FINANCE, POLICY & MARKETSPublished by Paul Ivinskas
fc.The Financial CurrentDAILY INTELLIGENCEWhat matters across finance
Deep-dive library

Appraisal-data modernization: UAD 3.6, comparable properties and the cost of a common format

7 min read · estimatedAI-generated analysis · Methodology
Current version · 1 version · Publication details

First published . This version published .

Initial full article. Primary sources checked October 4, 2026; historical research retains its dates, and numerical illustrations are hypothetical.

Related research, policy & entities ↓

At a glance

Excerpts from this version
What it covers
UAD 3.6 changes how appraisal information is structured, transmitted and reviewed. A common format can improve comparability, but it does not settle property value; the September 30, 2026 exception also makes migration status more nuanced than a single deadline.
Why migration can affect financing economics
The example shows why migration involves a chain of participants: appraiser, software provider, appraisal-management company, lender, aggregator and final buyer. A file can appear correct to one participant and fail a later system's validation. An exception available to a Seller also does not mean every investor in the chain will accept every legacy-format report. Contractual eligibility, operational readiness and successful transmission are separate questions.Read in context
0% through article

Tap a dotted-underlined term for a definition; terms are highlighted once per section. Use Aa in the navigation for reading preferences.

In this article

The same property, a different information structure

An appraisal contains both observations about a property and an opinion about its value. Modernizing the information structure changes how those observations are recorded and exchanged. It does not mechanically determine the valuation conclusion. Freddie Mac describes the Uniform Appraisal Dataset 3.6 and Forms Redesign as an update aligned with the MISMO reference model and a move toward a flexible, data-driven appraisal report. MISMO provides a common mortgage-data vocabulary; UAD specifies appraisal information used in the GSE process. [1]

The distinction is similar to changing a financial report from a scanned image into consistently labeled fields. A computer can identify bedrooms, living area or a comparable sale without first guessing where the information appears on a page. But a correctly labeled field can still contain an incorrect observation. A standardized address does not establish that a property was inspected accurately, and a valid data file does not establish that the selected comparable properties support the final value.

A report driven by characteristics rather than a form number

The redesigned Uniform Residential Appraisal Report, or URAR, uses a flexible structure driven by property information. Freddie Mac's FAQ explains that legacy appraisal form numbers are being replaced by this data-driven approach, alongside related update and completion reports. This is broader than replacing a PDF's visual style: downstream systems must understand the new data structure and the circumstances in which different information appears. [2]

An invented example illustrates the operational gain. Two appraisers might previously describe an additional living space in different free-text locations. A reviewer searching one familiar box could overlook the second description. A structured representation can make the characteristic easier to find and compare. Yet the reviewer still needs to understand whether the spaces are genuinely comparable in legal use, access, condition and market appeal. Standardization makes the question more visible; it does not supply the answer.

Dynamic reporting also changes the meaning of absence. A section might be absent because it is inapplicable, because a value was not observed, or because the file is incomplete. Those explanations have different consequences for review. When software treats every missing field as a zero or every omitted section as evidence that a characteristic does not exist, it can introduce errors while appearing to simplify the workflow.

Comparable data does not guarantee comparable properties

Imagine three hypothetical nearby sales at $400,000, $420,000 and $460,000. Their average is about $426,667. That arithmetic is reproducible, but it is not an appraisal. One sale might include a major renovation, another might have occurred before a local price change, and the third might involve a materially different site. A common data format can make these distinctions easier to analyze without deciding how much each should affect the subject property's value.

Suppose the subject property's reported living area changes from 1,900 to 2,050 square feet after a measurement correction. A simple price-per-square-foot calculation changes even if no sale price moves. If the old and new values are silently merged, a time series may appear to show a change in market pricing that actually arose from better measurement. This is an analytical illustration, not a claim that UAD 3.6 itself mandates that particular correction.

The same issue arises when a broad category is divided into several more specific categories. The new categories may be more informative, but a historical comparison needs a bridge. An apparent rise in a newly recorded feature could reflect improved reporting rather than physical changes in the housing stock. Preserving the distinction between observation, classification and market movement makes the richer data useful rather than merely larger.

