Extraction Method

[alebrewer]

Using the subject to extract its own site value is akin to using the subject to develop a grm or cap rate in a commercial assignment.

in general, I'd agree Greg. In this case, however, I do not believe it is circular. It's kind of like A+B=C, where A = land/site value, B = depreciated improvements value, and C = overall market value of the property. If you know the value of either A and C, or B and C, you know the value of the unknown...

 

[alebrewer]

BTW, "ambiguous" usually comes up as a result of reading something into an established and considered explanation that isn't otherwise there.

Ambiguous: capable of being understood in two or more possible senses or ways (Merriam Webster)

 

[George Hatch]

So this is the example given from the text: "..., assume an appraiser is estimating the value of the land under an aging, deteriorated automobile service garage that was recently sold for $575,000. No vacant lots have been sold in the market area recently. The appraiser estimates the cost of the improvement at $200,000 and total depreciation at 80%, indicating that the depreciated cost of the improvement is $40,000. Deducting $40,000 from the $575,000 sales price, the appraiser obtains a residual land value indication of $535,000 by the extraction technique." The text does not specify whether the property under consideration is the subject property, or a comparable property. Again, in my opinion, it infers that it is the subject property - but that is my opinion. I have absolutely no argument if someone infers that it applies to a comparable sale...

In that example you have a known sale or list price and you're working on the indicated land value from that one sale or listing transaction.

The process is for a single datapoint and uses its own internal references. There is no distinction being made about it being applicable to only a subject transaction or only the comparables, but in either case the utility of the result is dependent on how confident the appraiser is that the sale or list price they're using to get to the underlying land value was reasonable to begin with.

We work with data all the time, so we understand that assumption (THAT sale price was reasonable) doesn't always pan out. In our sales comparisons we understand the idea that no one sale makes or breaks. Likewise, I wouldn't expect anyone to rely on just one sale for an extraction analysis. I don't do that with extracting cap rates or price/sf indicators or excess land values or any other type of market-based extraction I perform. I always do multiples of that analysis to see what the prevailing trend among them all looks like. I *commonly* see outlier transactions that aren't supported by the trend.

Even if you have a known sale or contract price for the subject, so what? Those don't always pan out, do they? I return value conclusions that are other-than-the-contract price on a regular basis. So that factoid all by itself rules out an assumption on my part that every contract price is a reliable measure of that property's market value within the context of its market segment.

 

[alebrewer]

In that example you have a known sale or list price and you're working on the indicated land value from that one sale or listing transaction.

The process is for a single datapoint and uses its own internal references. There is no distinction being made about it being applicable to only a subject transaction or only the comparables, but in either case the utility of the result is dependent on how confident the appraiser is that the sale or list price they're using to get to the underlying land value was reasonable to begin with.

We work with data all the time, so we understand that assumption (THAT sale price was reasonable) doesn't always pan out. In our sales comparisons we understand the idea that no one sale makes or breaks. Likewise, I wouldn't expect anyone to rely on just one sale for an extraction analysis. I don't do that with extracting cap rates or price/sf indicators or excess land values or any other type of market-based extraction I perform. I always do multiples of that analysis to see what the prevailing trend among them all looks like. I *commonly* see outlier transactions that aren't supported by the trend.

Even if you have a known sale or contract price for the subject, so what? Those don't always pan out, do they? I return value conclusions that are other-than-the-contract price on a regular basis. So that factoid all by itself rules out an assumption on my part that every contract price is a reliable measure of that property's market value within the context of its market segment.

Agreed 100%. I personally would prefer to use my own opinion of market value regarding the subject property, and perform extraction based on that, rather than using a sales price for a comparable sale that may, or may not, actually reflect the true market value of that property. I'm MUCH more confident in my own conclusion of value than I am of some unknown parties consummating a transaction. And as I've said repeatedly (ad nauseam you you might say), I've NEVER said you HAVE to only use one property to perform extraction, only that it is ACCEPTABLE to so so. As a state investigator, I'm ecstatic to see ANY support for the appraiser's conclusion of site value - much less a detailed analysis of even ONE property using the extraction technique...

