Channeling Deep Blue Versus Garry Kasparov in Home Valuations – It’s time to standardize how real estate appraisers make their adjustments
Source: Corelogic blog
Excerpts:
It might be time to reengineer the process appraisers use to make adjustments to comparable homes. The current approach does not appear to be defensible.
To investigate this issue, we obtained a set of relocation appraisals from an appraisal management company and conducted our own analysis. (As background, for relocation deals, two or more appraisals are ordered at the same time; and the two appraisers often choose an identical comparable property).
For the experiment, we:
1) Obtained a set of relocation appraisals—two appraisals on each property prepared at roughly the same date for which the appraisers choose an identical comparable.
2) Looked at the adjustments made to that comparable and noted discrepancies in the amount of the adjustments, and even the direction (positive or negative).
It’s time that we re-engineer the process so that appraisers can focus on the things that may benefit from the human element, such as defining the neighborhood and selecting comparables. But for the pieces of the puzzle that need to be standardized, like adjustments to comparables, we should be harnessing the power of our machines. Too much is at stake to maintain the status quo. Appraisers would benefit by being allowed to focus on things they do well; loan applicants would benefit from a less-costly, streamlined process; and lenders would benefit from valuations that were standardized and more reliable.
Read the full article and see how they analyzed the data, with graphs etc. at:
http://www.corelogic.com/blog/authors/michael-g.-bradley/2014/10/channeling-deep-blue-versus-garry-kasparov-in-home-valuations.aspx?WT.mc_id=crlg_131016_0OGMz#.VFERojTF8zr
My comments: of course, it would be very helpful if appraisers had access to all the Big Data being compiled by Fannie and Others. Fannie will be looking at appraiser adjustments but don’t offer any help from their Big Data to appraisers making the adjustments.
This was emailed to me by Ron Napier. I am posting it, with his permission, so everyone can read it!!
In the description of how they compared adjustments,
they lack any explanation or whether they had the foresight,
“To compare the quality and experience level of the two appraisers”.
Perhaps they had a well experienced appraiser and another with limited experience.
It has usually been discussed in the past, that two “good” appraisers
will be less than 5% apart in values, and many times 2-3% apart.
The difference being the judgement applied to various amenities and their contribution.
I can recall lenders saying, we had 2 reports, and you fellows were only a few thousand
dollars apart. (and they seem amazed)
We have noticed this in our Metro area, ” years ago”. Of course, with everyone
pressing Federal agencies to include Licensed, not certified, and Delete the
Apprentice/training section of the rules in the early 90s, things were bound
to change.
Just like with Assessors, we probably have a lot of people with
limited training and bad habits, teaching other people the same bad habits.
They should pull out 2 relocation appraisals from the 1980s and compare.
I agree with the author. I’ve seen this sort of thing in reviews myself. I think we should be using multiple regression to come up with our adjustments in situations where enough data exists. It’s not hard to do, takes only a few minutes, and there is appraiser-friendly software out there for it. The main obstacle is that lending UWs typically do not understand it and will not accept it mainly because it sometimes produces adjustments that seem counterintuitive. Government agency staff reviewers will also push back. Fannie needs to take the lead in encouraging its use.
All we have to do is hold the various AVM solution models accountable to the same standards that appraisers are. So let the AVMs ‘produce’ their values and they’d be put out of business in one week after numerous appraisals of theirs were deemed unreliable. All of these alternative methods of valuation rely upon that the fact that they’ll never be held accountable for their lack of reliability. I’ve always said that you can read a large city newspaper and find a property that just burned down- and then run an AVM on that property to see what the value is. That’ll tell you how good their verification methods are.
I’ve often wondered about the validity and reliability of appraisals. Validity: are we measuring what we purport to measure? Reliability: would multiple appraisals of the same property tell us the same thing? CoreLogic has tested the reliability and found it lacking. As for validity, appraisers and clients can’t even decide what it is they are trying to measure. Probable price or intrinsic value? Today’s value or the value three years from now, when the homeowners are divorced or bankrupt due to job loss or health costs? (That, of course, is what clients REALLY want.) The residential appraisal business is doomed, and I just hope the commercial business lasts until my home is paid off in 10 years.
