The 10-year Treasury yield broke through 5% on September 23 for the first time, then surged all the way to 5.31%. Wall Street offered various explanations — oil prices, AI, Japan, France.
But Philip Marey, Senior US Strategist at Rabobank, in his latest report titled "Higher Ground," splits the rise in the 10-year Treasury yield into two distinct phases. He argues that the Federal Reserve pushed yields to 5%, and US debt pushed them to 5.31%.
Marey wrote in the report: "The sustained rise in the 10-year Treasury yield in 2026 largely reflected the Fed's shift from considering rate cuts to actually beginning hikes in September... However, the recent sharp break above 5% was mainly attributable to a rise in the term premium, which stemmed from market concerns about public debt."
Phase One: The Fed Drives Treasury Yields Higher
Marey used the New York Fed's ACM model to decompose the 10-year yield into two components: the risk-neutral yield and the term premium.
Data shows that during the first three weeks of September, the 10-year yield rose from 4.80% to 4.96%, while the term premium actually fell from 0.78% to 0.57%. Meanwhile, the risk-neutral yield jumped from 4.09% to 4.43%. The risk-neutral yield roughly equals the market's average expectation for the federal funds rate over the next decade, reflecting expected monetary policy path. The term premium refers to the extra compensation investors demand for holding long-term bonds versus rolling over short-term bonds. A rising term premium means investors perceive greater risk in holding long-dated bonds and require higher yields to be willing to buy them.
During this period, the Fed hiked rates by 25 basis points and signaled through its dot plot that another hike was coming this year. Marey's assessment: "This means the Fed's expected rate path was the primary driver of the 10-year yield climbing toward 5%."
Phase Two: After September 23, Debt Takes Over Treasury Yields
On September 23, the script flipped.
The term premium began surging from its low of 0.57%, reaching 0.98% by October 7. Meanwhile, the risk-neutral yield actually edged down slightly from 4.43% to 4.39%.
The bond market stopped worrying about the Fed and started worrying about the borrower itself.
Marey pinpointed the turning point to a Treasury auction on September 23. That day, the US Treasury auctioned $70 billion in 5-year notes, with a high yield of 5.033% — the first time above 5% since 2007. The tail was 3.1 basis points, the second-largest on record. The indirect bidder share (typically representing overseas central banks and other institutions) fell to its lowest since March 2020, forcing primary dealers to absorb 15.8% of the allotment.
It was the worst single-day performance in the bond market since "Liberation Day."
Rabobank's diagnosis: "This suggests that concerns about public debt played an important role in the 10-year Treasury yield's sharp break above 5%. The synchronized move in the term premium and the MOVE index (a measure of implied volatility in the US Treasury market) indicates that the unsettling 5-year note auction that day was the main trigger."
Inflation Expectations Stable, Real Rates Rising
Marey used another tool — TIPS (Treasury Inflation-Protected Securities) — to decompose the 10-year yield into the breakeven inflation rate and the real yield.
The results were not kind to those arguing the Fed is losing control.
From September 1 to October 5, the real yield rose from 2.44% to 2.95%, while the breakeven inflation rate barely budged, inching up from 2.35% to just 2.36%.
Rabobank's interpretation: "Energy price increases triggered by the Middle East war and a resilient US economy pushed up the Fed's (real) policy rate path and the 10-year real yield. At the same time, the Fed's hawkish response kept inflation expectations anchored."
In other words, the market believes the Fed can control inflation, but confidence in fiscal sustainability is wavering.
Goldman Sachs: Mostly the Fed, a Small Part "Hard to Explain"
Goldman Sachs rate strategist William Marshall, in a report published the same day titled "What's Behind the Rise in Treasury Yields," reached similar but not identical conclusions.
Goldman's macro model shows that since the eve of the Iran war, the 10-year yield has risen a cumulative 130 basis points, of which about 70% can be attributed to hawkish Fed repricing. Marshall wrote:
"So far this year, the market's consensus expectation for 1-year short rates has risen by about 125 basis points, contributing roughly 55 basis points to long-end forward rates, accounting for 70% of the cumulative increase... Overall, about 10 to 15 basis points of the year-to-date yield increase cannot be explained by fundamental factors."
