T10Y2Y: Daily 10-Year Minus 2-Year US Treasury Yield Spread from FRED (1976–2026)

This dataset covers T10Y2Y daily since June 1976 — over 12,000 observations spanning every modern curve inversion. Each inversion since 1976 has been followed by a US recession, with a lead time of 6 to 24 months.

Dataset: Yield Curve Spread 10Y–2Y (T10Y2Y) · Updated 2026-08-03

Source: FRED series T10Y2Y · Federal Reserve Bank of St. Louis

Macro Takeaway

Dataset Overview

Dataset Variables

The CSV and Excel files contain the following columns. Each row represents one business day.

Negative values indicate yield curve inversion.

Download the Complete Dataset

The full T10Y2Y dataset covers nearly 50 years of daily yield curve observations, including every major inversion episode. Each option is weighed on its own terms in 2-year versus 10-year, by the data.

FRED Direct CSV Access

The spread is directly available from FRED under series code T10Y2Y:

https://fred.stlouisfed.org/graph/fredgraph.csv?id=T10Y2Y

The Eco3min dataset provides this same series with consistent column naming and stable download URLs, suitable for automated scripts and academic citation.

Direct CSV Access — Eco3min Structured Dataset

https://eco3min.fr/dataset/yield-curve-10y-2y.csv

This URL returns the complete T10Y2Y dataset in CSV format.

Using the Dataset in Python

import pandas as pd

url = "https://eco3min.fr/dataset/yield-curve-10y-2y.csv"
df = pd.read_csv(url, parse_dates=["date"])

# Count inversion days
inversions = df[df["spread_10y_2y"] < 0]
print(f"Total inversion days: {len(inversions)}")
print(f"Current spread: {df['spread_10y_2y'].iloc[-1]:.2f}%")

Using the Dataset in R

library(readr)

url <- "https://eco3min.fr/dataset/yield-curve-10y-2y.csv"
df <- read_csv(url)

head(df)
summary(df$spread_10y_2y)

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

Methodology

T10Y2Y is computed by subtracting the 2-year Treasury constant-maturity rate (DGS2) from the 10-year rate (DGS10). Both input series are interpolated by the US Treasury from the daily yield curve of outstanding marketable securities, using the same constant-maturity methodology described in the 10-year Treasury yield dataset.

The spread is expressed in percentage points. A value of 1.50 means the 10-year yield is 150 basis points above the 2-year. A value of −0.50 means the curve is inverted by 50 basis points. The “zero line” — where T10Y2Y crosses from positive to negative — is the threshold that defines inversion. For the broader picture: the interactive tool that replays the curve’s shape.

T10Y2Y captures the slope of the intermediate portion of the yield curve. An alternative measure, the 10Y–3M spread (T10Y3M), uses the 3-month Treasury bill instead and is preferred in some academic literature — notably the Federal Reserve Bank of New York’s recession probability model. Both are included in the Eco3min dataset library.

This Eco3min dataset is updated weekly (Saturday 08:00 UTC) via automated pull from the FRED API.

Data Quality & Provider Notes

T10Y2Y is one of the most reliable high-frequency series in the FRED catalog: it is a deterministic subtraction of two constant-maturity yields, both published in the Federal Reserve H.15 release with no revisions and no seasonal adjustment. The series is internally consistent across the full 1976-onwards sample.

• Release latency. The Federal Reserve publishes H.15 with the prior business day’s DGS10 and DGS2 quotes around 4:15 p.m. ET; FRED computes T10Y2Y within minutes. Eco3min mirrors the FRED endpoint with a weekly Saturday pull, so the dashboard typically lags the FRED feed by up to one week between scheduled refreshes.
• Revisions policy. T10Y2Y is not revised once published. It is a same-day subtraction of two end-of-day quotes — vintage analysis can use the current download as a faithful real-time series.
• Alternative sources. Bloomberg (USYC2Y10), Refinitiv/LSEG and Haver Analytics distribute the same H.15-derived spread with no material differences. ALFRED preserves vintage versions for real-time research.
• Known gaps. Weekends and US federal holidays are excluded. Because T10Y2Y is derived from DGS10 and DGS2, any one-day gap in either input creates a gap in the spread. The series is otherwise continuous from June 1976 onwards.

