US Equity Risk Premium: S&P 500 Earnings Yield Minus 10-Year Treasury Yield Monthly Since 1871

The US Equity Risk Premium is an Eco3min monthly composite calculated as the S&P 500 trailing earnings yield minus the 10-year US Treasury yield. It measures, in percentage points, how much additional income equities deliver per dollar of price relative to risk-free bonds. Available since January 1871 using Robert Shiller’s long historical series, the US Equity Risk Premium turned negative in 2023 for the first time since the dot-com era, a development that has anchored an ongoing debate about whether equities are mispriced or whether the historical “premium” concept itself requires recalibration in a higher-rate environment.

Dataset: US Equity Risk Premium — Earnings Yield minus 10Y Treasury (1871–2026) · Updated 2026-05-01

Macro Takeaway

Construction & Components

The US Equity Risk Premium follows the simplest and most widely used formulation: the inverse of the S&P 500 trailing P/E (i.e., the trailing earnings yield) minus the long-term government bond yield. The Eco3min composite uses Robert Shiller’s harmonized historical series, which extends back to 1871 with consistent methodology.

Formula:

ERP = (E_ttm / P) − GS10
where  E_ttm = trailing 12-month S&P 500 earnings
       P     = S&P 500 price (month-end)
       GS10  = 10-year government bond yield

Components:

• S&P 500 trailing earnings yield — Robert Shiller’s ie_data.xls dataset (Yale) — monthly. Earnings are reconstructed from Cowles Commission data pre-1926 and Standard & Poor’s data thereafter. The yield is computed as trailing 12-month real earnings divided by the month-end real S&P 500 price.
• 10-year US Treasury yield (GS10) — Shiller’s ie_data.xls, column GS10 — monthly average yield on long-term US government bonds. For the pre-1953 period, Shiller harmonizes earlier long-bond series; from 1953 onward, the series matches FRED’s GS10.

Frequency reconciliation: Both components are monthly with end-of-month alignment, so no resampling or interpolation is applied. The composite is monthly throughout the full sample.

Coverage: January 1871–2026. The 1871 start date reflects the earliest reliable reconstruction of S&P-equivalent earnings in Shiller’s data. The pre-1926 segment is academically reconstructed and should be read with care; the post-1926 segment uses original Standard & Poor’s data.

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.

Direct CSV Access — Eco3min Structured Dataset

https://eco3min.fr/dataset/equity-risk-premium.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/equity-risk-premium.csv"
df = pd.read_csv(url, parse_dates=["date"])

print(df.head())
print(df.describe())

Using the Dataset in R

library(readr)

url <- "https://eco3min.fr/dataset/equity-risk-premium.csv"
df <- read_csv(url)

head(df)
summary(df)

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

Methodology

The US Equity Risk Premium is rebuilt monthly by an Eco3min pipeline that pulls Shiller’s ie_data.xls from Yale (Shiller’s public academic page), reads the relevant monthly columns (trailing earnings, S&P 500 price, GS10 long-bond yield), computes the trailing earnings yield as E/P, and subtracts the GS10 yield to obtain the premium in percentage points.

The pipeline runs after each update of Shiller’s spreadsheet, which Shiller refreshes approximately monthly. Because Shiller’s data uses month-end values for prices and contemporaneous monthly averages for yields, no further alignment is required. The historical series back to 1871 is rebuilt at each refresh to incorporate any retrospective revisions Shiller publishes (these are rare but occur when earnings data for recent quarters are restated by Standard & Poor’s).

Data Quality & Provider Notes

The US Equity Risk Premium inherits the characteristics of its source dataset, which is academic rather than commercial:

• Shiller cadence. Robert Shiller refreshes ie_data.xls approximately monthly. The composite latency is therefore on the order of 4–6 weeks after month-end, longer than typical FRED-sourced datasets. The exact lag depends on when Shiller posts each update.
• Earnings restatements. Trailing 12-month earnings can be revised retrospectively when Standard & Poor’s restates corporate earnings for prior quarters. Any such revision propagates to the historical ERP at the next Eco3min refresh.
• Pre-1926 caveat. The pre-1926 segment of Shiller’s data relies on academic reconstruction of earnings from Cowles Commission research. Readers should treat the 1871–1926 ERP values as indicative rather than directly comparable to modern accounting earnings. The post-1926 segment is more reliable.
• Multiple ERP formulations exist. The E/P−GS10 formulation used here is the simplest and most commonly cited (Damodaran, Wharton). The Excess CAPE Yield uses 1/CAPE minus a real (inflation-adjusted) yield. Both are valid; they answer slightly different questions. Forward-looking ERP estimates (e.g., Damodaran’s implied ERP from analyst forecasts) are not provided in this dataset.

