Nfr Esters. Photo by Jessica McKenzie

Nfr Esters knew she couldn’t save the family house through conventional means. She had inherited the duplex on the East Side of Detroit from her great-grandfather in 2010, but a pipe burst before she moved back to the city, flooding the basement and racking up a $4,000 water bill that the city added to her property taxes. Esters would surely lose the house to tax foreclosure—but there was still hope.

Every year, Wayne County auctions off tax-foreclosed properties for a starting bid of $500. Since the auction went online in 2010, anyone can participate remotely. Esters planned to bid on her own home, but she worried a speculator would find it and decide it was a good investment property after all the time and money she spent rehabilitating it. One day, Esters’s husband came inside while their girls played on the front lawn and said a woman was taking photos of their home. Esters worried she had been sent by the county treasurer to “build a database” to market her house.

Although the woman said she was just there to check occupancy, Esters felt vulnerable. She made the woman take a photo of the family in front of the house so people would know that they were not just bidding on a house—they were bidding on a family to lose their home.

Although Esters didn’t know it at the time, the database she imagined was already available. Since 2011, a for-profit civic tech company called Loveland Technologies has published public property data in map form, starting with the lots and houses available in the county’s tax foreclosure auction. The company’s mission, in part, is to help Detroiters “battle a plague” of foreclosures by “arming” them with information. Co-founder and CEO Jerry Paffendorf says his company’s guiding principle is “making public information freely available to the public.”

That reflects a central philosophy in the open-data movement—that more information is inherently in the public’s interest. But in Detroit’s troubled housing market, it’s also a case study in how an open-data initiative can do as much to reinforce power imbalances and social inequity as it can to correct them.

Mural commissioned by the Tricycle Collective. Photo by Jessica McKenzie

The “we” in “Why Don’t We Own This?”

In 2011, Loveland Technologies launched “Why Don’t We Own This?”, a map of all the properties available in the Wayne County tax foreclosure auction. The goal was to make the auction more accessible and to encourage investment in Detroit. But some worried—in part because of its name—that the website was selling the city to outsiders.

“[It was] one of the most sophisticated property speculation tools—free property speculation tools—that had been put on the web at that point,” says Joshua Akers, a professor of geography and urban studies at the University of Michigan at Dearborn.

Through Loveland, it’s easy to identify which houses are likely occupied, an indication that they would be easy to flip or to rent out. Early users swapped information on property conditions and occupancy status—and sometimes tips on evicting residents. The company says the information it provides isn’t especially valuable to speculators. “They already have it,” Loveland co-founder and COO Mary Carter says. “They’ve always already had it.”

This is likely true of companies that scoop up properties in the tax foreclosure auction. But Loveland’s high-quality information expanded that opportunity to “every wannabe investor,” says Ted Phillips, executive director of the United Community Housing Coalition (UCHC), a Detroit nonprofit that works on foreclosure prevention and relief.

Meanwhile, Michele Oberholtzer, UCHC’s lead on foreclosure work and a former surveyor at Loveland, says Loveland is too challenging for most of her clients to learn, though she sometimes uses it herself to help them. After Esters successfully bought back her house at auction in 2014 with help from UCHC, she joined the board of the Tricycle Collective, a group Oberholtzer founded to raise money for families with children trying to buy their home in the auction. They use Loveland when they canvass neighborhoods to find families at risk of foreclosure and let them know how they might keep their homes.

That highlights a core challenge with open data, identified by Utah Valley University’s Jeffrey Johnson, who coined the phrase “information justice.” Johnson distinguishes between “citizen” and “enterprise” users, arguing that enterprise users’ resources and proficiency with information gives them an advantage over average citizens, even if the data is theoretically available to everyone. In Detroit, that gap is wide: Roughly 40 percent of residents live without access to broadband internet.

