Appraisers' Free Forum

You might have mentioned that assumption up front,

I did: "For unique properties, a buyer has two choices:"
I said "unique," not "rare". And you're right to point out that this approaches value-in-use, if it doesn't run right up its tail lights.

I get really irritated at the ambiguity this profession exhibits, and which many of its senior practitioners tolerate and embrace. I don't have the depth of experience and education that enables you to detect and highlight them so readily. In the past, you've voiced serious concerns when someone describes appraisal as "more art than science." I think you'll have to relax that criticism somewhat, though. Art involves suspension of disbelief. If one relies on the cost approach for MV, one certainly has to suspend disbelief for a time.

My own rural appraisal reports sometimes remind me a lot of a Matisse painting. Those blobs of color look a lot like flowering shrubs, but don't get too close.

As I attend AI's classes on report writing, I'm more and more coming to believe that your Unified Approach Theory of appraisal is correct: There is only one approach to value, and it is frequently calculated three interdependent ways. Maybe much of this controversy would go away if they would just say that, instead of blathering about "three independent approaches" so much.

that enables you to detect and highlight them so readily

I am asking the same questions I asked during and after my first appraisal course. If I appear to detect that a large chunk of appraisal theory is all holes and no cheese, it may be because I already had concluded that ALL emporers are naked. And clearly what is going on here, is that 70 years of appraisers have simply assumed that 6 guys on a 1936 committee were infallible.

I'm more and more coming to believe that your Unified Approach Theory of appraisal is correct: There is only one approach to value, and it is frequently calculated three interdependent ways.

I don't know if that's my theory, or I thought of anything hasn't been thought of already. I may have streamlined the explanation.

The middle part of my "theory" is that in each appraisal, the appraiser creates a theory of why the subject is worth what it is worth. Three theories cover about 90% of all appraisals:
1) subject is worth Y, because other gismos like it sell for Y,
2) subject is worth Y, because investors pay multiple X for investment property, and
3) subject's is worth Y because that's the discounted present value left after investing the A dollars necessary to transition the property and selling off the finished product for B dollars) or selling off the finished pieces (for b dollars each).
In short, marketable non-investment property should be appraised by sales comparison, investment property should be appraised by capitalization, and transitional property is appraised by development residual (in US books called the subdivision method or sometimes DCF).

You figure out what value theory explains the property and use the associated method. Unless, there is a second theory of value,which one would expect would produce a different value than the first, there is no need for a second, less reliable, or faked method.

Example, one office building has 10% more capacity to produce income than another. It is worth 10% more. It doesn't matter if they are both on the same street and one has 10% more square footage; or if they are the same size, but on different streets, and one has 10% higher rental rate. There is only one value theory here. Subject is worth ten percent more because it is ten percent better. It's actually moot whether you want to attribute the 10% better to 10% more size, or 10% better location. Ten percent is ten percent. I published this in an article which is not available on line any more. Using variables, I went on to write the basic steps of the income capitalization and sales comparison approaches, but they both simplified to the exact same equation. It is inevitable that this would happen, but nonetheless, that is mathmatical proof that there is noreal independent "second approach," just one core calculation that display the same value theory two differerent ways (line adjustments sales comparison and IRV capitalization) reaching the same answer. I could display it several other ways as well. Now that I see it in print, I can see why you would think I am saying there is only one approach, but what I am saying is that there is only one value theory (subject is 10% better and worth 10% more), and it could be displayed with either of the two "approaches." but displaying it with both is reduandant. The second does not "confirm" or "support" the first.

The alternative I find to this theory-centric thinking is approach-centric thinking that seems to treat approaches like three magic formulas that all lead to Rome. There really logical foundation to this. Multiple methods came into recognition because there were multiple types of property, and multiple types of value. Whoever decided all roads are supposed to lead to Rome was wrong. That's why appraisers end up leading them to Rome (by backing in and fudging).

A second method becomes necessary if the highest and best use of the property is not clear, or the property appeals to both users and investors, or perhaps the assignment asks multiple question. That is, you could take the same type of office building I mentioned before and find (by method 1 above) that office building users pay $100/sf, but that based on (method 3 above) market rent, rent-up delay and cost of TI, the value to an ivestor is only $80/sf. Then your most probable price becomes either the one that is more probable or a weighted average. In the three magic formulas school, someone is likely to draw the reasonable inference that they are supposed to take the two numbers and "reconcile" them into one number, because for example, one magic formula had better data. The reality is, there are many split purposes properties that appeal to more than one profile buyer.

Of course, we would spend 90% of our time, talking about how to appraise the other 10% of properties.
But simply put, the primary protocol is to focus the power of the best method on the problem. Secondary methods, to the extent they are real, are generally a waste of time, until the absence of data forces the use of them. For example, with commercial users figure land prices by residual based on gross potential sales. You can do that to, but the range is wide, and has little credibilty when it misses the actual sales prices.

