[*PG1]THE ETHNIC AND RELIGIOUS PROFILING OF NONCITIZENS: NATIONAL SECURITY AND INTERNATIONAL HUMAN RIGHTS
Abstract: In response to the terrorist attacks of September 11, 2001 the United States has taken exceptional national security measures, particularly with respect to noncitizens. Many of those measures either expressly or at least arguably entail profiling young male Arabs and Muslims. They are summarized in the first section of this Article. The next section offers an analytical framework for evaluating a profiling program from a policy standpoint; it maintains that no profiling practice is justified unless it satisfies certain minimum requirements of rationality and weighted cost effectiveness. The final section suggests that some of the national security-related profiling practices raise serious issues of U.S. compliance with its obligations under the Convention on the Elimination of All Forms of Racial Discrimination.
In response to the terrorist attacks of September 11, 2001, United States policymakers have sought new and more imaginative ways to protect national security and public safety. Many of the strategies operate specifically on those who are not United States citizens. Primarily in the immigration arena, these measures entail detention, intelligence-gathering, expanding the substantive criteria for removing noncitizens, contracting their procedural rights, tightening visa and other overseas policies, and upgrading apprehensions and inspections at ports of entry.
Several such practices utilize what has come to be known as profiling. I shall use that term here to mean specially targeting individuals who possess identifiable attributes that are believed to bear positive statistical correlations to particular kinds of misconductin this case, involvement in terrorism.
[*PG2] This sanitized definition makes profiling sound innocuous. But the specific profiling attributes at issue here have included ethnicity, country of nationality, religion, gender, and age. In practical terms, people who appear to law enforcement officials to be of Arab descent or Muslim (or both) have been the most thoroughly scrutinized, especially when they are also young and male. Are these practices a sensible policy response to the threats posed by foreign terrorists? If so, are they legal?
Feeling obliged to disclose my personal biases, I shall say at the outset that I am opposed to ethnic and religious profiling, even in the present national security context. My route to that conclusion differs somewhat from the road traveled by some other commentators, however, and along the way I am prepared to concede perhaps more ground.
Section I of this Article summarizes the various national security-related initiatives that the United States has undertaken in relation to immigration specifically or non-U.S. citizens generally. Section II suggests a framework for analyzing the policy tradeoffs that these various profiling practices present. Section III considers whether current U.S. profiling strategies comport with its obligations under the International Convention on the Elimination of all Forms of Racial Discrimination (Race Convention).1
A few words of limitation: the United States is also a party to other, more generic international human rights regimes that include, among the many human rights they cover, prohibitions on specified forms of discrimination.2 This Article does not explore the analogous [*PG3]issues that these more generic instruments present, nor does it analyze the many human rights treaties that the United States has not ratified, but which also contain broadly worded anti-discrimination provisions.3 Finally, the Article does not assess the compatibility of the current U.S. profiling practices with domestic constitutional constraints.4
As the Introduction suggests, the United States efforts to combat terrorism have been far-reaching, particularly with respect to immigration policy and the general regulation of non-U.S. citizens. Em[*PG4]phasis here is placed upon the post-September 11 measures, but a few preexisting programs are discussed as well.
Among the many immigration components of the U.S. national security program, preventive detention has come to occupy center stage.
During the months or years that removal proceedings often consume, the government generally has the discretion to release the person on bond if he or she will not endanger persons or property and is likely to appear for the removal proceeding.6 By way of exception, however, Congress has made detention mandatory when the removal charge involves terrorism.7
Detention has also been made mandatory in certain asylum cases. On March 18, 2003, the Department of Homeland Security (DHS) Secretary Tom Ridge announced an initiative called Operation Liberty Shield. The initiative was accompanied by an administration list of thirty-four countries designated as harboring terrorists. The list consisted almost entirely of countries that are predominantly Arab, Muslim, or both. Under one key provision of the short-lived Operation Liberty Shield, if a national of a listed country applied for asylum at a U.S. port of entry but lacked valid entry documentsa common condition for asylum seekers at ports of entrydetention was mandated until the person either received asylum or was removed.8
Apart from formal mandatory detention, national security considerations have recently influenced discretionary detention decisions. First, as elaborated upon in the next subsection, the FBI arrested and interviewed hundreds of noncitizens in the months [*PG5]following the September 11 attacks. Many arrested individuals were believed to have violated immigration laws. In those cases, the Immigration and Naturalization Service (INS) instituted removal proceedings, and the FBI routinely sent boilerplate memoranda to the Executive Office for Immigration Review (EOIR) opposing discretionary release on bond pending the removal decisions.9
Second, when an immigration judge releases an individual on bond, and the government appeals, the individual will be detained while the appeal is pending.10 DHS need not demonstrate, or even allege, any reasons for doubting the immigration judges decision to release. As a result, simply by filing an appeal to the BIA, the prosecuting official can unilaterally effect the persons confinement for a period of months or even years.11
Immediately after the September 11 attacks, the FBI, in cooperation with several other law enforcement agencies, launched the Pentagon/Twin Towers Bombing Investigation (PENTTBOM). Officials questioned and frequently detained individuals possibly involved in the September 11 attacks or other terrorist activities, as well as anyone who might have information of use to the investigation. Sometimes, however, undocumented immigrants were encountered fortuitously during the course of the investigation. Even if no link to, or knowledge of, terrorism was suspected, they too were arrested, turned over to the former INS and detained.12 Of the detainees held on immigration-related grounds (whether or not suspected of terrorism), the overwhelming majority were males, between ages twenty-six and forty, from Arab or Muslim countries. About one-third were from Pakistan.13
[*PG6] To compound the controversy, the government has steadfastly refused to disclose to the public the names or whereabouts of the detainees, a policy that the D.C. Circuit has upheld.14 In addition, a recent federal Bureau of Prisons regulation rooted in the PENTTBOM investigation authorized the monitoring of inmate-attorney conversations, sometimes with the inmates knowledge and sometimes without it, in certain cases of suspected terrorism.15 The American Bar Association has been deeply troubled by these developments.16
The PATRIOT Act, enacted on October 26, 2001,17 requires the Attorney General to certify and detain any noncitizen when there are reasonable grounds to believe the person is either inadmissible or deportable on certain national security-related grounds. Unlike the general detention powers discussed in the preceding subsections, this power expressly allows indefinite detention as long as the case is reviewed at least once every six months.18 As of March 26, 2003, the certification provision had yet to be invoked,19 most likely because the Department of Justice has found it simpler to detain noncitizens suspected of terrorism either through the PENTTBOM program or through its ordinary power to detain noncitizens during the course of removal proceedings.
