Real Interest Rates: 10-Year US Treasury Yield Minus CPI Inflation Annual Since 1953

Real Interest Rates — the 10-year US Treasury yield minus annual CPI inflation — is one of the most consequential variables in macroeconomics. It sets the true cost of long-term borrowing, the real return on savings, and the discount rate applied to all long-duration assets. This dataset is an Eco3min annual composite that pairs the 10-year nominal Treasury yield (FRED series GS10) with year-over-year CPI inflation (FRED series CPIAUCSL) to deliver a clean, four-column file of Real Interest Rates from 1953 to 2026 — structured for monetary policy research, quantitative analysis, and academic use.

Based on FRED series: GS10 (10Y Treasury) and CPIAUCSL (CPI-U) · Updated March 2026

Source: FRED series GS10 · Board of Governors of the Federal Reserve System

Macro Takeaway

Construction & Components

The Real Interest Rates composite is the annual ex-post real long-term rate — the average yield on 10-year US Treasury notes during a calendar year minus the realized year-over-year CPI inflation rate for the same year. This follows the approximate Fisher equation and matches the standard academic convention used in long-horizon real-rate studies.

Formula:

Real Interest Rate = GS10 − CPI YoY

where:
  GS10    = 10-Year US Treasury Constant Maturity Rate, annual average, %
  CPI YoY = (CPIAUCSL_dec / CPIAUCSL_dec-1y) − 1, annual average, %

Components:

• 10-Year Treasury Constant Maturity Rate (GS10) — FRED series GS10 — monthly observations published by the Board of Governors of the Federal Reserve System (H.15 release). The annual value is the simple average of the twelve monthly observations.
• CPI for All Urban Consumers (CPIAUCSL) — FRED series CPIAUCSL — monthly seasonally adjusted index from the US Bureau of Labor Statistics. The annual inflation rate is the December-over-December percentage change, the most common convention for annual real-rate series.

Frequency reconciliation: Both inputs are native monthly series. The composite is annual: GS10 is averaged over the calendar year, and CPIAUCSL is differenced December-on-December. No interpolation is required.

Coverage: 1953 to present. The series starts in 1953 because the GS10 constant-maturity yield series begins in April 1953; CPIAUCSL extends much further back, but the binding constraint is the Treasury yield data.

Dataset Overview

Dataset Variables

The CSV and Excel files contain the following four columns.
Each row represents one calendar year.

Column names match the CSV headers exactly.
The real rate is calculated as yield_10y − cpi_inflation (approximate Fisher equation).
Negative values indicate periods where inflation exceeded the nominal 10-year yield.

Latest Real Interest Rate

Latest available annual average. 2026 value will be updated when full-year data is published.

Real Interest Rates by Year (1953–2025)

The table below presents the complete annual dataset.
Each year shows the 10-year Treasury yield, CPI inflation rate,
and the calculated Real Interest Rates series.

Sources: Federal Reserve Bank of St. Louis (FRED series GS10), U.S. Bureau of Labor Statistics (CPI-U).
Real rate = 10Y yield − CPI inflation (approximate Fisher equation).
Negative real rates indicate periods where inflation exceeded the nominal yield.

Download the Complete Dataset

The full Real Interest Rates dataset is available in CSV and Excel formats.
Both files contain identical data: year, 10-year Treasury yield, CPI inflation, and real interest rate.

FRED Direct CSV Access

This composite is built entirely from publicly available FRED series. Both components can be downloaded directly from FRED for users who prefer to compute Real Interest Rates themselves:

Source FRED series for this composite:

GS10      — https://fred.stlouisfed.org/graph/fredgraph.csv?id=GS10
CPIAUCSL  — https://fred.stlouisfed.org/graph/fredgraph.csv?id=CPIAUCSL

FRED does not publish a native annual real 10-year Treasury yield series. The raw GS10 file provides only the nominal yield, and the inflation rate must be derived from CPIAUCSL. The Eco3min composite combines both inputs into a clean four-column file (year, yield, inflation, real rate), structured for immediate use in research or quantitative models.

Direct CSV Access — Eco3min Structured Dataset

https://eco3min.fr/dataset/gs10-us-10-year-treasury-yield-history-dataset-1953-2026.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/wp-content/uploads/2026/03/GS10.csv"
df = pd.read_csv(url)

# Display first rows
print(df.head())

# Average real rate over the full period
avg_real = df["real_rate"].mean()
print(f"Average real interest rate (1953–2025): {avg_real:.2f}%")

# Negative real rate years
negative = df[df["real_rate"] < 0]
print(f"Negative real rate years: {negative['year'].tolist()}")

Using the Dataset in R

library(readr)

url <- "https://eco3min.fr/wp-content/uploads/2026/03/GS10.csv"
df <- read_csv(url)

head(df)
summary(df$real_rate)

# Negative real rate years
negative <- df[df$real_rate < 0, ]
print(negative)

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

Methodology

The Real Interest Rates dataset is recomputed annually by an Eco3min pipeline that pulls GS10 and CPIAUCSL from the FRED API after each full calendar year of CPI data is available (typically mid-January of year Y+1, once December CPI is released). The pipeline averages the twelve monthly GS10 observations to produce the annual nominal yield, computes the December-over-December percentage change of CPIAUCSL for the annual inflation rate, and writes the four-column file (year, yield_10y, cpi_inflation, real_rate) to disk.

The series is ex-post real: it uses realized year-over-year CPI inflation, not market-implied breakeven inflation from TIPS. An ex-ante version of the long real rate — available from the TIPS market since 2003 (DFII10 on FRED) — offers a different angle on real rates but does not extend the historical coverage back to 1953.

