04.6 · Per-share metrics & building a ratio dashboard

Why this matters: you own a slice of a company, not the company, so a company-level number only matters to you after it has been divided by the share count. And one ratio on its own tells you almost nothing. Thirty ratios across six years, read in the right order, tell you a company's story in ten minutes. This lesson builds both habits: per-share thinking, and a dashboard you read from top to bottom.

Learning objectives — after this lesson you can:

• Compute and reconcile basic, diluted, adjusted, trailing-twelve-month and "cash" EPS, and say which one a quoted P/E is built on.
• Break book-value-per-share growth into retained profit, ESOP credits and share issuance, and link EPS growth to BVPS growth and ROE.
• Compute DPS, payout ratio (paid basis and declared basis), dividend yield, buyback yield and shareholder yield, and spot a buyback that only offsets ESOP dilution.
• Build Kaveri's six-year, 28-row ratio dashboard with tools/fi/ratios.py and read it from top to bottom as one causal story.
• Build a peer dashboard on consistent definitions and say what it cannot tell you.
• Reproduce Screener.in's headline numbers from an annual report and name the definitional quirks that make them differ from yours.

Prerequisites: 01.2 Shares, market cap & enterprise value, 04.3 Returns on capital, 04.4 Working capital & cash conversion, 04.5 Leverage, solvency & liquidity · Time: ~120 min

---

1. Per-share thinking, and the engine that links the per-share numbers

01.2 defined earnings per share (EPS), profit attributable to equity holders divided by the weighted-average share count, and book value per share (BVPS), equity attributable to owners divided by period-end shares. It also showed that "profits up 60%" can mean "EPS down 8.6%" once an acquisition is paid for in shares. This lesson treats the per-share numbers as a system. They are linked by identities, and the identities are what let you read them together.

For a constant share count \(N\), write opening equity as \(E_{t-1}\):

\[ \text{EPS}_t=\frac{\text{PAT}_t}{N}=\underbrace{\frac{\text{PAT}_t}{E_{t-1}}}_{\text{ROE on opening equity}}\times\underbrace{\frac{E_{t-1}}{N}}_{\text{opening BVPS}} \]

In words: this year's EPS is last year's book value per share times the return earned on it. Three more links complete the engine:

• DPS = payout ratio × EPS. DPS is dividend per share; the payout ratio is the share of profit paid out.
• BVPS grows by retained EPS, plus anything that adds equity (share issues, ESOP credits) and minus anything that removes it (buybacks).
• Price = P/E × EPS = P/B × BVPS. The market's multiples sit on top of the accounting engine.

flowchart LR
  B0["Opening BVPS"] -->|"× ROE"| EPS["EPS"]
  EPS -->|"× payout"| DPS["DPS (cash to you)"]
  EPS -->|"× (1 − payout)"| RET["Retained EPS"]
  RET --> B1["Closing BVPS"]
  ISS["Share issues, ESOP credits (+)<br/>Buybacks (−)"] --> B1
  B1 -->|"next year"| B0
  EPS -->|"× P/E"| P["Price"]
  B1 -->|"× P/B"| P

The engine gives the decomposition you will use most:

\[ 1+g_{\text{EPS}}=\big(1+g_{\text{opening BVPS}}\big)\times\frac{\text{ROE}_t}{\text{ROE}_{t-1}} \]

EPS can only grow if book value grows (by retaining profit) or if the return on that book rises. A company that keeps growing EPS faster than BVPS must be raising its ROE. That can't go on for ever, so find out why it is happening: better margins, faster asset turns, more leverage (04.3), or buybacks shrinking \(N\).

2. EPS: four honest versions and one misleading one

Version	Numerator	Denominator	Where you find it	Use it for
Basic EPS	PAT attributable to owners of the parent	Weighted-average shares outstanding	Face of the P&L (Schedule III requires it) and the EPS note	The reported number; trailing P/E
Diluted EPS	Same, adjusted for convertibles' interest	Basic + dilutive potential shares (Ind AS 33)	Same	A conservative accounting per-share figure
Adjusted EPS	PAT excluding one-offs, post-tax	Basic (or diluted)	Your own work; sometimes the company's presentation	Comparing years; forecasting
TTM EPS	Sum of the last four quarters' PAT	Shares	Quarterly results (03.4)	Up-to-date P/E for a seasonal business
Forward EPS	Forecast PAT (yours or consensus)	Shares	Broker notes, data vendors	Forward P/E; thin for Indian small caps
"Cash EPS"	PAT + depreciation & amortisation	Shares	Indian broker reports	Almost nothing (see §2.4)

2.1 Basic and diluted

Section 6 of 01.2 covers the mechanics: weighted-average counts, the treasury-stock method and anti-dilution. For Kaveri the tools market data (fi.data.kaveri_market_data()) carries a diluted count for each year after the FY24 ESOP grant: 6.02 Cr (FY24), 6.05 Cr (FY25), 6.07 Cr (FY26) against 6.00 Cr basic. The gap is about 1%, small enough to ignore for most purposes. For a company whose ESOP pool is 5–10% of equity it is not (07.4).

2.2 Adjusted EPS

Adjusted EPS removes items you judge to be non-recurring, after tax. Kaveri has two:

\[ \text{Adj. EPS}_{FY25}=\frac{97.4-14.0\times(1-0.2517)}{6.00}=\frac{86.92}{6.00}=₹14.49\quad(\text{reported } ₹16.23) \]

\[ \text{Adj. EPS}_{FY21}=\frac{32.8+7.5\times(1-0.2517)}{6.00}=\frac{38.41}{6.00}=₹6.40\quad(\text{reported } ₹5.47) \]

Three rules keep adjusted EPS honest.

1. Adjust symmetrically. Take out one-off gains as readily as one-off losses. Management presentations tend to remove only the losses. The FY25 land gain flatters reported EPS by ₹1.74; an "adjusted" number that kept it would be pure marketing.
2. Tax-effect the adjustment. A ₹14.0 Cr pre-tax gain is worth ₹10.5 Cr to shareholders at 25.17%. The course uses the normal corporate rate for simplicity. In reality a land gain may be taxed as a capital gain at a different rate, so check the tax note (03.3).
3. Use the same share count as the reported figure. IAS 33 (paragraph 73) lets a company present extra per-share amounts based on another component of profit, provided it uses the standard's weighted-average share count, shows basic and diluted amounts with equal prominence in the notes, and states the basis of the numerator. Ind AS 33 follows IAS 33 here. We could not retrieve the MCA-notified Ind AS 33 text online on 21-Sep-2026, so check the notified standard if the point matters (IFRS Foundation, IAS 33).

Whether an item really is non-recurring is a judgement. 04.7 gives the tests.

2.3 TTM and forward EPS

TTM (trailing twelve months) EPS adds up the last four reported quarters. After Kaveri's Q1 FY27 results on 8-Aug-2026, TTM PAT was Q2 FY26 + Q3 FY26 + Q4 FY26 + Q1 FY27 = 12.7 + 17.1 + 34.0 + 22.3 = ₹86.1 Cr, so TTM EPS was ₹14.35, below the FY26 figure of ₹15.08 because Q1 FY27 profit fell 16.5% year on year. For a seasonal business TTM is the right trailing measure: it always holds one of each quarter, whereas annualising a single quarter (×4) builds the season into the answer.

