The post-debacle analyses have fingered many possible problems with the Democrats presidential campaign, one of which was John Kerry himself. I don't think there's a lot of value picking over Kerry's bones here, but this reminds me of a basic question, which is how the hell did he get the nomination in the first place? Basically, he won tough races in Iowa, with their rather odd caucus system, and again in his neighbor state of New Hampshire, after which he was the front-runner, coasting with decreasing opposition, losing only South Carolina and Oklahoma. As such, he's one more example of a pretty firm rule, which is that after Iowa and New Hampshire, the candidate with the most money left wins the nomination. George W. Bush, as you should remember, actually lost New Hampshire in 2000, but had so much money left over that McCain couldn't compete in the rapid-fire following primaries. (New Hampshire actually rarely picks the nominees, but it weeds out a lot of wannabes.)

On the surface of it this is a really stupid way to select a party nominee. New Hampshire is less representative of America as a whole than any other state (possibly excepting Alaska, Hawaii, and maybe Wyoming): it is basically a tax haven which runs its pathetic state government by selling discount hootch to its neighbors. Iowa is a bit less skewed, but it's a very WASPish, New England-settled farm state, not all that much of a contrast. Candidates spend a year or more microcompeting in those two forums, after which the primaries come so fast and furious, and all you get at that point is a count of money for TV advertising and anticipatory piling-on by the local politicos: everyone wants to get in good with the winner early. In other words, we spend a year or so with no votes (nothing but PR), then get a rush of votes that decides the nomination in a few weeks, then we wait 3-5 months for the convention, then 2-3 months for the election.

What I want to propose here is a complete revamping of the primary system. We want more people to vote. We want to compress the campaign season so it's worth following. And we want to limit the decisive effect (and expense) of individual money-raising. I'll get back to the latter point after we go through the proposed structure. What I propose here is to have three rounds of primaries, each separated by three weeks. Each primary round takes place in 1/3 of the states (plus the District of Columbia, so 17 each round). The first round covers the 17 smallest states (so New Hampshire still gets first crack, but not uniquely). The second round is in the middle 17 states, and the third round is in the 17 largest states. The first round is tiny percentage wise (7%) but diverse geographically, and straddles the red/blue divide (10 red, 7 blue). The first round has intrinsic interest as the first primaries. It starts to weed out the failures, but the broad distribution offers a lot of niches for candidates. The middle round, with 23% of the population, is also diverse but introduces most of the South, another niche (or two). But the third round, with 70% of the population, remains decisive. The three-week period between the rounds allows for adjustments without dragging on too long. Regional splits in the first two rounds seem very likely. If this continues in the third round you actually get a convention that decides the nominee, where people can look back at data from the whole nation.

The schedule looks like this:

• Round 1 (E - 15 weeks): Alaska, Delaware, District of Columbia, Hawaii, Idaho, Maine, Montana, Nebraska, Nevada, New Hampshire, New Mexico, North Dakota, Rhode Island, South Dakota, Vermont, West Virginia, Wyoming; total population 18,785,867 (6.7%).
• Round 2 (E - 12 weeks): Alabama, Arizona, Arkansas, Colorado, Connecticut, Iowa, Kansas, Kentucky, Louisiana, Maryland, Minnesota, Mississippi, Oklahoma, Oregon, South Carolina, Utah, Wisconsin; total population 65,624,977 (23.3%).
• Round 3 (E - 9 weeks): California, Florida, Georgia, Illinois, Indiana, Massachusetts, Michigan, Missouri, New Jersey, New York, North Carolina, Ohio, Pennsylvania, Tennessee, Texas, Virginia, Washington; total population 197,011,062 (70.0%).
• Convention (E - 6 weeks): Have both conventions the same week, in different cities. Don't give either party a bounce advantage, and make them compete for air time.

One more change is that voters should be able to vote for more than one candidate. This does two things: it keeps equivalent candidates from splitting the vote and cancelling each other out, and it lets fringe candidates register support without voters throwing their votes away or practicing lesser-evilism. Same thing should happen in the general election, which would encourage third parties without the sort of recriminations we've seen viz. Ralph Nader.

The way I see the money working is that a high percentage of the money raised goes into a common pool. Individual candidates can raise up to a fixed budget to cover their operating expenses, but not TV advertising, "ground game," stuff like that. Campaigning is basically done through common forums: debates, joint rallies, chunks of TV and radio time (e.g., a 30-minute chunk where each candidate gets a slice to make a presentation), and publications (e.g., a pamphlet where each candidate gets a few pages, there's some side-by-side comps, and maybe some generally agreed up background info; the whole thing is then broadly distributed by the party). The party manages these things and ensures fair representation for all candidates. The idea here is to get a lot of information out to voters, to get a lot of exposure for the candidates, to organize volunteer help and resources in a way that carries the party through past the convention, and to not waste a lot of money doing all this.

I think you get three main things out of this: a lot more people participating in the selection of the nominee; a much more battle-tested nominee; an advantage for campaign skills over moneyraising skills. I'm generally leery of structural solutions, which are often what bad managers do to make the same things merely look different, but the current structure is skewed to the point of ridiculousness.

One thing I wonder about is whether the Democrats could do something like this on their own. If they did, I think that setting an example of how campaigns can be waged with minimal money would be a good lesson for all of us, even if the final stretch against the Republicans is a spend-a-thon. I also think that putting more information out much more broadly is essential to inoculate us against the next round of Rove lies and distortions and innuendos. Getting better candidates is important; getting better voters is essential.