How's the Connecticut economy doing? Better, thanks, but we're not exactly sure -- in part because of a quirk in the way the state rolls out job reports.
Although the national jobs picture brightened considerably in January and February, Connecticut has yet to issue any data for 2012. Every March, the state Department of Labor churns out a delayed figure for January along with revisions for the previous year.
It's happening Monday afternoon -- the biggest release of the year for state job-watchers -- and economists are optimistic not only for January but for the 2011 revisions.
Preliminary data show that employers added 9,000 jobs in 2011. That's not terrible, but hardly what we need to knock the unemployment rate down below 7 percent any time soon. The 2010 gain was 13,600.
Economists Nicholas Perna and Donald Klepper-Smith both expect to see upward revisions for 2011 and positive numbers in January because indicators such as new unemployment claims, not to mention the general feel of the job market, point to a stronger picture than what we've seen.
"At last it looks like things are coming in better than expected for a change," said Perna, a Yale lecturer and adviser to Webster Bank. Although there are threats to the recovery, he added, "better to be in this situation than the situation we found ourselves in toward the middle of last year."
Later in March, the state will catch up and issue its report for February. That, too, is expected to show solid gains because the national report issued Friday tallied 227,000 new jobs for February -- though no change from the unemployment rate of 8.3 percent.
Connecticut's 8.2 percent jobless rate in December could drop to 8 percent or even lower with positive reports for January and February.
But the biggest news could be in the revised job total for 2011, because Monday's revisions could be very large.
Just a small adjustment in the underlying numbers could cut that 9,000 figure down by half, if we're not lucky, or double it to 18,000, which would be more than 1 percent growth -- still not spectacular, but very respectable for a high-cost, high-wage state that doesn't have much population growth.
"I'd say right now if I were to be guessing at a number, I wouldn't be surprised to see it somewhere between 12,000 and 15,000," said Klepper-Smith, of DataCore Partners in New Haven, an adviser to Farmington Bank.
Here's how it works: The government surveys 5,000 employers each month for job totals. State labor officials, working with the U.S. Bureau of Labor Statistics, crunch the numbers to show a total of jobs in the state, adjusted to smooth out seasonal variations.
For any given month, the survey-based numbers can be unreliable. Each month's figure is revised a month later. Then in March, numbers are revised to reflect actual payroll data. But the March revision only gives an accurate total for the previous March, and the rest of the year is adjusted accordingly.
Unemployment numbers are based on a separate household survey, which is why they sometimes don't line up with the job totals.
Looking back at last year's job totals matters for two reasons. First, a better number would change the collective psychology in Connecticut, and that can lead to real gains going forward as consumers and businesses feel better about spending money. And second, any improvement brings us closer to returning to the pre-recession peak of March 2008.
Between that month and January 2010, Connecticut's economy shed 119,200 jobs, a hair under 7 percent -- second in the modern era to the recession of 1989-92, which took 10 percent of the state's jobs.
This has been a painfully slow recovery, of course, as the state has regained just 34,300 of the lost jobs, or 29 percent, compared with the nation's recovery, 39 percent as of Friday, Klepper-Smith pointed out.
That means, as Klepper-Smith illustrated at a recent talk, that even a recovery averaging 2,000 added jobs per month would not return Connecticut to the 1,713,000 jobs we had in early 2008 until well into 2015.
Often, but not always, the revisions push the totals further in the same direction they're moving. In 2005, the wildest recent example, a yearly total of 8,400 new jobs for 2004 was revised to 22,500. Still later, in yet another revision, the 22,500 figure was revised back down to 15,500.
That sort of swing won't happen again, said Andy Condon, director of research for the state Department of Labor. The Bureau of Labor Statistics discovered "a methodological procedure they had never seen before," and corrected it, he said.
Condon, adhering to federal law, declined to even hint at which way Monday's revisions will move. "The benchmarks are plus and minus each year; there is no particular pattern," he said.
All of which begs the question, in Perna's view: If we have the actual payroll data in hand, why does it take a year or longer to get the numbers right?
"We should spend more money nationally on gathering numbers, because these things are important," Perna said.
He mused that it isn't happening because politicians like to point to tangible achievements. "You can't cut ribbons on ... IT infrastructure."
Reprinted with permission of the Hartford Courant.
To view other stories on this topic, search the Hartford Courant Archives at