Overview
This project surveys the expanding literature on long-run asset returns, focusing on historical return premia across major asset classes—stocks, bonds, and real assets—over the very long run. It examines the benefits and pitfalls of long-run data sets and sets out best practice for compiling and using such evidence. The research reports the magnitude of risk premia over the current and previous two centuries, compares estimates from alternative data compilers, and identifies promising directions for future research.

Project researchers
David Chambers
Cambridge Judge Business School
Elroy Dimson
Cambridge Judge Business School
Antti Ilmanen
AQR
Paul Rintamäki
Frankfurt School of Finance & Management
Research summary
This survey brings together the rapidly growing evidence on how major asset classes have performed over the very long run. The authors review historical return premia on equities, government and corporate bonds, real estate and commodities, drawing on data that in some cases extend back to the early nineteenth century and beyond.
A central contribution is to show how sensitive long-run estimates are to data construction. The article highlights four key challenges: “easy-data” and survivorship biases (focusing on markets or periods that look good), lack of macro-consistency in index design, difficulties in replicating historical portfolios, and the frequent omission of income (dividends, coupons, rents) from return measures. The authors set out best-practice guidelines for building and using long-horizon datasets.
Using improved data, they revisit several classic questions. They confirm that the equity risk premium was exceptionally high in the twentieth century but much smaller in the nineteenth, and that the recent era of very low real bond yields is historically unusual. Evidence on the credit premium from corporate bonds is positive but modest and harder to measure. For housing, newer work suggests total real returns are lower than once thought, while a diversified portfolio of commodity futures appears to earn a small but persistent premium over cash.
For students, the key message is that long-run return estimates are powerful but fragile: understanding how the data are built is as important as the headline numbers when drawing lessons for asset pricing and long-term investment.
Related paper
David Chambers, Elroy Dimson, Antti Ilmanen and Paul Rintamäki, ‘Long-Run Asset Returns’ (2024) 16 Annual Review of Financial Economics: https://www.annualreviews.org/content/journals/10.1146/annurev-financial-082123-105515


