What is “Centripetalism” and Why Does It Matter?

A philosophy of democracy, at least for periods of dangerous polarization

“Centripetal” is a word to describe forces that bring dispersed objects together, towards the center. It is the opposite of “centrifugal,” which describes forces that drive objects apart, away from the center.

In politics, then, “centripetalism” is a philosophy that advocates institutional structures and arrangements that bring divergent groups together, so that governance of society can be based on a broad coalition encompassing the political center. In a democracy, centripetalism advocates electoral procedures that produce winning candidates whose broad support among voters is anchored in the center of the electorate.

Centripetalism, to the extent that it has been developed as a distinctive philosophy of democracy, has been focused on deeply divided societies with emerging and fragile democracies, which are often riven by seemingly intractable racial or religious conflict. In this context, centripetalism often advocates constitutional provisions designed to prevent the dominance of one racial or religious group over others by reducing the salience of race or religion in politics.

Recently, however, the idea of a centripetalist philosophy of democracy has been applied to so-called “advanced” and longstanding democracies, like those in Europe and the United States. As these democracies increasingly suffer from acute polarization, political scientists (and others) are turning to centripetalist theories as a way to counteract this polarization and protect these democracies from the centrifugal effects of ascendant extremism. The Australian political scientist Ben Reilly has been a leading proponent of this approach.

Centripetalists, including Reilly, have argued that ranked choice voting can be an effective means to combat polarization and elect candidates who are closer to the center. The evidence from Alaska’s new electoral system, which combines all-candidate primaries (where candidates from all political parties, plus independents, compete against each other) with ranked choice voting in general elections, suggests that this kind of electoral reform can indeed have centripetal effects, causing the defeat of more polarizing and extreme candidates. Most prominently, in 2022 the defeat of Sarah Palin and the election of Mary Peltola to Congress (and, separately, Lisa Murkowski in the Senate election), seemed to show how ranked choice voting can serve the cause of centripetalism.

As promising as these preliminary results are, centripetalism faces additional issues in its quest to counteract polarization and extremism. As Common Ground Democracy readers already know, there are different forms of ranked choice voting. Alaska uses “Instant Runoff Voting,” which operates by eliminating the candidate with the fewest first-choice votes on the ranked ballots and then redistributing the ballots that ranked the eliminated candidate first to whichever other candidate is ranked second on those ballots. By contrast, Convergence Voting would use the ranked ballots to elect a candidate who is preferred by a majority of voters to each other candidate on the ballot when candidates are compared one-on-one. (As one of Alaska’s 2022 elections demonstrated, the Instant Runoff Voting winner is not necessarily the same as the Convergence Voting winner.)

There are reasons to believe that Convergence Voting better serves the values of centripetalism than Instant Runoff Voting. Consider the example, which I have often used, of Ohio’s 2022 Senate election. There, JD Vance—now Donald Trump’s running mate—defeated Tim Ryan, the Democrat. If Ohio had used Alaska’s electoral system for that race, and if then-incumbent Rob Portman had been willing to run for reelection with that system in effect, Vance almost certainly still would have won. Portman would have received the fewest first-choice votes on the ranked ballots, given the nature and degree of partisan polarization in Ohio, and thus he would have been eliminated using the Instant Runoff Voting method to tabulate the ranked ballots. The result would have been the same polarized victory for the extremist Vance.

By contrast, if Ohio had used a modified version of Alaska’s system that employed Convergence Voting as the method for tabulating the ranked ballots, and if Portman had run in that system, then Portman almost certainly would have won. He would have been preferred by a majority of voters to Ryan, gaining the support of Vance’s voters as well as his own. Likewise, he would have been preferred by a majority of voters to Vance, gaining the support of Ryan’s voters as well as his own. In this way, Convergence Voting—unlike Instant Runoff Voting—counteracts polarization and results in the defeat of the extremist candidate, Vance. This example shows how Convergence Voting is designed to achieve the most centripetal result, electing the candidate who is closest to the center of the electorate by winning the votes of different majorities that converge upon this candidate in the middle. Instant Runoff Voting, however, achieves the most centripetal result (when it does) only by happenstance, not intrinsically, and fails to do so when the electorate is especially polarized (and thus when there is arguably the greatest need for an anti-polarization outcome).

Notwithstanding the attractiveness of Convergence Voting from the perspective of centripetalism, there are important philosophical issues to consider.  Convergence Voting can elect candidates with little first-choice support. Convergence Voting in effect favors compromise by denying victory to either side of a deeply polarized divide, requiring instead that both sides of the divide converge upon an acceptable second-choice candidate. Instant Runoff Voting, conversely, eliminates the compromise candidate in the middle, who has few first-choice supporters, and then requires those few first-choice supporters of the centrist candidate to pick one side or the other of the polarized divide (between the two candidates with the most first-choice support). Is Convergence Voting’s structural preference for compromise the best way to operate a democracy? Or is it better to determine which of the two most popular candidates, measured by their first-choice support, is preferred by a majority of voters when less popular, including potentially compromise, candidates are eliminated?

