Extracting an Adjustment – One Way to Measure


WRE, Working RE Magazine, Appraiser News, Appraiser Magazine, Real Estate Appraisers, Volume 35
>> See Past News Editions
>> Click to Print
Just Updated
2015 Expert’s Guide to Defensible Workfile

“Bulletproof” your Workfile
protect your license and your livelihood with expert help. Make your time count: learn quick tips from an expert on how to produce highly defensible reports that keep you safe and raise the quality of your product. Click now.

Just Updated

2015 AMC Guide
The JUST UPDATED 2015 AMC Guide helps increase orders so you can turn down low-fee AMC work, negotiate fees from a position of strength and confidence and refuse to work for the AMCs that undervalue you. Start 2015 off with a new book of clients! Work with the best and fire the rest. Click now.











OREP E&O: Rates lower for most insureds.

Editor’s Note: Author Rachel Massey offers more help on how to make supportable adjustments.

Extracting an Adjustment – One Way to Measure
By Rachel Massey, SRA, AI-RRS

Because I often get calls from both Realtors and homeowners asking how much a certain feature in a home is worth, I thought a brief discussion of one method of extracting an adjustment from the market might be worthwhile.

This method is described in detail in The Appraisal of Real Estate, 14th Edition (as “grouped data analysis” starting on page 398) and is not a new technique, but one that appraisers may find useful in their daily practice. It can work well because if the appraiser uses care in the isolation process, the sheer number of sales will provide a range of answers, which can then be used for extraction, and support of that particular adjustment.

Instead of writing about theory, I think a simple example from my market is a good starting point. I work in a market where there are usually enough sales to use this method, but it can be useful even in markets where data is more limited. I have to go back two plus years for most of my studies to get enough data points for an adequate sample. This is not perfect but it does work for me when determining certain adjustments, such as basement finish, basement versus no basement, garage stalls, and swimming pools. I have not found it to be very effective with gross living area and it has had mixed results with bathroom counts. There are drawbacks to using it, mainly that the underlying site value is not extracted, but if the sales that are used for the study are relatively similar, the volume of sales generally resolves the issue.

The following show two different examples of an extraction for basement finish; one in my main market big-city area, related to a generally newer house in the $400,000 +/- price range, and the other in an outlying district about ten miles away, in the under $200,000 price range. Both use the same methodology and show substantial differences in results, which is why an appraiser cannot just provide a number or a percentage when asked. Instead, the appraiser has to look at the market.

(story continues below)


(story continues)

For the first example, I went back over two years and narrowed my market data to houses between 2,000 and 3,000 sq. ft., built between 1990-2010, on the west side of my market area, and then downloaded all these sales to Excel and segmented the sales between houses with finished basements and those without. The results included 37 sales without finished basements and 62 identified with finished basements. Here is what it looks like on a spreadsheet:

I then looked at median and average sales price differences and median and average amount of basement finish, and came up with between $21,647 and $24,500 difference in price, favoring those with the basement finish and between $24.24 per sq. ft. and $27.75 per sq. ft. of basement finish. This provided me with some support for whatever adjustment I considered most reasonable. This would be anywhere from $21,500 to $25,000 based on sales price differences, or between $24 and $28 per sq. ft. of finished space, if used in that manner.

From experience, I know that basement finish typically costs around $40 per square foot in this market, which suggests that both physical depreciation and functional obsolescence are playing a role here, since the difference is more than what would be expected from physical depreciation alone.

(story continues below)

(story continues)

For the second example, I used data from another proximate market with my target properties between 1,200–1,700 sq. ft. in size and built between 1985-2010. I also went back just over two years. I had 48 sales without basement finish and 36 with basement finish, and the median difference in price was $8,953; the average price difference was $14,420. Here is what it looks like on a spreadsheet:

The median size of finish was 625 sq. ft. and the average size of finish was 703 sq. ft., supporting adjustments per sq. ft. of $14.32 to $20.51. This means I could be comfortable using adjustments anywhere from $9,000 to $14,500 for the basement finish as a whole, or between $14 and $21 per square foot if I chose to address it that way. This data gives me something to work with and in the end, I use my experience in the market and what the comparables are telling me for my final determination, but I have support for whatever I do.

As you can see, there are differences in price between the areas and the sizes, as would be expected. Cost remains about the same to complete. Each appraisal may be different, and the numbers presented in these two examples could change depending on how far back the appraiser goes when collecting data and what they set as the perimeters for the data search. I offer this to fellow appraisers as a simple study showing how I often go about trying to extract an adjustment from the market.

How to Support and Prove Your Adjustments
Live Part 2 Tomorrow!
Part 1: Available Now
Part 2: March 12th, 10 – 12 p.m. PST

It was a great webinar, now I need to redo all my reports for the last 30 years!” – Sharon L.

