Monthly Archive for December, 2005

Suzanne Spaulding on the Domestic Surveillance Program

Suzanne Spaulding has written an excellent article in today’s Washington Post, discussing (among other things) the administration’s key legal arguments as to why the president was not bound by FISA in authorizing large scale domestic NSA surveillance. She finds:

  1. That the AUMF of September 18, 2001 fails to provide the president with statutory authority to circumvent FISA; and
  2. That Article II of the Constitution fails to provide the president with the inherent authority to order domestic surveillance.

As to (1):

The law clearly states that the criminal wiretap statute and FISA are “the exclusive means by which electronic surveillance . . . and the interception of domestic wire, oral, and electronic communications may be conducted.” If these authorities are exclusive, there is no other legal authority that can authorize warrantless surveillance. Courts generally will not view such a clear statutory statement as having been overruled by a later congressional action unless there is an equally clear indication that Congress intended to do that.

As to (2):

The administration’s ultimate argument is that “the president has the inherent authority under the Constitution, as Commander-in-Chief, to engage in this kind of activity.” This is the same argument outlined in the infamous torture memo, which concluded that the president can effectively ignore any statute that appears to infringe on this broad authority. That memo was withdrawn after it became public and was roundly criticized. The legal reasoning behind the arguments, however, has never been repudiated and appears to have resurfaced here. We cannot know for certain how the Supreme Court would rule on the legitimacy of the spying program. However, the court rejected President Harry Truman’s similar claim of broad presidential power in seizing control of the nation’s steel mills to avert a strike during the Korean War. The court, in a 6-to-3 ruling , stated that the president’s inherent authority is at its weakest in areas where Congress has already legislated. It ruled that to find inherent presidential authority when Congress has explicitly withheld that authority — as it has in FISA — “is not merely to disregard in a particular instance the clear will of Congress. It is to disrespect the whole legislative process and the constitutional division of authority between president and Congress.”

Also check out Marty Lederman’s posts over at Balkinization.

Applying Marginal Analysis to the Balancing of Constitutional Rights

The balancing of competing principles is one of the most common yet most controversial practices in constitutional interpretation. A number of authors have raised the objection that balancing is both irrational and subjective (e.g., Habermas), yet courts have no choice but to engage in it all the time. Others are uncomfortable with the notion that rights (e.g., the right to privacy) are optimization principles as opposed to rules. This note proceeds from the assumptions

  1. that most, if not all, constitutional guarantees are, in fact, optimization principles as explained by Dworkin and Alexy; and
  2. that balancing, while it may ultimately involve an existential, pre-rational commitment, can at least approximate a satisfactory level of discursive rationality.

The case of MacWade v. Kelly may serve as an example of where the court weighed the intrusion imposed upon subway riders against the government interest in preventing a terrorist attack. As a result of that balancing, the court held that the NYPD’s subway container search program was constitutional. As a predicate for its decision, the court found that the search program was effective in deterring and detecting potential terrorists. The correctness of that claim was contested, because the searches are pre-announced and de facto voluntary. Anyone who doesn’t want to be searched can simply walk away from the checkpoint and enter the subway system at a station where no searches are being conducted. The plaintiffs contended, not unreasonably, that someone with a bomb in their bag would not agree to be searched, so that the program would only serve to inconvenience harmless passengers while failing to identify and deter terrorists. The interesting feature here is the balancing of “the compelling government interest in preventing a terrorist attack” (P1) against “the relatively limited level of intrusion imposed upon subway riders” (P2). For purposes of the following discussion, I will identify the competing principles as safety (P1) and privacy (P2).

The structure of the balancing of P1 and P2 is that of a zero-sum game. An increase in P1 can only be realized at the cost of a decrease in P2. At the heart of balancing is therefore a tradeoff, which is very similar in structure to tradeoffs analyzed by microeconomic theory.

Constitutional marginalism

The core of modern microeconomics is marginal analysis. Marginal analysis breaks an action or a change into small steps and evaluates the benefits and the costs of each step. As long as the benefits exceed the costs, the step is taken. Once the costs exceed the benefits, the step is not taken. The critical insight of economics is that for rational action, only the marginal costs and benefits count. It makes little sense to weigh the total value of one good (G1) against the total value of another good (G2), for example, the value of (all) water against the value of (all) diamonds. The outcome of a balancing of two totals is arbitrary. In our example, (all) water would always win. Of course, that’s not our experience at the margin, where one additional unit of water is in almost every circumstance worth less than one additional unit of diamonds. Similarly, it makes even less sense to balance marginals against totals, e.g., all diamonds against one additional glass of water. In almost every instance, the totals will win. These insights can be applied to constitutional balancing. For example, in the subway case, the court wrote:

Against the compelling government interest in preventing a terrorist attack, the Court has weighted the (relatively limited) level of intrusion imposed upon subway riders.