Forward EPS is a forecast, and a forecast is only as good as whoever made it. For large caps, consensus (the average of analysts' forecasts) is a usable anchor. For small caps like Kaveri, coverage is often one or two brokers or nobody (03.4). Then the forward P/E is your own forecast dressed up as a market number.

2.4 "Cash EPS": the one to distrust

Indian broker reports often quote cash EPS = (PAT + depreciation & amortisation) ÷ shares, sometimes adding deferred tax and other non-cash charges. It adds back the cost of using up assets but never subtracts the cash spent replacing them (capex), and it ignores working capital completely. Cash EPS relates to free cash flow the way "adjusted EBITDA" relates to profit. It is always the flattering version.

Worked example 1 — the Kaveri EPS ladder, and six P/Es for one price

All ₹ per share; shares 6.00 Cr basic; diluted 6.02 / 6.05 / 6.07 Cr; tax 25.17%.

₹ per share	FY24	FY25	FY26	Construction
Basic EPS	14.40	16.23	15.08	PAT ÷ 6.00
Diluted EPS	14.35	16.10	14.91	PAT ÷ diluted count
Adjusted EPS (basic)	14.40	14.49	15.08	(PAT − exceptional × 0.7483) ÷ 6.00
Adjusted diluted EPS	14.35	14.37	14.91	same ÷ diluted count
"Cash EPS"	19.88	23.62	23.05	(PAT + D&A) ÷ 6.00
Free cash flow per share	(10.73)	0.48	2.18	(CFO − capex) ÷ 6.00

Growth. Reported EPS fell 7.1% in FY26 (16.23 → 15.08); adjusted EPS rose 4.1% (14.49 → 15.08). Over two years the two agree, because FY24 and FY26 have no one-offs: +4.7% in total, a 2.3% compound annual growth rate, while revenue grew 31.0%.

P/E at ₹390. The same price gives six different multiples:

P/E built on	EPS, ₹	P/E
FY26 basic EPS (what the peer table uses)	15.08	25.9x
FY26 diluted EPS	14.91	26.2x
TTM EPS to Q1 FY27	14.35	27.2x
FY26 "cash EPS"	23.05	16.9x
FY26 FCF per share (P/FCF)	2.18	178.6x

The one to remember is the last. On FY26 cash, ₹390 buys a 0.56% free-cash-flow yield (13.1 ÷ 2,340). The 16.9x "cash P/E" looks cheap only because it ignores the ₹52.0 Cr of capex and the ₹89.7 Cr of working capital that the business consumed. Whenever someone quotes a P/E, ask which EPS, and for how many shares.

2.5 Bonus issues and splits: restate the history

A 1:1 bonus issue (one free share for each share held) doubles the share count and changes nothing economically (01.3). EPS for all periods shown is restated retrospectively, as if the new count had always existed (IAS 33 paragraph 64, which also covers bonuses and splits between the year-end and the date the accounts are approved). If Kaveri made a 1:1 bonus in FY27, its FY26 EPS would be re-presented as ₹7.54, BVPS as ₹58.84 and DPS paid as ₹2.00. The share price would halve to about ₹195, and the P/E would stay at 25.9x. The trap is in your spreadsheet: mixing a pre-bonus EPS history with a post-bonus price, or the reverse. Aggregators generally restate past per-share data; old annual reports and your own notes do not.

3. Book value per share and its growth

BVPS is equity attributable to owners ÷ period-end shares. Tangible BVPS also deducts goodwill and other intangible assets, which a lender or acquirer might not treat as capital. Kaveri's only intangibles are ₹8.0 Cr of software, so its FY26 tangible BVPS is (706.1 − 8.0) ÷ 6.00 = ₹116.35, barely below ₹117.68. For an acquisitive company or a bank the gap can be large (P/ABV for lenders is in 07.1).

Book value per share is built by an exact roll-forward:

\[ E_t = E_{t-1} + \text{PAT}_t - \text{Dividends}_t + \text{Share issues}_t - \text{Buybacks}_t + \text{ESOP credit}_t \pm \text{OCI}_t \]

The ESOP credit is the share-based payment expense credited back to equity (the expense reduced PAT; the credit adds it back to reserves, so on its own it leaves equity unchanged, but it shows the value handed to employees). OCI (other comprehensive income) is gains and losses that bypass the P&L, such as actuarial losses on gratuity or translation differences (02.8).

Worked example 2 — Kaveri's book value, rolled forward and decomposed

Opening equity at 31-Mar-2020 was ₹374.0 Cr (share capital 30.0 + other equity 344.0). Kaveri issued no shares, bought none back and has no OCI in the reference data, so the roll-forward has only three moving parts (₹ Cr):

₹ Cr	FY21	FY22	FY23	FY24	FY25	FY26
Opening equity	374.0	397.8	436.0	486.6	558.8	637.2
+ PAT	32.8	47.2	62.6	86.4	97.4	90.5
− Dividends paid	(9.0)	(9.0)	(12.0)	(15.0)	(21.0)	(24.0)
+ ESOP credit (share-based payment)	0.0	0.0	0.0	0.8	2.0	2.4
Closing equity	397.8	436.0	486.6	558.8	637.2	706.1
BVPS (6.00 Cr shares), ₹	66.30	72.67	81.10	93.13	106.20	117.68
Book-value growth	6.4%	9.6%	11.6%	14.8%	14.0%	10.8%
ROE on opening equity	8.8%	11.9%	14.4%	17.8%	17.4%	14.2%
Retention (1 − dividends ÷ PAT)	72.6%	80.9%	80.8%	82.6%	78.4%	73.5%
ROE × retention	6.4%	9.6%	11.6%	14.7%	13.7%	10.4%

Every column ties to the balance sheet exactly. Book-value growth equals ROE × retention, plus the ESOP credit (FY26: 10.4% + 2.4 ÷ 637.2 = 10.4% + 0.4% = 10.8%). Note that the ROE here is on opening equity, which is why it is higher than the dashboard's ROE on average equity (14.2% vs 13.5% in FY26).

EPS vs BVPS. From FY21 to FY26 basic EPS compounded at 22.5% a year (₹5.47 → ₹15.08) but BVPS at only 12.2% (₹66.30 → ₹117.68). By the identity in §1, the gap had to come from ROE, which rose from 8.8% to 14.2% on opening equity. FY26 shows the identity working in reverse:

\[ 1+g_{\text{EPS}} = \frac{106.20}{93.13}\times\frac{14.20\%}{17.43\%} = 1.140\times0.815 = 0.929 \;\Rightarrow\; -7.1\% \]

Opening book value grew 14.0%, but ROE fell by a fifth, so EPS fell 7.1%. (FY25's ROE was flattered by the land gain. On adjusted numbers the FY26 fall in ROE is milder.)

Book-value growth as a speed limit. \(g = \text{ROE} \times b\), where \(b\) is the retention ratio, is the sustainable growth rate: the rate at which equity, and so a business with constant asset intensity and leverage, can grow without new shares or extra borrowing. Kaveri's FY26 figure is 10.4% (10.8% with the ESOP credit), but revenue grew 12.5% and working capital grew much faster. The gap was filled with debt: net debt rose ₹28.5 Cr in FY26 (112.0 → 140.5). When growth runs ahead of ROE × retention, look at the balance sheet to see who is paying for it (04.5).

Share issues move book value per share too. Nirmal Finance's FY24 QIP issued 1.0 Cr shares at ₹300, which was 3.5 times its opening BVPS of ₹85.3. Its equity rolled forward as 853.4 + 169.0 − 20.0 + 300.0 = ₹1,302.4 Cr. Without the QIP, BVPS would have been (853.4 + 169.0 − 20.0) ÷ 10.0 = ₹100.24. With it, BVPS was 1,302.4 ÷ 11.0 = ₹118.40. Selling shares well above book lifted every existing holder's book value by ₹18.16 while diluting their EPS (₹16.90 on 10.0 Cr shares vs ₹15.36 on 11.0 Cr). An issue below book does the opposite. For a lender, whose value is anchored on book, this is a first-order effect (07.1).

Book value is a scorecard with a known flaw. For decades Warren Buffett opened Berkshire Hathaway's letter with the change in per-share book value. In his 2018 letter he dropped the practice, calling it "a metric that has lost the relevance it once had". His reasons generalise: operating businesses sit in book value at far below their worth, and buybacks above book but below intrinsic value lower BVPS while raising intrinsic value per share. For a lender or a holding company, BVPS growth plus dividends is still a good scorecard. For a capital-light compounder it understates value creation, and ROE and the P/B multiple carry the message instead (justified P/B = (ROE − g)/(r − g) is derived in 06.5).

4. Cash returned to shareholders: DPS, payout and yields

4.1 DPS and the timing trap

Dividend per share (DPS) comes in two flavours, and mixing them produces nonsense:

• Declared for the year: interim dividends paid during the year plus the final dividend the board recommends after the year-end, which shareholders approve at the AGM months later. This is what companies announce and what Screener's "Dividend Payout %" uses (§8).
• Paid in the year: cash that left in the financial year, which is what the cash-flow statement shows. For a company that pays mostly final dividends, this is largely last year's final dividend.

Kaveri's dashboard row is DPS paid in the year. Look at what happens to the payout ratio under each lens:

₹ Cr	FY21	FY22	FY23	FY24	FY25	FY26
Dividends paid	9.0	9.0	12.0	15.0	21.0	24.0
PAT, same year	32.8	47.2	62.6	86.4	97.4	90.5
Payout, paid ÷ same-year PAT	27.4%	19.1%	19.2%	17.4%	21.6%	26.5%
Payout, paid ÷ previous-year PAT	–	27.4%	25.4%	24.0%	24.3%	24.6%

The same-year ratio swings from 17.4% to 27.4%. The lagged ratio sits between 24.0% and 25.4% from FY23 onwards, in line with management's stated "~25%" payout. We infer that Kaveri pays roughly a quarter of each year's profit as a final dividend in the following year. The reference data don't say so explicitly, but the numbers fit that and nothing else. The general lesson: before computing a payout ratio, work out when the dividend relating to a year's profit is actually paid.

4.2 Payout, and what pays for it

The payout ratio can be measured against PAT, adjusted PAT, or free cash flow. The last is the one that tells you whether the dividend is earned in cash:

• FY21–FY26 cumulative: dividends ₹90.0 Cr against PAT of ₹416.9 Cr, a 21.6% payout, which looks conservative.
• Against cumulative free cash flow of −₹20.3 Cr (02.5, worked example 1), every rupee of dividend was borrowed or came from the treasury.
• FY26 alone: dividends ₹24.0 Cr, FCF ₹13.1 Cr, so FCF covered 0.55x the dividend.

Meanwhile DPS paid rose 60% from FY24 to FY26 (₹2.50 → ₹4.00), against EPS growth of 4.7%. A rising dividend reads as confidence. Funded by working-capital debt, it is a choice worth questioning (05.5).

4.3 Dividend yield

Dividend yield = DPS ÷ price. Trailing yield uses the last twelve months' DPS; forward yield uses expected DPS. Kaveri's trailing yield on 31-March prices was 0.71%, 0.47%, 0.56%, 0.41%, 0.45% and 0.77% for FY21–FY26 (₹1.50 ÷ ₹210 … ₹4.00 ÷ ₹520). At ₹390 it is 1.03%, or 0.71% after tax for a shareholder paying 31.2% (01.3 covers dividend tax as of Sep-2026). Compare that with the 6.5% ten-year G-sec yield in the reference valuation. For a stock like this the dividend is a rounding error. The return has to come from growth in value per share. A high dividend yield matters for mature, low-growth companies, where it is most of the return, and there you check it against FCF, not PAT.

4.4 Buyback yield, net issuance and shareholder yield

Companies also return cash by buying back shares, and they take cash (or give value) back by issuing shares to employees, acquirees or investors. Three measures capture the net effect:

\[ \text{Buyback yield}=\frac{\text{Cash spent on buybacks}}{\text{Market cap}},\qquad \text{Shareholder yield}=\frac{\text{Dividends}+\text{Buybacks}-\text{Cash from share issues}}{\text{Market cap}} \]

and the one that cannot be dressed up: the net change in the share count. Kaveri has done no buybacks, so we use a fictional company.

Worked example 3 — Tapti Infotech (fictional): a 3% buyback that shrank the share count 1.7%

Tapti starts the year with 100.0 Cr shares at ₹400 (market cap ₹40,000 Cr) and earns PAT of ₹2,400 Cr (EPS ₹24.00). During the year it:

• pays a dividend of ₹8.00 a share: ₹800 Cr;
• buys back 2.5 Cr shares by tender at ₹480, a 20% premium: ₹1,200 Cr;
• issues 0.8 Cr shares to employees exercising ESOPs at ₹100: ₹80 Cr received.

Measure	Computation	Value
Dividend yield	800 ÷ 40,000	2.0%
Gross buyback yield	1,200 ÷ 40,000	3.0%
Net buyback yield (cash)	(1,200 − 80) ÷ 40,000	2.8%
Shareholder yield	(800 + 1,200 − 80) ÷ 40,000	4.8%
Total payout ratio	(800 + 1,200) ÷ 2,400	83.3%
Year-end shares	100.0 − 2.5 + 0.8	98.3 Cr
Net share-count reduction	1.7 ÷ 100.0	1.7%

The company spent 2.8% of its market value, net, but the share count fell only 1.7%. Had it bought and issued at ₹400, the same cash would have retired 3.0 Cr and issued 0.2 Cr shares, a net 2.8%. The 1.1-point leak comes from buying at a premium and issuing at a deep discount. In value terms, the tender premium paid 2.5 × (480 − 400) = ₹200 Cr to shareholders who sold, and the ESOP exercises transferred 0.8 × (400 − 100) = ₹240 Cr to employees (at the ₹400 market price). Much of the "buyback" was simply mopping up employee dilution.

What happens to EPS? Suppose the ₹1,200 Cr had been earning 7% pre-tax in liquid funds. After tax at 25.17% that is ₹62.9 Cr a year of lost other income. Next year's EPS on unchanged operating profit would be (2,400 − 62.9) ÷ 98.3 = ₹23.78, 0.9% below ₹24.00. Without the ESOP issuance it would be (2,400 − 62.9) ÷ 97.5 = ₹23.97, roughly flat. The reason is the rule from 01.3: a buyback raises EPS only if the earnings yield at the buyback price (24 ÷ 480 = 5.0%) beats the after-tax yield on the cash used (7% × 0.7483 = 5.24%). At 20x earnings it doesn't.

Three habits follow. Track the share count over five to ten years, not the buyback headlines. Add ESOP issuance to the dilution side before crediting a company with "returning capital". And remember that a buyback creates value for the holders who stay only if it is done below intrinsic value. EPS accretion is a different and weaker test.

