HOUSINGNEWS REPORT
HOUSING PRECOGS: BIG DATA PREDICTIONS BEYOND HEURISTIC HUNCHES
will build. This will lead to optimal real estate options tailored to the specific future needs of each area. “Also, the government on a county level could benefit from AI predictive tools. One example would be predicting future property tax revenues, based off which properties will sell. This will allow them to have more accurate tax revenue forecasts and budgets.” AI Meets Mortgage Lenders Mortgage lending — like real estate brokerage — is an enormous business. In 2017, according to the Urban Institute, first lien originations totaled
$1.8 trillion. That’s a huge number but it’s 14 percent less than the $2.1 trillion originated in 2016. It also represents a major shift in lender profitability. The Mortgage Bankers Association (MBA) reported that in 2017 “independent mortgage banks and mortgage subsidiaries of chartered banks made an average profit of $711 on each loan they originated in 2017, down from $1,346 per loan in 2016.” The reason for lower profits in 2017 is the combination of reduced volume and fixed costs. Production per company fell from 11,106 loans in 2016 to 8,882 originations in 2017 according
to the MBA. In the same period costs per loan rose from $7,209 to $8,082.
Millions of homes have been financed and refinanced at 4 percent and below during the past few years so why would owners refinance in 2018 when rates are higher? Refinancing represented 49 percent of the market in 2016, a figure which is expected to reach 27 percent this year and even less in 2019 according to the MBA. Loan originations on residential properties decreased 19 percent in Q4 2017 led by a 34 percent drop in refinance originations, according to ATTOM Data Solutions. While declining volume is a concern for lenders, an equally fundamental issue involves the product they sell. While all real estate is different — in theory it’s indestructible, unmovable, and always unique plus it reflects such psychological and social values as status and ego — that’s not the
“I don’t think big data will impact inventory shortages in the short term, but combined with AI prediction forecasting tools, it will be a huge help to developers moving forward when conducting feasibility studies to making final decisions about what product mix they should build, and where they will build.”
BRAD MCDANIEL CO-FOUNDER & CEO, LIKELY.AI
6
MAY 2018 | ATTOM DATA SOLUTIONS
Made with FlippingBook Online newsletter