The reliability of Automated Valuation Models (AVMs) has been thrust into the spotlight in recent weeks, following groundbreaking changes in the way these models are now being tested. The implications for home equity lenders, who rely heavily on AVMs to determine property values, could be far-reaching, especially as the industry grapples with new regulatory questions about risk management and model accuracy.

Having spent decades working in the valuation industry, including introducing the first AVM to the marketplace in the 1990s and later fighting states who wanted to ban AVMs, I can tell you this shift in testing is seismic. The release of independent testing results showing poor performance from some of the top AVM models means not only that the existing approach to AVM testing may be outdated and flawed, but that home equity lenders now must scramble to respond as a new federal AVM regulatory framework goes into place on October 1st.

At the heart of this disruption is the decision by AVMetrics, the nation’s only independent AVM testing firm, to stop allowing AVM model companies to use list prices in their AVM testing methodology. This decision challenges an industry norm that has long relied on the list price of properties to help determine AVM valuations. But testing by AVMetrics and others has shown that many AVM models don’t just use list price, they “anchor” to it.

This is especially concerning for home equity lenders who, as an industry practice, don’t lend to borrowers who have listed their properties for sale. Given this reality, many in the industry, including myself, think it’s time to remove list prices from AVM testing. 

For some AVM providers, particularly those that have long relied on list prices, no doubt this decision feels like a significant blow to their models. These AVM providers argue that more data is always better, and list price data provides a clearer picture. Yet, those who support the change believe using list prices in AVM testing is nothing short of “cheating,” and a quick fix that fails to represent the true, real-world dynamics of home equity lending.

It’s not just about the technicalities of how AVMs are tested. It’s about ensuring these models provide accurate, reliable information that reduces risk for lenders and protects borrowers. The stakes here are high as Federal agencies get ready to formalize new AVM regulatory guidelines that go into effect on October 1st, guidelines that include proper quality control monitoring of AVMs.

Lee Kennedy, CEO of AVMetrics, articulates the crux of the matter when he says, “It’s not important how AVMs perform in test environments, but how accurate they are in real-world production situations, such as in home equity lending or refinance, where no list or sale price is available.” This insight underscores the real concern: AVMs that depend on list prices may be delivering inflated or inaccurate results that don’t reflect market realities, especially when applied to home equity lending, where most properties are not actively for sale.

For years, AVMs have been heralded as a tool for reducing risk in lending by providing quick, data-driven estimates of property value. But as the American Enterprise Institute’s 2024 report on AVM performance highlighted, the use of list prices can lead to what the study calls “springiness”—a term used to describe how AVMs “spring” toward the list price when available. This phenomenon distorts the true value of a property and diminishes the AVM’s usefulness for lenders who are focused on real-world scenarios, not theoretical models.

At the center of the debate is whether AVMs, as they are currently tested, can truly capture the real value of a property in scenarios like home equity lending. I have always supported the use of AVMs as an effective tool in property valuation, but it’s critical that we now refine the methodologies behind these models. AVMs should provide accurate representations of market conditions, not just rely on list prices, which can distort their reliability, especially if markets start changing and list prices become a lagging indicator. With billions of dollars in home equity loans on the line, the industry must ensure that the data we use is accurate and representative of real-world conditions.

Some AVM providers continue to defend the use of list prices in their models. But this viewpoint is not universally accepted. There’s a growing consensus among lenders and valuation experts that AVMs that rely on list prices are inherently flawed and risky. As one valuation executive at a large regional bank recently told me, “Not using list price, especially in home equity AVM testing, is the way to go. We are going to be looking very hard at this.”

The response from the industry is clear: More and more lenders are reevaluating their use of AVMs and beginning to adopt the new testing methodology. Appraisal management firms like Accurate Group say they have already shifted to AVMetrics’ new approach, signaling a potential tipping point for the industry.

Home equity lending in particular stands to benefit from this shift. As banks like Bank of America, Wells Fargo, and Chase, as well as top regional players, rely on AVMs to underwrite home equity loans, it’s critical that the models they use accurately reflect market conditions, not artificially inflated list prices. Failure to adapt to these changes could expose lenders to greater financial risks and create problems for consumers who depend on fair valuations.

The implications of this testing shift extend beyond AVMs alone. They also touch on the broader question of how we, as an industry, use data and technology to make critical lending decisions. Are we relying on shortcuts that could undermine the integrity of our processes? Or are we striving to develop solutions that are accurate, trustworthy, and aligned with the realities of the market?

The changes to AVM testing are not just a technical matter, they represent a turning point for the entire home equity lending ecosystem. By ensuring that AVMs are based on real-world data and not reliant on list prices, we can create a more robust, reliable system that better serves lenders, borrowers, and the industry at large.

A promising future for AVMs—and perhaps for home equity lending itself—depends on this conversation. Let’s ensure we’re all part of it.

Mark Sennott is the CEO of Sennott Consulting.

This column does not necessarily reflect the opinion of HousingWire’s editorial department and its owners.

More: