Of Polls and Predictions

It appears that one of my appointed tasks this year is to cheer up discouraged Democrats. Ruy Teixeira has effectively rebutted the Gallup polls, but some folks might also be concerned about the predictions of political scientists at this year's APSA (The Gadflyer mentioned this a few days ago, but I can't find the link). The models indicated that Bush should receive about 54% of the vote. So we can expect Bush to cruise to victory, right?

Wrong. These models are principally economic models: if the economy is growing, the incumbent usually wins. Since we have had GDP growth over the last year, these models predictably forecast a Bush victory. Unfortunately for these models, however, the nature of economic growth has changed a great deal in the last few years. Remember, many of these same models predicted a landslide Gore victory. The problem is a de-coupling of overall economic growth and median incomes. To put it in laymen's terms, the rich people are eating the whole pie. Any income gains are being concentrated at the top, therefore the economy may seem to be doing well even while the middle class continues to suffer. There is an interesting article in the American Prospect on this very point. This problem bedevilled the forecasters last time, when the economy was doing fine measured as GDP growth. But when you looked at median income, the 2000 economy wasn't as swell.

So don't take these forecasting models too seriously. They've been wrong before, and for very much the same reason.