Well, kinda, sorta.

In the discussions I often have with real estate professionals and users of their services (and mine) is the topic of precision in valuation. Its not a direct topic of discussion. It’s kind of stumbled into by accident. Coming up with a reasonable value with undefined expectations of precision can be frustrating for all parties (do we hear “managing expectations” anyone?).

Here are a few examples:

• Example 1 My parents bought a house in a subdivision in Framingham Massachusetts in 1963 (the day President Kennedy was assassinated but after it happened) for something like \$19,250 and then sold it for \$21,050 in 1967. I am not sure if I have the sales prices perfectly accurate but I know it was in increments of \$50. That is the particular level of precision that existed in that market at that time.

• Example 2 I have reviewed appraisals where the result of the valuation might be a very specific number like \$975,228, mainly because the appraiser was lazy and averaged the comps presented (not supposed to do that). The problem with inferring this type of precision is you are essentially saying as a valuation expert that you are so accurate; that you can come within one dollar of what the property is worth. I contend that there is not a human being on the planet that can be that accurate in valuation. \$1 / \$975,228 = 0.0001%.

• Example 3 Very high-end properties in my market are usually negotiated in 6 figure increments. In Manhattan, properties over \$20,000,000 tend be sold in increments of \$500,000 to \$1,000,000. Our firm had inspected properties last year that sold for \$41,000,000 and \$52,000,000 (and both need about \$10,000,000 worth of renovations) so this price range has been fairly active.

One of the first things I look at is the general price range of the properties being observed.

In my subdivision house example 1, \$50 in a \$20,000 (0.25%) housing market was the precision that the market expected. That’s pretty tight. In my second example, the precision at the near \$1,000,000 price point in my market tends to be about \$25,000 (2.5%). In my third example, the expectation is about the same at 2.5% (\$500,000 increments at \$20,000,000 price levels).

So in my market there is an expectation level of a precision level of about 2.5%. Our firm does relocation work where we are asked to anticipate the market value of a property when sold. The last time we were given a report by one of the firms we do business with, our accuracy level is 3% (the next closest firm was 10%). So this means that we are pretty consistent with the precision expected in the market. If I was valuing property in 1963 where I grew up (for the record, I would be 3 years old), then I would be horribly inaccurate at 2.5% because the increments would be in \$500 steps rather than \$50 steps.

The reasons for valuation inaccuracy are numerous but here are the biggies:

• Blinded by one-sided information – the appraiser or broker relies on information provided by the property owner, who is already biased towards their property being worth more. Appraisers who only use comps provided by the broker in the sale are not providing an independent valuation for the lender, who hires them to access the collateral.

• Lack of information – limited current data, or access to relevant data like listings and contracts in addition to closed sales make the results much more inconsistent.

• Little understanding of amenity differences – For example, understanding locational differences such as neighborhoods, subdivisions, cul-de-sacs, busy streets, school districts, etc.

• Using out of date rules-of-thumb – There are some who use rules or experience gathered long ago and do not continue to modify their experience in understanding variances in the contributory value of amenities.

• Inability to read buyer and seller’s minds – I’ve been working on this by taking vitamins but it hasn’t worked. The message that buyers and sellers give a broker or an appraiser can be very different than what actually motivated them to agree to the sales price.

• Lack of experience – Raw data doesn’t tell the whole story. Someone who is immersed in a market will stumble on information that less experienced “experts” would have. Data is data but interpretation separates the hacks from the professionals. Automated valuation models (AVMs) [Soapbox] are data crunching programs that spit out values for a property for lenders and online services such as Zillow that provide online values for consumers. They may or may not take you to something close to a reasonably accurate value. If they do, then it’s more likely to be a coincidence and that’s probably not good enough for the user IMHO. However, both services provide a “number” and the assumption by the user is usually made that if a valuation result is in writing, then it must be accurate (read the National Enquirer lately?)

Sellsius makes a strong case for using your senses in valuation in their post I See, said the Blind Man to the Deaf Lady [Sellsius]. However, I feel strongly that the post should be called “I see said the blind man as he picked up his hammer and saw”, but that’s another topic. 😉

Here’s an excerpt:

There are 6 things an AVM (including zillow) cannot do but which impact value are:

* See
* Hear
* Smell
* Taste (this might be a stretch, but some air can be tasted)
* Touch
* Interpret

Precision is a big topic to cover and it affects virtually every real estate transaction. There is plenty more to discuss on the subject, but what the heck, this post is good enough.

Tags: