Barron’s is predicting Republican victories in both the House and the Senate. Why? Well, GOP candidates have raised more money than their opponents, and… that’s it. The Barron’s analysis is based entirely on fundraising data.

Ramesh Ponnuru is skeptical. So am I; the blithe dismissal of polling strikes me as particularly misguided:

Polls can be far less reliable [than following the money]. Remember, they all but declared John Kerry president on Election Day 2004.

It’s true that botched exit polls skewed to Kerry, as I chronicled here. But the polls leading up to the 2004 election were actually very useful in predicting the outcome. Bush won with 286 electoral votes; I predicted he’d get 276, and David Hogberg hit it right on the button. How did we do so well? Read the columns for yourself; while we fine-tuned based on our senses of the state of play on the ground in each state, we both started with the state-by-state polling data.