New Insights on Borrowers From the National Survey of Mortgage Originations
The 2019 APPAM Fall Research Conference and a recently published issue of Cityscape highlighted the National Survey of Mortgage Originations, a source of data on the mortgage market.
The Association for Public Policy Analysis and Management’s Fall Research conference included a session featuring research recently published in Cityscape based on an exciting new data source for the mortgage market.
The National Mortgage Database (NMDB) is a collaboration between the Federal Housing Finance Agency (FHFA) and Consumer Financial Protection Bureau (CFPB). NMDB is based on a 1-in-20 sample of closed-end first lien residential mortgages active since January 1998 obtained from one of the national credit bureaus. The credit files are then merged with administrative records of federal housing agencies as well as information from public and private sources. FHFA has made aggregated NMDB data publicly available, and CFPB has created an interactive website of mortgage performance trends.
Since 2013, new mortgages have also been randomly sampled every quarter for the National Survey of Mortgage Originations (NSMO). Because NMDB already collects detailed information on the loan, the survey can focus instead on the borrower’s knowledge and experience. A public use file consisting of 24,847 responses from 2013 through 2016 is available from FHFA.
Tim Critchfield, Jaya Dey, Nuno Mota, and Saty Patrabansh use a special release of NSMO that oversampled originations in lower-population counties to examine the perspective of rural borrowers. “Mortgage Experiences of Rural Borrowers in the United States: Insights from the National Survey of Mortgage Originations” finds that borrowers in “completely rural” counties pay higher interest rates and are less satisfied with the borrowing experience than borrowers from metropolitan and other nonmetropolitan areas. Accounting for other aspects of the property, loan, and households, rural borrowers pay an average of 14 basis points more than urban borrowers. Rural borrowers are also less likely to report that they received the best loan terms. Rural borrowers are less satisfied with the loan closing process and report being less confident in their knowledge of key aspects of mortgages. Finally, NSMO responses reveal a different pattern of mortgage shopping among rural households. Rural borrowers are more likely to apply directly to a lender on their own initiative and value an established banking relationship over builder or real estate agent recommendations.
“Perceptions and Expectations of Mortgage Borrowers: New Evidence from the National Survey of Mortgage Originations” reviews the accuracy of borrower perceptions about house price changes. Chad Redmer compares survey responses to questions about how house prices have changed in the past few years (“backcast”) as well as the future direction of house prices (“forecast”) against actual changes in house prices over that time. Borrower opinions yield statistically significant effects even when controlling for common housing market conditions, such as mortgage delinquency rates, unemployment rates, and income growth. Except for first-time homebuyers, however, the demographics associated with more accurate retrospective perceptions are not the same as the demographics associated with more accurate prospective perceptions. Backcasts were strongly correlated with the direction of forecasts.
Brian Bucks, Tim Critchfield, and Susan Singer use NSMO to evaluate the effect of a change in how homebuyers receive information under the Real Estate Settlement Procedures Act. Lenders had been required to provide applicants with a copy of “Shopping for Your Home Loan: Settlement Cost Booklet.” The Dodd-Frank Act transferred responsibility for borrower education to CFPB, which issued a new “Your Home Loan Toolkit: A Step-by-Step Guide” in October 2015. “National Survey of Mortgage Originations Survey Data on Your Home Loan Toolkit” finds that the share of borrowers who report recalling that they received this information increased from 22 percent using the booklet to 42 percent using the new toolkit. In addition, more borrowers report asking questions based on this information, although the rate of questions conditional on recalling receiving the booklet or toolkit is unchanged.Finally, Robert Argento, Lariece Brown, Sergei Koulayev, Grace Li, Marina Myhre, Forrest Pafenberg, and Saty Patrabansh examine the effect of homeownership counseling in “First-Time Homebuyer Counseling and the Mortgage Selection Experience in the United States: Evidence from the National Survey of Mortgage Originations.” Among first-time buyers questioned in NSMO, 17 percent report receiving counseling. Half of those received counseling online, and most reported being very satisfied with their counseling. Counseling is more prevalent among borrowers with lower credit scores, less education, and lower income as well as among black borrowers. The authors use propensity score matching to control for these differences and find that counseled borrowers report being more aware of their credit history and knowledgeable of the differences in mortgage types and interest rates. Perhaps because they are more informed, counseled borrowers are also more likely to report an unpleasant surprise at closing. No difference is found in loan performance yet, although NSMO/NMDB easily allows for longer-term evaluations in the future.