5. Building the ratio dashboard

A ratio dashboard is a fixed-format table of ratios across years, in a fixed order, with fixed definitions. Its job is to make change visible. Design rules:

1. Fixed row order, grouped by question: growth → margins → returns → working capital → cash conversion → funding → per share → valuation. §6 explains why this order.
2. Definitions travel with the table. Averages or closing balances? Is other income in ROCE's numerator? Days on revenue or on cost? Write it down once (the course conventions are in tools appendix §3.2 and tools/fi/ratios.py).
3. One basis. Consolidated or standalone, the same for every year and every peer.
4. Six years at least, ideally ten. It takes that long to see a cycle, a capex programme and its payoff.
5. Levels and changes. Add a "what changed" view (latest year vs a reference year) to force your eye to the inflections.
6. Flag one-offs in the per-share rows (reported and adjusted EPS).
7. Units in every label: %, x, days, ₹ Cr, ₹ per share.
8. A source line: which annual reports, which data vendor, and the price date for the valuation rows.

With the course tools this is a few lines. Run from the repository root:

import io, sys
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8")  # Windows: lets you print ₹
sys.path.insert(0, "tools")                                          # run from the repository root

import pandas as pd
from fi.data import load_kaveri
from fi.ratios import ratio_dashboard, format_dashboard

pd.set_option("display.width", 200)
df = load_kaveri()                 # line items x FY21..FY26, ₹ crore (+ prices, DPS, shares in df.attrs)
dash = ratio_dashboard(df)         # 28 ratios x 6 years: decimals, multiples, days, ₹
print(format_dashboard(dash).to_string())

# The "what changed" view: FY24 (the peak) against FY26
key = ["ebitda_margin", "roce", "rec_days", "ccc", "cfo_to_pat", "fcf", "net_debt", "adj_eps", "bvps", "pe"]
change = pd.DataFrame({"FY24": dash.loc[key, "FY24"], "FY26": dash.loc[key, "FY26"]})
change["change"] = change["FY26"] - change["FY24"]
print(change.round(3).to_string())

ratio_dashboard needs no extra arguments for Kaveri because load_kaveri() carries the FY20 opening balance sheet (for FY21's averages), the 31-March prices, DPS paid and the share count in df.attrs. For your own company, pass opening=, prices=, dps= and shares= explicitly. The script tools/examples/kaveri_ratio_dashboard.py prints the same table plus a DuPont split and the working-capital days. The output, grouped into blocks:

Kaveri Pumps (fictional)	FY21	FY22	FY23	FY24	FY25	FY26
Growth & margins
Revenue growth	–	23.9%	15.3%	15.1%	16.5%	12.5%
Gross margin	36.0%	33.0%	34.2%	36.8%	35.6%	34.9%
EBITDA margin	13.0%	12.0%	13.1%	15.4%	14.7%	13.8%
EBIT margin	8.7%	8.3%	9.7%	12.1%	10.9%	10.2%
PAT margin	5.4%	6.2%	7.2%	8.6%	8.3%	6.9%
Returns
ROE (PAT / avg equity)	8.5%	11.3%	13.6%	16.5%	16.3%	13.5%
ROCE ((EBIT + other income) / avg capital employed)	12.5%	14.1%	16.4%	19.2%	17.0%	15.9%
ROIC (NOPAT / avg invested capital)	10.1%	11.6%	13.4%	15.1%	13.0%	12.0%
Working capital (days, closing balances)
Inventory days (on material cost)	82	74	76	75	78	80
Receivable days (on revenue)	64	57	61	70	84	96
Payable days (on material cost)	66	60	62	63	64	61
Cash conversion cycle	80	71	75	82	98	115
Cash conversion
CFO / EBITDA	63.8%	72.2%	58.2%	55.9%	35.3%	35.8%
CFO / PAT	154.9%	139.2%	106.4%	100.2%	62.5%	71.9%
Free cash flow (CFO − capex), ₹ Cr	32.8	31.7	(36.4)	(64.4)	2.9	13.1
Funding
Net debt (debt − cash − liquid inv.), ₹ Cr	(27.8)	(47.0)	5.9	95.9	112.0	140.5
Net debt / EBITDA	−0.3x	−0.5x	0.1x	0.6x	0.7x	0.8x
Interest cover (EBIT / finance costs)	7.8x	10.0x	10.1x	9.0x	7.9x	7.9x
Debt / equity	0.2x	0.1x	0.2x	0.3x	0.3x	0.3x
Asset turnover (revenue / total assets)	1.0x	1.2x	1.1x	1.1x	1.1x	1.2x
Per share (₹)
EPS (basic)	5.5	7.9	10.4	14.4	16.2	15.1
Adjusted EPS (ex-exceptional, post-tax)	6.4	7.9	10.4	14.4	14.5	15.1
Book value per share	66.3	72.7	81.1	93.1	106.2	117.7
Dividend per share paid in year	1.5	1.5	2.0	2.5	3.5	4.0
Valuation (31-March prices)
Share price at 31-March, ₹	210	320	360	610	780	520
P/E (trailing, on reported EPS)	38.4x	40.7x	34.5x	42.4x	48.0x	34.5x
P/B	3.2x	4.4x	4.4x	6.5x	7.3x	4.4x
EV/EBITDA (EV incl. lease liabilities)	15.6x	20.7x	19.0x	24.3x	27.9x	18.0x

This is exactly the ratio table on the Kaveri reference page (§6); the test suite checks that the function reproduces it.

6. Reading the dashboard from top to bottom

Read a dashboard the way you'd read a chain of cause and effect. Each block answers a question, and the answer to one question tells you what to look for in the next. Don't start with valuation. The multiple is the market's conclusion, and you want your own before you look at it.

flowchart TD
  G["1 · Growth<br/>Is the business getting bigger?"] --> M["2 · Margins<br/>Is growth profitable at the P&L level?"]
  M --> R["3 · Returns<br/>Is profit high relative to the capital it ties up?"]
  R --> W["4 · Working capital<br/>Where is the capital being tied up?"]
  W --> C["5 · Cash conversion<br/>Is profit becoming cash?"]
  C --> F["6 · Funding<br/>Who pays for any gap?"]
  F --> S["7 · Per share<br/>What reaches the shareholder?"]
  S --> V["8 · Valuation<br/>What is the market paying, and has it noticed?"]

