Inventory
info icon
Single family homes on the market. Updated weekly.Powered by Altos Research
682,150-7,865
30-yr Fixed Rate30-yr Fixed
info icon
30-Yr. Fixed Conforming. Updated hourly during market hours.
6.91%0.02
Appraisals and ValuationsSponsored

How data and technology can combat appraisal bias

May 23, 2023 1:10 pm  By
Data & AnalyticsFair HousingTavantTechnology
appraiser appraisal

In the past few years, appraisal bias has come to the forefront of conversations about appraisals and how they can be improved. According to many reports, it’s been an issue for a long time, but has only recently come into the spotlight with a high profile case in 2022 and increased discussions around fair and equal housing.

“You can’t ignore it. It’s a problem, it’s real, it’s documentable,” said Shannon Johnson, Touchless Lending product manager at Tavant. “Today, we actually have an opportunity to combat it in a way that we didn’t have five or 10 years ago, which is leveraging automation, artificial intelligence (AI) and machine learning.”

At Tavant, they’re taking a look at the past and present of appraisals to prepare for a better future.

Problems with the current state of appraisals

One of the current problems with appraisal is the barrier to entry, said Dawn Svedberg, vice president of sales for Tavant. The profession is not widely known, and there are significant barriers to entry, she said. We often hear the statistic that the average appraiser is a 57-year-old white male.

To break into appraising, you need to essentially apprentice with an existing appraiser, and for them to spend that much time training you often requires them to already have a close relationship with you.

Additionally, appraisals are conducted in much the same way today as they’ve been done for the last 60 years, she added.

“Appraising can be a subjective process in a lot of ways — as appraisers are always working within dollar ranges and imperfect information. Increased data access has improved the process, but there is still a lot of ‘gut’ that can perpetuate bias,” she said. “We’re trying to inject more analytics — regression, time value, and other — to increase objectivity. There’s been progress, but there’s still a long way to go.”

Recent appraisal changes

Luckily, changes are being made to appraisals even today. In March, Fannie Mae released an update to its Seller’s Guide that outlined more options for property valuations, including value acceptance (formerly appraisal waivers), value acceptance plus property data and hybrid appraisals.

“These additional valuation options will lessen the opportunity for appraisal bias.” Johnson said. “The need to physically inspect each property will be replaced with data or other inspection options in many cases. I think that’s a really great step.”

Appraisal concerns have also been mitigated by more easily accessible data. In the past, real estate data was hidden behind paywalls and subscriptions, and any consumer who wanted to know what was available or what a home sold for had to reach out to a Realtor.

Now, anyone can idly browse real estate listings and find out what homes have sold for and what inventory is available. Consumers can even look at comps for their own home and are quite accurate in estimating what their home is worth when they’re going to either purchase or to refinance,” Svedberg said.

This gives them an idea of whether they’re being discriminated against in the appraisal process, reducing the risk of appraisal bias — or at the very least, increasing the consumer’s ability to spot if it’s happening to them.

Connecting the data dots

According to Tavant, the next step toward eliminating appraisal bias is to connect the dots of all this data and information.

“I think it will give lenders the comfort in knowing they have a solid appraisal and then the loan can move on,” Johnson said. “95% of the time there isn’t bias present, but we need  to identify it and determine the impact of it so we can solve that 5%.”

Tara Dugan, who works in marketing and real estate for Tavant, said that technology should focus on integrations and aggregations.

“The information is becoming more accessible, but we need to integrate and aggregate all the different platforms and be more collaborative with the data to be effective and cohesive in real time on a very granular level — that, and standardize everything,” she said. “On an industry leader level, that is where we need to be. I think that we’ve seen that [it may not happen] on a government level, so it is maybe going to take the private sector to lead the way.”

Tavant’s approach and Touchless Lending

Johnson said she and her team at Tavant are approaching the issue in a few different ways. The first involves data — looking at the subject property and its comps, then overlaying that with neighborhood and demographic information to make sure the comps make sense.

They also use image analysis to make sure there are no potentially bias-creating items in the house, and natural language processing to read freeform comments for context and to relate them back to appraisal adjustments.

“Everything’s like a spiderweb, it’s all connected,” Johnson said. “When you pull all of these pieces of information together, you can start doing regression testing. As more appraisals run through, you can see patterns for a particular area, maybe even down to a particular appraiser.”

The end goal is to make sure everyone in the process is confident that appraisers, lenders and tech companies are doing their best to keep lending fair and equitable.

Tavant’s work to eliminate appraisal bias is tied into its Touchless Lending platform, where it is building in its Collateral Analysis assessment, Svedberg said.

Many lenders set up manual appraisal review teams using the same people who are doing the appraisal work, which can lead to issues with objectivity and scale. But Tavant’s Touchless Lending Collateral Analysis assessment will allow lenders to review appraisals at scale and know which may require a closer look.

“Tavant’s Touchless Lending Collateral Analysis will provide lenders an additional quality and consistency check utilizing information from numerous data sources and will include an appraisal bias flag,” Svedberg said. “Allowing lenders to focus on the outliers and streamline their operations while providing a better, and unbiased experience for their borrowers.”

“It’s information, automation and transparency,” Johnson said. “Tavant has subject matter expertise to make sure that everything we’re developing related to mortgage is relevant to the people who are going to be using it. That’s the journey that we’ve been on, and after decades of discussion, it is fantastic to see lenders and borrowers benefitting today.”

To learn more about Tavant’s Touchless Lending platform, click here.

3d rendering of a row of luxury townhouses along a street

Log In

Forgot Password?

Don't have an account? Please