Before reproducing inversion statistics from third-party sources, confirm whether they used end-of-day, end-of-week, or monthly averages — the same calendar inversion can produce different counted days depending on the aggregation.

Common Pitfalls When Using T10Y2Y

T10Y2Y is one of the most-cited macro indicators in modern markets, which makes recurring interpretation errors particularly costly. Four pitfalls dominate the literature.

1. Confusing intraday touches with durable inversion. A T10Y2Y reading briefly below zero on a single day is not equivalent to a sustained inversion. Historical recession-signal research typically defines inversion as monthly or quarterly averages below zero, not isolated daily prints. The 1989, 2000, 2006 and 2022 episodes were all sustained inversions of several months.
2. Equating inversion date with recession start. The NBER recession start dates lag the T10Y2Y inversion by 6 to 24 months. Reading the inversion as a “recession now” signal misrepresents the historical record — the inversion announces the conditions that precede recession, not the recession itself.
3. Reading T10Y2Y as a price. T10Y2Y is a yield spread, not a bond price. It cannot be “long” or “short” in the same way a bond can; trading exposures linked to the spread (steepener / flattener) are structured via derivatives or yield-curve futures, not by holding the spread itself.
4. Over-interpreting weakly positive spreads. A T10Y2Y reading of 0.10% (10 basis points) is closer to inversion than to a normal 1.5% spread. Many analyses bin T10Y2Y too coarsely (positive vs. negative) and miss the information in the near-zero regime, which has historically been the leading edge of every inversion cycle.

Historical Regimes

1978–1980 — Volcker inversion. T10Y2Y inverted deeply (below −2%) as the Federal Reserve raised the fed funds rate above 20% to break the inflation regime. The 1980 and 1981–1982 recessions followed within months. This remains the most extreme T10Y2Y inversion in the dataset and the clearest example of intentional monetary-policy-induced recession. The episode is documented in detail in the yield curve inversion history study.

1988–1990 — Late-cycle tightening. A brief T10Y2Y inversion in 1989 preceded the 1990–1991 recession by approximately 14 months. The inversion was shallow (around −0.2%) and short-lived, but the signal was confirmed by the subsequent downturn. This episode coincided with the savings-and-loan crisis and the Gulf War oil shock.

1998–2001 — Dot-com era. T10Y2Y inverted in 2000 as the Federal Reserve tightened into an overheating economy. The recession began in March 2001, roughly 12 months after the first inversion. The 9/11 attacks deepened the downturn but did not cause it — the yield curve had already signalled the underlying weakness, a pattern explored in the five recession rules tested study.

2005–2007 — Pre-GFC inversion. T10Y2Y turned negative in late 2005 and remained inverted through 2007. The Great Recession began in December 2007, consistent with the historical 12–24 month lead time. Notably, many commentators dismissed this inversion as a “Greenspan conundrum” driven by foreign central bank Treasury purchases rather than a genuine recession signal — a view subsequently invalidated by the 2008 downturn.

2019 — Brief pre-COVID inversion. T10Y2Y inverted briefly in August 2019. The subsequent COVID-19 recession (February–April 2020) followed the historical lead-time pattern, although the proximate cause was exogenous rather than the typical late-cycle dynamic. The NY Fed recession probability model registered the signal at the time.

2022–2024 — Deepest inversion since 1981. T10Y2Y reached −1.08% in July 2023, the deepest inversion in over 40 years. The inversion lasted approximately 26 months — the longest in the dataset — driven by an unprecedented combination of Federal Reserve tightening and persistent long-end demand. Whether the subsequent re-steepening will be followed by recession remains the central macroeconomic question as of early 2026. The 10-year yield trajectory and the Federal Funds Rate path will likely determine the outcome.

Related Macroeconomic Datasets

T10Y2Y is most informative when read alongside the rest of the US Treasury curve and the Federal Reserve’s policy stance — the alternative slope measures and the underlying maturities provide cross-checks on the recession signal.

Macroeconomic Dataset Hub

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

Sources

• Board of Governors of the Federal Reserve System — H.15 Selected Interest Rates
• Federal Reserve Bank of St. Louis — FRED series T10Y2Y
• US Department of the Treasury — Daily Treasury Yield Curve Rates

Dataset Reference