The Fanniegate Gate Map: Every Clock, Gate, and Unlock Between Now and the Recap, Tied Out

If you're new here: I'm Glen Bradford. I'm long Fannie Mae and Freddie Mac junior preferred shares and have written the full Fanniegate thesis for years. Yesterday's post did one piece of arithmetic: the President's $200 billion MBS order does not fit under the PSPA's $225 billion portfolio caps. This post is the whole board. It reconciles the retained portfolios month by month so you can check every number against the filings, lays out the back history of how each constraint got here, maps every gate between today and a recapitalization with who holds the key, dates the calendar, and then says what I think it means for the price. It is long on purpose. Skip to the part you need.

Part I. The tie-out: retained portfolios, month by month

Two sources, both from the companies: the monthly summaries (Fannie, June 2026; Freddie, June 2026) and the Q2 10-Qs (Fannie, Freddie). All figures $ millions of unpaid principal unless marked. "Cap basis" is the balance the PSPA cap is tested against: UPB plus 10% of the notional of interest-only securities.

Fannie Mae retained mortgage portfolio

Ties: every month rolls forward (prior balance + purchases − sales − liquidations = end balance) and the year-to-date columns sum. The December and June balances are the 10-Q figures to the dollar ($132,461M and $175,048M). The 10-Q's "for this purpose" cap balance of $177.5 billion is the June cap-basis figure. The 10-Q's own agency-securities line ($70,416M → $103,487M) runs about $1.0 billion below the monthly summary's because the 10-Q table excludes reverse-mortgage securities and wrapped private-label paper; the totals tie exactly, the sub-line differs by definition, and the H1 change is the same either way ($33.1B vs $33.0B).

Freddie Mac mortgage-related investments portfolio

Ties: same roll-forward, same YTD sums. Freddie's 10-Q Table 9 discloses the cap-basis line directly: "Mortgage-related investments portfolio for purposes of Purchase Agreement cap" $159,750M at June 30 vs $161,222M at December 31, with the 10%-of-IO add-on at $21,179M and $21,995M. The 10-Q's mortgage-related securities ($45.5B → $57.1B) equals the monthly summary's agency plus non-agency securities. Inside the loans line, unsecuritized multifamily fell from $37.9B to $21.9B while single-family rose from $55.8B to $59.6B; the multifamily runoff is what funded the MBS growth without moving the cap balance.

What was actually bought, and where the room went

The tie-out says three things plainly. The $200 billion order never fit: $154.6 billion of room on the day it was given. The buying stopped: +$45.1 billion in Q1, −$0.5 billion in Q2, with Fannie a net seller of agency paper in May. The room is being eaten by other things: Fannie's retained loans (delinquent-loan buyouts and lender-liquidity purchases) grew $9.6 billion in the half on their own. Roughly $155 billion of the order remains against $112.7 billion of room that is also shrinking from the inside.

Part II. The back history

How the cap got here

How the PSPA gets amended, and every time it has been

A PSPA change is a letter agreement between Treasury and each company "acting through the Federal Housing Finance Agency as its conservator." No Congress, no comment period, no Federal Register. One letter can amend the purchase agreement, the senior preferred stock certificate (where the liquidation preference lives), and the warrant at once; the January 2025 letter did all three in thirteen kilobytes. Every one has landed as an 8-K, and I checked every 8-K both companies have filed since: none in 2025 after January, none in 2026.

The next one will be the first of Bessent's Treasury. It is contemplated by the last one.

How the lawsuit got here

How the policy drumbeat got here

The promotion timeline has the full record. The short version for this post: May 2025, the President says he is "working on TAKING THESE AMAZING COMPANIES PUBLIC" with the "implicit GUARANTEES" kept. Bessent, December 17, 2025: an offering is "not off the table. We're working on it very deliberately," conditioned on MBS spreads. Pulte, February 6, 2026: the companies are "definitely ready" for a 2.5 to 5 percent sale, "up to the president when or if he decides to do it." Then the calendar went quiet: Wedbush on March 13, "the only discussion we are expecting from the administration regarding these two until after the midterms, if at all, is around the issue of lowering mortgage cost"; Pulte named acting Director of National Intelligence in June while keeping FHFA; the President on June 5, "It's not a rush." July and August: no underwriters, no filing, no date.

