Agency Multifamily Debt: Fannie, Freddie, Ginnie

Introduction

The word "agency" hides three very different machines. When a multifamily borrower takes agency debt, they are tapping one of three government-linked channels, each with its own way of moving credit risk off the originator and into the capital markets. Fannie Mae shares the loss with the lender who made the loan; Freddie Mac sells the risk to bond investors through securitization; Ginnie Mae wraps loans the government has already insured with its full faith and credit. The three together are the deepest, most reliable source of apartment financing in the country, and they anchor a US multifamily debt market that now tops $2 trillion. Fannie Mae's multifamily guaranty book alone reached $499.7 billion in 2024. Understanding how the three differ, and why a borrower would choose one over another, is core to reading agency debt.

The three agencies share a public mission, to keep multifamily credit flowing through the cycle, but they pursue it through fundamentally different credit structures. The distinction that matters is where the credit risk ends up. Fannie keeps lenders on the hook for part of every loss; Freddie passes nearly all the risk to private bond buyers; Ginnie carries government-insured loans where the credit risk was already absorbed by the FHA before securitization. That single design choice cascades into different loan terms, different pricing, and different best-use cases.

This is also why agency lending is the countercyclical floor under apartment financing: when banks and other private lenders retreat, the agencies are chartered to keep lending, which is the demand-side story told in the agency multifamily debt article and the broader US multifamily market overview. Here the focus is the supply-side machinery: how each agency actually structures and prices its debt.

Fannie Mae: the DUS loss-share

Fannie Mae's flagship is the Delegated Underwriting and Servicing program, and its genius is delegation. Rather than reviewing every loan itself, Fannie approves a small group of DUS lenders to originate, underwrite, close, and service loans under Fannie's standards without a pre-purchase review, in exchange for the lender retaining a meaningful share of any loss.

DUS (Delegated Underwriting and Servicing)

Fannie Mae's multifamily program, in place since 1988, in which approved lenders originate and underwrite loans to Fannie's standards and retain a portion of the credit risk (commonly a one-third first-loss share). The loss-sharing alignment is why Fannie can delegate underwriting and still post very low losses.

The loss-share is the alignment mechanism: because a DUS lender eats roughly the first third of any loss on a loan it underwrote, it has every incentive to underwrite well, which is why the program has produced strikingly low losses since 1988. The result is speed and certainty for borrowers, since a delegated lender can commit without waiting on Fannie's review. The program's dominance is near-total: in 2024, 99% of Fannie's multifamily acquisitions came through DUS, and most are pooled into single-loan DUS MBS sold to investors.

Freddie Mac: the K-deal capital-markets model

Freddie Mac reaches the same goal by the opposite route. Through its Optigo seller/servicer network, Freddie aggregates loans and then securitizes them into multiclass structured bonds it calls K-deals, with a separate small-balance loan (SBL) channel for smaller properties.

K-deal

Freddie Mac's multifamily securitization, first issued in 2009, that pools loans into a multiclass structure and sells the subordinate, first-loss bonds to private investors. Freddie guarantees only the senior classes, transferring the bulk of the credit risk to the capital markets rather than retaining it.

The K-deal is, in effect, a CMBS-style structure run by a GSE: senior bonds carry Freddie's guarantee, while subordinate first-loss bonds are sold to private B-piece investors who absorb the credit risk, the same tranche-and-subordination logic detailed in CMBS structure, tranches, and subordination. Where Fannie keeps lenders aligned through loss-sharing, Freddie offloads risk to the bond market through structure. For a borrower the two often look similar at the closing table, but the machinery behind the rate is entirely different.

Ginnie Mae: the FHA full-faith-and-credit wrap

Ginnie Mae is the odd one out, and the most often misunderstood. It does not buy loans, underwrite, or set credit standards. It guarantees mortgage-backed securities built from loans the federal government has already insured through the FHA, and that guarantee carries the full faith and credit of the United States, the strongest backing in the market.

That trade-off defines Ginnie's niche. A developer pursuing ground-up construction or a long-term hold who can tolerate a slow, paperwork-heavy process gets unbeatable terms; a borrower who needs to close in 60 days does not use FHA. The FHA construction programs in particular have no real conventional equivalent, which is why Ginnie execution dominates agency construction lending even though it is a fraction of total agency volume.

The choice among the three is a practical optimization, not a matter of prestige:

Underneath those three executions sits a remarkably consistent product. Agency multifamily loans are non-recourse (with standard bad-boy carve-outs), typically run 5, 7, 10, or 12-year terms against a 30-year amortization schedule, size to roughly 65% to 80% loan-to-value, and require a minimum 1.25x debt-service coverage ratio. Most allow supplemental loans, a second agency loan layered on later as the property's income grows, and most carry yield-maintenance or defeasance prepayment protection rather than open prepayment. Those terms are why agency debt is the default for stabilized apartments: longer, cheaper, and more flexible than a bank will offer on the same building.

The agencies compete loan by loan, and a good agency lender quotes all three before a borrower commits. For a banker, the deeper point is that agency debt is the structural backbone of multifamily, and it sits within the broader lender landscape mapped in the CRE debt universe. No private lender can match the agencies' through-cycle reliability, because none operates under a government mandate to keep lending when markets seize. That is why apartment financing kept flowing in 2023 and 2024 even as banks pulled back hard.

The mandate also constrains them. Because Fannie and Freddie operate under conservatorship, their regulator, the FHFA, sets annual multifamily loan-purchase caps and requires that a large share of each agency's volume support mission-defined affordable and workforce housing. The FHFA set each agency's 2025 cap at $73 billion ($146 billion combined) and raised the 2026 caps to $88 billion apiece, while requiring that at least 50% of each agency's multifamily business be mission-driven affordable housing; workforce housing loans sit outside the caps entirely. Those caps and mission tests shape where agency capital flows: a market-rate luxury deal competes for capped capacity against affordable housing that carries no such limit, which is one reason affordable and workforce properties often find agency execution easiest to secure.

The reach of these three machines extends well beyond the borrower. Agency multifamily MBS, Fannie's DUS bonds, Freddie's K-deal certificates, and Ginnie's project-loan securities, form one of the deepest, most liquid corners of the US bond market, bought by the same insurers, money managers, and banks that hold Treasuries and single-family agency paper. That deep, standing bid is precisely what lets the agencies lend through a downturn: because investors will always buy government-backed multifamily paper, the originator never has to warehouse risk it cannot move, so credit keeps flowing even when private securitization markets freeze.

Three Machines, One Mandate

The unifying point is that "agency" is not one lender but three risk-transfer models. Fannie shares loss with its DUS lenders so it can delegate underwriting and close fast; Freddie sells the credit risk to bond investors through its K-deal securitizations; and Ginnie wraps already-FHA-insured loans in a full-faith-and-credit guarantee to deliver the lowest rates and longest terms, at the cost of speed. The scale runs through one channel above all: roughly 99% of Fannie's multifamily volume moves through DUS, a guaranty book of about $499.7 billion. What ties the three machines together is the mandate to keep lending when private capital retreats, which is why, for all their structural differences, the agencies remain the floor under multifamily financing through the cycle.