Working Paper

Algorithms as Shadow Regulation: Secondary Market Access Overrides Home Buyer Credit Risk

Abstract: Using confidential HMDA data on 30 million purchase applications, we show how automated underwriting distorts credit allocation. We document a severe 7.5-percentage point jump in denials at the 50% debt-to-income threshold. This "cliff" is unique to Fannie Mae’s software, whereas Freddie Mac’s algorithm exhibits no matching friction. By exploiting institutional routing to isolate secondary market access as the causal mechanism, we find a stark price-quantity asymmetry: the interest rate penalty is a mere 3 basis points, yet the threshold suppresses $7.7 billion in conventional originations annually, diverting 40,000 households into higher-cost financing.

JEL Classification: G21; G28; D14; L86;

https://doi.org/10.20955/wp.2026.011

Bibliographic Information

Provider: Federal Reserve Bank of St. Louis

Part of Series: Working Papers

Publication Date: 2026-05-29

Number: 2026-011