Updated and expanded, Richard Hagar, SRA shows you how to properly support your adjustments- the foundation of good appraising! Regulations state that appraisal adjustments cannot be based upon an appraiser’s opinion. Failure to provide proper proof and analysis to support your adjustments means a rough road ahead: state board complaints, panel removal, lawsuits, even license revocation. Fannie Mae cites “the use of adjustments that do not reflect market reaction” as the number one reason an appraiser can be “blacklisted.” This training is critical in helping appraisers avoid catastrophic appraisal failures. Sign Up Today!

February (recorded): Fannie Mae’s CU: Understanding Quality and Control Ratings

About the Author

Rachel Massey, SRA, AI-RRS has been appraising full-time since 1989. She is a certified residential appraiser in Michigan, specializing in review work for various clients, as well as lake properties, and other residential properties, in and around the Washtenaw County market. She is an AQB Certified USPAP instructor.

>Click to Print
>New: Collateral Underwriter Blog: Find answers, offer solutions.
>Opt-In to Working RE Newsletters

Send your story submission/idea to the Editor: dbrauner@orep.org


Tags: ,

Comments (18)

  1. Use your comp grid page. In my opinion, paired sales/regression analysis is most relevant. In a perfect world, the only difference in 2 or more sales would be finished and unfinished basements of similar sq.ft. however, we do not see that often. In regression analysis, after all other warranted adjustments are applied, the remainder of the two paired sales is compared. ie $275,000 finished and $260,000 unfinished. The indicated extractable adjustment would appear to be $15,000. Using multiple, relatively small data sets is recommended as this helps to provide a more accurate, measurable market reaction to differences in utility due to variable levels of construction quality, materials, etc.

    - Reply
  2. by Dave Harper, SRA

    Dear Rachel, this is intended only to bring a different perspective. Statistically, it is more correct in example one to say $24.24 per square foot of Average Sq.Ft. of GLA and $27.75 per square foot of Median Sq.Ft. of GLA, respectively. It is more appropriate to calculate the Sold Price per Sq.Ft. of GLA for each transaction, then summarize by Average (Mean) and/or Median, and then to compare the Mean Sold Price per Sq.Ft. of Unfinished to the Mean Sold Price per Sq.Ft. of Finished, and to follow suit with the Median Unfinished vs the Median Finished. Even without the entire data set to vet, it is reasonable the results will be different. After all, when the data changes, the answer changes.
    If the Mean Sold Price per Square Foot Unfinished is compared to Mean Sold Price per Square Foot unfinished, it would be appropriate to state the difference as Average price per square foot difference. Similarly, if Median Sold Price per Square Foot Unfinished is compared to Median Sold Price per Square foot Finished, it would be appropriate to state the difference as Median price per square foot difference,
    In Example one the difference of in averages of $21,647 is divided by the average GLA which is only a statistic summarizing GLA.. Similarly, in Example one, the difference in medians of $24,500 is divided by the median GLA, which too, is only a statistic indicating that 50% of the GLA’s are above 883 Sq.Ft. and 50% are below 883 Sq.Ft..

    My two cents.

    Respectfully submitted,

    Dave Harper, SRA

    - Reply
  3. Hello Michael, yes our market has been appreciating but by including enough data points this should be dealt with under the median and average, in particular if you get rid of sales that you know are outliers. It isn’t perfect, but it is simple, and is a means of support. Ultimately it is the appraisers judgement that comes into play as to what is most supported, but this provides some level of support.

    My next blog post is going to be about another method which can be used as well. Using more than one method is helpful, and the different methods often lead to similar results that the appraiser can then reconcile.

    Thanks for your post!

    - Reply
  4. As a retired MAI, it is heartwarming to see so much interest in conducting the research necessary to make objective adjustments. Keep up the good work.

    - Reply
  5. Rachel, I like this method for its simplicity & I commend you for presenting it so well. You went back 2 years for sufficient data. How then do you account for time ? I see this working well in a stable market, however, there could be significant change in median value and a disproportionate number of sales during market spikes. I’ll definitely put this to work, especially during stable market conditions.

    - Reply
  6. Excellent article. Thank you.

    To those that point out flaws. The system is not 100% foolproof. It is simply one more tool in our box that we can (sometimes) use to determine& support specific adjustments. One can always find a flaw in others methods, but unless one also has an alternate method without that flaw, then the criticism rings hollow.

    I lean toward percentile adjustments converted to dollars based on experience and other participants views for subjective variables like view amenity. For easily quantifiable areas like GLA I use combined or compared extraction and depreciated replacement cost estimates. The extracted value is market based since I CAN separate land value based on recent reassessments which SPECIFY land value (or rate). What remains is a price per SF of improvement area. I can quibble abut what percentages of that are attributed to view etc but unless YOU can prove to me what those percentages are, my method is fine when coupled with a comparison of depreciated replacement cost. IF I have CORRECTLY analyzed DRC then my value per SF IS a market value rather than just a replacement cost. Usually I find that number and the extracted number to be reasonably close-like the range cited in the article above. In my area except in very few circumstances at the very low end (size and value wise) individual room count adjustments just are not determinable outside of GLA.

