Today’s AVMs do not suffer from the same shortcomings as yesterday’s real estate technology
When you ask some people about an Automated Valuation Model (AVM), they will bring up some of the headaches that this trailblazing prop tech caused in the early and mid-aughts. It’s not uncommon to hear the following complaints about this real estate technology:
- AVMs aren’t very accurate.
- AVMs don’t understand the trends in the property market.
- AVMs can’t tell me the overall condition of the property.
- AVMs operate using property data; they don’t know how good the view is from the porch, they can’t hear if the deafening decibels of the highway reach the backyard, and they can’t see if there’s a pool.
While all of this may have been true, innovation in real estate prop tech has come a long way, especially when it comes to AVMs. In fact, improvements in this real estate technology have only accelerated since the MBA Valuation Analytics Workgroup published a report in January 2019 showing a consistent improvement in AVM performance over the preceding decade.
So, what exactly are the technological advancements that the past decade has brought to this segment of real estate technology? Innovation has occurred primarily in several critical areas to redefine the answer to the question, “What’s my house worth?”
- Reliable determinations of property condition. Today’s AVMs employ tactics such as computer vision artificial intelligence, machine learning, digital imagery, and real estate agent comments to assess the overall condition of a property.
And these valuation models are amassing data more quickly than in the past, which comes in handy when valuing flipped homes. Today’s AVMs can now identify properties where remodels and significant upgrades are present.
- AVMs can use geographic information to accurately situate a property. Imagine two properties on the same street. One has an ocean view. One has a view of the neighbor’s backyard. In the past, an AVM probably wouldn’t have distinguished between the two. Today, this real estate technology can combine detailed map data with artificial intelligence to account for the difference in views and value properties accordingly.
- Property details are now important differentiators. Does the property have a pool? What about solar panels? Is there an outstanding view from the backyard? The availability of advanced datasets allows modern AVMs to understand when these sorts of elements are present on a property and add or subtract value accordingly.
- Incorporate future housing trends. Many advanced AVMs now account for home price forecasts. This approach can help the model not only determine current home values, adjusting for price changes from recent sales, but it allows for a better understanding of pricing trends in the near future to provide more accurate values even in fast-changing housing markets.
- AVMs score themselves and show their work. Most AVMs return a confidence score with every valuation to indicate accuracy and help users determine how much confidence users should have in each AVM output.
Some models are even taking things a step further and redefining valuations to increase user confidence in valuation accuracy. To help determine if additional research or valuations are needed, solutions like Total Home ValueX proactively provide multi-page reports showing comparable sales, market conditions, detailed subject property data, and other insights that explain how the model arrived at a particular valuation. By showing its work, this modern AVM helps users keep pace with the fast-changing property market with precision.
- Consistent testing and monitoring improve accuracy. Frequent testing and validation ensure that AVMs maintain their performance and that users receive early warning signs about any issues that would materially change performance.
In addition to testing, it is imperative that AVM vendors seeking to have lenders adopt their models must meet numerous other obligations such as adhering to the 2010 interagency guidelines, following auditing requirements, and engaging in model governance reviews.
There are many options when it comes to selecting an AVM since they all have a similar goal: to try to estimate the current market value of the subject property. However, not all models are created equal. The difference lies in the data.
Whether you’re looking to accomplish verifying an owner’s estimate of value during prequalification, determine the collateral value for certain loan types, identify properties with equity, or monitor TLVs in servicing, it pays to have the assurance that your AVM builds its valuations on top of models with decades of property data covering 99.9% of U.S. properties.
And don’t limit yourself to thinking that an AVM only has a single use. CoreLogic’s Total Home ValueX TM analyzes comparable sales alongside the other relevant statistics for an individual property, and it does this all on a single AVM that can be leveraged across the entire loan lifecycle.
©2023 CoreLogic, Inc. The CoreLogic® statements and information in this blog post may not be reproduced or used in any form without express written permission. While all the CoreLogic statements and information are believed to be accurate, CoreLogic makes no representation or warranty as to the completeness or accuracy of the statements and information and assumes no responsibility whatsoever for the information and statements or any reliance thereon.