Kaveri, block by block:

1. Growth. Revenue compounded at 16.6% from FY21 to FY26 (612.0 → 1,318.0), and 14.7% over FY23–FY26. FY22's 23.9% was partly a recovery from a COVID-hit FY21 base (04.1). Growth slowed to 12.5% in FY26, and the dashboard can't show what came next: Q1 FY27 revenue grew 3.5% year on year.
2. Margins. Everything peaks in FY24: gross margin 36.8%, EBITDA margin 15.4%, EBIT margin 12.1%. By FY26 they are 34.9%, 13.8% and 10.2%. EBIT margin fell further than EBITDA margin because D&A jumped 45% (₹32.9 Cr → ₹47.8 Cr) once the Hosur plant started depreciating while running at 57% utilisation (04.2).
3. Returns. ROE, ROCE and ROIC also peak in FY24 (16.5%, 19.2%, 15.1%) and fall to 13.5%, 15.9% and 12.0%. Put the last number next to the reference WACC of 12.19%: at the margin, Kaveri now earns about its cost of capital on the capital in the business (04.3). Growth at ROIC ≈ WACC creates no value.
4. Working capital. Inventory days (74–82) and payable days (60–66) are stable. Receivable days are not: 57 in FY22, 70, 84, then 96, and the cash conversion cycle goes from 71 to 115 days. The capital is going into receivables, and the segment table says where: solar systems sold to state agencies (04.4).
5. Cash conversion. CFO/EBITDA halves (72.2% in FY22 → 35.8% in FY26); CFO/PAT falls from 139.2% to 71.9%. Profit is being booked much faster than it is collected. 04.7 asks whether that is risk or worse.
6. Funding. FCF was negative in the Hosur years (FY23–FY24), which was expected. Afterwards it recovered only to ₹2.9 Cr and ₹13.1 Cr, below dividends of ₹21.0 Cr and ₹24.0 Cr. Net debt went from ₹47.0 Cr of net cash (FY22) to ₹140.5 Cr of net debt. Leverage is still modest (0.8x EBITDA, interest cover 7.9x), so this is a trend, not a crisis (04.5).
7. Per share. Adjusted EPS went 14.4 → 14.5 → 15.1: two years of near-stagnation while revenue grew 31%. BVPS rose 26% (₹93.1 → ₹117.7) because profits were retained, but much of that book is now receivables. DPS paid rose 60%.
8. Valuation. At its FY25 peak the market paid 48.0x reported EPS (53.8x on adjusted EPS of ₹14.49) and 27.9x EBITDA. It then de-rated to 34.5x and 18.0x at ₹520, and to 25.9x and 13.7x at ₹390. The market noticed. The question for Module 06 is whether it has noticed enough: the house DCF says ₹320.

The story in one paragraph. Kaveri grew well and profitably until FY24, when the new motors plant, a margin peak and a growing solar business came together. Since then growth has continued but has turned into receivables rather than cash: margins and returns have slipped back to the cost of capital, cash conversion has halved, and the company has borrowed to keep paying a rising dividend. Per-share earnings have gone nowhere for two years. The market has cut the multiple by almost half. Whether the stock is now cheap depends on one question the dashboard can't answer: will the state agencies pay?

That is what a dashboard is for. It doesn't give you the answer. It tells you which question to take to the notes, the concall and the scuttlebutt. Here: the receivables ageing note (overdue > 6 months doubled to ₹62.4 Cr), the state-wise payment record of PM-KUSUM tenders, and management's "75–80 days by end-FY27" promise (03.3, 05.7).

The "what changed" view from the code makes the inflection impossible to miss:

FY24 → FY26	FY24	FY26	Change
EBITDA margin	15.4%	13.8%	−1.6 pts
ROCE	19.2%	15.9%	−3.3 pts
Receivable days	70	96	+26 days
Cash conversion cycle	82	115	+33 days
CFO / PAT	100.2%	71.9%	−28.3 pts
Free cash flow, ₹ Cr	(64.4)	13.1	+77.5
Net debt, ₹ Cr	95.9	140.5	+44.6
Adjusted EPS, ₹	14.40	15.08	+0.68
BVPS, ₹	93.13	117.68	+24.55
P/E (31-March)	42.4x	34.5x	−7.9x

Only FCF improved. That is because FY24 was the capex peak, which is why a "change" row always needs its context.

Trader's lens — the dashboard is a risk report

A desk head reads the morning risk report in a fixed order: P&L, then positions, then Greeks, then limits, then what moved overnight. They look for changes and inconsistencies more than levels: a book whose P&L is up while its cash margin posted is ballooning gets a phone call. A ratio dashboard is the same instrument for a company. Growth and margins are the P&L. Working capital is the collateral the business has had to post. Cash conversion is settlement. Net debt is the funding line. The valuation rows are the market's implied view. Kaveri's report shows P&L steady, collateral posted up 60% in two years (receivables 192.9 → 346.7), and funding drawn to cover it. No risk manager would sign that off without asking who the counterparty is and when it pays.

7. The peer dashboard

A single company's dashboard tells you how it has changed. A peer dashboard tells you how it compares. The rules are stricter, because definitions that are consistent within one company's history are often inconsistent across companies:

• Same definitions, rebuilt from primary data where you can. Two vendors' "ROCE" figures are rarely comparable.
• Same fiscal period and basis: consolidated FY26 for all, with the market-data date stated.
• Medians, not means. With four or five peers, one outlier moves a mean a lot.
• Choose peers by economics, not by label: same end-markets, customers, capital intensity and business model (06.5 covers peer selection in depth).
• Add the columns that matter for this thesis, even if you have to build them by hand.

Worked example 4 — Kaveri against its (fictional) peers

From the reference peer table (§10; market data as of 18-Sep-2026, Kaveri at ₹390), plus three derived columns:

Company (fictional)	EBITDA margin	PAT margin*	ROCE	Rev CAGR FY23–26	ND/EBITDA	P/E	EV/EBITDA	PEG**
Nilgiri Pumps	16.8%	10.5%	24%	13%	−0.4x	39.8x	24.5x	3.1
Deccan Flow Systems	11.5%	5.9%	14%	9%	1.2x	36.2x	20.1x	4.0
Sabarmati Motors & Drives	15.2%	9.8%	21%	17%	0.1x	33.1x	21.4x	1.9
Konkan Solar Pumps	12.1%	6.8%	18%	38%	0.9x	29.6x	17.8x	0.8
Peer median (excl. Kaveri)	13.7%	8.3%	19.5%	15%	0.5x	34.7x	20.8x	2.5
Kaveri (at ₹390)	13.8%	6.9%	16%	15%	0.8x	25.9x	13.7x	1.7

*PAT ÷ revenue from the reference table. **P/E ÷ revenue CAGR in percentage points. A true PEG uses expected EPS growth. The peer table has only historical revenue growth, so treat this column as a rough cross-check.

Reading it. Kaveri is middle of the pack on EBITDA margin (3rd of 5) and growth (3rd), below the median on ROCE (4th) and PAT margin, a little more levered, and the cheapest on both multiples: 25% below the peer median P/E and 34% below the median EV/EBITDA.

A cross-sectional sketch. Fit EV/EBITDA against ROCE across the four peers:

\[\text{EV/EBITDA} \approx 11.8 + 0.48 \times \text{ROCE (in \%)} \quad\Rightarrow\quad \text{at ROCE } 16\%:\; 19.4\text{x}\]

At 19.4x FY26 EBITDA of ₹181.9 Cr, Kaveri's EV would be ₹3,529.5 Cr, its equity (less net debt ₹140.5 Cr and leases ₹18.2 Cr) ₹3,370.8 Cr, or ≈₹555 per diluted share. The house DCF says ₹320 and the market says ₹390.

Reconciling the three. Either the peer group is expensive (34.7x median P/E for 15% growth and ~20% ROCE is not a bargain, and relative valuation will happily tell you a stock is cheap against an expensive group), or Kaveri's discount reflects something that isn't in the peer table. Both are probably true. The table has no column for cash conversion (Kaveri's CFO/EBITDA is 35.8%), receivable days (96), customer concentration (state agencies), related-party purchases or the promoter pledge. Add those columns and Kaveri stops looking like the cheapest stock in the group and starts looking like the riskiest. (Four data points also give you almost nothing: the fit's R² is 0.53, and one peer moved would change the slope a lot.)

