How to Analyse a Stock Exchange or Broker Stock

Between FY25 and FY26, BSE's revenue grew about 60% — from ₹3,030 crore to ₹4,834 crore (Inve data, FY25 and FY26). Its net profit grew 88%, from ₹1,323 crore to ₹2,487 crore, and its operating margin widened from roughly 48% to 64% over the same stretch (Inve data, FY26). Read those three numbers together and you have understood the entire investment case for a market-infrastructure business in one breath: profit grows faster than revenue, and revenue grows because other people are trading more. The exchange did not build a new factory. It carried more volume across rails that were already paid for. (Illustration of how to read the numbers, not a view on the stock.)

That is the thing to hold onto before any ratio. A factory turns steel into a margin. An exchange or a broker turns somebody else's transaction into a margin — and almost the entire cost of doing so was sunk years ago in matching engines, clearing systems and a regulatory licence that is very hard to replicate. So when volumes rise, the incremental rupee falls to profit at a rate a manufacturer can only dream of. And when volumes fall — because the regulator changed a rule, not because the company did anything wrong — that same leverage runs in reverse, fast.

This is how to read the businesses that sell access to the market itself: the half-dozen numbers that decide the outcome, where the important ones hide (rarely in the income statement), and the single risk that can halve volumes between one quarter and the next without anyone at the company making a mistake.

A boundary first. You can model these businesses better than most sectors, because the unit economics are clean. What you cannot model is the regulator's next move — and in this sector that is not a footnote, it is the main variable. Hold that thought; we return to it.

What actually drives the economics here?

Strip a stock exchange down and it is a toll bridge. It owns a piece of regulated infrastructure that every trade must cross, and it charges a tiny fee per crossing. The bridge cost a fortune to build and costs very little to run, so once traffic passes the level that covers fixed costs, almost every additional toll is profit. IEX, the power exchange, is the purest version: it ran an operating margin of about 85% in FY26 (Inve data, FY26). That is not a typo and it is not a moat made of brand — it is a toll bridge with one lane and a near-monopoly on the route.

A broker is the toll collector standing at the booth. It does not own the bridge; it brings the traffic. Angel One, a discount broker, earns when its 35.7 million clients (Angel One Q3 FY26 investor presentation) place orders, and on the cash it holds for them. The broker has thinner economics than the exchange and far less pricing power, but it has something the exchange does not: a direct relationship with the customer, which it can sell other products to — lending, mutual funds, wealth management. Listed names worth studying across this spectrum run from discount platforms like Groww through full-service and wealth-led houses such as Motilal Oswal, 360 ONE and Nuvama.

Two consequences follow, and they organise everything below. First, volume is the master variable — for both, revenue is essentially "how much activity crossed the rails" times "what we charged per unit." Second, the cost base barely moves with volume, so operating leverage is the headline feature in good times and the headline risk in bad ones. MCX, the commodity exchange, shows the upside: FY26 revenue roughly doubled (₹1,112 crore to ₹2,302 crore) while net profit grew about 138% (₹560 crore to ₹1,331 crore) (Inve data, FY25 and FY26). Costs did not double. The bridge just got busier.

The metrics that matter — and where they hide

Here is the discipline. For each metric, learn what it is, why it matters in this sector specifically, and — the part most retail analysis skips — where to find it. Several of these never appear on the income statement. They live in the investor presentation, on a single slide, and you have to go and get them.

ADTO — average daily turnover (the revenue engine)

ADTO is the value of trading that crosses the platform on an average day. It is the closest thing this sector has to "units sold," and it is the number from which almost all transaction revenue is derived.

Two cautions make ADTO trickier than it looks. First, always separate cash from F&O (derivatives), because they are different businesses. Cash turnover is the underlying value of shares changing hands; F&O turnover is measured on premium (the option price), which is a fraction of the notional value and is the segment that exploded. Second, ADTO is not on the income statement — you pull it from the exchange's monthly volume disclosures or the quarterly investor deck. BSE's average daily premium turnover (the F&O measure) ran to roughly ₹15,084 crore in Q1 FY26 (Inve data, Q1 FY26), up from about ₹8,758 crore two quarters earlier (Inve data, Q3 FY25) — the derivative volume ramp that drove that 60% revenue jump. MCX's F&O average daily turnover reached about ₹7.5 lakh crore in Q3 FY26 (Inve data, Q3 FY26), after ADT grew roughly 103% year-on-year in Q2 FY26 (Inve data, Q2 FY26). What's "good" is not a level — it is durable, broad-based growth, not one hot segment carrying everything.

