"State-Level House Prices Uncorrelated With Polling"
Yesterday, I had some doubts that Greg Mankiw’s factoid about Obama leading in states where home prices have declined really “explained the election” (as Marc Ambinder said) or, indeed, explained much of anything at all. I had the argument, but David Shor’s got the regression and, indeed, there’s nothing here:
Since nothing tops off a Sunday night like a regression, I ran an analysis using the house-price change data Professor Manikiw provided, and as Matt suspected, it was very insignificant( Rejected at 5, 10, 15, 20, and 40% confidence).
He’s promising a county-level examination of the issue once the voting data is in after the election.