Positioning for Financials as Treasury Backstops the Long End
Every difficult story in the bond market this year traces back to the same starting point. When the Iran war broke out at the end of February, it did more than unsettle oil markets and shipping lanes. It kicked off a rise in government bond yields that has since spread across the entire developed world, not just the United States. Look at a chart of G7 ten-year yields since 2020 and the pattern is unmistakable: a sharp collapse during the pandemic, a long climb back through 2022 and 2023 as inflation forced central banks to tighten, a brief plateau, and then a fresh leg higher starting almost exactly at the moment the war began.
The United States has felt this more acutely than most of its peers. By mid-August, the thirty-year Treasury yield had climbed to levels not seen since before the 2008 financial crisis. On August 17, it touched 5.31% intraday, a genuine nineteen-year high, before easing slightly in the days that followed. For a government financing itself with trillions of dollars in outstanding debt, that is not an abstract number. It is the difference between manageable interest costs and a budget that starts to strain under its own financing burden.
It was against this backdrop, two days after that nineteen-year high, that the Treasury Department made its move.
On August 19, the Treasury announced it would increase the maximum size of its liquidity support buyback operations for longer-dated nominal coupon securities, covering the ten-to-twenty-year and twenty-to-thirty-year sectors of the curve. The current ceiling of two billion dollars per operation would rise to at least four billion dollars per operation. The change takes effect September 9 and runs through November 4, the end of the current refunding quarter, with Treasury indicating it will provide further guidance on future sizes at its next quarterly refunding announcement.
Coverage of the announcement moved fast, and much of it reached for a familiar comparison. Commentators and strategists drew a direct line to the Federal Reserve's Operation Twist, the 2011 and 2012 program in which the Fed sold short-dated Treasuries to fund purchases of longer-dated ones, flattening the yield curve in an effort to bring down long-term borrowing costs. Some went further and called what Treasury is doing now a form of yield curve control, the practice by which a central bank commits to buying however much of a security is required to hold its yield at a specific level. We think both comparisons deserve more scrutiny than they have received, and the data available on Treasury's own buyback history makes a fairly clean case for why.
Yield curve control has a specific meaning, and it requires three things to be true. The size of purchases has to flex with market pressure, growing larger as yields rise to defend a target. The buyer has to pay whatever price is necessary to hold that level, meaning it typically pays above fair value when the selling pressure is heaviest. And the policy has to actually work, in the sense that the targeted yield stops moving in the unwanted direction. Looking at the full record of Treasury's long-end buybacks since Bessent took office in late January 2025, none of the three holds up.
Start with size. Across all forty-two operations in the ten-year-plus buckets over that period, the maximum purchase amount stayed fixed at two billion dollars, set months in advance on a published schedule. Treasury bought the same two billion dollars when the thirty-year yield sat at 4.4% as it did when the same yield had climbed to 5.3%. If this were yield defense, the size should have scaled with the pressure. It did not move at all. On one occasion in March, dealers offered thirty-six billion dollars of bonds into a single operation, and Treasury accepted just two hundred and five million of it. You do not walk away from thirty-six billion dollars on offer if you are trying to hold a line.
Then look at price. To defend a yield, a buyer generally has to overpay relative to fair value, because that is what it takes to induce sellers to part with bonds they would otherwise hold. Treasury's buybacks did the opposite in the large majority of operations, purchasing at a discount to fair value rather than a premium. That discount showed no tendency to widen when the market had just sold off hard in the days before an operation, which is precisely the pattern you would expect from a program under no pressure to chase a falling price, and precisely the opposite of what a defender of a specific yield level would need to do.
And finally, the yield itself never behaved like something being defended. The thirty-year rose from 4.65% at the very first operation under Bessent to 5.28% at the last one before the size was doubled, a climb of roughly sixty-three basis points across the entire life of the program, hitting a nineteen-year high along the way. A policy of yield curve control that allows its target to rise by more than sixty basis points, unchecked, across a year and a half, is not yield curve control. It is something else, and the honest answer is that it looks much closer to routine debt management than monetary intervention.
