Every month the Pennsylvania Department of Labor and Industry and the U.S. Bureau of Labor Statistics release details about the previous month’s employment totals for the Pittsburgh metropolitan statistical area (MSA). These releases elicit responses from analysts who try to make sense of the numbers one way or another.
But what hardly ever gets mentioned in the media is that these figures will be revised at least twice as the original release represents preliminary estimates which will be adjusted in the subsequent month. And then every March the entire previous year, and sometimes two preceding years, are re-benchmarked resulting in dramatic changes. Yet the media only focuses on the preliminary releases which are oftentimes very misleading, especially for policy makers trying to assess the state of the economy.
And now, right on schedule, numbers for 2015 have been re-benchmarked and the final tally is quite different from those preliminary figures.
The count of total nonfarm jobs for 2015 reported on the original releases (1,177,483) placed the average yearly increase compared to 2014 (1,161,125) at 16,350—a rise of 1.4 percent. The monthly increases compared to the twelve month earlier figures ranged from 4,800 (March) to 32,800 (July). February (9,200) and March were the only months to post year-over-year increases of less than 10,000 total nonfarm jobs. On the surface, it was a respectable year for job growth. But that was before the re-benchmark.
March 2016 revisions lowered the number of nonfarm jobs in 2015 (1,161,177), compared to 2014 (1,159,575), to an annual average gain to just over 2,140 jobs—an uptick of only 0.18 percent. The after-revision gains range from a low of negative 1,500 (December) to a high of 5,400 (January). There were three months with year-over-year decreases—December, November (-1,000), and June (-400). Thus the re-benchmarking wiped out nearly 14,000 of the presumed job growth in 2015 turning the year from respectable to anemic. Again long after the media reported the rosier picture painted by the preliminary releases.
The re-benchmarking also affected the number of total private jobs. The yearly average difference, based on the original press release data, came in at 16,925 jobs—a growth rate of 1.62 percent over the 2014 reading. The range spanned a low of 6,000 (March) to a high of 33,500 (July). Only March had an increase of less than 10,000 jobs. After re-benchmarking the data, that average has fallen to just over 3,280—increasing only 0.32 percent over the 2014 figures. The range now covers a low of 100 (December) to a high of 7,300 (January).
Revisions by industry groups were once again mixed, with a few having their job counts being revised upward, but most with downward adjustments. As reported by the original releases the goods producing groups—mining and logging, construction, manufacturing—had an average annual increase of 2,208 jobs compared to 2014—a rise of 1.42 percent. Unfortunately, the re-benchmarked data shows an average monthly decrease of 1,200 jobs. This swing of over 3,400 jobs shows the degree of volatility there is between the originally released data and the final re-benchmarked data.
Of the three industry groups in the goods producing sector, mining and logging went from having average monthly year-over-year gains of 1,158, to having average monthly year-over-year losses of 525. This may be in large part due to the sharp drop in the price of natural gas which has hampered development and drilling in the Marcellus Shale formation.
The construction industry and the manufacturing industry also had reductions to their jobs counts, but not nearly as wide a swing as in the mining and logging sector. Based on the original releases, the average monthly 2015-over-2014 difference for construction was 1,325. However, after re-benchmarking the data for 2015 and 2014, this industry’s average annual growth was reduced by more than 930 jobs on a year-over-year basis in 2015 to 392. The manufacturing sector, based on the original press releases, had shown average monthly year-over-year losses of 275 jobs. The re-benchmarked data not only confirmed these losses, but expanded them to 1,075 jobs.
For the service providing industries, nearly all sectors showed reductions to their average monthly year-over-year jobs totals. Overall the service providing industries had shown an increase to the number of average monthly year-over-year jobs of 14,150. With the re-benchmarking of the data, those gains fell sharply to 3,350. The growth rate fell from 1.4 percent to 0.33 percent. We will look at some of those sectors below.
The one sector that had increases to its average monthly year-over-year jobs totals is the professional and business services sector. Based on the original press releases, the average monthly year-over-year gains to employment for 2015 compared to 2014, was 583. After the March 2016 re-benchmarking, that number increased to 2,950. This was aided by an increase to the re-benchmarked data for both the 2015 and 2014 data, with the average monthly increase to 2015 jumping by 3,250 and for 2014 by more than 880 jobs.
On the other end of the spectrum is the education and health services industry group. Based on the original press release data from 2015, this industry group had a monthly year-over-year average increase of 4,658 jobs. However with the March 2016 re-benchmarked data that difference became a decrease of over 830 jobs. Thus a nearly two percent increase in education and health jobs became a loss of 0.35 percent with the re-benchmarking of the data.
The two main sectors in the industry group, education services and health care and social services, both took sharp losses with the revised data. The education services sector had an average monthly boost of 650 based on the original 2015 press releases to having a monthly average drop of 725 –a swing of 1,375 jobs for 2015 compared to 2014. The health care and social assistance sector took an even bigger hit. Based on the original press release data, this sector had an average monthly year-over-year rise of 4,008. However, after the re-benchmarked data, that monthly year-over-year average is now a negative 108—a swing of more than 4,100 jobs.
For most of 2015 the leisure and hospitality sector accounted for a large share of the job growth in the area. We warned numerous times that the growth being reported was untenable and the re-benchmarking would likely wipe out most of that growth. We were correct. The average monthly jobs in leisure and hospitality in 2015, based on the original press releases, came in at 120,675. This was revised downward to 116,333 in the re-benchmarking. The data for 2014 was also re-benchmarked slightly, dropping from 114,633 to 114,325. Thus, the original difference dropped from 6,042 to 2,008—a fall of over 4,000 jobs. While this is still respectable growth, these are typically lower paying jobs and not in the area where you would want to see the strongest growth.
The re-benchmarking of the 2015 data has wiped out most of the originally reported job gains from the press releases—total nonfarm and total private jobs grew by only 0.22 and 0.32 percent respectively. What could be the reason for such anemic growth?
Two reasons come to mind. As we wrote in a previous Policy Brief (Volume 15, Number 28), the Kauffman Foundation’s Index of Startup Activity ranked the Pittsburgh area second to last in startup density, rate of new entrepreneurs, and opportunity share of new entrepreneurs, placing the region last overall amongst the forty largest metro areas in the country. Furthermore, the Pittsburgh area, and Pennsylvania overall, has fared poorly on business friendly indicators for decades. This does not bode well for job growth and that is manifesting itself in the actual data.
Secondly the latest Census data shows that the Pittsburgh region has lost population since the last official census in 2010. Is the contraction of the area’s population driving the weak job gains or is it the other way round. Is it more appropriate to say that the weakness in job growth is more responsible for the outflow of migrants from 2014-15? After all, people are typically drawn to an area for economic opportunities more than for any other reason, including environmental amenities. Maybe policy makers should address the factors inhibiting business startups and business expansion. Marcellus Shale activity must not be depended upon exclusively to pull the region forward.