Chapter 5 · Valuing a Company

5.1 Discounted Cash Flow and Dividend Models

What is a company worth? Forecast the cash it will generate, discount it at a rate that reflects its risk, add a terminal value for the years beyond the forecast — and then find out how much the answer moves when the assumptions do.

📍 The situation

Kabir, a final-year commerce student in Nagpur, is reading a research note on a hypothetical listed company, Sahyadri Foods Ltd, which makes packaged snacks. Its shares trade at ₹200. The note says: "Our DCF gives a fair value of ₹248; margin of safety about 19%." A second note on the same company, from a different analyst, says ₹206. His uncle, meanwhile, holds shares of Konkan Power Ltd, a hypothetical electricity utility that has paid a steadily rising dividend for years, and values them with a single line: "dividend next year divided by (required return minus growth)".

Kabir wants to know where such numbers come from, why two careful analysts can disagree by 20%, and how much faith to put in either. The answer is the subject of this section. A company is worth the present value of the cash it will generate for its owners — the discounting of §1.1 and the NPV of §1.3 applied to a business instead of a deposit. Three ingredients go in: a forecast of free cash flow, a discount rate that reflects risk (the cost of equity from CAPM, §4.2), and a terminal value for all the years after the forecast, which turns out to be the sum of a geometric series. Then comes the honest part: a valuation is a function of its inputs, and with calculus we can measure exactly how sensitive it is — which is why careful investors insist on a margin of safety.

⚠️ About the numbers on this page

All companies on this page are hypothetical, and their financial statements, share prices, growth rates, betas, risk-free rates, equity risk premiums and tax rates are illustrative inputs chosen to make the arithmetic realistic. They are not data about any real company, forecasts or recommendations. A value computed on this page is the output of a model under stated assumptions — change the assumptions and the value changes. Corporate tax rates and accounting rules change; check the Income Tax Department and current accounting standards for the actual position. Nothing here is investment advice.

Value Is the Present Value of Free Cash Flow

If you buy a fixed deposit, you know the cash it will pay. If you buy a whole business, you are buying the cash it will generate in the future — after paying its staff, suppliers and taxes, and after reinvesting what it needs to keep running and growing. That surplus is the free cash flow. It is uncertain, so we work with forecasts (expected values) and discount them at a rate that reflects the risk.

📘 Definition 5.1 — Intrinsic value and discounted cash flow (DCF)

The intrinsic value of an asset is the present value of the cash flows its owner expects to receive, discounted at a rate $k$ that reflects how risky those cash flows are: $$V_0=\sum_{t=1}^{\infty}\frac{E[CF_t]}{(1+k)^t}.$$ A discounted cash flow (DCF) valuation computes $V_0$ from an explicit forecast of $CF_t$. The market price is what the asset trades at today; intrinsic value is an estimate, and two people with different forecasts will estimate different values.

Accounting profit is not cash. A company's net profit is computed after depreciation (a non-cash charge spreading the cost of machines over their life) and before the money it actually spends on new machines (capital expenditure, capex) and on extra stock and receivables as it grows (the increase in net working capital, NWC). Free cash flow corrects for both.

📘 Definition 5.2 — Free cash flow to the firm and to equity

For one year, with operating profit (earnings before interest and tax) EBIT, tax rate $t$, depreciation and amortisation D&A, capital expenditure capex and increase in net working capital $\Delta\text{NWC}$:

  • Free cash flow to the firm — the cash available to all providers of capital, lenders and shareholders together: $$\text{FCFF}=\text{EBIT}(1-t)+\text{D\&A}-\text{capex}-\Delta\text{NWC}.$$
  • Free cash flow to equity — what is left for shareholders after lenders are paid (after-tax interest) and new borrowing is added: $$\text{FCFE}=\text{FCFF}-\text{interest}\,(1-t)+\text{net borrowing}.$$ Equivalently $\text{FCFE}=\text{net profit}+\text{D\&A}-\text{capex}-\Delta\text{NWC}+\text{net borrowing}$.

Here $\text{EBIT}(1-t)$ is sometimes called NOPAT (net operating profit after tax): the profit the business would make if it had no debt.

✏️ Example 1 — Sahyadri Foods: from profit to free cash flow
Sahyadri Foods' last year, in ₹ crore: revenue 2,000; EBITDA (operating profit before depreciation) 300; D&A 60; so EBIT 240; interest 50; tax at an illustrative 25% of profit before tax. During the year it spent 90 on capex, its net working capital rose by 25, and it borrowed a net 10. It has 10 crore shares. Find net profit, EPS, FCFF and FCFE (both ways).
Profit Profit before tax $=240-50=190$; tax $=0.25\times190=47.5$; net profit $=142.5$ crore. EPS $=142.5/10=$ ₹14.25 per share.
FCFF $\text{EBIT}(1-t)=240\times0.75=180$, so $$\text{FCFF}=180+60-90-25=125\ \text{crore}.$$
FCFE From FCFF: $125-50\times0.75+10=125-37.5+10=97.5$ crore. From net profit: $142.5+60-90-25+10=97.5$ crore. ✓ The two routes agree because net profit $=(\text{EBIT}-\text{interest})(1-t)$.
Interpret The business generated ₹125 crore of cash after taxes and the reinvestment it needed; ₹97.5 crore of it belonged to shareholders. Net profit (₹142.5 crore) overstates the cash available to them, because capex (90) exceeded depreciation (60) and working capital absorbed 25 — typical of a growing company.

The discount rate. Cash flows that belong to shareholders are riskier than a bank deposit, so they are discounted at a higher rate: the return shareholders require. The standard way to estimate it is the Capital Asset Pricing Model of §4.2 (CAPM, Factors, Asset Allocation and Rebalancing): the required return is the risk-free rate plus beta times the equity risk premium. A company financed partly by debt pays its lenders a lower rate, and interest is tax-deductible, so the cost of the company's capital as a whole is a weighted average.

