Fannie and Freddie Just Published Their Own Recap Math: The Q2 2026 ERCF Capital Disclosures

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. Every quarter, both companies are required to publish their capital positions under FHFA's Enterprise Regulatory Capital Framework — the ERCF. The Q2 2026 disclosures are out, and they are the closest thing you will ever get to the government publishing the recap math itself. All four primary sources are linked at the bottom. Every number below is from those documents.

Fannie Mae published its Q2 2026 Capital Disclosures report on August 11, 2026. Freddie Mac's Q2 2026 ERCF Public Disclosure covers the same June 30, 2026 quarter-end. These are dry regulatory documents. Nobody reads them. You should, because they quantify — in the companies' own reconciliation tables — exactly why the senior preferred stock is the entire ballgame.

The headline numbers

Both companies are profitable, growing net worth every quarter, and still show negative regulatory capital. Here's Fannie as of June 30, 2026:

Fannie Mae	Available	Minimum required	Required incl. buffers
CET1 capital	($32.7B)	$66B (4.5%)	$143B (9.8%)
Tier 1 capital	($13.5B)	$88B (6.0%)	$165B (11.3%)
Adjusted total capital	($13.5B)	$117B (8.0%)	$194B (13.3%)
Core capital (statutory)	($4.1B)	$111B (2.5% leverage)	—
GAAP net worth	$116.5B	—	—

Risk-weighted assets: $1.46 trillion. Adjusted total assets: $4.44 trillion. Fannie's own words: "As of June 30, 2026, we had a deficit in available capital for purposes of our risk-based adjusted total capital requirement under the ERCF."

And Freddie:

Freddie Mac	Available	Minimum required	Required incl. buffers
CET1 capital	($13.6B)	$57B (4.5%)	$117B (9.2%)
Tier 1 capital	$0.5B	$76B (6.0%)	$136B (10.7%)
Adjusted total capital	$0.5B	$101B (8.0%)	$161B (12.7%)
Core capital (statutory)	$5.1B	$98B (2.5% leverage)	—
GAAP net worth	$77.8B	—	—

Risk-weighted assets: $1.26 trillion. Adjusted total assets: $3.91 trillion. Freddie's framing: capital levels "were below the required levels as we are still in the process of building our regulatory capital base."

So: combined GAAP net worth of $194.3 billion, and combined CET1 of negative $46.3 billion. How does a company with $116 billion of equity have negative $33 billion of core regulatory capital?

The sign flip is the whole story

Walk through Fannie's own reconciliation table (Exhibit 1.1). Start with $116.5 billion of GAAP stockholders' equity. Then:

• Subtract the senior preferred stock: $120.8 billion. It doesn't count. Fannie's footnote says it plainly: the funds Treasury provided are excluded "as the senior preferred stock does not qualify as core capital due to its cumulative dividend provisions."
• Subtract the junior preferred: $19.1 billion (it comes back one line later — hold that thought).
• Subtract $9.2 billion of deferred tax assets above the 10% CET1 threshold.

Result: CET1 of negative $32.7 billion. Freddie's table works identically: $77.8 billion of equity, minus $72.6 billion of senior preferred, minus $14.1 billion of junior preferred, minus $4.5 billion of DTAs and change, equals CET1 of negative $13.6 billion.

Read that again. The companies are not undercapitalized because they lack equity. They're undercapitalized by classification — because the single largest item on both balance sheets is an instrument the capital rule refuses to count. The ERCF looks at $193.5 billion of Treasury's senior preferred sitting in equity and scores it as zero.

Which means the "capital shortfall" and the "senior preferred problem" are not two problems. They are one problem, printed on one line of one table, in the companies' own quarterly disclosures. I made this argument in June in It's All One Move — these disclosures are that argument with the companies' signatures on it.

$193.5 billion — the number that matters

The combined book value of the senior preferred as of June 30, 2026 is $193.5 billion: $120.8 billion at Fannie (the $119.8 billion actually drawn from Treasury plus the $1 billion initial liquidation preference) and $72.6 billion at Freddie. That's face value — the cash that actually went in.

The liquidation preference is a different, much larger number — roughly $373.5 billion combined as of March 31, 2026, and it ratchets up every quarter, because the January 2021 letter agreement raises Treasury's claim dollar-for-dollar as the companies retain earnings. Not one dollar of that gap above face value is new cash. It's a contract term that Treasury and FHFA wrote, and can rewrite.

My base case has been consistent, and it matches how CBO scores this: in a recap-and-release, the senior preferred gets resolved at face value — converted to common, or deemed repaid — with Treasury exercising its 79.9% warrants for its upside. The liquidation-preference-enforced wipeout is a receivership-flavored scenario that would rattle a $14 trillion MBS market for no policy gain, which is why I consider it unlikely.

