Proceduralizing detention through evidence-based risk assessments

Immigration imprisonment is on the rise everywhere. As with many fast growing industries, its expansion is often chaotic and seemingly rudderless. Evidence-based risk assessments offer the potential to discipline the decision to detain—to provide a reasoned anchor that guides the state’s decision to deprive a person of her liberty. In making detention decisions more regimented, I’m concerned that we might simply be adding procedure without altering the nature and extent of confinement. More importantly, I fear that proceduralizing detention through ostensibly objective criteria masks the need to ask whether immigration imprisonment is even defensible.

In the United States, we have become accustomed to seeing more than 400,000 people locked behind barbed wire every year while the federal government decides whether they will be allowed to remain. Our willingness to stomach immigration imprisonment is so deep that we have resumed a practice we seemed to have largely abandoned just a few years ago: detention of children and families. These days, toddlers are literally locked away for having done nothing worse than ask the United States for refuge.

While the United States certainly claims the largest number of people confined due to a suspected or confirmed immigration law violation, we’re far from alone. Much of Western Europe also relies on detention. Italy’s role in responding to Middle Eastern and African migration has made headlines in recent months because of the large number of deaths that have resulted in the Mediterranean, but lesser known are the long-standing detention practices that many countries throughout the continent use. From Austria to the United Kingdom, almost every Western Europe country has some immigration detention practice. After Spain and Italy boosted their enforcement apparatus, migrants turned toward Greece where, when caught, they were frequently treated so poorly that multiple decisions of the European Court of Human Rights concluded that immigration confinement in Greece amounted to degrading treatment. To the credit of the current Greek government, they have announced a desire to shift away from immigration detention. These examples could go on: in Libya, in México, in the Pacific island of Nauru that houses detainees on behalf of Australia.

Critics frequently complain that these facilities detain the wrong people. It’s hard to dispute that when the list of detainees worldwide includes people with no criminal history, asylum seekers, and even, as in the case of the United States, children.

Evidence-based risk assessments promise to change that. Instead of imprisoning everyone, risk assessments can help government officials tailor detention more carefully to target those individuals who should be confined. Like a surgeon wielding a scalpel, risk assessments can help separate the good from the bad. People who are likely to disappear can be confined while those with good reasons to appear for court dates can be let free. Similarly, people who pose a public safety threat can be locked up while those who show no indication of dangerousness can be released.

In this way, risk assessments hold out the possibility of better equipping immigration officials to realize the two readily accepted justifications for imprisoning migrants thought to have violated immigration law: flight risk and dangerousness. In the United States, this fits neatly with the Obama Administration’s repeated assertion that it is focused on enforcing immigration laws against “felons not families.” Putting aside the obvious fallacy that these two categories can be neatly divided (felons, after all, are often part of families, just like families often include felons) and that the reality of immigration enforcement under President Obama doesn’t reflect this rhetoric, this catchy claim illustrates the widely commended goal of using finite governmental resources against the worst migrants while leaving the rest at peace from detention.

Critics also frequently complain governments simply detain too many migrants. Risk assessments, it’s thought, might reduce the immigration prison population by making it clear that most people currently detained don’t need to be. As Mark Noferi put it in a 2014 article, “the logical extension of tailoring a detention system to its civil aims is less detention.”

But can evidence-based risk assessments actually do all this? Can they ensure that only people destined to flee or likely to endanger the community will be detained? And that, in the process, there will be fewer people locked up? Perhaps, but the early results and baseline assumptions of risk assessment analyses make me skeptical. In my view, there are a number of risks associated with risk assessments that need careful empirical scrutiny and normative justification.

To begin with, any risk assessment algorithm is only as good as the data it uses. Unreliable data will result in misleading outcomes that are nonetheless cloaked in the veneer of plug-and-play analyses. Tim Dekkers’ study of Dutch border police illustrates this risk. According to Dekkers, Dutch officials conducting intra-EU border surveillance using a risk assessment technology called @migo-boras regularly engage in national-origin profiling based on a vehicle’s license plate. Not surprisingly, this broad strategy results in many false positives. In the United States, we don’t even know how ICE weighs various factors in its risk assessment so it is impossible to know just how smart its algorithm is. As Mark Noferi and Robert Koulish wrote on crImmigration.com in 2013, If the risk tool says all immigrants are risky, detention reforms will be for naught.”

Relatedly, risk assessment frameworks can easily become beholden to political expedience or individual bias. In the United States, for example, risk assessments that turn on whether a person is subject to mandatory detention under the Immigration and Nationality Act reflect congressional will more than they do anything about the migrant’s likelihood to flee or endanger the community.

Secondly, risk assessments raise important normative questions about the political rhetoric of immigration law enforcement. At bottom, an assessment of a person’s risk suggests an element of dangerousness. There must be something inherently suspicious about migrants if we need to weigh the amount of risk they pose to the rest of us. This is familiar language of course. Criminal policing and judicial adjudications routinely use the language of risks and threats. By adopting this discursive frame, risk assessments run the risk of validating policymaking premised on that very security-focused frame in which conduct by “them” poses a danger to “us” and thus policymakers are right to protect the good from the bad through the segregative power of imprisonment.

Given this premise, the need for detention is assumed. There is a danger among us—a risk—that must be contained. Imprisonment offers the promise of brick and concrete protection. The only questioning left becomes the question at the heart of evidence-based risk assessment regimes: who should we confine? By the time we’re asking this question, we’ve given up on whether anyone should be confined at all. Given up on that initial inquiry with a solid normative foundation is problematic as a theoretical matter and destructive as a matter of human dignity.

A third problem risk assessments raise is very practical: they facilitate an expanded surveillance regime. Some folks will be analyzed according to the risk assessment algorithm and imprisoned. Others will be deemed unworthy of confinement. Rather than let them go free of governmental intrusions into their affairs, risk assessments have a tendency to expand surveillance to people who would not previously have been supervised. A form of this occurred in a noncustodial supervision pilot project that the Vera Institute of Justice operated on the INS’s behalf in the 1990s in which people who would not have been detained were required to submit to supervision.

By measuring a person’s characteristics and balancing that against some acceptable level of risk, we might quickly see risk assessments justify watching people who government officials decide pose a small amount of risk just not enough to detain. These individuals will then be let out of prison doors with ankle bracelets, reporting requirements, and the like. In the past, many of these individuals would’ve been let go without such restrictions on their liberty. Indeed, in the penal context, research suggests that noncustodial supervision alternatives to incarceration expanded the scope of governmental surveillance to people who would otherwise not merit much supervision.

Imposing restrictive conditions begs the question of what happens when people violate those conditions. In the U.S. context, for example, what should ICE do when a migrant placed into an alternative surveillance regime disables the ankle bracelet or fails to show up for one of the in-person reporting appointments? A February 2015 DHS Inspector General report suggests that the solution is to detain such individuals. This points in a worrying direction.

Ultimately, I’m concerned that the move toward evidence-based approaches to imprisonment won’t lead to a carceral regime where only those people who can’t be trusted to show up for court dates and avoid injuring others will be locked up. Instead, I’m concerned that risk assessments will simply sanitize imprisonment.

This essay builds on a talk I presented at the Law and Society Annual Conference in Seattle on May 29, 2015.

---

Find this information useful? Then let others know about crImmigration.com, as well as César’s Twitter, Facebook, and LinkedIn pages. And to make sure you don’t miss an update, subscribe to the blog by entering your email address in the subscription box that appears on every page.