Deep Dive - The Cost Approach

[CGinMN]

Actually I don't even have a beef with the CA

Not to be a dick but…

I believe the CA to be intrinsically flawed

The reason machine-developed values cannot surpass human-developed values is a theoretical argument, not a practical one. To me it is kind of like saying a desktops will never be more credible than a traditional appraisals. Will we create a marketplace where desktops test better than traditional appraisals? Possibly! But expanding the scope to involve more data collection and analysis should always theoretically result in a more accurate value. Appraisers can use AVMs and AI, and then some. AVMs cant assign tasks to humans. If they do, they’re not AVMs in my definition anyway.

 

[alebrewer]

But expanding the scope to involve more data collection and analysis should always theoretically result in a more accurate value.

This is exactly what I'm talking about. The appraisal profession should have dumped the '3 comps and done' mentality about 2 decades ago, but we didn't - and now we're gonna end up losing the war because we didn't. And to be honest - I don't think we were equipped to. Most appraisers don't know how to manipulate large sets of data into meaningful analysis. We didn't start teaching that until it was way too late. In addition, most appraisers aren't comfortable with quantitative modeling. All that to say that ANYONE should agree that incorporating more data into the analyses should always result in more meaningful analyses, but as a profession - we are ill equipped to be able to do that. Hence the reliance on 'judgement'...

 

[alebrewer]

Not to be a dick but…

to be fair (to me) - I really don't have a beef with the CA. I just think it's pointless - waste of time - spurious - you get the point.

 

[Ricky LaFleur]

The appraiser's judgment and reconciliation are worth something. If we buy into the argument that it is not, then what differentiates us from AVMs and data scientists who don't have local market knowledge?

All that to say that ANYONE should agree that incorporating more data into the analyses should always result in more meaningful analyses, but as a profession - we are ill equipped to be able to do that. Hence the reliance on 'judgement'...

"Even when adjustments are supported by comparable data, the adjustment process and the indicated values should reflect judgment. Small inaccuracies can be compounded when several adjustments are added or multiplied, and thus seemingly precise arithmetic conclusions derived from adjusted data might contradict the appraiser’s judgment. The sales comparison approach is not formulaic. It does not lend itself to detailed mathematical precision. Rather, it is based on judgment and experience as much as quantitative analysis." Appraisal of RE

I find myself quoting this more and more as some attempt to assert appraisal dominance by data 100% of the time, down to the dime, with no understanding or "judgement" from the user of the data necessary. Just ask the guy who adjusts all of his comps to within 1/10 of a cent and claims everyone else is below his IQ. Btw, I don't disagree with the notion that more data can help produce more reliable results. Rural and complex appraisal often leaves the appraiser groveling for more data, if only there was some.

 

[alebrewer]

"Even when adjustments are supported by comparable data, the adjustment process and the indicated values should reflect judgment. Small inaccuracies can be compounded when several adjustments are added or multiplied, and thus seemingly precise arithmetic conclusions derived from adjusted data might contradict the appraiser’s judgment. The sales comparison approach is not formulaic. It does not lend itself to detailed mathematical precision. Rather, it is based on judgment and experience as much as quantitative analysis." Appraisal of RE

I find myself quoting this more and more as some attempt to assert appraisal dominance by data 100% of the time, down to the dime, with no understanding or "judgement" from the user of the data necessary. Just ask the guy who adjusts all of his comps to within 1/10 of a cent and claims everyone else is below his IQ. Btw, I don't disagree with the notion that more data can help produce more reliable results. Rural and complex appraisal often leaves the appraiser groveling for more data, if only there was some.

Love me some Appraisal of RE quotes - I use it quite often to defend my position that extraction only takes one sale.

And, even though the 15th ed was released just 4 years ago - much has changed, and will continue to change, with respect to quantitative analysis in the appraisal profession. IMO - the primary reason models can't distill to a point value is because of: (1) the fact that the RE market is not a homogeneous market, and (2) there are very few truly comparable properties on the market - or to have sold - at any given point in time. As one would expect with that level of variability (to use mathematical terms), R's and R2's will typically not be as strong as we'd like, and residuals will typically be higher than we'd like. That doesn't excuse us from not having to pay attention to our analyses, though. IOW - I can't tell you how many folks aren't even familiar with the concept that the adjusted range should be tighter than the raw range. Even had a poster tell me - on this forum - less than 2 weeks ago - that he didn't care whether the adjusted range was tighter or not.

 

[Ricky LaFleur]

I can't tell you how many folks aren't even familiar with the concept that the adjusted range should be tighter than the raw range. Even had a poster tell me - on this forum - less than 2 weeks ago - that he didn't care whether the adjusted range was tighter or not.

I wasn't there for the context, but maybe you both have a point. How do you adjust sales that simply were high versus sales that somehow squeezed through the cracks and were rather low? Do you adjust under "non-homogeneity" or "over-optimistic buyer" to tighten that range up? As you pointed out in point #1, the market simply is not perfect, especially when you step outside of cookiecutterville. I see appraisals plenty often that have an incredibly tight range, and all that tells me for such an imperfect set of data is that they adjusted it to look nice and inspire confidence in the user. Well-intending, but misleading.

 

[alebrewer]

I wasn't there for the context, but maybe you both have a point. How do you adjust sales that simply were high versus sales that somehow squeezed through the cracks and were rather low? Do you adjust under "non-homogeneity" or "over-optimistic buyer" to tighten that range up? As you pointed out in point #1, the market simply is not perfect, especially when you step outside of cookiecutterville. I see appraisals plenty often that have an incredibly tight range, and all that tells me for such an imperfect set of data is that they adjusted it to look nice and inspire confidence in the user. Well-intending, but misleading.

Just so I'm clear: If asked whether the adjusted range should be tighter than the raw range - your answer would be 'maybe'?

 

[CGinMN]

The sales on the grid are typically just a sample of a data set, so it is possible that they would widen, while the larger set of data from which the adjustments were developed would tighten. I am in agreement with what you're trying to say though. We shouldn't just plug adjustments in and then let the results speak; this is yet another reason why human judgment and expertise play an important role.

 

[Ricky LaFleur]

Just so I'm clear: If asked whether the adjusted range should be tighter than the raw range - your answer would be 'maybe'?

Certain adjustments will tighten up the range by their nature. Sensitivity is accounting for a variable that makes all the comps fall more in-line. So that I understand why you're asking this, let me first ask: do you believe the point of adjusting the sales is to create a tight range or to make the sales more comparable to the subject?

 

[CGinMN]

Another example would be using both quantitative and qualitative adjustments, which is common in litigation. After applying transactional adjustments to the comps, it widens the indicated range. The rest is dealt with qualitatively via reconciliation. The same approach can be used in residential financing appraising.