Appraisal Statistics

[Michael P Jacobs MAI]

Isn't about time we have a separate forum group for "Appraisal Statistics"?...<SNIP>...And, it is time for me to say something about Appraisal Statistics as taught by the Appraisal Institute and others. The discipline has done a lot of harm to the profession!. The typical book in appraisal statistics is not written by a statistician or mathematician. It is most likely written by someone with a Ph.D. in Real Estate from some lower level state university. They know what they know, but don't really understand it to any depth. - What they don't know is appalling.

I am frustrated as well. Your first two statements are related - but IMO the answer is NO, a separate forum would most likely result in more harm than the poor courses. I couldn't agree with you more that the CE courses are too shallow to be worthwhile. Statistics ("econometrics" in our case) is a science, right up there with ROCKET science. The depth of understanding to produce credible results is beyond what most social-science PhDs get in their curriculum. We don't know what we don't know, and throwing statistics at the appraisal we learned in "trade school" is misleading at best.

I'd rather see a practical approach to our training - many of the posters here and in other threads have eluded to it. Use statistics to describe, not estimate. But, this discussion is for another thread.

 

[RCA]

I am frustrated as well. Your first two statements are related - but IMO the answer is NO, a separate forum would most likely result in more harm than the poor courses. I couldn't agree with you more that the CE courses are too shallow to be worthwhile. Statistics ("econometrics" in our case) is a science, right up there with ROCKET science. The depth of understanding to produce credible results is beyond what most social-science PhDs get in their curriculum. We don't know what we don't know, and throwing statistics at the appraisal we learned in "trade school" is misleading at best.

I'd rather see a practical approach to our training - many of the posters here and in other threads have eluded to it. Use statistics to describe, not estimate. But, this discussion is for another thread.

I was originally thinking about the Statistics course taught by the Appraisal Institute. It is rather rudimentary standard parametric statistics.

However, I was recently talking to another SRA who is very happy with R-Language. He was trying to sell me on the "Tidyverse" package. Note that I have used R-Language it in the past a bit, but that was maybe 10 years ago. It appears like R has improved a lot since then. My son-in-law who has a Ph.D. from Stanford in GeoStatistics says his company (a large oil company) uses R over SAS and that the R "packages" coming out of the universities, in particular the Chinese universities are very high quality. They also use R-Studio and he likes, most of all, the plotting routines (ggplot2).

Now, there are 8000+ packages in R - and there is no way I could begin go through them all, learn how to use them and test them. BUT, I am pretty sure that as far as extracting adjustments - there is nothing that can beat Salford-Systesm MARS. I may be wrong. If anybody out there finds an R package that can get an R2 of 0.90+ on the classic Boston Housing Data - please let me know.

The point is, I have no way of knowing what the SRAs and MAIs out there are using with respect to the 8000 packages in R. They may be using non-parametric methods for all I know.

Why should we not have a separate forum? Well, I'm thinking a lot of you probably don't want to have your freedom constrained through some kind of consensus - and yes that is a legitimate concern. We need to experiment and try new things. But I would like to suggest that we are not seeking conformity at this point. We are trying to find better methods.

Also, I see that some of us are more perfectionist than others. Some are happy to have all their adjusted prices within a 5% range; others want to try to get it within 1% (like having all your adjusted prices between $700K-$705K). And if they find something that is "good enough", they aren't too interested in someone elses method that may be superior - or it may be that their method has advantages they give a higher priority on, such as being easier to use or graph.

In conclusion, I think we have to give this some more time to digest. But - it would be nice if appraisers could kind of let others know what they have discovered and their experience using it. In particular how well their methods fit the data .... and so on.

We have some "cowboys" on the forum who don't see how relationships and patterns have anything to do with cause and effect. I'm scratching my head. If I go into a neighborhood and find that that home prices vary according to $200/sf living area, $15,000/bathroom, $150/sf for lot size, and that these adjustments account for over 90% of the variation in price (R2 value), are they going to ignore the statistics because they can't figure out whether it is because the builder is setting the selling price according those criteria or because that's what the buyers are generally basing their offers on because of competing properties in the surrounding areas. Cause and effect are good to know - but I think we can assume if we create a good fitting model based on all 20+ sales over the past year, and another potential sale comes along - it will probably fall into the same pattern, regardless of the cause. Of course, the appraiser is going to most likely make some other adjustments for things like condition, quality and/or view that may not be part of the statistical model.

