How to Value a Cyclical Stock in India (P/E Trap)

A commodity producer trading at 9 times earnings is not cheap. It might be the most expensive thing on your screen. That sentence breaks most of what retail investors are taught about the P/E ratio, and it is the single most useful idea you can carry into the metals, sugar, cement, commodity chemicals, auto, and real estate corners of the Indian market.

The reason is mechanical, not mystical. For a cyclical business the denominator of the P/E — earnings — swings far harder than the price. Profits triple at the top of a commodity up-cycle and collapse at the bottom. The market, half-anticipating both, refuses to fully extrapolate either. So the multiple compresses exactly when profits are unsustainably high, and balloons exactly when profits are temporarily crushed. The ratio inverts. You end up paying a low P/E for earnings that are about to disappear.

We'll use one real company to keep ourselves honest: National Aluminium Company (NALCO), a state-owned, zero-debt, fully-integrated bauxite-to-aluminium producer. It is named here purely as an illustration of how a cyclical's numbers behave — not as a view on the stock, and certainly not a recommendation.

Why does the P/E ratio invert for cyclical stocks?

Think of an alumina refinery. Its costs — bauxite, caustic soda, power, depreciation on a 40-year-old plant — are fairly sticky. NALCO's own management pegged its cash cost of alumina at "around Rs. 22,000 a ton, so around $260" (NALCO Q4 FY25 concall, May 2025). The selling price, though, is a global commodity that booms and busts. When the alumina price runs hot, almost every extra dollar of realisation drops to the bottom line, and margins don't rise — they explode.

Watch it happen in NALCO's own quarterly numbers. In Q2 FY23 the company earned an operating margin of 8% and a net profit of ₹125 crore — quarterly EPS of ₹0.68 (Inve data, Q2 FY23). Ten quarters later, at the cycle peak in Q4 FY25, operating margin hit 52%, net profit hit ₹2,067 crore, and EPS hit ₹11.26 (Inve data, Q4 FY25). Same plants, same ore body, same headcount. Earnings per share rose roughly sixteen-fold in two years. That is not a growth story. That is a price chart wearing an income statement's clothes.

The market is not naive about this. It knows ₹11 of peak quarterly EPS is not repeatable, so it refuses to pay 20 times for it. At the peak the trailing multiple looks low and the headline reads "cheap." Then the cycle turns, margins revert, and the same arithmetic runs in reverse: profit falls, the multiple balloons, and the stock suddenly screams "expensive" — having done nothing but sit where the cycle put it. For a stable FMCG or software name this almost never happens, because earnings grow in a smooth line and the multiple does the talking — the clean read the P/E ratio explained for Indian investors assumes. For cyclicals, the multiple actively misleads.

What is the peak-earnings P/E trap, exactly?

The trap is buying a cyclical on a single-year P/E when that single year happens to be the best year of the cycle. It feels disciplined — you are "only paying 9 times earnings," after all. You are actually paying 9 times a number the company cannot earn again until the next up-cycle, which may be years away.

Here is the inversion in NALCO's real, full-year figures. Watch the earnings, not the price.

NALCO full-year earnings across the alumina up-cycle

All figures Inve data, FY23–FY25 (sum of four quarters).

One fence before we go further, because the mirror of the peak trap is just as costly: a depressed-earnings year is not automatically "about to recover." A cycle low is not the same as a structurally dying business, and not every trough reverts. The discipline below — normalise, then check where the company sits in its cycle — is what tells the two apart; a low number on its own never does.

FY25 EPS of ₹28.69 is 3.7 times the FY23 figure. An investor who anchored on that peak number and slapped even a modest multiple on it was capitalising a margin the business had touched for exactly one year in the prior decade. The CMD said the quiet part out loud on the May 2025 call: Q4 net sales realisation for alumina "was around $600," but "now in Q1, the net sales realisation which we are getting is around $400. The last spot price was around $400" (NALCO Q4 FY25 concall). A one-third collapse in the price that drives the whole P&L — disclosed in the same breath as the record profit. The peak EPS had an expiry date stamped on it by management itself.

