Abstract
We analyze over 30 million home purchase mortgage applications from 2018–2024 using publicly available Home Mortgage Disclosure Act (HMDA) data to study the determinants of mortgage denial. We establish three primary findings. First, credit access is highly sensitive to monetary policy; the 2022–2023 tightening drove aggregate denial rates from 12.2% to 15.7% via the debt-to-income (DTI) channel. Second, we identify a critical nonlinearity in underwriting: While the 43% qualified mortgage (QM) threshold—below which lenders receive legal safe harbor from ability-to-repay claims—is non-binding in practice, denial rates jump by 15–17 percentage points at the 50% DTI mark, marking the functional market boundary. Third, substantial racial disparities persist; controlling for lender fixed effects and financials, Black applicants are 7.8 percentage points more likely to be denied than White applicants. Observable characteristics explain at most 41% of this gap. These results demonstrate how monetary tightening interacts with structural inequalities to disproportionately restrict credit access for vulnerable populations at the extensive margin.
Introduction
For most households, their largest asset is housing and their largest liability is the mortgage. Access to mortgage credit determines homeownership, location choices, and wealth accumulation. In 2024, lenders denied more than 526,000 home purchase applications, about one in six. Understanding which applicants are denied and why is central to housing policy, financial regulation, and macroeconomic analysis.
We leverage the near-universe of U.S. mortgage applications to provide a comprehensive examination of mortgage denial. Our dataset comprises more than 30 million home purchase applications from 2018–2024 as reported under the Home Mortgage Disclosure Act (HMDA). While these publicly available records are aggregated at an annual frequency—unlike the restricted-access, date-specific confidential HMDA files—they feature significant enhancements mandated by the Dodd-Frank Act. Beginning in 2018, these expanded disclosures include debt-to-income (DTI) ratios, combined loan-to-value (LTV) ratios, specific denial reasons, and interest rates on originated loans. These variables offer a more granular view of underwriting decisions than was possible in prior research using earlier iterations of the public data.
We focus primarily on home purchase mortgages for owner-occupied principal residences. Denial rates vary dramatically across loan purposes: Home improvement loans face rates above 40 percent, while home purchases average between 12 percent and 16 percent. Crucially, home purchases represent the primary margin of entry into homeownership. Unlike refinances or home improvement loans, which are used to restructure debt or enhance the physical characteristics of an existing housing asset, a home purchase mortgage facilitates a transaction for the dwelling itself. Consequently, this segment determines whether a household acquires housing collateral to enter homeownership rather than to simply modify the financing or quality of an existing holding.
