Evaluations inherited the proof appraisals used to supply

The October 2019 rule let banks skip appraisals on home loans up to $400,000 if they obtained an evaluation instead. Evaluations follow guidance, not appraisal standards, so each lender had to prove its own valuation method was sound, years before AVM rules arrived.

Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.

Two-minute summaryThe dek and the first sentence of each section, quoted from the article

The October 2019 rule let banks skip appraisals on home loans up to $400,000 if they obtained an evaluation instead. Evaluations follow guidance, not appraisal standards, so each lender had to prove its own valuation method was sound, years before AVM rules arrived.

  1. The week On October 8, 2019, the Office of the Comptroller of the Currency, the Federal Reserve Board and the Federal Deposit Insurance Corporation published a final rule in the Federal Register, at 84 FR 53579, amending their real estate appraisal regulations.
  2. How the threshold works Title XI requires appraisals for federally related transactions, and it lets the federal financial institutions regulatory agencies decide which real estate-related financial transactions do not need them.
  3. When to use an evaluation The threshold permits an evaluation.
  4. What goes into an evaluation An appraisal’s content is defined by USPAP.
  5. Who prepares and reviews it Appraisals for federally related transactions are prepared by state certified or licensed appraisers.
  6. What the borrower sees The threshold change also altered what borrowers receive.
  7. Banks, credit unions and nonbank lenders The rule applied to institutions regulated by the three agencies that issued it.
  8. The strongest objections The argument here is that the rule shifted the burden of proving collateral value onto lenders’ own evaluation programs.
  9. With hindsight The threshold held, and its first test came in the spring of 2020.

The week

On October 8, 2019, the Office of the Comptroller of the Currency, the Federal Reserve Board and the Federal Deposit Insurance Corporation published a final rule in the Federal Register, at 84 FR 53579, amending their real estate appraisal regulations. The agencies had announced it in a joint press release on September 27. The rule raised the threshold at or below which an appraisal is not required for residential real estate transactions from $250,000 to $400,000. It defined a residential real estate transaction as one secured by a single 1-to-4 family residential property. The change appeared in each agency’s own regulation: 12 CFR 34.43 for the OCC, 225.63 for the Board and 323.3 for the FDIC. The press release said the threshold was last changed in 1994 and that, given price appreciation since then, the change would provide burden relief without posing a threat to safety and soundness.

The rule did not remove valuation from exempt loans. For residential transactions exempted by the new threshold, regulated institutions must obtain an evaluation of the real property collateral that is consistent with safe and sound banking practices. The press release described evaluations as generally less burdensome than appraisals and noted that they had been required since the 1990s. The rule text itself treats them differently from appraisals in one central way: federal law does not require evaluations to follow the Uniform Standards of Professional Appraisal Practice, known as USPAP, or to be performed by state certified or licensed appraisers. The agencies’ expectations for evaluations are set out in supervisory guidance, chiefly the Interagency Appraisal and Evaluation Guidelines adopted in 2010 and an interagency advisory on evaluations issued in 2016.

The rule did two other things. It added to the list of exempt transactions those covered by a rural residential appraisal exemption that Congress created in 2018, in the Economic Growth, Regulatory Relief, and Consumer Protection Act, and it required evaluations for them. And it required regulated institutions to subject appraisals for federally related transactions to appropriate review for compliance with USPAP, implementing a requirement added by the Dodd-Frank Act. The threshold increase took effect October 9, 2019. The rural evaluation requirement and the appraisal review requirement took effect January 1, 2020.

The rulemaking had drawn more than 560 comments since the agencies proposed it in December 2018. Financial institutions, their trade associations and state banking regulators generally supported it. Appraisers, appraiser organizations, individuals and consumer advocates generally opposed it. Many commenters said automated valuation models, or AVMs, were unreliable, and some argued that the agencies should not widen the threshold before issuing the quality control standards for AVMs that Title XI of the Financial Institutions Reform, Recovery, and Enforcement Act of 1989 required. The agencies adopted the threshold anyway. As Title XI requires, the Consumer Financial Protection Bureau concurred that the new threshold provides reasonable protection for consumers who purchase 1-to-4 unit single-family residences.

