Chapter 5 · Valuing a Company

5.2 Ratios, Relative Valuation and Sensitivity Analysis

EPS, P/E, PEG, P/B, EV/EBITDA, dividend yield and ROE — what each ratio measures, how each one hides a DCF assumption, how to compare a company with its peers, and why a low multiple is not the same as a bargain.

📍 The situation

Priya, a software engineer in Hyderabad, has started using a stock screener. She is looking at three hypothetical consumer companies. Narmada Consumer Ltd trades at ₹540 with a P/E of 45; Godavari Retail Ltd at ₹252 with a P/E of 28; Kaveri Home Products Ltd at ₹162 with a P/E of 18. A friend says "Kaveri is obviously the cheapest". Another says "P/E means nothing — look at the PEG". A third says "for retailers use EV/EBITDA". The screener also shows P/B, dividend yield and ROE, each in a different colour.

Ratios are the everyday language of the stock market because they are quick: one division puts a ₹540 share and a ₹162 share on the same scale. But every ratio is a compressed DCF. A P/E of 45 is not "expensive" or "cheap" by itself; it is a statement about growth, return on equity and risk — and §5.1's Gordon model tells us exactly which statement. In this section we define the main ratios, derive the formulas that connect them (a little algebra and one telescoping product), learn to compare peers properly, and build the sensitivity and scenario tables that show what a price requires. The lesson Priya will take away is in the section's last heading: why cheap can stay cheap.

⚠️ About the numbers on this page

All companies on this page are hypothetical, and their profits, prices, ratios, growth rates and costs of equity are illustrative numbers chosen for realistic arithmetic — not data about any real company, sector averages, forecasts or recommendations. Screeners and companies compute some ratios in slightly different ways (trailing or forward earnings, year-end or average balances); always check the definition used. Accounting and disclosure rules change; check current accounting standards and SEBI's disclosure requirements. Nothing here is investment advice.

Earnings Multiples: EPS, P/E and What a P/E Assumes

A share is a slice of a company, so to compare companies of different sizes we divide everything by the number of shares. The most quoted per-share number is earnings per share, and the most quoted ratio is the price you pay for each rupee of it.

📘 Definition 5.7 — EPS, P/E ratio and earnings yield
  • Earnings per share: $\text{EPS}=\dfrac{\text{net profit}-\text{preference dividends}}{\text{number of equity shares}}$ (in practice the weighted average number of shares over the year). Diluted EPS also counts shares that could be created by employee stock options or convertible securities.
  • The price-to-earnings ratio is $\text{P/E}=\dfrac{\text{share price}}{\text{EPS}}$ $=\dfrac{\text{market capitalisation}}{\text{net profit}}$. It is trailing if EPS is the last twelve months' and forward if it is next year's expected EPS.
  • The earnings yield is $\text{EPS}/\text{price}=1/(\text{P/E})$: the profit per ₹100 invested.

P/E is undefined (or meaningless) when earnings are zero or negative.

✏️ Example 1 — Narmada Consumer's EPS and P/E
Narmada Consumer (hypothetical) earned a net profit of ₹480 crore last year; it has 40 crore shares and no preference shares; the price is ₹540. Employee options could create 1 crore new shares. Next year's EPS is expected to be ₹13.20. Find the EPS, diluted EPS, trailing and forward P/E, and the earnings yield. Compare the earnings yield with an illustrative 7% G-sec yield.
EPS $480/40=$ ₹12.00; diluted $480/41\approx$ ₹11.71.
P/E Trailing $540/12=45$; forward $540/13.20\approx40.9$.
Earnings yield $12/540\approx2.22\%$ — about a third of the G-sec yield.
Interpret A buyer at ₹540 accepts ₹2.22 of current profit per ₹100, against ₹7 of fixed interest from a G-sec. That only makes sense if profits are expected to grow strongly for a long time. How much growth, exactly? The next theorem answers that.

What a P/E assumes. Combine the Gordon model of §5.1 (Theorem 5.3) with the definitions of payout and sustainable growth: next year's dividend is $D_1=(1-b)E_1$ and $g=b\cdot\text{ROE}$, so $1-b=1-g/\text{ROE}$.

📐 Theorem 5.6 (Formula) — The justified P/E

For a company growing at a constant $g\lt k$ for ever, with cost of equity $k$, constant ROE and payout ratio $1-b$, $$\frac{P_0}{E_1}=\frac{1-b}{k-g}=\frac{1-g/\text{ROE}}{k-g}\quad(\text{forward}),\qquad \frac{P_0}{E_0}=\frac{(1-b)(1+g)}{k-g}\quad(\text{trailing}).$$ In particular:

  1. if $\text{ROE}=k$, the forward P/E is $1/k$ whatever the growth rate;
  2. if $\text{ROE}\gt k$, the P/E increases with $g$; if $\text{ROE}\lt k$, it decreases with $g$.

Proof. Divide $P_0=D_1/(k-g)$ by $E_1$ and use $D_1/E_1=1-b=1-g/\text{ROE}$; the trailing version uses $E_1=E_0(1+g)$. If $\text{ROE}=k$: $\frac{1-g/k}{k-g}=\frac{(k-g)/k}{k-g}=\frac1k$. In general $$\frac{d}{dg}\,\frac{1-g/\text{ROE}}{k-g}=\frac{-\frac{1}{\text{ROE}}(k-g)+\left(1-\frac{g}{\text{ROE}}\right)}{(k-g)^2}=\frac{1-k/\text{ROE}}{(k-g)^2},$$ which is positive exactly when $\text{ROE}\gt k$. ∎ The economics: retaining a rupee to earn ROE on it creates value only if ROE beats the return shareholders could get elsewhere ($k$). Growth is good only when it is profitable growth.

