04.3 · Returns on capital: ROE, ROCE, ROIC¶

Why this matters: a business is a machine that turns capital into profit. Growth and margins describe the output; return on capital describes the efficiency of the machine — how much profit each rupee tied up in it produces. It is the single number that most separates businesses worth paying up for from businesses that merely get bigger, because growth only creates value when the return on the new capital exceeds its cost.

Learning objectives — after this lesson you can:

Define and compute ROE, ROCE and ROIC correctly, including what goes in the denominator, why averages are used, and how cash, goodwill, leases and CWIP are treated.
Decompose ROE with the 3-step and 5-step DuPont identities and say which lever moved.
Explain why a high ROE can be a leverage artefact, and why ROIC vs WACC is the value-creation test.
Compute incremental ROIC and the reinvestment identity (growth = reinvestment rate × ROIC).
Apply all of this to Kaveri FY21–FY26 and explain the rise to FY24 and the decline since.

Prerequisites: 02.4 The balance sheet, 04.2 Margins & cost structure · Time: ~100 min

1. Three ratios, three questions¶

All return ratios have the same shape — a profit measure over the capital that generated it — and the art is matching numerator to denominator. Each answers a different question:

Ratio	Numerator	Denominator	Question it answers	Who uses it
ROE — return on equity	PAT (attributable to owners)	Average shareholders' equity	What do shareholders earn on the book value they have in the business, after everyone else is paid?	Equity investors, bank analysts (it is the bank metric)
ROCE — return on capital employed	EBIT (usually + other income; pre-tax)	Average capital employed = equity + all borrowings (+ lease liabilities)	What does the whole business earn on all long-term money, before tax and financing?	Indian sell-side and management (the dominant Indian convention)
ROIC — return on invested capital	NOPAT = EBIT × (1 − tax rate)	Average invested capital = net operating assets (fixed assets + net working capital), excluding cash and financial investments	What does the operating business earn, after tax, on the capital actually invested in operations? Compare directly with WACC.	Valuation work, DCF, quality investors

The relationship: ROE is the shareholder's view (after leverage and tax); ROCE is the lender-plus-shareholder view (before tax); ROIC is the operating view (after tax, excluding financial assets). All three should be computed on average capital (opening + closing)/2 because the profit was earned over the year while the balance sheet is a year-end snapshot; a fast-growing company's closing capital overstates what was available on average.

2. Getting the denominators right¶

This is where most published numbers go wrong. For Kaveri at 31-Mar-2026 (₹ Cr, from the running example):

Component	₹ Cr	Equity	Capital employed	Invested capital
Total equity	706.1	✓	✓	—
Non-current borrowings	92.0		✓	—
Current borrowings	96.0		✓	—
Lease liabilities	18.2		✓	—
Capital employed	912.3
Net block (PP&E)	495.3			✓
CWIP	2.0			✓
Right-of-use assets	16.5			✓
Intangibles	8.0			✓
Inventories + receivables + other current assets	574.3			✓
− Trade payables − other current liabilities	(209.3)			✓
Invested capital	886.8
Cash + current investments (excluded from IC)	47.5			financing side
Deferred tax liability (excluded; quasi-equity)	22.0			—

Check: invested capital 886.8 + cash 47.5 = 934.3 = capital employed 912.3 + DTL 22.0. The two views reconcile because sources (equity, debt, leases, DTL) must equal uses (operating assets plus cash). This reconciliation is your error check whenever you build these numbers.

The judgement calls, and the course's conventions:

Cash and investments are excluded from invested capital (they earn other income, which is excluded from NOPAT). Some analysts subtract only "excess" cash; the course excludes all financial assets and treats other income separately. ROCE, by Indian convention, includes cash in capital employed and other income in the numerator — consistent, just different.
Goodwill and acquired intangibles: include them (the company paid that money) for judging management's capital allocation; exclude them for judging the operating business's economics. Report both when goodwill is large. Kaveri has none.
CWIP: capital that is not yet producing. Include it (it is invested) but recognise that ROIC is depressed during a build. Kaveri's FY23–24 CWIP of ₹68 Cr → ₹0 as Hosur was commissioned is a good example: FY24 ROIC benefits from a plant that only started depreciating in FY25.
Leases: since Ind AS 116, ROU assets are in invested capital and lease liabilities in capital employed; EBIT is after ROU amortisation but before lease interest — consistent with treating leases as debt.
Deferred tax liabilities: quasi-equity (the company keeps the money until timing differences reverse). Exclude from capital employed or treat as part of equity; do not treat as debt.
Taxes: NOPAT uses the marginal or effective tax rate on operating profit — 25.17% for Kaveri.

