Fun insanity — time adjustment requirements made simple. A concise market conditions guide and summary.
Editor’s Note: This is part 7 of a new series on Time. Get caught up here.
Let’s outline the new GSE requirements:
- Time “adjustments” will be a required field, even if exactly zero! (Zero is a number.)
- Time-series analysis is an established method in economic measurement.
- The appraisal challenge is the varying relevant data sizes.
- The path to solution is simple and repeatable.
- Judgments required by the analyst are:
- What time-span is appropriate?
- Is there a change in the market trend?
- Are there any outlier sales to handle?
- How do I calculate dollar/time adjustment?
The established, accepted method in financial and economic analysis is called “time-series.” It’s different from other adjustments. It’s about the market segment , not location, not preferences! Not insanity.
- Time is measured in days, to the contract day of the comparable sale.
- Dollars is the time-indexed price, the same as other adjustments.
This series of blogs intends to consider each of the claimed “methods,” in the context of USPAP, GSE guidelines, licensing curriculum, and FHFA/regulatory realities. We will end where we start: What is the method which brings the best accuracy, relevance, and understandability to market conditions adjustments, in the context of good science, critical thinking, and ethical intent.
In this blog series on time-series, we intend to consider the different methods suggested by the GSEs (FannieMae and FreddieMac), the FHFA (Working Paper 24-07), The Appraisal Institute, and others (including some “word salad” prose), regurgitated from the use of artificial intelligence by those who do not have human intelligence nor education in economic/statistical time series analysis.
Most of the advice given today is questionable. At best, it is vague and confusing. It lacks step-by-step instructions on a reliable, repeatable method which meets the requirements below:
- Only the similarity-specific market segment is used. (Anything else measures something else!)
- The time-span included is directly related to the comparables (market segment) used.
- Trend changes must be clearly identified and reflected in the solution process.
- Outliers and anomalies must be identified and properly handled.
- Reporting methods which include quick illustrative explanation.
Finally, individual comp adjustments should be easily put into existing and coming “dashboard” forms.
We will focus on the GSE requirement of two separate calculations plus illustration:
- An “overall” trend in the neighborhood section 1004 form: “increasing, stable, or declining.”
- Specific “market derived time adjustments for changes in market conditions.”
- “Illustration of the methodology used to determine specific comparable sale time adjustments.”
We will investigate each of the claims of an “accepted” way of getting a time adjustment. Most are bogus. We will end with the simplest, most direct, and fastest way to “acceptable” market analysis.
Avoid the insanity of legacy magic, AI word salad, and clever make-believe “accepted” methods.
Go now — GeorgeDell.com to get this series, and access to free classes and fun and full CE education.
February 14, 2025 @ 7:48 am
I have been making Time adjustments since 2007. I use every sale of every home in the neighborhood boundaries in order to squash the influence of outliers. My logic is “if there is a change in trend, it is an over-arching reality that affects all sales so, all sales should be considered in the measurement.” I do a year-over-year comparison split into quarters, for example, Q1 2023 compared with Q1 2024 = the variance in median prices. I arrive at 4 variances (sometimes 5 if I am more than 1/2 way through the current quarter). I then determine a mean average of the variances indicated and calculate the rate of change on a monthly basis. When I do this, I get very reasonable results. No wild numbers because of big variances in sample size and no wacky indications because of outliers in the data set. And my conclusions are always put to the test in my reports because I always profile one or two listings in my sales comparison analysis. If my Time adjustments are off, it will show-up because the indicated values of the listings will be off. But they never are. This is a simple method, and it has served me well for almost 2 decades since I implemented it.