The Incidental State: Coercion in the Age of Big Data

The Incidental State: Coercion in the Age of Big Data

Inside the Charleston Police Headquarters, outfitted by IBM (ibmphoto24/Flickr)

When government officials exert their most coercive powers, the logic behind their actions should be obvious. Only autocratic states deprive people of their liberty without an explicit rationale, and this country’s political forebears took pains to circumscribe the government’s legal authorities, limiting their potential for abuse.

Yet today coercive government power, already strengthened by wars on drugs and “terror,” has been advanced further by innocuous-seeming databases sharing data on a scale, and with an agility, never before possible. While distressing in its own right, the recent unwieldy expansion of the state’s coercive powers—fraught with unintended consequences and counterproductive results—also makes us all less safe. At the same time, the new data imperative guiding coercive force shifts the balance of political economy, bolstering people, arguments, and institutions invested in punishment while silently and powerfully reinforcing society’s commitment to the same.

Our present-day deportation machine provides the most dramatic example of the pernicious effects of the data imperative. When local police officers make a citation or arrest someone, they run that person’s name through an FBI database to see if the person has any outstanding warrants; since 2008, under a program called “Secure Communities,” they also run the name through the Immigration and Customs Enforcement (ICE) database to check for any issues with immigration status, including expired visas. This simple act deputizes local police into magistrates of the federal government, a move that is itself an affront to the American political tradition of local law enforcement.

Moreover, while adjudicating immigration status is an administrative rather than criminal proceeding, under Secure Communities undocumented residents—many of whom have resided in the country for over a decade without committing any serious infraction—can be handcuffed and forcibly detained in what can only be described as a prison. Immigrant detention centers, many of them run by private contractors, are rife with abuse and sexual assault, as NPR’s Maria Hinojosa reported for Frontline in 2011. In that same year, operating under a congressional mandate to deport 400,000 people annually in order to maintain its funding, ICE paid private detention contractors $80 to $120 per detainee per day.

The New York Times recently disclosed that the vast majority of deportations only involve minor offenses.  A speeding ticket, for example, could to lead to deportation. Feasting off the low-hanging fruit, the U.S. deportation machine tears families apart, creates resentment, and drives a wedge between law enforcement and low-income immigrant communities.

To their credit, some state and local officials have begun to push back against this trend by instructing their police to not hold undocumented migrants they would otherwise release. (Most prominent among them is Maryland Governor Martin O’Malley, widely thought to be a contender for the Democratic Party presidential nomination in 2016.)  But among policymakers, the backlash against the “Secure Communities” program—albeit growing—remains small, and the discussion of it is by and large restricted to immigration politics, rather than a broader consideration of state power.

By automating what was once a sequence of discretionary decisions, law enforcement agencies lower the bar for the application of force.

Even when immigration status is not at stake, the new possibilities of data sharing foster a perversely punitive approach to law enforcement—particularly since, in some jurisdictions, not only the perpetrators but the victims of crimes have their names run through the scanner. As Joseph Goldstein recently reported in the New York Times, the New York Police Department has developed the practice of handcuffing shooting victims to their hospital beds if they have an outstanding warrant. This is done without any review of the warrant or the crime that prompted it, leaving recovering victims shackled to their beds for trivial offenses like failing to answer a charge of disorderly conduct.

Enhanced data collection enables the NYPD to research a victim’s background quickly and easily. Launched in 2012 through a partnership with Microsoft and soon to be deployed as a portable application—as well as marketed to other cities for a profit—the NYPD’s Domain Awareness System brought an end, in the words of then-Commissioner Raymond Kelly, to a time when the department was “stovepiped as far as databases are concerned.” In order to assess a victim’s credibility, or identify a suspect who may be known to the victim, officers can search records instantly, and they do so as matter of course. Before Domain Awareness, an officer would have to decide whether a search for outstanding warrants for the victim was worth the effort. Now, with no effort demanded, no such decision is made.

There is nothing inherently nefarious in the ability of a government agency to share information or plumb its own records. But as law enforcement agencies invest more and more resources into collecting and sharing data, particularly data about people and not about crime, they broaden the scope of their activities, and, by collapsing or automating what was once a sequence of discretionary decisions, they lower the bar for the application of force. Although it is tempting to think that all warrants should be heeded or, as some conservatives urge, that all undocumented migrants should be deported, it is important to realize the ways in which this strategy differs from the universal application of the law, a desirable component of justice. Because of differences in how law enforcement officers conduct themselves in various settings, leveraging a data set for maximum effect aggravates the disparities in criminal justice many of us already find troubling.

It is a worrying paradox: because, according to legal scholar Bill Stuntz, over 70 percent of Americans have committed an imprisonable offense, and because the nation relies on undocumented workers to perform essential work, universal enforcement remains unthinkable. Instead, data-driven law enforcement reinforces existing patterns in selective enforcement, be it undocumented residents who make no effort to shield their presence or similarly “easy targets” in neighborhoods subject to disorder policing or other aggressive strategies. If those with more privilege came under the same scrutiny, the country would become a giant penal colony.

In the past, bureaucratic disconnects and inefficiencies gave rise to horrible instances of failure; but, in a blanket and indifferent fashion, these same attributes also lent themselves to a kind of latent justice, acting as an inadvertent bulwark of federalism and an operational restraint on state power. Today, big data and mobile access lure law enforcement and security agencies with the possibility of finding the proverbial “needle,” but a relentless search conducted without discernment risks toppling the stack.

Since September 11, the government has raced to combine and upgrade datasets and reward the sharing of information. Despite these efforts, some “dots” remain unconnected, because the people and organizations that generate them remain highly imperfect. Yet gradually and for the most part unobtrusively, these efforts have produced countless uses of coercive state power that are more incidental than essential; guided more by what can be done rather than what would be smart to do; and biased toward data that can be readily submitted and searched, rather than information derived from a consideration of context and consequences.

One patient profiled by Joseph Goldstein was chained to his gurney while recovering from a gun wound because of an outstanding warrant for perching a cup of wine on a backyard fence at a barbecue (and his inability to find a babysitter for his day in court). Bound to his bed, the victim heard the sounds of the popular game “Candy Crush” coming from the phone of the police officer who stood guard outside his hospital room. In this Kafkaesque scene, technology is both a diversion and a driving factor, producing a moment of banality and detachment in the application of government power that should alarm any student of twentieth century history.


Kathleen J. Frydl is an award-winning scholar of the U.S. state. Her latest book is The Drug Wars in America, 1940-1973, published by Cambridge University Press in 2013.