Marion Fourcade and Kieran Healy (2017) introduced the wonderful concept of ubercapital, “a form of capital flowing from [individuals’] positions as measured by various digital scoring and ranking methods” with consequences for stratification. Two new papers provide clear (and terrifying) analysis of how ubercapital works in hiring and law enforcement.
In “The art of deciding with data: evidence from how employers translate credit reports into hiring decisions”, Barbara Kiviat explores how hiring managers use credit data to determine who (not) to hire. Despite an absence of data showing links between aspects of credit reports and workplace outcomes, Kiviat finds that hiring managers routinely rely on credit data to inform hiring decisions, using various ad-hoc rules along with moral storytelling. That is, hiring managers tried to create a narrative that contextualized credit information and thus justified its inclusion or exclusion from a hiring decision. For example, medical debt was often seen as more legitimate, an outcome of an unfair health system and necessary expenditures, not bad character:
The moral ambiguity of unpaid medical debt was so great that one company taking a calculative approach baked it into the rules. A recruiter for a financial data firm said that while company policy was not to hire anyone with more than two accounts in default, medical debts did not count toward the total.
Other debts were less forgivable. Perhaps the most telling quote comes from a fast food hiring manager explaining how she interpreted different classes of debt:
At one fast-food company, the director of human resources said that after the most recent recession she grew more sympathetic about late credit card payments, but that she would still never hire a job candidate who was behind on child support, since providing for one’s children should always be a top priority: ‘If you don’t have a job, get a job, you know? Figure it out, take care of your child.’
Here we see ubercapital’s stratification effects: certain kinds of debts (incurred, perhaps, because of the absence of work at a critical moment) lead one to become morally unworthy of employment in a vicious loop, showcasing how information spills over from one domain to another.
In “Big Data Surveillance: The Case of Policing”, Sarah Brayne examines the LAPD’s increasing reliance on new forms of data to guide policing. Brayne identifies five consequences of the increasing reliance on big data approaches:
First, discretionary assessments of risk are supplemented and quantified using risk scores. Second, data are used for predictive, rather than reactive or explanatory, purposes. Third, the proliferation of automatic alert systems makes it possible to systematically surveil an unprecedentedly large number of people. Fourth, the threshold for inclusion in law enforcement databases is lower, now including individuals who have not had direct police contact. Fifth, previously separate data systems are merged, facilitating the spread of surveillance into a wide range of institutions.
The fourth and fifth consequences are especially interesting to consider in the context of ubercapital. Both involve the merging of previously disparate sorts of data, and the creeping expansion of what data are deemed relevant to a particular context. Much as Kiviat shows how hiring managers turn to credit reports to infer moral worthiness, Brayne demonstrates how police officers begin to see mere presence in the database as a sign of criminality. For example, the number of times a person’s name has been queried in the database becomes a proxy for criminality:
When I asked [a detective] why he would want to know how many times someone’s name has been queried, he replied that “if you aren’t doing anything wrong,” the cops are not going to be looking you up very many times over the course of your life. He continued: “Just because you haven’t been arrested doesn’t mean you haven’t been caught.” In other words, in auditable big data systems, queries can serve as quantified proxies for suspiciousness.
Mere contact with the police (through routine street-level policing) becomes another quantified for suspiciousness, with each contact adding a point to someone’s total score used to assess whether the individual is at high risk of committing a crime. And of course the database – maintained by Palantir – isn’t limited to policing data, but includes “data from repossession and collections agencies; social media, foreclosure, and electronic toll pass data; and address and usage information from utility bills” with plans to expand security feeds from hospitals and universities, and even call data from pizza delivery chains.
Together these papers start to fill in the edges of ubercapital’s present, while at the same time pointing towards the possibilities for ubercapital’s even more dismal future as databases continue to expand, become further integrated, and creep into new decisionmaking contexts.