It turns out that a New York Police Department program that assembled large databases on the ordinary activities of innocent Muslims, infiltrated student groups, and monitored sermons wasn’t just controversial—it was useless. As the Associated Press reports, the head of the NYPD’s Intelligence Division recently confirmed in a deposition that the Department’s “Demographics Unit”—a delightful euphemism for a team dedicated to spying predicated wholly on ethnicity, language, and religion—turned up no useful leads and gave rise to no terrorism investigations in its six years of operation.

At the risk of being a broken record, this is a reminder of how misleading it can be to discuss these topics under the rubric of “balancing liberty and security.” If government surveillance performs as advertised and yields a substantial security benefit, there’s a debate to be had over how much government intrusion we’re prepared to countenance as the price of that security. But that security benefit has to be proven, not assumed. If it can’t be demonstrated—and a fortiori, if all available evidence demonstrates there is no benefit—then it just should not be a serious question in a decent society whether it’s acceptable for police to keep tabs on all Urdu‐​speakers of Pakistani origin on the premise (endorsed by this official) that “most” of them are people “of concern” to the government. Just to put that “most” in context: There are some 15 million Pakistani Urdu‐​speakers worldwide, and about 50,000 legal residents of Pakistani descent in the New York metro area. Treating them all, by default, as potential terrorists who need to be watched would be offensive and ugly even if the policy occasionally yielded a useful piece of information. But to squander scarce law enforcement resources targeting a minority population without any useful results over six years? How can that be anything but obscene?