No wonder that the court found the interest in preventing a terrorist attack to outweigh the limited intrusion imposed upon the passengers’ right to privacy, because the court incorrectly balanced totals against marginals. The correct inquiry would have been whether the marginal increase in safety is worth the marginal loss of privacy.

The privacy-elasticity of safety

Another feature of dealing with tradeoffs in a rational manner that economics has made explicit is the relative value of any good, that is, the definition of the marginal value of any good in terms of its opportunity cost. The value of one additional unit of G1 is equal to what I have to give up in G2…Gn in order to obtain G1. In a world with only two goods, diamonds and water, the value of one additional diamond is the amount of water that I have to give up in order to get it. The same reasoning can be applied to balancing competing constitutional principles. In the subway case, for example, the question that the court should have asked is how much privacy are we willing to sacrifice in order to achieve the gain in security, promised by the bag search program. The price of a marginal increase in security is the marginal loss of privacy. Of course, what economics can’t tell us is how much of P2 we are willing to sacrifice in order to get more of P1, which, in economic terms is a question of elasticity. When economists talk about elasticity, they look at two variables, one variable that changed (e.g., a 10% drop in quantity sold), and another one that caused that change (e.g., a 5% increase in price). By dividing the first by the second (the “effect” by the “cause”), we get the elasticity of the event. If the elasticity is < 1, we say that the first variable is inelastic with respect to changes in the second. (E.g., a big change in price has little effect on the quantity sold.) If the elasticity is > 1, we say that the first variable is elastic with respect to changes in the second. (E.g., a small change in price has a big effect on the quantity sold.) We can apply the same reasoning to constitutional principles. If we are willing to sacrifice lots of privacy for a small increase in safety, then safety is inelastic with respect to privacy and, by the same token, privacy is elastic with respect to safety. Put differently, safety is privacy-inelastic. How do we know whether safety is privacy-inelastic (we’re willing to give up lots of privacy for a small increase in safety) or privacy elastic (we’re willing to give up a little privacy, only in exchange for a big increase in safety)? In the constitutional context, text and precedents are our guides. Of course, establishing “constitutional elasticity” is not an exact science, much like economics by the way. But there are certainly different elasticities with respect to different principles. For example, safety, a compelling state interest, is likely to be less privacy-elastic than, say, an accurate economic census, which is a rational but not a compelling state interest. The analogy to economic reasoning demonstrates that the weight and the significance of any constitutional right, understood as an optimization principle, depends entirely on the “marginal rate of substitution” in terms of a competing right that we are willing to accept. Rights are not absolute, they are relative with respect to competing rights in a conceptually clearly defined manner.

The declining marginal value of privacy and safety

How much of a right we are willing to sacrifice in order to obtain an increase in another not only depends on the “nature”? or the “inherent significance” of the respective rights relative to one another (e.g., safety is generally more weighty than the accuracy of an economic census relative to privacy), but also, and maybe predominantly, on how much of each competing right we already have and enjoy. In economics, this phenomenon is known as declining marginal value. The assumption is that, with respect to most goods, having more is better than having less, but decreasingly so the more you already have. The value of the first glass of water after a day in the sun is greater than that of the second glass of water and so on. (For the first glass, I’d be willing to give you a nice diamond, for the second glass, you’d only get a smaller one, etc.) In other words, our marginal valuation of a good depends on the total amount of the good that we already possess. It is reasonable to assume that the same is true for most constitutional rights. Suppose I enjoy a great deal of privacy in the subway (e.g., no camera surveillance, no recording of my points of entry and exit) and very little safety (e.g., everyone can walk in with a bomb). In that situation, I’d be willing to sacrifice quite a bit of my privacy even for just a modest gain in safety. That willingness is likely to change, however, if my privacy in the subway is already severely restricted, e.g., by pervasive surveillance. The less privacy I have, the more valuable whatever remains becomes, measured, as always, in terms of safety as the competing interest. Economists use the law of declining marginal value along with other relatively modest assumptions of transitive rationality to model a universe of acceptable and unacceptable choices in the form of indifference curves.

Constitutional indifference curves

Suppose you presently enjoy a high degree of privacy (P2) and a low degree of safety (P1), and that, broadly speaking, safety is relatively privacy-inelastic. Based on that information, a judge could create a (mental) schedule of hypothetical tradeoffs among which the legal system would be indifferent.

