Common Errors In Cost Of Equity Estimation

Why Estimation Errors Matter

A cost of equity estimate can be arithmetically correct and still be misleading if its inputs do not match the model. A dividend-growth estimate may use a dividend that has already been paid; a CAPM calculation may substitute a stock’s total volatility for beta; or a quarterly growth rate may be combined with an annual return. Each error changes the meaning of the result, not merely its presentation.

The two common approaches in this topic are CAPM and the constant-growth dividend model. They estimate a required return from different inputs, so their assumptions and timing conventions need to be checked separately. Neither method turns uncertain inputs into a certain answer.

Definition: In CAPM, the cost of equity is estimated as $r_e=r_f+\beta_e E[R_m-R_f]$, where $r_f$ is the risk-free return for the period, $\beta_e$ is the equity’s beta relative to the market, and $E[R_m-R_f]$ is the expected market risk premium for that same period. The model describes expected return under its equilibrium assumptions; it does not guarantee the return that will be realized.

Definition: Under the constant-growth dividend model, $P_0=D_1/(r_e-g)$, where $P_0$ is the current ex-dividend share price, $D_1$ is the dividend expected one period from now, $g$ is the constant dividend growth rate per period, and $r_e$ is the required return per period. The model requires a stable perpetual growth assumption and $r_e>g$.

Deriving the Dividend-Growth Estimate

The Gordon growth model values a stream of expected dividends growing at a constant rate. Suppose the next dividend is $D_1$, each later dividend grows by $g$ per period, and each is discounted at the required return $r_e$. The price is the present value of that stream:

P_0=\frac{D_1}{1+r_e}+\frac{D_1(1+g)}{(1+r_e)^2} +\frac{D_1(1+g)^2}{(1+r_e)^3}+\cdots

This is an infinite geometric series. Its first term is $D_1/(1+r_e)$, and the ratio between successive terms is $(1+g)/(1+r_e)$. When $r_e>g$, that ratio is less than 1, so the series converges. Applying the sum formula for an infinite geometric series gives:

P_0= \frac{D_1/(1+r_e)}{1-(1+g)/(1+r_e)} =\frac{D_1}{r_e-g}.

Rearranging gives the implied cost of equity $r_e=D_1/P_0+g$. The dividend in the yield term must be the next expected dividend: $D_1$, not the dividend that has already been paid, $D_0$. If the only supplied dividend is $D_0$ and growth is expected to be $g$ for one period, first calculate $D_1=D_0(1+g)$.

Worked Example: Using the Next Expected Dividend

A hypothetical company’s shares trade ex-dividend at $48.00. Its most recent annual dividend was $2.40 per share, and the dividend is expected to grow at a constant 4.0% annually. Estimate the cost of equity under the Gordon growth model. Assume annual dividends, annual effective growth and required returns, and perpetual constant growth. State the common error avoided.

Method and timing. The supplied $2.40 is $D_0$, the dividend just paid. The model requires $D_1$, one year ahead. Calculate:

D_1=D_0(1+g)=\$2.40(1.04)=\$2.496.

The next-dividend yield is:

\frac{D_1}{P_0}=\frac{\$2.496}{\$48.00}=0.052=5.2\%.

Therefore:

r_e=\frac{D_1}{P_0}+g=5.2\%+4.0\%=9.2\%.

Check. The estimated required return exceeds growth by 5.2 percentage points, so $r_e>g$, as required. Substitution gives $D_1/(r_e-g)=\$2.496/0.052=\$48.00$, matching the stated price.

Common error avoided. Using the trailing dividend directly gives $\$2.40/\$48.00+4.0\%=9.0\%$. That calculation uses $D_0$ where the model requires $D_1$, understating the estimate by 0.2 percentage point.

Match the Periods Before Calculating

Every term in a return formula must refer to a consistent time period. In CAPM, the risk-free rate, beta inputs, and market risk premium must have compatible return periods. In the dividend-growth model, the dividend yield and growth rate must use the same period. For example, a quarterly dividend yield cannot simply be added to an annual growth rate.

