“Other commonly accepted methods” is the fifth and final “acceptable method” per the Fannie Mae Selling Guide, for time adjustments.
Editor’s Note: Read the entire series (so far) here.
The FHFA working paper 24-07 listed “common strategies” used by appraisers, whether reliable, understandable, or not. The GSE guideline simply notes some acceptable “methods.”
“Commonly accepted,” in the law, refers to standards, practices, or interpretations that are widely recognized and endorsed in the relevant community. (USPAP?)
In these several recent editions of this series on time adjustments we have learned some things:
- Appraisal standards (USPAP) say nothing about any specific method.
- The main industry text suggests two methods: “simple linear regression,” and “paired sales.”
- Current licensing education repeats the above, in a less-than-convincing and generalistic way.
- The most commonly accepted method, by far, has been avoidance. This by simply leaving blank the simplified single box on the form. Effectively NA = “not attempted.” (Thank you, Danny Wiley, from Freddie Mac for this observation.) This is in 80% to 90% of appraisals, depending on the then-current market direction.
What does this mean?
It means that the most “acceptable” methods (my % estimates) are:
- Avoidance. Leave blank, unless a number is needed to come closer to intended users expectations (per USPAP Scope of Work Rule). [80% – 90% – per the FHFA study.]
- Grouped–pair contrasting, regardless of the lagged effect and upward analytical bias. [1% -5%]
- Matched–pair comparison, regardless of the practical impossibility of relevant data. [1% – 5%]
- POMA or aggregated index, such as Zillow. [5%?]
- Time-series analysis, graph + regression. [10% – 15%]
IS IT POSSIBLE? Paradox! It appears, by far, the most “commonly acceptable” method is avoidance. In the past, time adjustments were required, but not enforced on lender reviewers and risk underwriters.
But now it’s official – avoidance is a generally accepted method, therefore it’s acceptable to the GSEs.
Contrarily, the most responsible, reliable, and relevant method is time-series analysis, as used by economists, statisticians, and asset analysts who follow established data science principles in their practice. Factual date, factual price, visual, reproducible, and straightforwardly intuitive.
The general rule for relevance is that the result should match the relevant market as closely as possible – the actual Competitive Market Segment (CMS)©.
But what do you do if you don’t have “enough” sales?
In the next post(s), we will consider the order of reliability, ranging from the best data set – the directly competitive sales – to a broader data set — to general published HPIs (Housing Price Indexes).
We will consider:
- A trade-off of sureness versus bias?
- How much data is enough?
- What about ‘outliers’?
- Explanations?