This past week, Roger Pilon of the Cato Institute and I wrote an op-ed in that gave a qualified defense of the controversial NSA surveillance program. Libertarians from the left and the right have come together in shrill opposition to the wiretapping program; they object to the government’s collection, retention, and examination of sensitive individual data.
Pilon and I agree that all government actions should be examined under a presumption of distrust. In many domains, that presumption should hold firm. It is for that reason that we have both defended constitutional regimes that afford strong protections to economic liberties and private property—at least in cases where matters of national security are not on the table. For instance, with the minimum wage, there is no coherent case for state intervention. The state should never spend public resources to degrade the efficiency of private markets.
Yet, not all cases are solved by our categorical rule concerning distrust of the government. According to libertarian theory, the bedrock function of government is to put in place institutions that limit aggression in both domestic and foreign affairs. The question is: Are the means chosen reasonably adapted to the stated ends? Any effort to squelch all government abuse will result in a field day for terrorists and criminals. It is never easy to set the right balance of government power versus individual liberty given that we have to contend with two kinds of error—letting terrorists and criminals escape or infringing upon the rights of innocent individuals.
Most libertarians are deeply suspicious of balancing tests because they do not have the hard-edged quality of fixed rules. But unfortunately, balancing is all that we have when it comes to having the government respond to uncertain future events. The only issue is how best to balance. One type of balancing involves an open-ended “facts-and-circumstances” test that leaves a lot to the imagination. This is the kind of test long used to determine whether police have “” to issue an arrest or search warrant, or whether testimony elicited in some custodial interrogation is sufficiently “voluntary” to be admissible into evidence for some purposes within the criminal system.
These tests tend to survive, not because they are perfect, but because courts develop certain niches in which a per se rule is applied: mere suspicion created by information from an untested informant does not count as probable cause in Fourth Amendment cases, but suspicion created after close observation of the subject by a reliable informant would.
In many cases, however, it is possible to use more categorical balancing tests that afford greater protection for individual liberties. In order to put the question of surveillance into context, it is useful to compare it with two other cases that raise issues of government trust. The first involves detention and possible criminal prosecution. The second involves the so-called “Terry Stops” of individuals in public places. As we progress from the first (detention) to the issue of the day (surveillance), the security interest becomes progressively stronger and the claims of individual liberty and privacy become progressively weaker.
Detention and Individual Liberty
A central question of American anti-terrorism policy is what should be done with individuals, often aliens captured overseas, who are detained on suspicion that they have engaged in terrorist activities. In these cases, the intrusion into individual liberty is always large: indefinite detention is a major limitation on personal liberty. Aggressive interrogation can often expose detainees to the risk of degrading treatment or even actual torture.
I dealt with this issue firsthand when I served as a member of the for nearly three years. Its report of April 16, 2013 was able to pinpoint major deficiencies in government detention, including instances of detainee torture that were painfully documented by a team led by Neil A. Lewis, its thorough and fair-minded lead investigator.
In these cases, it is hard to gainsay the magnitude of the interest in individual liberty. No one should ever underestimate the risk of government abuse by torture, much of which turns out to be virulently counterproductive. Clearly, much direct oversight is needed.
But once you move beyond oversight, the operational choices become harder and call for further balancing. The majority of the Constitution Project Task Force favored closing Guantanamo and releasing all detainees who were not tried (preferably tried in federal court). My view was the opposite. Closing Gitmo did not make sense if done only for symbolic reasons. Better to reform the facility than to send detainees off to Bagram Air Force Base or some CIA dark hole.
Nor does it make sense to release all detainees who cannot be tried. There could be risks of serious harm that cannot be established beyond a reasonable doubt. That high standard does not apply to these detainees, who are enemy combatants and not ordinary prisoners, who receive much stronger constitutional protections in the criminal justice system.
By the same token, we should be aware of the limitations of habeas corpus if it only requires the government to justify the initial detention and nothing more. For these reasons, so long as indefinite detention is appropriate, there must be a correlative obligation of periodic status reviews before independent parties (which can be done within the system of military justice) to see whether the reasons that brought detainees to Gitmo still applied. It is critical to invest substantial resources to make sure that no person, citizen, or alien is unduly detained.
“Stop and Frisk” and the NYPD
is a class action case now pending in the Southern District of New York before Judge Shira Scheidlin challenging the stop and frisk practices of the New York City police. In Terry v. Ohio, a 1968 Supreme Court decision, the liberal Warren Court, with much agony, decided that it was permissible for a police officer to stop and frisk a suspect on the public streets without a warrant to see if he was in possession of a gun, so long as there was “reasonable suspicion” that the person intended criminal activity.