Corelogics article is nothing more than another marketing ploy to continue with a domination that allows end users to pay a minimal fee for a “push button” answer. This gives said user the FALSE perception of accuracy and accountability. Simply put – If the lender actually reads the limiting conditions of Corelogics responsibility if the loan goes bad and the valuation is found to have been incorrect Corelogic has ZERO financial accountability to the user who paid the mere $50+/- fee. However, appraisers do have the accountability and liability if they are found to have made significant error. “Paired data analysis” is a term they may have to look into (see the Appraisal Institutes 4th edition dictionary page 207). Unfortunately a vast majority of the appraisers in the industry do not use this technique largely because they feel they are not paid well enough to utilize this (many don’t know how) and various other techniques that are truly the basis for accurate valuations. With that said – Appraisers not doing appraisals with these techniques are the main crux of the defense for replacement by data bases. If appraisers accept the assignment, regardless of fee you are required to complete it correctly. At the current fee’s most appraisers will not put in the effort even if it is wrong to not use these techniques (not saying that is correct but its a reality). Moving full circle – the fee’s remain low partially because of supposed reliable data bases provided by Corelogic and other providers giving a false impression of accuracy, reliability and faith in a large company. The speed and false sense of security of the data bases will not go away because it is profitable to get loan applicants approved right away. Not only that but hey – The fancy charts look really cool and users can’t see the pig under the pretty dress. When the dress comes off though the users will then again seek those appraisers who can PROPERLY see under the dress PRIOR to the dance – not after (Can you say no doc loans???). I will end the rant as I could go on for hours. There is a place for both appraisers and data bases BUT today’s business model and society prefer quick and fast over any other option – Good luck appraisers your going to need it!!
The only obstacle would be adjustment variations for hundreds of thousands of neighborhoods & local market areas. You just cannot have one standardized figure per item for the entire country, state or even a county. Adjustmet figures are sometimes different within a zip code.
The only onbstacle would be adjustment variations for hundreds of thousands of neighborhoods& local market areas. You just cannot have one standardized figure per iteem for the entire country, state or even a county. Adjustmet figures are sometimes different within a zip code.
As professional appraisers we are called upon to make logical sense of price differentials in comparable sales when the Realtor who priced the home had absolutely no basis for pricing the home in the first place. The truth is that Realtors do not value homes the way that appraisers do. Neither to buyers who rely on the advise that they get from their Realtor.
In my experience, Realtors typically look at a price per square foot range for an area and price the home somewhere in that range without analyzing and valuing the individual components that contribute to the appeal, marketability, and value of the home. How much does the view contribute? How much does the remodeling contribute? How much does the 3 car garage contribute?
Professional appraisers are trained how to value the individual components using comparable sale data, cost data, and rental data. Honestly, most Realtors don’t have a clue. Sellers and buyers rely on their Realtors for advise on pricing a property and for making a reasonable offer on a property. The simple truth is that the majority of Realtors do not know how to value a property in the same way that a professional appraiser does. In the end, we appraisers are called in to make logical sense of a transaction that may or may not have a sale price based upon factual data. And when the property does not have pricing based upon an a detailed analysis of the market data, who gets blamed? The professional appraiser, of course!
IMHO, all the UAD did was expand the addendum. Everything that the UAD tries to classify as the same is explained in the addendum why they are not the same. Unfortunately, the addendum is never included in Corelogics data and therefore Corelogics data is incomplete and misleading. They should not be advising on adjustments based on the incomplete data they collect.
If I could select qualified market data and use an algorithm market data to derive adjustments that would be great. I’d understand the data, the market and the adjustments but to have someone impose them on me? No thanks.
