Scenario Analysis for Dividend Discount Valuation

Why Use Scenario Analysis?

A dividend discount valuation is a result of assumptions about future cash distributions and the return investors require. A single estimate can conceal how much the result depends on uncertain inputs. Scenario analysis makes that dependence visible by calculating values under a consistent set of alternative assumptions, such as a lower growth outlook, a stronger dividend forecast, or a higher required return.

The purpose is not to make an uncertain forecast look precise. It is to show which assumptions drive the valuation range and how sensitive the result is to reasonable alternatives. A useful comparison changes one input at a time when the goal is to isolate its effect, and uses internally consistent bundles of assumptions when the goal is to describe coherent upside, base, and downside outcomes.

Definition: Scenario analysis is the calculation and comparison of model results under alternative, explicitly stated assumptions. In dividend discount valuation, a scenario specifies the dividend forecast, the applicable growth rates, and the required return. A scenario value is conditional on those assumptions; it is not a guaranteed price or a prediction that the scenario will occur.

Constant-Growth Scenarios

For a stock expected to pay dividend $D_1$ one year from now, with dividends then growing at a constant annual rate $g$ forever, the Gordon growth model gives the value $P_0$. Here $r$ is the required annual return. The model assumes annual dividend payments, a constant required return, and a perpetual growth rate that satisfies $r>g$. The dividends and rates must refer to the same currency and time units.

P_0=\frac{D_1}{r-g},\qquad r>g

The model does not give a valid finite perpetuity value when $r\leq g$. Scenario analysis must therefore check this condition in every scenario before interpreting the result. It is also important to forecast $D_1$, the next dividend, rather than substitute the dividend just paid without adjusting it for the forecast growth.

The formula follows from discounting each future dividend. Since $D_t=D_1(1+g)^{t-1}$, its present value is a geometric series:

P_0=\sum_{t=1}^{\infty}\frac{D_1(1+g)^{t-1}}{(1+r)^t} =\frac{D_1}{1+r}\sum_{t=0}^{\infty} \left(\frac{1+g}{1+r}\right)^t

When $r>g$, the ratio $(1+g)/(1+r)$ is less than 1, so the infinite geometric series converges. Its sum is $1/[1-(1+g)/(1+r)]$. Substitution gives $P_0=D_1/(r-g)$. The convergence condition is not a technical detail: it is what makes the infinite stream have a finite value under this model.

Worked Example: Isolating the Main Assumptions

A hypothetical stock is expected to pay $D_1=\$2.40$ next year. The base assumptions are constant annual dividend growth $g=3.0\%$ and annual required return $r=9.0\%$. Compare the base value with (a) a 0.5 percentage-point increase in growth, (b) a 0.5 percentage-point increase in required return, and (c) a combined upside scenario with a 5% larger next dividend, growth of 3.5%, and required return of 8.5%. All scenarios use the constant-growth model and annual rates.

Method and base case. First verify $r>g$: $9.0\%>3.0\%$. Then:

P_0=\frac{\$2.40}{0.09-0.03} =\frac{\$2.40}{0.06} =\$40.00

(a) Higher growth only. Keep $D_1=\$2.40$ and $r=9.0\%$, and set $g=3.5\%$. The spread is $0.09-0.035=0.055$, so $P_0=\$2.40/0.055=\$43.63636\ldots$, or \$43.64 to the nearest cent. Holding the other inputs fixed, higher growth raises the value.

(b) Higher required return only. Keep $D_1=\$2.40$ and $g=3.0\%$, and set $r=9.5\%$. The spread is $0.095-0.03=0.065$, so $P_0=\$2.40/0.065=\$36.92307\ldots$, or \$36.92 to the nearest cent. Holding the other inputs fixed, a higher required return lowers the value.

(c) Combined upside case. A 5% increase in the next dividend gives $D_1=\$2.40(1.05)=\$2.52$. The denominator is $0.085-0.035=0.05$, which is positive. Thus $P_0=\$2.52/0.05=\mathbf{\$50.40}$.

Check and trap avoided. The base denominator is $0.06$, so the base value is $\$40.00$, not $\$2.40/0.058$ or a value obtained using the dividend just paid. The first two alternatives change only one assumption; the combined case changes three, so its increase cannot be attributed to growth alone.

What Drives the Valuation Range?

The constant-growth formula shows why small changes can matter. With $D_1>0$ and $r>g$, its partial derivatives are:

\frac{\partial P_0}{\partial g} =\frac{D_1}{(r-g)^2}>0,\qquad \frac{\partial P_0}{\partial r} =-\frac{D_1}{(r-g)^2}<0

The strict inequalities above assume $D_1>0$. If $D_1=0$, both derivatives are zero. The derivatives describe the direction of a small change while the other inputs remain fixed. They also show that sensitivity increases as $r-g$ narrows: the squared denominator becomes smaller, so the magnitude of each derivative becomes larger. This is one reason to examine the spread between required return and growth rather than discussing either rate in isolation.