📘 Definition 5.3 — Cost of equity, cost of debt and WACC
  • The cost of equity $k_e$ is the annual return shareholders require. With CAPM (§4.2), $k_e=r_f+\beta\,(E[r_m]-r_f)$, where $r_f$ is the risk-free rate (for example a long-dated G-sec yield, §2.3), $\beta$ the stock's beta (§3.2) and $E[r_m]-r_f$ the equity risk premium.
  • The cost of debt $k_d$ is the interest rate the company pays on new borrowing; its after-tax cost is $k_d(1-t)$.
  • With market values $E$ (market capitalisation) and $D$ (debt), the weighted average cost of capital is $$\text{WACC}=\frac{E}{D+E}\,k_e+\frac{D}{D+E}\,k_d\,(1-t).$$

Matching rule: FCFF belongs to everyone, so it is discounted at the WACC; FCFE belongs to shareholders, so it is discounted at $k_e$.

✏️ Example 2 — Sahyadri's cost of equity and WACC
Assume an illustrative risk-free rate of 7%, an equity risk premium of 5%, and a beta of 1.0 for Sahyadri. Its market capitalisation is ₹2,000 crore (10 crore shares at ₹200), its debt ₹500 crore at 10%, and the tax rate 25%. Find $k_e$ and the WACC.
CAPM $k_e=7\%+1.0\times5\%=12\%$.
Weights $V=2{,}000+500=2{,}500$; $E/V=0.8$, $D/V=0.2$. After-tax cost of debt $=10\%\times0.75=7.5\%$.
WACC $0.8\times12\%+0.2\times7.5\%=9.6\%+1.5\%=11.1\%$.
Comment The weights use market values, because they describe what investors could get for their claims today. Each input is an estimate: a beta of 1.2 instead of 1.0 raises $k_e$ to 13% and the WACC to 11.9%. We will see in the last subsection what such a change does to the value.
📘 Definition 5.4 — Enterprise value, net debt and equity value

The value of the whole business — the present value of its FCFF at the WACC — is its enterprise value (EV). Net debt is debt minus cash and liquid investments. Shareholders own what is left after the lenders' claim: $$\text{equity value}=\text{EV}-\text{net debt},\qquad\text{value per share}=\frac{\text{equity value}}{\text{number of shares}}.$$ (In market terms, $\text{EV}=$ market capitalisation $+$ net debt; §5.2 uses this to build the EV/EBITDA multiple.)

Terminal Value and the DCF Value per Share

No one can forecast a company's cash flows year by year for ever. The standard approach splits the future in two: an explicit forecast for $n$ years (typically 5–10), during which growth may be high and changing, and a terminal value at year $n$ that summarises everything after, assuming the company then grows at a modest constant rate. The terminal value rests on one formula — a geometric series.

📐 Theorem 5.1 (Formula) — The growing perpetuity and the Gordon terminal value

A stream of cash flows $C_1,\ C_1(1+g),\ C_1(1+g)^2,\dots$ paid at the ends of years $1,2,3,\dots$ and discounted at rate $k$ has present value $$PV=\frac{C_1}{k-g}\qquad\text{provided } -1\lt g\lt k.$$ If $g\ge k$ the sum is infinite and the formula does not apply. The first $N$ years contribute the fraction $1-\left(\frac{1+g}{1+k}\right)^N$ of the total. Applied at the end of a forecast of $n$ years, the Gordon (constant-growth) terminal value is $$TV_n=\frac{CF_{n+1}}{k-g}=\frac{CF_n\,(1+g)}{k-g}.$$

Derivation. Put $x=\frac{1+g}{1+k}$. The present value of the year-$j$ cash flow is $\frac{C_1(1+g)^{j-1}}{(1+k)^j}=\frac{C_1}{1+k}\,x^{j-1}$, so the first $N$ years sum to the geometric progression $$\frac{C_1}{1+k}\,(1+x+\cdots+x^{N-1})=\frac{C_1}{1+k}\cdot\frac{1-x^N}{1-x}.$$ Since $1-x=\frac{k-g}{1+k}$, this equals $\frac{C_1}{k-g}\,(1-x^N)$. If $g\lt k$ then $0\lt x\lt1$, so $x^N\to0$ and the sum converges to $\frac{C_1}{k-g}$; if $g\ge k$ then $x\ge1$ and the terms do not shrink, so the sum grows without bound. The terminal value is this formula applied at year $n$ with $C_1$ replaced by $CF_{n+1}$; it is a value at year $n$ and must still be discounted by $(1+k)^n$. ∎

With $g=0$ this is the ordinary perpetuity $C/k$ of §1.2. The condition $g\lt k$ is not a technicality: a company cannot grow faster than the discount rate for ever, and long-run growth should not exceed the growth of the economy in nominal terms. Analysts usually take terminal growth close to long-run inflation plus a little real growth.

✏️ Example 3 — How far into the future does the value lie?
A business will produce ₹100 crore next year, growing 5% a year for ever; the discount rate is 11.1%. Find the present value of the whole stream. What share of it comes from the first 10, 30 and 50 years? After how many years has half the value been received?
Total $PV=\dfrac{100}{0.111-0.05}=\dfrac{100}{0.061}\approx1{,}639.3$ crore — about 16.4 times next year's cash.
Shares With $x=1.05/1.111\approx0.94509$, the first $N$ years give $1-x^N$: $N=10$: $43.1\%$ (₹707 crore); $N=30$: $81.6\%$; $N=50$: $94.1\%$.
Half-life $1-x^N=\tfrac12$ when $N=\dfrac{\log2}{\log(1.111/1.05)}\approx12.3$ years.
Lesson More than half of the value lies beyond year 10 — beyond any analyst's forecast. That is why the terminal value dominates most DCFs, and why its two inputs, $k$ and $g$, deserve the closest scrutiny.
📐 Theorem 5.2 (Formula) — DCF value of a company

With an explicit forecast $\text{FCFF}_1,\dots,\text{FCFF}_n$, a discount rate $\text{WACC}=k$ and terminal growth $g\lt k$, $$\text{EV}=\sum_{t=1}^{n}\frac{\text{FCFF}_t}{(1+k)^t}+\frac{TV_n}{(1+k)^n},\qquad TV_n=\frac{\text{FCFF}_n(1+g)}{k-g},$$ and $\text{value per share}=(\text{EV}-\text{net debt})/\text{shares}$. Equivalently, discounting FCFE at $k_e$ gives the equity value directly; with consistent assumptions the two routes agree.

This is just Definition 5.1 with the infinite tail summed by Theorem 5.1: after year $n$ the flows are $\text{FCFF}_n(1+g)^{j}$, whose value at year $n$ is $TV_n$, and moving $TV_n$ back to today divides by $(1+k)^n$.

