By Kresmion Research, October 6, 2026
The US 10-year Treasury yield rose 53 basis points between August 31 and October 2, from 4.75 to 5.28 percent, and 48 of those basis points were real yield. The 10-year TIPS real yield went from 2.44 to 2.92 percent, with 2.93 on September 30, a level not seen since November 2008, while the 10-year breakeven, the inflation rate the market builds into the nominal yield, moved from 2.31 to 2.36 percent (Kresmion fred_series_history, FRED DGS10, DFII10, T10YIE). Through September 28 the whole curve rose together, after the Fed's first rate increase since 2023. Hike odds for October went from about 71 percent last week to about 26 percent on October 5, and in the four sessions to October 2 the 2-year gave back 9 basis points while the real yield rose 2. Four sessions is a thin sample.
Key takeaways
| What | Number |
|---|---|
| 10-year Treasury yield, August 31 to October 2 | 4.75 to 5.28 percent, plus 53bp (Kresmion fred_series_history, FRED DGS10) |
| 10-year TIPS real yield over the same dates | 2.44 to 2.92 percent, plus 48bp, about 90 percent of the nominal move. 19 of 5,943 daily prints since 2003-01-02 were at or above 2.92, 17 of them in 2008; the high of this episode is 2.93 on September 30, and the series maximum is 3.15 (FRED DFII10) |
| 10-year breakeven inflation rate | 2.31 to 2.36 percent, plus 5bp over the whole window; the October 5 print is 2.36 (FRED T10YIE) |
| Three periods | To September 15 (before the decision): 2-year plus 33bp, 10-year plus 25bp, real yield plus 18bp. September 15 to 28: 2-year plus 25bp, 10-year plus 24bp, real yield plus 28bp. September 28 to October 2: 2-year minus 9bp, 10-year plus 4bp, real yield plus 2bp (FRED DGS2, DGS10, DFII10) |
| Fed target range, upper bound | 3.75 to 4.00 percent, the first increase since 2023 after six cuts between September 2024 and December 2025 (FRED DFEDTARU) |
| 2-year yield minus the policy-rate midpoint | Plus 95.5bp, about the 95th percentile since 2009 and the 93rd since 2016, a gap consistent with the market pricing more tightening than the Fed has delivered (FRED DGS2, DFEDTARU) |
| 10-year yield minus the policy-rate midpoint | Plus 140bp: the 50th percentile since 2009, and the 94th percentile on days with the policy upper bound at or above 1 percent, where the median is plus 18bp (FRED DGS10, DFEDTARU) |
What moved, and by how much
A nominal Treasury yield splits into two parts: the yield on inflation-protected TIPS of the same maturity, and the breakeven, which is the difference between the two. On October 2 the 10-year sat at 5.28 percent, the TIPS real yield at 2.92 percent and the breakeven at 2.36 percent, and 2.92 plus 2.36 gives 5.28. On August 31 the same sum was 2.44 plus 2.31, or 4.75.
The breakeven moved 5 basis points over the whole window, against 48 for the real yield. The 22-trading-day change in the real yield of plus 0.48 sits at the 97.3rd percentile since 2003, about 2.1 standard deviations. It is large, and moves of this size have happened before: 168 days including October 2 matched or exceeded it, in about 16 separate episodes, with 2022 alone holding 62 of them.
Three periods
The decision was announced on September 16, so September 15 is the last pre-decision close. The Fed raised its target range to 3.75 to 4.00 percent at the September 15 to 16 meeting, and FRED's series carries the new upper bound from September 17 (FRED DFEDTARU; Federal Reserve calendar).
| Close | 2-year | 10-year | 30-year | 10-year real | Breakeven |
|---|---|---|---|---|---|
| August 31 | 4.34 | 4.75 | 5.25 | 2.44 | 2.31 |
| September 15 | 4.67 | 5.00 | 5.36 | 2.62 | 2.38 |
| September 28 | 4.92 | 5.24 | 5.56 | 2.90 | 2.34 |
| October 2 | 4.83 | 5.28 | 5.63 | 2.92 | 2.36 |
From August 31 to September 15 the 2-year rose 33 basis points, the 10-year 25 and the real yield 18. A short end that leads is consistent with a repriced policy path, and the six prior changes since 2024 were all cuts, so a hike after six cuts can signal further increases.
From September 15 to 28 everything rose together: the 2-year 25 basis points, the 10-year 24, the 30-year 20 and the real yield 28. Odds of an October hike were about 71 percent by late September, so a policy path still priced higher fits this stretch.
