Steering Consumers to Affiliated Financial Services: Evidence from Brokered Housing Transactions
65 Pages Posted: 31 May 2019
Date Written: April 4, 2019
We use artificial intelligence to identify financial steering activity in housing transactions. Examining data that allow us to observe private information exchanges between listing and buyer agents, we find that over 16 percent of the homes for sale had negotiation constraints requiring financed buyers to be cross-qualified by an affiliated lender even if they had already been pre-qualified with another lender. Applying a quasi-experiment, we find that affected financed buyers encountered a premium of 3.7 percent (or $5,600) for the average home. Steering also seems to reduce access to credit for African Americans, Hispanics, and women.
Keywords: Steering, Regulations, Mortgages, Housing, Machine Learning
JEL Classification: G21, G24, G28, R22
Suggested Citation: Suggested Citation