Part III. The gate map

Every gate between today and a recapitalized, released, or partially sold company; who holds the key; what is actually required; where it stands. The order is roughly the order they have to open.

Read the key-holder column. Gates 1 through 6 have the same two names on them. Gate 1 opens by inaction on a fixed date. Gate 2 now has a deadline attached to the President's own promise. Gate 3 is the event. Gates 4 through 8 are downstream of 3.

Part IV. The calendar

Part V. The bid

Here is where the juniors have traded through all of it. FNMAS is the most liquid Fannie series and the market's yardstick; FMCKJ the same for Freddie. Both are $25 par. Daily closes:

The junior preferred is down about a third on the year and trades at roughly forty cents on the par dollar. On the day a unanimous appellate panel affirmed that the government breached its duty of good faith to these exact securities, the most liquid series moved up 1.3%. Every leg down in that table is an IPO-timing headline. Not one is a change in the legal or contractual position, which strengthened in every one of those months: the affirmance in July; the Q2 capital disclosures showing the juniors already sit inside Tier 1 and the senior preferred is the sole reason CET1 reads negative; DFAST showing the balance sheet breaks even through a worse-than-2008 crash; and a portfolio cap that now forces the PSPA question. The market sold the timing and threw out the position with it.

How this market reacts: the record

Before guessing how holders react to September 8 and October 22, look at how they have reacted to everything else. FNMAS daily closes; d1 = the event day's move, d5 and d20 = five and twenty trading days later, all measured from the prior close.

Three things jump out. Legal losses are priced instantly and violently: Collins took 62% off the preferred in one session. Legal wins are under-reacted on the day and re-rated over months: the verdict was +12% on the day and the preferred nearly doubled by the time the judgment was entered seven months later; the affirmance was +1.3% on the day and is +7% three weeks on, with the July 21 low holding. Political signals move it hardest: the election was +69% in a day and +113% in a month; one Trump sentence in May 2025 was worth 15%. Timing news bleeds it slowly rather than gapping it.

And that is the whole shape of the past year. From the October 2025 high of $17.57, after Lutnick's "largest IPO in history," Bessent's "very deliberately," and Pulte's "could be Q1 or Q2 2026," the most liquid junior preferred lost 47% to the July low without a single adverse legal or contractual development. It was all timing: the MBS pivot, the "not before the midterms" consensus, the Pulte-to-DNI headline, "It's not a rush." Every one of those is a statement about when, and not one of them changed whether or how much.

That is also the honest base case for what comes next. September 8 and October 22 are legal-finality events without a headline, which this market has historically absorbed slowly and then re-rated. A record-date order is a scheduled mechanical event, which this market has never had. And a letter agreement is a political-structural event, the category that gaps.

Three layers of bid, in increasing importance:

Mechanical, and least important. About $914 million with post-judgment interest, less fees, paid to holders of record pro rata to par (stated value for Fannie series, redemption price for Freddie series). Call it two to two and a half cents on the par dollar, gross. In the context of a recap that turns forty cents into par, it is a rounding error, and I want to be plain about that: nobody should own these for the check. It matters only as a technical: it goes to whoever holds the shares on a record date the court will set after October 22, so there is a reason to be long into that date and no reason to be short into it. It supports; it does not spike.

Narrative. After September 8 and again after October 22, every article, brief, and Treasury talking point about how the juniors should be treated in a restructuring gets a different sentence in it: not "a jury verdict the government is appealing," but an unappealed, unanimous appellate holding that the last time the government reached into these companies' capital structure it violated its duty of good faith to the shareholders. Bloomberg Law wrote up the affirmance the day it came out (Barash); Law360 and Inside Mortgage Finance did too. There are three more natural stories in the chain, and each lands in front of the readers who sold the timing: "government won't appeal," "checks going to current holders," and "Treasury amends the PSPA."

