July 30, 2006
To: Minneapolis Charter Commission and Minneapolis City Council
Fr: Rob Richie, Executive Director
Re: Comments on Cynthia Reichert’s July 21st memo on instant runoff voting
I am writing in my capacity as executive director of FairVote – The Center for Voting and Democracy. FairVote is the nation’s leading organization researching and advocating ranked choice voting methods. (I should note that while we greatly respect FairVote Minnesota, we are entirely separate organizations.) Our representatives have testified before city councils, charter commissions and state legislatures across the nation and have worked with cities like San Francisco (CA), Burlington (VT) and Cambridge (MA) on implementing ranked choice voting methods and educating voters. Our chair is former Illinois Member of Congress and presidential candidate John B. Anderson. Our board of directors includes Rep. Jesse Jackson, Jr., Advancement Project senior attorney Edward Hailes and Laura Liswood of Council of Women's World Leaders. Our advisory board includes many of the nation’s leading scholars on voting method such as Arend Lijphart, former president of the American Political Science Association, and Kathleen Barber, author of the definitive scholarly work on the history of the single transferable vote system in cities in Ohio.
I wanted to respond to the report presented on July 21st to the city council by Cynthia Reichert, the Director of Elections. Although impressive in its research, the report does not explore all reasonable options for implementing instant runoff voting and the single transferable vote (a system that we call “choice voting” in our materials, by the way) and makes statements that demand clarification. The most important points I want to make clear are that:
- Ranked choice voting methods can be implemented well: I agree that it is very important to focus attention on effective, efficient ways to implement ranked choice voting methods and that ideally advocates of reform and local election administrators will have a cooperative working relationship. At the same time, our experience suggests that where there’s a will (and sufficient community commitment), there can be a way. Entire nations have used ranked choice voting methods for their most important offices for years, including Australia, Ireland and Malta, while next year Scotland will implement the single transferable vote (STV) for all its city elections. Instant runoff voting (IRV) has worked out exceptionally well in the two cities that use it today – Burlington (VT) and San Francisco (CA) – and STV has a long and successful history in the city of Cambridge (MA). In March, South Carolina adopted IRV ballots for all overseas voters, and those military and overseas voters returning ballots in the state’s primary less than three months later seem to have handled the ranked choice ballots nearly flawlessly. The North Carolina state legislature just sent a bill to the governor to try out IRV in up to 20 localities and to use IRV for any future vacancies in statewide judicial elections – using the exact same ES&S M100 voting equipment used by Minneapolis.
One example of an approach your city might consider would be to lease a small number of central machines able to record voters’ full choices at a central counting place. Other cities currently have optical scan equipment that can tally ranked choice voting ballots, and given growing interest and support for ranked choice voting methods, we expect all the more to be in place by 2009. Adoption of vote-by-mail is a separate decision to consider, but one that typically leads to higher turnout in city elections and reductions in the costs of running and election while meaning ballots would automatically come to a central counting place.
- The single transferable vote is a popular voting system: The report makes a non-fact-based evaluative statement questioning voter acceptance of STV, as has been proposed for some of your elections. But there is no reference to the fact that STV is widely appreciated in the countries and cities used it today, without any talk of repeal in these places. There also is no reference to the fact that groups of scholars, elected officials and citizens that study comparative voting methods often come down in favor STV. For example, in 2003 a citizen’s assembly in British Columbia overwhelmingly recommended adoption of STV to replace winner-take-all elections after an elaborate and exhaustive review of options; the proposal then went before voters, winning 58% support and a majority in 77 of 79 legislative districts (only failing to be implemented because of a requirement that it win at least 60%). Scotland’s recent adoption of STV followed intense study and analysis of a full range of voting options by a commission established by the parliament; a similar commission in Wales came to the same conclusion.
I wanted to address several specific elements of the memo, moving through it chronologically and organizing my comments by page.
- The memo is misleading when it suggests Minneapolis might get sued if it were to decide not to implement instant runoff voting or the single transferable vote after passing the proposed charter amendment. The basis for its contention was the unsuccessful lawsuit brought against San Francisco in 2003 when its elections director decided not to implement instant runoff voting. But note that first, the lawsuit was unsuccessful and second, the charter amendment in San Francisco did not make implementation of IRV an option; the city was supposed to get ready to run IRV elections, period, and failed to do so despite viable options and 20 months to get ready. Despite chastising the city for its failure to prepare to implement its charter, the judge stated he was not equipped to order the elections director to do something he contended he could not do. Given that the proposed charter amendment in Minneapolis allows the city council to vote to postpone implementation, a judge would be all the more likely to rule that if both the director of elections and the city council believe it is not ready to implement its new system, its decision should stand.