If Treasury was not defending a yield, the timing of the August 19 announcement still needs an explanation, and we think the real one is less dramatic than the headlines suggested. Treasury bought its full two billion dollar allowance in forty of the forty-two operations we examined. That is a program running at its ceiling almost every single time it operated, in a part of the curve that had been allocated roughly half the buyback capacity given to shorter maturities. Treasury's own borrowing advisory committee flagged this imbalance as far back as July 2025, more than a year before the size was actually increased. Read that way, the case for doubling the program had been sitting on the table for a long time. The nineteen-year high on August 17 supplied the political urgency to act on it. It was not the reason the case existed in the first place.
Where the Treasury General Account comes in is worth explaining plainly, because it has been the source of some confusion. The TGA is simply the federal government's operating bank account at the Federal Reserve, funded by tax receipts and prior borrowing, and it currently holds close to nine hundred and fifty billion dollars, well above the five hundred fifty to six hundred billion dollar range the prior administration had targeted. Using part of that balance to fund bond purchases means Treasury does not have to immediately issue new short-term bills to raise the cash for each buyback. It can draw down existing cash now and replenish the account with bill issuance later, once market conditions allow. The mechanics matter here: this is not the Federal Reserve creating new money the way it does during quantitative easing. It is the government spending cash it already collected, and the practical effect is closer to reshaping the maturity profile of debt already outstanding than to expanding the money supply.
None of this makes the underlying dynamic irrelevant to a portfolio. It simply means the honest framing is fiscal housekeeping conducted at an unusually large scale, not a rescue operation aimed at a specific number on a screen. Whether that housekeeping still moves markets at the margin is a separate question, and it is the one that actually matters for positioning.
We want to be precise about where things stand as we publish this memo. The larger four billion dollar operations do not begin until September 9. As of today, not a single dollar has been purchased at the new size. Every argument in this section is therefore about what we expect to happen, not a description of something that has already played out, and we think that distinction is the entire opportunity.
Even without a fixed yield target, a doubled purchase program concentrated in the least liquid, most beaten-down corner of the Treasury curve can matter at the margin. It reduces the amount of long-duration supply that private dealers and investors have to absorb into their own balance sheets during a period when that supply has been genuinely hard to place, evidenced by the very discount Treasury has been capturing on its purchases. Removing a marginal buyer of stress from the market, even one operating on a fixed schedule rather than a reactive one, tends to show up first in reduced volatility and improved liquidity conditions before it ever shows up as a specific yield level being defended. That is a real, if modest, tailwind, and it has not been tested yet at the new scale.
There is also the matter of unused capacity. Treasury has been explicit that four billion dollars is a floor, not a ceiling, and has left open the possibility of going larger depending on conditions. Combined with a Treasury General Account sitting near nine hundred fifty billion dollars, the government has considerably more room to act than the market appeared to believe when it initially shrugged off the August 19 announcement and let yields drift back higher within days. We do not think that skepticism is unreasonable given the program's own track record of not defending levels. But skepticism priced in ahead of a program that has not yet run at its new size is exactly the kind of setup that has rewarded patience in every memo we have written this year.
We are not making this call because we believe Treasury is secretly running yield curve control and is about to force long-term rates sharply lower. We have just spent three sections explaining why that framing does not survive contact with the data. We are making it because financials, as a sector, are disproportionately sensitive to exactly the kind of marginal, incremental easing in long-end conditions that an expanded buyback program is realistically capable of producing, even without a fixed target.
Lower long-term yields, even modestly lower ones, ease pressure across a range of things that matter to bank balance sheets and lending economics: the value of long-duration securities holdings, the cost of long-term funding, and the general appetite for credit extension when the future looks less uncertain than it did during a nineteen-year high in borrowing costs. This sits alongside, rather than replaces, the earnings-driven case for financials we made in Memo #002 and the seasonal and political case we made in Memo #003. Three separate mechanisms pointing at the same sector is not proof of anything on its own, but it is the kind of layered evidence we look for before sizing a position with real conviction.
We are keeping this call at the sector level rather than attaching it to individual names, consistent with how we have approached the financial sector across this year's memos. The thesis here is about a structural, marginal improvement in conditions for an entire category of long-duration, balance-sheet-sensitive businesses, not a bet on any single company's execution.
Disclosure: Double Eleven Capital holds no position in the Financial Select Sector SPDR ETF (XLF) as of August 25, 2026. The partnership may hold positions in individual companies within the financial services sector.
September 9 is the date that turns this from a thesis into a test. We would rather be positioned ahead of that date than spend October explaining why we were not.