Approximately two months after the September 11, 2001 attacks, President Bush, acting as Commander-in-Chief, issued a Military Or[*PG7]der that requires the detentionand military trial if criminal charges are filedof noncitizens who the President has reason to believe: (a) are members of Al Qaeda; (b) are involved in specified ways in present or potential future activities with adverse effects on the United States, its citizens, national security, foreign policy, or economy; or (c) have knowingly harbored any of the former individuals.20 The Order imposes no time limits on the detention.21 Its main use thus far has been for Taliban and Al Qaeda fighters captured in Afghanistan and flown to the U.S. Naval Station at Guantánamo Bay, Cuba since January 10, 2002.22
If nothing else, the events of September 11 have exposed deep fault lines in the gathering, sharing, and use of intelligence data. In the case of noncitizens, three parallel initiatives will soon be folded into a comprehensive intelligence operation called US-VISIT. They occupy the first three subjects of this subsection. The fourth subsection elaborates upon the current status of the combined system. The remaining subsections discuss other miscellaneous intelligence operations related to national security that specifically affect noncitizens.
For many years, numerous observers bemoaned the United States governments failure to keep systematic records of all entries and departures of noncitizens. The concerns, however, were rarely based on national security; the notion was rather that without such records it was impossible to identify, apprehend, and remove overstayers effectively. In 1996, Congress gave the Attorney General two years to develop an automated entry and exit control system that would require the recording of every noncitizens entry into the United States, the recording of every noncitizens departure, and a matching of the two [*PG8]in order to identify all cases of overstay.23 There were strong and immediate objections that recording the tens of millions of annual border crossings would cause horrific congestion at border points and would jeopardize trade.24 After various delays and reductions in scope to assure proper implementation,25 urgency levels heightened sharply with the September 11 attacks. In 2002, Congress set October 26, 2004 as the date by which the administration must implement an integrated entry and exit data system; the same legislation specified the technology standards, including use of biometric identifiers, that Congress expected the system to incorporate.26
For decades, United States law has required almost all noncitizens who are ages fourteen or over and who remain in the country thirty days or more to register with designated authorities and to be fingerprinted; additionally, they must report any changes of address to the immigration authorities within ten days of moving.27 Failure to do so subjects them to both criminal prosecution and removal.28
Before September 11, the registration requirements were rarely enforced. Few noncitizens were even aware of the requirements, and the government made little attempt to publicize or enforce them. It was widely felt that the practical benefits of the information were outweighed by the administrative burdens and costs.
After September 11, priorities changed. On June 5, 2002, Attorney General John Ashcroft announced the National Security Entry-Exit Registration System (NSEERS).29 In 2002, on the first anniversary [*PG9]of September 11, this controversial program became applicable to nearly all male nonimmigrants sixteen years of age or older who were nationals of any of (by the end) twenty-five designated countries twenty-four Arab or other predominantly Muslim states and North Koreaas well as any individual nonimmigrants who otherwise presented national security or law enforcement concerns. The affected persons were fingerprinted and photographed at the ports of entry and, if they remained more than thirty days, were required to report periodically to DHS with additional information. At the time of this writing, the registration deadline for the last groups subject to NSEERS was April 25, 2003. Because the government intends to fold NSEERS into a new and more comprehensive program (US-VISIT, described below), it appears there will be no more special call-in announcements for entire countries; rather, future call-ins are likely to be individualized.30
As of June 7, 2003, approximately 16 percent of the 82,000 males required to register had been subjected to removal proceedings, in almost all cases because of immigration status violations. Many of those removed were individuals awaiting priority dates for family reunification; they had often assumed mistakenly that their cooperation in appearing for registration would not precipitate removal proceedings. The 82,000 special registrations, combined with the tens of thousands of additional registrations at ports of entry, have, thus far, yielded only eleven terrorism suspects.31
For reasons that have never been entirely clear, even before September 11 Congresss concerns about foreign terrorism have focused exceptional attention on foreign students, as distinguished from other nonimmigrant visitors. In 1996, Congress required the Attorney General, in consultation with the Secretary of State, to collect individualized information from colleges and universities on every enrolled foreign student. An institution that failed to provide the required [*PG10]information would be barred from further enrollment of foreign students.33 Following September 11, Congress expanded the required information and demanded prompt implementation.34 It simultaneously authorized schools to disclose individual student records to designated government officials pursuant to ex parte court orders issued on suspicion of terrorism.35 In addition, immediately after the attacks, the FBI began asking schools for information on their foreign students. The schools typically complied, even absent the court orders required by the new law.36 In 2002, Congress required the collection of additional information and mandated that the system be electronic.37
SEVIS (Student and Exchange Visitor Information System) is the name of the government-designed electronic system intended to satisfy the various statutory requirements.38 This Internet-based system stores the required data on foreign students and permits the transmission of the data, both among the relevant government agencies and between the agencies and the schools. It became fully operative, albeit with controversial technical glitches, on August 1, 2003.39
On April 29, 2003, Secretary Ridge announced a new initiative called the United States Visitor and Immigration Status Indication Technology System (US-VISIT). It is designed to integrate the entry-exit, NSEERS, and SEVIS programs into a single comprehensive electronic data system to monitor nonimmigrants. When it is fully operational (currently projected to be the end of 2005), all nonimmigrants who enter the United States, other than those tourists and business visitors who enter under the visa waiver program for nationals of selected countries, will encounter US-VISIT upon entry. Their fingerprints and digital photographs will be taken, and their travel documents scanned. Iris scanners might also be employed. The goal is to use biometric identifiers to create an electronic check in/check out [*PG11]system that can be used not only for national security purposes, but also for locating visa violators. When the visitors leave the country, DHS will verify their identities once more and input the departure information into the database.40 That information will then be available to inspectors at ports of entry, special agents and adjudications staff at DHS, U.S. consulates, and other law enforcement agencies.41
Yet another response to the events of September 11, 2001 was a massive Department of Justice scheme to interview thousands of noncitizens who might have information on potential terrorist activities. The idea was to talk with people whose demographics and immigration statuses resembled those of the nineteen hijackers. The interviews were purportedly voluntary, and the interviewees were to be regarded as potential sources of information, not as suspects. The project was coordinated by the Department of Justices Executive Office for United States Attorneys (EOUSA).