What did the cycle actually do next?

It rolled over, exactly as a cyclical does. By the Q2 FY26 call in November 2025 the CMD reported alumina averaged "around $380" in the quarter, "but as of now… the spot prices are around $320" (NALCO Q2 FY26 concall). From $600 to $320 in roughly six months — the realisation nearly halved.

Here is the tell that separates a cyclical from a compounder, and it came not from a bear but from NALCO's own Director (Commercial), unprompted, at the peak: "this type of spike in the prices do happen, and then these normalizes. Actually, that always happens in a cyclic manner" (NALCO Q4 FY25 concall). When a company's own commercial head tells you the good number normalises, the only mistake left is to pay as though it won't. Plot it: alumina realisation $600 → $400 → $380 → $320 across four prints, while the screen still flashed the fat trailing margin from the top. The number you were trusting was already three quarters stale.

How do you normalise a cyclical to mid-cycle earnings?

Normalising means estimating what the company earns in an average year — not a peak or trough one — and valuing that. The recipe is plain: take a sane through-cycle margin and apply it to a sane through-cycle volume.

Walk it in three steps. First, find the mid-cycle operating margin — average it across a full cycle, ideally 7 to 10 years so you capture at least one boom and one bust, and throw out the single best and single worst years if they look like outliers. NALCO is the clean illustration: averaging its ~52% peak quarter with its ~8% trough quarter, mid-cycle margin lands far closer to the high-teens-to-twenties it printed in FY23–FY24 than to the FY25 peak. Second, estimate mid-cycle revenue using normal capacity utilisation and a long-run realisation — alumina nearer its through-cycle band, not the $600 spike. Third, run that normalised margin down to a normalised EPS, and apply a multiple appropriate to a cyclical — usually below the market average, because the earnings carry higher risk, lower predictability, and lower through-cycle growth than a steadier business.

The discipline is to value the cycle, not the latest quarter. Two cross-checks keep you honest when the earnings number is too noisy to trust at all. Price-to-book often works better for asset-heavy cyclicals, because book value doesn't gyrate the way profit does. NALCO's reserves grew from ₹12,208 crore (FY23) to ₹16,887 crore (FY25) — a smooth climb, even as quarterly EPS leapt sixteen-fold (Inve data, FY23–FY25). One caveat: for capital-intensive cyclicals, reported book sits at depreciated historical cost — NALCO's plants are partly 40 years old — so treat book as a stability proxy, not a fair-value reading. That is exactly why the second cross-check, EV to replacement cost, sits alongside it: is the market valuing this plant above or below what it would cost to build it again? Buying well below replacement cost, when no rational person would add capacity, is roughly how the great cyclical investors talk about the bottom — but mind the survivorship bias in that telling. We remember the ones who timed it; plenty bought below replacement cost and sat dead-money for years, and some rode a permanently-impaired asset all the way down because the plant was obsolete and no one would pay to rebuild it. Below replacement cost can persist for years, or never re-rate at all. It is a margin-of-safety condition, not a timing signal — and a cycle low is not the same thing as a structurally dying business. Not every trough reverts.

How should you read management's "this up-cycle is structural" guidance?

Near every peak you will hear a version of the same line: demand is now structural, not cyclical; this time is different; the China-plus-one shift, the capex super-cycle, the housing tailwind makes this durable. It is the most seductive sentence in cyclical investing, and it is also the most testable claim management ever makes — because it is a forecast with a track record attached.

NALCO offers a smaller, sharper version of the test: a hard, dated commitment you can hold management to. On the May 2025 peak call the company said its new Pottangi bauxite mine "is expected to be opened by the end of this year" (NALCO Q4 FY25 concall). It reaffirmed the date through the down-cycle — on the November 2025 call the CMD was explicit: "our target is June next year we will be starting the mines" (NALCO Q2 FY26 concall). Then came the May 2026 (Q4 FY26) call. With the June 2026 deadline weeks away, Pottangi was not mentioned at all. The commitment did not get formally missed or withdrawn — it simply went quiet (Inve data, NALCO Promise Tracker). On the same call the volume guidance for that flagship 1-MTPA refinery was quietly revised down — the FY27 production target was cut from five lakh tons to three to two lakh tons even as the commissioning date itself held at June 2026 (Inve data, NALCO Promise Tracker). An analyst on the FY25 call had already flagged an earlier six-month slip in the refinery's commissioning, asking "what has changed in the last 3 months that has lead to this 6 months of delay?" (NALCO Q4 FY25 concall).

