Retail Sales and the Weather

Parija Kavilanz is excited about retail sales reports:

A surprising rebound in February sales gave retailers a much-needed respite after a very difficult winter sales season that had pointed convincingly to a pullback in consumer spending.
"It's very interesting that consumers are actually showing some signs of life," said Ken Perkins, president of sales tracking firm Retail Metrics.
Perkins said part of last month's sales strength, which came on the heels of broad-based softness in December and January, was in part because of "pent up demand."
"The sales numbers were just so weak in December and January that you almost had to expect consumers would come back at some point," he said.

Which prompted Glen Lineberry to email me:

Does anyone ever factor weather into an analysis of consumer spending?
This is admittedly purely anecdotal, but there were major storms across the midwest and northeast in both December and January. If I lived in Cleveland and it was zero degrees out, I'd skip going to Best Buy to check out the new flat screens, and I suspect most people are the same way.
Just curious if anyone out there compares sales stats to average temperatures or snowfall or such.

Which would give a whole new meaning to the term "seasonally adjusted". Retail sales in February are always much lower than retail sales in December, which is why the numbers are "seasonally adjusted", or else looked at only on a year-on-year basis. What Lineberry is asking for is some way of adjusting the winter months' sales volumes somehow in order to account for how cold it was outside. In order to calibrate that, you'd need a long history of weather reports - we have that - as well as an equally long history of same-store retail-sales reports - which would be much harder to find. I can't say I'd find it easy to have much faith in any such calibration.

Instead, then, we have the annual ritual of retailers blaming poor sales on the weather. Which is almost comforting, in its own way.