Common Errors in Return and Risk Calculations

Why Return and Risk Calculations Go Wrong

Return and risk calculations often involve familiar formulas, but a small mismatch in what is counted or how numbers are expressed can change the result substantially. A return based only on a price change omits cash distributions. Returns from successive periods do not generally add. Scenario outcomes must reflect their probabilities, and percentage rates must be handled consistently when calculating variance.

A useful habit is to identify the quantity being calculated before substituting numbers: is it a single-period total return, a compounded multi-period return, an expected return across scenarios, a variance, or a standard deviation? The calculation method and units depend on that choice. The procedures below are aligned with the SOA VEE Accounting and Finance syllabus.

Definition: For an investment with initial value $P_0$, ending value $P_1$, and cash distribution $D$ received during the holding period, the single-period total return is $R=(P_1-P_0+D)/P_0$. The price change and cash distribution are both part of the investment's return. For a discrete set of scenarios with returns $R_j$ and probabilities $p_j$, where $p_j\geq 0$ and $\sum_j p_j=1$, the expected return is $\mu=E[R]=\sum_jp_jR_j$, and the probability-weighted variance is $\operatorname{Var}(R)=\sum_jp_j(R_j-\mu)^2$. The standard deviation is the nonnegative square root of the variance.

Start With the Return Being Measured

A holding-period total return measures the change in investment value plus distributions received, relative to the initial investment. The denominator is the value at the start of the period, not the ending value. If a distribution is omitted, the calculation reports only a price return, not the total return.

For sequential periods, the return in each period applies to the value carried into that period. If distributions are reinvested at the end of each period, the wealth growth factors multiply. This is why adding periodic returns is generally incorrect.

1+R_{\mathrm{cumulative}}=\prod_{t=1}^{n}(1+R_t), \qquad R_{\mathrm{cumulative}}=\prod_{t=1}^{n}(1+R_t)-1.

The relation follows directly from the definition of return. If wealth at the start of period $t$ is $W_{t-1}$, then after that period's return it is $W_t=W_{t-1}(1+R_t)$. Applying this relation period by period gives $W_n=W_0\prod_{t=1}^{n}(1+R_t)$. Dividing by $W_0$ and subtracting 1 gives the cumulative return. Addition leaves out the cross-period effects represented by the product.

Worked Example: Including Distributions and Compounding Returns

A hypothetical fund unit is worth $50.00 at the beginning of Year 1, $48.00 at the end of Year 1, and pays a $3.00 distribution during that year. In Year 2, its price rises from $48.00 to $50.40 and it pays a $0.96 distribution. Calculate each year's total return and the two-year cumulative return, assuming each distribution is reinvested at the end of its year. Compare the result with adding the price-only returns. The method is to calculate total return period by period, then multiply the growth factors; this avoids omitting distributions or adding compounded returns.

Year 1 total return. The price change is $48.00-50.00=-2.00$. Including the distribution, the dollar gain is $-2.00+3.00=1.00$. Therefore,

R_1=\frac{48.00-50.00+3.00}{50.00} =\frac{1.00}{50.00}=0.02=2.00\%.

Year 2 total return. The price change is $50.40-48.00=2.40$. Adding the $0.96 distribution gives a total gain of $3.36$, so

R_2=\frac{50.40-48.00+0.96}{48.00} =\frac{3.36}{48.00}=0.07=7.00\%.

Cumulative total return. The two growth factors are $1.02$ and $1.07$:

R_{\mathrm{cumulative}}=(1.02)(1.07)-1 =1.0914-1=0.0914=9.14\%.

Check against price-only returns. The Year 1 price-only return is $(48.00-50.00)/50.00=-4.00\%$. The Year 2 price-only return is $(50.40-48.00)/48.00=5.00\%$. Adding them gives $1.00\%$, which is neither the two-year cumulative price return nor the total return. Even compounding the price-only returns gives $(0.96)(1.05)-1=0.008=0.80\%$, still excluding both distributions. The total-return result of $9.14\%$ includes them and compounds across periods. The arithmetic is exact to two decimal places for reported rates.