🧭 How to build a DCF in seven steps
  1. Understand the business and its last few years of revenue, margins, capex and working capital.
  2. Forecast FCFF for $n$ years: revenue growth × operating margin gives EBIT; subtract tax; add D&A; subtract capex and $\Delta$NWC.
  3. Estimate the discount rate: $k_e$ from CAPM, $k_d$ from the company's borrowing cost, WACC with market-value weights.
  4. Choose terminal growth $g$, well below the WACC and near long-run nominal growth of the economy; compute $TV_n$.
  5. Discount and add: EV = PV(forecast) + PV(TV). Note what share of EV the terminal value is.
  6. Bridge to equity: subtract net debt; divide by the number of shares (diluted for options, if any).
  7. Stress-test: build a sensitivity table over WACC and $g$, run bear and bull scenarios, and compare with the price only after doing so.
✏️ Example 4 — A DCF value per share for Sahyadri Foods
Starting from last year's FCFF of ₹125 crore (Example 1), assume FCFF grows 12% a year for 5 years and 5% a year afterwards. Use the WACC of 11.1% (Example 2). Sahyadri has debt of ₹500 crore and cash of ₹100 crore. Find the EV, the equity value and the value per share, and the share of EV that comes from the terminal value.
Forecast $\text{FCFF}_t=125\times1.12^t$: 140.00, 156.80, 175.62, 196.69, 220.29 (₹ crore).
Discount Factors $1/1.111^t$: 0.9001, 0.8102, 0.7292, 0.6564, 0.5908. Present values: 126.01, 127.03, 128.06, 129.10, 130.15; sum $\approx640.35$.
Terminal value $$TV_5=\frac{220.29\times1.05}{0.111-0.05}\approx3{,}791.9,$$ worth $3{,}791.9\times0.5908\approx2{,}240.2$ today.
Value EV $\approx640.35+2{,}240.21=2{,}880.6$ crore. Net debt $=500-100=400$, so equity $\approx2{,}480.6$ crore and value per share $\approx$ ₹248.06.
Check The terminal value is $2{,}240.2/2{,}880.6\approx77.8\%$ of EV — typical. The first five years, the only ones actually forecast, supply less than a quarter of the value. (This is the first note's "₹248".)
Interactive 1 — Build a DCF: Where Does Sahyadri's Value Come From?

Sahyadri Foods (hypothetical): last FCFF ₹125 crore, net debt ₹400 crore, 10 crore shares, market price ₹200. The blue bars are the present values of the five forecast years (the hollow outlines are the undiscounted cash flows); the purple bar is the present value of the terminal value. Move the sliders, or press Play to sweep the terminal growth rate and watch the purple bar — and the value per share — swing. (Illustrative inputs.)

12%11.1%5.0%
PV of years 1–5—
PV of terminal value—
enterprise value—
value per share—
terminal share of EV—
margin of safety at ₹200—

Dividend Discount Models

For a shareholder who never sells, the only cash a share ever pays is its dividends. (One who sells receives the price, which is in turn the value of the dividends after that.) So a share's value is the present value of its expected dividends — the oldest valuation model, still the natural one for steady dividend payers such as utilities and some banks.

📘 Definition 5.5 — Dividend discount model, payout and retention

The dividend discount model (DDM) values a share as $P_0=\sum_{t\ge1}D_t/(1+k_e)^t$, where $D_t$ is the dividend per share expected in year $t$ and $k_e$ the cost of equity. $D_0$ denotes the dividend just paid, $D_1$ the next one.

If $E$ is earnings per share, the payout ratio is $D/E$ and the retention ratio is $b=1-D/E$, the fraction of profit kept in the business. The return on equity is $\text{ROE}=\text{net profit}/\text{book equity}$ (studied in detail in §5.2).

📐 Theorem 5.3 (Formula) — The Gordon growth model

If dividends grow at a constant rate $g\lt k_e$ for ever, $$P_0=\frac{D_1}{k_e-g}=\frac{D_0(1+g)}{k_e-g}.$$ Rearranged, the return implied by a market price $P_0$ is $$k_e=\frac{D_1}{P_0}+g\qquad(\text{dividend yield}+\text{growth}),$$ and the growth implied by the price is $g=\dfrac{P_0k_e-D_0}{P_0+D_0}$. If the company keeps a constant ROE on its book equity, pays out a constant fraction of earnings and issues no new shares, earnings, dividends and book value all grow at the sustainable growth rate $$g=b\times\text{ROE}.$$

Why. The first formula is Theorem 5.1 with $C_1=D_1$, $k=k_e$. Solving $P_0(k_e-g)=D_1$ for $k_e$ gives the second; writing $D_1=D_0(1+g)$ and solving for $g$ gives the third. For sustainable growth, let $B_t$ be book equity per share at the start of year $t$: earnings are $E_t=\text{ROE}\cdot B_t$, a fraction $b$ is retained, so $B_{t+1}=B_t+bE_t=B_t(1+b\cdot\text{ROE})$. Book value grows at $b\cdot\text{ROE}$; earnings ($\text{ROE}\cdot B_t$) and dividends ($(1-b)E_t$) grow with it. ∎

✏️ Example 5 — Konkan Power: the uncle's one-line valuation
Konkan Power (hypothetical) has just paid a dividend of ₹6.00 a share. Its EPS was ₹10 and its ROE is 12.5%. Its beta is 0.6; take $r_f=7\%$ and an equity risk premium of 5%. (a) Find the sustainable growth rate. (b) Find $k_e$ and the Gordon value. (c) The share trades at ₹110: what return does that price imply if growth is 5%?
(a) Payout $=6/10=60\%$, so $b=0.4$ and $g=0.4\times12.5\%=5\%$.
(b) $k_e=7\%+0.6\times5\%=10\%$. $D_1=6\times1.05=6.30$, so $$P_0=\frac{6.30}{0.10-0.05}=\text{₹}\,126.$$
(c) $k_e=\dfrac{6.30}{110}+0.05\approx5.73\%+5\%=10.73\%$. At ₹110 a buyer who believes the 5% growth expects about 10.7% a year — more than the 10% CAPM requires, which is the same statement as "value ₹126 is above price ₹110".
Check A low beta gives a low $k_e$, and a steady 5% growth gives a value of 20 times next year's dividend: $1/(k_e-g)=1/0.05=20$. The whole answer depends on the 5-percentage-point gap $k_e-g$.

