“Automating Inequality” — Using Technology to Screw the Poor

Automating Inequality is a book by local researcher Virginia Eubanks; I attended a talk she gave. The focus was upon three initiatives ostensibly aimed at using technology to improve delivery of social services to needy people — that in practice do the opposite.

I’ve written about how it’s expensive to be poor in America — the many ways we actually penalize poverty. I discussed the criminal justice system actually preying upon the disadvantaged, extracting money from them. While banks and credit card companies exploit poorer people’s financial precariousness to load them with fees.

“Well, they’re mostly bad people,” remarked a guy sitting beside me at the talk. Referring to the poor. No, they are not mostly bad. They are mostly unlucky people — especially in their choice of parents. It’s easy to be smug if you’ve grown up with all the advantages (like me, and probably him). But if you’re born into lousy circumstances, there are huge obstacles (starting with rotten schools) to rising out of them, even if you are smart and responsible.

The bureaucrats in Eubanks’s reporting are mostly not bad people either. Most are well intentioned in trying to serve the public (somehow or other). Especially the “line workers” in actual contact with the disadvantaged people they’re tasked with helping. But it’s others who design the “advanced” systems she discussed.

One was Indiana’s, for processing applications for public benefits. It moved caseworkers from local facilities into regional ones, putting them in front of computers rather than the human beings they previously dealt with face-to-face. No more single point of contact; applicants would now speak to a different person every time they called. (Ever been in that situation? A recipe for frustration and run-arounds.) Meantime, the whole process was moved online. Fine if you have ready computer access; half of welfare recipients don’t.

The upshot was a million applications denied over three years. Mostly for some error in the process, often not the applicant’s fault. A notice of denial would give them ten days to fix the problem. Would the notice explain the problem? Nope!

Eubanks commented that the system couldn’t have worked better at kicking people off welfare if it had been designed to do exactly that.

Next was Los Angeles County’s “Coordinated Entry” system to evaluate homeless people for their vulnerability and match them with resources. Eubanks mentioned 58,000 LA County homeless people living in “encampments.” Only about a quarter get housing through the new system. A problem is that “higher functioning” homeless people get low vulnerability scores, so they’re de-prioritized. On the other hand, the kinds of things that give you a high score are often considered crimes, so people have to incriminate themselves to get a better chance at housing. And the info going into the system also goes to the police. But meantime, incarceration actually lowers one’s score — being in jail rates as “housing.”

Seems like one giant Catch-22. It’s really a way to ration — however irrationally — available housing resources that can accommodate only a fraction of the homeless.

The third case study was the “Family Screening Tool” used by Pennsylvania’s Allegheny County; here the scoring is to identify children at risk for abuse or neglect, based on information collected by social service agencies, incorporating factors that correlate with such risk. A family’s high score makes an investigation mandatory.

What actually results is a big feedback loop. Even if that investigation shows no problem, the fact that it occurred goes into a family’s score going forward. And the scoring really fails to distinguish poor parenting from parenting-while-poor. Non-poor and, especially, white families don’t even go into the database. And the system has real consequences — it’s all geared toward taking kids away from parents, in the guise of protecting them. Poor and non-white families are at constant risk for this.

And where do those kids go? To foster care. And the reality is that children are, generally, better off with biological parents, however less than ideal that situation may be, than in foster care, which tends to be far worse. The Nanny State on stilts. Here, it’s the Nanny from Hell.

Our entire system of public benefits and social safety nets is a crazy quilt of bureaucratic complexity that costs us way more — supposedly to make sure people are entitled to what they receive — than if we just handed a check to everyone who asks. Likewise, simply giving every homeless person an apartment would cost far less than we actually spend, not only on bureaucracy, but on the costs of people being on the streets, which include police, courts, and constant emergency interventions.

The system reflects our fundamental societal schizophrenia between, on the one hand, recognizing an obligation to help the needy and, on the other, seeing them as unworthy moochers (like that guy sitting next to me did).

This is a very rich country. We could amply afford to take care of every unfortunate person in the country if we would overcome that schizophrenia and decide to do it because it’s just humanely right. We give way more welfare to the well-off. Welfare for all the needy, without all the nonsense, would cost less than the waste in the defense budget. Less than we’ve thrown away in Trump’s tax cuts for the rich.