The rule came from the three federal banking agencies. The National Credit Union Administration, also a Title XI agency, was not a party to it.

On its face, the October 2019 appraisal rule changed a threshold. Its more lasting effect was on evidence. An appraisal brings its own standard: a licensed or certified professional, working under USPAP, signs a report that follows a prescribed structure. An evaluation brings no such standard by law. When the threshold rose, a larger share of residential loans could rest on a valuation whose adequacy the lender itself had to prove, against supervisory guidance, with no external standard to point to. The argument here is that the rule moved the burden of proof for collateral value from the appraisal profession to each lender’s own valuation program, and that the gap was filled only partly, and years later, by AVM rules. It starts with how the threshold works. It then takes four parts, one for each decision a lender owned once an evaluation could replace an appraisal: when to use an evaluation, what goes into it, who prepares and reviews it, and what the borrower sees. It then turns to how the rule differed by type of lender, and to the strongest objections.

How the threshold works

Title XI requires appraisals for federally related transactions, and it lets the federal financial institutions regulatory agencies decide which real estate-related financial transactions do not need them. The agencies have used that authority to exempt several categories, including transactions at or below set thresholds. Before the 2019 rule, the $250,000 threshold applied in principle to all real estate-related transactions, but in effect only to residential ones, because commercial and business loans had higher thresholds. Commercial real estate transactions of $500,000 or less and certain real estate-secured business loans of $1 million or less already needed no Title XI appraisal.

Commercial lenders were already used to evaluations below those thresholds. The 2019 rule brought a larger slice of consumer home lending into the same model. The difference is the borrower. A commercial borrower is a business. A residential borrower is a household, protected by consumer rules that assume a professional valuation is part of the process.

An exemption that carries a duty

The exemption is conditional. For transactions at or below the threshold, the regulations require an appropriate evaluation consistent with safe and sound banking practices, unless the transaction also qualifies for another exemption that needs no evaluation. Loans sold to Fannie Mae or Freddie Mac, or insured or guaranteed by a U.S. government agency, were already separately exempt from the agencies’ appraisal requirement, and those programs set their own valuation rules.

The agencies also kept a reservation of power. The OCC’s regulation says it may require an appraisal whenever it believes one is necessary to address safety and soundness concerns, and the other agencies’ regulations say the same. A lender cannot treat the threshold as a ceiling on what an examiner may expect.

The threshold measures the loan, not the house

For loans, the transaction value that the threshold tests is the amount of the loan, not the value of the property. A $400,000 loan on a house worth $800,000 qualifies. So does a $400,000 loan on a house worth $420,000. The two carry very different collateral risk, because the second has almost no equity cushion. The threshold treats them alike. That is one reason a lender’s selection policy has to bring loan-to-value ratio back into the decision.

Existing loans

The regulations also exempt transactions involving an existing extension of credit at the lending institution, but require an evaluation for them, provided there has been no obvious and material change in market conditions or in the physical property that threatens the adequacy of the collateral, even with new money advanced, or there is no new money beyond reasonable closing costs. Renewals, modifications and refinances with the same lender fall here. The lender has to judge whether a material change has occurred, and the evaluation program has to cover those decisions as well as new originations.

How many loans moved

The agencies estimated the reach using 2017 Home Mortgage Disclosure Act data, limited to first-lien single-family originations by FDIC-insured institutions and their affiliates that were not sold to the government-sponsored enterprises or insured or guaranteed by a federal agency. They found that about 214,000 additional transactions would be exempted, which they said represented only three percent of total HMDA originations in 2017 but 16 percent of the regulated transactions in their data. The headline share is small. The share of loans outside the enterprise and government programs, which include the loans banks keep and whose collateral they rely on, is not.

Why the agencies were comfortable

The agencies gave three reasons for finding the new threshold safe. Supervisory experience pointed to other causes of losses. Call report data showed the net charge-off rate for residential real estate loans remained relatively unchanged after the 1994 threshold increase through 2007, and the agencies attributed the losses of 2008 through 2013 primarily to weak underwriting. And evaluations would still be required. The third reason carries the weight. The rule’s safety rested on evaluations being done well.