Structural. The litigation excuse ends. Every remaining decision about the senior preferred is discretionary and belongs to two people, and the portfolio cap gives one of those decisions a deadline attached to the President's own promise. When the PSPA opens, the question of what else gets written in it is live, and it gets written by a government that just chose to stop fighting the last war. Nothing in the July 24 opinion prevents an ugly restructuring; it affirmed the denial of rescission and killed the $48 billion restitution theory. It just makes the ugly versions politically costlier, and it puts a dated, unappealed finding on the record for anyone drafting the next document.

September to remember: the sequence in one place

Why September 8 is the tell even though it is not the last legal date. A rehearing petition is not a prerequisite for cert; the government could let September 8 pass and still petition by October 22. But the two filings are not alike. En banc is a brief the private counsel who argued the case (Arnold and Porter) can file on their own. A cert petition for a federal agency runs through the Solicitor General, who does not spend the office's capital on a unanimous, fact-bound state-law contract ruling with no circuit split and $812 million at stake unless someone senior insists. The opinion is a poor vehicle besides: it already cut against shareholders on rescission and the $48 billion restitution theory, so there is no clean question presented. And petitioning means Pulte's FHFA asking the Supreme Court to keep fighting the holders of companies he told the market in February were "definitely ready" for a sale. If they will not make the cheap filing, the expensive one is unlikely. Not impossible. Unlikely.

Why October 22 is the trigger. The Plan of Allocation's definition does the work: "Final Non-appealable Judgment" means the D.C. Circuit is done and "the time for filing petitions for writ of certiorari has expired with no petition having been filed." Ninety days from July 24 is October 22. Nobody rules; the clock runs out. Two things could still move it without a petition: a Rule 13.5 extension application at the Supreme Court (up to 60 days, filed in the second half of October) or a rehearing petition on September 8, which resets the cert clock entirely.

Why the distribution can be fast, and why that is unusual. A normal securities class action pays through a claims process: forms, proof of trades in a class period, a year or two of administration. This one was built differently. The classes are holders as of December 7, 2021 "or their successors in interest to the extent shares were sold after that date and before any final judgment," and the Plan of Allocation pays holders as of a court-approved Record Date through their brokers, pro rata to par, with no claim form. The claim travels with the shares and gets paid like a dividend. Class counsel built that architecture in January 2024, before the appeal, and the defendants appealed the Order Governing Plan of Allocation itself along with the judgment. The D.C. Circuit affirmed both. The post-sweep-purchaser standing holding in the opinion rests on the same principle: the chose in action moved with the stock. So there is no "who gets paid" fight left to have. What remains after October 22 is mechanical, and some of it was staged a year ago: in April 2025, mid-appeal, class counsel moved for approval of a post-judgment notice and a scheduling order for the fee motion (ECF 432 on the D.D.C. docket); the defendants' only substantive objection was that the notice must spell out the deduction for opt-out shares, and they asked for a joint status report within ten days of the D.C. Circuit's ruling. Whether Judge Lamberth has ruled on that motion is not visible in any free source, and the class site's court-documents page shows nothing after the July 24 opinion. So here is the honest range. The steps left are: the mandate (mid-September); the notice-and-objection cycle, which can run in parallel with the cert clock because the Plan gates disbursement, not the process; the fee motion; a hearing and an order approving fees, the allocation, the distribution method, and the Record Date; the judgment actually being paid into a fund (interest runs at 5.01% until it is); and then A.B. Data through the brokers, which the record-date precedents (Lehman, GM, WaMu) put at four to ten weeks. If the papers are staged and the court moves, a Record Date in the first quarter of 2027 and cash a month or two later; if every step goes serially, mid-2027. Either way the money is a 2027 event and the finality is an October one, and only the second of those matters to this thesis. The first visible signals: a status report or notice order on the D.D.C. docket, a new PDF on the class site, and the Q3 10-Qs in late October. One loose end worth knowing: the class notice says recovery goes "only to those who are shareholders at the time of the final judgment," while the Plan pays on the Record Date; nothing public reconciles the two.