Active clients (the broker's base)

For a broker, the client count — specifically active clients who actually trade — is the engine ADTO sits on. More clients placing more orders is the whole growth algebra. Angel One reported a total client base of 35.7 million in Q3 FY26 (Angel One Q3 FY26 investor presentation), with gross client acquisition still running at 1.7 million in the quarter. The number you want is not the cumulative "ever opened" figure brokers love to headline — it is the active base and the order trajectory, because dormant demat accounts pay no toll. You find it in the broker's deck, not its P&L.

Market share (the moat made visible)

Market share tells you whether the moat is real. In market infrastructure it tends to be unusually sticky because of the network effect: liquidity attracts liquidity — traders go where the orders already are, and that self-reinforces. CDSL, one of two depositories, held roughly 79–80% of the depository market through FY25–FY26 and an incremental (new-account) share of about 87.5% in Q4 FY26 (Inve data, Q4 FY26) — meaning it is winning the flow, not just sitting on a stock. IEX held about 84% of the electricity exchange market (IEX Q2 FY25 concall). When you see a share that high and still climbing on incrementals, you are looking at a genuine network moat. When you see it slipping quarter after quarter, the network is fraying — and for these businesses the network is the asset.

Take rate / yield (what they keep per unit)

Take rate is revenue divided by the volume that produced it — the toll per crossing. It matters because volume can grow while the take rate quietly erodes (regulatory fee caps, competitive discounting, mix shifting to lower-yield segments), so rising ADTO with a falling take rate is not the win it looks like. This number is rarely printed directly; you compute it — transaction revenue ÷ turnover — from the deck and the volume disclosures. Watch it like a hawk in regulated venues, where the regulator can cap the toll directly.

Transaction vs non-transaction revenue mix (the quality of the earnings)

This is the most underrated number in the sector, and it decides how much you should pay for the profit. Transaction revenue rises and falls with daily volumes — it is real but volatile, a hostage to sentiment and regulation. Non-transaction revenue — listing fees, data and index licensing, KYC and corporate-action income for a depository, colocation rack rentals, and the float income earned on client and settlement cash — is stickier and recurring. A business leaning on annuity-like non-transaction income deserves a higher multiple than one that is a pure bet on this quarter's froth.

Angel One's gross income split makes the point concretely: in Q3 FY26, interest income was about 33% of gross income and brokerage the largest slice, with distribution, depository and other lines making up the rest, on ₹13.4 billion of gross income (Angel One Q3 FY26 investor presentation). A third of the income was interest on client funds and the lending book — not broking at all. That is the part a volume crash does not touch as hard, and it is why the mix slide, buried in the deck, tells you more about durability than the headline revenue line does.

Regulatory exposure (the variable that overrides all the others)

This is not a metric you read off a page; it is a risk you map. Every business here operates a licensed monopoly or duopoly at the regulator's pleasure, and the regulator's tools — fee caps, lot-size rules, expiry-day limits, market-coupling mandates — can reset volumes or yields directly, with no warning the financials would give you. It belongs on this list because it dominates the others: a perfect ADTO trend and a fortress market share are worth less if a single circular can halve the addressable volume. We give it the worked case below, because abstract warnings about "regulatory risk" are cheap and the specific mechanism is what's worth knowing.

How do you value a business that's mostly operating leverage?

The right lens here is P/E, but read through the operating-leverage cycle — and that "but" is the whole craft. A naive P/E badly misprices these businesses in both directions, because earnings are the most cyclical thing about them.

Think of it as a seesaw bolted to the volume cycle. When volumes are surging, margins are near their ceiling, earnings are at a cyclical high, and the trailing P/E looks deceptively cheap — you are capitalising peak earnings as if they were normal. When volumes crash, margins collapse faster than revenue (fixed costs don't flex), earnings fall off a cliff, and the trailing P/E spikes to a scary-looking number on trough earnings — which can be exactly the wrong moment to flee. The market chronically pays too little at the top of the volume cycle and too much fear at the bottom.

So three adjustments. First, normalise — ask what earnings look like at mid-cycle volumes, not today's. BSE going from a 31% operating margin in Q3 FY25 to 67% in Q4 FY26 (Inve data, Q3 FY25 and Q4 FY26) is a swing no manufacturer shows; capitalising the 67%-margin quarter at a fat multiple is capitalising the peak. Second, weight the multiple by the revenue mix — a rupee of listing or data or float income deserves a higher multiple than a rupee of expiry-day option-premium toll, because one survives a quiet year and the other does not. Third, and most important, discount explicitly for regulatory headroom — a venue earning a high take rate the regulator could cap tomorrow is worth less per rupee of current profit than one whose pricing is already light. The cheapest-looking exchange on trailing earnings is often the one with the most regulatory rope still to pay out.