✏️ Example 2 — Same growth, very different P/Es
Three hypothetical companies each grow at 8% a year for ever; the cost of equity is 12% for all three. Their ROEs are 25%, 12% and 10%. Find each one's payout ratio and justified forward P/E. Then redo the 12%-ROE company with 4% growth.
ROE 25% Payout $=1-0.08/0.25=68\%$; P/E $=0.68/0.04=17.0$.
ROE 12% Payout $=1-0.08/0.12=33.3\%$; P/E $=0.3333/0.04\approx8.33=1/0.12$.
ROE 10% Payout $=1-0.08/0.10=20\%$; P/E $=0.20/0.04=5.0$ — growth destroys value: to grow 8% the company must keep 80% of a profit it reinvests at only 10%.
ROE 12%, $g=4\%$ Payout $=2/3$; P/E $=0.6667/0.08\approx8.33$ again: with ROE $=k$ the growth rate does not matter (Theorem 5.6(1)).

Earnings growth compounding. If EPS grows at $g$ a year, after $n$ years it is $E_0(1+g)^n$ — the compound-interest formula of §1.1 applied to profits. A high P/E is a bet that this compounding will be fast and long. The quickest way to see the size of the bet is to ask how much growth a buyer needs.

📘 Definition 5.8 — PEG ratio, book value, P/B ratio and dividend yield
  • The PEG ratio is $\dfrac{\text{P/E}}{100\times\text{expected annual EPS growth}}$; e.g. P/E 30 with 15% expected growth gives PEG 2. It is a rough adjustment of the P/E for growth.
  • Book value (shareholders' equity, or net worth) is assets minus liabilities on the balance sheet; book value per share is $\text{BVPS}=\text{book equity}/\text{shares}$. The price-to-book ratio is $\text{P/B}=\text{price}/\text{BVPS}$.
  • The dividend yield is $\text{DPS}/\text{price}$, where DPS is the dividend per share for the year.
✏️ Example 3 — What growth does a P/E of 45 need?
Analysts expect Narmada's EPS (₹12) to grow 15% a year. (a) Find its PEG. (b) An investor buying at ₹540 wants a price return of 12% a year for 10 years, and thinks that by then the P/E will have settled at 25. What EPS growth does she need? (Ignore dividends.)
(a) PEG $=45/15=3.0$.
Target price $540\times1.12^{10}\approx$ ₹1,677.16 in 10 years.
EPS needed At a P/E of 25 that price needs EPS $=1{,}677.16/25\approx$ ₹67.09, so $$(1+g)^{10}=\frac{67.09}{12}\approx5.59,\qquad g\approx18.8\%\ \text{a year}.$$
Interpret Ten years of almost 19% EPS growth — more than five-fold — just to earn 12% a year, because the P/E is assumed to fall from 45 to 25 along the way. If EPS grows "only" the expected 15%, she earns less than 12%. Theorem 5.9 below turns this into a general formula.

Book Value, Dividends and Enterprise Value

The P/E is not the only multiple, and the others are tied to it by simple identities — useful for checking a screener and for seeing which business driver each multiple reflects.

📐 Theorem 5.7 (Formula) — Ratio identities and the justified P/B

With price $P$, earnings $E$, dividend $D$ and book value $B$ per share for the same year, and $\text{ROE}=E/B$:

  1. $\dfrac{P}{B}=\text{ROE}\times\dfrac{P}{E}$, $\quad\dfrac{D}{P}=\dfrac{\text{payout}}{P/E}$, $\quad\dfrac{E}{P}=\dfrac{1}{P/E}$.
  2. In the constant-growth model with $\text{ROE}=E_1/B_0$ and $g\lt k$, the justified P/B is $$\frac{P_0}{B_0}=\frac{\text{ROE}-g}{k-g},$$ so a company deserves to trade above book value exactly when $\text{ROE}\gt k$, at book when $\text{ROE}=k$, and below book when $\text{ROE}\lt k$.

Proof. (1) $\frac PB=\frac PE\cdot\frac EB$ and $\frac DP=\frac DE\cdot\frac EP$. (2) By Theorem 5.6, $P_0=\frac{(1-g/\text{ROE})E_1}{k-g}$ and $E_1=\text{ROE}\cdot B_0$, so $P_0=\frac{(\text{ROE}-g)B_0}{k-g}$. Since $k-g\gt0$, $P_0/B_0\gt1$ $\iff$ $\text{ROE}-g\gt k-g$ $\iff$ $\text{ROE}\gt k$. ∎ This is why banks and financial companies, whose assets are mostly loans valued close to book, are usually compared on P/B against their ROE.

✏️ Example 4 — Two banks on P/B
Coromandel Bank (hypothetical) has a steady ROE of 16% and trades at 2.5 times book; Deccan Bank (hypothetical) has ROE 9% and trades at 0.8 times book. Take $k=12\%$ and long-run growth $g=7\%$ for both. (a) Find each bank's justified P/B. (b) What ROE does each market P/B imply? (c) Find each bank's P/E from the identity.
(a) Coromandel: $(0.16-0.07)/(0.12-0.07)=1.8$. Deccan: $(0.09-0.07)/0.05=0.4$.
(b) Solving $\text{P/B}=(\text{ROE}-g)/(k-g)$ for ROE: $\text{ROE}=g+\text{P/B}\,(k-g)$. Coromandel: $7\%+2.5\times5\%=19.5\%$; Deccan: $7\%+0.8\times5\%=11\%$.
(c) $\text{P/E}=\text{P/B}/\text{ROE}$: Coromandel $2.5/0.16\approx15.6$; Deccan $0.8/0.09\approx8.9$.
Interpret Deccan "trades below book" and looks cheap, but at a 9% ROE (below $k$) its justified P/B is only 0.4: the market is already assuming ROE improves to 11%. Coromandel looks expensive at 2.5 times book, but the market is assuming ROE rises to 19.5%. Both prices embed optimism; which is more realistic is a question about the banks, not about the ratio.