When a periodic effective return $r_q$ applies for four quarters, its equivalent annual effective return is $(1+r_q)^4-1$. Annualizing a rate this way compounds it; multiplying by four is only a simple approximation and does not give the annual effective rate. The growth rate must also be converted if an annual growth figure is needed.

Worked Example: Converting a Quarterly Dividend Estimate

A hypothetical share is priced at $40.00 just after a quarterly dividend. The most recent quarterly dividend was $0.50, and quarterly dividends are expected to grow at 1.0% per quarter forever. Estimate the cost of equity as a quarterly rate and then as an annual effective rate. Assume the price and dividends are in dollars, the model’s growth is constant, and the required return exceeds growth. Identify the period-matching trap.

Method. First find the next quarterly dividend:

D_1=\$0.50(1.01)=\$0.505.

Use the quarterly yield and quarterly growth together:

r_q=\frac{\$0.505}{\$40.00}+0.01 =0.012625+0.01 =0.022625=2.2625\%\text{ per quarter}.

The model condition holds because $2.2625\%>1.0\%$. Convert the periodic required return to an annual effective rate:

r_{\text{annual}}=(1+0.022625)^4-1 =0.0936178\ldots =9.36\%.

The quarterly growth rate has an annual effective equivalent of $(1.01)^4-1=0.040604$, or 4.06%. This is consistent with the quarterly convention; it is not a reason to add 4.06% directly to the quarterly yield.

Check. The quarterly estimate is greater than the quarterly growth rate, and compounding it produces an annual effective return of 9.36%. Common error avoided: adding the 1.0% quarterly growth rate to an annual dividend yield or annual return mixes periods and has no coherent interpretation.

Beta Is Not Total Risk

CAPM compensates investors for systematic risk: risk associated with movements in the market that cannot be eliminated by broad diversification. Beta measures this market-related risk. For returns $R_i$ on the equity and $R_m$ on the market, beta is:

\beta_e=\frac{\operatorname{Cov}(R_i,R_m)} {\operatorname{Var}(R_m)}.

A stock’s standard deviation measures its total return variability, including both market-related and company-specific movements. It is not a substitute for beta. A highly volatile stock can have a comparatively modest beta if its returns do not move closely with the market; a less volatile stock can have a larger beta if its returns move strongly with the market.

CAPM uses expected returns and a market risk premium for the same period. It assumes that beta is the relevant measure of priced risk in the model and that the market portfolio and risk-free asset are appropriately defined. A calculated CAPM return is a model-based estimate, not a forecast that the stock will deliver that return in every period.

Worked Example: Using Covariance Rather Than Volatility

A hypothetical equity has an annual return standard deviation of 30%. Its covariance with the market’s annual return is 0.009, and the market return standard deviation is 15%. The annual risk-free rate is 4.0%, and the expected annual market risk premium is 7.0%. Estimate beta and the CAPM cost of equity. Assume all return statistics use the same annual measurement period and the stated market premium is an expectation.

Method. First calculate market return variance:

\operatorname{Var}(R_m)=(0.15)^2=0.0225.

Then calculate beta from covariance divided by market variance:

\beta_e=\frac{0.009}{0.0225}=0.40.

Apply CAPM:

r_e=4.0\%+0.40(7.0\%)=4.0\%+2.8\%=6.8\%.

Check. The implied correlation is $0.009/(0.30\times0.15)=0.20$, and $\beta_e=0.20(0.30/0.15)=0.40$, consistent with the covariance calculation. Common error avoided: inserting the equity’s 30% standard deviation as beta would produce $4.0\%+0.30(7.0\%)=6.1\%$. That is not a CAPM calculation: volatility is not beta.

Estimates Are Not Certain

A cost of equity estimate depends on inputs that may themselves be estimated. A beta may depend on the sample period and market index used; a market risk premium is not directly observable; a dividend-growth estimate may be sensitive to the historical window or forecast assumptions. Reporting many decimal places does not remove that uncertainty.

