As this merciless presidential election rounds its final bend, it has become commonplace to hear complaints of election rigging from the Republican nominee, Donald Trump. He has urged his supporters to be vigilant. Supposedly there are shenanigans going on at the polls, particularly in areas where lots of Democrats and nonwhite voters live.
Trump is one among millions of Americans unaccountably nervous about voter fraud. But it’s hard not to sympathize with citizens’ deeper anxiety that their votes don’t really count. A recent poll conducted by the Public Religion Institute finds that only 55 percent of Democrats and 44 percent of Republicans believe their votes will be tallied accurately. This worry extends beyond mere errors of tabulation, though. Voters say their representatives are “out of touch” with them and too greatly influenced by special-interest money. Across the country, Americans pine for nonpartisan redistricting; voters believe gerrymandering has also diluted their voice.
It should come as no surprise, then, that some constituencies are turning to changes in voter procedure on the theory that, if elections were run better, elected officials would better mirror the preferences of voters. The most significant such reform now under consideration is an initiative on this year’s ballot in Maine. If the latest poll is right, and the referendum on question 5 passes, the state’s current electoral system will be scrapped and replaced with a method called ranked-choice voting (RCV).
Also known as instant-runoff voting, RCV allows voters to rank candidates by preference. If one candidate wins a majority of first-place votes, she is declared the winner. If no candidate takes a majority, then a series of virtual runoffs commences (virtual because voting takes place just once, and “runoffs” occur in the tallying process). In each runoff round, the last-place finisher is eliminated from contention. All ballots on which she was ranked first are now repurposed. The second-place candidate from these ballots is given another first-place vote in the next runoff. This process continues, with the worst performer in each round eliminated and remaining ranked candidates promoted, until one candidate has won at least half of the first-place votes. Sometimes. More on that later.
The Yes on 5 campaign argues that RCV will provide voters “more voice and more choice,” as well as freedom from voting for the lesser of two evils. A voter can avoid the Ralph Nader effect and select a third-party candidate with little chance of winning, secure in the knowledge that a major-party candidate she can tolerate will benefit from her second-place vote. No more spoilers.
Backers of the initiative also argue that the new system will discourage negative campaigning because candidates risk losing valuable second- and third-place votes by alienating one another’s strongest supporters. Proponents of RCV argue that the process will even reduce legislative gridlock. RCV “forces candidates to try to appeal to a broader cross-section of the public,” political scientist Larry Diamond writes. This “makes it much more likely that the winner will be open to moderation, compromise and building governing coalitions.”
So transformative is RCV, according to the question 5 campaign, that it will restore the crumbling bedrock of democracy: majority rule. “Our leaders should be elected by more than half of us,” proponents say, noting that the winner in nine of the state’s last eleven gubernatorial races took less than half the vote. And when Portland, Maine’s largest city, used RCV for its mayoral election in 2011, turnout exceeded predictions by 15 percentage points, suggesting that the system may even boost participation. Anyone can get behind that.
But there are reasons for skepticism when it comes to RCV—and not just RCV itself, but the larger notion that what is broken in American politics, and therefore what will fix it, is procedure.
RCV hasn’t been used extensively in the United States. Nor has it been tested at the state level since the early twentieth century. But it has been used in municipal elections in California, Minnesota, Washington state, and elsewhere. And for nearly a hundred years, Australians have elected their lower house of parliament using the method.
A closer look at American and Australian RCV races suggests that the system probably won’t damage Maine’s voting process, but it probably won’t help much each either. Few of the touted benefits are likely to materialize.
RCV regularly falls short of Yes on 5’s headline goal: a majority winner. In a 2014 paper in the journal Electoral Studies, political scientists Craig Burnett and Vladimir Kogan analyzed some 600,000 votes cast using RCV in four local elections in California and Washington. In none of the four did the winner receive a majority of votes cast.
How is this possible? On paper, RCV ensures that no one can win unless she receives more than half of the vote. But what works on paper doesn’t necessarily on ballots.
The problem is exhaustion. Not the kind you’re experiencing now, as you cry yourself to sleep at the prospect of another day absorbing the pay-per-view punishment of “Clinton v. Trump: The Rumble in the Rustbelt.” No, this is ballot exhaustion, which happens when voters rank too few candidates to stay meaningful until the final runoff. Say there are five candidates running, but the voter ranks only three, and all three are eliminated prior to the last round. As a result, none of their votes will have gone to the winning candidate or the runner-up. In effect, their ballot doesn’t figure in the outcome.