Looking past the capital structure. Net profit is what is left after interest, so P/E depends on how much a company borrows. To compare the businesses rather than their financing, we compare the value of the whole firm with a profit measured before interest.

📘 Definition 5.9 — Enterprise value, EBITDA and EV/EBITDA

In market terms, $\text{EV}=\text{market capitalisation}+\text{debt}-\text{cash}$ (plus preference capital and minority interests where present), the market's price for the whole business (compare Definition 5.4). EBITDA is earnings before interest, tax, depreciation and amortisation — operating profit before the cost of financing and before non-cash charges. The EV/EBITDA multiple is $\text{EV}/\text{EBITDA}$. Related multiples are EV/EBIT and EV/sales.

✏️ Example 5 — Same business, different debt
Firms A and B (hypothetical) run identical businesses: EBITDA ₹500 crore, D&A ₹100 crore, so EBIT ₹400 crore; tax 25%. A has no debt and no cash; B has borrowed ₹1,500 crore at 8%. The market values each business at ₹4,000 crore. Find each firm's market capitalisation, net profit, P/E and EV/EBITDA.
Equity A: market cap $=4{,}000$. B: $4{,}000-1{,}500=2{,}500$.
Profit A: $400\times0.75=300$. B: interest $=120$, so $(400-120)\times0.75=210$.
Multiples P/E: A $4{,}000/300\approx13.3$; B $2{,}500/210\approx11.9$. EV/EBITDA: both $4{,}000/500=8.0$.
Interpret B "looks cheaper" on P/E only because of its debt; EV/EBITDA correctly shows the businesses priced the same. B's shareholders are also taking more risk (their profit falls faster if EBIT falls), so a lower P/E for B is exactly what theory expects. Compare leveraged and unleveraged companies on EV multiples.

ROE and the DuPont Decomposition

Theorems 5.6 and 5.7 put ROE at the centre of valuation: it decides whether growth adds value and whether a company deserves to trade above book. But the same ROE can come from very different businesses. The DuPont decomposition splits it into three drivers by an identity so simple it is almost a joke — and so useful that every analyst uses it.

📘 Definition 5.10 — ROE, ROA and the DuPont factors

For one year: return on equity $\text{ROE}=\dfrac{\text{net profit}}{\text{equity}}$; return on assets $\text{ROA}=\dfrac{\text{net profit}}{\text{total assets}}$; net profit margin $=\dfrac{\text{net profit}}{\text{sales}}$; asset turnover $=\dfrac{\text{sales}}{\text{total assets}}$ (rupees of sales per rupee of assets); equity multiplier (financial leverage) $=\dfrac{\text{total assets}}{\text{equity}}=1+\dfrac{\text{liabilities}}{\text{equity}}$. (Analysts often use averages of opening and closing balances; we use year-end figures for simplicity.)

📐 Theorem 5.8 — The DuPont decomposition

If sales, total assets and equity are positive, $$\text{ROE}=\underbrace{\frac{\text{net profit}}{\text{sales}}}_{\text{margin}}\times\underbrace{\frac{\text{sales}}{\text{assets}}}_{\text{turnover}}\times\underbrace{\frac{\text{assets}}{\text{equity}}}_{\text{leverage}},\qquad \text{ROA}=\text{margin}\times\text{turnover}.$$ If also EBIT and profit before tax (PBT) are positive, the five-step form is $$\text{ROE}=\frac{\text{net profit}}{\text{PBT}}\times\frac{\text{PBT}}{\text{EBIT}}\times\frac{\text{EBIT}}{\text{sales}}\times\frac{\text{sales}}{\text{assets}}\times\frac{\text{assets}}{\text{equity}},$$ i.e. tax burden × interest burden × operating margin × turnover × leverage.

Proof. In each product every intermediate quantity appears once in a numerator and once in a denominator, so it cancels (a telescoping product), leaving net profit/equity. The positivity hypotheses ensure we never divide by zero and that each factor has its usual meaning. ∎ The identity is trivial; its value is diagnostic. The tax burden (fraction of PBT kept after tax) and interest burden (fraction of EBIT left after interest) show how much of the operating profit reaches shareholders.

✏️ Example 6 — Two companies with a 20% ROE
Company X (hypothetical) has net margin 10%, asset turnover 1.25 and equity multiplier 1.6; Company Y (hypothetical) has margin 4%, turnover 1.0 and multiplier 5. (a) Find each ROE and ROA. (b) X's tax burden is 0.75 and interest burden 0.90; find its operating (EBIT) margin. (c) Suppose a bad year halves each company's ROA. What happens to each ROE, and how much of a fall in asset values would wipe out each company's equity?
(a) X: ROA $=10\%\times1.25=12.5\%$, ROE $=12.5\%\times1.6=20\%$. Y: ROA $=4\%\times1.0=4\%$, ROE $=4\%\times5=20\%$.
(b) Net margin $=0.75\times0.90\times\text{EBIT margin}$, so EBIT margin $=0.10/0.675\approx14.8\%$.
(c) Halved ROA: X's ROE falls to 10%, Y's to 10% too — the same percentage change. But leverage measures fragility: X's equity is $1/1.6=62.5\%$ of its assets, Y's only $1/5=20\%$. A 20% fall in asset values wipes out Y's equity entirely; X would need a 62.5% fall. (Liabilities/equity: X 0.6, Y 4.)
Lesson Equal ROEs do not mean equal businesses. X earns its ROE from operations; Y borrows its way to it. Theorem 5.6 says a high ROE justifies a high P/E — but the cost of equity $k$ must rise with leverage, which cancels part of the benefit.
Interactive 1 — The DuPont Map: Margin, Turnover and Leverage

Each point on the map is a business: asset turnover across, net margin up. The thin curves are iso-ROA hyperbolas margin × turnover = constant; the labels show the ROE each curve gives at the current leverage (ROE = ROA × leverage). Choose a hypothetical company, or move the sliders; press Play to tour the companies and watch the same ROE arise from different mixes. (Illustrative figures.)