Weight Scenarios Before Measuring Risk

A scenario distribution describes possible returns and the probability of each return. Its expected return is a probability-weighted mean, not necessarily the ordinary average of the listed returns. The probabilities must be nonnegative and sum to 1. If the scenarios are mutually exclusive and cover all outcomes, the weighted calculation uses all their probabilities.

Risk measured by variance is also probability-weighted: each squared deviation from the expected return contributes in proportion to the probability of that return. An unweighted average of deviations or squared deviations is appropriate only if all listed scenarios have equal probability.

Worked Example: Detecting an Unweighted Scenario Average

A hypothetical investment has three possible one-year returns: $-10\%$ with probability $0.25$, $8\%$ with probability $0.50$, and $20\%$ with probability $0.25$. Calculate expected return, variance, and standard deviation. Use decimal returns for variance. The method is to weight both the mean and squared deviations by scenario probabilities; this avoids treating unequal-probability outcomes as equally likely.

Check the probabilities. Their total is $0.25+0.50+0.25=1.00$.

Expected return. Write the percentage returns as decimals: $-0.10$, $0.08$, and $0.20$. Then

\mu=(0.25)(-0.10)+(0.50)(0.08)+(0.25)(0.20) =-0.025+0.040+0.050=0.065=6.50\%.

The unweighted average would be $(-10\%+8\%+20\%)/3=6.00\%$. That is wrong because the middle outcome has twice the probability of either other outcome.

Variance. The deviations from the expected return are $-0.10-0.065=-0.165$, $0.08-0.065=0.015$, and $0.20-0.065=0.135$. Their squares are $0.027225$, $0.000225$, and $0.018225$. Weight each square by its probability:

\operatorname{Var}(R) =(0.25)(0.027225)+(0.50)(0.000225)+(0.25)(0.018225) =0.00680625+0.00011250+0.00455625 =0.011475.

Standard deviation. Taking the nonnegative square root gives $\sqrt{0.011475}=0.1071$, or $10.71\%$. As a check, the probability-weighted deviations are $(0.25)(-0.165)+(0.50)(0.015)+(0.25)(0.135)=0$, as they should be when deviations are measured from the weighted mean. The variance is $0.011475$ in decimal-return-squared units; the standard deviation is in return units.

Keep Percentage and Decimal Units Consistent

A return of $8\%$ is $0.08$ in decimal form. Either representation can be used, but a calculation must not mix them. For example, subtracting a decimal expected return from a percentage-point return does not produce a meaningful deviation. The variance compounds the unit issue because it squares each deviation.

If all returns are represented in percentage points, such as $4$, $10$, and $16$, the resulting variance is in percentage-points-squared units. If the same returns are represented as decimals, such as $0.04$, $0.10$, and $0.16$, the variance is in decimal-return-squared units. The numerical variances differ by a factor of $100^2=10{,}000$; the standard deviations differ by a factor of $100$. The underlying risk is unchanged when units are converted consistently.

Worked Example: Finding a Variance With Mixed Return Notation

A hypothetical asset has returns of $4\%$, $10\%$, and $16\%$, with respective probabilities $0.20$, $0.50$, and $0.30$. A calculation sheet records the first return as $0.04$, the second as $10$, and the third as $0.16$. Find the correct expected return, variance, and standard deviation, and identify the error. Use decimal returns and round the final rate to four significant figures. The method is to convert every return to a common unit before calculating; this avoids mixing percentages and decimals.

Correct the inputs. The return $10\%$ must be written as $0.10$ when the other returns are decimals. The probabilities sum to $0.20+0.50+0.30=1.00$.

Calculate the expected return.

\mu=(0.20)(0.04)+(0.50)(0.10)+(0.30)(0.16) =0.008+0.050+0.048=0.106=10.60\%.

Calculate the variance using squared deviations. The deviations are $0.04-0.106=-0.066$, $0.10-0.106=-0.006$, and $0.16-0.106=0.054$. Their squares are $0.004356$, $0.000036$, and $0.002916$. Thus,

\operatorname{Var}(R) =(0.20)(0.004356)+(0.50)(0.000036)+(0.30)(0.002916) =0.0008712+0.0000180+0.0008748 =0.001764.

