Geoffrey Skelley at FiveThirtyEight recently wrote an article challenging the premise that primary elections reform will make politics less divisive. Skelley draws heavily on a new report from New America’s Lee Drutman.

To be sure, Skelley gets a number of points right. Our politics is suffering from highly partisan electorates as a result of partisan sorting. This means there is often no meaningful ideological gap between primary and general election voters. Primary reforms such as open primaries and top-two primaries fail to address many of the key problems with hyperpartisan politics.

But Skelley also makes several key mistakes in his analysis of top-four voting, a close relative of final-five voting, that works against his thesis. Unlike our current “first-past-the-post” elections or the top-two primaries in California and Washington state, these systems would place all candidates, regardless of party, on the same ballot in the primary before advancing either four or five of them to the general election, where ranked-choice voting would be used to determine the winner.

Skelley contends that:

“[A] top-two primary in a deep blue or red district sometimes sends two candidates from the dominant party to the general election. In that situation, reformers expected voters from the other party to support the more moderate contender, but that hasn’t really panned out. Instead, voters from the other party often don’t bother voting because they may struggle to differentiate between the candidates from the dominant party. In other words, a Democrat may see two Republican candidates as being two sides of the same coin and choose to abstain; similarly, a Republican may have the same reaction when two Democrats are on the ballot.”

This analysis is disappointingly incomplete and fails to account for evidence to the contrary.

First, undervoting (when a voter decides not to vote for a particular office) does not diminish the strategic incentive for candidates to reach out to crosspartisan voters for support. As Skelley notes, in a race winnowed down to, say, two Democrats under a top-two system, Republicans may be less likely to vote at all or may feel indifferent about the two candidates. But as long as some Republicans are willing to vote for the more centrist Democrat, each Democratic candidate continues to have a substantial incentive to seek out centrist Republican support.

While it is just one example, this phenomenon played out recently in Washington state, where two major offices winnowed down to two Democrats in the general election: lieutenant governor and the House seat for the 10th Congressional District. In both cases, the candidates that presented themselves as more centrist won the race.

Second, much of Skelley’s argument rests on the assumption that the effect of top-two voting would be the same as top-four voting (or final-five voting). It is unlikely, however, that undervoting would occur to the same degree under these latter systems. While it is true that many voters are likely to rank only candidates that share their own political persuasions, the true value of these systems is that they ultimately would encourage legislators to behave in ways that maximize their voter base.

For example, in a “noncompetitive,” highly partisan district utilizing a standard winner-take-all election, majority-party politicians are incentivized to focus on “red meat” policies that earn the favor of their partisan bases. That’s because, in many districts, the winner of the primary is virtually assured of winning the general election. As Skelley notes, districts are becoming increasingly polarized due to partisan sorting, so it is completely rational for candidates to ensure they don’t get outflanked on the partisan extreme.

In a top-four or final-five system, however, elected officials not only need to shore up their base, but they also must ensure that no candidate from their own party outflanks them from the middle. It is hard to overstate how significant a departure from the status quo this change would be.

The benefits extend well beyond the context of campaigns. Politicians elected to office by having to appeal to the middle would be incentivized to work more with centrists and others across the aisle once elected, too. If voters are judging candidates based on actions while in office and more voters have a say in who represents them, simply posturing to appeal to one’s base will no longer be sufficient to ensure reelection.

It is true that, in a reformed primary system, the most pragmatic candidate would not always win. And many politicians (especially incumbents first elected under first-past-the-post) might not respond fully to a change in their political incentives. But they would ignore the new reality at their own electoral peril. The purpose of primary reform is not to change every outcome — it is to put in place incentives that encourage elected officials to pass sensible policies while earning support from the broadest base possible.

No reform is a silver bullet that can fully resolve the problem of political polarization. But primary reforms can indeed nudge our politics in a less divisive, less partisan direction. And that makes them worth giving a shot.