Political science research on ranked choice voting (RCV) is not sufficient to draw conclusions regarding voter turnout, elected officials’ accountability to voters, or voter confidence in election outcomes, says Josh McCrain, assistant professor of political science at the University of Utah. However, there are useful insights to be drawn from scholarly work on RCV.

“The empirical political science research on RCV suggests mixed effects,” he says. The evidence in its favor includes support from voters in primaries and big-city elections, potential to reduce election administration costs in some cases, and more candidate diversity in big-city mayoral elections.

On the other hand, McCrain finds no empirical evidence showing that RCV produces better behavior from candidates or elected officials. That is a potential problem for expansion of RCV.

“It is entirely possible … that RCV would decrease the accountability mechanism between voters and elected officials, particularly in the elections that advocates are trying to get it implemented: small, low-salience local elections. This would be worse than maintaining the status quo,” he says.

However, McCrain sees some elections – including primaries – as a good fit for RCV.

“We do not worry so much about these [accountability] problems in primary elections, where RCV seems to be a net improvement,” says McCrain. “There’s some evidence that voters are supportive of the system in large, one-party elections” such as primaries or mayoral elections in major cities.

He concludes that “for high-salience elections that are dominated by one party … voters are more likely to get their preferred candidate in these cases with the fewest negative externalities possible. We should adopt RCV for those elections.”

The entire interview between McCrain and Sutherland Institute’s vice president of policy, Derek Monson, is included below.

Monson: What does the empirical research say about the arguments commonly used in favor of and opposed to RCV?

McCrain: The empirical political science research on RCV suggests mixed effects. There may be some evidence for a minor increase in turnout, but at the cost of increased confusion by voters. There’s some evidence that voters are supportive of the system in large, one-party elections (i.e., primaries and major city mayoral elections) because it enables a better match to vote preferences. There’s also some evidence that RCV can reduce the cost of elections through eliminating runoffs, but this depends on the setting. There’s also some evidence for an increase in the pool of diverse candidates in major city mayoral elections that use RCV, but it is not clear how well this would travel to other cases.

There is no research that I’m aware of that shows RCV produces better behavior in candidates elected under this system; indeed, the existing research would suggest that the complications from RCV would result in worse accountability of elected officials.

Monson: What elements of RCV (or elections in general) does the research not adequately address or cover – due to lack of good data, methodological limitations, etc. – that it would in the ideal world?

McCrain: Three things:

• It is impossible to know the effect of RCV on turnout, as we don’t observe the same election with and without RCV. The pool of candidates that runs for an office is a product of the electoral system, which in turn affects turnout and campaign strategies, and vice versa.
• We have very little evidence of any accountability (i.e., how responsive elected officials are to voters) effects. This is generally hard to analyze anyway, but especially so given the limited rollout of RCV. The prior work on accountability would suggest RCV will, if anything, produce worse behavior in incumbents because it is a more complex system. This, of course, is less problematic in primary elections, where RCV is likely a net improvement.
• Voter confidence in election results. There is, to my knowledge, no good systematic evidence that RCV increases or decreases the confidence of voters in the electoral system. This is primarily due to a lack of systematic, high-quality surveys. This is obviously a major concern given the current political climate, where a major national party is purposefully sowing doubt about the outcomes of elections.

Monson: What findings from the empirical research regarding RCV are receiving inadequate attention or emphasis in the public debate?

McCrain: The advocates of RCV make an argument that suggests it must be better than any alternative. The evidence would not support this.

My view is that we need a better understanding of how a new, complicated electoral system affects the behavior of elected officials before we make such a claim. It is entirely possible, if not likely given prior knowledge, that RCV would decrease the accountability mechanism between voters and elected officials particularly in the elections that advocates are trying to get it implemented: small, low-salience local elections. This would be worse than maintaining the status quo.

The important caveat to this is that we do not worry so much about these problems in primary elections, where RCV seems to be a net improvement.

Monson: Thinking in terms of election policy reform generally, what evidence-based election policy recommendations would you make to lawmakers, and why?

McCrain: I would advise lawmakers not to adopt RCV for small, low-salience elections (such as local elections). We simply do not understand what it’ll produce, and what it produces could be worse than the status quo, given existing knowledge.

For high-salience elections that are dominated by one party (primaries, large city elections) this is less problematic. Voters are more likely to get their preferred candidate in these cases with the fewest negative externalities possible. We should adopt RCV for those elections, especially statewide primaries (such as the presidential primary).