My prejudices when it comes to campaign regulations are pretty simple: more politics is good, less is bad.
Taking that to the campaign finance question...those familiar with these debates will know that in the 1970s, the Supreme Court ruled that "money is speech" -- in particular, campaign expenditures (but not contributions) were a type of protected First Amendment speech. Matt Yglesias has a good discussion of it today, in response to a a high-profile Goo Goo speech by Russ Feingold on the general subject of money in politics.
The trick here, as far as I'm concerned, shouldn't be how to keep the wrong sorts of people out of policies; it's how to encourage more participation, and meaningful participation.
Political scientists, to be sure, disagree on all of this. But to me, there are really two important things to keep in mind. One is that campaign spending has diminishing marginal returns. What that means is that we should worry far more about underfunded campaigns than about the potential malignant effects of really large sums of money in elections (as anyone who remembers last year's gubernatorial race in California might remind us). The other is the logic of influence. Feingold is afraid that, for Democrats, taking corporate money will be corruption (he says Dems will "lose" their "soul"). I just don't think that's right. On the one hand, Democrats are going to be influenced by business even if they never take a dollar of campaign contributions, because big corporations are important, influential constituents. On the other hand, as long as no single contributor dominates, there's quite a bit of evidence that the politicians, and not the contributors, have the upper hand: contributors give because they have to in order to avoid being cut off entirely, and then pols do what they would have done anyway.
I could get into details...for example, does money buy access? Perhaps -- but does anyone really think that Wall Street couldn't get an audience with Members of Congress if only they were prevented from giving campaign contributions? I sure don't (and I think the evidence is with me on that one).
What all of this means to me is that the best campaign finance rules create floors, not ceilings -- plus meaningful disclosure. And the place where floors are badly needed are in Congressional elections. If I had my ideal system, I'd probably give House major party nominees some substantial amount of public financing ($500,000? $1/constituent? $1/voter?), come up with a similar formula for the Senate, enforce strong disclosure rules on top of that, and call it a day (actually, I'd also keep the current system for presidential elections, with the assumption that no serious nomination candidates or major party nominees will ever use it, but just in case). You want to drop $10M into a House district? Fine. You're going to waste most of it, and the opposition can run against some outsider trying to tell the locals how to vote (you better make sure opposition research won't turn up anything about you), and that's the end of the story, as far as I'm concerned.
Beyond that, I can see the case for or against allowing direct corporate spending, and I take the point Yglesias makes about allowing groups with tax advantages to contribute. Mostly, however, a lot of this is just shell games caused by restrictive rules. Loosen up the rest of the rules on who can spend, and how much they can spend, and you get rid of the constant efforts by big money to stay one step ahead of the regulators. A contest, by the way, that big money is always going to win -- but one that's a problem for the system as a whole, since it may matter quite a bit if money trails get shunted from candidates to formal party organizations and interest groups to shadow party organizations and shadow interest groups and back again over time.
So that's what I'd like to see. Floors, not ceilings, plus disclosure. More politics, not less.