Long Island jobs data does an about-face

Revised jobs data for Long Island released last week amounted to one of the most dramatic do-overs for economic statistics in recent memory.

The sharp revisions to the 2011 data followed a shift in responsibility to federal officials from state workers in calculating the numbers.

Original monthly estimates showed the Island losing jobs in the last eight months of 2011. But revised figures showed a local economy gaining employment in every month last year -- sometimes at a brisk pace. For example, June estimates, originally showing a 6,800-job decline, were reversed to a gain of 17,300 jobs.

Those reported losses carried significance for the local economy, said Pearl Kamer, chief economist of the Long Island Association, a business group. "There is an economic impact when you have months of negative data," she said. "It discourages consumers from spending."

The biggest restatements for Long Island occurred after March, when the U.S. Bureau of Labor Statistics, in a cost-cutting move, took over estimating and disseminating state and local data, said Martin Kohli, a Manhattan-based BLS regional economist. State job-market economists now play a lesser role in analyzing the data, he said.

The BLS news release announcing the change raised the possibility of wider data variations.

"Introduction of the new estimation procedures may result in more month-to-month variability in the estimates, particularly in smaller MSAs (metropolitan statistical areas)."

The data now have fewer "atypical adjustments," or changes to reflect unusual occurrences at the local level, said Deborah Brown, a BLS assistant regional commissioner based in Boston.

State Labor Department economists, who used to fine-tune the data to reflect, for example, the unusual seasonal high point for an industry, have less of an opportunity for such fiddling now. "We had more flexibility at the local level," said James Brown, a state Labor Department analyst for New York City.

Even before that change, experts who compile the data and analyze it say the system for collecting it is notoriously less accurate when the economy is at a turning point.

Revisions are inevitable because the monthly reports are based on samples of businesses. As more sample data come in, the numbers are revised.

Local data are revised once a year, in March. BLS said it doesn't release monthly sample sizes for local areas, but said for the annual revision it used data from 102,271 businesses for Long Island, the total number of firms based on tax records.

State and federal economist said the revisions don't look very large in terms of percentages.