Reading the notes - where the real disclosures live

A pharma company's balance sheet looks sturdy: net worth of ₹10,500 crore, modest borrowings, healthy cash. Nothing on the face of the statement suggests a problem. Then you turn to the notes, and one of them lists the contingent liabilities - disputed tax demands, product-liability and price-fixing litigation in the United States, guarantees given for subsidiaries. Added up, they come to ₹3,150 crore, nearly a third of the entire net worth. None of that appears in any total on the balance sheet, because each item is only a possible obligation: it becomes real money owed only if a court rules against the company, a tax authority prevails, a subsidiary defaults. But "possible" is not "small". If even half of that cluster went the wrong way in a bad year, it would tear a hole in the balance sheet that the face of the statement gave no hint of. illustrative

That is the point of this module. The three statements are a summary, constrained by accounting rules and the auditor and the brute arithmetic that ties them together - which is exactly what makes them trustworthy, and exactly what makes them incomplete. The notes are where the summary is forced to admit what it left out: the obligations that are not yet certain, the profit that came from selling to the promoter's own company, the strong division quietly carrying a weak one, the risk the auditor thought important enough to name. "Revenue: ₹9,800 crore" does not tell you a seventh of it was sold to a company the promoter owns. "Net worth: ₹10,500 crore" does not tell you there is ₹3,150 crore of contingent claims hanging over it. The face gives the figure; the note gives the meaning.

The reason this matters is blunt: almost every accounting failure that later looked obvious was disclosed, in some form, in the notes before it blew up. The related-party dependence, the ballooning contingent claim, the segment losing money behind a healthy group average, the auditor's carefully worded warning - they were on the page. The face of the statement is where a company presents itself. The notes are where it is made to disclose itself. Learning to read them is the difference between being surprised and being early.

Four notes answer questions the face of the statement cannot, and each is read a different way.

Contingent liabilities - what could go wrong that is not yet counted. These are potential obligations that crystallise only if some event happens: a tax dispute lost, a legal claim upheld, a guarantee called. They are deliberately kept out of the balance-sheet totals because they are not yet real, and that is precisely why they are dangerous to ignore - they carry no total anywhere on the face, so a screener never sees them. The note is always read relative to net worth. A contingent figure that is a few percent of net worth is background noise. One that is a third of net worth, as below, is a live risk to the whole balance sheet - and one that exceeds net worth is an existential threat, because a single adverse cluster of rulings could take the equity below zero. A claim larger than the equity that would absorb it is not a risk you average out; it is one that can end the company outright, so the ratio, not the rupee figure, is the whole of the skill.

Related-party transactions - who is this company really dealing with, and could the profit be manufactured between insiders. These are dealings between the company and parties connected to it: the promoter, the promoter's other companies, key managers, their relatives. The note lists what was bought, sold, lent, or guaranteed, and to whom. It is one of the sharpest governance signals in the report because a sale to a related party is a sale the company controls both sides of - profit can be created by selling high to a captive buyer, or siphoned by buying dear from a promoter's firm. A little related-party dealing is normal; a lot of it, especially where the profit or the growth is concentrated, means the reported numbers may not be arm's-length and can be moved at an insider's discretion.

Segment reporting - which part of this business actually earns, and which is being carried. A consolidated profit is an average, and an average hides as much as it reveals. The segment note breaks revenue, profit and capital down by business line or geography. A group reporting a healthy 18% margin might have one segment earning 28% and its core business losing money - a fact the consolidated line erases completely, and the segment note is the only place it survives.

The auditor's report - what did the people who checked the books want me to notice, or refuse to sign off on. Most of the time it is a clean "unqualified" opinion and you move on. Two things make it worth reading every time. An emphasis of matter is the auditor signing the accounts as fairly stated while deliberately pointing at a specific risk - a going-concern doubt, a large contingent claim, a material uncertainty - that they want the reader to see. A qualification goes further: it says some part of the accounts could not be verified or does not comply, so that part cannot be relied upon. The first is a signpost; the second is a refusal. Neither is boilerplate, and both are written by the one party legally obliged to look.

Every company has all four notes, but the first one to open changes by business. For a pharma exporter it is contingent liabilities; for a bank it is the notes behind provisioning and asset classification; for a commodity maker it is the hedging and derivative note that can make or unmake a cyclical profit before it settles. The genuine inversion is the holding company. For most businesses the face is the main event and the notes are supporting detail; for a holding company the relationship flips - the consolidated face blends a bank, a cement plant and an insurance arm into one meaningless margin, and the segment and related-party notes are where the entire story actually lives. Reading the summary first, as you would for an ordinary company, is precisely the wrong instinct there.

Return to the pharma exporter. On the face of it, a fine business: ₹9,800 crore of revenue, a 22% EBITDA margin, ₹1,240 crore of profit, low debt. Stop at the statements and you would call it clean. Now read three notes in order. illustrative

First, contingent liabilities: ₹3,150 crore, thirty percent of net worth, concentrated in US litigation and disputed tax. That does not make the company uninvestable - these disputes often settle for a fraction and drag on for years - but it reframes the balance sheet. The question is no longer "how strong is the equity" but "how strong is the equity if a third of it is called", and you now read the cash and the borrowing headroom differently, because you know what might land on them.

