What we, as political scientists, saw in this election was a strong and predictable national backlash against the party holding power in Washington — another case of the phenomenon that has produced losses for the president’s party in 11 of the last 12 Virginia gubernatorial elections. (The state’s contests have been called “correction elections” because of their timing relative to the presidential race.) But it is easy to squander a favorable national mood by picking a problematic nominee. In 1994, a very poor year for Democrats across the country, scandal-plagued Sen. Chuck Robb was spared defeat because Virginia Republicans chose Oliver North, a central figure in the Iran-contra affair, as their standard-bearer.
Something similar happened in the 2013 governor’s race, which otherwise had much in common with this year’s. At the time, President Barack Obama had a 52 percent disapproval rating, roughly the same as President Biden’s rating today — a result of strong opposition to the Affordable Care Act and the botched rollout of the health-care exchanges. But that year, McAuliffe managed to win, 48 percent to 45.5 percent, over the Republican firebrand Ken Cuccinelli II.
2013 and 2021: Two elections, same Democratic candidate, similar low polls for the president. The difference in outcome, we believe, came down to the Republican nominee. In 2013, the Virginia GOP selected Cuccinelli at a convention swarmed by right-wing tea party delegates. Seeing the writing on the wall, the mainstream candidate, Lt. Gov. Bill Bolling, pulled out before the convention.
To choose candidates for state office, the Virginia GOP sometimes runs a primary and sometimes holds a convention. To reduce the chances that a Trumpist candidate would run away with the nomination, party insiders this year pushed for a convention. (Because of the pandemic, it ended up being an “unassembled convention,” with delegates voting in 39 sites around the state.) To be sure, extremist candidates can also prevail at conventions, as such gatherings attract the most ideologically committed party members. But Virginia Republicans offset that tendency by instituting a system that let the delegates — some 53,000 — list their preferences in ranked order for all seven gubernatorial candidates, and requiring that the nominee secure a majority of votes (not just a plurality). That process forced candidates to reach out to the broadest possible spectrum of delegates, putting a thumb on the scale against the most extreme populism.
Under a ranked-choice system, when the ballots are tallied, the last-place candidate is removed and the second choices of his supporters are distributed to the other candidates. This process continues until one candidate is left standing. There are several benefits to doing things this way. First, voters can vote sincerely for their favorite candidates — even sure losers — because they know that their subsequent choices will be counted. More important for the quality of the selection, requiring that candidates do more than eke out a modest plurality makes it less likely that a factional candidate disliked by most of the party will prevail.
As it happens, Youngkin did not have to rely on ranked-choice voting alone. He led by a plurality of 33 percent in the first round and held the lead throughout. His vote totals increased substantially by the last round, when he won with 55 percent of the delegates, signaling broad support among the people who would be pounding the pavement for him.
Youngkin was not the most moderate candidate; ranked-choice voting does not necessarily reward centrism per se. Kirk Cox, a state delegate, was the moderate in the race — the only candidate to state clearly that President Biden won the 2020 election fairly. But nor was Youngkin the most closely linked to Donald Trump. That description fits state Sen. Amanda F. Chase, who won 21 percent of the vote in the first round but failed to rise above 25 percent in later rounds.
Had the Virginia GOP held an ordinary primary, Chase might well have won — or at least, her attacks on Youngkin might have left him wounded in the general election. Chase might have also fared better in the pressure cooker of an in-person convention held under one roof — because Trump supporters are more likely to show up, and can be vocal and aggressive in support of their candidates. (That the pandemic prodded Republicans to distribute voting across the state was a lucky break.) In the end, Youngkin straddled quite deftly the Trump-loving base and pragmatic GOP-leaning voters in the suburbs, even as he made headway with persuadable independent voters.
RCV is not foolproof against bad candidates, and there might be other ways to choose nominees that incentivize the coalition-building on display in Virginia. In Alaska, voters recently passed a measure that requires the use of a nonpartisan primary system in which all candidates compete on a single ballot. The top four vote-getters move on to the general election (where voters use RCV to choose the winner). That process, too, may reduce the odds of extremist candidates winning.
Modern American primaries, in which rank-and-file party members choose candidates through direct election, have few parallels in other democracies. Indeed, research suggests that they likely contributed in recent decades to the decline in the quality of candidates running for office. In the past, party bosses wrangled among themselves to find the consensus nominee, which helped keep out extremists, incompetents and dangerous candidates. But smoke-filled rooms are relics, and voters have a rightful say in party matters, along with insiders. RCV offers a middle way. With RCV, an algorithm does the “bargaining” and the screening. It’s a reform that Democrats and Republicans in other states should be considering — but especially the GOP, given its perilous and electorally risky descent into Trump-style politics.