Appraisal Bias

[J Grant]

I don't think that is what he is saying. He is saying that within a range that cannot be reduced, then each value within that range is equally valid. It is only YOUR opinion, and not any FACT, that determines this final single value. The FACT you may choose an average, a median, a contract price, or select the exact value of a single sale within that range, either a raw sale price or an adjusted price, is based solely upon your heuristic judgment, not on any given facts or calculated number.

What he is saying is wrong because he is not looking at it as an appraiser, which is as you say, and as USPAP says, is our OPINION ( developed in the appraisal

Of course, it is not a real-world fact. This is a fundamental of appraisal. as a model of value. What goes wrong is when the reasoning Alebrewer uses conflates the two.

I disagree that in an appraisal each value within the range is equally valid. If our subject is highly upgraded, for example, and closer to the $ price of sale 1 which needed the least adjustment and was also highly upgraded, then it makes sense to opine our reconciling closer to sale 1, rather than the rote , generic well gee any number can be it they are all the same.
Our client is hiring an appraiser for their professional judgment where to reconcile, not some mindless dart thrower who can not comprehend why one reconciliation point is better or worse wrt support than the other.

 

[Terrel L. Shields]

some mindless dart thrower

actually that is what we become if there is no substantial similarity between the subject and at least one of the comps. Can we say a Colonial subject with no nearby Colonials is "more similar" to the 2 story conventional comps we are limited to? Perhaps we have to make the case that each sale is equally dissimilar to the subject. At any given moment the data available to us is better or worse than the data was a few months ago. 2 years ago, perhaps a model match to that Colonial sold...but how useful is that (especially considering the chaos of the market in the past 3 years) and would we even search that far back?

Cookiecutterville and remote rural property have little in common. The "typical" sale in a given population (or neighborhood) does little to describe the outliers. Where do we even start with those? As I pointed out it is your heuristic judgment that decides that final number, and it is nothing more. Your peers might well chose a different metric, a different comp, etc. to "support" their opinion. Again, they nor you, will ever "prove" your opinion, you can only "support" it with your judgment and do so by stating what you keyed upon to get to that conclusion. But it does not make it "right" nor "wrong"... which means it is a randomized selection based upon your past experience, gut feeling, curb appeal, intuition, or whatever you choose to call it. In other words, it is a choice you make in your head then seek the support for that decision. It is your Personal AVM in your head.

 

[J Grant]

actually that is what we become if there is no substantial similarity between the subject and at least one of the comps. Can we say a Colonial subject with no nearby Colonials is "more similar" to the 2 story conventional comps we are limited to? Perhaps we have to make the case that each sale is equally dissimilar to the subject. At any given moment the data available to us is better or worse than the data was a few months ago. 2 years ago, perhaps a model match to that Colonial sold...but how useful is that (especially considering the chaos of the market in the past 3 years) and would we even search that far back?

Cookiecutterville and remote rural property have little in common. The "typical" sale in a given population (or neighborhood) does little to describe the outliers. Where do we even start with those? As I pointed out it is your heuristic judgment that decides that final number, and it is nothing more. Your peers might well chose a different metric, a different comp, etc. to "support" their opinion. Again, they nor you, will ever "prove" your opinion, you can only "support" it with your judgment and do so by stating what you keyed upon to get to that conclusion. But it does not make it "right" nor "wrong"... which means it is a randomized selection based upon your past experience, gut feeling, curb appeal, intuition, or whatever you choose to call it. In other words, it is a choice you make in your head then seek the support for that decision. It is your Personal AVM in your head.

That is correct, it is our opinion based on the current data and the appraisal SOW as well as drawing on past experience in the market.

But we should be able to present it in a way that is clear to our readers why we choose X $ as the value and that shuts the door on other, less credibly supported options.

 

[Terrel L. Shields]

that shuts the door on other, less credibly supported options.

of our own. Other appraisers may disagree with the final number as they may apply different metrics

 

[J Grant]

of our own. Other appraisers may disagree with the final number as they may apply different metrics

I am not responsible for what another appraiser does.

I just need to support my own conclusions and value opinion.

One can gauge their work over the years in the market - did my trend hold up to events, did my market value opinion for a refi that sold 3 months later hold up , one can get confidence that way, we are not expected to be perfect and any of us can produce a dud , but overall our values and reason for choosing some comps and eliminating others should make sense and when that all makes sense it leads to a well and credibly supported market value opinion, and that includes a reason we reconciled at X $ point value ( and not the 1000 other price possibilities goes without saying)

I reject the argument that any value along the range is as good as another - Reconciling to a point value is not a complex math problem, and we are not trying to forecast from the myriad price possibilities. All we are doing is demonstrating in our appraisal there is a credible supported reasoning for our reconciliation point value. If an appraiser does not believe in point values as being able to be developed and supported in an appraisal then they should not be doing point value appraisals.