In both of these cases, the level of intrusion, while far from trivial, is lower than it is in detention and torture cases. Likewise, the police interest, although substantial, involves street crime, not national security.
Terry’s reasonable suspicion test is one way to balance the conflict. The phrase “reasonable suspicion” does not appear in the Constitution, but represents Chief Justice Warren’s recognition of the modest nature of the intrusion and the strong—some might say imperative—need to keep public order, balanced against invasions of privacy. Floyd is a prolonged effort of the Center for Constitutional Rights to obtain injunctive relief and individual damages for excessive police intrusions.
Passions are clearly high on both sides of the case. There is little question that a disproportionate amount of the New York Police Department Searches are targeted at black males aged 14 to 21. Jacob Sullum, writing in has made a powerful case that the program looks like unconstitutional overreach. In his words, “As the number of stop-and-frisk encounters initiated by the NYPD from about 100,000 in Michael Bloomberg's first year as mayor to almost 700,000 in 2011, the share of stops yielding guns fell from 0.38 percent to 0.033 percent.”
At the same time, the crime rate in New York City fell by about 50 percent. On that score, Heather MacDonald wrote an equally powerful op-ed in the claiming that successes of this sort are always prey to excessive judicial oversight that curtails police discretion and leads to a sharp increase in black-on-black crime.
So who is right? Well, it turns out that both are, in different ways. Empirically, proof of correlation is not proof of causation, so that the decline of crime in New York may be attributed to other initiatives, like more police officers on the street or strengthened community involvement, rather than stop and frisk. This is a natural social experiment of sorts. The NYPD, perhaps in response to the Floyd case, cut back on its stop and frisk activities by about 50 percent in the first three months of 2013 relative to the same period in 2012. The result, as reported by in the Wall Street Journal, is that “the city's overall crime rate continued to go down by 2.7% during that time and murders were down 30% from a year earlier.”
Two implications follow from this. First, it looks as though the critics are right about the practice. Second, it looks as though defenders of the NYPD like MacDonald are correct that no federal oversight program, such as that recently volunteered by an overly intrusive Attorney General Eric Holder, is needed, precisely because, given the massive publicity and debate surrounding the issue, self-corrective measures seem to be in place. If this evidence is right, the Terry problem should be self-limiting so long as the NYPD realizes the counterproductive nature of its intervention; in other words, a responsible balancing is already taking place.
Finally, the NSA surveillance program signals a strong categorical shift in the balance to the government side. In these cases, the government has collected information that might be useful in tracking down future terrorists.
So far, critics of the program have been far too glib, as in Jim Harper’s on Cato@Liberty. Harper suggests that the very act of data collection brings back the sordid British practice of issuing a “general warrant,” which is the very abuse that the Fourth Amendment was intended to curb. But such a statement is simply not accurate: What sparked the Fourth Amendment was the 1765 case of , in which Lord Halifax, a Secretary of State, ordered his agents, one of whom was Carrington, to ransack John Entick’s premises, which they did for four hours collecting charts and pamphlets, and causing damage of some £2,000, which Entick was eventually allowed to collect.
All the evidence suggests that nothing remotely like this outrage has taken place. In the NSA case, the key point is that it takes a search warrant to examine the contents of the phone calls collected. Yet, there is no evidence of any breach of the line between collection and use.
A recent says the government estimates that about 300 phone calls were probed. Why should we presume that this claim is false? Does anyone think that NSA agents are petty voyeurs, that they could even use their equipment for unauthorized use, or that the NSA is institutionally insensitive to privacy breaches? Besides, there are credible stories that tracking actually works in some cases, such as when the FBI uses cell phone transmissions from speeding escape cars to .
I don’t always agree with Alan Dershowitz, nor does he always agree with me, but I think that he is right on the money when he laments at that, with the outcry against the NSA program, we are witnessing a return to a form of paranoia that has too often marred American politics. Dershowitz here is not arguing whether we do or do not need a government program; he is describing the level of trust that we put in government.
In making that observation it is imperative to distinguish between cases. Nothing whatsoever should insulate the NSA from political scrutiny and legislative and judicial intervention. But nothing should allow us to equate the so-called NSA standard with the inexcusable IRS scandal that is rife with partisan politics and worse, precisely because of the utter absence of any institutional protections against partisan abuse.
Richard A. Epstein, the Peter and Kirsten Bedford Senior Fellow at the Hoover Institution, is the Laurence A. Tisch Professor of Law, New York University Law School, and a senior lecturer at the University of Chicago.