The downward sloping shape of the indifference curve, which is convex to the origin, is a function of the declining marginal value of both privacy and safety. The steep incline once the curve comes close to the y-axis reflects the increasing relative value of privacy once much of it has already been sacrificed. Any further decrease in privacy requires a significant increase in safety to be constitutionally acceptable.

Suppose that we change our assumption about the overall relative value of privacy and safety, e.g., because terrorism is no longer seen as a significant threat. That change would be reflected in the slope and the position of the indifference curve. The following graph depicts a world in which safety is privacy-elastic, that is, a small decline in privacy is constitutionally acceptable only if offset by a significant increase in safety.

As indicated above, the shape and the position of the constitutional indifference curve would be determined on the basis of precedent and of the present degree of realization of each competing principle, prior to the challenged regulation. The indifference curve thus identifies the realm of constitutionally permissible and generally impermissible trade-offs. Any change, likely to follow from the challenged regulation, that will result in a bundle of rights (P1/P2) on or above the indifference curve will be acceptable, assuming that the status quo ante was a constitutional one. Any change that will result in a bundle of rights (P1/P2) below the indifference curve will almost always be unacceptable and therefore unconstitutional.

For example, in my analysis of the subway case, I (somewhat charitably) assume a relatively high degree of privacy in the subway system and a low degree of safety. I further assume that safety is relatively privacy-inelastic, that is, I am willing to sacrifice “lots” of privacy for a “moderate” increase in safety. Lastly, I assume that the bag search program, as proposed, is largely ineffective. With these assumptions in place, we can draw the following argumentative space:

The regulation moves the bundle of rights from an (presumably) acceptable status-quo ante (x) to a state (y) where the realization of both P1 and P2 is below the indifference curve. The privacy loss is moderate and the gains in safety are minimal. This is the sign of a highly suspect regulation that results in a net loss of constitutional freedoms and that should therefore almost always be unconstitutional.

Note that the framework outlined above has less to do with law & economics proper than with discourse theory. It is not my intention to analyze constitutional law from an external economic point of view, but rather to import some of the argumentative strategies applied to economic tradeoffs in the context of price theory into the normative legal discourse that usually relies on concepts such as “compelling state interest,” “least restrictive means,” etc. As always, comments are welcome!
[tags]marginal analysis, economics, constitution, balancing[/tags]

How the Court Confuses Marginals and Totals in the Subway Search Case

In MacWade v. Kelly, the subway bag search case, the court found that the bag search program “is an effective measure to help deter and detect a terrorist attack against New York City’s subway system,” even though searches are pre-announced and anyone who doesn’t like to have their bag searched (a subset of passengers that presumably includes everyone with a bomb in their bag) can simply walk away. The basis of the court’s finding was the testimony of Commissioner Sheehan and Richard Clarke, the defendants’ experts. The reasons given for the effectiveness of the program were that

  1. Its random nature “adds uncertainty and unpredictability to the planning and implementation of a terrorist attack”; and that
  2. The program “reinforces the awareness of police officers, transit workers, and the public of the need to be alert.”

Reason (1) simply begs the question. A program that’s not effective to begin with doesn’t somehow become effective just because it its administered in a random fashion. Random strategies presuppose an effective program. And with respect to (2), one must seriously wonder if the witness wasn’t joking. The purpose of the bag searches is to keep the police awake alert? C’mon! (Dan Solove has some fitting comments on this issue.) And that’s it. That’s the basis for the court to conclude that the bag searches are an effective means to deter and to detect a terrorist attack. I am not convinced, and it appears that court took its own admonition to heart that:

[I]t would be inappropriate for courts to second-guess the judgments of law enforcement officials and other public officials who are charged with protecting the public and making difficult choices of resource allocation.

That, of course, is a ridiculously lenient standard of judicial oversight. I wonder if there are any public officials who aren’t “charged with protecting the public and making difficult choices of resource allocation.” If that were the standard, then judicial review as such would be nonexistent.

The most interesting issue raised by the opinion is somewhat less obvious. In the conclusions of law, the court writes:

Against the compelling government interest in preventing a terrorist attack, the Court has weighted the (relatively limited) level of intrusion imposed upon subway riders.

Borrowing language from economics, the court compares totals, i.e., a terrorist attack, with marginals, i.e., the intrusions imposed upon subway riders. The comparison is not between “preventing terrorist attacks” and “people’s privacy” (two totals), nor is it between “the increase in safety from the subway bag search program as challenged” and “the relatively minor intrusions imposed upon subway riders by the program” (two marginals). As between the two, only the latter is a reasonable approach to weighing competing interests, but be that as it may, comparing marginals to totals is simply not meaningful, because the totals will always win. The court’s faux balancing is merely a rhetorical gloss on a foregone conclusion.