Jamelle Bouie lays out several reasons why popular election of judges is a terrible idea: The need to fund-raise (especially in smaller districts) creates conflicts with the requirement of impartiality; judges start handing down longer sentences to appear “tough on crime” during election season; because judicial elections are usually low-priority, they create the illusion of accountability more than the genuine article… These are all excellent points. But the latter two, at least, seem like symptoms of the more fundamental underlying problem: Ordinary people are generally not in any position to evaluate how well judges do their jobs, and in any event, that job isn’t to satisfy majority preferences. A judge whose rulings always produced the outcome preferred by the majority would, in fact, be a terrible judge.

Though not a lawyer, there are a handful of narrow areas of law I’ve been researching, reading, and writing about heavily for a few years now. I’d like to believe I’ve managed to get a decent handle on those few pet areas. Child, I ain’t passed the bar, but I know a li’l bit. And seeing what a complex and detail-driven process legal decision making is in the areas I do know something about has made me acutely aware of how ill equipped I am to seriously evaluate the quality of legal arguments in other domains. An opinion in a typical case will run dozens of pages (or longer), citing scores of other cases, each with their own equally lengthy opinions and tables of citations. Have the right controlling precedents been identified? Are the fact patterns really relevantly similar, or is there a basis for distinguishing? Have the canons of statutory construction been applied correctly? Usually you can’t tell just by reading an opinion—to really tell whether the reasoning is correct, you’ve got to actually be pretty familiar with all those other cases—and perhaps a bunch more that are relevant but didn’t figure in the decision—as well as all the statutes involved, maybe their legislative histories, as well as all the rules about how, exactly, to apply all those precedents and reconcile conflicts or resolve ambiguities.

Most of us, of course, can’t actually do all that. What we can do is look at the outcome and decide whether we like it or not. But the point of the rule of law is precisely that judges aren’t supposed to make decisions that way: They’re supposed to apply the law neutrally, even if that requires throwing out the case against a criminal who’s probably guilty, or rejecting a sympathetic plaintiff’s claim to compensation, or (for that matter) letting a gang of bigoted sociopaths traumatize a grieving family. And while the Supreme Court enjoys a certain amount of latitude, other judges are supposed to be bound by the rulings of higher courts—even when they think those rulings are profoundly misguided.

A good politician, other things equal, is one who produces the outcomes the majority of people want, and notwithstanding the rich and depressing literature on political ignorance, voters are at least somewhat able to monitor political outcomes. But that’s not at all the test of a good district judge. A good judge is one who engages in the process of legal reasoning skillfully, even when—maybe especially when—the law doesn’t yield the result most people would find more congenial in a particular case. But voters typically don’t evaluate the process—and for the most part couldn’t even if they wanted to. They observe the outcomes, which is exactly the wrong way of assessing judicial performance. In that sense, I might rephrase Jamelle’s last point: Judicial elections aren’t a problem because they provide a mere illusion of accountability to popular will. They’re a problem because judges shouldn’t be accountable on the dimensions that elections naturally track.