|
COPYRIGHT 2003 Stanford Law School
I. TWO VIEWS
II. RATCHETS A. Ratchets, In General B. The Statist Ratchet 1. Conceptual problems 2. Institutional problems: Is there a mechanism? 3. Psychological problems: Adaptive preferences? 4. Normative problems: Is the statist ratchet bad? C. The Libertarian Ratchet D. Government Without Ratchets III. FEAR A. Preliminaries B. Two Views of Fear C. The Influence of Fear During Emergencies D. A Note on Cognitive Panics E. Institutional Problems 1. Intrapersonal self-restraint 2. Rules 3. Insulation of decisionmakers F. Strict Enforcement as a Precommitment Device CONCLUSION
I. TWO VIEWS
Events since September 11, 2001 have produced a new round of debate about law and the emergency powers of government. Citing the need for swift and resolute action in a national emergency, the government has both sought new legal authority to combat terrorism and has suggested, in various spheres, that some emergency powers are inherent in executive authority. On these bases the Bush administration has secured the enactment of the Patriot Act, expanding law enforcement's powers to fight terrorism; ordered the detention of both foreign nationals and American citizens as unlawful enemy combatants; proposed a system of military tribunals for such cases; and increased security at a range of public facilities. Yet the legality of all these policies is contested; even the proper analytic framework itself is hotly disputed. Does constitutional law--either in the sense of constitutional rules or of constitutional outcomes-change in "emergencies," or is the law constant over time? Is there even an emergency in the first place? Which institution or institutional process is authorized to determine whether there is? In what follows we aim to cut through these tangled debates to focus on the underlying concern about emergency powers.
There are two main views about the proper role of the Constitution during national emergencies. We label them the "accommodation" view and the "strict enforcement" view. The accommodation view is that the Constitution should be relaxed or suspended during an emergency. (1) During an emergency, it is important that power be concentrated. Power should move up from the states to the federal government, and, within the federal government, from the legislature and the judiciary to the executive. Constitutional rights should be relaxed, so that the executive can move forcefully against the threat. If dissent weakens resolve, then dissent should be curtailed. If domestic security is at risk, then intrusive searches should be tolerated. There is no reason to think that the constitutional rights and powers appropriate for an emergency are the same as those that prevail during times of normalcy. The reason for relaxing constitutional norms during emergencies is that the risks to civil liberties inherent in expansive executive power--the misuse of the power for political gain--are justified by the national security benefits. (2)
Many constitutions have had provisions that enhance executive powers during times of emergency. The most famous example is also the most unfortunate: the Weimar Constitution's provision granting dictatorial powers to the executive in case of emergency. (3) But this is just one example among many. (4) In American constitutional law, Article I, Section 9 of the U.S. Constitution authorizes Congress to suspend the privilege of habeas corpus "when in Cases of Rebellion or Invasion the public Safety may require it." (5) Although this emergency power was pointedly vested in the legislative branch rather than the executive, (6) a more important constitutional rule is implicit but widely recognized: During most major wars involving the United States, Congress and the judiciary have deferred to the executive more than they have during peacetime. (7)
The second major view about emergencies, the strict enforcement view, is that constitutional rules are not, and should not be, relaxed during an emergency. (8) This view starts from the observation that the Constitution already provides that the level of protection for civil liberties depends on the interest of the government. Consider, for example, "compelling interest" standards used to evaluate laws that discriminate against protected classes. When an emergency exists, the government has a "compelling interest" in responding to it in a vigorous and effective way. Thus, laws that would not be tolerated during normal times are constitutionally permissible during emergencies. Racial or ethnic profiling, for example, is likely to be seen as less objectionable when used to prevent a terrorist attack than when used to interdict the distribution of illegal drugs. (9) The Constitution should be enforced "strictly"--that is, the rules should be the same during emergencies as during normal times, even if outcomes differ--so that both civil liberties and government interests such as national security can be appropriately balanced, as they always need to be.
Although the strict enforcement view and the accommodation view can be given different doctrinal formulations, what interests us is the degree of deference that courts give the executive during an emergency. Stipulate that judges could provide "high" or "low" deference during emergencies, where high deference permits some aggressive executive actions that are prohibited by low deference. High deference could be implemented through a threshold test: Courts will apply strict scrutiny unless they first find that an emergency exists, in which case they will permit any executive action that has a rational basis. High deference could also be implemented through a relaxed version of the ordinary compelling interest test, where judges find that the government interest becomes more compelling whenever an emergency occurs, whether or not there is a formal declaration (by courts or other officials) that an emergency exists. Correlatively, low deference could result from the refusal to treat constitutional rights differently during emergencies and normal times, or it could result from the use of a strict compelling interest test. We do not take a position on these issues; the public and scholarly debate is not, for the most part, about doctrinal formulations but about the degree to which judges should defer to the executive.