 

[George Hatch]

Most appraisers I ever met - including myself - were initially "acclimated" on how to back into land value using the results of our sales comparison as a proxy for a known sale price. In my case I'm glad there's a statute of limitations which has enabled me to progress past my previous sins and to resolve to do better.

That's like using the subject's sales comparison results as the basis for backing into a cap rate or GIM that you're using in an income approach in lieu of extracting one from a handful or a bushel of sales data with such indicators. So long as we're talking about relevant data, more > less.

 

[Joe Flacco]

This is a little bit off topic but the reporting format in the 1004 is Site Value + Improvements - Depreciation. It is not Land Value + Direct Costs + Indirect Costs - Depreciation.

Raw land sales do not equal site value. It is probably different in different places but a lot with tear down is the best indicator of site value because it has utilities on site and impact fees are not necessary. The lot is developed to a point that is suitable for construction. Raw land has not even had a perc test or water pressure test if in a area where public utilities are not present. There is a lot that goes on to develop raw land into a site or sites and entreprenurial incentive associated with that.

 

[alebrewer]

Most appraisers I ever met - including myself - were "acclimated" on how to back into land value using the results of our sales comparison as a proxy for a known sale price. That's like using the subject's sales comparison results as the basis for backing into a cap rate or GIM that you're using in an income approach in liue of extracting one from a handful or a bushel of sales data with such indicators.

As I said to GregB earlier, in general I'd agree. I don't think it's the same here, though. Think of it this way: IF you have a solid and defensible estimate of the replacement cost new of improvements, and IF you have a solid and defensible estimate of the depreciation of those improvements, and IF you have a well supported opinion of overall market value, then what is left is, by definition, the site value. I, like you, was taught early on to 'back into' the cost approach. After careful analysis of the extraction method, though, I now believe that estimating the site value via subtracting the depreciated value of improvements from the overall value IS a solid basis for estimating said site value. The problem (currently) lies in documentation and support in the appraiser's workfile. If the appraiser says he/she used 'vacant site sales' to estimate the site value, yet none are in the workfile - there's a problem. IF the appraiser says that he/she performed extraction, yet no analysis is in the workfile - there's a problem. If, however, the appraiser offers a well defensible and supportable estimate of depreciated value of improvements, and performs extraction based on that analysis, that IS a defensible methodology for estimating the subject's site value. It is beyond the scope of this thread to discuss RCN estimation, depreciation, etc., but assuming that information and analysis is in the workfile, extraction CAN (as evidenced by The Appraisal of Real Estate) be performed using one sale - even if it is the subject property. Of course, the preferred method of estimating site value is via vacant site sales (if available).

 

[alebrewer]

This is a little bit off topic but the reporting format in the 1004 is Site Value + Improvements - Depreciation. It is not Land Value + Direct Costs + Indirect Costs - Depreciation.

Raw land sales do not equal site value. It is probably different in different places but a lot with tear down is the best indicator of site value because it has utilities on site and impact fees are not necessary. The lot is developed to a point that is suitable for construction. Raw land has not even had a perc test or water pressure test if in a area where public utilities are not present. There is a lot that goes on to develop raw land into a site or sites and entreprenurial incentive associated with that.

I agree with that as well, Joe. It's interesting, though, that the text uses the word 'land' instead of 'site'. I think it would have been much less muddy to use the term 'site', as you've very aptly pointed out the difference between the two.

 

[George Hatch]

I work in Southern Calif. I rarely have problems finding site sales data for any of my assignments, even in the urban areas. I almost never use extraction to get to land values. I will if I have to, though. I just won't ever do it on the subject.

 

[alebrewer]

I work in Southern Calif. I rarely have problems finding site sales data for any of my assignments, even in the urban areas. I almost never use extraction to get to land values. I will if I have to, though. I just won't ever do it on the subject.

Same here - I am in a fairly rural area in Texas where site sales are abundant. Not a lot of subdivision appraisal work where I am. Have to from time to time, but very seldom. I never use allocation...