There were two rounds of interviews, including a cross-section of male nonimmigrants ranging in age from approximately seventeen to forty-seven, who had entered the United States between January 1, 2000 and February 27, 2002 and were nationals of any of twenty-six countries where Al Qaeda was thought to have a strong presence. Roughly 3,200 people were interviewed. Some of the questions elicited personal information about the interviewee (e.g., employment, education level, whether the person has ever visited Afghanistan, etc.), while others focused on whether the interviewee had knowledge of suspicious persons or activities. Despite their invasive nature, there seems to be a consensus that the interviews were conducted respectfully and professionally.42
[*PG12] The program was controversial, for at least two reasons. First, although no evidence indicates anyone was coerced to participate, interviewees did not always perceive the conversations to be truly voluntary. They often reported that they feared there could be [immigration-related] repercussions to them for declining to participate.43 Second, since the overwhelming majority of the interviewees were Muslim, of Arab origin, or both, interviewees often felt singled out and investigated because of their ethnicity or religious beliefs, according to immigration advocates and attorneys who sat in on interviews.44
A separate interviewing program commenced one year later. In 2003, in connection with the then-impending Iraq War, the FBI interviewed approximately 2,000 Iraqi nationals living in the United States. That program gave rise to many of the same issues.45
Shortly after the September 11 attacks, Congress made permanent a previously expired temporary program that authorized nonimmigrant visas for individuals who share or have shared critical reliable information about a terrorist organization.46 Up to fifty of these so-called S-6 visas may be issued each year.47
On November 29, 2001, Attorney General Ashcroft launched the Responsible Cooperators Program, in which he officially encouraged various Department of Justice personnel to make liberal use of S-visas in appropriate cases. In those instances in which a person who is willing to share important terrorism-related information is ineligible for an S-visa, the Attorney General urged officials to consider parole or deferred action as an incentive to cooperate.48 The American-Arab Anti-Discrimination Committee has expressed the concern that these programs would make members of the Arab-American community suspicious of one another.49
Post-September 11 legislation has expanded retroactively the range of terrorism-related activities or associations that trigger inadmissibility or deportability.50 In addition, the Secretary of State has designated many new organizations as foreign terrorist organizations; the legal consequence is that members of such organizations are subject to removal without proof that they knew or even should have known of the organizations terrorist activities.51
Several provisions of U.S. law, some of them predating September 11, permit executive officials to bypass the usual procedural safeguards in certain national security cases.
Under INA § 235(c), an immigration officer or immigration judge who suspects that an arriving noncitizen is inadmissible on any of several specified national-security-related grounds (including terrorism) is required to order the person removed and to report the removal to the Attorney General. If the Attorney General concludes from confidential information that the person is inadmissible on any of the designated grounds, and that disclosure of that information would be prejudicial to the public interest, safety, or security, then the noncitizen may be removed without a hearing.
One year after the 1995 bombing of the federal building in Oklahoma City by domestic terrorists, Congress established a special removal court for removing alleged foreign terrorists. The court is available when the governments case rests on classified information that the government feels it cannot safely disclose. The details of the [*PG14]statutory scheme appear elsewhere.52 As of this writing, the government has yet to invoke the courts jurisdiction.
In the days following the September 11 attacks, the Department of Justice identified a number of noncitizens who, it believed, either were connected to terrorist activities or had information useful to the government investigation of terrorism. These were designated as special interest cases. On September 21, 2001, Chief Immigration Judge Michael Creppy sent a memorandum to all the immigration judges instructing them to close to the public all removal hearings held in such cases. There could be no visitors, no family, and no press.53 The directive also ordered the immigration judges not to discuss the cases or disclose any information about them to anyone outside the immigration courts. It barred the judges even from confirming or denying whether such a case is on the docket or scheduled for a hearing.54 As of May 8, 2003, closed removal hearings had been conducted in 641 cases; no future closed hearings were scheduled.55 Two courts of appeals have divided over whether, by denying media access to the proceedings, the Creppy directive violates freedom of the press.56
INA § 240(b)(4)(B) allows the government to use secret national security information either to oppose an individuals admission to the United States57 or, more importantly, to oppose his or her application for discretionary relief even in the deportation setting. Over the years, both courts and commentators have questioned the procedural fairness of secret evidence removal hearings, with some courts holding that the practice violates due process, at least insofar as it excludes even in cam[*PG15]era evidence.58 Of particular relevance is the secret procedures potential and realized capacity to be applied disproportionately to Muslims and to individuals of Arab descent or appearance.59
The aftermath of September 11 included intense public concern about the adequacy of the visa system as a device for detecting dangerous individuals before they reach U.S. territory. The principal challenge lay in how best to reconcile critical U.S. national security requirements with the need for openness to tourists, students, business travelers, and other visitors who apply for visas and enter the country by the tens of millions every year. The following are a few of the overseas programs that resulted.
In May of 2002, Congress prohibited the issuance of nonimmigrant visas to any noncitizen who is from a country that is a state sponsor of international terrorism unless the Secretary of State determines, in consultation with the Attorney General and the heads of other appropriate United States agencies, that such alien does not pose a threat to the safety or national security of the United States.60 For purposes of this requirement, a state sponsor of international terrorism is any state whose government has already been so classified for other purposes under any of several existing laws.61
By May of 2003, the State Department had designated seven countries as state sponsors of international terrorismfive Arab states, North Korea, and Cuba.62 Before a consular officer can issue a [*PG16]visa, an applicant from one of the designated countries must complete substantial additional paperwork and receive a Visas Condor clearance, which entails an intense security check by federal intelligence agencies. According to one senior visa officer at an overseas consular post, the resulting delays average three to four weeks, but many others . . . can run into months.63 The same officer explained that the statutory phrase from a country has been interpreted to cover anyone born in that country, whether or not still a national of that country.64 Thus, for example, a person who left the country as a small child and has lived in the West ever since would be subject to the same delays.