George Cox in the AI forum published an article that said essentially the same thing about using the method appropriate to the property. The TN RE Appraisers Commission republished it in their newsletter recently. Then when I applied to upgrade to CG, they wanted appraisal reports that demonstrated all three approaches in them. So I did'em a demo, and included language to that effect in the reconciliation, stating that the cost approach had been included only because it was required that I demonstrate it; and that the income approach didn't give the best indication of value because most local buyers purchased to occupy, relying on rent only to offset holding costs. The report was approved. (And that surprised me a little; I had expected at least a little resistance.)

I read the article but couldn't quite figure out what George was saying. I understand what you mean by demos. I wrote them, graded them, and advised others and was part of the chain.

You are right that "owner occupy predominance" is the "generally accepted" palabre for why you don't rely or develop the income approach for SFR's. Say that, and you will pass muster in perpetuity. In a demo, you are just showing you can develop it and make the normal "sensible" comment and thus show a professional capable of"due diligence."

But what does reality say? I use other principles, not in most appraisal texts.
Postive versus normative - Positive is concluding what the indicate. Normative means concluding how it ought to be, opining what is "normal" in absence of data. Appraisers do both. Post a thread about how much to adjust for a house painted pink, and someone will hit you that adjustments MUST come from the market. But you are going to have to best-guess, PFA, SOPE it, because the market will never give you that data. I put this right in my scope-of-work disclosure - how much of narrowing the range is the data, and how much is my guesstimate.
Inference versus simulation - Appraisers do both. Inference is saying these babies sell for $100/sf so subject is worth $100/sf. You are not walking in the footsteps of the buyer. You are not saying the buyer say, 'ooh another square foot, I am going to pay another $100.' It's a simple and primitive matter of statistical correlation. For example, if you use regression, a critic will hit you for a lack of simulation - 'buyer's don't use regression." But $100/sf is what they pay, predicting prices is the name of the game. On the hand, you could try to "simulate" the market, by doing what buyers do. That would include using the type of residual method that someone like McDonalds might use for land value or building value. However, you will find because of collinearity (Austin's favorite word) that even McD's has figured out a shortcut, all they have to do is count the cars. That leads to the next principle
Shortcuts - Most appraisals and some methods are shortcuts. Just about any inferential method is a shortcut. If they sell for $100/sf, then that's what they sell for and I dont' have to engage in the teduim of walking in the footsteps of the buyer. You were posting elsewhere about DCF and direct cap, but the fact is that direct cap is just a shortcut. You don't have to account for future value, changes in income stream and gaps in income stream, you just assume they are the same (perhaps not stating all these stability assumptions and forecasts, just saying, these babies sell at 9% cap rates).
Least variance and collinearity - This is why a little regression concept helps. Most appraisers will say that if one unit of comparison makes a tight range and the other makes a wide range, the first better explains the market. But this idea of least variance is not dilligently and rigorously applied. Collinearily, on the other hand, means that if you have similar properties, the price per anything is going to make a similar range. This is why "the three approaches" do not accurately or adequately break practice into its logical components. Price per square foot, or per doorknob, or per light fixture are one approach, but price per dollar of income (GIM to appraisers) is another, separate "approach?" Mathmatically, its the same approach based on a diferent property attribute. Price per dollar of cost is another separate approach? Well, if price per doorknob, or per light fixture don't really "confirm" price per square foot, then why does price per dollar of income confirm it as a "separate" approach?

Generally speaking, positive beats normative (comps rule), inference beats simulation (comps rule), and least vairance wins (that is, comps indicate 100-110 beats comps indicate 30-200).

Let's apply this to a real life case study - and go back to your comment about the income approach and the owner-occupied matra, while applying the above principles.

Many years ago, I appraised condo units in a small, diverse market. Some projects had hundres of units, some as few as 12. The local population is relatively transient, so there is probably a high number of rentals, anyway more than enough for income cap.

I found based on three years of sales, that the GRM's formed an amazing narrow range. I think two standard deviations was something like 118-132. This included local and resort properties (which included some units in short term rental plans), all improvement ages, all amenity packages, and everything from studio to 3-bed (least variance). These properties were not "investments," so even though this "inferential" method did not "simulate" what buyers did, it was highly price predictive. It was completely "positive" (exactly what the data indicate) with no appraiser "norming" (what some would mis-lable as "subjectivity"). And yes, it was a great "shortcut" to the problems of direct sales comparison. Apparently, the market saw the proportional quality among these properties as the same, whether they were going to live there for one year or several years (collinearity), and nothing gets around time adjustments better than this months market rent and this months multiplier.