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

The US Equity Risk Premium, in the E/P−GS10 formulation, is a backward-looking yield comparison, not a forward-looking expected return. Distinguishing what it measures from what investors often read into it is essential.

What it captures:

• The contemporaneous spread between current earnings yield on equities and the risk-free long-term bond yield
• A long historical reference (1871–2026) covering monetary regimes that no shorter dataset can span, from the classical gold standard through Bretton Woods, the Great Inflation, the Great Moderation, and the post-2020 era
• Episodes where the equity–bond yield spread compressed or inverted, providing context for the current reading

What it does NOT capture (common misinterpretations):

• Expected equity returns. Earnings yield is backward-looking. Forward expected equity returns depend on earnings growth, payout ratios, and re-rating, none of which appear in this dataset. The empirical relationship between current ERP and forward 10-year equity returns is positive but noisy, with R² typically in the 0.20–0.40 range depending on the sample and horizon.
• A timing signal. A negative or low ERP can persist for years before any equity drawdown materializes. The 1929 and 1999 episodes both featured prolonged periods of compressed ERP before the eventual correction.
• Real versus nominal distinction. The E/P−GS10 formulation mixes a real yield (earnings yield is roughly inflation-pass-through) with a nominal yield (GS10 includes inflation expectations). The Excess CAPE Yield addresses this by subtracting a real rate instead.
• Risk premium in the financial economics sense. The term “equity risk premium” in academic finance refers to expected excess returns over the risk-free rate, often estimated from forward-looking models (Damodaran’s implied ERP, dividend discount models, Fed model variants). The E/P−GS10 measure here is a proxy, not a structural estimate.

The composite is best read as a historical-context indicator that situates the current equity–bond yield spread against more than 150 years of comparable data, not as a tactical allocation tool.

Historical Regimes

The US Equity Risk Premium series traverses every modern US monetary regime. Reading the data in regime context shows that the simple E/P−GS10 measure has produced very different levels under very different inflation and interest rate environments.

• 1871–1913 — Classical gold standard. ERP fluctuated in a wide range, typically +2 to +5 percentage points, with significant cyclicality tied to railroad-era panics (1873, 1893, 1907).
• 1914–1929 — Inter-war boom. ERP compressed through the late 1920s as equity prices rose faster than earnings. The 1929 peak corresponds to one of the lowest pre-WWII readings.
• 1930–1949 — Depression and reconstruction. ERP widened sharply as equity prices collapsed faster than earnings, then normalized in the post-WWII recovery. Readings above +10 percentage points were common.
• 1950–1981 — Bretton Woods and Great Inflation. ERP exhibited high volatility. The 1970s stagflation produced a sustained period of double-digit ERP as bond yields and earnings yields both rose, with earnings yields rising faster.
• 1982–2000 — Disinflationary supercycle. ERP compressed steadily as equity multiples expanded and bond yields fell. The 1999–2000 dot-com peak produced one of the few sustained periods of negative or near-zero ERP in the sample.
• 2001–2019 — Post-bubble normalization and ZIRP. ERP widened to multi-decade highs during 2008–2012 as the 10-year yield collapsed below 2% while earnings yields stayed near 5–7%.
• 2020–2026 — Inflation surge and ERP compression. The 2022–2023 rise in 10-year yields toward 4–5% combined with elevated equity valuations pushed the US Equity Risk Premium negative for the first time since the dot-com era, generating extensive debate about whether the spread is mispriced or whether structural factors (passive flows, AI capex narratives) justify a lower historical mean.

For an alternative formulation that anchors the equity yield in smoothed real earnings rather than trailing nominal earnings, see the Excess CAPE Yield.

Related Macroeconomic Datasets

The US Equity Risk Premium is one of several measures comparing equity and bond valuations. The datasets below provide the components, alternative formulations, and broader market context.

Macroeconomic Dataset Hub

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

Explore the Eco3min Dataset Hub

Sources

• Robert Shiller, Yale University — ie_data.xls, monthly S&P 500 earnings, prices, and long-term bond yields, open academic data (1871–present)

Dataset Reference

Eco3min Research. (2026). US Equity Risk Premium — Earnings Yield minus 10Y Treasury (1871–2026).<br> https://eco3min.fr/en/us-equity-risk-premium-dataset/