In the December issue of The Urbanist, we published an article by University of San Francisco Associate Professor Corey Cook that questioned the impact of ranked-choice voting on San Francisco elections. Recently, San Francisco State Professor Emeritus Rich DeLeon asked us if he could present another point of view. His reply is below. Professor Cook has in turn responded with a follow-up post.
Professor Corey Cook’s article in the December 2011 Urbanist assesses San Francisco’s ranked-choice voting (RCV) system in the 2011 mayoral election. His opening statement concludes that “by most objective measures the system held up rather well: The election results were clear and uncontroversial, individual ballots contained fewer errors than in past contests and most voters chose to participate fully by ranking their first-, second- and third-choice candidates.”
This would seem to be an occasion for high-fives and popping champagne corks. But Cook sees problems with RCV, lots of them. He has “deeper questions” about the effects of RCV on such things as the degree to which the election outcome “accurately reflects popular opinion,” the voter turnout rate, the level of negative campaigning, the perceived legitimacy of election results viewed as a mandate to govern, the informational burdens placed on voters in ranking candidates, and the incidence of voting errors. He acknowledges that these questions “are difficult to answer. In addition to the voting system, the context of the election included generous public financing, an incredibly deep pool of serious contenders and a popular acting mayor who entered the race at the last minute. It’s impossible to disentangle the independent effects of ranked-choice voting.” Undeterred by these obstacles, however, Cook goes on to assert a giant non sequitur: “But it’s easy to see the deep flaws in this election.”
What are these “deep flaws,” and what’s his evidence for them?
1.Ranked-choice voting and popular opinion.
Cook writes that it is “unclear whether ranked-choice voting accurately reflects popular opinion.” Popular opinion about what? Measured how? Here one might reasonably expect a careful comparison of actual election outcomes with the predictions of pre-election polls, or perhaps an assessment of how accurately the observed voting patterns reflected the city’s diverse and complex demography. Instead, Cook reports statistics on the incidence of voting for only one candidate in the mayoral, district attorney and sheriff races. He misleadingly calls voting for only one candidate “bullet voting,” and he claims that it “remains prevalent” based on a reported 16 percent of voters indicating a preference for only one candidate in the mayoral race, with higher percentages in the district attorney and sheriff faces. (Cook’s definition of “bullet voting” is misleading because political scientists and campaign managers typically use the term to mean a form of tactical voting in which a voter is encouraged to vote only for his or her preferred candidate, despite having the option to vote for more, in order to deny votes to rival candidates.) He writes that it’s unclear whether a “sizable block of voters sincerely preferred only one candidate” or — the only alternative he suggests — “whether they [the voters] were unsure what to make of a ranked-choice ballot,” implying voter ignorance or confusion. Cook adds that 1.2 percent of voters marked more than one candidate as their first choice, a figure “higher than in standard ‘vote for one’ candidate races.”
Questions: First, how much lower than 1.2 percent can the comparable figure in “standard” races possibly be? Second, why is an explicit contrast made here but not elsewhere, except by implication, between RCV and “standard ‘vote for one’” methods (such as are used in the traditional runoff system, which some RCV critics would like to restore)? Third, and most important, what makes Cook think these “bullet voting” and voting error statistics are in any way an appropriate yardstick for gauging “popular opinion” and assessing whether RCV accurately reflects it?
2. Ranked-choice voting and voter turnout.
Cook writes that the “clear results from November’s election included abysmal turnout — right around 42 percent — the lowest in a contested mayoral election since at least the 1960s. Only Mayor Gavin Newsom’s 2007 landslide re-election was lower.”
What Cook deems an “abysmal” turnout of 42.5 percent of registered voters was only 3.2 percentage points lower than in the November 2003 election (45.7 percent) and a mere 2.5 percentage points lower than in the November 1999 election (45.0 percent). (The turnout rate was 35.6 percent in November 2007, truly “abysmal” by San Francisco standards, but Cook dismisses that election as anomalous.) Given the secular decline in U.S. voter-turnout rates generally and in urban electorates particularly, why does Cook think San Francisco’s 42.5 percent turnout is abysmal? Moreover, San Francisco has the highest voter turnout rate among the 22 most populous U.S. cities in their most recent mayoral elections. Most urban political scientists would want to explain why San Francisco’s turnout rates are so high relative to other large U.S. cities, not why the city’s turnout dropped 2 to 4 percentage points over the last few elections, which could have happened for any number of reasons.
Lacking such background for evaluating Cook’s claims, some readers might infer that RCV must be the culprit behind the drop in turnout. Voter turnout under the old runoff system was high. Voter turnout under RCV was low. Therefore, RCV must have caused the drop in turnout. To avoid even the suggestion of that kind of fallacious logic, Cook could have provided more context and clarification. Instead, he chose to double down on the thesis that RCV itself was in some way responsible for the lower turnout in the 2011 mayoral election.
3. Ranked-choice voting and the informational burdens of voter calculation.
To buttress that thesis, Cook makes a clever but rather dubious argument. First, the city’s voters, being rational, calculated the benefits and costs of turning out to vote. Second, the informational burdens of ranking as many as twelve serious candidates under RCV increased the perceived costs of voting. Third, those higher costs combined with lowered expected benefits of voting (in a predicted landslide victory for Lee) help to explain why, under RCV, voter turnout dropped from previous levels.
Let’s leave aside debates we might have about how consciously aware typical voters are in computing benefit-cost ratios of whether or not to vote, other than to say that pollsters and campaign managers tag many voters as “habituals” for a reason.