The same analysis in code (runs offline):

import io, sys
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8")
import numpy as np
import pandas as pd

# kaveri-pumps.md §10 (fictional peers, market data as of 18-Sep-2026; Kaveri at ₹390)
peers = pd.DataFrame(
    {"mcap": [6850, 2100, 3900, 1450, 2340], "rev": [1640, 980, 1210, 720, 1318],
     "ebitda_m": [0.168, 0.115, 0.152, 0.121, 0.138], "pat": [172, 58, 118, 49, 90.5],
     "roce": [0.24, 0.14, 0.21, 0.18, 0.16], "rev_cagr": [0.13, 0.09, 0.17, 0.38, 0.15],
     "nd_ebitda": [-0.4, 1.2, 0.1, 0.9, 0.8], "pe": [39.8, 36.2, 33.1, 29.6, 25.9],
     "ev_ebitda": [24.5, 20.1, 21.4, 17.8, 13.7]},
    index=["Nilgiri", "Deccan", "Sabarmati", "Konkan", "Kaveri"])

peers["pat_m"] = peers["pat"] / peers["rev"]                 # derived: PAT margin
peers["peg"] = peers["pe"] / (100 * peers["rev_cagr"])       # P/E per point of *revenue* growth
others = peers.drop("Kaveri")
med = others.median()
print(pd.DataFrame({"Kaveri": peers.loc["Kaveri"], "peer median": med,
                    "Kaveri vs median": peers.loc["Kaveri"] / med - 1}).round(3).to_string())

# Cross-sectional fit: EV/EBITDA against ROCE across the four peers (n = 4: a sketch, not a model)
slope, intercept = np.polyfit(100 * others["roce"], others["ev_ebitda"], 1)
fit = intercept + slope * 16                                 # Kaveri's ROCE ≈ 16%
print(f"EV/EBITDA ≈ {intercept:.1f} + {slope:.2f} × ROCE%  →  fitted {fit:.1f}x vs actual 13.7x")
# ... EV/EBITDA ≈ 11.8 + 0.48 × ROCE%  →  fitted 19.4x vs actual 13.7x

8. Screener.in: using its ratio pages without being fooled

Screener.in is where most Indian investors meet these numbers first (00.2 §4.1 introduces the page). It is fast, free and generally careful. It also has its own definitions, which are documented only partly and in more than one place. Reproduce its numbers once for a company you know, and you'll know how far to trust it everywhere else.

8.1 What Screener's rows mean

Sources: Screener's knowledge-base article Reconciling numbers and reporting the errors (dated 25-Feb-2022) and its guide Optimizations in calculating numbers, both read on 21-Sep-2026, tested against the Asian Paints consolidated page on the same day.

Screener label	How it is built (documented or reproduced)	Difference from this course
Sales	Revenue from operations	Same
Expenses	Total operating expenses (materials, employees, other)	Same
Operating Profit / OPM %	Sales − Expenses, before other income	= our EBITDA excl. other income. A company's own "EBITDA" line may include other income (Asian Paints' does)
Other Income	Other income plus exceptional items (KB), and (reproduced) the share of associates' profit	We keep all three separate. One-offs vanish into this row
Net Profit	Profit after tax including non-controlling interests (reproduced)	Use profit attributable to owners for EPS and P/E
EPS in Rs	Profit attributable to owners ÷ shares in issue (reproduced)	Close to, but not the same as, the reported Ind AS 33 figure
Tax %	Total tax ÷ PBT	Same
Dividend Payout %	Dividend declared for the year ÷ EPS (reproduced)	Our dashboard's DPS is paid in the year (§4.1)
Debtor Days	Closing trade receivables ÷ sales × 365 (reproduced; the KB wording mentions an average)	Same as ours
Inventory Days	On a cost-of-goods basis per the KB; we could not reproduce Asian Paints' figure	Ours: closing inventory ÷ material cost × 365. Don't compare the two
ROCE %	EBIT ÷ average capital employed; EBIT = PBT + interest, excluding exceptional items; capital employed = share capital + reserves + borrowings (incl. lease liabilities) (KB)	Very close to ours: (EBIT + other income) ÷ average (equity + borrowings + leases)
ROE %	Net profit ÷ average (share capital + reserves) (KB)	Same idea
Stock P/E	Market cap ÷ trailing-twelve-month profit, excluding extraordinary items (guide)	TTM, not last fiscal year
Book Value	(Share capital + reserves) ÷ shares	Same
Dividend Yield	Uses the higher of the latest interim dividend and last year's dividend (KB)	Check it against the actual declared amount

8.2 A reproduction on a real company

Worked example 5 — Asian Paints FY25: Screener vs the annual report

Method illustration only; not a view on the stock. Screener's consolidated page (link, accessed 21-Sep-2026) against the consolidated statements in the Integrated Annual Report 2024-25 (₹ Cr):

Screener row (FY25)	Screener	Rebuilt from the annual report	Check
Sales	33,906	Products 33,626.82 + services 170.54 + other operating revenue 108.26 = 33,905.62	✓
Expenses	27,899	Total expenses 27,899.41	✓
Operating Profit	6,006	33,905.62 − 27,899.41 = 6,006.21. The company's own "EBITDA" line is 6,578.82 because it includes other income of 572.61	✓ (and a trap)
Other Income	350	Other income 572.61 + share of profit of associates 140.71 − exceptional loss 363.10 = 350.22	✓
Net Profit	3,710	PAT 3,709.71, of which owners 3,667.23 and non-controlling interests 42.48	✓ incl. NCI
EPS in Rs	38.23	Owners' profit 3,667.23 ÷ 95.92 Cr shares (share capital ₹95.92 Cr at ₹1 face value) = 38.23. The reported basic EPS is 38.25	✓ (not the reported figure)
Tax %	27%	1,393.36 ÷ 5,103.07 = 27.3%	✓
Dividend Payout %	65%	FY25 dividend declared ₹24.80 (interim ₹4.25 + final ₹20.55; BusinessToday, 8-May-2025) ÷ EPS 38.23 = 64.9%	✓ declared basis
Debtor Days	46	Current trade receivables 4,313.65 ÷ 33,905.62 × 365 = 46.4 (on the average balance it would be 49.5)	✓ closing
Borrowings (balance sheet)	2,290	Borrowings 259.62 + 604.13 + lease liabilities 1,123.30 + 303.24 = 2,290.29	✓ incl. leases
Reserves (balance sheet)	19,304	Other equity 19,303.89. Non-controlling interests (659.24) are not in it	✓ excl. NCI

What the reproduction teaches.

• The FY25 "Other Income" row is net of a ₹363 Cr exceptional loss. Screener shows other income of 821, 350 and 728 for FY24–FY26. The dip is not a treasury problem; it is White Teak and Causeway goodwill impairments and a loss on exiting Indonesia, all folded into one row. For a company with an exceptional gain the same row would flatter PBT, as Kaveri's FY25 land sale would (its Screener "Other Income" would read 4.5 + 14.0 = ₹18.5 Cr; 00.2 worked example 3).
• "Net Profit" and "EPS" are on different bases. Net profit includes the minority's share, EPS does not. A P/E computed as market cap ÷ Screener "Net Profit" is slightly understated for any company with NCI.
• The payout is on the declared basis. That is sensible for a dividend policy, but it is not comparable with a cash-flow-statement payout.

