It is a familiar lesson of U.S. constitutional doctrine that the outputs of decision rules will sometimes depend on the level of generality with which their inputs are defined. This theme is perhaps most evident in substantive due process doctrine: Defining a liberty interest in broad terms can increase the likelihood of its qualifying as “fundamental,” just as defining the interest in narrow terms can reduce that likelihood. But generality levels can make a difference in other areas of the law as well. A “right” may be more likely to qualify as “clearly established” for purposes of a qualified immunity defense if we characterize that right broadly rather than narrowly, a “matter” may be more likely to qualify as “of a public concern” when the matter itself is defined abstractly rather than specifically, a “power” may be more likely to qualify as “great substantive and independent” (and hence not implied by the enumerated powers of Article I) if the power is described in general rather than specific terms, and so forth. In these and other contexts, the outcome of a doctrinal inquiry can depend not just on the content of its evaluative criteria (e.g., what does it mean for a right to be “fundamental”?, what matters are and are not of “public concern”?, what does it mean for a power to be “great substantive and independent”?, etc.), but also on how one defines/describes/characterizes the objects to which those criteria apply (e.g., what is the “liberty interest” whose “fundamentality” is at issue, what is the “matter” whose “public-concerned-ness” we are evaluating?, what is the “power” whose “greatness”/“independence” we are measuring?, etc.).
I’d originally figured that this insight applied with equal force to the various forms of means/ends analysis that pervade constitutional law. Means/ends analysis, after all, requires some estimation of the strength of a “government interest” said to justify a constitutionally suspect enactment, and the strength of that interest will in turn depend on the level of generality with which we define it. Think, for instance, of Holder v. Humanitarian Law Project. There the Court rejected an as-applied First Amendment challenge to the federal “material support” statute, brought by plaintiffs “seek[ing] to facilitate only the lawful, nonviolent purposes” of foreign groups designated to be terrorist organizations. One can define the government interest in HLP at different levels of generality. From most to least specific, the interest might be characterized as that of (1) “cutting off support for the lawful, nonviolent activities of foreign organizations designated as terrorist groups,” (2) “undermining foreign organizations designated to be terrorist groups,” (3) “undermining foreign terrorist groups,” (4) “combating terrorism,” or (5) “promoting national security,” with a bunch of intermediate options in between. And as the generality-level of the government interest increases, so too should the ease of demonstrating that interest’s overall importance. (All else equal, for instance, the government will have less difficulty in highlighting the vital importance of “national security” than that of “cutting off support for the lawful, nonviolent activities of foreign organizations designated to be terrorist groups.”) In that sense, means/ends analysis does indeed seem to be at least somewhat sensitive to changes in generality-levels, with increased generality levels yielding increased odds of a government-friendly result.
But things turn out not to be so simple, as something interesting happens when we proceed to ask whether the law is sufficiently closely related to the government interest we have identified. Here, we encounter something akin to the opposite relationship between generality-levels and justificatory ease: the more generally we have characterized the government interest, the more difficult it will become to show the requisite means/ends fit. It would not be difficult to show that the material support statute is “necessary to further” the government’s interest in “cutting off support for the lawful, nonviolent activities of foreign organizations designated as terrorist groups”—that objective, after all, is precisely what the material support statute purports to pursue. But would the law count as necessary to further the more generally defined interest in “promoting national security”? Maybe, but maybe not. The problem is that the government can “promote national security” in many more possible ways than it can “cut off support for the lawful, nonviolent activities of foreign organizations designated to be terrorist groups.” And the wider the range of potential means of achieving an interest, the more likely it becomes that a “less restrictive” or “less discriminatory” means will emerge from the heap—thus demonstrating that the chosen means was fatally over- or underinclusive with respect to the interest in question.
In other words, broadening our characterization of the government interest may make things easier for the government (and more difficult for the challengers) when evaluating the strength of the interest, but it will then make things more difficult for the government (and easier for the challengers) when evaluating the degree of fit between the interest and a challenged law.
That’s not to say that the government can’t ever win an argument about “means” where the relevant “end” has been defined in highly general terms. Indeed, the government did end up winning in HLP, notwithstanding the Court’s decision to characterize the government interest broadly rather than narrowly (the Court went with Option #4, “combating terrorism”). Rather, the point is that because it focuses attention on both means and ends, means/ends analysis may manage to mitigate the influence of generality-levels on the ultimate outcome of the test. One can ratchet-up the generality level of the government interest to assist in justifying the ends, but one must then pay a price when attempting to justify the means. And one can ratchet-down the generality level to assist in justifying the means, but one must then pay a price when attempting to justify the ends. As long as one maintains the same description of the government interest throughout the analysis, then there should be some level of equilibration across the two prongs of the test.
Now before anyone starts publishing banner headlines about this observation, let me identify a few grounds for skepticism. The hypothesis I’ve offered may turn out to be (a) false or (b) trivial:
– Why the hypothesis may be false: Even if everything I’ve said is right, it still might be true that the generality levels matter more at the first step of the inquiry than at the second. In other words, the chosen generality-level of the government interest may exert a major positive influence on whether or not that interest counts as “compelling,” “important,” “legitimate,” or what have you, while exerting only a minor negative influence on whether there is a sufficiently close fit between the interest itself and the law under review. If that is true, then means/ends analysis would remain highly susceptible to manipulation via characterizations of the relevant government interest, with the positive effects of high-generality at the “ends” stage of the inquiry drowning out its negative effects at the “means” stage of the inquiry.
– Why the hypothesis may be trivial: I can imagine two arguments to this effect—one grounded in cynicism and the other grounded in optimism, with both concerning the overall constraining effect of doctrinal rules.
- The cynic’s argument would maintain that even if means/ends analysis is not sensitive to fluctuations in the generality-level of a government interest, means/ends analysis as a whole remains malleable, manipulable, and ultimately non-constraining in a myriad other ways. A judge who is dead-set on upholding a law subject to strict scrutiny (or striking down a law subject to rational basis review) can always find always find the arguments necessary to justify the desired means/ends result, no matter what how generally or non-generally the interest has been defined. So the cynic would say, it doesn’t much matter whether my hypothesis is right or wrong. Even if it is right, the outcomes will be what they will be because judges can and will manipulate other parts of the means/ends test to go where they want to go.
- The optimist’s argument would focus instead on what has thus far been an unstated (and undefended) premise of my argument—namely, that choices among generality-levels are to some extent arbitrary and difficult to predict ex ante. But if that point turns out to be false—if, in other words, there does exist a coherent, predictable, and defensible way of identifying the operative generality level of a given constitutional input, then any “equilibrating” or “self-regulating” process built-in to means/ends analysis would be of only marginal significance. Either way, the optimist would maintain, means/ends analysis would operate in a principled fashion—either levels of generality matter, in which case means/ends analysis is influenced by a principled legal choice, or levels of generality do not matter, in which case means/ends analysis will still be influenced by other principled legal choices.
Posted by Michael Coenen on February 20, 2015 at 03:04 PM
Comments
Very interesting. I’ve been giving this some thought in the choice-of-law area. There, you run into the same problem(?) of levels of generality when courts engage in government-interest analysis. The purpose of a law can be defined very narrowly (“To protect tavern-owners in Montana against excessive liability due to lawsuits brought by family members of decedents who became intoxicated at the tavern then drove out of state, where they died.”) or very broadly (“To promote entrepreneurship in Montana.”). Courts seem more inclined to find a “government interest” where the purpose is broad; a broad purpose is more likely to catch the facts at hand.
In this area, however, there’s no countervailing means/end analysis.
Posted by: anon | Feb 21, 2015 10:07:53 AM