From September 28 to October 2 the picture changed. The odds of an October hike went from about 71 percent last week to about 26 percent on October 5, a reading taken after the last FRED close, so part of the drop may postdate these yields (CME FedWatch, Reuters via the Spokesman-Review). The 2-year fell 9 basis points, from 4.92 to 4.83 percent. The 10-year rose 4, the 30-year 7 and the real yield 2, from 2.90 to 2.92. A real yield that holds while the near-term policy path softens is what a term-premium or global-spillover reading predicts, but four sessions is a thin sample.
The 10-year gap to the policy midpoint is also high, the 94th percentile with the upper bound at or above 1 percent. Like the 2-year gap, it fits both a policy-path and a term-premium reading.
The weak payrolls report did not stop it
September nonfarm payrolls rose 29,000, below the 12-month average of 45,000. The unemployment rate held at 4.2 percent, and July and August were revised down by a combined 60,000 (July to minus 10,000, August to 133,000) (BLS Employment Situation). A cooling labour market is a reason for a lower policy path, and the real yield did not give the rise back: 2.88 percent on October 1 and 2.92 on October 2, against 2.93 on September 30. The ISM services prices-paid index reached 74.0 in September, the highest since July 2022 (Spokesman-Review, above).
The alternative explanations
At least three readings compete, and the data on hand cannot fully separate them.
The first is a rising term premium from Treasury supply. Issuance can lift the premium for holding duration without any change in the expected policy path. This week the Treasury sells $58 billion of 3-year notes on Tuesday, $39 billion of 10-year notes on Wednesday and $22 billion of 30-year bonds on Thursday, $119 billion in total (U.S. Treasury auction announcements; Kresmion treasury_auctions). Those sizes are unchanged from earlier in 2026 and smaller than the August refunding sizes of $42 billion and $25 billion for the 10-year and 30-year (Kresmion treasury_auctions), so supply volume does not itself explain the move, and the argument would be about demand for duration. The Kresmion series do not include a term-premium estimate, so this reading cannot be measured directly.
The second is global spillover. The Irish Times reported on October 1 that the 10-year Treasury had touched 5.34 percent intraday, its highest since 2002, as French 10-year yields hit roughly a 25-year high, UK 30-year gilts passed 6 percent for the first time since 1998 and Japan's 10-year yield approached 3.1 percent (Irish Times). The highest FRED close in the window is 5.29 percent, on September 30. The same article cites forced selling by hedge funds and other big investors as part of the October 1 move. Kresmion cannot separate spillover or forced selling from the term-premium reading.
The third is the oil shock. Brent closed at $102.31 a barrel on October 1 (Kresmion futures_prices). Higher energy prices could, in principle, lift the expected policy rate and with it the real yield. In that reading energy is the driver and the policy path carries it through to yields. A flat breakeven fits that only partly, since an energy shock should lift breakevens too, and the breakeven moved 5 basis points over the window.
The evidence against
The last-week reading rests on four sessions. The 2-year's 9 basis point fall and the real yield's 2 basis point rise are small, and hike odds were still about 71 percent the week before October 5, so most of the rise sits in a period in which the policy path was priced firmly higher.
A move of this size has a base rate. The 22-day change in the real yield was matched or exceeded on 168 days since 2003, so a month like this recurs. A level this high has been reached on 19 days, 17 of them in 2008 and the other two on September 30 and October 2.
On October 5 the Nasdaq Composite closed at a record, up 1.05 percent, while 206 Nasdaq-listed stocks made 52-week lows against 46 highs (StockTitan). The index did not sell off, and new lows outnumbered new highs that day. The regime engine's equity-momentum z-score is minus 0.14 (Kresmion macro_regime).
Kresmion's regime engine reads Risk-Off with high conviction, smoothed score minus 0.30, 2 days old. Its MOVE z-score of plus 1.40 and high-yield spread z-score of plus 1.14 are consistent with a rate-volatility shock. The engine's real-yield z-score is plus 0.91, a separate measure from the 19-of-5,943 rank above. The 10-year minus 3-month spread of 109 basis points is the 36.5th percentile since 1981, so the curve is not unusually steep. The regime engine's 2-year to 10-year curve z-score is plus 2.58, a different measure.
