Security, Uncertainty, and the National Security Administration: The President Should Defend, Not Revise, Current NSA Procedures

The question of national security and its relationship to individual privacy has always provoked a challenge for persons with strong libertarian inclinations, who rightly embrace a general presumption in favor of small government.  On most economic matters, the group that hold these positions tend to be relatively small, and thus does not count as a large political constituency with which the Congress and the President has to deal on general matters of economic regulation.

The political landscape, however, takes a very different turn on national security interests, when large portions of the American left vigorously endorse the kinds of small government position associated with libertarian writers on the right.  Thus the American Civil Liberties Union is all too true to its name when it criticizes the range of programs that are intended to impose comprehensive surveillance systems on the American public in the name of national security.

How Small is Small Government?  I yield to no one in support of the general proposition that all government programs should be examined under a presumption of error, and that national security issues are not an exception to that general rule.  But the presumption in this, as in all other cases, cannot be absolute. It has to be rebuttable on good cause shown.  To a libertarian, the class of cases that fall within good cause should be narrow.  But it is not empty.  Indeed the one kind of situation in which this initial presumption can be rebutted is those cases that involve an explicit or implicit threat of force against the population of the United States (not to mention people around the world).

As in so many cases, it is instructive to start with the private analogies to the public use of force.  Thus, it has always been held that individuals do not have to wait until they are attacked to use force against a stranger.  They may use force to repel the threat of force in at least some circumstances. Deciding which circumstances allow for the use of force is no easy matter, and the problem only becomes more difficult when the source of the threat is diffused and uncertain, even as the magnitude of the harm, should it come to pass, is enormous.

In a nutshell, this profile describes the issue with national defense.  To say in principle that the government should be paralyzed until it has information about a specific offense before it can throw its might into the ring is to my mind a grievous error because, by that time, it could easily be too late.  Yet to let the government root about for private information without rhyme or reason also poses serious threat to the personal tranquility and financial security of huge numbers of American citizens.

The solution that—until the Snowden leaks—had been widely adopted is that the government can engage in the collection of metadata without showing any form of probable cause for its actions, but that it can only peek inside the envelope once the collected information is shown (at the very least by reasonable suspicion) to be of importance in tracking a particular case.  The use of this information is thus subject to far more stringent standards than the standards that apply to the collection of that information.

The critical factual question is whether the government can develop those institutional safeguards needed to make these separations possible. On several occasions, first with Roger Pilon and then with Mario Loyola, I took the position that there was no credible evidence that the NSA had allowed for systematic breaches in that wall of separation, so that there was in principle no reason to pile on additional safeguards against the government use of information in its intelligence effort. Indeed, the case for that conclusion is unambiguously stronger today than it was when this issue first broke because there has been additional time to find systematic, uncorrected violations of the basic separation norm, none of which have yet been put convincingly forward.  In dealing with these critical error calculations, abstract objections, whether they come from the libertarian left or the libertarian right, get progressively weaker when the huge efforts to find systematic breaches in the current system turn up a succession of blanks.

The Obama Administration’s Two Stage Retreat  It is therefore a source of special regret that the Obama administration has retreated in two stages from its basic position. At stage one, the President commissioned a distinguished Task Force to deliver to him a White House Government Surveillance Report that provides an overview of the basic situation coupled with recommendations for possible improvement.  At stage two, the President made a public speech that proposed certain changes in security practices that, in my view, do nothing to improve the operation and justice or to figure out how to prevent the next generation of leaks, which in my view ought to be the first priority on this exercise.

With respect to the Report, it goes wrong in its first sentence when it insists that it is important to protect “two forms of security: national security and personal privacy.”  The key mistake here is to write as if both forms of security are of equal dignity when they are not.  To be sure, the Fourth Amendment does refer to “[t]he right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures,” but the word security in this context referred in all probability to the situation that the full scale search under a general warrant that had been held in Entick v. Carrington (1765) to expose the government to an action in trespass for damages caused by the mayhem that the defendants imposed on the plaintiff’s premises.

Keep matters in perspective. No one doubts that data protection is an important issue, but it is a flight of fancy to treat the collection of metadata as the equivalent to an all-out body cavity or house search, let alone the widespread devastation that an armed attack could wreak against the United States.  The danger of this comparison is that it leads to a false assessment of the relative cost of two kinds of error, such that people will now give the privacy interest a greater weight that in turn will lead to its greater protection.

At this point, the Report thus compounds two errors.  It attaches too much weight to privacy, and gives too little attention to the weak evidence on abuse that should be at the center of the Report. Many of the recommendations that follow from this incorrect way of setting up the problem thus lead to a weakening of the surveillance system without any material improvement in the overall operation of the system.

To mention just one, there is no reason whatsoever to strip the government of the power to store the metadata on its own computers and to enter into cooperative agreements with the major telephone and data companies to be the prime repository of information.  Divided control over secret information is a presumptive bad for two reasons: First, it leads to a higher probability of leaks given the added vulnerabilities that are always introduced at any new interface. Second, should that leak happen it becomes harder to find out its source, and harder therefore to make needed corrections within the system.  Yet, there is no offsetting privacy gain that justifies this departure for current practices.

The same can be said of the President’s own remarks.  He acknowledges that the Task Force found no evidence of actual abuse, but insisted that “potential” abuse was also a problem.  That observation is trivially correct and practically wrong because it encourages effort to fix things that are not broken in ways that can harm the overall operation of the system.  Thus the first changes by the President are these:

“Effective immediately, we will only pursue phone calls that are two steps removed from a number associated with a terrorist organization, instead of the current three, and I have directed the attorney general to work with the Foreign Intelligence Surveillance Court so that during this transition period, the database can be queried only after a judicial finding or in the case of a true emergency.”

But why? If the third link in a phone call chain should be tracked for operational reasons, what reason is there to stop its use when we already know that the separation between collection and use has been maintained correctly? Any breach in this system should be countered by keeping the separation secure, not in reducing the information that is needed. Second, there is no reason for the president to raise the standard for looking at information to that which requires a judicial finding or a true emergency, if that pairing is meant to slow down access. What is really needed are safeguards against leaks, which was not taken up in his speech.

An Orwellian Moment In conclusion, I am dismayed by the hostile reception that his speech has received from the libertarian side, where it is all too common to denounce his indecisive actions in Orwellian terms.  For those who believe that the collection of data American style reaches that Orwellian level, I suggest that they spend a bit of time in China where in fact every conversation can be tracked in real time by government operatives without any safeguards or compunction whatsoever, where the omnipresent government drives a massive level of civil distress. Or think of the nonstop Soviet systems of surveillance with the same totalitarian effects.  And recall that in these cases, surveillance routinely leads to arbitrary arrest, fine, or loss of gainful employment.  The problem is quite simply that the harsh, near hysterical, denunciations of current NSA practices are blind to these critical differences.

This thermonuclear attack on the current system means that we shall run out of superlatives to deal with actions that are truly dangerous. And it misdirects current popular criticism, which should be placed at the doorstep of the Internal Revenue Service whose serious and systematic abuses now get a free pass. When questions turn on relative rates of error, there is always room for fine-tuning key institutional arrangements. But the full throated denunciation of the NSA sheds far more heat than light on national security issues.  We need to do much better as a nation. 

About the Author(s)

Richard A. Epstein

Laurence A. Tisch Professor of Law at New York University School of Law, Peter and Kirsten Bedford Senior Fellow at the Hoover Institution, James Parker Hall Distinguished Service Professor of Law Emeritus and Senior Lecturer at the University of Chicago Follow him on Twitter (@RichardAEpstein).