For the record, I'm not a big advocate of public financing of elections. I'm not opposed to it, but I just don't think it will make our politics less corrupt (although it might look less corrupt), and I think the current system has quite a bit more virtue to it than is generally recognized. I figure that just letting anyone donate to anyone is fine, as long as there's rapid and public accounting of it. True scandals will out.
That said, one concern I do have with the current private system of campaign donations is the amount of time that officeholders need to devote to fundraising. The time spent hanging out with annoying rich people begging them for money is time not spent learning about issues, raising children, or talking with regular voters.
Luckily for me, I had the chance to discuss a paper right along these lines last week at the Midwest Political Science Association. Michael Miller looks at state legislative candidates in Arizona, Connecticut, and Maine, where candidates can opt in to a full public financing of their campaigns. He then uses a matching algorithm to compare these candidates with other similar ones who do not (or can not) use public financing. The differences are pretty striking. Miller finds that candidates who use full public financing devote 11 percent more time to door-to-door campaigning than those who must fundraise. The implication is that public financing gives candidates the opportunity to do things we consider not only good for their campaigns but also good for democracy.