Sigh. One step forward...

Much excitement on liberal campaign-focused blogs over the last few days about the DCCC possibly going millions into debt to take full advantage of this "historic" opportunity. As Jonathan Singer writes, though:

That said, if the committee is going to the bank to borrow money for these second and third tier races, the money should go to them -- not contests in which both parties have already sunk hundreds of thousands, if not millions of dollars already. (I'm looking at you, Illinois 6!) In races in which the price tag for each campaign has already topped one or two million dollars, the next $50,000 or $100,000 isn't going to have the same type of impact it would in an emerging race that has thus far been less costly. Even $500,000 for more television ads in a multi-multi-million dollar race could not possibly have the same effect that that same amount of money would have on one of the less-watched races.

So, what is the DCCC doing?

Yesterday the DCCC pumped a staggering $11,859,818.68 into 32 races in 17 states, new filings with the FEC show. ...

Notably, almost half of the cash — $5,042,707.45 — was spent on three House races in Pennsylvania alone, on ads targeting incumbents Jim Gerlach, Curt Weldon and Mike Fitzpatrick.

It's worth returning to the logic behind Jonathan Singer's unfulfulled hope. From an article by Jonathan Krasno and Donald Green, excerpted at Ruy Teixera's site:

Because of diminishing returns, we know that a large investment in an expensive race will bring few votes, while a small investment in a cheaper race may bring many. Parties shy away from the latter on the grounds that hopeless candidates are hopeless causes. But the math says different. Suppose that we could increase the odds of twenty candidates from 5 to 10 percent for the same cost of helping two candidates with 45 percent chances get to 50 percent. By helping the twenty hapless candidates, we would increase the expected number of victories from 20 x 0.05 = 1 to 20 x 0.10 = 2. By helping the well-heeled candidates, we would increase the expected number of victories from 2 x .45 = 0.90 to 2 x .50 = 1. The first investment portfolio has an expected return of 1 additional victory, while the second one is just one-tenth of an additional victory.

The DCCC is stuck in the same type of conventional logic that has NFL coaches punting on fourth down too often: Every play carries some nontrivial risk of costing you the game. If you make the play conventionally perceived to be riskier, even if you can show mathematically that it's less risky, and you lose because of it, you will be blamed. But nobody's going to fire you for sticking with conventional wisdom, even if you lose.

This logic doesn't only apply to elections and football games, of course.