Other angles worth holding in mind

"It's not a rush" is a seller's sentence, not a bystander's. The government owns the asset. A seller who says he is in no rush, and who says on the same days that the stake is worth "many times" what he could have sold it for and that people are wrong to value it at $500 to $700 billion, is describing a valuation problem, not a disinterest problem. The things that raise the mark are exactly the recap steps: resolve the senior preferred so the capital reads positive, clean the cap table, then sell. "No rush" on the sale is entirely consistent with doing the paperwork first. It is also consistent with the President putting his own voice on two Fannie Mae television ads this year, which is not something you do for a stake you intend to let sit.

The offering conditions and the MBS program are the same machine. Bessent's stated condition for an offering, in December, was that MBS spreads "remain the same or come down." The MBS buying is what tightened them. The buying stopped in Q2 because the room ran thin and the relative value went away. If Treasury wants tight spreads for a sale, it needs the companies buying again, and that runs straight into the cap. The offering itself generates the amendment.

The consensus is "after the midterms," and the price already reflects it. The Street said it in March, and it is the working assumption across the investor community. That is why the preferred sits at forty cents on the par dollar. Which is the point: the calendar in Part IV has three dates before November 3 that the government cannot move by staying quiet, and anything that happens on them is, by definition, not in the price. Optics cut the other way as well. A letter agreement that treats shareholders well is easier to sign on November 4 than October 4. Several prior letter agreements were signed at holidays or transitions: December 24, 2009; December 21, 2017; January 14, 2021; January 2, 2025. The natural windows after the midterms are the year-end and the January turn, which is also about when the cap wall arrives if the buying resumes.

Housing affordability is a midterm issue, and the only rate lever the administration fully controls runs through the cap. The Fed sets the policy rate; the market sets the 10-year; the President's one direct instrument on mortgage rates is the one he reached for in January: the GSE bid for MBS. The Street's own math says the January program was worth roughly 25 basis points of spread and 15 to 20 of it reached borrowers. If affordability is the argument the administration wants to make in October, the tool is sitting on the shelf with $112.7 billion of room, and using it at any real pace puts the cap in play before the election, not after. That is the pre-midterm case for the amendment, and it does not depend on anybody caring about shareholders. It depends on somebody caring about a rate.

The end of the term is a hard date even if the warrants are soft. Treasury said it expects to extend the September 2028 warrant expiration. The presidency ends January 20, 2029, and that is not extendable. A sale with a real syndicate takes six to nine months from restructuring to pricing. If the plan is a sale on this President's watch at a mark he is willing to defend, the senior preferred is resolved by 2027. Every strand of this post, the litigation clock, the cap wall, and the offering calendar, points at the same year.

And two that cut against me. The liquidation preference grows every quarter, so if the resolution is anything other than face value, waiting makes Treasury's headline "recovery" bigger, not smaller; there is a version of "compensated for its past support" (Bessent, January 2025) that likes delay. And a Congress that changes hands in November holds hearings, not keys, but hearings change the optics of a face-value conversion. Neither changes who signs. Both change when.

Part VI. What would change my mind

• A rehearing petition on September 8. The cert clock resets to 90 days from denial and this whole calendar slides into 2027.
• A cert extension application in late October. Same effect, smaller: up to 60 days.
• A one-paragraph letter agreement that raises the portfolio cap and touches nothing else. That is the bear case on the PSPA leg and it is entirely possible. It would tell me Treasury is willing to open the document for the President's housing number but not for the shareholders yet.
• The MBS program quietly dying at 22%. Then the cap never binds and gate 2 stops being a deadline.
• Any Treasury signal that the senior preferred will not be resolved at face. The opinion doesn't stop that; it only raises the price of it.

What the next six months look like, and what to do while waiting

I get asked this by people who hold these and by people who advise people who hold these, so here it is without hedging.