When growth is not constant. A young company may grow fast for a few years and then settle down. Its dividends cannot be valued with one growth rate: $g$ might even exceed $k_e$ for a while, which is fine for a finite stretch but not for ever. The fix is to split the future again.

📐 Theorem 5.4 (Formula) — Two-stage dividend discount model

If dividends grow at $g_1$ for $n$ years and at $g_2\lt k_e$ afterwards, $$P_0=\sum_{t=1}^{n}\frac{D_0(1+g_1)^t}{(1+k_e)^t}+\frac{P_n}{(1+k_e)^n},\qquad P_n=\frac{D_n(1+g_2)}{k_e-g_2}.$$ For $g_1\ne k_e$ the first sum equals $\dfrac{D_1}{k_e-g_1}\left[1-\left(\dfrac{1+g_1}{1+k_e}\right)^{n}\right]$ (the partial sum in Theorem 5.1); $g_1$ may exceed $k_e$, $g_2$ may not.

✏️ Example 6 — Tapti Chemicals: fast growth, then maturity
Tapti Chemicals (hypothetical) just paid ₹4 a share. Dividends are expected to grow 15% a year for 5 years and 6% a year afterwards; $k_e=12\%$. Find the value of a share, and compare it with a one-stage Gordon value at 6%.
Stage 1 Dividends $4\times1.15^t$: 4.60, 5.29, 6.08, 7.00, 8.05. Present values at 12%: 4.107, 4.217, 4.330, 4.446, 4.565; sum $\approx21.67$. (Closed form: $\frac{4.60}{0.12-0.15}\left[1-\left(\frac{1.15}{1.12}\right)^5\right]\approx21.67$ ✓ — both factors are negative.)
Stage 2 $P_5=\dfrac{8.0454\times1.06}{0.06}\approx142.14$, worth $142.14/1.12^5\approx80.65$ today.
Value $P_0\approx21.67+80.65=$ ₹102.32.
Compare One-stage at 6%: $4\times1.06/0.06\approx$ ₹70.67. Five years of 15% growth add about ₹31.65 (45%). Even here, about 79% of the value sits in the terminal price $P_5$.
✏️ Example 7 — Reverse engineering: what is the price assuming?
Instead of asking "what is Konkan Power worth?", ask "what growth is the market price assuming?". With $D_0=$ ₹6, $k_e=10\%$ and price ₹110, find the implied perpetual growth rate. Is it plausible next to the sustainable growth of Example 5?
Solve From $110=\dfrac{6(1+g)}{0.10-g}$: $11-110g=6+6g$, so $g=\dfrac{5}{116}\approx4.31\%$ (the formula $g=\frac{P_0k_e-D_0}{P_0+D_0}$).
Compare The price assumes 4.31% growth for ever; the company's own payout and ROE support 5%. If ROE and payout hold, the price looks undemanding; if ROE slips to about 10.8% ($0.4\times10.8\%\approx4.31\%$), the price is fair.
Why this is useful A reverse DCF turns an argument about one number ("the value is ₹126") into a question about the business that can be checked ("can it keep earning 12.5% on equity?"). Practitioners often find this more honest than a point estimate.

Sensitivity and the Margin of Safety

Every DCF and DDM depends on inputs that nobody knows precisely. Calculus tells us exactly how fragile the answer is. Hold next year's dividend (or cash flow) $D_1$ fixed and treat the value as a function of the two rates.

📐 Theorem 5.5 — Sensitivity of a Gordon value

Let $V(k,g)=\dfrac{D_1}{k-g}$ with $D_1\gt0$ fixed and $g\lt k$. Then $$\frac{\partial V}{\partial g}=\frac{V}{k-g}\gt0,\qquad\frac{\partial V}{\partial k}=-\frac{V}{k-g}\lt0,\qquad\text{so}\qquad\frac{\Delta V}{V}\approx\frac{\Delta g-\Delta k}{k-g}.$$ Exactly, $\dfrac{V(k,g+\Delta)}{V(k,g)}=\dfrac{k-g}{k-g-\Delta}$. Moreover $\dfrac{\partial^2V}{\partial g^2}=\dfrac{2V}{(k-g)^2}\gt0$: $V$ is convex in $g$, so an overestimate of growth raises the value by more than an equal underestimate lowers it. The smaller the gap $k-g$, the more sensitive the value.

Proof. Differentiate $V=D_1(k-g)^{-1}$: $\partial V/\partial g=D_1(k-g)^{-2}=V/(k-g)$ and $\partial V/\partial k=-D_1(k-g)^{-2}$; the first-order change is $dV=V_g\,dg+V_k\,dk$. Differentiating again, $\partial^2V/\partial g^2=2D_1(k-g)^{-3}\gt0$. The exact ratio is $\frac{D_1/(k-g-\Delta)}{D_1/(k-g)}$. ∎ With $k-g=5\%$, a 1-point change in either rate moves the value by about $1/0.05=20\%$ to first order — and the true change upward is larger.

📘 Definition 5.6 — Margin of safety

If an investor estimates intrinsic value $V$ and the market price is $P$, the margin of safety is $$\text{MoS}=\frac{V-P}{V}=1-\frac{P}{V}.$$ A rule "buy only with a margin of safety of at least $m$" means paying at most $(1-m)V$. The margin is a cushion against errors in the inputs, not a guarantee: if the true value is lower than estimated by more than $m$, the cushion is gone.