When to use an evaluation

The threshold permits an evaluation. It does not require one. Each lender decides, loan by loan or by policy, when an evaluation is enough and when to order an appraisal anyway.

What the data showed about discretion

The agencies addressed commenters’ fear that competition would push lenders to use evaluations whenever possible. They cited a five-year review of supervisory data from large institutions, drawn from the FR Y-14M reports that bank holding companies with $50 billion or more in assets file. It found that larger lenders obtained appraisals on 74 percent of portfolio residential originations at or below the old $250,000 threshold. The agencies read that as evidence that institutions are often exercising discretion and are not automatically using evaluations whenever permitted.

That evidence covered large lenders under detailed supervision. It said nothing about smaller banks, and it described behavior under the old threshold. A selection policy written down in advance is the only way a lender can show an examiner that its choice between appraisal and evaluation follows risk, not convenience.

What a selection policy should weigh

A defensible policy states the conditions under which an evaluation is acceptable and the conditions that call for an appraisal regardless of loan size. The relevant factors follow from what an evaluation can and cannot see. Loan-to-value ratio matters, because a thin equity cushion makes valuation error more costly. Property type matters, because unusual or rural properties have fewer comparable sales and give models less to work with. Market conditions matter, because rapid price changes make recent data less reliable. Data quality matters, because an evaluation built on stale or incomplete property records inherits their errors.

The policy also has to address the loans for which other rules require an appraisal. The 2019 rule itself pointed to the higher-priced mortgage loan rule, under which lenders must obtain an appraisal, in some cases two, with an interior property visit, for certain higher-priced loans secured by a consumer’s principal dwelling, and provide free copies. A threshold exemption does not override that rule. The selection policy needs a step that checks for it.

Cost and time

Supporters argued that the higher threshold would cut costs and closing delays for lenders and borrowers. The agencies were more cautious. They said the savings were difficult to quantify because of limited data and variation by type and complexity of transaction, that costs may be passed on to borrowers, and that the extent to which institutions would use evaluations instead of appraisals was uncertain, since they kept the option of ordering appraisals. They also noted that commenters in their earlier regulatory review had said certified and licensed appraisers can be difficult to find, particularly in rural areas. A lender deciding where to use evaluations is balancing a real time saving in some markets against the cost of building and defending its own program.

Recording the choice

Each loan file should show which valuation was used and why. If the policy allows an evaluation, the file should show that the loan met the policy’s conditions. If the lender chose an appraisal for a loan that could have used an evaluation, recording the reason helps show the policy is applied. Aggregated across loans, those records answer the examiner’s first question: how often does the bank use evaluations, for which loans, and has that changed.

A portfolio view

The same records support a portfolio measure. The share of the residential book valued by evaluation, broken down by product, region, loan-to-value band and preparer or model, shows where the bank’s collateral values rest on its own method. A concentration in one model or one vendor is a dependency like any other. If that model’s accuracy slips in a market, every loan it valued there carries the error. Reporting that share to the credit risk committee each quarter makes the evaluation program visible at the level where risk appetite is set.

What goes into an evaluation

An appraisal’s content is defined by USPAP. An evaluation’s content is defined by guidance and by the lender’s own standards. The 2019 rule restated the agencies’ expectation: an evaluation should contain sufficient information and analysis to support the institution’s decision to engage in the transaction, including information about the actual physical condition and characteristics of the property.

Condition and inspection

Physical condition is one place where evaluations can differ from appraisals. Some commenters said appraisals are more accurate than evaluations because they include a physical inspection. Some evaluation providers said theirs typically include one. The agencies said the individual performing the evaluation should determine whether a physical inspection is necessary to support the property’s value, and that inspections by appropriately trained individuals can provide useful information and help keep valuations consistent with safe and sound practices.

That leaves the inspection decision inside the evaluation process, not in a rule. A lender needs a standard for when its evaluations include an inspection, of what kind, and by whom. Without one, the content of its evaluations varies with each preparer’s habits.

