Bruce Hahn, MAI, SRA, ASA is sitting in for me with this week’s post, “Data Science Makes Appraisal Fun Again”. Bruce is the owner of Landmark Valuation and a past president of the Northern California Chapter of the Appraisal Institute. Find him on Linkedin or at www.LandmarkValuation.com. Thank you, Bruce!
2018 marks the start of my 30th year as a real estate appraiser. Yes, much has changed during that time – particularly with respect to comparable data. Three decades ago most data sources were still printed on paper (quarterly MLS sales books) or on microfiche. It was time consuming (and not really fun) to research and find a handful of sales. I remember having to use sales that might not necessarily have been that comparable, but they were nonetheless a sale I was happy to locate after considerable research!
Fast forward to today where you can download the entire market of sales from MLS into a single spreadsheet in a matter of minutes. Several years worth for an entire county that can include tens of thousands of sales! Yet not much has changed about appraisal methodology in the last three decades. Today the comparable data is readily searched, parsed, and available for analysis. But we continue to apply the same old “give ‘em three to five comps and trust me” because these are the “best and most comparable!” This begs more questions than it provides answers when anyone can use readily available free data gleaned from the internet to come up with dozens of comparable sales for any residential property. What about the two dozen or more other properties that might be in the Competitive Market Segment© that are not used in the appraisal?
I have to thank George Dell, MAI, SRA, ASA, for his outstanding and thought provoking ideas regarding the use of Data Science techniques with his Evidence Based Valuation© methodology. This has given me the tools to clearly identify the Competitive Market Segment© which positions all relevant sales for the subject property. The analysis tool with R statistical programming language allows the use of all identified sales within the Competitive Market Segment© to predict value conclusions and adjustments. All relevant sales are utilized and incorporated in the valuation. There are no questions about the other sales you did not use!
Evidence Based Valuation© takes a little work to set up and learn, but it is a reproducible product that makes appraisal fun again! The use of Data Science makes the valuation tasks easier, quicker, and more accurate! It makes full use of our valuation skills and knowledge and builds on top of the real estate valuation skills appraisers already possess. It makes you much more professional and provides a much more detailed and supported conclusion of market value. A more reliable valuation product!
Since “drinking from the firehose” of new information in Stats, Graphs and Data Science1 two years ago, I have reflected considerably on the appraisal profession. I have come to two major conclusions. First, the appraisal profession needs to make some major changes to continue to be relevant. Secondly, the growth of data availability and incorporation of Data Science into valuation practice will only increase the need for highly skilled real estate asset analysts that are up for the task! Thank you George! I’m having more fun as a real estate appraiser than I’ve ever had in the last three decades! And I’m delivering a higher quality valuation product to clients who are appreciative of the increased professionalism and detail in my work!
Steven Smith
March 7, 2018 @ 1:27 am
Bruce, good comments. For those willing to take it up a notch, the future is bright. For those who choose to repeat reports of the past, the future might be short lived as technology eliminates them.
I consider George Dell the best friend an appraiser could have, and I have known him since 1984. I remember his original class, and I remember the first offering of Stats and Graphs in Palm Springs in September of 2006. Most of the addendee’s of that class are a member of this forum.
So many of us were given a List of Adjustments and Red Flags, but never taught how to actually embrace and employ the processes and procedures of appraisal. For me, it was a road of hard knocks, which included a recession; that changed my course from the Vocational appraiser I had started out to be, to a path toward professionalism.
That led to degrees in this field, as well as credentials. Of all the courses I have taken, I have used what I have learned from George, more than anything else in my day to day work. I now do fewer reports in a year, than I used to do in a month, but make more money.
Bruce Hahn
March 8, 2018 @ 8:45 am
Steve, as usual you are right on! In fact I think the future is better for any appraiser that embraces Data Science. The need for true expert always seems to increase when technology changes a playing field.
And George is the only person to introduce significant new ideas to the real estate appraisal profession for decades!
Jerry Parsons
June 21, 2018 @ 12:05 pm
This is my impression. Real Estate appraisal is an art based on science. Statistics if completed correctly show ranges of values between sales in a chosen market area. What it does not take into effect, are the sales location differences, solar panels, interior upgrades (new/used homes, same home in a housing project can sell for $50,000 more on initial sale based on the upgrades). Site utility. Your in an market area of hilly topography. The utility of each site is different. The zoning is also into play. Some sites are zoned for multi-residential, some for agriculture, some can be subdivided, some not. Some homes have bathroom updates, some homes have bathroom and kitchen updates, some just have bathroom and windows updated. Some use tile, some use granite, some use quartz. Statistics are generalities on
a chosen market area as it applies to real estate appraisal. The funny thing about stats are that the numbers can be manipulated. A correct appraisal will have a mixture of both. For example, vacant land sales can give you the adjustment for site area (if any apply), view adjustment, location adjustment (busy road), etc.
What is a comparable? that, in my opinion is based on the Buyer’s expectation of a similar type of replacement property. The buyer (market) defines what is a comparable.
Bruce Hahn
June 21, 2018 @ 12:18 pm
Jerry – great comments. I think you illustrate exactly why appraisers bring the necessary skill sets to the Data Science table to make meaningful sense of all of what you describe. Statistical analysis does not render anything about the appraisal process less meaningful. Actually it makes it all the more important. The market knowledge of how to interpret and apply appropriate methodology to all you mention will be more important than ever. But using the entire Competitive Market Segment data available instead of just a handful of “trust me” sales will provide much better results any time it is applied!