Residential Real Estate is the largest asset class in the world- $190 Trillion. Congealed in this number is not only enormous value but also the collective energies of members of the world’s population for whom their “home” is their single largest investment and their sanctuary. For those who have no dwellings or permanent places to call “home,” there is still the aspiration. As such, residential real estate is not only the largest asset class but among the most important as well.
Ironic then –considering its size and importance—that this asset class is analyzed often in overly simplistic terms. Further, even when simplicity is fine, experts are often in disagreement about the fundamental drivers of even the basic parameters of the industry: value, consumer behavior, and growth (or recession) patterns. This might not be surprising- given the fundamental uncertainty of all socio-economic analyses—but it begs the question- “Can we find a better way to think about valuing residential real estate and predicting trends?”
When this question is posed (as we have hundreds of times), most people correctly surmise that we should “look at the data.” In a world awash with talk of “Big Data” and “Data-Enablement” this thought, while well intentioned, is gratuitous to the point of tautology. It’s as if we are saying that with regard to real estate, we should “look at the data to get the data.” Practitioners of Economics, Data Science, and Artificial Intelligence know that cursory analyses of Data might in some cases offer “Good Enough,” but in others they can obscure as much as they reveal.
And with an asset class of this size and emotional important isn’t “Good Enough” in fact NOT GOOD ENOUGH?
Let’s for a moment isolate the US Market in which there are 105 Million residential parcels. For an individual home buyer or seller, arriving at a “fair price” for a single property is a chore. For a bank or other financial institution that might finance and/or own hundreds of thousands of mortgages, this chore gets magnified to an unimaginable scale. Add to that the need to value the portfolio multiple times a year and to communicate and transact effectively with hundreds of thousands of individuals, and the picture gets even more complex. Then overlay fundamental questions like “which families might default, buy a new house, refinance their existing house, or take some other relevant action” and you begin to understand the challenges and opportunities inherent in this wonderful industry.
Taking this further is not difficult. Once we determine- through proprietary algorithms and techniques- what the “signals” for action are (and thus to be able to effectively act on these signals), the natural next step is to find other variables that themselves might be proxies for or enhancements to these signals. Going back to the scenario we used earlier, imagine the value of looking at “Utilities” data to determine signals for default. If the consumer has not paid his utility bill in 3 months, he is more likely to default than a consumer who is always on time. Is a homeowner about to refinance? Well, does she have kids going to college that year? If so, she might need some extra cash for tuition. The point is clear- which data points are relevant and of the relevant ones, how many do we really need to analyze? Remember, we need to do better than “Good Enough!”

We’ve been doing these analyses for years now and have concluded that there are four fundamental questions that inform our ability to exceed “Good Enough:”

Is the Data set large enough to be as good as complete?

·         Are the Data in this set accurate, clean and representative of reality?

·         Can we use AI to derive real and dynamic intelligence from these data?

·         Can we translate these findings into usable intelligence for businesses and consumers?

We add to this that each of these 4 questions has scores of questions embedded within it.
Information about prices, values, trends, and decisions in the Housing Market are fundamental to the health of the economy and of family life all over the world. Housing prices are on the minds of people in all cultures as they seek better lives.
With AI , we have the opportunity to understand these markets at a deep level in time scales heretofore thought impossible.
Homeowners, home seekers, and the entire ecosystem that supports the whole housing process deserve not only great data but the wisdom that can be derived from it. Good Enough doesn’t cut it.
Erick Watson is Director of Strategic Accounts at Quantarium and Samir Saluja is Co-Founder of DeriveOne.