Sensitivity Analysis For GLA

[George Hatch]

View attachment 37497

These are really comparable lots to you?

They have some things in common and they have some differences. You understand the point that outside of SFR appraising we often work with much smaller datasets and involving more variables, right? Unlike most SFR assignments, we often have to struggle to find enough data and are often compelled to make do with far less in both quantity and quality.

Why do you think I pay so much attention to the long term sales history for the market segment I'm working in? I do it in order to get some context in which to better understand the smaller number of recent sales data - what they do and don't mean.

FYI, even with the shape issues, the larger lot can still be built out with a significantly larger home, those criteria being established in that town's planning policies. That fact would register with an SFR developer.

 

[gregb]

GLA adjustment via Pizza analogy-

http://appraisersblogs.com/gross-living-area-GLA-adjustment

 

[Joe Flacco]

They have some things in common and they have some differences. You understand the point that outside of SFR appraising we often work with much smaller datasets and involving more variables, right? Unlike most SFR assignments, we often have to struggle to find enough data and are often compelled to make do with far less in both quantity and quality.

FYI, even with the shape issues, the larger lot can still be built out with a significantly larger home, those criteria being established in that town's planning policies. That fact would register with an SFR developer.

You don't take 2 POS lot sales and say that is all you have and assume that all regular lots are the same value. You need to find evidence that the value of a regular lot is not more than those POS lots. The sale price of the improved properties on regular lots suggest that it is likely that site values are higher than indicated by those POS lot sales. You already know this because you are a good appraiser.

 

[George Hatch]

I will admit that this wasn't the best example to use because I don't have the 3 model matches that you're accustomed to dealing with, so this leads you to indulge in the idea that this situation is some extreme outlier. I only brought it up because I just got done looking at this situation and had the data at my fingertips.

I showed you the regular lot in the other part of town AND the proximate sales in that neighborhood, but that doesn't seem to be good enough for you, either. I told you that other sale is literally the high sale for a small SFR lot in the entire sales history of that town but that isn't good enough for you either.

Remember, this whole conversation started when you said that the same structure will generate the same adjustments regardless of location. With which I and others in this thread disagree. You then asked for the example of the 1940s dogbox selling for that much more than the site values so I brought that example up. As well as the other example on the north side of town.

If you hate this one example so much then go back to post #69 - which demonstrates the same thing but with somewhat different numbers. Tell me how that example squares.

 

[Joe Flacco]

I have not looked into post 69. When I have time I will do it.

 

[George Hatch]

The properties in that example lie within 5 city blocks of each other.

 

[Joe Flacco]

I think I understand this phenomenon that you are talking about. The phenomenon of depreciated improvements contributing significantly more to the site (based on land or site sales) than the value of depreciated improvements.

I have not read appraisal texts like many of you guys probably have so let me know what you guys think. Maybe it is discussed somewhere.

1) Extraction method is a recognized technique for developing site value.
2) Comparison of site sales is widely considered the more reliable method for developing site value.

What we are really talking about is what is the site value when the value indication by sales comparison approach and extraction method are very different. Is the value indication by sales comparison approach always more reliable by default?

In looking at the data in the area of Glendale, North of 210, what jumps out to me is that there is not much new construction in the area and especially in the immediate area of the $430k lot on Stevens Street that George is talking about.

These are my general observations of the very immediate area .

• There is not much new construction
• One vacant lot sold for $430k
• There are many depreciated 1,000 SF houses on similar lots that trade for between $650k to $750k.
• There is one new home 2,400 SF that traded for $1.35 million one block north at 3316 Santa Carlotta Street.

So based on sales comparison approach, the site value indication is about $430k. Based on the extraction method, the site value indication is let just say about $600k.

What the data is saying is that based on the site value of $600k, new construction is not feasible. The sale price of the vacant lot for $430k is pricing in the fact that new construction is not feasible at $600k but is feasible at $430k. So if we rely on the site value of $430k then that means that new construction is always feasible.

So then should the site value always be the value that makes new construction feasible?

 

[J Grant]

I often see in my area a lot with a depreciated house that is later torn down for build selling for more $, or selling faster than a similar vacant lot. The lot with a depreciated house can often be financed as a res property, whereas the vacant lot often is a cash sale or owner finance at disadvantage rate. That's why it strikes me funny when appraisers rote adjust down to demolish a house when lots with houses on it frequently sell for more $. The minor cost to demolish is more than compensated for the fact that they often can be financed ( and also has sewer/elect service already )

An interesting topic, but ....we develop land value for a subject in the CA, rarely do we develop it for a comp...if our comps are in the market area the assumption is a similar ratio of land to value of dwelling.. The $ per sf between comps and subject on the grid is done too make the comps more equivalent to the subject and is one of the simpler adjustments to develop.

 

[Joe Flacco]

Is the site value of similar lots in the immediate area $600k? Or is it $430k?

What makes more sense to me is that the site value is $600k based on extraction method of hundreds of properties with depreciated improvements in the immediate area and there is depreciation to new construction on similar lots in the immediate area of around $170k ($600k-$430k).

It is either that or site value is $430k based on a couple vacant lot sales and hundreds of properties with depreciated improvements contribute more than cost new.

 

[George Hatch]

Several things:

AT BEST, land value by extraction is a secondary approach to valuing land. It's something we do when we don't have land sales data to work with. Or at least, not enough land sales data.

I've said repeatedly that the $430k land sale was the highest sale for a site in that size range without plans. But there have been other site sales, and if you don't like the $430k then you're *really* not going to like any of the others.

If you really think the opinion I expressed about land values in that area was based solely on 2 sales then you haven't been paying attention to how many comments I've posted over the years about the value of looking at everything I can get my hand on, going back several years, on virtually every assignment I do regardless of property type.

I want to know what you're going to do when your assignment is valuing one of those lots, and not an existing SFR. Are you really going to blow off the entire sales history of the land sales data because you think they're all outliers and resort to land value by extraction? I don't think so.

When SFR values are increasing (or decreasing) do you allocate the rate of change evenly to all the components of cost including profit, or do you recognize the possibility that some of those components will change at different rates?

For example, if developer profit always remained at 12% and hard costs only increased by 3%/yr it might make sense to allocate the rest of the change in value to the site as the residual. But IRL developer margins can range from an outright loss (project isn't feasible to build under the current conditions) to the-sky-is-the-limit.

I recently reviewed a proposed construction deal on a suburban city lot (formerly improved with a 2800sf home in avg condition) where the IRL developer's profit margin for the new house will almost equal their costs of construction. Assuming the market holds up for the 12 months it takes to build that house. There's nothing theoretical about that site value because there was a sale. Now on the Sales Comparison side it may be that the buyers in the current market shouldn't pay so much for those new homes, and it's possible than in the 2021 market they won't be paying that much for those new homes. But for now, they are paying that much even if you and I think they're idiots for doing so.

Lastly, in the end I don't really care why there's currently a disconnect between land sale values vs new home prices. Even though I have my opinions on the reason why, those opinions are immaterial to the job at hand, which is to observe and report. What I do need to do is to look before I leap in my analyses; and to comment on these situations in my report so my readers can see what I'm seeing. And when I say "look before I leap" that doesn't mean looking at only the most recent site sales in isolation of the long term sales history. Or reverting to land-value-by-I-backed-into-it because I don't like the quality of the limited amount of land sales data I do have to work with.

​