8.3 Reproducing Screener's ROCE, and a documentation trap

Screener's two help pages describe capital employed differently. The KB says share capital + reserves + borrowings. The Optimizations guide says capital not deployed in the business (CWIP, investments and non-current assets) is excluded. They give very different answers for Asian Paints, which carries large investments and CWIP:

Asian Paints	Screener shows	KB definition (average CE = equity + reserves + borrowings)	Guide definition (CE excl. CWIP & investments)
FY24	38%	38.6%	56.9%
FY25	26%	26.5% (after adding back the ₹363 Cr exceptional loss; 24.9% without)	36.0–38.4%
FY26	26%	26.4% (from Screener's own table; exceptional items not checked)	38.6%

(EBIT = PBT + interest from Screener's own P&L; capital employed from Screener's own balance sheet; FY25 inputs match the annual report as shown above.) The displayed numbers follow the KB definition to within about a percentage point in the years we tested. Across FY16–FY26 the gap is at most about 2.5 points, which we could not explain from the page alone (unseen reclassifications and rounding of inputs are likely). Two consequences:

• Screener's ROCE includes treasury. Cash and investments sit in capital employed and their income in EBIT, so for a cash-rich company Screener's ROCE is a blend of the operating return and a ~7% treasury return. That blend understates the operating business's return. The course's ROIC (04.3) strips treasury out. For Kaveri, which has little cash, the two ROCEs coincide: FY26 (121.0 + 17.0) ÷ 868.1 = 15.9% either way.
• Never quote a vendor ratio you haven't reproduced once. The documentation itself was inconsistent on the day we checked (21-Sep-2026). Screener's definitions may change, so reproduce them again when it matters.

Screener's cash-flow block also shows a "CFO/OP" row. For Asian Paints FY25 it reads 97%, but CFO ÷ Operating Profit from the same page is 4,424 ÷ 6,006 = 74%. We could not find its definition on 21-Sep-2026. Until you can reproduce it, don't use it.

8.4 A safe workflow

1. Toggle to consolidated (the /consolidated/ URL) unless you specifically want the parent alone, and use the same basis for every peer.
2. Screen and browse on Screener; rebuild the dashboard from the annual report for anything you'll act on (10.2). Screener's Excel export is a fine starting point, but check the exceptional items and other income rows first.
3. Split "Other Income" into treasury income, share of associates and exceptional items using the P&L and the notes.
4. Use owners' profit for EPS and P/E, not "Net Profit".
5. Check the dividend basis (declared vs paid) before computing yields and payouts.
6. Reproduce ROCE once per company, then decide whether you want ROCE (a blend) or ROIC (operations only).

India notes

• EPS on the face of the P&L. Schedule III (Division II, for Ind AS companies) requires basic and diluted EPS on the statement of profit and loss. The EPS note gives the weighted-average share counts and the potential shares. Read it whenever the count changed (QIP, ESOPs, bonus, merger).
• "Dividend %" is on face value. Indian announcements often say "a dividend of 80%": ₹4.00 on a ₹5 share, as in Kaveri's case. It says nothing about yield. Always convert to ₹ per share and divide by the price.
• Dividend distribution policy. The top 1,000 listed companies by market capitalisation must adopt a dividend distribution policy and publish it on their website, with a link in the annual report (SEBI LODR Reg 43A, in the text amended to 22-Jan-2026, checked 21-Sep-2026). Read it before forecasting DPS. Many are vague, but some state a payout range.
• Final dividends are paid after the AGM, so a year's "dividend paid" in the cash-flow statement is mostly the previous year's final dividend (§4.1).
• Buybacks. The tax treatment changed three times between 2024 and 2026 and the open-market route was restored from 1-Aug-2026. The current position and sources (as of Sep-2026) are in 01.3. Tax changes move companies between dividends and buybacks, so shareholder yield is the comparable measure across regimes.
• Promoter participation in tender buybacks raises the promoter's percentage holding; check the post-buyback shareholding pattern (01.4).
• Consolidated vs standalone EPS can differ a lot for groups with listed or loss-making subsidiaries. Screener shows both, so be clear which one you are comparing.

Common mistakes

• Quoting a P/E without saying which EPS: basic, diluted, adjusted, TTM or forward, and on how many shares.
• Adjusting out one-off losses but not one-off gains, or forgetting to tax-effect the adjustment.
• Mixing pre- and post-bonus per-share data in one series.
• Computing payout on the wrong year: dividends paid in FY26 against FY26 PAT when they mostly relate to FY25.
• Treating "cash EPS" as cash. It ignores capex and working capital. Use FCF per share.
• Crediting buybacks without netting ESOP issuance, or judging a buyback by EPS accretion instead of price vs intrinsic value.
• Reading the valuation rows first. You'll end up hunting for evidence to support the multiple.
• Comparing peers on vendor ratios with different definitions, or using means with four peers.
• Taking Screener's "Other Income" and "Net Profit" at face value when exceptional items or NCI exist.

Key terms

Term	Meaning
Basic EPS	PAT attributable to owners ÷ weighted-average shares outstanding (Ind AS 33)
Diluted EPS	Basic EPS adjusted for dilutive potential shares (options, warrants, convertibles)
Adjusted EPS	EPS excluding items judged non-recurring, after tax; your own construction unless disclosed in the notes
TTM EPS	Sum of the last four quarters' EPS; the right trailing measure for seasonal businesses
Cash EPS	(PAT + D&A) ÷ shares; ignores capex and working capital. Not a cash measure
BVPS / tangible BVPS	Owners' equity ÷ shares; tangible also deducts goodwill and intangibles
Equity roll-forward	Opening equity + PAT − dividends + issues − buybacks + ESOP credit ± OCI = closing equity
Sustainable growth rate	ROE × retention ratio: growth fundable from retained earnings at constant leverage
DPS (paid vs declared)	Dividend per share paid in cash during the year, vs declared in respect of the year
Payout ratio	Dividends ÷ PAT (or ÷ FCF, to test cash cover)
Dividend yield	DPS ÷ share price
Buyback yield	Cash spent on buybacks ÷ market cap
Shareholder yield	(Dividends + buybacks − cash from share issues) ÷ market cap
Ratio dashboard	A fixed-order, fixed-definition table of ratios across years, read as a causal chain
Peer dashboard	The same ratios across comparable companies for one period, on consistent definitions
Operating Profit (Screener)	Sales − Expenses, before other income, D&A and interest (= EBITDA excl. other income)
ROCE (Screener)	(PBT + interest, excl. exceptional items) ÷ average (share capital + reserves + borrowings incl. leases)

Check your understanding

1. Kaveri's FY26 basic EPS is ₹15.08 and its FY25 ROE on opening equity was 17.43%. Using only BVPS (FY24 ₹93.13, FY25 ₹106.20) and ROE on opening equity (FY26 14.20%), show that FY26 EPS growth was −7.1%. What does this say about the source of the fall?

Answer

EPS = ROE on opening equity × opening BVPS. FY26: 14.20% × 106.20 = ₹15.08 ✓. Growth: (106.20 ÷ 93.13) × (14.20 ÷ 17.43) = 1.140 × 0.815 = 0.929, i.e. −7.1%. Book value per share grew 14.0%, so the whole fall came from a lower return on that book. Part of it is FY25's land gain inflating the base: on adjusted FY25 PAT (86.92) ROE on opening equity was 15.6%, and adjusted EPS actually rose 4.1%. The rest is the margin and asset-turn deterioration seen in the dashboard.