JT, Q3 and Q4, and all the other Q’s and C’ are shackles that keep good appraisers from doing their job properly and diminish and degrade the quality of the appraisal. Prior to UAD, appraisers were able to be more precise and concise with The Quality of Construction and Condition adjustments. Sometimes adjustments fall somewhere in between these categories and one appraiser may lean on the conservative side with a high Q3 while another will be generous with a low Q4 rating. For Quality of Construction, I used to make adjustments for difference Elevations because the market reflected a premium for a brick or stone front but, a C2 versus a C3 rating is not appropriate because it rates the entirety of the construction not just one facet. One C3 with a stone front is worth more than a C3 without a stone front. I could go on and on about the flaws of UAD but, who cares if the quality of an appraisal is downgraded as long as lending industry gets the standardization they need to devalue and degrade the appraisal profession.
Of course, Core Logic would take this viewpoint… they are in the business of selling data. As with all AVM pushers, they get confused and don’t understand the difference between “price” (a fact) and “value” (an opinion). Even the relo definition of Anticipated Sale Price is a value opinion when the appraiser renders it; the eventual, actual sale price is how the appraiser is rated for these jobs. If you don’t get really close most of the time, you won’t be doing relos anymore. James had it right, it makes little or no difference if two different appraisers used slightly different methods to arrive at the same place.
Relocation appraisers are forecasting value 60 – 120 days in the future. They are primarily judged on how close they came to the actual sales price when the home sells. If they are over 5% more than occasionally they will lose their client. The fact that both appraisers took a different road to get to the same place does not trouble me. If they were not accurate overall they would have been terminated. What does trouble me is the perception that many believe you can provide appraisals simply using mathematical models. You can’t. There are far too many variables. For example, one home has a large newer deck and one has a smaller older deck. According to the mathematical models I have seen they have the same adjustment. But do they? In addition the models will never have enough accurate unbiased information to draw from. In my area its not unusual for some agents to list “updated kitchen” when all the owner did was paint the cabinets. Appraising is both an art and a science. Accurate appraising always will be.
I do not do REO appraisals simply because I am not a prophet; as the saying goes: “past performance does not guarantee future results”. Are appraisers opening themselves to potential liability if their forecast ends up being wrong?
: I guess if you average fannie and mba, you get a 1% decline – not much change!!
Just supports the theory that Big Data is also questionable……..
Adjustments are the perception of the Interpreter and therefore, Independent and Unbiased. You cant have “canned” adjustments in different markets, similar to Templates, someone will always forget change the information to reflect the current subject property & applicable language.
Was it Q3 or Q4 ?
Jim makes the point as to why, no matter how they try, eliminating the appraiser isn’t going to be beneficial to anyone except Realtors and originators, and that is what these concepts are trying to move towards. If someone wants to help the consumer and make mortgage loans less costly and more reliable, they should move away from commission based compensation for anyone involved with mortgages or purchases and that includes Realtors, mortgage brokers, loan originators etc. Standardizing the compensation would be THE start, something like $500 or less (give them an RMC to work with, Realtor Management Company, OMC Originator Management Company).
It is interesting how they selected perhaps one of the most ambiguous factors of site for scrutiny. The other focus of age adjustment from one when the other made a condition adjustment, isn’t necessarily wrong or incorrect, perhaps you shouldn’t apply both but one or the other could be adjusted based on the same observation. Either way they should have been addressed in the report and THAT is what should support or justify the application regardless of where they put it.
It’s not so much having identical adjustment values, if that is what they are looking for, the amount of adjustment applied should be reasonable and rational, and that an adjustment SHOULD be applied to a factor/amenity.
‘Defensible’ is this new focus of how to eliminate the appraiser. Keep them (appraisers) in court and they won’t be able to do any appraisals. Perhaps they should be trying to defend these insane commissions that are forced on the consumer for pushing paper, data entry, checking credit ratings, making false promises, sales. It may be the consumers best protection is coming from appraisers, home inspectors, termite inspectors, engineers etc. and not the ones focusing on ‘defensible’.
Standardized valuations are NOT more reliable; valuations are more reliable when competent appraisers make adjustments based on the market value of the amenity and no given amenity has a standard value from one market to another. A fireplace in a $250 – 300 K market area will have a different value in a $1,000-2,000 K market area.