04.2 · Margins & cost structure¶

Why this matters: growth tells you how big the business is getting; margins tell you how much of each rupee of sales survives to shareholders — and, more importantly, why. A company whose gross margin is stable through a commodity spike has pricing power. A company whose EBITDA margin expands as it grows has operating leverage. Both are worth more than the same profit earned by luck.

Learning objectives — after this lesson you can:

Build and read a common-size income statement and explain what each margin layer (gross, EBITDA, EBIT, PAT) is telling you.
Split a cost base into fixed and variable components, compute contribution margin, break-even revenue and the degree of operating leverage, and forecast how profit responds to a revenue change.
Explain raw-material pass-through, lags and mix effects, and separate them in a margin bridge.
Judge whether a margin is sustainable using peer levels, history and the pricing-power evidence in the gross margin.
Apply all of this to Kaveri Pumps FY21–FY26 and diagnose what drove its margin peak and decline.

Prerequisites: 02.3 The income statement, 04.1 Growth analysis · Time: ~90 min

1. The margin ladder and what each rung measures¶

A margin is a profit subtotal divided by revenue. The four that matter, and what each isolates:

Margin	Formula	What it measures	Typical India ranges (indicative)
Gross margin	(Revenue − cost of goods) ÷ revenue	Pricing power over customers and suppliers; product economics before any overheads	Commodity processors 10–20%; auto components 25–40%; pumps/engineering 30–40%; FMCG 45–60%; software 65–80%; pharma formulations 60–70%
EBITDA margin	EBITDA ÷ revenue	Operating efficiency after overheads (people, power, freight, marketing) but before the cost of the asset base	Contract manufacturing 5–10%; engineering 12–18%; FMCG 20–25%; IT services 20–28%; cement 15–25% (cyclical)
EBIT margin	EBIT ÷ revenue	Operating profit after depreciation — the honest operating margin for asset-heavy businesses	—
PAT margin	PAT ÷ revenue	What is left after financing and tax; mixes operating performance with capital structure	—

The ladder is read top-down. If gross margin moves, the cause lies in price, input cost or product mix. If gross margin is stable but EBITDA margin moves, the cause is overheads — scale, marketing, one-off costs. If EBITDA margin is stable but EBIT margin moves, it is depreciation (a new plant). If EBIT margin is stable but PAT margin moves, it is interest, other income, exceptional items or tax. This diagnostic order is the whole skill.

India notes

Indian P&Ls classify expenses by nature, so gross margin must be built: cost of goods = cost of materials consumed + purchases of stock-in-trade + change in inventories. Screener.in's "OPM" is EBITDA margin excluding other income; many broker reports quote EBITDA including other income. When two sources disagree by a percentage point, this is usually why.

2. Common-size statements: Kaveri FY21–FY26¶

Divide every line by revenue. Numbers are ₹ Cr from the running example; percentages are computed with tools/fi/ratios.margins.

% of revenue	FY21	FY22	FY23	FY24	FY25	FY26
Revenue (₹ Cr)	612.0	758.0	874.0	1,006.0	1,172.0	1,318.0
Materials	64.0	67.0	65.8	63.2	64.4	65.1
Gross margin	36.0	33.0	34.2	36.8	35.6	34.9
Employees	10.2	9.2	9.1	9.0	8.8	8.9
Other expenses	12.8	11.8	12.0	12.4	12.1	12.2
EBITDA margin	13.0	12.0	13.1	15.4	14.7	13.8
Depreciation & amortisation	4.3	3.7	3.4	3.3	3.8	3.6
EBIT margin	8.7	8.3	9.7	12.1	10.9	10.2
Finance costs	1.1	0.8	1.0	1.3	1.4	1.3
Other income + exceptional	−0.4	0.8	0.8	0.7	1.6	0.3
PAT margin	5.4	6.2	7.2	8.6	8.3	6.9

Reading it top-down:

FY22: gross margin fell 300 bps (36.0 → 33.0) — the post-COVID copper and steel spike hit a company that prices its pumps to dealers once or twice a year. Overheads fell as a percentage because revenue grew 24% on a largely fixed base: EBITDA margin fell only 100 bps. Operating leverage cushioned a pricing-power problem.
FY23–FY24: gross margin recovered to 36.8% as price increases caught up with costs and the industrial mix helped; employee costs kept shrinking as a share of sales. EBITDA margin peaked at 15.4%.
FY25–FY26: gross margin slid 190 bps from the peak — solar pumping systems (bought-in panels, tender pricing) went from 14% to 27% of sales. Depreciation stepped up when the Hosur plant was capitalised. PAT margin in FY25 was propped up by the ₹14 Cr land gain (1.2% of revenue); strip it and FY25 PAT margin is 7.4%, so the decline into FY26 is gentler than the reported 8.3 → 6.9.