Most assignments did not afford the oppurtunity to use "comps" from the same project. After all, if there are only twelve units, the last sale might be five years old (and what if the 12 units had different floor plans). To avoid time adjustments, most appraisers used comps from other projects with adjustments that had to be "normative." Sometimes, especially on the 2-bed and 3-bed units, which sold more rarely and showed greater project differences, appraisers would develop battle fatigue. (That's why I stopped appraising churches). I reviewed on assignment where the appraiser included a long narrative analysis and reconcilation based on three sales from three different projects. In one way, it was masterpiece. However, battle fatigue probably caused the appraiser to fail to notice that the abutting unit with the same floor plan and level of upgrade (all original) was listed for $25,000 less than the final conclusion of value. Ouch!

These appraisals and reviews brought three typical reactions: 1) very clever (rarely), 2) you must be insane (much more often) and 3) thanks for the appraisal, but I really thought it would have come in higher (most often). I can remember one conversation with the senior loan god at a big bank. He said it was hard to believe how low the value is. I said, there are two things. First, that's exactly what the data indicated that day, and second, about five days after the appraisal, one of those units sold for within $1,000 of my appraisal. After a pause he asked, is there any way you can appraise this for the higher number (the original loan value) so I don't have to book the loss in this quarter?

So, an income approach can be the "best method" for residential property, even if only in the rare circumstances, I described. I doubt it would ever have that little variance in SFR appraising.

-----
Using my "theory" we now unravel several sets of problems that further disconnect cost from market value. It's not just that at square one, cost isn't value, and cost might not even be "cost." The cost approach itself as promulgated by the AI, is a "simulation" method. As that course said, the buyer says why should I pay you $200,000 when I can build one for the same amount and get the same or better benefits. I'll avoid the math problems on this here, to point out that this premise holds that the buyer knows land values, and the cost to construct. But in real life, few do. That is, the cost approach to market value attempts to simulate buyers, but these actions are something buyers don't do. Thus, the process of doing depreciation is inherently "normative" and it is going to lose to sales comparison on the basis of postive-normative and least variance EVERY TIME. That's why there is no sense in developing it.

So, some appraiser takes one million in cost and "simulating" a process no one uses, "norms" that into $700k, and then there is bunch of comps at $850k (or the appraiser "backed in" to $850). Either way who cares? I would never rely on an appraisal for market value based on the alleged "cost" of the subject.

You take Cox's article a couple of steps higher, I think, and provide better reasoning. I'm an incurable cynic. I don't believe anybody just because he's got credentials and has survived a long time in any given field. I know too many old pilots who have merely survived several decades. Did they survive because of a high level of skill and good judgment, or did they survive by avoiding situations requiring skill and judgment? Many of them survived by using political skills to land low-risk contracts, on which they sat for a couple of decades. They built longevity, but not experience.

I think many very senior appraisers operate the same way. They have good interpersonal skills and highly-developed political acumen. They use these skills to get a foot in the door for high-level assignments, then parlay those assignments into professional recognition whether they were done correctly or not. A certain zone salesman comes to mind, but just about all professions contain this type of individual. And the main problem arises when these people start to believe their own propaganda. Then they write articles, and the rank and file believe what they say because of their "credentials" and who they know. Thus, fiction and bad practice become reality and peer practice, e.g., across-the-fence theory, and the three *independent" approaches mantra.

For your information and comfort, I don't recite mantras just because that's what everyone else does. I describe the market. In the case of my demo, the buyers of all three of my comps purchased their properties to have a place to work. There was extra rental space partitioned off already, and people wanted to rent it. So the owners rented it month-to-month at about $0.50 - $1.00/ft^2. Most equity cap rates were 3-6%. Of the 96 commercial properties in the CBD, 17 were vacant. No property in the CBD had EVER sold for more than $200,000. I think the top sale was recorded at $186K, and it was a sham value. The owner donated it to a local church, and certified that value as its worth for purposes of a tax deduction. I had appraised that one at $130K, two years *after* the "sale." In this small market, the SCA absolutely gives the best indication of market value.

Jim,
I thought I had the Cox article. I thought I had every cost approach article ever printed.
If you can send one along, I'd like to look at it again.

Here is one a like from a 1993 article written by a current ASB member.
"Though appraisers continue to make depreciation estimates for use in the cost approach, it is questionable whether such estimates help to estimate market value for most properties. The tenth edition of The Appraisal of Real Estate includes a chapter dedicated to accrued depreciation and endorses the philosophy of deducting depreciation from the replacement cost to derive a market value estimate by the cost approach (the cost approach actually involves an 11-step process according to the tenth edition, published in 1992).' There is growing evidence, however, that the cost approach is of little help in estimating market value."

I am amused by the phrase "growing evidence." I wonder what the evidence was. On the other hand, I don't recall any evidence in favor, and I wonder how we ended up with an obligation to prove the negative.