With the 2016 presidential election already dominating political news, it’s easy to forget that there’s an election day this year, too. Depending on where you live, on November 3rd, you may have the chance to vote for the mayor, or members of a school board, or some other local office. In a lot of cases, the ballot won’t mention these candidates’ party.

That makes local elections an interesting test case for the popular argument that partisan polarization is alienating voters and messing up democracy. Organizations like No Labels call for candidates and voters to set aside their differences and become “problem solvers,” pursuing straightforward, practical solutions to the nation’s problems.

In a 2001 paper, Brian F. Schaffner, Matthew Streb, and Gerald Wright examine how nonpartisan elections work in practice and find results that are a lot less inspiring than rhetorical promises.

Like the No Labels types today, Progressives at the turn of the twentieth century saw party labels as too divisive. They pushed for nonpartisan local elections, arguing that this would curb the power of party machines and allow informed citizens to choose the most competent managers for their cities and towns.

Schaffner, Streb, and Wright compared similar municipal and state governments that use nonpartisan and partisan ballots, as well as two cases where local governments switched between the systems. They found that potential voters in nonpartisan elections are more likely to skip the vote entirely. If they do vote, they’re more likely to go for the incumbent.

Why would this be? Pointing to modern political science theory, the authors note that researching candidates takes time and effort. Most voters look for shortcuts to determine whose positions are closest to their own, and party affiliation provides a pretty reliable indicator of where someone is likely to stand on a range of issues. This means partisanship can help average people—who, by and large, just aren’t going to spend hours learning about a candidate for state representative—collectively make a fairly rational decision.

If voters don’t have a party label to go by, the authors found, they’re likely to use another easily available piece of information: whether the candidate is an incumbent. They also suggest, based on previous research, that other ways people choose candidates in a nonpartisan election are ethnic or gender-specific names.

Schaffner, Streb, and Wright note that the relatively low turnout in nonpartisan elections probably wouldn’t have bothered the Progressives, who were more concerned with having a well-informed electorate than a large one. If we see elections as a way to choose the cleverest manager for the government, this may make sense. But if we see it as a contest of ideas, in which everyone ought to be involved in pushing for the policies they like best, partisanship has some clear advantages.