The mandate remains, with a conditional exception

As checked October 4, 2026, the November 2, 2026 UAD 3.6 mandate remains in place. The September 30 joint GSE announcement offers Sellers a temporary policy exception that must be obtained from each applicable GSE. It is not a blanket postponement. Approved exceptions allow new UAD 2.6 reports through May 19, 2027 and resubmissions through June 27, 2027. Sellers must follow agreed implementation plans; the announcement describes the exception as one-time and not extendable. Lenders that do not sell directly to a GSE must align with their investors rather than independently request the Seller exception. [3]

The updated timeline distinguishes several operational stages. Starting March 1, 2027, UAD 2.6 reports receive the designated scores of 99 in Freddie Mac's Loan Collateral Advisor and 999 in Fannie Mae's Collateral Underwriter, and the associated loans are ineligible for collateral representation-and-warranty relief for value. From May 20, only UAD 3.6 can be used for new submissions; previously submitted 2.6 reports have a resubmission window through June 27. UAD 2.6 retires June 28, 2027. These are the published transition terms, not predictions about a later extension. [4]

Freddie Mac's FAQ further explains that the mandate turns on the initial UCDP submission date, rather than the loan application date or appraisal effective date. It also distinguishes retirement of the submission format from the ability to sell a loan that already has a valid report, subject to the usual appraisal-expiration requirements. A timeline describing submissions therefore cannot be read as a universal loan-sale cutoff. [2]

Why migration can affect financing economics

Consider an invented lender funding a $300,000 loan before delivering it to an investor. If a data-format problem delays delivery for ten days, the funding need persists for those extra days even when the borrower and property have not changed. At a purely hypothetical 6% annual financing rate using a 365-day year, ten additional days cost approximately $493.15. That is an illustration of timing cost, not a quotation of warehouse financing or an estimate of actual UAD-related delays.

The example shows why migration involves a chain of participants: appraiser, software provider, appraisal-management company, lender, aggregator and final buyer. A file can appear correct to one participant and fail a later system's validation. An exception available to a Seller also does not mean every investor in the chain will accept every legacy-format report. Contractual eligibility, operational readiness and successful transmission are separate questions.

The reduced-functionality period adds a different kind of consequence. A lender may retain a path to submit a report while losing the specified collateral warranty relief for value. Submission success therefore is not equivalent to receiving every risk-management benefit associated with a fully supported report. This is an economic interpretation of the published transition terms, not a claim that a particular loan must be repurchased or that every other form of relief disappears.

Public research files are a separate product

FHFA's appraisal-level public-use files are not public copies of every confidential appraisal submitted to the GSEs. The documentation for version 2.1 describes five-percent samples of eligible Enterprise-acquired and FHA-insured mortgage appraisals, selected fields and disclosure-avoidance measures. It discusses removing identifying and precise-location information, recoding or limiting some values, and sampling considerations. The public file has a different purpose and information boundary from the operational submission system. [5]

The public page checked for this article lists Enterprise data for 2013–2022 and FHA data for 2017–2022, with the Enterprise release and documentation dated February 4, 2025. Those historical coverage dates matter: that listed file cannot directly demonstrate how UAD 3.6 production reports performed in 2026. The existence of a richer current submission standard does not mean its complete records are already available in public research data. [6]

A hypothetical researcher studying valuation differences also needs to distinguish the sampled mortgage-appraisal population from all homes. Properties bought with cash, homes without a transaction and loans outside the file's acquisition or insurance scope may be absent. A statistic can be accurately calculated within the available sample while answering a narrower question than a headline about the whole housing market suggests.

Better evidence still requires interpretation

UAD modernization can reduce ambiguity in transmission, support more consistent review and make property characteristics easier to analyze. Its benefits depend on the quality of the underlying observations and on participants interpreting fields consistently. A common vocabulary is infrastructure for evidence, not a replacement for judgment about comparable sales, local conditions or the appropriate valuation conclusion.

The transition's immediate significance lies in the combination of data design and operational eligibility. The mandate, approved exceptions, reduced functionality, new-submission cutoff and final retirement date describe different events. Keeping them separate makes the migration understandable for borrowers, lenders, appraisers and researchers without turning a complex conversion into a misleading claim that everything changes on one day.

Sources

  1. Freddie Mac, Uniform Appraisal Dataset resources; checked October 4, 2026SourceBack to text: ↑
  2. Freddie Mac, UAD 3.6 FAQ; mandate, submission and report-format explanations checked October 4, 2026SourceBack to text: ↑1↑2
  3. Freddie Mac and Fannie Mae, Temporary Policy Exception for Sellers Unable to Meet UAD 3.6 Mandate, September 30, 2026Source · PDFBack to text: ↑
  4. Freddie Mac and Fannie Mae, updated UAD redesign timeline; checked October 4, 2026Source · PDFBack to text: ↑
  5. FHFA, UAD PUF version 2.1 Data Documentation, February 4, 2025Official source · PDFBack to text: ↑
  6. FHFA, UAD Appraisal-Level Public Use File release page; checked October 4, 2026Official sourceBack to text: ↑

Flag an error or suggest a correction →Public corrections log →