Best Method To Determine Location Adjustment?

[NachoPerito]

Comparing site values is certainly one way, and it is certainly better than nothing, or PFA. However, based on 25+ years of observing/analyzing my market, using only the difference in site value would result only in the minimum difference and may not reflect the full difference for improved homes. YMMV

If you are a builder then the site difference is the location difference as long as other costs are consistent (utility hookups and mitigation fees can vary substantially).

The thing is, if a location is better, usually the quality of construction is also slightly better. A nicer neighborhood tends to have a higher standard for quality. This makes it seem like the location adjustment is larger than it really is.

 

[George Hatch]

I don't know what you mean by you can't test. The way I explained why site value comparison is best is that builders build the same product in difference locations. Site value for example ranging from $400k to $800k for the same size lot. Finished house selling for $1.4 million in one location and $1.8 million in another location or $1.6 million in another location with site value in between. That is paired sales with improved properties with identical improvement and also paired sales with site sales. Can't get more reliable than that.

25% difference in site value is definitely not 25% difference in total property value. Unless the site value makes up close to all of the total property value. Like 90%.

This is actually what i said:

And inasmuch as its home values that we're analyzing we wouldn't have any way of testing and refining your adjustment without circling back to the direct comparisons of improved properties to each other.

There's no point in developing an adjustment that you can't test.

That being the case, what would you call this process?

Finished house selling for $1.4 million in one location and $1.8 million in another location or $1.6 million in another location with site value in between. That is paired sales with improved properties with identical improvement and also paired sales with site sales.

I dunno about you, but I'd call that mode of testing an example of circling back to the direct comparisons of improved properties to each other. Which means it was the improved properties that actually demonstrated the adjustment. I mean, what would you have done if - for whatever reason - the sites had sold for the same price but the improved sales didn't? Of if the sites sold for different prices but the finished homes didn't? Ignore the improved sales because their results didn't fit your model? I doubt that.



I'm curious if you guys would voluntarily develop location adjustments on a 20-yr old home using new construction when you actually have comps involving 20-yr old homes that you could use instead? As if the location adjustment could always be assumed to be the same for both types of homes.

I sure wouldn't do that - not voluntarily.

 

[Joe Flacco]

This is actually what i said:



That being the case, what would you call this process?



I dunno about you, but I'd call that mode of testing an example of circling back to the direct comparisons of improved properties to each other. Which means it was the improved properties that actually demonstrated the adjustment. I mean, what would you have done if - for whatever reason - the sites had sold for the same price but the improved sales didn't? Of if the sites sold for different prices but the finished homes didn't? Ignore the improved sales because their results didn't fit your model? I doubt that.



I'm curious if you guys would voluntarily develop location adjustments on a 20-yr old home using new construction when you actually have comps involving 20-yr old homes that you could use instead? As if the location adjustment could always be assumed to be the same for both types of homes.

I sure wouldn't do that - not voluntarily.

The point is that when you compare apple to apple improvements in two different locations, the difference is site value. It is not a difficult concept. Any difference observed after that is a result of something else. If you take a big builder subdivision that is 20 years old and find the same builder subdivision with the same models in a different location, I can almost guarantee you that the sale price difference between the same models of 20 year old homes in two different locations will be very similar to difference in site values. This is not difficult to investigate in a large market area where big builders frequently build subdivisions with the same models in different locations. Try it. I am pretty sure that you will come to the same conclusion in your area as I do in my area.

What I believe is going on is that you guys think you are comparing apples to apples improvements but you are not.

 

[George Hatch]

What I believe is going on is that you guys think you are comparing apples to apples improvements but you are not.

That's a pretty incredible thing for you to say in this thread.

Nobody is arguing that the difference is attributable to the site location. The big variable is how to translate the 20% difference in site costs to the difference in improved pricing without actually surveying the improved properties themselves - which requires no separate site sale analysis. Not even the example you gave above was limited to a site-analysis-only process.


I'm pretty sure that if I'm incapable of discerning a difference in improvement attributes that it's probably not significant to the value. After all, I've seen a lot more SFRs in my career than any layperson home buyer.

Bear in mind a large percentage of appraisal assignments in many of the metro areas have very few relevant site sales to work with in the first place, so "site sale" analysis of any kind is going to be dodgey anyways, let alone collecting enough to pin a location adjustment on them for use in appraising an older existing home.

A large percentage of SFRs that I do appraise involve construction, and I'm one of the few appraisers I know in this region who regularly does land appraisals and who regularly does include actual site sales data in development of a cost approach. In fact, I just did an appraisal on a 30-yr old SFR last week that required a comprehensive Cost Approach and for which I did do a site sale analysis - in a grid - in support of my site value opinion. So this discussion ain't about a bias on my part to cut corners. i just think it's easier to compare improved SFR properties - which I've been doing for over 30 years - to other homes in order to isolate and support a location adjustment.

