Tee'd Up Since 2019: The Blueprint, CBO's Math, and the 2028 Clock on Fannie and Freddie

If you're new here: I'm Glen Bradford. I'm long Fannie Mae and Freddie Mac junior preferred shares — most of my net worth, not a side bet — and I've written the full Fanniegate thesis for years. This is analysis and opinion, not investment advice. I hold what I write about.

If you're back: my Declaration of Independence post made the political "when, not if" case, and It's All One Move put the capital mechanism under the hood. This one ties the whole thing together — back to 2019, forward to a hard 2028 deadline — and shows you the receipts at each layer.

---

Start with two facts that don't care about the news cycle

Fact one: there's a clock. Treasury's warrants to buy 79.9% of Fannie and Freddie's common stock for a nominal price expire on September 7, 2028. That's not my date — it's stated twice in the Congressional Budget Office's December 2024 letter to the House Financial Services Chairman. Trump's term ends January 20, 2029. Those warrants — which CBO values at anywhere from $37 billion to $271 billion depending on the scenario — evaporate roughly four and a half months before his term is up. You do not let a quarter-trillion-dollar asset expire on your watch. The incentive to act inside this term is structural, not rhetorical.

Fact two: the math flipped. In CBO's 2020 analysis, only 12% of its 250 recapitalization scenarios raised enough to fully repay the Treasury. In the December 2024 update, that jumped to nearly 60%. The driver was capital: the GSEs went from $24 billion (end-2019) to $125 billion (end-2023). And that letter is already stale — by Q1 2026 the two companies hold about $186.6 billion in combined net worth (Fannie $112.7B, Freddie $73.9B). So the real number today is better than 60%. The window isn't opening. It's open.

Now the part the market keeps missing: almost none of this is being improvised. It's the execution of a plan written in 2019.

---

It was all written in 2019

In September 2019, Treasury under Secretary Steven Mnuchin published the Housing Reform Plan, pursuant to a March 2019 presidential memorandum from Trump. Read it next to what's happening in 2026 and the resemblance isn't coincidental — it's a blueprint being run. The plan's four goals: end the conservatorships, foster competition, regulate for safety and soundness, and ensure the government is "properly compensated for any explicit or implicit support." Every one of those is load-bearing in what follows.

Two numbers in that 2019 document still matter enormously today:

• Treasury drew $191.5 billion into the two companies. Through June 30, 2019, the GSEs had paid Treasury $301.0 billion in dividends. Read that again: the government put in $191.5 billion and took out $301 billion — by its own accounting, about $110 billion more than it ever advanced. That is the entire moral and arithmetic basis for "deem the senior preferred repaid," and it's sourced to Treasury, not to a shareholder.
• The plan's recap menu (Section III.E.2) lists, verbatim, "Eliminating all or a portion of the liquidation preference of Treasury's senior preferred shares or exchanging all or a portion of that interest for common stock" and "negotiating exchange offers for one or more classes of the GSE's existing junior preferred stock." The original blueprint explicitly contemplates dealing with a portion — and addressing the juniors by design. Hold that thought; it settles a debate below.

---

The Fitzgerald bill is the 2019 plan, codified

On June 25, 2026, Rep. Scott Fitzgerald introduced a three-bill housing package. The centerpiece is the Sustainable Homeownership Act (H.R. 9460) — and when you read the actual text, the administration's fingerprints are all over it, because it's the 2019 Treasury plan turned into statute. The mapping is almost one-to-one:

• SEC. 8 — "Stock of Each Enterprise; Plan to Terminate Conservatorship." Treasury may convert the senior preferred to common (8(a)); shall not resume the periodic commitment fee (8(b)); shall exercise the warrants (8(c)); FHFA determines capital standards within 90 days (8(d)(1)); Treasury and each enterprise restructure within 1 year (8(d)(3)); Treasury sells the warrant stock within 2 years (8(e)). That's the 2019 plan's "preconditions for ending the conservatorships," written as a timetable.
• SEC. 5 — Core Capital. Gives the FHFA Director authority to add components and adjustments to the statutory definition of "core capital." The 2019 plan asked Congress to "repeal the existing statutory definitions relating to the GSEs' regulatory capital that restrict FHFA's discretion." Same idea.
• SEC. 6 — Risk Transfer. Defines CRT as "economically feasible" only when the enterprise "remain[s] profitable," and amends the statutory capital provision so capital "align[s] with the actual credit risk" and stops penalizing CRT. The 2019 plan: capital relief for "a diverse mix of economically sensible CRT."
• SEC. 7 — Capital Framework and Return Regulation. Keeps Treasury's line of credit as an explicit "catastrophic risk backstop," priced with an annual commitment fee benchmarked to credit-default-swap spreads of comparable institutions — and sets a return-on-equity band (initially 9–13%), remitting earnings above the upper bound to Treasury "to compensate the Federal Government for its implicit risk-bearing role." That is the 2019 plan's "explicit, paid-for guarantee" and its "properly compensated" goal, made concrete.

A few honest precision notes, because the details matter and I'd rather you hear them from me: the bill's "redirect to housing" provision moves the §306(l) guarantee-fee stream (currently routed to deficit reduction) toward housing-supply initiatives — it is not a redirect of IPO/sale proceeds. The CRT capital language amends the statutory risk-based capital authority (§1361), not "the ERCF rule" by name. And the 8% retained-portfolio cap is a "greater of" floor — 8% of assets or an amount Treasury and FHFA jointly deem necessary — with construction loans carved out, not a hard ceiling. The 9–13% ROE is the initial band; the Director can move it.

Here's the part that makes the bill bullish whether or not it passes — and I don't think it passes Congress this year. The administration doesn't need it to. If the legislation stalls, the executive can run Section 8's logic administratively — convert or restructure the senior preferred, exercise the warrants, set capital standards — because after Collins v. Yellen, FHFA and Treasury both answer to the President. The bill is the codification and the tell; the break-glass is administrative. It's tee'd up either way.

---

They're already executing — and it's in the Federal Register

This is the proof that it's not just talk. On March 13, 2026, Trump signed two housing executive orders: EO 14393, "Promoting Access to Mortgage Credit" (which directs FHFA to revise capital requirements and report within 120 days — that report is due around July 11), and EO 14394, "Removing Regulatory Barriers to Affordable Home Construction."

Then, on June 24, 2026, FHFA published a Duty to Serve proposed rule that names EO 14394 by number as its basis — verbatim, the rule is "consistent with policies of E.O. 14394" and "E.O. 14394 directs FHFA to take action to reform and eliminate unduly burdensome… standards." In other words: it's not speculation that the agency is acting on the orders. FHFA put its compliance in the Federal Register last week. When people tell you nothing is happening, point them here.

And six days after the EOs, the bank regulators dropped the other shoe: three joint NPRMs, over 1,500 pages, revising bank capital downward — LTV-based tiering replacing the flat 50% residential risk weight, the securitization floor dropping 20%→15%, mortgage-servicing assets moved to a flat 250% weight. That's the move that makes the GSEs' current capital rule, the ERCF, analytically incoherent: you can't hold Fannie and Freddie to capital anchored to a bank framework that's being relaxed.

---

The ERCF is the swing variable — and CBO proves it

Here's why the capital-rule fight is the whole ballgame, in CBO's own table. Take the 2027 recapitalization case and vary only the capital requirement:

• 6% of assets: Treasury recovers $0 on its senior preferred, juniors get impaired, warrants worthless.
• 4.5%: Treasury recovers $171 billion of its ~$190 billion, warrants near zero.
• 3%: Treasury recovers the full ~$190 billion plus $206 billion in warrant value.

The capital requirement alone is the difference between a failed recap and the government collecting ~$400 billion. That makes the capital rule the whole ballgame — and it's worth being precise about where it actually stands, because the bears aren't wrong that the machinery is slow.

