After reading the reporting on the January job numbers, it seems that not many economists pay attention to the weather. If they did, they might have been somewhat more pessimistic on the January employment numbers. By itself, 111,000 is not a bad performance, especially when it follows two months in which the economy added more than 400,000 jobs. But, it seems likely that the January figure was inflated by extraordinary good weather across the Northeast and Midwest over the reference period. (The survey asks the number of employees on the payroll for the pay period that includes January 12th.) While we always look at seasonally adjusted data to assess the economy, there actually are very large fluctuations in employment and output associated with the weather. Typically, bad winter weather in the Northeast and Midwest leads to large falloffs in employment in some sectors like construction and restaurant employment. However, the seasonal adjustment factors offset this effect. That is why a loss of 295,000 jobs in construction was reported as a gain of 22,000 in the seasonally adjusted data. Similarly, a loss of 196,000 jobs in restaurant employment was translated into a gain of 21,000 in the seasonally adjusted data. These seasonal adjustments are essential to making sense of our data, but what happens when people don�t get laid off because a construction site is shut for two weeks due to heavy snow? Or, what happens when restaurants continue hiring because it is not too cold for people to go out for dinner in early January? In these cases, our seasonal adjustment factors would be overcorrecting. This is likely what happened this January. How big is this effect? We�ll find out when we get the February data. [Late addition: The final version of the NYT article included comments about the weather from Goldman Sachs economist Jan Hatzius. Of course, you can always get the real story about the state of the economy from CEPR data bytes.
--Dean Baker
