T10Y3M: 10-Year Treasury Minus 3-Month Bill Yield Spread Daily Data (1982–2026)

The T10Y3M series, published daily by the Federal Reserve via FRED, tracks the spread between the 10-year US Treasury constant maturity yield and the 3-month Treasury bill yield since January 1982 — over 11,000 daily observations. T10Y3M is the recession indicator used by the Federal Reserve Bank of New York in its official recession probability model, and academic research has repeatedly found it to outperform the more popular 10Y-2Y spread as a leading indicator across post-1980 US cycles.

Dataset: Yield Curve Spread 10Y–3M (1982–2026) · Updated 2026-05-19

Source: FRED series T10Y3M · Federal Reserve Bank of St. Louis

Macro Takeaway

Dataset Overview

Dataset Variables

The CSV and Excel files contain the following columns.

Column names match the CSV headers exactly.

Download the Complete Dataset

The full dataset is available in CSV and Excel formats.

FRED Direct CSV Access

The underlying data is available from FRED under series code T10Y3M:

https://fred.stlouisfed.org/graph/fredgraph.csv?id=T10Y3M

Direct CSV Access — Eco3min Structured Dataset

https://eco3min.fr/dataset/yield-curve-10y-3m.csv

This URL returns the complete dataset in CSV format. It can be used directly in pandas, R, curl, or any data tool.

Using the Dataset in Python

import pandas as pd

url = "https://eco3min.fr/dataset/yield-curve-10y-3m.csv"
df = pd.read_csv(url, parse_dates=["date"])

print(df.head())
print(df["spread_10y_3m"].describe())

Using the Dataset in R

library(readr)

url <- "https://eco3min.fr/dataset/yield-curve-10y-3m.csv"
df <- read_csv(url)

head(df)
summary(df$spread_10y_3m)

Both examples load the dataset directly from the URL — no download or API key required.

Methodology

T10Y3M is computed by the Federal Reserve as the arithmetic difference between the 10-year US Treasury constant maturity yield (FRED series DGS10) and the 3-month Treasury bill secondary market rate (FRED series DTB3), expressed in percentage points. The constant maturity yields are themselves derived from the Treasury’s daily yield curve, which fits a piecewise model to outstanding Treasury issues; the 3-month bill rate is the market-quoted secondary rate, not the auction rate.

The series is published daily by the Board of Governors of the Federal Reserve System as part of the H.15 statistical release, typically around 4:15 PM Eastern Time on each US business day. The series is not seasonally adjusted, is not revised once published, and contains no observations on weekends or US federal holidays. Eco3min mirrors FRED with an automated pull cadence.

Data Quality & Provider Notes

T10Y3M is among the most reliable and stable series in the FRED catalogue: it is computed mechanically from two daily Treasury reference rates published by the Federal Reserve under the H.15 release, with no judgment, no estimation, and no revision policy. The series is suitable for automated quantitative pipelines without manual reconciliation.

• Release latency. The Federal Reserve publishes the underlying H.15 components around 4:15 PM Eastern Time on each US business day; T10Y3M is typically available on FRED within minutes after the underlying components are released. Eco3min mirrors FRED with a daily automated pull.
• Revisions policy. The series is not revised once published. Both components (DGS10 and DTB3) are market-observed reference rates, so corrections beyond same-day data-feed fixes are extremely rare.
• Alternative sources. Bloomberg, Refinitiv (LSEG) and Haver Analytics distribute the same H.15 components under their own tickers, with intraday refresh rather than the once-daily FRED snapshot. ALFRED (FRED’s vintage archive) provides historical real-time snapshots of T10Y3M for reproducibility of backtests.
• Known gaps. The series contains no observations on weekends or US federal holidays — expect roughly 251 to 253 observations per calendar year. The 3-month bill leg uses the secondary market rate; in episodes of severe bill scarcity (e.g., late 2008, early 2020), the secondary rate can dislocate briefly from the auction rate.

Before running historical regressions on T10Y3M, verify the last observation date on FRED to confirm the most recent business-day update has been pulled into your local copy.

Common Pitfalls When Using T10Y3M

T10Y3M is widely used as a recession indicator but several recurring interpretation errors distort the signal, particularly in tactical or short-horizon analyses.

