indicates where we are on the confidence scale at the moment. But at least outside the Fed, there is some evidence confidence in the labor market forecast is increasing. The following chart shows year-over-year averages of the interquartile range of four-quarter-ahead unemployment rate forecasts from the Federal Reserve Bank of Philadelphia's Survey of Professional Forecasters:

The interquartile range is essentially the difference between the most optimistic one-fourth of the forecasts in the Philadelphia Fed's panel and the most pessimistic one-fourth of forecasts. It is thus a measure of dispersion, or forecast disagreement.

The trend in this measure of forecast disagreement is clearly—very clearly—downward. That doesn't exactly say that each individual forecaster is becoming more confident about his or her individual outlook (though this type of dispersion measure is often used as a proxy for overall uncertainty). Even less does it mean that forecast uncertainty has fallen to the level President Dudley, President Lockhart, or any other Fed official would deem sufficient to alter policy in any way. The FOMC minutes, for example, include this...

A number of participants expressed willingness to adjust the flow of purchases downward as early as the June meeting if the economic information received by that time showed evidence of sufficiently strong and sustained growth; however, views differed about what evidence would be necessary and the likelihood of that outcome.

... and following his congressional testimony Wednesday, Chairman Bernanke engaged in a Q&A, and ABC News summed up the state of the policy discussion this way:

When asked by Kevin Brady, the Panel's chairman, whether the Federal Reserve could start winding it back before before September's Labour Day holiday, Mr Bernanke responded, "I don't know. It's going to depend on the data."

It really need not be emphasized that I don't know either. But the narrowing of opinions of where things are headed must signify some sort of progress.

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

 

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 participation...by 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:

(enlarge)

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.

(enlarge)

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:

(enlarge)

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...

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

Building a Better Jobs Calculator: Choose Your Own Payroll/Household Employment Ratio

To provide even greater flexibility, the Federal Reserve Bank of Atlanta's Jobs Calculator has been enhanced to allow the user to adjust another statistic used in the calculations. The statistic is the ratio of payroll to household employment and is a necessary component that links the target unemployment rate with the resulting required payroll employment growth.

The fact that estimates of payroll and household employment numbers reported by the U.S. Bureau of Labor Statistics (BLS) differ each month received a lot of attention in the fall of 2012 when, in October, the BLS reported a whopping 0.3 percentage point drop in the unemployment rate, accompanied by a rather tepid growth of 114,000 jobs in payroll employment. The culprit in that apparent incongruity is that the Household Survey (from where we get the unemployment rate) reported a gain of 873,000 jobs. That particular employment report (and its divergent statistics) received extra attention since it was the last employment report before the November 2012 election.

As Atlanta Fed Research Director Dave Altig pointed out at the time (in this blog post) and as others discussed (here and here), the two most important measures of labor market conditions come from two different surveys—the Establishment Survey, which produces the payroll employment number from the Current Employment Statistics (CES) program, and the Household Survey, which produces the unemployment rate from the Current Population Survey (CPS). Both surveys claim to estimate the number of jobs in the economy. However, the employment numbers they produce are different for several reasons, detailed in one of the Jobs Calculator's FAQs.

The good news is that even though there may be wide discrepancy in the change in employment reported by the two surveys in any particular month (as we saw in October 2012), any one-month divergence does not persist. In other words, the two employment series closely track each other.

This is good news for the Jobs Calculator, since a conversion needs to be made between the CPS employment implied by the target unemployment rate entered into the Jobs Calculator and the average monthly change in payroll employment (CES) needed to achieve the target unemployment rate. Since the two series closely track each other, wide deviations in month-to-month reported growth numbers will not severely affect the ability of the Jobs Calculator to make longer-term projections (within the limits of the other assumptions of the calculator). (In fact, unanticipated changes in the labor force participation rate are much more potentially problematic in making longer-term projections than are any potential variations in the conversion rate between CPS and CES employment numbers.)

The Jobs Calculator uses the average ratio of CES/CPS employment over the previous 12 months as the default conversion factor and now allows the user to see what happens if that ratio were to be different.

(enlarge)

The following example illustrates how innocuous that conversion factor is.

Suppose the goal is to attain a 6.5 percent unemployment rate in two years. Entering 6.5 in the unemployment rate target box and 24 in the box (for the number of months you want to take to get there) yields 164,917 as the average monthly change in payroll employment needed to achieve that goal (holding everything else constant).