✏️ Example 8 — Konkan Power: how wrong can we be?
For Konkan Power, $D_1=$ ₹6.30, base case $k_e=10\%$, $g=5\%$. (a) Build a $3\times3$ table of values for $k_e\in\{9,10,11\}\%$ and $g\in\{4,5,6\}\%$. (b) Compare the first-order estimate of a 1-point rise in $g$ with the exact change, and do the same for a 1-point fall. (c) At the price of ₹110, what is the margin of safety in the base case? What price would give a 25% margin?
(a) $V=6.30/(k_e-g)$, in ₹:
$k_e$ \ $g$4%5%6%
9%126.00157.50210.00
10%105.00126.00157.50
11%90.00105.00126.00
Values depend only on $k_e-g$, so the table is constant along diagonals. The range is ₹90 to ₹210 — a factor of 2.3 from rates that each move by just one point.
(b) First order: $\Delta V\approx V\,\Delta g/(k-g)=126\times0.01/0.05=$ ₹25.20 either way. Exact: up to ₹157.50 ($+$₹31.50, $+25\%$), down to ₹105 ($-$₹21, $-16.7\%$). Convexity: the upside is bigger than the downside, and the linear estimate sits between them.
(c) $\text{MoS}=1-110/126\approx12.7\%$. A 25% margin needs a price of at most $0.75\times126=$ ₹94.50. Notice from the table that one bad combination ($k_e=11\%$, $g=4\%$, value ₹90) already wipes out a 12.7% margin.
Interactive 2 — The Gordon Cliff: Value, Growth and the Margin of Safety

The blue curve is the Gordon value $V(g)=D_0(1+g)/(k_e-g)$ of a share that just paid ₹6 (Konkan Power, hypothetical), as the growth assumption $g$ varies; it shoots up as $g$ approaches $k_e$ (dashed line). The red line is the market price; the green band marks the growth assumptions under which the price still leaves your chosen margin of safety. Press Play to move the base-case growth and watch the value accelerate. (Illustrative inputs.)

10.0%5.0%₹110
value at base $g$—
margin of safety—
growth implied by price—
$g$ 1 point higher, same $D_1$—
$g$ 1 point lower, same $D_1$—
implied return at price—
⚠️ Common mistakes and myths
  1. "The DCF says ₹248, so the share is worth ₹248." A DCF value is a function of assumptions. Sahyadri's value ranges from about ₹182 to ₹378 when WACC and terminal growth each move by one point (📈 Market case). Quote a range, not a number.
  2. Discounting profit instead of cash. Net profit ignores capex and working capital; a fast-growing company can report profits and still consume cash (Example 1). Use FCFF or FCFE.
  3. Mixing flows and rates. FCFF goes with the WACC and gives EV; FCFE goes with $k_e$ and gives equity value. Discounting FCFE at the WACC (a lower rate) overstates value; forgetting to subtract net debt from EV does too.
  4. Terminal growth at or above the discount rate. If $g\ge k$ the Gordon formula gives a negative or infinite number. Even $g$ just below $k$ gives absurd values: no company can outgrow the economy for ever.
  5. "The forecast years are where the value is." Usually 60–85% of a DCF is terminal value (Example 4: 78%). Spend your effort on the terminal assumptions and on the discount rate.
  6. Using book-value weights in the WACC, or a cost of debt equal to an old coupon rather than today's borrowing rate.
  7. "A margin of safety makes the investment safe." It protects against moderate estimation error, not against a business whose cash flows collapse. It is a cushion, not insurance.
  8. "Higher growth always means higher value." Growth paid for by retaining profit adds value only if the company earns more than $k_e$ on what it retains (ROE $\gt k_e$, §5.2). Growth at ROE $=k_e$ adds nothing.
  9. Precision theatre. A value of ₹248.06 is not more accurate than "₹200–₹300". Two decimals in the output never fix uncertainty in the inputs.
  10. "The market price is wrong because it differs from my DCF." The price is the market's aggregate estimate. A reverse DCF (Example 7) shows what it assumes; you need a reason to believe your assumptions are better.
📈 Market case — Sahyadri Foods: from a value to a decision

Kabir has the first note's base case (Example 4): FCFF ₹125 crore growing 12% for 5 years, then 5%; WACC 11.1%; net debt ₹400 crore; 10 crore shares; value ₹248.06 against a price of ₹200. He works through four checks, all with the formulas above.

  1. Sensitivity table (value per share, ₹; forecast growth fixed at 12%):
    WACC \ terminal $g$4%5%6%
    10.1%258.0306.1377.8
    11.1%214.7248.1294.5
    12.1%182.1206.4238.6
    Seven of the nine cells are above ₹200; the two pessimistic corners are not.
  2. The second analyst. The second note's ₹206 is exactly the cell WACC 12.1%, $g=5\%$: that analyst used a beta of 1.25 instead of 1.0, so $k_e=13.25\%$ and WACC $=0.8\times13.25\%+0.2\times7.5\%=12.1\%$. The 20% disagreement is a disagreement about one input.
  3. Reverse DCF. Holding everything else at the base case, the ₹200 price is matched by terminal growth of about 3.45%, or by forecast growth of about 7.6% (instead of 12%), or by a WACC of about 12.3%. So the price assumes either noticeably slower growth or a higher risk than the base case — not a collapse.
  4. Margin of safety. Base case: $1-200/248.06\approx19.4\%$. Against the bear cell (₹182.1) the price is about 10% too high; against the bull cell (₹377.8) it is about 47% below value.

What he learns. "Fair value ₹248" really means "₹180–₹380, most likely in the low ₹200s, and the price of ₹200 assumes terminal growth near 3.5% or a slightly higher risk". Whether that is attractive depends on whether he trusts the 12% growth forecast and the beta — questions about the business, not about arithmetic. The mathematics did its job: it turned a single confident number into a map of what has to be true. Interactive 1 lets you redraw that map.

Exercises

All companies are hypothetical and all inputs illustrative. Amounts in ₹ crore unless stated; give values per share to the nearest paisa and rates to two decimals.

