What the Party Primaries Stopped Measuring

In a large share of legislative districts, the general election is a formality. The seat is safe, the district's partisan lean is stable, and the meaningful contest happened months earlier in a nominating election with a fraction of the turnout.

This makes the primary the actual selection mechanism for a substantial part of the legislature, which is a heavy load for an institution that was not designed to bear it.

The electorate is small and unrepresentative

Primary turnout runs well below general turnout, and the people who show up differ from the party's general election coalition in consistent ways. They are older, more ideologically committed, more likely to follow politics as an ongoing interest rather than an occasional civic duty.

A candidate optimizing for that electorate is solving a different problem than one optimizing for the district. Intensity beats breadth. A position that thrills fifteen percent and irritates forty percent can be a winning position in a low turnout multi candidate field.

The general election electorate never gets a chance to weigh in on the choice that mattered.

Plurality winners in crowded fields

Most primaries are decided by plurality. In a field of six, a candidate can win the nomination, and effectively the seat, with under thirty percent of a small electorate. That is a mandate from a few percent of the district's adults.

The obvious responses exist and are in use in a handful of places. Runoffs, ranked ballots, and open or nonpartisan primaries all raise the threshold of support a nominee needs. Each has costs and each has produced arguments about whether the outcomes are better, which is a fair debate to have on the evidence rather than on theory.

The incentive to govern narrowly

The effect does not stop at selection. A legislator whose only real electoral threat comes from a primary challenge governs with that threat in mind, which means the constituency that disciplines them is the primary electorate, not the district.

This is a sufficient explanation for a great deal of behavior that gets attributed to personality. Refusing a compromise that most of the district would support is rational if the compromise is what generates a challenger. The incentive is doing the work.

What reform can and cannot fix

Changing primary rules would help at the margin, and the margin is where a lot of seats sit. It would not resolve the deeper problem, which is that districts have become partisan enough that general elections do not function as a check.

That is partly a drawing problem and partly a sorting problem, and the sorting part is not solvable by any rule. People have moved toward neighbors who vote as they do, and no districting commission can undo geography.

What remains available is the rules of the contest that actually decides. Those are set by state legislatures, which are made up of people the current rules selected, which is the difficulty in a sentence.

The original purpose is worth restating because it explains the failure. A primary was meant to find the candidate with the broadest support inside a party, replacing a selection made by officials in private.

It succeeded at the replacement and the measurement drifted. What a low turnout primary now measures is intensity within a small and unrepresentative slice of the party's voters.

How states structure them varies enormously and the variation drives the result. Closed, open, semi-closed and top-two systems each select a different electorate, and each produces a different kind of winner.

That is a design choice made at state level, usually decades ago, and it is almost never argued about in the terms of what it selects for.

The election code sets the mechanics: who may participate, when, and how a candidate qualifies. Each provision is administrative in appearance and consequential in effect.

Timing does more work than any of them. A primary held separately, in a month with nothing else on the ballot, draws the voters with the strongest attachment, which is precisely the group whose preferences diverge most from the general electorate.

The reporting obligations make the money side traceable, and the traceable pattern is that primary spending has grown while the electorate deciding those contests has not.

The consequence is a general election contested between two candidates each selected by a committed minority, which is a different arrangement from the one the reform was meant to produce.

It is worth being careful about the conclusion, because the obvious one is wrong. This is not an argument for returning selection to party officials, and the failures of that system were real.

The narrower point is that a primary is an instrument with a specific measurement property, and the property has changed while the institution has not.

The administrators who run these contests have almost no say in any of it. Turnout in a primary is a function of the date, and the date is set by the legislature.

Which puts the fix in an awkward place. The people who could change what primaries measure are the people primaries selected, and they were selected by the current arrangement working exactly as it does.