Jobs Report Short on Reliability

Member Group : Allegheny Institute

(February 11, 2016)–The December jobs report for the Pittsburgh metro area was just released. At first glance, the report on balance seems quite positive. Data from the household survey show a lower unemployment rate, while the employer payroll survey shows a 14,000 increase in nonfarm jobs compared to December 2014. But do the data in the report paint an accurate or reliable picture of the area’s labor market?

A deeper look into the household data indicates there might be questions regarding the validity and reliability of the report. In December 2012 the initially reported civilian labor force for the region (not seasonally adjusted) stood just over 1.261 million people, later revised in the spring of 2013 to 1.227 million—a drop of 34,000. Then the initially reported figures for December 2013 showed labor force to be to 1.245 million. Again that number was revised downward, this time to 1.201 million—a decline of 44,000. Then in keeping with the pattern, the initial report twelve months later showed the labor force count for December 2014 to have risen once again to a year to year jump of more than 20,000 to 1.229 million. And once again, a downward revision reduced that number to just 1.195 million, a drop of 34,000.

Amazingly, from an initially reported 1,261,000 participants in the labor force in December 2012 to a revised December 2014 count of 1,195,000 the Pittsburgh metro area in just two years has seen the reported number of labor force participants slide by 66,000.

Now comes the initially reported figures for December 2015 that place the labor force at 1.222 million. Given the pattern of downward revisions to the numbers reported over the last three years means the latest labor force count is almost certainly going to be revised to a lower number. However, even if the December 2015 labor force is not lowered by a benchmark revision later this year it will be very close to the revised 2012 level, which means there has been no net increase in the metro area labor force for three years.

If the pattern of downward revisions continues this year, it will mean that in three years, the area’s labor force has actually gone down, assuming of course the revised figures are more reliable than the initially reported numbers. At this point who can be sure? One thing is certain, putting a lot of confidence in the initial reports on the labor force and household employment numbers as a good indication of the state of the labor market is simply not warranted because of the pattern of large revisions. Of course this is a problem inherent in a methodology that relies on a small monthly sample of households on which to base the estimates.

Meanwhile, in addition to the problems in the household survey, there are several glaring issues in the establishment (employer) payroll survey, especially in some of the service producing categories.

For December 2015, the recently released report indicates a gain of 17,500 private sector service jobs over the twelve months since December 2014. The largest employment gain occurred in the leisure and hospitality sector (which represents about 11 percent of all private jobs) with a 7,700 rise that accounted for a hugely disproportionate 53 percent of all private job growth between December 2014 and December 2015 (14,500). The initial monthly reports of rapid jobs growth in leisure and hospitality is a long standing trend for this sector. Sadly, major downward revisions in these jobs are also a long standing trend as we have documented in earlier Policy Briefs (Volume 14, Number 14 and Volume 15 Number 23).

Part of the problem is that much of the employment pickup is accounted for by arts, entertainment, and recreation for which there is not enough sampling to provide reliable estimates, so the Labor Department is truly guestimating the job count. As a result substantial downward revisions are necessary almost every year. This has gone on so long it would seem someone at the Labor Department would acknowledge and deal with the problem rather than simply continuing this monotonous pattern of reporting huge gains and then taking them away later with revised data.

Another area of questionable data occurred in the education and health services group which added 5,100 jobs over the twelve months ending in December 2015. Education services (does not include public K-12 schools) employment rose by 1,200 jobs while health care and social assistance rose by 3,900 jobs.

However, the education services job count looks suspiciously high and warrants closer inspection. While this sector is up 1,200 jobs year-over-year, its largest and most reliably documented component, colleges and universities, showed a drop of 2,100 jobs. Thus, employment in the combined remaining education services components (data for individual components are not published) must have risen by 3,300 to offset the decline in college jobs. This group is made up principally of private K-12 schools, education support services firm, and technical schools. As just mentioned, this group is reported to have had year-over-year growth of 3,300. This is quite astonishing considering that this group would have had to grow from 14,300 jobs in December 2014 to 17,600 jobs in December 2015—a rise of 23 percent in only twelve months. The absence of any explanation of exactly where such growth is occurring and why it’s so strong makes it hard to accept the numbers as credible. Downward revisions would seem to be inevitable.

Another concern about the jobs data and one we documented in a Policy Brief (Volume 15, Number 44) last year is the fact that the combined employment gains in Pennsylvania metro areas is almost twice as large as the state as a whole. And adding in the micropolitan area labor markets and county job markets makes the picture even worse with the sum of all the individual labor markets greatly exceeding state totals. Obviously, sampling problems are at the heart of the issue. But the upshot is that downward revisions in most metro area numbers are virtually certain.

Finally, it is noteworthy that state government employment across the Commonwealth was reported to have risen by 5,900 jobs from December 2014 to December 2015 after 4,000 job increases in both October and November. In the Pittsburgh metro area state government employment climbed by 700 jobs over the year ended December 2015. Ironically, the state government jobs growth has occurred even though there was no state budget during the last six months of 2015. It is bad enough that the state continued to spend on all departments and kept everyone employed but it was actually adding to staff and no doubt giving raises while the state funding for public schools in the state was on hold.

In summary, the monthly labor market and jobs reports for the Pittsburgh metro area—and doubtless most others across the Commonwealth —must be used carefully because of all the flaws and the enormous amount of revision that occurs every year in the preceding year’s data.

Jake Haulk, Ph.D., President
Frank Gamrat, Ph.D., Sr. Research Associate

If you wish to support our efforts please consider becoming a donor to the Allegheny Institute.The Allegheny Institute is a 501(c)(3) non-profit organization and all contributions are tax deductible.Please mail your contribution to:

The Allegheny Institute
305 Mt. Lebanon Boulevard
Suite 208
Pittsburgh, PA15234