Fundamental Analysis: The Ratios That Matter (And How They Lie)

Direct Equity — Arc 2, Article Two

The Right Order

Business first, price last

Forget the stock market for a moment. Imagine a friend offers to sell you his kirana shop. What do you ask, and in what order?

You ask what the shop earns in a year. You ask how much of that is real profit after rent, staff and electricity. You ask whether he has borrowed against it and what the interest eats. You ask how much stock sits unsold on the shelves and how many customers buy on credit and pay in ninety days. And only after all of that do you ask his price — because until you know what you are buying, no price can be judged cheap or expensive.

That is the entire logic of ratio analysis, and beginners reverse it. They start with the P/E ratio because it is on the front screen of every app. The professional sequence runs the other way: establish what quality of business you are looking at, then ask what that quality should cost. A low P/E on a deteriorating business is not cheap — it is the market’s verdict arriving before yours.

Order	The question	The ratios that answer it
First	Is this a good business?	ROE, ROCE, operating and net margins
Second	Is it safely financed?	Debt-to-equity, interest coverage
Third	Is it efficiently run?	Working-capital days, cash conversion cycle, asset turnover
Last	What does it cost?	P/E, P/B, EV/EBITDA, dividend yield

Question One — Quality

ROE and ROCE meaning, in plain language

ROE — Return on Equity. In one line: what the owner earns on his own money. Your friend put ₹10 lakh of his savings into the shop and it earns him ₹2 lakh a year — that is a 20% return on equity. The formula is profit after tax divided by shareholders’ equity. Persistently high ROE, and the folk threshold in India is around 15% sustained across years, is the statistical signature of a business with some advantage — a brand, a network, a cost position — because ordinary businesses get competed down toward ordinary returns. That is not a rule someone invented; it is what competition does.

How ROE lies: debt inflates it. Suppose your friend put in only ₹5 lakh of his own money and borrowed the other ₹5 lakh. The same ₹2 lakh of profit — less the interest — now sits on a much smaller base of his own capital, and his “return on equity” leaps. Nothing about the shop improved. He simply used someone else’s money, and he will keep looking brilliant right up until a bad year arrives and the interest still has to be paid.

The antidote is a decomposition worth memorising, because it turns one number into a story. ROE = net margin × asset turnover × leverage. How much profit per rupee of sales, multiplied by how many rupees of sales per rupee of assets, multiplied by how many rupees of assets per rupee of owner’s capital. Two companies can report exactly the same ROE and be completely different animals:

ROCE — Return on Capital Employed. In one line: what the business earns on every rupee put to work in it, no matter who lent it. The formula is operating profit divided by equity plus debt. This is the leverage-proof sibling of ROE, and it is the reason the trick above stops working. Your friend’s shop earns ₹2 lakh on ₹20 lakh of total capital — 10% — whether he funded it with his own money or the bank’s.

The pairing to hunt for is high ROE and high ROCE together, both stable across years. High ROE with mediocre ROCE is leverage wearing a costume. And there is one more idea inside ROCE worth carrying permanently: a company that consistently earns ROCE well above its cost of borrowing creates value with every rupee it reinvests, while a company earning below its borrowing cost destroys value by growing. Growth is not automatically good. Growth at a poor return is expensive theatre.

Margins. In one line: how much of each ₹100 of sales survives to the bottom. There are four worth knowing, and they are best read as a ladder — each rung stripping away one more layer of cost. Take Bharat Widgets Ltd from the previous article — ₹1,000 crore of revenue, ₹820 crore of total expenses, ₹210 crore of EBITDA and ₹82 crore of net profit — and assume ₹560 crore of those expenses were raw materials:

Margin	How it is calculated	Bharat Widgets	What it tells you
Gross	(Revenue − cost of materials) ÷ revenue	44%	Pricing power against input costs — can it charge more than it pays?
EBITDA	EBITDA ÷ revenue	21%	Cash operating efficiency — what survives salaries, power, freight and overheads, before the asset base is charged for
EBIT (operating margin proper)	(EBITDA − depreciation) ÷ revenue	15%	True operating profitability, after the wear and tear on the assets that produced the sales
Net	Profit after tax ÷ revenue	8.2%	What finally reaches the owner, after lenders and the government

Reading down the ladder tells you where the money leaks, which is far more informative than any single margin. Here, 44% becomes 21% — twenty-three points consumed by employees, power, freight and administration. Then 21% becomes 15%, the six points being depreciation: the annual charge for wearing out the plant and machinery that produced those sales. Then 15% becomes 8.2%, taken by lenders and the government. Now watch those steps across five years. If the gross margin holds but the operating margin falls, the problem is overheads and the company has a cost-discipline story to explain. If both hold but the net margin falls, the problem is debt or tax, not the business. If the gross margin itself is sliding, the company is losing pricing power against its suppliers or its customers — the most serious of the three, because it is the hardest to reverse.

The gap between the two middle rungs deserves its own attention, because it is where a great deal of promotional accounting lives. EBITDA margin is the number managements quote — in presentations, in earnings calls, in press releases — precisely because it is the flattering one. It says nothing about the cost of the factory, the fleet or the fibre network that generated the sales, since depreciation has not yet been charged. For an asset-light services business the two margins sit close together and the distinction barely matters. For a cement plant, a telecom operator or an airline, the six or ten or twenty points between EBITDA and EBIT are a real recurring cost of staying in business, and a company that only ever discusses EBITDA is choosing which half of its economics you see. EBIT is also the number that matters mechanically: it is the numerator of ROCE and of interest coverage, both of which you met above. When a screener shows you “operating margin”, check which of the two it means — platforms differ, and the difference is not small.

One more idea makes margins click permanently: a high margin is not automatically better than a low one. A jeweller may earn 25% on every sale but turn his stock over twice a year. A kirana shop earns perhaps 5% but turns its stock over twelve times. Multiply margin by turnover and the two businesses can land in exactly the same place — which is precisely the DuPont decomposition above, seen from the other side. So the useful question is never “is this margin good?” but “is this margin normal for this kind of business, and what is it doing over time?” A supermarket at 4% net margin may be excellent; a software company at 4% is in trouble.

How margins lie, in four familiar ways. A mix shift can move the headline margin without anything improving or deteriorating — a company selling proportionally more of its premium product looks better on margin while its underlying businesses are unchanged. Commodity pass-through distorts percentages in both directions: when input prices spike and the company passes them on rupee for rupee, revenue rises, the margin percentage falls, and absolutely nothing has gone wrong. Capitalising costs — pushing expenses onto the balance sheet as assets rather than through the P&L — flatters operating margin today at the cost of depreciation tomorrow. And lease accounting under current Indian standards moves rent out of operating expenses and into depreciation and interest, which mechanically lifts EBITDA margin for retailers, hotels and airlines without any change in economics. Compare a company against its own history and its direct peers, and each of these distortions becomes visible rather than invisible.

Question Two — Safety

Will it survive a bad year?

Debt-to-equity. In one line: how many rupees the company has borrowed for every rupee its owners put in. Context and trajectory beat thresholds here. A debt-to-equity of 1.5 is unremarkable for a power utility with contracted cash flows and alarming for a consumer company that should not need it. Learn a sector’s normal range before judging any member of it.

Interest coverage. In one line: how many times over the year’s operating profit covers the year’s interest bill. This is the ratio that tells you whether a bad year is survivable. Coverage below about three times deserves attention in any sector; coverage falling for three years running deserves considerably more. And read the maturity footnote in the notes to accounts while you are there: ₹2,000 crore of debt spread over ten years and the same ₹2,000 crore due next March describe two completely different companies.

