Mergers between Lenders and Real Estate Agencies Could Be Good for the Mortgage Market

Since the Great Recession, nonbank lenders, including credit unions and independent mortgage banks, have increased their market share dramatically in the mortgage lending space.

Today, most homebuyers work with a real estate agent, many of whom make recommendations for lenders. The National Survey of Mortgage Originations shows these recommendations are highly valued by homebuyers, particularly first-time homebuyers and those with lower credit characteristics. By merging with real estate agencies, nonbanks may be able to guarantee this recommendation and widen their customer base significantly. This study examines what happens when real estate agencies and mortgage lenders merge, with attention to both market-level and consumer-level outcomes. This work can help regulators consider what constraints to put on these mergers or what consumer protections to improve or create, if any, to ensure homebuyer welfare does not decrease as lenders and real estate agencies merge.

The author compiled and matched data to create a novel dataset of purchase loans from 2011 to 2019 to examine the effect of mergers on housing market and individual outcomes. They merged proprietary data on property characteristics, lender details, and loan-level information including interest rate, FICO score, and loan-to-value ratio at the time of origination, and performance of the loan with listing data, which included real estate agent name, and with Home Mortgage Disclosure Act data on borrower characteristics, like race and gender. The author then identified relevant transactions in proprietary merger and acquisitions data. Ultimately, they collected, cleaned, and compiled data on more than 100 mergers and related loans for analysis.

Using this new dataset, the author examined lending shares overall, by real estate agency owner, and by core-based statistical area (CBSA) market over time, to figure out national market share and geographic concentration or expansion. The author also analyzed changes on interest rates, time to close, and borrower characteristics from mergers. To estimate causal effects of mergers, the author exploited natural variations in ownership structures. Lenders that are ever merged with agents are compared both with lenders who never merge and buyers that use an agent, and real estate pairs that ever merge are compared with pairs that never merge. The analysis also includes controls for observables and fixed effects, as well as a lack-of-search variable to differentiate between a price effect from a merger versus on account of a borrower choosing not to search or shop around for the best deal.

Overall, the research suggests that mergers lead to lenders growing their loan share within the now jointly owned agency, but the research found little effect on the overall market share of the lender in a given CBSA. In terms of borrower welfare, buyers that use a jointly owned lender paid more in interest rates—a nontrivial average of $225 per year on the average loan in the sample, which has likely only increased in recent years, as interest rates and home prices have shot up dramatically—but the credit characteristics, performance, and cost to close changes are not significant.

Key findings

• Lenders that merge gain market share, but they mainly gain share of the partner agency’s home sales relative to what that share was before the merger, by 15 to 28 percentage points. They are not becoming dominant in the market at national or local levels.
• Lenders that have weaker relationships with real estate agencies were found to be more likely to merge or bring an agency “in house.” Real estate agents having a new “inside track” to lenders could improve loan financing availability for borrowers who would otherwise struggle to have a relationship with a lender, but it could also lead to agents cutting potential borrowers out of the process earlier to protect their relationship with the lender.
• Buyers that used a merged lender-agency pair and went directly to the lender without searching paid 9 basis points more in interest for their loan, even when controlling for FICO score and loan-to-value ratio. In the sample, this represented an extra $225 per year in interest payments, on average, in 2019.
• Though there were statistically significant changes in time to close, the differences in average days were very small: 1 or 2 days in an average of 41 days to close. This suggests that efficiency gains for borrowers from using a merged pair of lender and agent do not come from faster closing times.
• There were no significant changes in borrower characteristics nor loan performance following mergers. This suggests that soft information from agent to lender is not changing issued loans.

The author estimates and examines three counterfactuals:

• If the government banned joint ownership and borrowers had to choose an outside option, the author finds that interest rates would fall, but overall utility—or the consumer’s ability to best satisfy their preferences—decreases slightly.
• If the government banned joint ownership but borrowers were not forced to choose the outside option, the author finds that consumer welfare decreases from consumers choosing products with less desirable characteristics.
• If the government allowed joint ownership but stipulated that lenders must offer all products to consumers regardless of which agents they chose, the author finds that consumers would opt for products with more desirable attributes, increasing consumer welfare.

Policy implications

• The findings suggest that regulating product offerings of merged agents and lenders could increase consumer welfare, but simply banning mergers would harm consumers.
• Examining differences in effects of mergers for subpopulations of borrowers, like first-time borrowers, borrowers of color, and women, is important because these populations face unique barriers in the mortgage market. Additional research to understand how mergers affect subpopulations will be important to determining the best regulatory measures.