 

[J Grant]

I can only speak for myself, but the final reconciled 'point value' is usually the result of rounding or weighting some external factor (contract for sale for example). I will also reconcile to the median or modal tendency of the adjusted comparables. That said, any other 'point' within my reconciled range would be just as supportable as any other point.

Maybe think about it this way and see if it helps. Would you adjust a $500k sale by $1 for some element of comparison? The answer should be a resounding 'no'. If we are in agreement with that, then what is the 'minimum' adjustment you'd make for an element of comparison at the $500k price point? $100? $500? $1,000? For me, rounding adjustments to the nearest $100 tends to work pretty well. If that is the case, how can someone state that they're capable of figuring out a property's value to the nearest dollar when they're not even adjusting comps that precisely?

Clearly, I and others do not adjust by 1$, or $100, why are you presenting tiny amounts and silly examples? Nobody makes such tiny adjustments, and yes they are rounded, the market does not slice and dice that thin and adjustments reflect overall contributory value..

You keep conflating different things and different steps on the appraisal with each other. Making and applying Adjustments is one step, and reconciliation of point value at the end is another step .

. Just because our adjustments are not that "precise" to a dollar, does not carry over to well gee then how can we be that precise in our reconciliation? We have to pick one finite number for a point value, and it should make sense both in rounding and in why we made it. I reject the argument that any number is the same along a range ( see my above post on why)

Do you write in the report, your actual thought process, one number is as good as another? And that you don't believe in point values? Or you write something different than what you really think? ( appraisal / reporting USPAP)

Statement of reconciliation" "I don't believe in point values, they are spurious, but I have to pick one to get paid so here it is. I have no reason for picking my point value since any number in the range is as good as the other one., but I had to pick a number to get paid so here it is."

The above would be an honest statement, yet I doubt it is presented. I, on the other hand, write the same on an appraisal as I argue here, wrt the reasons behind why I choose a point value and that it is the better-supported choice (from the range ) and that yes I believe in point values as a thing in appraisals, so how I think and how I report is consistent.

 

[Zoe]

CFPB and FHFA could come down on TAF, ASC, State Board. Literally nobody is out of the reach of that panel on these appraisal bias hearings. Look who is on the panel. Every bank regulator, etc, etc.

 

[Tom D]

i agree with a value range. however, the underwriter, lender, has lines drawn in the sand. rate changes at 75%, pmi is asked for over 80% ltv. for this reason, with a range only, an underwriter would have to pick a point to determine ltv ratio. i'm not doing it as an underwriter, above my pay grade. that is the reason for a pov by a 3rd party. with waivers the ltv is so low that the risk is acceptable not having an appraiser pov. with avm, it's FNMA ltv problem.

 

[J Grant]

i agree with a value range. however, the underwriter, lender, has lines drawn in the sand. rate changes at 75%, pmi is asked for over 80% ltv. for this reason, with a range only, an underwriter would have to pick a point to determine ltv ratio. i'm not doing it as an underwriter, above my pay grade. that is the reason for a pov by a 3rd party. with waivers the ltv is so low that the risk is acceptable not having an appraiser pov. with avm, it's FNMA ltv problem.

Another one who should not be appraising since you do not understand why a point value is needed - thanks to you and your ilk argument against the lifeblood of our profession ( point values ), the GSE have taken away half the work and the AVM provides the range you love so much, and a lender picks the point value (for a waiver now called value acceptance) .Since so many appriasers it turns out, do not believe in a point value. Who knew?( sarcasm ) since they signed their name next to a point value all these years that wink wink they did not believe in. Shoujld have all turned in your license years ago then. But you didn't, just kept undermining he profession with your nonsense about it should be a value range.

Are you happy now ? All of you? You got your wish. The res side of the profession did as you suggested, and the appraisal profession has lost now arpox half of its volume.to waiver/value acceptance, the range you kept agitating for, I warned on this board years ago that if appraisers would not provide a point value or kept agitating against it, somebody else would pick the point value -it has happened
Thank you so much for decimating the appraisal profession on the res lending side. Because that is where this argument led. .

 

[djd09]

...with a single point value you are either a number hitter or racist...pick one