Photos of the three candidates for city commission, from the Daily Mining Journal Dec. 3, 1928. (Photo courtesy of the Marquette Regional History Center)

In most elections, regardless of the number of candidates, the candidate with the most votes wins. This is true even if they received less than 50% of the vote. This is known as plurality voting. When there are more than two candidates this sometimes means the winner is someone that a majority of the voters did not choose.

To address this problem, a number of jurisdictions, from small cities to entire states, have implemented “ranked choice voting.” This allows voters to pick a second choice if their first choice does not win more than 50% of the votes.

On these ballots, if no one has over 50% of the vote, there is no immediate winner. The system calls for the person in last place to be eliminated and their votes to be redistributed to the candidate ranked second on each ballot. This process repeats until a candidate crosses the 50% threshold.

Although much of the movement toward ranked choice voting has been recent, the idea is not new. From 1918 until 1933 Marquette used a unique ranked-choice voting system in municipal elections.

A number of cities in the United States experimented with ranked choice voting in the early years of the 20th century. Marquette, however, was the only one to adopt “Nanson’s System.” This voting method was devised in 1882 by Professor E. J. Nanson at the University of Melbourne in Australia. The Marquette city attorney described its use here,

“For each ballot cast, five points are distributed among the three candidates. The candidate designated on the ballot as the voter’s first choice gets one point. The candidate marked as second choice gets two points and the third-choice candidate gets three points. If a voter marks his ballot for only one candidate, that candidate gets one point and each of the other two gets two and one-half points. If a voter expresses two choices, the first choice candidate gets one point, the second choice two points, and the candidate not marked is regarded as the third choice and gets three points.”

After all points had been counted, the grand total was obtained and divided by three (the number of candidates) to obtain the average vote and those candidates having more than this number of points were eliminated. After a candidate was removed for being over the average, the entire calculation process had to be repeated with new values for the votes for the two remaining candidates.

In April 1917, Marquette voters empowered the City Commission to adopt by ordinance “a majority preferential ballot system of voting for officers of said city…” The ordinance was adopted by the Commission in August 1918, although the municipal elections in December of that year only had two candidates, so the counting system was not needed.

The system was put to its first real test in 1920 with a three-person race for the office of supervisor. Some mild criticism was directed against the form of the ballot and the inadequacy of the voting instructions rather than against the method.

The fact that the ballot provided a space, rather than the familiar box, confused the voters and at least 100 ballots were thrown out although the election officials counted all improperly marked ballots from which they were able to ascertain a clear intention. Fortunately, the race was not close, and voters and candidates alike accepted the results of the count without question.

Over the next seven years (1921-1927) the system attracted little or no attention, as there were no campaigns with more than two candidates. But an election in 1928 brought a storm of criticism due to suspected vote manipulation.

The candidates for city commissioner were Simon Anderson, Ralph Eldredge and William Janzen. Anderson and Eldredge were considered to be the better qualified candidates. Although Eldredge won, Janzen, the third and supposedly weaker and less qualified candidate, scored just a few more points under the Nanson count than were recorded for Anderson, the second leading candidate.

Although unsubstantiated, it was widely believed that there had been a silent understanding, if not actual collusion, between Eldredge and Janzen. It was alleged that the followers of each were instructed to exchange second-place support, thus ensuring an accumulation of third choice points sufficient to defeat Anderson.

The system survived this round of criticism and remained in place. As before, there were no campaigns with more than two candidates between 1929 and 1931. But in 1932, the city saw a repetition of the 1928 election, with Eldredge and Anderson again competing in a three-person race for city commissioner, this time against Peter Schon.

This time, Anderson won the election, followed by Schon and then Eldredge. Anderson announced that he would immediately move to repeal the ordinance under which the system had been adopted.

In addition to the objection that voting could be manipulated to prejudice the position of an otherwise-strong candidate, there were also complaints that the system was too difficult for voters to understand, and too costly and complicated for election officials to implement.

After 15 years of experience with preferential voting, the ordinance was repealed in February 1933 and Marquette returned to plurality elections. The extended duration of the experiment may be ascribed to infrequent election contests rather than to any real satisfaction with the method.

During the 15 years, only three elections provided contests in which it became necessary to use the Nanon count to determine the winner. In each of these elections, the winning candidate polled a plurality of first-choice votes. Each contested election was followed by a storm of criticism and demands for the abandonment of preferential voting.