The 20-year auction cleared at 5.469%, the highest yield since 1986

I was watching the 20-year auction on Wednesday. I do that more than I should. It's not healthy. But every few months, an auction tells you something that the stock market won't.

This one told me a lot.

The U.S. Treasury sold $13 billion in 20-year bonds. The yield came in at 5.469%, the highest at a 20-year auction since 1986. It tailed by nearly a full basis point. Dow Jones called it softer than expected. That's polite.

What happened was simple. Not enough people showed up to buy.

So I started pulling the thread. And the thread led me to one of the more important shifts happening in markets right now. It's about who buys Treasuries… and what happens when the biggest buyer starts to leave.

For four years, hedge funds were the marginal buyer of U.S. government debt. Not because they loved America. Because they loved the basis trade.

The trade works like this. A fund buys a Treasury bond in the cash market. It sells a futures contract on that same bond. The spread between the two is tiny… a few basis points. So the fund borrows through the repo market, pledges the bond as collateral, and levers up. Sometimes 50 to 1. Sometimes more.

At that leverage, a few basis points becomes serious money. Multiply it across Millennium, Citadel, ExodusPoint, Capula, and a dozen other multi-strategy funds, and you get a trade that the Federal Reserve estimated at $830 billion by September 2025. That was already double its early-2020 peak.

By early 2026, Morgan Stanley put it at $1.26 trillion.

Now it's around $900 billion. That's $360 billion gone in ten months.

The trade isn't failing. It's just getting crowded out. A Dallas Fed paper from October 16 showed that swap-spread trades now offer a comparable risk-adjusted return. Funds are rotating. Not panicking. Rotating. But the effect on the Treasury market is the same: the largest leveraged buyer of government debt is stepping away.

And here's what worries me. The U.S. government needs to sell $2 trillion in new debt this fiscal year. Debt held by the public is over $32 trillion. The Fed ended its runoff last year. Foreign central banks have stopped growing their holdings. The IIF reported in May that foreign demand for Treasuries has stalled.

So who fills the gap?

I have no idea. And that's the problem.

When leveraged hedge funds step back from buying Treasuries, someone else has to absorb the supply. Right now, that someone is the 25 banks and broker-dealers authorized to trade directly with the Federal Reserve.

New York Fed data shows primary dealer Treasury holdings have risen to their highest level since 2007. They're warehousing bonds that used to move through quickly. And they're doing it while the government keeps adding supply.

Meanwhile, the roster itself is shrinking. On October 16, NatWest withdrew from the primary dealer list. The bank stopped primary dealing in U.S. and European government bond markets. That leaves 25 dealers, down from a peak of 46 in 1988. Fewer seats at the table. More bonds to hold.

The FSB reported in February that hedge fund borrowing through the repo market has grown to nearly $3 trillion… about 25% of total hedge fund assets. All of that borrowing uses Treasuries as collateral. When dealers hold more bonds and hedge funds pledge more bonds, the same collateral gets stretched across more trades.

This is how plumbing breaks. Not from a headline event. From too many claims on the same pipe.

And here's where the strain shows up in price. Wednesday's 20-year auction tailed by 0.9 basis points. The high yield was 5.469%, up from 5.420% at the previous sale. That's the highest yield for this maturity since 1986… the year the Treasury stopped issuing 20-year bonds altogether.

A tail means buyers demanded a higher yield than the market expected. The auction came in below what dealers were willing to offer. It's a soft rejection.

The 20-year has always been the least popular maturity. It sits in the thinnest part of the curve, with limited natural demand. But even bad auctions used to clear close to the when-issued yield. This one didn't.

The Treasury needs to sell $2 trillion this year. Every auction is a test. This one got a C-minus.

The RBA flagged the same risk in its October Financial Stability Review. It called out hedge fund leverage in the U.S. Treasury market as a growing vulnerability. The FSB said the same thing in February. Two central banks on opposite sides of the planet, both watching the same trade. That's worth paying attention to.

Meanwhile, this isn't just an American problem. The bond selloff is global. And the banks that hold all these bonds are starting to feel it.