The lesson: the level of Kaveri's margin is average for Indian engineering; the direction since FY24 is down and the cause is mix (a lower-margin segment growing fastest), not costs running out of control. That distinction decides whether you expect mean reversion.

3. Fixed and variable costs¶

Every cost line is somewhere between purely variable (moves one-for-one with volume: raw materials, freight, sales commissions, payment-gateway fees) and purely fixed (does not move with volume in the short run: plant depreciation, salaried staff, rent, software licences). Reality is a mix — "semi-variable" costs with a fixed core and a variable top-up — and most costs are fixed over a quarter and variable over five years.

Why it matters: the split determines how profit responds to a change in revenue.

\[\text{Contribution} = \text{Revenue} − \text{Variable costs} \qquad \text{Contribution margin (CM)} = \frac{\text{Contribution}}{\text{Revenue}}\]

\[\text{EBIT} = \text{Contribution} − \text{Fixed costs} \qquad \text{Break-even revenue} = \frac{\text{Fixed costs}}{\text{CM}}\]

3.1 Estimating the split for Kaveri (FY26)¶

Annual reports do not label costs as fixed or variable; you estimate. A reasonable first cut for a pump maker: materials 100% variable; employee costs and other expenses ~60% fixed (salaried staff, plant overheads, rent, IT) and ~40% variable (contract labour, freight, power for production, dealer incentives); depreciation 100% fixed.

rev, mat, emp, oth, da, ebit = 1318.0, 858.0, 117.3, 160.8, 47.8, 134.1
variable = mat + 0.4 * (emp + oth)          # 969.2
fixed = 0.6 * (emp + oth) + da              # 214.7
contribution = rev - variable               # 348.8
cm = contribution / rev                     # 26.5%
breakeven = fixed / cm                      # 811.2
dol = contribution / ebit                   # 2.60
print(round(contribution,1), round(cm,3), round(fixed,1), round(breakeven,1), round(dol,2))
# 348.8 0.265 214.7 811.2 2.6

So Kaveri earns about 26.5 paise of contribution on each extra rupee of sales, carries ~₹215 Cr of fixed costs, breaks even at ~₹810 Cr of revenue (it did ₹1,318 Cr — a comfortable 38% cushion), and has a degree of operating leverage of 2.6: a 10% change in revenue moves EBIT by roughly 26%, if margins per unit hold.

3.2 Degree of operating leverage (DOL)¶

\[\text{DOL} = \frac{\%\Delta \text{EBIT}}{\%\Delta \text{Revenue}} = \frac{\text{Contribution}}{\text{EBIT}} = 1 + \frac{\text{Fixed costs}}{\text{EBIT}}\]

You can measure it two ways: structurally (contribution ÷ EBIT, as above) or empirically from history. The empirical version is noisy because price and mix move at the same time as volume:

Period	Revenue growth	EBIT growth	Empirical DOL
FY23 → FY24	+15.1%	+44.3%	2.9
FY24 → FY25	+16.5%	+4.8%	0.3
FY25 → FY26	+12.5%	+4.8%	0.4

FY24 shows the structural leverage working (2.9 ≈ the 2.6 estimate). FY25 and FY26 show it being overridden by gross-margin erosion from mix — which is the point: DOL tells you what happens to profit when volume changes with everything else constant; the gross margin tells you whether everything else is constant.

Trader's lens

Operating leverage is the gamma of a business. A high-fixed-cost company (airline, hotel, steel mill, exchange) has convex profits: small revenue moves produce large profit moves, in both directions. A variable-cost business (distributor, contract manufacturer) is nearly linear. When you pay a high multiple for an operationally levered company near its break-even, you are long gamma at a bad strike — the payoff is explosive but the theta (fixed costs) bleeds every quarter that volume disappoints. Cyclical troughs are where that convexity is cheapest; 07.3 builds on this.

3.3 A second worked example — a QSR chain¶

Tandoor Express Ltd (fictional) runs 200 restaurants. Per store per year (₹ lakh): revenue 240; food & packaging cost 30% of revenue; delivery commissions 8% of revenue; staff 60 (mostly fixed); rent 36 (fixed, under Ind AS 116 this is D&A + interest, but treat as fixed cost here); utilities 12 (half fixed); depreciation 18.

rev = 240.0
var = 0.30 * rev + 0.08 * rev + 6          # food, delivery, variable half of utilities = 97.2
fixed = 60 + 36 + 6 + 18                   # 120
contrib = rev - var                        # 142.8; CM 59.5%
ebit = contrib - fixed                     # 22.8; margin 9.5%
be = fixed / (contrib / rev)               # 201.7 lakh per store
dol = contrib / ebit                       # 6.3
for g in (-0.10, 0.10, 0.20):
    print(g, round((rev*(1+g) - var*(1+g) - fixed), 1))
# -0.1 → 8.5 ; +0.1 → 37.1 ; +0.2 → 51.4

Store-level EBIT is ₹22.8 lakh (9.5%) with a DOL of 6.3: a 10% fall in same-store sales cuts store profit by 63%; a 20% rise more than doubles it. This is why QSR stocks trade on same-store sales growth (SSSG) prints and why a chain that keeps adding stores while SSSG is negative can show revenue growth and collapsing profit at the same time. 08.3 uses these unit economics.

