Our recent essay “The Rules of the Game: A New Electoral System” [NYR, January 19] provoked thoughtful responses from many readers—in letters to The New York Review, in blog postings and columns, and in private communications. We are grateful to the Review for giving us the chance to reflect on some of the ideas that came up, and also to say something about the French presidential election.
Our essay proposed two improvements to US presidential elections. First, in both presidential primaries and the general election, we would replace plurality rule (in which each voter chooses a single candidate, and the candidate with the most votes wins, even if he or she falls short of 50 percent) with majority rule (in which voters rank candidates, and the candidate preferred by a majority to each opponent wins). Second, we would reform the Electoral College so that nationwide vote totals rather than statewide totals determine the winner.
Currently, all but two states rely on both plurality-rule voting and a winner-take-all system to award Electoral College votes: the candidate with the most votes, no matter how far short of a majority, wins the state and gets all of its electoral votes. By contrast, two states, Maine and Nebraska, use plurality-rule voting but a proportional system to award Electoral College votes. In either case, however, plurality-rule voting is seriously vulnerable to vote-splitting, which arises when candidate A would defeat candidate B in a one-on-one contest, but if candidate C (who appeals to some of the same voters as A) also runs, then A splits the vote with C, giving B the victory.
Vote-splitting has had a profound influence on many presidential elections, for example, in 2000, when Ralph Nader took votes from Al Gore, enabling George W. Bush to win; in 1992, when Ross Perot cut into George H.W. Bush’s support, allowing Bill Clinton to prevail; and in 2016, when Republican candidates such as Marco Rubio, John Kasich, and Ted Cruz divided the mainstream Republican vote in the early primaries and thus gave outsider Donald Trump a path to the nomination.
In view of the unhappy history of plurality rule, some readers have suggested instead using runoff voting, another well-known voting system. Under runoff voting, each voter again chooses a single candidate, but if no candidate gets a majority, the two top vote-getters face each other in a second round. This is the method used for electing presidents in France, but as French history shows, it too is highly subject to vote-splitting.
On April 23, Emmanuel Macron and Marine Le Pen finished first and second in the first round of the French election, and as a result faced each other in the May 7 runoff. However, most available evidence shows that if the third-place finisher, François Fillon, had faced Le Pen head-to-head, he would easily have won (even the fourth-place finisher, Jean-Luc Mélenchon, would quite possibly have beaten her one-on-one). Thus the fact that Macron faced a runoff against Le Pen, as opposed to against Fillon or Mélenchon, seems anti-democratic. (And Le Pen’s post-election claim that she is the main opposition to Macron is clearly inaccurate.) As an extremist, she had been able to “divide and conquer” her way into the final round.
Macron, who was elected president decisively in the second round with 66 percent of the vote, seems likely to be the true majority winner; one-on-one, he defeated Le Pen and probably would have done the same against the other candidates. But French elections don’t always produce a winner who has the most overall support among voters. In 2002, for example, Socialist candidate Lionel Jospin failed to advance to the runoff because he split the left-wing vote with several others and finished third, while incumbent president Jacques Chirac and National Front leader Jean-Marie Le Pen (Marine’s father) came in first and second, respectively. Chirac handily defeated Le Pen in the second round, but the shocking thing was that Le Pen was in the runoff rather than Jospin. Not only would Jospin have easily defeated Le Pen in a two-man race, but he might have beaten Chirac head-to-head as well. There’s a good chance that the wrong man—in this case, Chirac—was elected president.
By contrast, majority rule avoids such vote-splitting debacles because it allows voters to rank the candidates and candidates are compared pairwise: if a majority of voters rank candidate A ahead of B, this ranking holds whether or not C runs too, and so there is no sense in which C can take votes away from A. Several readers have suggested going a step further by having voters grade candidates (say, on a scale of 1 to 5) and electing the candidate with the highest average score—much as gold medals are awarded in Olympic diving. But there is a big difference between grading in the Olympics—where standards are clear and judgments reasonably impartial—and grading in politics, where criteria are highly variable and personal. Thus we doubt that grading schemes could work successfully in political elections: grades would have no common meaning, and voters would have strong incentives to distort the grades they award candidates.
The most obvious rationale for reforming the Electoral College is to make it conform to the principle of “one citizen, one vote” (as one reader put it). The Electoral College under current rules violates this principle; a vote by a Californian doesn’t count the same as one by an Ohioan. A number of other readers have pointed out, however, that there is a more subtle reason for reforming the Electoral College, one connected to majority rule.
Because it reduces vote-splitting, majority rule would encourage more major candidates to run in the general election. For example, under the existing system, Michael Bloomberg and Bernie Sanders had a powerful disincentive to run as independent candidates in the general election last fall because of the overwhelming likelihood that they would have split Hillary Clinton’s vote and handed the election to Donald Trump. But under majority rule, they could have run without this fear.
There is a risk that the presence of additional major candidates might prevent any one of them from getting 270 votes in the Electoral College. This could be avoided by amending the Electoral College system so that the winner is the candidate who wins the nationwide vote under majority-rule voting. Such a change could be instituted, for example, by revising the National Popular Vote Interstate Compact initiative, in which a state pledges to award its electoral votes to the winner of the national popular vote as long as states totaling at least 270 electoral votes make the same pledge. (The compact has already accumulated states worth 165 electoral votes.)
Specifically, we suggest that the national popular vote winner be defined as the national majority-rule winner (not the plurality-rule winner). Such a winner can be said to truly reflect voters’ preferences. In our view, this is the most important reform to aim at.