Professor Cook’s assertion about the costly informational burdens of RCV seems to be the critical causal link he wants to make between the adoption of RCV and lower voter turnout. Two points:
First, Cook’s argument, such as it is, ignores the voter’s rarely mentioned burdens of calculation under the old runoff system. Examples: I prefer candidate X. But X is unelectable. So should I vote for X anyway, on principle? Or should I vote for my less-preferred but minimally acceptable and more electable candidate Y, thus avoiding the risk that my principled vote for X will spoil Y’s chances and thus perversely help elect candidate Z, whom I despise? Or should I vote for the despised Z, a weak candidate, to place him in the runoff against my favored candidate, X, thus increasing X’s chances of victory? Etc. As these examples illustrate, if the informational burdens of voting under the old runoff system seem so much smaller than under RCV, that is only because they are so familiar.
Second, Cook’s argument also ignores relevant findings from a San Francisco State exit poll he and his colleagues conducted to assess RCV in the November 2004 Board of Supervisors elections. In that poll, 2,610 non-first-time voters who voted in the Board of Supervisors election were asked: “Compared to past elections for the Board of Supervisors, how much information did you gather about the candidates before voting today: More than in past elections, no difference, less than in past elections?” About 31percent of sample voters said they gathered more information than in the past, about 7 percent said less, and the rest reported no difference. Given these results and Cook’s claims about the burdens of information gathering under RCV, one would expect that voters so burdened would much prefer the less demanding old December runoff system to the new one. Not so. An estimated 71 percent of voters who gathered “more” information on the candidates said they preferred RCV; only 11 percent favored the old December runoff system. Of those who said there was no difference, 68 percent said they preferred RCV and 12 percent the old system. And of those who said they gathered “less” information, “only” 52 percent said they preferred RCV versus 21 percent who favored the old system. Some readers may find these results surprising, especially if they view information-gathering strictly as a burden or cost of voting rather than as a benefit.
One last comment on this point: Some critics of RCV complain that it allows voters too few rankings under the technical limitations of San Francisco’s existing voting machines and software. In other words, the problem, as they see it, is one of not enough choice rather than too much choice. A recent lawsuit challenging the city’s RCV system on just those grounds was rejected by the Ninth Circuit Court earlier this year. Unfortunately, Professor Cook doesn’t consider this side of the debate in his assessment.
4. Ranked-choice voting and negative campaigning.
Cook disputes the claim made by some RCV advocates that it discourages negative campaigning. I have questioned that claim, too, in my independent analysis of Cook et al’s 2004 exit poll data. (PDF download, see pages 5-7.) Yet Cook characterizes the recent mayoral race as “exceptionally nasty in the final month,” implying that negative campaigning reached new heights of vituperative meanness under RCV. He concedes, however, that RCV apparently did discourage negative campaigning among second-tier candidates, whose ballots were likely to be redistributed to the quarrelsome front-runners. That, it would seem, is a glass half full, and a point (or at least half a point) in favor of RCV.
Like the contrary arguments some critics have made about RCV’s informational burdens (too much choice! too little choice!), critics also seem divided on whether hard-hitting negative campaigns are a bad thing or a good thing. Some lament the muted differences and the lack of blood and gore under RCV. Others, like Professor Cook, regret to inform us that the new system may actually foster exceptionally nasty campaigning, at least among top-tier candidates. Based on my own observations, I’ve seen little evidence in this latest election or in earlier ones that RCV discourages candidates from taking strong and clear positions on the issues or from engaging in sharp debates. Politics remains a blood sport in San Francisco, and the old saying that “truce is stranger than friction” in this city still applies, even under the civilizing inducements of ranked-choice voting.
5. Ranked-choice voting and the contested concept of “majority” rule.
I confess I had a hard time following the logic of Professor Cook’s concluding arguments that RCV has failed to produce a “majority” winner (and thus governing legitimacy) in most San Francisco RCV elections. He doesn’t accept the charter language defining a “majority” winner under RCV as the candidate receiving a majority of continuing votes — a definition that has served well over many elections since 2004 to produce “uncontroversial” election results. For Cook, a majority is at least 50 percent plus one of total votes cast in a given election, and that is that, no matter how many more votes might be cast for the winner in the typically high-turnout November RCV elections than in the typically low-turnout December runoff elections.
Cook’s denial of governing legitimacy to majority winners under current RCV rules leads to some unfortunate polemics. (Most disturbing to me is the gratuitous swipe at Oakland’s new mayor, Jean Quan, challenging the legitimacy of her election, which she won fair and square, and insinuating that her current struggles to govern her city under the most trying circumstances stem from a tainted victory under RCV.) I have great respect for Professor Cook, his scholarly publications and his political expertise. However, I believe his assessment of RCV would be more credible as a fair and objective analysis if he had actually acknowledged and engaged those who offer other perspectives on these issues. I trust that his forthcoming study of the city’s ranked-choice voting system will be more comprehensive in scope, more inclusive and respectful of different points of view, and more illuminating.
One last thought: It seems to me that the local political culture has adapted very well to RCV since the city’s first ranked-choice elections in 2004. More political groups and clubs are making ranked endorsements. More media organizations and campaign managers are using ranked-choice formats in their polling. More candidates, at least those serious about winning, are paying close attention to the new rules of the electoral game in calculating strategies and tactics. The city’s election administration is operating more smoothly, quickly and efficiently in processing ballots and reporting election results. Indeed, the most recent election results were, as Cook writes, “uncontroversial” — a word rarely spoken about San Francisco politics. In general, San Franciscans appear to have become quite comfortable and familiar with RCV. That’s why I’m so puzzled by the timing of recent attacks on RCV and calls for its repeal. Could it be — and I’m just speculating here — that some critics fear the last chance is slipping away to smother RCV in its cradle?
* Rich DeLeon is professor emeritus at San Francisco State University, where he taught political science and urban studies for 35 years. He is author of Left Coast City: Progressive Politics in San Francisco, 1975-1991.