Over the years, with the approval of their regulator/conservator, the Federal Housing Finance Agency (FHFA), Fannie Mae and Freddie Mac (the GSEs) put in place an opaque system of cross-subsidies that has enabled them, year in and year out, to fund their statutory affordable lending obligations without direct aid from Congress. However, with housing finance reform legislation still stalled, the Trump administration and the new FHFA leadership both committed to reducing the GSEs’ market dominance by cutting back their mortgage purchases, which, depending how they go about this task, could gut the enterprises’ affordable housing activities by sapping their funding capacity.
The GSEs generate cross-subsidies in a two-step process. First, they target a lower, but still positive, economic return on their low-income mortgage purchases by charging a lower guarantee fee than would be warranted purely by risk-based pricing. Step two, they make up for the lower return on these loans by charging selected low credit risk borrowers a higher guarantee fee than is justified by their individual risk profile and loan type or purpose.
In our essay, A Missing Piece of the Administrative Reform Puzzle: How the GSEs Generate Cross-Subsidies, we provide original analysis of how the GSEs’ multibillion-dollar cross-subsidy system plays out across their product landscape. We map out, for selected loan types, the sources and magnitudes of guarantee fee net overcharges (relative to risk-based costs) that generate the GSEs’ internally generated subsidy pool. This information is critical to policymakers and regulators intent on paring back Fannie and Freddie’s market footprints without gutting their affordable housing lending. In our paper, the bottom line message to the FHFA director is to not buy into the siren song from conservatives that “if only the GSEs would step back, private capital will swoop in.” Following this advice would result not only in a reduction in market liquidity, but also in a loss in affordable lending.
Read the accompanying FAQ for more detailed explanation of the modeling assumptions and findings.