Expected Return: The Average Outcome You Will Almost Never Actually Get
Investors routinely treat "expected return" as a number the market owes them each year, then feel cheated or vindicated depending on whether that year came in above or below the figure. In reality expected return describes the center of a wide distribution of possible outcomes, and in any given year the actual result almost always lands somewhere else on that distribution.
The core principle
Expected return is the probability-weighted average of every possible outcome an investment could produce, where each possible return is multiplied by the probability of that outcome occurring, and the resulting products are summed. In formula terms, expected return = sum of (probability of each outcome x that outcome's return), across the full range of scenarios being considered. It is a statistical concept borrowed directly from probability theory, and like any average, it can be a number that no single observation actually equals.
The critical distinction is between expected return and realized return. Realized return is what an investment actually did over a specific, completed period, a single fixed number with no uncertainty left in it. Expected return is a forward-looking estimate built from assumptions: historical patterns, current valuations, interest rate expectations, and earnings growth forecasts. Because it depends on assumptions, expected return is not a single objective fact the way a past year's realized return is; different analysts feeding different assumptions into the same framework can reasonably produce expected return estimates that differ by several percentage points for the identical asset class.
Underlying every expected return figure is a full distribution of possible outcomes, not just the average itself. Two investments can share an identical expected return of 8% while one has outcomes clustered tightly between 6% and 10% and the other ranges from losing 30% to gaining 50% depending on the year. The expected return number alone says nothing about that spread, which is precisely why it must always be considered alongside a measure of volatility or standard deviation, never in isolation.
How the math works
Example 1: a simple two-outcome scenario. Suppose an analyst estimates a stock has a 60% chance of returning 20% over the next year and a 40% chance of returning negative 10%. The expected return is (0.60 x 20%) + (0.40 x −10%) = 12% − 4% = 8%. Note that 8% is not one of the two actual possible outcomes; the stock will return either 20% or negative 10% in this simplified model, never exactly 8%. The 8% figure only becomes meaningful as an average across many repetitions of a similar bet, or as a planning input for a long time horizon, not as a prediction of what this specific year will look like.
Example 2: a more realistic multi-scenario estimate. A financial planner building a 30-year retirement projection for a diversified portfolio might use four scenarios: a 15% chance of a strong decade averaging 12% annually, a 45% chance of a typical decade averaging 8%, a 30% chance of a weak decade averaging 4%, and a 10% chance of a recessionary decade averaging negative 2%. The expected return is (0.15 x 12%) + (0.45 x 8%) + (0.30 x 4%) + (0.10 x −2%) = 1.8% + 3.6% + 1.2% − 0.2% = 6.4%. This 6.4% figure is a defensible planning assumption for a multi-decade projection, useful for estimating how much to save each month, but it says nothing about whether the next twelve months specifically will look anything like it. A retiree who built a spending plan assuming a steady 6.4% every year, rather than a 6.4% average across a lumpy sequence of good and bad years, is planning around a fiction the math itself never promised.
How it shows up in real portfolios
The S&P 500's long-run average annual return is frequently cited near 10% before inflation, a figure built by averaging roughly a century of calendar years. Yet in any individual calendar year over that history, the index has returned anywhere from a decline of more than 35% to a gain of more than 30%. An investor who treats "the market returns about 10% a year" as a guarantee for the current year, rather than as a long-run average built from wildly different individual years, will be confused and possibly panicked the first time an actual year deviates sharply from that number, which happens routinely.
A concrete scenario involving a high-earning professional: an executive maxing out a 401(k) and a backdoor Roth IRA each year builds a retirement projection assuming a steady 7% real return, compounding smoothly every single year until retirement. When a market downturn produces a negative 18% year midway through their accumulation phase, the executive concludes the plan has failed and considers shifting entirely to cash. In fact, a single sharply negative year is fully consistent with, and even expected within, a long-run 7% average; what matters for the plan's actual success is the average and the sequence across the full multi-decade horizon, not any single year's deviation from the average. The mistake here is not the market's behavior, it is treating an average built for planning as though it were a promise for every twelve-month period along the way.
Expected return estimates also shift meaningfully with starting valuations, a relationship well documented in the empirical literature on stock market returns. When broad market valuations are historically elevated, forward-looking expected returns over the following decade tend to run lower than the long-run historical average would suggest, and the reverse holds when valuations start out depressed. An investor building a retirement plan today using yesterday's long-run historical average, without adjusting for current starting conditions, risks anchoring to a number that no longer reflects the most defensible forward-looking estimate.
A related scenario involves a young professional automatically enrolled in a target-date retirement fund who never actually looks at what expected return assumption the fund's glide path is built around. Two target-date funds aimed at the same retirement year can hold meaningfully different mixes of stocks and bonds, and therefore carry different embedded expected returns and different amounts of risk, even though both are marketed with an identical year in the fund's name. Comparing the underlying asset allocation, not just the label, is the only way to know whether the fund's implicit expected return assumption is a reasonable match for that investor's actual goals and time horizon, rather than simply accepting a default that happened to be pre-selected by an employer's plan administrator.
Actionable breakdown
- Use expected return for:
- Setting long-term savings and contribution targets.
- Comparing asset classes over multi-decade horizons.
- Sizing a diversified asset allocation.
- Do not use expected return for:
- Predicting next year's specific performance.
- Timing entry or exit points in the market.
- Justifying a concentrated, undiversified bet.
- Before trusting any expected return figure, ask:
- What time horizon does this number assume?
- What historical period, or which scenarios, built it?
- Is it net of fees and inflation?
- Does it account for today's starting valuations?
- Always pair the number with a measure of volatility around it.
Common pitfalls
- Recency bias: extrapolating the last few strong years forward into the expected return, when a hot recent stretch does not raise the true long-run expected return; it more often means those years borrowed from future gains.
- Ignoring the variance around the average entirely, treating a headline expected return figure as though it came with no uncertainty attached, when the spread of possible outcomes is often the more decision-relevant number.
- Confusing nominal and real figures: a 7% expected return before inflation is meaningfully lower after subtracting even a modest inflation rate, and that gap compounds significantly over a retirement-length horizon.
- Treating expected return as a promise the market must deliver on schedule, rather than a probability-weighted average that individual years will routinely miss in both directions.
Related concepts
For the spread of outcomes around this average, see standard deviation and risk tolerance. For how expected return is used to build a portfolio, see the guide on asset allocation. For how the order of returns interacts with a spending plan, see sequence of returns risk and the guide on withdrawal strategies.
The bottom line
Expected return is a useful compass for decades-long planning but a poor guide for any single year, so build a plan around a range of outcomes rather than around one number.