Glossary/Performance

Expected Return

Also known as Mean return, Forecast return, E(R)

Expected return is a probability-weighted estimate of an investment’s future return over a specified horizon. It can come from scenarios, valuation models, market-implied information, historical evidence, or capital-market assumptions and is an estimate rather than a promised outcome.

Editorially reviewed 2026-07-29

Why expected return matters

Portfolio decisions require a forward-looking view of compensation for risk. Expected returns help compare opportunities, set strategic allocations, test whether objectives are feasible, and value assets. Even when highly uncertain, stating the assumption exposes the trade-off behind a recommendation. The estimate should be considered with risk, correlation, liquidity, fees, taxes, and estimation error, not used as a ranking that automatically sends capital to the highest number.

How it is applied

A discrete scenario model multiplies each possible return by its probability and sums the results. Asset-class forecasts often combine income yield, growth, valuation change, and currency effects. Equilibrium or factor models provide other starting points. Analysts should define whether returns are arithmetic or geometric, nominal or real, gross or net, and in which currency. Sensitivity ranges and alternative regimes are generally more decision-useful than a single precise forecast.

Formula

E(R) = Σ pi × Ri
E(R)
Expected return over the stated horizon
pi
Probability assigned to scenario i
Ri
Return if scenario i occurs

Portfolio example

An investment has a 50% chance of returning 12%, a 30% chance of returning 4%, and a 20% chance of losing 15%. Its probability-weighted expected return is 4.2%. No investor will necessarily earn 4.2%; each outcome is different. If the probabilities or downside estimate are overly optimistic, the calculation gives a falsely attractive answer. Liquidity and fees may reduce the expected return actually available.

How to interpret it

Expected return is the mean across modeled possibilities, not the most likely result and not a guarantee. Two investments with the same expectation can have very different distributions and loss potential. Arithmetic expected return is useful for one-period modeling, while long-term compound growth is lower when returns vary. Confidence, data quality, valuation, and sensitivity to assumptions should be visible whenever the estimate influences portfolio weights.

Limitations and common misconceptions

Future probabilities are unknown, historical averages are noisy, and valuation relationships can take years to normalize. Forecasts are often correlated and may fail together during regime changes. Model precision can encourage extreme optimization. Expected returns may omit taxes, fees, trading costs, capacity, or illiquidity. Scenario ranges, reverse stress tests, robust constraints, market-implied information, and governance are needed to keep uncertain forecasts from dominating decisions.

Sources and further reading

← All glossary terms