What it means
Economic models connect interest rates with how people value consumption at different times and under uncertainty, and a safe payoff tomorrow has a price today. Someone expecting higher future consumption may require compensation for giving up consumption now.
The strength of that preference, expected growth and uncertainty can affect the predicted safe rate, so a model can produce a rate substantially above the historical observation it is meant to explain. The equity premium puzzle concerns the difference between risky equity returns and relatively safe returns, while the risk-free rate puzzle examines whether the same modelling framework can explain the safe return itself.
Fitting the premium while missing the safe-rate level leaves an important part of the evidence unexplained. Philippe Weil's 1989 NBER working paper studies preferences that separate risk aversion from intertemporal substitution, and its abstract says that simply separating those parameters does not resolve the equity premium puzzle under plausible calibration and can add a risk-free rate puzzle.
Risk aversion describes dislike of uncertain outcomes, while intertemporal substitution concerns willingness to move consumption between dates when the terms change. A model that changes one parameter to match a statistic should be checked against other implications.
It may explain a large equity premium while implying an implausibly high safe rate or other consumption behaviour. The word real matters, because a nominal interest rate includes a different purchasing-power comparison from a rate adjusted for inflation.
Mixing nominal bond yields with real model predictions can manufacture a discrepancy before the theory is tested. The safe asset also needs identification, since a short government bill, a long bond and an inflation-linked security have different exposures and measurement conventions.
Historical observation depends on the country, period and return calculation, and an old paper's dataset is not a statement about today's bond yield. Researchers can alter preferences, market assumptions or risks to investigate the mismatch, but such changes need their own empirical support.
For managers reading valuation commentary, keep the academic puzzle separate from a chosen discount rate. A current financing decision needs appropriate market evidence and project assumptions, and the existence of a model puzzle does not dictate one correct hurdle rate.
In practice
Real-world examples.
Example
A fictional model predicts a 4% safe real rate while its historical comparison dataset shows 1%. The 3-percentage-point gap raises a model-fit question. It does not by itself prove the observed securities were mispriced.
Example
An analyst raises assumed risk aversion to match a historical equity premium. The model now predicts an implausibly high safe rate. The team examines both statistics instead of declaring success from the premium alone.
Example
A report compares a nominal bond yield with a real-rate prediction. A reviewer first aligns inflation treatment and the asset definition. Only the remaining mismatch can inform a discussion of the puzzle.
Formula
Calculation
An illustrative model-fit gap is predicted safe real rate minus the comparable observed safe real rate. It is a diagnostic subtraction, not a full asset-pricing equation.
For fictional values, a model predicts a 4% safe real rate while the observed rate is 1%, so the gap is 4% - 1% = 3 percentage points. If equity returns are 7% in the same simplified illustration, the observed equity premium is 7% - 1% = 6 percentage points.
The 3-point level gap and 6-point return premium are different quantities. Neither calculation explains its own size or proves a model's assumptions.
A full analysis must specify preferences, consumption dynamics, information, asset payoffs and market conditions. Do not use the simple subtraction as a trading signal.Case study
Seen in the real world.
Fictional case study: Alder Valuation selects a model because it reproduces the equity premium in a historical dataset. Its research lead asks whether it also reproduces the safe real rate. The safe-rate prediction is much too high.
The team reviews the preference assumptions, inflation treatment and benchmark security before testing alternative specifications. The final report distinguishes a partial fit from an explanation of both statistics. Its corporate clients receive separate sensitivity analysis for their discount-rate choices rather than a claim that the academic puzzle guarantees future returns.
Watch out
Common mistakes.
- Defining the puzzle only as stocks outperforming bonds. The central issue is the safe-rate level implied by a model.
- Mixing nominal and real observations or different safe assets. Align the evidence before evaluating a discrepancy.
- Treating a historical model mismatch as a current arbitrage. The puzzle does not establish executable mispricing.
Questions
People also ask.
Is it identical to the equity premium puzzle?
No. They are related model-fit problems, but one concerns a return difference and the other the safe-rate level.
Does it say government bonds have no risk?
No. The safe-asset assumption and observed benchmark must be specified. Actual securities can carry several risks.
Does solving it set a company's discount rate?
No. Practical valuation still requires suitable current evidence, project risk and explicit assumptions. A research model is not a universal rate recommendation.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%