Next, go down to the new line showing, "Average monthly CES/CPS employment ratio." You'll see the current default value for the ratio is 0.940. Click on the chart box on the far right of that line. You'll see that since 1980, that ratio has ranged from a low of just under 0.900 in about 1984 to a high of 0.969 just before 2000. Close the box.

Now, enter the low ratio number of 0.900 in the employment ratio box. At that low ratio, only 157,899 payroll jobs are needed to achieve your 6.5 percent unemployment rate in two years.

Next, enter the high ratio number of 0.969 in the employment ratio box. At that high ratio, 170,005 payroll jobs are needed each month to achieve your goal.

The current default CES/CPS ratio provides an estimated number of monthly payroll jobs needed to achieve your specified goal of a 6.5 percent unemployment rate in two years within a 5,000–7,000 job margin, based on the highest and lowest ratio values since 1980.

The bottom line? When it comes to factors that can derail a longer-term projection of the number of jobs needed to attain a specific unemployment rate in a given period of time, the degree to which household employment estimates deviate from payroll employment estimates is just not that important, nor are monthly discrepancies in these series’ reported growth, since the discrepancies aren’t absorbed into the trends. And there you have the reason we added this flexibility to the Jobs Calculator: to allow users to see this for themselves.

By Julie Hotchkiss, research economist and policy adviser in the research department of the Atlanta Fed

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."

(enlarge)

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.)

(enlarge)

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:

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

 

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

 

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):

(enlarge)

(enlarge)

(enlarge)

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."

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

 

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.

(enlarge)

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).

(enlarge)

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.

By Patrick Higgins, an economist at the Atlanta Fed, and



Mark Carter, a senior economic analyst at the Atlanta Fed

What the Dual Mandate Looks Like

Sometimes simple, direct points are the most powerful. For me, the simplest and most direct points in Chairman Bernanke’s Senate testimony this week were contained in the following one minute and 49 seconds of video (courtesy of Bloomberg):

At about the 1:26 mark, the Chairman says:

So, our accommodative monetary policy has not really traded off one of [the FOMC’s mandated goals] against the other, and it has supported both real growth and employment and kept inflation close to our target.

To that point, here is a straightforward picture:

I concede that past results are no guarantee of future performance. And in his testimony, the Chairman was very clear that prudence dictates vigilance with respect to potential unintended consequences:

Highly accommodative monetary policy also has several potential costs and risks, which the committee is monitoring closely. For example, if further expansion of the Federal Reserve's balance sheet were to undermine public confidence in our ability to exit smoothly from our accommodative policies at the appropriate time, inflation expectations could rise, putting the FOMC's price stability objective at risk...

Another potential cost that the committee takes very seriously is the possibility that very low interest rates, if maintained for a considerable time, could impair financial stability. For example, portfolio managers dissatisfied with low returns may reach for yield by taking on more credit risk, duration risk, or leverage.

Concerns about such developments are fair and, as Mr. Bernanke makes clear, shared by the FOMC. Furthermore, the language around the Fed’s ultimate decision to end or alter the pace of its current open-ended asset-purchase program is explicitly cast in terms of an ongoing cost-benefit analysis. But anyone who wants to convince me that monetary policy actions have been contrary to our dual mandate is going to have to explain to me why that conclusion isn’t contradicted by the chart above.

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

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.

By Dave Altig, executive vice president, and



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

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.

By John Robertson, vice president and senior economist, and

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

Visualizing Improvement

The December employment report has come and gone without much of a splash. The roundup of commentary at MarketWatch included these observations:

"In short, not especially strong nor weak. While a 150–170K per month trend in payrolls is far from booming, it is strong enough over time to keep the unemployment rate moving down given slowing in the secular trend in labor force growth. Unemployment was flat in December, but it is down 0.4 points in the last six months."
—Jim O'Sullivan, chief U.S. economist, High Frequency Economics

"The overall picture is that the labor market remains lackluster. If this state of affairs continues throughout most of this year, as we expect, then it is hard to see the Fed dialing back or stopping its [quantitative-easing] purchases as some officials currently envisage."
—Paul Ashworth, chief U.S. economist, Capital Economics

Today's Job Openings and Labor Turnover Survey (JOLTS) report from the U.S. Bureau of Labor Statistics (BLS) did little to change that impression.