Whatever the doctrinal formulation, during normal times the accommodation and strict enforcement view permit the same kinds of executive action, and during war or other emergency the accommodation view permits more aggressive executive action than the strict enforcement view does. We assume that courts provide extra deference during an emergency or war because they believe that deference enables the executive to act quickly and decisively. Although deference also permits the executive to violate rights, violations that are intolerable during normal times become tolerable when the stakes are higher. Now, one could criticize courts for getting the liberty-security tradeoff wrong during emergencies, just as one could criticize courts for getting the liberty-security tradeoff wrong during normal times. But in the current debate civil libertarians make a more ambitious claim: That executive action is more likely to be worse during emergencies than during normal times--and that therefore the accommodation position taken by judges is a mistake.
On what grounds could one argue that judges have erred at the wholesale level? Civil libertarians argue that even if judges are right that the benefits from decisive executive action rise during emergencies, judges systematically overlook two costs or risks that are special to emergency. The first overlooked cost is the long-term, postemergency institutional damage from accommodating aggressive executive action during an emergency. The second is the risk that during an emergency fear leads to bad policy.
The institutional argument is that emergencies work like a ratchet: With every emergency, constitutional protections are reduced, and after the emergency is over, enhancement of constitutional powers is either maintained or not fully eliminated, so that the executive ends up with more power after the emergency than it had before the emergency. With each successive emergency, the executive's power is ratcheted up.
The other argument is psychological: During an emergency, people panic, and when they panic they support policies that are unwise and excessive. (10) Relaxation of constitutional protections would give free rein to the panicked reaction when what is needed is constraint. Normal constitutional protections hinder the enactment of bad laws during emergencies. To the critic who argues that normal constitutional protections also prevent needed concentration of power in the executive, the defender of strict enforcement argues that any hindrances on forceful executive action would be justified by the benefits, the avoided laws and acts that reflect fear rather than reason.
The ratchet theory and the panic theory have become fixed points in the debate about emergency powers yet have escaped rigorous analysis. As we will show, both theories have conceptual, normative, and empirical difficulties. The ratchet theory lacks a mechanism that permits constitutional powers to rise and prevents them from falling and makes implausible assumptions about the rationality of individuals who consent to constitutional changes during emergencies. Those who fear the ratchet's power point to constitutional trends--such as the rise of executive power--that are more plausibly the result of long-term technological and demographic changes, not of recurrent emergencies; they ignore the possibility of constitutional trends in the opposite direction, such as the rise of individual rights. (If there is such a trend, it is not a ratchet process either; we include a critique of an optimistic variant of the emergency ratchet, in which a succession of emergencies causes government to display ever-increasing respect for civil liberties.) As for the panic theory, it relies on a psychologically unrealistic conception of fear and on dubious empirical assumptions about the influence of fear on public policy. Finally, defenders of either theory do not examine their normative premises sufficiently: It is not clear that panics and ratchets, if they occur, are bad things. Fear is often the correct response to a threat; panics can shatter constitutional structures, but sometimes constitutional structures should be shattered. Ratchets put the status quo out of reach, but sometimes that is where it should be.
II. RATCHETS
In this Part we critique accounts of emergency that posit a ratchet effect, in which a succession of emergencies produce a unidirectional, and irreversible, increase in some legal or political variable. Part II.A provides a brief overview of ratchet accounts in legal theory, questioning their general utility. Part II.B critiques the most common version of the emergency ratchet account, which holds that emergencies produce a statist ratchet: an irreversible trend towards increased state power and official suppression of civil liberties, free speech, and political association. We suggest that the statist ratchet account is implausible on conceptual, institutional, psychological, and normative grounds. Part II.C critiques the opposite account, which holds that emergencies produce a libertarian ratchet in the form of ever-increasing governmental respect for civil liberties. Part II.D suggests that the ratchet idea has little utility for positive or normative argument about emergencies, principally because ratchet accounts posit an implausible amount of friction in the lawmaking system. In our picture, by contrast, the lawmaking system adjusts fluidly, if unpredictably, to emergencies, exogenous shocks, and other changes in the political and social environment; few changes are unidirectional and irreversible in the strong sense that ratchet accounts suppose.
A. Ratchets, In General
The "ratchet" (or, redundantly, the "one-way ratchet") is a favored analytic tool of legal theorists. Too much so, in fact. Ratchet accounts are invoked in a bewildering array of settings, ranging from the theory of regulation and bureaucratic behavior (11) to racial profiling (12) and sexual mores. (13) For a genuine ratchet to occur, however, highly specialized conditions must obtain. The essential features of a ratchet are unidirectional and entrenched change in some legal variable. First, the policy space in which the ratchet occurs is assumed to be one-dimensional, so that the ratchet produces ever-increasing values of a variable--more and more and more of something. Second, the incremental increases are fixed once they occur. Note that the change need not be literally irreversible; although strong ratchet accounts posit irreversibility, weak ratchet accounts merely posit that change is sticky, because more or less costly to undo. Weak ratchet accounts seem more plausible than strong ones, but they also pack less punch: The less costly it is to undo a given change, the less important is that change.