Net debt to EBITDA. In one line: how many years of operating earnings it would take to clear the borrowings. Take total debt, subtract cash, divide by EBITDA. Lenders and rating agencies reach for this before they reach for debt-to-equity, and the reason is sound: it measures debt against the company’s ability to service it rather than against an accounting figure that can be revalued. As a rough map, under one time is comfortable, two to three times is unremarkable for a capital-heavy business, and above four times the lenders are effectively running the company — every major decision needs their blessing. It lies in two ways. Because EBITDA sits above depreciation, an asset-heavy firm looks stronger on this measure than its cash reality supports. And it uses a single year’s EBITDA, so a cyclical business at the top of its cycle can show a reassuring 1.5x that becomes 6x the moment earnings normalise.

Current ratio and quick ratio. In one line: can it pay next year’s bills out of this year’s near-cash? Current assets divided by current liabilities; the quick ratio strips out inventory, on the reasoning that unsold stock is not money. These are the classic textbook liquidity checks, and they deserve a smaller place than textbooks give them. A ratio below one is not automatically alarming — the strongest consumer businesses in India run negative working capital by design, collecting from customers before paying suppliers, and that is a sign of power rather than distress. Treat these as question-raisers: a deteriorating current ratio in a company that is also stretching its payables is worth chasing down.

One special case is worth naming, because it splits opinion. A zero-debt company sitting on a large cash pile is either genuinely conservative or genuinely out of ideas. High ROCE, no debt and growing is elite. No debt, hoarded cash and a stagnant ROE is a piggy bank — and piggy banks deserve piggy-bank valuations, not compounder valuations.

Question Three — Efficiency

How hard does the money work?

Back to the shop. Stock sitting on the shelf is money you have already spent and not yet earned. Customers who buy on credit are money you have earned and not yet received. Suppliers who let you pay in sixty days are, in effect, lending you their goods for free. Put those three together and you have the cash conversion cycle: inventory days plus receivable days minus payable days, or how many days of cash are locked up in each loop of the business.

What you are watching for is movement. A cycle lengthening across years means weakening bargaining power — customers taking longer to pay, goods sitting longer, suppliers demanding cash sooner. It is the aggressive-accounting tell from the previous article expressed in days rather than rupees. A cycle wildly better than every peer is either a genuinely superior model that you should be able to explain in a sentence, or numbers to distrust.

Asset turnover. In one line: how many rupees of sales the company squeezes from each rupee of assets. Meaningful mainly within a sector and across time. A declining turnover during heavy capital expenditure is normal if the new capacity eventually produces revenue. “Capacity expansion” that never lifts turnover is the stuck capital-work-in-progress red flag wearing a ratio.

Question Four — Price

PE ratio, price to book and the rest — asked last

P/E — price to earnings. In one line: how many years of current profit you are paying up front. A P/E of 25 means you are handing over twenty-five rupees for each rupee the company earns annually — or, flipped around, an earnings yield of 4%. It is universal, useful, and the most abused number in investing, for three reasons.

First, the E is manipulable and cyclical. One-off gains, accounting choices, or a cyclical peak flatten the P/E artificially. A commodity company at “P/E 6” on peak-cycle earnings is frequently more expensive than it looks, because the E is about to halve. Cyclicals routinely look cheapest at tops and dearest at bottoms, which is the single most expensive lesson new investors buy. Second, growth and quality context is everything: a P/E of 40 for a business compounding profits at 25% with high ROCE can be more rational than a P/E of 12 for a shrinking one. Comparing P/Es across businesses of different quality is comparing prices without reading the labels. Third, sector norms differ persistently — IT, consumer staples and lenders occupy structurally different neighbourhoods, so cross-sector P/E comparisons mostly measure which sector you are in.

P/B — price to book. In one line: what you pay for each rupee of accounting net worth. Most meaningful where book value means something — lenders and asset-heavy businesses — and nearly meaningless for asset-light franchises whose real assets are brands, code and relationships that never appear on a balance sheet. The pairing that matters is P/B alongside ROE: a high-ROE business deserves a high P/B, because each rupee of book value earns a great deal. A P/B of 0.6 with an ROE of 4% is not a bargain; it is fairly priced mediocrity. “Below book value” alone has lured generations into value traps.