The interest in the Fed's reaction, always acute given the employment half of the Fed's dual mandate from Congress, has been heightened since the Federal Open Market Committee (FOMC) announced, first in September, that it will continue its asset-purchase programs as long as "the labor market does not improve substantially." But what constitutes substantial improvement is a matter of some art, as Fed Chairman Ben Bernanke made clear at his press conference following the December meeting:

In assessing the extent of progress, the Committee will be evaluating a range of labor market indicators, including the unemployment rate, payroll employment, hours worked, and labor force participation, among others. Because increases in demand and production are normally precursors to improvements in labor market conditions, we will also be looking carefully at the pace of economic activity more broadly.

A terrific gallery that includes "a range of labor market indicators" is available at the Calculated Risk blog—you might also check out the Cooley-Rupert Economic Snapshot—but here at the Atlanta Fed, we have been experimenting with our own method for summarizing the general state of the labor market. Though this project is very much a work in progress, the idea is to highlight variables that look at employer behavior, signals of employer and employee confidence, measures of labor resource utilization, and leading indicators of labor market conditions.

As a first pass, we've organized a collection of variables we find interesting, grouped in the categories I just described. In the category of employer behavior we include payroll employment (from the BLS's Establishment Payroll Survey, also known as the Current Employment Statistics survey), job vacancies or job openings, and hires (from JOLTS). Variables in the confidence category include hiring plans (from the National Federation of Independent Business's jobs report), job availability (from the Conference Board's Consumer Confidence Survey), and quits (from JOLTS). The utilization group contains unemployment (from the BLS's Current Population Survey, or CPS), marginally attached workers (from the CPS), the job finding rate (defined as the ratio of short-term to long-term unemployed as described in work by University of Chicago professor Rob Shimer), and workers who are part-time for economic reasons (from the CPS). Finally, we capture leading indicators with initial claims for unemployment insurance (from the U.S. Department of Labor), difficulty in filling jobs (from the NFIB small business jobs report), and temporary help services employment (from the CES).

We've based one prototype for how all of this information might be visualized at once on the following "spider chart":

(enlarge)

Here's how to read this chart: Think of each point on the inner orange circle as representing the value of each of our labor-conditions variables in the fourth quarter of 2009. Point A, therefore, would represent the value of "temporary help services employment" in 2009:IVQ, point B would be "payroll employment" in 2009:IVQ, and so on around the circle. (We've chosen 2009:IVQ as a benchmark because that's the last time we experienced two consecutive quarters of negative employment growth. You can thus think of 2009:IVQ as the quarter just before the beginning of the current "jobs recovery.")

The chart's outer dark-red circle represents the value of each of the labor-conditions variables in the first quarter of 2007—the beginning of the last recession, according to the National Bureau of Economic Research Business Cycle Dating Committee. Moving out from the inner orange circle to outer dark-red circle tracks the progress each variable makes from its value at the end of the recession (i.e. 2009:IVQ) toward its prerecession (i.e. 2007:IVQ) level. For example, being at point C today would mean initial claims for unemployment insurance have fallen by half relative to the amount they increased between 2007:IVQ and 2009:IVQ. ("Improvement," of course, involves lower numbers for bad things, like unemployment and initial claims, and higher numbers for good things, like payroll employment and hires.)

As of the December 2012 employment report, here's where we stand:

(enlarge)

The chart tells a familiar, but not too happy, story. Only one of the variables in the collection of employer behavior, employee and employer confidence, and labor resource utilization categories has recovered even half the gap from its prerecession benchmark. The labor resource utilization variables look particularly bad, with one variable—marginally attached workers—actually getting worse over the recovery as a whole. On the brighter side, our leading-indicator variables are looking relatively strong, perhaps portending improvement ahead.

The interpretation of these spider charts comes with several caveats. 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. Thus, the charts by construction are about visualizing the transition to some fixed benchmark, not a device for monitoring labor markets over the long run.

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. For example, the demographics associated with the aging of the baby-boom generations have arguably slowed the long-term trend in employment growth, meaning that a return to pre-recession payroll jobs will be slow even in circumstance that we would want to characterize as a "substantially" improving labor market.

Finally, signs of labor market improvement sufficient to alter the pace of FOMC asset purchases may be more about momentum or steady progress than about the return to a specific target or threshold. In fact, this chart depicts signs of such progress over the past three years...

(enlarge)

...but that progress has been very modest in some cases, notably along labor utilization dimensions.

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