The Downside Scenario Was the Conservator: FHFA Just Dropped the One Stress Case That Could Still Send Fannie and Freddie Back to Treasury

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. Last week I covered the 2026 Dodd-Frank stress test results — a worse-than-2008 housing crash, and the companies break even. I mentioned in one sentence that the report dropped its usual alternate case. Hat tip to Rule of Law Guy, who wrote it up this week and made me realize that sentence deserved its own post. He frames it as FHFA finally shedding an Obama-era habit. I think it's bigger than that, and the reason why runs straight through the 2008 draws, the 2012 sweep, and the jury verdict the DC Circuit just affirmed.

What got dropped

Every year since 2014, FHFA's DFAST report for the Enterprises has shown two sets of results. The first is the plain severely adverse scenario. The second is the same scenario plus one more assumption: that mid-crisis, Fannie and Freddie establish a full valuation allowance against their deferred tax assets.

Here's the 2025 report, Table 1, combined:

2025 DFAST, combined	Without DTA allowance	With DTA allowance
Total comprehensive income (loss)	+$8.6B	($6.7B)
Impact of the allowance	—	($15.3B)
Post-stress net worth	$162.8B	$147.5B

And here's the 2026 report, Table 1: one column. Comprehensive loss of $0.1 billion, post-stress net worth of $179.4 billion. The results bullet reads, in full: "Fannie Mae projected comprehensive losses while Freddie Mac projected comprehensive income, without establishing a valuation allowance on Deferred Tax Assets (DTA) in the severely adverse scenario." No second table. No "with" column. The alternate case is gone.

What a DTA valuation allowance actually is

A deferred tax asset is a future tax deduction you've already earned — from past losses, from timing differences, from reserves you've booked but haven't deducted yet. It's worth something only if you'll have taxable income to deduct it against. Under the accounting rules, if it's "more likely than not" that you won't, you take a valuation allowance: you write the asset down, and the write-down runs through the income statement as a loss.

So a full DTA valuation allowance is not a loss on anything. No mortgage defaulted. No security got marked down. It is a judgment that the company will not earn enough money in the future to use its tax deductions. For a company with a guarantee-fee business that throws off $71 billion of pre-provision earnings in a prescribed catastrophe, that judgment has exactly one plausible meaning: someone has decided the company doesn't have a future. It is receivership, expressed in tax accounting.

That's the scenario FHFA carried as its official downside case for twelve years. Not "what if the housing market is worse than we think." The housing market is already in the base case — a 30% home price decline, 10% unemployment, a 58% equity crash. The alternate case was "what if, on top of all that, the conservator decides the companies are finished."

Why this is the assumption that built the senior preferred

This is where I part ways with the "housekeeping" read. The DTA valuation allowance is not some abstract stress-test convention. It is the single accounting entry that did the most to create Treasury's claim on these companies in the first place.

In the third quarter of 2008, one quarter after conservatorship, Fannie Mae booked a $21.4 billion charge to establish a valuation allowance against its deferred tax assets — the single largest item in a $29 billion quarterly loss, dwarfing the $9.2 billion of actual credit expenses. Freddie Mac recorded a $22.4 billion allowance over the course of 2008. Combined: roughly $44 billion of paper losses in a single year, produced by a judgment that the companies would never be profitable enough to use their tax deductions.

Under the Senior Preferred Stock Purchase Agreements, every dollar of negative net worth triggers a Treasury draw, and every draw adds a dollar to Treasury's liquidation preference. The allowances went straight into the draws. The draws went straight into the senior preferred balance that eventually reached $191 billion. And those allowances kept growing through 2011 as the companies reserved for losses they expected but hadn't taken. By the end of 2012, Fannie alone was carrying a valuation allowance of roughly $59 billion.

Then the judgment flipped. In the first quarter of 2013, Fannie Mae released its valuation allowance and booked a $50.6 billion tax benefit — a $58.7 billion quarterly profit, the largest in the history of American business. Freddie Mac followed in the third quarter of 2013 with a $23.9 billion release. Combined: $74.5 billion of profit from reversing the same entry that had helped justify the draws five years earlier.

Same lever. Pulled down in 2008, it manufactured losses that sized the government's claim. Pulled up in 2013, it manufactured profits that — by then — flowed entirely to Treasury under the Net Worth Sweep.

The sweep was timed to the reversal, and a jury said so

None of this is my inference. It's the trial record.

On August 9, 2012, Fannie Mae's CFO, Susan McFarland, told senior Treasury officials the company was in "sustainable profitability" and that it expected to reverse the valuation allowance, generating roughly $50 billion of income. Treasury's own internal memos from late 2011 and mid-2012 anticipated the "golden years of earnings." The Third Amendment — the sweep — was signed eight days later, on August 17, 2012, and publicly justified as protecting taxpayers from a "death spiral" of circular dividend draws at companies that could never repay.

On August 14, 2023, a federal jury in the District of Columbia unanimously found that by agreeing to the sweep, FHFA breached the implied covenant of good faith and fair dealing it owed shareholders, and awarded $612.4 million. Judge Lamberth entered final judgment at $812 million with interest. On July 24, 2026, the DC Circuit affirmed with no dissent. The rehearing window closes September 8.

