This article was originally published by HousingWire, FinLedger’s sister-publication covering the U.S. residential housing and mortgage market.
From thousands of feet above your home, a high-tech camera snaps pictures of your roof, the mess of debris in your yard, and that pool you hate cleaning. Pictures like these could cost – or save – you thousands of dollars when deals are about to close. It could also have big implications for appraisers, who are already nervous about the continued rise of automated valuation models.
With ValPro+, Cape Analytics and Weiss Analytics have created what they say is the first automated home valuation engine that uses geospatial imagery and artificial intelligence to integrate current home conditions – like a damaged roof or a pool – into valuations.
Cape Analytics’ AI system instantly extracts property condition from images, which is then run through Weiss’ valuation model, Raj Dosaj, Cape Analytics’ head of real estate, explained. The systems have condition data sets for 110 million buildings across the United States. Previously, an appraiser would need to make an in-person visit to note the condition of the property, Dosaj said.
The rise of AVMs over the years has waxed and waned based on the perceived risk of the product, but it’s caught fire over the last year with the COVID-19 pandemic and new initiatives by the Federal Housing Finance Agency. Condition data is the latest frontier, according to Cape and Weiss.
“One of the key missing ingredients which this now answers is, how do you know what the condition of the house is?” said Allan Weiss, head of Weiss Analytics and co-creator of the Case-Schiller Index. “Because most of the data is very out of date that you can get in computerized form – you can get assessor data, which could be a year more out of date, and does not tend to focus on condition of the house.”
Cape Analytics and Weiss Analytics are primarily targeting real estate investors, real estate brokers and mortgage originators with the ValPro+ tool to help them identify undervalued purchase opportunities. The two companies claim their ValPro+ engine has shown a 7.7% improvement in PPE-10 predictions of on-market valuations. The model can also flag homes in distressed condition and will accurately predict they will sell at a 10% off-market discount, the companies said.
But the companies are especially interested in hooking loan traders, refinance and HELOC loan originators and iBuyers, who often want to bid on a home that isn’t on the market and doesn’t have an asking price. Overall, off-market houses sell for about 2.5% below their on-market peers, but Cape and Weiss say investors using their product achieved a median discount of 10% compared to their on-market peers.
Weiss wasn’t shy in saying the product represents a direct threat to appraisers and broker price opinions (BPOs), whom he says are both slow and expensive.
“There’s been a decades-long process by which the traditional appraisal has been, in some cases, augmented, in many cases, replaced by technology,” said Weiss. “And that’s been something that’s been going on since at least around the mid 1990s. Because of the cost of appraisals, and because of the delay that appraisals create and various transactions, they’re just not practical for all the ways that investors, lenders, and homeowners require in order to go about transacting the way they want to.”
Cape Analytics and Weiss Analytics said they spent about a year integrating the image data into the the home valuation models. Their model allows investors to get lists of all houses discounted below market value due to condition factors. The system also automatically scans real-time MLS listings and screens houses so bids are priced appropriately to maintain credibility with local agents.
“We now know if we have a house in South Florida, that has a pool and would otherwise be worth $600,000 versus a house that doesn’t have a pool, what that means in terms of the value of the house,” Weiss told HousingWire. “We didn’t know things like that before. We didn’t know what it meant if there was a lot of yard debris, or there was a lot of vegetation overgrowth or what it meant if the condition of the roof was very good versus very bad in a given market.”