Putting these conditions together, a well-formed ratchet account must have something like the following shape: At Time 1, some legal rule or practice emerges endogenously from political processes, including the legal system; at Time 2, the rule or practice is cemented by some mechanism and has become an exogenous constraint; at Time 3, some dimension of the rule increases endogenously; at Time 4, the increase is cemented into place; and the process repeats indefinitely. These conditions are rare, perhaps even nonexistent. The danger here is that the "ratchet" label is being bandied about too freely and is often confused with a simple trend that happens to extend over time or with endogenous but reversible change in some variable that would quickly revert to its original value if other legal or social conditions changed.
Despite the ubiquity of ratchet accounts, few such accounts are fully specified, and often there is no plausible way to cash them out. Take a popular idea, or intuition, in constitutional theory: If conservative judges respect precedent while liberal judges freely overrule precedents, and conservative courts alternate with liberal ones, then a ratchet effect is created, whereby the existing stock of precedents becomes increasingly liberal over time. But this account is either out of equilibrium or arbitrarily assumes that the two camps have wildly disparate preferences. If the implicit picture is that both liberal and conservative judges are political, seeking to embody their preferences and attitudes in legal decisions, then conservative judges are myopic in refusing to overrule liberal decisions; they are repeatedly, and inexplicably, duped by the equally unprincipled but more cunning liberals. So the picture must instead be that liberal judges are political while conservative judges have a strong and principled preference for adhering to any past decision, whatever its political valence; but this seems arbitrary.
Perhaps history contains a few genuine ratchet processes. The continual, and doubtless irreversible, expansion of the political franchise in liberal democracies over the course of the nineteenth and twentieth centuries might qualify. But even here there are problems; although it seems intuitive that voting rights, once granted, are difficult to revoke (assuming the recently enfranchised may themselves vote on any revocation proposal), it is not obvious why enfranchised groups would continually admit politically powerless groups into the political system. In general, ratchet arguments are methodologically suspect and are invoked with far greater frequency than is warranted by theory or evidence. This pattern holds true for ratchet arguments about emergency powers, to which we now turn.
B. The Statist Ratchet
The statist ratchet identifies a putative tendency of emergency policies to "become entrenched over time and thus normalized and made routine.... The maintenance of emergency powers may be accompanied by expansion over time of the scope of such powers. At the same time, built-in limitations on the exercise of emergency authority and powers tend to wither away." (14)
As it turns out, however, the statist ratchet account has only a surface sheen of plausibility, and no core. It assumes that emergencies produce unidirectional and irreversible change in the direction of official intrusion on civil liberties. But there is no obvious reason to think that any such process occurs; the statist ratchet fails to supply a mechanism that would explain such a process if it were to occur; and, if there is such a mechanism, it is not clear that the resulting ratchet process is bad. We will organize these points into four critiques of the statist ratchet: conceptual, institutional, psychological, and normative.
1. Conceptual problems.
The statist ratchet, like all ratchet accounts, assumes a finite, one-dimensional policy space. In this space, government policies vary from minimally to maximally intrusive; the statist ratchet assumes that emergencies produce a continual increase that is unidirectional on this dimension, moving steadily from less official oppression to more.
But this picture is too crude. The policy space is not one-dimensional but multidimensional: Official policies, whether instituted during an emergency or not, can intrude more (or less) on some margins while intruding less (or more) on others. At Time T the government policy for airport security is to search passengers who fit a given ethnic and religious profile. At Time T+1 the policy changes to random searches; the new policy, let us say, imposes a cost (at least in an expected sense) on a greater number of people but reduces the stigma of being searched. Here it is senseless to ask whether liberty has been increased or decreased; instead it has been redistributed, by imposing a smaller deprivation more widely. In addition, there is the standard problem of conflicts or tensions between and among libertarian rights, arising from budget constraints on the government that funds the institutions needed to protect those rights. (15) More money for airport searches may reduce the need for ethnic profiling, but it may mean less money for public defenders or a longer court queue for citizens asserting constitutional liberties against government.
These two problems--controversial choices about the distributive profile of libertarian rights and the interdependence of budgeting choices that affect rights--mean that officials face the difficult problem of aggregating incompatible liberties across different individuals. In rare cases, Pareto-improving moves will enable greater security at a given level of official intrusion or less intrusion with a constant level of security; (16) but in most cases more liberty for some means less liberty for others. Because aggregative judgments are inescapable, it is not so much wrong as incoherent to speak generally of "society" having "more" or "less" liberty.
To be sure, these conceptual problems are not dispositive in and of themselves. We might discover, empirically, a decrease of liberty on all dimensions or on some suitably weighted or aggregated combination of dimensions. What is true, however, is that ratchet...
Read the full article for free courtesy of your local library.
|
Bookmark this article
Print this article
Link to this article
Email this article
Digg It!
Add to del.icio.us
RSS
