The Labor Department's January employment report came out today. Economists' reactions were all over the map. And well they should be.
When the Bureau of Labor (BLS) asked firms how many people were on their payrolls, the results indicated that jobs grew by a fairly modest 111,000 in January. When its surveyors phoned households, they found that 4.6 percent of the workforce said they were unemployed, up from December's 4.5 percent.
Conclusion: based on those bare facts, the economy is growing but somewhat weakly. Hence the Wall Street Journal's headline: "Heading for a Soft Landing."
But you need to read the fine print to get the real shock! The statisticians on Massachusetts Avenue rebenchmarked their samples to an actual count of firms who are registered with the state employment offices.
There were 981,000 more jobs in December than BLS had estimated in its previous report. The monthly data were revised back almost two years. (The revision is "only" 933,000 on a seasonally adjusted basis.)
This means that the year over year employment growth in December was 1.66 percent. A month ago, we thought employment had grown only 1.37 percent. Apparently the economy was stronger than the 2006 data initially led us to believe. On Wednesday we got a surprise. The Commerce Department reported the Gross Domestic Product grew a strong 3.5 percent at a seasonally adjusted annual rate in the last quarter. In light of these revisions, it should not have been such a surprise.
Fed watchers should note that the economists who work for Dr. Bernake also read footnotes!
What is a Forecaster to Do?
This episode gives you a little insight into the real difficulties of business forecasting.
Consider the Postal Service which not only has one of the most sophisticated business forecasting systems in the world, but also shares it with the public through its rate filings. When you read the testimony of Thomas E. Thress one of the Postal Service's experts with RCF Consulting, you will find that a similar measure of employment in the forecasting equation for First-Class single piece mail. One of the attractions of this variable in forecasting and econometric work is its robustness against revision and statistical reporting quirks by the reporting agency (BLS.)
Unfortunately, the gnomes in the bowels of BLS are not cooperating.