Algorithm as Real Estate Agent Algorithmic Buyers Purchased 1 percent of U.S. Homes in 2021

Reading time
2 min read
Different charts related to AI-assisted home purchases in the U.S.

House sales priced by algorithms account for a small but growing portion of the real estate market.

What’s new: Companies that use algorithmic pricing models to buy and sell houses, known as iBuyers, purchased around 1 percent of homes sold in the United States in 2021, roughly double the volume of such transactions in 2019, according to Core Logic, a real-estate data company. However, these deals may not benefit typical home buyers.

How it works: Unlike traditional real estate agents who determine a property’s value by considering the selling prices of similar properties nearby, iBuyers use models that estimate prices based on a variety of factors including national real-estate listings, mortgages, reviews of local businesses, and human assessments.

  • The top four iBuyers accounted for 95 percent of all iBuyer purchases between 2017 and early 2021, the most recent time frame for which data is available. Opendoor bought 56 percent of those homes, Zillow 24 percent, Offerpad 18 percent, and Redfin 2 percent. (Zillow shuttered its iBuying division after incurring a huge loss amid the pandemic, which disrupted housing prices and confounded its algorithm.)
  • In that time, 75 percent of iBuyer purchases took place in five states: Arizona, Texas, Florida, Georgia, and North Carolina. Their models can have difficulty estimating the value of properties that are older or atypical, so they tend to operate in cities with large numbers of newer, homogenous homes, according to CoreLogic.
  • In February, Opendoor told MIT Technology Review that its model could assess homes in locales that are harder to price, such as gated or age-restricted communities, and in cities that offer more varied property types including older buildings, multiplexes, and condos, such as San Francisco.

Yes, but: iBuyers sell 20 percent of their stock to institutional investors like banks and private equity funds rather than individuals or families, according to a January analysis by Bloomberg News. These investors, in turn, often sell the houses to landlords as rental properties.

Why it matters: Automated pricing can make markets more efficient. It can also bring unintended consequences. While iBuyers pitch their services as a way to streamline the Byzantine process of selling and buying houses, they often end up funneling homes into the rental market. That can make it harder than ever for individuals and families to find an affordable home.

We’re thinking: While automated commerce may increase the market’s efficiency in aggregate, we should work to make sure that systems we design don’t inadvertently shut out some buyers.


Subscribe to The Batch

Stay updated with weekly AI News and Insights delivered to your inbox