12.0%1.671.50
ROA—
equity multiplier—
ROE—
liabilities / equity—
asset fall that wipes out equity—
justified P/B (k 12%, g 6%)—

Relative Valuation, Scenarios and Why Cheap Can Stay Cheap

Most valuations in practice are relative: instead of a full DCF, an analyst asks what similar companies trade at and applies that multiple. It is quick and uses the market's own information, but it inherits any mispricing of the peer group, and it is only as good as the similarity of the peers.

🧭 Relative valuation in five steps
  1. Choose peers with similar business, growth, risk and size — not just the same industry label.
  2. Choose multiples that fit: P/E for stable profitable companies; P/B for banks and financials; EV/EBITDA or EV/EBIT when leverage differs or to compare before depreciation policies; dividend yield for utilities.
  3. Compute the peers' multiples on the same basis (all trailing or all forward) and summarise with the median, which one outlier cannot drag.
  4. Apply to the target: price $=$ multiple × EPS (or BVPS); for EV multiples, value per share $=(\text{multiple}\times\text{EBITDA}-\text{net debt})/\text{shares}$.
  5. Adjust and explain any premium or discount by fundamentals — ROE, growth, leverage (Theorems 5.6–5.8) — and report a range, not a point.
✏️ Example 7 — Kaveri Home Products against its peers
Kaveri (hypothetical): EPS ₹9, EBITDA ₹420 crore, debt ₹900 crore, cash ₹50 crore, 20 crore shares, price ₹162. Five peers trade at P/Es of 22, 25, 18, 30 and 16, and EV/EBITDA of 12, 14, 10, 16 and 11; their median ROE is about 20%, against Kaveri's 12%. Value Kaveri on each multiple and decide whether its discount to peers is a bargain.
P/E Median of $\{16,18,22,25,30\}$ is 22, giving $22\times9=$ ₹198; the peer range gives ₹144–₹270.
EV/EBITDA Median 12: EV $=12\times420=5{,}040$; equity $=5{,}040-850=4{,}190$; per share $\approx$ ₹209.50.
Price check At ₹162 Kaveri's own multiples are P/E 18 and EV/EBITDA $(3{,}240+850)/420\approx9.7$ — below the peer medians, so "relatively cheap" by 18–23%.
Decide Its ROE is 12% against the peers' 20%. With $k$ around 12%, Theorem 5.7 says a company earning $\text{ROE}\approx k$ deserves a P/B near 1 and (Theorem 5.6) a forward P/E near $1/k\approx8.3$; Kaveri trades at P/B 2.16. The discount to peers is explained — arguably more than explained — by its lower ROE. Relative valuation says "cheaper than peers"; fundamentals say "not obviously cheap".

Where returns come from. A share's price is its P/E times its EPS. Over $n$ years, therefore, the price can rise because earnings grow, or because the market is willing to pay a higher multiple — and fall because the multiple shrinks.

📐 Theorem 5.9 (Formula) — Decomposing a share's return

If EPS grows from $E_0$ to $E_n=E_0(1+g)^n$ and the P/E moves from $m_0$ to $m_n$, the price changes by the factor $$\frac{P_n}{P_0}=(1+g)^n\cdot\frac{m_n}{m_0},$$ so the annual price return is $\left(1+g\right)\left(\dfrac{m_n}{m_0}\right)^{1/n}-1$; adding the average dividend yield gives an approximate total return. To earn a price return $r$ a year, EPS must grow at $$g=\frac{1+r}{(m_n/m_0)^{1/n}}-1.$$

Proof. $P_t=m_tE_t$ at every date, so $P_n/P_0=(m_n/m_0)(E_n/E_0)$; take the $n$-th root, and solve $(1+r)^n=(1+g)^n m_n/m_0$ for $g$. ∎ This is the arithmetic behind Example 3, and behind one of the most important warnings in investing.

📘 Definition 5.11 — Re-rating, de-rating and the value trap

A rise in a stock's multiple ($m_n\gt m_0$) is a re-rating; a fall is a de-rating. A value trap is a stock whose low multiple reflects genuinely poor prospects — low ROE, shrinking earnings, or high risk — so that it stays cheap, or gets cheaper, instead of re-rating.

✏️ Example 8 — Why cheap can stay cheap
Stock L (hypothetical) trades at a P/E of 8 with a 5% dividend yield; its EPS shrinks 4% a year and in 5 years the market pays a P/E of 6. Stock Q (hypothetical) trades at a P/E of 30 with a 1% dividend yield; its EPS grows 15% a year and its P/E falls to 24. Find each stock's price factor and annual return over 5 years.
L $0.96^5\times\frac68\approx0.8154\times0.75\approx0.612$: the price falls about 39%; annual price return $\approx-9.4\%$; with the 5% dividend yield, roughly $-4.4\%$ a year.
Q $1.15^5\times\frac{24}{30}\approx2.0114\times0.8\approx1.609$; annual price return $\approx+10.0\%$; with dividends roughly $+11\%$ a year — despite a de-rating.
Lesson The "cheap" stock lost money and the "expensive" one made money, because over five years earnings growth (or shrinkage) mattered more than the starting multiple. Of course the opposite can happen: if Q's growth disappoints, its de-rating can be brutal. The multiple is the start of the analysis, not the end.
Interactive 2 — The Sensitivity Table: Growth vs the Exit Multiple

Each cell is the annual price return (Theorem 5.9) from buying at the entry P/E you choose, if EPS grows at the row's rate and the market pays the column's P/E at the end. Green: at least your required return; amber: positive but below it; red: a loss. The dashed line marks the entry P/E (no re-rating). Press Play to sweep the entry P/E from 10 to 60. (Price return only, dividends excluded; illustrative.)

