Credit History and the FHA–Conventional Choice

Abstract

Models explaining whether households choose conventional or FHA mortgage financing typically use differential insurance premiums, loan‐to‐value (LTV) and payment‐to‐income underwriting standards, and local economic conditions to explain household behavior. Using a large and geographically diverse sample, we expand the standard choice model by including measures of borrower credit history. We find that the ability of a homebuyer to avoid credit problems is an important part of the FHA–conventional choice. In addition, credit scores of FHA borrowers are worse on average than those of conventional borrowers, but as LTV increases credit scores of conventional borrowers deteriorate.

Borrower Self-Selection, Underwriting Costs, and Subprime Mortgage Credit Supply

“…This ignores the income and wealth constraints imposed by lending standards. Following Pennington-Cross and Nichols (2000), to determine the unconstrained demand, we estimate a reduced-form, house-price equation over unconstrained homeowners, defined as households who purchase a home with down payments greater than or equal to 30 percent of the value of the home, PTIs of less than 20 percent, and FICO scores above 700. Using the estimated non-constrained coefficients, the desired house price is calculated for all remaining homeowners.…”

Credit Rationing in the U.S. Mortgage Market: Evidence from Variation in FHA Market Shares

“…8 Figure 1 illustrates the distribution of applicant outcomes in this market. The marginal density function of credit risk Φ is drawn, reflecting the empirical fact noted by Pennington-Cross and Nichols [15] that the peak of the marginal density function falls at a level of credit risk less than the maximum credit risk acceptable given conventional underwriting standards, indicated by Φ .…”

“…However, some households, fearing possible rejection for conventional loans, apply for FHA insurance, although their credit risk is below the conventional standard (Φ Φ ), whereas other households, less concerned with possible rejection than with insurance cost, apply for conventional loans with Φ Φ . Pennington-Cross and Nichols [15] demonstrate that although the cumulative distribution functions (CDF) of credit scores reported by the Fair, Isaac and Company (FICO) for FHA and conventional borrowers have considerable overlap, the 5 For borrowers with high loan-to-value mortgages (greater than 95%), FHA insurance premiums are lower than PMI premiums. For borrowers at the 95% LTV cutoff, the relative cost advantage or disadvantage of FHA over PMI depends upon a number of factors, including whether the applicant qualifies for the first-time home buyer credit or the central city discount credit rationing probability of applying to FHA decreases monotonically with increasing FICO scores.…”

“…Thus, if we assume that Φ can be interpreted as the inverse of the credit score, then we can define Φ; Φ as the share of FHA applicants at a given risk level indexed by Φ, given current conventional lending standards (Φ ) Thus, Φ; Φ is an increasing function of Φ. Following Pennington-Cross and Nichols [15], we note that because mortgage insurance charges for lower-risk conventional mortgages are generally lower than FHA insurance rates, Φ; Φ is approximately 0 when Φ ≪ Φ and increases monotonically until it equals unity at some value Φ ̓ , where Φ Φ ̓ Φ and Φ is the upper limit of credit risk acceptable to the FHA.…”

“…However, because the analysis is confined to purchase mortgage loans (not refinance loans), the number of subprime loans in the dataset is extremely small (approximately less than 1%). 15 Because FHA loan limits vary within each year, we use the weighted average loan limit, where the weight is the number of months the limit is in force, as indicated by a HUD Mortgagee Letter or the Federal Register. 16 The FHA defines loan limits by MSA and country, grouped into high and standard cost areas.…”