I’ve written this column for the Washington Post. Building upon ELB posts I’ve written recently on this topic, it argues in part:

Duverger’s law does not work in party primaries. Political science has not settled on why; one theory is that competition in primaries is not stable enough for strategic dynamics to exert sufficient force. Whatever the reason, it is evident that using the plurality-winner rule for party primaries does not yield two-candidate competitions. Instead, it delivers the kind of irrational outcomes we are witnessing this year, where a political party’s nomination goes to the candidate who lost two-thirds, or more, of the party’s votes.

The harm this irrationality causes is not confined to the party; it affects the public as a whole. Pennsylvania offers an illustration: If Oz ekes out a primary win with 31 percent of the vote, he might go on to win the general election. But doing so wouldn’t vindicate him as the most preferred candidate. November’s voters might have preferred McCormick even more. But they will never get a chance to express this preference, assuming McCormick is knocked out in a primary that fractures among multiple candidates with none receiving even a third of the total. …

… Self-government can tolerate aberrational outcomes if the opportunity for self-correction in future elections is preserved. But when a political faction threatens to repudiate future elections upon attaining power, it is essential that the existing electoral system not award victories to this faction’s candidates who aren’t genuinely the majority’s choice.

Thus, add this to the list of electoral reforms having utmost urgency: eliminate plurality-winner primaries — which either state legislatures or Congress can do. Otherwise, we may lose our democracy not because we made that choice, but because an irrational and anti-majoritarian system produced that result.