The Homeland Security Act of 2002 left visa applications in the hands of the State Departments consular officers, but, as an added safeguard, Congress introduced DHS into the process.65 The Secretary of Homeland Security now has the authority to issue regulations about, and to oversee, the visa work of the consular officers.66 The Secretary may also prevent issuance of a visa but may not reverse a consular officers decision to refuse one.67 In addition, the Secretary may station DHS employees at the consulates to review individual visa applications, to perform investigations, and to advise and train consular officers on issues of national security.68
Elevated security precautions have caused long and controversial delays in visa issuance, especiallybut not exclusivelyfor students. Visa investigations now require a good deal of additional information.69
Moreover, until August 1, 2003, interviews were not required for most applicants for B-2 tourist visas and certain other visa categories. [*PG17]A State Department interim regulation now requires such interviews for all nonimmigrant applicants except those who fall within any of six narrowly defined groups, including young children, persons age sixty or older, and other miscellaneous categories.70
In addition to the overseas enforcement initiatives just discussed, a series of steps have intensified the scrutiny of arriving travelers at land borders, airports, and seaports. These steps include increased funding for additional immigration inspectors, investigators, and support staffs, improved training, and new technologies,71 as well as requiring passenger manifests, which contain certain prescribed details about each arriving passenger and crew member.72
For purposes of both the policy analysis in the next section and the Race Convention analysis in Section III, there are at least two ways to classify the national security initiatives discussed above. First, they can be classified on the basis of the groups of individuals on whom they operate. Thus, some of the programs apply only to nationals of select countries or, in one variant, to individuals who were born in select countries. Other programs, in contrast, apply to all noncitizens who exhibit particular behaviors, often with systematically different practical effects on nationals, natives, or descendants of particular countries. Second, the programs can be classified according to the individual interest affected or the resultant deprivations or level of intrusion. Some deprivations, such as detention, denial of a visa, or removal from the United States, are potentially quite serious because the effects can be highly invasive, long-lasting, or both. Others, such as closed removal hearings and secret evidence removal hearings, implicate the loss of procedural safeguards that are meant to reduce the risk of erroneous substantive deprivations. Still others, particularly those aimed at gathering intelligence, can involve anything ranging from minor inconveniences to costly delays, significant disruption, or loss of privacy.
The Introduction defines profiling as specially targeting individuals who possess identifiable attributes that are believed to bear a positive statistical correlation to particular kinds of misconductin this case, involvement in terrorism.73 Thus, customs officials have long made productive use of drug smuggler profiles based on appearances or habits. When investigating serial killings and serial rapes, police often employ profiles. So long as we live in a world where law enforcement resources are finite and the experts have good reason to believe that certain criminal behaviors tend to be disproportionately present in a discrete identifiable group, profiling seems rational. Credible statistics and probabilities can enhance the efficient deployment of scarce law enforcement resources.74
The problems arise when the profiling attributes include race, ethnicity, religion, gender, or other disfavored forms of discrimination. First, since profiling is premised on the notion that it is a rational way to allocate limited administrative or enforcement resources, the profiling is irrational if the particular attributes fail to accurately predict criminal conduct. Profiling is indefensible when its criteria are nothing more than the product of irrational prejudice. Because the profiling attributes discussed here coincide with popular prejudices toward Arabs and Muslims, there is cause to scrutinize their rationality with particular care.
Second, even when solid empirical evidence substantiates the assumed correlation between the profile and the criminal conduct, rational does not necessarily mean justified. A balancing is required. On one side of the ledger are the perceived gains in national security from the governments efficient use of limited law enforcement resources. On the other side are the considerable harms associated with government-sponsored discrimination.
In recent years, much has been done to expose the systematic police practice of targeting African-American motorists for vehicle stops.75 The phrases driving while black and, more recently, driving while Latino have made their way into our vocabulary.76
[*PG19] Since September 11, 2001, Congress, the executive departments, and federal agents who are engaged in both terrorism investigations and garden-variety immigration enforcement have become increasingly prone to target individuals thought to be Arab, Muslim, or nationals of Arab or Muslim countries.77 The expression flying while Arab has crept into our vocabulary.
Several distinct types of profiling should be distinguished. Officials might consciously and explicitly target people who appear to be Muslim or of Arab descent. The same practical results can occur indirectly, or de facto, as the byproducts of profiling on the basis of some third attribute that correlates positively with race or religion. Thus, policymakers or law enforcement agents might target nationals or natives of those countries that are predesignated as harboring terrorists. Many of the post-September 11 federal programs utilize that form of profiling; the populations of the listed countries are overwhelmingly Arab or Muslim or both.78
There are special reasons to scrutinize even the latter forms of profiling. First, proponents might well use country of nationality or birth as a pretext for ethnic or religious discrimination. Second, even when the architects of the program sincerely desire to focus on governments rather than ethnic or religious groups, or to proceed solely on the basis of statistical correlations between terrorism and particular countries of nationality or birth, such practices have obvious disparate impacts on Arabs and Muslims. Fair and responsible solutions to the legal and policy problems must take that negative consequence into account.
Some of these programs superimpose two other profiling attributesgender and age. An example would be any strategy that targets male nationals of particular countries who fall within a prescribed age range.79
My view is that law enforcement profiling is justifiable only when two conditions are met. First, the practice must be rational. At a minimum, this means that the particular profiling attributes must correlate [*PG20]positively with the relevant danger or misconduct, so that their use will increase the probability of detection. Second, any gains in the efficacy or efficiency of the inspection process must be balanced against the substantial harms of government-sponsored discrimination. Since the present discussion implicates policy rather than constitutionality, readers can decide for themselves the relative weights that these competing interests should command. Certainly, however, by analogy to settled constitutional traditions, the weight assigned to the harmful effects of government-sponsored discrimination should increase when, as is the case with many of the current practices, the bases for special targeting are race, ethnicity, religion, gender, and the like.
Do the national security-related profiling practices discussed here meet those two conditions?
