Steven Hill of the Center for Voting and Democracy has written a thought-provoking
essay at CommonDreams.org called "How to Handle Nader.
Basically Hill proposes that swing states that are under Democratic control (New Mexico, Maine, West Virginia and Tennessee) should eliminate the "spoiler" effect by implementing a runoff for Presidential electors. But to avoid having to hold two elections, he suggests they implement Instant Runoff Voting (IRV).
His essay contains a lot of food for thought; go give it a read. But I don't agree with his conclusions. My own thoughts are under the "There's more" link.
Hill briefly explains IRV:
Used in Ireland and Australia and recently adopted for city elections in San Francisco and for congressional and gubernatorial nominations by the Utah Republican Party, IRV has drawn support from Howard Dean, Jesse Jackson Jr. and John McCain. By allowing voters to rank the candidates (for example, a [Nader supporter could vote] 1 for Ralph Nader and a 2 for John Kerry), IRV can resolve the spoiler problem.
and then goes on to ask:
With one vote of the legislature and a stroke of the governor's pen, these states could accommodate the reality of the Nader candidacy. The question is: what is stopping them?
While Ralph Nader may be ready to risk a repeat of 2000--and could do much more to make multi-party democracy a viable option by highlighting reforms such as IRV--most Greens don't want to be spoilers. They consistently support reforming winner-take-all elections, and their presidential frontrunner David Cobb promises to focus this fall on safe states, in recognition of Greens' interest in defeating George Bush.
But only Democrats and Republicans have the power to change the rules of the game. Democrats' failure to use that power begs the question: would they rather engage in name-calling and suppressing candidacies, even at the risk of costing themselves the presidential election, than allow new political voices to join the fray? More people, Democrats and non-Democrats alike, should begin asking party leaders: why not IRV?
It's a fair question. Why not IRV? Are Democrats really willing to hand the election to Bush just to avoid the embarrassment of Nader garnering a disconcertingly large share of first-choice votes?
With all due respect to Steven Hill, I don't think so. It's true that IRV solves the "spoiler" problem without risking a second, runoff, election, but IRV has problems of its own.
First, IRV is going to require quite a bit of voter re-education. Voters in the US have been marking ballots the "old-fashioned" way for a long time; changing the rules to allow them to rank candidates instead of merely marking their first (and only) choice isn't going to be easy, especially with less than four months to go until the election. I'd expect massive confusion on Election Day if IRV is implemented for this election at this late date.
Second, our media infrastructure isn't set up to handle IRV. The temptation would be to report the candidate with the most first-choice votes as the "winner," but that's wrong! To determine the winner in IRV, you have to add in the second choices of eliminated candidates like Nader! A clueless and hostile media would probably blow their misunderstanding of the process up into the "scandal" of the "stolen election of 2004" (unlike the actual stolen election of 2000)!
Finally, most voting equipment isn't designed to handle IRV. Optical-scan ballots would have to be completely redesigned. Punch-card ballots could probably be made to work, but the machines that count them would have to be reprogrammed. And lever machines are obviously hopeless.
The only types of ballots that could easily deal with a ranked-candidate format are old-fashioned paper ballots and touchscreen voting machines. IRV with paper ballots would be a logistical nightmare, and touchscreen machines raise their own well-known concerns.
So I can't see any of these states implementing IRV for 2004. A "traditional" runoff is a better approach, at least in the short term, but it has problems of its own: e.g., the schedule for both elections would have to be well-publicized, to ensure good turnout in the likely event both the first election and the runoff have to be held.
Fortunately, there's a better alternative: acceptability voting.
Acceptability voting is simple: it works just like our current system, except overvoting is allowed. If you want to show Ralph Nader (or, more likely, David Cobb) some love, but don't want to risk a Bush win, you can just vote for Kerry, Cobb, and Nader! You don't need to "rank" them; just vote for every candidate you find acceptable as President (hence the name). The winner is simply the candidate with the most votes, as with the current system.
Absent the need to rank candidates, most current ballots would work just fine, with only minimal tweaking to the ballot-counting machinery. Lever machines might still be a problem, since they're designed to "lock out" overvotes; but unlike IRV, one can at least conceive of a lever machine being used in acceptability voting. And the media can report the results and project the winners just as they do now: the person with the most votes wins.
Acceptability voting still leaves the voter re-education problem, but at least an un-reeducated voter would still cast a valid ballot for a single candidate, so (s)he wouldn't be disenfranchised.
So my question to both the states and Steven Hill is, why not acceptability voting?