Give me a break

[RCA]

Unfortunately the appraisal performed by that licensed professional more than likely will not be acceptable for lending purposes in the not so distant future, one appraiser being replaced instead by these hybrid assembly-line style valuations.

?? I'm not quite sure what you mean. Do you mean that it (AVMs) will likely be acceptable for lending purposes at some point in the near future? Without the kind of regulations that appraisers are subject to, the AVMs would be pretty much free to do what they want and you can be sure the lenders will start doing a lot of fishy stuff. That will not likely happen. They are going to want E&O, a signature, a license, someone they can really hold accountable.

However, instead of licensing just appraisers, they might consider licensing the companies that produce the AVMs, achieving the same purpose. In fact, they would probably have to consider licensing the actual software, which would be very interesting. It would be like the FDA coming in and doing these deadly QA audits that BioTech and pharmaceutical companies have to get through to get their drugs approved. That process is bloody murder, as anyone in biotech knows (my wife has worked for decades in biotech since she started with Genentech in the 1980's). And that would mean that every change to the software would have to be approved. And that would be a real bottleneck. How are you going to fix bugs and make small changes?

--- No, they probably will just license "Valuation Engineers" who will have to sign on off on the valuations and bridge any open gaps by whatever means. -- But, as I've said before on this forum, the current bottleneck on that type of thing is USPAP which doesn't have a clue about how mass appraisal works.

You are going to have to have a cadre of engineer level appraisers (valuation engineers) to manage AVM based appraisals in the future, because it is a damn complicated process - and you simply will never regulate it that well via pure automation; flexibility will be needed. If you could just get in a plane and fly over all of the housing developments in California, imagining the $X,+++,+++ values over the homes, just how many and varied they are individually and as neighborhoods and the vast amount of wealth they represent, to begin to understand the magnitude of the problem. [And you can maybe begin to understand how "clever" finance people can make a genuine fortune by engineering the loan process to take advantage of of home owners, buyers and the system the GSE's and TAF (et al.) have created.]

 

[George Hatch]

I never claimed that, which is why I always say appraisals provide a value AND other information about the property and the market.

What do you propose we "add" to appraisals?

And it is more than about " the number" - the appraisal value has the signature behind it of the appraiser who can support the number when challenged and do so for years - and who is the party behind the AVM estimate/ who gets to pick the value from the AVM range or defend the value ?? If nobody signs it , then nobody is responsible.

Well for one thing, the majority of appraisals I see, residential and commercial alike, pin their entire valuation of the limited number of sales and rental data in their comparisons, with only lip service being paid to long term trend analysis or macro analysis that would provide the context for the micro of their SC or IA. When done appropriately that works, but when done inappropriately it's more difficult for the reader to discern without going directly into the data themselves. You cannot see what isn't there.


It's one thing to cite some regional study about pricing trends year over year, and quite another to track a the long term trends of the physical neighborhood in which a subject fits vs its market segment. For example, if my subject is part of a subdivision of similar homes built by the same builder then that neighborhood will present a much tighter range for sizes and pricing than will a neighborhood comprised of a mix of ages and sizes. How my subject fits into the one neighborhood of similar can be quite different than how that same home fits into a more mixed neighborhood. As so on. It can help explain why I might have been compelled to go outside the immediate area to find my more direct comparables and it can provide the means to reliable develop and demonstrate any location adjustments (or the lack thereof) which may be applicable to those comparables.

These other summaries and analyses are not difficult to do but can add a tremendous amount of context for what we're doing in a Sales Comparison with your direct comps. When the micro is part of the macro that's demonstrating similar trends that's a much more convincing analysis that being wholly reliant on just the 5 sales which are taken out of the context in which they occurred.

 

[J Grant]

?? I'm not quite sure what you mean. Do you mean that it (AVMs) will likely be acceptable for lending purposes at some point in the near future? Without the kind of regulations that appraisers are subject to, the AVMs would be pretty much free to do what they want and you can be sure the lenders will start doing a lot of fishy stuff. That will not likely happen. They are going to want E&O, a signature, a license, someone they can really hold accountable.

