Fannie Mae “CU” Scoring is a Danger for Appraisers
Many know by now that the GSE’s, primarily Fannie Mae, have instituted a new ‘appraisal scoring’ procedure based on an electronic read of your reports, specifically on a SFR 1004 or the Condo 1073. Those are the only forms currently being analyzed by the CU process.
On Nov. 18, 2014, FNMA released a document named “UCDP Fannie Mae Appraisal Messaging Change Notification” which you can find here.
I encourage all appraisers to actually read this document … all 11 pages.
When you do read this document, you will learn that your reports are being compared to your peer’s reports, and to your other reports, and to some unidentified ‘model’ FNMA uses.
Some of the ‘checks’ being performed by the CU process include these:
- The reported GLA is materially different than what has been reported by other appraisers.
- The reported lot size is materially different than what has been reported by other appraisers.
- The condition rating is significantly different than what has been reported by any other appraiser.
- The quality rating is significantly different than what has been reported by any other appraiser.
Here are a few that can cause real concern among appraisers:
The GLA adjustment is larger than peer and model adjustments.
The GLA adjustment is smaller than peer and model adjustments.
The view adjustment is materially different from peer and model adjustments.
And I just love this one:
The appraiser-provided comparables are materially different than the model-selected comparables.
It’s time for appraisers to get serious about meeting your peers in person, compare notes, and develop a regional adjustment chart for all variables, much like that yellow legal pad paper you were handed when you got in this business…that paper with the ‘required’ adjustment amounts on it for almost all items.
Oh…and when you get that knuckle slapping letter from FNMA saying your adjustments or comparables don’t match the ‘model’ be sure to get the specifics and pass on ‘model info’ to your peers.
Yep, appraising real property and developing an opinion of value is actually a science with absolutely no variability. Bet you didn’t know that!
- Speed Regardless of Accuracy Under the Banner of Modernization - March 8, 2023
- Marin City Discrimination Case Settled - March 7, 2023
- Exactly How Are Property Data Collectors Professionally Trained? - March 3, 2023
The appraiser is the new whipping boy in the real estate profession. The only one who is unbiased and who gets paid the least amount must be made to pay (tongue in cheek if you didn’t get the sarcasm) for their wrong doings. The powers that be are trying to fit the square peg that is the variability of appraising into the round hole of check list underwriting. Thanks for the informative post Dave.
Good advice Dave; and thank you for pointing out the particularly worrisome aspects of it. For my part I am going to ADD A SCOPE AND USE LIMITATION in the body of the text on all my reports:
“The appraiser recognizes the necessity of addressing reasonable underwriter questions; and/or responding to those issues that may arise as a result of a USPAP compliant field review appraisal. The appraiser will not however be obligated to respond to non compliant desk reviews or AVM comparisons; or any other algorithmic derived ‘scoring’ systems that do not fully comply to the review appraisal requirements of the Uniform Standards of Professional Appraisal Practice. Users of this appraisal are strongly cautioned that such systems have inherent flaws which render them unreliable as value indicators or appraisal report quality measuring tools.”
The again, I don’t particularly CARE if I do FNMA work. I will not be bullied into “aiming” either my results or my appraisal particulars to someone else’s idea of what is the norm in a particular market at a particular time based on unidentified source data. Similarly, expecting MY appraisal to be the same as another appraisers that may have been performed three or six months earlier does not take into account real world factors like tearing a structure down and building it as new; or the possibility that the prior report may have been incompetently completed.
Dave I LIKE your latter idea about shared adjustment trends. Some regression software sellers note that samples of less than 200 properties cannot be deemed reliable. Few of us ever analyze that many sales in an assignment. Doing so quarterly might be a good idea.
Also, those of you making rote room count adjustments (you know who you are) be prepared to support your $2,500 for a half bath; $5,000 for a full bath or common room, etc.. unless you can PROVE your market actually does the same.
“Proving it” by regression or any other quick data model is not going to work. This new scoring program essentially throws the cash based adjustments for material quality out the window, and also means we should not independently verify data. We all know by now that the majority of the appraisers out there don’t actually verify and adjust data as appropriate. I’m staring down at whole developments where the assessor does not report room and bed counts correctly, and the listing salespersons often just auto import that data and call it a day. The manually orientated appraiser who adjusts and verifies data in a logical manner will be penalized. The appraisers who put blind faith in other parties data will be the winners. I’ve got hundreds of examples where appraisers or sales persons used financial calculators instead of straight line calculation, to come about rounded figures for land and other considerations. That’s another warning for me. If a computer could have successfully predicted real property value, it would have done so a long time ago. When I credit an adjustment for multiple thousands for those superior materials via thoughtful picture and field based research, that will create a differential in data, where the automatic form fillers would have used simple extrapolation and regression analysis instead. Our numbers won’t match up, and the appraiser who allows his process to be controlled and automated will be victorious. This program looks like it’s developed by big data supporters who do not want legitimate checks and balances systems in place. It’s one thing to analyze comparative points, but it’s quite another to hold appraisers to the standard of the lowest common denominator. Looks like I’m working too hard with my manual detailed methods, and FNMA would rather have me just get automated and let the pre defined program and MLS systems tell me what ‘number to hit’. This will definitely create a lot of conflict with Realtors and such, since they won’t understand why I can’t report the data correctly, and instead have to rely on written record data, even if it’s not accurate. This system won’t even begin to be effective, until such time as every listing salesperson in the USA, has to comply with the same terms regarding their MLS entries. Also counties need to properly report. FNMA is putting the cart before the horse. And they’re mysteriously still reluctant to base rejection on the lenders track record of repurchases or the appraisal firms track record of higher than average defaults. Tracking successful loan activity is obviously not the goal of this program.
Sadly, FNMA and users of the CU, will miss out on what could be an excellent tool.
Every appraisal and corresponding report is different and each carry a level of risk based on available sales data, level of comparability, market trends and market participants reaction.
Often we are faced with sales that appear to make sense, but cannot be supported by sales data. In these cases the parties may have to alter the terms of the sales contract or depending on the appraiser the report may opine a market value that is equal to the contract price, using data that may not on the face appear as the most recent or best available.
If the CU were used to truly address and assign risk, it would price these loan products based on the risk score of the appraisal and the applicant’s finance.
Unfortunately the tool will most likely be used to bully appraiser’s to adapt to the use of inaccurate data and poorly chosen comps. This is already evident in the implementation of the UAD system.
The unintended consequences of this practice will lead to a false market controlled by computer programs, which will determine the amount of money you will be able to sell your home for or purchase a home.
In my humble opinion the actions of FNMA and their UAD initiative along with the CU and UCDP is to herd appraiser’s to adopt practices that mirror AVM’s in order to demonstrate their validity.
All consumers and politicians should be aware and fear the obvious actions of FNMA in stripping all independence from the appraisal practice.
I guess we can forget about appraiser independence with the advent of CU. And let’s not forget that Fannie still holds bad appraisals responsible for the mortgage meltdown. All this by an organization still under receivership. Do you ever wonder why? It is becoming more apparent that a no confidence issue exists toward Fannie. Hopefully, intelligent minds can see the bias of this transformation of the appraisal process and will be viewed as a feeble attempt at getting the Fannie house in order.