The recent Appraisal Institute appeal to HUD to apply Evidence-Based Valuation Enforcement is a (nice) move in the right direction.  If regulators and users begin to apply it, demand it — the appraisal profession will finally have to comply, and provide modern evidence-based analysis. Read this earlier post here.

Today’s “pick comps, then adjust” model is terribly outdated.

Today’s data, computer power, visualization, and artificial intelligence enable a much better valuation product than was possible even just a few years ago.  The change moves us from an art of appraisal, to the science of valuation.  From belief-worthy opinion, to measurable reliability.  It is called the science of data analysis.  Data Science!

The main conceptual change from judgment-based appraisal practice is around data selection.

Traditional practice forces data selection to rely on judgment and convenience, or even luck.  For example, a returned phone call from an agent creates more usable, in-depth information. An unreturned call may limit you to a restricted sale price and an exterior view.

Comparable selection was good enough when you had four or five comps, with various levels of verified “richness.”  It helped that the human brain does nicely in comparing up to 4 or 5 data inputs.  Beyond 7 or 8, some “technology” help is needed, even if it is as simple as a spreadsheet table or 1004 form.

EBV^© is “nice” because it has advantages and benefits. 

As I started to write down a quick outline of these advantages, this list seemed to grow — way beyond my self-imposed blog limit for today.

The ‘large’ benefits are that proper use of data provides greater accuracy and sureness.  It: 1) clarifies the scope and identification of the problem; 2) optimizes to the data/information set; and 3) clearly defines the prediction (adjustment) to the “most probable price.”

Here we look at just three of these numerous benefits:  AI integration,  qualifying education, and regulatory policy and enforcement.

Artificial Intelligence integration.  The data science process provides the structure needed for acceptable use of artificial intelligence.  This extends to issues of confidentiality and security.  It prevents AI hallucinations and senseless artifacts.  It enables consistent and clear insertion into the use of AI for workflow (agent) solutions.

Trainee appraisers.  My experience teaching data science methods is that younger career-seekers are attracted to the data science approach. To the emphasis on computers, including the game theory aspect of the bilateral oligopoly economics nature of the problem.  There is appreciation of the “art” involved with structured critical thinking and clarified judgment that are involved in data-centric analyses.  And an ability to provide concise analysis in a clear flow.  (This applies to the patterns in the UAD3.6 new residential reporting path — as well as providing a similar understandable, consistent reporting structure for commercial and other non-residential property type appraisal narrative reports.)  “Qualifying” Education emphasizes today’s technology, instead of yesterday’s measuring-tape philosophy.

Regulation and policy becomes easier to write, to enforce, encourage, and understand!  It removes the onus on a “reviewer” declaring “violation!” – where there may be just a difference of judgment.

Evidence-based enforcement can only work with evidence-based analysis.  It is nice for everyone! Learn to Measure Markets, not just Adjust Comps here.

© Copyright is Creative Commons, Princeton.  To prevent monetization of open-source intents.