Here's what these disclosures let you do for the first time this quarter: run the base case against the companies' own capital tables.

Convert the senior preferred at face value and combined CET1 goes from negative $46.3 billion to positive $147 billion. The combined CET1 minimum requirement is roughly $123 billion. They clear it. The same day. No mega-IPO, no $250 billion capital raise, no decade of waiting. One signature on a PSPA amendment and the two most "undercapitalized" companies in America are above their regulatory minimums with about $25 billion to spare.

What about the buffers? Fully buffered, the combined CET1 requirement is about $260 billion — so post-conversion there's still roughly $113 billion of buffer to build. But buffers are not minimums, and the documents say so themselves: buffer compliance isn't required until conservatorship exit, and falling short of a buffer doesn't make you non-compliant — it restricts dividends, buybacks, and executive bonuses until it's rebuilt. The companies are earning somewhere in the neighborhood of $25–30 billion a year combined. That's roughly a four-year glide path to fully buffered, during which nobody expected them to pay common dividends anyway.

The juniors aren't in the way. They're load-bearing.

Now the part that matters for what I actually own. Look at where the junior preferred sits in both reconciliation tables. It gets subtracted to compute CET1 — and then added right back as additional Tier 1 capital. Fannie: Tier 1 = CET1 of ($32.7B) plus $19.1 billion of qualifying junior preferred. Freddie: Tier 1 = CET1 of ($13.6B) plus $14.1 billion of qualifying preferred.

Under the ERCF, the junior preferred is not a claim competing with the recapitalization. It is part of the recapitalization — $33.2 billion of combined Tier 1 capital, already issued, already outstanding, already counting. Redeeming or impairing the juniors would destroy Tier 1 capital at precisely the moment the entire project is building it. Every dollar spent taking out the juniors is a dollar of capital requirement that has to be raised again from somewhere else.

That's the asymmetry that has me in the preferred rather than the common: in any non-receivership resolution, the juniors' $33 billion is more valuable to the government left intact than taken out. The commons get diluted by the warrants and any conversion. The juniors just... keep sitting there in Tier 1, until someone turns the dividends back on or converts them on negotiated terms.

The bear headline, handled fairly

Somebody is going to add up the biggest columns and write that the GSEs are "$369 billion short" — the combined adjusted-total-capital shortfall against fully buffered requirements ($208 billion at Fannie, $161 billion at Freddie). It's a real number; Freddie prints its version in its own table.

But it conflates three different things:

1. Minimums vs. buffers. The $369 billion measures against 13.3% and 12.7% effective capital ratios — minimums plus stress buffers, stability buffers, and countercyclical buffers. The actual minimum shortfall is much smaller, and buffers phase in as a payout governor, not a solvency test.
2. Classification vs. cash. As shown above, roughly $193.5 billion of the "shortfall" evaporates with a senior preferred resolution that costs Treasury zero new dollars and hands it warrants worth tens of billions.
3. A static snapshot vs. two earnings machines. Both companies retain every dollar they earn. The disclosures' own eligible-retained-income lines show about $26 billion combined over the trailing four quarters.

The honest version of the bear case is that release requires the administration to actually execute the senior preferred resolution — a policy decision, not an arithmetic one. That's true. It's also exactly what the entire 2026 policy apparatus has been teeing up.

What I'm doing

Nothing new. I've been long the junior preferred for years, these disclosures don't change my position or sizing — they just put the companies' own numbers behind the thesis. The next quarterly disclosures land in November. Between now and then, the numbers only move one direction: net worth up, and the "shortfall" mechanically down.

---

The Sources

• Fannie Mae Q2 2026 Capital Disclosures report (PDF) — published August 11, 2026
• Fannie Mae capital disclosures hub
• Freddie Mac Q2 2026 ERCF Public Disclosure (PDF)
• Freddie Mac ERCF disclosure hub

If You Want to Go Deeper

• Full Fanniegate thesis — the whole story
• The 2026 DFAST Stress Test Results — the day-after companion: FHFA ran them through a worse-than-2008 crash and they broke even
• It's All One Move: ERCF Reform, the Senior Preferred, and Freeing Fannie and Freddie Without a Mega-IPO — the June framework these numbers now confirm
• ERCF Capital Requirements vs. Stress Tests — why the requirement itself is ~2.4x what stress tests imply
• The Trump GSE Promotion Timeline — the policy drumbeat, dated and sourced
• 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. All capital figures are from the companies' own Q2 2026 ERCF disclosures, linked in full above, so you can check my math and disagree with me. Do your own research. The full thesis is at glenbradford.com/fanniegate.