![Sample our Health and Social Care journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5938/TOC-Subject-Sample-2021-Health-Social-Care.jpg"})
Abstract
Rural homeownership is promoted in the United States by mortgage insurance programs administered by the federal government. We analyze the choice between assistance offered by two such agencies: the Federal Housing Administration and the Rural Housing Service (RHS). We find applicants are sensitive to the relative annual mortgage insurance premiums and guarantee fees. However, there are also persistent racial differences as well as institutional effects. We also find the application and origination process is substantially longer in the RHS program, but variation in closing times does not clearly impact mortgage choice.
Acknowledgments
The authors thank William Reeder for invaluable comments and suggestions. Any omissions and errors belong solely to the authors.
Notes
1. In addition, areas with population not greater than 35,000 and classified as rural in a previous census may be grandfathered until the next census if the area is rural in character and has a serious lack of mortgage credit. Eligibility for RHS’s community development loans and grant programs is based on a different statutory definition of rural.
2. RHS recently revised its method of estimating credit subsidies, using an econometric model of loan performance based on economic conditions and borrower characteristics instead of relying on historical averages (GAO, Citation2016a). Consequently, the latest reestimates (Office of Management & Budget, Citation2017) drastically increased the estimated economic value of recent RHS books of business (i.e., decreased the credit subsidy). These revisions increased the economic value of the 2011 to 2015 Section 502 single-family program books of business by $5.4 billion and converted an estimated positive credit subsidy (2.24%) into a negative one.
3. The financing of upfront premiums is common for low-downpayment borrowers. From 2012 to 2015, more than 80% of all FHA forward loans on existing homes fully financed the upfront premium.
4. TRID (TILA-RESPA Integrated Disclosure) creates greater consistency between the Truth in Lending Act (TILA) of 1968 (P.L. 90–321; 15 USC Sec. 1601) which pertains to the disclosure of mortgage terms and costs, and the Real Estate Settlement Procedures Act (RESPA) of 1974 (P.L. 93–533; 12 USC Sec. 2601) which pertains to closing costs.
5. Depository institutions are covered if their assets exceed a threshold specified by the CFPB ($42–44 million between 2012 and 2015). Coverage of nondepository mortgage companies depends on a combination of assets, volume of loan originations, and office/branch locations.
6. Unfortunately, reporting does not distinguish between the RHS direct loan and guarantee programs.
7. Financing of manufactured homes is also important given their prevalence in rural areas. However, HMDA does not distinguish between real and chattel property with respect to manufactured homes. In addition, each program has different requirements for insuring mortgages secured by manufactured homes.
8. Applications are not reported in HMDA until an action has been made (e.g., loan origination, denied by lender, withdrawn by applicant, etc.); consequently, some applications made late in one calendar year will not be reported until the following year. We exclude the final three months of 2015 to account for this delay in reporting.
9. RHS-eligible geographies have changed somewhat over time because of population recounts and legislative changes. Changes caused by population recounts in the 2010 census became effective October 2012. Changes caused by the Agricultural Act of 2014 were phased in during 2014. As noted, areas previously classified as RHS-eligible may be grandfathered if the area is rural in character and has a serious lack of mortgage credit, reducing the impact of these changes.
10. The Oster (Citation2016) test for omitted variable bias is implemented using the user-written Stata command psacalc, which is only available for linear models.
11. The logistic regression model was unable to converge. This statement reflects the results after 100 iterations.
12. Both FHA and RHS face potential adverse selection problems from the conventional mortgage market; surprisingly, however, Park (Citation2016) finds that FHA does not suffer from substantial adverse selection after accounting for observable credit risk characteristics, despite lack of risk-based pricing.