Here's the honest status. No proposed rule to revise the ERCF has been filed. EO 14393 directs FHFA to consider revising capital rules and to report to the White House by about July 11 — a direction and a deadline, not a rulemaking. Rewriting the codified ERCF on the books — the durable standard CBO's recap math depends on — requires full notice-and-comment, which historically runs from ~6 months (if FHFA reuses an existing record) to ~2.5 years. Former Freddie Mac CEO Donald Layton's argument for why it could be fast is specific: FHFA "does not need to start from scratch… it can instead start with [the] proposal developed by the FHFA in 2018," which already went through a comment round — a "shortcut" that could "considerably reduce the time". That's his recommendation, not yet FHFA's commitment.

But — and this is the part the bears miss — FHFA doesn't need a finished rule to start cutting guarantee fees. As conservator under HERA, FHFA holds all the powers of the companies' boards; binding capital requirements have been suspended since October 2008; and the GAO has held (B-335424) that FHFA's g-fee changes are conservator directives, not rulemaking. So the near-term lever — lower g-fees ahead of the midterms — can be pulled by directive, immediately, while the durable ERCF rewrite runs on its own clock. Guarantee fees are a direct function of required capital, so it's the same lever from both ends: cut pricing now as conservator, lock it in as a revised rule later.

---

The Warsh pivot: the Fed steps back, the GSEs step in

This is the pillar almost nobody is connecting. Kevin Warsh is the sitting Fed Chair (confirmed in May 2026), and he wants to shrink the Fed's balance sheet — "active and prudent" quantitative tightening, a dedicated balance-sheet task force, and he's called Fed housing policy "somewhat restrictive." The Fed still holds around $6.3 trillion in Treasuries and MBS.

Follow the logic. The Fed has been the marginal buyer of agency MBS. If Warsh lets that book run off, MBS-to-Treasury spreads widen and mortgage rates rise. The logical replacement buyer is Fannie and Freddie, via a larger retained portfolio — and it's already in motion: Trump's June 12 National Homeownership Month proclamation directs the GSEs to buy $200 billion in MBS "to further drive down borrowing costs." They're already near their ~$225 billion caps, so doing this at scale requires raising the portfolio caps — which is exactly why the Fitzgerald 8% cap is written as a floor-with-escape-valve, and exactly the SPSPA-mod lever that lets the companies expand the book, tighten spreads, and juice earnings.

This is the one place 2026 evolved from the 2019 plan. The 2019 plan was a footprint-shrinker (small-government DNA): reduce the retained portfolios. 2026 flips it — use the GSEs to hold MBS and push mortgage rates down. That evolution is bullish for shareholders: a bigger, earning, spread-tightening Fannie and Freddie is worth more than a shrunken one. And it satisfies Treasury Secretary Bessent's one stated guardrail — that release must not push mortgage rates up. The lever that helps shareholders is the same lever that protects his red line.

---

Thirteen months, one direction

The principals whose signatures and actions actually move this — Bessent, Pulte, Lutnick — have a trajectory, and it only points one way. None of them ever retreats toward "never." The hedge is always timing or terms, never whether.

• Bessent (Treasury), the careful one: from "no conservatorship should be indefinite" and "privatization is a goal" (2025) to "we're working on it very deliberately… we want the maximum amount of value," with the offering "hinges on MBS spreads" (Dec 2025). Goal → deliberate → value-and-spreads. His "maximize value" plus "spreads" framing is the ERCF/retained-portfolio thesis in his own mouth.
• Pulte (FHFA), the operator: from "Bessent and I are very aligned" and a $500–700B valuation, to "20 to 30 items presented to the president… executive action, then codify in Congress," to "ready to go," "more likely than not," a 2.5–5% secondary structure, and a floated $1T. Alignment → readiness → mechanics and a rising number. And note his sequencing: executive action first, codify after — which is precisely the Section-8/Fitzgerald relationship.
• Lutnick (Commerce), the salesman: flat conviction — "I think a deal is going to be struck… it could be the largest IPO in history" (Sept 2025), "a sooner rather than later story" (Dec 2025).