So the sequence is: the conservator's discretion over one accounting judgment helped create the senior preferred; the reversal of that same judgment was known to Treasury before the sweep; the sweep captured the reversal; and the courts have now found, with finality approaching, that the sweep was made in bad faith.

And for every one of the twelve years after that reversal — after the allowance had been released, after the companies had proven they could earn $20-plus billion a year, after the DTAs had been used exactly as a going concern uses them — FHFA kept presenting "what if we write them off again" as the official downside scenario. Rule of Law Guy calls that an Obama-era tell, and he's not wrong about where the habit started. But it survived Watt, Calabria, Thompson, and the first year of Pulte. It only died in 2026.

Why it matters now: the draw trigger

Here's the part I think is material, and it has nothing to do with accounting history.

A Treasury draw happens when net worth goes negative. That's the PSPA mechanism. Draws are how the senior preferred got built, and a new draw is the only thing that could grow it further or revive the "they're a ward of the state" story.

Read the 2026 base case again: worse-than-2008 housing crash, combined loss of $0.1 billion, post-stress net worth of $179.4 billion. The actual book can't get to a draw. Not close. It would take a loss 1,800 times larger than the one FHFA's scenario produces.

The only thing in FHFA's entire modeling history that could manufacture negative net worth at these companies was the DTA write-off — because it's not a loss on anything, it's a decision. That's exactly why it was the alternate case. It was the one lever that could still produce a draw when the housing market couldn't. And even it was running out of room: the 2025 version only got to a $6.7 billion combined loss and $147.5 billion of net worth. The alternate case had become a scenario in which the conservator does 2008 again and it still doesn't matter.

FHFA is no longer presenting any scenario — market or discretionary — in which Fannie Mae or Freddie Mac draw on Treasury. The official downside now ends with $179 billion of net worth and zero new senior preferred.

That's material for three reasons:

• It retires the last official artifact of the receivership path. You cannot put a company with $179 billion of post-stress net worth into receivership without first doing the thing FHFA just stopped modeling. The regulator's own document no longer contemplates it.
• It weakens Treasury's negotiating leverage. The value of the funding commitment — the thing Treasury charges for, the thing it cites as the taxpayers' ongoing exposure — rests on the idea that a draw could happen. FHFA's own stress test now says that in a scenario worse than the one that caused conservatorship, it can't.
• It's the same lever, viewed from the other side. The entry that created roughly 40% of the senior preferred's original $191 billion is the entry FHFA has now retired as a downside case. Nobody at the agency is prepared to put their name on the assumption that these companies have no future — not even hypothetically, not even in a document that says on its cover that the projections "are not expected outcomes."

The one-sentence version

For eighteen years the worst case for Fannie Mae and Freddie Mac was never the housing market. It was the conservator's discretion over their accounting. The 2008 write-offs, the 2012 sweep, the twelve alternate-case tables — all the same lever, in the same hands. In 2026, the regulator stopped modeling its own discretion as the downside. The only way these companies ever go back to Treasury is if FHFA decides to send them, and FHFA just stopped pretending it would.

What I'm doing

Same as always — long the junior preferred. This doesn't change the position; it removes the last hypothetical in which the position gets hurt by anything other than politics. The seniors sit at $193.5 billion of face and a liquidation preference pushing $390 billion, built substantially on a judgment the regulator no longer makes. The companies can't draw. The courts have ruled on the sweep. The next dates are September 8 for the rehearing deadline and October 22 for finality, and I've written up how I'm thinking about the window between them.

---

The Sources

• Rule of Law Guy — GSEs Pass Dodd Frank Severely Adverse Stress Test with Flying Colors, Again — August 19, 2026; the post that flagged the dropped DTA case
• FHFA 2026 Dodd-Frank Act Stress Test Results — August 14, 2026 (PDF mirror); results bullet and Table 1
• FHFA 2025 Dodd-Frank Act Stress Test Results — (PDF mirror); Table 1 "with" and "without" valuation allowance columns
• Fannie Mae Q3 2008 results — $21.4 billion DTA valuation allowance charge
• Freddie Mac 2008 Form 10-K — $22.4 billion valuation allowance for the year
• Fannie Mae Q1 2013 results — $50.6 billion release, $58.7 billion quarterly net income
• Freddie Mac Q3 2013 results — $23.9 billion release, $30.5 billion quarterly net income
• Boies Schiller Flexner — trial verdict announcement — August 14, 2023 jury verdict, $612.4 million
• Forbes / Richard Epstein — deposition testimony on the 2012 DTA discussions — the August 9, 2012 McFarland meeting

If You Want to Go Deeper

• Full Fanniegate thesis — the whole story
• The 2026 DFAST results — last week's post on the base case
• The September 8 bid thesis and gate map — what happens between the rehearing deadline and finality
• The Q2 2026 ERCF capital disclosures — why negative CET1 is a classification artifact of the seniors
• Fanniegate timeline — 2008 to today