The Euro Stoxx Banks index fell about 8% in two weeks through October 10, dropping to its lowest level since June. Deutsche Bank and UniCredit each fell more than 4% on October 7. Bloomberg called it more like a reset than a crash. But the direction matters more than the label.

The UK's 30-year gilt yield hit 6.029% on October 1… the highest since 1998. Germany's 10-year Bund sits around 3.36%. Japan's 10-year is around 2.8%. Every major sovereign bond market in the world is repricing at the same time.

When bond yields rise, the bonds that banks and insurers hold lose value. European banks hold large portfolios of sovereign debt. Rising yields mean unrealized losses. And unrealized losses, as we saw with Silicon Valley Bank, stay unrealized right up until the moment they don't.

The IIF said in May that foreign investors are diversifying away from U.S. Treasuries and toward Japanese and European government bonds. So capital is rotating into markets that are also selling off. The safety net is moving. We'll see where it settles.

Three more things worth keeping track of this week.

CCC-rated corporate bond spreads reached 1,282 basis points this month, a three-year high, per the ICE BofA index. But investment-grade spreads are still sitting near 80 to 85 basis points. The lowest-quality borrowers are paying almost 13 percentage points more than Treasuries to borrow, while blue-chip companies can still raise money cheaply. That gap tells you the stress has arrived, but it stays contained to the weakest credits. For now. When the gap widens from the bottom up, the riskier stuff fails first, and then the dominoes start. We're in the first stage.

The Baltic Dry Index fell from 2,917 on October 9 to 2,511 on October 16… a 14% drop in a single week. The Baltic Dry measures the cost of shipping raw materials like iron ore, coal, and grain. It's not a financial product. It can't be faked with leverage or derivatives. When it falls this fast, it usually means factories somewhere are ordering less. I don't know where yet. But a 14% weekly drop is worth watching.

Stablecoins now hold nearly $200 billion in U.S. Treasury bills. The GENIUS Act, signed into law last year, requires stablecoin issuers to hold reserves in short-dated Treasuries, bank deposits, or repo backed by T-bills. Tether alone reported about $115 billion in Treasury exposure in its latest disclosure. Three years ago, nobody would have named a crypto company as a meaningful buyer of government debt. Now stablecoins rank alongside some foreign central banks. I find that interesting. We'll see.

Robert Citron won seven elections.

He was the treasurer of Orange County, California. His job was to manage the county's investment pool… money deposited by 187 public entities. School districts. Water authorities. Cities. By 1994, the pool held $7.6 billion.

Citron's returns were extraordinary. He beat the state pool by about two percentage points a year, year after year. Depositors loved him. He kept winning elections on that track record alone.

But the returns weren't extraordinary because Citron was a genius. They were extraordinary because he was leveraged. He used reverse repurchase agreements to borrow against the bonds in his pool, took the cash, and bought more bonds. Then he borrowed against those. He bought structured notes and inverse floaters that paid more when interest rates fell.

By 1994, the $7.6 billion pool had been leveraged into more than $20 billion in positions. Merrill Lynch helped structure the trades.

The strategy had one requirement. Interest rates had to keep falling. Or at least stay flat.

In February 1994, the Federal Reserve started raising rates. It raised them six times that year. Citron's inverse floaters and structured notes lost value. His repo counterparties demanded more collateral. He didn't have it.

By November, auditors told the county it had lost $1.5 billion. On December 6, 1994, Orange County filed for bankruptcy. It was the largest municipal bankruptcy in American history.

School districts couldn't access their money. Cities froze hiring. Merrill Lynch paid $400 million to settle in 1998 without admitting fault.

The formula was simple. Leverage plus repo plus a one-directional rate bet. The returns looked steady right up until they weren't. Nobody asked how the extra yield was made. They just cashed the checks.

I don't know whether today's leveraged Treasury trades will end the same way. The instruments are different. The mechanics are different. But the shape is familiar. Borrow short, lever up, bet on a spread staying stable, and count on the plumbing holding together. When the plumbing held, Citron was a genius. When it didn't, 187 school districts and cities lost access to their own money.

We'll see.