3. Kaveri FY21–FY26: the numbers¶

Computed with tools/fi/ratios.ratio_dashboard (which reproduces the running-example table). Averages use the opening balance sheet at 31-Mar-2020 for FY21.

	FY21	FY22	FY23	FY24	FY25	FY26
PAT (₹ Cr)	32.8	47.2	62.6	86.4	97.4	90.5
EBIT + other income	58.1	69.4	92.0	129.0	132.5	138.0
NOPAT = EBIT × 0.7483	39.7	47.3	63.3	91.4	95.8	100.3
Avg equity	385.9	416.9	461.3	522.7	598.0	671.7
Avg capital employed	463.3	493.0	560.0	672.6	780.3	868.1
Avg invested capital	394.9	408.1	470.9	607.0	738.8	836.8
ROE	8.5%	11.3%	13.6%	16.5%	16.3%	13.5%
ROCE (pre-tax)	12.5%	14.1%	16.4%	19.2%	17.0%	15.9%
ROIC (post-tax)	10.1%	11.6%	13.4%	15.1%	13.0%	12.0%

from fi.data import load_kaveri
from fi.ratios import ratio_dashboard
dash = ratio_dashboard(load_kaveri())
print(dash.loc[['roe', 'roce', 'roic']].round(4))

Three observations:

The FY24 peak. ROIC of 15.1% against a WACC of about 12.2% (06.2) — a spread of ~3 pp. Kaveri was creating value, modestly.
The decline since. By FY26, ROIC of 12.0% is at WACC. The company is growing 12–16% a year but the new capital — mostly receivables and a motors plant running at 57% utilisation — earns roughly its cost. Growth like that adds size, not value (06.1 proves this).
ROE flatters. ROE (13.5%) is above ROIC (12.0%) because Kaveri carries ₹140 Cr of net debt at an after-tax cost (~6.5%) below its operating return. That is leverage doing the work, not the business getting better.

4. DuPont: which lever moved?¶

4.1 Three-step DuPont¶

\[\text{ROE} = \underbrace{\frac{\text{PAT}}{\text{Revenue}}}_{\text{net margin}} \times \underbrace{\frac{\text{Revenue}}{\text{Avg assets}}}_{\text{asset turnover}} \times \underbrace{\frac{\text{Avg assets}}{\text{Avg equity}}}_{\text{leverage}}\]

	FY21	FY22	FY23	FY24	FY25	FY26
Net margin	5.4%	6.2%	7.2%	8.6%	8.3%	6.9%
Asset turnover (on avg assets)	1.06	1.22	1.23	1.20	1.20	1.21
Leverage (avg assets / avg equity)	1.50	1.49	1.53	1.61	1.63	1.62
ROE	8.5%	11.3%	13.6%	16.5%	16.3%	13.5%

from fi.ratios import dupont_3
dupont_3(pat=90.5, revenue=1318.0, avg_assets=1089.7, avg_equity=671.7)
# {'margin': 0.0687, 'turnover': 1.2095, 'leverage': 1.6224, 'roe': 0.1348}

The story is unambiguous: asset turnover has been flat at ~1.2 since FY22 and leverage crept up from 1.5 to 1.6; everything else is margin. ROE rose with the net margin to FY24 and fell with it after. Kaveri is not becoming a more capital-efficient business; it is a margin story wearing a capital-efficiency costume. That tells you where to spend your research time — on the solar mix and receivables, not on the plant.