When a model should not be used

Every valuation model has conditions under which its estimate deserves little weight. A property with few recent comparable sales, an unusual size or design, a recent major renovation, damage, or a location at the edge of a market can all defeat a model that works well on typical homes. Where a model reports a confidence measure alongside its value, the method document should say what level is acceptable. It should also say which property characteristics rule out a model-only evaluation, and what happens then: an inspection-based evaluation, a different model, or an appraisal. Writing those rules down in advance is what separates a governed use of models from a convenient one.

Models as evaluation tools

Automated valuation models were at the center of the debate. Many commenters said AVMs were unreliable and that the threshold increase would lead to greater reliance on them. Others said AVMs could provide valuable information and that better technology and data had improved evaluations. One commenter said AVMs were more predictive of default than appraisals, and another that evaluations based on AVMs were more objective because they were not skewed by knowledge of the contract price. Title XI already required the agencies to issue quality control standards for AVMs. In 2019 those standards did not exist, and the agencies adopted the threshold without waiting for them.

A lender using AVMs in evaluations in 2019 was therefore governing a model with no AVM-specific rule. The obligations came from the evaluation guidance and from the lender’s own risk management. That meant knowing which AVM was used for which loans, how accurate it was in the lender’s markets, when its estimates should not be relied on, and how its results were tested against later sales or appraisals.

A method document

The practical answer is a written evaluation method, maintained like any other model or procedure. It names the data sources, the AVMs and their use conditions, the inspection standard, the content of the report, the confidence measures that trigger escalation to an appraisal, and the testing program. It records versions and approvals. An examiner who asks how the bank knows its evaluations are adequate can be shown the document, the test results and the exceptions.

Testing whether evaluations are right

An evaluation program needs a feedback loop. The most direct test compares evaluation values with later independent evidence of value: subsequent sale prices, appraisals ordered on a sample of loans that also received evaluations, or appraisals ordered when a borrower disputed an evaluation. A lender can track how far evaluation values fall from that evidence, whether the gap is larger for some property types, price ranges or areas, and whether it is moving over time.

Sample appraisals are the most informative and the most expensive part of that loop. Choosing the sample deliberately, with more weight on loans where the policy’s conditions were close calls, makes each one count. The results feed back into the selection policy. If evaluations prove less reliable for a property type, the policy should send more of those loans to appraisal.

Property data

An evaluation is only as good as the property data behind it. Records of square footage, condition, recent improvements and comparable sales come from public records, listing services, prior loan files and inspections, and each source has gaps. A lender’s method document should say which sources it uses for which fields, how it handles conflicts between them, and when missing data triggers an inspection or an appraisal. Data lineage from source to valuation is what lets a reviewer reproduce a value months later.

Who prepares and reviews it

Appraisals for federally related transactions are prepared by state certified or licensed appraisers. Federal law does not require that of evaluations, so they can be prepared by a wider range of people and firms. The rule and the guidance it cited addressed the gap through expectations about competence and independence.

Independence

The agencies noted that the Guidelines and the 2016 advisory provide that individuals preparing evaluations should be qualified, competent, and independent of the transaction and of the institution’s loan production function. They also pointed to the Valuation Independence Rule, which implements Dodd-Frank Act provisions for valuations of a consumer’s principal dwelling and bars covered persons from causing the value assigned to be based on anything other than the independent judgment of the preparer, through coercion, inducement, instruction or similar means.

For a lender, the independence question is structural. Who orders the evaluation, who can see the target value, who can ask for a revision, and who the preparer reports to. A community bank where a loan officer orders an AVM report and reads the result has a different independence profile from a bank where a separate collateral unit orders and reviews evaluations. The first can work if controls compensate. The guidance asks the bank to show that they do.

Pressure and compensation

The Valuation Independence Rule names compensation and instruction, as well as coercion and inducement, among the ways a covered person may not influence a valuation. For evaluations, the risk points are practical. A preparer paid more for faster turnaround, or a vendor whose contract renewal depends on loan officers’ satisfaction, faces incentives that can bend values. An AVM whose output can be rerun with different inputs until a target value appears is a pressure point of its own. Controls follow: preparer compensation that does not depend on the value or the loan closing, vendor scorecards owned by the collateral function rather than production, and logs that show every AVM run for a loan, including the ones that were not used.

