MacroRoom

Monitors / Rates and credit

Yield driver

What is moving Treasury yields: rate expectations or compensation for risk?

MacroRoom·Measure by MacroRoom daily model; Federal Reserve DKW benchmark

Last updated: October 31, 2026·Update frequency: Daily; Fed benchmark released monthly

A Treasury yield can rise because the expected path of rates changes, or because investors require different compensation for risk. Start with the move in market yields, then look beneath it with daily economic estimates and the Fed’s published decomposition.

Treasury par yields · Oct 30, 2026 · Changes since Oct 23, 2026

Select a maturity to inspect its move. Red raises the yield; green lowers it. 1 basis point (bp) = 0.01 percentage point.

Loading the full daily history…

Data

01 · Market split

Where did the yield move?

Daily · 5 to 30 years

Separate the nominal yield move into the TIPS yield move and the change in the nominal–TIPS spread. Both contain expectations and risk premiums.

Market comparison dates

Presets use the last available close on or before the calendar date, within 7 days. Actual dates are shown. Custom dates are exact.

Oct 23, 2026 → Oct 30, 2026

10-year Treasury

−6.0 bp5.00% → 4.94%

The move, step by step

MacroRoom · U.S. Treasury · par yields

The curve before and after

MacroRoom · U.S. Treasury · par yields

Each step adds to the preceding change. The final bar is the net nominal move. The spread is a par-yield proxy for inflation compensation; it is not a pure inflation forecast.

Market componentStartEndChange
TIPS yield2.66%2.60%−6.0 bp
Nominal–TIPS spread2.34%2.34%0.0 bp
Nominal yield5.00%4.94%−6.0 bp

How the move accumulated

Changes from the selected start date; the TIPS and spread changes add to the nominal change.

MacroRoom · U.S. Treasury · change from start, bp

02 · Economic decomposition

Expectations or compensation for risk?

MacroRoom daily · 5 to 30 years

Our implementation of the DKW joint affine framework estimates four economic components from daily Treasury quotes. Coupon basis + fit reconciles the model’s zero-coupon yield to the quoted par yield.

Daily estimates through Oct 30, 2026. Parameters calibrated through Sep 30, 2026.

MacroRoom estimates, calibrated to published DKW and Fed yield curves. The 20- and 30-year components extend beyond the Fed’s published decomposition; their calibration sensitivity is shown below.

Economic model comparison dates

Presets use the last available close on or before the calendar date, within 7 days. Actual dates are shown. Custom dates are exact.

Oct 23, 2026 → Oct 30, 2026

10-year Treasury · par yield

−6.0 bp5.000% → 4.940%

MacroRoom · daily joint affine model · DKW framework

Changes in model estimates, not identified causes of a market event. Premiums can be negative. A positive contribution raises the yield even if that component’s level is negative.

ComponentStartEndChangeCalibration range
Expected real short ratesAverage expected real short rate over the bond’s life.1.317%1.286%−3.1 bp−3.1 bp to −2.5 bp
Expected inflationAverage expected CPI inflation over the same horizon.2.849%2.834%−1.5 bp−1.5 bp to −1.1 bp
Real term premiumCompensation, which can be negative, for real rate risk.0.825%0.807%−1.9 bp−2.0 bp to −1.9 bp
Inflation risk premiumCompensation, which can be negative, for inflation risk.0.114%0.110%−0.4 bp−0.5 bp to −0.4 bp
Coupon basis + fitQuoted par yield less the model zero-coupon yield; includes coupon convention and curve fitting error.-0.105%-0.097%+0.8 bp+0.1 bp to +0.8 bp
Observed nominal par yield5.000%4.940%−6.0 bp—

Calibration range spans the change estimated using history from 2010 and a shorter window from Jan 29, 2016. It is a sensitivity comparison, not a statistical confidence interval.

Latest model fit (Oct 30, 2026): 9.6 bp RMS across nominal and TIPS par quotes. The reconciliation row includes fitting differences.

The model’s TIPS liquidity premium at the end date is 0.51%. It helps separate inflation expectations from the nominal–TIPS spread; it is not an extra additive component of the nominal yield.

MethodologyWhat adds up, and what can be inferred

Two consistent accounting views

Daily market view: nominal par yield = TIPS par yield + nominal–TIPS par-yield spread. The same identity holds for changes and for differences between maturities. This is an exact subtraction of the published par rates, not a matched zero-coupon breakeven calculation. Coupon weights, liquidity, risk premiums and TIPS indexation can affect the spread.

Economic model: fitted nominal zero-coupon yield = expected average real short rate + expected average inflation + real term premium + inflation risk premium. MacroRoom estimates the daily state of a joint affine nominal/real model using the DKW Model LII pricing equations, including a separate TIPS liquidity factor. State updates use an iterated extended Kalman filter and price semiannual par coupons directly.

Parameters are calibrated when the Fed DKW release changes, using overlapping monthly observations of Fed zero-coupon curves and published DKW components. This minimum-distance calibration uses the published Model LII normalization and fixed innovation covariance. It is our implementation; it does not reproduce the authors’ full estimation from CPI and survey data. Daily changes are inferred from Treasury quotes.

The four components sum to a model zero-coupon yield. The daily view adds an explicit “coupon basis + fit” difference to reconcile to the observed Treasury par yield. The published-Fed view uses its original zero-coupon basis and fit residual. Comparisons always use the same model at both endpoints.

What happened to the recession discount?

Recession concerns can change expected real rates, inflation and the compensation investors accept for risk. They are not a separately identified fifth component. Subtracting today’s short rate from a long yield does not isolate recession risk: the relevant comparison is with the expected average future short rate. The fit residual is not a measure of fiscal concerns, supply pressure or recession probability.

