My prediction on 3.6

[George Hatch]

Gladwell said it takes 10,000 hours of practice to be skill in what you do. I passed that years ago.
You still need couple more thousands of hours to be competent.

You're squabbling over someone else's competency?

 

[RCA]

Help me understand this point of view, as UAD is solely a reporting format. It has nothing to with the STD 1 part of the appraisal process. What exactly do you think UAD 3.6 changes about the development part of an appraisal assignment?

UAD is not "just" a reporting format.

Any sales grid without columns for "value contributions" is a defamation of the concept of appraisal, IMO.

In any case, my appraisals have a column for "value contributions" in addition to "value adjustments' - and UAD does not allow that.

That is just one restriction. There are numerous others, such as dropdowns with a discrete set of values that may not be sufficient for a valid appraisal.

Don't you understand that? How is that even possible? Simple logic. UAD 3.6 is NOT just a reporting format. It is a TRAVESTY of the sales grid concept.

TRAVESTY: "A false, absurd, or distorted representation of something."

 

[OSU Beavers]

UAD is not "just" a reporting format.

Any sales grid without columns for "value contributions" is a defamation of the concept of appraisal, IMO.

In any case, my appraisals have a column for "value contributions" in addition to "value adjustments' - and UAD does not allow that.

That is just one restriction. There are numerous others, such as dropdowns with a discrete set of values that may not be sufficient for a valid appraisal.

Don't you understand that? How is that even possible? Simple logic. UAD 3.6 is NOT just a reporting format. It is a TRAVESTY of the sales grid concept.

TRAVESTY: "A false, absurd, or distorted representation of something."

I believe the word you are looking for is


Abomination

 

[RCA]

I just had another one of those conversations with Claude Fable 5:

Me to Claude: "If the new Fannie Mae UAD 3.6 form blocks appraisers from doing what they consider a valid appraisal, is that legal?"

Claude Fable 5: "Short answer: almost certainly yes, it's legal — because of who Fannie Mae is and how the legal duties are actually distributed. But the question has an interesting structure worth unpacking.

The GSEs are buyers setting purchase conditions, not regulators. UAD 3.6 isn't a law or regulation. Fannie and Freddie are (nominally private) corporations in FHFA conservatorship, and the UAD is a condition of sale: lenders who want to sell loans to the GSEs must deliver appraisals in the prescribed format. The mandatory sunset of legacy form submissions is November 2, 2026, but "mandatory" means mandatory for delivery to the GSEs — nothing legally compels an appraiser to accept GSE-destined assignments at all. A buyer of loans can lawfully specify what data package it will accept, just as it can specify LTV limits or ADU eligibility rules. Courts have consistently declined to treat GSE Selling Guide requirements as government action subject to constitutional or APA challenge, even under conservatorship.

The legal obligation that actually binds the appraiser runs the other way. Under FIRREA Title XI, state licensing law incorporates USPAP, and USPAP puts the burden squarely on the appraiser: if an assignment condition would prevent development or reporting of credible results, the appraiser must either modify the scope of work, supplement the report, or decline the assignment. So the legally coherent framing isn't "the form blocks a valid appraisal, is that legal?" but rather "if the form blocks a valid appraisal, the appraiser is legally prohibited from signing it." The GSEs know this and have built the escape valve into the spec itself: appraisers remain responsible for USPAP compliance and must supplement the information in any report form — via text addenda, exhibits, maps, sketches, or whatever else is necessary to produce a credible and understandable report. Notably, 3.6's dynamic structure was pitched partly as fixing this problem — Fannie's own newsletter concedes that the old UAD 2.6 framework limited appraisers to two Location/View factors chosen from a small enumerated list, entered as cryptic abbreviations, and "limits an appraiser's ability to communicate clearly".

Where a real legal issue could arise is narrower than form design. Dodd-Frank's appraiser independence provisions (15 U.S.C. §1639e) and TILA's AIR rules prohibit coercing, influencing, or encouraging an appraiser toward a targeted value — but they protect against pressure on the conclusion, not against constraints on the reporting container. A structured-data schema that forces enumerated responses doesn't violate AIR so long as commentary fields and the appraiser's ability to reconcile remain available. If a lender or AMC used a 3.6 validation hard-stop to pressure an appraiser into changing a supported conclusion (rather than fixing a data-format error), that could implicate independence rules — the misconduct would belong to the lender/AMC, not to the form.

There's also a practical failure mode you'd appreciate from the compliance-engine side: the UAD Compliance Rules include severity levels, and a fatal-severity rule that mechanically rejects an unusual-but-defensible reporting choice functions as a de facto scope restriction even though it's formally just data validation. The appraiser's recourse there is commentary/override paths where the spec provides them, and declining where it doesn't — economically painful, legally clean.

Usual caveat: I'm not a lawyer and this is the analytical lay of the land, not legal advice. But the core answer is that "the form prevents a valid appraisal" is a claim USPAP resolves against signing, not a claim that renders the GSE requirement unlawful — the GSEs sit on the demand side of the market, and the law regulates the appraiser's signature, not their spreadsheet schema."

