By Philip C. Mead
The problem is a common one: So called official sources constantly revising economic data.
The CT News Junkie reported in October about the wonderful news on job creation in Connecticut. We had created over 1100 jobs in October. The original report had to be, well - revised. The original author, Mr. Hugh McQuaid, wrote a virtual cut and paste piece that sure read a lot like the previous article. He did put in the correction, thought. The state had reported only a 500 job gain. Even that isn’t quite right.
Meantime, our friends over at the federal Bureau of Labor Statistics had a different estimate. October 2022 and November 2022 had increases of around 1600 jobs. They also put total job growth at 36,000, year over year for November, while Hugh McQuaid of CT News Junkie claimed 50,000 jobs for 2022 as of October 2022. If you use 50,000 as the denominator, that is a 28% difference. If there is that great a difference in the two figures, such as that, not only is someone wrong, but the methods for estimation are also highly suspect. Were these projections? Extrapolations? Or were they the bastard child of some poorly constructed models? “I didn’t do it”, said some analyst.
The State of Connecticut Labor Situation report for December 2022 is a classic case of needing to find the buried lead – aka, the real headline. The stated headline is about nonfarm industry payrolls increasing by 6,200 positions in November 2022. However, then the buried lead, oops:
“ The originally released small October 2022 job gain of 500 (0.03%) was revised down by 1,600 positions to a loss of 1,100 (-0.07%). Connecticut's unemployment rate fell to 4.2% over the month from the 4.3% rate in October 2022 and is now lower by one percent from the 5.2% rate a year ago. “
More over, the estimate for November 2021 is some 1600. That makes for flat growth in hiring. Yet, Mr. McQuaid noted the claim of decreases in unemployment by the Connecticut Labor Department. How can hiring be flat while unemployment goes down? This makes for another suspect piece of data.
Messy data in economics is a professional hazard. Tagging data as “preliminary” is a common practice. Nonetheless, they should be more forthcoming in explaining why there can be such dramatic variations and inconsistent patterns in the data.
It should be pointed out that even the incorrect job gain estimate of 500 (and how did they screw this up? We are still waiting for an answer) versus the 1,600 total job loss, is roughly a 400% difference. People get fired for less.
What makes this confusion and error all the more troubling is that the Connecticut Department of Labor supposedly shares data and resources with the U.S. Department of Labor. Neither can be trusted with so many problems in their reporting. Subsequent revisions make it no better. When will the next revision be coming?