 

[George Hatch]

At any rate and circling back to the original question, the "best" method to use in any given situation is the one that has the most relevant data for you to analyze. As I say, it's pretty hard to argue the comparability of the subject property to itself, so looking for its prior sale and how that did against the similarly dated sales in neighborhoods where I'm pulling my current comps from will usually be the first alternative I'll be looking for.

 

[Joe Flacco]

That's a pretty incredible thing for you to say in this thread.

Nobody is arguing that the difference is attributable to the site location. The big variable is how to translate the 20% difference in site costs to the difference in improved pricing without actually surveying the improved properties themselves - which requires no separate site sale analysis. Not even the example you gave above was limited to a site-analysis-only process.


I'm pretty sure that if I'm incapable of discerning a difference in improvement attributes that it's probably not significant to the value. After all, I've seen a lot more SFRs in my career than any layperson home buyer.

Bear in mind a large percentage of appraisal assignments in many of the metro areas have very few relevant site sales to work with in the first place, so "site sale" analysis of any kind is going to be dodgey anyways, let alone collecting enough to pin a location adjustment on them for use in appraising an older existing home.

A large percentage of SFRs that I do appraise involve construction, and I'm one of the few appraisers I know in this region who regularly does land appraisals and who regularly does include actual site sales data in development of a cost approach. In fact, I just did an appraisal on a 30-yr old SFR last week that required a comprehensive Cost Approach and for which I did do a site sale analysis - in a grid - in support of my site value opinion. So this discussion ain't about a bias on my part to cut corners. i just think it's easier to compare improved SFR properties - which I've been doing for over 30 years - to other homes in order to isolate and support a location adjustment.

Nobody is saying anybody is cutting corners.

When you have the examples in the market area with improved sales demonstrating that the price difference in improved property sales is very similar to the difference in site values then that is market evidence that the difference in locations is site values. When the price difference in improved property sales are a lot different than the price difference in the site value then there is something else causing that difference.

Obviously, if you have two properties with nearly identical improvements in the two locations you are trying to compare, then you just compare the two properties which would give you the same result as comparing site values. The reality is you probably don't have identical improvements in the two locations. The debate here is if it is better to compare the grouped data from the two locations or to isolate the location variable itself which is site value. It is basically the same debate as the one in another thread about regression with large set of data versus paired sales.

 

[J Grant]

The site as improved occupies a location thus comparing houses is the most relevant method, because any difference in location is baked into the package. If you have additional support from site /land sales that's good, but site prices are not always reflective $ for $ and there are often none or few to be found.

Put the comps on the grid, adjust for all other factors except location . Whatever is left is the location adjustment. Check that against grouped sales for more support, and or ask RE agents who work the area how much more buyers pay to live here then there for a similar house. What a builder charges for a site and then for the home does not always turn out to be the contributory value of a site when the home resells on the open market.

 

[NachoPerito]

Put the comps on the grid, adjust for all other factors except location . Whatever is left is the location adjustment.

I can't imagine this goes well for you. It would mean that all comps will be adjusted to the same number.

The comps range from $249,612 to $249,612. Comp 2 at $249,612 required the most gross adjustments and given less weight. Comp 4 at $249,612 adjusted is the most recent sale and most similar in type and quality. A value for the subject property of $249,612 is well justified.

 

[J Grant]

I never find all my comps adjust to the same number. Adjusting the comps for other features to isolate a differential such as location provides the location adjustment ( among other ways of doing it ). Applying that location adjustment to those comps that need it does not result in all the comps adjusting to the same number ( at least in my work it doesn't). After I apply the location adjustment, I might fine tune others, and some comps won't need a location adjustment.

I try to stick with equivalent locations and avoid location adjustments wherever possible, because a comp needing a location adjustment can appeal to a different buyer group

I still can't see how you came out with the same number for every comp...( theoretical of course)....If I extract out a location adjustment as described, I would use it as a basis for an average...and base it off market knowledge as well. I'd hope 1 of the comps or at most 2 out of 4 need a location adjustment.... I'd rather go back in time and find an older sale in same or equivalent location and make a time adjustment if needed. I don't mind going out of an immediate geo area, but going out of an area for a comp with an equivalent location not needing an adjustment would be ideal.

 

[Terrel L. Shields]

Land value and "site" value can differ. A site is prepared from a bare land sale. A site may include all utilities, landscaping, driveways, sidewalks, curbs, etc.

In my area, when a small subdivision is built, the developer prices the lots and people buy them. And, because of our economy, the typical ratio is about 15% site, 85% house. And that might mean only 10% was a bare land value.. but it isn't bare land. It is developed. So the bare lot value of $20,000 will translate into $40,000 or more once you include driveway, landscaping, and utility installation. And you are likely to stick a house of 2200 SF and RCN of $100/SF on it... When the land price is double or more the RCN...well, that's a problem I don't face in residential work here.

For rural property, I find valuing the site "as if vacant" works for me. Since I extract the same way for my comps, no real issue or inconsistency. Some will parse out the house lot (or in a poultry farm the minimum acreage required by the poultry firm - usually 40 acres) from the "excess" acreage. It works too, but it isn't going to change the value imho.