Three temperaments, one vector. And a structural shift worth naming: over thirteen months they stopped describing a splashy capital-raise IPO and started describing administrative recap-and-release mechanics — the ERCF-plus-SPSPA path, not a Wall Street roadshow.

---

They staffed for this — on both sides of the table

Here's the argument that doesn't lean on a single quote: look at who got hired into the seats that control a release. A recap is, at bottom, an FHFA-Treasury negotiation — and the roster now reads like it was assembled for one.

• Kate Tyrrell became Treasury Secretary Bessent's Chief of Staff in June 2026 — and she has worked the senior-staff layer on both counterparties to a GSE exit. She was Deputy Chief of Staff at FHFA under Mark Calabria (2019–2021), when the recap-and-release machinery was built, and Deputy Executive Secretary at Treasury in the first Trump term — the department that holds the senior preferred and the 79.9% warrants. In between, she ran agency multifamily lending at Greystone. You do not put that exact résumé in the Secretary's chair by accident.
• Jonathan McKernan, Under Secretary for Domestic Finance, runs the desk that holds Treasury's instruments; Luke Pettit, Assistant Secretary for Financial Institutions, came from Senator Hagerty's office. (More on both in my Declaration post.)
• Mark Calabria — the man who, as Trump's first FHFA Director, started the retained-earnings recap in September 2019 and built the administrative path out of conservatorship — is back in the administration with a housing portfolio at OMB. Be precise: that's budget-side oversight, a step removed from the transaction, not a Treasury seat. But the architect is inside the building.

And the outside chorus moved too. Donald Layton, former CEO of Freddie Mac, spent 2020–2025 as the most credentialed technocratic critic of the capital rule — too high, too distorting, too complex, ~2.4x the stress-test-indicated need. In March 2026 he did something he'd never done: he embraced a Republican President's executive order as the mandate to finally cut it, lowering the difficulty bar from "heavy lift" to "it might even be relatively easy." Be honest about what that is and isn't — Layton pitches the cut as mortgage affordability, not recap-enablement, and he remains a skeptic of a fast IPO. So he's establishment cover for cutting the capital rule — the thing that makes the recap math work — not a cheerleader for the deal itself. That's the honest version, and it's still powerful: the ex-Freddie-CEO expert and the administration now want the same capital cut.

Inside the building and outside it, the people who wrote the plan, ran the 2019 recap, and prosecuted the capital rule are now holding the pens. The roster is the tell.

---

The honest part: the senior preferred, face vs. liquidation preference

Here's where I lead with the loss, because the bull case has a genuinely contested spot — and it's about which number even matters.

Treasury's senior preferred has two figures. The face value (the draws) is about $190 billion combined — roughly $120.8B at Fannie, $73B at Freddie. The liquidation preference is about $373.5 billion — ~$230.5B Fannie, ~$143B Freddie, climbing every quarter because the January 2021 letter agreement raises Treasury's claim dollar-for-dollar with retained earnings. The gap above face value isn't new cash; it's a contractual construct.

So which gets dealt with? CBO answers it. In its December 2024 letter, CBO models the senior preferred at "roughly $190 billion" — the face value — and the juniors at $35 billion face, "retired, to the extent possible, as part of the sale of new common shares." The nonpartisan scorekeeper runs the base case on face, not on the liquidation preference. The 2019 Treasury plan said the same thing in different words: deal with "all or a portion." Converting the full liquidation preference is the "wipeout" scenario, and it really only fits a receivership — which is the one outcome 17 years of conservatorship was built to avoid, because it would roil the $14 trillion MBS market. I don't expect it. Anyone telling you the overhang is $373 billion is using a number CBO doesn't use for a recap-and-release.

And the single most important line in the CBO letter, footnote 5, writes my junior-preferred thesis for me: in any scenario short of receivership, "the Treasury would take a reduction in the value of its senior preferred stake before requiring junior preferred shareholders to do so… changing the GSEs' commitments to junior preferred shareholders would be difficult outside a receivership scenario." Translation from the official scorekeeper: the juniors get redeemed at full $35 billion face, ahead of Treasury's own senior, in everything short of receivership. That's the par anchor, sourced to CBO.

