Advisory Opinion 37 (AO 37), A Primer

Examining Paperwork

Examining Paperwork

Advisory Opinion 37, Computer Assisted Valuation Tools will be included in the 2018-2019 version of the Uniform Standards of Professional Appraisal Practice (USPAP). So, what does it mean for you? AO-37 clarifies the scope of computer assisted valuation tools related to appraisers and their responsibilities. These tools gather and analyze market data, and when used correctly, can add credibility to completed appraisals. AO-37 also reminds appraisers of their responsibility for the entire valuation process. Appraisers cannot use software or valuation tools to shield themselves from scrutiny.


The history of Advisory Opinion 37 starts way back in 1997 when the Appraisal Standards Board (ASB) adopted Advisory Opinion 18. This AO set forth guidance for the use of Automated Valuation Models (AVMs). AO-18 clarified that appraisers can use AVMs as the basis for appraisals, although the output alone is not sufficient as an appraisal. AO-18 went into effect in the 1998 version of USPAP. As the ASB reviewed AO-18 for the upcoming version of USPAP, they considered replacing it. However, after receiving feedback from key stakeholders, they decided against replacing because it is specific to AVMs and still applies.

They decided to issue a new advisory opinion to address a broader range of technologies. The new AO became Advisory Opinion 37, Computer Assisted Valuation Tools. AO-37 clarifies appraisers’ responsibilities while completing an appraisal using online or stand-alone software tools. AO-37 cites fourteen sections of USPAP, advising appraisers on more sections of USPAP than many other Advisory Opinions. Here are some important takeaways.

Use this tool to see the rules related to AO-37.

Advisory Opinion 37 Interpreted

AO-37 cites issues appraisers face when relying on software or online resources to complete their appraisals. The AO addresses two broadly-used technologies — Regression Analysis and Multiple Listing Services — to illustrate the role of software tools in the valuation process. The main message of this AO is that technology can build credibility in appraisals as long as appraisers apply it correctly to support their decisions. Conversely, they can hurt credibility and valuation results if used incorrectly.

Supporting Adjustments

As mentioned in the AO, tools can support adjustments if appraisers are aware of the information used as the input and how to properly apply the output. Some tools offer data analysis or data visualization to assist appraisers in the valuation process while providing support for their decisions and conclusions.

Scope of Work

One major responsibility for appraisers is determining the scope of work for the appraisal assignment. Because of this, appraisers must understand the assignment to make informed decisions regarding the process required. Appraisers must control the process, selection and input of information based on their knowledge and experience. If appraisers allow their software or tools to control the entire process the output may result in less credibility and/or misleading and unreliable results. Instead, appraisers should control determination of the process, data selection, parameter analysis and interpretation of the output.


According to the USPAP Competency Rule, appraisers must identify the problem being addressed; maintain the knowledge and experience to complete the assignment; and recognize and comply with laws and regulations that apply to the appraiser or the assignment.

muliple regression formula

Multiple Regression Formula

The AO states that the Competency Rule applies to appraisers’ familiarity with specific types of property or assets, market, geographic areas, intended use, specific laws or regulations, or analytical methods, etc. Appraisers must also understand how to apply their knowledge with or without computer assisted tools. Basically, appraisers must understand the basic tenets of appraisal. Appraisers must be able to use tools properly. Luckily, appraisers don’t need to know the formulas used by their tools or any algorithms and/or proprietary information about, or contained in it. AO-37 states “However, the appraiser should be able to describe the overall process and verify that the computer assisted valuation tool is consistent in producing results that accurately reflect prevailing market behavior for the property that is being analyzed.”


Appraisers are responsible for selecting the input parameters AND for affirming the inputs are entered correctly. This demonstrates that appraisers understand the market and has tested the software to make sure it works. Appraisers must also understand how to apply the results of the calculations. For example, appraisers must be able to determine what constitutes a strong relationship between variables from supporting materials such as graphs and tables. Appraisers cannot simply rely on the results of a given software or other tool unless they have reviewed it to ensure its reliability.


Appraisers must understand terminology used by their computer assisted valuation tools. Terms like standard deviation, coefficient of variation, degrees of freedom, etc., have specific meanings and specific applications. The best-case scenario if you use terms like these incorrectly is loss of credibility. The worst-case scenario is incorrect or unsupported value conclusions and reports. One review appraiser recently told me that he often sends reports back to appraisers because they use terms incorrectly or in a misleading way.


When using computer assisted valuation tools make sure the outputs or results make sense with respect to prevailing market behavior.  They’ve got to pass the eye test. Remember, no software can replace your knowledge or experience. The AO states “Regardless of the tool chosen, the appraiser is responsible for the entire analysis including…controlling input, the calculations, and the resulting output.”

Now that you're finished reading, try DataMaster

Get a 14-day Free Trial

3 Responses to “Advisory Opinion 37 (AO 37), A Primer”

  1. Stuart Groten

    The AVM is not an accurate tool for estimating the value of a property. It assumes the individual sitting at a desk can hit a few computer keys and magically provide an estimated property value. Understanding that the information provided in the report may or may not be correct how can a desk bound individual estimate any value sight unseen. It can never take the place of a physical inspection and personal verification of all information to be placed in the report. Granted it may save time but it is no more than the robo signing of the mortgage foreclosure actions during the 2008 depression. Because of the questionable information in an AVM every AVM must be verified by a personal review of all information provided.

  2. Mike Ford, AGA, GAA, RAA

    I agree with Stuart Groten comments. Most (all?) regression software either fails to account for all or even an adequate number f significant market variables. Typically, regression only accounts for 70%+- of the significant property characteristics. It ‘apportions’ the remaining 30% among the 70% thereby creating a KNOWN FALSE indication of from 1% to 30% depending on the specific allocation. Additionally, the sub markets that most appraisers have to work in may well be too small for an adequate amount of market-significant data sets to be available or identifiable. Expanding the competitive market area to obtain enough data sets now changes the results from what the market perceives adjustments to be in the subject neighborhood or competitive area to what is perceived for a different area. AVMs do not and likely never will adequately account for the myriad views; externalities that benefit or penalize a property, topography, site utility, specific interior upgrades, and good old fashioned curb appeal. Give me the opinions of three to six well informed, area agents over regression any day of the week! They may not agree completely with each other but collectively they can and will give me a range of market perceptions and approximations that is generally more reliable than “big data” is or ever will be. THEY are the ones that talk directly with buyers and sellers. Who would know better what their own customers have to say? I can rate their credibility on my own.

  3. Don Cox

    Statistics is a greatly undersold product, few other professionals rely on the appraisers archaic methods Appraisers have been relying on this anecdotal method for so long.
    Longevity is not proof that anecdotal comparisons have worked. Its is evidence that Appraisers have made it work for their special circumstance. Real Estate values are local, trends are general. Comparing a house on a river when is about to fall in against one well founded with a great view, is a much to common. properly used anecdotal comparison used appropriately with statistics are a believable technique for reports or forecasts of local value.
    Honesty in reporting is difficult and expensive to prove, but is a good reason for higher fees