With these questions in mind, let’s consider the current presidential race. As of today, on the eve of the Democratic National Convention, this most volatile of presidential election years has yielded a neck-and-neck contest between ex-president Trump and Vice President Kamala Harris (after President Biden’s exiting the race). Although Harris has clearly captured the enthusiasm of previously dispirited Democrats, and reversed the momentum of the race, it is essentially a toss-up at this point, as virtually all pundits proclaim, and either candidate could win.

One astute commentator observes that the election could be decided by where Nikki Haley’s voters end up. Haley’s voters were Republicans who refused to support Trump in the primaries, often in significant percentages even after Haley dropped out. Haley endorsed Trump at the Republican convention, but it is unknown what percentage of her voters will follow her.

In considering these, we can imagine a ranked-choice ballot with Trump, Haley, and Harris all on it. I described a redesigned “Top Three” presidential election system in a previous Common Ground Democracy essay, and I’m asking us to contemplate the hypothetical situation in which the system were used for this year’s election—and those three candidates were the ones in contention under that system. What would be the first-choice votes for these three candidates? We don’t have perfect polling on this point, but given the current polling averages of the head-to-head between Trump and Harris, we can surmise that it might be something like Trump and Harris with 45% each, and Haley with the remaining 10%. It would of course vary somewhat in each of the battleground states, but this simple conjecture is useful for our thought experiment.

If Instant Runoff Voting were used to tabulate these hypothetical ranked-choice ballots, then Haley would be the candidate eliminated, and the election would be decided based on whether more Haley voters support Trump or Harris. On the other hand, if Convergence Voting were used to tabulate these ranked-choice ballots, then Haley presumably would win: Haley would get the second-choice votes of Harris’s supporters to be majority-preferred against Trump, and Haley would get the second-choice votes of Trump’s supporters to be majority-preferred against Harris.

This example illustrates the crucial question that a centripetalist philosophy of democracy must address. Should Haley really win the election if she is the first-choice candidate of only 10% of voters, and 90% have a first-choice preference of either Harris or Trump? It seems odd to elect Haley when so many more voters have a different first-choice preference.

Yet in a three-way race that includes Haley along with Harris and Trump, if neither Harris nor Trump have a majority of first-choice votes, then shouldn’t Haley being a compromise candidate prevail as long as she is preferred by a majority of voters against each opponent considered separately one-on-one? Of course, if Harris’s momentum through the remainder of the campaign were to cause her to have over 50% first-choice votes against both Trump and Haley on the ranked-choice ballots, then Harris clearly would deserve win as the most centripetal candidate. But if the polarization between Harris’s supporters and Trump’s is such that neither candidate has the outright support of a majority, then from the perspective of centripetalist philosophy is it not better to elect a depolarizing compromise candidate who is the second-choice preference on both sides of the divide? (To be sure, had a Top-3 system with Convergence Voting been in place at the beginning of this presidential election year, the three finalists might have ended up not Trump, Harris, and Haley but instead some other combination—like Trump, Joe Manchin, and Gavin Newsom—but we need not consider that possibility for the purposes of this simple thought experiment.)

Additionally, in evaluating this fundamental philosophical issue about the best way to choose a winner when the electorate is polarized in this way, we must also consider the risk to democracy itself—specifically, of electing a candidate who is hostile to the very premise of declaring winners based on an honest counting of the ballots cast. The consequence of electing Trump over Harris is not only a matter of whether the voters are best represented by that outcome (given their electoral preferences among the candidates who ran for president this year), but also a matter of whether the ongoing operation of democracy can sustain itself from one election to the next if a would-be autocrat regains the powers of the presidency. Adopting a Top-3 presidential electoral system with Convergence Voting, which had it been in effect this year would have hedged against this type of risk by electing Haley rather than Trump or Harris if neither of the latter two were to reach a majority against both opponents combined, might be better for preserving democracy over the long run than an electoral system (like either the existing one or Instant Runoff Voting) that potentially elects Trump over Harris once Haley is removed from the race. Again, if Harris can achieve over 50% against both Trump and Haley combined, this risk doesn’t exist, and it doesn’t matter which electoral system is used. But if Harris is under 50% when up against both opponents together, then eliminating Haley from contention leaves the risk of ending up with Trump as the winner when a majority of voters would have preferred Haley to him.

We obviously cannot change the system for this year at this late date, and so we will need to see how the Trump-Harris match plays out with Haley no longer in the mix. After this year’s election is over, hopefully the country will be in a position to consider what electoral reform would be best to pursue for the future. If and when that occurs, the centripetalist philosophy of democracy—including its potential implications for choosing winners in multi-candidate elections—should be part of that conversation.