A useful check is to vary plausible inputs and observe the resulting range. This is a sensitivity analysis, not a confidence interval unless a statistical method specifically justifies interpreting it as one. The range should be described as conditional on the chosen input bounds.

Worked Example: Reporting a CAPM Estimate With Input Ranges

An analyst estimates a hypothetical company’s annual cost of equity using a 3.2% annual risk-free rate, beta of 0.90, and expected annual market risk premium of 5.5%. For sensitivity analysis, beta is considered plausible from 0.80 to 1.00 and the market risk premium from 5.0% to 6.0%. Assume the risk-free rate is held fixed and the input ranges are scenario bounds, not statistical confidence limits. Calculate the point estimate and the range.

Point estimate. Using the central inputs:

r_e=3.2\%+0.90(5.5\%)=3.2\%+4.95\%=8.15\%.

For the low scenario, combine the lower beta and lower premium:

r_{e,\min}=3.2\%+0.80(5.0\%)=7.2\%.

For the high scenario, combine the higher beta and higher premium:

r_{e,\max}=3.2\%+1.00(6.0\%)=9.2\%.

Check. With a positive market risk premium, CAPM increases as either beta or the premium increases, so these combinations produce the stated low and high values. Report the point estimate as 8.15% under the central assumptions, with a scenario range of 7.2% to 9.2%. Common error avoided: calling 8.15% a certain or uniquely known return ignores the sensitivity of the result to estimated inputs.

A Practical Error Check

Before relying on a cost of equity estimate, verify what each input represents and what period it covers. The following checks address the most frequent calculation and interpretation errors without treating either model as universally appropriate.

Identify the model and its assumptions.
For CAPM, confirm the risk-free rate, market premium, and beta are appropriate and period-consistent. For constant growth, confirm dividends are expected to grow at a stable perpetual rate and that the estimated return exceeds growth.

Check dividend timing.
In the Gordon formula, use the next dividend $D_1$. If given the latest paid dividend $D_0$, forecast it forward for one period using the stated growth rate.

Match the time unit.
Do not add a quarterly yield to an annual growth rate or combine periodic CAPM inputs with an annual premium. Convert effective rates by compounding when changing the period.

Use the right risk measure and report uncertainty.
CAPM uses beta, not total volatility. State the result as an estimate conditional on its inputs, and use sensitivity analysis when important inputs are uncertain.

Key takeaway: A defensible estimate uses the next expected dividend when applying Gordon growth, keeps rates and cash flows on a consistent time basis, uses beta rather than total volatility in CAPM, and acknowledges that uncertain inputs make the result an estimate rather than a certainty.

Check Your Understanding

For each problem, state the relevant period and apply the model’s conditions before interpreting the result.

A hypothetical share costs $36.00 just after its latest annual dividend of $1.80. Dividends are expected to grow at 5.0% annually forever. Under the constant-growth model, what is the implied annual cost of equity? (Answer: 10.25%)
A quarterly required return is estimated at 2.0%. What is its equivalent annual effective rate? (Answer: 8.2432%)
A hypothetical equity has covariance with the market of 0.009, while the market return standard deviation is 15%. If the annual risk-free rate is 4.0% and the expected annual market risk premium is 7.0%, calculate beta and the CAPM cost of equity. (Answer: $\beta_e=0.40$; cost of equity $=6.8\%$)
A CAPM estimate uses a 2.5% risk-free rate, beta between 0.80 and 1.00, and an expected market premium between 4.5% and 6.5%. Holding the risk-free rate fixed, what is the scenario range for the estimated cost of equity? (Answer: 6.1% to 9.0%)
A hypothetical company has a current share price of $30.00, latest annual dividend $D_0=\$1.50$, and expected annual growth of 4.0%. Its CAPM inputs are an annual risk-free rate of 3.5%, beta of 1.10, and expected annual market risk premium of 5.0%. Calculate both annual estimates and their difference. Assume constant perpetual dividend growth for the Gordon estimate. (Answer: Gordon estimate $=9.2\%$; CAPM estimate $=9.0\%$; difference $=0.2$ percentage point)

What You'll Learn