(to be continued)

 

[RCA]

I just had another one of those conversations with Claude Fable 5:

Me to Claude: "If the new Fannie Mae UAD 3.6 form blocks appraisers from doing what they consider a valid appraisal, is that legal?"

Claude Fable 5: "Short answer: almost certainly yes, it's legal — because of who Fannie Mae is and how the legal duties are actually distributed. But the question has an interesting structure worth unpacking.

The GSEs are buyers setting purchase conditions, not regulators. UAD 3.6 isn't a law or regulation. Fannie and Freddie are (nominally private) corporations in FHFA conservatorship, and the UAD is a condition of sale: lenders who want to sell loans to the GSEs must deliver appraisals in the prescribed format. The mandatory sunset of legacy form submissions is November 2, 2026, but "mandatory" means mandatory for delivery to the GSEs — nothing legally compels an appraiser to accept GSE-destined assignments at all. A buyer of loans can lawfully specify what data package it will accept, just as it can specify LTV limits or ADU eligibility rules. Courts have consistently declined to treat GSE Selling Guide requirements as government action subject to constitutional or APA challenge, even under conservatorship.

The legal obligation that actually binds the appraiser runs the other way. Under FIRREA Title XI, state licensing law incorporates USPAP, and USPAP puts the burden squarely on the appraiser: if an assignment condition would prevent development or reporting of credible results, the appraiser must either modify the scope of work, supplement the report, or decline the assignment. So the legally coherent framing isn't "the form blocks a valid appraisal, is that legal?" but rather "if the form blocks a valid appraisal, the appraiser is legally prohibited from signing it." The GSEs know this and have built the escape valve into the spec itself: appraisers remain responsible for USPAP compliance and must supplement the information in any report form — via text addenda, exhibits, maps, sketches, or whatever else is necessary to produce a credible and understandable report. Notably, 3.6's dynamic structure was pitched partly as fixing this problem — Fannie's own newsletter concedes that the old UAD 2.6 framework limited appraisers to two Location/View factors chosen from a small enumerated list, entered as cryptic abbreviations, and "limits an appraiser's ability to communicate clearly".

Where a real legal issue could arise is narrower than form design. Dodd-Frank's appraiser independence provisions (15 U.S.C. §1639e) and TILA's AIR rules prohibit coercing, influencing, or encouraging an appraiser toward a targeted value — but they protect against pressure on the conclusion, not against constraints on the reporting container. A structured-data schema that forces enumerated responses doesn't violate AIR so long as commentary fields and the appraiser's ability to reconcile remain available. If a lender or AMC used a 3.6 validation hard-stop to pressure an appraiser into changing a supported conclusion (rather than fixing a data-format error), that could implicate independence rules — the misconduct would belong to the lender/AMC, not to the form.

There's also a practical failure mode you'd appreciate from the compliance-engine side: the UAD Compliance Rules include severity levels, and a fatal-severity rule that mechanically rejects an unusual-but-defensible reporting choice functions as a de facto scope restriction even though it's formally just data validation. The appraiser's recourse there is commentary/override paths where the spec provides them, and declining where it doesn't — economically painful, legally clean.

Usual caveat: I'm not a lawyer and this is the analytical lay of the land, not legal advice. But the core answer is that "the form prevents a valid appraisal" is a claim USPAP resolves against signing, not a claim that renders the GSE requirement unlawful — the GSEs sit on the demand side of the market, and the law regulates the appraiser's signature, not their spreadsheet schema."

(to be continued)

Me: "So, for the appraiser, the general method around the madness is to do the report, as he sees fit in commentary and then extract the parts that fit into the UAD 3.6 as best possible, referring the user to the commentary for full explanation."


Claude: "Yes — with one refinement in emphasis: it's not so much a workaround as the legally correct order of operations, and 3.6 arguably formalizes it rather than fighting it.

The appraiser's USPAP obligations run: develop credible results per Standard 1 using whatever methodology the problem demands, then report per Standard 2 in a way that isn't misleading to intended users. The UAD schema only ever touches the second step. So "do the full analysis as you see fit, then project it into the schema" isn't a hack — it's the only sequence that was ever compliant. The appraiser who works the other way around, letting the form's fields define his analysis, is the one on thin ice. The schema is a lossy serialization format for the assignment results, and the commentary plus exhibits are where the loss is documented and recovered.

Three constraints keep the method honest rather than mad:

The projected values must be independently true. Whatever lands in a structured field has to be accurate and supportable as stated, not merely "the closest available approximation, see commentary." If a field genuinely cannot be filled truthfully — not awkwardly, but truthfully — that's the decline-the-assignment trigger, and commentary can't cure it.

Grid and commentary must reconcile, not compete. A report where the structured data says one thing and the narrative says "actually, ignore that, here's the real analysis" is misleading in exactly the way USPAP prohibits. The commentary's job is to show the grid is a faithful summary of something richer — your RCA-deltas case — never to contradict it.