EV/EBITDA. In one line: what the whole business costs, debt included, against its operating cash generation. Enterprise value is market capitalisation plus debt minus cash, which makes this the capital-structure-neutral comparison — the right tool when comparing a heavily borrowed company against an unlevered peer, exactly where P/E misleads. Its blind spot: EBITDA ignores depreciation, which for asset-heavy businesses is a real recurring cost rather than an accounting fiction.

Free cash flow yield. In one line: the actual cash the business throws off each year, as a percentage of what it costs to buy. Free cash flow divided by market capitalisation — or, more rigorously, divided by enterprise value. This is the P/E’s more honest cousin, because the numerator is cash that arrived rather than profit that was reported, and it connects directly to the free cash flow you learned to derive in the previous article. A business trading at 25 times earnings while generating a 6% free cash flow yield is telling you something quite different from one at the same P/E generating 1%. How it lies: a single year of unusually low capital expenditure inflates it handsomely, and a genuinely good company investing heavily in growth may show a poor or negative yield for years by choice. As always, five years, not one.

Dividend yield. In one line: the cash the company hands you each year as a percentage of what you paid. Honest cash and a discipline signal — but very high yields are frequently distress signals, where the price collapsed faster than the payout will. India’s best compounders often pay little precisely because reinvesting at high ROCE beats distributing. Yield is a fact to explain, not a score to maximise. The companion number is the payout ratio — dividend divided by earnings per share — which tells you whether the dividend is comfortably covered or is being paid out of borrowings and goodwill. A payout ratio drifting above 100% is a dividend with a countdown on it.

Three more ratios appear constantly in commentary and screeners without earning a section of their own. Know what they mean and where they fail, and you can read anyone’s note without being impressed by vocabulary:

Ratio	In one line	Where it fails
PEG	P/E divided by the profit growth rate — an attempt to price growth	Entirely dependent on a growth forecast, and it treats all growth as equally valuable regardless of the ROCE it is earned at
Price-to-sales	Market value per rupee of revenue — used where there are no profits to divide by	Revenue without margins is not a business; it flatters loss-makers precisely when scepticism is most needed
ROA	Profit per rupee of total assets	Useful for lenders, weak elsewhere — ROCE answers the same question better by using operating profit and excluding non-operating assets

The Whole Mesh

A worked contrast, and the summary to keep

Two fictional companies, same sector, same ₹100 crore of annual profit:

Measure	Steady Consumer Ltd	Levered Infra Ltd
ROE	22%	19%
ROCE	24%	9%
Debt-to-equity	0.1	2.2
Interest coverage	18x	2.1x
Operating cash vs profit, 5-year	~105%	~55%
Receivable days, trend	Stable, around 30	45 rising to 78
P/E	34	9

The app’s front screen says Levered Infra is “cheap” at a quarter of the P/E. The full table says something else entirely: its ROE is a leverage artefact, since ROCE of 9% is probably below what it pays to borrow; its interest coverage leaves no room for one bad year; its profits are not converting into cash; and its customers are paying later every year. Whether Steady Consumer at a P/E of 34 is attractive is a further judgment this article does not make. But the ordering of the two on quality is not close, and that is the whole argument: ratios work as a mesh, never as single numbers.