1. Confusing the daily reading with the NY Fed model input. The Federal Reserve Bank of New York’s official recession probability model uses the monthly average of T10Y3M, not the daily reading. The daily series is considerably noisier, and brief intraday or single-day inversions historically have not coincided with the model’s recession signals.
2. Treating mild and deep inversions as equivalent. A T10Y3M reading of −0.10pp and one of −1.50pp carry very different historical signal weight. Empirical studies of post-1968 US cycles have documented that both the depth and the persistence of inversion correlate with subsequent recession probability, not just the binary inverted/non-inverted state.
3. Reading the inversion as a market timing signal. The lead time between T10Y3M inversion and NBER recession start has historically ranged from approximately 6 to 18 months across post-1968 cycles, with significant variance. Using inversion timing for directional positioning has produced inconsistent results across observers; the signal is about regime change, not timing.
4. Forgetting that the spread does not measure recession depth. T10Y3M is a probability signal, not a magnitude signal. The 1980 and 1981 inversions were among the deepest on record but were followed by recessions of very different durations and depths. The series captures the conditional probability of recession occurrence, not its eventual severity.

Historical Regimes

1982–1989 — Post-Volcker normalization. Following the deep inversions of 1980 and 1981 (which preceded the 1980 and 1981–1982 recessions), T10Y3M recovered into positive territory as the Volcker disinflation succeeded. The spread averaged approximately +2 percentage points through the mid-1980s, consistent with a normalizing term structure and falling inflation expectations.

1989–1990 — First post-Volcker inversion. T10Y3M inverted briefly in mid-1989 ahead of the July 1990 NBER recession. The inversion was shallow (−0.3 to −0.5pp) and short-lived, but the recession that followed validated the signal. This episode established the post-1980 pattern of brief inversions preceding US recessions.

2000–2001 — Dot-com inversion. T10Y3M inverted in mid-2000 as the Fed raised rates to slow the equity bubble. The March 2001 recession followed within roughly 9 months. The inversion coincided with the equity peak in March 2000 by a few months, though the relationship between the spread and equity markets is not mechanical.

2006–2007 — Pre-GFC inversion. T10Y3M inverted in mid-2006 and remained inverted for over a year before the December 2007 NBER recession start. This was one of the longest pre-recession inversions on record. The 10-year Treasury yield remained anchored around 4.5–5% while the 3-month bill rose to over 5% under the Bernanke tightening cycle.

2008–2019 — Extended positive regime. Following the GFC and the introduction of zero interest rate policy plus large-scale asset purchases, T10Y3M remained positive for over a decade, often in the +1 to +3pp range. The 3-month leg was pinned near zero by Fed policy while the 10-year yield reflected term premium and growth expectations.

2019 — Brief pre-COVID inversion. T10Y3M inverted briefly in mid-2019, ahead of the February 2020 COVID recession. The signal was noisy and was largely dismissed at the time given the absence of obvious cyclical strain; in hindsight, the inversion preceded the recession by approximately 8 months, in line with the historical range.

2022–2024 — Deepest inversion since 1981. The Fed’s aggressive 2022–2023 hiking cycle drove the 3-month bill above 5% while the 10-year yield lagged, producing a T10Y3M inversion that exceeded −1.5pp at its trough — the deepest reading since 1981. The inversion persisted from late 2022 through mid-2024, longer than any prior post-1980 episode. The relationship between this inversion and subsequent growth dynamics is analysed in the NY Fed recession probability model study, which uses T10Y3M as its central input.

2024–2026 — Post-inversion re-steepening. As the Fed began its rate-cutting cycle in late 2024, T10Y3M re-steepened sharply. Historically, the re-steepening from a deep inversion has often coincided with the onset rather than the avoidance of recession, though the post-2024 sequence is still being analysed and remains an open empirical question.

Related Macroeconomic Datasets

T10Y3M is mechanically constructed from two Treasury reference rates and sits at the heart of a network of related series — its two components, the broader yield curve, and the Fed’s policy rate that drives the short leg.

Macroeconomic Dataset Hub

This dataset is part of the Eco3min macro-financial data repository.

Explore the Eco3min Dataset Hub

Sources

• Federal Reserve Bank of St. Louis — FRED database (series T10Y3M, DGS10, DTB3)
• Board of Governors of the Federal Reserve System — H.15 Selected Interest Rates
• Federal Reserve Bank of New York — Yield curve and recession probability research

Dataset Reference

Eco3min Research. (2026). T10Y3M — Yield Curve Spread 10Y–3M (1982–2026).<br> https://eco3min.fr/en/yield-curve-spread-10y-3m-dataset/