Vote early, vote single transferably

Voters in British Columbia, when asked on a 2005 ballot whether they supported switching to a Single Transferable Vote (STV) system like the one used in the USG’s U-Council elections, chose not to make the switch.

Though 57.7 percent of voters opted to adopt the STV system, the number fell short of the 60 percent required to pass the proposal. In a subsequent poll of the people who had voted against the STV referendum, a large majority said they hadn't supported the new voting system because they didn't understand it.

U-Councilors echoed a similar sentiment of confusion at last at Sunday's USG meeting, when several candidates said they didn't fully understand the STV algorithm used to fill their seats. USG president Josh Weinstein ’09 has suggested possibly reforming the system in the future, but perhaps clearing the confusion really requires the explanation of a simple mathematical algorithm rather than radical change in procedure.

The idea of transferring votes, which underlies all STV algorithms, dates back to the 1819 election for the Committee of Birmingham Society for Literary and Scientific Improvement. Thomas Hill, a schoolmaster, devised a voting method for the election whereby each participant voted for one candidate, and anyone who received five or more votes was automatically elected. For each vote that a candidate received beyond the five-vote quota, one randomly selected person who had voted for that candidate was allowed to vote again, this time for a different candidate, thereby "transferring" their vote, since their first-choice candidate no longer needed it. By this process, votes were transferred to other candidates until all the seats on the committee were filled.

Hill's ideas were later formalized into the single-transferable vote system by Thomas Hare, an English barrister, in his 1857 pamphlet “The Machinery of Representation,” in which he advocated letting voters rank candidates in order of preference and then transferring surplus votes in a fashion similar to the one suggested by Hill. Hare made several changes to the system, most notably deciding that in cases in which no candidates had received the necessary quota of votes, the candidate with the fewest votes would be eliminated from the running, and the votes cast for the eliminated candidate would then be transferred to the voter's next choice.

Fewer than five years after Machinery's publication, Hill was praised by philosopher John Stuart Mill for creating this STV system. "The more [that writings about STV] are studied, the stronger, I venture to predict, will be the impression of the perfect feasibility of [STV]," Mill wrote. "In my conviction, they place [STV] among the very greatest improvements yet made in the theory and practice of government."

Mill was perhaps a bit hasty in his glowing assessment of STV — a number of issues surrounding its effectiveness have since been discovered — but to this day, STV is one of the most popular and reliable methods of ensuring proportional representation.

Proportional representation electoral formulas aim, as their name suggests, to select groups of elected officials that proportionally reflect the diversity of opinions within the electorate. Many elections operate on a "winner-takes-all" basis that awards political parties who win the majority of the vote a disproportionately large number of elected offices.

Cincinnati, for instance, chose to adopt an STV system in 1925, after a city council election awarded the Republican Party 97 percent of the seats on the Council though the party received only 55 percent of the vote. In the first STV election in Cincinnati, Republican candidates won 33.3 percent of the Council seats based on 27.8 percent of the vote.

Ironically, the success of STV in awarding proportional representation to different political parties proved to be its downfall in America. Several cities, including Cincinnati, New York, Cleveland and Cambridge adopted STV during the first half of the 20th century. Political parties, however, found that as a result of the reform, they were losing power in cities in which they had once been clearly dominant. As a result of party lobbying, almost every city in the country that adopted STV had repealed it by the 1960s.

Proportional representation of political parties can be achieved in many different ways, not just with STV. Governments can, for instance, simply choose to allot a party a number of elected offices directly proportional to that party's percentage of the vote. STV is unique, however, in that it allows for a more proportional representation of the candidates' qualities (rather than political party) that matter to the voters and therefore makes the most sense for elections like the USG U-Council that do not involve any political parties.

Weinstein said STV was "likely used [in U-Council elections] to ensure diverse representation ... and eliminate ‘strategic’ voting." STV has indeed been quite successful in electing diverse groups of candidates. In three cities in Ohio — Cincinnati, Hamilton and Toledo — African-American candidates were first elected to city government after STV was adopted, according to research by Mount Holyoke politics professor Douglas Amy.

Get the best of the ‘Prince’ delivered straight to your inbox. Subscribe now »

"Significantly, after these cities abandoned [STV], African-Americans again found it almost impossible to get elected," Amy wrote. In Ireland and Australia — where STV is routinely used for elections — the system has been hailed as facilitating the election of women.

STV’s tendency to increase diversity of elected officials has also worked against its widespread implementation in the United States. In Cincinnati, STV opponents campaigned against the system by telling city residents that continued use of the system might lead to the election of a "Negro mayor."

In New York, voters repealed STV elections in fear that it might allow communist candidates to come to power.

Apart from its tendency to diversify elected officials, STV has also won praise for its minimization of "wasted votes," or votes that do not actually serve to elect a candidate to office since the candidate could have been elected without them.

STV systems allow some voters who choose either the most or least popular candidates to "recast" their vote for their next choice candidates rather than waste their votes. One problem with Hill's initial implementation of this vote transfer is its arbitrary nature. In 1819, Hill chose to randomly select the surplus voters from the groups of people who voted for the most popular candidates. This meant that some voters were offered the chance to support their next choice candidates while others weren’t. Besides the obvious unfairness of this system, it meant that the election was non-deterministic. The same votes could be cast by the same people for the same group of candidates, and the outcome of the election could be different each time the votes were counted, based on which randomly chosen votes were transferred.

More recent versions of STV algorithms have solved this problem by transferring a fraction of each vote cast for a successfully elected candidate. In this fashion, each voter can contribute equal support for their next choice candidate.

STV has also been criticized for its oddly unstable behavior since its results vary depending on the number of seats open for election.

In 1880, a clerk for the United States Census Bureau discovered that the state of Alabama would be allotted eight seats in a 299-seat House of Representatives but only seven seats if the House were expanded to 300 seats. The Alabama paradox, as this was called, comes about because the House of Representatives increases the fraction of seats available to larger states much faster than it does for less populated ones.

Similarly, in an STV election for a group of n individuals, the winning candidates will not all necessarily be elected if the same ballots are used to choose n + 1 people. This is because the quota of votes needed to elect an individual candidate shrinks as the available seats increase, so individuals who were initially eliminated for having too few votes may now meet the new quota and be elected, replacing some of the candidates to whom their votes were transferred when they were originally eliminated. In other words, there is less compromise between voters when more seats are available and more elected candidates who are strongly supported by smaller groups, rather than mildly supported by larger ones.

Many voting experts have differed over the choice of which quota must be reached for a candidate to be successfully elected. A quota that is too big can result in not enough candidates meeting it to fill all the open seats, an outcome which can also result from failure or inability of the voters to rank all possible candidates on their ballot, as in the recent USG elections, when students were allowed to vote for only two candidates.

The current confusion surrounding the recent USG elections may stem largely from the lack of clarity in the USG constitution that states that "U-Council [and Senatorial] elections shall be held by the system of the single transferable vote." No provisions are made for what quota or vote redistribution algorithm must be used, or how many candidates should be ranked.