Bad advice from Fannie--"Multiple Parcels" from Dec. 2019 'Appraiser Update'

[J Grant]

Here's an example and the reasoning behind the manner in which we develop discount factors for mixed use properties.

If i know that the value of the SFR parcel if sold individually is $200k and the value of the adjacent vacant parcel is $50k then in an assemblage they both get discounted, not just the vacant lot. That's because *from an investor-buyer's* perspective the short term profit motive to which we normally attribute to a profit-driven investor lies in flipping both parcels, not just the one.

Long story short, unless you actually have direct comparables with the same unit mix (1 SFR + 1 Vacant) you have to get to the retail of each first before you can develop a discount for the assemblage.

And if you have to go to the effort of doing a land sale value anyway, then that makes the "discounted" value an extra and arguably unnecessary step. For most mortgage lenders, anyway. IRL, if you give a portfolio lender the retail on each, they're usually going to do a single LTV on the improved property and either do no loan at all on the land or else simply take it as an abundance of caution.

As for having to do two reports, I don't see that. Nobody squawks about it when you do a proposed construction appraisal and actually use site sales to develop your opinion of site value - and then proceed to use that land value indication as your opinion of the "as is" value of that proposed construction. We do multiple values in appraisal reports all the time without it being a problem.

Now these lenders and the GSEs might have to kick these appraisal reports into manual review because their machine isn't set up to handle an auto-review of this type of appraisal problem, but they should still be able to use that report without going into meltdown mode.

Agree with all except the first...this property's only buyer is not necessarily a flipper - though it might be. Could be an owner occupant looking to live in the house and keep the lot for future investment or possibly build on the lot for themselves or family one day. Such as live in the older house, build a new one to occupy on the vacant lot, then sell the older house. Whatever their plan for the lot, buyer got a good terms low payment mortgage covering it out of the deal.....sweet- so maybe they need not discount by much?.

A vacant lot sold alone is either all cash buy, or owner fiancee at big chunk down /repay in short balloon terms. so ability to finance a vacant lot sale into a 30 year SF house mortgage has its own value.

IT is an oddball property for sure.

 

[George Hatch]

What if both lots have the same value as vacant? Sell separate (discounted) or as a whole? H&BU?

I do a lot of assignments like that, where my subject consists of multiple lots. I just had one earlier this year consisting of 9 separate and finished estate lots in a gated project. Each one of them had a different retail value. The value of the group included consideration of the length of time to sell them all off, what market conditions could be forecasted (guessed) to occur over the length of the holding term, the holding costs over the length of the holding term, the costs of sales and so on. By the time I got done with that analysis the value of the group was way less than the sum of the retail values. And i could not have done that analysis without using a retail value for each lot as my starting point.

When it's just a single use like that my chances for finding directly comparable sales goes up significantly, so I had some of those sales to use, too. And yes, those assemblages virtually always sell for less/unit than if the units were sold off individually.

 

[Carnivore]

I think the

View attachment 42854I do not follow the logic of a "value in use" statement? The value is what the value is...and they are implying the value is zero (in use) but you need to value it at market. Again, as stated however, you do include the value in the grid and you are not doing 2 appraisals although you certainly need to support the value of the second lot with sales or extraction preferably.

In the last paragraph above it says how to add the contributory Value as a single line item adjustment. To do that you first must determine the market value of the extra lot just as the same process you determined the site value of the subjects site. Then you go through the gyrations of determining its actual contributory value, Assemblage

So in effect you have performed essentially three appraisals. Unless you don't consider developing Site value an appraisal, but just a component of the total SFR Appraisal

Not sure where they are getting Value in Use?

Back to assemblage. Example; Your assemblage is to be able to have enough room for a four unit apartment building which you could not do on any of the individual lots
There are many other reasons to use assemblage..

So i don't see how assemblage is proper for the argument thats taking place. AKA SFR BUT maybe I can create a scenario that would be legitimate for a SFR like they are talking about. Maybe the guy is refi to build a garage, His current site will not accommodate a garage. So he wants to assemble the two lots to get enough area to build his garage. Thats a legitimate purpose.