4512%
growth needed, exit P/E 20—
growth needed, P/E halves—
return: 12% growth, exit P/E 20—
cells that lose money—
⚠️ Common mistakes and myths
  1. "Low P/E means cheap." A low P/E may reflect low ROE, shrinking earnings or high risk (Theorem 5.6, Example 8). Cheap can stay cheap.
  2. "High P/E means expensive." A high P/E can be justified by a long runway of profitable growth (ROE well above $k$). The question is whether the growth the price requires (Theorem 5.9) is plausible.
  3. "Growth always deserves a higher multiple." Only when ROE exceeds the cost of equity. At ROE $=k$ the justified P/E is $1/k$ at any growth rate; below $k$, growth lowers it.
  4. Treating "PEG below 1" as a law. PEG ignores ROE, risk, the length of the growth period and interest rates. Two companies with PEG 1 can deserve very different P/Es (Exercise II.6).
  5. Comparing P/Es across different capital structures. Leverage lowers P/E mechanically (Example 5). Use EV/EBITDA or EV/EBIT.
  6. "EV/EBITDA captures everything." EBITDA ignores capex: a capital-hungry company and an asset-light one with the same EBITDA are not worth the same.
  7. Admiring ROE without DuPont. A high ROE built on leverage is fragile (Example 6). Look at ROA and the equity multiplier too.
  8. Mixing trailing and forward numbers, consolidated and standalone profits, or one-off gains, when comparing peers. Use one basis and strip one-offs.
  9. Averaging peer multiples with an outlier. One P/E of 60 among five peers lifts the mean far above the median (Exercise IV.1). Use the median and a range.
  10. "The whole sector is cheap, so this stock is cheap." Relative valuation can only say "cheaper than its peers"; if the peers are mispriced, so is the answer.
📈 Market case — Three consumer companies on one table

Priya puts the three hypothetical companies side by side (₹ crore unless stated). Market capitalisations: Narmada $540\times40=21{,}600$; Godavari $252\times30=7{,}560$; Kaveri $162\times20=3{,}240$.

metricNarmadaGodavariKaveri
EPS (₹)12.009.009.00
P/E45.028.018.0
expected EPS growth13%16%6%
PEG3.461.753.00
P/B13.55.042.16
EV/EBITDA30.612.49.7
dividend yield1.56%0.71%2.78%
net margin × turnover × leverage12% × 1.67 × 1.53% × 3.0 × 2.06% × 1.0 × 2.0
ROE30%18%12%
sustainable growth $b\times$ROE9.0%14.4%6.0%

The inputs behind the table: Narmada — revenue 4,000, net profit 480, assets 2,400, equity 1,600, payout 70%, EBITDA 700, no debt, cash 200. Godavari — revenue 9,000, profit 270, assets 3,000, equity 1,500, payout 20%, EBITDA 650, debt 600, cash 100. Kaveri — revenue 3,000, profit 180, assets 3,000, equity 1,500, payout 50%, EBITDA 420, debt 900, cash 50. (Check with Theorem 5.7: Narmada's P/B $=0.30\times45=13.5$ ✓, Kaveri's dividend yield $=0.5/18\approx2.78\%$ ✓.)

  1. Read the DuPont line. Narmada earns its 30% ROE from margins with little leverage; Godavari from turnover (a retailer: thin margins, fast-moving stock); Kaveri's 12% ROE is about equal to a 12% cost of equity, so by Theorems 5.6–5.7 its growth adds little value and it deserves a P/B near 1, not 2.16.
  2. Scenario table. For each company she writes down a bear, base and bull case for 5-year EPS growth and the exit P/E (her own assumptions), with probabilities 0.25, 0.5, 0.25, and applies Theorem 5.9:
    companybearbasebullexpected
    Narmada (P/E 45)8%, P/E 30: −0.4%13%, P/E 38: 9.2%18%, P/E 45: 18.0%9.8%
    Godavari (P/E 28)8%, P/E 18: −1.1%16%, P/E 25: 13.4%22%, P/E 30: 23.7%13.7%
    Kaveri (P/E 18)0%, P/E 12: −7.8%6%, P/E 16: 3.5%10%, P/E 20: 12.3%3.9%
    Each cell is an annual price return; "expected" is the probability-weighted 5-year price factor ($1.595$, $1.898$, $1.209$) turned into an annual rate. Adding dividend yields gives roughly 11.3%, 14.4% and 6.6% a year.
  3. What the prices require. To earn 12% a year in price over 5 years with an exit P/E of 30, Narmada's EPS must grow $1.12/(30/45)^{0.2}-1\approx21.5\%$ a year — far above its sustainable 9% unless ROE or payout changes. Godavari with an exit P/E of 20 needs about 19.8%; Kaveri with an exit P/E of 15 needs about 16.2%.

What she learns. Under her assumptions the lowest-P/E company has the worst outcomes, the middle one the best, and the most expensive one needs everything to go right. Change the assumptions and the ranking can change — that is why she writes them down. The ratios did not answer "which is cheap?"; they told her which questions to ask: how long can Narmada's 30% ROE last, can Godavari keep turning its stock over three times a year, and why does Kaveri trade above book with ROE equal to its cost of equity? None of this is a recommendation; it is a method.

Exercises

All companies are hypothetical and all inputs illustrative. Amounts in ₹ crore unless stated.

