Recent GSE (Fannie Mae and Freddie Mac) guidelines for appraisers and lenders make some assumptions.  In particular we can use, for example, the requirements for market analysis and time adjustments.

USPAP (Uniform Standards of Professional Practice) Rule 1-4 states that an appraiser must “collect, verify, and analyze all information…and must analyze such comparable sales data as are available.” [Emphases added].  Nowhere does USPAP say use the best 3 or 4.

In short, an appraiser must use all relevant data.  The key words are:  all and relevant.

And the GSEs?  The entire large set of guidelines make quite the contrary statement.  The GSEs claim they cannot require a particular method.  However, they fall back to historical, outmoded methods based on the “Pick comps, make adjustments” paradigm.  (On the other hand, the modern, evidence-based approach provides superior accuracy and precision, and can even be made to measure reliability/risk on individual appraisals.)

We can note in their (GSE’s) defense, that the “pick comps, make adjustments” paradigm continues to be the “established practice” in education promoted by the professional organizations (such as the Appraisal Institute), as enforced by the state regulatory agencies.

Even this assumption of using 3 comps – is embedded in the FannieMae Selling Guide.  In the section entitled “Minimum Number of Comparable Sales,” the required minimum is “three closed comparables must be reported.”  “The subject property can be used as a fourth comparable…”

These statements do not preclude the appraiser’s use of more sales.  But these statements do not positively encourage the analysis of “all comparable sales data as are available” – as required in USPAP.

A following section states a controlling “should”:  use closed sales within the past 12 months.  What is ignored (if we apply modern data science methods) is that, properly done, market-analyzed time adjustments are highly reliable, and can present otherwise much more similar sales.

To sum up:  The USPAP development standard requires the use of all available relevant sales, but qualifies it by “worthy of belief” (credibility).  It does not mention the need for reliability or accuracy.

GSE guidelines, however, do ask for reliability and accuracy.  Maximum reliability and accuracy require the use of  all available relevant information.  Unfortunately, the GSE guidelines are built on the outdated paradigm of “pick comps, make adjustments.”

The answer?  Evidence Based Valuation (EBV)^©.  EBV is based on simple data science principles:

• Include all relevant data (the “population”).
• Apply the three basic predictive (adjustment) algorithms.
• Measure or estimate reliability, using repeatable methods.
• Deliver more than a point-value opinion.

Judgment, knowledge, and skills do not go away.  The new needed judgment and skills are just different.

Market Price Indexing (MPI)^© as taught by Valuemetrics.info, does not require specialized software.  Existing spreadsheets or analytic software such as R (free open source) can be learned.  Once understood, reliable and presentable graphs and results can be done in minutes or seconds.

However, appraiser skills and knowledge are needed.  Judgment is still required.

The needed complete data judgment, and software are available in free webinars GeorgeDell.com/free-stuff.  Or in the complete introductory class Stats, Graphs, and Data Science 1 CE class.

Graduates of a recent SGDS1 class can join us in the CAA (Community of Asset Analysts).