TheMReport — News and strategies for the evolving mortgage marketplace.
Issue link: http://digital.themreport.com/i/152295
Feature T echnology has radically enhanced our working and living conditions to the point where it is difficult to imagine life without it. That said, while it delivers speed and efficiency to many processes in the mortgage industry, there are times when technology fails to deliver. Automated valuation models (AVMs) are arguably one such example. Data-based solutions that provide an estimate of value for properties by using public records and various proprietary algorithms, AVMs take a back seat in terms of accuracy to the more traditional approach of using appraisers. Estimates, Not Values A VMs are to be used as a supplement, not as a substitute, for an opinion of value. There is no comparison to the experienced appraiser or agent. As neutral third-party valuators, their specialized knowledge of the market is vital to ascertain an accurate property value and to sell that property in a timely manner— something technology can't do. AVMs reap flawed results, begging the question of whether or not they should continue to serve the valuation process. But there is a time and a place for appropriately using AVMs, albeit that time and place is rare. AVMs are viable only if their strengths and weaknesses are understood and they are used correctly—as a backup, rather than as a replacement, for an appraiser or agent. An AVM can be utilized as either a starting point or as a second opinion. With rapidly rising home values, there may be temptation to use AVMs. But the fact is AVMs are only effective in a market with stable prices. If an agent is listing a property for sale and has not approached an appraiser yet, he or she can use an AVM to get a general ballpark idea for how much it might be worth. Or, if a seller thinks the property's value has changed drastically since a prior valuation, that would be another appropriate instance to utilize an AVM. Even if the seller wanted to check to see if he or she was being taxed appropriately, an AVM would be a good starting place to develop a better perspective. Another appropriate usage for an AVM would be to validate the value of an appraisal or a BPO. Notice that AVMs databases relying on other unknown sources for the information. And this regurgitated data is typically outdated information; it can be up to three years old. Just think of how much can change in the course of three years that can markedly affect that value. There is also a strong risk that AVMs will pick the wrong sales comps to review. The system automatically takes into account criteria that could be rejected when making certain decisions. If the criteria on the assignment ask for information about the most recent property closing near the home in question, the AVM might use data from an REO property to determine a value for a new or a fair market property. That purchaser would be robbed of potential value, and the AVM will not have the ability to catch that. With an AVM, you are not getting the vetted data that you would get from working with an appraiser or agent, carefully analyzing and interpreting the presented information. You are simply getting collected data, drawn from unknown sources for the information. should never be used as a sole means to generate value, but must be paired with a licensed agent or appraiser if used at all. Data Drawbacks A VMs' weaknesses far outweigh their strengths. A glaring problem is their lack of data integrity. With an AVM, you are not getting the vetted data that you would get from working with an appraiser or agent, carefully analyzing and interpreting the presented information. You are simply getting collected data, drawn from AVMs lack their own judgment. An appraiser or agent can qualify or disqualify a sale, saying these properties are more similar or these are clearly less similar, and align the value after filtering through the data. An AVM cannot do that. It simply looks for checkpoints or qualifications based on certain criteria. It will not know, for instance, to take into account that this property is completely remodeled like the subject. These are similarities that you would only notice by going inside the property. An AVM is based off the exterior data only because computers are incapable of inspecting homes. Therefore, AVMs do not include pictures of the property. The property's interior can tell a completely different story, affecting the value in a negative or positive way. And this may lead to an AVM's comparable properties differing greatly from the subject in terms of updates, quality of construction, and marketability of the home. The AVM product also disregards market influences, which can have a sizable impact on market value. A valuation that does not consider a property's market influences is insufficient and should not be accepted. Geographical location, for example, is going to have a major impact on a property's value. There is a reason for the adage "location, location, location," and there is also a reason the word is repeated several times—for emphasis. It is a basic rule in real estate, and it makes a difference when weighing market value. Only an appraiser with deep familiarity with the surrounding area can provide this level of insight. A Good Start, but That's It A lthough working with an appraiser or agent clearly trumps working with an AVM, AVMs can serve limited purposes. For a quick way to get a broad idea on a price for a property, they serve as a good starting point but require further interpretation. And they are excellent tools to verify opinions, but users must scrutinize their output carefully. Despite advances in AVM technology such as expert systems, neural networks, and artificial intelligence, relying on AVMs as the standalone means for generating a value is a dangerous practice that could lead to expensive mistakes. AVMs can be more reliable in markets that consistently have minimal fluctuation in values. The buyer or user of an AVM must be experienced in using the product, otherwise too many errors can occur, leading to mistakes that could cost the investor thousands of dollars. The M Report | 21