Meager job growth in September sends a clear signal of a slowing economy, right? It ain't necessarily so
October 6, 2006
Once again we are reminded that economics is not an exact science. The employment report released on the morning of Oct. 6 by the Bureau of Labor Statistics does little to clarify the true state of the economy, no matter what the Wall Street economists and Washington politicians try to tell you. If anything, the report's true message is how little we know about what's going on.
On the surface, the story looks like a simple case of a slowing economy. The number of jobs went up by only 51,000 in September. Leaving aside the Hurricane Katrina months of last year, that's the lowest monthly job growth since mid-2004. Great news for the Democrats to trumpet going into an election.
But wait! The same report also showed the number of employed workers rose by 271,000, which helped push the unemployment rate down from 4.7% to 4.6%. Great news for the Republicans to trumpet going into an election.
But wait! How could the increase in jobs be different than the increase in employed workers? The short answer is that the jobs number is based on a survey of companies, while the data on employed and unemployed workers is based on a survey of households. Very often these two numbers diverge, on a month-by-month basis and even longer.
But wait! It gets even worse. Usually economists are more willing to trust the jobs numbers, because it is based on a broader sample than the household numbers. The survey of households includes only 60,000 households, which is only a small fraction of the U.S. population.
WHICH IS MORE RELIABLE? But wait! Buried deep in today's report is a bit of a nugget. As it does every year, the BLS counted all the jobs reported to state unemployment insurance offices in March—the so-called "benchmark" month. This year they announced that they found an extra 800,000 jobs in March, 2006, that they didn't know about—and they have no idea why.
This benchmark revision for March, 2006, has no direct impact on the change in jobs in the latest month. But it does raise questions about the reliability of the business survey compared to the household survey. Before the revision, the jobs data showed a gain of 2 million from March, 2005, to March, 2006, while the household survey showed an increase of 3 million. The revision, however, moves the jobs data much closer to the household survey.
Does this mean we should trust the good news from the household survey more than the bad news from the jobs report? Did the economy add 271,000 jobs in September, or 51,000? Should the Federal Reserve raise rates to counteract inflationary pressures from a falling unemployment rate? Or should the Fed cut rates to head off a stalling labor market or an imminent economic slowdown? All good questions, no good answers—at least for now.
Email This Page!