Goldman also acknowledged that the most recent selloff was concentrated in the long end and that its own term premium estimates tell the same story as Rabobank's MOVE chart.
The disagreement lies in: Why is the term premium rising?
Rabobank's answer: Fiscal debt.
Goldman's answer is more diffuse: Rising macro uncertainty, plus the fact that Treasuries are no longer an effective hedge for equities — "bonds have become a poor diversifier for stocks."
On fiscal issues, Marshall's language was almost understated:
"Although the US fiscal trajectory is unsustainable, there has been no substantive shift this year, and privately held Treasuries as a share of GDP have been broadly stable this year. What has truly changed this year is the market's expectations for AI-related spending and borrowing."
Diverging Outlooks: New Floor or Buying Opportunity?
The two banks show clear divergence in their forward-looking assessments.
Rabobank: Yields Have Found a New Floor
Marey's logic: Before the war, the Fed was expected to cut once this year and once more in 2027. Now it has shifted to a September hike and another by year-end, with the first cut pushed to 2028, and neutral rate expectations also being revised up. He wrote:
"Because the AI boom is transformative and may continue to support the economy beyond this year, we believe downside from monetary policy expectation revisions is limited, and the 10-year Treasury yield is unlikely to return to the sub-4.30% levels of the six months before the war. Note that we are now a full percentage point above that level."
On the term premium, Rabobank's warning is more direct: "Because the recent peak in the term premium is related to public debt, and we do not expect any progress on that front, the term premium may occasionally retreat when public debt or geopolitical concerns ease, but it is likely to continue trending higher. Therefore, even if market expectations for Fed hikes contract somewhat, the downside impact on the 10-year yield will be limited, because the public debt problem will not go away. In fact, it may become the most persistent driver pushing rates higher in the years ahead."
Rabobank forecasts: The 10-year yield at 5.20% by year-end, 5.10% by mid-2027, and 5.00% by end-2027. Even under its "optimistic scenario," the end-2027 yield would still be about 70 basis points above pre-war levels.
Goldman Sachs: Time to Build Duration
Marshall believes now is the time to "establish a foothold in duration positions." He calculates that the 5y5y forward yield (the 5-year rate five years from now) is in the 5.5% to 5.6% range, 40 to 50 basis points above fair value, approaching one standard deviation cheap.
Goldman's bond fund duration indicator has slid to -2, representing an underweight position. "Historically, such underweight levels have tended to favor lower yields over the following 3 to 6 months."
Additionally, Goldman estimates that the negative convexity effect from MBS (mortgage-backed securities) has forced approximately $60 billion in 10-year equivalent selling since early September. MBS convexity is now close to zero overall, and this amplifier is essentially exhausted.
Goldman's year-end 10-year target: 4.75%, a full 45 basis points below Rabobank's 5.20%.
This Week's Focus: CPI, Retail Sales, and Warsh's Speech
The bond market reopened after the Columbus Day holiday. According to Bloomberg, the 10-year yield fell as much as 6 basis points to 5.18%, after briefly touching 5.36% the prior week — the highest since 2002.
Key events this week:
CPI (Wednesday): Goldman expects core CPI at 0.24% month-over-month, above the consensus of 0.2%; JPMorgan economist Feroli forecasts 0.18%. Goldman's mortgage division warned that "if the data comes in slightly above expectations, it could pose trouble for the rates market."
Retail Sales and PPI (Thursday): JPMorgan's market intelligence desk said that if both core inflation and retail sales soften, "it could be enough to trigger a tactical bond market rally," while noting that some clients believe the 10-year yield is on an "unstoppable path toward 5.50%," a level JPMorgan flags as "extremely unfavorable for stocks."
Warsh Speech (Thursday): The market is currently pricing about 5 basis points for an October hike, with the main bet on the next move in December.
Meanwhile, Apollo Chief Economist Torsten Slok, using another Fed term premium model (D'Amico-Kim-Wei), arrived at the same two drivers as Rabobank: an upward shift in the expected path of real short-term rates, and "investors demanding more compensation for holding duration."
IMF official Athanasios Vamvakidis told Bloomberg that the surge in government bond yields has become "the biggest emerging risk to the global financial system."
Comments