Diversity can be a goal in itself when diversity means breaking down barriers of groupthink and apathy. An expansive concept.
Fairness, on the other hand is elusive, demanding definition. Fairness depends on the very point of view it attempts to define. “Fair” is in the eye of the beholder.
EBV (Evidence Based Valuation)© (or EBA, Evidence Based Appraisal)© does not directly address fairness, diversity, prejudice or personal (conscious or unconscious) bias. EBV does significantly uncover a foundation where both types of bias are identified, addressed, and even eliminated.
Traditional appraisal (and “evaluation”) is vulnerable to bias. Its foundation, “picking comps”, is subjective and personal from the start.
AVM (“Automated” Valuation Models) have evolved as an industry, not as a particular type of algorithm. The internal algorithms are considered proprietary secrets to be kept from competing AVM companies. Whoever keeps the best secret wins. Unfortunately, any bias, assumption, or algorithmic result is hidden, even if it perpetuates unfairness and conceals any prejudicial bias.
The solution is simple, yet elusive – gridlocked by the five forces of friction.
The mass elusion is built-in. Our prior Analogue Blog articles identify the five friction sources which collectively block any real hope of change. These include the appraisal process, appraisal standards, appraisal education, client expectation, and appraisal regulation (including 54 jurisdictions, each with different forms, fees, administrative efficiency, and competency).
It is well to note that the five frictions are actually five forces of vortex, mutually pulling each other toward the old subjective ways of doing things. Changing any one or two forces will simply return to the fixed, obsolete, and outdated way of providing a product which does not really solve the needs of the public trust. (Recall, that we have had a crisis of finance and real estate every 12-15 years, regularly.)
The noted subjectivity and secrecy perpetuate the valuation process in ways vulnerable to personal bias. Worse yet any personal bias (intentional or unintentional) is hidden by the analytic bias which can be used to cover or hide or simply ignore actual religious, race, or other prejudices.
So long as data is personal-judgment based, vulnerable to analytic bias, diversity is discouraged.
In my stats and graphs classes, I often ask: “Is it possible to lie with statistics?” Heads shake yes. Then I ask: “Is it possible to lie with graphs?” More shaking heads. Finally, I ask: “Is it possible for lie with words?” Silence. Then. The question: “What is the one thing in common with all three?”
Only the person doing the communication. The person.
So long as the valuation process starts with subjective data selection – or via secret preprogrammed neighborhood search algorithms – the result carries the original data selection bias.
The vortex of subjectivity and bias resides in every one of our five forces of friction: 1) the process; 2) the standards; 3) education; 4) client expectation; and 5) regulation. These reinforce and stabilize each other.
Until this vortex is resolved, it will be difficult to convince many toward diversity, objectivity, and real consideration of the three social/economic impacts: consumers, taxpayers, and opposing viewpoints on social justice.
May 4, 2022 @ 10:42 am
Traditionally, we were given three approaches to solve the appraisal problem. By applying each one independently, in theory, we should be presented with a set of answers to weigh the strengths and weaknesses of the approaches, data, etc., and hopefully, come to reasonable (and logical) conclusions as to the value of the property. Over time, in the name of efficiency and cost-cutting, the cost and income approaches (for residential) were generally set aside. The sales comparison approach was considered the only meaningful (and market-based) method to reflect the actions of buyers and sellers.
By abandoning the other two approaches, we gained efficiency but lost the “checks and balances” originally inherent in the valuation process. In my opinion, that’s where we went off the rails. The similarities in the results of the three approaches singled were “evidenced-based valuation,” simply from other sets of criteria. If the three approaches yielded similar results, you have multiple methods supporting a narrow potential value range. If the results were dissimilar (and each approach was developed independently and objectively), you better understand the forces acting upon value in the marketplace and the issues behind those forces.
The problems caused by a lack of checks and balances in valuation are evident. I think Georges’ suggestion of using data (EBV) stands as a “check and balance” methodology for valuation problems. I don’t see it as a replacement for the sales comparison approach (and I don’t George promotes it as such) but rather a supplemental approach to augment the process. I don’t consider processed data alone an answer, and I don’t think George would advocate that either. However, I think we all need to recognize the lack of checks and balances in what we do and the use of processed data to supplement our judgment.
Processed data will tell you that two Toyota Camrys cost the same and are worth the same amount of money. An experienced salesman will tell you the white one will sell quicker and for more money than the purple one.
Just my two cents … George, as always, thanks for making us think about things.
May 5, 2022 @ 5:31 am
I agree and if we have a checks and balances, then that should also help to prove if BIAS was or was not a part of the sales comparison approach to value (picking comps). I still do a cost approach on most of my reports. Not the income since most of what I do is owner occupied. The most important part of that cost is land valuation, so sales come into play but it is imperative to know that land value as that makes the difference between where that Toyota sits! Same Toyota, different area and land values.
February 24, 2024 @ 11:20 am
Yes, so long as the analytic bias enables the personal bias, we will not have clarity on the underlying issues.