Reporters should be asking this question, or perhaps more accurately, is the Labor Department understating job loss. The issue here is the imputation of jobs for new firms not included in the survey. Since firms are constantly coming into or going out of existence, the Bureau of Labor Statistics (BLS) must incorporate changes in employment in these firms in its estimate of job growth. However, since these firms are not answering its survey of employers, it has to guess how many jobs they are creating. Of course, BLS does not literally guess, they use a model to predict job gains/losses in these firms. The model generally does a respectable job when the economy is moving along at an even pace, but it tends to miss turning points. When the economy speeds up, it understates job growth. When the economy slows, it overstates job growth. Last year, it overstated job growth by 300,000 jobs, an average of 25,000 month. This year, the problem might be even worse. For February and March, the model imputed a total of 31,000 more jobs into the BLS estimate than in the same months last year (it imputed far less for January of 2008 than January of 2007). This could mean that the true job decline in the last two months is even larger than the data imply. My candidate for the location of these phantom jobs is the restaurant sector. Including the imputed jobs, the restaurant sector added 48,000 jobs over the last three months, an increase of 0.5 percent. But, if we check the latest data from the Commerce Department, we find that nominal spending in restaurants was down by 0.6 percent from December to February. When we adjust this figure for inflation, we get that real sales in restaurants were down by 1.4 percent for these two months. Okay, so let's assume all these numbers are accurate. We have employment growing at a 2.0 percent annual rate for the first three months of 2008, while output in restaurants is declining at an 8.5 percent annual rate for the first two months of the year. This implies that productivity in the restaurant industry is falling at more than a 10 percent annual rate. Therefore, we can either believe the job growth reported for the restaurant sector, and that productivity is now plummeting, or we can believe that BLS' imputation is overstating job growth and that productivity in the restaurant sector is following a more normal pattern. It's your call.
--Dean Baker
