fb-pixelBig data changes the way you buy a home - The Boston Globe Skip to main content
data dive

Big data changes the way you buy a home

Where you stand on big data depends on where you sit, or at least where you live. That’s one of the findings tucked into the first national survey on open government data issues, published by the Pew Research Center.

Posting performance data about individual teachers online? Sixty percent of respondents said yes. Individuals’ criminal records online? Sure, said 62 percent. How about real estate transactions? Some 54 percent were comfortable with putting those online.

But the transparency brakes started squealing loudly when it came to personal mortgage data — only 22 percent favored publicizing it. Higher-income people were more in favor of putting sales transaction data online than lower-income Americans were, but wealthier people were marginally less in favor with regard to individual mortgage information.

The volume and availability of information about every aspect of real estate can, for many, seem downright creepy, as Internet-based companies such as Zillow and Trulia package together sales figures and taxes, current value estimates, future value predictions, neighborhood demographic data, school ratings, and crime rates. Sales prices and mortgages are typically publicly recorded data, available at the town clerk’s and county assessor’s offices, although in about a dozen states, data is restricted in various ways.

Montana, one of those low-data states, notes in its law, “This is because the legislature finds that the demands of individual privacy outweigh the merits of public disclosure.” Also on that list are Alaska, Idaho, North Dakota, Texas, and Wyoming.

Advertisement



Ubiquitous property data — like many data sets — are wonderful tools for society in the aggregate, yet a threat to the individual. “More information leads to more efficiency, but at the same time, you are infringing on what people do not want to reveal,” says Sebastien Gay, a University of Chicago researcher who studies both real estate and privacy issues.

Privacy rules set at the federal level will be key to balancing these competing notions, but the technology is moving far faster than any legislative machine.

Given so-called big data analytics, the future could hold all kinds of additional slicing, dicing, and scoring of real estate, neighborhoods, and the individuals in them.

Advertisement



It’s also created a “naked before your neighbors” phenomenon for new homeowners everywhere and changed social dynamics, blurring public and private all the more. With instantaneous access to sales figures, new homeowners who pay a premium increase the value of comparable houses nearby and might rightfully be greeted by neighbors with more than just cookies as a housewarming gift.

These trends are all of a piece, observers say, with the growing culture of media-fueled voyeurism about property, or what many have called “real estate porn.” (Think television shows like “House Hunters,” “Flip or Flop,” “Rehab Addict,” and “Love It or List It.”)

These fetishes have had all sorts of consequences for the real estate business, including positive ones like producing more informed buyers and sellers. But Lu Han and William C. Strange of the University of Toronto note in their research that Internet usage by home buyers is associated with a greater probability of sales taking place through bidding wars. And the rise of the Internet corresponds with a much higher share of sales taking place through bidding wars generally, even through the Great Recession.

Beyond the worries over diminished privacy, there are also looming concerns about discrimination, because demographics data give off not-so-subtle signals about neighborhoods. Some sites such as Movoto.com, a realty brokerage site, have attracted criticism for displaying racial and ethnic data. Banks have long been accused of redlining — the active discrimination against minority groups based on geography — and that’s still with us, as a recent Zillow analysis suggests. Now, it’s easier than ever for bias to slip into the location choices of home shoppers, too.

Advertisement



Making things potentially worse, Web companies themselves are using algorithms to sort their own customers, steering them towards different products and even pricing differentially. It’s almost inevitable that this will play a bigger role in online real estate commerce in the years to come. Data and security expert Bruce Schneier calls this “Web-lining,” and he says that data brokers are now assigning customer scores, in the way credit scores have traditionally been assigned by private third parties.

These phenomena have the “potential to be much more pervasive and much more discriminatory than traditional redlining,” Schneier writes in his new book, “Data and Goliath.” He says that real estate data collection by companies is also just one part of a much larger story of ubiquitous surveillance.

