Veros Say Its Automated Valuation Model Isn’t Biased

Veros Real Estate Solutions, an enterprise risk management and collateral valuation services company, released new research that shows its Automated Valuation Model does not discriminate based on redlining.

Experts have recently sounded the alarm on bias in home appraisals. Studies have shown that homes occupied by Black Americans are consistently undervalued compared to similar homes with white owners.

Automated home appraisals are supposed to reduce the problem by removing human biases altogether, but concerns persist that the algorithm, fed human-produced data, can reproduce these problems. Further, they may be harder to spot because of the apparent objectivity of the technology.

The Consumer Financial Protection Bureau introduced a rule intended to safeguard homeowners against this issue earlier this month.

“While machines crunching numbers might seem capable of taking human bias out of the equation, they can’t,” the CFPB said. “Based on the data they are fed and the algorithms they use, automated models can embed the very human bias they are meant to correct.”

The proposed rule would create basic safeguards to minimize the potential harm of automated valuation, though the CFPB notes it can only mitigate the problem, not solve it. Companies employing automated valuation models would need to have policies in place to avoid conflicts of interest, conduct random sample testing and reviews, and comply with nondiscrimination laws.

Veros contends that its AVM is already protected from this type of bias, however.

The company analyzed its AVM’s predictions for single-family homes in redlined and non-redlined neighborhoods near Los Angeles. Redlining is the historical practice of racist lenders designating non-white neighborhoods as risky. Its legacy reverberates in housing today.

After controlling for the homes’ physical attributes, Veros says its AVM produced similar estimates regardless of the neighborhood.

Any differences in values had more to do with the type of home than its location, according to Reena Agrawal, Veros’ Research Economist.

“The research also found that a modern-day outcome of historical redlining is that homes in neighborhoods that were redlined generally have smaller lot sizes and living areas compared to homes in non-redlined areas, leading to lower median values than larger homes,” she said.

While this is an unfortunate byproduct of redlining, it is unrelated to bias in its AVM, Veros says, making its valuation model effectively bias-free.

Veros is a mortgage technology company focused on enterprise risk management and collateral valuation services, including automated valuation, fraud and risk detection, portfolio analysis, and forecasting. It also supports the FHA’s Electronic Appraisal Delivery portal.