 

[Artemis Fowl]

The most important relationship, IMO, is the appraiser's thought process to the typical buyer's thought process. Statistical analyses is not a typical action. I've seen some intellectual-types spend inordinate amounts of effort on stats when prepping for an appeal...often they become lost in the methodology and some very common sense considerations get missed. The old sit across the street and ask yourself what would you pay for it method is closer to how buyers do it. I've watched my wife educate buyers about their market and lead them to an informed decision...no stats involved. Don't get me wrong...I love analyses and finding patterns and relationships. I just think they belong in a more supportive role. The confidence I would place is also dependent on the buyer type...what may be appropriate for a large operation with beancounters involved in the decision making process will likely be less so to a single family residential occupant-type.

 

[Michael P Jacobs MAI]

I was originally thinking about the Statistics course taught by the Appraisal Institute. It is rather rudimentary standard parametric statistics. <SNIP> I have no way of knowing what the SRAs and MAIs out there are using with respect to the 8000 packages in R. They may be using non-parametric methods for all I know. <SNIP>
Why should we not have a separate forum? Well, I'm thinking a lot of you probably don't want to have your freedom constrained through some kind of consensus - and yes that is a legitimate concern. We need to experiment and try new things. But I would like to suggest that we are not seeking conformity at this point. We are trying to find better methods.
In conclusion, I think we have to give this some more time to digest. But - it would be nice if appraisers could kind of let others know what they have discovered and their experience using it. In particular how well their methods fit the data .... and so on...

Gotcha on your objective for the "statistics forum" and agree with your thinking, but it's not a functional open-forum topic. Maybe a forum open to reading but closed to posting except invitation-only would fit your concept better.

Why do I say this? You see how easily threads are hijacked here (and on every board like this), people can't resist making distracting segues into jokes, slightly off-topic replies, etc. and you can only sustain a cohesive line of logic a short time before a thread disintegrates. This concept is deep - really DEEP. Everything is math - there's little judgement involved and appraisal professionals are used to relying heavily on - and expressing in terms of - judgement. Keeping a thought on track here long enough to resolve a complex issue is like, as we say, "Herding cats."

The theory needs to be developed like scientific theory - publishing brief but complete manuscripts, and then defending challenges or making adjustments for the next article. Repeat. A forum like this is too informal.

That's why I say it won't work (in my opinion) although I am as eager as you are to see the theories and processes advanced.

As for the Appraisal Institute, I agree so much with you that the AI courses need to be re-thought... maybe individual instructors have rounded their courses out, since I'm sure that they want to provide relevant education, but as you know, we need a lot more than a course to make econometrics useful and credible. Maybe the people who authorize courses didn't actually understand this topic's course content well enough, and it was put into practice too soon. Or, maybe it is going according to plan - baby steps; we don't know the overall plan if there is one. So, that may be a place to begin - get involved in an educational platform like that and improve the system. Publish... just not only here if you want progress.

 

[Charles West]

I for one would like to see a healthy discussion of the subject matter suggested by the OP, can't see any downside.

 

[Meandering]

I would also like to read such discussions.
As long as they avoid political overtones.

.

 

[Slappy]

Interesting topic. I have found regression analysis useful in neighborhoods of high conformity for deriving a GLA adjustment. Otherwise, if the neighborhood is too diverse, you have too many factors at play which will bias the results. Then you get into manipulating the data which will bias the results futher.... Lately, I have been thinking about how pricing decisions are made. Stat analysis presumes that buyers are evaluating property characteristics and a making decisions based on these characteristics. But that does not seem to be the way it works all the time. Pricing is often controlled by realtor decsions. Realtors base pricing offers on their interpretations of what the perceive as "comps". Other agents look at these comps and price accordngly. That is how pricing operates in neighborhoods. And we all know typical realtor analytical thinking ( the good ole $/GLA benchmark) is more self serving and arbitrary than useful. I am not trying to dismiss regression or statistical analysis as it can be very useful. But we have to keep in mind that residential pricing decisions involve a lot of emotion and subjectivity.

 

[Eli]

I will interject this fact that is relative to the variables outlined here in this thread. VA loans do 100% financing and places huge emphasis on appraised values and has never needed any bailout.

 

[Mr Rex]

I thought ads had to be paid for around here. This thread reads like an ad for MARS from the get go. MHO.

 

[Eli]

I thought ads had to be paid for around here. This thread reads like an ad for MARS from the get go. MHO.

It reads worse than that to me.

Like somebody who don’t know their butt from a hole in the ground or is under severe pressure.