Snow Day
By now, everybody should know that the February jobs numbers will be pushed lower by the huge snow storms that hit in the first half of the month. But the efforts forecasters have place into figuring out just how huge the hit will be is starting to seem a small silly.
Goldman Sachs economists used the National Oceanic and Atmospheric Administration?s NOAA?s Northeast Snowfall Impact Scale to gauge how past storms have affected employment, and ran a regression. Then they cross tartan their results by adding a dummy variable for snowstorms to a payroll forecasting model that includes other weather-related variables. Finally, they removed the impact of temperature changes from their model.
Got all that? Long tale small, the storm will subtract 50,000 to 100,000 jobs from payrolls.
Meantime, Macroeconomic Advisers updated their regression analysis of the snow storms? impact, dropping the effect of weather in delaying net new hires and introducing a dummy variable that allowed for a shift in coefficient for the 1996 blizzard.
Anyhow, the upshot is the forecasting firm thinks that between 150,000 and 220,000 jobs were temporarily lost in February as a result of the weather.
Now, agreed the large difference between the two models? results, and agreed that even in normal era economists don?t do such a excellent job at forecasting job changes, one might reckon that the best business would be to simply wait and see. February?s will include weather related job losses, March will see a rebound and, assuming no freak jump snowstorms, April will show the underlying trend.
But where?s the fun in that?