What you will see, in order. Late August: the July monthly summaries, which say whether the MBS buying resumed. September 8: either a rehearing petition on the D.C. Circuit docket or nothing; nothing is the news. Mid-September: the mandate. September 30: the liquidation preference ticks up again in the Q3 numbers. October 22: either a cert petition or an extension application at the Supreme Court, or nothing; nothing means the judgment is final. Then, over the following months, a notice and eventually a record-date order on the D.D.C. docket, mirrored on the class site, and a small payment through your broker in 2027. Late October: the Q3 10-Qs, with the next cap-basis read and the next word on "further action." November 3: the midterms. Then the windows in which every prior letter agreement has tended to land: the year-end and the January turn.

What to do. Nothing, which is harder than it sounds. The whole point of this post is that the timeline is now driven by dates the government cannot quietly move, and the reaction record says this market prices legal wins slowly and political signals fast. If you own these for the recap, the check in November is irrelevant and the September and October dates are confirmations, not catalysts. The catalyst is a letter agreement on EDGAR, and it has never come with notice. The only actionable thing is to be positioned before the dates rather than after them, and to know in advance which filings would change the picture: a rehearing petition, a cert extension application, a cap raise that touches nothing else, or a Treasury signal that the senior preferred will not be resolved at face. Absent those, the calendar is doing the work.

What I'm doing

Nothing new. Long the junior preferred, same as I've been for years. What changed this month is not the position but the calendar: for the first time in a long time there are dates on it that the government cannot move by staying quiet. September 8 is a decision even if nobody makes it. October 22 is a trigger even if nobody pulls it. And the portfolio cap is a wall that the President's own order is driving toward.

The Sources

• Fannie Mae Form 10-Q, Q2 2026 · Q1 2026 · Form 10-K, 2025 · Freddie Mac Form 10-Q, Q2 2026
• Fannie Mae Monthly Summary, June 2026 · Freddie Mac Monthly Volume Summary, June 2026
• January 14, 2021 letter agreement · January 2, 2025 letter agreement (8-K exhibit) · Treasury release, Jan 3, 2025 · FHFA/Treasury Sept 14, 2021 suspension · Third Amendment, Aug 17, 2012
• D.C. Circuit opinion, Fairholme Funds v. FHFA, No. 25-5113 (July 24, 2026) · Final Judgment, Mar 20, 2024 · Order Governing Plan of Allocation · Official class site (A.B. Data) · FRAP 40 · FRAP 41
• Scotsman Guide, Jan 8 · Bessent on matching Fed runoff, Jan 9 · Fortune/AP, Jan 24 · Fox Business, Feb 25 · NMN, Feb 6 (Pulte "definitely ready") · NMN, Mar 13 (Wedbush) · NMN, Mar 28 (Barclays) · NMN, May 18 (JPM/BofA) · Rithm, Jun 4 · Bloomberg Law, Jul 24
• H.R. 9460 text (govinfo) · CBO, "Seven Things to Know" (July 2025)
• Price data: daily closes, Yahoo Finance, FNMAS / FMCKJ / FNMA.

If You Want to Go Deeper

• Yesterday: the $200B order vs the $225B cap — the arithmetic, in full
• Full Fanniegate thesis — the whole story
• The Q2 2026 ERCF Capital Disclosures — the companies' own recap math
• The 2026 DFAST stress test — break-even through a worse-than-2008 crash
• The Trump GSE Promotion Timeline — the drumbeat, dated and sourced
• D.C. Circuit oral argument recap, April 21 — the hearing that produced the affirmance
• Fanniegate Timeline — interactive milestone history

I hold long positions in Fannie Mae and Freddie Mac junior preferred shares. This post is my personal opinion and is not financial advice. Portfolio, cap, judgment, and calendar figures are from the filings, court orders, and rules linked above; the reconciliations, the gate map, and the reading of what they imply are mine. Price history is from public daily closes. Whether the government files anything by September 8 or October 22, whether Treasury amends the PSPA, and what it writes when it does are open questions; this post argues they are now questions with dates on them. Do your own research. The full thesis is at glenbradford.com/fanniegate.