This may sound like a marginal problem, but its effects can be substantial. Of the four elections Burnett and Kogan studied, none produced an exhaustion rate lower than 9.6 percent. In one case, the 2011 San Francisco mayoral race, just over 27 percent of valid first-round ballots were exhausted before the last tally. “Voters who cast these discarded ballots had no say in the final round of vote redistribution, which decided the election outcome,” Burnett and Kogan write. This is akin to saying that, thanks to RCV, 27 percent of voters who cast primary ballots sat out the general.
When RCV does produce majorities, they may be unconvincing. In 2010 the Australian Labor Party won the House of Representatives with just 38 percent of first-place votes on the initial ballot, while the second-place Liberal-National coalition captured 43 percent. That hardly sounds like a firm mandate.
So much for guaranteed majority rule. What about a more pleasant campaign atmosphere, no-guilt third-party voting, and legislative moderation? Experience suggests there isn’t a lot to look forward to on these fronts, either.
For one thing, much of campaigning in America isn’t done by the candidates themselves but instead by ideologically driven political action committees. A candidate may lay off a near competitor in order to court second-place ballots, but Heritage Action, Planned Parenthood, and other issue organizations in the scrum don’t have anything to gain from compromise.
Quite to the contrary, the system may give life to more strident candidates, hoping to siphon first-place ballots from extreme voters who will give second preference to whichever major party is closest to them. This could result in more comity between the major-party candidates, as fringier competitors blot the airwaves with attacks. Or it might produce strategic coalitions sniping at each other, leaving us effectively back where we started.
But we needn’t rely on hypotheticals. Negative advertising is all over Australian elections. David Crowe, a columnist for The Australian, apparently didn’t get the word about his country’s gentle electioneering. He likened the scare tactics of this year’s federal campaign to those of 2010, another recent “display of pure political desperation.”
Don’t expect ad buys to fall under an RCV system, either. When Oakland first tried RCV for its mayoral race in 2010, candidates spent $1 million; the 2014 race cost them nearly $1.8 million. This may reflect the sense that RCV makes viable a wider range of candidates, so more people run. One way or another, it doesn’t sound like a recipe for a smaller TV war or reduced bickering.
There is also little reason to believe that RCV will promote legislative moderation—or new campaign tactics—at the federal level, because it usually produces outcomes similar to what one would expect from a standard plurality system. In the 2013 Australian federal election, 90 percent of constituencies elected the candidate with the most first-preference votes, which suggests that choice ranking had little effect on the outcome.
And it is hard to ignore the resemblance between the Australian and U.S. governments, as far as partisan divisions go. Despite RCV, just two governments have led in Australia for almost the entire history of the current Federal Parliament: Labor and Liberal-National. (Technically the Liberal and National parties are separate, but they have been allied since the 1920s, and, at least at the national level, a vote for one is effectively a vote for the other.) Every time there is a federal election in Australia, one of the two major parties wins, RCV be damned.
Australians do vote for third-parties at a greater rate than Americans. But this makes sense in a parliamentary system, where small parties can wield outsized influence by joining governing coalitions. That is extremely rare in the American system, so there is little potential for third-party influence, even if RCV could bring more independents into legislatures. It is not a bad thing if RCV enables no-guilt third-party voting, but doing so won’t wrest power from Democrats and Republicans and turn it over to independents.
None of this is to say that RCV is sure to be hazardous. Maybe it is even an experiment worth trying. But it is notable that, in the midst of a presidential campaign that has unmasked deep and dangerous fissures in American politics, concerned citizens are looking to procedural minutiae as their savior. It is difficult to avoid the conclusion that voters are grasping for a solution as simple as the problem is daunting.
The appeal to procedural tweaks reflects a belief that, through technique, genuine differences of opinion in the electorate can be overcome. This might have some merit if those differences were based in policy views, presumably amenable to revision. But policy has never been a major driver of voting decisions. Rather, what moves voters is whether they think the incumbent party was good for their bank accounts and whether a candidate promotes their group interests, which are powerfully delineated by race.
If we can’t engineer our way around fundamental differences, the voting system could at least ensure that they are represented in legislatures. But failures of representation don’t stem primarily from flawed voting procedure. Even the effects of gerrymandering are overblown. Instead, political scientist Martin Gilens has figured out why legislators aren’t listening to you: It’s because you’re not rich. His analysis of decades of public opinion polling and subsequent Congressional action finds, “In most circumstances, affluent Americans exert substantial influence over the policies adopted by the federal government, and less well off Americans exert virtually none.” The economy of influence surrounding campaign finance is considerably to blame for this.
We are thus left with a plutocracy insulated by election laws that confuse corruption with free speech and by voters who don’t care about policy details, leaving legislators to continue doing the bidding of donors while riling the troops with identity politics.
If this seems rather hopeless, try technical meddling. That may dull the despair.