My colleagues Ilya, Jim, Roger, and Walter have said most of what needs to be said about the Supreme Court’s recent decision in the DNA sampling case Maryland v. King. So let me just hover for a moment on a point Roger makes. Everyone seems to agree that Justice Kennedy’s majority opinion strains the bounds of language by arguing that the state purpose of “identification” served by DNA sampling arrestees includes establishing a “context” for understanding “who the person really is,” including their “past conduct.” By the same logic, we might justify searching the homes of every drunk driver for evidence of unrelated crimes, since this too would give us a sense of “who they really are,” and whether they have reason to jump bail lest other crimes be discovered. The real argument, disingenuously shoehorned into this rubric of “identification,” is that this is indeed a warrantless search for ordinary investigative purposes, but that once a person has already been legitimately detained, the marginal intrusion involved in a cheek swab is trivial—and the benefit to society of enabling serious crimes to be solved so great—that an exception to the normal Fourth Amendment rules is justifiable. This is, as Roger suggests, a closer call.

Let’s go further and make the argument that Justice Kennedy, determined to cast this as a matter of “identification,” didn’t bother with. He could, after all, have cited to the Supreme Court’s major dog-sniff cases, Place and Caballes, in support of the following argument: The limited DNA profile actually entered into the CODIS database is only useful for matching, not for revealing other sensitive facts about medical conditions or genetic predispositions. In essence, then, this is a search that only reveals whether one is the unidentified perpetrator of a crime—which, like possession of contraband, is a fact in which a person has no “reasonable expectation of privacy.” So one might argue.

The first point to make is that the narrow “if you have nothing to hide, you have nothing to fear” argument doesn’t really work. A murder investigation will naturally involve collection of foreign DNA samples on the victim, which may well belong to persons that had nothing to do with the crime. Thus a search of an innocent arrested persons DNA could easily reveal the existence of, say, an unrelated but secret sexual relationship with the victim, or merely the presence of the searched person at the scene of a crime they had no involvement in. So this is not really a search with no realistic risk of exposing innocent but legitimately private information.

The larger point, though, is that the provisions of the Bill of Rights were meant to avoid precisely this kind of granular case-by-case “balancing” process, to the extent possible. An analogy to the First Amendment may be helpful here. Let’s concede: It is totally plausible that prohibiting Nazis from marching through a community of Holocaust survivors, or the grotesque Westboro Baptist Church from picketing military funerals with signs that read “Thank God for Dead Soldiers” and “God Hates Fags,” would suppress particular instances of speech with no real social value and spare decent people anguish they do not deserve. In a vacuum, probably neither instance of speech would survive a “balancing test.” But the courts correctly protected both nevertheless, because the First Amendment articulates a meta-balancing judgment that we do not want the government engaged in this kind of specific case-by-case balancing analysis of which speech is valuable enough to be protected. The Framers of the Constitution had already done a balancing test about when it is better not to engage in balancing tests.

So it is, I would argue, with the Fourth Amendment. In the short term, it is easy enough to say that a few cheek swabs are a trivial marginal intrusion, even if they sometimes expose innocent private information, compared with the social benefit of catching murderers and rapists. But especially as DNA testing technology evolves, what are the consequences of establishing a massive repository of genetic information about the one-third of Americans who will be arrested by the age of 23—especially if that database disproportionately encompasses poor minorities, many of whom are never convicted of any crime? (Those who ARE convicted, as Justice Scalia’s dissent in King observed, get sampled anyway—so the policy in question here only really makes a difference to the innocent.) How do you “balance” the crimes solved at the margin when samples are taken from people arrested though ultimately acquitted against the creation of an architecture of genetic information-gathering, which may in itself encourage pretextual arrests for trivial offenses to circumvent the need for search warrants for genetic material, whose long-term uses are impossible to foresee? The general attitude of the courts, after all, is that once information or evidence has been legitimately acquired by police, there is no Fourth Amendment barrier to further analysis of that evidence, even if unrelated to the purpose for which it was acquired. (There are, I think, good theoretical reasons to regard this as a mistake, but that’s how things presently stand.)

In this case, then, the Court has invoked the idea of “identification” to obscure what is fundamentally an application of a “balancing test” to a warrantless investigative technique. But the Court is balancing benefits it can see reasonably clearly with costs it cannot. Perhaps, even in the long run, the benefits will outweigh the costs. But the point of the Fourth Amendment is to provide a basis for limiting governmental information gathering that, to the extent possible, avoids saddling the Court with the responsibility for engaging in this sort of utilitarian calculus. It cannot be avoided entirely, of course—that much is implicit in the inclusion of the normative term “unreasonable” in the text of the Fourth Amendment—but it should not be the ordinary grounds for deciding which particular searches are permissible. Sometimes, as Hayek understood, we stick to simple rules, not because they are truly optimal, but because we are not clever or prescient enough to develop more nuanced rules that do better.