A worked case: when the rule changed, not the company

Take the clearest regulatory shock the sector has lived through recently — SEBI's tightening of the equity-derivatives segment through late 2024 (larger contract sizes, fewer weekly expiries, tighter intraday limits), aimed at cooling retail F&O speculation. For a discount broker whose revenue leans on derivative order flow, this is the existential variable made real. (Illustration, not a view on the stock; figures as reported by the company.)

Watch what Angel One's own slide says about it. In its Q3 FY26 deck, under a heading that reads, almost defiantly, "Business Maturing With Regulations," management plotted average daily orders against each successive rule change and concluded the franchise had "demonstrated better growth post all regulatory changes" — citing 2.2x and 2.4x order growth through earlier cycles (peak-margin norms, cash-collateral segregation), and on the F&O rules specifically: "Transient impact on average daily orders during the implementation of F&O regulations… average daily orders normalizing from Q2 '26" (Angel One Q3 FY26 investor presentation).

Now hold that confident framing against the numbers the same period produced. Look back at the quarterly profit path: Angel One's net profit fell to about ₹114 crore in Q1 FY26 and ₹175 crore in Q4 FY25 (Inve data) — well below the roughly ₹281 crore it earned in Q3 FY25 (Inve data, Q3 FY25), the quarter before the rules bit, when F&O made up about 77% of gross broking revenue (Inve data, Q3 FY25). The recovery is real — profit climbed back to about ₹320 crore by Q4 FY26 (Inve data, Q4 FY26) — but "transient" is doing a lot of work for a swing that more than halved quarterly profit at the trough and took four quarters to mend. The honest reading is not that management lied; the orders did normalise. It is that a broker calling a regulatory volume shock "transient" while it is happening is exactly the management voice you should weigh against the cash, because they are paid to sound unshaken — don't ask the toll collector whether the new tax will hurt the booth.

The discipline this teaches: when a regulator moves, read the operating numbers (orders, ADTO, the revenue-mix slide) yourself across the next three or four quarters, and treat management's "transient" as a hypothesis to test, not a finding. This is precisely the work Inve's Promise Tracker is built to spare you — pinning each forward statement to the quarter and quote it was made in, then marking it as later results confirm or quietly contradict it, so you see the sequence rather than re-reading every deck by hand.

The red flags specific to this sector

Five things that should slow you down, none of which a single good quarter rules out:

• One segment carrying all the growth. If ADTO growth is entirely F&O premium turnover — the most speculation-driven, most regulator-targeted line — the franchise is one circular away from a reset. Breadth across cash, F&O and segments is durability.
• Take rate quietly eroding while volumes rise. Rising turnover masks a falling toll per trade. Compute the take rate yourself; a venue celebrating volume while its yield slips is renting growth it doesn't keep.
• Earnings flattered by float income at the top of the rate cycle. Float income (interest on client and settlement cash) is real but rate-sensitive; a profit line fattened by a high-rate environment will thin when rates fall, independent of volumes.
• Market share slipping on incrementals. The stock of share lags reality; the flow (new accounts, incremental volume) leads it. A depository or exchange losing the incremental battle is losing the network, even while the headline share still looks dominant.
• A pending regulatory change management calls "well understood." The most dangerous sentence in this sector is a confident one about a rule not yet implemented. IEX, sitting on ~84% electricity-market share, faced exactly this: a CERC order to implement market coupling — which would pool bids across exchanges and erode IEX's liquidity moat — that management acknowledged was mandated "by January 2026 as per the order" (IEX Q1 FY26 concall) while noting a quarter later it was "not aware about any developments which have taken place so far" (IEX Q2 FY26 concall). A near-monopoly whose monopoly is a regulator's pending decision is not as safe as its 84% suggests. (Illustration, not a view on the stock.)

Frequently asked questions

The discipline comes down to refusing to be impressed by a peak. These are wonderful businesses when volumes run and unforgiving ones when a rule resets them — and the difference between the two states can be a single circular, not a management mistake. So invert the question you bring to a quarter. Don't ask "how fast is it growing?" Ask: if a regulator capped this venue's best fee or curbed its busiest segment next quarter, how much of this profit survives — and does the revenue mix tell me it would? A business that is all F&O premium toll and float income does not survive that test well; one with real listing, data and recurring income does.

And the owner's question, the one to sit with before buying a single share of any market-infrastructure business: what must I believe about the next regulatory cycle — not this quarter's volumes — for this toll bridge to still be standing, still busy, and still pricing its own toll five years from now? If the honest answer leans on a regulator's forbearance rather than the business's own breadth, you have read the volume chart, not the business.

If it helps to start from a sibling balance-sheet business with a very different risk shape, how to analyse an NBFC walks the lender's version — where the killer is credit, not a circular.

Related reading

• How to Analyse a Bank Stock (NIM, CASA, GNPA, ROA)
• How to Analyse a Building Products Company (Pipes, Tiles)
• How to Analyse a Cables and Wires Stock