Review

The 2019 rule added a review requirement for appraisals: from January 1, 2020, appraisals for federally related transactions had to be subject to appropriate review for compliance with USPAP. It did not add a parallel rule for evaluations. The Guidelines cover review of both appraisals and evaluations, and the agencies referred institutions to them. In effect, the regulation required review of the more standardized product and left the less standardized one to guidance.

A sound review program applies to both. For evaluations, review checks that the evaluation followed the method document, that its data and inspection were adequate for the property, that the value is supported, and that any model output was used within its limits. Review is also where a lender gathers data on evaluation quality: how often reviewers find problems, of what kind, and from which preparers or models.

Review depth can vary with risk. A model-only evaluation on a low loan-to-value loan may need only a check that the model was used within its conditions. An evaluation near the threshold, on a high loan-to-value loan or an unusual property, deserves a fuller review of data, comparables and condition. Setting those tiers in the review procedure lets a small review team spend its time where errors would cost most.

Vendors

The rule changed nothing about the lender’s responsibility when it buys evaluations or AVM results from a third party. If a vendor supplies evaluations, the lender’s method document has to cover what the vendor does, its contract has to give the lender the information needed to test quality, and its oversight has to include the vendor’s preparers and models. A vendor’s marketing claim of accuracy is not evidence of accuracy in the lender’s own markets.

Owners and evidence

The responsibilities divide naturally. The chief credit officer owns the selection policy and its exceptions. The collateral or appraisal function owns the evaluation method, the ordering process and vendor oversight. Whoever validates models owns the AVMs, including vendor models used in evaluations. Compliance owns the borrower copy process, the dispute route and fair lending testing. Internal audit tests the whole. The evidence an examiner would ask for follows the same lines: the policy, the method document and its version history, the testing results, review findings by preparer and model, dispute records, and the share of below-threshold loans that received evaluations rather than appraisals.

What the borrower sees

The threshold change also altered what borrowers receive. A borrower whose loan once required an appraisal might now receive an evaluation instead, with different content and less standardized form.

The copy requirement

The agencies answered consumer protection concerns in part by pointing to the Equal Credit Opportunity Act rules in Regulation B. Lenders must provide a copy of all appraisals and written valuations developed in connection with an application for a first-lien loan secured by a dwelling, and the agencies said that includes both appraisals and evaluations. A borrower who receives an evaluation therefore sees the lender’s valuation document. If that document is a bare model output with little explanation, the borrower sees that too.

Second liens and home equity lines

The copy requirement the agencies cited covers applications for first-lien loans secured by a dwelling. The appraisal threshold covers any residential real estate transaction secured by a 1-to-4 family property, which includes junior liens. A home equity line or second mortgage at or below $400,000 can rest on an evaluation, and the copy rule the agencies relied on does not reach it. A lender can choose to share its valuation with those borrowers anyway. If it does not, the borrower has less to dispute and the lender hears less about errors.

Disputes

Commenters raised the question of recourse when an evaluation is wrong. The agencies said information from entities that use or provide evaluations indicated that lenders often order appraisals when disputes arise with evaluations, so they did not expect the rule to materially affect consumer recourse. That puts the dispute path inside the lender’s process. A lender that uses evaluations needs a defined route for a borrower to challenge one, a rule for when a challenge leads to an appraisal, and a record of outcomes.

Disputes are also data. Each challenge that leads to an appraisal produces a second, independent value for the same property. Recording both values, and the reason for the challenge, adds to the testing sample at no extra cost. A rising share of disputes in an area, or for a vendor or model, is an early sign that the evaluation method needs attention there.

What a borrower can understand

An evaluation that is clear to a borrower is also clear to an examiner. Stating the value, the method, the data sources, whether the property was inspected, and the main comparable information in clear language costs little and reduces disputes. A lender that designs its evaluation report for the borrower as well as the credit file meets both audiences.

Fair lending exposure

A valuation method that is systematically less accurate in some neighborhoods than others creates fair lending risk, because a low value can mean a smaller loan, a higher rate or a denial. In 2019 no AVM rule existed, and the appraisal regulation said nothing about valuation accuracy across neighborhoods. Later federal actions did, as the hindsight section describes, including an AVM standard on compliance with nondiscrimination laws. A lender in 2019 could already have tested its evaluation accuracy by area and investigated gaps, and the same test serves both safety and soundness and fair lending.

