What to do with populations in which we can identify threats ex ante only probabilistically? This is a central issue in the U.S. debate over gun rights and gun control, and, in many respects, this debate parallels the argument over refugee admissions policy. Some commentators, mainly on the left, argue that the threat of gun violence, and the risk of accident, is so high that, even though the vast majority of gun owners never shoot someone else, the level of gun ownership should be restricted significantly relative to today’s levels. Other commentators, mainly on the right, argue that gun ownership rights are important to allow people to defend themselves and others. So the right needs to be protected despite the cost in lives and injuries from criminal gun violence and accidents.
If we knew for sure in advance who would use guns to commit crimes, then we could tailor policies to disarm only those individuals. But we don’t, and we can’t. We can probabilistically identify some people with traits who are more likely to use guns for criminal violence relative to others. Denying guns to felons convicted of violent crimes in the past, for example, seems to be a wise policy, as is denying them to people with manifest psychological traits that put them at obvious risk to themselves or others. But even then, people without these traits, or at least without manifest or sufficient evidence of these traits, will use guns to commit violent crimes. So broadly aimed gun-control policies seek to reduce the number of guns in the population overall.
How to respond to risks and threats when one cannot effectively drill down beyond the level of the population itself?
For gun rights, many advocates conclude gun rights need to be protected because of the significance of the right, the current social cost in gun violence and accidents notwithstanding. Liberals, too, are willing to trade away some safety for some rights, albeit, usually for rights other than gun rights. And both right and left would overwhelmingly oppose, say, random house-to-house searches for contraband or evidence of criminal activity. This, despite that police activity of that sort would undoubtedly uncover or deter many crimes.
The intuition about tradeoffs and probabilistic threats applies to policies more broadly than judges apply it, nonetheless, the intuition here is one behind the approach judges often take when they apply the Fourteenth Amendment’s Equal Protection Clause.
At least four variables are involved in asking about the reasonability of a policy classification: What is the trait to be regulated? What is the harm to be avoided? How many people who do not commit the harm will be burdened, and what is the magnitude of the burden placed upon those individuals?
Consider a proposal to limit guns in the population at large: The trait to be regulated is gun ownership. The harms to be avoided are gun violence and gun accidents. The issue is the “fit” between those who hold the regulated trait and those who do the harm. As I mentioned above, the problem with gun control policies is that they are massively overinclusive: The vast majority of people who have the trait (i.e., they own guns) do not cause the harm (their guns are not involved in criminal violence or accidents), yet they would be burdened by policies denying them guns.
Much of the policy disagreement over gun rights and gun control in the U.S. results from different weights that different people place on the different variables. Even if everyone agreed on the objective data – how much gun violence and accidents would be reduced and how many people would be criminally victimized because putative gun owners cannot defend themselves or others – different people can weigh the respective variables differently, and come to different conclusions about the reasonability of the “fit” between the classification and the harms.
I chose a “conservative” example on purpose. With respect to guns, conservatives are willing to tolerate a level of social violence and risk to protect a right. The same reasoning can apply when considering policies related to admitting refugees fleeing of necessity from violence in their own countries. The problem is that terrorists might immigrate under the pretext of being a refugee, and attack people in the nation in which they relocated.
The point is not that refugees must always be admitted regardless of the risk they pose. Far from it. But magnitudes and probabilities matter. As with gun rights, conservatives are willing to accept a level of social harm because the corresponding right is so important and the difficulty of identifying individuals who will commit the harm is too high.
To be sure, the first response of many will be, “But nations do not admit refugees as a matter of right, therefore the initial premise of your argument is wrong.” I agree as a matter of positive law that nations need not admit as a matter of right refugees who, out of necessity, must leave their home country. But that’s tautologically true. There is no positive law right if there is no positive law right.
But the question is not quite as easily answered for people, or for nations, who, or which, affirm commitments to natural rights, such as those articulated in the Declaration of Independence. As I’ve discussed in recent posts, here, here, and here, Hugo Grotius is one of the important natural rights philosophers who developed the tradition articulated in the Declaration. He argues as a matter of natural right that refugees created by “necessity,” such as escaping violence in their home country, have a right to immigrate to other nations, a right that stems from the original collective ownership of the earth.
But even the natural right to take refuge, under necessity, in other countries is not, for Grotius, and absolute right. Grotius writes:
Nor ought a permanent residence to be refused to foreigners, who, driven from their own country, seek a place or refuge. But then it is only upon condition that they submit to the established laws of the place, and avoid every occasion of exciting tumult and sedition. . . . To drive away refugees, says Strabo, from Eratosthenes, is acting like barbarians; and a conduct like this in the Spartans was also condemned. St. Ambrose passes the same sentence of condemnation upon those powers, who refuse all admission to strangers.
An argument we hear often today with respect to refugees from Syria (and from other countries) pertains to Grotius’s proviso that the duty to admit refugees depends on the “condition that they submit to the established laws of the place, and avoid every occasion of exciting tumult and sedition.” In the case of Syrian refugees, leaders and citizens fear that terrorists enter their nations hidden among, and under the pretext of being, refugees, but with the intent to attack and injure people in those countries.
Grotius recognizes threat of “tumult and sedition” as reasons that governments can justly to exclude threatening refugees. On the other hand, Grotius also condemns, in St. Ambrose’s words, nations “who refuse all admission to strangers.”
The question for Grotius’ theory is what threshold of risk must be met before exclusion of all refugees from a given country or area is reasonable?
Obviously any individual known to be a terrorist, or even suspected to be a terrorist with any specific evidence or warrant, can be excluded. The problem is when we deal with a large numbers of individuals, the vast majority of whom hold no threat and seek merely to escape violence in their home country, and there is no evidence to suspect one person over another.
It again deserves emphasis nations accepting refugees can consider both the risk and the magnitude of the prospective harm. There are parameters under which exclusion is reasonable. But the notion that that any threat at all, no matter how remote, justifies exclusion of all refugees from a country or region, is simply inconsistent with Grotius’s argument. The class of excluded individuals – all refugees from Syria – is overinclusive relative to the individuals who threaten harm.
To be sure, if admission of refugees were merely a matter of charity, then an overinclusive policy of exclusion wouldn’t matter as an issue of natural justice. But Grotius argues nations must admit as a matter of right refugees who flee out of necessity. If Grotius is correct, then the U.S. (and other countries) need to ask whether the “fit” between the excluded trait (namely, being Syrian) and the harm to be avoided (terrorist activity) is close enough to be consistent with the demands of justice. To be sure, the danger to be avoided, terrorist activity, is a danger with potentially a great magnitude of harm. But the excluded class is also massively overinclusive relative to the group we really want to exclude, those who intend us harm.
Again, the point is not that exclusion is perforce unjust. The point rather is that, if Grotius is correct, nations do not have the option of playing it entirely safe, excluding all refugees from a country if the probability and magnitude of harm are extremely low, but nonetheless greater than zero.
Rather, in a conflict of right with right, policy makers who acknowledge the obligations of Grotius’s form of natural right need to balance the harm to the refugees by excluding them as a class relative to the threat to the policy makers’ nation by admitting them as a class. (And, again, individuals who are known threats can be excluded. There is no problem at all with careful vetting of refugees.) This balancing does not require a commitment to an extreme form of cosmopolitanism in which U.S. policy makers are not allowed to weigh citizens’ lives more heavily than non-citizens’ lives. But even if the level of acceptable risk is appropriately lower with respect to admitting refugees from other nations, the form of a nation’s inquiry should be recognizably related to tradeoffs those nations make when weighing the risks and burdens of overinclusive policies that affect their own citizens.