Bessent Bond Buybacks Fail to Cap Yields as Chips, Bitcoin Slide

The treasury secretary's bet on the bond market has not worked

Will Treasury Secretary Scott Bessent's campaign to bring down long-term borrowing costs lift stocks? So far the honest answer is no, and the problem goes beyond the buybacks being too small. Two months after he promised to sharply expand Treasury's debt repurchases, long-dated yields are higher than when he started. Rising yields, not falling ones, have become the main pressure on stocks.

The effort started in August, when Bessent surprised markets by pledging to "at least double" Treasury's usual repurchases of government debt. Officials paired that with steps to slow the yen's fall against the dollar1. Under the new setup, per-auction buybacks of 10-, 20- and 30-year securities rose from the $2 billion level set under Janet Yellen to at least $4 billion4. A later announcement put the cap at $6 billion, triple Yellen's limit, and the bond market moved the wrong way right away. The 30-year yield jumped about 5 basis points to roughly 5.30%, and the 10-year rose to 4.85%4.

Treasury has since run three operations: $5.2 billion on September 10, $4.1 billion on September 24 and $6 billion on October 13. Yields rose on the days of the first two. They dipped on the day of the third, then rose again the next session3.

Where yields stand now

Reports give slightly different peaks, but they all point the same way. A Senate Banking Committee letter to Bessent dated October 7 says the 10-year yield hit 5.34% on September 30. That was its highest since 2002 and more than a full percentage point above its level on February 27, just before the U.S. war with Iran began3. Other coverage has the benchmark at 5.31% on October 59 and 5.35% on October 72. A weekly economic review says the 10-year briefly topped 5.36%. In that same week, a $39 billion auction cleared at 5.30%, the highest auction yield since 2000. Decent demand at the sale pulled the 10-year back to about 5.24% by week's end, while the 30-year finished near 5.61%6.

Those small differences don't change the picture. The 10-year has moved from about 4.8% when Bessent spoke at Southern Methodist University on September 8 to well above 5.3% in early October9. The September 30 reading also marked the 10-year's biggest quarterly rise since early 19943.

Bessent's own tone has changed. In September he said "I am the house now"5. By early October he was conceding that investors betting against him can win "a few hands." He said he cannot control the bond market and blamed higher oil prices tied to the Iran war, not a loss of faith in Treasuries8.

Why the buybacks backfired

Critics broadly agree on why the program hasn't worked, and the argument holds up. The first issue is size. Ed Yardeni called $6 billion of purchases "little more than a rounding error" in a $32 trillion market5. On the same day Treasury offered to buy back $6 billion of longer bonds, it also sold $39 billion in new 10-year notes1. The second issue is funding. Treasury cannot create money, so buybacks are paid for with new debt410. They swap old low-coupon bonds bought at deep discounts for new debt carrying much higher coupons. That can leave interest costs slightly higher, not lower4.

The third issue is signaling, and it may matter most. Analysts quoted by NBC described a "Streisand effect": a forceful effort to push yields down showed how worried officials were, which encouraged traders to test them1. Stanley Druckenmiller, Bessent's longtime mentor, wrote that once markets think Treasury is defending a price, each rise in yields becomes a test of official resolve1. He has also called the expansion a "mistake"7. Wall Street analysts have compared the effort to 1942, when the Fed agreed to hold rates down to fund wartime deficits1.

Underneath all this are fundamentals no buyback can fix. Those include deficits near $2 trillion a year on more than $40 trillion of debt7, an oil shock from the Iran war, and a quarter-point Fed rate increase driven by inflation worries3. One analysis also points to heavy corporate borrowing to fund AI spending as part of a worldwide repricing of sovereign debt2.

The chip trade: record profits, falling shares

For stocks, the knock-on effects are already showing up most clearly in semiconductors. On October 8, an Investing.com screen counted 34 U.S. chip stocks worth more than $10 billion that were down over 1% on the day. Arm fell more than 6%, Intel and Marvell lost about 5%, and Micron, Broadcom and AMD each dropped around 4%17. The next day, the Nasdaq fell 1.3%, its worst session since mid-August, as questions about OpenAI's revenue hit AI stocks11.

The gap between business results and share prices is the key point. Samsung reported third-quarter operating profit up about 780% to $80.2 billion12, yet its stock is roughly 25% below its June high11. TSMC said September sales rose 54.6% from a year earlier12. Micron's latest quarterly revenue was nearly five times its year-ago level18.

Opinions split here. One Investing.com columnist argued that chip strength is meeting a "more supportive" bond market and called the group a near-term buy12. The yield figures above don't support that view. The better reading is that higher discount rates are making investors less willing to pay up for future AI earnings. That matters a great deal right now, because Nvidia and Micron alone are expected to deliver more than a third of the S&P 500's third-quarter earnings-per-share growth18. Chips and chip equipment are projected to post 130% earnings growth, the most of any industry11.

The next big test is TSMC's earnings before the open on October 15, with consensus at about $4.48 a share15. ASML reports a day earlier20. Bulls point to TSMC's raised guidance for full-year growth slightly above 40%14. Still, the stock fell an average of 3.61% the day after each of its last two earnings beats14.

Bitcoin is trading like a rate-sensitive asset

Crypto has followed the same rate-driven pattern. Bitcoin fell more than 4% this week after failing near $87,60022. It dropped to about $80,300–$80,400 on Thursday before bouncing back toward $83,000 on Friday2123. U.S. spot bitcoin ETFs lost $484.9 million on Wednesday, their worst day since June 25, and another $244 million on Thursday as the 10-year hit 5.32%21. Liquidations across crypto topped $2.4 billion for the week, mostly from traders betting on higher prices30.

Crypto coverage consistently names higher Treasury yields and a stronger dollar as the main drags2224. That is a sharp change from early October. On October 5, bitcoin was above $86,700 and traders were talking about $90,00029. On a prediction market, traders put a 67% chance on bitcoin trading at or below $80,000 before October ends21. Analysts say $82,800 is the level to reclaim. Failing that, they see risk toward $77,00023.

Our reading

On the question in the headline, the coverage points one way. Bessent's buybacks have not lowered yields, so they cannot boost stocks through that channel. They may have done the opposite by making Treasury look worried. Treasury's options are shrinking. Bessent could expand purchases using the roughly $1 trillion Treasury General Account, a possibility senators have already asked about3, or cut long-bond issuance1. Either step would be a bigger and riskier move.

For now, oil, inflation and the Fed are driving the long end of the curve. That leaves chip stocks and bitcoin, two of the most rate-sensitive risk trades, carrying a burden that strong earnings have not yet offset. The October 27–28 Fed meeting29 and TSMC's report will show whether that changes.