Banks, credit unions and nonbank lenders

The rule applied to institutions regulated by the three agencies that issued it. Its effect varied with the lender’s charter, size and business model.

Large banks

Large bank holding companies already reported granular residential loan data to the Federal Reserve, and the agencies’ evidence on appraisal use came from those reports. For them, the rule mainly changed the boundary of an existing valuation program. Their task was to extend the program’s governance to the larger evaluation population and to show, with data, that selection followed risk.

Community banks and rural lenders

For smaller banks, the rule offered relief where appraisers are scarce. The agencies noted that commenters in their earlier regulatory review had said it can be difficult to find certified and licensed appraisers, particularly in rural areas, leading to delays and sometimes higher costs. The new threshold also made the separate rural exemption largely unnecessary for residential loans up to $400,000. That exemption required, among other conditions, that the property be in a rural area and that the mortgage originator or its agent contact at least three certified or licensed appraisers on its approved list in the market area and document that none was available within five business days beyond customary and reasonable fee and timeliness standards. A community bank could now skip those steps and obtain an evaluation. It also inherited the full burden of proving its evaluations were sound, possibly with a small staff and a heavy reliance on vendors.

A small bank does not need a large program to meet that burden. A defensible minimum has five parts: a short selection policy with clear appraisal triggers, a documented evaluation product from a vendor whose method the bank has read and accepted, a review of each evaluation by someone outside loan production, a small annual sample of appraisals on loans that received evaluations, and a log of borrower disputes and their outcomes. Each part produces a record. Together the records let a community bank answer the same questions an examiner would put to a large one, at a scale it can sustain.

Credit unions

The NCUA was not a party to the 2019 rule. Credit unions follow the NCUA’s own appraisal regulations, and in the AVM rulemaking years later, the agencies described credit unions’ alternative to an evaluation as a written estimate of market value. A credit union reading about the bank threshold had to check its own regulator’s rules rather than assume the same treatment.

Nonbank mortgage lenders

The rule applied only to institutions regulated by the three agencies. Nonbank mortgage companies were outside its scope, and their valuation requirements came mainly from the investors and insurers whose programs they used. That changed with the AVM rule, which applies to mortgage originators and secondary market issuers broadly, whether or not they are banks. A nonbank lender that used AVMs for principal dwellings came under the AVM quality control standards when that rule took effect.

Loans sold to the secondary market

For loans sold to the government-sponsored enterprises or insured or guaranteed by federal agencies, the bank threshold did not decide the valuation. Those programs set their own valuation requirements. A lender with both portfolio and secondary market lending therefore runs two valuation regimes side by side, and its systems need to know, at the time of valuation, which one a loan will fall under.

The strongest objections

The argument here is that the rule shifted the burden of proving collateral value onto lenders’ own evaluation programs. Five objections deserve an answer.

Evaluations were not new

Evaluations had been required below the threshold since the 1990s, and banks had been doing them for years. That is correct, and the agencies said so. The 2019 rule did not create the burden. It enlarged it, by moving a larger slice of residential lending, 16 percent of regulated transactions in the agencies’ own 2017 data, into the evaluation regime. A program that worked for small loans now had to work for larger ones, where valuation errors cost more.

Appraisals are not perfect either

A second objection is that appraisals have their own problems, and some commenters said models were more objective because they were not anchored to the contract price. That may be true for some loans. The argument here does not claim appraisals are better. It claims that an appraisal comes with an external standard and an evaluation does not, so the lender has to supply the standard itself. A lender that does so well may produce better valuations than an appraisal would. A lender that does not has nothing to show an examiner.

Most loans never relied on the bank threshold

A third objection is that the bank threshold mattered less than it seemed, because many residential loans are sold to the enterprises or insured by federal agencies, which set their own valuation rules. For those loans, the 2019 rule changed nothing. That is right, and it narrows the argument to portfolio lending and other loans that the threshold governs. The 16 percent figure in the agencies’ data applied to loans outside those programs, which include the loans whose collateral the bank keeps. Narrower is not smaller where the bank bears the loss.

