Collateral Underwriter (CU)

[Joe Flacco]

The problem with CU is that they are not collecting enough data. They are collecting maybe 5% or 10% of data needed to make any sense. Same with AVM's.

 

[Meandering]

Wait, wait, let's think about this.

IThis means that if Fannie's algorithm indicates $100 per sqft is the correct adjustment, but the appraiser, through the use of paired sales, grouped data, depreciated cost, sensitivity and whatever other means they use says it is $70 per sqft, Fannie wins. This applies to repurchase demands, but that does not mean it won't apply to other issues as well.

JT's rhetoric has been that Fannie has been whining about appraisers making $25 per sf adjustments.

Fannie did not use $10 as an example of GLA adjustments, but rather used, $100???

So, if the majority of comps are larger, requiring downward adjustments, appraisers are over valuing property by $75 a square foot, in Fannie speak?

They aren't whining about under-valuing properties for repurchase demands.

Is this now incentive to under value properties?

There could be major repercussions when the market tanks again and repurchases start resulting in lawsuits against appraisers, who are not privy to Fannie algorithms, which by definition are not appraisals.

Sounds like there may be a need of new boilerplate disclaimers in residential appraisals.

.

 

[TRESinc]

In the time I have been getting CU feedback from lenders/clients I have never received an appraisal risk score...they must keep it to themselves.

Are appraisers really receiving a risk score when a client asks for explanation from CU feedback ?

every CU report comes with a score from 1.0 (best) to 5.0 (worst) and everything in between. one of my long standing clients told me months and months ago that their policy is that if a score comes in at 4.0 or lower they push it through; 4.1 or higher triggers a manual review at their location of the CU messages and they determine if there is anything sufficient to send it back to the appraiser. messages like "condition adjustment for comp X is materially different than peer or model adjustments" are thrown out the window but say the appraiser uses the same comp and changes the C or Q rating or GLA or room/bath count then they bring it back to the appraiser for a correction. every lender is different and has their own parameters.

 

[timd354]

The problem with CU is that they are not collecting enough data. They are collecting maybe 5% or 10% of data needed to make any sense. Same with AVM's.

While CU is certainly far from perfect, not collecting enough data is not one of its problems

 

[Joe Flacco]

While CU is certainly far from perfect, not collecting enough data is not one of its problems

They are collecting everything they can. Not even close to collecting enough to put a reliable value on a property unless it is a newer development.

 

[timd354]

They are collecting everything they can. Not even close to collecting enough to put a reliable value on a property unless it is a newer development.

They have literally built the biggest and best property/appraisal database in the world. Aside from the data collected from over 15 million appraisals (including all of the comps used in those appraisals), they have pulled in public records data and data collected from CoreLogic and FNC. In most areas, any problems with CU do not result from a lack of data collection.

 

[Eli]

They have literally built the biggest and best property/appraisal database in the world. Aside from the data collected from over 15 million appraisals (including all of the comps used in those appraisals), they have pulled in public records data and data collected from CoreLogic and FNC. In most areas, any problems with CU do not result from a lack of data collection.

I agree with that but real estate is so different from other products that there is no substitute for qualitative analysis and knowledge from a local appraiser who is skilled and competent. Their data is no match for my MLS combined with my experience and knowledge of techniques and methods of appraisal practice.

For one thing I actually talk to people and perform much more analysis (quantitative and qualitative) than they do. My data from the MLS is good enough that when I apply all other recognized techniques and methods of appraisal practice, they are still in second place.

They know it. That's why they hire me. I have done a ton of work for them directly.

 

[Joe Flacco]

They have literally built the biggest and best property/appraisal database in the world. Aside from the data collected from over 15 million appraisals (including all of the comps used in those appraisals), they have pulled in public records data and data collected from CoreLogic and FNC. In most areas, any problems with CU do not result from a lack of data collection.

I am sure you know better than me. I am just saying data is not being collected for a lot of things that have a significant impact on value. It is not collected anywhere on MLS or tax records or any public records. There are many things that are addressed when selecting the comparable properties that are not being collected. If you are comparing two different three bedroom two bathroom above grade house that are the same size and one house has one bathroom on the main level and one bathroom on the upper level and the other house has two bathrooms on the upper level they are functionally very different. Or two bedroom, two bathroom condo units in the same project. One unit is a corner unit and the other is not. This info is not available in any database and the information is typically not provided in appraisal reports but these things have a significant impact on value. These are the things that market participants notice. Not data. This list can go on and on.

 

[J Grant]

At present time, the data collected by CU is used to vet appraisals, not replace appraisals.

 

[Joe Flacco]

All they can really do with it is ask questions. I do not care how much data CU has. CU does not know my market, values, or adjustments better than me.