The reliability of automated valuation models (AVMs) has been put into the spotlight in recent weeks, after groundbreaking changes in the way in which these models are now being tested. The implications for lenders for the home, which are highly dependent on AVMs to determine real estate values, can be reached, especially because the industry is struggling with new regulatory questions about risk management and model accuracy.
After having worked for decades in the valuation industry, including the introduction of the first AVM on the market in the nineties and later fighting states that wanted to forbid AVMs, I can tell you that this shift in testing is seismic. The release of independent test results that demonstrate poor performance of some of the best AVM models does not only mean that the existing approach of AVM tests can be outdated and defective, but that the lenders now have to clamber to respond as a new Federal AVM -regulatory framework is in place on 1 October.
The core of this disruption is the decision of AVMETRICS, the only independent AVM test company of the nation, to stop allowing AVM model companies to use catalog prices in their AVM test method. This decision challenges an industrial standard that has long been based on the catalog price of properties to help determine AVM valuations. But testing through AVMetrics and others has shown that many AVM models not only use catalog price, they “anchor” them.
This is especially worrying for lenders who, as an industrial practice, do not borrow to borrowers who have mentioned their property for sale. Given this reality, many in industry think, including myself, that it is time to remove catalog prices from AVM tests.
For some AVM providers, especially those who have been familiar with catalog prices for a long time, this decision undoubtedly feels an important blow to their models. These AVM providers claim that more data is always better and data price data offer a clearer picture. Nevertheless, those who support the change believe the use of catalog prices in AVM tests nothing less than ‘cheating’, and a quick solution that does not represent the true dynamics of the Real-World of their own power loans.
It is not just about the technical details of how AVMs are tested. The point to ensure that these models offer accurate, reliable information that reduces the risk of lenders and protects borrowers. The commitment here is high because federal agencies are preparing to formalize new AVM control guidelines that will come into force on 1 October, guidelines that include the correct quality control monitoring of AVMs.
Lee Kennedy, CEO of AVMETRICS, articulates the core of the matter when he says: “It is not important how AVMs perform in test environments, but how accurate they are in real-World production situations, such as in-house equity or refinancing, where no list or selling price is available.” This insight underlines the actual care: AVMs that are dependent on catalog prices can inflate or produce inaccurate results that do not reflect market reality, especially when applied to stock loans, where most properties are not active for sale.
For years, AVMs have been announced as an aid to reduce the risk in the loans by offering rapid, data -driven estimates of the real estate value. But if the The 2024 report of American Enterprise Institute On AVM performance, the use of catalog prices emphasizes what the study “Springiness” calls – a term that is used to describe how AVM’s “Spring” in the direction of the catalog price if available. This phenomenon distorts the actual value of a characteristic and reduces the usefulness of the AVM for money lenders aimed at Real-World scenarios, not theoretical models.
In the center of the debate is whether AVMs, as they are currently being tested, can really record the actual value of a characteristic in scenarios such as Home Equity Lending. I have always supported the use of AVMs as an effective tool for the valuation of real estate, but it is crucial that we now refine the methods behind these models. AVMs must offer accurate representations of market conditions, not only depending on the catalog prices, which can disrupt their reliability, especially when the markets begin to change and the list prices become a lagging indicator. With billions of dollars on its own power loans, the industry must ensure that the data we use are accurate and representative of real circumstances.
Some AVM providers continue to defend the use of catalog prizes in their models. But this point of view is not universally accepted. There is a growing consensus among lenders and valuation experts that AVMs who depend on the catalog prices are inherent and risky. As a appreciation director of a large regional bank recently told: “It is the best way to use catalog price, especially not when testing equity, the best choice. We will look at this very hard.”
The response of the industry is clear: more and more lenders are evaluating their use of AVMs again and are starting to adopt the new test methodology. Assessment management companies such as Accurate Group say that they have already shifted to the new approach of AVMETRICS, which indicates a potential turning point for the industry.
Especially stock credit for equity is to benefit from this shift. Because banks such as Bank of America, Wells Fargo and Chase, as well as best regional players, are dependent on AVMs to endorse loans of equity, it is crucial that the models they use accurately reflect the market conditions, not artificially inflated catalog prices. Not adjusting to these changes can expose money lenders to greater financial risks and cause problems for consumers who depend on fair ratings.
The implications of this test shift only extend further than AVMs alone. They also discuss the broader question of how we as an industry use data and technology to make critical credit decisions. Do we trust shortcuts that can undermine the integrity of our processes? Or do we strive to develop solutions that are accurate, reliable and are in line with the reality of the market?
The changes to AVM tests are not only a technical issue, they represent a turning point for the entire ecosystem of the Ecosystem for Equity. By ensuring that AVMs are based on practical data and are not dependent on the catalog prices, we can create a more robust, reliable system that better serves donors, borrowers and industry in general.
A promising future for AVMs – and perhaps for equity that lends itself – can be rejected from this conversation. Let’s make sure that we are all part of it.
Mark Sennott is the CEO of Sennott Consulting.
This column does not necessarily reflect the opinion of the editorial department of Housingwire and the owners.
Related
#AVM #Testing #Fire #Methodology #challenges #standards #industry #increases #risk #lenders