4.2 Five-step DuPont¶

The five-step version splits the net margin into tax, interest and operating components so that financing and tax effects are visible:

\[\text{ROE} = \underbrace{\frac{\text{PAT}}{\text{PBT}}}_{\text{tax burden}} \times \underbrace{\frac{\text{PBT}}{\text{EBIT}}}_{\text{interest burden}} \times \underbrace{\frac{\text{EBIT}}{\text{Revenue}}}_{\text{EBIT margin}} \times \frac{\text{Revenue}}{\text{Avg assets}} \times \frac{\text{Avg assets}}{\text{Avg equity}}\]

(EBIT here includes other income so that PBT/EBIT captures only interest and exceptional items.)

	FY24	FY25	FY26	Comment
Tax burden (PAT/PBT)	0.748	0.748	0.748	Constant 25.2% tax — nothing to see
Interest burden (PBT/EBIT)	0.895	0.983	0.877	FY25 inflated by the ₹14 Cr exceptional gain; FY26 shows finance costs taking 12% of EBIT
EBIT margin (incl. other income)	12.8%	11.3%	10.5%	The real driver
Asset turnover	1.20	1.20	1.21	Flat
Leverage	1.61	1.63	1.62	Flat
ROE	16.5%	16.3%	13.5%	0.748 × 0.877 × 0.1047 × 1.2095 × 1.6224 = 13.5%

fi.ratios.dupont_5(pat, pbt, ebit, revenue, avg_assets, avg_equity) returns all five factors.

India notes

Indian brokers and Screener.in lead with ROCE, usually defined as EBIT ÷ (equity + debt) on closing balances, sometimes including other income, sometimes not. Screener's ROCE for a company can differ from your average-based number by 1–3 pp in a fast-growing year. Management often quotes "ROCE" on pre-tax EBIT, which makes 20% sound better than the 15% post-tax ROIC it implies. Always restate to one definition before comparing companies — fi.ratios.roce and fi.ratios.roic are the course's definitions.

5. Why a high ROE can be a leverage artefact¶

Financial leverage raises ROE whenever the after-tax return on assets exceeds the after-tax cost of debt:

\[\text{ROE} = \text{ROA} + (\text{ROA} − k_d(1−t)) \times \frac{D}{E}\]

With ROA (NOPAT/assets) of 12% and after-tax debt cost of 6.75%:

D/E	0.0	0.5	1.0	2.0
ROE	12.0%	14.6%	17.3%	22.5%

The 22.5% ROE at D/E = 2 is the same 12% business with more risk; if ROA drops to 5% in a bad year, ROE at D/E = 2 becomes 5 + (5 − 6.75) × 2 = 1.5%, and at D/E = 4 it is negative. This is why a 20% ROE from an NBFC (leverage 6–8x) and a 20% ROE from an FMCG company (no debt) are not comparable, and why for non-financial companies you judge the business on ROIC and the financing choice separately.

Trader's lens

Leverage is a constant-notional position financed with borrowed money: it scales the P&L in both directions and adds a fixed carry (interest). ROE at high D/E is the return on the margin posted, not on the exposure. A lender's 15% ROE at 7x leverage is a ~2% return on assets — a thin edge that a small credit-cost surprise wipes out, exactly like a high-Sharpe strategy run at 7x. 07.1 builds the RoA tree that makes this precise for Nirmal Finance.

6. ROIC vs WACC: the value-creation test¶

A business creates value only when it earns more on invested capital than the capital costs. The spread ROIC − WACC, multiplied by invested capital, is the annual economic profit (EVA):

\[\text{Economic profit} = (\text{ROIC} − \text{WACC}) \times \text{Invested capital}\]

Kaveri FY26: (12.0% − 12.19%) × 836.8 ≈ −₹1.6 Cr — essentially zero. FY24: (15.1% − 12.19%) × 607.0 ≈ +₹17.6 Cr. The company went from creating ₹18 Cr of value a year to none, while PAT grew from 86 to 90. That divergence is invisible in the P&L and is the entire reason this ratio exists.

The corollary, proved in 06.1: when ROIC = WACC, growth is worth nothing; when ROIC < WACC, growth destroys value; only when ROIC > WACC does faster growth deserve a higher multiple.