The referral has to be findable. Under 2.6 the honest appraiser's full explanation lived in a back addendum nobody read. 3.6's section-level commentary means the pointer to the fuller analysis sits next to the data it explains, which is the one part of the redesign that works for this method rather than against it.

Framed that way, the appraiser isn't routing around the form — he's treating it as what it legally is: the GSE's required data extract of a report whose scope and content he, not the schema, is responsible for. The madness only enters when anyone (appraiser, reviewer, or underwriting engine) mistakes the extract for the report."

(to be continued)

 

[RCA]

Me: "So, for the appraiser, the general method around the madness is to do the report, as he sees fit in commentary and then extract the parts that fit into the UAD 3.6 as best possible, referring the user to the commentary for full explanation."


Claude: "Yes — with one refinement in emphasis: it's not so much a workaround as the legally correct order of operations, and 3.6 arguably formalizes it rather than fighting it.

...

(to be continued)

So, in summary, the UAD 3.6 is considered a legitimate extract of whatever the appraiser decides to put in the commentary, if that is what he wishes to do!

My earthUI program actually produces a grid - and from that I can make a UAD extract - which is always what I have done. And for the dropdowns, I just consider them required, useless (for me) commentary that they require and that has no impact on the value conclusion.

But, in the end, for me, the appraiser, the UAD 3.6 is something not required for a valid appraisal; only because some client insists on it for GSE compliance.

Another way to put this is that the UAD form is often not sufficient for a valid appraisal without a possibly much much larger commentary that includes the real and valid appraisal report. The UAD 3.6 is an extract that "floats on the surface" of the real appraisal report.

 

[RCA]

Me to Claude: " The UAD 3.6 can be considered an extract that floats on the surface of the real and valid appraisal report, via the commentary."

Claude: "That's a good formulation, and it's worth noticing it inverts the official ontology. The GSEs' documents treat the UAD payload as the report with commentary as supplementation; your framing treats the developed analysis as the report, with the UAD as a buoyant extract — structured, machine-readable, floating on top, tethered to the depths by commentary. For any appraiser doing real methodology, your version is the epistemically accurate one.

The metaphor earns its keep because it gets the dependency direction right. Surface phenomena are determined by what's beneath — the extract is valid only insofar as it's generated from the analysis, and the tether has to hold: every structured value must be derivable from the underlying work, and the commentary is the visible line showing the derivation. An extract that floats free — grid values that don't reconcile to any documented analysis — is exactly what a review appraiser or a state board would call unsupported. Most weak appraisals are, in this picture, all surface and no depth: the form filled directly, nothing underneath.

Two caveats keep the metaphor honest. First, the surface is where all the automated readers live — the UCDP rules, Collateral Underwriter, the lender's engines all consume only the extract, and their verdicts have real consequences regardless of what's below. So the surface, while ontologically derivative, is operationally primary for everyone but the humans. The appraiser's craft under 3.6 is making the projection to the surface as informative as the schema permits, not treating it as a chore. Second, USPAP won't let the depth be private. The workfile plus report together must let a reader reconstruct the reasoning — the deep report has to be in the delivered document (exhibits, section commentary), not merely in the appraiser's files with the extract floating above an undisclosed ocean.

There's also a clean software-architecture reading you've essentially already built ValEngr around: the analysis is the source of truth, and the UAD XML is a view — a materialized, schema-constrained projection, regenerable and verifiable against the model. Views can't contain what the source lacks, and a source that only exists to populate its view isn't a source at all. The appraiser who understands the form that way is compliant almost by construction; the one who mistakes the view for the database is the one the review process is designed to catch."

 

[Non Sequitur]

My prediction as an appraiser.... use AI or perish and die....

Yep, 100%.

Is it possible to complete 3.6 reports without a.i. assistance? Sure, and it takes much longer and is just flat out more difficult by several measures without using a.i.. But "is it possible without a.i." isn't the issue, the issue is exactly what you posted "use a.i. or perish and die." Fees aren't going up to cover the additional time and effort, not to mention the ever increasing software and portal costs. So either use a.i. to level up, or as you posted eventually fall behind and "perish and die" as an independent residential appraiser.

Either way, I'm just a guy behind a keyboard trying to find his way in this new world too. Screw the classes taught by people who haven't been in the field for decades (or ever) and will never have to make a living accepting 3.6 assignments. Start test driving the clunky software now. Start on your own home, then do it again, and again, from scratch each time. If after a few times you think 3.6 is no big deal, like some in the echo chamber want you to believe, then you're good to go. If not, make a decision that's best for you.

 

[Non Sequitur]

Skip to the 2 minute mark. An AMC chief appraiser talking about 3.6 fees:

 

[chad hampton]

What planet does that woman live on? I haven’t had AMC ask me for a fee in quite a few years. These podcasts are painful to sit through.