- Note that nearly all of the $13 million price tag for Alameda County’s new equipment doesn't relate to ranked choice ballots. The county plans to get new equipment regardless.
- The report estimates that printing costs would double with ranked choice voting elections. I would like to see evidence to support this assertion, which seems dubious. Different ballot designs would seem quite unlikely to double the number of cards necessary to print, and even if there twice as many cards, that might well not mean double the costs. I also am surprised to see some of the projected increases in costs for ballot programming and “accumulative software.” Your city’s vendor ES&S already has run instant runoff voting elections, and there are numerous examples of software able to count both IRV and STV ballots reliably and accurately. At the very least it is fair to say that these estimates are based on worst-case scenarios, not necessarily what they actually would be.
- Studies from San Francisco and the Burlington’s successful implementation of instant runoff voting show that the imperative for extensive voter education is overstated. We support more voter education for voters in general, but note: 1) A recently released report from the Public Research Institute at San Francisco State University shows that less than half of voters in the first citywide elections in San Francisco knew about instant runoff voting before they voted but only 2% found it "very difficult" and more than 99.6% of voters cast valid ballots. Similarly, exit polls during Burlington’s first IRV elections for Mayor in March 2006 show that the vast majority of voters learned about IRV from free media, far more than learning from publicly-funded voter education activity; still, the voter success rate was 99.9%. We believe the biggest keys are good ballot design and effective pollworker training.
- The report suggests nothing indicates that instant runoff voting boosts voter turnout. We do not have many IRV elections to consider in the United States, but note that: 1) our analysis of San Francisco’s first citywide IRV elections in 2005 resulted in turnout in the decisive round of counting in the one race needing an instant runoff that was nearly three times what it likely would have been in the old runoff system; 2) the Burlington elections this year showed that turnout was much higher than its most recent mayoral race, and significantly more voters cast ballots for mayor than for city council.
In addition, Minneapolis’ current system reduces the number of choices available to voters significantly in a primary where turnout is significantly lower and the electorate is significantly whiter, wealthier and older. This less reflective electorate in primaries is likely a national problem; a long-time political consultant estimated that the median age of voters in primaries is 61, which is at least 10 years older than the median age of voters in general elections, and a recent study of Oregon’s primary voters this year showed that fully two-thirds were over 57 and eligible voters over 70 voted at a rate of three and a half times the rate of eligible voters under 45. Eliminating the need for a low-turnout primary with IRV and STV should not be dismissed as an irrelevant consideration.
- The memo makes far too much out of the fact that ranked choice voting methods don’t necessarily change the outcome. First and foremost, the value of election systems cannot be evaluated by such a simplistic measure, as they are affecting a full range of important aspects of democracy, from demands on campaign finance to voter turnout. Relating to part of that point, the first count winners in IRV and STV elections might have been different if held under traditional plurality rules. Another related point is that establishing that a winner in a single-winner race has majority support and establishing incentives to build electoral coalitions are important considerations separate from determining who wins and loses elections. Secondly, the basic implication is misleading. True, starting off with more first choices is better than starting off with fewer first choices, but there are numerous high-profile examples of elections where, if the majority vote had been left split, the winner would not have been representative. One example is Ireland’s 1990 presidential election when Mary Robinson became the nation’s first woman president. She trailed after the first-count, but won handily in the instant runoff count.
- The report asserts that voters in the choice voting election may believe that they have been denied a voting right previously enjoyed. As stated early in my memo, this statement is an unsupported political assumption that does not seem appropriate for a fact-based analysis. Looking just at the facts, the current at-large voting system allows gross distortions of the spirit of one person, one vote -- some voters have all six of their votes elect someone, some voters might only elect two and some voters might elect none. The principle of choice voting is that as many people as mathematically possible will have one full vote count for the top-ranked candidate who can win with that vote. It is that power of everyone having one vote that is likely to count that is central to the effectiveness of STV – and indeed in the best spirit of the principle of one person, one vote. We believe that voters will like STV, just as they did recently in the British Columbia referendum on using it for their highest offices; but surely that is for them to decide if the charter commission and/or city council believe it is the best system for the offices where it is under consideration.
I trust that these comments are helpful. I also hope that advocates of reform can maintain a dialogue with your elections officials, as that is important both as reforms are being considered for adoption and as they implemented. If you have any questions about this letter, please do not hesitate to contact me at (301) 270-4616 or by email (as I am on the road this week) at firstname.lastname@example.org.