David Cole is one of the leading critics of government programs that target the Arab and Muslim communities. In his groundbreaking article, Enemy Aliens,80 he questions whether the recent spate of profile-based counter-terrorism initiatives is even effective. He first observes that effective law enforcement requires the cooperation of local communities. In his view, it makes more long-term sense to work constructively with those communities than to alienate them through stereotypes and special targeting. He would apply the same principle on the international plane. To fight terrorism, he argues, the United States sorely needs the cooperation of the Arab and Muslim states; antagonizing them by targeting their nationals is counterproductive.81
Up to that point, I am largely in agreement. Professor Cole goes on, however, to deny the effectiveness of ethnic profiling in other ways. He argues that the profiling programs are poorly tailored to their national security objectives. In particular, he suggests that the vast majority of people who appear Arab or Muslimhe estimates 99.9%have no involvement with terrorism at all, and therefore that Arab or Muslim appearance is a hopelessly inaccurate proxy for terrorism.82
[*PG21] This is where I must respectfully part company. By the reasoning used, no one should be questioned or inspectedeven when boarding a planeunless there is something suspicious about that specific individual. After all, only the most minute proportion of air travelers wants to bring down the plane.
The fallacy is in looking at the wrong percentage. Certainly, I agree that only a minuscule percentage of noncitizens who appear to be Arab or Muslim are involved in any way with terrorism. But that is not the point. The more relevant figure, I maintain, is the conversethe percentage of those noncitizens involved in terrorism who are Arab or Muslim. If there is credible evidence that this percentage is higher in this subgroup than in the general population, then it seems rational for the government to focus particular attention on that group.83 It is simply a matter of channeling inspection resources to places where they are statistically most likely to detect real terrorists.
Professor Cole also offers the following argument: by treating an entire group of people as presumptively suspicious, agents are likely to miss the truly dangerous types who dont fit the profile. He gives the example of the British shoe bomber, Richard Reid.84
Again, I respectfully disagree. The government does not have the resources to scrutinize everyone with the rigor that would be required to ferret out all terrorists. At the other extreme, no responsible government could decide not to scrutinize anyone. That leaves only one other possibility: the government must be selective. But selectivity requires another decision. Either the government selects people randomly, or it selects people based on attributes that it thinks correlate positively with terrorist involvement. Either option has the unavoidable disadvantage that Professor Cole identifiesthat the government will overlook a certain number of terroristsbut the latter option at least mitigates that risk.
Lastly, he argues, since the vast majority of the targeted population will prove to be innocent, agents relying on these profiles will in[*PG22]evitably let their guard down. In this way, overbroad ethnic generalizations . . . may actually undermine effectiveness.85 Yet that problem would be even greater without profiling. Unless the proportion of terrorists that the government would encounter through random sampling (or even nonrandom sampling that avoids ethnicity and religion) is higher than the proportion of terrorists in the Arab and Muslim sample, then the problem of officers letting their guard down because the sampled population is overwhelmingly innocent would be even greater without the ethnic and religious profiling. On the issue of rationality, the real question is whether a positive correlation exists between terrorism and the attributes the government employs in its profiles. If there is not, then all can agree the profiling is irrational; if there is, however, the practice seems rational.
For all the above reasons, it is perilous to rest the case against ethnic and religious profiling in the context of national security on claims that the correlation between the targeted population and the feared misconduct is too minimal. Such claims require a demonstration that those foreign terrorists who seriously contemplate inflicting harm on the United States are not disproportionately Arab or Muslim. In my view, the more convincing arguments against the current profiling practices are those that recognize that rational does not mean justifiable. The countervailing social harms inherent in government-sponsored discrimination are the subject of the next subsection.
One other qualification is necessary. As Leila Sadat has noted,86 many of the current profiling practices crudely lump together people of Arab ethnicity when more nuanced distinctions are available and useful. Not all Arab nations possess identical policies or cultures, and not all Arab populations share common ideologies. It follows that not all nationals, natives, or descendants of all predominantly Arab countriesmuch less all nationals, natives or descendants of all countries that are predominantly either Arab or Muslimcan be gainfully aggregated. The larger point is a more general one. In evaluating a particular profiling practice, one might find it irrational to aggregate dissimilar groups when more refined breakdowns that consider relevant distinguishing features are attainable at costs that are reasonable in relation to the benefits. Nonetheless, this Article will employ a less demanding definition of rationalone that is satisfied so long as [*PG23]the profiling criteria positively correlate to the relevant danger or misconduct. Although the practice might not be optimal, as the magnitude of the positive correlation could be increased by adopting finer distinctions at reasonable data collection costs, this fact does not disturb what I am calling the rationality of the practice. Since the lack of optimal precision diminishes law enforcement gains and increases law enforcement costs, however, it does enter the equation during the second, cost-benefit phase.
The importance of national security in our post-September 11 world is self-evident. If the kinds of ethnic and religious profiling in use today truly yield substantial national security benefitsand whether they do should be determined through the framework suggested abovethen such profiling cannot be dismissed as irrational. Again, however, even if that is the case, rational does not necessarily mean justifiable.87 The harms that flow from government-sponsored discrimination are perhaps equally self-evident. A few of those harms are nonetheless worth emphasizing.
They fall into two categories. In the first category are those harms associated with particular substantive deprivations: the persons removal, detention, arrest, questioning, registration, prosecution, and so forth. Second are those harms associated with the discrimination itself. These includes not only the core injustice of being treated less favorably than others for inadequate reasons, but also personal feelings of humiliation, unfair treatment, a loss of dignity, a loss of confidence, and the sense of being seen and treated as an outsider. These harms have been eloquently articulated in the literature.88
[*PG24] Some might feel that the harmful consequences of discrimination should be discounted when, as is true of most of the practices discussed here, the profiled population consists almost entirely of non-U.S. citizens. The targeted group might, for example, consist only of those non-U.S. citizens who are nationals of designated foreign countries, of Arab descent, or Muslim. Or it might consist of subcategories of those groups, defined by gender and age.
That too is an issue on which opinion is divided. To the extent that noncitizens are thought of as mere guests, rather than full members of the national political community, their interests might be seen as deserving less weight relative to national security needs. And if the government perceives that the particular ethnic subcategory of noncitizens is deriving terrorist-related support from the country or countries of origin, then the sensitive foreign policy concerns might prompt some observers to say the U.S. government should have greater latitude than in a purely domestic context.89 Moreover, as Professor David Martin has long argued, it might be that within the class of noncitizens the content of ones due process rights (and perhaps other constitutional rights) should vary with the differing levels of membership that our immigration laws implicitly assign.90 Others might feel, however, that racial discrimination is racial discrimination, and that it is no more acceptable simply because its victims are not U.S. nationals.