1. On 8-Aug-2026 a broker note says "Kaveri trades at 25.9x earnings"; a second says 27.2x; a third says 16.9x. At ₹390, which EPS is each using, and which is the most useful? Why is the third misleading?

Answer

25.9x = FY26 basic EPS ₹15.08; 27.2x = TTM EPS to Q1 FY27, ₹86.1 Cr ÷ 6.00 = ₹14.35; 16.9x = FY26 "cash EPS" (PAT + D&A) ÷ shares = ₹23.05. For a seasonal business the TTM figure is the most current fair trailing measure. The diluted FY26 P/E (26.2x) is the conservative accounting version. Cash EPS adds back ₹47.8 Cr of D&A but deducts none of the ₹52.0 Cr of capex or the ₹89.7 Cr of working-capital build. On actual FCF (₹2.18 a share) the price is 179x.

1. A company reports PAT of ₹500 Cr, including a ₹60 Cr pre-tax profit on selling a building and a ₹20 Cr pre-tax restructuring charge it says is "one-off". It has 50 Cr shares and pays tax at 25.17%. Management's presentation shows "adjusted EPS" of ₹10.30. What is the correct adjusted EPS, and what did management do?

Answer

Net one-offs pre-tax = +60 − 20 = +40; after tax 40 × 0.7483 = 29.93. Adjusted PAT = 500 − 29.93 = 470.07; adjusted EPS = ₹9.40. Management's ₹10.30 = (500 + 20 × 0.7483) ÷ 50 = 514.97 ÷ 50 = ₹10.30: it added back the loss and kept the gain. Reported EPS is ₹10.00; the honest adjusted figure is 6% lower, not 3% higher. (And check whether "restructuring" recurs; see 04.7.)

1. Nirmal Finance's FY26 equity is ₹1,695.3 Cr and FY25's is ₹1,485.8 Cr. FY26 PAT was ₹239.5 Cr and dividends paid ₹30.0 Cr. Does the roll-forward tie? Compute BVPS growth and the sustainable growth rate, and compare with loan growth of 19.7% (5,780 → 6,920).

Answer

1,485.8 + 239.5 − 30.0 = 1,695.3 ✓ (no issues or other items). BVPS: 135.07 → 154.12 (11.0 Cr shares), +14.1%. ROE on opening equity = 239.5 ÷ 1,485.8 = 16.12%; retention = 1 − 30.0 ÷ 239.5 = 87.47%; sustainable growth = 16.12% × 87.47% = 14.1%, equal to BVPS growth. Loans grew 19.7%, faster than equity, so leverage has to rise (total assets ÷ equity went from 4.2x to 4.4x on the reference ratios) or capital adequacy falls. Nirmal's CRAR fell from 28.7% to 27.6%. A lender growing faster than ROE × retention will eventually need equity (as in its FY24 QIP).

1. Tapti Infotech (worked example 3) announces next year's plan: the same ₹800 Cr dividend, a ₹1,200 Cr buyback at ₹400 (no premium), and 1.5 Cr ESOP shares issued at ₹100. Market cap ₹40,000 Cr, 98.3 Cr shares at the start. Compute the shareholder yield (cash basis) and the net share-count change. What would you tell a client who describes this as "a 3% buyback"?

Answer

ESOP cash = 1.5 × 100 = ₹150 Cr. Shareholder yield = (800 + 1,200 − 150) ÷ 40,000 = 4.6%. Shares bought = 1,200 ÷ 400 = 3.0 Cr; net change = −3.0 + 1.5 = −1.5 Cr, i.e. −1.5% of 98.3 Cr. Half the buyback just offsets employee issuance. The 3% headline is gross; what reaches continuing shareholders is a 1.5% reduction in the share count, and the ₹450 Cr (1.5 × (400 − 100)) of value transferred to employees is real compensation cost that sits in "buyback" cash.

1. In Kaveri's dashboard, ROIC is 12.0% in FY26 and the reference WACC is 12.19%. Revenue grew 12.5%. A colleague says, "Growth is good, ROE is 13.5%, so value is being created." Respond using two rows of the dashboard and one identity from this lesson.

Answer

Value is created only when the return on the capital invested exceeds its cost. ROIC (12.0%) is at or just below WACC (12.19%), so FY26's growth was roughly value-neutral at best. ROE of 13.5% is a return to equity helped by leverage (net debt is up to ₹140.5 Cr) and is not compared with the 12.8% cost of equity *per unit of risk taken*. The growth also consumed cash: CFO/PAT 71.9%, FCF ₹13.1 Cr against ₹24.0 Cr of dividends. And the sustainable growth rate (ROE on opening equity 14.2% × retention 73.5% = 10.4%) is below the 12.5% revenue growth, so the gap was borrowed. Growth that needs borrowing to fund receivables at ROIC ≈ WACC does not create value.

1. Screener shows a company's FY26 "Other Income" at ₹210 Cr against ₹60 Cr in each of the previous three years, and ROCE of 22%. What would you check before using either number?

Answer

Screener's "Other Income" includes exceptional items and (from our reproduction) share of associates' profit. Open the P&L and notes to split the ₹210 Cr into treasury income, associates and exceptional items: a ₹150 Cr jump is probably an exceptional gain (asset sale, fair-value gain) or an associate's one-off. Screener says it excludes exceptional items from ROCE and P/E, but other rows (Net Profit, OPM-based views, "Other Income") still carry them. For ROCE, check whether the company has large cash/investments (which Screener's capital employed includes) and whether it has NCI (profit includes it, reserves exclude it). Then recompute ROCE and ROIC yourself on consistent definitions.

1. Build a peer-dashboard column list for a thesis on a solar-pump company selling to state governments. Name six columns that are not in the standard peer table, and say where you would find each.

Answer

For example: (1) CFO/EBITDA and CFO/PAT over three years (cash-flow statements); (2) receivable days and the ageing schedule's > 6-month bucket (receivables note, Schedule III ageing); (3) share of revenue from government/state agencies (segment note, MD&A, concall); (4) order book and its state-wise mix (investor presentation, concall); (5) ECL allowance as a % of overdue receivables (financial-instruments note); (6) promoter pledge % (shareholding pattern, SAST Reg 31 disclosures); (7) related-party purchases as % of material cost (related-party note); (8) contingent liabilities, e.g., GST classification disputes (contingent-liabilities note). Any six with correct sources.

Go deeper

• Berkshire Hathaway 2018 shareholder letter: Buffett explains why he stopped using per-share book value as the scorecard. Useful for thinking about BVPS vs intrinsic value.
• IFRS Foundation, IAS 33 Earnings per Share: the standard Ind AS 33 follows. Read the paragraphs on retrospective adjustment and additional per-share amounts.
• Screener's Reconciling numbers and Optimizations pages: read them yourself, then reproduce one company's page as in §8.
• tools/examples/kaveri_ratio_dashboard.py and the tools appendix: the dashboard code, a DuPont split and working-capital days, ready to point at your own company's numbers.
• Aswath Damodaran, Investment Valuation (3rd ed., Wiley, 2012), the chapters on relative valuation: how multiples relate to fundamentals, and why peer comparisons need the fundamentals alongside the multiples.

---

← Previous: 04.5 Leverage, solvency & liquidity · Module index · Next: 04.7 Quality of earnings →