Does this (in-process) Congressional Act prevent or cement bias-vulnerable appraisal practices?

[The Appraisal Industry Improvement Act (S.1635) was introduced on May 7, 2025, by Sen. Kevin Cramer (ND) and co-sponsored by Sen. Ruben Gallego (AZ).]

Traditional, legacy appraisal practices were developed in ancient ‘paper data’ times.  Simply put, the process was:  1) pick some “comps”;  2) make adjustments;  3) reconcile disparities.

Modern, data-science based practices exploit today’s complete, electronic data.

Here is the problem with this bill:  It writes the obsolete “pick comps – make adjustments” scheme into law.

If this were an automobile mechanic’s law, it would require the mechanic to make careful and “credible” adjustments to the carburetor, i.e., how to turn the manual tuning screws and set the choke range.  This even as today’s carburetors are electronic, and are calibrated and set electronically using sensors and actuators!

Manual carburetor setting relies on the trained mechanic’s skill.  But the carburetor law requires something different.  Even as the modern electronic data and controls are available — the mechanic is in violation of the “established” and enforced (by law!) method of manual tweaking and judgment.

Modernization of the appraisal process cannot be fully accomplished with a forced, outdated appraisal “opinion” practice.  It must come from today’s tools, algorithms, and intelligence (artificial), and competence (human).

If the goal is to reduce or eliminate minority bias issues, then the analytics/practice must match the modernized data construct.  Legislating legacy process and science onto modern-day data technology precludes innovation!

Technology-based methods and data — combined — are the solution to bias.  Note the two types of bias:  1) personal bias (conscious or unconscious); and  2) analytic bias, (model or algorithm).

The analytic methods required in this bill, as in the draft copy SIL25837 VGM, constitute outdated legacy practices.  In particular, bias (both types) is instigated by subjective “picking of comps.”  Similarly, “adjustments” are substantially subjective, and cannot be mathematically calculated.  (Any data scientist or statistician or economist will confirm that bias is best concealed via data selection, not the analysis which follows.)  Simply put, remember George Dell’s Rule #1: “You cannot get objective results from subjective data.”

Historically, legacy appraisal practice worked well for legacy (paper and typewriter) data.

Legacy appraisal practice works poorly with modernized, electronic data and instant availability.

Modern, data-science-based analytics – based on the proposed “searchable and downloadable appraisal-level public use file” enables appraisers and others (those not subject to ethical or performance standards) to optimize accuracy and reliability.  More importantly, a measure of risk/reliability will be feasible, which accommodates the true need of the industry, and the public good.

Evidence-based data science methods work best on optimized data selection not “cherry picked” comps. An appraisal must analyze “sales data as are available” and “all information necessary for credible assignment results” [from USPAP – Uniform Standards of Professional Appraisal Practice, SR 1-4].  Not a convenient 4 or 5.

Evidence-based data science methods do utilize and optimize “Modernized Appraisal Data.”

Any enforcement or validation of legacy analysis practices (picking comps on a “trust me” basis) defeats the entire purpose and intent of this bill.

The intent and purpose of this bill is best served by clear usage of terminology, such as “competitive market segment” in lieu of looser legacy terminology (e.g., “comps”).  Another undesirable example is “market area,” which implies/restricts that competitive sales must come from a single area or neighborhood.

Modernized data availability must be matched with modernized, evidence-based analytic methods.

We Measure Markets, Not Compare Comps.