I
Free cash flow and the discount rate.
  • 1 A company has EBIT 80, tax rate 25%, D&A 20, capex 35 and an increase in NWC of 10. Find its FCFF.
  • 2 The same company paid interest of 12 and borrowed a net 5 during the year. Find its FCFE.
  • 3 With $r_f=7\%$, an equity risk premium of 5.5% and $\beta=1.3$, find the cost of equity.
  • 4 A company has market capitalisation 3,000, debt 1,000, $k_e=13\%$, $k_d=9\%$ and tax rate 25%. Find its WACC.
  • 5 Explain why FCFF must be discounted at the WACC but FCFE at the cost of equity. What goes wrong if FCFE is discounted at the WACC?
  • 6 A fast-growing company has had negative FCFF for three years while reporting profits. Is that necessarily a bad sign? Explain using Definition 5.2.
  • 7 In part 4 the company's book equity is only 800. Compute the WACC with book weights and say which figure should be used, and why.
1 $\text{FCFF}=80\times0.75+20-35-10=60+20-45=35$.
2 $\text{FCFE}=35-12\times0.75+5=35-9+5=31$.
3 $k_e=7\%+1.3\times5.5\%=14.15\%$.
4 $V=4{,}000$; $\text{WACC}=0.75\times13\%+0.25\times9\%\times0.75=9.75\%+1.6875\%\approx11.44\%$.
5 FCFF pays both lenders and shareholders, so its risk and required return are a blend — the WACC. FCFE is what shareholders alone receive after lenders, so its rate is $k_e$. The WACC is lower than $k_e$ (debt is cheaper and tax-deductible), so discounting FCFE at the WACC overstates the equity value — and counts the benefit of debt twice.
6 Not necessarily: FCFF is NOPAT plus D&A minus capex and $\Delta$NWC. Heavy investment for growth (new plants, more inventory) makes FCFF negative now in exchange for larger cash flows later. It is a bad sign only if the investment earns less than the WACC — the question is the return on that capex, not the sign of today's FCFF.
7 Book weights: $E/V=800/1{,}800$, $D/V=1{,}000/1{,}800$, $\text{WACC}=0.444\times13\%+0.556\times6.75\%\approx9.53\%$. Use the market-value WACC (11.44%): investors' required returns apply to what their claims are worth today. The book weights understate equity and give a rate nearly 2 points too low — which would inflate any DCF built on it.
II
Terminal value and DCF.
  • 1 Find the present value at 10% of FCFF of 100, 110 and 121 at the ends of years 1, 2, 3.
  • 2 After year 3 the FCFF in part 1 grows 4% a year for ever. Find the terminal value at year 3 and its present value.
  • 3 Net debt is 300 and there are 5 crore shares. Find the EV and the value per share.
  • 4 What fraction of the EV in part 3 is terminal value?
  • 5 Find the present value of ₹10 a year starting next year, growing 3% a year for ever, at 8%. What happens to the formula if the growth is 8%?
  • 6 Instead of the Gordon formula an analyst sets the year-3 terminal value at 8 times year-3 EBITDA of 200 (an "exit multiple"). What perpetual growth rate does this terminal value imply, given FCFF$_3=121$ and WACC 10%?
  • 7 With WACC 10%, compute the terminal multiple $TV_n/\text{FCFF}_n=(1+g)/(k-g)$ for $g=2\%,4\%,6\%$. Comment.
1 $100/1.1+110/1.21+121/1.331=90.91+90.91+90.91\approx272.73$ (each flow grows 10%, exactly the discount rate).
2 $TV_3=121\times1.04/0.06\approx2{,}097.33$; present value $2{,}097.33/1.331\approx1{,}575.76$.
3 EV $\approx272.73+1{,}575.76=1{,}848.48$; equity $1{,}548.48$; per share $1{,}548.48/5\approx$ ₹309.70.
4 $1{,}575.76/1{,}848.48\approx85.2\%$.
5 $10/(0.08-0.03)=$ ₹200. At $g=8\%=k$ the formula divides by zero: every year's cash flow has present value $10/1.08$, and infinitely many of them sum to infinity (Theorem 5.1 needs $g\lt k$).
6 $TV_3=8\times200=1{,}600$. Solve $1{,}600=121(1+g)/(0.10-g)$: $160-1{,}600g=121+121g$, $g=39/1{,}721\approx2.27\%$. An exit multiple is a hidden growth assumption; checking it against Gordon keeps it honest.
7 $1.02/0.08=12.75$; $1.04/0.06\approx17.33$; $1.06/0.04=26.5$. Each extra 2 points of terminal growth raises the multiple — and most of the EV — sharply, and ever faster as $g$ approaches $k$.
III
Dividend discount models.
  • 1 A share has just paid ₹5; dividends grow 6% a year; $k_e=11\%$. Find its Gordon value.
  • 2 A share trades at ₹150, next year's dividend is expected to be ₹6, and growth is 7% for ever. What return does the price imply?
  • 3 A company has ROE 15% and pays out 40% of its profit. Find its sustainable growth rate.
  • 4 $D_0=$ ₹2; dividends grow 20% for 3 years and 5% afterwards; $k_e=12\%$. Find the two-stage value.
  • 5 A company pays no dividend at all. Does the DDM say its shares are worthless? Explain.
  • 6 A share trades at ₹80, $D_0=$ ₹4 and $k_e=12\%$. Find the perpetual growth rate the price implies, and check your answer.
1 $D_1=5.30$; $P_0=5.30/0.05=$ ₹106.
2 $k_e=6/150+7\%=4\%+7\%=11\%$.
3 $b=0.6$, $g=0.6\times15\%=9\%$.
4 Dividends 2.40, 2.88, 3.456; PVs at 12%: 2.143, 2.296, 2.460, sum $\approx6.90$. $P_3=3.456\times1.05/0.07=51.84$, PV $=51.84/1.12^3\approx36.90$. Value $\approx$ ₹43.80.
5 No. The model values expected dividends, and a company that retains everything now (to reinvest at a good ROE) is expected to pay larger dividends later, or to be valued on its free cash flow to equity. For such companies the FCFE version of the DCF is more practical. A company that will never return cash to its owners in any form would indeed be worth nothing to them.
6 $g=\dfrac{80\times0.12-4}{80+4}=\dfrac{5.6}{84}\approx6.67\%$. Check: $4\times1.0667/(0.12-0.0667)=4.2667/0.0533=$ ₹80 ✓.
IV
Sensitivity and margin of safety.
  • 1 $D_1=$ ₹10, $k_e=12\%$, $g=7\%$. Find the value, and the values (and percentage changes) when $g$ is 8% and 6%.
  • 2 Use Theorem 5.5 to estimate the value in part 1 if $g$ rises to 7.5%. Compare with the exact value.
  • 3 You estimate a value of ₹250; the price is ₹190. Find the margin of safety. What is the highest price that keeps a 30% margin?
  • 4 Bear, base and bull values of ₹180, ₹250 and ₹320 have probabilities 0.25, 0.5 and 0.25. Find the expected value and, at a price of ₹190, the margin of safety against the bear case.
  • 5 In part 1, suppose $g$ is equally likely to be 6% or 8%. Compare the average of the two values with the value at 7%, and explain the difference with Theorem 5.5. Why is this "convexity bonus" a trap rather than a gift?
  • 6 A mature company has FCFF next year of 50, growing at a constant rate for ever; WACC 11%; net debt 100; 4 crore shares; price ₹150. Find the growth rate the price implies (a reverse DCF).
1 $10/0.05=$ ₹200; at 8%: $10/0.04=$ ₹250 ($+25\%$); at 6%: $10/0.06\approx$ ₹166.67 ($-16.7\%$).
2 $\Delta V\approx V\Delta g/(k-g)=200\times0.005/0.05=$ ₹20, so about ₹220. Exact: $10/0.045\approx$ ₹222.22 — the linear estimate is slightly low because $V$ is convex.
3 $1-190/250=24\%$. For 30%: price at most $0.7\times250=$ ₹175.
4 $0.25\times180+0.5\times250+0.25\times320=$ ₹250. Against the bear case the price is above value: $1-190/180\approx-5.6\%$, a negative margin. A 24% margin on the base case does not cover the bear case.
5 $(166.67+250)/2\approx$ ₹208.33 $\gt$ ₹200: because $V$ is convex in $g$, averaging values over uncertain growth gives more than the value at average growth (the upside branch gains more than the downside loses). It is a trap because it rewards uncertainty about $g$ — and the most uncertain, story-driven companies then look most valuable. It is also exactly where forecasters are most likely to be optimistic.
6 Market equity $=4\times150=600$, so the implied EV $=600+100=700$. $700=50/(0.11-g)$ gives $0.11-g=0.0714$, $g\approx3.86\%$.
🎯 Practice corner NISM-styleCase studyInterview

