Ranked-choice voting in Minneapolis was supposed to transform the city’s politics. As Minneapolis prepares this year for its second round of mayoral and City Council elections with RCV, it makes sense to look back at 2009 to assess what we can learn and what we should consider. As author of a study of the city’s first use of RCV in 2009, I can offer some insights.
To evaluate the 2009 implementation of RCV, one of my graduate students and I were given full access to documents and city officials and others involved with election administration. The report, completed in June 2010, identified five goals for RCV’s adoption: 1) save money by eliminating a primary; 2) increase voter turnout; 3) ensure that winning candidates in single-member districts receive a majority of the vote; 4) encourage voters to support third-party candidates, and 5) encourage third-party candidates to run for office and break the monopoly of the two-party system.
First, survey results demonstrated that voters were generally pleased with RCV. But it was unclear whether there were real cost savings in 2009 by eliminating the primary. There were many upfront, one-time costs associated with the implementation. Many of these costs will not be borne again, but it is fair to say that in 2009 there were probably no cost savings. In 2013, it may become more clear whether the city saves money.
Second, there is no indication that RCV affected voter turnout in 2009. Election officials planned for 70,000 voters; there were only 45,968. This was a significant downturn from 2005. RCV was not the cause. Given that incumbent Mayor R.T. Rybak was an overwhelming favorite, the election simply drew less interest and, therefore, voter turnout suffered.
Yet 2013 is different from 2009. There is no incumbent or clear favorite running for mayor, suggesting perhaps a more competitive and open race that may generate more turnout.
Third, all of the candidates who won did receive a majority of the votes. Yet that occurred in the first round of voting in 17 races, with three races decided in the second round. Even without RCV, almost all candidates received majority votes — and would have even under the old voting system. In all races, the first-round leader eventually won.
In 2013, this may change. In the mayor’s race alone, there are six or so viable candidates. A first-round winner might not receive more than about 20 percent of the vote. In this election, the eventual winner might actually be a second- or third-place first-round finisher. Multiple rounds of balloting may also make it more difficult to finish counts in time for the Canvassing Board to meet and certify the results.
Fourth and fifth, there is little indication that RCV either encouraged more third-party candidates or third-party voting. Minneapolis, like many cities adopting RCV, is essentially a one-party town. There is no clear indication in other cities that RCV has changed one-party domination, candidacy or voting patterns. After one election in Minneapolis, the same can be said.
We did note two concerns with RCV in 2009. First, while there were worries about voter confusion leading up to the election, the worst fears were not realized. Of the 45,968 total ballots cast, there were 1,888 spoiled ballots and 2,958 voter error ballots with mistakes specific to RCV.
In 2005, there had been 755 spoiled ballots out of the total 70,987 absentee and in-person voters. Thus, in 2005, 1.06 percent of all ballots cast were spoiled, as opposed to 4.1 percent in 2009. Explaining why the percentage of spoiled ballots increased is not clear. It is possible that voter confusion over RCV was a cause, but there seems to be no specific answer.
There was no voter error category in 2005. In 2009, 6.43 percent (2,958) of the ballots cast had errors specific to RCV. They were ultimately counted because voter intent could be ascertained. There is no indication that spoiled ballots and voter error ballots changed the outcome of any election in Minneapolis, but they accounted for 10.54 percent of all the ballots casts in 2009.
In close 2013 elections, this type of combined spoiled ballots and voter error could affect the outcome, running the risk of litigation and delayed results.
Moreover, the incidence of spoiled ballots and voter error was significantly greater in areas with higher concentrations of minority voters. Why this occurred the report could not ascertain, but it recommended more attention be given to evaluating the voter education and outreach strategies used in 2009 in preparation for 2013.
Overall, after one election the report concluded that it was simply too soon to tell whether RCV secured its five goals. Conversely, there was little to say that it failed. The 2009 elections were not a great test of what RCV could do. The spoiled ballots and voter errors raised concerns, but there is no indication that they changed any outcomes. In 2013, they could matter.
David Schultz is a professor and editor of the Journal of Public Affairs Education at Hamline University.