This article is focused on an analysis of the recent trends involving increased utilization of what has come to be known as “big data” by the various market participants, how the use of that “data” has evolved, the impact it has had in the short term, and a look at the potential impact long term.
We begin our discussion with the question raised by various industry participants: “What, exactly, do we mean when we speak of big data, and what effect does it have on me?”
Big data is not new. It has been the topic of white papers, research articles and tutorials for well over a decade. Its growing significance can be traced to increasingly complex computer programs producing more meaningful output results for users.
For several years, the American Enterprise Institute and Collateral Risk Network have co-hosted a two-day conference in Washington, D.C., attended by stakeholders from all around the world. During discussions related to property valuations, investors in attendance consistently ask for “more meaningful data” from appraisers than they currently receive via form appraisal reports with a point-in-time value, a half dozen plus-or-minus number of sales, and listings displayed on a grid.
The broader question is not about “big data” but rather the “relevant/meaningful results” the sophisticated software programs are producing to enhance the decision-making of the users.
Therefore, it is really about the analytics – which, defined, is the process of examining large and varied data sets, such as big data, to uncover hidden patterns, unknown correlations, market trends, customer preferences and other useful information that can help organizations make more informed business decisions.
When it comes to analytics and big data within the valuations sector of the mortgage markets, there is no greater influence than that of the government-sponsored enterprises (GSEs) – in particular, Fannie Mae.
Here are some timelines and perspective: In September 2011, Fannie Mae launched the Uniform Mortgage Data Program, the Uniform Appraisal Dataset (UAD) and the transmission of standardized appraisal data through the Uniform Collateral Data Portal (UCDP).
I suspect very few industry participants really understood the significance these deployments would have for the future of the industry. These mandated protocols forced market participants (correspondents and appraisers) into a standardization of data elements that significantly enhanced the capability of the GSEs to collect appraisals and the property information contained in those electronic appraisal reports on a mass scale.
By January 2017, after five-plus years of gathering electronic appraisal information through the UCDP, Fannie Mae extracted market and property information from over 30 million appraisal reports, which include not only subject-property information, but also data and photos of each comparable sale and listing, resulting in unquestionably the largest repository of residential property data and photographs in the country.
Now comes the analytics piece of the GSE equation. While gathering the massive amounts of property information, the GSEs had analysts using that data to build artificial intelligence that, by 2015, was powerful enough for Fannie Mae to introduce Collateral Underwriter and make it available to its correspondent lenders.
For the first time in the history of the valuation industry, the GSEs and their correspondents utilizing the Collateral Underwriter system had access to more data and more powerful analytics than most of the residential appraisers completing the mortgage-lending appraisal assignments.
Freddie Mac followed Fannie Mae into the realm of data standardization and electronic appraisal data submissions, and by 2015, the two GSEs successfully deployed their data-sharing capabilities.
What is the impact?
In 1995, I attended the Mortgage Bankers Association’s (MBA) national conference in San Diego, along with a half-dozen other appraisers who were marketing appraisal services. During the conference, Transamerica introduced an automated valuation model (AVM), demonstrating the capability of entering a property address and having the “black box” produce a property value.
By the second day of the conference, we had real estate agents and mortgage lenders approaching our booth and telling us that within a year, appraisers would be gone.
That was 22 years ago. A recent query of the Appraisal Subcommittee National Registry revealed over 76,000 appraiser credentials.
As it turns out, the Transamerica AVM did not replace appraisers as predicted, nor did other iterations over the 20-plus-year time period following that MBA conference. It is undeniable that AVMs have had an increasing usage by mortgage market participants for various validation and quality control (QC) functions – but nonetheless, it has been somewhat limited.
That is, until the technological advancements that were achieved by the GSEs, using hundreds of millions of bits of data from electronic appraisal reports to develop artificial intelligence that generates virtual instant messages to participating correspondents regarding appraisal-quality flags, data integrity messages, comp selection messages, adjustment messages and reconciliation messages.
In their marketing materials, the GSEs promote the improved efficiencies the industry is experiencing via the front-end validation of appraisal quality for loans submitted through Fannie Mae’s Collateral Underwriter system. Fannie Mae correspondent lenders utilizing the Desktop Underwriter and Collateral Underwriter system are now able to identify and resolve potential appraisal deficiencies and red flags prior to submission, significantly reducing post-submission delays involving appraisals.
Correspondents doing business with Freddie Mac experience similar improved efficiencies when using their Loan Advisor Suite, which also analyzes appraisal reports and provides a view of appraisal quality and valuation risk using the massive amounts of big data provided by those millions of electronic appraisal reports.
Having access to such massive amounts of data enabled the GSEs’ analysts to identify common threads of information from a broad population of appraisers, as well as defective appraisal reports and outliers containing information contradictory to the broader population of data now available and used by analysts.
