In the Nation's Interest


By Joe Minarik

As just about everyone knows, the monthly employment situation report, produced by the Bureau of Labor Statistics (BLS), is one of the most important looks (if not the most important look) at where our economy is headed.  It is both a snapshot of activity in a given month, and an indicator of consumer demand in the months to come.  It has an additional status as a reliable monthly indicator, whereas some other important economic measures are quarterly – and so provide information only with a longer lag than the jobs report.

What is the Employment Situation Report?

The employment situation report is based on two different surveys – one of households, and the other of employers.  In years gone by, the household survey provided the “headline” number that grabbed all of the attention.  That headline was the unemployment rate – the percentage of all persons who are either (i) working or (ii) looking for work who do not have jobs.  However, over time, observers have wised up to the fact that the household survey is less reliable, in part because of its less-than-ideal sample size.  In addition, some people are reluctant to bother to participate – sometimes because they do not want to share information with the government.  And some people do not grasp all of the subtleties of employment status, as hard as the BLS tries to explain them, or they might try unsuccessfully or inaccurately to answer for another member of the household.

Accordingly, the focus has shifted from the survey of households to the survey of employers, whose headline number is the count of non-farm payroll jobs.  The employer (“establishment”) survey includes more than twice as many businesses as the household survey includes dwellings.  Considering also that each surveyed business represents on average more than three work sites, and each work site might include many employees, the employer survey is far more reliable and stable than the household survey.  Furthermore, business representatives who provide those data become more knowledgeable and accurate over time than the household respondents, who tend to quickly rotate out of the survey.  This factor adds another measure of reliability to the inherently more reliable establishment survey.

The establishment survey has its own weaknesses, however.  Notably, the sample of businesses, for practical reasons, simply cannot be kept up to date with new businesses as they are created in real time.  (Newly closed businesses simply do not respond, and so are identified more promptly.)  Accordingly, the survey numbers must be adjusted with estimates of new business formations.  And obviously, such estimates are least accurate at precisely the times when they are most important – that is, when the economy is changing direction, and new businesses are beginning to be formed more rapidly or more slowly.  (A further weakness is that moonlighting employees are counted twice.)

In a perfect world, both surveys would accurately tell us separate things that policymakers need to know.  In this veil of tears, what we lose most because of statistical shortcomings and realities is probably lost from the household survey.  This is especially true now, in the aftermath of the financial crisis.  The steep economic downturn of the last decade clearly crunched employment, with a disproportionate impact on some workers and households.  People who lost their jobs at the wrong time and in the wrong place – especially in localities that were the most badly hurt by the collapse of the housing market – are the big question marks of today’s economy.

Deep Recessions and Recoveries

Over the last few years, some observers have repeated the truism that deep recessions have tended to spawn sharp recoveries – and so those observers have argued that the current tepid recovery must result from recent economic policymaking errors.  What those observers have missed in their superficial analysis is the deeper finding of economists Carmen M. Reinhart and Kenneth S. Rogoff, who have shown that economic downturns caused by financial crises have been followed by slower and more-protracted recoveries, likely caused by systemic damage from the financial crises themselves.  Clearly, if and when, this recovery reaches the breakout stage and begins to resemble the sharp rebound people expected, there will necessarily be some significant departures from the lackluster trends that we have hitherto seen in the labor market.

May 2014 Employment Report – Released on June 6, 2014

The May employment situation report, to my eyes at least, shows no signs of such a breakout.  There is continuing forward momentum that stems any concern about a near-term reversion to recession.  But there is no real departure from trend in any of the indicators that most clearly showed the pain of the financial crisis half a dozen years ago.

Here is a brief rundown of the key numbers in the report:

Non-farm payroll employment.  The headline number from the establishment survey shows a plus of 217,000 jobs.  That would be a strong number in the middle of an established expansion, after the economy already had worked off most or all of its job losses.  But it is just a mild positive now, following the weakest job market that the nation has endured in decades.  There was significantly more job growth in the recoveries of 1983-84 and 1993-94 than there has been in this cycle – even though the population was smaller then – and those more-robust recoveries came sooner after the downturns then where we find ourselves today.  (See the following chart.)

Labor-force participation.  Although the labor force grew in May, the amount of growth was so small that the labor force remained smaller than it was in March.  The increase in participation was small enough that the labor-force participation rate – the percentage of the adult population that is either working or looking for work – merely remained constant.  The labor force has shrunk so much since the financial crisis that merely ending its decline is not newsworthy – though, of course, ending the decline is a necessary first step to an ultimate turnaround.  In early 2008, the labor-force participation rate was more than 66 percent, whereas now it is less than 63 percent.  (See the following chart.)

Discouraged workers.  Associated with the labor-force participation rate, and the key story of the current economic recovery, has been the extended exit of discouraged workers from the labor force.  Technically, a discouraged worker is someone who has recently looked for work, and who wants a job, but has given up the search after having drawn the conclusion that looking would be fruitless.  (There is further category of persons who are “marginally attached to the labor force,” but who have most recently looked for work only at some time in the preceding 12 months.)  The data on discouraged workers are not seasonally adjusted, and so comparisons from one month to the next can be misleading; but the number in May of this year is not much below the number in May of last year, indicating little change.  (See the following chart.)  If and when the economy breaks out, we should see the number of discouraged workers drop while the size of the labor force and the labor-force participation rate jump.  That is not happening yet; the number of discouraged workers, though declining, remains about double what it was before the financial crisis.

Long-term unemployment.  The number of persons unemployed for 27 weeks or more was essentially unchanged last month.  Again, in a breakout recovery, that still-elevated number should drop – noting, however, that success requires those long-term unemployed to find jobs, not to become discouraged workers and leave the labor force.  The failure of the number of discouraged workers to drop more (see chart above) makes the stubborn decline in the number of the long-term unemployed look a little limp.  The number of long-term unemployed has declined by about half from its peak in 2010, but it remains almost triple its level of early 2008, before the full explosion of the financial crisis.  (See the following chart.)

There are other important indicators in the monthly report, but these four together tell us what we need to know.  The simple non-farm payroll employment count is the best single-number summary, but the details that we care about come in the household survey.  And as was noted at the outset, the household survey is inherently the less reliable of the two.

Possible End Games

We need to read the household survey’s tea leaves, because the difference between the economy’s success or failure now probably comes in the disposition of (i) the labor force departures and (ii) the long-term unemployed.  The Federal Reserve is trying to discern how long it can keep its collective right foot on the accelerator pedal.  If the unemployed keep searching for work and the discouraged workers sense opportunities and begin looking again, the economy will have plenty of room to grow, and the Fed can keep pushing for that growth.  But if the labor force continues at an historically low percentage of the adult population, then the Fed will have to lift its accelerator foot – and perhaps even tap the brakes – and the feared “new normal” may become a vivid reality.

The May employment situation report did not show an exciting level of job growth, nor did it show an encouraging response from the “shadow labor force” – yet.  The household survey is inherently noisy, and so no single month is definitive.  But the United States clearly needs definitively favorable data – not a repeat of the moderately positive May employment release – before we can be confident of a breakout in the economic recovery.