Moreover, building search engines for registries of deeds across counties makes it easy now to match up names and addresses, leading to worries about everything from stalking — the 1994 Driver’s Privacy Protection Act addressed this same issue in a largely pre-Web era — to commercial profiling, as personally identifying open data makes it easier than ever to estimate an individuals’ net worth.

Interestingly, though, creepiness can cut many ways in the online real estate universe. Both a 2008 study in the American Economic Review and a 2014 study in Real Estate Economics analyzed the effect of posting of online information about the residences of registered sex offenders. Having an offender living within 0.1 miles can lower nearby home values by 4 to 7 percent and lengthen sales time on market.

Advertisement



These titanic changes in the way property is bought and sold come against a counterintuitive narrative: The real estate industry’s core business model has basically endured. And it remains relatively solid even as Americans, who took such a pounding in the financial crisis, are now again bullish on real estate as a long-term investment.

A decade ago, the real estate industry saw the beginnings of what looked like a revolution: the founding of online platforms Trulia and Redfin in 2004, and Zillow in 2005; and a major antitrust lawsuit in 2005 by the Justice Department against the National Association of Realtors, resulting in the industry opening up listings data to Internet-based brokers.

Odds were that the residential real estate business would encounter the same level of disruption as the travel agency, newspaper, and music industries ultimately saw: New entrants. Massive open platforms. Democratization of data. Lower margins. The Internet wins again.

But in real estate, that’s not the whole story. Most transactions still involve a buyer’s and seller’s agent. Commissions have stayed around 6 percent, split between the agents. Although the advertising placements have shifted from traditional media to the Internet, on the whole the business hasn’t radically changed. The what-Uber-did-to-the-taxis story has not unfolded, at least not yet.

Advertisement



Perhaps the amount of money involved in real estate transactions makes people far more risk-averse and willing to pay for a steady, professional hand. And in some ways, Zillow and Trulia merely decided to facilitate the traditional business model and not get into brokerage themselves.

But there’s also a lesson about levels of data quality and control.

Although the newer portals and platforms dominate in terms of Web traffic and have national scale, having a lock on the most updated, accurate, high-quality localized data remains a decisive advantage for brokers. The Multiple Listings Services, or MLS, model — sites operated by associations of local brokers — has been hard to best. Having high quality ensures a role for brokers in the marketplace, even as brokers across the country lament the lack of technical progress on MLS issues, and some wish to develop a national, consolidated platform.

According to a recent report commissioned by the National Association of Realtors, one of the biggest threats to the traditional brokerage model is the potential loss of a control of their data, as well as waning consumer confidence in that data generally, because the massive new online portals often feature outdated or conflicting listings.

Research by B. Douglas Bernheim of Stanford and Jonathan Meer of Texas A&M suggests that, despite claims that they add special value through their knowledge and insight, it appears that brokers don’t always act in the best interest of sellers. They studied properties around the Stanford campus that were not listed on MLS and estimated that a seller’s use of a broker reduced the selling price by 6 to 8 percent. This, they say, speaks to the advantages of allowing open access to MLS data.

Ironically, perhaps brokers provide the last available fig leaf of privacy in an age of data nakedness, carefully managing issues such as disclosures about property condition defects and confidentially handling financial information.

The University of Chicago’s Gay recently coauthored a paper suggesting that broker-massaged online house profiles — sharp pictures and detailed descriptions — may bring higher prices. “There’s potentially a great deal of future work for (sell-side) brokers,” he says. “They have the knowledge to pick and choose the right attributes of a given home.”

The research on selling houses online is like online dating, Gay notes. If you don’t serve up quality data, he says, “people will be less likely to engage with you.”

John Wihbey works at the Harvard Kennedy School’s Shorenstein Center on Media, Politics and Public Policy.

Related:

The race to preserve disappearing data

The case for an absent-minded Internet

2013: Is cyberwar really war?

Address: Your real estate guide to buying, selling, living