E1. NISM-style A share is expected to pay a dividend of ₹12 next year; dividends are expected to grow at 5% a year and investors require 13%. By the constant-growth dividend discount model its value is (A) ₹92.31 (B) ₹150.00 (C) ₹157.50 (D) ₹240.00.

Idea $P_0=D_1/(k_e-g)$ with $D_1=12$ already next year's dividend.
Compute $12/(0.13-0.05)=12/0.08=150$. (A) is $12/0.13$ (growth ignored); (C) grows the dividend a second time; (D) divides by $g$ instead of $k_e-g$.
Answer (B).

E2. NISM-style Other things equal, which change increases the value of a share given by the constant-growth dividend discount model? (A) A rise in the stock's beta (B) A fall in the expected dividend growth rate (C) A fall in the risk-free rate (D) A rise in the equity risk premium.

Idea $V=D_1/(k_e-g)$ with $k_e=r_f+\beta\times\text{ERP}$: $V$ rises when $k_e$ falls or $g$ rises.
Decide (A) and (D) raise $k_e$ — value falls. (B) lowers $g$ — value falls. (C) lowers $k_e$ — value rises.
Answer (C).

E3. NISM-style A DCF gives an enterprise value of ₹5,000 crore. The company has debt of ₹1,200 crore, cash of ₹300 crore and 20 crore shares. The value per share is (A) ₹175 (B) ₹190 (C) ₹205 (D) ₹250.

Compute Net debt $=1{,}200-300=900$; equity $=5{,}000-900=4{,}100$; per share $4{,}100/20=$ ₹205.
Distractors (A) subtracts the cash as well as the debt; (B) ignores the cash; (D) forgets the debt altogether.
Answer (C).

E4. Case study Ganga Logistics Ltd (hypothetical) is expected to produce FCFF of ₹180, ₹200 and ₹220 crore in the next three years, then grow 5% a year for ever. Its beta is 1.2; take $r_f=7\%$, an equity risk premium of 5%, a pre-tax cost of debt of 9% and tax of 25%. Market capitalisation is ₹3,000 crore and debt ₹1,000 crore; cash is ₹200 crore; there are 25 crore shares, trading at ₹120. (a) Find $k_e$ and the WACC. (b) Find the EV and the value per share. (c) What share of EV is terminal value? (d) Compute the margin of safety, and find the terminal growth rate the price implies.

(a) $k_e=7\%+1.2\times5\%=13\%$; $\text{WACC}=0.75\times13\%+0.25\times9\%\times0.75=11.4375\%\approx11.44\%$.
(b) PVs of the three flows: 161.53, 161.05, 158.97; sum $\approx481.55$. $TV_3=220\times1.05/(0.114375-0.05)\approx3{,}588.35$, PV $\approx2{,}592.99$. EV $\approx3{,}074.54$; equity $=3{,}074.54-800\approx2{,}274.54$; per share $\approx$ ₹90.98.
(c) $2{,}592.99/3{,}074.54\approx84.3\%$.
(d) $\text{MoS}=1-120/90.98\approx-31.9\%$: the price is about 32% above the base-case value. Solving for the $g$ that makes the value ₹120 gives $g\approx6.34\%$: the market is assuming perpetual growth about 1.3 points higher than the base case — or a lower discount rate. The case does not say the share is "overvalued"; it says what you would have to believe to pay ₹120.

E5. Case study Malabar Bank Ltd (hypothetical) earned EPS of ₹30 last year, pays out 40% of earnings and has a stable ROE of 15%; its cost of equity is 13%. (a) Find its sustainable growth rate. (b) Find its Gordon value. (c) At a price of ₹260, find the growth rate the price implies and the margin of safety. (d) Recompute the value with $k_e=12\%$ and $14\%$, and comment.

(a) $b=0.6$; $g=0.6\times15\%=9\%$.
(b) $D_0=0.4\times30=$ ₹12; $D_1=12\times1.09=13.08$; $V=13.08/(0.13-0.09)=$ ₹327.
(c) $g=\dfrac{260\times0.13-12}{260+12}=\dfrac{21.8}{272}\approx8.01\%$. $\text{MoS}=1-260/327\approx20.5\%$.
(d) $k_e=12\%$: $13.08/0.03=$ ₹436; $k_e=14\%$: $13.08/0.05\approx$ ₹261.60. One point of cost of equity moves the value by $+33\%$ or $-20\%$, because $k_e-g$ is only 4 points. At $k_e=14\%$ the 20% margin of safety has vanished. For a bank, where ROE and risk are both uncertain, a single Gordon value is a starting point, not an answer.

E6. Interview "Why do growth stocks fall more than value stocks when interest rates rise?" Answer with a number: compute the duration $-\frac{1+k}{V}\frac{\partial V}{\partial k}$ (the Macaulay duration of §2.3) of a Gordon stock with $k=10\%$ when $g=6\%$ and when $g=2\%$, and the effect of a 1-point rise in $k$ on each.