Calculate and check the standard deviation. The square root is $\sqrt{0.001764}=0.04200=4.200\%$. The result is plausible because the three possible returns are within 6.60 percentage points of the mean, and the calculated standard deviation of 4.200% is smaller than that maximum deviation.

Common incorrect calculation. If $10$ is used in place of $0.10$, the supposed weighted mean becomes $(0.20)(0.04)+(0.50)(10)+(0.30)(0.16)=5.056$, or an apparent $505.6\%$. This is not a valid expected return; the input mixed a decimal with a percentage-point number. If all three returns had instead been entered consistently as $4$, $10$, and $16$, the variance would be $17.64$ percentage-points-squared and the standard deviation would be $4.200$ percentage points. These match the decimal results after converting units: $17.64/10{,}000=0.001764$ and $4.200/100=0.04200$.

What Variance Units Mean

Variance is an average of squared deviations, so its units are squared return units. It is not a return rate and should not be described as a percentage return. Standard deviation reverses the squaring by taking the square root, so it is expressed in the same units as the returns themselves.

This distinction matters when communicating risk. If a variance calculation using decimal returns gives $0.001764$, the corresponding standard deviation is $0.04200$, or $4.200\%$. Reporting $0.001764$ as “0.1764% risk” is incorrect: that treats a squared unit as an ordinary return. If the variance is reported in percentage-points-squared units, label it that way.

A Practical Error-Checking Sequence

Before accepting a return or risk result, use a brief audit. It checks the structure of the calculation as well as the arithmetic.

Name the measure.
Decide whether the question asks for a holding-period return, cumulative return, expected return, variance, or standard deviation.

Include all return components.
For total return, include both the investment's value change and any cash distribution received during the holding period.

Check the scenario structure.
Confirm that probabilities are nonnegative and sum to 1, then use them in both the expected return and the variance.

Standardize units.
Choose decimals or percentage points and express every return and deviation in that same unit before calculating.

Check interpretation.
Variance has squared return units; standard deviation has return units. For multiple periods, compound growth factors rather than adding returns.

These checks are useful because many incorrect answers look numerically tidy. A result can be arithmetically consistent with the substituted numbers and still be wrong because a distribution was omitted, an outcome was given the wrong weight, or a unit was changed halfway through the calculation.

Key takeaway: Include distributions in total return, multiply growth factors across periods, weight scenario outcomes by their probabilities, and keep return units consistent. Interpret variance in squared units and standard deviation in the original return units.

Check Your Understanding

For each problem, keep the units explicit and check whether the requested quantity is a return, variance, or standard deviation.

A hypothetical share costs $80.00 at the start of a year, ends at $84.00, and pays a $4.00 distribution. Calculate its total return and its price-only return. (Answer: Total return $10.00\%$; price-only return $5.00\%$.)
A fund has annual total returns of $12\%$ and $-10\%$ in consecutive years. Assuming distributions are reinvested and the returns apply sequentially to wealth, calculate the cumulative two-year return and compare it with their arithmetic sum. (Answer: Cumulative return $0.80\%$; arithmetic sum $2.00\%$.)
A hypothetical asset returns $-5\%$, $5\%$, or $15\%$, with probabilities $0.20$, $0.30$, and $0.50$. Calculate its expected return, variance using decimal returns, and standard deviation. (Answer: Expected return $8.00\%$; variance $0.006100$ in decimal-return-squared units; standard deviation $7.810\%$.)
Two scenarios have returns of $2\%$ and $12\%$, with probabilities $0.40$ and $0.60$. Calculate the expected return, variance in decimal-return-squared units, and standard deviation. (Answer: Expected return $8.00\%$; variance $0.002400$; standard deviation $4.899\%$.)
A hypothetical investment returns $6\%$ with probability $0.25$ and $14\%$ with probability $0.75$. Calculate its expected return and variance in both decimal-return-squared and percentage-points-squared units, then give its standard deviation. (Answer: Expected return $12.00\%$; variance $0.001200$ in decimal-return-squared units or $12.00$ percentage-points-squared; standard deviation $3.464\%$.)

What You'll Learn