CHAPTER 5
INSTANT RUNOFF VOTING--THE TECHNICAL DISSENT
The following three items are taken from what I would describe as the professional literature on voting systems as opposed to the polemical literature. As such they highlight issues which, as indicated in the first chapter, I believe supporters of IRV fail to understand.
THE FOLLOWING, taken from Brams and Fishburn's book "Approval Voting"
(1983), OUTLINES SOME ADVANTAGES OF APPROVAL VOTING OVER PREFERENTIAL VOTING, also known as IRV.
IT (Approval Voting) IS SUPERIOR TO PREFERENTIAL VOTING. An election system that is used in a few places in the United States (for example, Cambridge, Massachusetts) shares many of the advantages of approval voting. It is called majority preferential voting, or "single transferable vote," and it allows each voter to rank the candidates from best to worst. If no candidate receives a majority of first-place votes, the candidate with the fewest first-place votes is dropped and the second-place votes of his supporters are given to the remaining candidates. The elimination process continues, with lower-place votes of the voters whose preferred candidates are eliminated being transferred to the candidates that survive, until one candidate receives a majority.
There are several practical and theoretical problems with preferential voting. First, it is difficult and costly to implement, and many voters do not understand how it works. For example, in New York City school board elections, in which this system has recently been used, it takes a week to count the ballots, make the transfers, and determine the winners. This delay could be alleviated by computerizing the system, but the capital expenditure needed to do this would be substantial. Moreover, it seems to be a system that voters do not comprehend very well, at least measured by extremely low turnouts of less than 5 percent in recent school board elections under preferential voting.
Second, under preferential voting, the candidate with the most first-place votes may be displaced after the transfers have been made to determine the majority winner. This may greatly upset that candidate's supporters, particularly if they are a large minority, and lead to questions about the legitimacy of the system--and ultimately its rejection by voters, as occurred in 1976 in Ann Arbor, Michigan.
This challenge cannot be mounted against approval voting since approval votes are indistinguishable: whether these votes are first-place, second-place, or whatever is not recorded, so no portion of the winner's total can be judged "inferior."
Third, and more serious, is the possibility that preferential voting may eliminate the candidate acceptable to the most voters. For example, Charles Goodell would have been eliminates as low man in the
1970 United States Senate race in New York with 24 percent of the plurality vote (Buckley got 39 percent and Richard Ottinger 37 percent), though there is evidence he, like candidate in the hypothetical example given in Section l.2.3, could have beaten each of his opponents in face-to-face contests and would have won under approval voting.
Fourth, and probably more serious, is preferential voting's violation of a fundamental democratic ethic: a candidate may do better under this system by getting fewer, rather than more, first-place votes.
For example, a candidate can win an election when, say, 10,000 voters rank his first but lose when an extra 5,000 voters move him from a lower rank into first place (without changing their ordering of other candidates). In other words, it may be in the interest of a candidate to tell some of his supporter not to rank him first to ensure his victory--a truly perverse result. Without going into details, this perversity arises not because additional votes are counted against a candidate but rather because of the successive eliminations of the ostensibly weakest candidates and the transfer of their votes to ostensibly stronger candidates, as we shall illustrate later.
Because of this perversity, preferential voting is said in the technical literature to violate a "monotonicity condition." It is a pathology that does not afflict approval voting, because the more approval votes a candidate received relative to the other candidates, the better he does. Under no circumstances can a candidate be hurt under approval voting by receiving additional first-place votes, which seems to us an extremely damaging flaw of preferential voting.
THE REPORT OF THE INDEPENDENT COMMISSION ON THE VOTING SYSTEM
(U.K. Official Document October 1998)
The following excerpt is taken from the Note of Reservation by Lord Alexander of Weedon QC:
Suppose within a constituency, Conservatives receive 40% of first preferences, Labour are second on 31% and Lib Dems third on 29%. Lib Dems second preferences happen to be split 15/14 in favor of Labour.
The Conservatives are therefore elected with 54% of the total vote (i.e. 40% + 14%).
But now suppose the position of Labour and Lib Dems had been reversed on first preferences, with Lib Dems 3l% and Labour 29%. If Labour second preferences were split 20/9 in favor of Lib Dems, the Lib Dems would be elected with 51% of the total vote (i.e. 31% + 20%). So the result would be different depending on which horse was second and which third over Becher's Brook first time round. This seems to me too random to be acceptable.
ANOTHER VIEW ON THE SINGLE TRANSFER VOTE, AKA IRV
The following is taken from: Rudolf Fara and Maurice Salles, "An interview with Michael Dummett: from analytical philosophy to voting analysis and beyond," SOCIAL CHOICE AND WELFARE, Vol. 27, No. 2, Oct., 2006, pp. 347-364. Sir Michael Dummett is a prominent British philosopher who has published important works on voting analysis (systems).
p. 353. There have been advocates of specific electoral systems, in particularly single transferable vote (STV) [on which IRV is based--jhw], who have claimed that this system makes it impossible for anyone to gain any advantage by strategic voting. I thought it absolutely essential to prove that there can't be any such system, before discussing any particular system, because, if people think that it is possible to have a voting system that always denies advantage to strategic voting, they will have a wrong approach to the whole question.
p. 354. It seems to me that the great advantage of STV is what it does for minorities; it guarantees that the minorities get represented.
Otherwise it has no advantages at all. It's an almost chaotic system in the sense that a very small change in preferences would have results that change completely the outcome because it would affect the order in which candidates got eliminated, and so which votes were distributed at each stage. So I think that it is really a very bad system...