This is what Class Valuation appears to be doing to justify their actions. According to the article written by Richard Torne in Mortgage Professional America, Class Valuation has announced the acquisition of DataMaster.
From the article:
“Were not replacing the appraiser, we’re empowering them”
Appraisers have had their fair share of discussions on Class Valuation; just search on any appraisal related publication. It’s a love / hate relationship. DataMaster has also been discussed heavily. It’s more of a love / like relationship. Are the two companies a good match?
Remember when Dave Biggers sold alamode to CoreLogic? Is it Deja vu? How will appraisers react? Will DataMaster lose market share? Will DataMaster be a requirement for appraisal completion with Class Valuation? Will it be business as usual?
“The deal, announced on Monday (September 20), will see the two companies work towards fully digitizing the entire appraisal process, Class Valuation said in a statement.”
“Digitizing the entire appraisal process”… what does this even mean? No more tape measures? No more typing? What exactly is being digitized that is not already digitized? We use laser measuring devices that import the measurements directly into the sketch, we use apps to gather our field data and take pictures, we import data from outside sources directly into our reports, (most MLS systems do this without the use of DataMaster, Spark or other services); we analyze our data through computer programs ( Excel/ R Studio, etc) ; we send and save our reports electronically. So what exactly is being digitized that is not already? How exactly is the appraiser being “empowered?”
Who is thinking hybrid appraisal?
Thank you for being part of VaCAP!
Thank you for my morning laugh!
Class, “empower this!”
DataMaster LIKE ALL other online commercial services providing so-called “assessor data” or “public records data” is replete with extensive, routine errors in nearly all categories.
It’s not their fault. The 600lb gorilla (CoreLogic) has the same problem. The extensive errors stem from a variety of causes. Most commonly the original zoning data; living area and land use is from very old county records that are rarely updated.
In other cases, it is from the commercial providers’ own reinterpretation of the data. A 100′ x 100′ lot is 10,000 sf. Online data services like to restate this often, as fractions of an acre. The actual acreage is 0.2295684113865932. They round it as .23 or .22. When .22 is converted back to sf it is 9,583.2 sf. If .23 is converted its 10,018.8 sf. WE all know it is a rounding issue. Computerized form fillers do not.
WE know to check the actual plat map which shows 100′ x 100′ and we (hopefully) all correctly report the site area accurately. DataMaster and CoreLogic do not. The end result unless you are in a very high land value area (Beverly Hills, Hilton Head, South Beach, NYC, SF etc. is not usually signficant BUT any disparity triggers a purported UW originate UW ‘correction’ request.
Much worse are the living area errors and misinterpretations. CoreLogic is infamously the worst for these. Systemically though, they are all similar.
Take the Cincinnati Ohio HUD case recently touted in various publications as proof of systemic racism. The house involved is approximately 2,450 sf of a legitimate legal living area (ANSI). There is a 500 sf garage and a 1,420 sf basement of which approximately 1,200 sf is reportedly ‘finished’.
CoreLogic is now ‘augmenting’ thier public records data (for ‘added value’?).
They show GBA as appx 4,370 sf. No problem. GBA is everything. Main Building, garage, outbuildings, basements. lofts etc.
They show LIVING AREA as appx 3,650 sf. In fact the default auto search for the subject and all potential comparables in the neighborhood ALSO include ANSI defined gla PLUS non GLA varied size finished basement areas as gla.
An appraiser searching for verified closed sale comps will normally pull all the ‘same size as subject’ potential comps of 3,650+-sf.
ANY AVM produced from this data will be completely wrong since they are calling non-living areas GLA with the same value as ANSI GLA. The ANSI-defined GLA is appx 2,450 sf still.
MOST (or many) appraisers, (especially novices or less experienced appraisers) subjected to FNMA ‘tolerant-misguidance/acceptance’ or lenders’ deliberate misinformation; or agents’ desired preferences would be confused by the above numbers. Model matches are reported as 3,650sf. Shouldn’t model matches be used?
The COMPETENT appraiser uses 2,450 sf as the correct GLA. He or she treats the finished basement area separately. Typically at a lower value than legal, or ANSI defined living area.
The competent appraiser’s value opinion is much lower than the less experienced, more easily confused appraisers value opinion. Its also very much lower than AVMs which keyed off the WRONG GLA.
These are only two of the error-prone areas. Zoning, land use, even things like whether or not the property is in the incorporated City of Los Angeles or the unincorporated County of Los Angeles; or in Torrance instead of Redondo Beach 90270 (The USPS puts Redondo Beach “Hollywood Riviera” area property in RB instead of Torrance…which is the actual city. Zoning is affected again.
Another huge discrepancy is that of reporting sales data for an address involved in multiple parcel transactions. Only the first APN keyed parcel is picked up under site area.
Neither Class nor anyone else is going to figure out how to catch and properly handle all these significant discrepancies, so Class can brag all they want about how wonderful they are, but they still don’t know their left from their right when it comes to data integrity and accuracy of appraisal analyses.
Appraisers would be better off marketing their services to all the consumers with (future) litigation claims arising from incompetent valuation services . Not just for Class AI reliant valuations, but all AVMs or hybrids built on bad data. Garbage in-garbage out.