The City of Louisville, Kentucky, will soon begin using a newly developed algorithmic tool that aims to stop city funding from going to housing projects that would displace local residents. The tool was mandated by a tenant-led 2023 law and, after a year of development by researchers at Boston University, was approved for use in November 2024. Now, it's ready to be deployed on upcoming development proposals, explains Next City.

The open-source tool analyzes whether any given project meets the neighborhood's housing needs and income levels, ensuring that rents match local residents' income. If the development does not meet these standards, then the city cannot subsidize it. Researchers say that the tool's criteria are as robust as tenant advocates had wanted to prevent residents from being priced out of their neighborhoods.

"We got a strong set of recommendations passed," says Kenton Card, a researcher at the Center for Urban and Regional Affairs at the University of Minnesota, who was hired by the city to work on the tool and the guidelines for its use.

"This is going to make a small lever of government more equitable. It's not a revolution of the housing system, but it's a legislative success."

Meeting Louisville's housing needs

"This has been four years in the making," says Jessica Bellamy, a co-founder of the Louisville Tenants Union, who helped push for the law mandating the new tool. She believes that it will limit the type of government-funded projects that she blames for destroying communities in Smoketown, the neighborhood where she was raised.

Bellamy grew up in a multigenerational house in Smoketown with her mother, brother, grandmother and three of her aunts. Her family was displaced when she was a child, but she currently owns the home she grew up in.

In 2012, the HOPE VI program demolished public housing that housed many other people in the area, with the intent of building mixed-income developments. Speculators began buying up property when investment came in, and houses started being flipped.

"So many people were pushed out … there's a number of people that I knew and grew up with who are cousins and family that began to live a life of housing precarity," Bellamy says. She says she tried to move back to the community and organized to fight rent increases with the Smoketown Neighborhood Association.

"That gentrification was really being funded by the city," Bellamy says of this period. Tenants began brainstorming in 2020 during the early months of the pandemic on a bill to address gentrification.

Initially introduced as the "Historically Black Neighborhoods Ordinance," the bill was intended to address displacement from Black neighborhoods. But legislators and housing advocates changed the ordinance's aim so that it could apply to any neighborhood with displacement concerns—and withstand potential legal challenges.

The tenants teamed up with Louisville council member Jecorey Arthur, a former activist and music teacher who was elected to the city council in 2020. Arthur says he was elected as a result of the George Floyd uprisings and the protests that emerged after the killing of Breonna Taylor in Louisville. (Taylor's killing has been tied to redevelopment and gentrification in the area where she lived.)

Arthur says that during his first budget cycle as a council member, he looked at the city's housing needs assessment—an analysis of Louisville's housing stock—and made amendments to ensure funding prioritized the housing that was most in demand: housing for people making 30% of the area median income.

Those recommendations were approved by the council and "forced developers to build that housing that we needed," Arthur says.

While the new tool is designed to restrict city subsidies for certain projects deemed harmful, Arthur says that he hopes it makes it easier to fund housing that is affordable for the most low-income people.

"When developers apply for city subsidies or they want financial incentives, this displacement assessment will show who's actually building housing for the poorest of the poor and who wants to build something luxury," he says.

While some might protest that middle-income people also need housing to relieve displacement pressure, researchers stressed that the tool doesn't prevent any private sector development. It's only applied when housing is subsidized by the Louisville city government.

The tool relies on Louisville's own housing needs assessment, released in March 2024, which found that there were no homes affordable to more than half of the city's lowest-income residents, those making 30% of the area median income. But there was an oversupply of homes for someone making 80%-150% of the area median income.

The researchers say the point of the tool is to prioritize housing subsidies that match the city's needs.

"It's not anti-development, it's pro-development—if that development is going to build truly affordable housing," says Loretta Lees, an urban geographer and director of the Initiative on Cities at Boston University, who also worked on developing the tool.

While there has been growing conversation across the country about zoning and bureaucratic red tape as factors holding back housing production, Arthur says that city-subsidized developers have largely operated without accountability when it comes to affordability.

"Developers … have kind of played the political system, whether it's through access, whether it's through donations, whether it's through people not being civically engaged, or people being really bamboozled by developers," he says. "They've been able to get projects approved without much vetting."

How the displacement tool works

Next City/Shelterforce experimented with the tool, which has a range of inputs for proposed developments including the project's address, number of units, and median incomes for those units, ranging from 30% to 80%.

We tested the tool with a hypothetical project in Louisville's Downtown area, where a quarter of the units are affordable to extremely low-income people and half are market rate. The tool said the area already had a high displacement risk: Half of residents who live there can only afford a rent of around $500 a month. Because of this, the hypothetical building where half the units are market rate would not qualify for city subsidy, according to the tool.

The tool also provides charts showing the changes in asking rent, median rent, racial demographics and number of residents with college educations over the previous years.

The tool takes on a wide range of data, including publicly available census data, when determining the underlying displacement risk of a project, researchers say. That includes changes in the share of college-educated adults, changes in household incomes, and changes in rent and home values.

"All those things that we know from decades of research are associated with higher risk of displacement," says Andre Comandon, a researcher at the University of Southern California who helped work on the tool. Comandon says that it looks at the difference between the proposed developments' number of units and rent amount and the demographics of the area.

Comandon acknowledges that in areas with household growth but without a lot of new development, rent costs can be higher due to the lack of housing. But he says that new subsidized development should be affordable to people who have lower incomes so they can find housing in their own neighborhood.

To some tenant organizers, there's a hope that the tool will lead to more subsidies and even investments in social housing. "We're hoping the ordinance exposes contradictions," says Josh Poe, a member of the Louisville Tenants Union. He says that elected officials will have to grapple with developers' profit motives and acknowledge that affordable housing requires government funding.

Algorithms and bias

Proponents of the anti-displacement tool say it will "take the politics out" of the process of determining how to allocate public funds, and that the tool will be objective in its determinations.

"We're looking at facts, not feelings," Arthur says.

Yet the use of algorithmic tools in government has been criticized in recent years, particularly tools that help judges determine bail or those used by police officers to determine deployment. Because algorithms rely on broad sets of data, critics point out that the data can be faulty and that using datasets as proxies for certain predictive outcomes can reproduce inequality.

Arthur says he's not concerned about this, because he believes the researchers put in enough work and because the tool can be tweaked if there are errors.

There is also the main difference: Many algorithmic tools that have been criticized determine risk associated with people. The anti-displacement tool determines risk created by developers and the potential harm to people, specifically people who earn low incomes.

Lees, one of the researchers who worked on the tool, says it is more than just an algorithm since it works together with the rubric that the researchers created to determine whether projects will be funded.

"Any tool and any kind of data is limited, is biased, but of the tools that are out there at the moment, this is the best that could be done," Lees says.

The new tool hasn't been utilized yet, but because it's now a mandatory part of the development process when city subsidies are in play, housing advocates believe it will be put to use soon. For her part, Bellamy hopes it will end decades of displacement she has witnessed in Louisville.

"From the perspective of someone from this community, watching this process … you are building homes that people in the community can't afford, and then evicting en masse," Bellamy says. She's optimistic that the open-source tool will not just carry weight in Louisville, but that other cities will try to adopt it.

"We want everyone to be able to use this," Bellamy says.

This story was co-published in collaboration with Shelterforce, the only independent, non-academic publication covering the worlds of affordable housing, community development and housing justice.