Second, related parties: fourteen percent of revenue is sold to promoter-linked entities. Not automatically sinister - many groups have genuine intra-group supply - but it means fourteen percent of the top line is not fully arm's-length, and if the margin on those sales is unusually fat, some of the reported profit may be a transfer rather than a market outcome. You would want the terms, and you would watch whether that share grows.

Third, segments: the US generics business is the largest at ₹5,400 crore of revenue but earns only a 12% margin under price erosion, while the Indian branded business earns 28% on ₹3,100 crore. The consolidated 22% is a blend of a big thin-margin engine and a smaller fat-margin one. That tells you exactly where a shock would hurt - a further US price collapse would drag the whole group down, because it is the largest slice - and where the quality really sits. None of it is visible on the face of the P&L, which shows one margin for one company.

The habit to build: never form a view from the face alone. For any company, read these three notes before you trust the summary - contingent liabilities scaled to net worth, related parties as a share of revenue and profit, segments to see which business actually earns. The face tells you the company's story about itself; the notes tell you whether that story survives contact with what the company was obliged to disclose.

An analyst who has read the notes brings them into the call, because the notes are where management is least rehearsed. On the pharma exporter, the question sounds like this: "Your contingent liabilities rose to about ₹3,150 crore, a third of net worth, led by the US litigation. What's your best estimate of the probable cash outflow, and over what period?" The aim is to turn a disclosed amount claimed into an estimated amount likely to be paid.

A good answer, verbatim-style.

"Of the ₹3,150 crore, the large item is the US price-fixing matter at ₹1,400 crore claimed. Our external counsel's view, and our provision, is that a probable settlement is in the ₹300–400 crore range, already provided for in other liabilities - so it's not additional to what's booked. The disputed tax of ₹900 crore we expect to win at appellate level based on precedent, but it could take three to four years. Guarantees for subsidiaries are ₹600 crore and those subsidiaries are cash-positive. So the incremental unprovided cash risk we'd flag is modest, and we've laid out the phasing in note 34."

A split by item, a probable figure against each claimed figure, what is already provided, the timeline, and where to read it. It converts the note into a risk you can size.

An evasive answer, verbatim-style.

"These are contingent liabilities, so by definition they may never materialise. We're advised we have strong cases on all of them and we're contesting vigorously. This is fairly standard for a company of our scale operating in regulated export markets, and our auditors are comfortable with the disclosure. We don't see a material impact."

Every sentence is defensible, and none of it is a number. It leans on "contingent" and "strong cases" to imply the exposure is theoretical, without giving a probable outflow, a provision, or a timeline for any item. "Standard for our scale" is a deflection; "auditors are comfortable" confirms the disclosure is adequate, not that the risk is small; "no material impact" is an assertion with nothing behind it.

The follow-up nobody asks. "For each contingent item over ₹500 crore, what have you actually provided, and what's the unprovided worst case?" That forces the vague reassurance into a table - claimed, provided, unprovided - for each big item. If "these may never materialise" is allowed to stand, the reader is left treating a third of net worth as theoretical when part of it may be months from settling. The silence is the tell: either the unprovided worst case is uncomfortable, or nobody following the company has done the reading.

The notes tell you what a company was required to disclose. They do not tell you what it was not required to disclose, and the gap between the two is where the hardest problems hide. A related-party structure routed through entities that fall just outside the disclosure threshold will not appear in the note, however real the dependence. Nor can a note tell you how a contingent liability will resolve: it gives the amount claimed, not the probability of losing or the likely settlement. A ₹1,400 crore exposure might cost nothing or cost the lot - the note cannot price it. What it does is tell you the exposure exists and how big it could be, so you can size it against net worth. The judgement of how likely and how much is yours to make, from the history of similar cases and the company's own track record, not the note's to hand you. Reading the notes carefully makes you far harder to surprise; it does not make you impossible to surprise, and treating a clean set of notes as proof of a clean company is its own mistake.

Three traps recur. The first is judging a contingent liability by its absolute size instead of against net worth. ₹1,000 crore sounds identical in a press summary whether the company has ₹800 crore of net worth or ₹40,000 crore. Against a small equity it is existential; against a large one it is a footnote. The number means nothing until you divide it by what would have to absorb it.

The second is being reassured by the wrong signals. "Our auditors are comfortable" tells you the disclosure is adequate, not that the risk is small. "One of the Big Four signs our accounts" tells you who checked, not that related-party sales were arm's-length. Investors reach for these brand signals precisely because reading the actual notes is work - but they are a shortcut to a different question than the one that matters.

The third is treating the segment note as detail rather than as the real P&L. A single consolidated margin is an average, and averages are where a weak core hides behind a strong division. The company that reports the same group margin as its healthy peers, but earns it by having one segment subsidise a loss-making core, looks identical on the face and is a completely different investment. Anyone who reads only the consolidated line will never see the difference - and it is exactly the difference that decides what happens when the strong segment has a bad year.