One of the trickier variables is the way the profiling attributes are formally articulated. Suppose, as is true of some of the strategies described earlier, that the targeted group of noncitizens is defined by country of nationality or of birth, not by ethnicity or religion.91 An example would be any policy that singles out nationals of those countries that the President has designated as harboring terrorists. From a policy standpoint, is anything wrong with such criteria? After all, the United States has several other country-specific immigration programsa visa waiver program for certain nationals of certain coun[*PG25]tries,92 the special border-crossing privileges for certain Canadian and Mexican nationals under NAFTA,93 and the per-country numerical ceilings on the admission of immigrants,94 to name a few. Some might say this policy is no different.
Thus, by basing its differential treatment on country of nationality rather than on ethnicity or religion, the government at the very least puts its policy on more respectable footing. Assuming that countries that harbor terrorists are inclined to provide logistical or other support to their own nationals who are planning terrorist acts in the United States, the U.S. policy of targeting the nationals of such countries has a rational basis. Moreover, foreign policy grounds might validate taking a hard line against nationals of a particular country. In contrast, policies that explicitly discriminate on the basis of ethnicity or religion are inherently more suspect.
Still, the distinction should not be overstated. First, almost all the countries that the United States has designated as harboring terrorists are predominantly Arab, Muslim, or both.95 Even if, as it asserts, the administrations policy truly is not driven by ethnicity or religion, the practical impact falls disproportionately on one ethnic and one religious group. Second, so long as the impact is disproportionate, large segments of the population will perceive, rightly or wrongly, that the country-of-nationality distinction is merely a facade for anti-Arab or anti-Muslim sentimentsor at least that the same action would have been politically infeasible had other ethnic groups been singled out.
The justifications for profiling become even more suspect when the existence of profiling attributes must be inferred from appearances rather than more objective criteria. Yet, as a practical matter, such determinations become inevitable once ethnicity or religion become acceptable profiling attributes. Thus, the government might decide to interrogate subclasses of those who appear to be either Muslim or of Arab descent, on the supposition that in each of those subgroups the number of terrorists is higher than in the general population. To that end, interrogators might decide to identify Arabs and Muslims by physical features, clothing, accent, name, neighbor[*PG26]hood, place of worship, and any other discernible correlates of ethnicity or religion.
Of course, many of the same concerns would still be present, including the offensiveness of treating people as suspects based solely on ethnicity or religion. Now, however, additional problems present themselves. First, the police can be wrong about the persons ethnicity or religion. When they are, the policy causes them to waste time on individuals whom there is no reason, either individualized or statistical, to suspect. The national security justifications weaken correspondingly. Second, every police force is comprised of individuals possessing vastly different views and attitudes. Those police officers whose biases incline them toward harassment of particular ethnic or religious subgroups would now have an official excuse.
Yet another obvious variable that directly affects the degree of harm flowing from discrimination is the specific action taken against members of the targeted group. Although some might tolerate profiling when the government simply conducts voluntary interviews of members of the targeted group, they might not be so amenable to the governments employment of profiles as criteria for deportation, detention, special registration, or heightened visa requirements. Others might feel that, while the degree of injustice increases as government actions become more onerous, ethnic profiling is unjust even when the government intrusion is otherwise relatively minor.
Given the preceding background and policy framework, one can now consider whether the various U.S. immigration-related national security measures are compatible with the United States obligations under the Race Convention.96 Adopted by resolution of the U.N. General Assembly in 1966, and entering into force as a treaty in 1969, the Convention reflects the worlds shared understandings that racial discrimination is wrong and that collective international efforts to eradicate it are essential.97 The United States ratified the Convention in 1994, but with several significant reservations, understandings, and declarations.98
[*PG27] After defining racial discrimination in article 1, the Race Convention prescribes several actions that the states parties agree to take (or refrain from taking) to combat either racial discrimination generally or particular manifestations of it. The Convention also provides several means of enforcement, including domestic remedies, an expectation of dispute resolution by the International Court of Justice (ICJ), and the creation of a body called the Committee on the Elimination of Racial Discrimination (CERD).99
The role of CERD is critical to the Conventions success. Consisting of eighteen experts of high moral standing and acknowledged impartiality elected by States Parties,100 CERD is entrusted with several supervisory functions. It receives and reviews reports submitted by each of the states parties, detailing the measures they have taken to implement the Convention.101 It also issues general recommendations to the U.N. General Assembly. Additionally, CERD has the power to receive complaints by one state party that another state party has violated the Convention.102 CERD may also receive communications filed by individuals or groups against any states parties that declare their recognition of CERDs competence over individual complaints.103
In the case of the United States, these enforcement mechanisms have been constrained. The U.S. Senate made ratification subject to a declaration that the provisions of the Convention are not self-executing.104 Some doubt exists as to the validity of the non-self-executing declaration. The Vienna Convention on the Law of Treaties105 prohibits reservations that are incompatible with the object and purpose of the treaty, and the Human Rights Committee, the supervisory body established by the ICCPR,106 has questioned the [*PG28]compatibility of non-self-executing reservations with human rights treaties.107 If the reservation is invalid, then the Human Rights Committees view is that the state is bound to the treaty without the reservation.108 If the reservation is valid, however, its practical effect under U.S. domestic law is that U.S. courts have no jurisdiction to hear private causes of action arising under the Convention.
Even if domestic enforcement proves unattainable, however, the United States remains bound by its Convention obligations on the international plane. Although subject to neither nonconsensual dispute resolution before the ICJ109 nor the individual communications authority of CERD,110 the United States remains subject to the other powers of CERD. Those include the authority to review U.S. country reports, file reports with the General Assembly, and receive interstate complaints.