Idea By Theorem 5.5, $\partial V/\partial k=-V/(k-g)$, so the duration is $\dfrac{1+k}{k-g}$ — a growing perpetuity behaves like a very long bond.
Compute $g=6\%$: $1.10/0.04=27.5$ years; $g=2\%$: $1.10/0.08=13.75$ years.
Rate rise Exactly, $V(11\%)/V(10\%)=(k-g)/(k-g+0.01)$: $0.04/0.05=0.8$ ($-20\%$) for the growth stock and $0.08/0.09\approx0.889$ ($-11.1\%$) for the value stock.
Answer More of a growth stock's value lies far in the future (Example 3), so like a long-duration bond it is more sensitive to the discount rate. A good interview answer adds the caveat: rates rising for a reason that also raises expected growth (say, higher inflation) can partly offset this.
🧑‍🏫 For the teacher

Suggested plan — 2 lectures × 60 min.

  1. Cash flows, rates and terminal value. Open with the 📍 situation: two analysts, ₹248 and ₹206 — ask the class to guess why. Profit vs cash (Definition 5.2, Example 1 on the board). Cost of equity from CAPM and the WACC (Definitions 5.3–5.4, Example 2). Derive the growing perpetuity as a geometric series (Theorem 5.1) and do Example 3 — the "half-life" of value is a memorable surprise. Then Theorem 5.2 and Example 4. Project Interactive 1: show the purple bar dominating, then sweep terminal growth.
  2. Dividends, sensitivity and judgement. Gordon and sustainable growth (Theorem 5.3, Example 5); two-stage (Example 6); reverse DCF (Example 7). Differentiate the Gordon value (Theorem 5.5) and build the table of Example 8 together. Project Interactive 2 and press Play: let students watch the "cliff" as $g\to k_e$. Close with the 📈 Market case — the two analysts' numbers both appear in one sensitivity table.

Class activity. Split the class into groups, each assigned a different beta (0.8, 1.0, 1.2) and terminal growth (4%, 5%, 6%) for Sahyadri. Each group computes its value per share; collect the results on the board as a sensitivity table, then reveal the price and ask each group for its margin of safety.

Misconceptions to watch for. Discounting net profit; FCFE at the WACC; forgetting net debt; $g\ge k$; believing the forecast years carry most of the value; treating a DCF output as a fact; thinking more growth always means more value.

Suggested homework. I(1, 2, 4, 5), II(1–4, 6), III(1, 3, 4, 6), IV(1, 3, 5, 6); Practice corner E1, E3, E4.

🛠️ Do it yourselfSpreadsheet mini-project · about 60 minutes

Project. Build a reusable DCF sheet with a sensitivity table and a reverse-DCF button, test it on Sahyadri Foods, then use it on the published annual report of any listed company as a learning exercise (not to pick a stock).

Why it matters. Reading a valuation critically means being able to rebuild it. Once the sheet exists, you can see in seconds which input a "fair value" depends on.

What to do.

  1. Inputs in B1:B9: last FCFF (B1), forecast growth (B2), years $n$ (B3, use 5), terminal growth (B4), $r_f$ (B5), beta (B6), equity risk premium (B7), net debt (B8), shares (B9). Add $k_d$, tax, $E$ and $D$ in B10:B13 and the WACC in B14: =B12/(B12+B13)*(B5+B6*B7)+B13/(B12+B13)*B10*(1-B11).
  2. Years 1–5 in A17:A21; FCFF in B17: =$B$1*(1+$B$2)^A17; discount factor in C17: =1/(1+$B$14)^A17; PV in D17: =B17*C17. Copy down.
  3. Terminal value: =B21*(1+B4)/(B14-B4) in B23; its PV =B23*C21 in D23; EV =SUM(D17:D21)+D23; value per share =(EV-B8)/B9; terminal share =D23/EV. Add a check cell =IF(B4>=B14,"g must be below WACC","ok").
  4. Sensitivity: put WACC values down a column and terminal growth values across a row, the value-per-share cell in the corner, and use What-If Analysis → Data Table (row input B4, column input B14). For Sahyadri you should reproduce the Market case table.
  5. Reverse DCF: Goal Seek "set value per share to the current price by changing B4" (then B2). Record the implied growth rates.
  6. For a real company, take revenue, EBIT, D&A, capex and working-capital changes for the last five years from its annual report (available on the company's website and on the NSE/BSE filing pages), compute historical FCFF, and choose — and write down — your own assumptions.

Python version: F=[F0*(1+g1)**t for t in range(1,n+1)]; pv=sum(f/(1+k)**t for t,f in enumerate(F,1)); tv=F[-1]*(1+g)/(k-g); ev=pv+tv/(1+k)**n; vps=(ev-net_debt)/shares; for the reverse DCF use scipy.optimize.brentq(lambda g: vps_of(g)-price, -0.02, k-1e-4).

Hand in: the sheet, a 5×5 sensitivity table, the three implied rates from the reverse DCF, and one paragraph: "for the price to be justified I would have to believe…".

Going further. Add a mid-year convention (discount year-$t$ flows by $(1+k)^{t-0.5}$); replace the Gordon terminal value by an exit EV/EBITDA multiple (§5.2) and report the growth rate it implies (Exercise II.6); add a two-stage DDM sheet and compare.

Instructor note: with Sahyadri's inputs (B1 = 125, B2 = 12%, B4 = 5%, WACC 11.1%, net debt 400, 10 crore shares) the sheet must give EV ≈ ₹2,880.6 crore and ₹248.06 per share, with terminal share 77.8%. Goal Seek to ₹200 gives terminal growth ≈ 3.45%, forecast growth ≈ 7.58% or WACC ≈ 12.28%. A common sheet error is discounting the terminal value by $(1+k)^{n+1}$ instead of $(1+k)^n$.

🧪 Self-check quiz

0/5 correct

Q1. Free cash flow to the firm equals

Q2. Next year's dividend is ₹8, the cost of equity 12% and constant growth 8%. The Gordon value is

Q3. With next year's dividend fixed, the gap $k_e-g$ falls from 4 points to 2 points. The Gordon value

Q4. You estimate a share's value at ₹500 and it trades at ₹400. The margin of safety is

Q5. Which pairing is consistent?