Let me say publicly that DonBoy’s answer exudes a combination of intuitive genius and confidence that make me think DonBoy is going to do big things in his life. -- Steven D. Levitt (Freakonomics blog)
Sunday, August 01, 2004
First of all -- see the Electoral Vote Predictor graphic over on the right-hand side? That's the new feature from the data miners at electoral-vote.com. It gives results at the time this page is viewed, so even if you come back here through a permalink it'll show you the "now" value, not the August 1 value.
Also from that site is this explanation of a proposed ballot referendum in the state of Colorado:
A group of Colorado citizens have proposed a change to the state's constitution specifying that Colorado's nine electors be apportioned strictly in proportion to the popular vote. Currently Bush is ahead 48% to 43% there, so under the proposed system, Bush would get five electoral votes and Kerry four electoral votes, instead of nine to zero. The group has turned in petitions containing 130,000 signatures. If about 68,000 of these prove to be valid, the question will be a ballot referendum in November. If it passes, the change takes effect for this year's election. If it makes the ballot, on the evening of Nov. 2, the TV news anchors will probably be saying: "President Bush won Colorado with 55% of the vote, but we don't know how many votes he will get in the electoral college until they finish recounting the closely fought referendum on changing the Colorado state constitution." Whoever loses will appeal to the U.S. Supreme Court, which once again may have to rule on the sensitive issue of state's rights. To learn more about what may be the sleeper issue of the year, start here.This leads us into some pretty strange territory. Someday I'll write up my mostly-partisan reasons for opposition to the Electoral College, but even I think that changing the system all-but-retroactively can't serve any democratic interest. More interestingly, in a world of perfect information and perfect self-interest -- classic game-theoretic conditions, in other words -- this can never pass. Colorado went rather solidly for Bush last time, 51/42/5; as mentioned above, current (well, June 18) polling has it similarly, Bush by 48/43/3. Assuming a known frontrunner, on Election Day, the supporters of the frontrunner will vote against the rule change, and the other side will vote in favor. There are more voters in favor of the frontrunner (duh), therefore, it loses.
There are other considerations, though; for a start, there's the Nader voters. You can imagine a situation where it's 49/44/5, and the Nader voters -- let's imagine they're not literally trying to get Bush elected -- figure they get two bites at the apple. They can vote for Nader, thus splitting the vote enough to give Bush the state; but they can also combine with the Kerry voters to change the rule, thus giving Kerry 4 of those 9 EVs anyway. With different motivations for the Nader voters, of course, you'd get different results.
Another possibility, at least in theory, is for each side to try to trick the other side's voters about the state of the race. If the R's can fool the D's into thinking that the D's are winning, the D's will vote against the change; the same applies in reverse. The challenge here is in tricking only the other side; you can "leak" a bogus poll, but everyone will hear about it. This tactic therefore seems better suited to logic puzzles involving tribes who always lie or always tell the truth, rather than 21st-century elections.
Finally, we have that most subtle of categories: the principled voter, whose opinions on the referendum are not constrained by mere partisan gain. This is fantasy and will not be considered. However, if the race were to become close enough by Nov. 2, voters might find themselves willing to pretend they were in this class.
UPDATE: Luis Toro at The American Street has more on this, and better.