Advancing Collateral Valuation: A Data-Driven Approach with Standardized Property Data Collection

Single-Family Business

Helping lenders serve homebuyers and homeowners with affordable mortgages

Multifamily Business

Financing for quality, affordable rental housing in every market, every day

Capital Markets

Reducing risk and enhancing housing finance liquidity

Manage Financial Uncertainty

All Resources to Manage Financial Uncertainty

Help for Homeowners

Help for Renters

Protect Yourself from Fraud

Recover from a Disaster

All Resources for Recovering from a Disaster

Recovery Assistance for Homeowners

Recovery Assistance for Renters

Protect Yourself from Fraud

In the rapidly evolving residential real estate landscape, there is a resounding demand to continue modernizing the property valuation process. As the industry begins to embrace digital transformation, data-driven methodologies, and artificial intelligence, significant changes in how collateral is evaluated in the mortgage process are underway. These advancements are aimed at enhancing accuracy, objectivity, transparency, and efficiency, serving the needs of all stakeholders in today's dynamic residential real estate market.

Collecting data about the subject property is a foundational element of the mortgage system, serving a crucial role in various aspects such as appraisals, market analyses, investment decisions, regulatory compliance, automated valuation modeling (AVM), and quality control. Over the past six years, Fannie Mae has worked with industry partners to create a standardized property data collection process designed to produce factual, accurate, and consistent property data and exhibits. These efforts culminated in the introduction of the GSE-aligned Uniform Property Dataset (UPD) on December 1, 2023. Fannie Mae made the UPD mandatory for new loans that involve a property data collection beginning April 1, 2024. Moving forward, property data collection conforming to the UPD will be integral to Fannie Mae's suite of valuation offerings. While this standardized approach may be unfamiliar to some industry stakeholders, Fannie Mae has diligently tested and refined it on hundreds of thousands of loans in our valuation modernization pilot program. As industry usage and adoption grows, questions naturally arise. I want to address some of those questions and share valuable insights gained from our experience over the last six years.

What is Property Data Collection? Property data collection involves physically observing and documenting property characteristics, which typically includes 120 standardized property attributes, 40-60 photographs, and an ANSI® (American National Standards Institute®) compliant floor plan. This task is carried out by trained and vetted property data collectors, and the collected data must adhere to the standards defined within the UPD. Importantly, the property data collection process is deliberately objective, focusing solely on capturing factual information about the property. It does not involve subjective elements such as comparables, opinion of value, or condition/quality ratings, and it should not be confused with an appraisal¹.

How is the data collected? Property data collectors utilize intuitive mobile applications assessed by Fannie Mae that facilitate data capture as they navigate through the property. These applications guide them through the necessary data fields, enabling them to capture high-resolution photos and generate a floor plan of the property. Advanced technologies, such as 3D scanning and LiDAR (Light Detection and Ranging), are also becoming more widely used. These technologies allow for the creation of detailed 3D property models, enabling downstream users to virtually tour the home. Additionally, LiDAR technology integrated into mobile devices can rapidly generate floor plans with room labels, interior walls, and accurate measurements of gross living area within minutes.

What are the requirements and expectations of a property data collector? Fannie Mae has established clear policy requirements in our Selling Guide. Lenders must ensure that data collectors are selected in accordance with Fannie Mae requirements, vetted through an annual background check, professionally trained, and they must also possess the essential knowledge to competently complete the property data collection.

Additionally, upon completing data collection, collectors must sign certifications affirming that the data was objectively captured, free from personal bias, and is accurate and reliable. Fannie Mae has also implemented Property Data Collector Independence Requirements to safeguard the independence, objectivity, and impartiality of property data collectors in the lending process.

Can this data be trusted? Ensuring data integrity is a cornerstone at Fannie Mae, essential to every facet of our business. We view it as our responsibility to foster industry-wide progress and maturity in property data collection.

Our commitment begins with a rigorous onboarding process for each service provider, where thorough vetting is conducted before they can receive, assign, and fulfill property data collection orders. This process includes a comprehensive solution review, examining the entire process from lender order, quality control, API submission, and lender delivery. It includes extensive test case reviews covering various property types and scenarios. We also conduct a detailed assessment of mobile applications to ensure a positive user experience and compliance with UPD standards.

Once onboarded, service providers undergo extensive quality control measures. New service providers attend mandatory quality control training sessions where we share common findings and best practices. During an assessment period, we review 100% of their property data collections, providing timely feedback based on internal reviews. Additionally, and once a service provider is through our assessment period, we regularly sample and report on property data quality through compliance reports and we host meetings to discuss compliance trends, issues, and case studies.

To bolster data integrity, we leverage technology such as our Property Data API, which serves as the interface for transmitting data to Fannie Mae. Additionally, we offer stakeholders access to our Property Data API Review Tool, a web-based utility for visualizing and reviewing property data attributes, photos, and floor plans. These technologies provide messaging regarding compliance, submission status, and flags on data points that could affect loan eligibility.

What are some statistics associated with the property data collection process? As of August 2024, we have approved 59 service providers to conduct property data collection for loans eligible for delivery to Fannie Mae, with over 300,000 property data collections processed to date. To ensure we maintain the highest quality standards, we have conducted extensive analyses in four critical areas and the results demonstrate consistency between the data collected in property data collections versus data collected in appraisals.

