Expected Return and What a Weighted Average Tells You
Calculate expected return from signed outcomes and estimated probabilities, then see why changing the odds can change your forecast before you invest.
In this chapter
Intermediate
Suppose a fund could return +24%, +6% or −16% over the next year. If those three results have different chances of happening, a simple average treats them too much alike. Expected return gives each outcome a weight equal to its estimated probability. It answers what the return averages to under the assumptions you entered.
Give each possible return its own probability
Use a single measurement period and a common basis across outcomes. If the investment makes a distribution during that period, decide how you will account for it. A total return includes price changes and investment income. In the examples here, the scenarios are for one-year nominal USD total returns before fees and taxes. Each return covers the whole year. The fund outcomes and their probabilities are hypothetical.
\[ \mathbb{E}[R]=\sum_{i=1}^{n}p_iR_i \]
\(R\) is the investment’s uncertain one-period return. \(\mathbb{E}[R]\) is its expected return, \(R_i\) is the signed decimal return in scenario \(i\), \(p_i\) is that scenario’s probability in decimal form, and \(n\) is the number of scenarios. A 25% chance is 0.25. A 16% loss is a return of −0.16. The sum adds the contributions of all \(n\) scenarios. Each probability is between zero and one, and the probabilities sum to one when the scenarios are mutually exclusive and cover the possible outcomes in the model.
Worked example. Three ways the same year could end
Imagine investing $2,000.00. The model has three outcomes.
| Scenario | Estimated probability | One-year total return | Contribution to expected return |
|---|---|---|---|
| Strong year | 25% | +24% | +6.00 percentage points |
| Ordinary year | 50% | +6% | +3.00 percentage points |
| Weak year | 25% | −16% | −4.00 percentage points |
| Expected return | 100% | +5.00% |
The contribution of the weak year is \(0.25\times(-0.16)=-0.04\). Add all three signed contributions to get \(0.06+0.03-0.04=0.05\), or 5.00%. The simple average gives about 4.67% and would fit a model where each outcome had a one-third chance.
Which balance could you actually end up with?
At 5.00%, the probability-weighted ending balance is $2,100.00. The three modelled balances are $2,480.00, $2,120.00 and $1,680.00. The $2,100.00 average lies between them, but no state in this model pays exactly that amount. Once the year ends, the return you actually earned is your realised return. In the weak state, it is −16%, even though the model’s expected return was positive. A benchmark gives you another way to judge that realised result.
A useful error check comes from the weights. Since they are nonnegative and sum to one, an expected return for this model has to lie between its lowest and highest scenario returns. If the answer falls outside, check the signs and the probabilities. For a portfolio of several holdings, combine their expected returns using each holding’s share of the starting portfolio.
Where the probabilities come from
Start with comparable evidence or explain how a forecast was built. Before interpreting the average, judge how much confidence to place in those probabilities. Make the states distinct. “Sales fall” and “the business loses money” could happen in the same year. If you use them as separate rows without accounting for their overlap, you can count the same year twice. For a scenario set such as strong, ordinary and weak, define which events qualify for each row and account for the remaining cases.
Worked example. The return scenarios stay fixed but the odds change
Suppose new orders are weaker than the original forecast. For this check, keep all three possible returns fixed and update the chances to 15% strong, 40% ordinary and 45% weak. The chances still sum to 100%.
\[ 0.15(0.24)+0.40(0.06)+0.45(-0.16) =0.036+0.024-0.072=-0.012 \]
The estimate moves from +5.00% to −1.20%, a drop of 6.20 percentage points. The weak outcome’s return is unchanged. It contributes −7.20 percentage points now because it gets a bigger weight. The updated probability-weighted balance is $1,976.00, and the three actual balances the model allows are unchanged. If weaker orders also change what a weak year costs, rerun the model with that return updated too. This check isolates the effect of the odds so you can see which assumption drove the change.
Practice questions
Choose one answer per question. Use the stated assumptions and round final dollar amounts to cents.
1. Weight signed outcomes
A one-year investment has a 60% chance of +15% and a 40% chance of −10%. What is its expected return?
2. Check the model before using it
A model claims to cover all outcomes, with 40%, 35% and 35% probabilities in three mutually exclusive states. What should you check first?
3. Find an expected balance
A $400.00 stake has a 75% chance of a +20% return and a 25% chance of a −20% return. There are no outside flows. What is the expected ending balance?
4. Move probability toward the loss
A position pays +20% in the gain state and −10% in the loss state. The loss probability rises from 25% to 40%. How does the expected return change?
5. Interpret an average of zero
An investment can return +30% or −10%. Its model assigns 25% to the gain. Which statement is true?
6. Check the range
A model has valid probabilities and all possible returns lie from −8% to +12%. Which reported average needs correction on that evidence alone?