No, I'm not going to go all Jonathan Swift on you. No cannibalism required, just a simple and yet radical reform of the primary process. It seems rational to me, so it won't to you. It's probably a political non-starter. But here it is...
So we have a few basic problems with the candidate selection process.
The first is the caucus. I went to the WA caucuses. It was a good experience, but it's a far cry from the turnout of a primary. While it may be helpful to select a candidate based on who the most faithful party members like, I think it's far better to get a broad cross-section of the electorate. Better by far to get people used to going into the voting booth and putting their mark next to a D.
I'm also in favor of open primaries. I doubt that there is much crossover mischief making, and including independents early is a good way to get our guys elected. It may make our general election candidate somewhat more centrist and less progressive, but I think that even that is less important, since progressive values are widely held.
The next is the primacy of Iowa and New Hampshire. Why shouldn't someone else have first shot for once? Why can't I get a presidential candidate out here in Seattle for more than an ATM call? I'd like to think that the other 48 states might like a chance at having the candidates spend the better part of a year talking specifically to their issues. On the other hand, we don't necessarily want the first primaries to be in states with expensive media markets. That makes it harder for less well-funded candidates to succeed, and may reduce the chances of some good candidates getting traction in the first primaries and using them to leverage donations to keep the campaign alive through later contests.
The last problem is the length of the process. Do we really need 6 months to choose a nominee? I know that most primary campaigns would be long dead by now, but this year shows the need to get it over with sooner. A more rapid schedule does potentially hurt the less-well-funded campaign, because there will be less time to drum up support before the next big vote. However, as more and more fundraising moves to the internet, the money bump will come faster.
So here's my idea to solve all of these problems. The root of the idea is stolen from the comments in another diary. I wish I remembered the name or the diary, but I can't recall. If you think it was yours, please let me know in the comments and I'll be glad to credit you.
First of all, we divide the states into regions. In general, the regions are intended to have somewhat similar issues, perhaps similar voting patterns, and physical proximity. In addition, the regions are intended to be roughly equivalent in electoral votes without being so large as to be unwieldy. I've started a rough list as shown below, with the total electoral votes shown for each. I've also listed the total number of 2008 D delegates for each region, not including superdelegates.
Appalachia: KY, TN, WV: 24 EV, 184 delegates
California: CA: 55 EV, 441 delegates
Grain Belt: IA, KS, MN, MO, NE, WI: 49 EV, 397 delegates
Great Lakes: IL, IN, MI, OH: 59 EV, 558 delegates
Mid-Atlantic North: NJ, NY, PA: 67 EV, 596 delegates
Mid-Atlantic South: DC, DE, MD, NC, VA: 44 EV, 394 delegates
Northern Interior West: ID, MT, ND, SD, WY: 16 EV, 109 delegates
Southern Interior West: AZ, CO, NM, NV, UT: 34 EV, 238 delegates
South Central: AR, LA, OK, TX: 56 EV, 388 delegates
Southeast: AL, FL, GA, MS, SC: 65 EV, 442 delegates
West Coast: AK, HI, OR, WA 25 EV, 209 delegates
Note: I thought about lumping the two interior West regions into one, but there were so many states that I thought it was better to have two.
The idea of a regional primary is that the candidate can focus their time and effort into a single area. This will focus the message the candidates give on a particular region (eg NAFTA issues in the Great Lakes region). It also may give a less-well-funded candidate a leg up, since the total travel in trying to reach the voters in a region will be less than reaching voters in primaries spread all over the country.
The regions would vote by turn in a regularly-scheduled progression, so that the first region to vote rotates over the years. The progression would also be set up to balance demographic divisions so that one candidate doesn't run away with a couple of early regions that favor them demographically and get so much momentum that they can't be stopped.
But what about the early states? We pick six states each year that are reasonably representative of the country as a whole and are from different regions. This list of states rotates every election as well, so everyone has a chance at glory. My only proposed restriction on which states are early states is that they must have fewer than 20 electoral votes. This limits the number of large media markets in the state, and limits the cost of campaigning. For those of you in big states, tough luck. :) Seriously, the candidates will spend time in your states, just later in the season. They'll need the votes. These early states have a rapid-fire series of primaries, two per week for three weeks, starting on the second Tuesday in January. This allows several candidates to start gaining momentum from wins, and spreads the momentum across the demographic spectrum.
The first two regional primaries would be separated by two weeks from each other and the early primaries. Then there would be a three-week gap followed by the largest states voting (CA, NJ, NY, PA). Then come the primaries for the regions that contain the early states, two per week with two weeks' separation each. This balances the power of the early states by placing the rest of their regions last. Keeping the largest states in the middle will ensure that they have some say, and will also decide the nominee in most elections.
So a typical schedule would look something like this, with Week 1 being the first week of January with a Tuesday:
Week 2: WA and NC (Yeah, WA is first. I'm biased. :) )
Week 3: AZ and ME
Week 4: MN and WV
Week 6: Northern Interior West & Southeast
Week 8: Great Lakes & South Central
Week 11: California & Mid-Atlantic North (always at this week)
Week 13: Grain Belt & Appalachia
Week 15: Southern Interior West & New England
Week 17: West Coast & Mid-Atlantic South
There, we're all wrapped up by late April. We have a nominee, and can start focusing on cutting the Republicans off at the knees.
Another reasonable schedule would have Weeks 6, 8, and 11 above combined into a Super Tuesday vote, taking place sometime between Week 7 and 11, and the remaining regions moved up to suit.
What about states that don't want to play? What if PA still wants to have their primary in April, and the heck with this system? I would suggest that any state that doesn't want to fully participate can have their primary the same week as California or later. Fully participating means taking the early primary as well as the late ones. If states try to move up too further (I'm talking to YOU, MI and FL), strip them of 90% of their delegates. They get a seat at the convention, so their voices are heard, but there's pretty serious pain in delegates and their ability to attract the candidates.