4. Raw-material pass-through, lags and the gross margin¶

For manufacturers, the gross margin is a tug-of-war between input costs and selling prices. Three questions:

Can the company pass cost increases on? Evidence: gross margin percentage stable while raw-material prices swing (paints, FMCG leaders, branded pumps). If gross margin per unit in rupees is stable but the percentage falls when inputs rise, the company passes through cost but not margin — common in auto components with contractual pass-through clauses.
With what lag? Contractual quarterly resets (auto ancillaries), annual price lists (pumps, appliances), tender-locked prices (solar, EPC, defence — no pass-through at all until the next tender). Lag creates a predictable margin squeeze and rebound: for Kaveri, copper spiked in FY22 and the margin recovered in FY23–24.
What is mix doing? A rising share of low-margin products lowers the blended gross margin even when every product's own margin is unchanged. This is Kaveri's FY25–26 story. You can separate it with a margin bridge.

4.1 A margin bridge (mix vs rate)¶

Suppose Kaveri's segment gross margins in FY24 were agri 40%, industrial 36%, solar 22% (fictional detail consistent with the blended 36.8%), and in FY26 they were agri 40%, industrial 36%, solar 24%. Segment shares moved from 57/29/14 to 46/27/27.

fy24 = {'agri': (0.57, 0.40), 'ind': (0.29, 0.36), 'solar': (0.14, 0.22)}
fy26 = {'agri': (0.46, 0.40), 'ind': (0.27, 0.36), 'solar': (0.27, 0.24)}
gm24 = sum(w*m for w, m in fy24.values())                    # 36.32%
gm26 = sum(w*m for w, m in fy26.values())                    # 34.60%
mix_effect  = sum(fy26[s][0]*fy24[s][1] for s in fy24) - gm24   # new weights, old margins: −2.26 pp
rate_effect = gm26 - (gm24 + mix_effect)                        # +0.54 pp
print(round(gm24*100,2), round(gm26*100,2), round(mix_effect*100,2), round(rate_effect*100,2))
# 36.32 34.6 -2.26 0.54

The bridge says the ~170 bps decline is entirely mix (−226 bps) partly offset by rate (+54 bps: solar margins actually improved). That is a different — and less alarming — conclusion than "Kaveri is losing pricing power". It also tells you what to model: if solar keeps growing faster than the rest, the blended margin keeps falling even if nothing goes wrong.

5. Are margins sustainable? Mean reversion and the evidence for pricing power¶

Margins mean-revert. Above-average margins attract competition and capacity (the capital cycle, 05.4); below-average margins drive exits and price discipline. The exceptions — companies that hold 20–25% EBITDA margins for decades — are the ones with a moat (05.3). Before extrapolating a margin, check:

Test	What to look at	Kaveri
History	10-year range and where the current margin sits in it	EBITDA 12.0–15.4% over six years; FY26 at 13.8% is mid-range
Peers	Same-segment peers' margins (the running example's fictional peer set: 11.5–16.8%)	Mid-pack; Nilgiri Pumps at 16.8% shows what a stronger brand/mix earns
Gross-margin stability through input cycles	Std. dev. of gross margin vs std. dev. of key input prices	±1.5 pp swings vs copper ±30%: decent but lagged pass-through
Mix trajectory	Which segments are growing and at what margin	Lower-margin solar growing fastest → structural pressure
Scale economies	Fixed costs as % of sales falling with growth	Employees 10.2% → 8.9%: yes
One-offs inside opex	Exceptional items, provisions, ESOP charges	FY25 land gain (below EBITDA); ECL provision arguably too low (flatters other expenses)

A last, quantitative sanity check: what is one percentage point of gross margin worth? For Kaveri, 1% of revenue is ₹13.2 Cr, which is 9.8% of FY26 EBIT. A 2-pp mix-driven slide — the FY24→FY26 experience — is a fifth of operating profit. Small margin numbers, big profit numbers: that asymmetry is why margin analysis comes before growth analysis in most analysts' heads, even though this course teaches them the other way round.