I
EPS and P/E.
  • 1 A company earned net profit 250, paid preference dividends of 10 and has 24 crore equity shares. Find its EPS.
  • 2 The share in part 1 trades at ₹180. Find its P/E and earnings yield.
  • 3 The company issues a 1:1 bonus (one new share for each share held, §2.1). What happens to its EPS, share price and P/E, other things equal?
  • 4 Before the bonus, employee options could add 1 crore shares. Find the diluted EPS.
  • 5 Next year's EPS is expected to be 20% higher than ₹10. Find the forward P/E at ₹180.
  • 6 Find the justified forward P/E of a company with ROE 18%, growth 9% and cost of equity 12%.
  • 7 Why is the P/E of a loss-making company meaningless, and which multiples could be used instead?
1 $\text{EPS}=(250-10)/24=$ ₹10.
2 $\text{P/E}=180/10=18$; earnings yield $=1/18\approx5.56\%$.
3 Shares double to 48 crore, so EPS halves to ₹5 and the price adjusts to about ₹90; the P/E stays 18. A bonus changes the number of slices, not the company.
4 $240/25=$ ₹9.60.
5 Forward EPS ₹12; forward P/E $=180/12=15$.
6 Payout $=1-0.09/0.18=0.5$; P/E $=0.5/(0.12-0.09)\approx16.67$.
7 With $E\le0$ the ratio is negative or infinite and says nothing about value. Use EV/sales, EV/EBITDA (if EBITDA is positive), P/B, or a DCF of the expected future free cash flows.
II
PEG, P/B and dividend yield.
  • 1 A share has a P/E of 36 and expected EPS growth of 18%. Find its PEG.
  • 2 A share trades at ₹450 with book value per share ₹150 and ROE 20%. Find its P/B and, from Theorem 5.7, its P/E.
  • 3 Find the justified P/B of a company with ROE 15%, cost of equity 12% and growth 6%.
  • 4 A share pays ₹6 a year and trades at ₹240; its EPS is ₹15. Find its dividend yield and payout ratio, and verify $D/P=\text{payout}/(P/E)$.
  • 5 A bank with ROE 10% and cost of equity 13% trades at 0.7 times book. With long-run growth 5%, is it cheap?
  • 6 Two companies each have P/E 10 and expected growth 10% for ever (PEG 1); their ROEs are 30% and 11%, and $k=13\%$. Find their justified forward P/Es and comment on the PEG.
1 $\text{PEG}=36/18=2$.
2 $\text{P/B}=450/150=3$; $\text{P/E}=\text{P/B}/\text{ROE}=3/0.20=15$ (EPS ₹30).
3 $(0.15-0.06)/(0.12-0.06)=1.5$.
4 Yield $6/240=2.5\%$; payout $6/15=40\%$; P/E $=16$ and $0.40/16=2.5\%$ ✓.
5 Justified P/B $=(0.10-0.05)/(0.13-0.05)=0.625$. At 0.7 it is slightly above its justified value: trading below book is not a bargain when ROE is below the cost of equity.
6 ROE 30%: payout $1-0.10/0.30=2/3$, P/E $=0.6667/0.03\approx22.2$. ROE 11%: payout $1-0.10/0.11\approx0.091$, P/E $\approx0.091/0.03\approx3.0$. Same PEG, justified P/Es seven times apart: PEG ignores the profitability of the growth.
III
EV/EBITDA and DuPont.
  • 1 A company has market capitalisation 6,000, debt 1,500, cash 500 and EBITDA 875. Find its EV and EV/EBITDA.
  • 2 Peers trade at 10 times EBITDA. A company has EBITDA 300, net debt 600 and 12 crore shares. Find its implied value per share.
  • 3 Net margin 8%, asset turnover 1.5, equity multiplier 2. Find ROA and ROE.
  • 4 Sales 1,200, net profit 96, total assets 800, equity 400. Find the three DuPont factors and check that their product is the ROE.
  • 5 A company with ROA 8% raises its equity multiplier from 1.5 to 3 by borrowing to buy back shares, keeping ROA unchanged. What happens to ROE? What is the catch?
  • 6 Tax burden 0.75, interest burden 0.8, EBIT margin 15%, turnover 1.2, multiplier 2.5. Find the ROE.
  • 7 Two companies have EBITDA of 100 each and trade at the same EV/EBITDA. One must spend 70 a year on capex to stay in business, the other 10. Why might the same multiple be wrong?
1 $\text{EV}=6{,}000+1{,}500-500=7{,}000$; $7{,}000/875=8.0$.
2 EV $=3{,}000$; equity $=2{,}400$; per share $2{,}400/12=$ ₹200.
3 ROA $=8\%\times1.5=12\%$; ROE $=12\%\times2=24\%$.
4 Margin $96/1{,}200=8\%$; turnover $1{,}200/800=1.5$; multiplier $800/400=2$; product $24\%=96/400$ ✓.
5 ROE doubles from 12% to 24% (in practice ROA also falls a little, because interest reduces net profit). The catch: liabilities/equity rises from 0.5 to 2, a 33% fall in asset values (instead of 67%) now wipes out equity, and shareholders' required return $k$ rises. Higher ROE from leverage is not free value.
6 $0.75\times0.8\times0.15\times1.2\times2.5=0.27$, i.e. 27%.
7 Free cash flow (roughly EBITDA − tax − capex) is about 30 for the first and 90 for the second, ignoring tax. EBITDA treats both as equally profitable; owners receive cash after capex. Use EV/EBIT, EV/FCFF or a DCF for capital-intensive businesses.
IV
Relative valuation, scenarios and return decomposition.
  • 1 Peer P/Es are 14, 19, 22, 25 and 60. Find their mean and median, and the value of a target with EPS ₹8 on each. Which would you report?
  • 2 EPS grows 12% a year for 5 years while the P/E falls from 30 to 20. Find the annual price return.
  • 3 A share trades at a P/E of 50. If the P/E will be 30 in 5 years, what EPS growth gives a 12% annual price return?
  • 4 Bear, base and bull values of ₹150, ₹220 and ₹300 have probabilities 0.3, 0.5 and 0.2; the price is ₹200. Find the expected value, the expected gain and the probability of a loss (if the value is eventually realised).
  • 5 A share's earnings yield is 4% while a G-sec yields an illustrative 7%. Explain, using the Gordon model, how buying the share can still be reasonable, and what must be true.
  • 6 A "cheap" share at a P/E of 6 sees EPS fall 10% a year for 3 years while its P/E stays at 6. Find the price change, and roughly the total return if it pays a 6% dividend yield.
1 Mean $140/5=28$, median 22; values $28\times8=$ ₹224 and $22\times8=$ ₹176. Report the median (and the range ₹112–₹200 without the outlier): the single P/E of 60 lifts the mean by 6 points.
2 $1.12\times(20/30)^{1/5}-1$ $\approx1.12\times0.9221-1$ $\approx3.28\%$ a year.
3 $g=1.12/(30/50)^{1/5}-1$ $=1.12/0.9029-1$ $\approx24.0\%$ a year.
4 $0.3\times150+0.5\times220+0.2\times300=$ ₹215; expected gain $215/200-1=7.5\%$; a loss occurs only in the bear case, probability 0.3.
5 By Theorem 5.3, $k_e=D_1/P_0+g$. A low current yield can still give an adequate $k_e$ if growth is high and profitable: with payout 50% the dividend yield is 2%, so a 12% return needs about 10% growth for ever, which (Theorem 5.3) needs ROE about 20% on retained profits. The bond's 7% is fixed; the share's return depends on that growth materialising.
6 Price factor $0.9^3=0.729$: a 27.1% fall. Dividends at roughly 6% a year add about 18 points in total, so the investor still loses roughly 9% — a value trap.
🎯 Practice corner NISM-styleCase studyInterview