However, instead of licensing just appraisers, they might consider licensing the companies that produce the AVMs, achieving the same purpose. In fact, they would probably have to consider licensing the actual software, which would be very interesting. It would be like the FDA coming in and doing these deadly QA audits that BioTech and pharmaceutical companies have to get through to get their drugs approved. That process is bloody murder, as anyone in biotech knows (my wife has worked for decades in biotech since she started with Genentech in the 1980's). And that would mean that every change to the software would have to be approved. And that would be a real bottleneck. How are you going to fix bugs and make small changes?

--- No, they probably will just license "Valuation Engineers" who will have to sign on off on the valuations and bridge any open gaps by whatever means. -- But, as I've said before on this forum, the current bottleneck on that type of thing is USPAP which doesn't have a clue about how mass appraisal works.

You are going to have to have a cadre of engineer level appraisers (valuation engineers) to manage AVM based appraisals in the future, because it is a damn complicated process - and you simply will never regulate it that well via pure automation; flexibility will be needed. If you could just get in a plane and fly over all of the housing developments in California, imagining the $X,+++,+++ values over the homes, just how many and varied they are individually and as neighborhoods and the vast amount of wealth they represent, to begin to understand the magnitude of the problem.

If they are going to do all that , why not just order an appraisal?

The software companies offer AVM products at low cost because they DO NOT take liability for results of each one (why would they), nor do they, or any company offering them cheap, have a personal signature/on them ( if they did it would approach a desktop appraisal )

If you regulate these alt products with a signature and liability of a analyst or software engineer, might as well be an appraisal then.... lol.....AVM's "value " property so very differently than an appraisal does using SCA with the highly qualified but limited in number comps.... an AVM by contrast might run more comps but their qualification is rote and thus the comps in any AVM can range anywhere from good, to mediocre, to terrible - the problem is impossible to predict which .

The appraisal is the product that models a buyer and seller behavior. I have not heard of a buyer choosing a house from an alogrithm.... though one day it might happen lol. The AVM has its uses to check appraisals and , appraisals also can act as a check on AVMs - a reason a waiver ( which uses an AVM ) per fannie needs an appraisal on file for the property done within X years ? to qualify for the waiver..

 

[J Grant]

Well for one thing, the majority of appraisals I see, residential and commercial alike, pin their entire valuation of the limited number of sales and rental data in their comparisons, with only lip service being paid to long term trend analysis or macro analysis that would provide the context for the micro of their SC or IA. When done appropriately that works, but when done inappropriately it's more difficult for the reader to discern without going directly into the data themselves. You cannot see what isn't there.


It's one thing to cite some regional study about pricing trends year over year, and quite another to track a the long term trends of the physical neighborhood in which a subject fits vs its market segment. For example, if my subject is part of a subdivision of similar homes built by the same builder then that neighborhood will present a much tighter range for sizes and pricing than will a neighborhood comprised of a mix of ages and sizes. How my subject fits into the one neighborhood of similar can be quite different than how that same home fits into a more mixed neighborhood. As so on. It can help explain why I might have been compelled to go outside the immediate area to find my more direct comparables and it can provide the means to reliable develop and demonstrate any location adjustments (or the lack thereof) which may be applicable to those comparables.

These other summaries and analyses are not difficult to do but can add a tremendous amount of context for what we're doing in a Sales Comparison with your direct comps. When the micro is part of the macro that's demonstrating similar trends that's a much more convincing analysis that being wholly reliant on just the 5 sales which are taken out of the context in which they occurred.

Good lord George, stop acting like the limited number of sales is a problem !! I cant stand it when people start moving the goal posts for an appraisal and thus declare an appraisal deficient .

The reason, as you hopefully know, that the comps are limited to 3-5 or so ( I do min 4 sold and a listing for 5), - reason the number is limited is because the comps represent the most competitive substitutes to a buyer for the subject property and often those subsitutes of truly comparable properties are limited in number. Appraisers usually do talk about wider searches and other sales in area not used. I spend hours and sometimes days paring down the sales to that small number of comps

As far as other studies, regional trends ...they are done to some degree in an appraisal but a lot of them are filler. To do them in greater depth, well I dont see most clients willing to pay for it on lender side or frankly in most cases what use it would be. Lenders and clients have lots of online tools now and they can run that kind of thing for themselves ( or expand SOW and fee to add them in appraisal )

 

[J Grant]

BTW I see lots of this macro/micro from RE agents and it is a lot of bull** most times and frankly how much different is it from an appraiser running a similar program - some of the RE agents (I appreciate them trying) hand me a CMA in a thick binder - full of charts, statistics, graphs, of price trends and market activity, .of the county, the town, the subdivision, going back five years and so on, -- Then, they put in their "comps" which are usually TERRIBLE, and all that other stuff turns out to be drivel, even if impressive in volume...too much information much of it not relevant or distracting.. .