Do the national security measures summarized in Section I of this Article comport with United States obligations under the Race Convention? As will be seen, most of the critical Convention provisions refer to state obligations that relate to racial discrimination. The starting point therefore, is article 1.1, which defines that term to mean:
any distinction, exclusion, restriction or preference based on race, colour, descent, or national or ethnic origin which has the purpose or effect of nullifying or impairing the recognition, enjoyment or exercise, on an equal footing, of human rights and fundamental freedoms in the political, economic, social, cultural or any other field of public life.111
As applied to the recent U.S. national security initiatives that operate specifically on non-U.S. citizens, this definition of racial discrimination gives rise to two sets of broad issues. First, do the actions of the United States truly embody distinctions based on national or ethnic origin?112 Second, do theyand, to be covered by the Convention, must theyimplicate human rights and fundamental freedoms?113 [*PG29]The following commentary attempts only to identify and synthesize possible strategies for assessing United States compliance with the Race Convention, not to offer definitive solutions.
One could construct at least three theories for characterizing some of the immigration-related national security programs as resting on national or ethnic origin distinctions for purposes of article 1.1 of the Convention. For some programs, one might theorize that the statute, regulation, agency policy, or actual practice expressly and facially embodies such distinctions. Other programs, while facially neutral with respect to nationality and ethnicity, might be claimed to be purposeful pretexts for nationality- or ethnicity-based enforcement. Still other programs, although neither expressly nor pretextually intended to distinguish by nationality or ethnicity, might nonetheless have the effect of doing so within the meaning of article 1.1. Of course, there might be programs for which all, some, or none of these three arguments are possible.
With respect to express intentional distinctions based on national or ethnic origin, articles 1.2 and 1.3 are instructive. Article 1.2 sets out the commonsense proposition that a state does not engage in national or ethnic origin discrimination simply by treating its own nationals more favorably than those of other states. Indeed, if it were otherwise, virtually all nationality classifications, including all immigration laws, would be prohibiteda proposition one can safely assume the States Parties did not intend to adopt.114
[*PG30] Article 1.3, however, says nothing in this Convention may be interpreted as affecting [states provisions on nationality], provided that such provisions do not discriminate against any particular nationality.115 CERD has confirmed that article 1.3 was meant to qualify article 1.2.116
As Section I demonstrated, however, many of the post-September 11 national security initiatives do precisely what article 1.3 warns against; they single out nationals of selected countries (generally, countries predesignated as linked to terrorism) for special, additional restrictions, some substantive and others procedural. Does article 1.3 suggest that such measures constitute national origin distinctions? And if so, what is the fate of the many other U.S. immigration provisions mentioned earlier that provide for the special treatment of nationals of specific countries, outside the national security context?117 Does article 1.3 prohibit them as well?
Such a conclusion is hard to imagine, for states commonly enter into free or preferential travel pacts, as well as other immigration arrangements with particular neighboring states, former colonies, and important trading partners. The European Union and other regional associations see nothing wrong in making special arrangements for fellow members of their associations. Surely, states with such extensive commitments did not agree to a convention that would render all preferential arrangements illegal.
It is true that the national security programs discussed in Section I tend to make adverse treatment exceptional, in contrast to the other cited examples in which favorable treatment is the exception. But there is no apparent policy reason that the validity of distinctions between nationals of some foreign states and nationals of others should hinge on whether the favored states are a majority or a minority. In each case, nationals of some states are treated more favorably than similarly-situated nationals of other states. Either the Convention permits that practice, one would think, or it does not.
My view is that the cleanest and most sensible way to analyze this problem is to give article 1.3 its literal meaning. If a nationality law does not discriminate against any particular nationality, then article 1.3 makes clear that nothing in the Convention prohibits such a law.118 But [*PG31]the converse is not implied; the fact that a nationality law does discriminate against a particular nationality does not necessarily mean the law violates the Convention. It means only that other considerations become relevant. For instance, even if one finds the distinction to be based on national origin within the meaning of article 1.1, one must ask whether it impairs a human right or fundamental freedom. If so, does sufficient countervailing justification exist? Conversely, even if the distinction is found not to be based on national origin, might it be a pretext for distinguishing according to ethnic origin? Or might its effect be so ethnically disparate that the law violates the Convention?
While the preceding discussion focuses on laws, regulations, and other official departmental policies that expressly differentiate based on national or ethnic origin, express discrimination would also violate the Convention if effected by the individual agents or employees implementing the general laws or policies. That much is evident from article 1.1, which makes no attempt to distinguish between those who discriminate in forming a broadly applicable policy and those who discriminate in carrying it out.119 Thus, even facially neutral U.S. immigration enforcement programs would violate article 1.1 if it could be shown that in practice government agents seek out suspects according to ethnic origin.120
A second theoryor, more accurately, a second species of intentional discriminationis that a facially neutral policy is a pretext for national or ethnic origin discrimination. When that is the case, then no obvious legal issue exists, other than possibly the question of whether the particular individual interest rises to the level of a human right or fundamental freedom. The only real question would be factual: whether there is persuasive evidence of pretext.
In the present context, this theory could surface in either of two ways. First, many of the national security programs discussed in Section I expressly prescribe adverse treatment for nationals of selected states. Even if those distinctions are not themselves characterized as national origin distinctions, they might be found to be pretexts for ethnic origin distinctions. Second, even those national security programs that are facially neutral with respect to country of nationality might be found to be pretexts for national or ethnic origin discrimination if the officials who formulated the programs anticipated that enforcement personnel [*PG32]would in fact apply the requirements more vigorously to individuals of particular national or ethnic origin. In either case, the question would be one of fact, and the burden would be to establish a hidden motive for the policy or the enforcement practice.