E1. NISM-style A company earns a net profit of ₹120 crore and has 15 crore shares, which trade at ₹200 each. Its P/E ratio is (A) 25 (B) 8 (C) 0.04 (D) 15.

Compute EPS $=120/15=$ ₹8; P/E $=200/8=25$. (B) is the EPS; (C) is the earnings yield; (D) is the number of shares.
Answer (A).

E2. NISM-style A company has a net profit margin of 5%, asset turnover of 2 and an equity multiplier of 2.5. Its return on equity is (A) 9.5% (B) 25% (C) 10% (D) 12.5%.

DuPont $5\%\times2\times2.5=25\%$. (A) adds the three numbers; (C) is the ROA; (D) multiplies margin by leverage and forgets turnover.
Answer (B).

E3. NISM-style A company has a market capitalisation of ₹800 crore, debt of ₹300 crore, cash of ₹100 crore and EBITDA of ₹125 crore. Its EV/EBITDA is (A) 6.4 (B) 8.0 (C) 9.6 (D) 7.2.

Compute $\text{EV}=800+300-100=1{,}000$; $1{,}000/125=8.0$. (A) uses market cap alone; (C) adds the cash instead of subtracting it ($1{,}200/125$); (D) adds the cash and ignores the debt ($900/125$).
Answer (B).

E4. Case study Two hypothetical banks: Bank A has ROA 1.5% and assets 11 times its equity, and trades at 2.6 times book; Bank B has ROA 0.9% and assets 12 times equity, and trades at 0.9 times book. Take $k=12.5\%$ and $g=7\%$ for both. (a) Find each ROE. (b) Find each justified P/B and the ratio of market to justified P/B. (c) Find the ROE each market price implies. (d) Find each P/E. Which bank's price asks more of the future?

(a) ROE $=$ ROA × leverage: A $1.5\%\times11=16.5\%$; B $0.9\%\times12=10.8\%$.
(b) A: $(0.165-0.07)/0.055\approx1.73$, market/justified $\approx2.6/1.73\approx1.51$. B: $(0.108-0.07)/0.055\approx0.69$, ratio $\approx0.9/0.69\approx1.30$.
(c) $\text{ROE}=g+\text{P/B}(k-g)$: A $7\%+2.6\times5.5\%=21.3\%$; B $7\%+0.9\times5.5\%\approx11.95\%$.
(d) P/E $=$ P/B ÷ ROE: A $2.6/0.165\approx15.8$; B $0.9/0.108\approx8.3$.
Decide Both prices exceed the justified P/B, so both assume improvement. A's price asks more: ROE must rise by about 4.8 points (16.5% → 21.3%), against about 1.2 points for B (10.8% → 11.95%). Note how leverage of 11–12 times turns a modest ROA into a double-digit ROE — banks are leveraged by design, which is why their $k$ is not low.

E5. Case study Sutlej Auto Parts Ltd (hypothetical): EPS ₹20, price ₹360, EBITDA ₹400 crore, net debt ₹300 crore, 10 crore shares. Peers trade at P/Es of 15, 18, 20, 24 and EV/EBITDA of 8, 9, 10, 12. (a) Find Sutlej's P/E and EV/EBITDA. (b) Value it on the peer medians. (c) Over 5 years: bear case EPS growth 5% and exit P/E 12; base 12% and 16; bull 18% and 20, with probabilities 0.25, 0.5, 0.25. Find the price in each case, the annual return, and the expected price. (d) Summarise.

(a) P/E $=360/20=18$; EV $=3{,}600+300=3{,}900$, EV/EBITDA $=9.75$.
(b) Median P/E 19 → ₹380 (range ₹300–₹480). Median EV/EBITDA 9.5 → $(3{,}800-300)/10=$ ₹350 (range ₹290–₹450). Central range about ₹350–₹380, a few percent around the price.
(c) Prices $P_5=20(1+g)^5m_5$: bear $20\times1.05^5\times12\approx$ ₹306.3 ($-3.2\%$ a year); base $20\times1.12^5\times16\approx$ ₹563.9 ($+9.4\%$); bull $20\times1.18^5\times20\approx$ ₹915.1 ($+20.5\%$). Expected price $\approx0.25\times306.3$ $+\,0.5\times563.9$ $+\,0.25\times915.1\approx$ ₹587.3, an annualised $\approx10.3\%$.
(d) Relative valuation says "roughly fairly priced against peers"; the scenarios say the outcome depends mainly on EPS growth, with a one-in-four chance (under these assumptions) of a loss. Both are conditional on the analyst's inputs, which should be stated.