Nfr Esters stands outside her home, which she successfully bought back at auction. Photo by Jessica McKenzie

“Garbage in, garbage out”

Even in a world with perfect access, the information itself can be tainted. Maps can show factual information about homes facing foreclosure, but they “decontextualize” the situation, says Shane Bernardo, a founding member of People’s Platform, a grassroots civic coalition in Detroit. He says the way Loveland presents property information obscures Detroit’s history of redlining and disinvestment in predominantly black communities, as well as more recent concerns about how it conducts tax foreclosures.

These are legitimate concerns of which Loveland is well aware. In a 2015 bid for interim county treasurer, Paffendorf cited evidence that Detroit and Wayne County failed to fairly assess property taxes when home values plunged after the 2008 financial crisis. Loveland also provided a copy of its data to the American Civil Liberties Union of Michigan and the NAACP Legal Defense Fund. Last year, the groups sued Wayne County and the treasurer, alleging they violated the Fair Housing Act by foreclosing on homes predominantly owned by African Americans for failure to pay improperly assessed taxes. (In October, the court found that the Michigan Tax Tribunal had exclusive jurisdiction over the case, but they are appealing the decision.)

With such major questions about the data itself, presenting it without proper context risks further codifying—and justifying—social inequalities. “‘Garbage in, garbage out’ is a central concept in data ethics,” Johnson writes, adding that in many cases that means “injustice in, injustice out.”

Loveland’s map of properties at risk of tax foreclosure in 2017.

Garlin Gilchrist, Detroit’s deputy technology director of civic community engagement, wrote Detroit’s open data policy and often grapples with what city data should be made available and how. “It is very important to try to be conscientious about what problems you’re trying to point transparency toward,” he says.

Gilchrist says he thinks about this when the city shares crime data, worried that it might portray crime as a neighborhood’s defining characteristic. When it is necessary to publish such information, Gilchrist says he asks himself, “What are we doing to put that into some broader context and perspective?”

“For open data, in order to really engender trust, it needs to tell the most complete story that we can tell,” he says.

To the extent that Loveland shone a light on the tax foreclosure crisis in Detroit, helping draw national attention to the problem, Gilchrist says its work has been valuable. However, he says substantive change can only come from within government—the city doesn’t need “interesting technology innovation” so much as “reforms with eyes towards fairness and reasonableness.”

Recently, some small steps were taken: Last year County Treasurer Eric Sabree hired Loveland to survey occupants of foreclosed homes and distribute information about payment plans. (Sabree credited the outreach for reducing tax foreclosures by 36 percent—although now that more than a quarter of Detroit properties are owned by public entities like the Detroit Land Bank Authority, there are fewer properties at risk of tax foreclosure in the first place.)

But the city and county have not made significant policy changes. In fact, Phillips says one of the few laws passed in recent years to crack down on speculation works more like a sieve, preventing homeowners from buying their house back after it has been foreclosed upon while speculators simply create a new LLC and proceed with business as usual. (One option that could help, Oberholtzer says, is to give residents the first option to buy.)

Other data scientists in Detroit have sought to present housing data with information justice in mind. Last year, Akers launched a project called Property Praxis, in which he and his colleague Alex Hill mapped the 20 percent of land parcels in Detroit owned by speculators. Wherever possible, they name the actual owner of the property—not the shell corporation or LLC that obscures the owner’s identity. So, while Loveland lists the owner of 8030 Townsend as CE Detroit LLC, Property Praxis lists it as Manuel Matty Moroun, and shows that he owns 599 Detroit properties through various companies.

Akers said he and Hill wanted to create “a map that no developer can use to build shit.”

As property data geeks—and neighbors—Akers and Paffendorf have had plenty of opportunities to debate whether open data is the solution to problems like foreclosure and blight. “We’ve got 100 years of information. Like, we understand how these markets work and how they function,” Akers says. “What we have is a lack of political will to actually deal with these issues.”

The reality in Detroit undermines one of the most deeply held beliefs of open-access evangelists: that more transparency is better than less. At its best, open data can hold the powerful accountable. But it can also expose the weak and the powerful alike, with mixed results.

Jessica McKenzie is the managing editor of Civicist.

Co-published with CityLab.