The base rate argues for treating every confident peak-cycle line this way. Across the management commitments Inve tracks over 1,500-plus listed Indian companies, only roughly half are delivered as stated; a large share are quietly dropped — stated once, then never mentioned again on any later call — and a sizeable minority of companies carry at least one such silently-dropped commitment (Inve data, as of 2026-06-12). One honest caveat on that base rate: it reads how managements communicate over the recent transcript window — only about two years deep — not a full through-cycle delivery record. It is a prior, not a verdict, which is the same reason this whole article argues you need seven to ten years to judge the numbers of a cyclical. A "demand is structural" claim made at a cyclical peak is exactly the species of guidance that gets dropped when the cycle turns. And remember whose incentive is whose: a CMD presenting record numbers is not paid to tell you they won't last — which is why his Director's offhand "that always happens in a cyclic manner" was the most valuable line on the call.

How do you check the through-cycle record without reading ten years of transcripts?

You read the claim against the receipts. When a metals or cement management says a margin level is the "new normal," the honest test is whether they said something similar at the last peak and what happened to it — and whether the dated targets they gave you (a mine by June, a refinery by December) actually landed or quietly slid. That is tedious to reconstruct by hand across a portfolio, which is exactly why it usually doesn't get done.

This is the gap Inve's Promise Tracker is built to close: it pulls a management's stated guidance out of each earnings call and follows it forward, so you can see whether the dated commitment from a year ago turned into delivery — or went silent when realisations fell. The evidence, not a verdict, is the point. NALCO's Pottangi date didn't get debated on the FY26 call; it simply vanished from it, and that silence tells you more about the gap between a peak-cycle slide and an operating reality than any single-quarter margin. You can also screen for where cycle-sensitive earnings sit using the KPI screener, then go read the actual transcript before you trust a peak number.

Where this could be wrong

The honest counter-case has to be made, because NALCO does not fit the lazy "cyclicals are cash traps" story. Its peak earnings were real cash, not paper: FY25 cash from operating activity was about ₹5,806 crore against ₹5,267 crore of net profit, and FY26 operating cash was roughly ₹6,438 crore (company filings, consolidated, FY25–FY26). The company is essentially debt-free and pays a meaningful dividend. So the trap here is narrower than the cliché: it is not that the profit was fictional — it is that the profit is not repeatable at the peak rate. A genuine bull can argue that aluminium demand in India really is compounding at 9–10% a year (management's own figure), that NALCO's bauxite integration makes it one of the lowest-cost producers in the world, and that some of the margin uplift is structural cost advantage rather than pure price. That case is coherent. It does not survive paying a peak multiple — but it does mean "value the cycle" cuts both ways: normalising a low-cost integrated producer to the industry's mid-cycle margin can understate it just as badly as anchoring on the peak overstates it. We have not modelled the long-run alumina price or the replacement cost of an integrated complex here; doing that is the actual work, and a single illustration cannot stand in for it.

What must a five-year owner believe?

Strip away the quarter and the question for any cyclical resolves to one thing: what do you have to believe about the next full cycle, not this print? For a name like NALCO, an owner is really making three bets — that mid-cycle alumina settles high enough to clear the cost curve comfortably, that the dated expansion commitments (a mine, a refinery) eventually land even if they slip, and that the price you paid was struck against a normalised number rather than a $600 spike that management's own commercial head told you would normalise. Get the normalised earnings roughly right and the multiple takes care of itself. Anchor on the peak, and the lowest P/E on your screen will have quietly handed you the highest-priced earnings in the market.

Frequently asked questions