In a given case, a good starting point for establishing pretext would be a showing that the stated justification does not withstand serious scrutiny. In Habassi v. Denmark,121 CERD held that in that event it is appropriate to initiate a proper investigation into the real reasons behind the . . . policy.122
A third theory for challenging many of the U.S. profiling strategiesand perhaps even those that do not officially rely on profiling is based on disparate impact. As noted earlier, the Conventions definition of racial discrimination includes national or ethnic origin distinctions that have either the purpose or the effect of impairing the exercise of human rights and fundamental freedoms on an equal footing.123 CERD has made clear that the test for effect is whether that action has an unjustifiable disparate impact upon a group distinguished by race, colour, descent, or national or ethnic origin.124
Thus far, CERD has provided little guidance as to the standard that unjustifiable demands. On the one hand, even a legitimate motive can leave a disparate impact unjustified. In considering a country report filed by Denmark, CERD disapproved of the Danish policy of bussing non-Danish-speaking schoolchildren to special reception classes when the effect is discriminatory, even though CERD acknowledged that it understands the reasons.125 On the other hand, in reviewing Australias policy of administering additional medical exams to graduates (Australian or otherwise) of foreign medical schools, CERD found no discriminatory effect, even though the pass rates varied considerably from one foreign country of birth to another.126
[*PG33] After the attacks of September 11, 2001, and at the request of the U.N. High Commissioner for Human Rights, CERD issued an official statement on the general subject of discrimination in the context of the war on terror.127 Condemning the terrorist attacks unequivocally, CERD emphasized that counter-terrorism measures must nonetheless respect international human rights law, that the prohibition on racial discrimination is a peremptoryand thus, non-derogable norm, and that states must ensure that their counter-terrorism programs do not discriminate in purpose or effect on grounds of . . . national or ethnic origin.128
This is vague language indeed, and the question is how to apply it to the recent spate of U.S. counter-terrorism strategies that operate in the immigration arena. Certainly, if the implementation of any of the policies can be shown to target Muslims, then a disparate effect on noncitizens of Arab descent would follow and would arguably constitute ethnic origin discrimination.129 Similarly, even the programs that single out either nationals or natives of countries thought to support terrorism are likely to have a disparate effect on individuals of Arab ethnic origin. Whether they violate the Race Convention again requires an assessment of whether they are justifiable. Finally, even the facially neutral programs might have unjustifiable and disparate effects if they leave the enforcement officials with impermissibly broad discretion that promotes discriminatory enforcement. CERD expressed just such a concern over French policies that gave police officers broad discretion to check the identities of foreigners in public, a practice that CERD feared could foster discrimination.130
Even when the particular national security strategy is ultimately characterized as embodying distinctions based on race, colour, descent, or national or ethnic origin, the definition of racial discrimination in article 1.1 additionally requires the impairment of human rights and fundamental freedoms.131 That qualifier distinguishes ar[*PG34]ticle 1.1 from a general equal protection clause, where protection does not require a showing that the interest being selectively denied rises to any particular level. Since article 1.1 defines racial discrimination in this limited manner, other provisions of the Convention that refer to that language are correspondingly restricted.132 One such restricted provision is article 2, which obligates states parties to eliminate racial discrimination and to refrain from engaging in it.
Article 5 reads:
In compliance with the fundamental obligations laid down in article 2 of this Convention, States Parties undertake to prohibit and to eliminate racial discrimination in all its forms and to guarantee the right of everyone, without distinction as to race, colour, or national or ethnic origin, to equality before the law, notably in the enjoyment of the following rights . . . .133
The provision then goes on to list various rights, some of which are discussed below.
One uncertainty is whether the second prong of article 5 (and to guarantee the right of everyone), like the first (undertake to prohibit and to eliminate racial discrimination), is qualified by the introductory phrase referring to obligations laid down in article 2.134 If yes, then both prongs of article 5 are confined to distinctions that impair human rights and fundamental freedoms, because article 2 is expressly limited to addressing racial discrimination as defined in article 1.135 If no, then article 5 ushers in a broad equal protection guarantee, one that prohibits unjustifiable discrimination based on race, colour, or national or ethnic origin even when internationally recognized human rights or fundamental freedoms are not otherwise at stake.
The choice of interpretations could have implications for the U.S. national security initiatives discussed earlier.136 As noted previously, some of the programs affect such important human rights as family unification, freedom from arbitrary detention, asylum from persecution, and the like, while others affect only lesser interests.137
[*PG35] The arguments favoring the narrower reading of article 5 are substantial. After the opening qualifier, the subject States Parties appears only once and the two prongs that follow it are not separated by a comma or other punctuation. Therefore, a literal grammatical reading leaves the opening qualifier applicable to both prongs. Moreover, if the second prong of article 5 were read as a general equal protection clause, there would have been no point in confining the racial discrimination definition in article 1 to distinctions that impair human rights and fundamental freedoms and then confining articles 2 through 7 to situations embraced by that term; in any case in which a lesser substantive interest was at risk, article 5 would make that limitation superfluous.
But if that is the case, then what is the point of article 5? The most likely explanation is that by listing examples of rights that cannot be denied selectively, and by strategically using the word notably, article 5 supplies a non-exhaustive catalog of the kinds of human rights and fundamental freedoms envisioned by the definition of racial discrimination.138 This interpretation finds support in CERDs General Recommendation XX, handed down in 1996. There CERD noted that many of the rights listed in article 5 emanate from other international human rights instruments and that the listed rights do not constitute an exhaustive list.139
A brief glance at the rights enumerated in article 5 reveals the breadth of individual interests which the Convention regards as worthy of protection from discrimination. An impressive array of rights is covered, ranging from the civil and political to the economic, social, and cultural. The first ones listedfreedom of movement and residence within the borders of the Stateare directly relevant.140 They correspond, respectively, to those U.S. national security strategies that entail preventive detention and loss of procedural safeguards in the removal process.
CERD, reviewing a French country report in 1994, has reinforced that view. In an apparent reference to article 1.1 of the Convention, [*PG36]CERD expressed concern that a recently enacted French immigration law could have racially discriminatory consequences, because it restricted the right to appeal expulsion orders and prescribed preventive detention for excessively long periods.141 The report also voiced concern over procedures concerning identity controls that allowed broad police discretion, expressing the fear that such unrestrained discretion could promote discrimination.142 To describe the feared consequences as discrimination prohibited by the Convention, however, CERD must have regarded the police checks of foreigners identities as implicating human rights or fundamental freedoms. If that is the case, then several of the U.S. intelligence-gathering strategies, even as applied to non-U.S. citizens, are also potentially vulnerable to challenges under articles 1.1 or 5.
The United States has responded to the threat of international terrorism in controversial ways. Some U.S. counter-terrorism strategies have entailed one form or another of ethnic or religious profiling. While this package of responses cannot be dismissed as inherently irrational, many of the specific profiling programs carry social costs that this Article finds excessive in relation to any law enforcement gains they might be thought to yield. Perhaps more significantly, the various profiling programs have international human rights implications, raising serious questions of compatibility with U.S. obligations under the Race Convention.