That artificial intelligence is generating such high confidence levels in the accuracy of the output that, in 2016, Fannie Mae introduced Day 1 Certainty, providing eligibility for relief from reps and warrants on property value for Collateral Underwriter risk scores of 2.5 or lower on a 1-5 scale, with 1 representing the lowest risk and 5 representing the highest risk. Fannie Mae also introduced the Property Inspection Waiver, providing eligibility for relief from appraisal requirements for certain refinance transactions. Freddie Mac followed suit in April 2017, delivering protections on certain purchase and no-cash-out refinance loans.
It is estimated that once the systems have been in place and given time to stabilize, the number of appraisals impacted is estimated to be as much as 25%. The GSEs are processing an estimated 20,000 appraisals per day. Just do the math.
The actual significance of that September 2011 deployment of standardized data, standardized delivery systems and the analytics of that big data is becoming clearer, and the appraisal community is responding. Mortgage market participants are beginning to see greater amounts of data and analytics in appraisal reports being delivered to virtually all mortgage market channels.
This is accommodated in large part by software providers’ continued enhancements to the analytics engines within the software that appraisers use to generate their appraisals. This usage has increased to such an extent that the Appraiser Standards Board of the Appraisal Foundation introduced Advisory Opinion 37, AO-37 Computer-Assisted Valuation Tools, into the 2018-2019 updated version of the Uniform Standards of Professional Appraisal Practice, which becomes effective Jan. 1, 2018. AO-37 provides guidance to appraisers using sophisticated software programs, as well as multiple listing services, to ensure they have competencies necessary to understand the information being used by the models and the results the models generate.
This is good news for the mortgage lending and investment community because the appraiser regulators are recognizing the increased use of big data by appraisers and the increasing sophistication of the software programs being used by appraisers to analyze and reconcile that data, and they are raising awareness of the significance of competent and compliant development of credible appraisals using the advanced technology programs. For the users of appraisal services, this should result in more meaningful and accurate information to use in lending and investment decisions.
Other industry changes
Lenders have been experiencing appraiser shortage challenges, not just in primarily rural markets, but also in major metropolitan markets in Colorado, Oregon, Washington, Idaho and New Jersey, to name a few. In response, on May 31, 2017, the Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corp., the National Credit Union Administration and the Office of the Comptroller of the Currency issued an interagency advisory on the availability of appraisers. In that advisory, they highlighted sections of FIRREA (the Financial Institutions Reform, Recovery and Enforcement Act of 1989 [Title XI]) that allow appraisers to temporarily practice in a state different from their “home” states.
The advisory also referred to temporary waivers relating to the certification or licensing of individuals to perform appraisals under Title XI in states or geographic political subdivisions where there is a shortage of appraisers and significant delays in obtaining appraisals in connection with federally regulated transactions.
Expanded use of the temporary practice permits and use of the temporary waivers should provide some relief for the immediate challenges in certain markets.
Another change in the market is the increased trend toward appraisers completing desktop appraisals utilizing a third party to conduct the field inspection and verification of the property condition and conformity to the neighborhood. This is not something new. Appraisers have been completing desktop appraisals for years. It is the broader acceptance of third-party inspectors replacing appraisers that will be the driving force behind this market transition.
In 1997, the GSEs introduced the Exterior-Only Inspection Appraisal Report, form 2055, in an effort to expedite the appraisal process for rate and term refinance transactions by not requiring interior inspection appraisals. Twenty years later, we are seeing a movement toward a bifurcated appraisal process that will facilitate appraisers moving toward the role of analyst, remaining in their offices, analyzing massive amounts of data to develop a desktop appraisal, and utilizing property information provided by a third-party property inspector to develop appraisals. By eliminating time spent driving to and from appraised properties, such a system will reduce the amount of time it takes an appraiser to produce an appraisal and will, no doubt, have broader usage in the short term, as well as the long term.
As mentioned earlier in this article, big data has been around for a very long time. The relevance of that data as a beneficial tool for lenders and investors has only recently expanded, as the software programs capturing and analyzing the data have become significantly more sophisticated, producing far more meaningful output for lenders and investors to make increasingly more informed decisions. The benefits also include increased efficiencies due to the ability to electronically capture and aggregate borrower information from numerous electronic resources and capture and aggregate property information, demographic information and market-trending data in real time for pre-submission facilitation and QC, as well as post-submission QC.
The role of the appraiser as analyst rather than form-filler will increase. The bifurcated appraisal process will continue to increase.
All of this is the direct result of the increased availability of massive amounts of data and the enhanced technology developed to facilitate the proficiency and accuracy of the artificial intelligence producing increasingly beneficial information for market participants to make more informed decisions in significantly shorter time frames.
Greg Stephens, SRA, MNAA, CDEI, is chief appraiser and senior vice president of compliance with Metro-West Appraisal Co., which has been providing compliant residential real estate valuation solutions for 30 years.