What is happening out there is the assemblage has no purpose other than to merge two sites into one for the purpose of a loan. What they are actually doing is they are not assembling the sites. Except in the Appraisal and in the deed of Trust.... BUT NOT THE DEED/Legal Description. the official document

 

[George Hatch]

Agree with all except the first...this property's only buyer is not necessarily a flipper - though it might be. Could be an owner occupant looking to live in the house and keep the lot for future investment or possibly build on the lot for themselves or family one day. Such as live in the older house, build a new one to occupy on the vacant lot, then sell the older house. Whatever their plan for the lot, buyer got a good terms low payment mortgage covering it out of the deal.....sweet- so maybe they need not discount by much?.

A vacant lot sold alone is either all cash buy, or owner fiancee at big chunk down /repay in short balloon terms. so ability to finance a vacant lot sale into a 30 year SF house mortgage has its own value.

IT is an oddball property for sure.

The presumption for a flip transaction includes a profit incentive whether the seller is partially occupying the property or not.

You don't omit the potential market rents for the owner's unit from your Income Approach just because the property is partially owner-occupied. Same principle applies to assemblages where the components are subject to being sold off individually.

 

[BRCJR]

I do a lot of assignments like that, where my subject consists of multiple lots. I just had one earlier this year consisting of 9 separate and finished estate lots in a gated project. Each one of them had a different retail value. The value of the group included consideration of the length of time to sell them all off, what market conditions could be forecasted (guessed) to occur over the length of the holding term, the holding costs over the length of the holding term, the costs of sales and so on. By the time I got done with that analysis the value of the group was way less than the sum of the retail values. And i could not have done that analysis without using a retail value for each lot as my starting point.

When it's just a single use like that my chances for finding directly comparable sales goes up significantly, so I had some of those sales to use, too. And yes, those assemblages virtually always sell for less/unit than if the units were sold off individually.

I understand DCFA. The thread is much more simplistic than that.

 

[alebrewer]

Not sure where they are getting Value in Use?

I think the 'value in use' refers to their take that, if the land can't be sold off (as it is encumbered by the first lien), the owner's only option is to 'use' the land at whatever value there is in the utility of the additional lot which, in most situations, would be inferior to the value of the lot if it could be sold (e.g. MV).

 

[George Hatch]

I understand DCFA. The thread is much more simplistic than that.

The buyers and sellers for SFRs tend to be less sophisticated, but the underlying principles remain the same even if the actual reactions in a given market segment are more muted. If you're buying both together you're going to pay less for each component.

 

[George Hatch]

I think the 'value in use' refers to their take that, if the land can't be sold off (as it is encumbered by the first lien), the owner's only option is to 'use' the land at whatever value there is in the utility of the additional lot which, in most situations, would be inferior to the value of the lot if it could be sold (e.g. MV).

That's one reason I always do "all of the above" whether the client is savvy enough to ask for it or not. That way, they can pick-n-choose which parts they want to encumber, or alternately, if they want to release one of the parcels later on they already have a value opinion for it.

I've never once gotten hassled by a client for answering a (obvious) question they didn't think to ask.

 

[George Hatch]

I should mention - it is appraisal problems like this where these clients and users need a good appraisal. The machine can't do stuff like this. The machine can't tell where the existing improvements sit relative to a lot line or setback boundaries. The machine can't handle all the covariables and the comparatively tiny datasets. The machine can't provide an explanation to the users that they'll understand.

We exist to solve problems and provide meaningful solutions. We can't blow off situations like this on the basis of them being unprofitable in one breath and then complain in the next that the machine is replacing us. The way forward for us does not lie in attempting to outcompete the machine on its terms and against it's strengths; the way forward for us is to excel at the things we can do which the machine cannot. We don't compete with a calculator, we use it.

Even if some of you do choose to take the path of least resistance it would still behoove you to developing an understanding of how to provide a meaningful result to a user that's actually competent enough to see past the expediency factor.

 

[alebrewer]

I should mention - it is appraisal problems like this where these clients and users need a good appraisal. The machine can't do stuff like this. The machine can't tell where the existing improvements sit relative to a lot line or setback boundaries. The machine can't handle all the covariables and the comparatively tiny datasets. The machine can't provide an explanation to the users that they'll understand.

I know this technically isn't functional, economic, or physical, but I, too, contend that there are just some things that AI (or machine learning, or whatever moniker you use) can't handle - yet...