A scenario table should make clear which inputs changed. For example, an upside case might combine a larger dividend, faster sustainable growth, and a lower required return. That combination can be useful if the assumptions tell a coherent story, but it cannot identify the separate contribution of each input. For that purpose, compare one-at-a-time sensitivities around the same base case.

Comparison	Inputs held fixed	Input changed	Expected direction for value
Dividend sensitivity	$g$, $r$	$D_1$	Higher $D_1$ raises value
Growth sensitivity	$D_1$, $r$	$g$, with $r>g$	Higher $g$ raises value
Required-return sensitivity	$D_1$, $g$	$r$, with $r>g$	Higher $r$ lowers value

Scenarios with an Explicit Forecast Period

A company may have dividends forecast individually for an initial period before reaching a stable growth assumption. In that case, value the explicit dividends separately and calculate a terminal value at the end of the forecast period. The terminal value is based on the first dividend after that period, not the last dividend within it. The same terminal assumptions must be used consistently when comparing scenarios.

Model setup: If dividends $D_1,\ldots,D_n$ are forecast explicitly and dividends grow at constant rate $g$ from year $n+1$ onward, with constant annual required return $r>g$, then $$ P_0=\sum_{t=1}^{n}\frac{D_t}{(1+r)^t} +\frac{P_n}{(1+r)^n}, \qquad P_n=\frac{D_{n+1}}{r-g}. $$ The terminal value $P_n$ is measured at time $n$, and $D_{n+1}$ is the first dividend in the continuing-growth period.

Worked Example: Changing Terminal Growth in a Multistage Forecast

A hypothetical company is forecast to pay $D_1=\$1.50$, $D_2=\$1.65$, and $D_3=\$1.80$ at the ends of years 1, 2, and 3. Beginning in year 4, dividends grow at a constant annual rate. The required annual return is $r=9.0\%$. Compare terminal growth assumptions of 3.0% and 2.5%. Use annual timing and round only final values to the nearest cent.

Method. Discount the first three dividends individually, calculate the terminal value at time 3 using $D_4$, and discount that terminal value three years. The discount factors are $1.09$, $1.09^2=1.1881$, and $1.09^3=1.295029$.

At $g=3.0\%$. The first continuing-growth dividend is $D_4=\$1.80(1.03)=\$1.854$. The terminal value at time 3 is:

P_3=\frac{\$1.854}{0.09-0.03} =\frac{\$1.854}{0.06} =\$30.90

The year-3 cash flow for present-value purposes is $D_3+P_3=\$1.80+\$30.90=\$32.70$. Therefore:

P_0=\frac{\$1.50}{1.09} +\frac{\$1.65}{1.1881} +\frac{\$32.70}{1.295029} =\$1.3761468+\$1.3887745+\$25.2504 =\$28.0153\ldots

The value is \$28.02 to the nearest cent.

At $g=2.5\%$. Now $D_4=\$1.80(1.025)=\$1.845$, and $r-g=0.09-0.025=0.065$. Thus $P_3=\$1.845/0.065=\$28.384615\ldots$. The year-3 amount is $\$1.80+\$28.384615\ldots=\$30.184615\ldots$, giving:

P_0=\frac{\$1.50}{1.09} +\frac{\$1.65}{1.1881} +\frac{\$30.184615\ldots}{1.295029} =\$1.3761468+\$1.3887745+\$23.308061\ldots =\$26.072982\ldots

The revised value is \$26.07, a decrease of $\$28.02-\$26.07=\$1.95$ after rounding each reported value to cents. The change comes from the terminal-growth assumption; the first three dividends and required return were held fixed.

Check and trap avoided. Both terminal denominators are positive, satisfying $r>g$. A common error is to use $D_3$ in the terminal-value formula. The correct input is $D_4$, because the continuing-growth stream starts after year 3. Another error is to discount $P_3$ for only two years; a value measured at time 3 is discounted by $1.09^3$ to time 0.

Probability-Weighted Scenarios

Sometimes an analyst assigns probabilities to a finite set of scenarios. If the scenarios are mutually exclusive, exhaustive, and have values $P_{0,j}$ with probabilities $p_j$, their probability-weighted value is $\sum_j p_jP_{0,j}$, where $\sum_j p_j=1$. This is the expected value of the scenario-specific valuations under the stated scenario probabilities. The probabilities are additional assumptions; they do not arise automatically from the dividend discount model.

Because the valuation formula is nonlinear in growth and required return, valuing the probability-weighted average inputs generally does not give the probability-weighted average of scenario values. If the task is to report an expected scenario value, calculate each scenario value first and then weight those values.