Lenders kept ordering appraisals

A fourth objection uses the agencies’ own data: if large lenders obtained appraisals on 74 percent of eligible portfolio loans under the old threshold, perhaps they would keep doing so, and little would change. If that were true, the evaluation population would stay small. But the data described large lenders, under the old threshold, before the cost and time pressures the rule was meant to relieve. A lender that keeps ordering appraisals by policy has a simple story for examiners. A lender that uses the new room has to tell the longer one, with evidence, which is the point of this article.

Guidance was enough

A fifth objection is that the 2010 Guidelines and the 2016 advisory already told lenders what to do, so no new burden arose. Guidance sets expectations, but it does not specify methods, and it leaves each lender to decide how to meet them. That flexibility is useful. It also means the lender, not the guidance, owns the method and has to defend it. The AVM rule adopted later, described in the hindsight section, kept that same flexible approach.

Where the argument is weakest

The weakest point is evidence of harm. The agencies cited data that the 1994 increase did not materially change loss experience, and this article has no data showing that the 2019 increase did. The argument is about governance, not outcomes: a lender relying on evaluations has to be able to prove they were adequate, and the rule increased the number of loans for which that proof was needed.

With hindsight

The threshold held, and its first test came in the spring of 2020. In an interim final rule effective from April 17 through December 31, 2020, the same three agencies deferred the requirement to obtain an appraisal or evaluation for up to 120 days after closing for certain residential and commercial real estate transactions, excluding acquisition, development and construction loans. They said institutions should make best efforts to obtain a credible valuation before closing and otherwise underwrite loans consistent with the agencies’ Standards for Safety and Soundness and Real Estate Lending Standards. The agencies said the relief was meant to let institutions extend liquidity quickly to creditworthy households and businesses. It also meant that for some loans the valuation followed the credit decision instead of supporting it.

The agencies then turned to the questions the 2019 commenters had raised. On July 26, 2024, five agencies, including the NCUA and the CFPB, published final guidance on reconsiderations of value for residential real estate valuations. It described how institutions may build processes for a request to an appraiser or other preparer to reassess a valuation, including requests arising from a consumer, and how to address the risk of discrimination in valuations. The agencies stated that the guidance does not have the force and effect of law or impose new requirements.

On August 7, 2024, six agencies, including the NCUA, the CFPB and the Federal Housing Finance Agency, published the AVM quality control rule that Title XI had long required. Effective October 1, 2025, it requires mortgage originators and secondary market issuers to adopt policies, practices, procedures and control systems so that AVMs used in credit decisions or covered securitization determinations meet five standards: high confidence in estimates, protection against data manipulation, avoidance of conflicts of interest, random sample testing and reviews, and compliance with nondiscrimination laws. It covers AVMs used to value a consumer’s principal dwelling, and excludes portfolio monitoring, reviews of completed valuations and use by a certified or licensed appraiser. It sets no specific method, an approach most of its roughly 50 commenters supported. Its scope also reaches secondary market issuers’ determinations of whether to waive an appraisal requirement for a loan they may buy. The standards appear in the OCC’s regulations today, and the residential threshold there is still $400,000.

Automated valuation has since moved into policy goals. Executive Order 14438, signed October 22, 2026, directs the Federal Housing Finance Agency’s Director, if he determines it is in the best interests of the enterprises, the agency and the public, to have Fannie Mae and Freddie Mac establish a pilot that uses AVMs and other data to identify borrowers who appear to qualify for cancellation of mortgage insurance. In expanding a title acceptance program, the Director is also to consider waiving appraisals for limited cash-out refinances that reduce the borrower’s monthly payment, where the enterprise’s existing data support doing so.

What held up was the 2019 bet that evaluations could carry more of the load. What took years was the rulebook for the models inside them. Between October 2019 and October 2025, a bank relying on AVM-based evaluations for principal dwellings had to supply its own standard, and even now the AVM rule tells lenders what their controls must achieve, not how.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • appraisals
  • evaluations
  • AVM
  • Title XI
  • mortgage lending
  • valuation