1. Data Consistency²: We evaluate property characteristics such as gross living area (GLA), lot size, view, and location by comparing current appraisals to prior appraisals on the same property, as well as current property data collections to prior appraisals on the same property. We measure the data agreement rate between both comparison groups, and the results are encouraging: GLA measurements and reported lot size align within 10%+/- in 85% and 90% of cases in both comparison groups. The reported view aligns 90% of the time, and the reported location aligns 87% of the time for both comparison groups.
2. 
   
3. Collateral Underwriter® (CU®) Performance³: We compare traditional appraisals to hybrid appraisals, which rely on property data collection, using CU risk scores and flags as benchmarks. CU generates a risk score on a scale from 1 to 5 with risk scores of 2.5 and below indicating lower risk and qualifying for representations and warranties relief on property value⁴. Using risk scores 2.5 and below and risk scores 2.6 and above as a way to compare the risk of the two appraisal types, the differences observed are minimal. CU risk scores of 2.5 and below differ by just 1.2%, and risk scores of 2.6 and above differ by 1.3% between the two appraisal types. The occurrence of the four main CU flags-Eligibility & Compliance, Undervaluation, Overvaluation, and Appraisal Quality-also shows a negligible difference of 1.2% between the two appraisal types.
4. 
   
5. Fannie Mae Loan Quality Center (LQC) Insights⁵: We assess collateral-related loan defect rates between valuation options that use property data collections and traditional appraisals. The difference in defect rates is a mere 0.6%, indicating that valuation options that use property data collections are on par with traditional appraisals in maintaining loan quality.
6. 
   
7. Loan Performance⁶: We analyze serious delinquency rates⁷between loans that use property data collections and those that use traditional appraisals. Over the time period we observed, the significant delinquency rates are the same between the two methods.

These objective measures help ensure property data collection meets the stringent standards required to execute our mission safely and soundly. As our processes evolve and technology advances, we anticipate continuous improvements in the quality and efficiency of property data collection, further enhancing our ability to drive business outcomes.

How is property data collection utilized? There are two primary uses today:

1. Value acceptance + property data: This innovative approach is akin to our value acceptance (appraisal waiver) offering. Here, we accept the lender's estimated property value (or contract price for purchase loans) at the loan submission to Desktop Underwriter® (DU®) with the requirement to obtain a property data collection. The lender engages an approved service provider to conduct a property data collection. Once completed, the service provider submits the data to Property Data API for compliance verification. After confirming compliance, the data is sent to the lender for review to ensure loan eligibility for delivery to Fannie Mae. Contrary to some lenders' concerns about increased repurchase risk associated with property data collection, we offer representations and warranties relief on property value with this option, similar to value acceptance (appraisal waiver). Lenders are responsible for confirming property condition and eligibility, as they would with traditional appraisals, which can significantly reduce repurchase risk with value acceptance + property data. This option provides lenders and consumers with value certainty earlier in the mortgage process and saves consumers on average $350-$400 over the traditional appraisal.
2. 
   
3. Hybrid pilot: This option, which has been in pilot phase for several years, starts with a participating lender ordering a property data collection from an approved service provider when a loan is eligible for the pilot. Upon completion, the data is submitted to Property Data API for compliance verification. After confirming compliance, the data is sent to an appraiser who uses it to fulfill their professional obligation to identify relevant characteristics of the subject property⁸and complete a hybrid appraisal. To date, we have received more than 200,000 hybrid appraisals, and by comparing CU risk scores and flags and our LQC findings and defects, hybrid appraisals perform very similar to traditional appraisals. We are currently exploring whether this option would be appropriate as a permanent Selling Guide offering available to all lenders.

In conclusion, Fannie Mae's commitment to standardizing property data collection with the UPD represents a significant advancement in how collateral is evaluated in the mortgage process. By embracing this initiative, the industry not only improves the mortgage process, but also promotes greater objectivity and transparency in the property valuation process. Through a rigorous service provider onboarding process, service provider training, technology controls, and quality assurance measures, Fannie Mae promotes the integrity of the data collected, thereby building trust and confidence in this modernized approach. As standardized property data collection gains broader adoption, it has the potential to reshape the property valuation process, benefiting industry stakeholders and improving the overall experience for both homebuyers and lenders.

Trademarks are the property of their respective owners.

¹ See 2024 Uniform Standards of Professional Appraisal Practice (USPAP) appraisal definition: APPRAISAL: (noun) the act or process of developing an opinion of value; an opinion of value. (adjective) of or pertaining to appraising and related functions such as appraisal practice or appraisal services.

² Fannie Mae analysis based on appraisal and property data collection submissions from January 2023 through July 2024.

³ Collateral Underwriter is Fannie Mae's proprietary appraisal risk assessment tool. Fannie Mae analysis based on loan deliveries that used traditional appraisals and hybrid appraisals from January 2021 through May 2024.

⁴ See Fannie Mae Selling Guide section A2-2-06

⁵ Fannie Mae analysis based on loan deliveries that used property data collections and traditional appraisals from May 2023 through April 2024.

⁶ Fannie Mae analysis based on loan deliveries that used property data collections and traditional appraisals from January 2020 through December 2023.

⁷ Our single-family serious delinquency rate is expressed as a percentage of our single-family conventional guaranty book of business based on loan count.

⁸ See 2024 Uniform Standards of Professional Appraisal Practice (USPAP) Standards Rule 1-2, Problem Identification (e) - "In developing a real property appraisal, an appraiser must: identify, from sources the appraiser reasonably believes to be reliable, the characteristics of the property that are relevant to the type and definition of value and intended use of the appraisal."

Public link

Please use the above public link if you want to share this noodl on another website.