Worked Example: Weighting Three Coherent Cases

An analyst considers three hypothetical constant-growth cases for a stock. The low case has probability 0.20, $D_1=\$2.10$, $g=2.0\%$, and $r=9.0\%$. The base case has probability 0.50, $D_1=\$2.40$, $g=3.0\%$, and $r=9.0\%$. The high case has probability 0.30, $D_1=\$2.75$, $g=3.5\%$, and $r=9.0\%$. Assume annual dividends and returns, and use the constant-growth model.

Step 1: Check the model condition and value each case. In all cases, $r>g$. The low-case value is $\$2.10/(0.09-0.02)=\$30.00$. The base-case value is $\$2.40/(0.09-0.03)=\$40.00$. The high-case value is $\$2.75/(0.09-0.035)=\$2.75/0.055=\$50.00$.

Step 2: Apply the probabilities. They sum to $0.20+0.50+0.30=1.00$, so the three cases are a complete probability set for this calculation:

E[P_0] =0.20(\$30.00)+0.50(\$40.00)+0.30(\$50.00) =\$6.00+\$20.00+\$15.00 =\$41.00

The probability-weighted scenario value is \$41.00. It is conditional on the supplied scenarios and probabilities; it is not automatically the market price.

Check and trap avoided. The weights sum to 1 and each scenario satisfies the convergence condition. Do not average the inputs first and then assume the model gives the same result: the model is nonlinear in $g$ and $r$. The expected value requested here is obtained by valuing each scenario first, then weighting those values.

Using Scenario Results Carefully

A valuation range is most informative when the assumptions are transparent. State the next dividend, the growth pattern, the required return, and the timing of each cash flow. For a multistage forecast, identify which forecast years change and which assumptions determine the terminal value. Report the resulting value alongside the assumptions that produce it rather than presenting a range without explanation.

Several common errors can make a scenario comparison misleading:

Changing several inputs but attributing the result to one. Use one-at-a-time comparisons to measure an isolated effect; label bundled cases as combined scenarios.
Ignoring the spread $r-g$. A small spread can generate a large value and high sensitivity. Check $r>g$ before calculating a constant-growth value.
Mixing timing conventions. Use $D_1$ in the Gordon formula, and use $D_{n+1}$ to calculate a terminal value at time $n$.
Confusing a scenario range with a probability statement. A set of low, base, and high cases has no implied probabilities unless probabilities are explicitly assigned.
Averaging inputs when the task asks for an expected value. Calculate scenario values first, then apply the given probabilities.

Scenario analysis is useful because it turns model uncertainty into a structured comparison. It cannot eliminate uncertainty in dividends, growth, or required return, but it can show which assumptions matter most, whether the valuation is robust to plausible changes, and where the model becomes especially sensitive.

Check Your Understanding

For each problem, use the stated annual timing and assumptions. Round dollar values to the nearest cent where needed.

A hypothetical stock will pay $D_1=\$1.80$, with perpetual growth $g=2.5\%$ and required return $r=8.5\%$. Verify the model condition and calculate its value. (Answer: $\$30.00$)
A stock has $D_1=\$3.20$, $g=4.0\%$, and $r=10.0\%$. Calculate its base value, then calculate the value if only the required return rises to 10.5%. (Answer: Base $\$53.33$; revised $\$49.23$)
A company is forecast to pay $D_1=\$1.20$, $D_2=\$1.32$, and $D_3=\$1.44$. From year 4 onward, dividends grow at 3.0% forever. If $r=9.0\%$, calculate $D_4$, the terminal value at time 3, and $P_0$. (Answer: $D_4=\$1.4832$; $P_3=\$24.72$; $P_0=\$22.41$)
Three exhaustive scenarios have values of $\$30$, $\$40$, and $\$50$, with probabilities 0.20, 0.50, and 0.30, respectively. Calculate the probability-weighted value and verify the probabilities sum to 1. (Answer: Probabilities sum to 1.00; weighted value $\$41.00$)
A proposed constant-growth case has $D_1=\$2.10$, $g=8.0\%$, and $r=7.0\%$. Should the Gordon growth formula be used to report a finite value? State the relevant condition. (Answer: No; $r>g$ is required, but here $r<g$)

Comparison	Inputs held fixed	Input changed	Expected direction for value
Dividend sensitivity	\(g\), \(r\)	\(D_1\)	Higher \(D_1\) raises value
Growth sensitivity	\(D_1\), \(r\)	\(g\), with \(r>g\)	Higher \(g\) raises value
Required-return sensitivity	\(D_1\), \(g\)	\(r\), with \(r>g\)	Higher \(r\) lowers value

What You'll Learn