- Donald Saari
With less than three weeks before the election, all eyes are on the polls. But there's a high likelihood the voting results won't reflect the will of the people, according to Don Saari, a professor of economics and mathematics at the University of California-Irvine.
The problem isn't electronic voting machines or misinformation campaigns. Saari says the most nefarious form of voter disenfranchisement is the very system we use to make our choices: plurality voting. Saari recently spoke at N.C. State University to discuss the ideas in his book, Chaotic Elections! A Mathematician Looks at Voting, and to propose what he says is a better system.
You're critical of the voting system we use.
It's horrible. When I started examining these things, I was quite upset to find out that the plurality vote, the system we will be using on Nov. 4, is the system least likely to represent the views of the voters. In fact, in closely contested three-candidate elections, you have to expect that 70 percent of the time, something will go wrong.
Let me give you an example.
Suppose we're going to have a party, and a committee of 15 is going to select which beverage we're going to have. Suppose six of them are health nuts, and they choose milk; wine is second place, and beer is last place. The next five come from Wisconsin, which means beer is first place, wine is second place, and milk is last place. The last four come from California, so wine is in first place, beer second and milk is third.
When they vote, milk gets six votes, beer gets five and wine gets four.
If you check, you'll see that nine of these 15 people, 60 percent of them, prefer last-place wine to first-place milk. Those are landslide proportions. If you compare second-place beer to first-place milk, you see the same thing. Milk is in first place, but these same voters strongly prefer any other beverage to milk. And if you compare beer to wine, you'll see that two-thirds of these voters prefer wine to beer. Wine came in last place, and yet it is these voters' beverage of choice.
Real-world examples of this problem are all over the place.
In Louisiana in 1991, Buddy Roemer was running for re-election as governor. He was a reasonably popular governor who made some controversial decisions. But he was running against Edwin Edwards, a former governor who had been disgraced with alleged illegal activities during his time in office, and David Duke of the Ku Klux Klan. Roemer could have beaten either of them in a head-to-head election, but because of the plurality vote, he came in last—like wine. So we had a crook-or-Klan runoff election; the crook won. By the way, Edwards is in jail now.
In the Florida 2000 election, for example, if we'd had any method that took second-place votes, the outcome would have changed and Gore would have won.
We look at these things, and we say, 'What happened?' The pundits say, 'Maybe people are changing their opinions.' No, that isn't it. Our voting system doesn't reflect what the voters think.
What system would best reflect voters' views?
This is a mathematical problem that took well over 200 years to be resolved. I solved it within the last several years. It turns out that the system that most accurately reflects the views of the voters is the Borda count.
In 1770, a French mathematician by the name of Jean-Charles de Borda did not like the voting system used to elect candidates to the French Academy of Sciences. Because of the plurality vote, they were getting mediocre choices. He proposed this system: If you have three candidates, you give two points to your first, one to your second and zero to your third.
North Carolina recently tried a pilot program of instant-runoff voting in some local elections. How would that address this problem?
The trouble with instant-runoff is that the first stage is flawed, because the first stage uses the plurality vote.
In the beverage example, wine will get dropped in the first stage of instant-runoff, and between beer and milk, beer will win. If we use the Louisiana example, Roemer would have been re-elected with the Borda count. With instant-runoff, we still end up with Edwards, the crook.
Suppose the local school system said, 'From now on, all students are going to be ranked according to the number of A's they receive.' Sounds good, until you realize that the student who got an A in gym and an F in everything else will be ranked above the student that got solid Bs. That is exactly what happens with the plurality vote. If you just do it in terms of A's and ignore everything else, the two people who might jump to the top might have a lot of F's.
Have you tried to advocate for changing the voting system?
Not beyond giving public lectures and informing people. I'm a theoretician, a mathematician. I get criticized for not [advocating]. My response is: If people believe very seriously in what I'm saying, they should be out there.
As far as changing the voting system, that's trivial. The machinery is there for instant-runoff [voting] or the Borda count. We do have the capability.
Everyone's watching the polls closely now. Do the polls give us an accurate picture of the will of the people?
Polls are indicators for where to put your resources if you're a candidate, but I worry about polls. Suppose I'm a McCain supporter; I may be discouraged from going to vote. Suppose I'm an Obama supporter; I may be overly confident and not vote. Voting is a right and a responsibility. It's also an emotional type thing. We may be lulled into a sense of security or frustration, and people won't vote. That bothers me.