Essays on Ranked-Choice Voting: Chapter 1

With another election behind us, I have been planning to write one or more essays promoting and analyzing ranked choice voting. I have long been an advocate of this method of conducting elections—see The People’s Choice (Round Two), originally written November 11, 2000—and there seems to be increased interest in it of late. By no means does the essay below start at the beginning, but I start here in response to an interview heard recently on NPR.

WAMU’s public affairs show1A presented a program titled “The Future and Feasibility of Ranked-Choice Voting.” Readers who have never heard of ranked-choice voting (RCV) should first read my essay mentioned above. More comprehensive information about RCV can be found on the FairVote Web site. (Fair Vote offers a concise description of RCV here.)

What most interested me on the 1A episode was the criticisms made of RCV by one of the guests, Jason McDaniel, a political science professor at San Francisco State University. Professor McDaniel’s role seemed to be that of naysayer invited to balance the opinions of RCV advocates. His fundamental criticism of RCV is that ballots are complex—more complex than the usual ballots, at any rate—which leads to two basic problems. First, some people are discouraged from voting because they do not understand the system. McDaniel claims to have empirical evidence of lower turnout due to ballot complexity, though he admits that the effect—if it exists at all—is small. Second, McDaniel asserts that RCV produces more ballot errors.

I believe that the benefits of RCV are substantial and remain so even in light of McDaniel’s criticisms. I do not intend to argue for those benefits here but merely to address McDaniel’s objections.

It is undoubtedly true that the ballots for RCV are more complicated than those of conventional American elections. Rather than marking a vote for one’s favorite candidate, it is possible, though not necessary, to indicate candidate preferences in order: first choice, second choice, etc. RCV ballots may be structured in various ways, and some may be more confusing than others. Typically, however, ballots include clear instructions as to how to cast one’s vote. And the changeover to RCV voting needs to be publicized and explained to the voting public. Will some people still misunderstand how to fill out a ballot? Surely, but people ruin ballots using conventional, “simpler” voting schemes—they mark their vote for two opposing candidates, they skip races by mistake, and they fail to follow instructions for mail-in ballots.

The instructions on an RCV ballot should make it clear that one can vote for a single candidate, i.e., treat the ballot as for a conventional election. If we believe even this is too confusing for some folks, we can offer alternative ballots, one that allows voting only for a single candidate in each race, and one that allows for ranking voter preferences. Let the default ballot be the conventional one, and let people confident in the use of the RCV ballot request such a ballot. The conventional ballots are then treated as if they were RCV ballots with only the most preferred candidate chosen.

As I mentioned, voters already have ways of spoiling their ballots. RCV provides more ways to cast an incorrect vote, but many errors can be “cured” by applying some straightforward rules:

• If two or more candidates are given the same rank (e.g., second choice), the votes for those candidates are disregarded.
• If the ballot requires the voter to indicate rank by entering numbers, rather than, say, filling in bubbles, marks other than numbers are disregarded. Likewise, illegible entries are disregarded.
• Candidates ranked without duplication have their preferences renumbered, if necessary, to maintain the indicated ordering. The candidate with the lowest indicated rank (i.e., the most desirable rank) is considered the voter’s first choice, the candidate given the next higher rank is considered the voter’s second choice, etc. This re-ranking can be done by computer before ballots are tabulated.

These rules do the best job possible of capturing the intention of the voter, even in cases where the voter has marked a ballot improperly. Is not this the goal we would like to achieve?

Some ballot formats are susceptible to a greater range of errors than others. Paper ballots requiring the voter to enter numbers are most susceptible to errors and should be avoided. From the point of view of catching errors at the source, a computer-based voting system is most desirable. Given the need for hardcopy voting records that can be audited, however, computer-based systems need also to generate a paper copy of the voter’s choices. The costs of such a system must be weighed against the cost of spoiled ballots and possible extra costs to “cure” ballots (if desired) as described above.