Dates and units

1D compares with the previous published observation. 1W subtracts seven calendar days; 1M, 3M and 1Y use calendar offsets, clamped at month end. YTD starts at the last published close of the previous calendar year. Presets choose the last available close on or before that target, within seven days, and display the actual dates. Longer gaps remain unavailable. Custom dates use exactly the dates shown. Levels are percent per year; changes are basis points.

A slope is the longer maturity minus the shorter maturity. Its contributions are differences of the corresponding contributions. Rounding may prevent displayed numbers from adding exactly. Historical estimates are from the latest source vintage, not an archive of what investors could see on each date.

Coverage and model uncertainty

The market history begins in 2010; the 30-year TIPS series begins during that year. Each input retains its own latest observation. Missing quotes and failed source pulls are not filled forward. The DKW series is a staff research product, subject to delay and revision, and is not an official statistical release.

Twenty- and thirty-year components are priced at those horizons by the model. They are not published Fed component observations. Calibration includes nominal zero curves through 30 years and TIPS zero curves through 20 years; the daily filter includes quoted 30-year TIPS par yields. Long-horizon expectations depend on the model’s mean reversion and identifying assumptions. The two calibration windows illustrate one source of sensitivity; they do not capture all model uncertainty.

TIPS indexation lag, seasonality and deflation floors are not separately estimated here; their effects can enter the liquidity factor and measurement errors. Estimates can be revised after recalibration. Reconciliation does not establish that the economic attribution is uniquely identified.

Comparison with published DKW

The check below freezes calibration before Sep 30, 2025 and compares subsequent daily estimates through Sep 30, 2026. It uses today’s revised source history, so it is a holdout comparison, not an archived real-time backtest. Twenty- and thirty-year component benchmarks are not published.

Maturity / componentLevel RMS errorWeekly-change RMS error
5y · Expected real short rates2.4 bp0.9 bp
5y · Expected inflation1.6 bp0.3 bp
5y · Real term premium2.8 bp0.8 bp
5y · Inflation risk premium3.7 bp0.5 bp
10y · Expected real short rates4.7 bp0.6 bp
10y · Expected inflation1.1 bp0.3 bp
10y · Real term premium1.7 bp0.7 bp
10y · Inflation risk premium1.3 bp0.2 bp
Inputs and sourcesSeparate observations and release schedules
SourceThroughLast fetchedPublication
Treasury nominal par curve5y: Oct 30, 2026 · 10y: Oct 30, 2026 · 20y: Oct 30, 2026 · 30y: Oct 30, 2026Oct 30, 2026Oct 31, 2026Daily
Treasury TIPS par curve5y: Oct 30, 2026 · 10y: Oct 30, 2026 · 20y: Oct 30, 2026 · 30y: Oct 30, 2026Oct 30, 2026Oct 31, 2026Daily
MacroRoom economic estimates5y: Oct 30, 2026 · 10y: Oct 30, 2026 · 20y: Oct 30, 2026 · 30y: Oct 30, 2026Oct 30, 2026Oct 31, 2026Daily
Fed DKW decomposition5y: Sep 30, 2026 · 10y: Sep 30, 2026Sep 30, 2026Oct 31, 2026Monthly

Treasury curve methodology: constant-maturity par curves derived from market quotations. Nominal and real curves are separate inputs.

DKW model notes and data definitions: jointly models nominal yields, TIPS yields, inflation and survey expectations, with an explicit TIPS liquidity adjustment.

Fed nominal zero-coupon curves andTIPS zero-coupon curvesinform parameter calibration. Their weekly publication schedule does not delay daily state updates from Treasury par quotes.

DKW paper and supplementary parameter tables ·MacroRoom model specification.

Notes shipped with the current model file

"Components of 5- and 10-year nominal yields and inflation compensation and 5-to-10-year nominal forward rate and forward inflation compensation."

" "

"Source: "

"D'Amico, Kim, and Wei (2018), ""Tips from TIPS: The Informational Content of Treasury Inflation-Protected Security Prices,"" "

"updated by Kim, Walsh, and Wei (2019), FEDS notes, ""Tips from TIPS: Update and Discussions."""

" "

"The latest data can be downloaded from :"

"https://www.federalreserve.gov/econres/notes/feds-notes/DKW_updates.csv"

" "

"Notes: "

"Data are updated monthly; all rates are in percentage points."

"This release is based on a model re-estimation using data through November 12, 2025. Other changes relative to the Kim, Walsh, and Wei (2019) FEDS note include using 1-year survey inflation forecasts from Blue Chip Financial Forecasts (monthly) instead of 1-year survey inflation forecasts from Survey of Professional Forecasters (quarterly) and adding 15- and 20-year nominal and TIPS yields to the estimation."

"This is not an official Federal Reserve statistical release. Because this is a staff research product and not an official statistical release, it is subject to delay, revision, or methodological changes without advance notice."

" "

DownloadMarket construction and original model output

Market split · CSV — MacroRoom’s construction from Treasury par yields. One row per date and maturity, in percentage points; blanks preserve missing observations.

Economic model · original Federal Reserve CSV — published components and source notes. The download is hosted by the Fed.

Daily economic estimates · CSV — MacroRoom model components, coupon basis + fit, liquidity estimate, fitted zero and par yields, and the observed nominal par yield.

Parameters and validation report — calibration dates, source hashes, sensitivity model and holdout errors.

Market through Oct 30, 2026; daily model through Oct 30, 2026; Fed model through Sep 30, 2026. Retrieved Oct 31, 2026.