One only needs to track if relevant subject area is stable, rising, or increasing, explain why, where your subject niche of properties fit in the wider picture. Most competent appraisers are already doing that. And a huge problem not allowing a deeper dive is the tight turn times, down to 48 hours now from many clients from inspection, where is an appraiser going to find the time to develop the kind of analysis you reference it in more depth, let alone a set of res lender clients with volume work who would pay for it?

 

[George Hatch]

I'm just suggesting that some of the value conclusions appraisers come up with would probably be different if they went from macro to micro with their analyses instead of going directly to a boilerplated and meaningless neighborhood description followed by the limited number of direct comparables.

For instance, in the Marin county appraisal where the borrowers are alleging racism, I can guarantee you that if I added a really good analysis of the subject neighborhood and compared that sales history to the other neighborhoods where the appraisers would be getting their comps that the chances of distortion from one of two sales gets almost completely mitigated. No matter what the subject is still part of its neighborhood and it still fits in there somewhere, even if it's overbuilt for that neighborhood (which isn't the case for the property in question).

And yes, I'm familiar with the time and fee constraints and so forth; just as I'm familiar with brokers using price as their sole filters in their analyses. That doesn't alter the point that if we're only talking about appraisals in terms of a value conclusion then -as far as that percentage of assignments the machine can return a usable result - we cannot compete with the machine in terms of time and cost.

The machine will continue to gobble up the easiest and most profitable assignments and appraisers will increasingly get stuck with the assignments that the machine cannot do - by definition more difficult and time consuming to do. The trend in that direction is never going to change. We will continue to get squeezed as the machine's effective capabilities increase (which is inclusive of them getting access to better quality data).

 

[djd09]

future?


323.3 Appraisals required; transactions requiring a State certified or licensed appraiser.


(a) Appraisals required. An appraisal performed by a State certified or licensed appraiser is required for all real estate-related financial transactions except those in which:

(1) The transaction is a residential real estate transaction that has a transaction value of $400,000 or less;

https://www.FDIC.gov/regulations/laws/rules/2000-4300.html

 

[djd09]

I think good ole Joan dumped her certification because she was tired of taking USPAPs two year continuing “education “ course.

https://www.FDIC.gov/regulations/laws/federal/2018/2018-real-estate-appraisals-3064-ae87-c-019.pdf

here is a good laugh. AVmetrics. oh they are an IAC member.

 

[J Grant]

George, it is hypocritical for anyone to complain about boiler plate with comps dumped in when AMC and clients choose those appraisers doing that because they offer low fees and fast turn times. Good appraisers dont do that and there are plenty of good appraisers doing residential lender work, commenting with context , but the preference of hiring does not see them used in cases where an AMC or big box lender accepts the boiler plate of competition appraisers who are faster and cheaper.

IDk what future holds but it is easy to see when those in a position of influence undermine appraisals for the fact that an appraisal based on SC is , by design, focused on a limited number of highly qualified comps ) , then of course they will use that very thing an appraisal is as the reason not to use appraisals.

I personally believe the ideal product is a blend of what big data offers and what a traditional appraisal offers in one product, but no matter what, a good product takes some kind of quality time and fee, so if the powers that be want cheap cheap and fast fast they can not have that increased quality or blend of the best of both, if AVM's serve their purpose with a robo inspector collection of "data" at a site then so be it, it has not been done on mass scale in res lending so outcome unknown.

 

[DWiley]

Good lord George, stop acting like the limited number of sales is a problem !! I cant stand it when people start moving the goal posts for an appraisal and thus declare an appraisal deficient .

The reason, as you hopefully know, that the comps are limited to 3-5 or so ( I do min 4 sold and a listing for 5), - reason the number is limited is because the comps represent the most competitive substitutes to a buyer for the subject property and often those subsitutes of truly comparable properties are limited in number.

Actually, the reason for three comps was primarily because that is how many the forms designers could squeeze across a piece of paper 8 1/2 inches wide. I am 100% not kidding