E6. Interview "A company earns an ROE exactly equal to its 12% cost of equity. Management announces it will retain more profit so that earnings grow faster. What happens to its fair P/E?"

Idea Theorem 5.6: forward P/E $=\frac{1-g/\text{ROE}}{k-g}$, and with ROE $=k$ this is $\frac{(k-g)/k}{k-g}=\frac1k$ for every $g$.
Answer Nothing: the fair forward P/E stays $1/0.12\approx8.33$. Each retained rupee earns 12%, exactly what shareholders could earn elsewhere, so faster growth is exactly offset by lower dividends today.
Follow-up If ROE were 10%, retaining more would lower the P/E; at 15% it would raise it. A strong interview answer ends with "growth creates value only when the return on the reinvested capital exceeds its cost".
🧑‍🏫 For the teacher

Suggested plan — 2 lectures × 60 min.

  1. Multiples and what they assume. Open with Priya's screener and take a class vote: which of the three is cheapest? EPS and P/E (Definition 5.7, Example 1). Derive the justified P/E from Gordon (Theorem 5.6) and do Example 2 — the "ROE = k ⇒ P/E = 1/k" result surprises almost everyone. PEG and EPS compounding (Example 3). Identities and justified P/B (Theorem 5.7) with the two banks (Example 4); EV/EBITDA and leverage (Example 5).
  2. Drivers, peers and scenarios. DuPont (Theorem 5.8, Example 6). Project Interactive 1, tour the presets, and ask why Godavari and Company Y need very different leverage for similar ROE. Relative valuation (strategy box, Example 7), return decomposition (Theorem 5.9) and the value trap (Example 8). Project Interactive 2 at an entry P/E of 45, then 18. Finish with the 📈 Market case and redo the vote.

Class activity. Give groups the same three companies but different scenario assumptions (optimist, pessimist, realist). Each group fills in the scenario table and announces its ranking. Discuss why the rankings differ and what evidence would settle the disagreement.

Misconceptions to watch for. "Low P/E = cheap"; PEG < 1 as a rule; comparing P/Es across leverage; EBITDA as cash; ROE without DuPont; averaging peer multiples with outliers; treating ratios as recommendations.

Suggested homework. I(1, 3, 6, 7), II(2, 3, 5, 6), III(1, 2, 4, 5, 7), IV(1, 3, 4, 6); Practice corner E2, E4, E6.

🛠️ Do it yourselfSpreadsheet mini-project · about 60 minutes

Project. Build a peer-comparison sheet that computes every ratio of this section from raw financial-statement numbers, decomposes ROE, values a target on peer medians, and produces a scenario table — first for the three companies of the 📈 Market case, then for any five listed companies of one sector, as a learning exercise.

Why it matters. Screeners show ratios without showing their inputs or definitions. Rebuilding them from the annual report teaches you exactly what each number means — and catches inconsistent definitions.

What to do.

  1. One column per company; rows for price, shares, net profit, preference dividends, DPS, revenue, EBITDA, EBIT, PBT, total assets, equity, debt and cash.
  2. Ratios: EPS =(profit-pref)/shares; P/E =price/EPS; P/B =price/(equity/shares); dividend yield =DPS/price; EV =price*shares+debt-cash; EV/EBITDA =EV/EBITDA; ROE =profit/equity.
  3. DuPont rows: =profit/revenue, =revenue/assets, =assets/equity and a check row =product-ROE that must be 0. Add the five-step version with =profit/PBT and =PBT/EBIT.
  4. Peer summary: =MEDIAN(range), =MIN(range), =MAX(range) for each multiple; implied target values =median_PE*EPS and =(median_EV_EBITDA*EBITDA-net_debt)/shares.
  5. Justified multiples: payout =DPS/EPS, sustainable growth =(1-payout)*ROE, justified P/B =(ROE-g)/(k-g) with your own $k$ (from CAPM, §4.2).
  6. Scenario/sensitivity table: EPS growth down a column, exit P/E across a row, and in each cell =(1+$A5)*(B$4/entry_PE)^(1/n)-1; add conditional formatting (red below 0, green above your required return). You have rebuilt Interactive 2.
  7. For real companies, take the numbers from annual reports or the results filed on the NSE/BSE websites; use consolidated figures throughout and note any one-off items you remove.

Python version: import pandas as pd; df["PE"]=df.price/df.eps; df["ROE"]=df.profit/df.equity; df["dupont"]=(df.profit/df.revenue)*(df.revenue/df.assets)*(df.assets/df.equity); df[["PE","PB","EV_EBITDA"]].median().

Hand in: the sheet, a one-paragraph DuPont comparison of the companies, the implied value range for one target, and the scenario table with your stated assumptions.

Going further. Plot each company's P/B against its ROE and draw the justified line $(\text{ROE}-g)/(k-g)$; companies far above the line need the most growth to justify their price. Repeat with five years of data to see whether ROEs are stable.

Instructor note: for the Market case the sheet must give P/E 45 / 28 / 18, P/B 13.5 / 5.04 / 2.16, EV/EBITDA 30.6 / 12.4 / 9.74, ROE 30% / 18% / 12% with DuPont check rows equal to zero, and sustainable growth 9% / 14.4% / 6%. The commonest errors are using book debt but market equity inconsistently, and adding cash to EV instead of subtracting it.

🧪 Self-check quiz

0/5 correct

Q1. A share has ROE 20% and P/E 15 (same-year earnings and book value). Its P/B is

Q2. Two companies have ROE 18%. Company P has net margin 9%, turnover 1.0, multiplier 2; Company Q has margin 3%, turnover 1.0, multiplier 6. Which statement is right?

Q3. A share has a P/E of 30 and expected EPS growth of 20% a year. Its PEG is

Q4. Why is EV/EBITDA preferred to P/E when comparing two companies with very different debt?

Q5. In the constant-growth model, a company whose ROE is below its cost of equity should trade at a P/B that is