Christine E. Webber

Mary Louis had a housing voucher and a rental history she was willing to document. When an apartment manager rejected her application in May 2021, she offered references from two landlords and from her employers to establish at least sixteen years of paying rent early or on time. In July, the manager replied that it did not accept appeals and could not override the screening outcome.

Louis had applied at Granada Highlands in Malden, Massachusetts, where her voucher would have covered about 69 percent of the rent. SafeRent Solutions supplied the score used to reject her. The plaintiffs alleged that it counted debts unrelated to tenancy while leaving out the voucher’s contribution to payment.

Christine E. Webber represented Louis, Monica Douglas and the Community Action Agency of Somerville with lawyers from Cohen Milstein, Greater Boston Legal Services and the National Consumer Law Center. They filed the class action in May 2022. Webber described a practical obstacle that reached beyond the accuracy of any one credit entry: without knowing the score’s inputs or weights, she said, “housing providers cannot exercise any independent judgment as to the merits of housing applicants.”

Credit History, Rental History

Webber had spent decades working with the records behind discrimination claims. After earning her law degree at the University of Michigan in 1991, she received a Women’s Law and Public Policy fellowship that funded the first of four years at the Washington Lawyers’ Committee for Civil Rights and Urban Affairs. She participated in the 1995 trial of the Neal sexual-harassment class action involving the District of Columbia’s Department of Corrections.

As a partner and co-chair of Cohen Milstein’s Civil Rights & Employment practice, she works with economic and statistical experts on class claims. In Louis, the plaintiffs connected differences in conventional credit scores among Black, Hispanic and white consumers to inequalities in income, wealth and access to credit. They alleged that using those scores to screen tenants imposed an unjustified racially disparate burden.

The voucher presented a separate question about what the model was measuring. Credit history was being used to estimate whether an applicant could meet future rent obligations. For Louis, however, a public program had committed to paying most of that obligation. A score that heavily weighted credit information while excluding that commitment could understate her ability to pay.

Webber and her co-counsel sought to link the model’s inputs, the racial distribution of its scores and the resulting housing denials. Their Fair Housing Act claim concerned discriminatory effects, including those produced without an explicit race variable or an instruction to discriminate. Describing the case at settlement, Webber told the Associated Press that the choice and weighting of data could have “the same effect as if you told it to discriminate intentionally.”

Who Controlled the Calculation

SafeRent supplied an accept, conditional or decline decision. A landlord selected a minimum threshold, but the plaintiffs alleged that it could neither inspect nor change the algorithm or its internal weighting. The vendor converted the underlying information into the score on which the housing provider relied.

SafeRent argued that the Fair Housing Act should not reach a company supplying information to landlords who made the ultimate housing decisions. On January 9, 2023, the Department of Justice and the Department of Housing and Urban Development filed a statement of interest addressing the Act’s application to tenant-screening companies and the connection between SafeRent’s score and access to housing.

Judge Angel Kelley allowed the Fair Housing Act claims against the vendor to proceed on July 26, 2023. The complaint plausibly alleged that SafeRent controlled the calculation and that housing providers relied on its output to accept or reject applicants. A landlord’s formal authority did not automatically break the causal connection between the vendor’s scoring practice and the alleged discrimination.

The ruling gave Webber and the plaintiffs’ team a path forward against the company that produced the score. The court had found their allegations sufficient at the pleading stage; the parties would resolve the claims through settlement before trial.

A Condition on the Next Score

Kelley granted final approval to the $2.275 million settlement on November 20, 2024. It provided monetary relief to defined classes and required SafeRent to remove the challenged score from specified reports used in covered voucher screenings.

SafeRent would stop supplying the score or a score-based accept-or-decline recommendation through its Affordable Model. For its Market and No-Credit models, housing providers would have to certify that the applicant was not receiving a publicly funded housing voucher. Without that certification, SafeRent would provide underlying information subject to the agreement’s score restrictions.

Webber and her co-counsel secured a condition on any replacement tenant score intended for the covered voucher screenings: it could be used only after the National Fair Housing Alliance or another qualifying organization agreed to under the settlement had found it valid for voucher holders. Removing the old score would not permit an untested equivalent to take its place.

The court retained jurisdiction to enforce the agreement for five years after SafeRent confirmed that it had implemented the required practice changes. That term gave the parties a forum for disputes about removal, replacement and continuing compliance after implementation.

The payment arrangements accounted for the recipients’ circumstances too. To help avoid disrupting eligibility for public benefits, eligible class members could choose to receive their money in two equal installments over two years.

The administrator mailed check payments on June 30, 2025 and sent digital-payment instructions on July 1. It reported issuing the second installments by February 17, 2026. The distributions and the five-year enforcement period served separate purposes: compensation for eligible applicants who had been denied housing, and continuing court supervision of the agreed screening changes.