    - Reply
  7. Rachel,

    Your excellent article describes an efficient, practical method to support adjustments. Practical tips, such as yours, are more essential than ever as residential Appraisers adapt to CU. I especially agree with you that using appropriate methodologies are very important; however, the market knowledge, judgment and experience of the Appraisers who reconcile the results are absolutely essential to developing credible results.

    I think your advice is much better than the $20, $10, and even $5/SF basement adjustment factors that are still in use today… because that is what has always been used. As Fannie Mae directs Appraisers to apply market-based adjustments supported with logic and reasoning, unrestrained by artificial adjustment guidelines, you are providing yet another tool that can improve appraisal quality.

    From another who loves the appraisal profession, KUDOS to you, RACHEL, for providing yet another useful technical tip for the benefit of all!

    Jim Baumberger

    - Reply
  8. Hello Bob, you are right that land value and site amenities are not extracted off this methodology, but if you are using similar sales in many respects, they don’t need to be. It is one tool in a toolbox full of methods, and I suggest trying it in conjunction with others. My suggestion is to try it with depreciated cost, which does segment out the site value as well as site amenities.

    If your data is homogeneous it should work. For instance, in the first grouping, the site values were all within probably a 10% range or so, and the sizes were generally similar. Ages were similar and overall the houses appeal to the same market segment. By getting enough data points into the method, there is a range, and that range is what you can use, or decide to reject. Having finished a couple basements in this area myself I am very familiar with costs, and the difference in value was somewhere close to 50% – 60% of cost new, which makes sense to a large degree.

    Give it a try and see if it works. If it doesn’t, then don’t use it, but do try it a few times.

    Personally I am all for appraisers sharing what they can, in order to help each other with a better more supported work product. Best way to know if it works or not is to try it. Best, Rachel

    - Reply
  9. Bob, you’re correct that different site sizes and other amenities can skew the results of this method. However, this would only be the case if it was true for the sample that the finished or unfinished basement properties tend to have larger sites or more site amenities. If it is true that the samples are similar, which can be controlled in the search parameters, then the adjustment would be based in sound logic. Here is a link to a blog post that demonstrates this using regression (which is just another form of grouped data analysis) http://www.aqualityappraisal.com/Significance%20of%20R%20Squared%20in%20Real%20Estate%20Appraisal%20Linear%20Regression

    - Reply
  10. Hello S skennett,

    Yes it is taking into account all major components but what you want to try and do is ratchet the data down to most similar, and then exclude those that you know are not competitive. If you can have 30 or more data points on each grouping, there should be enough data to get a range. I look for a range, and then choose what I think is most appropriate within it. This doesn’t always work but it is a place to start often, and if you couple it with something like depreciated cost, and the comps that are within the appraisal then hopefully it does give an indication and provides support.

    My thought for the article is to provide some practical applications that could work. They won’t necessarily always work, but often they will. Give it a try and see how it works and whether it gives you a good range that is reasonable. Then try it with cost (depreciated cost using your comparables as well if you can) and also try it using sensitivity, as well as a ranking analysis and common sense. It might be very useful, or maybe not, but try it on a few different units of comparison. As stated in the article, I haven’t had success with it related to GLA adjustments and only mixed results with baths, but I have had some people tell me it worked really well for them with bathrooms.

    - Reply
  11. Hello Kathy,

    Unfortunately I don’t know other spreadsheets. My MLS does have a function called “statistics” which is basically a function that will break out median and average prices plus highs and lows. The way I use that is simply define what I am looking for and what I am excluding, and get rid of sales that I know to be outliers, then run that search. My MLS is Rappatoni, so I know that function is there, and we had something similar with MLXchange when we had that previously. You could see if that is a function within your MLS.

    Hope this helps.

    - Reply
  12. This is all fine, but there is a flaw in your methodology. It does not appear that you extracted the land value or other site amenities from the sales price prior this process. Different site sizes and their amenities (shop, pool, green house, barn, etc.) would likely be a factor in the overall sales price. Without extracting these first, your cost or value of the dwellings specific amenity is skewed.

    - Reply
  13. Hi Rachel,
    I found your article to be very helpful. My question to you is, what if I don’t know how to use excel. What other type of spreadsheet can I use to set up my comps?

    - Reply
  14. While I i understand the formula used for extracting the cost of basement finish vs no finish, what about all the other components in the house which contributed to the sale price? What about buyer personal taste, type of basement finish, etc. I can’t understand how this is being made such a science when it is not one. While we need to make adjustments to major components and to extract in the best way possible those contributions, how is it possible to really know the value of one component over the other when there are so many issues in a house to be considered ? Even with 2 “like” properties, you can have a wide value range which may be due to buyer emotion which cannot be measured. How do you adjust that?

    - Reply
    • And yet here we are still struggling 5 years later to figure a way to support adjustments! Many different ways but most all have flaws that basically render them no better than a guess that we do. If Fannie wants support then why don’t they teach or show ways to get these market-driven adjustments?

      - Reply

Leave a Reply

Your email address will not be published. Required fields are marked *