Rules Check Software
If the ‘story’ can be told using 3 comps, why are 4 or more really necessary?
As you are probably aware, all of our appraisal reports receive an initial examination electronically. There are a number of businesses which have written and provide to the industry various types of ‘rules check software.’ This actually began in the early 1980’s by the appraisal computer software pioneers who wrote programs to ‘check’ reports for inconsistencies.
I just learned about one such ‘rules check’ company, and the results they’ve seen using their software – below. They even admit to users that the users can ‘mine’ reports for data.
I’ve taken the liberty to add my own comments using this typeface:
Interesting report data from review of 1.5 million appraisals:
4.72% of the appraisals Market Value is higher than the adjusted comp values
Yea, we all know ‘our value’ ideally must fall between the comp sale prices, and the adjusted values of those – or U/W’s will come unglued. But real estate is not perfect, and sometimes it does make sense, due to the properties used for comparables, that an Opinion of Market Value is outside those ranges. This is something that happens often in rural areas, but U/W’s think those must have the same pattern as urban properties – which seldom occurs.
12.10 % of the appraisals use comps from different Cities
And the concern is? In urban areas where the entire landscape is covered with homes, it’s entirely possible that comps might not be in the exact city (or ZIP Code) where the subject is located. One problem with this is Census data, which can track average income, age, etc. for the cities & ZIPs. However if the report is done correctly (i.e., comps have similar characteristics and prices), all the appraiser need do is explain why this was done.
14.94 % of the appraisals use comps from different ZIP codes
See the above. My city has a ZIP code boundary along a primary arterial street. This boundary line was added AFTER the development in that area. Similar properties are on either side of the ZIP code line.
1.69% of the net adjustments are under 15%
Probably a misprint. More likely the concern is NET above 15%. But let’s remember FNMA has eliminated their guidelines about line, net and gross adjustment percentages. However, other lending entities stubbornly refuse to change this… even though the resulting property value may be inaccurate because the appraiser also stubbornly makes sure adjustments magically fit that outdated percentage mold.
4.05% of the adjusted comps have a 14% low to high value range
Again, this gets back to the perception that real estate is “perfect.” It is not, never has been, and never will be. There are too many variables… and it’s not the same as buying the same national brand boxed pizza from multiple grocery stores in your area. Depending on property circumstances, location, etc., having a wide adjusted range is entirely consistent with the reality of imperfect real estate. Some will say that ‘adjustments’ are wrong if the range is too wide. Maybe… if enough time is taken to analyze every morsel of sales data over spans of time to try to justify location or site size differences and narrow the range. But in the real world of home mortgage lending, there seldom is enough time available, or fee, to do everything necessary to SUPPORT super accurate adjustments to try to narrow the range. These should be done separately for every assignment, because adjustments for one report might not be accurate for a different report.
7.58% of appraisals indicate that adverse location adjustments range from 5% to 15%
See the above. A high adjustment percentage may be appropriate based on assessed land values, property sale trends, school district boundaries, geography, access roads, etc.
6.25% of the reports did not use 5 or more comps
And the concern is?? Did the assignment instructions say 4 or more were required? If the ‘story’ can be told using 3 comps, why are 4 or more really necessary, especially if the lender does not mandate more than 3?
8.32% of the appraisal reports didn’t include current listings
Probably because the assignment instruction did not mandate listings be included. However, I happen to believe one or two listings SHOULD be included and I do that, voluntarily. Listings tell the current ‘story’ while the retrospective sales reveal the prior history.
5.44% of the appraisals had inconsistent neighborhood price ranges
I see this when I review reports. It means too wide a price range. Too many appraisers include ALL sales (and listings) within a radius around the subject. Many of those properties are not competitive/comparable to the subject, and SHOULD NOT be included – on the top of report form page 2 or in the MC form (which flows back to form page 1 Neighborhood check boxes), and certainly not as comps. Appraisers should concentrate on the characteristics of the properties, and not the NUMBER of all properties in a particular area.
This relates to the ‘imperfect’ MC form. Many appraisers get all twitter-pated if the “number” of sales across the first MC grid line is limited. So what? You are asked to report the number of COMPETITIVE / COMPARABLE properties. If there are few of those, so be it. I sometimes have 1 sale in 12 months (that the form mandates). It is what it is, where it is, when it is. Do not rely on the ‘imperfect’ MC form to justify market trends. Instead, use your own market derived data from other sources to supplement the MC.
7.15% of the reports with new construction PUDs did not use comps from competing PUDs
This is reported because lenders want to see supportable market evidence for similar homes in similar PUD’s. But this again gets back to imperfect real estate. Sometimes there are no similar PUD’s near the subject’s location. Builders build and buyers buy, sometimes irrespective of what else might be in the market area. Appraisers then need to fully explain any differences and perhaps location adjust for non-PUD homes used as comps.
All of the above relates to ‘big data’ that so many are fixated on and enamored with these days. ‘Big data’ is useful when the originating data is factual, accurate and consistent. Sometimes the originating data is not.
That’s the problem with ‘big data’ relating to imperfect real estate. With imperfect real estate, there are often ‘outliers’ which can skew results. If those are not removed from the ‘big data’ sample, the resulting output is flawed… and imperfect.
Unfortunately, appraisers are caught up in this mess. When our property data is actually imperfect to begin with as it relates to the subject, but needs to be used for a variety of reasons, we then are judged to be imperfect when our reports don’t magically produce results ‘big data’ expects.
All we can do is be careful to properly analyze all properties used as comps, and use supportable adjustments as necessary, and explain what we did.
Quit saying you have used the ‘best comparables available.’ That is too easily challenged, and is trite. Instead, say you have used the “most appropriately comparable properties” as your comparables. (“Appropriate” is the word used in Dodd-Frank where it says lenders can submit other properties to the appraiser for consideration.)
Include necessary and additional commentary on the pre-printed form pages, in the Addendum, and other exhibits. Strive to generate reports that are well written, factually correct, and free of inconsistencies from one section of the report to another… and USPAP compliant in terms of required additional reporting items.
The better reports you write, the easier they will slide through the electronic and additional review process. That means fewer correction notices received, and you’ll be much happier!
- The Unbiased Reality of Census Tract Data Dump… Open Letter to FHFA - August 15, 2023
- Isn’t that Just PAVEY! PAVE Response and Marcia Fudge - August 8, 2023
- ‘Pesky’ Words. Keeping Appraisers out of Language Purgatory - July 26, 2023