For immediate release September 16, 2009For information: Alice Tibbetts, 651-399-7329
Donald Saari, author of the book, Disposing Dictators, Demystifying Voting Paradoxes, will speak at the University of Minnesota on Tuesday, September 22 at 7 p.m. in 125 Willey Hall, 225 19th Avenue South. His talk is titled "Chaotic Elections; Why don’t voters elect whom we really want?" He is a distinguished professor of mathematics and economics at the University of California-Irvine and the first speaker in the Institute of Mathematics public lecture series.
"The plurality vote, as used here and in many countries, is a terrible way to determine an election," he says, "because a voter cannot register who they prefer in second or third place.” Whenever there are three or more candidates in a race, this election tool is flawed. Jesse Ventura, for instance, would not have won the three-way vote in the 1998 governor’s race if voting allowed a way to rank candidates. Similarly, the close tally for the Coleman/Franken race may not have happened with a better voting system." In his lecture, he also discusses why instant run-off voting is not a solution.
Saari proved through mathematics that a 300-year-old tool called the Borda rule provides a more accurate voting system. It allows voters to rank candidates by number: 2 for the voter's first choice, 1 for second choice, and 0 for last. He compares the current system of win/lose elections with ranking students only by the number of A’s they receive in college.
"Let's say the University of Minnesota decided to rank students this way. That means a student with all Bs is ranked below the student with one A and Fs for all other grades. This system throws away information needed to determine the quality of students' performances. The same thing happens with plurality voting."
He cites the 2002 election in France as an example in which the voter's presumed choice may not reflect reality. Voters appeared to rank Jean Marie Le Pen as second choice in the first round of voting for president. His success created worldwide concern about whether French voters supported right wing candidates. “In fact, this outcome reflected the serious flaw in the plurality vote; only about 18 percent of the voters supported Le Pen; the other 80 percent ranked him at, or near, the bottom,” Saari said.
Instant run-off voting (IRV, also called ranked choice voting or RCV) is not a much better solution, he says. "For me, IRV is like trying to cure a serious disease by masking the symptoms. Yes, at times, IRV provides a more accurate outcome, but the likelihood of a bad outcome remains high."
He cites the 1991 election for Louisiana governor as a failure of the plurality vote and IRV, had it been used. “Buddy Roemer, the incumbent, would have beaten either of the two opponents in a head-to-head race.” he said. "One was former governor Edwin Edwards, who many claimed was a crook; the other was David Duke, the head of the Ku Klux Klan. But, because of the problems with the plurality vote, Roemer came in last in the first stage of the process and Edwards won the 'Krook or Klan' runoff election. Clearly, he was not really the public’s choice as governor."
Jean Charles de Borda, a French mathematician, created the 2-1-0 rule in 1770 because he believed that weaknesses in plurality voting led to the election of mediocre members to the prestigious French Academy of Sciences. His method was used for several years by the academy until Napolean Bonaparte found he could not manipulate elections that used the Borda rule, so he eliminated it.
Others have known about the Borda rule for years, Saari said. "My recent work validates the mathematical properties that determine what can go right or wrong in elections." With these tools, any voting system can be analyzed for validity.
The Institute for Mathematics and its Applications (IMA) brings together the best minds in math and the sciences to solve pressing problems facing our society, our industries, and our planet. It receives major funding from the National Science Foundation and the University of Minnesota.
Speaker's webpage: http://math.uci.edu/~dsaari/