7. Incremental ROIC and the reinvestment identity¶

Average ROIC blends old, depreciated, cheap assets with new ones. What matters for the future is the return on the next rupee — incremental ROIC:

\[\text{Incremental ROIC} = \frac{\Delta \text{NOPAT}}{\Delta \text{Invested capital}}\]

Window	ΔNOPAT (₹ Cr)	ΔInvested capital (₹ Cr)	Incremental ROIC
FY23 → FY26 (three years)	100.3 − 63.3 = 37.0	886.8 − 523.4 = 363.4	10.2%
FY24 → FY26 (two years)	100.3 − 91.4 = 9.0	886.8 − 690.6 = 196.2	4.6%

Since the FY24 peak, Kaveri has invested ₹196 Cr of new capital (₹77 Cr of it in receivables) and earned an extra ₹9 Cr of NOPAT on it — a 4.6% return, far below WACC. This is the number that should worry a shareholder more than any margin: the marginal rupee is being wasted, and management guides to more of the same growth. Use 3-year windows at least; single years are dominated by timing (a plant capitalised, a working-capital swing).

Finally, the identity that links returns to growth — the engine of every DCF:

\[g = \text{Reinvestment rate} \times \text{ROIC}, \qquad \text{Reinvestment rate} = \frac{\text{Capex} + \Delta\text{NWC} − \text{D\&A}}{\text{NOPAT}}\]

Kaveri FY26: net reinvestment = (52.0 + 89.7 − 47.8) = ₹93.9 Cr, which is 94% of NOPAT (100.3). At a 12% ROIC that supports g ≈ 0.94 × 12% ≈ 11% — close to the ~12.5% actually delivered. But the reinvestment is 70% working capital (receivables from state agencies), not plant, and it consumed nearly all of NOPAT, which is why free cash flow was just ₹13 Cr on ₹90 Cr of profit (04.4). A business that must reinvest 94% of its profit to grow 12% at a 12% return is running to stand still; one that reinvests 30% to grow 12% is earning 40% on new capital and can pay you the rest. That comparison — reinvestment needed per unit of growth — is the practical definition of business quality.

8. A second worked example — two companies, same ROE¶

Vindhya Consumer Ltd and Narmada Steel Processors Ltd (both fictional), FY26, ₹ Cr:

	Vindhya Consumer	Narmada Steel
Revenue	2,000	2,000
EBIT	400	180
Tax rate	25%	25%
Avg invested capital	1,000	1,800
of which cash & investments (excluded)	600	20
Avg equity	1,500	600
Net debt	(600)	1,200
PAT	330	110

t = 0.25
v = dict(ebit=400, ic=1000, eq=1500, pat=330); n = dict(ebit=180, ic=1800, eq=600, pat=110)
for name, c in (('Vindhya', v), ('Narmada', n)):
    print(name, 'ROIC', round(c['ebit']*(1-t)/c['ic'], 3), 'ROE', round(c['pat']/c['eq'], 3))
# Vindhya ROIC 0.30 ROE 0.22
# Narmada ROIC 0.075 ROE 0.183

Both report ROEs around 18–22%. Vindhya's is depressed by ₹600 Cr of idle cash sitting in equity (its operating ROIC is 30%); Narmada's is inflated by ₹1,200 Cr of debt on a 7.5% ROIC business. Vindhya could double its ROE by returning cash; Narmada's ROE turns negative in a steel down-cycle. Any comparison on ROE alone gets this backwards.

Common mistakes

Closing instead of average capital (overstates growth-year returns; understates them when capital shrinks).
Mixing a pre-tax numerator (EBIT) with an after-tax comparison (WACC). ROCE compares to pre-tax WACC; ROIC to WACC.
Including other income in NOPAT while excluding cash from invested capital (or the reverse).
Judging a company mid-capex on average ROIC — CWIP earns nothing until commissioned; look at incremental ROIC over a full cycle.
Comparing ROE across different leverage without a DuPont.
Taking management's "ROCE" at face value — check numerator (pre/post tax, with/without other income) and denominator (closing/average; with/without leases and cash).
Ignoring goodwill when judging acquisitive management: the price paid is capital too.

