(March 7, 2008) – Today’s jobs report sucks. Nonfarm payroll employment decreased by 63,000 jobs and the unemployment rate remained essentially unchanged at 4.8% in February, according to The Employment Situation: February 2008, released this morning by the Department of Labor. This marks the second month in a row of falling employment; jobs fell by 22K in January, not a good sign for first quarter GDP Growth.
The chart shows now jobs created each month over the past 10 years. As you can see, the trend in job creation is down, although there is a long way to go before things get as bad as they were in 2001, when GDP shrank for 2 quarters.
Government jobs increased by 38K in January, which means that private sector jobs actually fell by 101K. Goods producing jobs fell by 89K, 39K in construction (shocking as that may be) and 52K in manufacturing. The big surprise was in the service sector, which usually carries the load for rising employment, where private-sector service jobs fell by 10K. Hours worked were flat for the month.
The household survey numbers make no sense to me at all. At face value they say that 450K people decided to leave the labor force during the month of February. 255K of them lost their jobs and decided not to look for work; the other 195K people were apparently looking for work in January but stopped looking in February. Those numbers don’t square up with the 644K additional people they (separately) report as “not in labor force,” unless there was a sudden increase in population during the month.
So where is the slip in the numbers? It may partly be in two strange seasonal adjustment procedures. First, the data collectors don’t actually seasonally adjust the total job numbers. They seasonally adjust each of the major components (young people, women,…), then add them together. Second, in the household survey, they re-estimate the seasonal adjustment factors every month but only report the new month using the new seasonal adjustment factors, i.e., the prior month (in this case December) was calculated using an entirely different set of seasonal adjustment factors. For practical purposes, this means that a comparison of a current month with the prior month–all the numbers in the previous paragraph) means absolutely nothing.
Finally, the estimates need to be taken with a big grain of salt. Although the data are collected from a large number of people (160K businesses and 400K total establishments in the payroll survey; 60K households in the household survey) the margins of error are huge. The 90 percent confidence interval for the payroll survey is +/- 104K; for the household survey it is +/- 430K!
So don’t get too excited about one month’s numbers. Still, the trend in employment seems to be down for the time being.
My final point–why are the job numbers holding up pretty well even though home prices, stock prices, and bond prices are falling? It’s because so far we have been lucky. The sharp restriction of credit that began with leverage loans last summer, then spread to mortgages in the fall, has so far not caused banks to stop making working capital loans to the private, mom and pop companies that make up more than half of GDP and provide more than 100% of all new jobs. It’s these credit lines that are the life blood of GDP. That’s where we should focus our attention.