The U.S. monthly employment and unemployment totals have been in the news in recent months, mostly for startling declines in the former. The employment numbers are compiled by state departments of labor, based on surveys of businesses, and then aggregated by the federal Bureau of Labor Statistics. They are based on the unemployment insurance tax records, but that information doesn’t come in until long after the surveys, when all the taxes are paid. Every March the past survey data is “rebenchmarked” back to what it should have been based on the unemployment insurance tax data that had arrived afterward. It is very difficult for the survey to completely identify a big short run changes in the economy, not the least because those doing the surveying can’t know how many new businesses have started recently, and are thus not included in the list of businesses to be surveyed. This year’s rebenchmarking, as shown by the attached spreadsheet, produced some stunning revisions.
The data in the spreadsheet is for major metro areas, in state order. Looking from the top, BLS preliminary data (the “p” in the table) had showed a loss of 86,800 jobs (4.5%) for metropolitan Phoenix in 2008, which is bad enough. With more updated information, however, it appears the MSA actually lost 117,600 (6.1%). That is like happened to New York City in 1991. Los Angeles County was thought to have lost 41,500 jobs (1.0%) in the year to December, then found to have lost 151,500 (3.6%), an 110,000-job negative surprise. The worst surprise on a percentage basis was in California’s Inland Empire, the desert counties east of the mountains from coastal southern California that boomed when more established areas became unaffordable. The previously expected 3.0% employment decline (38,500) was bad enough, but “after further review,” as they say in the NFL, the loss was put at 6.6% (84,300). Orange County got clobbered in the rebenchmarking a year ago, as its big subprime employers went down, and got clobbered again this year, as a loss of 41,400 jobs (2.7%) morphed into a loss of 61,600 (4.0%). Also out west, Las Vegas went from a loss of 15,500 jobs (1.7%) to a loss of 36,600 (3.9%). And Seattle swung from a previous estimate of a gain of 6,200 jobs (0.4%) to a new estimate of a loss of 23,500 (1.6%).
Elsewhere in the country, Detroit and Florida faced huge employment declines of 4.0% to 6.0%, roundly speaking, that had previously been expected to be somewhat less terrible. In Memphis, the previous estimate of 15,700 jobs lost (2.4%) was surprisingly right on the nose.
Just about the only place employment went up from December 2007 to December 2008 is in Texas metropolitan areas, but even there it went up a lot less than had been expected when the numbers were first compiled. One wonders if those gains will turn to losses after further rebenchmarking next year. Austin gained about the same number of jobs (9,900 or 1.3%) as had previously been estimated, and that is the highest percentage gain among the Texas metros included.
In general, in fact, state capitals (like the federal District of Columbia) seemed to be doing better than major metro areas for most of last year. For example, Raleigh-Cary NC lost 2.1% of its jobs (11,100) December to December while Charlotte lost 3.6% (31,900). But fiscal crises are bound to arrive in state capitals sooner or later. I wouldn’t expect 2009 to be a good year for Sacramento, capital of bankrupt California.
There were some surprising surprises on the Northeast Corridor. In the Washington area, both the District of Columbia and the suburbs had been thought to have gained a handful of jobs from December 2007 to December 2008. Updated information shows a gain of 5,900 (0.8%) in the District, better than previously estimated, and a loss of 19,700 (0.8%) in the suburbs. One finds a similar pattern in the Baltimore area, in which the troubled City of lost 500 jobs more than previously estimated (for a total of 3,300 or 0.9%) but the suburbs lost 11,800 more than previously estimated (for a total of 21,600 or 2.2%). On the Pennsylvania side of metropolitan Philadelphia, the City of lost 4,800 jobs (0.7%), or 2,400 less than previously expected, while the Pennsylvania suburbs lost 48,700 jobs (1.2%), also less than expected but a higher percentage loss than in the City of Philadelphia. Across the river in Southern New Jersey, the loss of 16,500 (3.0%) was worse than on the Pennsylvania side, and 9,400 worse that previously expected.
Some of the out-performance of central cities may be because older cities have less left to lose, with the Health Care and Social Assistance sector (the only one growing) accounting for a large share of the remaining jobs. Housing construction, retail stores, and real estate and (often subprime) mortgage brokers accounted for a higher share of suburban jobs in many places, and these were it hard in 2007 and 2008 — one reason San Francisco’s West Bay lost just 1.9% of it jobs while Oakland-East Bay lost 4.0% (42,400), a pattern repeated by greater losses in Orange County than in Los Angeles. On the other hand, younger generations do seem to favor urban locations, older suburban housing, office buildings and shopping centers are reaching the age at which similar urban areas went into decline in the 1960s and 1970s, and the soaring gas prices of a year ago might have influenced location on the margin. We’ll have to see what patterns this real estate bust produces.
Preliminary data had both New York City and the surrounding suburbs losing 1.4% of their jobs from December 2007 to December 2008, with 53,500 jobs lost in the city and 66,700 lost in the suburbs. But according to rebenchmarked data New York City lost 7,800 fewer jobs, for an newly estimated decrease of 1.2% (45,800) while the suburbs lost 39,900 more jobs, for a decrease of 2.2% (106,600).
It should be noted that the reason I did this little exercise is because I wanted to see what the downward revision was for New York City, but got an upward revision instead. It’s always interesting to be surprised, especially by a good surprise. But I’m not in the mood to celebrate yet. New York City’s economy turned down much later than the rest of the country, and many of the gut wrenching changes that have already occurred elsewhere — including a massive reduction in housing values back to some relation to income — have just gotten started and mostly have yet to occur here. New York City’s 2010 rebenchmarking could thus be like other metro’s 2009 rebenchmarking, something to keep in mind this year. How many new businesses are starting? Is it more or less than the models of the New York State Department of Labor would predict?
I received third quarter 2008 employment and wages data, which is based on the actual unemployment insurance tax records, from the New York State Department of Labor on March 20th. The state had sent me the rebenchmarked data for New York on March 5th. That might mean the data was rebenchmarked based on the hard third quarter numbers from July to September 2008, with the collapse that started in September barely recorded. Or it might mean that only second quarter (April to June) data was available for rebenchmarking, meaning the subsequent collapse wasn’t included at all. Yikes! Wait until next year.