How much does money matter?

The authors of Freakonomics argued that campaign money doesn't buy popularity nearly as much as popularity buys campaign money. In other words, candidates who have huge war chests have huge war chests because they're winning candidates. They don't become winning candidates simply because they have a lot of money. (As Mike Huffington, Steve Forbes, and Ross Perot can attest.)

Dafydd at Big Lizards links to a piece suggesting there's a campaign saturation point (CSP) above which further spending is at best ineffective, and more likely, hurts the candidate. As a candidate advertizes above the CSP, he winds up irritating voters.

I emphatically believe that every campaign in every election generates a campaign saturation point (CSP), beyond which further campaigning -- ads on TV and radio, appearances on talk shows, billboards, posters, signs, rallies, debates, GOTV, and door-knocking electioneering -- diminish, rather that augment a candidate's electoral performance. This factor should be measured in campaign density, not duration: You don't want to stop campaigning two months before the election, but you might want to throttle back on your campaigning to avoid oversaturating the market (inundating voters).

Past that point, no amount of money a campaign has on hand will help... and it can hurt a candidate badly, since there is an almost irresistable impulse for a campaign to burn through every penny it raises... even if doing so hurts rather than helps. Thus, Obama's "advantage" over McCain in campaign cash won't be as big as the raw figures naively indicate... and may not exist at all, depending where Obama's CSP lands.

So where does the CSP land? That's the hard question.

CSP is a very hard factor to measure, not least because the CSP depends upon several variables, including (a non-exhaustive list):

• The intelligence of the campaign: A smart campaign has a higher CSP than a stupid one;
• The importance of the underlying issues: If the contested issues impact the lives of ordinary voters, they will have a greater tolerance for the candidates campaigning on those issues;
• The likability of the candidate himself: Voters will be more tolerant of a candidate they like than one they dislike;
• Competing interests: If there are many other stories competing for voters' interests, they will be less tolerant of a candidate campaigning.

This phenomenon looks a bit like other invariants that show up in finance and economics, such as the finding that the revenue from a taxed population is about 20% of their income, pretty much regardless of what the income tax rate is. Another one Dafydd mentions is disaster aid.

The concept of CSP is homologous to a similar phenomenon I learned about anent reconstruction money in areas devastated by war or natural disaster: You can only pump so much money into reconstruction, an amount determined by the available infrastructure: Beyond that, money is simply flushed away. In Iraq, for example, there are only so many people available at any one time, based on skill and security, to rebuild an electrical grid or sewer lines; even if you have more money in your pocket, it won't do any good to throw it around.

Maybe campaign spending is subject to the same sort of forces that make other invariants so reluctant to vary. Once you saturate the system, further increases in the independent variable (tax rate, dollars thrown into a campaign) have no further effect on the dependent variable (tax revenue, votes).