Key terms¶

Term	Meaning
ROE	PAT ÷ average shareholders' equity
ROCE	(EBIT + other income) ÷ average capital employed (equity + borrowings + lease liabilities); pre-tax
ROIC	NOPAT ÷ average invested capital; post-tax, operating assets only
NOPAT	Net operating profit after tax = EBIT × (1 − tax rate)
Capital employed	Equity + all borrowings + lease liabilities (the sources view)
Invested capital	Net fixed assets + CWIP + ROU + intangibles + net working capital (the uses view, ex-cash)
DuPont analysis	Decomposition of ROE into margin × turnover × leverage (3-step) or with tax and interest burdens (5-step)
Asset turnover	Revenue ÷ average total assets; capital intensity's inverse
Financial leverage (DuPont)	Average assets ÷ average equity
Economic profit (EVA)	(ROIC − WACC) × invested capital
Incremental ROIC	Change in NOPAT ÷ change in invested capital over a multi-year window
Reinvestment rate	(Capex + ΔNWC − D&A) ÷ NOPAT; the share of profit ploughed back
Reinvestment identity	Sustainable growth g = reinvestment rate × ROIC

Check your understanding¶

Kaveri's ROCE (15.9%) is higher than its ROIC (12.0%) in FY26. Give the two definitional reasons.

Answer

ROCE is pre-tax (EBIT + other income) while ROIC uses NOPAT after 25.17% tax — the larger effect; and the denominators differ (capital employed includes cash-funded capital and excludes DTL; invested capital excludes cash). Roughly: 15.9% × 0.748 ≈ 11.9%, close to the 12.0% ROIC once the denominator difference nets out.

Using the 3-step DuPont, what would Kaveri's FY26 ROE have been with the FY24 net margin (8.6%) and everything else unchanged?

Answer

0.086 × 1.2095 × 1.6224 = 16.9%. The entire ROE decline is the margin.

A company reports ROE of 25% with D/E of 3.0 and after-tax cost of debt of 6%. What is its ROA, and what happens to ROE if ROA falls by 3 pp?

Answer

25 = ROA + (ROA − 6) × 3 → 25 = 4·ROA − 18 → ROA = 10.75%. If ROA falls to 7.75%: ROE = 7.75 + (7.75 − 6) × 3 = 13.0%. A 3-pp fall in ROA cuts ROE by 12 pp.

Compute Kaveri's FY26 economic profit if you use ROCE-style pre-tax numbers with a pre-tax WACC of 16.3% (12.19% ÷ 0.7483).

Answer

(15.9% − 16.3%) × 868.1 ≈ −₹3.5 Cr — the same conclusion (≈ zero value creation) as the post-tax version, which is the point of matching pre-tax with pre-tax.

Why is the FY24→FY26 incremental ROIC (4.6%) so much lower than the average ROIC (12.0%)?

Answer

The ₹196 Cr of new capital went mostly into receivables (₹77 Cr) and inventory while the Hosur plant runs at 57% utilisation; the old asset base still earns well, but the marginal rupee earns little. Average ROIC is history; incremental ROIC is the trend.

A company grows revenue and NOPAT at 20% a year with a reinvestment rate of 50%. What ROIC is it earning on new capital? If instead it needs to reinvest 120% of NOPAT, what does that imply?

Answer

g = RR × ROIC → ROIC = 20% / 50% = 40% — a very high-quality business. At 120% reinvestment the implied ROIC is 16.7% and the company is consuming more than it earns (negative FCF, must raise debt or equity to grow) — fine only if 16.7% comfortably exceeds WACC and the funding is sustainable.

Go deeper¶

McKinsey & Company, Valuation: Measuring and Managing the Value of Companies — chapters on ROIC and the value-driver formula; the canonical treatment of invested capital.
Michael Mauboussin, Return on Invested Capital: How to Calculate ROIC and Handle Common Issues (Credit Suisse, 2014) — free, practical, deals with goodwill, leases, cash and R&D.
Screener.in's ratio definitions page — read exactly how its ROCE and ROE are computed before you compare.
Warren Buffett's 1979 Berkshire letter — on why return on equity capital, not EPS growth, is the test of management.

← Previous: 04.2 Margins & cost structure · Module index · Next: 04.4 Working capital & cash conversion →