Common mistakes

Comparing EBITDA margins across companies with different lease intensity: since Ind AS 116 (FY20), rent sits in depreciation and interest, so retailers' and airlines' EBITDA margins jumped without any economic change. Use EBIT or pre-Ind AS 116 EBITDA for cross-period comparisons.
Treating a mix-driven margin decline as a pricing-power problem (or the reverse). Do the bridge.
Applying DOL to a period where gross margin also moved — it only isolates volume.
Ignoring other income when comparing "EBITDA" from different sources.
Extrapolating peak margins from the top of a commodity cycle (chemicals FY22, steel FY22).
Forgetting that a percentage margin can rise while rupee profit falls (shrinking revenue with fixed costs cut).

Key terms¶

Term	Meaning
Common-size statement	Each P&L line expressed as a percentage of revenue
Gross margin	(Revenue − cost of goods) ÷ revenue; the pricing-power layer
EBITDA / EBIT / PAT margin	Successive profit subtotals ÷ revenue
Variable cost	Cost that moves proportionally with volume (materials, freight, commissions)
Fixed cost	Cost that does not change with volume in the short run (depreciation, salaried staff, rent)
Contribution margin	(Revenue − variable costs) ÷ revenue; the profit on each incremental unit before fixed costs
Break-even revenue	Fixed costs ÷ contribution margin
Degree of operating leverage (DOL)	%ΔEBIT ÷ %Δrevenue = contribution ÷ EBIT; profit's sensitivity to volume
Pass-through	The ability and speed with which input-cost changes are reflected in selling prices
Mix effect	Change in a blended margin caused by the weights of segments/products shifting
Rate effect	Change in a blended margin caused by segments' own margins changing
Margin bridge	Decomposition of a margin change into mix, rate (price/cost) and volume/scale components
Mean reversion	The tendency of abnormal margins to move back toward industry norms as competition responds
SSSG	Same-store sales growth — revenue growth from stores open more than a year

Check your understanding¶

Kaveri's employee costs fell from 10.2% to 8.9% of revenue between FY21 and FY26 while revenue more than doubled. Is this cost cutting or operating leverage? How would you tell?

Answer

Almost certainly operating leverage: absolute employee cost rose 88% (62.4 → 117.3) while revenue rose 115%, so the base is growing but slower than sales, as a partly fixed cost should. Cost cutting would show absolute costs flat or falling, headcount down, or a VRS (Kaveri did have one in FY21). Check headcount (2,310 → 2,480) and revenue per employee.

Using the FY26 fixed/variable split above, estimate Kaveri's EBIT if FY27 revenue falls 8% with unchanged contribution margin.

Answer

Contribution = 348.8 × 0.92 = 320.9; EBIT = 320.9 − 214.7 = ₹106.2 Cr, a 20.8% decline (≈ 2.6 × 8%). Reported Q1 FY27 already showed EBITDA margin down to 12.6% on +3.5% revenue — margin slippage is compounding the leverage.

Company A: gross margin 55%, EBITDA margin 12%. Company B: gross margin 25%, EBITDA margin 12%. Which likely has pricing power, and which has the higher operating leverage?

Answer

A has pricing power (high gross margin) but spends 43% of revenue on overheads (marketing, staff) — typical of a consumer brand or software firm; its DOL is high because so much cost is fixed. B is a lean processor/distributor with little pricing power and mostly variable costs; low DOL. Same EBITDA margin, very different businesses and risk.

Redo the margin bridge if solar's FY26 margin had stayed at 22%. What would the blended margin be?

Answer

0.46×0.40 + 0.27×0.36 + 0.27×0.22 = 0.184 + 0.0972 + 0.0594 = 34.06%; the whole decline (−2.26 pp) would be mix, with zero rate effect.

A retailer's EBITDA margin rose from 8% in FY19 to 13% in FY20 with no change in sales or gross margin. What happened?

Answer

Ind AS 116 took effect for FY20: store rents moved out of other expenses into depreciation (ROU assets) and finance costs (lease liabilities). EBITDA rose mechanically; EBIT and PAT barely moved. Compare EBIT margins, or add lease payments back to get a pre-116 EBITDA.

Why can a company report a rising EBITDA margin and falling EBITDA in the same year?

Answer

Revenue fell and management cut variable and some fixed costs faster than revenue fell (or mix shifted toward higher-margin products as volume shrank). Margin percentages are ratios; rupee profit is what pays dividends. Always look at both.

Go deeper¶

Michael Mauboussin & Dan Callahan, Operating Leverage (Credit Suisse/Morgan Stanley research note, 2016) — the clearest practitioner treatment of DOL and why it is hard to measure.
Any Indian auto-component company's investor presentation — look for the raw-material pass-through clause language and lag.
Asian Paints and Pidilite annual reports across the FY21–FY23 crude-derivative spike — a live study of pricing power in the gross margin (see case I2).
Screener.in "Profit & Loss" tab for any company — the OPM row is the EBITDA margin excluding other income; check it against your own common-size table.

← Previous: 04.1 Growth analysis · Module index · Next: 04.3 Returns on capital →