If you remember nothing else from this article, remember this table:

Ratio	In one line	The lie it tells
ROE	What the owner earns on his own money	Debt inflates it — decompose it before believing it
ROCE	What the business earns on all money put to work	Hardest to flatter, but distorted by revalued or idle assets
Gross margin	What survives after paying for materials	Mix shifts and commodity pass-through move it for no real reason
EBITDA margin	What survives after running the business, before depreciation	The flattering one — ignores the cost of the assets and is lifted by lease accounting
EBIT margin	True operating profit per ₹100 of sales	Distorted when depreciation policies differ from peers
Net margin	What finally reaches the owner	A healthy level hides a falling trend; compare only within a sector
Debt-to-equity	Borrowed rupees per rupee of owner capital	Meaningless without the sector’s normal range
Interest coverage	How many times profit covers the interest bill	A single year hides the maturity wall next March
Net debt / EBITDA	Years of operating earnings needed to clear the debt	Sits above depreciation, and a peak-cycle EBITDA halves the apparent burden
Cash conversion cycle	Days of cash locked up per business loop	Suspiciously better than peers is a question, not a prize
P/E	Years of current profit paid up front	The E can be a one-off or a cyclical peak
P/B	Price per rupee of accounting net worth	Low P/B with low ROE is a trap, not a bargain
EV/EBITDA	Cost of the whole firm against operating earnings	Ignores depreciation, which asset-heavy firms really pay
Free cash flow yield	Cash generated per rupee of company value	One year of low capex flatters it; deliberate growth spending depresses it
Dividend yield	Annual cash returned per rupee invested	A very high yield often means the price collapsed

The Discipline Rules

Five rules that stop ratios from misleading you

1. Five years minimum, and always against peers. A single ratio is a word; the sentence is its trend and its peer context. Every serious platform shows both, and reading one number in isolation is how confident mistakes get made.
2. Quality ratios veto valuation ratios. No P/E is low enough to overrule collapsing interest coverage and cash conversion. Cheapness is not a defence against deterioration.
3. Check the inputs behind any surprising ratio. One-off income inside the E, revalued assets inside the B, peak-cycle EBITDA in the denominator. Surprising ratios are usually questions, not opportunities.
4. Sector first. Learn a sector’s normal ranges before judging any member against them — and treat lenders as a separate curriculum entirely, with different statements and different ratios.
5. Ratios screen; they do not decide. They tell you where to look. Our research process article covers what looking actually means. Buying on a screenshot of ratios is buying on a tip you generated yourself.

Key Takeaways

---

Frequently Asked Questions

Your questions answered

What is a good P/E ratio for the Indian market?

There isn’t one, and anyone offering a confident single number is guessing. Index-level P/Es have ranged widely across decades, and company-level P/Es are only interpretable against growth, ROCE and sector. The productive question is never “is 28 high?” but “what growth and return profile does 28 imply, and is this business likely to deliver it?”

Which single ratio would you check first?

The five-year ROCE trend. It is the hardest ratio to flatter, being leverage-neutral and built on operating profit, and it is the fastest read on whether a business creates or consumes value. But “first” is doing real work in that sentence — it tells you where to look next, not what to conclude. The mesh principle stands.

Where do I find these ratios already calculated?

Screener platforms compute ten-year ratio tables free, and company filings give you the raw inputs. Two habits worth building: verify decision-critical numbers against the annual report, and compute ROCE yourself at least once for any serious candidate. Knowing exactly what sits in your numerator and denominator is half the protection. SEBI’s investor education portal is a useful neutral reference for the underlying concepts.

How do these apply to loss-making companies and recent listings?

Mostly they don’t. There is no E for the P/E, and ROE is meaningless at negative equity or near-zero profits. The analysis shifts to revenue growth and its quality, the credibility of the path to profitability, cash burn against the balance available, and unit economics — a genuinely harder game with much wider error bars. A beginner’s honest option is the “too hard” pile, which is a legitimate position rather than an admission of defeat.

Do professionals use anything beyond these?

Sector-specific extensions — same-store sales, average revenue per user, order books, net interest margins — plus discounted cash flow models and forward estimates. But the core mesh in this article genuinely is the shared foundation underneath all of it. The retail investor’s edge was never a fancier ratio; it is the patience to require five years of quality and the discipline to size positions as though any single analysis might be wrong.

Keep Learning