, , 1

Time adjustments are now required by Fannie Mae and Freddie Mac.

Time adjustments were previously required “when necessary.”  Now a “market conditions” adjustment is required on every appraisal.  This means that even if personal judgment says the subject market is “stable,” an analysis must be performed to arrive at the zero adjustment conclusion.

An FHFA (Federal Housing Finance Agency) paper (24-07) mentions three “common methods” used for time adjustments:  “grouped data,”  “paired sales,” and “indexed similar-sales prices.”

We will consider each of these methods, one-at-a-time, in coming editions, followed by the “acceptable” methods per the Fannie Mae Selling Guide for lenders.  Finally, we will list specific Selling Guide requirements, as well as the “ideal” analysis for time-series analysis in the real estate context.

Note:  The ideal/optimal model emphasizes accuracy and reliability, not just credibility/believability.

First, we consider “paired sales.”

Per the “Appraisal of Real Estate” 15th edition, “paired sales” assumes you can find two properties, equivalent in all respects but one – the difference in price then measures a single difference amount.

This appears as a good theory.  Does it work in practice?

No.  Even if it’s a re-sale of the same property.

Yet another magical theory.

No, each property is always unique and different from every other property, even if it is just a matter of location.  Also, transaction circumstances are always different.  The buyer’s motivation is different.  The seller’s motivation is different.  And the color of the kitchen is different.

The best any valuation method can ever do, is to admit that there is a “transaction zone” within which individual buyers/sellers negotiate.  A game, not an economic equilibrium point value.  Variation.  In fact, differences in motive alone, will overwhelm any slight trend or change over a few months.  A transaction zone will typically be 2% to 5%  wide – even in a case where the trend change is less than 1% over (say) a 3 month time difference.  Darn.  The variation overwhelms any hope for a reliable number.

The ARE, 15th edition, goes on:  “Paired data analysis should be developed with extreme care to ensure that properties are truly comparable and that other differences do not exist”!  Only “truly comparable” data is allowed.  Got it:

  • Show extreme care. (So, less care is OK in other areas of appraisal?)
  • Use only “truly comparable” sales. (Sorta comparable is a no-no).

The text goes on to explain:  “Appraisers try to use several paired sales to support the adjustments and prevent an unknown factor from contributing to a misleading conclusion.”  So:

  1. Be extremely careful, for all adjustments;
  2. Use several pairs, to prevent unknown factors;
  3. Use truly comparable sales. (A term truly undefined anywhere!)

Finally, the ARE, 15th edition, states:  “These techniques can be time-consuming and sometimes difficult to apply with a high level of mathematic precision.”  [Actually, mathematic precision does not exist here.  It is magical “point value” thinking, applied to justify a theory of “support.”]

The next “common technique” is “grouped data,” which is essentially the use of several pairs, then averaging the differences.

All this assumes that each “pair” time span coincides nicely to the comparable date span.

A final note:  If a “perfect pair” exists … at least one sale would have been a perfect comp in the first place.

“WE MEASURE MARKETS, NOT COMPARE COMPS.“