What I don't know — and won't pretend to — is the exact mechanism or the precise effective amount. That's a Treasury decision, and Treasury has given no public signal. That's the real risk: not that the companies aren't valuable, but that the terms haven't been announced.

---

The reality check (because you should distrust a clean story)

The bull case is real. So are these, and I'm not cutting them to make the pitch prettier:

• Pulte just took a second job. Effective June 30, 2026, he becomes acting Director of National Intelligence while keeping FHFA. The bear read — TD Cowen's Jaret Seiberg put it bluntly — is that you can't execute an IPO if the one official responsible is running national security. It's the strongest near-term overhang, and it's fair.
• The President said "not a rush." On June 5, asked directly, Trump said it's still on the table but "It's not a rush." Pulte's January "decision in a month or two" didn't land.
• The Street is skeptical. KBW calls the pre-midterm window "narrowing" and pegs fair value at $200–250B, a fraction of the $1T talk. BTIG downgraded Fannie and Freddie to Neutral on no release visibility. And the $200B MBS directive cuts both ways — some read it as the GSEs becoming a policy funding arm, which argues against privatization.
• Even the megabulls say later. Bill Ackman — as long this trade as anyone, calling it his "best idea for 2026" — argues for Q4 2026, ~$30B raised to hit a 2.5% capital ratio, shares around $34 at an offering, and explicitly against a rushed sale.
• The tape hurts. As of June 26: FNMAS $11.29 (~55% below par), FMCKJ $10.37 (~59% below), and the commons (FNMA $6.88, FMCC $6.17) are down 30–40% on the year.
• The political cover is messier than it looks. The 21st Century ROAD to Housing Act passed Congress (Senate 85–5, House 358–32, June 22–23), but the signing has been contested — so the "affordability win banked in hand" isn't fully banked, and that act doesn't release the GSEs anyway.

Now the rebuttal, and it's why I'm still here. "Not a rush" is not "not happening" — Trump keeps every option open at all times. The distraction argument cuts against a splashy IPO, but the path I've laid out is administrative: an ERCF revision and a senior-preferred restructuring don't require a roadshow or Pulte's undivided calendar. And the deepest answer is the one the bears never address: I don't need a 2026 IPO date. Per CBO's own footnote, the junior preferred get made whole short of receivership — by retained-earnings recap if not by a deal. The clock (September 7, 2028) creates the urgency; the 60% math creates the means; the Federal Register proves the motion. The vision is a capstone deal; the tape says deliberate. Both are true today.

---

Where I sit

Nothing here changes the trade. The common get diluted — Treasury's warrants for 79.9%, the senior preferred, a fresh raise. The junior preferred have the par anchor: a fixed liquidation preference, trading at a deep discount, sitting in the right seat when the priority of claims matters again — and now with CBO on record that they're redeemed at face short of receivership. That's where my net worth is, and that's where it stays.

The case has always been simple, and 2026 made it concrete: the plan was written in 2019, the math went from 12% to 60%, the executive orders are being executed in the Federal Register, the capital rule that decides everything is squarely in the administration's sights with the July report as the next concrete marker, and the warrants that make it all worth doing expire in September 2028. It's tee'd up. The last decision is Treasury's, and I'll tell you the moment it's real.

Focus on the positives. There are a lot of them.

Disclosure: I own Fannie Mae and Freddie Mac junior preferred shares and have been long for years. This is my opinion and analysis, not investment advice. Quotes are attributed to the linked sources; where a timeline has slipped or a fact is contested, I've said so. Carrying values and liquidation preferences are labeled distinctly on purpose. Do your own work.

Keep reading

• The GSE preferred cheat sheet — the thesis, the risks, and the ticker map in about 60 seconds.
• My actual positions — yes, really, 100% of my net worth in GSE junior preferred.
• It's All One Move — the ERCF + senior-preferred mechanism.
• When, Not If: the Declaration of Independence — the political case.
• The full Fanniegate thesis — ten-plus years, the whole case, in one place.