The equity premium (or equity risk premium) is the additional return that stocks are expected to deliver above the risk-free rate. It is the compensation investors demand for bearing the systematic risk of equity ownership, including the possibility of permanent capital loss, deep drawdowns, and correlated losses during recessions.
The equity premium connects directly to the yield curve through several channels:
The equity premium is the single most important parameter in long-run asset allocation. Small changes in the assumed premium compound into large differences in recommended equity weights over a multi-decade horizon.
Two methods dominate in practice, and they often produce different answers.
Historical realized premium. The Dimson-Marsh-Staunton dataset, which spans U.S. equity and bond returns from 1900 to the present, puts the geometric equity premium at approximately 4.4% per year over long-term government bonds and 5.0% over Treasury bills. The arithmetic mean is higher, roughly 6.0%, because volatility inflates the arithmetic average relative to the compound return. For asset allocation work, the geometric figure is more relevant as a long-horizon expected return, while the arithmetic figure is appropriate for single-period mean-variance optimization.
Forward-looking dividend discount approach. This method estimates the implied equity return directly from current prices. The Gordon Growth model gives:
E(r_equity) = D/P + g
where D/P is the current dividend yield and g is expected long-run earnings growth. With the S&P 500 dividend yield near 1.3% and a consensus nominal earnings growth estimate of 4.0% to 5.0%, the implied equity return is approximately 5.3% to 6.3%. Subtracting the current 10-year Treasury yield gives an implied equity premium of roughly 3.0% to 4.5%, depending on yield level and growth assumption. This approach sidesteps the hindsight problem of historical data but requires explicit assumptions about sustainable growth rates and payout ratios.
The historical premium is highly sensitive to the measurement period. From 1966 to 1982, the realized equity premium over bonds was negative: high and rising inflation destroyed real bond returns while simultaneously compressing equity multiples. From 1982 to 2000, the realized premium was unusually high as P/E multiples expanded from roughly 8x to over 30x, a one-time valuation re-rating that cannot repeat by definition.
These distortions mean that a simple trailing average carries substantial period-selection bias. A portfolio manager who calibrated return expectations in 1999 using the prior 18-year history would have embedded an equity premium estimate well above any plausible forward-looking figure.
Forward-looking models solve the hindsight problem but introduce different assumptions. The dividend discount approach requires a stable long-run payout ratio, a constant growth rate, and an accurate current yield, all of which fluctuate. Small changes in the assumed growth rate g produce large swings in the implied premium. Shifting g by 100 basis points changes the estimated premium by 100 basis points, one-for-one.
Real yields are the most direct link between the yield curve and equity valuations. When the 10-year real yield rises sharply, all risky asset discount rates rise. Unless earnings expectations rise in proportion, equity prices must fall to restore an equilibrium premium.
The 2022 episode illustrated this cleanly. The 10-year TIPS yield rose from approximately -1.0% in January 2022 to approximately +1.7% by October 2022, a swing of roughly 270 basis points. That repricing drove losses in both investment-grade bonds and equities simultaneously, the exact outcome you would expect when the real rate component of the discount rate increases across all asset classes at once.
The real yield page on YCP tracks the TIPS curve in real time. The term premium page shows the ACM decomposition, which separates the nominal yield into the expected short rate path and the term premium. Together, those two pages give you the building blocks needed to assess whether a given equity premium estimate is rich or cheap relative to current fixed-income pricing.