Reserves are low and funding rates are firm against the interest rate on reserve balances. The overnight reverse repo facility held $1.0 billion on October 5, bank reserves were $2.948 trillion on September 30, the 6.7th percentile of weekly prints since 2021, and SOFR minus that rate was minus 2 basis points on October 2, the 83.4th percentile of 1,293 days since July 2021 against a median of minus 9 (Kresmion fred_series_history, FRED RRPONTSYD and WRESBAL, and Kresmion money_markets). It is not a stress signal in this data, and it fits tighter liquidity.
What would change the read
Each item below is observable this month.
Whether the 10-year real yield holds near 2.9 percent if the 2-year keeps falling. Holding supports the term-premium or spillover reading. A real yield that follows the 2-year down sends the story back to the policy path.
The 10-year auction on Wednesday, October 7. A large tail against the pre-auction yield would point to weak demand for duration. A clean auction leaves the question open.
The FOMC minutes of the September 15 to 16 meeting, scheduled for October 7 (Federal Reserve calendar). Language that describes further increases as likely would put the policy path back in charge. Language that treats the hike as a one-off would weaken it.
The 10-year breakeven. At 2.36 percent it has moved 5 basis points in a month. A climb of 15 or 20 basis points while the real yield holds would turn this into an inflation-compensation story.
September CPI on October 14 at 8:30 AM (BLS release schedule). A hot print would test whether the breakeven stays flat. The next FOMC meeting is October 27 to 28.
Frequently asked questions
What is a real yield and why does it matter?
A real yield is the return on an inflation-protected Treasury, quoted after inflation compensation is stripped out. The nominal 10-year yield equals the 10-year TIPS real yield plus the breakeven inflation rate, so a move in the nominal yield can be traced to one or both. On October 2 that sum was 2.92 plus 2.36, which gives 5.28 percent.
Does a flat breakeven mean inflation risk has gone away?
It means the 10-year market is not adding compensation for inflation. The breakeven prices inflation protection in the bond market; the ISM index reports what firms pay, and it stood at 74.0.
Why did yields rise after a weak jobs report?
Payrolls of 29,000 argue for a lower policy path, and the real yield rose over the same window. Daily closes show no relief on the report day: the 10-year went from 5.24 on October 1 to 5.28 on October 2, and the 2-year from 4.78 to 4.83. The move had started earlier, with the 10-year at 5.29 percent on September 30.
What is on the calendar this week?
Auctions of 3-year, 10-year and 30-year Treasuries run Tuesday to Thursday, with the 10-year on October 7, the same day the FOMC minutes are scheduled. September CPI follows on October 14, and the next FOMC meeting is October 27 to 28.
Sources
Kresmion fred_series_history, FRED DGS2, DGS10, DGS30, DFII10, T10YIE, DGS3MO, daily observations 2003-01-02 to 2026-10-05; last FRED print for DGS10 and DFII10 is 2026-10-02.
Kresmion fred_series_history, FRED DFEDTARU, federal funds target range upper limit, change from 3.75 to 4.00 on 2026-09-17.
Kresmion macro_regime, regime, score, factor and z-scores including equity momentum, data as of 2026-10-06.
Kresmion fred_series_history, FRED RRPONTSYD (2026-10-05) and WRESBAL (2026-09-30), and Kresmion money_markets, SOFR minus IORB (2026-10-02).
Kresmion futures_prices, Brent (BZ=F) settlement close 2026-10-01.
Kresmion treasury_auctions, announced 2026-10-01; auctions 2026-10-06 (3-year $58bn), 2026-10-07 (10-year $39bn), 2026-10-08 (30-year $22bn); August refunding 10-year $42bn and 30-year $25bn.
U.S. Treasury, upcoming auction announcements: https://www.treasurydirect.gov/auctions/upcoming/
BLS, The Employment Situation, September 2026: https://www.bls.gov/news.release/empsit.nr0.htm
Reuters via the Spokesman-Review, US services sector cools in September, price pressures building (ISM services and October hike odds, October 5): https://www.spokesman.com/stories/2026/oct/05/us-services-sector-cools-in-september-price-pressu/
Irish Times, Global bond sell-off pushes 10-year Treasury yield to highest since 2002 (October 1): https://www.irishtimes.com/business/2026/10/01/global-bond-sell-off-pushes-10-year-treasury-yield-to-highest-since-2002/
Federal Reserve, FOMC meeting calendar 2026: https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm
StockTitan, Why small caps are lagging the Nasdaq record in 2026: https://www.stocktitan.net/articles/nasdaq-record-small-caps-lagging-october-2026
BLS, Schedule of CPI releases: https://www.bls.gov/schedule/news_release/cpi.htm