Data Quality & Provider Notes

Latency for the annual Real Interest Rates composite is dictated by the slowest input: a given year’s value is finalized only once that December’s CPIAUCSL print is released by the BLS, typically in mid-January of the following year. GS10 monthly observations are available much earlier, but the annual average obviously requires the full twelve months.

Revisions to either input propagate to the composite. GS10 monthly averages rarely revise; CPIAUCSL undergoes seasonal-adjustment recalibrations once a year (typically in February), which can shift the most recent annual reading slightly. The Eco3min pipeline overwrites historical values on each run to reflect the latest FRED vintage.

No standardized alternative source exists for a long-run annual real 10-year rate. Bloomberg, Refinitiv, and academic datasets (Damodaran, Shiller) publish nominal yields and inflation separately, but the composite itself depends on convention choices (calendar-average vs. period-end yield, headline vs. core CPI, December-on-December vs. annual-average inflation). The Eco3min specification — annual-average nominal yield minus December-on-December headline CPI — matches the convention most widely used in long-horizon real-rate research.

What This Index Captures (And What It Doesn’t)

The Real Interest Rates composite is the most commonly cited long-run measure of the inflation-adjusted long-term US rate, but its construction imposes specific boundaries on what it can be used to measure.

What it captures:

• The ex-post real cost of long-duration US dollar borrowing for a default-free benchmark issuer
• The realized real return for a holder of constant-maturity Treasury duration over the calendar year
• The long-run discount rate applied to fixed nominal cash flows once realized inflation is netted out

What it does NOT capture (common misinterpretations):

• The market’s expected real rate. This composite uses realized CPI inflation. The ex-ante real rate, derived from TIPS breakevens (DFII10), can diverge meaningfully — especially when inflation surprises occur. The 2021–2022 episode is a textbook case: realized ex-post real rates collapsed deeply negative while ex-ante TIPS real rates were near zero.
• The full real rate curve. This is a single point at the 10-year tenor. The real Fed Funds rate, the real 2-year yield, and the real 30-year yield each behave differently across regimes; the slope between them carries information that any single tenor cannot.
• Credit-risk adjusted real returns. The composite uses default-free Treasury yields. Real corporate borrowing costs are higher by the credit spread (which itself varies with the cycle), and the relevant real rate for a leveraged borrower can be substantially different.
• Duration risk. A negative real interest rate does not mean a 10-year Treasury holder lost money in real terms — bond prices also move with the yield. The annual real total return on the 10Y Treasury can diverge sharply from the annual real yield calculation shown here.

Used responsibly, the series is a clean descriptive measure of the long-run real rate regime; analytical interpretations of its level, trend, and relationship to equity valuations are explored in the companion US Real Interest Rates History study.

Historical Regimes

1953–1965 — Stable positive real rates under Bretton Woods.
Real Interest Rates averaged approximately +2.5% during this period, anchored by low and stable inflation within the Bretton Woods fixed exchange rate system. The 10-year yield hovered between 3% and 4%, while inflation rarely exceeded 2%. This was the baseline real rate environment that would not be seen again for decades.

1966–1980 — The Great Inflation and negative real rates.
As inflation accelerated through the late 1960s and 1970s, nominal yields failed to keep pace. The 10-year real rate turned deeply negative during the oil shocks — reaching −3.4% in 1974 and −2.0% in 1980. The Federal Reserve under Arthur Burns and G. William Miller was consistently behind the curve, allowing inflation expectations to become unanchored.

1981–2000 — Volcker disinflation and historically high real rates.
The Volcker shock produced the highest Real Interest Rates in modern US history. As inflation fell from 13.5% in 1980 to 3.2% in 1983, nominal yields declined more slowly — pushing the real rate above 8% in 1984. Real rates remained elevated throughout the 1980s and 1990s, averaging approximately +4%, as markets demanded a persistent inflation risk premium even as actual inflation remained contained.

2001–2019 — Secular decline toward zero.
Real Interest Rates fell steadily over two decades, driven by the global savings glut, aging demographics, declining potential growth, and successive rounds of quantitative easing. By the late 2010s, the 10-year real rate hovered near zero — a historically unprecedented configuration that compressed risk premia across asset classes.

2020–2026 — Pandemic shock, negative real rates, and normalization.
The pandemic triggered the most deeply negative real rates since the 1970s: −3.25% in 2021 and −5.05% in 2022, as CPI inflation surged while the Fed initially kept rates near zero. The subsequent tightening cycle restored positive real rates by 2024 (+1.31%), and 2025 closed at +1.69%. The structural question of where the “neutral” real rate settles — the r* debate — is examined in the companion US Real Interest Rates History study and in the Real Interest Rates vs CAPE Ratio analysis.

Related Macroeconomic Datasets

The Real Interest Rates series at the 10-year tenor is the most-cited long-run real rate measure, but it sits within a broader rate ecosystem. Combining it with shorter-tenor real rates and the underlying nominal and inflation series maps the full real rate curve.

Related Research

The interpretation of long-run Real Interest Rates — their relationship with equity valuations, currency dynamics, and inflation regimes — is examined in the following Eco3min studies.

Macroeconomic Dataset Hub

This dataset is part of the Eco3min macro-financial data repository.
Explore all available datasets including inflation, interest rates, equity returns,
and currency indicators.

Explore the Eco3min Dataset Hub

Sources

• Board of Governors of the Federal Reserve System — 10-Year Treasury Constant Maturity Rate (FRED series GS10)
• U.S. Bureau of Labor Statistics — Consumer Price Index for All Urban Consumers, seasonally adjusted (FRED series CPIAUCSL)

Dataset Reference

Eco3min Research. (2026). Real Interest Rates — 10-Year US Treasury Yield Minus CPI Inflation, Annual (1953–2026).

Real 10-Year Treasury Yield — Daily CSV Download (Inflation-Adjusted Rates)